; ; Senet aia ath k PP ADL re Pi dad eid le odtad how eee vA sitene wy ae \ Pa he hawt txt * Rae vs a POT Wier “iesdiaaies: at om J ° aaa 998 Ay + BON Oi) i A ye i) VER Ee as 62 - Cee reatave: ey + ¥eT2 Cree) Ari é toutsivee” #:% oie Fy} hee tig ob ke wha Sti tach dae tare 4 ae Pah Noh A a) Agi ‘i 04 Pe tee Sep hok we ' 798 far ovine > ee nity : BAN Suen ” ites aoe os oLAw v3 5 ) 14 DSP haarbanty tin PURPORT MAL tas Fei WU MC bo @ Save: ' aeaitaes area nts Ag yvve OWOEN! re Aud are lhe i) hie > Lr eee i Sannin 53 hk qT eee Wye aN eS ee needed 14 4 4-4. et Deri a4 wil aoa 'b aitataty ceils sa? ait at #4 Fy Mek Sth ee Shasarires yn ace, 14 Hae vey Ee raat tr ball AS te eau Lg Wh cen is bed abet Be Bee " ay ey MH a an Gai ee Ay Ty CUB fy a tend Eee ire AN * vty Peeietl ainsi ik i) iy set 13 Ria y) oe Se AL aa mahai tend 7 piRoaiaany 19 eee: . Manes Wate d Hd4 Roa ae We es tT TL HEM is Pie Pi ka Were Ot Mh ek § wad * en he t, wleceenca heared aunt ' ‘ HAE Re egtprasuswed) bea d Uy si Seni tia Ua bapa ee WR Ae , Wiech ee aches yathst : heh v ASA Yate re ie ae ' ae ae . ¥g sa) a) ay ek ite @ Myf) Tk i Pip of a0 LA ord she ny SS itaretsan Tw fot ee * . +a 98) a “gi um) aliens Gh ead Se: bye ate eid Api un A Ke * ait solece Let g ene a behets 8 y fre el Psa orgs eva bes laterite Hl ae nitions he etal pa a mane rh gee ee Weigle Halo) as "eo 4 Ah Pr Fa br nee Wyo es hae rts HP ie ocd ahaha AL Per eew etry a 1 2 ae Wows Wake be Bieta ae Gatehesti riteb cy even anes i lies: or ar Se oka ny 4 Pe i ead I a iy La het paper te aha eas be aA saniae! ped pagers a oF Oe ALES hehe Heeves MGA, $y :8) presi itia. Ae Se Ae I brah eae a He Nhs SF Minter) Acts + The ‘A seit oder miaen sds 2B it-4, ber ot al i hie pa Hi by aise Cheer yey Bad aa no peed Les} smebiyige ba alah ve Gl We We Weald. sean > + sate 4 ‘ Hive ees ny he ee) hs ey ry: Hogtencl #. A hes Le Ky ¥ Kew ele a rapa teed fi +4 dtnivan! ty aaran st tese nae 19) i aa eh has oratasucaths aoe Whe Con ORE Hy te iv Debary Hee et LAG eS Ce a Prien : aN 1a ‘itv aH ty “ot4 Nd) iin ih vartatee vn pit isies aie a irre ee Aad pe oi Be ane in epee a: ae ihe A ee Bs et 5 i SPOOL) 1 Bot he Wty a ihe Vy ee) Fs 4 Pinal Wit et wahitet: sate ihe igri sith ret aed Ssjaslice's: vee LA. Oi2- 4 cur ewe e Lae a 7a Se egret: hee Seaton APA anther bcty Sh ch ead per ; : ed 849 4 rede 8 a) Dias Hy sett LUBA ely SWE Na Patek Ora SAID oka . vin sag Peper ee "ata ymca 4 sineaaaentnie see rete A LUD Sa NaN ine ‘4 Mey UT chk de Fan ANAi 5, Sota 9-4, Ret 19,10) asin one Var APSNENT Seat sa mayenas a) Tha GUA, yh i “4 ei Ah, HY ; ' Hier bo sony cir 149i me) mm Vid ney Pitre peeve Be, c ad “Moat ros es i euehe “i uh HY atieitarene Co S sath cies arhe eae ary wart A Ra. ffl ey ave Tt ae a) a? ptiy ieee ity 4 ies PA Aa eet bat eee ¢ yi Cae hae Hp i a» esata Hee Powe ei Fe Bi nea Bent? ai my ie rie e% or Sse rte oF de wt ‘iracine: aA + : Sarees ev Aa i PN ah iain We far? elt As He sre ego had ey CoA area doe ei ives Apes ie cage sien ma Grit Ma ers): in be gel dt ae feet ps beat me tad) eh ak et ate oe * Lape hes! ge f Pirin Ady sity i} eee ieee ies Tie TH Tanti bar bsteheeEbeas eae rye beh a . beh ya Adee be Gane HEL fale iantatenerte ee th bei Asker eas up’ sane ne ean ered bead : ws de hep ii ee iD VP oe: Me , reir astsreg cae ef aed Wonatatad He att bi hia oy ae beryurit Teh poe ranma § rat ee 2) or ae mee fH asesbed ot oie ame ett bat re. aware aiid bite bin i rated sae Leta sae UUs sien aay aes eae b> i aio Petros gee awe gia nite! eis OF erate HOTS tT NY cs eae hee puiher role an SL ae eae ee «eae Lo ie Fr OTS th oh alah Skok Shoe pines Weak Do) AL Tak ed Ach 7) Hs Pa yp Gwes . ry) sae LA tie setae ‘ siete Ha es bvarqagedt are ard ifs whee ences tu yriesl es hie mereerirren tt etna) feat itl irate are Arion re ser soo F ieee hs BBE ST ce ra Was rh 2 Ait bee pay Rh er rare eh. Wa st iin Hs > athe baie phe p states , , ry a a erences) ran 2 ye nds ag hee AAs i om un Sweet acme Tamia! pring Med emt aH fe gs any ae Resteahounen aa apegs Ure Bebe seeds Ben ply yas rater Wa neber Tee ee eet ’ aft eae iets: ett ode nas tt ot Tok pee ‘ ssonas Parts BARE Nt ere” t yacdre® peer interests +9 soit ys cee ny rs ea Has Ranta ea era vied, ah rhe va it ioe s Vibe ip ay riee rds Be ike D7 Rrat 687 Owe ne ’ teehee on Rompe a ler ot eeae Lae ti Bete irae ert _. ee tonne ip Lawetal ieee nee , Rr nearetatat A abe thet Se seh teeatetetitatatecac hi os ee sia seis LSE ay cee ral i ‘f Mik, wr re ie pebebeke rea ies arth tee Wiiniaee ate Aah dy eas ehh Pras per deageag es a bear mare Fis eset ataaere rae pera te y Aoi ae + UAL ; Br er Prtag bie). %: aitititee nae hehe Rede a yrart: oie he td bra eo te a, K pehealght Uae the ats sae AX eas in Phe hI R 4 oy Thier bt it wed a perce tah sacey H » rice ae ” 4 jaan? 144 Ate ? vi pe dy bul bs ser het a 1 Se MHerarh rare te PE sete im Ry WD tht sete naoes batt ae wth Penni Or Mask lade al ee So me via nee’ Tua Hyde ete arcane ath Lad wie: aa 2 bey tx ge acetbasacarbait a edema At} 5 Keats orem Vainacatanl veya itstd Legely Ahi A citys eens ewes elacasiesrenate abt i iesieaameeaenr dolerern sl Reem oR Nh gs ht soa: woipe delenit n ‘ar ele 4 ra . nee he Pieuuae nthe tartanmtahe BPO Se ane yey pe eat neneet gree tte yee ee es bevene re’ a ies nei oe i Lea a ive Ci iferatrececat rte os ih Marae! De adh lod Carat re aos fsttiete ae oe i aeiehae rhs he ed te iesiten inns Cea eet a secre va ns ba Ses eeu i tte ae 41 ai Hi i ere ra renee Ae ase yok ct oe isa Hears tre Be 4 Cath Hl Bho ne Sas ay Wc we CA arth lit: ) wean need nue Ware Ay aati yy ty ta Dhaene sae ; pron. Ae at yng . ye’ a Vay haa eee 8,49 a as eit arias et Tas 4) ARN A BM Sento naeh ihe . 4 4:40 eae REN TG Jel irusene ar aries eat nee ' eriewrn bee Yh, veo rasta PDT aph one : Tey: pacha ah raat a heanenaan uf te MA n fey in HY aaa 4 ae Rohr ences ts wy) ratte Ft per hee wie porwr Se Ct uk Oy Witetataryacacaeict mM ANN Aes bt ie Rhye YA ’ ’ rE EE eG ashe ey a gaame Ne NA ecb Br te athe DY etn see A eye ek WMO, hein ae eek ach bs hctade sr TT Me + Oe eater’ rast fas ek saree? 4 Mu ikke welt vee RearWartits i) titer iets Cy hyd 14st 88 re oT ' " Us Heb: ease Th sasheeostetetets » mit: brea er ee Poe ‘ 2A Hens arb egy 7. hyd a yi ty i gale Re “i mitoses + pea ety Wy Ext) ‘bees piecaga: aT ‘A wridirigh Wap teit Ae yt te ih Sih vs vA MH Hrneareo yp i Oe peatede bee fe th Dae he Wee Gloss te tes) * ie Be hae OF 08 sf cei na theds ate ara seem arts phe ae able cae ft it tao ce w ey im nits nay mareenier: ey Gidy ar ee Po ahy eae DBAS Beith »¥ ea ee very Ae bacine fo th Wa Arvin V4 be vee by 9, Oty ff iy Wad i) aaitsie ng #84 DEH deb ayavinaz Tp44 f LM rae seater keses Peer it See aed ndohehs Ot: Gor ts a rele ts Ores ee pian Ehret bee ye Obs ies age ee Rhee HOD + 1A WEN he “ Pavan De Reed AS ty ve he ; VashehoarAsiehcbuaeae arWeArayh cue 42 WA a rth ee? ye * a ' ch da Sh ie vasa e) tire a wire ave ee Sanna ae eh read id Pens eeae sea tis ols aut Ys Batis aye ait Fad u "4 ents (hh Wikre eas iu) rier a” ra ie Sean an aueroeti 3h oe eis oS yehsaie-aetyat ‘ enna ane hy onie¥r Ar ts tere Gonna ie PLS Tippee he Aas fant ais oS ue be see dy Arye eater : _ oo bey init 19 cei yh) aren ba ” if p "ag iy fm we oo he ee at vee: ‘eocpeoaene Arann We ths Go viet he 4~ Set sales Wate reheat ae ase an ty Beh? yA sting atone a sarieraets ib Do Be fate ee yee aD eke bees Raidole Melby ob Be Bete Ae Wo Oy Tee bahg habe t ele te hed DOM oe Ae a) es bry ws Meee yates gts Riba Rew BB oy ype Vay pds Th wr wee ek hese ne Asem roa ci a iiee he de ne ie SM Ae! * rt spn Wy Tea a elke Warrant bets Sy Che he ethstns eth pf us Dey Ne eis Fre fey me Mahe meals Nets me 9 Weg we a My "Y Tgte (he Fe aig ha einen ity he ee » ya, ihe omens he eat h(t Oo FWA Noh ay Saabs ne 1a Oy ite Ihe a te Lah aie ate i be ha ite apy Bh Wiebe 4 2 i bt Cau Moo a) Merunt ithekits patra sb, Ne rr Betts 448 ere + A hint OF Phe a vibe Rar Hh har ihe ah a the Ube esha Ont fraelenaee ¥.! Senate Hab ne fe ones 244 ee WS Wa lbe B it fasetley aE eal a thet 0 a Yh Gy pater a oo fs ei Seas Wty haa i. rr wh Ub ve mh ey kd = Uy ie se eae ae ahaa t hh 1s Hy Os Oe 8 te Baie ie ete oe _ : Bee Say > hy rye Av Be ily De i toy the (hn 49 Dn Oe hn te HVE 3a AACE a ee woe = Cehibeuaoh , ee Heated noch inaltyaos an iskedverasimartodsdah a Ye Jha bo fh ee s IPD eatset beets Pe MR Meee Rete APNE, hein | Aa bre . Vimar yay aie eV beh why? Wik OP ary br ORAS WEE he pis hs Si Be SORE BEST! ‘ 4 yh ebee ries ihe eh MAM bay my We ee Poh AM, he ihe Wear les bathe Pe fh Phe Da dh ths bles ee a Anse eh Wee elt Be seit bere & fee et Rane nepereaNn Arty’ * Wis iets Be Bertie Ge be Oy teh joarhatyhehs bse bea G01 ler os ete Me be + oe Ae po ott site ore welt Nero: Obs te Rha Bho tog RY Mh 4 ao Gait neh aa hy eee ane ata wir Writs abet iy hen 1 Ue (ea hm Oe hy he the 8 hy Be Biermann be Mien rea —— rit Fy Re Ber Bete Ag 8 Hy on Aloha ‘THE UNIVERSITY ‘OF ILLINOIS LIBRARY 9) (5,64 % 6 | x eae (Teolo Qy RR, The person charging this material is re- sponsible for its return on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. University of Illinois Library L161— O-1096 Di, . 90 fF EXPLORATIONS IN [TURKESTAN PAPEDEFIONFOFR 1904 PREHISTORIC CIVILIZATIONS OF ANAU Origins, Growth, and Influence of Environment EDITED BY RAPHAEL PUMPELLY DIKECTOR OF, [THE EXPEDITION IN TWO VOLUMES—VOLUME TWO WASHINGTON, D. C. Published by the Carnegie Institution of Washington 1908 oe Oe CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 73 [VOLUME II] f J : PRESS OF GIBSON BROS. 1908 oem Del MLB Ancient Anau and the Oasis-World, and General Discussion of Results By RAPHAEL PUMPELLY Archeological Excavations in Anau and Old Merv . By Husert SCHMIDT Note on the Occurrence of Glazed Ware at Afrosiab, and of Large Jars at tinge see eee eee 6) ou By Homer HaAKIpper Description of the Kurgans of the Merv Oasis . By ELLSwoRTH HUNTINGTON Chemical Analyses of Metallic Implements . . . . . By F. A. Goocu Physiography of Central-Asian Deserts and Oases.. By R. WELLES PUMPELLY Animal Remains from the Excavations in Anau, and the Horse of Anau in its Relation to the Races of Domestic Horses . . By J. ULRICH DUERST Description of some Skulls from the North Kurgan, Anau . . By G. SERGI Some Human Remains found in the North Kurgan, Anau . By TH. MOoL.ison Wheat and Barley from the North Kurgan, Anau . By H. C. SCHELLENBERG Stone Implements and Skeletons excavated in Anau . By LANGDON WARNER iy at} oe ae : Hise On TABLE OF CONTENTS. SECOND VOLUME. Part V. PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OaAsES. R. WELLES PUMPELLY. PAGE, eed YEE GEE eee ete cg. Scion Mensa sp eto gla, 6 ONY dl oak sink mye Fed me ae 243-298 SP tenor OC ASIIt te ail dar GETINS W INCI ory Mr suse ese ik Sica ed vii hay cf aad = Bue wales bie bes wma 243-250 Lidice Corti Git CRATaEeete C4 A LPSOLE DASIT) al Sukh esl tck anid He sete Ea ee phe ee aD S 243-244 The three agencies of erosion and transportation, ice, water, and wind; and the five deposition zones, glacial, alluvial, lacustrian, flying sands, and loess arising CLLET GLY OLIN ee een rere eM See eee me Sieh Bichake AO See Me Benne are cea 244-246 The interlapping of deposition zones effected by climatic oscillations.................. 247 ee mycncal aevenmment ak diriceal gesert baste 12.09 og oie ve Ae es des oe deat 247-250 eM Gee ena aiken ae OWES AL, ne a et en Mae ASS CHS oc iow BE eS ahah 6 ole Oe o> 251-259 ream tCatUlresfOle LGw Palit Ler ee. Fae Ore at curd acto Siesta UeRlaats. cus Maia ae oe 251 MCA ASIMEO PO TeAtwat tak tl tan arena aren sat Oe weak hee oins Arata eo aceon wt he oychoe helen ds 251-258 Tentative reconstruction of Quaternary sequence of events................ 0000000 eee 259 Ai eemre ieee RSAITE IIS elt Ghd SN hens a eg OE eg aha Kereoraca RAS SMe, ca wale adie ads 259-264 Ram CtE Ye OL VaNeV Ton amr iA ion cis tre fal ain 64 tl Soe pre ee ee ols nia A Sk awe es aco s 259-261 Sliaciology ang evinences Of mouttain movement «2... 55.5 eee de ei eels dae dees 261-264 A tentative reconstriction of events in the Alai Valley... 2... 00... 0.6.2 eens ete 264 DRG Scat Ea pS Ree ee a yh ee, Se er eee eek or ar 265-274 SPSAR ACO CITIEAS ASIA DEM DIVA Les CIZ1l Uren’ Mngt Ren Mae (che A arte ac Peewee A On oars eee 265-270 Cred tered ciireseOiathe (Elissara Vale YE. tient pues tele sis Ot lereeles Suk S ewhis Saas AAS Ste oes 270-274 Tentative reconstruction of erosion cycles in Karategin and Hissar................... 274 ene a AeA ea nt cM PO eR to. cid ger a ee WOR Peck Khe aye eek wae me AON boTa pated e% hp 275-282 ie entender IOnoVeudinn! VAlICY. fo.c5 sate «is eww & wn tie cig wie en dhe Hom wets SN eS 276 ner eee RE CM CES MOE ae ce od Pe enh itiam Sa hints Sais eee es BR bie toe WSs ca e's yh ee RE ee RN ee Re Aa AN ie EE Suhaa os ra ue Pt Se. ong baid Sods oh ve ea ale 4 < 277 RAD OS NCC CO cae eae ote, 2088 5) e- ne a eng hares SC Waks Tighe. 6 pe aid ayes vn a em te wR os 277 lees aee tC MMEOMO MCLG Tk ss Crp eth Aap leee cin Sdn ncn ice gece BA args Cag eee Ac as We A 278-279 Pr CERIO COMM Car Rr, Mata el. hohe oak Aiea ce det Ovo odd ae hewn es hehe 279-282 aeaeaey Sea MR PS iat gk ers es A a A 2 © ie ee gre esha a so Se EWR ees o's sh aes os 282-286 Pao SE ay te Ta a ey ee a a ee 282-283 The uptilted piedmonts of northwestern Tarim as a key to the past.................. 283-286 Regier 6 serene COnnCe GS CRIA OY oe ecient Rok Sew a me Meh oa eee ae ey 286 sents eye recoiariction OL fhe: past tay Lariting 0). gcc dec GAs ade Ge Boa Save dis ae ess 286 Bare Ler ek sere a enrages er ia See 4 /s-crsiie waa Baar D Aeohvade eave 287-290 TES) AEOAC SO LUELIT CS awe RRR IES ven coer pet SAIN Bis ORR er Sc ek oe Gn A eo 5 287 PiGE. CCOMOME YCICS BASEL Gt TROT I BIGIG: POTS SA. ogc oe 5s coded ites sieve Kien ee eh 287 pigcleks All gist ad BERS Ct hoon ic eg ho ar Mig ah gta Mire 9 a, Saar en RA a 287-289 Rasvitial Ceforina tons Ole. PIAINS con. ts ca tyre wien ice ins. wad Mel Soup eS ae oy et 289 Present mectiation in the GeposiGom zones ten. eas mea os < ont la a vase siden ed ae 290 Reconstruction of past events of the Pergana Basin... (cl < de gs wie ee ee wee vs bee 290 Eien tease DASITT heirs s'0 th alk on He ro. RRM Paha h 6 OED icia! hn G 6 ba wh mate. » Se) eee bin &s 291-298 Complications and functional peculiarities arising from shape and size................ 291-292 Phesacustrian, zone (Aralo-Caspian Sea: expansions)<.0°.) 25.5 tha ery see wee age aes» 292-294 Recent developments in the alluvial and the flying-sands zones................0-0005 294-295 Recentenauces ia the, course of the Oxs. oi; ....g0cc, ries Mek atime «pies Cone Ry 295-297 Tentative reconstruction of the past in the Aralo-Caspian Basin..................... 298 CHAPTH RENCE ACHS MM ey Tame eile selene Shh a ee ae taete a te eames oben ea 299-337 Bt EG Siete. COP MICA TIPO PRCIIL Sh ys Sree ho 5.o ws nics « ul nih vac t. Seton arent pad kee Meret nigh Soha lng 299-307 Man as a geologic factor of excavation, transportation and deposition, and a director of TATU CROSICONS soc. 554 cso lbh Slash eat Cte BUNA ary 5 eo oS 299-301 Vv VI CONTENTS. PAGE, CHAPTER XV. OasEs.—Continued. The oasis as a geological problem.—Continued. Physitographic classification of oases. 2... 2. si 60s 43 Hewes or » tileie ee eee ees See 301-303 T; Delta oasesys\.... Facer elites ous Svante abe aeeksunkenees okey t ace tebe Meee bce meee eae 301 (a) OF fiver 5 ge.se iho ales eitigala op 00 Seine Skee Sine Shane 301 (b) Of-small streams, 6.0.15. ccs one ci cieteceker ee ania. cite sy eto en 301 II. River-bank and. flood-plain oases i: .-% 5. canis wisps eyes Woke p one npe 302 IIL. High-valley oases.» «3055 ms. ba « she pone li aot ihe a ayaa dion ale a ae 302 IV. Spring and well oases ....62....2 0 viscrnp + sities ge eben vip ce = ote eee 302 V.‘Lake-shore 0aseSisicf5. 268 oc hai em) «suet 0 eke ee le outlier ieee 302-303 Relations between erosion and burial in the obliteration of kurgans (oasis culture deposits) 303-307 The ancient Alai Valley route:from Bactta to*Kasho ars gnats yeisuseiea siv eisieecrsn terrier eae 307-310 Objective criteria of ancient long-used routes....2...2.5% 9.5 5 sevens sey a ee 307 ‘The three controlling factorsiof trade ,outes gem 6 is eee tae eee eae 307-308 Advantages of the Alai Valley route is .6 cs. icc 5 ince ole 0 «pnw ple et eee 308-310 Oases of the Zerafebran s «06 iss. tesBiance ok fa bea 6 dee nk RE ce 310-316 River-bank (type IT) oases of the lower Zerafshan. 04. amines ... -ei-e> e ee e B22 Recent decrease of surface drainage . 0... Oases. of the Murg-ab, Delta ci a2) save avcintcrsvetate s civen cele least el seneucians tee ete eet eae or 330-337 The river Murg-ab and the type peculiarities of the Merv oases, past and present...... 330-333 Irregularities of the delta surface wrought by man’s débris of occupation and his control of ‘allirvial Gepositions ao. acy cag Sepa ciate mile abe gel coer tenant 333 The great mound-of Odontche Tepe: 2.7 cccsn se sanmcirlak eis =e fo eee eee eee 334 Analysis of shafts at Ghiaur Kala. The deposits from man, water, and wind.......... 334-336 The stratigraphic order [(1) loess, (2) dune-sand, (3) alluvium] explained by climatic change to dry and recession of the delta .. 7.5.5.5. «obs tee ee ee 336-337 Part VI. ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU, AND THE Horse OF ANAU IN ITS RELATION TO THE RACES oF Domestic Horses. J. ULRICH DUERST. CHAPTER XVI. INTRODUCTION AND DESCRIPTION OF MATERIAL...........000ceecceeceeweees 341-343 CHAPTER XVID io ais ies ciSiapin- wi wlsrsciecdn Sue ae ee le Oe 345-354 GCarativora «ois iss as cig 65 os uc v alecelae plo laieie sles can Sn GOURD SNe eRe RRA ene Foxes of Anau J and Is 02 cc vows os dca ow ee nee ee 345-346 Wolf of Anau I anatomically compated..... opis epee es eee eee 346-347 Domestic Dog of Anau II anatomically compared with prehistoric and recent domestic and wild dogs <\o.:d5 5 ots, ecm eared rte ie ce ene Sect ene a ee 348-354 Probably importéd into Anat..2:.c.< inpue eek urtes fxn Seen ee ae eee 350 Rodentia ....6 05 abicxdd,cthcle Bev o wld de eae eI cee ee Core gy eer eae 354 CONTENTS. VII PAGE, See eRe aN LO er We ee ee eee raged ra Pes ks wma ia od ea evan zh mune wee Meng a vind we 355-399 CR Mae AM ee er nk, eres hae Cie shots cis tls, se a roe wes on eiw ho apa aa 355-358 The pig of Anau I (Sus palustris Riitimeyer) anatomically compared with prehistoric and FECEIL WANE COMCELCIDING Sem aan vie Coie muasiita ares és se MS. ned 355-358 Identical with Torfschwein (turbary pig) of European neolithic and bronze-age stations. . 355 UR Rag SRT Bo a ere Be ag ee Oe eC a a eee er 359-382 PeieenOteens Wid OX ( Os MMAGSCHS oes sie esl b ye keep cos vi cise ade Raw Swe veda ces 359-364 Anatomically compared with fossil and prehistoric bovids of Eurasia and Egypt.... 359-369 Domestic long-horned cattle (Bos taurus macroceros) of Anaul................2-005- Anatomically compared with wild and domestic cattle of Eurasia and Egypt....... 364-369 Domesticated out of the wild Bos namadicus of Anaul................00.0 cece 369 Identical with the Jong-horned Apis of early Egypt. :.. 22.0... 2.0000 e se enn naan nn 369 Ovina. Wild sheep of Anau I (a) (Ovis vignet arkal) anatomically compared with Eurasi- BW aH) 68) Sigg es cud ate A Sie Git ROW SAS ao BAO DE OO) ENE ey nook ah ene 370-372 IWoTMEStcESH CC sO A 12 lace ue eee ee TRON Tare hyn a RNS deere aryo resins adiic=s eegese) ss 372-380 Progress of domestication out of Ovis vignei arkal shown in changes in size, and in CHATACTEIZOL DOMCISLEMCE MLC ern erature te iemie exe rei Sein delicate ai oeuey 372-373 Towards end of Anau I a smaller breed established identical with Ovis aries palustris Riitimeyer, the Torfschaf of European neolithic and bronze-age stations. Anatomically compared with the wild ancestor and with domestic sheep of Wi atte RT ODM tee ns Mette es Cehigie, Met eee ey «sec cate ss 374-376 Rocher perte mee TTTLOO Mereees ee ee isla hg. Po dd Acide s orale eas en's mime ape esta) «as 376 EUG eSce LOCO LOLeA Tall il Meneame sprees orate) oer Lactate chat hays ince ais «hao pia eae wae nies 376-380 es eae eA CONICAL OTTERS ISOITE coos oo cn tai he w Sas hw ve ele wil PAO 374, 375, 377-379 Capra hircus riitimeyert Duerst. Domestic goat of Anau II, anatomically compared..... 380-381 Seat meae hr ete OLANE A TUELET Recs, Picts vik dace ake Sse os ely dliw’h sie Qin oleanin oe» oo 382 ested eM RELL EPS ed rege, & eo) Mead esicew ag bd PtP Shea WO wee 8 Sig dine 04.4 a Whe irigse ace » 382 Cafmraivles, (Ceri! ei 2 Urey 08 oo 5 des cancio-c 6 canis Ding a cic Dicey Pee iene, See nace create 383-384 PSISSOU UCLY Umm LO TSETOL GANT UL Yr uses cede Freee ye Gy foe esos iti ous: cxoyaus Be foy teh eeareecs 7 eke aigG poten? ate 384-399 Distinguishing anatomical characteristics of Western and Oriental horse groups and asses. 386 tIeGHesteet itn bel DLCS) mein deere nia tie erodes ier cane alec tetsretcl eta aregs .csval ci eusteyanauays love 387, 388, 390 PES (2e0S) era seen Gt) re ae i I cern aM eg ee 391-395 yee anutemuons oF the horsé, from Anatt Moy. oe. 0. ogee rae we pss easiness we 396-399 The diluvial horse differentiated into three types: Desert type (Equus caballus pumpellit), Steppe type (Equus caballus germanicus seu robustus) and Forest type PERsuee CLRAROELLS SECRUBADE 2k ONS. be neh sts Aa cashed Waco Kage Secs ork Boi sb) kis 399 Re MOe ANI Garn a0 PARTE OTE STAY Nc iisisgies a'arm a'r wais',+ wale 001449 tine vue wieie aes wc gee pe we 399 CHAPTER XIX. THe Horse oF ANAU IN ITS RELATION TO HISTORY AND TO THE RACES OF ADO AHO JeWernsoe 4s opin ae Be ee pe Gee ae tke CRED Onn Gh oe Deeg oe Cree 401-431 eR Ea aie OCCII OEE IO LIG POO UTI it ioe a nae 3 ese met Se =! i, ra Miwa ete oral Msg Fi ce nigidd vce. 401 OSS atl Cl ist DLOSSI< HOTSES meepemen Mer orem ture or tons ens ts lcdiia a! Ja voasy avitdmray e. guys sUeirletele nce tGP meld eoreah wats 's 402 FIOCSe Of Liter Uaterinar y pier ieee Ol, KOPN a2 alge ath tlk odes natin vgn age ha iat rh ew wee 402 SL tG: ORSES O11 DEC OIS OEIC AItKLOS i are sis, oa a waned oily chain Wei ss vw eee 403 ‘he horses ofethe nN eolitinicvA oe see racrs sretcts cette os chokes fe maces © ote ie nearer wiles, ove'el 403 Cramotovical differences’ between the ass.and horses do. joa.00 8. 5 se eteie sis Fe Fee Se ce av as 404-410 The horses of Europe in the Copper, La Tene, Hallstadt, and Roman periods............... 410 ees te Nrenistoric HOTSOS. cee ae meres oma ie fle ee eee aa eek ads et oa oles 41 Ream ry aie tans ee scat ca Moet so ath ates eth PBs Bic wie ah eat aces GR Sine be edhe 8 si ove hala a 411-414 Sree ee SMO TICL AES TOE OTE TIONS en che! ieee chen sins reg AS a Aaa tal) Hie eae leeds PR A emia AEN 414-417 Risen A CREO OX TOTALS ete 707 = dine sista ors «ae Bae apc rolkega ke a rteter tee a odes wari ates ooo, 417-426 Pace ated aires, CNG PTCHIstOric: NOTSCS aids sate a. ots oon es Spell egias Mp a with ainda oe a mea 426 PPT IIGRONUA eRe ta ee che eee oe toe a cs Reng Bape 5, avs AE AORN alk RURAET 19. 0 Mc nde ies, wry aes 427 Application of the measurements to determination of the relationship of the various prehistoric horses among themselves and to the horse of Anau.................-4-. 427 Pe. Pefciar es NOCSER OL TCULODE cain: s cited Se vids icc! ola sia nie Se Bilis w medeie ain, «Mae ep laty as ns ale 428 Only one type of wild horse over Europe and Asia in Pliocene and Pleistocene time; probably ES ANUIES SCN OIE CORO oma siete csi ce «aps Slee atv ahaa tg faleitenta vigor ae are Gs 428 Equus przewalskit its last wild representative. . 1.2.0... ccc sees eee eee ween e eee e nee 428 VIil CONTENTS. PAGE, CHAPTER XIX. THE HorsE oF ANAU IN ITS VARIOUS RELATIONS.—Continued, With post-glacial changes of climate began parallel regional differentiation of vegetation and types of horsesco. 3 aa Cavers ss helsie a els alee oa ee aes eho eee 429 The horse of Anau (Equus caballus pumpellii) brought to Europe in the bronze age......... 430 Used dater in ennobling:Romian horses. sto... pitw,- sw ce Lies 5 oR ee eee ee ee 430 Genealogical chart of the races of domesticdiorses! su. 240-2 «snes ees eee 431 CHAPTER XOX. CONCLUDING REMARKS 9 oj. e cgeleel tier eee ete te On ot ee ee 433-442 Modern stock-raising in ‘Transcaspia ov, .G\e. wo be ie we pint ses Ric Santee ena i eee 433 Mucke’s theory of domestication 2.5 2)4 e126: scia 9 os ving sds = 0.0 us Ne ee 436 Changes in relative percentages of different animals bred, due to secular climatic change...... 438 Importation of turbary sheep:into’ Huropeé: 4. 09) eee ee oe ee 439 Their occurrence at Schweizersbild discussed ./sa)-— 1.) ee ae 440 Dispersion of Anau cattle over the prehistoric world 2%.) 002 2 enact oe) ere oaue ieeee 440 Significance of the ox-cult: 2. 25.5 .0¥ 5.) oils oben ee eee ee 2 i en ee, 441 Part VII. DESCRIPTION OF SOME SKULLS FROM THE NORTH KuRGAN, ANAU. G. SERGI. CHAPTER X XI. SKULLS FROM ANAUss .. % eos c des nb covets kin one 445-446 Representatives of the two oldest cultures differ absolutely from Mongolian type; closely resemble Mediterranean type. «ca 4 aires fu a overs ne 446 © Table of measurements «2% :cc sc aie oc cedk saute Dw wo eet > a Kee ee 446 Part VIII. Some HuMAN REMAINS FOUND IN THE NORTH KuRGAN, ANAU. ‘TH. MOLLISON. CHAPTER XXII. DEscRIPTION oF SOME HUMAN REMAINS FOUND IN THE NORTH KURGAN, ANAU © ao ooilel a a) Saty wine atapetdeg gino wae bye oe ti ahi ee ede ie cae eee 449-468 Enumeration of the bones from five adult individtials.........3..) 4 sus se nas) ee 449 The bones of individual I, from Culture II, described and racially compared................ 450-461 The bones of individual-II, III, [V;and.V described... ....: .c.0% + ee ee 462 The bones: of children/from Culture’ 0. 2.) a 2. ote ste oe cates cel) eee ee ee 463 Estimation of the height:of the individuals... 2.65. i040 .s%s0 gos tae Eee ee ee 463 Tabulation of the measurements’. 2...) deci oscr sistanetie cc cie. heceic.o ecg etene eee ee 465-467 Bibliography . 2. 2v/a.« o: alsin Ge De Seg Seal oe oyu ow ele ha gle a le ey pwr 1g ne OO 468 Part IX, WHEAT AND BARLEY FROM THE NORTH KURGAN, ANAU. H. C. SCHELLENBERG. CHAPTER XXIII. THE REMAINS OF PLANTS FROM THE NORTH KuRGAN, ANAU............-. 471-473 Charcoal ‘fronmi Cultures Tand lL sisauhe soe als Sete nt greiner tere cease een ane 471 Casts and siliceous skeletons of wheat and two-rowed barley from Culture I............... 471-473 Appendix to Professor Schellenberg’s report, and-note by R. P..............00-e eee eees 473 Part X. STONE IMPLEMENTS AND SKELETONS EXCAVATED IN ANAU. LANGDON WARNER. CHAPTER KXTV oes. ows ose lvid cn a gigis swan oie « Wim rust edie ib REL NORe Bl cist a 477-494 Report on the larger stone implements of the Kurgans at Anau....................-..- 477-482 Summary of work done in Terrace Il, North Kurgai. 9.0.2 5 ess. oe ey eee ee 482-484 Skeletons‘excavated in North Kawgati. . cs ue. wep ates eee eee oe eee re 484-491 Skeletons excavated in South Kurgan... 1. 2aes ee ee ee ee ae ee 491-494 REsumMe 2 ican cae as vieaae Seale ve Be 0B cerse Ameen plete anes yan aan orien 494 LIST OF ILLUSTRATIONS, SECOND VOLUME. PLATE. FACING PAGE, Pamenrae ig enital Asia: (oatiarkane tOnG ASHGAL) ofcg ch wee ee se fs arate wn O8 co yee ehh eee 243 fica twelve cross-sections of the Zerafshan) Valley oo. O). og 0. 6c cick Sad oc p See okie eee cele cu ee bees 278 63. Profile of valley terraces and moraines from Osh to Kara Kul (1903)... .........-0 00200 ee eue 298 PEMD OrmerOl Kilt ate in Mergarin ohios ey. cer th hes ihin O% chek Ss whe he Oe PEN Eek cos be sb wees 314 ea ree Ctl 16 10 ere Rater ae A eerrivis ai Sites in bs na a Sul he 2S Od ete Wie ah Pen 320 ee Lea em Oe cn nr EN ale Bie See sorcerers ings tied GOW A Aa aS WE RA A wee a BD gm Licinem Kea ArShiclt Caf ll OL enue Weave eae tee Meee ei Barra car cee Wien.’ oe See area celal EMM Senne sa, « 334 See enn es | Vee GUC) SOLES TOML AAU. or ow ae Os sis hail Se CN wie Ho ae es Ee od ee ee Sa ee 346 7 pee Canis matris optime from Anau compared with same from Bohemia. Bones of Sus from QM 5 AIC 6 oi oohs yciostp ob 6 Gan OIENO © Oi LIB © Cig hol BIC Res Corr Re ata oi Ce es ee 346 73. Bones of Bos namadicus from Anau I, Bos brachyceros (Bohemian turbary),andcamelfrom AnauII 360 74. Comparisons of basioccipital bones of Ovibos, Bubalus, and Taurus, and teeth of Yak, Bison, Bos, EREMOU WERTH OTATNS OS 2 ay 0 caitereyceaeet cation Cte Geer acca Ga OE er aI Oey ORCA an Ace REC ee ae 360 PDB cRO RCI SLOT PATIANY IOAtl allen manent eck. ered oko! eit hae OR cre le ohio tre cite plete pie ahs He 370 76. Horn-cores of Ovis, goat, and antelope. Metacarpus of Ovis and goat. Basal part of antler of MRE sOL NAM CUP LITIGATION testis ten dae DSR. serge wospe ten SENS hs wh scehes Mists sels) « sunray aide etiathaes 370 77. Bones of Equus from Anau I and II and Solutré, and camel from Anau II................... 384 78. Bones of Bos from Anau and Bizino, and skull of goat from Crete................0.000ee ears 384 79 Egyptian Canis (marble, in the Louvre) and long-horned cattle (tomb of Manofer, 2600 B. ¢.).... . 384 80. Skulls of Sus cristatus, after Rolleston, and Sus scrofa var, ferus, Germany..................-. 392 81. Skull of European urus in British Museum, and Hungarian bull skull in Vienna Museum....... 392 82. Skulls of long-horned ox from Nepal, India, and Ovws vigner from Ladak...................... 392 83. Assyrian and Babylonian representations of wild bull and long-horned ox. Horn-cores and skull Se euataty COST PACE EIT OLE, MOV IAO) oi cece ae 2a ee kN oe OSE Ne LSU be a ee ae 396 See C men eaeenesreciimure fromoNine velo .\.oc 2. . ten. oe Sie Aloe Se vedas he eed aed lew es bs 396 85. Map showing distribution, ancient and modern, of Bos tawrus macroceros Duerst............... 400 86. Map showing prehistoric and modern distribution of Ovzs aries palustris Riitimeyer (turbary sheep) 400 87. Norma lateralis of skull of horse from pile-dwellings at Auvernier, Switzerland, compared with same view of skull of Egyptian prehistoric ass of Abadieh (IV dynasty).................. 410 88. Norma verticalis of skulls of Equus przewalskit and various prehistoric horses................. 410 89. Norma lateralis of skulls of Equus przewalskit and two prehistoric horses..................... 410 go. Comparisons of metacarpi medii and metatarsi medii, Anau horse..................-.-0 000. 420 PIMC atisuG Olle x Letucy WOUES wALIATIBNOTSE. «tice cuveidehe ts Hela fo eis Tp ale sda soa dials «ura 420 See ends kuls Jrom. North Kurgan, #AnAt Sh ss teats ys ac lee ee epee eres mes Sie ae Se he 446 94-96. Human extremity bones from Anau in comparison with those from other localities....... 464 97. Glumes of wheat and particles of barley in potsherds from Anau, compared in microscopic views Wi theeGenta wile tall (aDAL leva matin amination Raly. teas tara) ork aie, s) cts Pele wiieatiavs adehevel sono 'sia)s¥eicvere seve 474 TEXT-FIGURE. PAGE. et Nea ee Ee CER PCIICNE RSET EL Ut oe We 10 anf sia yep vlogs inks a leon evita Rash Sawa le ba Bure Sow ee eee 251 432. A granite ridge at Kara Kul (showing the secular deflation of the Pamir).................... 252 emmoand agnceareose tesidiiim, or defhation (kata Kil) iin. oc e.c. ccs sts 6 sms aS ws se wala seas 253 434. Glacier of the Kara-at and a section of its valley 1.5 miles below (Kara Kul)................ 255 435. A pool, where Kara Kul sediments have caved in over melting lobes of glacier ice............ 255 436. An ice dome and its third and fourth epoch moraines in the foreground (Kara Kul)........... 256 MeymnAT lacie Nottheastion Karay Kut see saan t letter cta: sie) apaoys, sek hada. sche Sse aks tes tue sles Rs alinie @rlaens. oreieieye 4 257 ener eIPLLINIe 1th Ther A fas, Valle yy oh Neal ns bg gv titye a NT cae, ale wy he Wath wien Sates aes Sw Siatane 260 er wuninanaking felt in/the Ala Valley facts. . shin vite iy pm re ose eee 'ayy io eeele Salen oes oe 261 Peme orcLarar i iurieet Stacier (Aldi: VAMOY ) ren are tat de a noi hee s oa Nios ews Red viel ce oS oR Gi 262 Wixoninvial terraces in the lower end. of the Alar Valley: 227. 6.2 ects cs wich oe mes hae ee ns kas 263 Rea 11741 pu where it leaves the Alai: Valley toma aap axetscmigne a we certs ee PSG yes sz eS 5 688 265 aaeern Galcna Bee of Karatepin with his hunting eagles. i. see aint eee be ie eee nee sees 266 ume aCes CL ATI Ml Zi) Stl it) MR ATALCOIN ey sie alte ciate circa eile nop muted, GEM vas aca sat Gece 267 meemahnridge over tie tird-cycle terrace 1m Karate gin oii amine ak ats wield eiiicia, cw Ve ve 35 Ole ils 268 Feb NMCREA TESS as URE EN Dak alee OR Ri a Ne eR AA Oe deter ue boMe Wari ra evecare MR Stat ee 269 Pee SW er Onta DIS (RK ATAte Li irc swe de ries wie euge Me eh ee eeitele Pan Ns adhe AS wy 6 270 Seem RE ES Riis RUG SAUL SSN eS So enn 924, 0 ice ly ge oe vw iplg shell 8 ee nm, ad ee PTE. ease se 271 cto dh, GINS ENUF gi ae Ver iN 0 Sek sry ee ORE)” Sree wom Rea AT cre Ok aches ce PERM aor eae aa ea oe we 27 Peper eeerroaie of tue Piissar VaUeys.ra ay: sel ee eas) cw

ao nlerae swe See ge aie ees cere > eee ee eee Cliff-trail up. the Zerafshan gorge’. o. 22 sSs2. sca ob to es ee Se ee ee A “high ‘valley oasis” of the: Zerafshanporge. <. = sc. ssn )... eee ee ee . Sagittal sections through the tibia-talus: joint... ..25 ss". «csc . vas we ae STONE IMPLEMENTS FROM THE NORTH AND SOUTH KURGANS. 496-502. Stone implements from North and South Kurgans.................2cccceccccsvccscses 503-508. Stone,implements from North Kurpams 32-0 6 ae4.. sees ne oie ee 509-516. Stone implements from South Kurgan. (202-7. oe i a waitin ie ie ee 517-524. Pestles and’ mullers from: South Kurgan 7). 4... 0. eee eee 525-528. Stone implenients fron1 South Kurgan: o.0ct 22. 2. eis scieie oe ieielc Sars oe tee Shee eee 529, 530. Stone implements from Poth Kirgan. 27 < peee ose ee eee ee a ea Type of mealing-stone im both kurgans..3 « ... .s omc ds = 5.45 75 22s se eee 532,533. stone mortars from South Kairos, Cil tyre Vee ye ree cee SKELETONS FROM THE NORTH AND SOUTH KURGANS. 534-536. Nos. 5,2, and 3, Terrace I, North Karpanio.) c. goa ore a, «ne ee ee 537-539. No. 4, Terrace III, and Nos. 6 and 7, Terrace II, North Kurgan...................-00%- 540-542. No 8, Terrace VIII, and Nos. 11 and 14, Terrace II, North Kurgan..................-. 543-545. Nos. 16 and 17, East Gallery, and No. 18, North Digging I, North Kurgan.............. 546-548. Nos. 19 and 23, Terrace B, and No. 28, Terrace C, South Kurgan..................0028- a = a 7 i a = is - ' 4 - oar ~ + 4 : a. we + aa _ i am) a 7 8 a és Pete 7 7 Ms . <= * . ~ . PART V. PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. - - | : A : ~ ; “i iy, \ * Ps e ; ait Pej F N ae 3 ra % + B- u Win Gg ss 2 O~, < ze A i vi “ a3 % he Ss dy 7 5 Ree: AS o-— 3 j ‘ ¢ A y . owe * op -. ¥ \ 4 4 Qe RJ | é Zaratshtin i . . ” ‘ vie} 8 A . 0 > i\ ? — ie 5/2) . Rt ; Ray y f a | SS y 6 7 ns ‘\ : te y\ PAE 2 A ek ? = : Re. Lf OAT nar 4 , é o < J ; ; P, J , as ., = ae iG 75 70° east From a map issued by the Military MAP OF CEN TRAL Bn Lene, SAMARKAND ip QO KASS E. GAR Topographic Division of the General Staff of the Russian Army, 1886 () 10 20 30 40 50 60 70 80 90 100 miles Route of R.W. Pumpelly, /904 Ss “Save Tchinaz ¥% s ’ emcee } ‘ ¢ ae AA v Tan ee UW) ‘ ‘., age ~ , @ ~~ “T A oe _ CHAPTER XIV.—DESERTS. As the earliest traces of man in Central Asia show him living in oases, as he does to-day, the archeology of that region is centered on a study of the oases of antiquity, and this study must include surrounding deserts, for climatic and topographic changes have been everywhere so pronounced and have left records so intimately interwoven with the records of man that no comprehensive search into the conditions of his past life in Central Asia can be undertaken without some fundamental ideas about the cause and effect of desert changes. For our purpose a map of Asia should be regarded as no more than a passing picture of a struggle between land and sea, and mountains and storms—but one phase in the evolution of a continent. And it is not only the archeologist and physiographer or those who probe into the remote past who must assume this dynamic point of view. The pace of continental change is so fast that a man in his lifetime can watch the trend of great events, the change of great features, and know that hardly a branch of human affairs exists but must feel the effect. THE DESERT BASIN AS AN ORGANIC WHOLE. Already in early Pliocene time Asia in her immensity had developed a vast interior region of desert basins, into which were spread the wastes of then large existing mountain masses, the cores of which still remain as worn-down granites of the Tian Shan, Pamir, and Tibet. When treated as an organic whole a desert basin forms one of the most interesting features of our planet, and the laws of Nature, under which its oases are controlled, must be of vital interest to man. THE ESSENTIAL CHARACTERS OF A DESERT BASIN. For our purpose this organic whole includes the area that drains thereto and may be divided into two parts—the inclosing mountains, ever worn into new relief by the storms of geologic time, and the plains they have created by the slow building, layer on layer, of their débris brought down by the waters and winds of those storms. ‘The first requisite of a desert basin is aridity, sparsity of vegeta- tion, and too little rainfall to carry the products of erosion away to the ocean. From the beginning of its definition as a desert basin self-contained, or hydro- graphically isolated from the ocean, each basin was destined to accumulate on its plains the débris of erosion sent from its inclosing ranges, and the inevitable sinking of the earth’s crust that yielded as that load increased to a thickness of many thousand feet appears to have resulted in a corresponding upward displace- ment of its border ranges. How long this great process has been in operation may be conjectured from the immensity of beds of fine red desert strata that reach a thickness of many thousand feet where exposed on the uptilted borders of the 243 244 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. Gobi basin. ‘There, as in Fergana and along the base of the Kopet Dagh, these uptilted piedmont formations are a magnificent corroboration of the idea of dis- placement, as well as of Suess’s theory of the encroachment of mountains on plains. A desert basin, then, is organically divided between mountains and plains, but this is only the beginning of our classification. Functionally, the plains are a vast geologic mill in which the material received is differentially assorted into layers of fine and coarse alluvium, whereof the surface is further sifted into loess and flying sands ere it comes to rest under succeeding layers; and this mill is worked by wind and water. It is a plain whereon the muddy floods of spring and fall give rise to momentary shoals of water spread over many scores of square miles, welcome lakes that vanish under the burning sun, to leave mirage and wind- swept barrenness of sandstorms and yellow days. But of this water some remains more permanently wherever the supply is in excess of evaporation. ‘The life of a desert lake or sea is, according to circumstances, anything from a day to a cycle of geologic time; anything from the momentary existence of a thin watery sheet far out among the dunes to the history of an Aral Sea. Thus arise four marked subdivisions of deposition—alluvial, lacustrian, flying sands, and loess—two of water and two of wind. Lastly, but perhaps most important in records of Quaternary change is the fifth subdivision of deposition, glacial ‘‘till.’”’ A desert basin is thus divided into areas of erosion and deposition, mountains and plains; the plains are divided into four zones—alluvial, lacustrian, flying sands, and loess—and the loess zone includes a portion of the mountains, while a fifth subdivision of deposition, glacial, is found on the higher mountains, THE THREE AGENCIES OF EROSION AND FIVE DEPOSITION ZONES. There are three agencies of erosion and transportation, ice, water, and wind, and five deposition zones, glacial, alluvial, lacustrian, flying sands, and loess arising therefrom. Rising among the glaciers and snow-clad peaks of the inclosing mountains, small and large silt-laden streams discharge upon the plains. Flood gives way to the drought of a burning sun that stirs the atmosphere into vast cyclonic storms and spiral dust-whorls—tall, shadowy forms that come and go in ever-changing shape, born out of the horizon to wander a while and vanish. By these atmos- pheric disturbances the surface materials are consumed and sifted over, digested into drifting sands and far-blown dust. Most of the dust is borne far away to rest as loess in the grass of high valleys and plateaus flanking the peripheral mountains, for it can not survive a wind on barren surface; but sand moves slowly to and fro in the shifting winds, and only that which gets beyond the ultimate shores of alluvial activity accumulates to form the larger masses we call dunes. Probably the most important source of this sand lies in the more or less impure sandspits that are invariably to be found after flood along distributary channels of silting streams. Any shifting aggregations that have not found their way onto an area more or less permanently free from alluvial activity must suffer rearrangement by the next flood, but in the vast nuclei of flying sands that characterize the desert plains we have ample proof of the large scale at which the wind has been successful. DESERTS. 245 Interior brackish seas and lakes, fed by the more constant flow of larger rivers, have played an important rdle in the history of the two great desert basins of Eastern and Western Turkestan. The surface area of a landlocked sea is a direct function of the climate of its basin. Its salinity varies (1) according to antiquity, (2) in some instances according to whether overflow took place, and (3) the relations between its surface area and amount of salt in its supplying streams. Although at first thought these appear to be the only controlling factors of salinity, there are yet four considerations, four ways in which an interior sea may diminish in salinity or totally lose its salt. It may dry up entirely for a while and the salt thus precipitated over its bottom be either (4) blown away or (5) so perfectly buried by sand or silt that future water is unable to redissolve it; (6) much of the salt may be precipitated in a gulf with such a narrow and shallow strait that water evaporates faster than it comes in from the mother sea, as it is now doing in the Kara Bugas and other gulfs of the Caspian. There the super- saturation is death to all sea life that comes in with the current—a graveyard of floating fish. Lastly, (7) as a theoretical possibility, sufficiently rapid sinking of the bottom or falling-in of the earth’s crust, such as appears to have taken place in the southern half of the Caspian, would give rise to a greater volume for the original surface area. ‘Therefore, since the surface area is a definite function of climate, it is obvious that if the volume were increased rapidly enough dilution would take place if the rainfall remained constant over the catch-basins of its supplying streams. To sum it up in a more general way, it may be said that, with a given topography and given chemical nature of rocks drained to start with, both the salinity and surface area of a landlocked sea depend upon the variations of climate and crustal movement that may take place over its basin. These facts are of fundamental importance in the consideration of lake-shore oases or type V.* It is from a study of the ancient shore-lines, sediments, soundings, and fauna of now shrunken seas and dried-up lakes that important records bearing on the archeology as well as physiography of Central Asia are to be drawn. The whole peripheral area of surrounding mountains is, in the long run, sub- jected to erosion and worn down to build the plains; but for closer analysis it is functionally subdivided into areas of erosion and deposition, with limits more or less temporary or unstable in definition. First, the high crests and valleys of greater ranges give rise to glacial zones of ice mantles and valley glaciers of which the interior portions are seats of glacial erosion, the margins and termini seats of glacial deposition. Second, the lower plateaus and foothills of pasture land have been the chief seat of loess accumulation—the settling-ground of dust blown up from the plains and down from the more arid heights of deflating rocks. Third, the greater areas of peripheral mountains have developed intricate basin- systems, some as half-closed valleys tributary to the great plains, others isolated or wholly inclosed, imitating on a small scale the greater scheme of which they form a part. *Chapter xv, Physiographic Classification of Oases, p. 301. 246 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. Wind, water, and ice erode the mountains and have particular deposition areas within the highlands, where part of their products accumulate more or less temporarily to form loess, alluvium, and moraine. But as these positions are, in the course of long time, unstable, practically all the products of erosion must ultimately find repose in strata of the great interior plains. Moraines, however, are very resistant to transportation from their zone, and massive remnants of those deposited in even the beginning of the glacial period still survive in situ; and it is still possible to recognize sections of alluvium and its wind-blown deriva- tives, sand and loess, deposited during the glacial period on now dissected high- lands and broad valley terraces. Within the latitudes of Central Asia, glacial deposits are confined to the peripheral mountains of desert basins. Areas of alluvium, loess, flying sands, and lacustrian deposits are found more or less unstable on the highlands, while the great alluvium, flying sands, and lacustrian deposits have their special, well- defined, and concentric zones respectively one within the other on the great interior plains of a basin. Loess and flying sands are in large part the wind- blown derivatives of spread-out alluvium, but they are undoubtedly much added to from direct deflation of the more arid highlands—a fact to be demonstrated in the section on the Northern Pamir. As lake deposits are simply modified alluvium, there are only four primary derivatives of the original mountain masses— moraine, alluvium, loess, and flying sands; and much of the alluvium is modified moraine and the direct charge of glacial grindings. An ideal desert basin, not over about 100 miles across from crest to crest, would be completely inclosed by mountain ranges. Its glacial zone would be a periphery of ice mantling the crests and extending into the heads of high valleys; its loess zone would flank the base of the mountains thus encircling the plains; its alluvial zone would lie next inside as a piedmont belt of the plains; while the flying-sands zone would lie next within as a wide belt surrounding the lacustrian zone or brackish sea of the middle of the basin, reached by large streams rising in the glacial zone and crossing the loess, alluvial, and flying sands. In Western Turkestan the lacustrian, alluvial, flying-sands, and loess areas are now four well-defined zones, respectively one within the other, loess on the outside, while glacial deposits are naturally confined to the higher mountains and nowhere reach below an elevation of 7,000 feet. As an exception to this generality, the rivers Amu and Syr penetrate to their inland sea, the Aral, thus dividing the zone of flying sands with two narrow extensions of the alluvial. The lacustrian is thus united with the alluvial zone, and this is more often the case in Eastern Turkestan, where flying sands are divided into several wide nuclei by long rivers that traverse the basin. Although it is to future exploration that we must look for comprehensive records, a general outline of past conditions may be construed from our observations on the five zones, together with sections in earlier layers, exemplified and checked by a study of the topography of erosion. DESERTS. 247 THE INTERLAPPING OF DEPOSITION ZONES EFFECTED BY CLIMATIC OSCILLATIONS. Owing to their low altitude and the hot column of air rising from them, rela- tively little rain falls on the plains and precipitation is for the most part confined to snow upon the mountains. Obviously, a regional decrease in precipitation involves a general shrinkage of glaciers, lakes, and areas of alluvial activity, and a sympathetic expansion of flying sands over abandoned portions of both lacustrian and alluvial zones, while the alluvial zone would recede mountainwards, encroaching on the loess zone, itself undergoing shrinkage for lack of rainfall on areas where grass had scarcely existed under the old supply. And it would be vice versa with increased precipitation. Continued oscillations, then, would bring about a column with alternating lacustrian sediments and modified dune-sand on the inner belt of overlap, one of alluvium and dune-sand on the middle belt, and one of alluvium and loess on the outer belt of overlap, while buried erosion surfaces of dead loess should be indicated in sections of loess where it has felt the change. And there would be successive moraines of different epochs overlying each other in the glacial zone. If these oscillations were sufficiently great, the middle belt of over- lap would alternate with loess, alluvium, and dune-sand repeated in that order, unless the topography was such that an interior sea would expand to consume the whole area. ‘Thus would climatic change record itself. The task of finding records in the mountains is in some ways easier than on the plains, in others harder; records there are on a large scale, but those of climatic variations are so tangled with those of crustal movements that, if it were not the constancy of upward movement, which in itself seems to involve a peculiar kind of climatic change, the task would be well-nigh impossible. Moreover, data like that of the shifting of man’s abode, so often found on the plains, are almost lacking in the mountains. It is to the topography and glaciology that we must turn. If uplift of the mountains had only been so simple as an equal and unbroken uplift of all the ranges together, it would be an easy thing to trace the stages of topographical developments; but unfortunately it is the inequality of recent uplift that gives the mountains some of their most striking features, fault- scarps and high-tilted blocks. THE CYCLICAL DEVELOPMENT OF AN IDEAL DESERT BASIN. To throw any light on those changes enacted by the deserts of Central Asia since the advent of man upon them, it is necessary to incorporate a reconstruction far back into the four controlling and more or less interdependent variables—uplift, erosion, aggradation, and climate. The most vital question is, What was the climate at any given time? But its solution depends much on the other three. Beginning with a theoretical development of these variables, let us picture the life of an ideal desert basin of the simplest kind. Born under the impulse of terres- trial forces, spontaneous adjustments in the stresses of our planet’s crust, its complete periphery of high mountain ranges would then be left to the tools of solar energy; and the aspect of such a basin as a whole would alternate between 248 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. the extremes of a grand cyclical evolution worked by solar energy under the direction of gravity, a geologic drama enacted by mountains and storms round the plains whereto the mountains crumble. Picture its massive ranges slowly wearing away into their rivers and the steady building of the plains till naught but the low relief of a gently rolling surface, half-buried in its own piedmonts remains of their once colossal heights. The shores of its wide interior sea are after this not so desolate but that sufficient rain falls to nourish grass over the surrounding steppes, rain from the moisture that would have been combed out by the mountains when they were higher. It is a basin worn to low relief and, therefore, receives a precipitation more evenly distributed over space and time. But though this does mean that some of the water which formerly found its way into the rivers and sea must suffer direct reevaporation on the steppes, and thus expand the loess zone, we must believe that so much more will find its way from the oceans to the basin that its rivers and sea would swell to relatively high shores. This brings us to a critical point, the end of the first cycle—a point of minimum relief, when so many thousands of feet have been unloaded from the mountains and loaded onto the plains that the ultimate strength of the already warping earth’s crust is reached, and the shearing stresses set up along the borders of the plains result in far-reaching faults. Then begins the second cycle with a long period, during which the plains sink slowly and the mountains rise by displacement. We can watch the gently graded hydrographic systems thus uplifted changed into torrential streams deep canyoning the rising peneplain and gently rolling slopes of worn-down mountain cores, while on the higher masses the acceleration of glacial conditions is reexpanding remnant ice-domes and even giving birth to advancing valley glaciers. This process continues till the crust has almost reached an equilibrium and regains its old rigidity. The first cycle is completed; and we have returned to a relatively greater concentration of precipitation on the mountains and intense aridity of the plains. Less moisture finds its way from the oceans to the basins and a shrinkage of rivers and sea has lowered their shores. Once more there is a desolate expanse of flying sands and relatively little grass and loess on the steppes. Our second cycle now is drawing to a close; a rapid carving of the mountains into deep canyons soon widened into immense gorges floored with broad flood-plains develops into the graded conditions during the crustal rest in the end of the second cycle. The continued shifting of load thus brought about ultimately results in a second yield more or less near the old lines of weakness, and a third cycle is ushered in as the plains go down and the mountains rise again. New canyons thus incised in the old valley-floors have cut down, leaving a terrace above. The mountains are perhaps higher, but sharper in relief, and the interior sea has shrunken. Con- tinued cycles result in manifoldly terraced valleys and a series of abandoned shores or terraces along the coasts of our interior sea, and an interlapping of the deposi- tion zones. The actual course of Central Asia’s development has been more complex than that of ideally simple basins. The Eastern or Tarim basin is, to be sure, DESERTS. 249 practically closed around by high ranges of the Tian Shan, Pamir, Karakoram, and Altin Tagh, and comes near to an ideal type. But the Western basin, the one most studied by us, still lies open out over the low steppes of Siberia and Southern Russia, with but little protection from the Ust-Urt, or low-domed ending of the Urals. It is, therefore, a basin less isolated and less independent in its variations of climate than Tarim and other divisions of the Gobi. We must expect to find that it was more directly influenced by the Quaternary ice mantle over Russia, as well as by the mundane change that brought about that ice. And, as already hinted, the cyclical uplifting of mountains was by no means simple, but took place in huge block-masses rising and tilting in various degrees and relations. Other facts not brought out by our imaginary basin, but of marked importance in all large ones, are the disturbances resulting from unequal loading of different areas and the encroachment of mountains on plains by migration of the tilted piedmont belt. This is especially well exampled in the western part of Tarim, where the strata have been tilted up, forming long ridges several hundred feet high with fault-scarps facing mountainwards and surmounted by gentle slopes towards the center of the basin. There are sometimes several parallel to each other and 30 or 4o miles apart, the innermost youngest in appearance as though migrating from the mountains inwards, stage by stage, to follow the zone of maxi- mum deposition-load as it moved inwards from cycle to cycle. Why this encroach- ment took place we can only guess, but that guess must fit in perfectly with the whole scheme. I have spoken of the mountains rising by displacement, and there may have been some doubt as to how that displacement took place. Was it flow in the hard crust of the earth, which is thought to be miles deep? Or was it a displace- ment in the supposed magma beneath? There has been a prevalent idea that some ranges piled up from overriding folds supplied by horizontal thrusting from both sides, and that others formed by simple folding on lines normal to horizontal compression, thrust in either case arising from a general contraction of the earth and wrinkling of its crust. That such was the primal origin of our Central-Asian ranges may be assumed, as far as this report is concerned. Indeed, the structure of their interior portions would lead to that conclusion. But without entering at all into a discussion of their primal origin, a thing that lies back in one of the earth’s great mountain-building epochs, we are still confronted by that series of peculiar secondary movements of a desert basin with its mountains. Then, to explain the hypothesis of shifting-load displacement hitherto assumed: The plains sank with their load of sediments; of that we have ample proof both in the fault-scarps of high-uplifted blocks along their borders, and in the Askhabad well-boring, 2,200 feet deep, which penetrated the plains to 1,400 feet below the sea-level, and even then remained in red alluvium and loess like that laid down on them to-day. It is absurd to suppose they sank into a cavity; we must assume that matter was displaced. The obvious corollary is that what- ever displacement took place was to the nearest zone of weakness of the earth’s crust, and there is twofold reason for supposing that the mountains lay over 250 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. that zone. ‘To begin with, their existence resulted from a tendency to yield along that line and a great load once borne by them has been taken off. So the plains sink and it would seem the mountains tend to rise and that the displacement is to beneath the mountains. Assuming this, we still have to account for uptilting of piedmonts along the border, and often two or more such lines nosed up with fault-scarps facing the mountains. This fact is the key to a deep-seated process that forms an important corol- lary to such displacement. Deep-buried, flat formations are under an enormous pressure that will squash the strata into thinner layers, if horizontal expansion is possible. There is, therefore, reason to believe that such a sinking of the plains of a desert basin involves a deep-seated squashing of the layers, a flat squashing out into folding in the nearest zone of yielding weakness, 7. e., under the bordering ranges that rise therefrom. And the crust being broken along the'shearing lines, this zone dies out into bending moments somewhat under the plains, so that minor folding may take place under there also and bend up the piedmont edge. It is also conceivable that more outlying folds would result with their surface phenomena; that is, the fault-scarps of our uptilted piedmonts probably pass beneath into monoclinal continuity or other folding of our squash- thrusted layers. One of the most striking facts about Nature, especially geology, is the perti- nacity of habit and the accumulative power of habit to overcome its obstacles. It is to this that we must attribute the fact that in reality each succeeding cycle of our process appears to have occupied less time than the one before, and that the first was vastly the longest. Moreover, it is only logical to suppose that it took a much greater shifting of load and, therefore, a much longer time to shear the crust on a virgin line than to renew that shearing in an already weakened zone. We do not intend to assume that no other terrestrial movements took place during all this time, nor that some more regional or mundane change of climate did not have its influence on these more local changes. All that is possible now, even in a tentative way, is to trace the influence of a desert basin on itself and try to superimpose its climatic cycles on those phases that it felt from the mundane glacial period. j Having offered these explanations and methods of attack with the above working hypothesis, we may analyze a series of notes on land forms. Beginning with the Pamir (Roof of the World), the following sections of this chapter will take up certain valleys and the chief basins of Central Asia, and each section will close with a tentative physiographic reconstruction of the past. DESERTS. 251 THE NORTHERN PAMIR. GREAT FEATURES OF THE PAMIR. As a general key to what has been happening in the mountains, the Northern Pamir with its border ranges stands paramount. On a map of Asia it appears as a massive knot of intersecting ranges, where high members of the Tian Shan system conflict with a northwestern extension of that immense plateau of moun- tains called Tibet. Offhand, we should expect to find a heavy precipitation of snow upon such a high uplifted mass. But looking down upon it in reality we behold a desolate expanse of barren clay and stone, with only here and there a small white blotch of snow, and some few desert lakes; a very high plateau crisscrossed by mountain ranges inclosing a multitude of broad barren steppes that sweep in graceful curves from range to range. While some of these are trav- ersed by streams, many of them are undrained depressions with or without lakes. We look upon a vast, extremely arid wilderness, void of trees and almost without any vegetation; a nude expanse of gray desert steppes and worn-down mountains with many-colored cliffs, of which the higher rise to white-crusted domes of ice. Fig. 431.—Lake Kara Kul (North End). This remarkable aridity is perhaps the most emphatic demonstration of Central Asia’s isolation from moisture. In the Pamir we have a region whose depressions lie from 13,000 to 15,000 feet above the ocean, and whose mountains rise to from 18,000 to 24,000 feet in height. Similar latitudes elsewhere record a snow-line of 10,000 feet elevation, but now we are dealing with an interior region surrounded by the greatest mountains and deserts of the world. It is, therefore, logical to find the snow-line at 16,000 feet with Sven Hedin’s report of less than 1-inch precipitation over Kara Kul, the salt lake of its widest basin. The Southern Pamir is less arid, as it receives about all the precipitation of southern winds left after the Hindu-kush and Karakoram have had their share. THE BASIN OF GREAT KARA KUL. The basin of Great Kara Kul was studied on two expeditions; the first in 1903, under the auspices of the Carnegie Institution of Washington, the second in 1904, an independent exploration, from which much of my data on other regions will be drawn. And though the basin has been discussed in my first year’s report some repetition is important for the sake of correlation. 252 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. It is an undrained depression 20 miles in diameter, with a lake 8 miles wide nearly divided by two hilly peninsulas of ledge rising from the sediments of its northern and southern shores. Whether it is wholly the result of moraine damming or in part a genuine structural basin is not certain. The bottom of the eastern half of the lake slopes as a continuation of a 3-mile-wide belt of abandoned sediment on that side, to a depth of only 50 feet near the peninsula. But a deep trough of 700 feet of water with steep ledge shores forms its western half and appears to be a continuation of a narrow gap in the mountains to the south. The inclosing mountains are of granite and highly tilted shales and crystalline limestones, while the peninsula is of granite and vertical slate. We are dealing with the core of an ancient mass. Kara Kul is a lake of bitter salt water. Its sloping shores are white with salt accumulated into low ridges, where the brine from each wave wetting has dried out after the recession. And behind some of these there are lagoons of Vo x4 Fig. 432.—A Granite Ridge at Kara Kul (showing the secular Deflation of the Pamir). brine, collected from the overflow of large waves, thus extending the white salt belt 100 feet or more on shore. During summer there are ducks and water-fowl that feed on the wide-bladed slimy grass growing in shallow water. On this high desert no man lives, and those who cross by caravan have diffi- culty in finding fodder and water, as but little grass is found below the water-. courses on high moraines, and even larger streams are dry by day. It is 10 o’clock at night ere the glacial water melted by day has accumulated and reached the steppes to run off before sunrise. A few small areas of thin, scattered wire-grass are found in shallow depressions near the lake, and a scrubby desert weed with long roots serves for fuel. Otherwise, the plain is void of life in summer. During winter large herds of Ovis poli, the great-horned wild sheep for which Kara Kul is famous, descend from their snow-bound mountains, to hunt for these rare bits of grass. By summer they live with the rabbits and marmots high up under DESERTS. 253 the snow-line. I have counted over a hundred in sight at once, ten of them big old rams standing apart, and the rest all ewes and younger ones. Save for the Kirghiz hunters who appear at rare intervals to stalk them with medieval fuse- guns, they live unmolested. Nowhere is there a more desolate land. It is a desert of unexpected forms, time-crumbled mountains and wind-worn cliffs, strange hollow and pitted bowlders, and sand-polished stones, efflorescent salt-plains and drifting dunes, with here and there the scattered remnants of an old bleached skeleton with sun-cracked horns. Limestone bowlders dropped on the plain by floating ice, when the lake stood higher and glaciers came far down, have cracked in the sun and crumbled to conical piles, while whole mountains of the same rock stand shrouded in their own remains. Perhaps the most remarkable example of desert disintegration is found in the granite mountains ranging on the east. ‘There whole mountains are fast crumbling to arkose and sand from which some few honeycombed slabs project as remnant wind-worn ridges. Such are the features wrought by an arid sun and shade, with a range of 80° F. from day to night; the records of diurnal change revolving through long time. Fig. 433.—Sand and Arkose Residuum of Deflation (Kara Kul). And what has become of all the fine stuff, the dust inevitably given off in such a colossal crumbling of the land? It is nearly absent from the surface, as, indeed, it could not long remain on barren, wind-swept ground. ‘The few inches of loess found here and there below the ice and in tiny patches of grass along its streams can not account for the dust of ages. It must be somewhere, and, if not here, we must conclude that it was ever blown away by the storms that come and go, blown away to settle in the grass of other, less arid, regions. Around the lake we find evidence of its former wide expansions in beaches, respectively 60, 120, and 200 feet above the present surface, and apparently one at 320 feet nearly obliterated. The lower ones are comparatively fresh and indi- cate but short existence at their levels. These expansions seem to correspond in relative magnitude and antiquity to the moraines that now lie in front of various surrounding valleys, and which are clearly divided into at least three glacial epochs, and a fourth much more recent advance. 254 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. My report of 1903 has set forth many reasons for correlating the expansions of Lake Kara Kul with its glacial epochs. Avoiding a repetition of detailed explanations, the trend of events may be summarized as follows: During Pliocene time the Pamir was a region of high mountains which by early Quaternary time had been eroded to the core with a topography of worn-down, rather gentle slopes and wide valleys. ‘Then took place that vast uplift which throughout most of Central Asia’s mountains seems to have been the first event of Quaternary time, as its completion marks the advent of the glacial period. Streams cut down, deep-gashing the old topography; the Markan Su evidently developed its gorge back where Kara Kul is now. They had apparently graded and begun to widen their valleys, when everything on the high Pamir was interrupted by the glacial period. Elsewhere dissection has nearly obliterated the old topography, but on the northern Pamir it never got beyond the early stage of a system of gorges perhaps from 1,000 to 3,000 feet deep, with wide intervening areas left intact. It is quite possible that, as Professor Davis suggests, the first epoch expansions of which we find such immense moraines were ushered in by a series of one, two, three, or more epochs of increasing magnitude, though all smaller than that great one by which they might have been obliterated. However complex may have been the transition to this maximum expansion, it seems unlikely that we shall ever know of them in this region, for even its moraines have lost all trace of their topography and_are recognized only by their structure of mixed-up till with huge sub-angular bowlders and occasional striated fragments. ji During this great ice-epoch, which for our purposes may be named the first, the mountains around Kara Kul and the Trans-Alai, and I suppose all high areas of Pamir, were wholly mantled with ice comparable to small continental ice-caps of whose marginal moraines there still remain masses over 1,000 feet thick. When at length these widespread glaciers withered, deep gorges were left choked with “till” and the Northern Pamir thus isolated into basins and blocked around into a zone of held-up detritus. And though we may suppose that in succeeding glacial epochs some detritus may have escaped even from Kara Kul, the aridity of interglacial times, if at all comparable to the present, could not allow of trans- portation from there by water. The Northern Pamir is thus characterized by a persistence of old topography. In its colossal isolation from moisture-bearing storms and with its glacier-made obstructions to stream erosion, it has stood in shape scarce altered through a period of geologic time; it has defied change while lands all around have suffered fast development of gorges, fast erosion of the old (first-cycle) topography which now remains elsewhere only on mountain tops and high spurs flanking them. Only the wind can succeed in getting much of anything out of the region. Otherwise, no débris can have been transported far since the first glacial epoch, excepting that shifted a few miles by glaciers and increasing the obstruction to subsequent erosion. Even part of the Alai valley and the first 20 miles of the Markan Su are no deeper than during the first ice-epoch. ‘Thus one of the world’s highest mountain regions was given long ago a shape so nearly dead to change that it DESERTS. 255 has stood from early glacial time till now with one fixed expression of rock features, sun-crumbled and wind-worn. It seems possible that when the ice of that first expansion had so far melted that the moraine-blocked depression it left, where Kara Kul is now, was free of glacier ice, the lake thus created may have risen to the older terraces 320 feet above its present surface. But they are so much rounded off and altered by erosion that it would not be safe to say definitely that the lake ever stood at that level. ) A following interglacial epoch was of such long duration that the first-epoch moraine had lost its old topography, was gashed by wide valleys, and in the lake depression lay buried by lake clays when glaciers of the next expansion pushed down upon them in places, where piled-up and distorted layers may still be seen in front of overriding moraine. end Of Sia Cie. sere ea Ss com, @>, 300 yards to half WY ROCoeR & SSD burfed moraine Present stream iw. Cl] Z, LY = Seo a —Glaciers1ce- See ene ie Ps Bluff - section A am A and B= glacier ice renerecesects®, Bed-rock Firstepoch Secondandthird Fourthepoch Glacier ice mmr Bluff shores upto 20feet high moraine epochs moraines moraine Fig. 434.—Glacier of the Kara-at and a Section of its Valley 1.5 miles Fig. 435.—A Pool where Kara Kul Sediments have caved in below (Kara Kul). over Melting Lobes of Glacier Ice. The general limits of ice in the second epoch are sufficiently traceable to indicate a remarkable difference of outline between it and the first. It may be doubted if much, if any, of the region of Kara Kul was free from ice during the first epoch’s maximum expansion. A snow-line 3,000 feet lower than its present would cover all the Northern Pamir with ice. During the second epoch mountains around Kara Kul were mantled with ice apparently reaching about 1,000 feet lower than their present ice-dome margins; and from this extended valley-flows through gorges cut in massive moraine heaped along the mountain flanks during the first epoch. Reaching the plain they spread out as coalescing fans, which in some way or other became buried by lake clays. There appears to be no doubt of the glacial origin of these lake clays, as they are characteristic of glacial grinding, light gray or non-oxidized, with flakes of mica and layers containing small angular fragments. ‘They are finely laminated and cover a wide area east of the lake. A large proportion lie horizontal, as 256 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. shown in sections of deflation, but there are significant fan-shaped areas of much- disturbed stratification spreading out from the main valleys. Within these, the surface is broken into irregularly distributed mounds where wind-carving has exposed an arched structure, conformable to the original surface. Among them are miniature lakes of clear, fresh water, up to a few hundred yards in width and 10 to 20 feet below the plain, their walls extending vertically into the water. All these facts, together with the utter lack of external hydrographic relations, led Professor Pumpelly to suggest their origin to be a caving-in of sediments on to underlying lobes of melting ice. And that seemed a logical explanation of the whole disturbance. It was corroborated in 1904 by my discovery of actual exposures of that ice, sections of ancient buried lobes in the bluffs around some of these pools; characteristic glacier ice lying beneath 5 to ro feet of lake sediments Fig. 436.—An Ice Dome, and its Third and Fourth Epoch Moraines in the foreground (Kara Kul). where exposed. At first it may seem too extraordinary that the ice of an ancient glacial epoch should exist to-day, but, when realizing that the temperature over © these steppes falls to 10° and 12° F. at night during the warmest part of the sum- mer, it appears more natural. We shall have to attribute it to the second of our glacial epochs, having found it towards the limits of that expansion and 7 miles from the end of the glacier to-day. So in some way or other Kara Kul rose to drown the piedmont sheets of ice-and bury them with its sediments of glacier-ground stuff. The corresponding shore-lines are on the peninsula well-preserved beaches of wave-action 200 feet above the present level. The open water for wave-action and sedimentation of glacier-ground stuff, together with vegetable life—grass like the present—all this during the second glacial epoch is of great significance. We must believe it DESERTS. 257 was a period of heavy precipitation, but no colder than now, Otherwise, no lake could have existed, for even now it is frozen nine months out of twelve. We have to face a climate as warm, if not warmer than now, and ascribe the greater ice accumulation to more snowfall. Indeed, it appears unlikely that any moisture to speak of could reach a region so isolated by range beyond range of the world’s highest mountains, if the climate were much colder. On the other hand, there would probably be more precipitation now if it were warmer, and it would have to be a great deal warmer to raise the effect of melting seriously. An ordinary temperature of 10° F. in midsummer nights on the lake-shore means a very low average for higher portions. It is sometimes difficult to separate moraines of the third epoch from those of the second, whose higher portions they overlie. Those on the east barely reached the plain and did not quite fill the second-epoch gorges in the first-epoch moraine, but must have lasted nearly as long as the second, for their moraines Fig. 437.—A Glacier Northeast of Kara Kul. reach a great thickness. During this epoch the lake appears to have stood at the 120-foot level, and these beaches are worn fully as long as those of the second epoch. ‘The third epoch, therefore, appears to have lasted for the same order of time as the second, though its glaciers were only half as long. In fig. 436, the fourth-epoch moraine is seen as a mass of loosely piled angular blocks overlying the third-epoch moraine of which the smooth surface forms the base of the picture. Long after these three had come to a close, there came a fourth advance to about one-half the distance of the third, but of far less than half the same duration, for its moraines are insignificant when compared to them. And this one came to a close so recently that its moraines have suffered scarcely any surface weathering, whereas those of preceding epochs have been much worn by wind; even the third one can be ridden over, while granite moraines of the first epoch have been smooth- planed into mosaic floors, the interspaces filled with the arkose residuum of long 258 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. time. One can ride all day over the old ones, but no horse dares set foot on the loose blocks of the recent fourth-epoch moraines. ‘The very fresh beaches from 60 feet down to the present lake level might be coupled with this last advance. All the Kara Kul glaciers of to-day lie in part on moraine, over which they have advanced from some more contracted stage, and they all terminate with a vertical or overhanging ice-cliff, from which large cakes of ice break off from time to time. I am inclined to think that their overhanging termini indicate that they are at present advancing. Most of them are mere valley tappings or tongue- shaped flows from high ice-fields, smooth-mantling the old worn-down granite mountains. Nowhere has nature been more graceful than in the making of these ice-fields, with high domes and troughs, and winding flows, all glare and clean and coalescing into the meandering striae and cross-ribbing of more rapid motion where - they are drained by valley glaciers. In this region there is no true snow-field, such as in the Alps. At no time did I find more than the 1 or 2 inches of snow of some recent fall, and under that was glare ice. Climbing on the domes is for the most part from step to step, while, with gun and camera, one balances and chops. When considering the power of solar rays felt directly through an arid atmosphere of 16,000 to 20,000 feet elevation, we understand why snow changes so rapidly that in a few days after storms all is transformed to a crust of ice that coalesces with the underlying dome. Another remarkable feature of these mantles is the utter lack of surface débris over their clean, white expanse. This must be attributed to the gentle slopes of their underlying topography, none of which can rise above to shed débris. We shall have more to say about that topography. ‘Their expansions during the older epochs were, as stated, somewhat similar to that of a continental ice-sheet and resulted in widespread accumulations of moraine now characterizing the slopes of all high mountains around the basin. Obviously no reconstruction of past events in the Northern Pamir can be undertaken without data from without its zone of held-up detritus; for, to decipher its orogeny, we must study its surroundings of erosive activity, where terraces of rejuvenated valleys fall under distinct erosion cycles -belonging to periodic uplifts, and, where these valleys debouch, look for marginal deformations of the flanking piedmont zones. Everywhere beyond the borders of the Pamir’s rela- tively dead topography deep gorges and canyons, torrential valley systems, and sharp land forms are met with. Already the Markan Su has cut through the last tributary moraines that blocked its course, and below them changes into a corrading stream. The Pamir’s old preglacial topography is only 500 feet above the stream at the junction of its headwaters of Kizil Kul, but diverges rapidly with the stream to several thousand feet above only 60 miles eastward, where only the flat backs of mountains and massive spursof soft red strata flanking the Trans- Alai remain of the ancient continuities of gentle slopes and inflected surface. This is in general the state of valley systems developed into border ranges of the Pamirs. My studies of several of these valleys (Kizil Su west, Kizil-Art Darya, Markan Su, Kizil Su east, Taldic Darya, and Zerafshan) follow under the headings Alai Valley, Karategin and Hissar, Zerafshan, Tarim Basin, and Fergana Basin. DESERTS. 259 In order to close our section on the Pamir with a tentative reconstruction of its Quaternary history, we may, in prevision, assume a subdivision into erosion cycles justified by these valleys. TENTATIVE RECONSTRUCTION OF QUATERNARY SEQUENCE OF EVENTS. First cycle (Pliocene). Pliocene mountains with the present Aralo-Caspian and Gobi basins defined in a general way. Erosion to low relief of Central Asia’s peneplain and piedmont stage. Remnant ridges still rising out of Pamir’s worn-down topography of rolling uplands. Second cycle (Quaternary). High uplift of mountains. Deep gashing of old Pliocene topography. Expansion of first epoch of glacial period with alluviation of valleys. Recession of first epoch of glacial period, leaving transportation in the Northern Pamir blocked around by moraines and its valleys converted into lake basins. Third cycle (Quaternary). Uplift (block-tilting of Alai). Narrower gorges incised in second-cycle flood-plains of lower valleys, but Pamir still isolated with second-cycle topography intact. Expansion of ‘second epoch of glacial period; lake Kara Kul filled to about 200 feet above present shores; alluviation of valleys of border ranges. Recession of second epoch, shrinkage of lakes, but plains aggrade back into valleys bearing old moraines. Fourth cycle (Postglacial). Uplift, warping, etc. Narrower gorges incised in third-cycle flood-plains of border ranges, but Pamir’s second-cycle topography intact. THE ALAI VALLEY AS A BASIN. A DISTINCT TYPE OF VALLEY. In passing north, east, or west off the high central mass of Pamir, we encounter a remarkable type of valley—a wide basin-like trough contained, often nearly inclosed, by longitudinal ranges and floored with a steppe of gently sweeping con- cavity. This is a type well exampled by the Alai valley and Keyak Bashi and common throughout the highlands of Asia, giving rise to their most fertile pastures. On a large scale it resembles the basins of Pamir, but differs from them because of lower elevation. The origin of these valleys may lie far back in Pliocene time, as a result of migration of the zone of piedmont upheavals, periodic encroachments of mountains on plains through successive upheavals parallel to the primal high- lands. As their streams average about 10,000 feet in elevation, they are within the zone of held-up detritus and have the aspect of deeply filled troughs. The Alai valley—the high eastern catch-basin of the Kizil 5u (the first great branch of the Oxus)—is a long, broad, east-west depression between the Pamir’s northern border ranges or Alai and Trans-Alai Mountains. As a basin, it is 110 miles long from the Taun Murun divide to its canyon outlet at Katta Kara Muk and attains a width of 30 miles from crest to crest of its inclosing ranges. The 260 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. even steppe resulting from its deeply filled state has a width of 15 miles for a dis- tance of 75 miles. Less isolated from moisture than the higher Pamir, it receives sufficient precipitation to form one of the most luxuriant pastures of Asia. In the higher Alai valley there are but two seasons—winter and spring. For nearly ten months of the year it lies deep-buried in snow, a vast expanse of white from range to range. ‘Then no man lives upon the plain and its gray wolf-packs are free to hunt the wild sheep and wandering ibex. By July the snow has melted and like magic the grass turns green; a myriad of marmots leave their holes to visit one another in the warm sunlight, uttering their shrill notes of warning when the caravans come down. A thousand Kirghiz families descend from the passes round about, with their long camel trains caparisoned and rich-laden with nomadic wealth, and each caravan with its flocks of sheep and goats, herds of camels and cattle and horses, proceeds to its traditional camping-ground. For a while it is all life and merriment in a world of grass and wild flowers, a wonderful valley of green with poppies and buttercups and peopled by men and animals, with here and there a group of round felt-domed kibitkas; a land whereover days of mist give way to skies of blue purity. But through it all the mountains stand colossal and cold, reminders of soon-coming snow, and from their ice-domes, 15,000 to Fig. 438.—Springtime in the Alai Valley. 23,000 feet in height, it creeps fast down upon the grass. Ere spring has finished bloom, winter has come and the valley is left frozen in snowbound emptiness. But, though it is a remarkably rich pasture, there is scarcely rain enough— too little over the western or lower part. It is to the shortness of summer and relatively heavy snow of winter that the richness of its grass steppes must be attributed. There is enough water from melted snow in the ground to keep things fresh with the occasional help of mountain mists; enough in the upper half and all is full green there when winter falls to bury it, but in the lower half, the sun dries out all the water and leaves a plain of parched grass. The Alai valley is thus a semi-arid type of desert basin. Of all the basins to be considered by us, it is unique in that it yields four of the deposition zones, so interwoven and interlapping that alluvium, moraine, and loess are found one over the other, and correlation becomes relatively easy. Kettle-hole pools are found on its widespread moraines, but they do not belong, organically speaking, in the lacustrian division of a desert basin: so that the lacustrian zone is here lacking. But the alluvial, flying sands, loess, and glacial zones are all especially well represented. Its nuclei of dunes, though small, are not very disproportion- DESERTS. 261 ately so, and become of great interest as a demonstration of the origin of such nuclei in greater basins. Its wide gravel-plains sloping from the Trans-Alai are strewn here and there with dunes and other areas of loess, while in the lower half of the valley there are sand nuclei several miles in length. Several feet of loess have accumulated over its old moraines and along the lower slopes of its border ranges. GLACIOLOGY AND EVIDENCES OF MOUNTAIN MOVEMENT. The glaciology was described in my first report. The same oscillations and the same greater advances and recessions took place here as in the Kara Kul basin. During the two first epochs much of the valley steppe was occupied by wide pied- mont ice-flows extending all the way across it from the Trans-Alai. Moraines laid down during the first have lost all topographical characteristics and their worn-down surfaces now form broad, low, transverse undulations of the valley floor and massive foothills of the Trans-Alai. Those laid down during the second Fig. 439.—Kirghiz making Felt in the Alai Valley epoch are now spread in the form of broad lobes covering immense areas and made up of vast numbers of conical mounds. In one remarkable instance, a second- epoch glacier spread all the way across the valley, piling up the slope of its northern side, just west of the Kashka Su. Through this moraine the Kizil Su has cut a channel, exposing a section of till resting on alluvium barely above the level of the stream. This section is of especial importance, for it is clear proof that the Kizil Su flood-plain was at about the same level there during the second glacial epoch as it is to-day, and that terraces, which leave the flood-plain about half- way down the valley and attain a height of 300 feet near its outlet, belong to some earlier age. On a visit to the glacier of the Tokuz Kungei, one of the greater Trans-Alai tributaries to this valley, some ideas were formed about the third and fourth epochs and their relations to the older. This glacier still terminates piedmont fashion, deploying—over a massive accumulation of moraine filling the mouth 262 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. of its mountain valley—as a flat sheet of clear ice, perhaps over half a mile wide and roo feet thick. Its waters discharge into a valley over 1,000 feet deep with an estimated width of three-quarters of a mile. This valley is excavated entirely in moraine composed, for the most part, of huge subangular blocks with glacial strie, and belonging to the first epoch, as it has lost all topographical character- istics, having been graded with a surface sloping towards the middle of the Alai and gashed by this great tributary valley. In this valley of discharge there are two terrace levels bearing kettle-holed moraines, and a less definite third terrace below. These, it would seem, belong respectively to the second, third, and indefi- nite fourth epochs. From out of the valley of discharge its second-epoch moraine spreads clear across the Alai with broad lobes, each composed of conical heaps 100 to 300 feet high. Between the first and second epochs, time enough elapsed for the excavation of this wide valley of discharge to within 200 or 300 feet of its present depth. a Fig. 440.—The Tokuz Kungei Glacier (Alai Valley). As a severe buran (snow blizzard) blew up and darkened the region, I was unable to complete my observations on the third and fourth epochs, but it is a locality worthy of careful exploration. The glaciers of to-day are of a different character from those about Kara Kul, To begin with, the snow-line on this side of the Trans-Alai is only about 13,000 feet, and we find true snow-fields—that feature for which the mountains of Kara Kul are so remarkably in want. There the snow turns to ice as fast as it comes down; but on this side there is more of it and a less arid sun to transform it. “Two of the present glaciers studied might almost be called piedmont flows. They descend with a steep grade and spread out fan-shaped upon massive accumulations of moraine, in part at least belonging to the third epoch (a mass that must be from 1,000 to 2,000 feet in thickness), and spread upon the flat floors they have planed over these old moraines, to terminate in wide sheets of pinnacled ice, possibly less than 100 feet in thickness. And these glaciers, like those of Kara Kul, are DESERTS. 263 remarkably free from surface moraine; and for the same reason, viz, that the mountains above are almost wholly mantled with ice. We must believe that the old epoch moraines, in spite of their immensity, were accumulated and brought forward under the ice and by its margins. In my report of 1903 it was shown that considerable mountain movement took place apparently between the first and second glacial epochs. The northern flank of the Trans-Alai is truncated by what appears to be a fault-scarp, displacing the first-epoch moraines and broad-trough valleys and dissected by the narrower trough valleys of later glaciation. The Kizil Art valley, principal Trans-Alai tributary, has the twice-troughed form due to uplift, from its heading cirque nearly to its mouth in the Alai; there the bottom of its lower trough sinks under the flood-plain of later alluviation, half-drowning the great second-epoch moraine that rises at the mouth of the valley to stretch 10 miles out into the Alai. —— / Fig. 441.—Alluvial Terraces in the Lower End of the Alai Valley. Portions of its second-epoch moraine lie apparently faulted up on the terraces of its upper, broad-trough stage. Half-way down the Alai there is an uptilted mass of alluvium jutting out from its northern side as though more complicated deformation took place. In the eastern half of the valley tributaries from the north are deeply drowned in alluvium, as contrasted with those from the south, uplifted Trans-Alai, which have been canyoned postglacially. Terraces begin to flank the Kizil Su toward the outlet of the valley; and of these the higher are, as shown above, the result of cutting down since the first glacial epoch, while the lower appear to have been since the second. By early explorers these were taken for shores of an ancient lake; but they incline gently downstream, diverging some- what above it, ascend into tributary valleys from the north as characteristic alluvial terraces, and in no way resemble those of a lake. It is, therefore, obvious that the Alai valley has suffered two distinct periods of mountain movement— 264 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. one between the first and second glacial epochs, another after the second glacial epoch. ‘The first movement appears to have given a relative uplift to the southern range, the second to have been an uplift of the whole valley, raising its western end at least 200 feet, probably several times that amount, and causing its outlet to cut down a narrow canyon. The Kizil Su is still corrading there rapidly on rock bottom. The first movement, it was shown, might be explained as a transverse block- tilt of the valley plain with the Alai Mountains north, on a dislocation along the Trans-Alai scarp; and one north of the Alai Mountains tilting the Alai valley down while the mountains north of the rotation axis were raised (see plate 63). This might have resulted because of the heavy load of moraines and alluvium piled in the Alai valley during the first epoch. Then, if the whole block was set in motion, it necessarily sank down on the Alai valley side, while its northern side was raised by displacement from the sinking Fergana plains. If the Alai valley had been larger, it probably would have gone through an independent set of movements; but, being a basin of small dimensions, it appears to have acted as a dead weight, an increasing load against surrounding mountain movements. There appears to have been some deviation from this general geodynamic scheme. Midway between the two ends of the valley and by the northern edge of its steppe the Kizil Su is truncating a low spur composed of displaced Alai valley alluvial gravels with a slight dip diagonally to the valley axis. This local departure merely indicates that though the Alai Range probably moved as a whole it was at the same time (or perhaps afterwards) more or less broken up. TENTATIVE RECONSTRUCTION OF EVENTS IN THE ALAI VALLEY. First cycle (Pliocene). Uncertain as to whether the valley was defined. Second cycle (Quaternary). Uplift of Pamir with differential block uplift of border ranges. Alai valley defined. Expansion of first-epoch glaciers, and piedmont ice-flows from the Trans-Alai reach all the way across the valley. Third cycle (Quaternary). Transverse tilting-down of the Alai Mountains and Alai valley with dislocation along the northern base of the Trans-Alai. Expansion of second-epoch glaciers and deepening of glacial valleys with troughs narrower than those of the first. Fourth cycle (Postglacial). Uplift of whole region. Cutting-down in the lower half of the valley. Loess and sand-dunes now accumulating. Glaciers have recently (probably during the last few hundred or a thousand years) receded between 200 and 300 feet. DESERTS. 265 KARATEGIN AND HISSAR. The highlands of northeastern Bokhara are portioned between Karategin and Hissar, two remote provinces still surviving as feudal tributaries of the ancient khanate. Ethnographically, this is a region of high valley oases similar to those of the Zerafshan, remarkably isolated, and preservative of a distinct type of man, the Galcha, still speaking a relatively pure Aryan dialect—rare survivals of a primitive sedentary stock, elsewhere diluted or exterminated by the nomads. Physiographically, it is an important portion of the Oxus drainage system, which, in Quaternary time alone, has enacted a series of both hydrographic and topo- graphic changes of astounding magnitude. a Fig. 442.—The Kizil Su where it leaves the Alai Valley. KARATEGIN AS SHAPED BY THE KIZIL SU. The Kizil Su now follows a course of three physiographic divisions: (1) The Alai valley, (2) the valley of Karategin (Katta Kara Muk to Obu-garm), (3) the Vaksh valley (Obu-garm to the Oxus). We have studied the Alai valley as a basin, and as a whole it is a basin from which but a small proportion of the materials of its Quaternary erosion can have escaped; but in the western half there begin alluvial terraces flanking the Kizil Su flood-plain and gaining height toward the outlet. At a point 12 miles east of Katta Kara Muk, the old floor narrows rather abruptly from a width of 8 miles to about 1 mile and the valley becomes of normal aspect. And at Katta Kara Muk, where the true outlet canyon begins, there are three terrace-levels, respectively 70, 100, 200, and 300 feet above stream and continuing up the tributaries. There can be no question as to their 266 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND_OASES. alluvial origin, as they slope with the stream at a rate of over 30 feet per mile and continue up its tributaries, and on the Trans-Alai side are composed of trun- cated fans sloping from high up in the mountains. When the river flowed at its highest or 300-foot terrace-level (first glacial and part of first interglacial epoch), its wide valley-floor was continuous into Karategin. The 200-foot terraces are also continuous, forming a narrower stage, while the present outlet gorge, into which the river plunges in a torrent state, is carved in the 200-foot floor. This narrow gorge is cut in the strike of a south- easterly dipping series (hard limestone on the north, soft gypsiferous beds on the south) and continues for 15 miles, whereupon the valley opens out and its terraces widen into the broad grade-plains of Karategin. Fig. 443.—A Galcha Beg of Karategin with his Hunting Eagle. The Kizil Su valley of Karategin is, in a general way, 25 miles wide and over a mile deep, as measured from the lower passes of its border ranges—a valley on the large scale characteristic of the highlands of Asia. Flanking it on the north is a branch of the Alai, on the south the Peter-the-Great Range, whose giant ice-clad peaks stand 10,000 feet above the oases scattered along the terraces and grade plains of its ancient valley-floors. The southwestern half lies open to a scanty precipitation of moisture from storms born up the Oxus embayment to nourish a moderate pasture over the grade-plains and flanking mountains below their snow. And there even grain is raised in fields patching the high slopes, with no irrigation; but the eastern half of the valley is barren of vegetation, a desolate land of sharp red and gray or black declivities rising from the dazzling gravel-plain of its ‘“braided’’ stream. There the higher peaks of the Peter-the- Great Range are less often seen mantled with clouds than as naked pyramids of white outlined against an arid blue. Karategin is essentially desert. DESERTS. 267 Physiographically, the great features of this spacious valley arise from its ancient terraces and grade-plains. In general, there are three past erosion stages represented and these with the present river channel make four erosion cycles. But for over 30 miles, midway between Obu-garm and Katta Kara Muk, the third and fourth stages merge into one; or rather the third-cycle flood-plain warps under present alluviation and the river spreads out with no channel. Often the second-stage terrace is found obliterated and the valley becomes a simple wide gorge with tributaries cut into the old topography of its first stage, the pene- plain stage of these regions. gee ee SO ee Ist cycle Grade plains flanking Limestone the Peter The Great mts. SECTION THREE MILES BELOW THE ALAI VALLEY. 29 cycle terrace J | Pe and 4th cycles (Present flood plain) ase Yi} FIVE MILE SECTION NEAR HAUI. 1st cycle 1St cycle Y Y > 2nd cycle //j 5 4theycle x= TYPE FIVE MILE SECTION WHERE THIRD CYCLE FLOOD-PLAIN IS NOT WARPED UNDER THE PRESENT FLOOD-PLAIN-~ Fig. 444.—Terraces of the Kizil Su in Karategin. As the first-stage topography has been uplifted and dissected by this great gorge and all its tributary systems to a depth of from 3,000 to 4,000 feet, only limited areas of its original slopes now remain. It is to be inferred from a con- formity of grade-plains and flat-topped spurs flanking the Peter-the-Great Range and dissected remnants of a one-time half-peneplained belt of the Alai Mountains north. When uplift began, it was by no means a peneplain, but rather a mature topography grading into the wide, shallow valley of the Kizil Su and peaked here 268 PHYSIOGRAPHY OF CENTRAL-ASIAN- DESERTS AND OASES. and there with low monadnocks and bordered by the Alai and Peter-the-Great Range. But to attain even that stage of maturity in the heart of a great moun- tain region; to erode what in Pliocene time must have been a region of colossal ranges to its metamorphic cores, must have taken vastly longer than all three suc- ceeding cycles of erosion taken together. The sum total of these later cycles has resulted in no more than an immature dissection of the ancient topography, and, though for our purpose they should be termed erosion cycles, they are by no means comparable to that which closed the Pliocene and should be regarded as mere phases of a Quaternary striving towards base-level. We shall find corre- sponding phases of Quaternary erosion over other regions and term them cycles for the sake of a tentative correlation. Fig. 445.—A Bridge over the Third-cycle Terrace in Karategin. Terraces of the second stage lie about half-way up the valley sides, but are found only at rare intervals, usually where tributaries join, and badly preserved because the third-stage valley-floor has widened to nearly obliterate the transition. But those of the third stage are in remarkable preservation and form the great feature of Karategin, the spacious plains and gentle slopes of its oases. Broadly speaking, it gives a concave sweep to the valley bottom, for the most part 4 miles across and traversed by the present river channel, about half a mile wide, of rec- tangular section, varying up to nearly 300 feet in depth. At Damburachi (junction with the Muk Su), this stage widens into a trian- gular junction-terrace* of over 30 square miles area, traversed by abandoned distributary channels of the Muk Su between 200 and 300 feet above stream. Two *A terrace in the angle or junction-spur where two rivers join. I venture to offer ‘‘junction-terrace”’ and ‘‘junction-spur” as terms I have found essential. DESERTS. 269 or 3 miles below Damburachi the river flows through a narrow cut, skirting the north side of the valley, while the lower terrace rises as a plateau on the south with an old channel of the Kizil Su on it. Below there it converges with the present flood-plain and just below Pildona, 6 miles above Hawee, it sinks under. From there on nearly to Garm, it lies buried in the present flood-plain and the river at high water swings against the valley sides it scoured during the third erosion cycle. Just above Garm the lower terrace rises out again and the river enters a channel that deepens steadily to Obu-garm, the westernmost point of observation, where this terrace is about 300 feet high as cut into by the present channel, that of our fourth erosion cycle. In defining the third stage we inevitably defined the fourth, the channel of to-day. But some important facts should yet be noted. The present channel is cut in alluvial gravels all the way, excepting for occasional glimpses of bed-rock bottom. Obviously, then, the stream had cut down as far as it is now during the third erosion cycle, which, however, closed with a re- filling and valley-widening to the third stage. The earth-movements which wrought its fourth and last erosion cycle were of sucha nature that about 35 miles of the valley, that portion between Garmand Pildona, was negatively affected, that is, aggraded instead of corraded. This would appear to indicate warping, and that idea is reinforced by observations on tribu- taries. Two large tributa- ries converge to Hawee, where they debouch in the Fig. 446.—A Galcha of Karategin. Kizil Su. One comes in from the west, the other from the north. If the great valley suffered a longi- tudinal warping, we should expect to find the one from the west rejuvenated, the one from the north aggrading in sympathy with the main river. This is precisely the condition at Hawee. One more important fact remains: the Kizil Su has not yet graded its fourth-cycle channel, excepting over the 35-mile stretch of relative depression. The present flood-plain is divided into several long stretches where the river splits into a braided stream. Between these the channel narrows into short shoots, where it plunges over sills of bed-rock. The fourth erosion cycle is, therefore, the result of a warped uplift still in process, or so recently completed that corrasion has not yet attained an even grade. 270 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. GREAT FEATURES OF THE HISSAR VALLEY. Hissar becomes of interest because of the extraordinary hydrography of its great valley. Opposite ancient Bactra a 15-mile wide strip of steppe sweeps up from the Oxus embayment into this valley, continuing northward up the Surkhan River as far as Karatagh, then bending due east into the wide open valley of Hissar. This portion of the valley-floor averages 2,500 feet in elevation, and is distinguished for its utter lack of a trunk-stream. It is, on the other hand, crossed by three tributaries, the Kanaka, Dushambeh, and Kafirnigan, converging to near the city of Hissar, where they break through the southern side and flow to the Oxus. The valley has a mixed population, divided between Usbeg camps and Tadjik villages. Throughout the old khanate it is famed for its wealth of pasture and the grace of its horses. The streams descending from the mountains north are diverted to irrigate a wide continuity of rice and grain-fields, while the silk woven in Karatagh and Hissar is prized throughout the cities of Central Asia. Ten days were spent in attempting to decipher the remarkable physiography of the Hissar valley. The more open part of the valley east from Karatagh is about 40 miles long and floored by a grass plain with an average width of 5 miles, but of irregular definition. On the northern side this plain is often bounded by a loess cliff, surmounted by a narrow belt of steppe rising north. From this it inclines transversely or southwards on a grade of about 20 feet to the mile, and sweeps up again to meet the mature topog- raphy of the southern side. The loess cliff varies up to 100 feet in height, and running along the northern side of the main valley truncates tributary spurs and tributary valleys in one plain; but is interrupted by the broad flood-plains of larger tributary valleys dissecting it; in general, it runs east and west, sometimes perfectly straight for several miles; sometimes gives way to a dissected deformation of the plain, and has the appearance of a recent fault-scarp. From near Dushambeh it runs straight east for 8 miles, dissected by local streams at rare intervals. The Hissar valley is all loess, except where crossed by the gravel flood-plains of the three larger streams. These streams cross in wide channels, beginning with a depth of about roo feet, decreasing till near the southern side, where flood-plains Fig. 447.—A Swimmer of Rapids, with an Inflated Goat-skin (Karategin). DESERTS. a7 merge into the great valley-floor. It is essentially a valley of deep loess accumu- lation, a grassy settling ground of dust blown up from the Kizil Kum. On the other hand, its great dust-beds undoubtedly alternate in depth with the flood- plain deposits of its cross-streams. Out of ten days, during seven there was a Fig. 449.—Silk Factory in Karategin. yellow haze, and of these days two were so gloomy that a few hundred yards was the range of vision. It is, therefore, a region of still living loess. The kurgans and burial mounds left by its more ancient inhabitants are now mantled with 2 feet or more of loess. Diya PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. These questions arise: What is the explanation of this valley, of which a part has no trunk-stream? Was it ever a true river-valley? Is it the lower portion of a valley of which the head-waters have been captured by a higher branch of the Oxus? Was it possibly the valley of the Kizil Su, now captured at Obu-garm by the Vaksh River? East of the open plains the valley-floor has been uplifted and dissected by the Hyak Darya, which plays the part of a true trunk-stream for that portion. Following it up to about 18 miles above Faizabad, we find the channel made by this stream, at first deep with high terraces, decreases in depth and finally opens out onto the ancient valley-floor again. Here the floor is a sweeping concave about 3 miles wide between the mountain sides, with spacious tributary fans and talus cones, and there are no terraces. It stands at an elevation of 6,100 feet and forms a rich summer pasture, resorted to by Usbeg nomads. The remarkable fact about this floor is that it continues rising east beyond the heading tributaries of its stream, and we thus confront another portion of the valley with no trunk-stream. For 3 miles it is a wide grass plain with not even a tributary descending to its borders. Then, while it is still rising east, there begin Ss Fault scarp Fig. 450.—Cross-profile of the Hissar Valley. gully systems developed back into it from a stream which joins the Kizil Su near Obu-garm. These gullies join into a gorge deepening rapidly with terraces inclin- ing east, while the old floor above apparently still rises east. The gorge finally develops into a canyon 2,000 feet deep in hard limestone and softer rocks; then widens again at Obu-garm and debouches into the Kizil Su with a total depth of 3,000 feet. The high-grade plains or first-stage terraces of the Kizil Su appear to have a height of about 3,000 feet at this point, and, if so, conform with the old uplifted floor of the Hissar valley. Here there is no doubt that the gorge of Obu-garm has captured the head of the Hissar valley. The valley at the divide is so broad that it must have been formed by a large stream. It is possible that the Kizil Su was that stream and that the mountain movements that first broke up the mature topography wrought this change in its course. In connection with the valley of Hissar it is important to know the valley form of its northern tributaries. One of these, the Sardai-miona, was briefly studied in descending from Kak Pass over the Hissar Mountains. For 40 miles it is a gorge over 2,000 feet deep, of which the last 25 miles narrows to a granite DESERTS. 278 canyon with cliffs rising often 1,000 feet sheer and sometimes 500 feet overhanging. Its population in the more open portion, the first 15 miles, is divided between shepherds and villagers. There the valley sides rise 2,000 feet at an angle of 25° to 30°, but are luxuriantly clothed with grass. On terraces along the bottom are its villages of square huts with flat mud roofs and cobblestone walls cemented with clay. These are usually seen incased with cakes of dung fuel, piled carefully up against the walls or heaped in cones on the roof to dry. The granite gorge beginning about 15 miles below, though for the most part a canyon with sheer walls, widens out occasionally, especially where tributary valleys come in. Here we find villages of sloping thatched roofs, recalling those of Europe, and a variety of trees. The mountains above the granite walls are sparsely forested with scrubby cedars. Willow, sugar-maple, yellow birch, hickory, poplar, cedar, black elm, wild cherry, and wild plum grow along the bottom. It is an extraordinary fact that all excepting the cedar had a familiar aspect, Fig. 451.—Kak Pass over the Hissar Mountains. recalling similar varieties in America; and even more surprising was it to find true maple-sugar, made into flat cakes like ours. Their sugar-maples and yellow birches would pass for the rock-maples and yellow birches of New England. In its upper portion the Sardai-miona is a terraced valley. Terraces diverge from the flood-plain near its head and increase in height downstream till the granite canyon begins, and there they vanish to reappear about 8 miles below Romette, where the valley opens out again. There we find them of loess and loess-capped alluvium in bluffs rising over 100 feet above stream. Below Romette the mountains are mantled with loess and patched here and there with grain-fields, some of them on slopes too steep for a horse to plow, and the land has to be hoed. The crops are brought down on wooden sleds that slide easily down these steep slopes of smooth loess. All the hillsides are scarred by tracks running straight down, as the same ones are used year after year. On the plains of Hissar similar sleds are used, but drawn by oxen instead of descending by gravity. 274 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. The Sardai-miona valley has no high terraces, no terraces comparable to those of Kizil Su. It is essentially a simple gorge, beginning with a V shape with low terraces at the bottom, then changing to a canyon and finally opening again to a V shape with low terraces at the bottom. These low terraces which mark its last transition undoubtedly resulted from the recent uplift which faulted up the northern side of the Hissar valley, breaking off the border of its floor with the scarp of loess. The river, which is of nearly clear water, as its region is of crys- talline and metamorphic schists, is a rapid and copious flow, but hardly a torrent. Its grade is fairly even, though at intervals it becomes more rapid and passes chutes. It may, therefore, be said that its last phase is still uncompleted. Any explanation of the Hissar valley must explain why the Kafirnigan with the other streams breaks through its southern side, instead of flowing east to join the Karatagh, thus continuing the valley to its proper outlet. It may be that the Hissar valley is not a valley, but no more than a belt of the Oxus embayment, shut off by uplift of the plain south, and that the Kafirnigan cut down as the mass rose. This is not likely, because the hills south of Hissar are by no means sharp in outline, but are, on the contrary, of mature form. It would seem that a regional tilt, raising the northern side with the Hissar Mountains, would account for the canyons of tributaries there and the cutting back of an Oxus tributary to capture the cross-streams at Hissar. The same process might account for cutting back of the Vaksh to capture the Kizil Su at Obu-garm, thus leaving the Hissar valley from there west without the great stream that belonged to it. TENTATIVE RECONSTRUCTION OF EROSION CYCLES IN KARATEGIN AND HISSAR. First cycle (Pliocene). Erosion of Pliocene mountains to low relief, while Kizil Su flows straight through the Alai valley, Karategin, and the valley of Hissar, and then south to the Oxus. Cycle closes with valley widening and a many-mile wide flood-plain continuous from the eastern end of the Alai valley to the Oxus embayment, a distance of nearly 500 miles. Second cycle (Quaternary). -High uplift; Hissar region tilted, raising north more than south. The Kizil Su cap- tured at Obu-garm, the valley of Hissar at Hissar; the Kizil Su deepens the western end of the Alai valley and canyons its valley in Karategin; the Sardai-miona develops its canyon cycle. Closes with the valley widening. Third cycle (Quaternary). High uplift. The Kizil Su deepens the western end of the Alai valley and canyons the second-cycle floor of Karategin to a depth of 2,000 to 3,500 feet below the first-cycle grade plains. Cycle closes with prolonged valley widening and glacial alluviation. Great accumulation of loess. Fourth cycle (Postglacial). Warped uplift of a few hundred feet. Faulting up of northern side of Hissar valley. Present channels incised in flood-plains of third-cycle alluviation. Now up to 200 or 300 feet in depth and still cutting down. Loess still accumulating. DESERTS. i) “I on THE ZERAFSHAN VALLEY. On the Northern Pamir and in the Alai valley we found a good field for glaci- ology, and would, off-hand, expect to find record of corresponding climatic change on outlying ranges, nearby members of the Tian Shan. But although no such extreme difference as the variation of from one to six glacial epochs, found by Fig. 452.—Thatched Roofs in the Sardai-miona Gorge. Mr. Huntington, was met with on my journey, there was an unmistakable discord- ance between certain valleys. We hope to show that a differential glacial record was inevitable on mountains subjected to the differential uplift such as we find recorded by various degrees of block-faulting and tilting. With the Alai Moun- tains, we have a region that has ats been uplifted some thousands of i | feet, faulted on the north and bor- | dered there by rows of uptilted | piedmonts. It isa significant fact that Mr. Huntington found a uni- versal correspondence of variations inclimate (by attributing valley ter- races to climatic change) and yet no correspondence at all between vailey glaciers. Wecan not, how- ever, believe that the glaciers of Central Asia were independent of Central Asia’s climatic change. If it were merely a disagreement be- tween valleys of different elevation, between high valleys now occupied and low ones now glacier empty and between empty valleys of different height, the matter might be argued independent of uplift. But such is not the case. Out of twenty-four valleys scarcely any two of the same height agree; and there are instances of valleys near together and of the same height disagreeing several epochs. It will be understood that most = ee Fig. 453.—A Sled in the Hissar Valley. 276 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. of Central Asia’s valley terraces have resulted from widespread cycles of uplift, in some parts locally interrupted, and that just such a variable glaciology would arise from a differential uplift. THE ZERAFSHAN AS A LONGITUDINAL VALLEY. The Zerafshan valley is perhaps the most valuable of longitudinal valleys for our purpose. Rising in the ice cave of its wonderful glacier, amid Alpine peaks up to 18,000 feet in height, at the forking of two westernmost members of the Tian Shan system, its river flows west for 150 miles as a thundering torrent, between the rock walls of its canyon carved in the bottom of a gorge several thousand feet in depth. Out of this it abruptly emerges onto the broad steppes, to nourish the great oasis of Samarkand and those bordering it for 200 miles, till the last of its waters filter away in the gardens and rice-fields of old Bokhara. Once it probably joined the Oxusand only about a thousand years ago filled the canals of Pai- kent, then the most powerful city of Central Asia, but now aban- doned to the desert dunes, from which project its ruined walls. As a longitudinal and structural valley that of the upper Zerafshan has responded to uplift differently from those carved transversely in up- lifted ranges. To begin with, it could not much feel any transverse tilting suchas so affected the Taldic valley to the east and, since it de- bouches from between the ends of two ranges where they die out and seem to have risen but little, it responded slowly up the 150 miles to its source. Moreover, there is more chance for a warp in a long valley than in ashort one. Lastly, the grade of such a long longitudinal valley is necessarily much less than that of transverse valleys heading at the same height on the same range. It therefore had more tendency to fill with the waste of glacial alluviation, especially during long interruptions of crustal movement when aggra- dation of the plains could raise the base-level back upstream, either case giving rise to massive terraces of alluvium after the cutting-down of a succeeding uplift. All the above distinctions are characteristic of the Zerafshan as well as the Kizil Su gorge of Karategin. In general, there seem to have been three cycles of erosion before the present, which makes a fourth, as the stream is now rapidly corrading. Fig. 454.—A Peak South of the Zerafshan Glacier. DESERTS. 277 SECTION OF EROSION CYCLES. First cycle (Pliocene). High relief (Pliocene). Base leveled in part (the peneplain stage of Central Asia). Second cycle (Quaternary). Uplifted. Graded to G — 25 feet (first epoch of glacial period ?). Alluviation to G+ 500 feet. Third cycle (Quaternary). Warped uplift. Graded to G — 300 feet (extensive landslides). Alluviation to G — 100 feet (close of second epoch of glacial period). Aggraded to G (third and fourth epochs obliterated by present glacier). Fourth cycle (Quaternary). Warped uplift. Cut down to G — 300 feet. Still cutting down. (Present glacier and tributaries have made considerable advance in this cycle. During the present oscillation glacier has receded about 250 feet.) FIRST EROSION CYCLE. The vertical degree of cut and fill referred to G level, or the broad floor into which the canyon has cut, varies throughout the valley in such a way as to indicate warping. The values given are about average for the exaggerated portion of the valley, that from Urmitan to Oburdon. Of those far-reaching gradual slopes forming the 20-mile wide valley established by the close of the first cycle, but little now remains and must be looked for surmounting high spurs and whole mountains dissected from it between the present gorge and its two containing ranges. That uplift which ushered in the second cycle of erosion seems to have been especially great through this region and the rest of the Alai Mountains and, though probably contemporaneous with a general breaking up of our first cycle’s topography throughout Central Asia, may have been somewhat sooner here where comparatively little of the old topography has survived. The Kopet Dagh, on the other hand, appear to have lagged behind the general uplift, while the remark- able peneplain of the Bural-bas-tau in the Tian Shan, so well described by Pro- fessor Davis, may have risen still later. Indeed, it is unlikely that all ranges throughout a region so vast as Central Asia would rise simultaneously. If these uplifts, all secondary movements posterior to the birth of the mountains, were, as we suppose, connected with loading and consequent sinking of their adjacent plains, the mountains originally highest ought to have risen first. And this appears to have been the case, for the Pamir and its border ranges, the Trans-Alai, Alai, and those on its east were already deeply gashed with well-developed valley systems long before the first glacial epoch. SECOND EROSION CYCLE. The second cycle appears to have lasted a long time after its uplift had ceased, for during it the river widened its floor, which seems to have been somewhat below the surface of terrace G, till it was in places even wider than that terrace is now. Even after this, erosion continued till by aggradation of the plains the base-level had risen up into the valley, filling it toa depth of some 500 feet with waste. But probably this refilling was partly a glacial alluviation in the first glacial epoch. PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. to NI oe) THIRD EROSION CYCLE. To us the third cycle is most interesting, as it overlaps the second glacial epoch and less time has elapsed to obliterate its records. As a result of its uplift, apparently warped, the river cut down through the alluvial fillings of the second Fig. 455.—Ice-cave of the Zerafshan, Looking from within. Fig. 456.—Terraces of the Zerafshan. cycle and into the old rock bottom to a depth of from 50 feet in the lower to 300 feet in the mid-warp part of the valley with a canyon about as wide as the present and reaching about the same depth. Then the uplift appears to have stopped PLATE 62, Ss. Dy f/ Hi yy Px. Berg schrund vi Uf » Moraine Py ==> River cave—_—— Glacier ice Re Q SS ——S ES OO —Eee oe —— E ZERAFSHAN GLACIER Cobble Plain MAIN VALLEY JUST BELOW THE GLACIER Village of Tro / / f 7 = Vertical and horizontal scale (0) 500 1000 1500 2000 feet ~ “XN ss AT KODISHAR ’ \ G AT PACHURD ABOUT THREE MILES BELOW PACHURD G AT VARSEMINOR UD, Uy %s: TWO MILES BELOW VISHIST AT MINDONA Obu Siab Samarcand Pans Rice fields <—— Flood Plain of cobble stones elds MMM TTT am aU LAAARURLLRAT AN LLU ELLALAD Re rt a SR 6000ft. ec Vertical and horizontal scale tala ceed fe) 1000 2000 + 3000feet tt LEGEND Cco0%S 02 “4 oO S 3 Seo5020 Moraine Loess Uy Slates OSES8GR Li ' e PX VY, SY, Ailuvial conglomerates USA (limestones G (The Great Lower Terrace) is referred ; LAP gs 4 (Unconsolidated) WEES to as the 3rd.erosion cycle Twelve Cross-sections of the Zerafshan Valley. ha > DESERTS. 279 short and the base-level aggraded back into the valley again, refilling it with waste. During this process, and when it had refilled to a height of about 250 feet some three-quarters of the way upstream, the second-epoch glacier advanced to 45 miles below the present ice front. This ultimate point is near the oasis of Mad- rushkent. ‘There, in the face of a 300-foot deep canyon section, may be seen a thickness of some 20 feet of finely stratified light-gray clays, contrasting in lightness and texture with 20 feet of overlying and 200 feet of underlying coarse gravels. But the important feature is its distorted stratification, evidently having resulted from a pushing of the ice front, which also beveled the distorted layers with a sur- face declining upstream or against the river’s grade, the very thing to be expected under the frontal margin of an advancing valley glacier. At intervals for about 20 miles above this point there are moraines scattered over terrace G standing half-buried in its alluvium, and in a tributary canyon section (see plate 62) oppo- Post glacial\ Fes=2== alluvium (cobbles) Thrusted alluvium Glacial_=> hfs : boulders (fine light gray clays) Thrust ———————_» West,down stream Fig. 457.—Section of Glacier-thrusted Alluvium in the Zerafshan Valley, 45 Miles below the Glacier. site Packshiff some of this moraine is seen standing on an irregular surface of allu- vium scraped over by ice and partly buried by later waste. So the second glacial epoch came to a close as the third-cycle gorge continued filling. While aggrading, the valley had widened and, ere the next uplift came, established the present terrace G. FOURTH EROSION CYCLE. Cycle 4, with the last uplift, has resulted in the present canyon, a channel incised from the last meander held by the river at stage G and thus crossing often the old-filled valley of the third cycle. And in its present torrential fall of nearly 6,000 feet in 150 miles it must be rapidly cutting down. Indeed, the deep rumble and grinding of cobbles heard beneath the river’s roar is ample indication of corrasion. The Zerafshan glacier, with its ancient moraine, its relation to other glaciers and uplift, the fine grindings it has supplied to loess steppes, and its influence on civilization, becomes of great interest. Only one epoch of abandoned moraine could be distinguished, and that remarkably far-reaching and of such antiquity that it must be classed as belonging to the first or second of the glacial periods. I have attributed it to the second, because it still rises from terrace G in good pres- ervation. Nowhere has the first epoch moraine been seen with its topography preserved. Recently the glacier has advanced into a part of the valley that had been ice-free for so long that its sides had struck an even slope to the flood-plain 280 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. and received a thick coat of loess mixed with talus. Directly in front both valley sides come down to the flood-plain in this fashion. ‘Time enough has elapsed for accumulation of huge deltas and gradual loess-mixed talus-cones, some of them truncated by the river, and the cutting-down of tributary once-hanging glacier valleys to sharp V-sections and canyons on a regular grade to the river since any time when the ice was farther forward than now. As an exception to this it may be said that the glacier is at present in the process of a minor oscillation that makes it about 200 feet short of a cross-ridge of moraine deposited probably some few years ago. Owing to the depth of its gorge and alpine character of surrounding mountains, the Zerafshan glacier, unlike those of the Trans-Alai and Pamir, is wholly covered with moraine longitudinally banded with the various colors of its different tribu- tary glaciers. But the most significant difference between it and those others lies in its being the only one that yielded no evidence of more than one expansion greater than the present. If the old moraine found half-buried in the alluvial terraces down to 45 miles below the present ice belongs, as seems most likely, to our second epoch, it is easy to understand how that of the first epoch was washed Fig. 458.—The Zerafshan Glacier. away by the river during the latter part of the second and early part of the third cycle of erosion. It is not so easy to understand how the fourth erosion cycle could have obliterated third- and fourth-epoch moraines without obliterating that of the second epoch. Indeed, it seems impossible under the conditions involved. There remain two alternatives: either local conditions were such that little if any expansions took place, or the present glacier is greater and obliterates them. But since the glacier is now less than 15 miles long, it seems necessary to assume that there were local reasons why no considerable advance corresponding to those of the third epoch of Trans-Alai and Pamir took place. It seems quite likely that our fourth-cycle uplift of the Zerafshan glacier and its surroundings took place after the third glacial epoch farther east, where glacial conditions may have been acceler- ated by uplift during or before that epoch. The Zerafshan is now much more actively cutting down than any other large stream met with, and its glacier is advancing* as though the uplift were still in process and accelerating glacial con- ditions also. * Excepting the minor oscillation of recent years. DESERTS. 281 The Zerafshan River is one of the few black rivers of Central Asia. Most of them are red from the gypsiferous beds folded into all those ranges; oxidized sediments apparently deposited under arid conditions similar to the present. Its charge of black stuff comes from the black slates and shales around its glacier, from which fully two-thirds to half its water springs. One of the most striking facts about the river is the increase of volume up- instead of downstream, because after about two score miles from its source more water is lost by evaporation than is gained from tributaries. Therefore, most of its sediment is glacier-ground stuff directly from the ice-cave, the rest from tributary glacial streams taken in a few Berg - schrund ——— N<—__—___—— Valley side—— of loess slope— Glacier ice covered with moraine eae Og eee NOE ee IL TO ag 7 Da. ae) | SY alee ar SS) VA Re ge SS Vqa0s7e vols ee VAG = edt ey iy S ~~ zs a Sy oa O° ¢=—,.**hy 7_—~ a ae NA aes r =2%y 7 eB °Go O5 CAs — aks Qep tte ilies Tie g —— =~ . Soae oy, B25 * 0% °9 fea, ae aear \Q een 2 oP 35 A S9.2% N\ 7 SN = Se ~o 2 Ose =2 09007288 eases ae, WO CO ea SLES eo °S3 Ft = g. SSS ces \\\ ——— Alluvial eee \ oy TTI, = cobble plain fo) Via, V2 mile Fig. 459.—Sketch-map of the End of the Zerafshan Glacier (R = Recession). miles below, and that from cutting-down of its channel. It may therefore be assumed that in total its deposits throughout glacial time till now are mostly of glacial origin, and that the loess derived from it was mostly fine stuff of glacial grinding. In fact, one may attribute a large part of Central Asia’s loess to wind- work over the flood-plains of glacial alluviation, especially those of earlier epochs. The loess steppe of Samarkand has been warped up into a dome about 40 miles across and 1oo feet high in the middle and dissected by old distributaries and irrigation canals of the Zerafshan, some of which, notably the Dargum (fig. 461), cross and join again beyond. ‘The interesting structure thus exposed in clean sec- tions up to 100 feet in height reveals an interlapping of pure loess of vertical cleavage 282 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. with stratified and cross-bedded alluvial grit (fig. 462). Upstream the proportion of alluvium and size of its fragments steadily increase, and about 30 miles above Samarkand the river is flanked by a con- glomerate cliff 50 feet high, with an over- lying coat of loess as well as some thinner beds of it lower down. Another fact of interest over that same distance is the grad- ual transition of color from yellow loess at Samarkand to light-gray loess there. Fig. 460.—Section near the End of the Yarkich Glacier. Since the flood-plain is necessarily much wider at Samarkand, dust has to drift about more there before coming to rest as loess and thus has more chance to oxidize. THE TARIM BASIN. EVIDENCES OF PERIPHERAL UPLIFTS. Passing to the Tarim, western basin of the Gobi, we find its border ranges gashed by gorges with high terraces. The Kizil Su gorge of Tarim and what was seen of its tributary topography is so remarkably similar to that of the Markan Su that one can not help drawing the conclusion that this type is persistent through that region. Uppermost in both we find the old graded-down topography and uplifted piedmont gravel-plains, while between this and the present flood-plain there are various terraces, badly preserved because of the gorge’s narrowness. There are also massive remnants of an alluvial conglomerate several hundred feet thick, recording a refilling that took place, a backing-up of waste into the graded widths established after the first uplift and before the second. The region of Aikart Pass, on the Markan Su route to Kashgar, is a massive piedmont formation of half-consolidated gravels from which project some remnant peaks of the under- lying highly-tilted red gypsiferous series. It appears to be a piedmont series, laid down during the long preglacial erosion that resulted in the old topography referred to. The whole is now dissected by well-developed valley systems to a depth of over 4,000 feet, having been refilled with 200 feet or more of waste, after- wards reexcavated. Some idea of the amount of silt in these streams was obtained in seeing two basins over half a mile wide, that were formed by landslides only thirty years ago, but now filled with red silt. At about 25 miles east of Aikart Pass the Aikart valley is confronted by a high transverse fault-scarp of uptilted piedmont. After making a short bend it cuts through this with a narrower and flat-bottomed valley with vertical sides, exposing a piedmont conglomerate inclining gently east downstream. In the mountain valley above there has been a broad refilling terrace cut on one side by a narrow gorge 220 feet deep. But in the uptilted piedmont those traditions do not hold and we have a flat-bottomed channel with narrow terraces rising, perhaps, 30 feet above stream and converging with it downwards. DESERTS. 283 The principal river which traverses the Tarim basin is the Yarkend Darya, which flows to Lob-nor after having been joined by the Kizil Su from the west, Ak Su from the north, and the Kotan Darya, which heads in Tibet and crosses between 200 and 300 miles of flying sands of the desert nucleus before reaching that trunk-stream. All other streams are consumed in the piedmont zone encir- cling its vast nucleus of flying sands,and through whose uptilted margins of more ancient alluvium they have carved prolongations of their valleys. The piedmont zone thus becomes of special interest in its exposures of various Quaternary hori- zons and stands as a structural key to the crustal movements peculiar to its basin as a whole. With its uptilted margins often composing half its width, this zone varies up to a hundred miles across with irregular limits, here and there containing an isolated area of flying sands, no doubt derived from the sifting of its silt. peer rrr rare eS ——————— Fig. 461.—The Dargum Canal in the Up-warped Loess-steppe of Samarkand. THE UPTILTED PIEDMONTS OF NORTHWESTERN TARIM AS A KEY TO THE PAST. In northwestern Tarim alluviation of the piedmont zone is nearly confined to that from smaller streams, while the larger systems tributary to Lob-nor traverse or cross it in channels slowly decreasing in depth downstream. If it be crossed on a trail over portions independent of these larger streams and where alluviation from smaller streams is building the playas and gravel-plains of to-day; if we proceed mountainwards over such areas, in the course of 15 or 20 miles from the edge of the great sand behind we come to where no deposition is going on, and there begin shallow channels debouching from the margin of abandoned piedmont to spread below. From there on to the mountains this ancient piedmont slopes up, ever higher, above the streams dissecting it. Riding on to this, we find its dry red silt, left prey to wind, has withered into varied and fantastic forms—a vast deflated area of flat-topped remnants ranged in rows, with wide intervening trenches that are half-choked with heaps of sand. These miniature monadnocks of deflation stand to a general level, while the trenches vary up to 15 feet in depth, 284 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. exposing red layers of laminated sandy clay, and doubtless range across the direc- tion of prevalent wind, as there is a constancy of leeward overhanging sides. Everywhere they are associated with heaps of sand derived from the silt, of which all finer material has been drifted away, doubtless to settle as loess in grassy mountain valleys. Anyhow, wherever the finer material is now, it has been totally removed by the wind that excavated the trenches and left their sand constituent behind. Another interesting feature is the frequency of large masses of sand piled on top of these ridges, to occupy spaces of calm in the eddies of windwork. Proceeding still mountainwards, we soon find these trenches of deflation floored by hard gravel-beds, and in the course of a few miles the silt deposit thins out and dwindles into spits and iso- lated areason the gravel-plain, giving it a mottled aspect as seen from a Me distance—mottled only inshadeand 934 texture, as both are red. ‘This is e the transition from silt to gravel, for in a short distance it is all one 75 vast expanse of gravel or cobbles 69 varying up to,4 or 5 inches in size. 647 grass x Here, therefore, is record of two significant changes of conditions suc- ceeding each other—first; a moun- tainward recession of alluviation bringing its zone of fine deposits over its more ancient zone of coarse de- ote posits; second, a dissection of both 20st ick preceding zones by the channels now 15 Ss gy ‘ne cross -oeaee occupied, moving alluviation again to a zone farther out than before the treba ra ete first change. It may be that the page aL first resulted from a decrease of Fig. 462.—Vertical Section of Interlapping Loess and Alluvium in precipitation corresponding to that eee ata Canal Comin an extreme reaction which followed the glacial period, as evidenced by moraine underlying the glaciers of Pamir. That the second resulted from an increase of grade caused by an uptilting of the margins of Tarim will be shown as we proceed. Now we are perhaps 25 miles from the great sand, and our abandoned pied- mont develops into a bad-land topography, an inclined table-land dissected into a desert of red mountains rising ever higher above us as we ride slowly up the bottom of a canyon. At first the canyon walls are built entirely of piedmont conglomerates with here and there a layer 1 to 3 feet thick of silt, and all in slope conforming to that of the plain above. Then towards the bottom of the wall appears a surface beveling the tilted strata of a still more ancient piedmont series, Heater y Mate, le! DESERTS. 285 perhaps the unconformity between Tertiary and Quaternary time. This uptilted series contrasts with that above in being of fine silts and sandstones, with only an occasional bed of conglomerate, and is much more consolidated. But its detailed structure is that of alluvium and appears exactly like those silts deflated on the surface farther out and whose thickness up to 4o feet is exposed in channels cut by larger streams. The beveled strata rise more and more to view with occasional masses rising above their general surface of erosion as monadnocks buried in the later piedmont, and, in time, some ridges rise out into open air, thus taking part in the topography of to-day. Fig. 463.—Terraces of Markan Su. The canyon is now perhaps 500 feet in depth and cut half in the older series. Its depth increases to 1,000 feet and more while the older series rises, outcropping from the now much-dissected piedmont more often, until at length we find it as spurs projecting from border masses of ranges entirely composed of it (the older series), and its tilt has steadily increased. These mountains, thus composed of desert piedmont strata built as these now forming, were in Quaternary time the Fig. 464.—Deflated Silts and Residual Sand-dunes (Tarim). source of the red silts and conglomerates composing the more recently dissected piedmont and are together with them the source of materials laid down on the oasis playas and piedmonts of to-day. Continuing the journey into the mountains there were seen remnants of a high-uplifted and much-dissected topography of mature form and doubtless corresponding to the dissected piedmont derived from them. This topography, none of which may be termed a peneplain, but which nevertheless conforms to that found on the border ranges of Pamir, is doubtless an equivalent of the mature stage found by Davis in the Tian Shan and Bailey Willis in Central China. Such in general are the great features of Western Tarim. Yet one character- istic of importance should be considered. We have seen how a wide margin of 286 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. the Tarim basin has been uptilted in Quaternary time, and it may be said that this tilt gradually increases mountainwards, so that a horizon nearly flat in its far-out extension is bent up to 10° or 15° and more near the mountains. Besides this general marginal tilt, there have been more local movements in the form of broad anticlinal arches nosed up, each with a transverse fault-scarp facing moun- tainwards and surmounted by its anticlinal surface sloping gently back into the plain. These ‘‘up-nosed’’ piedmont strata, bent behind and faulted out in front, some of them rising as high as 500 feet out of the plain, range about parallel to the bordering mountains and are found even as far out as Kashgar. EVIDENCE OF RECENT CHANGE TO EXTRA DRY. The last great change over the Tarim basin has been one of desiccation. Of this we have both physiographic and historic records, which tell that it became serious about a thousand years ago, when some hundreds of cities were over- whelmed by sand. Some of these ruins were excavated by Stein, and Mr. Pumpelly found mention of them in Chinese literature in the imperial archives of Pekin. It is also believed there were then expansive bodies of water of which Lob-nor and other shrunken lakes and brackish tarns are the withering survivals. During the time of greater precipitation much of the great area of dunes throughout Tarim was doubtless grassed over, and we may thus ascribe its burial of cities to sand set free when rainfall had so seriously decreased that grass failed and left the dunes bare and free to drift. TENTATIVE RECONSTRUCTION OF THE PAST IN TARIM. First cycle (Pliocene). The Tarim basin defined with high border ranges eroding to mature topography and building immense piedmonts of gravel, sand, and silt. Second cycle (Quaternary). Uplift of border ranges, with deep gashing of old topography and sinking of plains, with upbending margins worn down by streams beveling their tilting strata, with an erosion plain and the building of a later piedmont over that. Third cycle (Quaternary). Second uplift of border ranges, with terracing down of valleys partially alluviated ere the close of the second cycle, and marginal tilting up of plains with dissection of their second-cycle piedmonts? Shrinkage of alluviation at close of the glacial period and recession of silt zone over gravel zone. Fourth cycle (Postglacial). Third uplift of border ranges with stream-channeling of valley flood-plains of glacial alluviation during the third cycle, and more sinking of plains with tilting and dissection of their third-cycle piedmonts. Recent decrease of precipitation, shrinking of rivers and lakes, and desolation of dune pastures, setting free the sand that buried the cities of Tarim a thousand years ago, DESERTS. 287 THE FERGANA BASIN. ITS BROAD OUTLINES. The Fergana basin lies north of the Pamirs as a deep embayment between the Alai Mountains and Tian Shan, or rather between two far-western members of the Tian Shan that branch out and nearly join again in the west to form a structural depression. Its plains, about 200 miles long and 50 miles wide, are thus nearly closed around by high mountains and connect with those of the Aralo- Caspian basin, of which it is a tributary, through a western gap only 15 miles wide, of which over 10 miles is blocked by half-buried mountains. Though it has doubtless accumulated most of its erosion products, an important portion must have escaped in the Syr Darya, through the outlet to be accumulated in the greater basin. With its border ranges the Fergana basin attains a maximum relief of 18,000 feet, while passes stand ordinarily about 13,000 feet above its lower plains of 1,000 feet elevation above sea. As an organic whole it approaches nearer the ideal type of basin than any other one considered by us. Its high mountains have responded to a varied series of glacial changes, and its nearly self-contained state has resulted in the differential crustal movements characteristic of such basins. The Alai Mountains ranging along its southern border rise rather abruptly as seen from Marghelan. From there they appear as high snow-mantled pyramids and giant peaks with cliffs truncating broad sloping fields of crevassed snow and ice. This high crest ranges east and west behind a flanking mass of rather flat- backed mountains with but few projecting horns—a high, uplifted, outlying mass, once base-leveled, but now deeply gashed by gorges that end abruptly to open out in a fairly even line. It was shown in my report of 1903 that the Alai Range appears to have been thrice uplifted in Quaternary time. ALAI EROSION CYCLES (BASED ON TALDIC PROFILE). First cycle (Preglacial). Pliocene, Alai worn to low relief and half-buried in piedmont deposits with projecting monadnocks. Second cycle (Quaternary). Uplifted about 1,500 feet and gashed with valleys that widened after uplift had ceased and partly refilled as the plains aggraded, raising their base-levels. Third cycle (Quaternary). Uplift tilting the transverse horizontal till the dissected northern side was about 3,500 feet higher than the buried southern side; valleys then alluviated again as in the second cycle. Fourth cycle (Postglacial). Uplift of about roo feet with canyoning of third-cycle flood-plain still continued. UPLIFT OF THE TIAN SHAN. In the Bural-bas-tau of Tian Shan, northeast of the Fergana basin plains, Professor Davis found a key to the history of that side. He states: The evenness of the plateau-like highland, all snow-covered at an estimated height of at least 12,000 or 13,000 feet, was most remarkable. . . . It must have gained its present altitude with comparative rapidity, and in geologically modern time. . . . . When it still lay low, the lowland of which it was a part must have been much more extensive than the present highland; for lowlands can not be worn down on resistant crystalline rocks without the very general reduction of all neighboring and quiescent structures. 288 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. And, as he goes on to say, the fact that some of the Tian Shan ranges are now gashed into more alpine form means that “they were somewhat less worn- down in the previous cycle of erosion, or, like the Dongus-tau, already more con- sumed (owing to greater uplift or to weaker structure) in the present cycle of erosion or both.’’* te us alt sat - Wy Se a Wl | a \\ Hh n pi il ie tlh) pil ; yi MEN HWS Hil MWh { \\ \ me i W wl Bp Ha Araya Be Hi Mi ih \\\ < = \\ SMW) Hii i TZ. aa HH i ii wee WS Aue \ys Is ‘é Tw’ V4 WN MA eee ve Till mM bl ¥ LS i Fr gE sii ii i piso Ch nk at of : yi Hi oe ie yt > ‘4% :, g \ Mr a i WA “ sng os ah he Zany satis wayuill fie a a va : why, AANA Zi Z Y : ail ey ex Aare SZ ky) ““ BB 2\ Ve Wa, gs a ; o \\ Al / NS sie ii \ Z gil aN e 10 ceased ae Kurgan ----- Trend of faults Fig. 465.—Uptilted Piedmonts of Marghelan. In the Fergana basin we have, as already hinted, a remarkably good example of the differential earth-movements characteristic of a sinking deposition area with uplifting of eroding border ranges. And there, as in the Tarim basin, this movement has involved an encroachment of mountains on plains through uptilting of their margins. We have seen how, with the Alai Mountains, uplift was periodic with intervening times of quiescence, which with the Taldic gave rise to four erosion cycles, the first one closing in Central Asia’s peneplain stage, and how * Carnegie Institution of Washington Publication No. 26, Explorations in Turkestan, 1903, p. 73. DESERTS. 289 that low relief was dissected during the following three uplifts with their erosion cycles. It is gratifying to find corroboration of this threefold division of uplift in the structural deformations of plain-deposits derived therefrom. ‘There are three rows of uptilted piedmonts ranging parallel to the Alai Mountains, as the three respective marginal deformations corresponding to the threefold uplift both in magnitude and degree of subsequent erosion. It is further interesting to find that these three belts of uplifted piedmont, which converge towards the mountains opposite Khokand, near the western or lower end of the basin, widen eastward to include a considerable area of the eastern end of the plains, where deposition has necessarily been much heavier, as that portion lies before a vast mass of high-uplifted mountains. Another feature affected by this process and characteristic of the Fergana plains, especially on their Alai side, is the threefold, and sometimes fourfold, division of distributary systems. Larger streams descend- ing from the Alai cross all three belts of uptilted piedmont, having cut down as they were uplifted, and apportion their depositions over areas between them, some escaping beyond the last to spread towards the middle of the basin. Each stream, therefore, gives rise to from two to four successive groups of delta-oases, thus giving an interesting variation of type Ia of my classification. In several instances a stream escapes from its first delta in two or more distributaries to form other deltas beyond, so that the intervening uplifted ridge of piedmont has been dissected by two or more channels separated by many miles (fig. 465). The oldest belt of uptilted piedmonts bordering the Alai Mountains probably contains products of the first erosion cycle and was thus upheaved during the uplift which caused the breaking up of its peneplain stage. On the Terek trail between Osh and Gulcha it attains a height of over 3,000 feet above the present Gulcha River flood-plain. There it is a mass of loosely cemented conglomerates with confused dips, and of which the old surmounting piedmont topography has nearly disappeared. The later upheavals of plain-deposits rarely rise more than a few hundred feet above present deposition. Where we approached the northern margin of the Fergana plains in the regions of Chust and Khojent, only one belt of uptilted piedmont was observed, probably because the mountains on that side are much lower and doubtless have been so in the past. This one belt, however, is of especial interest as it can be traced all the way from Namangan to Khojent, a distance of about 140 miles, and crosses the Syr Darya, which has cut down as it was uplifted. From the standpoint of hydrography alone, the Fergana basin expresses the same series of cyclical events deduced from the topography and structure of its mountains and plains. In the mountains larger streams join at oblique angles and generally inherit the courses held before the first uplift of the low relief then drained by them. Their smaller tributaries contrast with this rule by joining the larger nearly at right angles and have come to be since that uplift, We have seen how streams descending from the Alai split into distributary systems in a manner determined by the marginal deformations of the plains and that feature may be recognized with a glance at the hydrography on a large-scale Russian map. 290 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. PRESENT OSCILLATION IN THE DEPOSITION ZONES. Four out of the five zones of deposition characteristic of a desert basin are found in Fergana, namely, glacial, alluvial, flying sands, and loess, while the lacustrian is lacking. An area of about 500 square miles of the central portion of the basin forms its nucleus of flying sands, while two other smaller nuclei of dunes are found in its western half—one just west of Khokand, the other crossed by the Syr where this basin narrows near Khojent at its western end. Most of the alluvial zone is barren gravel steppe and only its small areas of silt give rise to oases. Loess is nearly confined to its bordering uplands, a portion of which is composed of uplifted piedmont structures. The nuclei of flying sands are now expanding, steadily encroaching on the alluvial zone as its perennially flooded areas decrease, a fact evidenced by Russian and native records as well as our own observations. A considerable part of the loess zone, its lower areas, is now dead and suffering erosion, while living grass-covered areas have shrunken to pastures ranging above 4,000 to 6,000 feet. RECONSTRUCTION OF PAST EVENTS OF THE FERGANA BASIN. First cycle (Pliocene). Pliocene basin defined with high relief of border ranges eroding and building the plains of waste under desert conditions. Cycle closes with the low relief of Central Asia’s peneplain stage. Second cycle (Quaternary). High block uplifting of border ranges with gashing of their mature first-cycle topog- raphy, and corresponding marginal upbendings of the piedmonts constructed by the first cycle. Closes with well-widened mountain gorges partially alluviated. Third cycle (Quaternary). Similar to second, but of less magnitude. In the Alai Range, characterized by tilting of uplifted blocks raising that side next the plains. Nothing is yet known about it in the Tian Shan to the north. Its mar- ginal deformations of piedmonts built by the second cycle were on new lines farther out on the plains than those upheaved by the second cycle. Fourth cycle (Postglacial). Similar to the second and third, but of much less magnitude and still in the uplifting stage. Marginal deformations of third-cycle piedmonts on new lines still farther out on the plains. Climate in process of desiccation, flying sands expanding, alluvial activity shrinking, living loess areas shrinking to higher limits. This reconstruction of the past is of course purely tentative and subject to subdivision with cycles of less intensity or duration. DESERTS. 291 THE ARALO-CASPIAN BASIN. COMPLICATIONS AND FUNCTIONAL PECULIARITIES ARISING FROM SHAPE AND SIZE. The broad definitions and most of the great features of the Aralo-Caspian basin have been so well set forth by Professor Pumpelly that little more than a description of its four deposition zones remains to be here undertaken. In the Tarim and Fergana basins we found that plains of deposition and their deformed margins express fully as much as their bordering mountains do, and perhaps even more, of the history of their basins. This is especially so of the Aralo-Caspian, where lacustrian zones become an important aid. Its peculiar shape and but half-isolated state, with naught but low divides north and west, have given it a history so involved and complicated with outside influences that we must be contented with explaining only a very few of its greater features and direct our attention to those concerning the archeology of its southern part. An important part of its physiography has already entered this report in undertaking that of the Pamir, Alai valley and Kizil Su, Karategin and Hissar, Zerafshan, and Fergana basin, which taken together compose most of its high eastern drainage area. In considering the immense area of high mountains drained r : -—— a 2a en - - —-- ae Fig. 466.—Manish Valley Terraces (in the Kopet Dagh Mountains). by the nine rivers—Ural, Chu, Syr, Zerafshan, Amu, Murg-ab, Tedjen, Atrek, and Kur—and the great flow of water brought in by the Volga, draining all of Eastern Russia, we must marvel at the degree of aridity signified by such a small area of water exceeding evaporation as that which survives in the Aral and Caspian— its two shrunken seas. Only half its larger rivers reach their seas at all, while the remainder with scores of other streams, by no means small, dwindle away on the plains, and others fail to reach even the plains, but die far up in the valleys they excavated when precipitation was heavier. Central-Asian hydrography is thus often the reverse of drainage systems in regions of ordinary rainfall familiar to most of us. Many of its rivers have the aspect of a drainage system reversed as they decrease in size downstream and finally split into distributaries resembling tributaries on the plain. The evolution of a basin so large as the Aralo-Caspian can not be expected to have been through the cyclical uniformity followed by a small basin with a complete periphery of mountains. Indeed, it is so large that there is no surprise in finding that different portions entered from time to time into fairly independent series of changes, or developed in a way causally connected only in a broad sense. 292 PHYSIOGRAPHY OF CENTRAL-ASIAN~DESERTS AND OASES. To begin with, the mountains directly bordering it have contributed but a small portion of the sediments spread before them. The Amu, for instance, brings most of its load from many hundred miles behind the border ranges, whereas these border ranges probably had the benefit of rising to give way for sinking of most of that load, if the differential movements of sinking plains of deposition and rising of worn-down mountains resulted from such a displacement. Another deviation from the ideal type is a serious difference in height of mountains from east, where they rise to over 20,000 feet, to west, where the Kopet Dagh attain a maximum of only 9,000 feet. And this difference appears to have been no less in Tertiary time, for no crystallines have yet risen to view in the Kopet Dagh, though the Pamir and Tian Shan have weathered to their granite cores. More sediments were, therefore, loaded onto the plains in the east, and we find that mountains there were uplifted higher and more often than the Kopet Dagh, where only one really great erosion cycle appears to have followed the peneplain stage, though the second and third uplifts are recorded on a small scale. Besides all this, the plains have been seriously warped, a fact evidenced by channels of the Oxus and Zerafshan and other streams, probably as a result of the great difference of deposi- tion load over different areas. Perhaps even more serious than these internal organic peculiarities must have been the climatic complications arising from external sea connections. A recent connection between the Caspian and Black Seas through the Manitch, north of the Caucasus, and a more remote overflow of an Aralo-Caspian Sea through Siberia to the Arctic, limit the possibility of climatic calculations based on determinations of sea-water areas. The moment an interior sea either overflows or receives an influx from some other basin or of tide-water, its surface area ceases to be a function of climate. And, on top of all this, the river Don, behaving in the uncertain manner of a large aggrading river, may have flowed alternately into the Caspian and Black Seas. It is, however, hoped that a more complete study of the past and present marine fauna and flora of the Aralo-Caspian Basin may decipher most of these remarkable hydrographic variations. THE LACUSTRIAN ZONE (ARALO-CASPIAN SEA EXPANSIONS). Its lacustrian or marine history, though involved, affects such a large area and throws so much light on its climate, especially during archeological time, that it becomes of first importance. Russian and other geologists have naturally directed special attention to this aspect of the basin, but their work has so far been lacking in quantitative results. After the early Tertiary upheavals of its peripheral mountain ranges, a Tertiary sea, perhaps at first connected with the Mediterranean and Arctic, appears to have shrunken till in late Tertiary time it washed the oldest shores indicated on Konshin’s map (fig. 467). From that it withered into an Aral and Caspian with shores below those of to-day. This fact, recognized first by Davis, is demonstrated on both sides of the Caspian, where valleys dissect- ing its high-level Tertiary sediments are contoured up to 200 feet and higher by shore-lines of its later Quaternary expansions, but extend down under water or DESERTS. 293 are drowned by the present level. This reasoning is enforced by Walthers’s state- ment that in a well-boring on the Caspian shore, southeast of Krasnovodsk, dune- sand was penetrated to a depth of 35 meters below the sea-level. Therefore, a long interval of subaerial erosion elapsed after the Caspian had shrunken from its great Pliocene Aralo-Caspian expanse to below its present level, and before it rose to its higher Quaternary shores. Assuming that it expanded from this low level, called early Quaternary by Davis, to unite again with the Aral and transgress the Kara Kum, we have an early Quaternary cycle of desiccation MAP OF THE ANCIENT COURSE OF THE OXUS (AMOU-DARIA) RC WSs . and the several phases Sh ¥; SAQws \ of the recession of the os : ARALO-CASPIAN SEA ISEKe . from the beginning of the quaternary epoch to the present, 50 100 200 of VOus -iourt D | e Scale of Kilometers Lonzitude of Poulkova (27°58’ W.Paris) 42 . | \ Puits/Teherychly Nee y Wess. i Plateau of the | Kara-Koum | SATS SS | | a . Ziandin Pehardjoui ] aa] ON ial / -\- ¢Bourdalik ; v4 O\\_ | [| J ofuitsiouten-aat | “>. Eletan 3 Soultan-bent RS Ancient limits of the Aralo-Caspian Basin bas @S¢rask at the beginning of the quaternary period. t KS QQ Limit of the Aralo-Caspian Basin during the quaternary period. PON Limits of the Caspian Sea and of the Aralo-Sarykamych Sea bP Bes at the beginning of the present period. i / 2 Pende 32 o.. 2 27 28 29 33 After Konshin Fig. 467.—Map of the Aralo-Caspian Expansions (Konshin). as distinct from the Pliocene. During this it appears to have fallen to lower and lower shores till but a narrow connection existed between the Aral and Caspian through the Sari Kamish Basin. In the course of desiccation both seas shrank till this connection became an overflow of the Aral through the Usboi channel to the Caspian, which is probably an historic stage. Davis’s recognition that “the Quaternary sea resulted from an expansion of a smaller early-Quaternary sea, to which the waters had shrunk from their great Pliocene extension,’ becomes of importance in that it demonstrates a pre- glacial aridity more pronounced than that of to-day. And if we place the great 294 PHYSIOGRAPHY OF CENTRAL-ASIAN. DESERTS AND OASES. Pliocene sea contemporaneous with the peneplain stage of its peripheral mountains, and the very shrunken early-Quaternary sea with the following high-uplifted stage of those mountains, still preglacial, the phenomena fall into organic accord; for the surface area of a landlocked sea is a direct function of the climate of its basin—the climate of its basin varies with the general continental geography and all mundane climatic change. Of the geographical factor, most important is variation in the shape of the basin; and, as shown in preceding sections, this shape changed through a series of erosion cycles with uplifts giving its periphery alternately low and high reliefs. Assuming, then, that the great Pliocene Aralo-Caspian Sea belonged to the peneplain stage at the end of our first erosion cycle, and that the early Quaternary, low Aral, and Caspian shrunken survivals of that sea belonged to the high-relief stage of its uplifted periphery of mature mountains dissected during our second erosion cycle, we are next confronted by the later wide-expanded Aralo-Caspian Sea. Knowing that during the third and fourth erosion cycles this basin suffered a great mundane change of climate in the glacial period, it is natural to correlate phases of its high Quaternary sea with assumed increases of precipitation belonging to the glacial epochs. Although we do not, as yet, know how many phases there were to the high Quaternary sea, it might roughly be called a glacial sea, leaving the epochs to future exploration. It may be that Konshin’s next lower Aralo- Caspian shores, when the Gulf of Kara Kum had dried out and only a strait running south of Ust-Urt connected the Aral and Caspian, were postglacial. This brings us into the third erosion cycle of its high eastern drainage; and the following uplift, ushering in our fourth erosion cycle, might account, in part at least, for the last historic shrinkage, severing the Aral and Caspian from their recent Usboi overflow connection to the two low seas of to-day—withered survivals of a glacial mediterranean. RECENT DEVELOPMENTS IN THE ALLUVIAL AND FLYING-SANDS ZONES. Turning to the other three lowland zones of deposition, we find additional data, especially on the more recent developments of this great basin. Glancing at a large-scale map we see most of that area from the Caspian to the high eastern peripheral ranges covered by dune-sand. Russian geologists have ascribed that of the Kara Kum to deflation of its Quaternary sea deposits. But, as we have shown, vast nuclei of flying sands inevitably accumulate from wind-work over the silted flood-plains of a desert basin, and all areas of its plains in the neighborhood of alluviation, whether or not far removed from ancient sea deposits, are charac- terized by them. We, therefore, differ by attributing much, if not most, of the Kara Kum sands to wind-work over the flood-plains of late Quaternary time. Some of it was undoubtedly derived from deflation of marine deposits, but those deposits are much more resistant than fresh-dried alluvium, especially than the sandspits of its distributary channels. In either case all the sand is ultimately of alluvial origin. The wide-expanded zone of flying sands we find surrounded by an alluvial zone, narrow between it and the southern peripheral ranges, but widening east- DESERTS. 295 ward over the Hunger Steppe and Fergana plains, 7. e., widening in proportion to the height of mountains drained. This alluvial zone, furthermore, extends into the great Sand, where it is penetrated by the rivers Tedjen, Murg-ab, and Zerafshan, and where it is divided by the rivers Syr and Amu crossing to the Aral Sea. Now, it is a fact of significance that all five of these large rivers, as well as many smaller ones that still reach, or have recently reached, well out onto the plains, have cut channels from ro to 100 feet or more in depth to where they debouch over deltas. It is, moreover, characteristic of these channels that they vary in depth in such a way as to indicate a varied warping of the plains. And though most of them are still occupied by streams, there are many instances of channels now always dry, but so recently abandoned by the streams now ending many miles above in a shrunken condition that ground-water still survives, obtain- able in shallow wells of the nomads. On our large-scale Russian maps there are remarkable fragments of such channels so far removed from present alluviation that it is difficult, sometimes impossible, to say what river they belonged to. Others appear to have been the work of distributaries cutting into the plains they had once overflowed. Where distributaries have been thus incised, we have definite proof of crustal movement. Our most striking examples of distributaries cut into a warped plain are afforded by the Zerafshan, while of those cut into the zone of uptilted piedmonts we find most remarkable examples along the southern border of the Fergana plains. The vast alluvial zone of this basin was built by its rivers when they wan- dered freely. Now most of them are relatively fixed. That the Turkoman Trough was at one time the Amu’s flood-plain, when that river flowed to the Caspian, building the immense deltas characterizing the coast south of Krasno- vodsk, seems more than likely. That would be postglacial. Then it and doubt- less most of the large rivers were unconfined and spread a large portion of their load on the plains, whereas silt of the Amu and Syr of to-day is mostly in transit to the Aral. This period of free-shifting rivers with unconstrained alluviation was followed by warping. Here we must remember the postglacial uplift of mountains, the peripheral uplift of our fourth erosion cycle. The warping of plains, uptilting of their margins, and uplift of their border ranges fall logically together into one cycle of a basin’s differential movements. As a confirmation of this idea we have the corresponding increase of aridity, shrinkage of sea-water area, contraction of streams, shrinkage of living loess, and expansion of flying sands, and, finally, depopulation of oases. RECENT CHANGES IN THE COURSE OF THE OXUS (AMU DARYA). The archeologic and historic period of this basin is treated under “ Physi- ography of Oases,’’ chapter xv, this report, but there has been so much dis- cussion about historic changes of river courses, especially of the Oxus, that a physiography of the basin must take up the problem. Elisée Reclus states: The great changes that have taken place in the course of the Oxus within the historic period are amongst the most remarkable physiographic phenomena, comparable in recent times only to the periodic displace- ments of the Hoang Ho. . . . . In the days of Strabo the Oxus, ‘‘largest of all Asiatic rivers except 296 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. those of India,’’ flowed to the Caspian, and the trade between the Euxine and India followed this river, continuing the valley of the Kur eastwards of the Hyrcanian Sea. But in the time of the first Arab and Turkish writers, the Oxus, described by Edrisi as ‘‘superior in volume, depth, and breadth to all the rivers of the world,’”’ had been diverted northwards to the Aral. In the fourteenth century it had again resumed its course to the Caspian, towards which there is a relatively steep incline, for the bifurcation of the present and the old bed below Kunya-Urgentch is 140 feet above the level of the Aral, and 380 feet above that of the Caspian. The new channel was blocked for about 200 years; but towards the middle of the sixteenth century the Amu, for the second time during the historic epoch, shifted its course from the Caspian to the Aral. If so, it has followed the present course for only about 350 years. These facts, based on the writings of classical and medieval travelers, and ancient maps, make it appear as though the Oxus were normally an affluent of the Caspian. It was not until the last few decades that actual physiographic study of the region opened up another side to the question. Konshin, Mushketoff, Sievers, Hedroitz, Lessar, and Somkoff have made special study of the problem. The now dry Usboi channel, from just south of Krasnovodsk, skirting around southeast of the Ust-Urt northwards to the tarn of Sari Kamish, has thus been a great subject for controversy. Few geographical problems have become more familiar than the question as to its origin. At first it was naturally taken for the historic course of the Oxus. Elisée Reclus, in reviewing explorations up to the time of his great work, was sure that it was. Conshin, after exploring it for two years, decided that the Oxus had never flowed that’ way directly, but that it was an ancient channel through which the Aral overflowed to the Caspian. The data now at hand are as follows: The Usboi is a channel in the uncon- solidated sediments of the steppe, starting southwestward from the Sari Kamish basin and thence skirting around the Ust-Urt escarpments down into the Balkhan Gulf of the Caspian, a distance of over 200 miles, with a total fall of about 234 feet; it averages 60 to 70 feet in depth, about 3,000 feet in width, and resembles a river-bed with occasional islands and rapids, and in it still survives a series of brackish “‘shores’’ or pools. Three ancient distributaries of the Amu, channels now dry, run from the Amu’s present delta into the Sari Kamish basin. Elisée Reclus states that during the inundations of 1878 the river discharged 31,500 cubic feet per second to the Sari Kamish. That it formerly flowed there regularly is evidenced by two epochs of ruined towns and cities along the abandoned courses. As there are no ruins along the Usboi, its water is supposed to have been brackish. The Usboi is, therefore, supposed to have been an overflow channel from the Sari Kamish, into which both the Amu and Syr have emptied. As the divide between the Aral and Sari Kamish basins appears to be at least 60 feet above that between the Caspian and Sari Kamish, the Aral was doubtless dried up when both its rivers were thus diverted, and only a small lake at whatever time the Oxus alone flowed west. As a full Sari Kamish sea would be of much less surface area than the present Aral, or about 130 by 70 miles, the Usboi would still receive an overflow if both rivers were again diverted there. The Usboi, therefore, throws no light upon the climate of our basin. But the more ancient stage of a wider strait or continuity of level between the Aral and Caspian, that stage which DESERTS. 297 Konshin believes to have followed the high Quaternary sea, is another, more ancient, and indicative of a different climate. Though thus forced to disbelieve that the Oxus ever flowed through the Usboi, we can not utterly discredit the writings of geographers and travelers so renowned as Ptolemy, Strabo, Pliny, and others. Assuming that during one or more periods of the past two millenniums water of the Oxus did flow to the Caspian, there are two alternatives: First, that the Usboi overflow, as a continuous waterway from the Caspian up the Oxus, might have been referred to as the Lower Oxus; second, that the Oxus may, in historic times, have flowed to the Caspian through some other channel. As both the Aral and Sari Kamish were through whole centuries omitted from writings and maps, it would seem that whatever waterway there was must have been far south of them or that they were dry. On the other hand, there appears to be little doubt that the Oxus recently flowed west from near Charjui through the Turkoman Trough and so into the Balkhan Gulf. This channel, the Kelif Usboi, or Ungus, indicated on Russian maps and known to the Turkoman, has not attracted government exploration as an engineering project, such as the Usboi of Ust-Urt, and must therefore remain only a possibility. However often the Oxus may have shifted, or whatever course it may have followed in reaching the Caspian or in contributing overflows to that sea, the total surface area of sea-water in the Aralo-Caspian basin would have been but little, if at all, affected by such oscillations. The River Don problem is more serious from this point of view. The Don, converging with the Volga to a point about 350 miles north of the Caucasus, now bends sharply away from there and flows to the Black Sea. There appears a possibility that it was once a branch of the Volga. The change of course may have resulted from faulting across the channel, and the river’s grade is very slight—only about 5 inches to the mile. But if it ever did flow to the Volga, the change to its present course was so long ago that time enough has elapsed to cut the present wide Don valley in con- solidated rocks. Assuming that Don water flowed to the Caspian before earth- movements forced it westward and to cut a deep channel in the plains, the sur- face area of Aralo-Caspian sea-water would not have been so much increased as might first appear the case. It is much smaller than the Amu and yet the Amu and Syr together maintain a surface area of only 26,300 square miles—that of the Aral. If the Don were now diverted to the Caspian, it might raise it till its surface area increased by perhaps 10,000 square miles. But that would change its present outlines but little except over the low marshes of its northern end, while the Kara Kum would be transgressed by some few miles. The Don, therefore, can not have effected any of those great changes we are discussing. 298 PHYSIOGRAPHY OF CENTRAL-ASIAN_ DESERTS AND OASES. TENTATIVE RECONSTRUCTION OF THE PAST IN THE ARALO-CASPIAN BASIN. First cycle (Pliocene). Opens with high relief of Pliocene mountains. Closes with low relief of Central Asia’s peneplain stage; great Pliocene Aralo-Caspian Sea; broad continuity of grass over the plains and gently rolling steppes of their worn-down peripheral mountains. Second and third Cycles (Quaternary). Preglacial: High uplift of peripheral mountains and general sinking of plains with upbending of their margins; change to high relief brings intense aridity and shrink- age of the great Pliocene sea till separated into an Aral* and Caspian as small if not smaller than the present. Most of what was water and grass in the first cycle is changed to barren desert. Glacial: Toward the end of the second cycle there begins the first epoch of the glacial period with its increase of precipitation f effected by mundane change; expansion of valley glaciers and ice-domes in the mountains, and birth of the great Russian ice-cap; swelling of rivers and seas till all unite into one hydrography, an Asiatic mediterranean reached by all the rivers of its basin; broad continuity of grass steppes and moderate climate with vast accumulation of loess. Between the second and third cycles a second peripheral uplift takes place with a sinking of the plains and marginal deformations of the piedmonts. he Gay or the rate of aggradation multiplied by the time since foundation minus the time that elapsed between then and the beginning of burial. Changing the equation of obliteration somewhat in form, we get our third and most important equation. (3) 1+ 2 when h=o which means that on aggrading areas any town, not occupied more than the ratio eee +e times the number of centuries since it was founded, has vanished from sight beneath the aggrading plain. The depth to which the eroded top of its accumula- tion has been buried can be found from equation (2). Assuming Professor Pumpelly’s values obtained at Anau, we have G=2 and A =o.8, and since it is from erosion the growth of plains is supplied and since the areas of erosion and aggradation seem to correspond in a general way and our cul- ture mounds probably erode as fast as anything, we may for experiment assume o ef. A+E 1.6 E=A or E=o.8. Then Conese error than equation E = A, because E partly compensates itself by division. =0.57 as a conservative ratio of much less OASES. 305 _ It may, therefore, be said that the kurgan remains of most favored oases, those where water was easily led and found its way in flood, have been eroded and buried, wholly obliterated unless they were occupied over 0.57 of the time since their foundation. Or, any such city, founded 5,000 years ago and not occupied at least 3,000 years, has vanished. Such must have been the fate of those where the plains are always aggrading; but along the borders in the region of uptilting piedmonts, where oases of type 1b and type IV abound, aggradation was apparently so counteracted by crustal movement that during dry periods some areas rose above it altogether. It is to that process we owe the preservation above ground of both kurgans at Anau. The piedmont on which they rest appears to have been uplifted so nearly as fast as the plain aggraded that during all of 10,000 years no more than 20 feet of sediments has risen around the North Kurgan, though during that time it seems likely that about 80 feet have been deposited farther out in the desert. As a check upon the logic of our equations, it is interesting to solve for occupation on the North Kurgan. Knowing the total remnantthickness of culture, 60 feet, and assuming Professor Pumpelly’s estimate of time since foundation, 100 centuries, we have: taking the remnant height, h=!1G—E (t—l), 64=1 X 2—0.8 (100—1/) or 2.8/=144 and /=51.5 centuries. Then its original thickness, /G, would have been 2 X 51.5 =103 feet, of which it has lost by erosion 43 feet. This, taken in view of the deformation testified by its profile, the upper part having withered to a rounded form leaving a base under ground of twice the diameter of that above ground, seems a conservative estimate. There is, however, no reason to suppose that it had ever attained a thickness of 103 feet at any one time. This accumulation may have resulted during several periods of occupation, between which it was abandoned to erosion. That part of its growth was during reoccupa- tion of comparatively recent times is evidenced by quantities of late pottery and débris, deep-buried in the wash under its surrounding plain, having drifted down there from above, though now absent on top. An examination of the surface and gully-sections of 20 or 30 kurgans on the zone of constant aggradation, revealed no indication of anything over 1,000 to 3,000 years old, and such was Mr. Huntington’s experience in his study of 20 or more kurgans north of Merv. To one having a general idea about rates of erosion and aggradation in this region, it is no surprise that nearly all the kurgans of Central Asia seem to belong in the last 2,000 or 3,000 years. ‘There can be very few as old as the North Kurgan at Anau still above ground. But the very fact that Central Asia has been progressively drying up has helped prevent a universal burial of oases of type Ia, those on far-out deltas, because streams and their canals contracted, leaving their oases beyond both water supply and sedimentation. Many kurgans, then, still rise above the desert for the very reason for which they were abandoned. Another set-back against burial has been warping. The region of Samarkand has been warped into a low dome about 40 miles across, through which the Zerafshan distributaries—many of them artificial—have cut their channels to a depth of 50 to roo feet in the highest or middle part, some of them reuniting beyond. Now, it is only with canals many 306 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. miles long that water can be led onto the old plateau and on such a gentle grade that scarcely any sediments are introduced, so that the surface there to-day is practically that of antiquity, the uplifted and dissected horizon of many thousand years ago. Type III or high-valley oases, because of their position on terraces high above the river, have perhaps been most free from natural burial; and owing to isolation from hostile tribes and their excess of water-supply, were less often abandoned to erosion. Favorable topography, crustal movement, and shrinkage of alluvially active areas have thus conspired against a universal burial, but we still face erosion. The question is, what is the reason why Nature’s wind and water have not alto- gether obliterated whatever she left unburied? It can not be from absence of the agencies of erosion; they are universal. There must have been protection. We have seen in the equation h=/ G—[E(t—l)+At] that as long ast< ress there is still a discoverable remnant. If /=1#, that is, if the kurgan is still occupied and never was abandoned, then the factor E(t{—/) =o. That is, no erosion has taken place. For several reasons, climatic and protective, it has always been advan- tageous to reoccupy old culture-mounds. In other words, the value of / was from time to time increased and so the time necessary for obliteration, eireaat increased whether E alone was involved and A =o, or when A alone was involved and & =o, or when both together were involved. In a grassy region the value of E is greatly diminished and vanishes altogether where loess is precipitated. Such a region is found in the wide Hissar valley, where kurgans are mantled with a foot or more of loess and the great citadel of Hissar itself, rising about 100 feet above the plain, stands as evidence of that pro- tection. It is not unlikely that other kurgans, now bare, were once favored with grass. And there is one other protective agency that has played an important role in the preservation of ruins to discovery; and that is flying sands. In Chinese Turkestan many cities have been reexposed by the shifting dunes that swallowed them over a thousand years ago. Such must have been the fate of most of type Ia, or far-out delta-oases. But do not be too encouraged by this display of protections against oblit- eration. We still must face the fact that few remains of very great antiquity have as yet been found. Erosion and aggradation have done their work and, where time allowed, have successfully obliterated. Whatever may be their actual values in any given region, the agencies of Nature, where continuous, have com- G+ ATE believe that most of what they have buried shall remain to us unseen. And high- valley oases, type III, though in other ways more favored, had to be on or near the edge of such steep slopes and receding canyon walls that many of them have been long since blown and washed away or caved into their valley torrents. On pleted the task in ¢, =/ And though flying sands have preserved, we must OASES. 307 the highlands streams are ever and anon accelerated by mountain uplift and their valleys changed so fast that with them some topography, once inhabited, exists no more; while the plains, where habitable because of river water, must with that water receive its sediments, so that with them the horizons of antiquity lie buried. Such are the problems to be faced in exploration for the oases of antiquity and interpretation of what remains of their culture depositions. And although so little of what once has been can still be found, there seem great chances in what is left, especially when one year’s exploration has revealed thus much of the remote past. If the caves of Central Asia were occupied by early man as in Europe, still more ancient records may be looked for in them. THE ANCIENT ALAI VALLEY ROUTE FROM BACTRA TO KASHGAR. After communication was established between oases, trade routes and their intersections must have had an influence so important that it becomes necessary to find out all we can about them. Asa contribution, I may be able to give new light on one of the ancient ways between Bactria and China. This will be purely objective evidence from observations on a journey over that route from Hissar to Kashgar. OBJECTIVE CRITERIA OF ANCIENT LONG-USED ROUTES. The question arises, how is it to be decided which of several routes was the important one? Must it remain a purely theoretical discussion, based on maps, traditions and writings, or are there objective criteria by which the traveler may recognize an ancient, long-used route of trade? There is reason to believe that all important trails of antiquity were, where crossing mountains, well-engineered with bridges and embankments, long rock-cuttings in cliffs and declivities, and carefully drained fillings where intersected by gulches. We should expect that steep ascents were, if possible, avoided by zigzagging grades and otherwise eased by long flights of steps hewn in the rock, while certain passes and defiles as well as stopping-places were fortified. Moreover, any trail long used must have worn deep into loess steppes where they were crossed, and even into bare rock. Of all this there should remain a recognizable trace. It is hard for Nature to obliterate long trenches worn in her loess plateaus, and harder still to wear away hoof-worn paths in rock and long cuts hewn midway up high cliffs. Where crossing aggrading plains and deserts of sand, there would be no marks except for caravansaries or fortresses, clay structures that fast crumble to low mounds and are soon buried in drifting dunes or obliterated by the growth of plains. Fortunately the nature of Central Asia has been such that few important routes could have been wholly on aggrading ground or drifting dunes. THE CONTROLLING FACTORS OF TRADE ROUTES. Trade routes are determined by three controlling factors, (1) needs of trade, (2) natural conditions, and (3) attitude of intervening peoples; and not only were they determined by these three factors but they also depended on them 308 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. for existence and shifted or died out at the hand of their change. Needs of exchange were always changing in degree and kind; new centers of trade were ever springing up and newly discovered routes opened in competition. Scores of miles of trail and sometimes whole routes were shifted by the work of rivers and avalanches, and others abandoned for lack of water and decrease of fodder during cycles of desiccation. And history and tradition and Oriental romance have emphasized how robber khans and nomad tribes made raid and others levied toll on passing caravans. It is easy to draw a line on the map, but quite another matter to follow it out in reality with pack-animals that can neither fly over canyons nor live without fodder and water, neither scale mountains nor swim bad rapids, nor pass unknown to watchful enemies; and it is hard enough to make such a risky enterprise profitable by having the right merchandise for the right market. ADVANTAGES OF THE ALAI VALLEY ROUTE. It is no surprise that within historic time there has been much shifting of the main trade-way between far East and West. Whether or not some of Bactria’s early trade passed over the Southern Pamir by way of Tash Kurgan to China, it does not seem likely that the much-spoken-of Terek Davan route to the north was ever of great importance. On it there are eight passes to be crossed within a distance of 250 miles—eight passes, most of them involving a steep rise of over 2,000 feet, and to cross the Terek itself from the north, one must climb 6,000 feet in one day, all in less than 12 miles, 5 of which are up the bowlder-strewn bed of a torrent in which many a pack-horse has stumbled to drown. Of all the passes I crossed during 2,000 miles of travel through those mountains, the Terek Davan stands out as one of the two or three worst, not so much because of its height (13,500 feet, which is perhaps low for passes on those high ranges), but because of bad trail, or rather the utter lack of trail over many miles, especially on its northern side. Moreover, there is no indication of its ever having had a good trail or any trail at all. Beyond Irkeshtam (going westward), where it branches from the Alai valley route, there is on the Terek trail no trace of what we should expect to find on an important pathway of antiquity, no resemblance to one’s expectations of a one- time main way from Bactria to China. Beyond this junction there is no remnant stretch of graded way or cliff-cut, nor hoof-worn path in rock or loess, nor sign that ever caravansary or castle stood upon it. Though with their railroad built to Andijan and the military post at Irkeshtam, the Russians have had to better it as much as possible; it is still a decidedly bad trail. If such a route was used for Bactrian caravans instead of that through the Alai valley, over all its length contrasting in goodness, whole mountains must have heaved and gorges carved since then. From Kashgar to Bactra, via the Alai valley and Hissar, is only 600 miles, all of it good trail, well-engineered, and over 300 miles shorter than via the Terek Davan and Iron Door—3oo miles shorter, or but two-thirds of the distance by the other, and so good that, with a little improvement at a few points, light artillery OASES. 309 could pass all the way. Moreover, there is no doubt about the Alai valley route having been one of long use. The ruins of its caravansaries and fortresses still remain beside its time-worn paths still followed by Afghan caravans and travelers. On it there are but three passes, none of them at all bad; and fully one-third the way lies over steppe, most of the rest along smooth, broad terraces such as those of the Kizil Su. Nowhere is there lack of water, and beyond the borders of Tarim, that all these routes must cross, it lies throughout in a region of remarkably good grazing. Lastly, it is perhaps the easiest route of all to police. The worst stretch is from Kashgar to Irkeshtam, over which it is one with the Terek route that branches off at that point. Through these first five days it leads over a desert of low red mountains, sharply sculptured in a gently rising plateau, as with the Western Bad Lands of the United States. There, as farther on, were seen long hoof-worn trenches in hard sandstone and notches worn 6 and 8 feet deep into ledge ridges crossed by it. At Ming Yole and Ulugchat the traveler still puts up in fortified cereis and at Shur Bulak Pass must ride through the Fig. 471.—A Fortress in the Alai Valley on the Ancient Route from Bactra to Kashgar. battered gate and wall that crosses its defile beside a ruined castle. One day from Irkeshtam leads over the Taun Murun, its last and highest pass, but only 11,200 feet in elevation, easily crossed, and down into the Alai valley, famous for its pasture. From there on for 100 miles this valley opens out a restful stretch for the caravan. In it there is still a ruined fortress, the relatively modern struc- ture at Daraut Kurgan, where a trail branches off into Fergana. Leaving the Alai valley, it continues along the Kizil Su through Karategin, as a well-engineered way where engineering was needed, but most of the way in this region is along broad, smooth terraces. The next ruined fortress is met with at Haui. Others may have been obliterated, but, as with the Alai valley, the valley of Karategin is so isolated, except for the route in question, that it is not likely they had to fortify its caravansaries. From there we enter the lands of ancient Bactria and find the trail worn sometimes 40 feet into loess steppes leading down to old Hissar. What kind of trail it is from Hissar to Bactra must be judged from maps, as I did not follow over that part. On a large-scale Russian map it follows south along the flat bottom of a valley to the Oxus, and beyond the ferry it lies in open country a few miles to the end. 310 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. There is, therefore, reason to believe that at one time the famous way from China to Bactria lay along the Alai valley and past Hissar, which city may have owed its importance to it. Moreover, there is a well-engineered trail from Hissar over the Mura Pass to Samarkand, and another one to Bokhara or Pai-kent, which would make at Hissar a point of intersection of three important routes. This would not interfere with the idea that direct communication between Bactra and Samarkand took place via the Iron Door. Perhaps it varied with the attitude of intervening people. And the Tash Kurgan route over the Southern Pamir may have been used for communication between Southern Bactria and China, while the more important trade of Bactra itself passed direct by way of the Alai valley. Fig. 472.—Ruins of the Tomb of Bibi Khanum (Samarkand). OASES OF THE ZERAFSHAN. RIVER-BANK (TYPE II) OASES OF THE LOWER ZERAFSHAN. Having been a river fed almost exclusively by glaciers for all archeological time, the Zerafshan has necessarily given a fairly constant supply of water—that is, its oases were never affected by the sudden droughts and minor oscillations of precipitation that ever and anon wrought famine to oases depending on streams fed by unconsolidated snow or rain. Everywhere along its lower course and beyond the limits of its now living oases, rise the mound remnants of past civilization. From Paikent to Samarkand is a land no less favored than the long stretch of the Jaxartes where there was an unbroken belt of gardens, of whose houses it is said the roofs were so joined through continuous villages of covered-over lanes that a cat might find his way throughout and never come to ground. Even now, for 200 miles along the Zerafshan it is mostly oasis, though crept upon by intervening deserts, and still stands unparal- leled in Central Asia. Such a gifted land was naturally preyed upon by the plundering hordes that ranged the steppes of Asia from Manchuria to the Caspian, and enters history as a goal of conquering armies. Even now the story-tellers of OASES. jit its bazaars recite the feats of Iskander (Alexander), Genghis Khan, and Timur; and from the foliage of Samarkand, once capital of that last world-conqueror, still rise glittering faience domes and towers, the earthquake-shattered ruins of his colleges and tombs. AFROSIAB. Just to the east of this old city of Timur and on a loess plateau about 70 feet above the big Obu Siob, bounding it north and east, lie the ruins of Afrosiab, still more ancient Samarkand. Tradition tells us this was founded by a Persian, Prince Afrosiab, while some believe it to have been the Maracanda cf Alexander. Its great areas of ruins have crumbled to a barren surface of low mounds with several depressions connected by canals and moats, the remnants of a water-system. Water entered from the south and split into secondary canals, two of which ran just outside the inner walls, those south of its citadel, to supply various neigh- boring basins. The surplus of these two emptied into the Obu Siocb on the north, at a level about 15 feet higher than the Obu Siob water of to-day, as though the canal had deepened 15 feet since these tributaries were abandoned. It was hoped CULTURE DEBRIS ZOFEEr DEEP DRY CANALS VERTICAL AND HORIZONTAL SCALE 2 800 1000 FEET Fig. 473.—Profile of Afrosiab. that some light on the antiquity, and especially on the introduction, of glazed ware might come from a study of gully sections through its culture remains. This work proved difficult, if not impossible, without excavation. In the gullies two habits have conspired against reliable sections; first, creeping down of muddy débris during wet weather; and second, refilling cf narrow parts choked up in various ways, after which reexcavation leaves sections of washed-down débris in which all horizons of the culture-strata are mixed together. Some gullies, 30 feet in depth, are so narrow from top to bottom that when a wall caves in, refilling takes place behind. The large gully running to the Obu Sicb canal opposite the mill widens and deepens downwards with several terraces, and has been artificially dammed across at regular intervals, thus refilling to form cross-terraces that are cultivated. Reexcavation of this valley would leave sections of most unreliable data. But though unreliable in general, there are a few gully sections in Afrosiab that give clean exposures of undisturbed strata. In two or three, through its central plateau of débris there appears to be a total depth of 30 to 35 feet of culture-strata resting on the original loess foundation. Through its northern wall along the Obu Siob cliff a tunneling gully shows the culture débris thinning out to but a few feet in thickness. HIGH-VALLEY (TYPE III) OASES OF THE UPPER ZERAFSHAN. So often conquered and swept by migrating hordes, the lowland oases of the Zerafshan now present a mixture of races, though according to the ethnologist there is still a predominance of Tadjik, excepting perhaps in Bokhara. And this mixture continues some way up into the mountain valley, where for about 16 312 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. miles from its entrance the traveler meets with Usbeg camps of felt kibitkas side by side the Tadjik villages. The last kibitka is passed below Iori, above which the type strengthens into a pure Tadjik; and, as my Tadjik caravan men testified, Fig. 475.—A High Valley Oasis of the Zerafshan Gorge. all speak a good Galcha, a language said by philologists to be a pure remnant of Aryan. ‘They are broad-shouldered, of medium stature, averaging something like 5 feet 9 inches, usually with black hair, bushy black beards, dark eyes, and clean- OASES. 213 cut features of dark complexion, high foreheads, strong chins, and prominent noses. If it were not for their dark complexions, many of them would pass for Europeans. Some are lighter in complexion and some freckled, and one or two were seen with reddish hair and blue eyes. They live in houses grouped into village oases, the high-valley type, from 200 to 600 feet above the river, where the waters of tributary streams may be diverted for irrigation. Standing in contrast to the desolate slopes of barren rock that surround them, these oases with their gardens and apricot orchards and grain-fields are a welcome sight to the traveler who has struggled over miles of the rough trails that wind up their desert gorge. Lying as they do, surrounded by a wilderness of cliffs and bare declivities, each is isolated and self-supporting, separated from the next by miles of dangerous trail often cut as a half-tunnel in the canyon wall hundreds of feet above its river; and there are long rock-hewn flights of steps up which pack-animals must struggle. Sometimes the cliff-cuts were so narrow and low-roofed that our packs had to be taken off and carried by hand. For about a hundred miles above Samarkand all houses are built of sun- burnt brick. They are rectangular in plan and sometimes two-storied, with a courtyard for the horses and stalls on the ground floor; but most of them are Fig. 476.—A Village built of Cobble-stones laid with Sun-dried Brick (Zerafshan Gorge). smaller and only one-storied, about 8 feet high. All have flat roofs of ordinarily 8 inches of clay over brush laid on split saplings and hewn timbers. Proceeding upstream we find occasional courses of cobbles built into the house walls, and the proportion increases as we proceed till in the upper part of the valley we see houses built entirely of cobbles, cemented with clay, while even this cement is lacking in the last two or three villages near its glacier, where many of them are mere squalid huts with rounded corners and brush roofs, usually protected with felt. An important fact about these people is that they have no tradition of arrival in the land, but boast of having been there from the beginning of man. All the old mullahs questioned insisted upon this, and it points to a very ancient Aryan civilization of the valley. For thousands on thousands of years they may have lived there, undisturbed and isolated from the rest of Asia, building up a simple civilization uninterrupted, hardly feeling an echo from the tumultuous struggles that so often destroyed all culture on the plains. 314 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. The mullahs say their forefathers were Christians, but were conquered by a great Mussulman general, Hodja Mussaii Ashari, who came over the Mura Pass from Hissar a thousand years ago. There are comparatively few abandoned culture-mounds in the mountain valley, but many of its oases appear to lie on a considerable thickness of accumu- lated débris; in other words, most of the village sites of antiquity there are still occupied, whereas most of those on the plains have been abandoned. ‘This differ- ence is a good illustration of the characteristic distinctions of high-valley oases, type III, especially the difference of water-supply and degree of exposure to hostile people. Towns on the plains were from time to time abandoned for lack of water as their distributary streams con- tracted because of a general pro- gressive desiccation of Central Asia, and others were destroyed by armies that plundered and passed on, leaving their ruins to the desert. Still others may have lost their water to pirating canals of other oases. Most of the oases of the high valley have always had an excess of water-supply, their size being limited by topog- raphy only, and their inaccessi- bility hasalways been a protection against invasion; one man can guard a trail in the Zerafshan. In many of the towns débris of occupation has accumulated in the form of terraces, in successive steps from 4 to 6 feet high, down slopes of the old alluvial terraces and doubtless extending toadepth of several feet below. The thick- Fig. 477.—Zerafshan Galcha Spinning at Yarum. ness varies from town to town, according to the amount of sediments in the waters drawn upon for irrigation, the proportion of stone used in construction, and the time of occupation. The few abandoned sites observed are in positions relatively more exposed to neigh- boring oases and intersecting routes. Their positions were evidently chosen as the easiest to fortify in their neighborhoods, and, in some cases, seem to have been abandoned for other points nearby that are agriculturally more advantageous. Of abandoned villages there are three of especial interest: One at Iori, one at Urmitan, and one at Kadushar (figs. 481-483). Iori Kurgan (fig. 481) is an old citadel, about 100 feet by 200 feet long, running north and south and resting on gypsum beds rising from the eastern edge of a PLATE 64. tee? ne ROM Kara tepe (eastern) ae Shallpast Kurgan ae ee eee Seepandsh tepe (near Djisak) Kalia tepe (near Djisak ) Kizil-ar tepe (near Djisak) Kurgan near Djisak station Kurgan near Djiseak station Kurgan near Djisak station Wi m eae MTUTT) LTT Ty Remains of walls 10feet high SUUIO UU UP ily, Dshisak = Djisak Plain O to2 feet Tritt VT} CHC 600 feet Plan of Kara Tepe (western) (near Obruchevo station) View of Eeman Tepe. Citadel of Kara Tepe (Western) taken from northwest. Type-forms of Kurgans in Fergana. ~~ Pe “ OASES. B15 tributary delta descending onto the Zerafshan’s great lower terrace. It is a mass of remarkably rich culture remains, about 25 feet thick, and composed of sun- burnt bricks, ashes, and bones, and very much pottery well exposed in pits dug out for fertilizer. Three kinds of pottery were found—two of fine red texture, wheel-turned, of which one was dull and the other polished, the third kind a large, coarse, brown jar. One piece of glass was found 5 feet below the top. Urmitan Kurgan, though small, becomes of interest in its relations to Zeraf- shan terraces and the tributary Vaushan Darya. Standing in an easily fortified position on the southern side of the canyon, it rises somewhat above the level of terrace G, from which it appears to have been partly severed by erosion since it was abandoned (see fig. 482). A portion of the Vaushan Darya’s flood-plain of a higher terrace age, belonging to the ultimate height of alluviation at the close of the Zerafshan’s second cycle of erosion, has now been cut down on both sides, Fig. 478,—Zerafshan Galchas near the Glacier. leaving a remainder standing as a high inclined table at that tributary’s valley mouth. ‘There still remains a shallow channel, once occupied by Vaushan water, leading to the kurgan, but now the Vaushan debouches into the Zerafshan through a canyon in terrace G on the other side of the ancient table. Kodishar Kurgan (fig. 483), or the ruins of ancient Kodishar, is physiographic- ally by far the most interesting abandoned oasis of the valley. Lying on terrace G and just outside the present oasis, it is bounded on two sides by an impassable cliff of the meandering canyon, while round the other two it is bounded by a triple row of moats, ranged one within the other. Altogether its ruins cover about 100,000 square feet, with about 4 feet depth of culture remains, composed of clay- mixed cobbles rich in pottery, both glazed and not glazed, with some glass and iron fragments. ‘Tradition places it over a thousand years old and mullahs say the Zerafshan flowed on a level with it, splitting through its moats then spanned 316 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. by bridges. Besides stone implements other slabs were found to have Arabic inscriptions, but without dates, one of our men being able to read them with some difficulty. Its moats and the native tra- dition that the Zerafshan once flowed through them make Kodi- shar Kurgan of interest. If we grant truth to this tradition, there are two possibilities; water may have stood at this level because of a landslide across the canyon or because the river then had not cut down below terrace G. Although there are remains of landslides that appear to have wrought a comparable change in other por- tions of the valley, no such re- Fig. 479.—Zerafshan Galchas (Gentlemen). mains are found near Kodishar. Perhaps the chances are in favor of a landslide, but it seems barely possible that the Zerafshan, now so actively corrading, has cut down its narrow channel to a depth of 280 feet in say 2,000 years, but that would be 1.5 inches per year. HISSAR. In discussing the natural processes of obliteration, the remarkable height of the citadel of Old Hissar was attributed in part to a mantle of loess protecting it from erosion. It rises to a height of 100 feet or more, in : > . the form of a crouch- ing lion facing east. A part of its highest end (the eastern) is occupied by the palace of the vice- roy (koshbegee), while the rest is bare except for hisstables. Culture-strata of loess mixed with pottery, bones, and charcoal are exposed to a depth of 20 feet in a pit on its western half. It is possible that a portion of its roo-feet thickness above ground is composed of loess deposited during periods of abandonment, 7 Fig. 480.—Zerafshan Galcha with his Plow. OASES. 24 but that would not detract from the antiquity of its deep layers. The base of culture may be at a considerable depth beneath the plain. Hissar was probably a city of importance when Bactrian trade with China came past it via the Alai valley. Suitable presents to the viceroy would make it possible to dig there, and labor is only 35 cents a day (fig. 484). There are many small kurgans and remains of ancient fortifications, canals, and dikes in the Hissar valley. ABANDONED OASES OF FERGANA. A rapid reconnaissance of chances for excavation in Fergana was made by. stopping at intervals along the railroad in a private car, lent through the kindness of General Ussakovsky and General Ulianin. Near Osh there are two small mounds with no surface indication of culture remains. No others were seen in that region and its natives say there are none, There is a kurgan at Marghelan, but west of there it is not until the railroad strikes Fig. 481.—loni Kurgan. loess steppe near the kurgan Karaul Tepe, between Balyakandose and Rojevat, that kurgans are seen from it. As Kavast junction with the Tashkent Railroad is approached they become numerous, and from there on to Samarkand abound, because the area is all loess. The following sketch outlines are typical ones selected from a larger number. As their forms are in general rounded, it would seem that few of these kurgans are not so recent as those for instance on the Murg-ab delta; and it seems possible, since it is a region of uptilted piedmonts, there may be examples of great antiquity among them, though only one (Kara Tepe, western) of those examined was free from glazed ware on the surface. Four of these—Kara Tepe (western), Eeman Tepe, and two river-cut mounds at Millitinskaya—deserve special mention; the first two because of their great size, the second two (fig. 485) because of their physiography. KARA TEPE (WESTERN). Kara Tepe (western) lies a few hundred feet south of the railroad, a little over 1.5 miles west of Obruchevo Station. Its citadel rises over 70 feet above its northern base and about 65 feet above the general level of the plain, and stands in 318 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. part surrounded by a broad moat between it and a crescent-shaped plateau of culture débris left open on thesouth. The long embankment, 1o feet high, appears to be the remains of a wall built late in the city’s history when it had expanded to that extent. In size these ruins approach those of the Merv delta, while their older portions are much older in appearance. The only pottery found in small exploration pits to a depth of 2 or 3 feet on top of the citadel was wheel-turned, red, vesicular black, and light-gray, while similar pits on the plateau west exposed human skulls. Here may be a column of records down through the Greco-Bactrian age into horizons contemporary with Anau’s South Kurgan culture. EEMAN TEPE. SEES | Heman Tepe, about half a mile | southeast of Dshisak Station, stands ) 65 feet out of the plain as a citadel of special importance to us, for, though of an ancient round-worn form, it rises into glazed-ware time. Its culture was explored with a few shallow exploration pits on top and has been exposed in terraces, cut in near its base, where natives have taken débris, presumably for fertilizer. In these were found two or three specimens of glazed ware and glass and much red and gray wheel-turned pottery, some of it withinciseddesigns. It isamound of rich culture, abounding in bones, ashes, and hearths. Here may bea column through the period of which so little is known, that between Mohammedan and early Greco- Bactrian times, perhaps overlap- Fig. 482.—Urmitan Kurgan. ping part of Kara Tepe (western). RIVER-CUT MOUNDS OF MILLITINSKAYA. The valley of Djillan-ooti Darya, otherwise known as Timur’s Gaie, is a remarkable example of the hydrographic complications brought about by uplift of mountains, whose round-worn, outlying, terminal spurs were well-nigh buried in waste. Before its uplift alluviation from the Zerafshan appears to have over- flowed, or nearly overflowed, the worn-down Millitinskaya spur and possibly coalesced with the Syr Darya plains, so that the Djillan-ooti Darya found its way out to the Syr Darya. Uplift resulted in the Djillan-ooti excavation of Timur’s Gate, widened during quiescence to a flood-plain about two-thirds of a mile wide. Even after this the ancients led Zerafshan water from Pendyakent through an old OASES. 319 canal-way to the Djillan-ooti, and it is thought the Zerafshan might still be diverted north to join the Syr. If so, Bokhara lies at the mercy of the rulers of Samarkand. When the two now river-cut mounds were built the Djillan-ooti was alluviating its whole flood-plain with a sluggish flow of water carrying fine silt and spreading through a dense mass of rushes on either side of a gravel-strewn channel. Sescce2e Socce TOY HH NHB \ \\ y Chi — 6.0 + of cul- ture debris of cobble stone / a " Y] il) " My ny Matt! Miele -—\ Za Much eocos Seeess SEs Scocos”? ef Soesce Section online A B. re) 200 400 FEET Vertical = horizontal \ iy i \ Hy) Ath vy 74\\ = LY tipi TAHT if, My ASS {fia Se mes wa S SSS pero AY. FSS 2 | \ WINS SS //| I} | il } 7 / ee {i LINWSS LA elaydl f \ GES a S \ NNT MNS S aS AAU IS oy) INS SY {aint ; * ie petigae’ vale i . i 2eeines ye ow ve. f S50% tia a , ia (3 4 age yorar we +s — 4 fad | ' —— aw a | as 2 ‘4 A id : a=, tet ' r o 4 Ve Rene ee aon ‘ = oe re : 4 * - 2 =< é ayets epee sa lal 9 : ; eerie: ‘ - ‘ yr re iy 7 + @e > = Phe, = pil ‘ =? _ } Tip ets TS dies re A tthe Dee Sa = ev - : ey AAG | ~~ qel! Lal vy amen) “7 a ; pare * / ; ; Pa MSFT NM Wb hos © re ee fa 120 On he ‘ 4 iy ee s seal i yy wy 7 ill ad some 5 \ Me Bes : , ee 2 hee t \ 3 2 “ , : } Bs ey = ty Bee ee, = aan : [Pane Ae ote) v2 | ay ' . | » : J ¢ ~ te : ! i 4 - . a en ee et . } : + . , j et ij . . | =i “ = ‘ : oy a] TURKES TAN ed 5 Ss i i Shaft E : Sheft B SORT > Stratified sand = aA pacecieciv ces S.K. Shaft B — 30 Irrigation LAN Lumpy loam is, Sand Lumpy loam Sandy ey with irregular rare layers of fine cross-bedding Unstratified sandy clay Hard unstratifiéd sandy clay === Yellow, gray and red banded - Clays, and sandy clay =x Gravel and sand Homogeneous Laminated clays Compact finely banded and laminated gray, white, and light brown clays Clays, loess Loess and clay Loess -20 S20 20 Shaft D 7 7 Starts 42 ft. below S.K.datum Z S.K.Shaft Fo Gray unlaminated sandy clay Clay cross-bed striking wavy lamination Clays and sandy clays with disturbe stratification Fine light yellow clay Pure Very compact yellow gray clays Kurgan wash Cornpact clays, very compact light gray clays, irregularly banded Brownish cross-bedded silica Banded clays, streaked with sand Interbedded streak of gravel Sands and clays St. compact Sandy clays and sands Light buff Fenn i a OF About //00 feet KEY PROFILE 0 50 roo 150 FEET {a ae ee Se ee Hl) Datum of $.K. Loose earth with little pottery Sandy, unstratified with some pottery Phe PLATE 66 Culture strata -10 Base of irrigation ? == Banded red and gray clays, _Base of ee re No.l pier aoe ee == —— Pure banded green ——-—-—-= and gray clays Sandy clay, gray oe PERS sendy ceyeeey. =15 > Freeredpottery 2 ees lays = = Stratified sand zs y coarse sand Pale brown } cigs und Sern —20 = Pure clay not stratified Li bi ; = ene orenh Compact fine clays with Rich culture disturbed banding and lamination, containin Pee Traces of some pottery : 25 Iture No.2 charcoal and Se : Slightly sandy pottery Light yellow Tati : : and gray Yellowish banded clays Laminated light yellow clays, loose Light yellow and gray banded clays Looseand fine clay banded Loess Clay banded —38/r° SHAHTS Kurgan wash,with pottery of S.K.older culture Banded clays ed Light red, yellow, gray, compact banded clays Saom-anr= Soft banded clays manip ileg adh immer PLATES Ov 190 Ff. WEST OF N.K B.M.Go epee ae a Lev: oT N.K. datum poe 0.3 ft - a < la} to) —- Zz Vo natural size Fine silica’ Bie tei eet ght brown - Saft fines vee. Da ere caer white clay 5 ES ; 1 Light brown cross-bedded silica 2 Brown sand Irrigation Pottery rare 3 Pure white clay 4 Pure light buff colored clay with surface-weath ered ...07 + | Base of irrigation 342 Wash-gravel andsand with pottery Horizontal section of clays andsands ~ ** in. NaK.W:S: 1, at — 35.2 Hearth Hearth Vertical section in N.K.W.S.1 showing weathered surface at —35 vu —10 5 2 a w Vv = ts Bu eo -35Ft =15.F » Op aw ve Vv Lo S20 ec Ey 30 oc 2 [pe =< 20 Ca g ag ) = a of RATE o = s " | = White sate Hearth Base of culture rs a9 Su eS oq .— oo ca KEY PROFILE Soe morte Crossbedded sand a =30 seen ee Crossbeddéd sand SAP SCE Horizontal section of clay- banded Ped ee loess at -18 ft.in N.K.W.S.2 Parided' clays N.K. datum : 2%) Very fine light brown Saks) crossbedded sand Sands and clays Compact light a : gray clays Homogeneous mass Sie Cine Butt I yrypypSe| Cross-bedded silica of sandy clay sabes Nini VV ccs allie —5 Irrigation =e : {Buff cross-bedded cla o porceer ets as x —20 poe ee with pottery, bones,charccal i 2 atene 2) y More sandy, <—| Pure green clay A aeeaee Se ne ee ween ars = mip s lenses of pure Clay el fe) - 2 C , . ‘ iF: 8 Laminated sand =22 Pic ashi gees = Silicious gray clay o Lamin.blue cla structure not deter. OS 2 Irrigation Crossbed send sere) Charcoal iy a2) Coreen =23 Lamin.gray Clay 4 Saae v OS x Base of irrigation Lamin blue clay. j=ss====> = -15 Rl -e====—-—=— —24 Lamin.blue and) p= === Cross-bedded sand N a Eoess (2) gray clays © oe Fine loose clay o5 i a a a pEeess!) Sper bell cole er potter =n aint anded = -20 adsociated with ie oes Irregularly banded, 26 ere dulerty charcoal SS GSS == very compact Upper blue- gray clays es ‘potter y : CU : See a ae === Same hard clay.Light gray ~- O - Ss AG) Resembles cross bedded sand 28 = oe lower pottery == = y E Buff color silicious Loe =25 =-Very compact clay, pure = clays with faint i~3B -=Brecciated blue-gray irregular banding Clay aj ==s-=-- = pune Clays ~30 Fine. clay ——— ——-—— = — ~~ —— _—__ — —— Valley run- m*5==-=Buff colored silicious clays loess N.K.W.S.2. Detail at — 15% ning west NW ris == =.with faint irregular bands esate See Vertical section of interbedded 30 clay and loess(flood plain) Swe eae suens and tran- sitions from one to the other showing cuts running N.65°W. North side on left x =pot sherd N.K.W.S. 2. Anau Shafts. PLATE 68. ANAU CITY Mosque: shaft 2 Graveyard shaft. Horizontal scale O 100 200 300 400 500 FEET — + tt Vertical scale (e) io 20 30040 50 FEET B.M.+17.35 TTI R 777 KWY7) B.M, +30.8 Datum of N.K. 4 Irrigation culture | } ! ; Pott. ; ee aah | Evident ottery, bones irrigation irrigation \e ace & (B culture and charcoal L 2H o> Sandy clay = o Pure clay te s a Banded irregularly “2 es anded irregular f Le aan ey low glass be d4_ _Datumof N.K. i> i) . . . o Gentimetratiiike Ss : pee oe itetion (Garden culture) e ray sandy clay, irri, mar 4t*== Pure clay-irregular rs) grays y Clay, irri Datunn oo Net FS rot——= brecciated banding Ss | ee Pure clay finely a Fee Coarse sand “e¢ ee aminaved -S- _Level of irrigation ae E FEES Coarse sand Sass = outside of walls Gray sandy clay, irri. = === Fine light yellow clay Se Te == == == Reddish =======-== Fine loess, yellow, c re Reecd oats Soe i | See oe = are Sir See Wins ciny === with layers of — Mosque shaft 2 See F x Base of irrigation PSS ZiS finely laminated = Clays and sandy clays === === pure clay S aes Concretions veers === === Stratified clays Pee |____SSS Mosque shaft Graveyard shaft Just east of Citadel. Datum ——+t-—~ Baseofculture At station 2485 east of N. K. === Laminated clays on profile to Citadel. (ae ee Mosque shaft 3 Anau Shafts, es nas yey - Tape, he n eede ee PLATE 69. Level of stream bed“2e9e~ t 2 Canal Irrigation ,Sediments Section of irrigation canal Bearings or Mosque, N.20£, Bearings on S.K., N,37°W = ct) Irrigation | = Ny | TTT ee — + Section of same 200 4yards N Cana! sediments SS ET tia Cross-bedded sands and clays On Mosque canal above Mosque shafts Bearings to Mosque N,/5E. Bearings to S k,N.72°W. Irrigation” Very soft structure indefinite Faintly banded clay sand Clay and sand, soft [282] Pure sand ere} Grit and gravel St.sands and clays Sand Mosque shaft 6 About 200 yards south, upstream from road Irrigation, bones, glass and pottery Cross-bedded and banded fine silica ee Clay and sand Irrigation canal deposits on Mosque stream ee : Detail of LEZ cross-bedding Fine silica FIVE SECTIONS OF THE MAIN CANAL GULLY Loose _un- stratified clay loam, with pottery ibones,and stones Pottery and stones ~ Hearth ele vel Cf) ae ae Datum 1 @ plain D=5 ae Clay cassetéte fragment of Metate x “Pottery R.W.P Well in Anau Valley =O -10-- eet ee Stone hammer and fragments of mealing stone Pottery of S.K.upper culture in pocket of loose earth = Coarse sand, occasional pebbles Mound 1 shaft Surface of mound /2 feet £ above S.K.0.0. Bearings on Mosque N.69°E. Surface Bearings on S.K. N./7°E, Hard loam, sandy Irrigation VERTICAL SCALE Oo 5 ite) SSS — dt FEET, Soft clay loam Clay, no grit r= Dehad io. Hearth and pottery Sand and gravel Hard, very pure banded clay Hard, very pure 30-- 35-—+ 40--4 20--- = Nearly homogeneous mass of clay, mixed loess { Pure loess diameters from 3o™M to 2&5 mm, Streak of pebbles Loess and clays Gravel i {om particles of =: Fine loess streaked with layers of clay Red = Banded clays{ Yellow “oO wOlO”e Fine loess,banded with fine clays ; Gravel Gravel Cross-bedded loess Cross- bedded loess Homogeneous loess (red, brown) Clay banded - Gray streaks Homogeneous loess Angular gravel of sizes from gritto4inch cobbles, filled with brown sand Sand Sand with lumps of clay Sandy clay with bones blue clay Se Pottery, red band ’einch Hard, very pure clay streaked with red Bandy clay Hard laminated clay, stratified Red st Saal = Red st. Sesksoee 4 inches of angular gravel Hard, light yellow, pure clay, with no visible stratification Red layer, sandy, | inch thick Stratified clay streaked with sand Middle well shaft JE0 yds. south, upstream from southern end of N. K. Anau Shafts. Koee Boll Pure clay, Mostly si J ‘ y silica faintly particles of banded ayo diameters = Clay Fine pure sand (Silica particles of diameters from i56mm.to 4mm.) & . “ i—_ , ) . a ¢ ~ a hf ~ a 7 t , a a , . 6 7 ae . $ Ly : J OASES. aso More careful observation shows the plain, at first so simple in appearance, to be of complex form. Its conical convexity, indicated on the map by radiation of distributaries, may be demonstrated by watching a rider who crosses its lines dis- appear as does a ship at sea, whereas one passing straight down the slope will slowly fade into the heat waves of mirage. Our map shows it to be a character- istic subaerial delta with the general outlines of a fan, encroached upon by dunes from the north. Except for about 2 miles from its apex, it is everywhere bounded by dunes, without which its radius would be over 10 miles, as prolongations of bare clay still reach that far into the desert. At present outlying sandhills stand 4 miles north of the apex, while its greatest width is but 5 miles. Careful study of the surface proves it by no means that of an even cone. It is everywhere broken with irregularities wrought by man; canals long since abandoned, mounds, and roadways, and most significant are its areas of many hundred acres several feet above the general surface, a difference caused by man’s control of alluviation, concentrating the stream with its depositions into limited areas of cultivation. These areas of concentrated deposition are bounded on the lower side with long bluffs varying up to 4 feet in height and of irregular plan, as shown on the map (fig. 486 and plate 65). For more complete explanation of these zones of concentrated deposition we may look to its present distribution. Except during exceptional flood, all the water of Anau Su is led into a system of canals irrigating fields with low dams on their down-slope sides. An irrigated area thus comprises a system of fields bounded round the lower side by an irregular, often more or less crescent-shaped, dam and merges above into the plain. This dam or dike may be only a foot or two in height, but it is always easier to rebuild or patch up the old one than to make it in a new place,so that a permanent barrier to deposition may exist for centuries on the lower border of irrigation; and since the whole stream is consumed in these areas, its depositions are concentrated therein and accumulate in the form of what we may term “‘irrigation terraces.’’ All phases of this process may be observed in the Turkoman grain-fields of to-day. We thus have an ancient delta surface surmounted by irrigation sediments concentrated into terraces near its apex. If man, their controlling factor, for any reason abandons them to carry on his agricul- ture elsewhere, they show their instability with the first flood; water-gates burst and dikes are rent by the stream thus set free to rush over terraces, falling down bluffs and gullying back. In the course of a score or so of years this channel will be carved to base-level and the excavated terrace material lie spread over the delta beyond. Pronounced irrigation terraces result only where a fixed area has been con- tinuously irrigated for a long time, and so the outlying, more erratically cultivated areas, comprising a large portion of the Anau delta, have aggraded in a less differ- entiated or more uniform way. The non-observant might cross such a terraced plain with never a doubt as to its uniformity. A sloping and often round-worn bluff, only 2 or 3 feet high and irregular in course, running perhaps a half mile and fading at either end into the plain, does not ordinarily arrest the eye nor does such a slight difference of level between two wide adjacent areas. To the 326 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. trained eye there are pronounced terraces and others which are distinguishable by careful sighting only. Some slight accumulations have resulted from Turko- man irrigation, while large areas formerly irrigated by the Anau-li and abandoned only 50 years ago are now dissected with gullies 15 feet deep and over that in width, and other areas of Anau-li fields are still growing under Turkoman irriga- tion. Asa whole the irrigation deposits of our Anau delta lie within an area spread out from its apex to irregular limits near the railroad. In its central portion this accumulation, exposed by a gully down into its underlying floor of natural sedi- ments, attains a thickness of 15 to 20 feet, but from there outwards terraces down over broad fields with 2-foot drops. Plan HH) i) MH) IX 97 /!) Tif) Wy Wy WWM, YY ee [I / I fi Wi HN) bh ) —Y IH I Wy [ill {| ——— oe — SS —_Terrace 3= 3 Terrace £-—1 | aaa ts E = =3 a — ee | | | ii | i) —— eo —— = acd Ve = ~ & anil = ip ATT IA OMT aren a ————— arre = errcee Ss ——————— a. ——— ae eee — a at ————— a- ae, 8 = eae gee. er Se - — | (EP ee A —-- AS: ee he | oe meeoees eee fica, \ My = Pr aks ae = en Section Fig. 486.—An Abandoned System of Irrigation Terraces. The old through route of East and West crossed the middle of the delta and so determined the boundaries of Anau’s ancient fields along it that a bluff from 1 to 4 feet high of irrigation limits follows much of the way along its southern side. Other irregularities are found in shallow rounded troughs, whose heading branches fade into the plain obliterated by more recent irrigation. ‘These are old canal- ways, bounded once by fields aggrading on either side while they remained unsilted till abandoned, now slowly losing shape. Each kurgan and the citadel has split the irrigation stuff around it, thus in part protecting an area directly north or OASES. B27 below from deposition and leaving it somewhat depressed. These are a few of the topographical variations wrought in time by human occupation of an aggrad- ing delta. THE ANCIENT AND THE MODERN DUNES OF KARA KUM AND INTERBEDDING OF THE DELTA MARGINS WITH THEM. Riding north from Anau, one passes from bare clay on to gently rolling sand- hills, ancient dunes that have long since lost their barkhan shape and now appear to be of great antiquity. For a few days in spring these are green with grass, soon withered brown by the arid sun. So old are these now fossil dunes that their firm, cross-bedded sand stands wind-carved in vertical and overhanging bluffs, while all around are seen resistant holes deep-burrowed by desert turtles, lizards, and hyenas. Over them lie fresh barkhans of sand, now drifting from the ever-shifting waves of Kara Kum beyond. One might expect to meet with naught but dunes in such a wind-built desert land, but far out among them lie small areas of smooth AES RT 4 epee Po tf Fig. 487.—A Canal Gully in the Abandoned Irrigation Terraces of Anau. flat clay, still bare—portions of the delta isolated from the rest at different times long past. And as they vary greatly in relation to each other, to the mother delta and to the dunes, these areas of bare clay become of interest. In some instances two plains separated by only a narrow ridge of dunes differ several feet in level. Thus we have preserved in open air the ancient delta surface, various horizons of antiquity escaped from burial. How old these more ancient dunes around Anau are may be conjectured from a section exposed by our shaft sunk through the ruins of the sand-buried oasis near Ball Kuwi (see plate 69). There culture rests on hard, extremely fine, lami- nated clay, light brown, the contact being practically on a level with the takir plain just north and doubtless a continuation of its horizon. This clay bed is 4.5 feet thick and rests on 1 foot of dune-sand, below which lies another sheet of clay 0.5 foot thick. From there down our shaft continued in dune-sand, cross-bedded on a large scale and so loose that it was unsafe to go deeper, and how far it is to the next clay layer we know not, but imagine that for great depth the structure would be large masses of dune-sand interbedded with clay. What we did suffices to show that here was an area where an aggrading delta surface was from time to 328 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. time encroached upon by dunes which it rose to bury. We have a value of 17 feet for the growth of the plain since it had risen to the level of the bottom of our shaft; a minimum value for the antiquity of the sand, because it is a long time since the last alluviation of this area. or at least many thousand years it must have been a region of outlying dunes. ANALYSIS OF SHAFT SECTIONS OF THE DEPOSITS FROM MAN, WATER, AND WIND. (See Plates 66-69.) The moment that deeper layers are explored a host of problems arise. In our shafts, ranged on curved lines on plate 2, and on the profiles on plates 66-68, we penetrated four distinct kinds of growth, that of man’s débris, that of his irriga- tion, that of natural alluvium, that of loess, and adding to them the flying sands of outlying dunes we find our plain is built of five divisions of deposition under the forces of man, water, and wind. It will be observed that there is a perfect grada- tion between the different divisions and subdivisions. ( Kurgan culture: Slow growth of clay débris rich in pottery, bones, and charcoal. Walled culture (that of fortified citadels): More rapid growth of clay débris less rich in pottery, bones, and charcoal. Garden culture: Growth of irrigated fields near a city; contains some artefacts. Undifferentiated sandy clay deposited under continuous irri- gation of annually plowed fields. Interbedded irrigation and natural sediments (the result of culti- vating a naturally flooded area at rare intervals), Series 3: Laminated clays, sandy clays, sand and gravel. Series 2: Pure, hard, banded clays and beds of angular gravel with grit. Series 1: Interbedded loess, blown sand, and alluvium. Soft, velvety clays, clay-banded loess, grit and gravel. Fresh dunes of pure loose sand (drifting). Dune-sand interbedded with the delta margins. Ancient dunes of firm sand mixed with loess particles (fixed). Loess mixed with wind-blown sand, Loess, pure homogeneous, of vertical cleavage, with small gypsum L L erystals. ( Culture débris......... | Manin. J Irrigation sediments. . . Natural sediments Water.... (alluvial igs eee: Wind-swept flood-plain deposits 4.2 pee ee Dune-sand:..7...:..<.. ~— Windia. ——— — — pee sees eh tt ee ee te Over 150 hours were spent underground in sketching the shaft sections here reproduced, and, in addition, fully half that amount of time was given to the study and comparison of samples taken out. Each shaft was scaled all the way down with levels scratched from a tape line, its layers cut clean by a long knife and studied with the light of an acetylene lamp, and characteristic samples taken out in solid blocks up to a foot high. Afterwards the material in these samples was studied under a microscope. Culture remains are easily distinguished by their pottery, bones, charcoal, clay bricks, and various artefacts. In amount they vary from isolated artefacts found as fossils in wind and water deposits to the massive accumulations left by towns. It was found that such remains when in situ of original deposition were invariably associated with bits of charcoal, whereas those afterwards shifted by water or wind, or gravity alone, are nearly always utterly without charcoal; a truth explained by the fact that a material so light as charcoal is inevitably borne far beyond its heavier associates, such as potsherds or bones. OASES. 329 Irrigation sediments are of course far more widespread than culture, while between them there is a gradation which may be termed garden culture—sedi- ments arising from the concentrated irrigation of gardens in and near a city and thus more rapid in growth than those of ordinary fields and containing many artefacts. These are all well exposed in gullies near the citadel of Anau. Irriga- tion stuff proper is at first hard to distinguish from natural alluvium, its strati- graphically differentiated equivalent, and even after long experience utmost care must be given to their separation. If all fields of cultivation had been continuously irrigated till abandoned, there would be no trouble, but this was not always so. In the graveyard shaft, we see how an area, after having aggraded g feet under irrigation, was abandoned to natural forces long enough for 2.5 feet of laminated clays to accumulate, after which it was again cultivated during the upper 11 feet of growth. It took about 1o hours’ hard work to make sure how deep irrigation was in that shaft. No stratification can result on a cultivated area unless it be aban- doned long enough for natural sediments to accumulate a greater thickness than is disturbed by subsequent hoeing or plowing, which is 4 inches and more. At its base irrigation stuff is often found containing fragments of natural sediment, below which are remnants of the original plow or hoe trenches. A characteristic mass of irrigation stuff contains all sediments utterly undifferentiated except for the gravel and coarser grit concentrated here and there in bottoms of canals from time to time abandoned and buried by irrigation through new canals. It is thusa homogeneous mass of sandy clay. The limits of such accumulations, both ancient and modern, have been described. Though pre in thickness from Io to 25, the average appears to be 15 feet. Natural alluvium directly underlies irrigation sediments. Of that pene- trated by our shafts there appear to be three epochs of growth, differing in struc- ture and kind and separated by two erosion intervals as indicated in sections of unconformity. During the first of these, our delta appears to have been a wind- swept flood-plain, which through inequalities of growth from time to time gave rise to shifting grassy areas left isolated from alluviation for so long that wind- blown material accumulated on them in various degree. Such appears to have been the first epoch state with its resulting interbedded gravels, grit and clays and homogeneous loess, which were penetrated with two shafts at the North Kurgan and two at the South. Some change took place and the delta was divided by a valley, how deep we do not know. Then began our second-epoch growth of pure, hard, laminated clays interstratified with beds of gravel. Probably ere the North Kurgan was founded this new epoch had aggraded its valley flood-plain and refilled its delta valley to within about 8 feet of the delta’s old first-epoch surface, as traces of culture are found in all shafts down to this horizon, but nowhere below. This growth was of pure, hard clay, banded and finely laminated; blue when wet and yellowish when dry; it appears to bottom on a basal bed of semi- angular gravel, and to have ultimately risen nearly high enough to overflow the whole delta; that is, till the delta valley or channel was filled practically flush. 330 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. A second reexcavation of the valley was followed by its third-epoch growth filling the last channel, and this growth was in process when Anau city was founded and the stream was probably wholly used for irrigation, and so the growth has continued till now in the form of irrigation sediments, raising the delta plain around our kurgans to a height of 20 feet above its ancient first-epoch surface on which the two kurgan oases were founded. In it there is more sand than in any of the preceding epochs and it is likewise more rapid in appearance. SUMMARY OF THE MORE IMPORTANT FACTS OF THE PAST PHYSIOGRAPHY OF ANAU. First cycle. Pliocene, Kopet Dagh worn to low relief. Immense deltas built. Second and third cycles (Quaternary). Two uplifts of Kopet Dagh and corresponding general sinking of plains, while the piedmont deltas of the first cycle are bent up along the base of the mountains. Valleys excavated and prolonged through old uptilted deltas to build new, smaller ones beyond. Fourth cycle (Postglacial). Mountains continue rising again and the new delta at Anau continues tilting, but so slowly that with sufficient alluviation it grows. Reaction to dry and formation of delta valley. Reaction to wet and alluviation of delta valley. North Kurgan founded. During oscillations in precipitation over the mountains, alluviation twice again falls below rate of tilt and our buried delta valley thus twice again comes to be. Precipitation now scarcely in excess of underground drainage. Dunes desiccated of grass and set free to encroach on the delta plains. OASES OF THE MURG-AB DELTA. THE RIVER MURG-AB AND THE TYPE-PECULIARITIES OF THE MERV OASES, PAST AND PRESENT. Rising in the mountains of Afghanistan and swollen by their melting snow and ice, the River Murg-ab issues north onto the plains of Transcaspia. There it splits over a large subaerial delta 100 miles from the mountains and inclosed by dunes of sand wind-sifted from its river silt. This far-expanded sea of drifting barkhans merges into the desert of Kara Kum, making the great delta a seat of oases extreme in desert isolation. At one time the Murg-ab may have joined the Oxus, and in still more remote antiquity may have flowed direct into an Aralo- Caspian Sea. Its chief oases may, therefore, in the past have changed through type V, type II, and finally into type Ia; from lake shore to river bank, and at last the oases of an isolated delta. From the time when it failed to join either Oxus or Tedjend or reach a sea, the Murg-ab flowed to build an independent delta, and its oases were delta-oases, which, in the course of centuries, moved with the delta out and back, or perhaps always back if the river shortened with continuous desiccation. Here and there upon the clay surface, and beyond the limits now attained by water or even where the sand is drifting in, are seen mounds of clay and crumbled walls, the ruins of ancient towns and cities. OASES. Su Prosperity of oases is wrought by two great factors, water-supply and geo- graphical position. With the rivers of Central Asia, Murg-ab ranks fourth in volume and, unlike the Zerafshan, which feeds a myriad of oases along its course above ere nourishing the last, Murg-ab water flows to spread in full among the gardens of its delta. In it we find a river whose life-giving has been concentrated into one veined mass of gardens safely housed far out amid the dunes. The cities of Merv have thus long been favored with abundant water combined with sound protection by its wide surrounding desert. Alexander found it populous and full of wealth, and so did the Persians and afterwards the Arabs. And once 4» Odontche Tepe: oe a5 ae Siecieaely here oS. f Sultan Kala ;, ¢ one Qo Ghiaur Kala : ° Bairaim-Ali Ne [3] Kurgan = P [Pees] High sandhills: | Low sand hills > Scale of miles 10 20 Fig. 488.—Map of the Murg-ab Delta. The Oasis of Merv. the caravans had opened out their routes across the sand, Merv ranked as one of the world’s great trade centers. Lying midway between the Caspian Sea and Roof of the World, it centers that whole region of ancient oases. By caravan from Meshed to Merv is less than 200 miles, from Khiva 300, from Paikent or Bokhara 190, from Samarkand 300, and from Balkh 275, measured along the route. In their peculiar state as oases of type Ia, those of Merv must have been especially sensitive to climatic change. To begin with, we must believe that 332 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. a sufficient general increase of precipitation or even some change that would give rise to a larger proportion over the plains would make the desert into grass and thus break down the protective barrier. Also a large volume of water would penetrate farther out in the desert, a less volume less far, so that the delta with its oases, ever since there were any, must have varied in distance from the moun- tains, shifting out or back; responding to all greater cycles of climatic change. It is also evident that a river whose grade is for roo miles that of a vast silt-made plain must have been extremely sensitive to any warping or tilting of its channel. But even greater must have been the changes wrought by floods, the ever-shifting of distributary systems or even bursting of the whole river out to build elsewhere anew. If we could look back through foreshortened geologic time, the Murg-ab would appear in course and kind fast changing, a river living through a marvelous variety; we might first see it flowing to the ancient Aralo-Caspian, and as that sea is cloven into shrinking remnants, and rivers wandering free-ended join, we see the Murg-ab now with the Tedjend, now with the Oxus, then shrunken alone and ever shifting, with meanders made to break into new straightness; a silt-laden flow that coils to burst and glide in some new wandering way; a river which with its season’s flood may spread rare water in a wide sheet far out among the dunes and from that flood subside into new channels; for millenniums be led far to one side, leaving what was garden so by chance transformed to desert. Thus were the oases of Merv controlled by Nature’s ways and, though man could not prevent the effect of long-changed climate or much alter that of serious crustal movement, if it happened, the capricious behavior of delta distributaries used by him so stimulated his ingenuity that in time he got them under control. The Murg-ab with her silting distributaries proved a costly school, but graduated engineers whose works—canals, barrages, water-gates, and meters—were a marvel to antiquity. An oasis so bountifully favored, and whose civilization was so stimulated by trade relations and the natural education forced upon its type, as well as the protective isolation of that type, bore a populous and wealthy growth with varied culture; a growth that always recuperated rapidly after falling to the power of such organized attacks as those of Alexander, Genghis Khan, and Timur. The present outlines of bare clay upon the Murg-ab delta are irregularly pronged and have the aspect of a change or shrinkage of alluviating area, upon whose abandoned parts sand-dunes are drifting. In a general way it resembles a long leaf about 30 miles across, with two prongs—the left-hand one longest, and main axis pointing northwest to follow a general slope of the Turkoman Trough towards the Caspian. The convexity of alluviation is beautifully emphasized by its radiation of distributary veining and indicates a permanence of the present position of the delta for many centuries. At present it is impossible to indicate the whereabouts of Murg-ab’s most ancient oascs. Knowing that for some thousand years all Central Asia has been undergoing desiccation, our first thought is to look north beyond the limits now attained by water. There is, however, no reason to doubt that a climate even OASES. Bo0 more arid than the present may have existed in still earlier time; indeed, we have seen that throughout the Pamir and its border ranges glaciers had receded back of where they are now, apparently having been subjected to a reaction of extreme dryness after the glacial period came to an end. The Murg-ab delta was then very likely south of its present position; and although the delta is the most favored spot, it is possible that some town of importance may have been upstream. ‘This, however, appears unlikely, for by tilt or warp its channel has long been cut beneath the plain and dunes have drifted to its banks where no water can be led now. It seems most likely that, if there were oases of importance to the south, they were of its delta there, and now cut through by its channel, since then prolonged. We have seen what a various existence the Murg-ab must have led through the long past, and shall therefore refrain from too much conjecturing as to the exact whereabouts of its more ancient oases. IRREGULARITIES OF THE DELTA SURFACE WROUGHT BY MAN'S DEBRIS OF OCCUPATION AND HIS CONTROL OF ALLUVIAL DEPOSITIONS. Turning to the ruins found upon the present delta and beyond, we find a field of great interest. Nowhere else in all Central Asia are ruins so abundant or so vast. In preservation they rank from Bairam Ali’s state of brick-robbed walls and still-standing battlements, with gates and inner streets that may yet be ridden through, to the round-worn mounds of far more ancient cities. In size they rank from mounds that count square miles and rise as platforms of stratified débris, one to five score feet above the plain, to low clay heaps that mark the ruins of past monuments and tombs. In all there are perhaps more than a hundred traceable towns and cities, some as much as 20 miles beyond the gardens of to-day; but for the most part they lie so far out on barren clay that only shepherd Turkomans know of them or wander among their heaps. Only Bairam Ali, Sultan Kala, and Ghiaur Kala are much visited or dug into by treasure hunters. They lie within the reach of cultivation, and through Ghiaur Kala’s outer walls, now trenched, water is led to irrigate a wide depression, which may once have been a market-place. Round the ruins of these three cities and Iskendar Kala the native romance dwells. Nearly a thousand square miles of the Murg-ab’s delta are still bare of sand and ruins are seen over all this wide expanse of clay. It isa field so vast of surface problems, mounds, depressions, walls half buried, and canals long since abandoned, that years of study might be carried on without digging. One ride of a few miles leaves the rider at a loss for explanations; he finds areas that stand 5 to 10 feet higher than the plain in general, as though irrigation had been carried on a long time there with the rest around left barren; others in irregularity resembling the mere tops of silt-buried towns; and low areas somewhat irregular in surface with small holes where water has leaked down as though to fill the loose débris of buried ruins. ‘Towards the delta margin he may come to a canal 5 to 10 feet deep and in the section thus exposed discover silt-buried dunes and find much sand has drifted, interlapping with alluvium. And the horizon round about 334 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. him, flat and desolate, is broken here and there by distant towers, mounds, and citadels, that range far out into the faintly outlined forms which rise and vanish in mirage. The most comprehensive point of survey is found on the Erke Tepe, a high knoll or crumbled tower of Ghiaur Kala’s inner walls. This point stands central and 120 feet above the plain, with all the citadels of ancient Merv ranged round for 20 miles. Nearly all are flat-topped plateaus of varied height and length in profile; but sometimes, as the dust haze clears, far to the north more rounded forms are seen. Highest of all on the far horizon stands Odontche Tepe, nearly 8 miles due north. A superficial examination of this citadel revealed common wheel-turned pottery and glass, such as found on Ghiaur Kala, but its great height of 78 feet seemed promising of old layers to- wards the bottom. We attempted to explore it to the foundation with a shaft, but owing to its distance from water no men would stay there, though offered high wages, and our shaft was abandoned at a depth of only 44 feet. The débris proved lean of remains, evi- dently having been rapidly built up with walls as a point of fortification and not by slow accu- mulation of occupation. It seems probable, Sai however, that in the construction of such a Gullies Sundriedbrick Clay walls citadel advantage was taken of some preex- Fig. 489.—Sketch-map of Odontche Tepe (Merv), iSting mound, so that old débris might have been penetrated at a greater depth. It never- theless proved a lesson in wind erosion. The adjacent area of culture mounds has been so rapidly worn down that a surface of erosion weathers pots and jars and cuts them clean off, flush with faintly outlined crumbling edges. Kit Y A | Ag ANALYSIS OF SHAFTS AT GHIAUR KALA.—THE DEPOSITS FROM MAN, WATER, AND WIND. As Ghiaur Kala was chosen for excavation it became important to concentrate physiographic shaft exploration upon that city and determine the relation between its growth and that of the plain through water and wind. For this work there was so little time that only a few shafts could be undertaken, and to assure the possi- bility of correlation it was necessary to have them fairly close together and in some simple relation to each other, to the main excavations, and to the oasis in general. With all these considerations in mind we located seven shafts on the profile line as shown on the map, three without and four within the city walls. Here, as at Anau, we find that culture-strata, irrigation silt, natural sedi- ments, and loess entered into the growth of the plain, but with the very important addition of large masses of flying sands, a fact to be expected with an oasis of PLATE 70. SE 9 Ray eee sea an eg edagaey Pea ane eel noes na eee amy sae SS SS SS SS SS = WALLS | WALLS CITADEL WALLS Canal Canal

. Ce pa | aS > lg. [o} 4 So" . -g50'— — — — — — — — — — — O——— — — — — — — — mm: ete Siete 0 =e Shaft 7- — — — —Brick- yards — — ——— — a} tee S yp aia es ) a 6 a J! AO = < 7 iy fiaceat : Seen $ W eis: {I OSOTEE ares) 4 —< dS Se ee ee See ay a Sa Nea 3 al a et cage y ee > 25 Rapid Rapid fe filling filling oO \ SSSSSANLE AES SS ne Sand dunes Irrigation sediments Stratified Bard filling of fine clays soft loam, with only occasional banding VERTICAL SCALE OF SHAFTS ie} 20 40 60 80 FEET aS as | ee es Loess ® Shaft 6 PLAN OF GHIAUR KALA SHAFT PROFILE SCALE Cn ae 6000 FEET Ghiaur Kala Shaft Profile. OASES. 335 type Ia. But the significant features in this section that may be comprehended at a glance are, first, the order of succession; second, the stratigraphic unconformity ; and third, the relatively small thickness of water-laid deposits. Beginning with culture-strata, we come face to face with several problems. Ghiaur Kala was a great city and one of fame. Its ruined walls inclose an area of more than 1.5 square miles, wherein there still remains a plateau of the débris of civilization rising to a height of over 50 feet above the surrounding desert, while the citadel itself attains a height of nearly 90 feet. Did occupation of this whole area begin on virgin ground, or was a part already occupied by some more ancient town when the greater city was laid out? Or, in either case, was the inner city founded at the same time as the whole? And after the whole city was laid out with its present outlines, was it occupied continuously till abandoned, or was it abandoned and reoccupied one or more times? In the shafts, as in the main excavations, there was found no sharp transition showing change of culture, and no one of them passed from culture into irrigation. Moreover, if there had been a town there before, we should expect to find its pot- tery or some trace, such as charcoal or ashes, in the natural sediments under irrigation and culture-strata, as was invariably the way at Anau; but such is not the case. Though our evidence is in part merely negative, the city of Ghiaur Kala seems to have been founded on a desert surface of sand-invaded loess-steppe, partially buried in alluvium, and irrigation seems to have started simultaneously withit. And the fact that genuine culture-strata attains the same thickness above the citadel’s foundation as it does in the plateau of the outer city is evidence pointing to a simultaneous occupation of both. With irrigation deposits we find that 12 feet is apparently the average depth, surprisingly little when considering the antiquity attributed to Merv; 15 feet was the average at Anau and we had thought of Merv as an oasis of such ancient importance that it must have introduced irrigation long before, and, with the whole Murg-ab to draw upon, been able to maintain bountifully rather than sparsely watered gardens from the beginning. Then what is the explanation of its shallow- ness? Obviously we must choose between three possibilities—either the rate of growth was less, or irrigation was introduced later, or it was in no given area carried on so uninterruptedly here, as at Anau. Surely there is silt enough in the river to give a growth as rapid as the Anau, and Ghiaur Kala was founded earlier than irrigation is supposed to have been introduced there. We are driven to the con- clusion that the gardens of Ghiaur Kala were of a wandering sort, shifting out and back and sideways around the city, according to complications in the canal system and conditions of soil. Large areas are in our days from time to time abandoned for fresh land on account of the efflorescence produced by prolonged irrigation with saline water. Natural sediments fall next in the order of antiquity. They lie directly under culture and irrigation silt and over dunes and loess. Here, again, we are surprised by shallowness. Except for the interesting masses of obviously rapid formation that appear to fill depressions of the old loess topography in shafts 6 and 7, the 336 PHYSIOGRAPHY OF CENTRAL-ASIAN DESERTS AND OASES. natural sediments are nowhere over 10 or 15 feet thick. Their delicate laminations and extreme fineness of material differentially colored in thin bands of clay, for the most part buff and brown, are evidence of slow accumulation over a surface exposed to the oxidation of desert conditions. Although in a general way the growth of natural sediments came to an end with the beginning of irrigation and was superseded by that new kind of growth, we should expect to find layers similar to the natural at any horizon and interstratified with both culture and irrigation beds. From the time when one of the Murg-ab’s distributaries which fed Ghiaur Kala came under the control of man for irrigation, it ceased to be natural, 1. e., became an artificial canal, and thenceforth any sediments deposited by its waters were other than natural. We have called those accumulating under the stratigraphically disturbing influences of cultivation irrigation sediments. Sedi- ments formed in choked-up canals, reservoirs, and abandoned fields may be termed canal sediments. The irregularities produced by occupation and irrigation of a plain, with such a gentle slope as the one with which we are dealing, inevitably result in the formation of extensive shallow depressions where, sooner or later, canal sediments accumulate. It is, then, canal sediments that have risen by the west wall and that form a 2-foot capping to the irrigation stratum cut by brick- yards southeast of the walls of Bairam Ali. Shafts 1, 11, and Iv penetrated characteristic dune-sand, so loose that shafts I and Iv had to be abandoned before the underlying loess was reached. But in shaft 11 we had just enough to give a key to the section and yet not enough to interfere with sinking, though the same mass attains a thickness of over 15 feet in the wall of the main excavation (lower digging), just to one side and a few feet above. Sand-dunes were evidently characteristic of the region before it was occu- pied and much of the city appears to overlie them. It is a significant fact that flying sands are found beneath both culture-strata and water-laid deposits, natural and artificial, and beneath it all is the loess. Every shaft that went deep enough found the great underlying mass of loess. Shaft 11 found it under dune-sand at —27 feet and sank 36 feet down into it, pure fine loess with vertical cleavage and calcareous concretions all the way down to where we stopped at water-level, 63 feet below the surface. How much deeper it goes may be guessed, but there is no reason to doubt that it might be many hundred feet. THE STRATIGRAPHIC ORDER: (1) LOESS, (2) DUNE-SAND, (3) ALLUVIUM, EXPLAINED BY CLIMATIC CHANGE TO DRY, AND RECESSION OF THE DELTA. Now we are in position to correlate. The direct neighborhood of Ghiaur Kala had long been a loess steppe with topographical relief amounting to at least 25 feet elevation between its summits and depressions. Asa result of some change in conditions it was invaded by flying sands, after which began the alluvial flooding and depositions over its lower porticns; and it was during this stage that the builders of Ghiaur Kala arrived to look upon a land of desert dunes and playas, with here and there a remnant of the old half-drowned loess topography. Upon one of these remnant loess masses rising about 16 feet above the flood-plain of a dis- OASES. SOL tributary canal, they built their citadel, piling it up with clods of clay to a height of 34 feet, and around it threw up the massive walls of the inner city, of whose colossal height more than 70 feet still remains. That the outer city with its walls was laid out at the same time we have shown to be likely. So they built their city and from that time the distributary stream they had chosen is for physiographic purposes to be regarded as an irrigation canal, and the sediments laid down upon its flood-plain, irrigation and canal sediments, according to whether the area considered was under cultivation or not. While the débris of occupation rose within, these sediments grew upon the plain without the walls and to a certain extent continued growing after the abandonment of Ghiaur Kala in the eleventh century; for it was then that a new Merv, whose ruins are now called Sultan Sanjar, was founded but a few hundred yards away and water still found its way into this region. Loess, dune-sand, alluvium, and human débris is, therefore, the stratigraphic order of our physiography at Merv, the record of Nature and man, the effect for which we seek a cause. And of all time-sections it has been our fortune to study, this one is the most beautiful illustration of the organic changes that constitute the process of a great interior desert region effected by climatic change. During the accumulation of loess there must have been a sufficient precipi- tation to nourish grass over this area, but it is now too arid. It was then doubtless a time of greater precipitation over the Murg-ab’s catch-basin which would enable that river to penetrate farther into the desert giving it a delta north of the present. The flying sands derived from wind-work over the delta were probably accumu- lated into more or less stationary dunes around it, while most of the finer material settled as loess between it and the mountains. Then, I believe, a decrease in precipitation demolished the grass, set free the dunes to drift over all neighboring areas free from alluviation, while the river shrank with its delta, receding moun- tainwards to build over the dune-strown loess topography of Ghiaur Kala, and at this stage the city was founded. — ws a ~ PART VI. ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU AND THE HORSE OF ANAU IN ITS RELATION TO THE RACES OF DOMESTIC HORSES. By Dr. J. ULRICH DUERST. Professor in the University of Bern. [CHAPTERS XVI-XxX. PLATES 71-91.] CHAPTER XVI.—INTRODUCTION. The rich material sent to me by Mr. Raphael Pumpelly for examination consisted of a very large quantity of small fragments of bones, the putting together of which was attended with great difficulties. Fortunately the collecting and the preserving of the bones had been conducted with extreme care. It was thus possible both to form a picture of the fauna which lived during the forming of the different culture-strata at Anau in Turkestan, and to complete the recon- struction of some skulls and extremities. The whole collection of bones can be best classified, according to derivation, into those which come: (1) from the North Kurgan; (2) from the South Kurgan; and (3) from the citadel of Anau. The material from the North Kurgan is by far the most abundant. It is classified according to the system established at first by Messrs. Pumpelly and Hubert Schmidt, viz: Culture I, from 20 feet below the level of the plain to 15 feet below the same; Culture II, from 15 feet below the plain to 25 feet above the same (—15 to +25 feet); Culture III, from +25 to +40 feet, 7. e., to the summit of the kurgan. This classification has, in the light of my investigations, shown itself to be justified, although these gentlemen have since abandoned it for archeological reasons, in favor of the more simple subdivision into two cultures, viz: (I) Afneolithic, —20 to +25 feet; (II) Copper Culture, +25 to +40 feet. Notwithstanding this change, I believe that the original classification should be maintained to the extent of dividing culture I as follows: (Ia) —20 to—1o feet; (1b) —10 to +25 feet. With this general explanation I shall present a review of the varieties of animals represented among the bones found in the kurgan. First of all, it is evident that these bones are wholly the remains from meals, this being shown not only by the manner in which they were broken, but also by the numerous traces of teeth and sharp instruments still to be seen on their sur- face. The bones of which I shall treat in the following pages are the best pieces only. All the indeterminable pieces and those of uncertain determination are wholly omitted. It would be difficult to give a trustworthy and convincing out- line of the approximate number of individuals, since the pieces are often too poorly preserved to permit us to see whether or not they belong to one and the same individual. Wecan, therefore, only estimate them as follows: Culture Ia@ contains about 150 good pieces of bones, including Equus sp., 20 per cent; Bos sp., 27 per cent; Ovis sp., 22 per cent; Anttlope sp., 20 per cent; Cams 1 sp;, 1 per cent. Of culture Ib there are about 1,850 good bones. Here the representation of species is as follows: Equus, 28 per cent; Bos,25 per cent; Ovis, 25 per cent; Sus, 12 per cent; Antilope, 7 per cent; Cervus, 1 per cent; Vulpes, 2 per cent. 341 342 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Culture II, with 1,300 bones, shows the following relative distribution: Equus, 25 per cent; Bos, 20 per cent; Ovrs, 20 per cent; Sus, 15 per cent; Capra, 10 per cent; Camelus, 5 per cent; Canis II, 2 per cent; Antilope, 2 per cent; various wild animals, 1 per cent. The same relation as in culture II holds good also for the South Kurgan and the mosque-shafts of the citadel of Anau, except that here the sheep and goat are more prominent, while cattle and pigs are diminished in importance. The number of bones determined and numbered by me amounts to about 3,500, of which unfortunately only a relatively small percentage, about 10 per cent, are skull bones, about 17 per cent lower jaws and teeth; about 5 per cent are vertebre and rump pieces and 71 per cent are bones from the extremities. As regards the preservation of the bones, we find here the same conditions as among the European occurrences. The greater part of the bones have a light vellow-brown color, though some from the very lowest layer, as for instance those of the wild ox, the gazelle, the wolf, and the horse, show a dark red-brown color. There also occur some burnt bones from the period 1b, which are calcined and colored greenish-black. Some bones are distinguished further by a rich content of saltpeter, which causes them continually to extract water from the atmosphere and remain in a constantly moist condition. The old fractures, which show the same coloration as the surfaces of the bones, in contrast to the yellowish-white color of fresh fractures, enable us to make certain observations concerning the way in which the Anau-li broke bones. But we must first mention a peculiarity of all the light-colored bones—their high porosity and capillarity. If, for instance, one takes the metacarpal or metatarsal bone of a horse, even as heavy as 200 grams, or a piece of any other bone with much substantia compacta, and touches the tongue to a fresh fracture, the bone will hang on so firmly that it can be removed only with difficulty; and a place so small as to be touched only with the point of the tongue is able to support a weight of 200 grams or more. This is a pecu- liarity which I have found to exist to a similar extent only in the teeth of the fossilized Siberian mammoth; and it indicates a very great age for the bones of Anau. The breaking of the bones was carried to a greater extent than among the neolithic Europeans; for while these last broke open only the tubular bones of the horse, ox, deer, sheep, and pig, to suck out the marrow, and rarely the plate bones, as the caps of the skulls, horn-cores, ribs, etc., this was always done by the prehistoric Anau-li. All bones were broken into several pieces and many still show the distinct traces of sharp cutting instruments as well as of crushing teeth. The phalanx bones of the horse, ox, sheep, and pig escaped this fate, as did the horn-cores of the Gazella subgutturosa, of which the structure is too hard and the texture too compact to offer any temptation to break them open for marrow. Little is to be seen here of a definite method of breaking bones, such as described by Rtitimeyer for the dwellers in the Swiss pile-dwellings, and by me for the Germans of the Schlossberg, as the tubular bones and plate bones, lower jaws, and other cranial pieces are of an entirely different shape. Of the tubular INTRODUCTION. 343 bones, chiefly the distal and proximal ends are found, and the shaft is broken in many pieces, which it is rarely possible to put together to complete the bone; so that it would seem that the bones were for the most part broken with a dull instrument. Only in culture periods Ib and II do we find more frequent traces of scraping and cutting done with sharp instruments. In connection with the cranial pieces it is to be mentioned that all skulls were split through the middle. What I just said also applies here. In culture Id for the first time, and in period II, we find a sharp cut through the skulls of horses and of sheep. Skulls of period Ia and the beginning of period Ib were probably all broken with dull instruments; and for this reason we do not find any linear fractures, but merely jagged breaks, which in young animals follow the sutures. Lastly, as regards the age of the animals of which we have the bones, it is to be noticed that really old animals do not occur, with the exception of some horses, an ox, and some pigs, sheep, and gazelles. The remaining animals are for the greater part adult, except that among the sheep and pigs there is a great quantity of bones of very young animals—small porklings and lambs. This is confirmed by the abundant occurrence of milk teeth and epiphyses of the extrem- ity bones. It is particularly important to observe that more than 95 per cent of the bones of very young animals belong to uppermost layers of the period Ib and period II; from which it would seem proper to draw the conclusion that this consumption of young animals indicates herds of considerable size. The contrast to the pile-dwellers of Europe and the Germans of Schlossberg is very evident, since among these Europeans the appearance of young animals on the table was relatively rare. The investigation of these bones lasted from the autumn of 1904 to the spring of 1907, and was carried on according to the methods applied in former researches. For their comparison with recent bones I have used my own collection, as well as those of the museums in Bern, Vienna, London, and Berlin. Besides this, my friend, M. Paul Gervais, in Paris, very kindly made for me any necessary meas- urements on the skeletons in the collection of the Myseum of Natural History in Paris. For comparison with subfossil bones I have, in addition to the results of my own previous investigations and the measurements given in the literature on the subject, the collection of the Museum Society in Teplitz, including bones from fifty localities in Bohemia, sent to me at the same time for determination and labeling. It only remains for me to express publicly to Prof. Raphael Pumpelly my warmest thanks for the pleasure he has given me by intrusting to me the study of the bones from such an extremely interesting locality, as well as for the kindness with which he assumed the tedious task of correcting and translating this report, which I had written partly in English and partly in German. I am also indebted to the directors and managers of the museums I have named, and especially to Professor Studer, director of the Museum fiir Naturkunde at Bern, for assistance in preparing this memoir. CHAPTER XVIL. Ordo CARNIVORA. CANID&. The Canide are abundantly represented among the bones from Anau; never- theless their determination is not always easy. The animal of which we find the best-preserved bones, both complete skulls and bones of the trunk and of the extremities, is the fox. Canis vulpes Linneus, Vulpes montana (2?) Pearson, (See plate 71, figs. 3-13.) We find in the Anau kurgan two skulls in a very perfect state of preservation and 12 bones of the extremities as well as cervical vertebre. Thus we can make an exact determination of this animal. As the measurements of the following list will show, the foxes of Anau were not as large as those of Germany, but are closely similar to a recent fox from Tor on the Red Sea, whose skull is preserved in the Museum at Bern. The neolithic fox skull of the Swiss pile-dwellings of Schaffis is also smaller and nearly agrees with another skull from Sinai. The size of these fox skulls must not be taken as indicating a difference in species, however, for it is quite possibly attributable to a difference of age and sex. Pearson,* in describing his Vulpes montana, which probably occurs also in Turkestan, was not able to show any osteological difference between it and the common fox; the only difference being in the skin. Thus we may assume, though without possibility of confirmation, that we have here also the mountain fox (Vulpes montana Pearson), which is surely only a variety of the common fox. The bones here shown are certainly not those of an interloper of modern times, as one might suppose from the perfect state of their preservation. They are, to judge from the structure of the bone material, as old as the other bones of the Table of dimensions (in millimeters.) NaN Tibia, Radius, Extremity bones. —6 feet, 426 feet, ey feet, +23 feet, No. 392. No. 871. Ds: No. 560. MRR Me Wee retary soe o Balas prne ile airs ew alee a8) «Sino hols 112 126 123 bgt Se ter iC LORUMNAL CNG arden ae Sperm daten os 58 3 21 24 19 14 Wemeterot proximallend. ony. sists se eta tei 7 8 21 20 Warttirot tiediai part same 2s yisesyriaia sels 3) ace eo. 6 7 5 6 Pemerer OF median Parl. <5 ..2.66 seep ewe a0 5 6 6 7 Witdthiotr distal "MG as) ae ct sietrs se sreke sinner e aie a 17 18 12 18 Mramietenordistalvend ty. cays yote eke enatetaieise cre et 18 21 8 13 *On the Canis vulpes montana, Bengal, Journal Asiat. Soc., Iv, 1835, Pp. 324. 345 346 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Table of dimensions (in millimeters).—Continued, l North Kurgan, Fox f Cc f Fox from Anau, +17 ft. a o ts Peaa inte OX! neolithic Fox from Skull. See ( cin.) from | pile-dwelling| Sinai (Mus. y Schaffis Bern). No. 95. | No. 94. | (Mus. Bern). |(coll. Duerst). (Mus. Bern) Basilarlengthic nemesis eee 123 124 122 128 114 113 Basicraniala xis ee 50 51 50 54 49 46 Bastiacialiaxise wearin wi ae 69 72 65 61 Lengthvotinasalian en ot oe 47 ay MS 46 49 39 42 Widthtof/nasaliacen cen. 8 9 8 9 7 9 Lengthiotppalatesnc, aes 66 Bayt 65 69 58 57 Width of palate behind pre- niolar 4s. See nee 22 22 21 23 20 oi Greatest width of skull...... 44 48 44 44 41 42 Width on meatus auditorius CXUCLNUNG fh ay weasel ieee ie 42 45 42 45 42 42 Width on arcus zygomaticus. . 66 ne 66 72 68 78 Least dimension of temporal. . 19 19 21 20 20 20 Width between proximal or- bitals A, Vea ee 29 28 30 32 30 31 Least dimension between the interior borders of orbits. . ok oe 22 Hy 24 BA Length of cavitas cerebralis. . 70 69 fi 76 69 72 Lengchiottacesn. ss seneeaie 62 64 +» 63 67 Scat 62 Widthvotsk ttl een 39 42 40 42 35 Widthiof occiput: 2... rae 42 45 42 45 42 Length of lower jaw......... 97 oe gI Length of molars of lower jaw 57 : 48 layers in which they are found; but their perfect condition seems to indicate either that the ancient inhabitants did not care to crush these bones or that the foxes died while searching for food after the destruction of the dwellings of the layer in which they were found. Canis lupus Linnzus (Canis pallipes Sykes[?]). (See plate 71, figs. 1 and 2.) In the eneolithic culture-period we find at +8 feet three remains belonging apparently to the same animal. ‘These are a half of a right upper jaw, a part of the right frontal bone, and the anterior part of the right branch of a lower jaw. The dark-brown color of these three bones seems to confirm the supposition that they belong to the same individual. The upper jaw is distinguished by a short row of teeth. In this again the premolars form a relatively short row and the carnassial tooth is relatively very large. If we make the dental row equal to 100, the length of the carnassial tooth is 30 per cent and that of the three premolars 44 per cent. Among wolves, both European, Indian, and American, Studer finds the first relation varying between 25 and 29 per cent; only in a Canis hodophylax from Japan is the relation 31.4 per cent. ‘The relation of the three premolars to the whole row of back teeth amounts among wolves to from 43.5 to 49 per cent, and in Canis hodophylax 40 per cent. In our case, however, it is 44 per cent, in which the proportion falls into line with those of the wolves. This Anau canine was without doubt a wolf whose muzzle is somewhat short, the premolars are small, while the carnassial tooth is large, although the anterior inner protuberance of the carnassial tooth is very slightly developed. Interstitial Fic. 1. Facial part of skull of wolf (Canis lupus Linn.). 2. Part of lower jaw of Canis lupus. 3, Lower jaw, left branch, of fox (C. vulpes). 4-6. Skulls of foxes (C. vulpes Linn.), 7. Left branch of lower jaw. 8. Atlas (first vertebra cervicalis), PEAT Eales 1. 12. 13, Scapula. Humerus of young individual, posterior epiphysis wanting. Femur of large individual. Femur of smaller individual, Tibia, UBRANY 2 2 4 OF THE UNIVERSITY OF 5 PEATEs 72: H { i i 4 | Fic. 4. Corpus of lower jaw of Sus. 5. Rest of right part of upper maxilla. . Posterior part of left branch of lower jaw of Sus. . Tibia of Sus with mark of gnawing. Right metatarsus medius of Sus. Fic, 1. Canis matris optime skull from Anau (restored). 2. Canis matris optime skull from Kutterschitz in Bohemia. Lower jaw from individual from Gross Czernosek a. Elbe. Bronze Period. Museum at Teplitz (Bohemia). 3. Brain skull of Sus palusirts. cristatus, plate 80, fig. 1.) Onn (Compare with Sus UBhaRY ; | . - OF THE ‘ UNIVERSITY OF ILUNOIS. — ORDO CARNIVORA. 347 spaces between the premolars are absolutely wanting, while they are very largely developed in the skull of a wolf from Peking. ‘This character is derived from the greater shortness of the muzzle, whose length is that of Canis pallipes. ‘The second molar is also strongly developed. Thus the remains of the upper jaw are charac- terized as belonging to a wolf. The lower jaw, however, contributes still more proof. So far the bones might still possibly belong to Cuon alpinus Pallas, since the dimensions of the teeth are similar, especially as regards the carnassial tooth, which for the most part is 21 mm. long (21.2 to 21.5, according to Nehring). But the lower jaw decides clearly in favor of the wolf, since the talon of the first molar shows two very stout conical points. In the Cuon there is invariably only one conical point. The accompanying measurements explain these relations. The Anau wolf stands apparently nearest to the Indian wolf or Canis pallipes Sykes. Indeed, it is not at all unlikely that Canis pallipes formerly existed in the district of Anau and on the Kopet Dagh. In any event it is certain that the Indian Table of dimensions (in millimeters). Length | Length of | Width of | Length | Width | Length | Width Upper jaw. of tooth | carnassial | carnassial of of of of | range. tooth. tooth. | molar 4.) molar 4.) molar 3.| molar 3.) PATI all ao O LOCUM Asatte) ole cis. sheers ss 70 21 10 14 16 8 12 Canis lupus Linnezus, Russia MUS ee ESOT ID) sheer vis Euselc, sows at ers 71 20 II i 17 7 12 Canis pallipes Sykes, India (Mus. London), after Studer........ Es 21 IO Hg) | Cuon alpinus Pallas (after STRIVE @) Rg. Sede a aaielier a ae en 67 21 Ser 15 15 7 10 Canis lupus Linneus, killed near Peking (coll. Duerst)......... 81 21 mT 17 18 8 12 Canis poutiatini (after Studer)... 65 18 abe Canis inostranzewr (after Studer).. 67 19.5 buffalo (Bubalus arnee Kerr) still existed in historical times in northern Persia and Mesopotamia; and if these, why may not Camis pallipes also have had a wide geographical range? According to the view of Th. Studer,* however, this is not at all necessary, for according to exact researches the different species of wolves present only local forms which are without profound differences. Therefore, we will characterize the wolf of Anau only in general terms as Canis lupus Linneeus. We will, however, emphasize the fact that it harmonizes best with Cants pallipes Sykes. We have not, however, done with the bones of the Canidz of the Anau kurgans. We find in the strata between +28 and +36 feet a perfectly preserved skull with the superior maxilla belonging to it, as well as two lower jaw branches, one right and one left, which possibly belong together. These bones differ in essentials from those of the wolf in numerous ways, for which reason they can not belong to the wild Canidze; we have here to do with a tame animal, the domestic dog. * The Prehistoric Dogs. Abhand. Schweiz. paleont. Gesell., p. 9. Zitirich, rgor. 348 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Canis familiaris matris optime Jeitteles. (See plate 72, figs. rand 2.) I have with great care put together the remains of the skull found at +28 feet and thus reconstructed the whole skull as shown in plate 72, fig. 1. The length of the skullis 164 mm. Therefore, this dog belongs to the dogs of medium size. The skull is relatively low and approaches in this respect Canis poutiatini Studer, and the dingo. It is but little developed in width and is arched. Also, in the temporal region it is but slightly concave. The sagittal-muscle ridge is strongly developed. The tympanic cavities are relatively small, blistery, and without the usual keel-shaped ridge. The face shows a slight zygomatic arch and relatively broad, short palate. The relation of the dental arch of the upper jaw is as follows: The length of the carnassial tooth is 25 per cent of the whole row of back teeth, that of the three premolars is 42.5 per cent. We have, therefore, to do with the genuine dog. The carnassial tooth, however, has only an absolute length of 17 mm. and a width of 8 mm., which indicates a house-dog. The other cranial measurements are easily seen in the following table, in which the skull is compared with various others of similar size, of wild and domestic Canide. Table of dimensions (in millimeters). s F Sere We . hoe 50 Australian Pariah dogs, = ola g| Bohemia, Canis | mes ae | Dingo, after Studer. 22/3 a i eee $5 2 oe Mm after © ao us. Teplitz, |.§°9 = 0 as Studer ‘so/2o BBL. SY > +. Male. ‘Female.; din = hte, T'schont- WHI So ee : > &|3 6 schitz ‘ ~@) a! S38 - wTsiRS : Briesen | d. ¢d| G&, 23 as| Ny a ere La Tene. | : %3| 2s 2a < De ~~ well- . =“ No. 4.|No. 3.|India| Egypt| Egypt.|O Se &| % Q2rxriaH] os ae 4 3.) Sypt) hsyp mS Ey ings. A IeZ2) es : 3 Basilar lengthier oe ee 164 166 | 165 |161 167 157 |165 |164 168 164 |164 |169 |208 Length Of palate... werner 88 gl 92 | 90 94 84 | 91 | 91 94 91 |:90 | [Lares Width of palate.......... 50 49 Sol eag 50 47 | 46 | 50 49 49.) 50 | ..0s3 | Greatest width of skull....) 57 Sa SS ae ay 58 56 | 62 | 58 56 56 | 56 | 570) S765 Width of meatus auditorius externus........... ree+| 57-5 56 574 5205 55 50 | 60 | 52 59 60 | 60 | 58 | 75 Widthonarcuszygomaticus |102 TOS hres seh ke Le 98 | .. |104 104 Pe eerie iio Least dimensions between inferior borders of orbits | 37 34 fell Wet 2 39 32.1 35 34 26 ae Heightofskull’ ei see ae 54 52 Sr St 59 53 | 56 | 58 55 57 | 58 Length of teeth pet teen 69 O44 "62" ha alan 65 63 | 66 | 66 68 61 | 61 Length of carnassial tooth,| 17 18 07 acute 17 18 | 18 | 18 18 15 | 18 Length of molars. one: 18 18 OA es 18 19 | 20 | 20 17 15 | 18 Width of carnassial tooth .| 9 | seal eeee}oees 9 Op Ito 8 7 elle Basicranial axis..........| 46 | 49 48 | 43 47 SA ASS ARS 48 rch ct dle 2. Basifacialtaxismem sss eae LI2S MSI 7 seer 1S 120 113 |120 |119 120 Hy tiee TOA | After long and careful consideration I give below certain conclusions which are based on these comparative measurements and on the direct comparison of the Anau skull with the collection of dog skulls in the museum at Bern, and also especially with the rich collection of subfossil Bohemian dog skulls, contain- ing abundant material from 50 different Bohemian localities, which the Museum Society in Teplitz had the kindness to intrust to me for determination and pub- lication. ORDO CARNIVORA. 349 As remarked by Prof. Th. Studer, to whom I gave simply the occipital bone and piece of skull of the still unreconstructed cranium for examination, this piece of the skull possesses in a high degree the peculiarity and character of the dingo or the pariah dog. However, after closer comparison made later with the com- pletely reconstructed skull, I was able to decide that it has a very strongly marked resemblance to the European shepherd-dog, especially in the facial part. The only differences that this skull shows from the Canis matris optime of Jeitteles are that the upper jaw is less pointed and the palate a little broader than in the latter. This means that the muzzle of the Anau dog was somewhat shorter and broader than that of the prehistoric European shepherd-dog. As appears from the measurements, indeed, the skulls of the dingo and pariah dog agree very well with those of the Anau dog. Here, too, the muzzle is always a little longer and narrower, and the zygomatic arch a little wider than in our dog. Nevertheless, these differences are not great. They are even smaller than those produced by the difference in sex between two upper Egyptian pariah dogs published by Studer. That the Anau dog belongs to the shepherd-dog or to the pariah dog, which resembles the shepherd-dog, will be made clear by the following relations of the basicranial axis to the basifacial axis. Table showing relations between the basicramal axis and the bastfacial axis expressed in proportions to 100, . | Anau dog, | gx opent pene Pariah dog. Dingo. Shepherd-dog. | North— ‘sma iste Kurgan. 30.5 36.3 36.5 Nepaul. 281 40 Germany. 41.8 36.8 38.3 39.0 Egypt. Ala3 41.9 France. B77 ONG, 39.1 Egypt. 41.9 41.9 Canis matris optime fossil. | 85.7 39.4 39.8 Bengal. B0n1 39.8 Sumatra. 40.8 Egypt. | Here again it is clear that the Anau dog can have nothing to do with the wolves or wolf-dogs, in which the basicranial axis is 36 to 39 per cent of the basi- facial axis, or with the hound, in which this is 35 to 37 per cent. What we deduced, from direct measurements as to the form and size of the skull, is thus clearly illustrated in these very constant ratios; and it appears that our dog, standing near to the dingo as well as to the small Russian fossil wild dog, (Cams poutiatint), must be assigned to the shepherd-dogs or to the pariah dogs which resemble them. This relation is very well shown further in the ratio between the cranial height and length, which is clearly expressed in the following table: Table showing the cranial height expressed in percentages of the length. Indian wolf. Dingo. Shepherd-dog. Pariah dog, Bibel Pe 30.9 32.7 Canis matris optima, fossil dog. 31.7 Nepaul. 27.8 hen 32.9 Anau dog. 33.8 Egypt. 28 ion 33.5 France. 35.0 Egypt. 30.2 Sialece 34.5 Germany. ae 350 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. According to the table the skull from Anau ranks directly behind the sub- fossil cranium from Olmiitz, the original skull of Jeitteles. The domesticated dog of Anau belongs accordingly to the subspecies of Canis jamiliaris matris optime in a form which stands craniologically very near to the dingo and to Canis poutiatint Studer, but which rs distinguished by a rather broad muzzle. But how does Canis matris optime come to be among the inhabitants of the Anau kurgans? Judging from the very scant material thus far in hand, there can be no question that this dog did not originate at Anau during the first culture period. It is much more probable that it came to Anau with the people who brought thither the camel and the goat. But how came the dog in the possession of that people, and whence came they? These questions can not be answered from the material in hand. One would have to know with certainty whether the pariah dog, the modern street dog of the East, which we know on the oldest monuments of Egypt, is directly related to the shepherd-dog. Th. Studer* assumes merely a parallelism in development and derives the pariah dog direct from the dingo, which he considers to have been distributed in former times over all southern Asia, where it still occurred even in most recent times in the Tengger Mountains of Java. He believes, further, that in Eurasia in the diluvial period there existed a small wild dog which he has recently designated Canis poutiatini. ‘This species is said to have shown itself to be more easily tamed than the wolf and from this he derives, on the one hand, Canis palustris and on the other the shepherd-dog. Jeittelest himself thought that the Canis matris optime might have been derived from the Indian wolf (Canis pallipes Sykes), and assumes that the domestication of this animal took place in ancient Iran. Our finds in Anau might seem to lend a greater degree of probability to this idea, but the direct measurements and ratios given above show that the Indian wolf stands very far from the Anau dog. We must, therefore, waver between two opinions, namely, that the theory advanced by Studer,t in which he derives the shepherd-dog from a paleolithic dog of Russia (Canis poutiatint) is correct; or that the dingo, which we have represented as being similar to our dog from Anau, must have lived in southern Asia in some form, and that from it both the pariah dog and the shepherd-dog have descended. Which view is correct we can not, as I have already said, decide with certainty from the Anau remains alone. Derivation from Canis poutiatini is favored not only by the similarity in the measurements but also by the fact that the dog was brought to Anau by a people who imported the camel and the goat. Considering the localities of fossil remains thus far found, and the present geographical dis- tribution of these animals, it is possible that the camel came from the south or east, and the goat from the south or west, since its wild form now lives in Persia *Ueber den deutschen Schaeferhund und einigen kynologischen Fragen. Mitteilungen Naturf, Gesell., [sp ease lefciash aeeenh J { Cf. Jeitteles, Die vorgesch. Altert. d. Stadt Olmiitz, pp. 56-80. Wien, 1872. Keller, Die Abstam- mung d. aeltesten Haustiere, p. 55. Ziirich, 1902. tZoolog. Anzeiger, Bd. xxIx, 1, pp. 27-30. ORDO CARNIVORA. 351 and the Caucasus. The dog, however, may have come either from the sphere of Indian culture or from Russia; although, according to Studer, a Siberian origin is possible. On the other hand an Iranian or Indian domestication or an autochthonous origin of the house-dog, shown by the presence in the lowest neolithic layers of the Anau kurgan of Canis pallipes or a similar wolf, might support a very plausible hypothesis based on former philological or archeological researches. ‘This would not apply especially to the domestication of Camis pallipes, but to that of the dingo or another wild dog of Turkestan as well. Perhaps later excavations by Mr. Pumpelly,; possibly in strata of a still earlier period, will bring to light the bones of house-dogs; for the fact that none are known from the first period is no proof that they may not be found. It would, in fact, seem almost necessary that the Anau-li should have had, with their great herds of sheep and cattle in eneo- lithic time, a domestic dog that originated in the same neighborhood. According to Hommel* the different peoples speaking the Turko-Tartaric languages must have had in common an autochthonous dog, which was designated by the radical word kuc. Budenzf also calls attention to the original character of this designation and concedes the hypothesis of an autochthonous domesti- cation of the dog in the earliest times of the development of the Altaic culture. Vambéryt also sees an evidence of the high age of Altaic domestication of the dog in the myth of the Kirghiz, who derive themselves from the dog through an unnat- ural connection with forty maidens. Not only is the shepherd-dog thought by some to have originated in Iran, but H. Kraemer$ and C. Keller|| attempt to derive most of the European mastiffs —at least the Canis molossus of the ancients and the St. Bernard—from Tibet. Albrecht,‘] however, shows from a large stock of philological data that the Tibetans were not responsible for this domestication but rather the people who lived to the west and south of Tibet; and that the name of the dog argues against a domestication in Tibet, for in the west it is kukurra, while in Tibet it is kh. Albrecht believes, therefore, in two domestications, one of which produced a large dog (kukurra), in the west, and a smaller one (khz), in Tibet, which were then exchanged and crossed among the respective peoples. If, lastly, we would look for the shepherd-dog of the East, which might pos- sibly have been derived from the dog of Anau, we must turn our eyes to where the earliest rays of the light of history penetrate the prehistoric darkness—to Babylon, Assyria, and Egypt. The Assyrian monuments do not introduce us to more than two varieties of the dog—the large and powerful mastiff, used in the chase of great animals, and the grayhound, used in coursing the hare. Other breeds, however, were * Die Namen d. Sdugetiere bei den suedsemit. Voelkern, p. 441. + Magyar-Ugor ésszehasonlito szotar, 1881, p. 74. ¢ Die prim. Kultur des turko-tatarischen Volkes, p. 197. Leipzig, 1879. §Die Haustierfunde von Vindonissa, Revue zoolog., tome 7, pp. 143-272; and Die Abstammung des Bernhardiner, Globus, pp. 171-188, 1904. || Die Abstammung der aeltesten Haustiere, p. 76. Ziirich, 1902. { Zur aeltesten Geschichte des Hundes, pp. 55-56. Miinchen, 1903. oe ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. doubtless known to the inhabitants of Assyria and Babylonia. In the bilingual lists which give all words in Accadian and Assyrian, we find the Assyrian word na-adh-ru, ‘‘the protecting dog,’’ with the Accadian equivalent, sega lik-ka-gab-a, which probably means “the good mouth-opening dog;’’ then follows the Assyrian cab-bi-luv, from “‘to tie up’”’ or “chain up;’’ represented by the same /lik-ka-gab-a. Houghton,* who gives these translations, thinks that na-adh-ru and cab-bi-luv both stand for some strong dog, which was used both as a watch-dog to guard the house and as a shepherd-dog to guard the flocks. The idea embodied in the Assyrian and Accadian words cab-bi-luv and lik-ka-gab-a, “the chained-up mouth- opening dog,’’ answers well to a house-dog; and the notion conveyed by the Assyrian and Accadian words, na-adh-ru and sega, “the good protecting dog,”’ is quite descriptive of the same kind of dog when used as a shepherd-dog. One of the best representations of “the good protecting dog’’ is on the cylinder seal of Bel-Bin (see tailpiece at end of chapter).f This dog seems to be of a large, powerful breed, with his tail rolled up and his ears drooping down. Another shepherd-dog is represented on a cylinder-seal of the Clercq collection.t The other dogs of the Babylonians and Assyrians were all intended for the chase, from the very large mastiff to the swift grayhound. Even to-day one still finds in those regions, extending as far as Asia Minor, a large shepherd-dog of wolf-like appearance, used, as was the case among the Assyrians, to guard houses and protect the flocks from wild animals. One can form some idea of the size and savage character of this recent form from the report of Diest: ‘In Delilerkoi I had a fight with a dozen savage shepherd-dogs which were about as large as my little horse and almost pulled me from my saddle.”’ We may then assume that the Assyro-Babylonian culture did not derive the Canis matris optime from the later inhabitants of Anau; the size of the Assyro- Babylonian dog favoring rather Albrecht’s opinion of the origin of both a large and a small dog in Central Asia, the large dog being that of the ancient Babylonians and Assyrians. If we turn now to the Egyptian culture, thinking perhaps to find there traces of Canis matris optime, the monuments show us several types which correspond in size and shape to that animal. The real shepherd-dogs of Egypt as we find them, for instance, in painting and sculpture at Sakarah,|| and in the pyramids of Gizeh,{ often represented with shepherds, seem to have belonged to the gray- hound form. On the other hand, there frequently occurs a somewhat larger, short-muzzled, smooth-haired dog which seems to correspond well with the Anau dog. Good pictures of this animal have been given by Rosellini, but the best representation is the statue executed in black marble in the museum of the Louvre in Paris, the dog represented here having erect ears and a rather short head (plate *On the Mammalia of the Assyrian Sculptures, Trans. Bibl. Arch., 1877. + Ménant, Rech. sur la Glyptique Orientale, partie 1, p. 205, fig. 137. tCatalogue de Clercq, plate 111, fig. 24; Description, p. 40. § Petermann’s Mitteilungen, Erg. Heft 94, p. 64. || Perrot and Chipiez,‘‘ Bergers aux champs,’’ Histoire de |’ Art dans 1’ Antiquité, tome 1,1’ Egypte, p. 36. {{ Lepsius, Bd. vir, Abt. 1, Blatt 9, Grab. 75, Ostwand. ORDO CARNIVORA. 953 79, fig. 1).* A mummied skull of the same animal is in the collection of the British Museum in London. Aside from these facts I am able to give exact cranial comparison because among the skulls of Jeitteles’s collection in the k. k. Hofmuseum in Vienna which Table of dimensions (in millimeters). Subfossil dogs Recent pariah dogs | (Mus. Vienna).* (Mus. Vienna). North |— —— — ee eee a ; Kurgan, Egypt. Skulls. renee sYP | +28 ft.) From | From | Greece. | Abydos. Wadi-Halfa. | rock-tomb | tomb at | at Sifit.t | Thebes.t{ ASIAte LOH OLS ea ce ay fore ie ease: ole estes ns 164 144, eat 173 161 162 Superior length of skull............:- 182 LGt ee LOG 197 178 185 et tte GlatrOMt als. ieye oc. shas exrei-de- 0) IA 55 57 61 53 55 pete Ula Ol PAxIetall sec 27 ef!s sis cias aie 5 oo + 46 47 43 52 45 46 rete Ob asa evs. Wy aiely sn Sos yenavareuevs aber vars Piste 59 66 | 72 66 71 WER O AIA TCI. cr a aieiids sucusierett ccs. & 50 48 46 | 53 46 47 Nirclenvorimarietall dai... rere cams uae se 57 56 56 58 56 55 Width between proximal orbitals of PEOU Ce om eee ducted ocr ckehe creat | 54 52 55 47 47 | (reatest width of nasals. .;......:¢... C7) 18 18 16 | 1%) Least width between inferior borders of CATAL S OY 12 eee TE Wedteh teers ee hisses ls eg; 38 38 37 31 31 Greatest height of occipital............ 46 46 47 50 | 47 51 Greatest width of occipital........... 60 62 62 67 60 63 Width on arcus zygomaticus........... 102 97 99 | 105 (?) 104 IOI Width of muzzle before front premolar 41 35 35 AT a 36 38 Ratio of width of ‘occipital triangle to |-————_| ——___| —— —- | height, expressed in proportions to 100. 77 74 70: 7 75 78 79 Anat. Pirets gaits Wolf fr om | Ree | ae | Lower jaw. ) poutia-- Canis airy | cent | Canis | Wolf- Canis matris optime. tint.§ | matris | P cf ut? | wolf, ‘inostran-| Tired 2am {| S08: optime. ||, 8s. 2ewt. Length from proximal angulare 145 140 139 | 152 197 145 | Length from proximal articu- | | | / LALGEEMS fei ys: oie se aveks escee TAF NTAAG Es 140 62 Frc Length of molars............ 72) 73 | 79 Sle teolllN o She 100 78 | Length of toothless space be- | fore premolar Th. .'.s.. ss 5 yy CE ee J pti Wha be = Gis 5 Height of vertical branch.... ST ese aws: te 54 62 78 | Height of horizontal branch Behind molar. .....-..05 6 30 a5 -125.|.aF | 23) 2 26.5 27 34 26 Bengt Of Molar... nk es ae 23 2O je 2T*| 522-227 20 5ia 2 375 34 21 Length of symphysis........ 25 Che Sai Mleexe 22 Be | * Determined as Canis matris optime by Jeitteles himself. 2 After Studer. After Jeitteles. {Collected by Dr. Lauth. al {Collected by Schneider, was open to me through the kindness of the Custos Kittl, there are several mum- mied skulls from Egypt which, according to the accompanying labels and written documents, were determined by Jeitteles himself as Canis matris optime. As a characteristic of these skulls, we may take the ratio of width to height of the * Rosellini, I., Monumenti dell’ Egitto, etc., Mon. Civile, plate xvi. 354 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. occipital triangle, which varies here between 1:0.74 to 1 0.78. In Cants mairas optime this ratio, according to Studer, is generally 1:0.74 to 1:0.81 and in the wolf 1 :0.91. That Canis matris optime occurred also in ancient Greece is shown by a skull in Jeitteles’s collection, which is said to have come from a find in Greece (see table). We have thus seen that the Assyro-Babylonian culture in all probability did not possess the dog of the metal period of the North Kurgan of Anau and therefore probably had no relation with the Anau people; but that, on the other hand, this dog is found in ancient Egypt—provided Jeitteles’s original determina- tions are correct—and that we have here no such parallelism of forms as exists in the opinion of Studer already mentioned. It is, therefore, not improbable that the primitive Egyptians, who, in the opinion of most Egyptologists as well as myself (cf. Die Rinder von Babylonien, Assyrien und Aegypten, p. 73, Berlin, 1899), migrated from Central Asia via the Red Sea to Egypt, brought with them this dog as well as the long-horned cattle which originated in Central Asia. This is an attractive conjecture which follows logically upon what has been said. "The appearance of Canis matris optime in Greece is not astonishing, but forms the connection of the Anau dog with Central European finds, which are especially abundant in Austria. Migrations of peoples and commercial intercourse had, therefore, at a remote time brought this dog from inner Asia into the heart of Europe. Ordo RODENTIA. MURIDA. Arvicola sp. Asa recent interloper we have the lower jaw with all the teeth of a mouse. Oy UN INC CHAPTER XVIII. Ordo ARTIODACTYLA. SUINA. Sus palustris Riitimeyer. (Plate 72, figs. 3-8, and plate 80.) The remains of the pig are very common in the Anau kurgan. ‘There are about 120 pieces, the greater part being remains of skulls. The hard frontal bones have shown themselves especially resistant. We have, therefore, parts of the frontalia of at least seven individuals, some older and some younger. In one of these pieces the bregma has a thickness of 2.5 cm., forming a real armor-plate over the brain, while this measurement in other individuals amounted only to 1.1 to 1.5 cm. This animal was probably a very old boar. Only one brain-skull has all the bones complete. Even then the skull was split in the middle along the suture so that a restoration was necessary. The skull is decidedly that of a small adult pig, whose front shows a slight convexity, which we usually find in the Indian Sus cristatus or S. vittatus Miller & Schlegel. We shall consider with Nehring (Katalog, 1886, p. 54) Sus cristatus as the continental variety of Sus vittatus and employ for the south Asiatic pig the general name of S. vittatus. In comparing with the parts of this skull the frontal, parietal, or occipital pieces of the other individuals mentioned, one recognizes that the other individuals can have been no larger than this. The relations of the skull to those of different other small Suide, as well as to two other small skulls from Anau, are shown clearly in the table on the following page. This table shows that the skulls from Anau stand nearest to those of a wild Sus vittatus from Sumatra or to a tame Battak pig, not only in form but also in dimensions, and that they possess the greatest similarity to the skulls of the Torjschwein (turbary pig) of Schlossberg and Ia Tene, as appears from their general form. I think, therefore, that I shall not go amiss if I pronounce these skulls to be the oldest known remains of the Torjschwein or turbary pig. According to the researches of Riitimeyer,* Rollestone,f Otto,{ and others, Sus palustris, the turbary pig, which first appears in the Swiss pile-dwellings during the later neolithic period, is derived from Sus vittatus, which would agree very well with our finding. Nehring,$ on the other hand, considers Sus palustris to have been autoch- thonous also in Germany and merely a starveling form (Kwmmerform) of Sus scrofa domesticus. Which one of the opinions is correct can be determined with * Riitimeyer, Einige weitere Beitraege, etc., Verhandlungen, Basel, 1876. + Rollestone, On the Domestic Pig of Prehistoric Time in Britain, Trans. Linn. Soc., ser. 2, vol. 1. tOtto, F., Osteolog. Studien z. Geschichte d. Torfschweines, Revue Suisse de Zoologie, 1901. §Nehring, Ueber das sog. Torfschwein (Sus palustris). Verhandl. Berl. anthrop. Gesell., pp. 181-187. 309 356 Table of dimensions (in millimeters). ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. * After Riitimeyer. 1 A ee North Kurgan, Sus vittatus. | Sus scrofa, Tunis Soha $s 8 Anau. (Mus. Bern). (wild boar). (Mus. Bern). “ g gig oR BSS Skull. ied Aa|3 20 Wild, | of the | Mus, | Coll, |Schloss| La (Bete eae | +6 ft.| —8 ft.+20 a re Bat- | Bern. | Duerst.| berg. | Téne. |'§ "9/2 9 5 f taks. | | 8a) So a” |Y) Length from middle of occi- | pital ridge to line of junc- tion between anterior bor- ders: of ‘orbitsie ene eee | 116 125 120 118 113 127 99 116 Length from same point to | coronal sutures ase oe ok 54 55 55 53 46 56 36 38 35 40 Greatest width of frontal | | | Done ke < Cee eee tn 82 96 IOI 86 92 85 102 84 86 81 102 Length of frontal bone to | line of junction between | anterior borders of orbits..) 61 7 75 67 is) 57 64 Height of fossa temporalis | 32 vee | 38 37 32 42 er ‘ 25 39 Greatest occipital height...) 92 | 98 99 eee 106 ake Ses: 60 102 Least occipital height...... | 62(?)} | sf te one 82 60 65-sales5 79 Greatest occipital width....)| 88 - -. Af ee) go 90 88 110 Least parietal width........| 32 | | 33 34 26 17 24 24 23 29 Greatest parietal width..... | 72 | Te bis ey 78 72 69 | 68 81 Width between anterior bor- | Gers ‘Of orbits sean eee ere aoe 58| 66| 54 58 53 74 60 74 Length from anterior to | posterior border of orbit. ; 3I 34 33 35 38 32 36 30 a0 Anau. A Vl: nau anlohsee | Vindo- os Lower jaw. | +12 to| +24| +19 | +26 Tae | Spain of pile- _ 1903. Adult, Adult, Adult|“ | , ee Length of symphysis........ | 62 62 59 123-144 Width of corpus between exterior borders of alveoles | | Of TUSKS\ titi wa aioe a eee | 45 J wees | wees 45 42 Width between inner borders | | Of ‘Saintes. on cet ere ee | 28 see | 31 30 Transverse diameter before premolar 24 ie cme ere | 3 HSA, ocOilves seaee 91) eens ets 34 ac Height of horizontal branch | before premolan3... 00. es 7 Aopen saree 37 38 Height of horizontal branch | | | | behind molatiaa.ce eee 51 A SO aed 7) | 47 47 60 Length of molar 3:7. 1.0 sees) Re re aie ee ie ac 26 30-31 43-46 Width of ascending branch, | 78 65. |. 70. | 69 jo leo ehh | | absolute certainty only by comparative experiments with living individuals of both of the respectively postulated ancestors. In the meantime we shall follow the recognized hypothesis of the majority of the authors, namely, that Sus palustris is derived from Sus vittatus, especially since this conception agrees with the requirements of our special case, and because, contrary to the requirements of Nehring’s hypothesis, no remains of the large pig (presumably Sus scrofa ferus) are found in the lower strata. Notwithstanding the fact that the contemporary climatic conditions were favorable to the breeding of swine, they appear only ORDO ARTIODACTYLA. Ley at a later period. If, therefore, a degeneration and stunting of the wild Sus scrofa had occurred through a domesticated condition, we should have found transitional forms as well as among sheep and cattle. The occurrence of Sus palustris among the remains of Anau is therefore no surprise, since it was logically easy to conclude, as had already been declared by C. Keller,* that the animal must exist in subfossil condition in Central Asia, since it came at so early a period from Asia into Europe. It is, however, important that the turbary pig does not seem to have been domesticated in Anau itself. In spite of what has been said, however, there remains the possibility that the turbary breed of pigs, if not domesticated at Anau, may have been formed on some other oasis of Turkestan, since it occurs at such an early period (at —8 feet) at Anau. If we do not carry this hypothesis further, it is because in the first place we find no bones of swine in the lowest layers of the wild animal period, and secondly because an importation of the tame turbary pig from Iran or India remains among the possibilities. It is, however, certain that the turbary pig reached Central Europe with the builders of the pile-dwellings and contemporan- eously with the turbary sheep that originated at Anau, since it occurs in the earliest pile-dwellings; and in this animal also we see proof of the influence that was exerted by the culture of Turkestan on that of Europe. It is interesting also to compare the lower jaw with that of the European turbary pig (see table on p. 356). Studer, who explained the form of the turbary pig’s skull as signifying a wild condition, owing to a freer life, thinks that the weakening of the lower jaw, which appears in the later bronze age in Switzerland, was due to a change in the manner of life to which the animals were subjected. Our comparison, however, shows that the turbary pig of the Anau kurgan, down to that of the Germans of the Schlossberg and the Romans of Vindonissa, underwent no weakening process. The measurements of the best of the extremity bones given in the following table need little further explanation. While the dimensions of the scapula, tibia, and ulna agree closely with those of the domestic pigs of the Sus indicus series, as for instance the China, Maori, and Siam pigs, the measurements of humerus, radius, and metatarsus correspond very well with those of the Eurasiatic Sus scrofa Linneus or the wild Sus vittatus of southern Asia. The actual presence of the Sus scrofa, the Eurasiatic boar,in Turkestan is known; the larger bones of a wild boar appear only in the higher layers of the North Kurgan. It may also be possible that the Eurasiatic wild boar (Sus scrofa) reached Anau only after the south Asiatic wild boar (Sus vittatus) had disappeared. The exact relations, can not be determined in the absence of fuller data. The best conception would probably be that the neolithic or neolithic Anau-li for a while killed and ate wild pigs. Nevertheless, the wild pig seems to have been very rare. Whether the region was too dry and the forests of the Kopet Dagh offered too little space, or whether the Anau-li found the chase of this animal too difficult, can not be stated, but it is certain that we find the bones of the wild pig only in later strata and very scarce among the enormous quantity of other bones. *C, Keller, Die Abstammung der aeltesten Haustiere, 18, 102. 358 ANIMAL, REMAINS FROM THE EXCAVATIONS AT ANAU. Table of dimensions (in millimeters). (The measurements are taken in comparison with those published by Padelt [Padelt, M., Skelettmessungen am Schwein; Inaug. Diss., Leipzig, 1892] of specimens from China, Siam, Germany, England, and the Maori pig.) = | : | 15 jx |e |S St (ale) eee eee ue) al Pe | Bein Ui oe 0138 a) re | —E | | So) oe ot el cio. |S wre — Pe el ok eee She |e clade | ae es on a= Sy 37 |° IRBs oro | S)8l8\4)8)6) 3/5 /selEsealeg bl kis | S| o/ 8] e] SS le slasiksis 5 Bl BO oe ere dh Shh ge ile OR oon Be . Hie SSl/AlLH | a/ALAIA |jO |m& JA | | aa - | SCAPULA: (30 1071, a Oe eS, cm ee ee ed | 68 North Kurgan, Anau ~+26ft.|..|..175|28|..|.. | op 2 135. 10., 380) os hg ARB ies oh ee ie BOS Nee China, temalertae seni ccm aes ee P34 NOS 5 tee ger en eae ile Wee ll Meet 5 Maori pig; male, .o5..c i aces oe 87 ITTF OG le37 1 tel) reer ste aaa fae Wg Siam male snes cet eo eee S85 1X08) BS") SG. ls. ee eos Pe te eee? Germany, female.........#..... G2 ltog tz i agit, Pee a op eee England, femalevnwica an see eee IOS lta3 6. wae) eGR ehh ee e AeeeO Wild’ pig, maletta.e.. .cicx a'cuaaeen [260 |143 |124 | 44 | | / / 86 HUMERUS: (20 Mod) ao A nO eC sone eo ' North Kurgan, Anad =< 4-28 10.) Vanja 9), WV a7 ee comin eee | ((?) ie eae ee ee 32 China, temaleoa. Asantne sarees 25 | SS lek es ble eee ee ee Maori pig. miales eee eet are 163,50 1 oo a ote ser ewe ena Siam, maleot..0- eee chee ator n63 b-@ 43%). Gai tae Ol ok eS Germany, temaleg:.... cscs ies 12760.) 47 49 [48 1) 2. ee England, femalew. 062... vers 257 49S | 35 04a tach pens ae oe) ieee et | RS 1 2G)-9 |} ascah SO Mngland c's) ca sap ent waar. 25S anaiea ral ST ATae2s 102 Wald pigs see ee eee ee 266 | .. | 18 | 18 1 26.) 18 TIBIA: | North Kurgan, Anau, +23 ft...:/154 1/37 | a7 | 2514. 1.. 1 24 1.. 14. | «2 (900 62 China face tee eae ee eee 122) 38 25.) Gr) Saal a eat te, ale eee Somme Maori pig V0 8 orc, dan aor ae ee 183 24°22.) AI 4 Rela eee 82 wel) a ae | Oo Slain: :). «cay eee eee 167 4 49 } 394 27-1) cee 5 bo eee OA German yin: «cae ome ee 202 | 49 |20 16261) =. 29 : .. {100 | 59 Bagland 1h. cps vu donee eae meee 248) 9h ae de Nae | 38 150 | 97 Wald ‘pig nen. areal ere nine 236) '56. 10a" | 38 | | 35 99 | 73 FOURTH METATARSUS: North Kurgan, Anau, +23 ft.:.-] 85/45 }474) .. 1) costae ie Ching «One ti: i ee ane eee AS \, oc. a| Se ale sae edemal eae 8 Maoti Pigic - sukicas es00 hates ye Mel eek || see 14 STAT 2. add see a ee tee OS cca CA eek in eee a oT Germanys <5 Pocn nee eis eee: coh OME Reta tesets ll GS eae n ts) aoe Erigiand .2picgr sie. hits een oe 85 22 07 Wild: pig sacra erg ence eee ee 99 17 15 CAVICORNIA. 359 CAVICORNIA. BOVINA. (Plates 73 and 74; plate 78, figs. 1-4, and 7; plate 81, fig. 1; plate 83, fig. 1; plate 85; text-fig. 490.) Bos namadicus Faiconer & Cautley (Bos macroceros Duerst) or the Asiatic form of the urus (Bos primi- genius Bojanus). The remains of this bovid are also not very complete, but next to those of the sheep and the fox they are certainly the best of the bone fragments. They at least answer the requirements for the exact determination of the species present and are sufficient for the reconstruction of the skeletons. Among the 68 fragments of parts of the skull, trunk and extremity bones, the most remarkable pieces are two proximal ends of the radius with the corresponding olecranon of the ulna, Nos. 13 to 18. Both bones show old fractures, about 10 cm. below the articula- tions, and the ulna, No. 13, has distinct traces of gnawing, probably by human teeth. The dimensions of these bone fragments compared with the homogeneous bones of other animals are given (in millimeters) in the following table of dimensions: Radius. Ulna. | —_—_—— - ee —— ee = es ee ss | Width Oi .: : Least | : ef cher ee Width (Diam- | Length | width | Height : : LeeLeL Of) : | of sig- | Proxi- | proxi- | ig. id- | of ole- of ole- | id mal mal | de | "ha ~ | cranon,| cra- “pales end end | a “ eC | non ossa, = = ( —24 feet... .| ror | 48 51 33 | ee aa North Kurgan, Anau {.—20 feet... .| 96 49 33 | 140? 64 As (after David*..... 99 | eh Bec ' 146 a7 50 Et Prmmigentis attics Bntmneyer< 9) £03 Foose fio tee Le 152 84 53 Egyptian Apis, a RLS erotica Oy SO e300 ras 20 Bubalus arnee, No. 6707, Paris Mus.. .| 78 28 40 22 cst Equus caballus (Clydesdale)... .....| 2 se aetaee |) yes 130 Metacarpus. Width of Diameter of W idth oe tinetes of distal end, distal end. middle. | middle. North Kurgan, Anau, — 20 feet...... | 70 36 37 We Bos primigenius, after T’scherski..... | 76.5 a 51 35 Egyptian Apis, Paris Mus........... J 70 37 38 | 27 Bubalus arnee, No. 6707, Paris Mus.. 67 34 a6 | 23 Bos taurus, Paris Mus. . So oto OPI 68 42 46 32 Bubalus bainii, Ambolisatra......... 2 | 45 62 | 36 5 4 * David, A., Beitrage zur Kenntnis der Abstammung des Hausrinds. Landw. Jahrbuch der Schweiz, Bern, 1897, p. 134. It is easy to observe that the dimensions of the subfossil bones from Anau greatly exceed those of the corresponding bones of one of the greatest modern horses and closely approach the bones of Bos primigenius of the ancient Kuropean culture-strata. It is not probable, considering the remains we find in the later periods of the kurgan, that the first inhabitants of Anau hunted a gigantic wild horse or had a horse of the size of a Clydesdale or a Percheron; and there can be no doubt that the bones mentioned belonged to a large bovine animal. Another extremity bone, a metacarpus medius, No. 121, confirms this conclusion. Marked 360 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. by its special form as a bovine metacarpus, its dimensions are nearly the same as those of the Apis bull of the ancient Egyptian tombs of Sakkara, of which there is a skeleton in the Paris Museum, and are even greater than those of the recent Indian wild bulls. It is, therefore, evident that the bone did not belong to a horse, but to a bovine animal, although it might be still doubtful whether it was of the ‘bubaline’’ or the “taurine’’ form. Anotker confirmation of the correctness of this view is found in the determina- tion of the bones collected by Professor Pumpelly from the Komorof trench in the North Kurgan,* during his previous visit to Anau in 1903. These were sent to Professor Zittel, in Munich, for determination; and among them Zittel deter- mined a small and well-preserved bone as a right scaphoid of Equus caballus domesticus Linneus. Mr. F. A. Iaicas, an American zoologist, better versed in recent comparative osteology than was the great German paleontologist, writes on the bone another and more exact determination: ‘‘Not Equus but Bos, from a large animal. We have no skeleton of water buffalo (Bos bubalus), but it is probably this.’’ It follows that we must first decide to which bovine form this animal belongs, whether to a bubalus or a taurus species. To do this we must first solve the question as to whether the bovid of the Anau kurgan was a wild or a tame animal. More prudence is now shown than formerly in the application of the charac- teristics given by L. Rtitimeyer in recognizing whether bones belong to a wild or a tame animal. I judge the Anau bovid in question to have been wild for the following reasons: (1) because it was much larger than all the other domestic bovine animals, which were found in great quantities in the higher layers of the Anau North Kurgan; (2) because the structure of the bones is much heavier and harder than that of bovine animals in domestication, whose bones are spongy and lighter; (3) because the other species in the same layers belong undoubtedly to wild animals, and because this large bovid seems to be wholly wanting in the higher layers of the kurgan. ‘These are my reasons, and I observe that in several recent publications on subfossil bones the authors have been contented with only one of these reasons in attributing bones to a wild or tame form of animal. Assuming that this bovid was wild, we will inquire what wild bovids were living in Turkestan or Northern Persia in prehistoric times. ‘The first indications are those furnished by the ancient Babylonians. In earlier publications I have established the fact that two wild bovine animals lived in Babylonia, and that the language and writing of the Sumero- Accadians, who are supposed to have immigrated from Iran or Northern Persia, before these people united with the Semitic race, have only one word for wild bull. The ideogram of bull was a two-horned bull’s head, written V= gud, in *Explorations in Turkestan, Expedition of 1903, Carnegie Institution of Washington Publication No. 26, p. 8. Washington, 1905. + Duerst, Die Rinder von Babylonien, Assyrien und Aegypten, Berlin, 1899; Notes sur quelques bovides préhistoriques, l’Anthropologie, 1900, pp. 129-158. INE 7/8). Fic. 1. Right"branch of lower jaw of Bos brachyceros of Fic. 4. Metatarsus medius of small breed of II or metallic medium size from a turbary of Tschontschitz period. (Bohemia) Museum ‘Teplitz. 5-7. Phalanx I, II, and III, of larger breed of Ia period. 2. Same bone (restored), Bos namadicus. 8-9, Phalanx I and II, smaller breed of II period. 3. Metacarpus medius of large ox, breed of Id period. 10-11. Phalanges of camel, PLATE 74. y GUGGSs . z= Poephagus grunniens and bison Ti) GRBs Bos namadicus Bos primigenius Bubalus and bison Taurus Bubalus occipitalis Fics. 1-3. Basioccipital bone in 1, Ovibos; 2, Bubalus; 3, Taurus. 4-10. Accessory columns in teeth of 4, Yak and Bison; 5, Bos namadicus, India; 6, Bos primigentus; 7, Bos namadicus (Anau); 8, Bos tawrus macroceros (Nepal); 9, Bos frontalis; 10, Bubalus occipitalis. CAVICORNIA. 361 Semitic = alpu; a wild bull was written YW =am, Semitic=rimu,in Hebrew = reem. The difference between tame and wild bull is, therefore, in the ideogram, only the sign for mountain = **. A wild bull was in the ancient Sumero-Accadian language “a bull of the mountains.” Several ancient Babylonian sculptures or cylinder seals and many later Assyrian sculptures show very realistic pictures of a wild bovine, which I formerly identified with Bos primigenius Bojanus (plate 83, fig. 1). My recent studies on fossil remains of the bovines of the Indian Pleistocene have shown me that the Indian (Narbada and Siwaliks) and China faurina are the exact equivalent of the European urus (Bos primigenius Bojanus), excepting some very slight variations produced by different geographical and local influences: so that the Bos namadicus Falconer & Cautley would represent the European urus for the Asiatic continent, especially the North Indian mountains and their neighborhood (compare fig. 490 with plate 81). Fig. 490.—Bos namadicus, after Lydekker. Indian Geological Survey. The buffalo, the other wild bull hunted by the ancient inhabitants of Persia, Babylonia, and Assyria, is Bubalus paleindicus Falconer, or the recent form descending from that Pleistocene species, Bubalus arnee Kerr. It is already represented on the cylinder seals of the kings of Shipurla and of Ur. ‘The best representation can be found on the cylinder seal of Sargon, King of Accad, who reigned B. C. 3800 to 3750. ‘This seal in the collection of M. de Clercq, of Paris, bears the following inscription, “Sar-ga-ni-sar-luh sar Agaddeki Ib-ni-sar tup-sar aradsu’’ (“from Sargon, King of Accad, Ibnishar the scribe, his servant).’’* * See Clercq et Ménant, Antiquités Assyriennes, p. 79, fig. 46. Paris, 1888. 362 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. We know further that great numbers of these large animals were killed by the Assyrian king Ashur-nasir-pal, on the hunting grounds near the Euphrates. Aristotle mentions the occurrence of the buffalo with horns curved back to the neck, in Arachosia, the Persian province Khokand; and Chosroes II (591 to 628 A. D.) is represented on a silver plate in the National Library of Paris as hunting this animal.* There can be no doubt that two large bovine animals lived in very ancient times in Northern Persia in proximity to the Anau kurgans. Which of these two animals furnished the bones we are discussing and was the one hunted by the ancient Anau-li? Fortunately there are among the bones from the deepest layers of the kurgan several which enable us to answer this question. These are fragments of a right branch of a lower jaw, which I have restored, a basioccipital bone, and a fragment of a horn-core. Table of dimensions (in millimeters). | North Kurgan, Bos namadicus Bos primige- Bubalus Lower jaw. Anau, Narbada nius arnee — 20 ft. | (Brit. Mus.). (Mus. Bern). (Paris). a ee ee ek fe HOES OE a 28 a = | } Lemptiiny: ceca siete a sieteue es Pico ot snatcka cee eee 438 443 ates 455 | Lengtiwot Gemtitign... $0 secur pe nee ee 162 171 161 125 Lengthiot,molarsy2acer ...ren eee nes 108 110 . 104 Length ofprentolarss a... .00 7 > sere ee 76 80 56 henets | Length ‘of toothless "parties... .ntsnis oer mee 1 ie 115 asiefre 154 Width of ascending bratich ya. eee 136 142 ae 139 | Height of horizontal branch at premolar1..... 53 | 5I 45 5I Width of:ineisivespartcca. = ae cee ee 78 83 | 84 102 North Kurgan, | Bos primi- i] Basioccipital bone. Anau, genius, Kirken-| Bos taurus, —17 ft. dorightshire Walthamstow. (Mus. London), | B4e: as - “ Ge ee sats = i a Length of cond jie!| 2: i wea a ee ee 110 130 105 | Width of foramen. = ceccc enn Ose ons ae re eee | 43 43 Width aboveicondyies o. ...> aeons See ae ees een | 56 65 Teg Width, on upper, protuberances)... ee eens 60 2 57 Width on tuberctla pharynves ...c.c ah var. eee ee 37 eae 26 Length from upper protuberances to tubercula pharyngee... . 44 46 oe Lateral lenoth of basioceipital eat. ce seen eee 2 94 76 7 | Circumf aie = Verti 1 7 r ienc . > erence Provenience of horn-core. Soy wp Pi brcibgeae | Length of core. Anat. oo.0 ne ce ga Sie es ican eit are cae ee ore | about 400 120 | aes Suen-Hua-Fu, China, Bos namadicus (Mus. Paris)...... oa 413 122 about 830 Ilford, Essex, England, Brit. Mus. (N. 25426) Bos primigenius.. 430 124 1000 Algeri: a, Bos opistonomus Pomel (Mus. Algeria)............ 340 90 800 Levallois- Perret (Paris), Bos primigenius (Mus. Paris)...... 415 110 725 Rhine, Bosprimigentus (Mus. Darmstadt)......../3........ ) 382 100 690 It is well known that one of the principal differences between the bubaline and the taurine form of the ox tribe consists in the slender form of the lower jaw and especially in the greater width of the incisive part (corpus) of the lower jaw in the buffalo, and the greater narrowness of the horizontal branch behind the * Duerst, Suciaaes, ruminants sur des ceuvres d’: art Ceaaiiates, Revue archéolog., I, pp. 239-244. CAVICORNIA. 363 incisive part. These qualities are easily recognizable in the table of dimensions above. It is clear enough that the present lower jaw belongs, not to a buffalo, but to a large taurine animal like the European urus. The well-preserved basioccipital bone shows equally well a marked difference between buffalo and taurus (see plate 74, figs. 2-3). In the buffalo it approaches more the type of Ovzs, is rather short, and the tubercula pharyngee predominate vis-a-vis the upper protuberances. In regard to size, the basioccipital of Anau corresponds very exactly with that of the skulls of Bos primigentus of the British Museum. Lastly, the fragment of a horn-core (plate 78, fig. 2) denotes a round-horned animal and not a bovid with flat horns of quadrangular or triangular cross-section. This fragment represents the basal part of a left horn-core with some frontal pieces attached. Its surface is granulous, therefore it seems to have belonged to an adult individual. The core is at the base filled with some spongy bone sub- stance, which gives the impression that it must have belonged to a gigantic indi- vidual with enormous horns, like those found by Abbé David in the Chinese loess near Suen-hua-fu. This is corroborated by the measurements compared with those of several other horn-cores of Bos namadicus and Bos primigenius, which are remarkable for their size. It is, therefore, proved beyond doubt that a large taurine animal furnished the bones in question, and in the light of our researches concerning the wild bovine animals of these regions, it must certainly have been the Bos namadicus Falconer & Cautley, 7. e., the Asiatic urus. Of the bovine group, there remain to be considered only the Taurina proper and the Protaurina, as well as the bison and yak. “Here, too, the teeth offer an easy means of discrimination. According to Rtitimeyer* (1): “Bison and yak have become so sharply characterized that their teeth can be distinguished from those of Taurus, Bubalus, and of the Bibovina (Protaurus mihi) through the weakest development of the accessory columns.”’ This difference is more clearly shown on plate 74. It will be seen from fig. 4 that in this specimen of Pephagus grunniens from Nepal, male, about 6 years of age, the construction of the teeth is very simple. It is No. 611a; 5ab 28, No. 152, British Museum. On the other hand, fig. 5, Bos namadicus (specimen 36672), called by Riitimeyer Bos paleogaurus (Paleontological Gallery, British Museum); fig. 6, Bos primi- genius Bojanus, Pleistocene, Grays (Essex) (No. 21296, 21647, Paleont. Gal., British Museum); fig. 7, a molar series from Anau; fig. 8, Bos taurus macroceros, long-horned cattle brought from Nepal by Hodgson, 1848 (?) (British Museum) (skull, plate 82, fig. 1); fig. 9, Bos frontalis Evans, from Assam (British Museum) ; fig. 10, Bubalus occipitalis Falconer (Probubalus triquetricornis Rtitimeyer, No. 16173, Paleont. Gal., British Museum) show more and different plications of the enamel-folds. * Versuch einer Natiirlichen Geschichte des Rindes, 1 Abteilung, p. 91. 364 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Since all these molar series belong to individuals of about the same age, the variation can not be due to difference in age.- The sex is also the same, at least in the recent specimens, although it is not certain whether fig. 8 is a male or a castrated male. ‘This, however, should not produce a perceptible influence upon the dental markings. Here, too, after an exact comparison, our Anau ox ranges itself between Bos namadicus and Bos primigenius; while the enamel plications are still more manifold than in Bos taurus macroceros of Nepal, which, in comparison with the early Egyptian long-horned cattle, has very numerous plications. Bos taurus macroceros Duerst. (Plate 79, fig. 2; plate 81, fig. 2; plate 82, fig. 1; plate 83, fig. 3; map, plate 85.) In the layers of the North Kurgan of Anau, beginning at —15 feet below the level of the plain and extending to the top of the hill, we find numerous bones of domestic cattle. These bones are found far more frequently than those of the other animals and are easily distinguishable by their lightness and highly porous character. In addition, the color of these bones is a light yellowish-brown, while that of the bones of the wild bull is a darker brown. To determine the origin of the domestic cattle, we must try, first, to recognize the breed and size; second, to see whether there is any connection between them and the wild cattle of the strata below —15 feet; and third, to determine the probable distribution of these animals over the neighboring parts of Asia. ‘THE BREED AND SIZE OF THE DOMESTIC CATTLE OF ANAU. In various earlier publications, I have several times expressed the opinion that the short-horned cattle of the turbary man of Europe (Bos brachyceros Rtitimeyer), which seemed to have been imported from Asia, were the oldest cattle of the world, although descended from a long-horned wild species, the Bos namadicus. It seems, however, that the bones of Anau tend to contradict this opinion. ‘The bones of the cattle in the layers from —15 feet to +25 feet show no concordance with those of Bos brachyceros. They are all distinctly larger in size and the few remains of horn-cores indicate a long-horned animal. We will now first compare the dimensions of these bones with those of other prehistoric cattle, especially with those of the mummified skeletons of a sacred bull, Apis, of the old Egyptian tombs of Sakkarah, now in the Museum of Natu- ral History at Paris, and with those of Bos taurus brachyceros Rtitimeyer, from the excavations of Schlossberg, near Burg (Brandenburg), and several localities of Bohemia. But we shall see that it is not possible that these marked differences were only those of sex or individuality. I begin with the extremity bones. Several fragments of scapula were among the bones from the kurgan, but only two fragments of distal ends of this bone are sufficiently preserved to serve for measurements. ‘The dimensions given in the table on pp. 366-368 show an exact concordance in size between the fragments from the +17-foot layer and the scapula of the Apis bull; the fragment from the +26-foot layer being smaller. CAVICORNIA. 365 There is also a number of humeri, mostly fragments of the middle part without articulation. Only one proximal part and four distal parts are in good condition. The measurements indicate the same important rule—that the bones of the lower layers are much larger in size than those of the higher layers, and that those from +25 feet are approximately of the same size as those of Bos brachyceros. The same rule is good for the radius, of which we have only four good pieces. We have the distal part, No. 1114, from the —15-foot layer, undoubtedly belonging to a younger specimen of the wild Bos namadicus, as is shown by the dimensions and the heavier and harder composition of the bones. As is proved by several other bones, such as phalanges, the wild animal appears from time to time in the layers near the level of the plain, but still decreasing proportionately in number, and seeming to disappear with the + 20-foot layer. All metacarpal bones were badly injured; only one has been restored and this shows a close concordance with the Apis of Paris, while another from the + 20-foot layer approaches in size the Bos brachyceros of the Schlossberg. I must here treat another question. It might seem probable that the smaller bones of the higher layers are those of younger animals of the same kind as the larger. This opinion is certainly true as regards the lower layers, as the smaller bones of these strata show very decided marks of youth; but this is not so with the smaller bones above +23 feet. All those which are mentioned in the meas- urement table are of adult animals. It seems very probable, therefore, that the higher layers contained a smaller breed of cattle which was formed there by the physiographical influence on the climate and on the production of food during the period of aridity at the end of culture period I, or which came into Anau at the same time as the camel, the goat, the hornless sheep, and the shepherd-dog. But the long-horned larger bovid does not entirely disappear in the metal period* of the kurgan; several larger extremity bones show his presence among the smaller cattle. Among the phalanges there are in the lower strata several which in size correspond wholly to those of Bos primigenius and which are even considerably larger than those of the Apis of Paris. The measurements of the femora from the North Kurgan, Anau, agree well, in the measurable dimensions, with those of Bos brachyceros and are considerably smaller than those of the Apis skeleton. The same applies to the tibie. The metatarsi show also, in part, dimensions which indicate a somewhat more slender-limbed cattle than was the Egyptian long-horned cattle. The measurements of the lower jaws of the first period correspond to those of the extremities and show the same dimensions as those of the family of the recent Bos macroceros, as is easily seen in the agreement of the few lower-jaw measurements with those of a mummy skull from Abadieh, and of the Hungarian bull of the Hofmuseum of Vienna (plate 81, fig. 2). * Culture II.—R. P. 366 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Table of dimensions (in millimeters). Width of articula- Saou ¢ .|2¢/5,|4 sie Me men oes £/ 25/58] < 2/8/3S|/ 8/5) 2] 88) ga] og | S ely AS tects aa E Ae 5 vs Ty ) —a|[—a|/FElalealS |] Gol e& | FS 1 ee a cs s B | co rs) oe fe) = iS iS a = — — =| BE oo = i Y tl\ Rl RIS |S |S) 8) 33/5 Biel elslBl\eB/ 8B) 22/88] @ Sle al al alAlAssS a SCAPULA: North Kurgan, Anau: Ey TeCt ee. ee enero : 57 “+56 feet oes eee te ae ies 55 Apis bull, Bos taurus macro- Ceres { Paris) (cee ee Sif Schlossberg, Bos taurus brachy- Ceros: (NOs 27)\ act eee 48 HUMERUS: North Kurgan, Anau: No: 618, -22 feets eae 110 |120 | .. xe eee No, 74:77 “7 teets woe ae 94 51 Pa No. 1081, +26 feet...... 3 69 46 37 No, 35, +26 fetta... - : 64 45 30 No, 1082, 26 fects... ee ee eee ee Be 38 30 Apis, Paris: = ene ene 350 |112 |117 | 42 | 57 | 87 | 80| 50 38 Schlossberg, NG™ 77.6. el 240 | G44'S5)| 28 7 375-63) i Ounce tte eee RADIUS: North Kurgan, Anau: No, aiitd4, 15 feet ae 60 | 35 | 91 | 58 NO? 6415 > ot 1eeh ac. tea. SW Ae ol skceel SOe AS No, 772,%4-2'5 feetwe sone 88 | 39 Bi NO). 808,25 CCC ea ee 89 | 39 Apis, Bos taurus macroceros (Paris): eh vere cere: 4 332 | 89 | 30 | 48 | 20 | 82 | 47 Schlossberg, Bos taurus brachy- ceros (No: 188)esee eee 174 54 We 33a BO ee Tes Om eo METACARPUS: North Kurgan, Anau: N0O26835 4 7ateec, ater P25 TIGA S Head Vesa aS No: 781; +26 feet... ee Sc ll Ged tml tote Male aes? NS No. 31, Komorof's trench] 73.fieax 1 Sei hosaehe aa eat a4 N0O.272; 4-26 feets...e. TT A OM eo Me 2 2 eee eer No; 257,)+-28 fects... Pay el eae cael Gree Sin peo Apis, Parts on oh as ees 234 1 62 f 43 1 384 37 | 90] a9 Schlossberg) = o.5..25.6. eee 190 | 534] 42) 25:) Fo)" 5041 30 FEMUR: North Kurgan, Anau: No,.1044, -b 28 feetar. 14: TINO NONS521; -- 32 feet eee 80 100 ‘25 eet oes ge ot eee eee 83 |112 Apis, Paris, Bos taurus brachy- CFOS: araeict. Tae Eee eee ‘104 134 TIBIA: North Kurgan, Anau: No. 899, +35 f6€ts. = aenical sos) ee ene ae ek “f= 32 feet. ee eee ee eee 307 175 7S ego SEO Sass Bos taurus brachyceros: Schaffis (after David), ...) .. 7 a5 9 <0) cee ee Oe eae Anis, Paris /o.... een 403 1107 | 90%) 47 |-54 1 7O "51 tion. 7O 50 Diameter of articu- lation. SS Table of dimensions (vn millimeters).—Continued. CAVICORNIA. 367 ist vu . | ¢ |. o © = | + : 44 Vv . ; rir Ca 2 = = is Oo = g is ae ce cd ee a ee E: 5 - cea SL Sy Sy Els Sets pens Pile ih a Sea ee oh ET 3 SH | ow | oR Sos e SS ~ ie} = 5 i) kota cra eRe Re a > er ey) ha a ee be = 6) H METATARSUS: North Kurgan, Anau: INOwO5O;0-1 al COUR atts sid crc eae ae eee OF 3o INGe Ore COMeCl tne sue clans COn eA te souls Dalene het. INOS OS 20; TCC laters wtemuneracts Meee meres) eta Wess I 2) 1 30 IN|), MeUNep, SoC PIN SAGs Bion Gan ee oor Seer cellieatcr hare bos cole BO! Te SS _ No. 1579, +24 feet...........- 221 At | 40 1 2351825)" 50 4) 28 IMS, Veh Cis 6 aig noo dn Gk ee oie tion oe bE RB.b 50 1 321 32.1 65. 37 PHALANX I: North Kurgan, Anau: IN} Ti aR Ss aa sie erod sare 7) 44.1-41 ) ou 28 | 34°) 26 Fea No 22, Komorof:sitrench’-.. -. 2. POz as seeee.4 261-43.) 23 Noja4G, Far feet sa ae ens ms 64 | 26 | 31 | 22 | 23 Eo 20 pteED EL CO Li sceen eevore sat dx ek eet shsl = ies", COMe2T RE STE 23a 2) Ve sAnia1LG oe joa - Pe a en 55 ' = aa ce ta eee 20 | TOT 2 Souced «Ati oe OeOartns Ce ee 34 | 22 | 2 2S; i 12 | NiomA S138 1226 TeGt yah ene onetern ey 60 | 26 | 30 | 22 | 22 | 27 | 20 INiOnE 7020-120 TEEl as yin se citer = 60 | 29 | 35 | 25 | 24 | 27 | 20 No. Belch Sets hts cee cud dma ae Gi 26 | 354025 fy!) 28) 28 aw Be 903, + 32 feet.......--+--- 63 30 | 33 | 26 25 28 | 22 » ti PIS, Faris... 2. ee eee ee eet eee Bos Si 36. 130.) 20. 8les 5,4) 2 Bos primigenius Bojanus: | Unter-Grumbach, after Schoe- | Edeae ‘ 5 | Pensack yn Gaerne bi aie A z : 79 |395 37 | 39 Robenhausen, after Riitimeyer 1| te - | ee Re / | | Bos brachyceros: | Schlossberg, No. 234.:...-.-.+-> | 58 | 30 | 33 | 23 | 22°1 38 | 20 Gross Czernosek, La Tene (Mus. ‘cd Ceo ev @licre eRe Ree SL@LIt Z,) Seeme talse ents tase, he 5 sees a| <8 PHALANX IT: * alle ck Saige ae a | ‘oe North Kurgan, Anau: | | No, LOOS'S Vester velere Srentioksse he sate. 43 | 33 | 37 | 27 | 26 | 29 ime NOMA DG store congas cis teers sree a AON SA e325 (e2Q) |e 32 1 33.1238 ; VS INOS 215 Poe. ie a nchavon Sees: A2e SOU 32a 1) oar 2Ouy 2S : INO 528: crac eevee sian wel 40 | 30 | 31 | 24 | 22 | 27 | 21 : No. GA Tw elarata cecie + teeadsts tive sls AT MeStuE SA 25 |, 240 fe? 5 oT INOW S4 ahd TeCbny ce oh een ne 40 | 29 | 31 | 24 | 23 | 24.| 29 ie Be ie a: S Sport aemreh ae: 40 33 S65) 27 (hey hath} 35 Os Oot OneeL ow eras tt. | 40 | 2 207 |t eee 2 eee INGHOT e--2 Sileetry ie ec- ares ie STAs ae sOe (27 e2o) | a2 | 35 ie eo JG. be ermine 5a. 28.1. 30r])-26 20) | 2141623 a) 26 DISSE ANIG Memenaart es on reacts eta AG") 32") 39] 26 | 26 |) 2 35 Hostomitz (Mus. Teplitz)........... 44. A034 Naz hres 128) ) 307) 35 Bos primigenius Robenhausen, , 6 a | ALCL RMTIMOY CL. cacmersvins, © cen! aan ae re «44 | 38 31 PHALANX IIT: North Kurgan, Anau: INO SO Fier t Meee AEN Sess et rote part 80 34 37 95 18 IN OROO Tee tela ae eer iis ee lore ene ho mele 77 29 38 90 17 NemSOR ews candi, tive sieatcys 71 25 a 85 17 IN Op A Opie res elev clorsieueleasiars sisis- shay 5 67 26 35 7 18 eNO OS are e wtetchangery x a eiecia': vies 65 23 33 73 T5 FA DISMPATIST ea tverae ota ondie crete awe «5 80 32 37 95 Soper 368 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Table of dimensions (in millimeters) —Continued. % rength of | Letgthof'| Width of ) ons Lower jaw. Length. Kaine toothless ascending ae 2 part. branch. glenoidalis. North Kurgan, Anau, No. 912, | de Srieet, Go ses we eee ee | ae tt Merce Deere 118 45 Mummied bull of Abadieh | (Egypt), Mus. London..... .| 375 135 109 119 44 Bos brachyceros, Walthamstow, coll, Duersts7... en eee : 318 oly 89 88 39 Bos taurus macroceros: Watussi, Mus. Berlin........ 340 122 104 114 re Hungary, Mus. Vienna....... 395 136 86 118 44 Somal, Mus: Beérlinz ns). .s. | 325 II4 93 Ill | Attee Length of bi of Ph of | gone fe! | Diameter of eater ee — we. | articulation. aperture. articulation. | North Kurgan, Anau, No. F507; 6 TCC. ee | 50 234 rs 143 52 112 Bos primigenius Riitimeyer 56 255 147 150 ine 115 The horn-cores of the tame bovid of Anau are represented only by fragments of the bases, and by several pieces from the inner curve of the cores. The one measurable horn-base is about 23 cm.; one of the longest fragments of the inner curve of the core has a length of 24.2 cm., which corresponds to an approximate length of horn-core of 32 cm. We find in some subfossil and recent macroceros bovids the following corresponding dimensions in millimeters, and the horn-cores from Anau would, therefore, correspond very well to the old Egyptian bovid: | Circumference Length of Longitudinal ‘Transverse at base. horn-core. diameter. diameter. Egyptian long-horns from Abadieh (Mus. London): INO. Gis Sis ata es wee eres anne eee 160 aac 55 44 INO} 5 nesee tele ae eee ee 190 270 64 56 Hungarian bull, Vienna............ 250 520 66 84 Apisiskulltof Hallesacnes i cee nee 270 405 As in my memoir on the Fauna of the Schlossberg zu Burg a. d. Spree (pp. 254, 255) and the size of the domestic cattle found there, so in this case I shall calculate the height of the withers from the separate bones, in order to make clear the form and size of the Anau bovid. For Bos namadicus, calculating the length of lower jaw, the length of skull along the base is 51 cm. and the height of the withers in the live animal is 149 cm. For Bos taurus macroceros we obtain, according to comparative measurements, and basing our calculation on the metacarpus, which has a distance of 60 cm. from its distal end to the proximal end of the radius, a height of the withers in the living animal of 153 cm. CAVICORNIA. 369 Some comparative withers heights are given below: Bos primigenius, skeleton at Berlin............... 168 Bos namadicus, Narbada Valley «India, calculated from the length of the skull (50.8 cm.) in the BritisheMiUseiiniaee cs ae he ie verte. f erie 149 PS OGMAMOLLCUS ANA mere eters wielee Gielen «sie each 149 Apis skeleton (Bos tawrus macroceros) Paris........ 154 BOSMMUTUS ANACHOCETOSSUATION to icloiie ce cee eve te steels 153 ERUNSATIAN Caltleemp eevee me Maree: catty ecrettrt races) © 150 SUAOTE VOI Callie re oc sans oa wo ew yee ae rh nie co 149-122 PactiPristan cattle fs ooo cates tee tee 149-137 Swisshommmentalicattlern comdcteie scsheres) se esettrarels oe 145-137 It is probable that the measurement calculated from the lower jaw for Bos namadicus is rather small, because the lower jaw is perhaps that of a smaller or female individual. We have unfortunately no means of checking this, unless we calculate the height of the withers from the width of the metacarpus. This small measurement, however, leaves a very uncertain result. We obtain 158 cm. in our special case. In any event, the two calculated withers heights suffice perfectly to show that we have to do with large, stout bovids, in both wild and domesticated forms. We can therefore recapitulate as follows concerning the results of our study of the bovids of Anau: In the lower layers of the period Ia, from —24 feet upward, there occur the remains of a wild Bos namadicus Falconer & Cautley. During period Id there originates from this wild form a domesticated bovid, large and stately, provided with long horns. Judging from the measurements of the preserved bones, this is absolutely the same ox that was possessed by the ancient Egyptians. In the period II the size of the animal seems to have somewhat diminished, unless possibly a smaller bovid may have reached Anau with the other newly imported domestic animals. It is, however, possible that this small form of cattle of the culture II originated in a decline of the cattle-breeding of the later Anau-li; as indeed the originally large, long-horned ox of the early Babylonians had already become small and short-horned in Assyrian times, and to-day, after a relatively shorter interval, shows a tendency to become hornless. The existence of the short-horned cattle in Western Central Asia is also shown by the discovery of a skull in a kurgan of bronze time in Bizino, near Tobolsk (plate 78, fig. 6). We find the long-horned form of domestic cattle already in the time of the Babylonians about 4000 to 5000 B. c., in Mesopotamia, as appears on a cylinder seal of those times. We see on this seal the representation of two oxen, moving through a field of grain. Still better known and more available for comparison, because of the greater quantity of existing bone, whole skeletons, skulls, etc., is the occurrence of long-horned cattle dating from the earliest times in Egypt. I have previously discussed the connection of these bovids with the African and European forms. It follows clearly from the distributions of the long-horned cattle over Asia and Europe that the Anau bovids also had an influence in forming the European domestic cattle, as we shall endeavor to show later. 370 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. OVINA. WILD SHEEP. Ovis vignei arkal Lydekker. (See plate 75, fig..1; plate 76, figs. 1-4 and 8; plate 82, fig. 2). Among the bones which are assignable to the sheep there are several frag- ments of very large horn-cores, which could in part be put together, forming then the calvarium (plate 75, fig. 1). This comes from culture Ia layers from a depth of —2o feet and, therefore, belonged to a contemporary of the oldest period. In order to identify this fragment of a skull we must first learn whether we have to do with a wild or domesticated sheep. This question, thanks to the better bones, is easier to determine than it was in the case of the Anau bovids. There is no domestic sheep which shows horns corresponding even approximately to these horn-cores. We find them, however, among wild sheep. Among the wild sheep that might come in question are those of the steppes, Ovis orientalis Gmelin (Ovis arkal Brandt) and those of the Kopet Dagh, which Lydekker calls Ovis vignet arkal. It is now evident that, even according to Lydekker, there is no great difference between these two forms of sheep, and that it will not be possible to show any differences from the few bones, since the species and subspecies are based only on characters of skin and horn. I would remark here that in my preliminary report of last year to Professor Pumpelly, without then knowing cf the occurrence of the urial in the Kopet Dagh, I wrote: ‘‘ These large spongy horn-cores seem to belong to the forms of Ovis orientalis Gmelin, seu O. arkal Brandt, although they are but little different from Ovis vigner Blyth.” The Kopet Dagh sheep was named Ovis arkal in 1857 by Blasius and is evi- dently allied to the urial of the Punjab race, with which Lydekker has proposed to identify it.* In a more recent treatiset Lydekker studies an adult skull of this animal and says: It will be remembered that the Punjab race of the urial (Ovis vignei cycloceros), at any rate as exempli- fied by the specimens from Peshawer and Afghanistan in the British Museum, differs from the typical Ovis vignet of Astor and Ladak in the much greater prominence of the two front angles of the horns, which are often raised into nodose beads, between which the front surface of the horn is depressed and carries bold and widely separated transverse ridges. In the Kopet Dagh urial the prominence of the front angles of the horns is still more pronounced, though the beading is somewhat less conspicuous. Moreover, the front surface of the horn is unusually broad and flattened, with the transverse wrinkles very low and indis- tinct. The length of the horn is 33 inches along the inner front angle, with a basal circumference of 11 inches, a basal width of 3 inches and a basal depth of 4 inches. The last two dimensions are considerably greater than in a skull of the urial, measured by Mr. Hume, in which the length along the curve is 35 inches. The Kopet Dagh urial appears decidedly to be a distinct form connected with the typical Ovis vignet by the Punjab race of that species. On these grounds I regard it as a local race, rather than a species; its name will accordingly be Ovis vignet arkal (or perhaps arca/). * Tydekker, Wild Oxen, Sheep, and Goat, p. 173. { Lydekker, Note on the Wild Sheep, of the Kopet Dagh, Proc. Zool. Soc., Feb. 3, 1903, pp. 102-3. PLATE 75. Fic. 1. Skull (posterior view) with incomplete horn-cores from Ovts vigner arkal of Anau, period Ia. 2. Skull (posterior view) of Ovis aries palustris, Anau, period I. (Compare plate 83, fig. 2.) 3. Skull (lateral view) of hornless sheep of period II (restored). gente pants ¥ Cee PLATE 76. Fic. 7. Horn-core of Ovts palustris form. Fic. 1. Median part of carbonized horn-core of Ovis vignet arkal, 8. Metacarpus of Ovis vignei arkal. 2. Basal part of another horn-core of same. 9. Metacarpus of goat. 3-4. Horn-cores of female individuals of Ovis vignet 10. Basal part of antler of Cervus. arkal. 11-13. Horn-cores of antelope (Gazella subgutturosa), 14. Horn-core of goat. 5-6. Horn-cores of large-horned breed of sheep. CAVICORNIA. Crt Now, Anau lies just at the foot of the Kopet Dagh and within the area of distribution of Ovis vignet arkal; it is, therefore, most probable that we have to do with the remains of a mountain sheep still living in the Kopet Dagh and not with those of an inhabitant of the broad steppes of Persia and Asia Minor. We can, therefore, reasonably exclude the Ovis orientalis, 7. e., the steppe form of the wild sheep, from further consideration and say logically that we have before us Ovis vignet arkal Lydekker. As I am not able to give a direct comparison of our horn-cores with the original skulls of the Kopet Dagh sheep in the British Museum, I shall take for comparison two skulls of Ovis vigner that are preserved in Berlin and London, and shall contrast them with the measurements of the Ovis ophion of Cyprus and the Persian Ovis arkal seu orientalis. From these measurements it results that the agreement is closest with Ovis vignet cycloceros of the Punjab, and that the difference in dimensions of the two recent Ovis vigner skulls from Ladak and the Punjab is more noticeable than that between the skulls of Anau and all the others here brought into comparison with them. Table of dimensions (in millimeters). North Kurgan, Anau. Punjab. | | Ovis pl oie | era oh lea 5 . us as VES Skull. secon e phion, | vignet, | arkal, —21feet. | —15 feet. | + 26 feet. Mus. : Mus. No. 666, Mus. Berlin, | London. | London. Paris. Greatest width of frontal bones......) 125 Lege Wr | ee Ua) 125 145 134 WeAStE WILE VOI Satie) ia a etter ce 4) «0 | OSM ees: auth ee 94 93 Cir 98 Front line between: base of horn-cores.| 2, Tate se 17* | 22% 18* 25% Back line between base of same...... | re tad oe 89* 103* 133* 118* Greatest width of occiput........... al Waotaes ae 78 79 96 | 88 Weastewidth: Of SAMC..c.6c.¢.c.00 2056 | 55 oer Arlee Sr 60 62 58 Greatest height of same............. om, era acanes 56 64 64 | 54 Wereteeroit Of SAMIE = 5.65/86. wine: sue 43 42 47 52 46 Greatest width of parietals.......... 91 82 80 89 83 Measts width of Sameiio...... Bm | 8g bp oO < = CS - e ae = ar 5 Skull of Ovts aries palustris Riitimeyer. ape ee | 23 : = oO | g ge) oe ) pees) oes a ae ae GE ae ahd ©} yn ey | se | seg|eee| EG | dee] & ee ae ers ed rh a gzA 7) iY Sz n S R= Oo x = a =e == =— = | = | —— 178= i Erse 110 80 110 | Distance between points.................00. 248." il voirae I pees Ae 169 Length | Length Width of : . Entire of of vertical Lower jaw of adult Ovis. "fe ee part be- | length. | dental toothless hires series. part. BS North Kurgan, Anau: | wm £5 Fett vcd wick ie 0s teenies Sie OR at aaa eae *180 (?) 76 45 *60 (?) = 15 feet ite. car ca hee ce eee nee eee Ree 78 45 see “6 feet 0. Seve earache on Aer cere ete Ties aes 54 +20 feet ccgptiia piv sauaists bx eae errant eee nena eee eee 167 71 43 59 +23 feet a3. Bi 2 aes Aen eee ee eee 152 71 42 46 +24 feet .. ete tguae ee ot ree aia ee bie ha arene 170 a1 48 50 Outs mignet, Ladaks Mus, London: see ee See 190 81 47 56 Ovis aries palustris, from Ireland, Mus. Paris..........| L73 71 42 5! | * Measurements on incomplete pieces. The appearance of the horn-cores of the turbary sheep and the goat-horned sheep mentioned in the eneolithic and copper-culture strata of Anau may, there- fore, easily be considered the remains of an autochthonously derived domesticated form, especially because of the quantity of transitional forms to the large-horned sheep. * Duerst, Versuch einer Entwicklungsgeschichte der Hoerner der Cavicornia, Frauenfeld, 1902. Sur le développement des cornes chez les Cavicornes. Bull. Mus. d’Hist. Nat., 1902, p. 197. CAVICORNIA. ans Table of dimensions (in millimeters). { Circum- | yonee Transverse Horn-cores. Length. ference tia a diameter at base, | “ameter | at base. at base. North Kurgan, Anau: Ovis vignet arkal Lydekker : pr CC UM cet ei tathe Rciant, vor wirene Aaa AEM ATS nee i 250 180 || 62 45 SPER WTSTE Saws Gia Chae tA OA Rea TAR ee ec ee ene 200 70 53 Tre Ol COLRM MERE Pars av atr cic oie rere en TO isl arose. os oe ane 170 67 43 Probably domesticated, small-horned specimen, form- | | ing connection with the real Ovis palustris: ; | " Fade) boa) Aes of NF gems NH CGA ee Wut ee. S7 ae r 110 94 2 20 SBOXON abe BINS ahs hoe ASS Or eNO CIEE Cate ae 130 120 41 26 Stee SLCC CMa Meee iy segs nea uitanehaaitebein sath ke shesiaiore fei 106 93 2 22 a POLIO BS Salotr SSO SO OR CCR 0 cate Ce Cie Sree 117 112 2 27 Ovis artes palustris Riitimeyer: ( 117 73 25 16 S AOIPM estooeas “fh bea Ch cin’ ORD CHGS Mee merepemraes cer ~ 97 72 2 14 (| 114 73 25 14 SEG Vs OS Sc he So EO OREO OER Ce EO 75 72 25 14 Types of Ovis aries palustris: England: WeaaValleya(collCornen\ens ose re eins cee 85 70 2 I4 Londonwall, Roman times (coll. Corner)...... | cate 105 32 a ( 105 85 36 16 Vea alluvittm; The Mills; London............'.. 70 54 15 15 Germany: Naegelstedt neolithic, Prehistoric Museum, ( 130 100 a0 22 JSYeSa Dal cnc, Seki Sema eb is BCE Mt eee ae ‘ 110 102 39 21 wciitossberes leaelenetan yas eset suse « clarelen oa iets 100 75 26 ars France: Grotte dtPontil, Nat. Hist. Mus..........4.... 120 100 40 20 Pile-dwellings of Paladru, Paris............... 75 80 23 16 Switzerland: ScHiais =neGltciiew: Delile sryeie cs heeds Se wv nee. 78 100 37 18 al Seer CCN premier te eres sielrcre car's = «ies 80 75 22 19 It may be objected, however, that the turbary sheep might have been imported from elsewhere in the culture period II, and that the different large-horned domestic sheep were only the products of crossing with the wild sheep or with some other large-horned domestic sheep, such being found in the pile-dwellings of the Lake of Bienne, as Ovis aries sluderi, or as products of the crossing of the wild Medi- terranean mouflon on Ovis palustris. The decision between these two possibilities would not be fully possible had not Professor Pumpelly conducted the excavation of the bones with the greatest care throughout. It is of the greatest importance to us that we know exactly the depth from which each fragment of bone was taken; and it follows that we are able to make the following statement concerning ‘the position in the stratified series, and consequently the relative time of appearance : Ovis vignet arkal Lydekker occurs practically throughout all the strata; the best-preserved piece comes from the lowest layer at — 20 feet, the second at —15 feet, other small pieces of horn-cores at +18 feet, +21 feet, and a good piece again at +26 feet. Therefore, this animal was hunted and eaten by the inhabi- tants of the North Kurgan down to the copper period. 376 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. At +8 feet we find for the first time a well-preserved core of a domesticated sheep, which belongs to the already mentioned large-horned form. It is the same as the Ovis vignet horn-cores, but a little smaller. But in the same layer there occurs a second core, which is not derived from the same animal, as it is notably smaller, more slender and also of two-edged form. Further core-pieces occur frequently in all the following layers, the best-preserved of these cores being found at +20, +25, and +26 feet. These cores disappear then entirely and from there on we meet only with the very slender cores of Ovis aries palustris, horn-core bases, with both attached frontal pieces and middle pieces, being very common. The cores of Ovis aries palustris begin to appear, however, before the + 28- foot layer, 7. e., during the occurrence of remains of the large-horned form. ‘The calvarium already mentioned, shown in plate 75, fig. 2, comes from this layer. Its horns, which are still rather stout, indicate that the now-developing turbary- sheep form is yet in the process of evolution. The turbary-sheep form continues until the uppermost layers of the kurgan are reached. It becomes considerably rarer, however, above +33 feet and about that time there enters a hornless sheep, which we shall presently consider. Thanks to Professor Pumpelly’s care these closely accurate stratigraphic relations afford us weighty evidence that the Ovis palustris is autochthonous in this part of Turkestan and was bred from the Ovis vignet of the Kopet Dagh; for it is clear that here a few feet of culture-strata represent centuries during which a very great transformation could take place in the Anau sheep. In the tables on pp. 374-375, I have brought together the measurements of the horn-cores and the cranial remains from the layer +28 feet, in comparison with such specimens from European culture-strata. It is easily seen that the measure- ments agree with each other, serving merely to strengthen the impression made by the agreement in form. THE HORNLESS SHEEP OF THE COPPER PERIOD. (See plate 75, fig. 3.) In the former section it was shown that at about the +33 to +34-foot culture- layer the horned palustris sheep was crowded into the background by a hornless sheep, which suddenly appears in numerous individuals. As is well known it is difficult to find among the hornless sheep such differences in cranial structure as to permit a determination of the breed. ‘The skulls of hornless sheep and of goats are distinguishable from each other with difficulty when the sutures of the parietal bone are no longer recognizable. It is, therefore, impossible to determine the exact relationship of the hornless sheep of the II or copper-culture period. The dimensions of the skulls of different hornless sheep, brought together in the following table, show a perfect agreement with the hornless sheep skull from a turbary at Abbeville in France, preserved in the Museum of Natural History at Paris; but, on the other hand, the agreement of the measurements with those of the skull of an Ovis platyura bucharica ewe is very strong. CAVICORNIA. 377 Table of dimensions (in millimeters), 5 ee ie oe &> Ovits aries, recent Seagal Ss 83 & specimen. Some SH | Se. Anau aod Ss Syeae , ro Ss wo = Skull. +34 ft. |< 8 oe SE A SE ey Cae. sues Ssg Be 3 | cogne. Ireland. 1S S S Greatest lencthrotp basen ca accent oe > ile s 185 okie 190 214 212 197 Lateral length of frontal (bregma to orbita).. .. 57 eee 63 80 74 75 Length of molars of upper maxilla........... 35 39 45 48 43 44 Kenguaror premolarsas 6. o5. seks ce erer sou: F 26 oT 21 25 19 19 Sagittal length of frontal bones............. : 82 7 80 2 g2 82 Erie OnRIgit ge cae. daa a sith ed 5 eee % 30 Sete oly) Bg 2 2 BALICL ALS WiC himeentey oie Sockets atact eel cans a es 59 52 61 71 53 65 Srealect Nersitt Of SKU mere mia are sete os Sap fH) Meee: 86 104 72 78 Greatest neightrOr OCCLpPilt ae vein teniene te «| «ie. 2 2 59 48 48 Weastaneic nt Ofcame sarees oe tee ete ele nck a aust 37 38 43 33 2 (sreatest width of occiput.) 0.06 vy aioe: wee os 65 66 76 67 67 east awidtiol cameras \isteree eters crs as, 42 | 43 55 44 41 eastawidth ol 1rOmbearienrs + tira aaeye pte, «et 65 ene 66 68 63 69 Greatestuwidthvorsamess ters. 6 as tae caertee. 104 108 109 119 117 119 Distaneeibetween orbitals. a. 6.20 2. 2266s +4 sie see 75 71 77 76 77 Width of palate behind molar y....5......... 48 39 45 58 50 52 Width of palate in front of premolar 1........ 2 AG 2 26 2 2 | This apparent resemblance does not, however, permit us to assume a rela- tionship to one of these forms, for it is readily seen from my former investigations into the influence of horns upon the shaping* of the skull, that the absence of horns produces uniform characteristics and that while considerable variance may exist in the absolute craniological dimensions, the relative dimensions always remain the same. Now, what can this hornless sheep form be, and whence can it have come? The bone remains give us no information on these points; and we must, therefore, resort to deduction and inference. Let us first examine the recent races of sheep of Turkestan. The Central Asiatic steppes harbor only two races of sheep, which are generally designated: (a) The fat-buttocked sheep (Ovis aries steatopyga Fitz); (b) the fat-tailed sheep (Ovis aries platyura Fitz). Ovis aries steatopyga is characterized by a posterior overloaded with fat, which on the buttocks projects upward in the form of a fatty protuberance which is split in the middle. The ram of this breed has horns of medium length which are thick and strong at the root and grow narrower towards the blunt point. The horns, without rising above the crown of the head, form, in winding, a double snail-shaped revolution back, down, and forwards. In the ewes and the wethers the horns are smaller and weaker, and curved only backwards and forwards. There occur at times four-horned and even five-horned rams in this race of sheep ; and, on the other hand, we find here and there hornless females. * Experimentelle Studien ueber die Morphogenie des Schaedels der Cavicornia. Vierteljahresschrift d. Naturf. Gesell. Ziirich, Jahrg. 1903, p. 360. 378 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. The second race is the fat-tailed sheep (Ovis aries platyura), whose long, limply hanging tail, reaching to the heel joint, is surrounded by a mass of fat, which has, however, no great size. The rams of this race are horned; the ewes for the most part are hornless. The horns of the rams are not very long, nor particularly thick; they are three-sided, with rounded edges, of which the inner one is always sharper. ‘They rise slightly above the crown of the head and wind sideways and backwards from the root, forming then a simple helical turn down, forwards, and upwards, while the points turn somewhat inwards. In the neighborhood of Anau the Afghan Maimene breed predominates. It is a sheep of excellent flesh of large growth. The wool is long and coarse, and the tail long. The animals are in part horned and in part unhorned. In the other parts of Transcaspia the fat-buttocked sheep (Kurdjak breed) predominates. If now we visualize these forms of sheep, we can express the following ideas concerning the derivation of the hornless sheep of the culture-strata of the copper period of the North Kurgan, though these ideas can be but speculations. The hornless sheep may be: (a) a female of the previously occurring form, which we have designated Ovis palustris and whose horns had reduced in size; or, (b) a hornless race introduced from somewhere else with the camel and the goat. The sudden appearance of several hornless frontal pieces argues for the last hypothesis. As we shall see, there appeared another domestic animal at about this time or in the + 32-foot layer—the camel, of which no trace was found in all the earlier layers which have been so carefully searched. At the same time we meet another domestic animal, the goat (Capra hircus riitimeyert Duerst). The camel, the hornless sheep and the goat are, however, the animals found in the later South Kurgan, where there are few traces of other sheep. Table of dimensions (in millimeters). Length Superior Width at Length of | Diameter of z : width. collum. articulation. | articulation. SCAPULA: North Kurgan, Anau: = 8 feet..cc nic hee eens Se sees 21 35 23 — 8 feets 22 unser meee ets 19 a2 21 ~$ feet. j<. hele eres ce ae 17 27 19 =" fet: nee oe ere ee 17 2 19 =i! feats iene en nro ero ins 19 30 20 4-1-7: feet 52: setae ors we 18 38 20 +19 feet. ieee eee =% 20 33 19 4-225 feet .8e.es ee Bee a 29 32 ie 4-23 feet. 5 leacks cede od, 21 32 20 +26 Teetiw:. ose cee ee $e 16 21 22 1-26 Teetvenk Gaderer a 17 28 20 Thala sel Calder, arbi wi Ae ae 21 35 oe 32 Leetonia ee 17I 110 20 35 22 $232 SECC cee ac cca eee aie 18 Bo 21 33 feeto nen ee 2 34 24 + 92: feet ey. sae eee 20 Bt 19 32. fet. tastin gee eas arate 22 33 25 +132) TREE ss pons, es ee 5 a 28 18 a3 2° fGGtE ae ee eee: 18 2 20 32) feet tern: sree ae ee 7, 27 20 | Ovis artes palustris, Nalps, coll Duérshet ee ee eee | 172 115 20 29 21 CAVICORNIA. 379 Table of dimensions (in millimeters) .—Continued. : Width Width Proxi- pe Me- oe Distal | Distal | of in- | of ex- Length. | mal ma" | dian |. 248 | ©2182" | diam-| terior | terior width | diam- width diam-| idth. eter. | troch-| troch- Wetter, fimeCtek, ; lea. lea. HUMERUS: North Kurgan, Anau: NOU 206 6 — Go LCC Rea crsie ens 138 38 43 18 20 30 Lz 13 INOWI 304 pe 8 1eCin oes aos Arye et Mae Be 26 it) 14 INOMLOOS5s 1222-05 feCU ers 30 19 14 INGWO20) -1-22).5 feet. enacts 27 17 14 INO TA S72 Selb nn «oct ae 27 17 1 NOMS 5H wictno LCCLe rien. 32 20 15 INom10325)-- 26 feetaannae. 32 20 13 INOm WOO.8 1-2 ONL Cen ae nye ce 30 18 14 NOMI O 7250-2 Or LeCte nares. 30 17 14 Turbary sheep, Schlossberg...... ae Sheed Poe OR Arsat Wes. aul ee 18 14 Recentisheep, Germany rie «i. 135 40 45 20 21 2 25 18 RapIivus: North Kurgan, Anau: NOP TeO8 4 —S teeta sass > Pa 35 17 5 dle’ Nos 1200)--—8 feet &. 3.30.00: sees 39 17 po tel eet wise a NOMS 30) eT Bulee@t ser. sae s.5 140 oF, 19 T4 8 29 14 INOre (Ol a127 215 TCC Came. eras 6 156 30 14 16 9 3I 20 INGE 1020.51 26) feet cm» atsncerais 160 30 14 14 9 28 19 INO LOT Tse 2Oukeete a2 ca. om aes sas. cl aistere [pletecens 2 18 INOS. 25°57 2s feet Macny eaten a3 14 15 9 See all pee INO OOSs.s2nreet.. oc tes. 30 14 14 9 Seat ae No. 943, + 32 feet, old animal. rites sein 7 9 30 20 | South Kurgan, Anau: SG Uh YS Ree in in eee 157 32 15 16 8 30 19 Londonwall, Ovis artes palustris, | 4142 33 15 16 9 26 17 | DDreConmer OUGCOR ise vada... (157 30 14 17 9 29 19 METACARPUS: North Kurgan, Anau: | Now 1202 — 15 feet Sack 23. 2: 26 18 15 12 fe | No. 12026, +15 feet........ 25 18 tere ef Te NOMEPS hI 2S feets 285.2501 < Le sie 22 16 INOS 04 Osuring 2a CCL. 2 clever. etl 25 Ty oe INGRO22, ato a 2 eleebers. . si sistet, 20 15 Aare ae NOP O43) 132 1eCles, awe See See Mey eel becesiees Pree 22 14 | Walthamstow, Ovis aries palus- III 19 14 sus 8 21 1 USHER ee eee ee cee 115 20 rr II 8 21 13 Ovis aries studeri, Lea alluvium. .| $135 26 18 17 II 30 16 | .... (161 29 20 UG 12 31 18 FEMUR: North Kurgan, Anau: INOmI TOO; 26nleetmaee saree 43* 19 ucrehocalpecstare’s INGIETOGS a 2OMNeCeL tates : toh or BF 45 INO#O62 5-2 7-1eeteniyac 2 Care 48* | 22 Me ee Ape sects MU rceeers. lat. te IN OMOSO 1 Sacleets yo tyes Sak AS lee 20 19 16 aS ry | Mac ING.3932, + 32 feet... aes etal oe since rs Mucaene re 35 43 Ovis arves, German breed........ PANG: 6259) 31 20 18 48 54 Ovis artes palustris, Nalps, coll. WDWErSU erat ity haere chee 185 AQ™ 22 18 15 36 44 TIBIA: North Kurgan, Anau: INGRY29 Ga —— oS teeta acter ac: 25 17 No. 1484, +23 feet......... Poe, [eva 36 21 INOH104357-1-26 feeti sy... « 44 45 Soa cee INOMI1034,---20.1eetemn eas. pec Wao 23 18 INO 688207-27 Teets a. os ae es 21 17 South Kurgan, Anau: INO 5 SO are. way tcc hee oe 34 38 Ovis aries palustris Nalps, coll. DILCLS Uae teeaere star cis a creicie« too oke: 220 38 41 26 19 * Proximal width on the caput femoris. 380 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. The latter alternative does not appear to me to be sufficiently justified, and I incline rather to the first idea, for the following reasons. We find in the layer at +30 feet the frontal piece of an adult animal with a little horn-core of 3.5 cm. length and a circumference of 7.3 cm. ‘This cranial piece must represent a transitional form to Ovis palustris, especially as no mistake as to the age of this animal is possible, because of the presence of a parietal piece which is connected with the frontal by the sutura coronalis. If this cranial piece had belonged to a young animal it would have broken open along the sutures when the skull was crushed, while in the old individual as a matter of fact the suture is still so firm that the bones would break before the sutures would open. Thus it seems that the long-tailed Ovis palustris form may have given rise to the long-tailed sheep (Maimene breed) which is still living in those regions, provided always that the formation of the fat tail, which was probably patholog- ical, did not originate until after the distribution of the turbary sheep to Europe, which possibly happened towards the end of the eneolithic period of Anau. The occurrence to-day of hornless female animals among the fat-tailed sheep and turbary sheep renders this explanation more probable. However, this view rests only upon speculation, for direct proofs are not to be had and probably never will be. In the tables on pp. 378-379, the extremity bones of the sheep of the Anau kurgan are brought together and compared with some accurately determined ex- tremity bones of subfossil or recent sheep. In these one can see that the larger wild sheep or its direct descendants occurred in the lower layers, while in the middle and upper layers the small palustris sheep predominated. Capra hircus rutimeyeri Duerst. (See plate 76, figs. 9 and 14.) The goat, of which we find the horn-cores and extremity bones among the bone remains from Anau, belongs, as already stated, in the uppermost layers of the North Kurgan. Really typical and well-preserved remains are very scarce. Of these there are some horn-core pieces and two perfectly preserved metacarpi, as well as the fragment of another. In these we can recognize a small short- horned goat, such as lives still, in a slightly differentiated form, in Central, Eastern, and Southern Asia, as well as in the Malayan Archipelago. One of the most primi- tive forms is without doubt the so-called wild goat of Crete, which is probably only a reversion from the domesticated to a wild state, very similar to Capra egagrus, and in which is embodied the exact type of the goat of the pile-dwellings. M. Evans has published from his excavations at Cnosse (Crete) of the second palace (about 1500 B. Cc.) a very perfectly preserved relief in faience representing a she-goat with her young.* The horns of this animal are much longer than those of the recent goat from Crete, figured in plate 78, fig. 5. The horn-cores differ from those of the sheep in the greater height to which the inner cavity extends, which leaves room for only a little dense substance at the point of the horn-core. * Salomon Reinach, La Créte avant |’Histoire, 1’ Anthropologie, 1904, p. 265, fig. 7. CAVICORNIA. 381 In their dimensions they are similar to those of the palustris sheep. A horn- core, unfortunately preserved only with the basal part, belonging to an old indi- vidual, and marked as an old he-goat by distinct frontal bumps, is the only one that shows any larger measurements. Table of dimensions (in millimeters). | Proxi- Me- : Longi-| Trans-| ,. Length, |e) mal) ME | dian [Distal BS! cadinal verse | eum Detain diam- | width diam- | width. ater diam- | diam- eehaee Veter. eter. DRpGLeE, Weeler. ial | | HORN-CORES: Anau ) Female cords ene TSM te ee! ees ack it -scs es 20 14 73 | Penile is Gan wares EOE re Abaco ae Naan es Piety MSE 19 85 Wale crepes sk. ie ssrcs Sele Pe Gi omnes ll Geese Mek ot ai al Oat ieee | re eg 40 28 |ca.120 Capra hircus, male, Tur- kestan, Mus. Paris.... 250 SCAG ot tees, alec Re eee eo ee 54 39 150 Capra hircus, Malay Archipelago, coll. by Hombrone and. Jac- GR TVSNO EA aia oe 100 ete Were vcaigerurce Theres Ms Gs fe ota « 26 15 74 Capra cretensis Brisson, Crete, Mus. Paris. 120 Ce Le MIE oreSs taal ros iat Wetsiohs s alyteaaces 22 13 65 METACARPUS: Anau Xda. Mewes eee 102 26: 16 15 9 Pig) 16 TENGE ONG: Rene re 6 re 98 20 14 14 8 23 13 Anau City, M.S. 1v, — 16 to —17 feet, recent....| .... “OOo Berke Cee 9 25 15 stan Gross Czernosek, Bohe- mia, Mus. Teplitz. . 99 20 14 x2 8 24 I4 Stockholm, Kungstrad- garden (King’s Gar- dens): : Subfossil goats*.... 102 23 Hours 16 eee 23 Oe SMAI poate Ie ees eames at Subfossil sheep | after Kinberg, ( 107 22-5 15 27 SAGE: yf el oe . 98 22 14 24 ( 102 18 ane 10 ia 20 Peco Se eet Pee Oh ft Schlossberg, Ovis aries DOLUSENUS arr eae ti 103 18 12 8 7 9 8 Sardegna, Ovis musimon, COUP Duerstem ree ee 135 28 21 14 12 25 17 *Kinberg, T. G., ‘‘Ossa metacarpi et metatarsi, Ovis et Capre.’’ Ofversigt Kongl. Vetenskaps-Akadem. Forhandlinger, 1869, pp. 359-433. Although Riitimeyer* states that only the hoof phalanges and joint-surfaces of the bones are useful in deciding between the goats and the sheep and that the goat shows a deerlike delicacy of the bones, later investigationst have shown that the metacarpalia and tarsalia are good characteristics for a diagnosis, since, contrary to the general view as well as to Rtitimeyer’s assertion, these bones are much broader and shorter in the goat than in the sheep. This appears clearly in the above dimensions, since the distal width of an equally long metacarpus of a goat and a palustris sheep is three times as great in the former as in the latter (27 mm.: 9 mm). * Fauna der Pfahlbauten, p. 127. +Cornevin et Lesbre, Caractéres ostéologiques differentiels de la chévre et du mouton. Bull. Soc. Anthrop., Lyon, 1891. Kriz, Martin, Die Hoehlen in den maehrischen Devon-Kalken, ete. Jahrbuch der k. k. geolog. Reichsanstalt, Wien, pp. 443-579. 382 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. ANTELOPINZ. Gazella subgutturosa Gueldenstedt. (See plate 76, figs. 1 I-13 and plate 84.) Several of the best-preserved bones of all the layers belong to the Gazella subgutturosa Gueldenstedt, the small but beautiful antelope that still lives in Persia and Turkestan. There are both horn-cores and teeth, parts of lower jaws and extremity bones, which make the presence of the animal very evident. It occurs frequently from the lowest layers below the level of the plain to the summit of the kurgan. It is clear that this swift game was successfully hunted by the inhabitants of culture period I of the kurgan. Assuming with Mucke,* that the wild animals of such an early period would show no great fear of man, it nevertheless seems strange that they could be killed without the aid of the dog, and it is probable that a dog, if not the same as the one we find in the higher layers, existed also at the earlier period, although no bones were discovered. The great hardness and the absence of cavities in the horn-cores have served well to protect the remains of this animal from destruction by tooth and time. The color of the horn-cores is different from that of the inclosing earth, ranging from dark-red to light-yellow. In the following table are given the dimensions of some of these in the order of their position in the kurgan and in comparison with the measurements of the head of a modern individual. These animals are also represented in the sculptures of the ancient Assyrians (plate 84). Table of dimensions (tn millimeters). ; 2 Width of Circum- Diameter Circum- Length. Horn-cores. Length. | anterior to Fraeaeies ference | frontal of ae ference posterior. lameter| on base. between sheet of horn- horn-cores. ‘ base. North Kurgan, Anau: =~ 2 feet... sci 45 vemme ees 160 30 22 89 14 m= Te TOCE ic Misctiee + bin wee eet 155 30 22 go elete —(9 Leet s.2. magic we bye eee 175 31 21 90 8.5 Te@tinc .t.ad.50 pee 145 38 27 105 Ais teeti... ocak acs is Mame 165 34 24 100 seo B7- TORE enews oy enn 165 32 24 92 27 feet sk Keen tee ae 210 at 23 90 -30 feet. xa siieseare eee 170 34 24 94 Gazella subgutturosa, stuffed male adult specimen, Mus. Betn fase ee eee eee eta ia ste 34 25 ates 15 CERVIDA. THE STAG OF PERSIA. Cervus sp. [maral Ogilby{(?)]. (See plate 76, fig. 10.) This great deer is represented by the remains of an antler. The main branch has been knocked off with a sharp instrument and only the crown or burr remains. The circumference is 25 cm. In the European stag from the Schlossberg I found 24.5 cm. in eighteen antlers and 19 cm. in twelve. It is probably C. maral Ogilby, but for lack of material this can not be proved. *Mucke, Urgeschichte des Ackerbaues und der Viehzucht. Greifswald, 1898. Gray, Cervus wallichit. Proc. Zool. Soc., 1850, p. 228, 1840, p. 11. CAMELID. 383 CAMELID&. Camelus sp. [bactrianus Erxel (?)]. (See plate 73, figs. 10 and 11; plate 77, figs. 10-12.) Bones of the camel are found only in the highest layers of culture II or the copper period of the North Kurgan. The phalanx secunda, No. 615, came from between +26 to +31 feet, and the fourth vertebra cervicalis, No. 1062, from +32 feet. No remains of the camel were found below these layers, though we naturally find them again in the shafts of the much later Anau citadel, where they can be but a few centuries old. It is therefore very probable that the camel was imported as a domestic animal at a much later time than the age of the lower strata of the North Kurgan. However, even the complete absence of bones of a wild camel in the layers of culture I is no reason to conclude positively that this animal was not then living in a wild state in this region, for Przewalski found it still in a wild state near Lob-Nor, south of the Tian Shan. Table of dimensions (in millimeters). . n 7) K bh h © 3 a A} Se ie i - _ 7 i ha a Bye Saeco Ne a(n s|/si/3/8ie|/e]os8|] 8) a.) 82) 83) = Gn Visca ins a ek i (= Ge ee g Se at ae a 1 — Lo} ow » v HS et Oe cay rs) 2 NP ca Pita ae RS a ah G9 RC Ms ee wage eS eee har Poke bog Fey Sr ane es bo ere he Aan hoe 5 © eae heer Si POA I 5 eS = = S PHALANX I: Anau City mosque shafts— 9 to—11fect.j104 | 44 | 34 | 21 | 22 | 37 | 28] .... | Camelus bactrianus, anult (Mae, Bern). «| 91 1740 | 32 | 20 [19 | 33 | 24] .... | coon PHALANX II: North Kurgan, Anau, eo letOnt=oOmeet.y nle74. 35 13027 tio. 1 AT | 19 Camelus bactrianus, adult (Mus. Bern)....| 56 | 28 | 21 | 21 | 14 | 30 | 11 FourRTH VERTEBRA CERVI- CALIS: PATIAI SNOUT OO2:.72 ste = Bi Neen lh gd hee age esc iaeal er 138 58 108 741 69 62 Camelus bactrianus Docs GUAE Mets e ed Mall or Saal gic] en oceh ea IRS 120) SOM ICO 69 68 60 It is impossible to determine exactly the species to which the Anau camel may have belonged; but historical reasons and considerations of geographical distribution make it seem probable that it was of the Bactrian race of camel, and therefore two-humped. This is only our opinion, however, for the differences between the skeleton of the one-humped dromedary and the two-humped Bac- trian camel are very slight and not perceptible in the well-preserved bones of the kurgan. The Anau camel was certainly a large animal, as will be seen from the preceding table of dimensions, where it is compared with the skeleton of a camel preserved in the Museum of Bern. Fossil remains of the camel have been found in the Siwalik Hills of Northern India and in later Pleistocene deposits in Lutschka, near Sarepta, on the Volga, north of the Caspian, the latter having been published by Nehring under the name 384 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. of Camelus knoblochi. Stefanesku discovered in Roumania the bones of a camel which he describes under the name of Camelus alutensis. Pomel describes Camelus thomasi from the Pleistocene of Algeria. Hence it seems to be highly probable, as was first pointed out by Nehring, that the one-humped and the two-humped camels were developed in different countries; and that while all descended from the ancestral form of the Siwalik Hills, one branch, reaching Western Asia and Eastern Europe, formed Camelus knoblochi and the Camelus alutensis, and probably also the domestic race of the camel found at Anau. This branch was two-humped; while the other branch, passing like the Indian buffalo (Bubalus paleindicus) into Africa, has formed the one-humped variety of Northern Africa and Arabia. That the camel was domesticated in very early times is proved by the repre- sentations and sculptures of the Assyrians and Persians. In the later Persian monuments of Persepolis and those of the Assyrians of Khorsabad and Nimrud, we frequently see one-humped dromedaries; but it is only on the black obelisk of Nimrud, which is inscribed with an account cf the campaigns of Shalmaneser II, King of Assyria from 860 B. c. to 825 B. c., that we see two-humped Bactrian camels, where several of them are represented under title of payment of tribute of the land of Musri. The land cf Musri, which belonged to King Asu of Gurzan, or Gilgani, was situated north of Lake Urmia, in the neighborhood of the Kara Dagh and Mount Ararat. Since our results seem to show that at the time of the oldest culture-strata of Anau the wild camel did not exist in this part of Turkestan, it is possible that the domesticated animal was imported with the goat from Bactriana or from the Iranian plateau. Ordo PERISSODACTYLA. EQUID A. Equus caballus Linneus. (See plate 77, figs. 1-9.) One of the animals of which we find the greatest quantity of well-preserved bones is a relative of the horse tribe. From the deepest layers, —24 feet, of the oldest period to the superficial remains of the latest habitations of the North Kurgan, we find great quantities of these bones in all the strata that have been opened. Therefore the equine animals must have been very abundant throughout the life of the kurgan. Notwithstanding this great quantity, it is not easy to form a picture of the equids to which these osseous remains belonged. We have only very few data concerning the bones of the Post-Tertiary and subfossil horses of China, Mongolia, and Central Asia,* which have been only slightly increased by Tscherskif for the Siberian horses. Nevertheless, despite the defective knowledge concerning the prehistoric horses, Central Asia is looked upon by many authors as the cradle of the European domestic horse, as well as that of the human race in general. Nehringtf has, it is true, proved that a domestic horse was formed out of the diluvial horse of Kurope on European ground, which took part in the creation *Gaudry, Journal de Zoologie, Gervais, 1872, t. 1, pp. 300-302; M. Wilckens, Nova Acta, Leop.-Carol. deutsch. Akad. d. Naturf., 1888, Bd. Lu, No. 5, p ; } Tscherski, Mémoires Acad. Imp. St.-Petersburg, VII série, tome XL, pp. 257-383, 1893. { Nehring, Fossile Pferde deutschen Diluvial Ablagerungen,. Landwirtsch. Jahrbuch, Bd. x11, 1884. Fic, 2a 3. 4. D5 6. Right part of upper maxilla, Eq. caballus pum pellit. Anterior part of left branch of lower jaw of same. Metatarsus of horse of II period. Metatarsus of horse of Ib period. Metacarpus of wild (?) horse of Ia period. Metacarpus of horse of II period. EIGo de PLATE 77- Metacarpus of wild horse from Solutré. Phalanges I-III of horse of Ib period. Phalanges I-III of horse of II period. Phalanx II of camel. Vertebra cervicalis of camel, Phalanx I of camel. 2 PLATE 78. (Mus. Paris.) Fic, 5. Skull of a female wild goat from crete. Fic. 1. Basioccipital bone of Bos namadicus Falconer and Cautley. 6. Short-horned cattle. Skull of a specimen from the 2. Remains of a long and large round horn-core of Bos kurgan of Bizino, near Tobolsk (western Siberia) taurus macroceros. (Mus. Paris). 3. Proximal part of radius with ulna of Bos namadicus. 7. Remains of basal part of long and large round horn- core of Bos namadicus. 4. Distal part of radius, , (Figs. 1-4 and 7 from period Ia.). y £ 5 - ; . py - ‘ He a PUA Eene. riche Bibs iy arte pence oe < i 4 4 2 Fic. 1. Egyptian dog, marble, Mus. Louvre, Paris. 2. Egyptian long-horned cattle, Tomb of Manofer, 2600 Bb. c., Mus. Berlin. OF THE UNIVERSITY OF ILUNOIS LIBRARY ‘a+ n ’ ORDO PERISSODACTYLA. 385 of the heavy draft-horse of Europe; and now Kraemer* comes to the conclusion that the horses of classical Rome and Greece represent a cross between the heavy European horse and the Asiatic type. Notwithstanding the very plausible fact of the domestication of the European wild horse, however, it can not be contra- dicted that this wild horse itself could have come from Asia. In considering the horse of the Anau kurgan, it is primarily worthy of note (1) That the horse from the lowest to the uppermost layers is represented by a great quantity of bones, to an extent which in the lower layers is only equaled by those of the bovid; (2) that in these bones we can recognize only one variety of horse, which thus occurs in the lowest layers with wild animals only and in the higher strata with the other domesticated animals; (3) that the percentage of the bones of the horse, as compared with those of other domesticated animals, also increases in period Ib. ‘This last fact permits the conclusion that the horse came to the table of the inhabitants more often in the later than in the earliest period of the development of the kurgan civilization, from which we might next conclude that the horse was then easier to catch and had, therefore, become tamed or domesticated. It is not possible to assert with logical certainty the correctness of our con- clusion that we have here, at least in the upper strata, a domesticated horse, as we were able to do in the cases of the bovid and sheep through a study of the skeletal remains. I hold that no one is able to determine with certainty, from the study of a few bones of a fossil or a subfossil horse, whether the individual was wild or domesticated. There are wanting in the case of the horse precisely the criteria which we have in the bovids, where in consequence of stabling or of restriction of freedom of movement, the substantia compacta of the bones is thrown into the background in favor of the spongiosa. Again, we are not able to base a distinction between the domesticated and the wild animal on a change in the skull, as we do in the sheep. On the contrary, the mode of life of the horse, especially among inhabi- tants of the steppes, remains the same as in the wild condition. Harnessing, and the use of the organs as in the wild condition, insures the stability of the bodily form and of the skeleton; and the influence of the weight of the rider carried by the animal is not further perceptible in the bones. Consequently, in the horse of a primitive people, such as were the inhabitants of the Anau kurgan in the neolithic age, the quality of tameness is wholly psychological and is therefore not perceptible in an anatomical investigation. The determinable remains of the horse from the kurgan number about 1,250. There are, however, but 120 well-preserved pieces, which repay an exact measure- ment and study. Beginning with the examination of the cranial remains, we find the best among them to be a right upper-jaw with the whole dental row and half of the bone palate. A comparison of the measurements of these pieces with other horse skulls shows a good agreement with a subfossil skull from Western * Die Rassen der Pferdeinder klassischen Staaten nach litterarischen und bildlichen Quellen. Deutsche Landw. Tierzucht, No. 37, pp. 433-437. 386 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Siberia, from the Warwarinskischen ‘‘ Yourts,’”’ on the River Tobol, and with the djiggetai (Equus hemionus). We must, therefore, first settle the question whether these cranial remains really belong to a horse or to a half-ass, like the dyjiggetai, or to a kiang. Investigation into the relation of the teeth of the horse to those of the djiggetai and ass has been carried out most thoroughly by L. Riitimeyer,* R. Owen,T J. C. Forsyth Major,t A. Nehring,§ T. Frank,|| and M. Wilckens.{ In consequence of these studies the following distinctive characteristics between the Western and the Oriental horse groups and the asses are available. ORIENTAL HorsE Group (BROAD-FRONTED HORSES). The premolars of the upper and lower jaws have a larger or equally large transverse diameter of the grinding surface with the longitudinal diameter. The plications of the enamel pattern are here considerably smaller than in western horses and the interior pillars of the anterior island appear rounded. OccIDENTAL HorsE Group (NARROW-FRONTED HorsEs). The premolars are here more drawn out in the length; hence the depth of the grinding surface is greater than in the transverse diameter. The enamel plications of the islands are considerably more folded, and the ant-external horn of the pos- terior island surpasses the post-external horn of the anterior island, projecting further outward, even on the molars, on which in oriental horses they stand almost even. Inthe same way the stronger plication of the enamel margin on the internal lobe causes in the Occidental horse the striking bifurcation of the internal lobule and the stronger development of the spur** in the ant-oblique valley. ASSES AND HALF-ASSES. In the asses and half-asses the longitudinal diameter of the crown is still shorter in comparison with the transverse diameter than in the oriental horse, the enamel plications are less prominent, and the spur is wholly wanting in the ant-oblique valley. Tscherskif{ gives a method by which he says the relationship to one of the groups mentioned can be expressed in figures. This is the determination of the index of projection of the anterior lobule of the interior pillar. If we take the distance from the posterior margin of the crown to the next point of the bottom *Beitraege z. Kenntniss d. fossilen Pferde u. z. vergleich. Odontographie d. Huftiere ueberhaupt. Verh. Nat. Ges. Basel, 1863, p. 538. + Description of the Cavern of Bruniquel. Phil. Trans. 1869, p. 517. t Beitraege z. Gesch. d. Fossilen Pferde. Schweiz. paleont. Gesell., 1877. § Fossile Pferde aus deutsch. Diluvial Ablagerungen. Landw. Jahrbuecher, 1844. ||Beitraege z. Rassenkunde unserer Pferde. Landw. Jahrb., 1875. {| Beitraege z. Kentniss des Pferdegebisses. Nova Acta, Leop.-Carol. deutsch. Akad. d. Naturf., 1888, P2587. **The ‘‘spur” of the German authors is the small enamel fold entering from the ant-oblique valley into the ant-oblique lobe (Owen’s terminology). TT Ob. ctt., p. 320. ORDO PERISSODACTYLA. 387 Dimensions in millimeters of teeth of upper maxilla. ow 5 rh N ae cid Equus caballus. Se : . = oi eal 2 | S186 cio Altai | Western E sf § S ea = , ic) 7 pi —24| +15] after vir i pi Tar- | 98°39] g | g mee trench. sect cet Seay Tscher- | Tscher- | P#7: | & 4 oe tag 3 cary skis, etski. Me Peps ie Premolar 2: | Mength of crown. ............ 38 ok eae eee Mae ‘Gerry 34-5 | 37 36.7 |35-7/36 | 40 RIG GO (CLOW flies cieccis se aise e 26 26 ore 30 eit 21 22.5; 24 (|20.7/25 | 21 Length of interior pillar from ! PROM ESCORDACK enets) oie ses strane 9 6 Bie eae 10.3 AAtse 8.5 8.5 8 7.5| 8 Length, posterior end of crown tonext point of bottom of an- | terior-interior enamel creek..| 19 16 aerate 22 | Sea 18.5 | 20 19 |17.7|20 Same measure to end of ante- rior lobus of interior pillar...| 21 17 Pate 23 Cer 19 21 1927 LOM I2Z0.5 Premolar 3: | Length Of CROW. ceo ocs be bs 32 Boer eras SYlSh ape5 27 29 26.5 |26 |28.5| 27 DRIAL NE STOW So6°5 Hoven VS didn. 28 SS aah Mb Hee Behe a3 e275 2555 27 27-5 |25 '26.5| 23 Length of interior pillar from | PLOMIUOLDACK:. een pee ttels siete 12 onstish Wpbewete Le See 12 a5 1285 10 10.5 |£0.7|11 Length, posterior end of crown tonext point of bottom of an- terior-interior enamel creek..| 19 eel iis. peng 22.8 Reece L7 20 18.7 |18 |19.3 Same measure to end of ante- rior lobus of interior pillar....| 23 Pi Beolit nytt 29 oa ene aM 21/28 ME2Or suet a n22 Premolar 4: WenetiOr crown. s<.0..55 600s s 28 27 27 30 27 26 26003) |\e2 See 2 One 26.7 hae WidtirofsCrOwl. voce patie. ss. 2G) 26 25 33 2 2S eS eeZ Ons ake 7a loz 26 ae Length of interior pillar from pee ee Week OL, ale oe. I4 ji) 103 17 13 Line ey LIeSir2: [12 Length, posterior end of crown tonext point of bottom of an- terior-interior enamel creek..| 18 7; 19 Pig ore 17 TS 5elel 7 D7 SLOMS Same measure to end of ante- ‘ rior lobus of interior pillar...) 23 22 22 2 Bina 21 21 2) 27, 343 Molar 1: AETISTM OL CLOW ce, se oes 25 26 ees 26.3 2287 23 23 220 22e aie Wadth of Crow... 2.25. fal) 23 26 eetats 31 ZnS 23 24. Sls Gaeta 52S Length of interior pillar from AEOL CR COLOACK ites cis) se es << oe 12 13 era 15 1255 13 IO 1635 12 Arb73 Length, posterior end of crown tonext point of bottom of an- terior-interior enamel creek..| 16 PRG taco 20.5 Ee ats iG fils 17 16 16 |17 Same measure to end of ante- rior lobus of interior pillar....} 19 20 cates 2425 Mer ere OMe | wie TOMNTOR SIL ONS Molar 2: Weuctarorcrowl sees a... o. 27 ae 23 28 25 22 SN 2403 P22 ne 1295 52525 Pe maRT OL CTOWN 5. 5. cat cy pae sé 24 25 24 30 ae 245° a 24 24 |25 (24.8 Length of interior pillar from ROORUILO WACK fons ie aie hx be! 12 I4 10 7 14.5 13250 | 10-7 tI oe) Nes Length, posterior end of crown to next point of bottom of an- terior-interior enamel creek..| 16 15 16 2009 Sian 16.5 17 16 Fea) Same measure to end of ante- rior lobus of interior pillar... .| 20 19.5} 19 27 AE 20.5 | 19 IO, aL, pr Molar 3: ene thcot CeOwnl sygeirce secs: « 25 24 sae 32 nee 25 s5 ea 0 25 27.5124 BEM OU CLOW 5 coc ods s,s x 21 20 eras 277(R) Mise eas Mie 21 Sous Length of interior pillar from RLOMELLO: PACK eit an cove sya +i peerene 10 HAE L7ia3 ants 13.8 II 13 T3253 Length, posterior end of crown tonext point of bottom of an- terior-interior enamel creek..| .... 19 mare DAES ae, 17 2208) NaI See i2On 515 Same measure to end of ante- rior lobus of interior pillar....| .... 22 ee 29.5 mae ZONA 2 565220024, S19 388 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. of the ant-oblique valley (Owen) =100, and call this =a, and call the distance from the same point on the crown to the end of the anterior pillar= 6, then a:100=6 ; index. By means of the data furnished by Frank and Wilckens, and this added method of Tscherski, we are able to compile the following illustrative table, based on the dimensions of the teeth of the upper jaw. Table of dental indices (in percentages). Premolar 2. Premolar 3. Premolar 4. Molar 1. Molar 2. Molar 3. | Index Index Index Index Index Index Provenience of |Index of pro-| Index| of pro-| Index) of pro- | Index | of pro-| Index of pro-| Index \of pro- dental series. length jection length) jection |length| jection | length) jection length jection length jection to of an-| to | of an-} to of an-|| to |of an-| to jofan-| to |of an- width.) terior | width.| terior | width.) terior | width:| terior |width.| terior | width. | terior lobe. lobe. lobe. | lobe. lobe. lobe. = = | Anau,Komoroftr’h) 78.8 | 110.5 | 87.5 | 121.0 | 96.5 | 128 104 119 88.8 125 | (63.3 ieee Y= 24 TOCEY Sai 83,6 WetOGn aul ae ep OO a S26 | 100 125 100 127 | 84.0 115 a= 1 5 TCCL salar eee, all eee mee 2-0. 1516 Sesh Ae avae?'| ges alse Vana Ruvenaevdss 7I-32>} 104.5, |TOO 12670 HLIO 122.2] 119 120 107 T2270) Shoe 120 Western Siberia..| 60.9 | 105.8 | 94.4 | 123.5 |100 123 || 100 123 109 125 | 88.8 123 Tarpan, after Tscherski, ae. o- 60.8 | 105.0 | 93.1 | 109. {100 113 104 ven 104 112 | 75.8 114 Equus hemionus...| 57.5 | 107.2 | 96.2 | 103 104.4 | 112 108 119 93 12g \2537.0 120 Equus przewalskit .| 50.2 5 oe | RCo SAS aca oe spies Ui isis aie rn rn Kis Equus onager.....| 69.8 | 100 93 115 96.4 | 115 | 100 119 96 123°], S255 132 It appears from the above table, as regards the length and width relation of the crown of the molars, which according to Frank and Wilckens is so characteris- tic, that in all the compared horse teeth the length exceeds the width in premolars 2 and 3, and molars 3 and 2, and only equals the width in premolar 4 and molar 1, or is shorter. Only on a skull from Yana River in Siberia do I find somewhat shorter teeth, in which the width is therefore relatively greater. This would be an indication that the horse of the Anau kurgan belonged to the Western race. If, now, we compare the relations under Tscherski’s method and their results, we find that the highest value for the projecting of the internal lobule belongs to the Anau horse, the Siberian horse, Equus onager and Equus hemionus. We have here in the especially demonstrative premolar 3=121, premolar 4=128 and 129, molar 1=119 and 125, molar 2=125 and 127. Tscherski* calls the measurements belonging to the Tarpan “high,’”? which in the sense of Frank’s method would assign this animal to the Oriental horse group. He mentions, however, that in an Arabian horse these measurements are 115.7, 116.6, 117.6, 121.2, and 118.3; and in a horse from Dongola 116.6, 122.7, 122.5, 125, and 125, which correspond excellently with those of the Anau horse. T'scherski says further that the maximum figures for this projection are found in the broad- fronted races of fossils and recent Siberian horses. In the ass T’scherski fourd strong variations, and in the half-ass a series of proportions which also were similar * Op: "cit. p. 307. ORDO PERISSODACTYLA. 389 to that of the’ Anau horse; thus premolar 3 =103.2, 116.6; premolar 4=112.5, 121.7; molar 1=144.7, 120.4; molar 2=117.3, 129.6. However, if we now consider the teeth in connection with the appearance of their form, as well as that of their patterns and enamel plications, we shall recognize at once the extremely slight plication of the enamel layers. The inter- nal lobule of the anterior island is further distinctly drawn apart into two horns and flattened. The “spur’’ (of the German authors) occurs only on the premolars. This eliminates the possibility that we have here an animal belonging to the group of asses or half-asses. This is further evidenced by the lateral expansion of premolar 2, which in the Anau specimen is 78.8 and 83.9, in the Siberian horse 71.2,and but 60.9 in the asses and half-asses; especially in the djiggetai, Equus hemionus, which otherwise, on account of the size of the extremity bones, would here come very much into question, where it sinks to 50 to 57 per cent. This, according to Tscherski, is a very typical occurrence. In general the premolars of the half-asses show a narrow isthmus and a strongly widened capitulum, which does not happen at all in the Anau horse. We can, therefore, assume with certainty that we have here, not a half-ass or an ass, but a genuine horse. It remains now only for us to determine to what variety this horse belongs. The shape and conformation of the enameled crown, as well as the projection of the lobes, indicate a horse belonging to the Oriental group, but the form of the anterior island and the relations of length and breadth of the teeth point to the Occidental races. We have no frontal bone pieces, which would easily enlighten us on this point, and must, therefore, seek some other method. If we compare fig. 4 on Tscherski’s plate 11 with the molar series of Anau (plate 77, fig. 1, and text-fig. 491, a ande, p. 413), we shall see that beyond question a similar form of teeth is represented in both. The dental system of the fossil as well as of the known recent Siberian horses is peculiar in that, notwithstanding the characteristics of the skull, which place it in the group of the Oriental (medium- browed) horses, it shows a character which is peculiar to the heavy, narrow-browed West-European races, and in addition reaches the highest degree of the develop- ment of this type, especially as regards the considerable anterior projection of the internal lobule in the teeth of the upper jaw. If we should judge only by the form and the before-mentioned dimensional proportions of the teeth our Anau horse would belong to this Siberian group. THE LOWER JAw. The remains of lower jaws which we have from the kurgan, although four in number, consist of very fragmentary pieces, in only one of which are the branches sufficiently preserved to permit a determination of the length of the diastemas. From the dimensions, however, we can easily recognize that only a horse of medium or small size is indicated. What has been said with regard to the peculiarities of the teeth of the upper jaw applies fully to those of the lower jaw. 390 ANIMAL REMAINS FROM THE EXCAVATIONS AT ANAU. Table of dimensions (in millimeters). Subfossil |Subfossil East La Téne} West |a.7 Prze- Tur- North | ‘horse, | Siberian Siberian | walski | kestan ; pea: from | Tartaria Hi horse, | horse, Upper jaw. nau, | Hosto- | horse, Paige after } after +12 f after _|ar rv mitz, after i epiears Tscher- | T'scher feet. Mus. | Tscher- ki ski. ski. Teplitz.| ski. ce Width of face at orbita maxillaris. 168 162 169 179 193.5 182 Width of face at diastema _ be- tween premolar 4and molarir... 118 5 4 TDS 9) 22400 gees Width of face be- | fore premolar 2. 58 64 59 69 | 68 Width of face be- | tween molars I / ands ewe 119 | 108 118 r20° | Length of dental | series of upper jaw from premo- lar 2 to molar 3. 174 169 164 T7255 aes 167 | 159 183 165 165.5 = SS = — = = | Equus / lus. fo Equus caballus hemné-w lt Edie nae Lower jaw and dental series. — | ee sate onager Ljachow.| Ireland. Teree | kiang Width of lower jaw between alveoli | of premolariottec 14, eee ae 36 Pee 38.3 38 4I 44 40.5 41 Length of diastemas2... aoe | go (?) 97 97 80 86 87 7308 Smallest part before premolar 2....... ar | hens 47 39.5 39 44 37 44 = (leapt, 225. yee ee 2 Ls Ae tae a3 30 30-5) 315 30 23 Ereta lat Awidth,< 4, sck sy eaeurne 15 sae 15.5 16 15 | 16.5 £6 13 Slength 2s 397 eae 27 | 26 30 28 27.3 | 28 28 27.5 Premolar 3 ( Width: wi sooo hare eB co.5 20 18 | 16 T7573) ete 18 16.5 Nlength 2.0 cae anes Sivek z Phe El eg) 26 he Bay 26.5 26 Premolac 4.7 wid ieee a aN aah eR Rela Se 18 | 317-5 | 18 17 (lengthier cere e 23 ed 2 HARES hee 24 25 Mola width: oa eee 16> Wty) ERB. See 16 15 3 , slength. cy wane saat oh 24 2738 2: 2555 sume 25 26.5 Molar 2 (widtharn cece eet eee 14 17 15 16 16 15 15 SCAPULA. The one almost perfectly preserved specimen of the scapula belongs also, as is shown by the comparative dimensions, to a horse of medium size. The slen- derness of the bones, a characteristic especially typical of the Anau horse, is here very perceptible. No further peculiarities are noticeable in the structure of the bones. HUMERUS. None of the bones permit a full measurement of the length of the humerus. Generally speaking, they were only distal pieces, the measurements of seven of which we have given in the table on p. 392. ‘The few measurements of width here given approach nearest to those of Equus przewalskii and Equus hemionus. ORDO PERISSODACTYLA. 391 RADIUS. This bone is represented among the Anau finds in a somewhat better state of preservation than were those we have considered. ‘The relatively slight lateral expansion of this shows already that the Anau horse belongs to the thin-footed type, as will appear clearly from the consideration of the well-preserved meta- carpalia and metatarsalia. METACARPUS TERTIUS. One of the four measurable metacarpal remains from Anau is very well pre- served; and from this it is recognizable that the Anau horse belongs to the narrow- footed horses, since it possesses long and slender metacarpi. The index of width of the Anau animal gives the following results in comparison with other fossil and subfossil as well as recent horses. Hor the Onset al (EGWUs REMIONUS) acc ie goss Kole nev seve ss 11.8 MOCLHNE ONAGCT A GWUS ONASET) 6 eo ow ccs ed div wee oS nee eulees os 12.0 ROG AD IR OU AOI) Fe nate d oieye'l on. ds % vince vio oe Anema eee eRe) RO tates AMAaly NOESE see ste sestegee thous b A f a 080800, Cae oe f ° ©,°% oo 7? 0% 20 °° COMO Ge ° 0.9 50.0) ae COC a ROR > AORN : CE eo it J oer ° . Cae [cists Se et ee wa see . 2 wheat 9° Peees CE CRyCl Sey gs kee Y Seas. So CID) OS SECS ets omer eer el nee SVC NOT CO ss Duigiic f . ee . . Lary Se eng echin eteT a & ane ota syere ee oe ee ig 9 ley) Sheree eed Olam Die ec eaeyhe rte Bieni@ei Se e “fone we eu eg hO) 6, Ye Oe Sg So *. A re Staleicg: ie) on Pe eee . Oe at Ene) we \e, ° « l, Lae Yay see . . e erates = Se ES) eg vor) palustris Rutimeyer (Turbary Sheep). nd Present Distribution of Ovis aries Map Showing Prehistoric a CHAPTER XIX. THE HORSE OF ANAU IN ITS RELATION TO HISTORY AND THE RACES OF DOMESTIC HORSES. (Plates 87-91.) Having, in chapter xvur, endeavored to prove that the equid of the North Kurgan of Anau is a horse and not an ass, and to show its relation to other Asiatic Equide, and further to picture the changes it underwent during the life of the civilizations of that kurgan, I shall now consider the relation in which the Anau horse (Equus caballus pumpellii mihi) stands to the subfossil horse and to some historical domestic breeds, as well as to the Equus przewalskii Polyakoff. As is commonly known, the domestic horses are generally classed in two groups: the Oriental and the Occidental. Frank* calls the first of these groups also Equus parvus, and the second Equus robustus, and discusses at some length the points of difference between the two types. In this connection I will state briefly some points which have not been sufficiently touched upon in the previous chapter. In the Oriental horse, especially in the Arabian, the brain-skull is, relatively, very strongly developed; the face less so. These horses are called broad-headed, because the width of the forehead is large in comparison with the length of the skull. We have already spoken of the teeth; regarding these we may here refer especially to the anterior and posterior crescentic islands (Owen’s terminology), in which the enamel-margin is not so wavy; the internal lobe is placed just in the middle of the grinding surface, and its division into two is not very clearly marked. The hollow bones are remarkable for their graceful shape and solid, hard texture; the metacarpal bones are relatively narrow. In many points, therefore, this group of horses resembles the ass. On the other hand, these same points distinguish the Occidental from the Oriental horse. In the Occidental horse the facial part of the skull predominates at the expense of the brain-skull, the skull appears long and narrow, the forehead is narrower, the rim of the eye-socket is but slightly prominent. The enamel- margin of the crescentic islands is very wavy, and the internal lobule is divided into two very distinct horns and flattened. The bones of the extremities are heavy and massive, while their texture is less dense and hard than in the Oriental horse. ‘The tarsal bones are generally broader than in the latter group. Again, Sanson,t applying Broca’s anthropological method, has proposed another classi- fication of horses based on dolichocephaly and brachycephaly, dividing them into four dolichocephalic and four brachycephalic races; but it has not been possible to maintain this division in practice, as it is too schematic. We can not consider here other attempts at classification. * Frank, Ein Beitrag z. Rassenkunde unserer Pferde. (Vortrag) Landw. Jahrbticher, Iv, 1875, pp. 33-52. A. Sanson, Nouvelle détermination des espéces chevalines du genre Equus. Comptes Rendus de l’Académie d. Sciences, t. LXIxX, pp. 1204-1207. 401 402 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. What concerns us before all else is the question: In what relation does the horse of Anau stand to the domestic horses of to-day, and espectally to their direct ancestors, the subfosstl horses? After the foregoing special investigation and the subsequent general comparisons, I have no hesitation in asserting that we must see in the horse of Anau the first representative of the Orvental race of horses. FOSSIL AND SUBFOSSIL HORSES AND THEIR CHARACTERISTICS. * I will not attempt here to trace the connection of the Equide of the later and middle Tertiary period, although since the wonderful results of Henry F. Osborn (in the American Tertiary) this would be a pleasant and profitable subject; nor can I here institute a comparison with the other diluvial Equide—the Equus stenoni Cocchi, Equus quaggoides F. Major, Equus speleus Owen, Equus mauri- tanicus Pomel, ete. In this place I will attempt only to establish the relationship of the horse of Anau, especially to the remains of those horses which, with more or less right, have been regarded as the ancestors of our domestic horse, and whose direct conversion to the domestic state has been assumed to be certain. We will here notice the principal types in question. THE HORSE OF THE QUATERNARY PERIOD OF EUROPE. During the glacial period the forested area became greatly restricted, grasses, herbaceous plants, shrubs, succeeded trees as the predominant vegetation. These steppe-like regions became the home of numerous herds of wild horses, together with scirtetes, jerboas (alactaga), spermophili, bobacs, lagomys, arvicole, and other characteristic inhabitants of the loess steppes of to-day beyond the Volga.t The horse that then lived in the northeastern part of Central Europe was, as shown by the remains found at Remagen, Westeregeln, and in other places, a medium-sized, stocky animal with thick bones and a large head. It may, there- fore, from this bodily shape be taken to have ranked very near the present Equus przewalskit, which according to Grum-Grshimailo has a withers-height of 153 centi- meters.{ In the southern part of this region there seems to have lived besides this horse, either at the same time or somewhat later, a smaller form of the same variety. ‘This is not surprising when we consider that Matschie recognizes a larger and smaller variety of the Przewalski horse. By this I do not mean the form of Equus caballus fossilis, which Woldrich has called “minor’’ and which Nehring, objecting, held should rather be called ‘‘ major,’’ since the horse of Nussdorf, with its 555 mm. basal length of skull, must be counted among the largest horses. No; it was the horse of Solutré, Cindré, and other points in France, that represents the small, broad-boned European wild horse. We do not know whether the paleolithic horse of Solutré was domesticated, as Toussaint asserts. Nehring also assumes that the horse was already domesticated in the glacial period, and I, *Dr. Duerst designates as ‘‘fossil” all occurrences of paleolithic age or earlier, and as ‘‘subfossil’’ all of later age.—R. P. tA. Nehring Fossile Pferde aus deutschen Diluvialablagerungen, etc. Landw. Jahrbiicher, x11, 1884, p. 148. {Compare p. 426. FOSSIL AND SUBFOSSIL HORSES AND THEIR CHARACTERISTICS. 403 too, influenced by many reasons, have expressed my belief in the probability of a very early domestication of the horse, probably in the paleolithic age of Europe. But this has here no bearing. ‘The essential fact is, and remains, that the horse of Solutré—which may have been the same that was hunted and pictured by the cave-dwellers of Dordogne—was smaller than the steppe horse of Central Europe, and had a withers-height of about 125 cm. if one may judge from the skeleton of a Solutré horse preserved in Lyons. ‘To what is the smallness of this horse due? In giving my conclusions I shall attempt an explanation; here I will remark only that the climatic and physiographic conditions under which the Solutré horse of the paleolithic age lived were probably essentially different from those surrounding the steppe horses of the northern lowlands and coastal lands of Europe. THE HORSE OF PREHISTORIC (LATER QUATERNARY) TIMES. (a) The Neolithic Age.—Remains of horses of the neolithic age are rare; never- theless, finds from Wohontsch on the Biela, Leitmeritz, Fouvent, and Louverné are evidence that at least in Bohemia and Gaul the horse had not disappeared in neolithic time. More complete remains of the horse than merely a few bones of the extremities seem to have been found at Schussenried, of which Fraas* has unfortunately given a very imperfect account, and it was not possible to determine its geologic age. Here belong, however, still other finds: Boucher de Perthes found in 1833 at the bottom of a turbary in the Department of the Somme in France, 5 to 6 meters below the water-level, two skulls of horses associated with late neolithic pottery and with flint implements. These were deposited in the Museum in Paris; and one of the skulls was later examined by Sanson himself and determined as “ Asinus africanus or the African ass.” Sanson remarks concerning it on page 133, t. mI, of his ‘‘Zootechnie’’: ‘‘ En le donnant comme étant celui d’un cheval, Boucher de Perthes s’était done trompé, erreur bien excusable d’ailleurs, de la part d’un trés-habile archéologue tout a fait étranger 4 l’anatomie zoologique. Ce qu’importe, c’est que la présence de ce crane dans le nord des Gaules, a 1’époque de la pierre polie, atteste que sa race y avait été amenée dés lors par des migrations de population humaine.”’ Sanson considered this an isolated case, but a no less eminent authority than Ludwig Riitimeyer described an equid skull from a pile-dwelling at Auvernier on Lake Bienne, which he ascribed to an ass.f Since, through the kindness of Professor E. Ray Lancaster and Oldfield Thomas, I was able to compare this African ass, so early an inhabitant of Europe, with its contemporary from the ruins of Abadieh near Kenia in Egypt, dating according to Professor Flinders Petrie from the IV dynasty, I came to doubt the correctness of the determination of the two authorities—Rtitimeyer and Sanson. Riitimeyer (p. 53), as well as Sanson, was led to its determination as Equus asinus, or the half-ass, on account of the small absolute size of the skull. But the greater extent of the diastema or toothless ridge of the jaws than is usual *O. Fraas, Beitrage zur Kulturgeschichte, etc. Arch. f. Anthrop., v, 1872. }Riitimeyer. Schadel von Esel u. von Rind aus den Pfahlbauten von Auvernier u. Sutz Pfahlbauten. vi. Bericht. Mitt. d. antiq. Gesellsch. Ziirich, Bd. xrx, pp. 50-56, 1876. 404 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. among asses distinguishes this animal from those. The transverse diameter of the skull and the form of the nasal bones conditioned thereby are like those in the horse, as is also the eye-socket. The perfectly preserved teeth show that the length of the upper jaw is 34 per cent of the length of the skull. Riitimeyer finds for the horse, elsewhere, 32 to 35.6; for the ass 35 to 38.5 per cent. Thus it should be a horse. Also, the relation of the premolar row of the lower jaw to the dental row, which in the horse is 51 to 53 per cent and 49 in the ass, is 52 per cent in the skull from the lake-dwelling, thus again as in the horse. Only the occiput, says Riitimeyer, looks like that of an ass. And he closes his obser- vations: “‘ Notwithstanding all the uncertainties which seem to attach to these measurements, not only on the teeth but on the skull as well, certainly derived from nature, there remains in my mind no doubt that the skull from the lake belonged to an ass.”’ The kindness of Doctor Lehmann, Director of the Swiss Landesmuseum in Zurich, enabled me to make a direct comparison of the skull from Auvernier with the mummified skull from Abadieh and with the skulls from the Somme which I studied in the Museum of Natural History in Paris. During this investigation there arose again the question which I had asked inyself before, during the study of the craniology of the ruminants: What are the really decisive criteria of species, and what the incidental characteristics brought into existence by causes acting during individual life? At last I came to the reali- zation that a conclusive method of discrimination did not exist; that all those in use might be said to be wholly empirical, in part, indeed, dependent on the personal perception and feeling of the individual student, and therefore not scien- tifically established. Nor have I succeeded—through lack of material, fresh heads and numerous skulls of asses—in adding much that is new; but I believe that I have thrown some light upon the causality of some of these relations, and have tried to incite to a more scientific treatment of the question. CRANIOLOGICAL DIFFERENCES BETWEEN THE ASS AND THE HORSE. The older authors, as De Blainville in his Ostéographie, and Cuvier, do not supply what is really needful for the comparison in question. [L. Rutitimeyer has opened the way here, too, as in many other branches of paleontology. His ‘“Beitrage zur Kenntniss der fossilen Pferde,”’ etc.,* was the first work worthy of note on the fossil remains of the genus Equus, but he did not treat of the differ- ences between the horse and ass till in his second treatise ‘‘The Horses of the Quarternary epoch’’t in the same year as the studies of Frank mentioned above. In the meantime this relation had been discussed by R. Owen in his “ Description of the Cavern of Bruniquel and its Contents,’’{ in which are beautiful plates representing the teeth characteristics of the horse and ass. *Riitimeyer, Beitrage zur Kenntniss der fossilen Pferde u. zur vergleichenden Odontographie der Huftiere ueberhaupt. Verh. Naturf. Gesellsch., Basel, Bd. 111, 4, 1863. {Riitimeyer, Weitere Beitrage z. Beurtheilung d. Pferde d. Quaternar Epoche. Abhandl. d. Schweiz. paleontol. Gesellsch., 1, 1875. t Owen, Philosoph. Transactions, vol. 159, 1869, pp. 517-557. CRANIOLOGICAL DIFFERENCES BETWEEN THE ASS AND THE HORSE. 405 Later J. C. Forsyth Major,* Mme. Pavlow,t Nehring,t Kowalewski,§ and Woldrich|| touched more or less at length upon the relation of the skulls of the ass and horse. Still later some of the French investigators discussed these very different and partly contradictory criteria, thus Duges,{{ Monfalet,** and before all X. Lesbre, ff the veterinary anatomist of Lyons. I will here state briefly those characteristics on which these authors lay the most stress: According to Owen, in the upper series of grinders the degree of oblique attri- tion of premolar 2 causes its working surface to appear more produced and acute anteriorly than in less worn and more evenly worn specimens. Besides the general inferiority of size of teeth, molar 3 is relatively less than in Equus caballus and is not bilobed behind: the outer channels are more evenly curved or concave; and as the same character prevails in the inner enamel-wall of the lobes these are more regularly crescentic in shape. The longitudinal ridge is relatively narrower. A slight excess of fore-and-aft over transverse diameter of grinding surface is recognizable in the ass—such excess not being seen in the permanent grinders, premolar 3 to molar 2, of the horse. Rtitimeyer regards as a constant characteristic for the teeth of the ass, as compared with the horse, the relatively slight length of the foremost as well as of the hindmost molar in both the upper and lower jaws. Also he considers the premolars and molars in the ass to be shorter than in the horse; the foremost premolar tooth is strikingly short. Riitimeyer declares irrelevant the circum- stance mentioned by Owen that molar 3 superior is less bilobed in its posterior circumference than in the horse. He ascribes to the ass, at least in the teeth of the upper jaw, more oblique enamel plications than occur in the horse, but he remarks that one can not disregard the fact that all these characteristics, in so far as they concern construction of the teeth, recur in very old horse teeth; there- fore, in the earlier stages of abrasion, the teeth of the ass show the characteristics which correspond to the deeper parts of the tooth lying nearer to the root. In addition to this is the relatively small extent of the toothless part between pre- molars and canines, as well as the slight width of the incisor crown. Thus, in the ass the whole construction is more compact and crowded. * J]. C. Forsyth Major, Beitrage z. Geschichte d. fossilen Pferdes, insbesondere Italiens. Abhandl. Schweiz. paleontol. Gesellsch., vr and vu, 1880. + Marie Pavlow, Etude sur I’histoire paléontologique des Qngulés. Bull. Soc. Imp. d. Naturalistes, Moscow, 1889. t Alfred Nehring, Fossile Pferde aus deutschen Diluvialablagerungen, Landw. Jahrb. 1884, Bd. 13, Ppp. 149 et seq. § Waldemar Kowalewski, Monographie der Gattung Anthracotherium Cuv. u. Versuch einer nattir- lichen Klassification d. fossilen Huftiere. (Paleontographica, N. F.u, 3, xxi.) Sur lAnchiterium Aur- elianense Cuy. et sur l’histoire paléontologique des Chevaux. Mém. de I’ Académie Imp. d. Sciences, St.- rar 8 Vile SDeLic ute Da 59 ko 73. || J. N. Woldrich, Beitrage z. Fauna der Breccien und anderer diluv. Gebilde Oesterreichs. Jahrb. k. k. Geol. Reichsanstalt, Bd. xxx, Heft 4. Wien, 1882. q A. Dugés, Paralelo de los craneos de caballo i de asino. Guanojuato, 1898. Actes Soc. Scient. d. Chilis tavillppay7. 7 «TD, Monfalet, tee complémentaire sur la communication de M. Dugés. Jbidem, pp. 79, 80. Ths. Lesbre; Observations sur la machoire et les dents des Solipédes. Bull. Soc. d’ Anthropologie de Lyon, t. XI, 1892, pp. 49 e¢ seq. 400 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. Frank states chiefly that the plication of the enamel-margin in the upper molars of the ass is less complex than in the Oriental horse, and that the so-called spur is here wholly wanting. He considers the best characteristic to be the fact that the distance from the anterior margin of the foramen occipitale to the median point of the vomer incision is shorter in the ass than the distance from the same point on the vomer to the end of the palatine suture. In the horse this dimension is much greater. Nehring also agrees with Frank as to the great value of this characteristic. Dugeés calls attention to the greater convexity of the forehead of the ass. According to him the face is shorter in comparison with the horse, and the orbits triangular. A perpendicular to the plane of support falls far behind the condyli of the occiput. The free part of the nasalia reaches to the posterior edge of the corpus maxillare and is therefore very long in the ass; then the spur is wanting on the teeth. In the horse, on the other hand, according to him, the forehead is flat, the orbits round and the occipital line touches the condyli. The free part of the nasalia does not reach to the middle of the incisive edge. The “‘spur’’ is characteristic of the teeth. Monfalet gives nothing on the characteristics of the face, but only on those of the brain-skull. The most accurate work, especially as regards dentition, is that of X. Lesbre. He is the first to distinguish between the teeth of the adult animal and of the young. He finds (p. 60) that in the horse, at the age of ten months, the ‘“‘spur,’’ which he calls “pli cabalin,”’ is already clearly developed. In the young ass this is always wanting. Lesbre confirms a shortening of the teeth of the adult ass through a kind of atrophy of the posterior pillar, but he considers the disposition of the enamel plications moreimportant. (1) In the upper molars the internal lobule is less devel- oped in the ass than in the horse, all proportions remaining the same; it is, in the first place, shorter towards the rear, so that its base stands median or almost median, and not, as in the horse, on the forward part of the tooth. In the first molar the plication is round in both animals, only in the ass it is less obliquely inclined toward the back than in the horse. (2) The exterior sides of the tooth seen from the grinding surface are narrow and simple in the ass, broad and com- pressed in the middle in the horse, especially the premolars. (3) The ‘‘spur”’ or pli cabalin is wanting in the ass at all ages, or is very inconspicuous, while in the horse it is often double, and disappears only at an advanced age, and earlier in the molars than in the premolars. (4) The crescentic islands are simpler in the ass than in the horse, less plicated and complex, but they often vary. (5) The outer channels are not so deep in the ass as in the horse. On the molars of the lower jaw Lesbre finds that the 8 formed by plications 1 and 2, has both of its loops generally rownd in the ass, and usually somewhat flattened and angular in the horse. In the ass both the loops of the 8 are equally large; in the horse the for- ward one is longer than the posterior, and they are separated by a sharp angle instead of by a curve. An exception is the first premolar, in which the posterior CRANIOLOGICAL DIFFERENCES BETWEEN THE ASS AND THE HORSE. 407 loop is often greater than the forward one. The external channel is much less deep in the ass than in the horse. To discriminate among whole skulls, Lesbre finds by comparison, empirically, the method of drawing a connecting line between the tuber molaris and the fossa articularis ossis zygomaticr; if the prolongation of this line strikes the crista occipi- talis, or anterior to this cuts the facial plane, the animal is an ass; if this line passes posterior to the crista occipitalis it is a horse. Tscherski* places much dependence on the construction of the nasal bones and on the form of the suture between the lacrimal and the nasal bone. ‘This, he says, is in the ass always curved outward in bow shape, but is straight in the horse. I find that this relation is very dependent on the breadth of the forehead. The foramina infra-orbitalia stand nearer (to the naso-maxillar suture) in the ass than in the horse. He says, further, that in the horse the processus zygomatici ossis frontalis is three-cornered, but in the ass oval and compressed. Salenskif asserts that Tscherski’s criteria are valuable and to the point, especially the form of the suture between the lacrimal and the nasal bone, but that a difference in the cross-section of the processus zygomatici ossis frontalis can not be verified in the ass and horse. On the other hand, Salenski proposes another method, which he considers extremely constant: the lower edge of the lower jaw is in the ass curved and furnished with protuberances, while in the horse it is smooth and straight. Unfortunately I can not agree with Salenski, as I consider this criterion of the lower jaw very variable, eminently dependent on the age of the individual; one has only to compare the lower jaw-bone of the horse of Auvernier with the one of exactly similar shape in the ass from Abadieh (plate 87, figs. 1 and 2). So far the authors. It is now time to investigate the conditioning factors that influence the shaping of the skull of the ass and of the horse. Starting with the law, which I have proved elsewhere,{ that the skull is a flexible product of the influence of skin and muscle, the longer ear and its generally different position seems to be decisive for the head of the ass. The greater burden imposed by the external ear must have had as a consequence a change in the musculature of the ear, which in turn, as is shown by Anthony$§ and Toldt| exerts its influence on the bones. Indeed the whole ear-musculature is different in the ass, as I was able to observe in agreement with Kirsten,{[ on the only fresh head that I could obtain. * J. D. Tscherski, Wissensch. Resultate d. Neusiberischen Expedition. Mém. Acad. Imp. St. Pétersbourg, t. XL, 1893. : +W. Salenski, Equus przewalski Polyakof. Wissensch. Resultate d. von Przewalski nach Central Asien unternommenen Reisen. Zoolog. Teil, Bd. 1, Abt. 2, Lief 1, pp. 27-47, St. Petersburg, 1902. tExperimentelle Studien ueber die Morphogenie des Schaedels der Cavicornia, Vierteljahrschr. Naturf. Gesellsch. Zitirich, 1903, pp. 360-374. §R. Anthony, Etudes experimentales sur la morphogénie des os. Modification craniennes consecu- tives a l’oblation d’un crotaphyte chez le chien. Journal de Physiologie et de Pathologie Gén., No. 2, Mars, 1903. 4 ed ol he || R. Toldt, Asymetrische Ausbildung der Schlaefenmuskeln bei einem Fuchs infolge einseitiger Kau- tatigkeit. Zoolog. Anzeiger, 1905, Bd. xxix, No. 6. : ; {F. Kirsten, Untersuchung ueber die Ohrmuskulatur verschiedener Saugetiere. Inaug. Diss., Bonn, 1902. 408 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. Not to go into anatomical details, this difference is shown, before all, in the significant enlargement and bipartition of the musculus retrahens aurts brevis, which Kirsten aptly calls musculus abductor auris brevis. Strikingly enlarged, too, in the ass are the musculus abductor (retrahens) auris longus; musculus adductor (scutulo auricularis) auris inferior; musculi levatores auris longi, medius, et brevis, and others. In consequence of the insertion, direct or indirect, of these muscles on the occipital crista and the linea nuchalis of the occiput, it is clear that in the ass a stronger pull is exerted on the occipital surface and the lever-arm which is formed by the crista. Through this action there must necessarily follow an inequality in the back part of the brain-skull. Variations may of course be caused by the size of the ears, as also by the manner of carrying them, as the flap-ears of the Sudanese domestic ass, in contrast with the Asiatic. In my opinion, therefore, the most important characteristic in the bony head for determining the specific difference between the ass and horse is the position of the occiput. This can be recognized by three different methods which are, however, of unequal uniformity and exactness in practice. (1) Lesbre’s method of the cheek-crista line, for the clearer expression of which I would make a suggestion later. (2) The Frank-Nehring method—distance from foramen occip- itale to vomer palatinum. (3) The inclination of the occipital plane mentioned by Duges, for the determination of which I would also propose a new method. Besides this there are other, mostly less regularly marked, features which can be traced to the action of the ear-musculature; the lateral edges of the occipi- talia lateralia are drawn more strongly towards the meatus auditoritus externus; the funnel of the meatus auditorius externus is more erect, etc. But, as was said before, there is here much room for the play of individual variation. It may occur in practice that one has only a skull of an equid without cheek- bones or maxtlle; in this case, to determine the inclination of the occiput, use a tangent, B, on the faczes-surface of the frontales and nasales. If, on the other hand, the maxillary is present, draw first Lesbre’s line; on this (A) or on the facial tan- gent (B) erect a perpendicular which touches the highest or most aboral point of the crista occipitalis, and draw a tangent from this same point, first, on the upper edge of the foramen magnum, and second, on the most aboral point of the condyli; we find, according to my measurements of these angles to date, on thirty- two skulls: re Aa. Bd; Bi2t HLOTSeS Scan eeet ee 14-25° 10-24° 15-30° 10-25° ASSES Sasa nit cetera aevere 35-50° 30-35° 35-50> 30-45° As is clear in this table, as great as is the difference between the ass and horse, the difference between the two relations A and B is small, and we can therefore use either the line A or B, according to the state of preservation of the material. I call this angle ear-load index (Ohrbelastungsindex) to indicate its dependence on the ear-musculature. In order to express the Lesbre line numerically in an index I have applied the following method: draw the Lesbre line and then a connecting line between the crista occipitalis and the intersection of the Lesbre line with the articular CRANIOLOGICAL DIFFERENCES BETWEEN THE ASS AND THE HORSE. 409 jossa of the lower jaw. The angle formed by the two lines is the positive or negative parvetal-crest-curvature index (Scheitelkruemmungsindex); the angle being positive (+) when the crista lies below the Lesbre line, and negative (— ) when the crista lies above the Lesbre line. According to this, horses should almost always have a negative, and asses a positive index. Since the ‘‘Lesbre line’’ leaves at times something to be wished for, I have chosen a second-control index which expresses the size of the acute angle formed between the prolongation of the facial tangent and that of a tangent on the brain part of the frontal and of the parietalia. In the ass this angle is about 40°, and in the horse 20° to 30°. As regards the bones of the facies, it is to be noted that the observations of the length of the free part of the nasal bones have absolutely no value, since this is wholly individual, or possibly also subject to racial variations. Also I do not find the shape of the sutures of the nasalia with the frontal and the lacrimal to be always characteristic; if they are useful, as stated by Salenski and Tscherski, they are too dependent on the width of the frontal to be decisive. Not more so is the shape of the lacrimal on which Riitimeyer seemed to place great reliance; nor can the triangular form of the orbital be used as a characteristic of species. Lastly, as regards the teeth characteristics, the occurrence of the ‘‘spur’’ must be used, notwithstanding Lesbre’s assurances, with some caution; I agree with Rttimeyer, Tscherski, and Owen, when IJ assign a higher value to the position of the internal lobule and its form, as also to the size of premolar 2 and molar 3. In this connection I have made some very instructive experiments by grinding on teeth of asses and horses, and have thereby arrived at the conviction that, as regards the ‘“‘spur,”’ by grinding to a sufficient depth, one may make out of every horse tooth an ass tooth, and sometimes from an ass tooth a horse tooth, without taking into account that in the first teeth of the ass ever published (Owen, plate Lvul, fig. 1) the spur shows on all the teeth and even double on one. Such cases are, however, extremely rare, and the ‘‘spur’’ is nevertheless to be regarded as a useful characteristic. After this review of the criteria in question we can pass now to a comparison of the ancient Egyptian mummified ass with the horse of Auvernier. The brain- skull of the ass of Abadieh shows the following relations: the ear-load index (Ohrbelastungsindex) is 42° with the foramen tangent, 35° with the condylus tangent; the same index on the skull of the Auvernier horse is 28° with the foramen tangent (B1); the condylus tangent (B2) is not measurable, as the condyli are broken off. The parietal curvature index (Scheitelkruemmungsindex) is + 22° on the ass of Abadieh, and —7° on the horse of Auvernier. Riitimeyer remarked that there was something asslike in the shortness of head and width of forehead of the equid of Auvernier. Teeth.—I have already discussed the teeth characteristics as described by Riitimeyer, and have repeatedly drawn the conclusion that the equid of Auvernier is a horse. The teeth are indeed very short and close-set. But we see clearly the horselike shape of the internal lobule which is very widely drawn out in two horns, while in the ass of Abadieh it is round and placed median. Then, too, 41O THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. in the ass the “spur’’ is wanting on all the teeth, while the equid of Auvernier has it on all the premolars and on the third molar. It is noteworthy, though without value in distinguishing between species, that the ass of Abadieh has the first premolar, which is not frequent at this age. I find no differences in the inci- sors, which are fully present in both skulls. The incisive part of the Auvernier equid is, with a width of 62 mm., somewhat wider than that of the ass, which measures only 52 mm. ‘The equid of Auvernier had during life a broader muzzle than the ass, again recalling the horse, which, especially in the diluvial horse, had a very broad muzzle, broader than is shown in recent horses (83 to 88 mm., Nehring, op. cit., p. 90). The equid of Auvernier is thus shown to differ in its principal points from the ass of Abadieh and must, therefore, be regarded as a horse and no longer as an ass. We may see confirmatory evidence that it is a horse that was used by the people of Auvernier, in the size of a bronze bit from the pile-dwelling of Mohringen on the same Lake of Bienne; this measures 9 cm., while the maxilla where the snaffle lies measures 5.1 cm. Moreover, Marek* has published a series of skulls from stations of the same age, as well as later (among them a plaster cast of the Auvernier horse), which are classed by him as Helveto-Gallic horses. If it has been shown that Rtitimeyer erred in his determination of the Auvernier skull, it is now easy to prove the same in the case of Sanson. ‘Two skulls in the Galerie de Paléontologie of the Museum in Paris belong, in dimensions and in form, to the same variety of horse as that of Auvernier, and their special charac- teristics will now be considered. We can take up the thread of our argument at the point where we left it (p. 403) in order to identify these two skulls, which, if they had been African asses, would have been of fundamental importance as to the distribution of domes- tic animals in prehistoric times. (b) The copper time of the bronze age brought to Europe the small, slender- limbed horses whose domesticated condition is for the first time certain; and which is proved also to have existed in: (c) The La Tene Period (iron) and which have been already treated by Studer,t Marek,{ Kraemer,§ and others. (d) The Hallstatt Period seems, probably on account of the heavier armor of the cavalry of that time, to bring us already the heavy type of horse that we find in Roman times at Vindonissa; at least the skull from Schuettarschen, which will soon be treated of, differs in several respects from those of the bronze and La Tene periods. Unfortunately we have from Schuettarschen no bones of the extremities to aid in more exact discrimination. (e) The Roman Times bring us better, more abundant material, as we shall see in the finds from Vindonissa. *T. Marek, Das Helvetisch-Gallische Pferd, ete. Abhandl. Schweiz. paleont. Gesellsch., Ziirich, 1898. 7 Th. Studer, Die Tierwelt in den Pfahlbauten des Bielersees. Mith. Berner naturf. Gesellsch., 1883. tJ. Marek, Das Helvetisch-Gallische Pferd. Abhandl. d. Schweiz. pal. Gesellsch., xxv, 1898. § Kraemer, Haustierfunde von Vindonissa. Revue Suisse de Zoologie, t. 7, 1899, pp. 143 et seq. PLATE 87: erie ol Fic. 1. Norma lateralis of skull (restored) of the horse (Equus caballus pum pellit) from pile-dwellings at Auvernier (Switzerland). Museum, Zirich. 2. Norma lateralis of skull of the ass of Abadieh, IV dynasty, Egypt; coll. by Flinders Petrie; British Museum, Natural History. Swiss National PLATE 88. 3 Fic. 1, Norma verticalis of skull of a greathorse (Eq. caballus Fic. 3. Norma verticalis of skull of an old Eg. caballus pumpellir robustus Nehring) of Hallstadt time. Schiittar- from the turbary of Kutterschitz (Bohemia), schen (Bohemia), Museum Teplitz. Museum of Teplitz. 2. Norma verticalis of skull of Eq. przewalskit Pol. 5-year- 4, Norma verticalis of skull.of a young animal from a old stallion. Photograph by Prefessor Noack, turbary at Sobrusan (Bohemia), PLATE 89. 3 Fic. 1. Norma lateralis of skull of horse from Alemannic tomb at Kénigsfelden (Switzerland). 2. Norma lateralis of skull of Equus przewalskii Polyakoff. Photograph by Prof. Noack. " 3. Norma lateralis of skull of young Sobrusan horse. THE SKULL OF THE PREHISTORIC HORSES. AII The horse of the La Téne period remains, however, till still later times, as is shown by a skull from an Alemannic grave at Ko6nigsfelden near Vindonissa. After this review of the most important finds at our disposal of remains of horses, dating from different periods, we may attempt to compare these finds with each other and with the Anau horse, which, with the horses of the Quaternary epoch and that of the paleolithic age at Solutré, ranks oldest among the horses we have considered. THE SKULL) OF THE. PREHISTORIC. HORSES. THE TEETH. The incisors of both jaws are abundantly represented in the collection before me. Still, notwithstanding the opinion of Wilckens,* I do not believe that their characteristics will be of value in classifying races and species; the phenomena of their growth and abrasion are sufficiently well-known and are widely used in determining age. Although Salenskif says that Equus przewalski has larger teeth than any other horse, this may be inconclusive, since this characteristic is essentially dependent on the age of the animal. Moreover, I have not been able to find out whether Salenski had in mind the whole length of a tooth or measured only the chewing surface. Indeed, it seems from the size of the dimensions that by length Salenski meant the projection of the tooth out of the alveola and called the length of the chewing surface ‘“‘width.’? The value of such measurements is naturally problematic, since with age the incisors can project from 1 to 7 cm., and especially since we do not know from Salenski exactly how old (according to the teeth) were the animals in question. More important than the size of the teeth is the width of the intermaxillary, which gives the means of determining the width of muzzle of the living animal. From Anau, however, we have some specimens in which the intermaxillary can not be measured because the teeth are broken off. I will use here, for com- parison with those from other localities, better preserved specimens of lower jaws. Dimensions of corpus of lower jaw (in millimeters). eee! Diastema Greatest | Seca Height of from in- width. bari até corpus. cisors to : canines. ( nee 40 32 Re PATICUIEMOLSE Guctayersdeelenehere cists 0 st-tns centers oo se 7 Soha 35 58 a Diluvial horse after Nehring, upper jaw....... 84 stots Seer: 18 Diluvial or paleolithic, Schellenken, upper jaw, piglet? WORST cok «eee ete ewan ane vam | 67 ce Boge 18 VOStonmtza male sory ears, . amet ees se ier ie 62 35 24 10 Gross Czernosek,smale;' 8 yearga se ee'\s10)5 5-101: 62 43 25 5 Liebshausen (La Téne), male, 14 years....... 64 40 29 9 Gross Czernosek (Rasch), 10 years........... 56 37 23 Asa PeePTAEZ 5 COLIC gr el, Disiata sinti Ae lsd sm =m Sete ws 8 We 38 27 PANUVERTIICL Wt ee tactics ster tkas srete.s serene ale aaetate cis 60 were avery Alemannic horse, Koenigsfelden............. 61 *M. Wilckens, Beitrage f. Kenntniss des Pferdegebiss, etc. Halle, 1888. Nova Acta, K. Leop.-Carol. deutsch. Akad. d. Naturf. TOp. cit., p. 48. 412 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. The molars both of the Anau horse and of the other subfossil horses are represented by hundreds, but a detailed study of them can be of use only when we possess sets of teeth which surely belong together, as we fortunately have in a complete half of a palate from Anau. After what has been said in reviewing the specific characteristics of ass and horse, it is superfluous to follow the course of each enamel plication. Here we will notice the principal differences; as regards the masticating surface the accom- panying drawings (fig. 491) will speak for themselves. Nehring and later Sal- enski affirm that the relations assumed by Frank, for the Oriental and Occidental horses, between length and breadth of masticating surface (teeth broader than long in the Oriental, and longer than broad in the Occidental horse) are not deci- sive; indeed this relation is greatly influenced by the age of the animal, since through progressive grinding off of the teeth these become smaller towards the roots, and the row of teeth shortens. Nevertheless a comparative tabulation of the size-relations of the whole row of teeth is probably worth making. Table of dimensions (in millimeters). e : e Pre- Pre- Pre- Molar | Molar | Molar y |% \\| S |molar 2.| molar 3.| molar 4. rs 2. 3. SB A. 8 E Grinding surface of a 8 al upper molars. 6 |S) 3 Sisssldlalaleilsislalelala/4ala tit |wi ele de | eee eee) ad Anau (Equus caballus pumpellir) 5-year-old specimen....... 176 |r52 |r00 | 38 |'26 | 32°) 28 | 28. 27 | 25 | 26.27 | Sar Saee pame,, —24 feets can cam ast ne Prt leckee | ameese hee 6 27 | 26 | 26 | 26 |°25 1.26 |) 25aleon Kesslerloch diluvial horse, 8 | Vearssok. Ue ekee eee sta, 1250.1 06,37 | 24-735 es 1.26) 25 bee rege eames | Auvernier domestic horse, bronze age, 6 years........ 153 |127 | 81 | 33 | 22 | 24.| 23 | 23 | 24 | 2r | 24") a1) oe) eee Konigsfelden, Alemannic time, 7 Years, chon oh Peete ees 162 |136-| or 1.37 | 225)°25 | 23 |' 24) 257) 21 |) oe ol tox a ono cme Hostomitz, 1) Vearsee eter 170 |144.| 89 | 35 |. 23.|' 27. | 25.| 22.1 23) 26, )"22.\°26")(23 0) soa 5mies Remagen, diluvial horse, after Nehring, 1O;yearssee eee 169}, 2 | ce | 37 24 29: 1-20) | 27 1h2Q 1 25 | 20 26") conlee vamos Remagen diluvial horse, after Schwatze 12 ite ec etee ae [ves He etigs i. 24 |-30%530))| 289-28 tees eases. Wooniseommee Thiede diluvial horse, after Nehring, 6 years. ..3.: 005% oe | ae | oa] 38 | 26] 32) 207 24 ve P9265] 27.526 | 26 1) Sonam Westeregeln diluvial horse, after! Nebrin golds. ee) een (3). | @) 1-29 | 29 1:28 | 30.) 26 }:27 | 26 | 27 eae Equus przewalskit, after Sal- (\170-) engi 2... teeters ou de Ur85 |b °° 37) 25 1 29 4 294 AT 29 4 FE ee a Equus przewalskit, Mus. Bern..| .. | .. | «. | 40 | 21 | 27 | 23] .. Indian horse, after Nehring, 7 years, Muss Berlina... 4: 154 | se | ool 37 ue23 i) 207126 2s h253| 22 2 es oak 2 caleo camo Arabian horse, after Nehring, 7 WCATSislis ja crelate a kepter steno iers 168 | we} os). 38 1°26 29-427 27 927 4 2427 foe 27 eon Grisonian horse, after Nehring, SeVEAarse oe uee atemueerneek I stelle ois Sale 28 | 2 2 2 20miee 263) 26m 2 a eter (Tyrol) horse, after My : se et a / ie Nehring, 10 yeafs.......0 PST) see iile. Sar 0.628320) 28ale278 270 27a e2Oulusonee Holstein horse, after Nehring, : 7 he : asa a? ; : MON CALS veueteneieselsiistetsie te ers 16 rare fare Saez 2 2 Zou s20n| 2O0\n2 2 2 2387102 Dutch horse, after Nehring, 30 : : ; f : : : ‘ Fears O10 vows aes heen ie II74.] o« | ae 19°37) 234 284126 Le7 82771 25) 279-20 et Laas THE SKULL, OF THE PREHISTORIC HORSES. A13 Fig. 491. (a) Teeth of Equus caballus pumpellii. Right part of upper maxilla. (b) Teeth of Ass, from the Mummy from Abadieh, Egypt. (c) Teeth of Horse, Pile-dwelling of Auvernier (Bronze Age). (d) Teeth of Equus przewalskii. (e) Teeth of Tscherski’s Siberian Horse. The form of the enamel of the teeth deserves further consideration, but since in horses the enamel is more wavy in younger than in older animals, while in the ass the reverse is the case, only specimens of similar age must be com- pared. Frank asserts that in the Oriental horse the enamel plication is less than in the Occidental. ‘The most marked enamel plication of the interior borders of the anterior islands is shown in the diluvial horses; among these especially in the horses of Westeregeln and Thiede. The horse of Remagen has also slighter rip- pling, and I find it still less in the horse of Kesslerloch and Solutré. The plication 414 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. of the enamel of the Anau horse stands midway between the Siberian horse of Tscherski and the diluvial horses of Rutimeyer, on the one hand, and on the other hand, the subfossil horses now before me from Hostomitz, Auvernier, Schuet- tarschen, and Konigsfelden, which have a much simpler enamel plication than animals of the modern heavy Occidental races of exactly equal age. THE SKULL AND ITS PROPORTIONS. In beginning the study of a horse’s skull one asks instinctively: was the skull large or small? Indeed the length of the skull gives a very good rule for deter- mining the withers-height, and thereby also a provisional classification of the horse. It is accepted that heavy horses have the largest heads, and that light Oriental horses and ponies have the smallest skulls. I shall not speak here of the methods of taking the measurements, and will only refer for these to the rule drawn up by me in association with Professor Kraemer at the request of the Deutsche Gesellschaft fur Zuechtungskunde, which will soon appear. Since it is possible to calculate with approximate accuracy the length of the skull from any of its measures of length (the longest measure possible being preferable), I have calculated this for all the incomplete subfossil skulls before me, from the proportions obtained from 50 skulls that the molar row stands to the basilar length as 10 : 28, and to the anterior length as Io : 31. Length of skulls (an millimeters). Length | Length | Length | Length of on base. anterior | on base. anterior face. | face. | Clydesdale horse, after Nehring. | 574 62: | Equus przewalskii, after Salenski: Horse from Boulogne, coll. | } NO:25218'5.c:0-04 9) sras.c mee on EO 543 Duerst:./:. cen = oe eee 543 618 | NQiS2 720 a oe ce a ee | 472 528 Diluvial horse from Nussdorf, | Neolithic horse, Kutterschitz....| 487 543 | after Woldmehiaee see ae 555 | Alemannic horse, Kénigsfelden..| 490 536 _ Diluvial horse from Remagen, La Téne horse, Hostomitz....... 470 O77 alter Nehring: “2051s ene 528 | 562 | Arabian horse, from Abassii, | Schuettarschen, horse of the | | | after Nehring 35. as sees 476 520 iron time (Hallstatt) .......| 506 | 560 | Indian horse, after Nehring...... 438 492 _ Arabian horse, after Nehring...) 500 | 540 | Auvernier, bronze time horse. ... 436 485 | Thoroughbred British race | | Subfossil horse from Gera, after | horse, coll. Duerst.........) 496 | 538 || Nehring sfiyknciue ocr | Anau horse, calculated, 5 years | | Exmoor pony, 15 years, after Old ...25 eae tee eens eeeeee 492 545 | Webring: tin. wer etc tt tee Equus przewalskii, after Salenski: _ Neolithic horse, turbary, Somme No. S213 cee oe eee | 485 542 | (France) a: ede te crea No, 52 8Ose ae ok ae oe ee | 495 547 1) Ass from Abadieh?. pe sei see ac No. 5214 484 538 | Ass from Aden, British Museum It follows from this comparison that the Anau horse had a skull of about the same size as Equus przewalsku and that the other Bohemian subfossil horses, like the Alemannic horse of Konigsfelden, stand very near the Anau horse and In contrast, it appears that the horse of the bronze age from Auvernier has a very notably smaller skull, but still smaller is that of the subfossil horse of Gera and Spandau mentioned by Nehring. And smallest of all is the skull of the neolithic horse from a turbary of the Somme in France, being smaller than that of the smallest Exmoor pony in the Berlin collection; Przewalski horse in size of skull. THE SKULL OF THE PREHISTORIC HORSES. AI5 and yet this little animal seems fully adult; in this we may see an explanation of the errors of Sanson and Rutimeyer. The next question relating to the skull of a horse must concern the indices of ear-load (Ohrbelastung) and of parietal curvature (Scheitelkruemmung). Unfortunately I am not able to produce here a very comprehensive tabulation of these indices on prehistoric horses, since I am able to verify them only on the skulls actually before me in natura. Table of indices. | | ‘ Far-load index. | Parietal crest | Facies tangent. Lesbre. SNe) Uke index: TN. te Be2. Br. } Be. Car. Gey Equus przewalskit, male, 3 years............. woes: 15 25 ee 23 —7 a REMC ELSC MIGZ a eee ctatac sr ctcgt eaeucia aie fee sts Ae Bees pre We eens se. 2 RE ESE Sige IIUU cee oh stale a ucls< 2 «Fao cal pie ule wishe od 20 17 i et a er 30 PCHVELTALSCH ETI = setae area cmasediok estes = Gs 20 ity) 22 18 —7 23 ANTERTES GON Clr es fs Sg RE ee een aaa 23 a: 7 Sled dae ag —7 25 MDOSSeed Ort IMIS OKI ieagaae in cisin tials 3) edeisoss 3 Ore) 15 24 a 20 —7 22 HPameNeHe: | MUIS4 DCL ec crinere ees sisies net « 21 15 Ae 9 ake —7 22 PHICSCEEZ eee FP. sie Crime rece Peas wie eTdiene wesc | 4 4 22 II —6 22 oC ERS ot ie CU Dec ne eb ane ir anne 21 10 25 14 —5 18 Still more important in discriminating between different races of horses is the relation of breadth of forehead to length of skull. We know that, as Sanson has shown, the ass and the Oriental horse as well as the ponies are broad-fronted, while the heavy Occidental horses are usually narrow-fronted, as are also the quaggas and zebras. The following table, based on Nehring’s method of deter- mining this relation, exhibits the order of the skulls of the different races, when arranged according to size. I must remark that, in inserting the frontal width of the Anau horse I have calculated this approximately from the palatal width, which is determinable. The palatal width is to the frontal width as 10:21, which I obtained from a series of 20 Oriental and 10 Occidental skulls; at the same time I ascertained that this ratio is always greater in the Oriental race than in the Occidental. Ratio of basal length to width of frontal bone. EGE: Ass from Abadieh (Egypt) mummified skull, 5 years, male........-........... 210 DScitt Onl PAG ene cAra ia) ssavegromiciial Cae) ire eran cl. einer rns ereltch ik 9222 HOrsentOmerceiancdmalter: Nelrino wOLd a Ialeoe ami eiernennett ena taiaicieleie n22 1, 7 Horse from attire of cbcipsee, alter Nehring c.c- Gch) k os be see oot wee ad 223 PopseirommatunoaLya Oracle? sommes (Hance) eeueri cies oe. cles vA ogee ans) e) sicpe as 230 Morsenromlceland tatters Nehring ,onvearsetialevyariia ss ceila iis sine ce ie cae. 230 HorsedromeArabiawalters Nehring: Si veats ailales migaeiel eases nei abe eres so 230.7 FLOTSEELt ONES 120 UES RV CALS ep tmnaicie a marae Oe ae Sia eeareie lakh? o. viscaieiers main casts 0.0232 Horsetron Gillbrandsdals(skatia)y alteneNeunnpn serra. sates cieere te re oe ees 232007 ES OUUSEDTZEWOISR ID walters Oalensklis ees owe flep goes ciate av. ke peice dueietals ais ies ese) «1 232-244 Horse from Kalmukia, after Nehring, 4 to 5 years, male...................... 239 Horse from Auvernier, Switzerland, bronze time, 6 years, female............... 240 Horse from Grisons (Switzerland), after Nehring, 8 years, male................ 241 4 Horse from Schuettarschen (Bohemia), Hallstatt iron time, old................ 241 Horse from Kutterschitz (Bohemia), 12 m. deep in earth, neolithic............243 Horse from Pinzgau (Tyrol), after Nehring, 10 years, male...................4. 242 Horse from diluvium of Remagen (Rhenania), after Nehring..................249 Horse from Hostomitz) (Bohemia), La Téne, 4 years...........-.2....5:0s-5 253 WorsefromsPimzranecryrol) atter Pranks O1yearsy malesuada sce oc 254.6 Horse from Grisons (Switzerland), after Nehring, old, female..................255 416 THE HORSE OF ANAU IN ITS VARIOUS RELATIONS. We see here that the skulls from the Somme, with a ratio of 230, stand in the same class with those of the Iceland ponies and the Anau horse, together with the Arabian horse and the smallest specimens of Equus przewalsku. Here, too, comes the Auvernier horse of the bronze age, and it agrees well with a Kalmuk horse of Nehring’s series; but the Bohemian horses all belong to the narrow frontal class. Nehring, in his table of this ratio finds the smallest index, on a real horse, on a Turkestan mare brought from Bushed by von Schlagenweit; her index was only 212. He calls broad-fronted all horses with indices up to 240, and narrow- fronted all above this. Tscherski, as we have seen already in the general discussion, distinguishes also medium-fronted horses, among which he includes all with indices between 226 and 240. If we would avoid the expressions broad and narrow-fronted, it would be better, as Eichbaum has mentioned, to use the terms “ dolichoprosopic”’ and ‘‘brachyprosopic’’ than ‘‘dolichocephalic’’ and ‘‘brachycephalic,’”’ since it is not the brain-cap but the face that varies. The diluvial horses of Germany appear at once to have been narrow-fronted, while those of Siberia, of Tscherski, and the Equus przewalsku of Salenski were the medium-fronted. With these we must rank the horse of Anau, and there remains only the question as to where the horses of Solutré and Kesslerloch belong. San- son wished to count them among the ‘dolichocephali;’’ he is, however, opposed by Fraas, who ranks them with the Wiirtemberg fossil horses as short-headed. The studies of Studer* and Heschelerf also, as regards the Swiss horses, have supplied no valuable data concerning the shape of head. BASILAR AND PALATAL REGION. I can not leave the discussion of the skull without touching upon Frank’s proportion of the basilar-palatal region, although this dimension can neither be directly measured on the remains from Anau, nor calculated with approximate accuracy. I have determined the value of this index (which is also influenced by the ear-musculature) in discriminating between ass and horse, and I remark that Salenski as well as Nehring considers it of the greatest value; yet Nehring asserts that a real ass, Equus teniopus, resembles in this respect a horse. But since this index is a weakened repetition of the ear-load and parietal-curvature indices, it is better to use the more delicate method; however this may be, a brief review of this index in our horses is given in the following table: From From foramen vomer to magnum sutura tovomer. palatina. Equus asinus from Sarepta, 5 years, after Nehring.................. he 95 Equus caballus, Exmoor pony, 35 years, after Nehring............. CR 9g! Equus asinus, East Africa, alter Nehting2). ic... «- 09. «as ea ede TOE: 88 Equus caballus Auvermieten -4 orcs ieee ee ethene ine eee 102005 8100 Equus caballus. Kitterschitz.4 css ce set at here een TOR: 94 Equus caballus Turkestan, 10 years, after Nehring ..................-. ise 96 Equus caballus Schuettarschen, Hallstatt time..................055 130. 28) 2 100 Diluvial horse; Remagen) after Nehring’. oe ee eee eae TAO Le *Studer, Die Tierreste aus den Pleistocaenen Ablagerungen des Schweizerbildes bei Schaffhausen. Neue Denkschriften Schweiz. Naturf., 1902. Die Knochenreste aus der Héhle zum Kesslerloch bei Thayngen, ibid., 1904. { Hescheler. A written communication from the author from a study not yet published, received with many thanks. THE BONES OF THE EXTREMITIES. 417 Table of dimensions (in millimeters). ¢ = 4 Equus caballus pumpellit. e janie vrs Pile- Al eer caballus 3 om Turbary | dwelling | 2Utbary| Bronze | Turbary ian ae Cath robustus Skull. cs". of yesh of yon as of Kut- | time of | of So- Kae “ie N 7 £0". |Hallstatt e208! Gomme ice terschitz} Hosto- | brusan Fac eso time of ies, : : (Bohe- | mitz, | (Bohe- oe ‘: | Schuet- 32.4) France. | (Switz- mia). |Bohemi . (Switz- |Salenski.|, h S erland). : ohemia| mia), erland). arschen. emma lett hin G840. 20). 6 sites ves 449 |407 477 543* 497* 528 560* ReMMRRMRRPOIO UN Se econ 8 cc's nO oe oo « 389. (396 |....1 436 487* erin 453* 472 506* Lateral length of frontal bone.......| ..-- |151 |170 | 169 185 ere: 157 I9I ate 198 BEL AMOIAIS 6 hie cn oss dns gst 6 as 62 | 59 67 84 79 rene 72 81 Penetn of premolars... ..... 66.0505 BO bess class 81 89 89 gI ee Aen ecu ol diastema c: «4 (neolithic) 33...) «..- 220 mne2 Ts 44 3 47 toy | ae: 32 13 .9 , 220 | 210 50 4 4 33 25 aa 16.3 Polutre ss Bele tea gece ier acts terete ce ; Pears 50 3x 7 32 23 aa i 000+. cenn aeons Manouvrier 62.8 Weddas, males). 6 hn ces mai ieee eet arene Sarasin 60.5 Thus our Anau tibia with an index of 61.5 approaches the extremest form that is found as a mean value in a race. As was remarked above, platycnemy is usually most pronounced in the upper third of the bone. On the other hand, P. and F. Sarasin found in the tibize of the Weddas that the flattening extended over the upper two-thirds. It is so also in the Anau tibia. Instead of diminishing from the foramen nutritivum (fig. 494, b), it increases downward, so that in the middle of the diaphysis (fig. 494, c) we find an index of only 61.1. The already mentioned curvature of the tibia forward, which is not rarely connected with platycnemy and has shared with this and with the pilaster formation of the femur the fate of being declared rachitic, is clearly expressed on our tibia (plate 95, fig. 5). Also, the posterior surface is more concave than is usual, for instance, among recent Europeans. In order to measure the amount of this curvature for comparison with others, one must lay a straight line from the lowest point (the tibia lying horizontal) below the tuberosity to the deepest point above the edge of the distal joint and measure the height of the ante or edge above this line. This method gives with our tibia a height of 7 mm. Since the distance between the two above-named points is about 230 mm., we find, by expressing the height in percentage of the chord, an index of 3.1. Unfortunately no investigations, so far as I can learn, have been made of this peculiarity of form, which we might use for comparison. Only Manouvrier mentions that he found it in the Guanches together with platy- HUMAN REMAINS FROM THE NORTH KURGAN. 457 cnemism, and he gives a figure of such a tibia (1888, p. 506). A measurement on this reduced picture can, of course, give no exact result, still the corresponding index seems to amount to 3.5. The two Wedda tibie Nos. 178 and 179, on plate Lxxxu of the frequently mentioned work of Sarasin, yield indices of 3.5 and 4.7. In contrast to these values the curvature of the Anau tibia must be considered moderate. A further character of great importance exists on the anterior edge of the lower joint-surface, where this goes over into the anterior sur- Pe a an: oe ere face of the bone. In European ir shin bones a sharp edge usually fi separates the joint-surface from the anterior surface of the tibia. In our tibia, on the other hand, the joint-surface extends in one place over the just-mentioned edge, and makes on the anterior Nees \~—Y” surface of the tibia a small facet whose condition shows clearly e d that it was originally covered with cartilage (plate 95, fig. 3, a). At this point the edge, instead of being sharp, is rounded softly off. On the right tibiathe facet is less large, but it is still clearly present. The difference in the eS Gores, form of the joint-surface from a that of a Europeanis best recog- nized in a sagittal section that was drawn with theaidof a lead Fig. 495.—(a) Sagittal section of joint surface of condylus lateralis tibiae : from Anau I. (b) Same section from a European tibia. wire through the place of the (c) Sagittal section through lower joint-section of left tibia facet (fig. 495, € and d). through facet, Anau I; thick line marks cartilage cover. (d) 2 gh’: Same section through European tibia. (e) Sagittal section This characteristic has been through medial edge of trochlea and medial neck-swelling. found among different lower (f) Sagittal section through trochlea and lateral neck-swelling. races, and also in anthropoids; thus by Thomson (1890, p. 213) in the gorilla, orang, and in the Weddas; by Martin in the Fuegians (1894, p. 198) and Senoi (1905, p. 635). The occurrence was already rightly interpreted by its first observer, Thomson (1889, p. 624), who sought the cause of this facet in frequent strong dorsal flexure of the foot which occurs especially in a squatting position. Let us look a little more closely into the mechanism of the object before us. As one can see in a skeleton, in strong dorsal flexure of the foot the talus presses with the anterior upper surface of its neck on the forward edge of the joint- surface of the tibia. In this action it is especially the medial and lateral edges drockiea Sea Proehien. (4 fi swelli rig 458 HUMAN REMAINS FROM THE NORTH KURGAN. of the neck which are subjected to the pressure, while the central depressed part is less affected. Accordingly the cartilaginous covering of the joint is spread especially over the two lateral swellings of the neck; the depression between these remains free (plate 95, fig. 3, 6). But now the two swellings act very differently. The medial swelling presents a direct continuation of the medial edge of the sur- face of the trochlea, as we can see by taking a corresponding impression with lead wire (fig. 495, e). The action of the lateral swelling is wholly different. It rises sharply from the surface of the trochlea (fig. 495, f). Consequently, when the foot is flexed in a dorsal direction, the medial swelling slides smoothly under the articular surface of the tibia, without altering essentially the edge of the latter; at most it deepens a little more the depression of the articular surface at the base of the malleolus. On the other hand, the lateral swelling presses itself against the edge of the articular surface, flattens it and produces the above-mentioned overspreading of the cartilage on the anterior surface of the bone. If this explana- tion is right, the edges of the cartilage-covered surfaces on the astragalus and tibia must fit each other exactly, when brought into an extreme dorsiflexion. Indeed that is so in our case. ‘This, too, gives a certain proof that our astragalus belongs really to the fragment of tibia. That the squatting position gives rise to this strong dorsal flexion is shown by the fact that the change just described on the lower joint is often accompanied by a corresponding one on the upper joint. We find there often a rounding off of the posterior edge of the joint-surface, which shows itself especially on the lateral joint-groove and causes the joint-groove to appear convex in its posterior segment. Also this characteristic is indicated on the left tibia head, even if not very marked (fig. 495,@). It would correspond about to No. 2-3 of Thomson’s scheme (1890, p. 211). The cause of this variation is doubtless to be found in the strong flexure of the knee, through which the posterior, upper surface of the condyles is brought against the posterior edge of the joint-surface of the tibia. It is at least doubtful whether the backward divergence of the head of the tibia (plate 95, fig. 2) is also produced, or increased, by strong bending of the knee, since it seems to occur in cases where such a function can not be shown as a cause. An examination of this divergence by Manouvrier’s method (1893, p. 231) gave an angle of inclination of 10°. The angle of retroversion could not be measured; with the considerable curvature of the tibia it is not possible to speak of a straight diaphysis axis so that we have no criterion for the position. An inclination angle of 10° lies perfectly within the range of variation of modern Europeans and exceeds but little the average value, 8.5°, found by Manouvrier for 72 European tibize (French) (1893, p. 236). The described curvature of the thigh bone, as well as of the tibia, by enlarging the space for the flexure muscles of the upper and lower part of the leg, facilitates squatting; this is so self-evident that one might consider whether this habit did not contribute to the increase of those curvatures. Also the low position of the condylus medialis tibiz seems to be more common among ‘primitive ‘peoples (plate 96, fig. 1). HUMAN REMAINS FROM THE NORTH KURGAN. 459 The relief of the bone is well elaborated, the crista interossea standing clearly out, especially in its upper part. In front and parallel with it there runs a depres- sion through the proximal two-thirds, which shows itself also in the cross-section (fig. 494, b). There is scarcely an indication of a crista posterior. The measurement of the torsion angle gave 26°; but in view of the extremely great variability of this characteristic it can scarcely have racial importance. That these upper and lower leg bones come from the same individual is so certain that I may base on it an investigation. If we compare the length of the tibia with that of the femur, we find a femoro-tibial index of 84.9. In doing this we have measured the length of the tibia without the spina intercondy- loidea, but have included the malleolus, and have used the length of the femur in its natural position. Ina similar manner the following values were found: Purapeans (Topinard)... ... ve. s. ose 80.8 | Andamanese (Flower)................ 84.5 iiropeans(hlowef): ie.kt od cose ae es 3 S21 Nevroesi(Hmmphtey hy rs cies ec ce eee 84.7 Pregians (Mattiny Giese 5 ov cease ce mss Somom MeAtStraanss(MIOWeE)ie te cele inc ie < ols.0) 5 84.9 For Senoi Martin found (1905, p. 642), by applying the condylo-astragal length of the tibia and using the results of Turner, Duckworth, Annandale, and Robinson, a mean value of 81.7. If we apply the method of these authors we shall have an index of 82.8. The race to which this Anau individual belonged is to be called moderately dolichocnemic; that is, the lower leg is relatively long in proportion to the length of the thigh, a peculiarity which presents itself as a primitive charac- teristic not only in that it occurs more often in primitive races, but especially because the new-born European has also a relatively long lower leg. Of the fibule we have unfortunately only a few short fragments from which but few inferences can be drawn as to their complete form. Therefore we content ourselves with the presentation of a statement of the absolute measurements, which will be found in the appended tables. These bones also show that sharply expressed modeling which seems to be a characteristic of these Anau individuals. The talus also shows several points of interest. In the first place the consid- erable development in width. If with Leboucq (1902, p. 144) we take the length of the talus from the highest elevation of the head to the sulcus pro musculo flexori hallucis longo =100, then the width from the lateral point of the fibular facet to the most medial point of the processus posterior left =82.5. This index amounts, according to Leboucq, in the mean among Europeans to 77.0. Martin found it for Senoi =79.5 and 80.9. The Spy talus has 91 (Leboucq). Thus we find here, too, an approach in this Anau skeleton to primitive forms. Next, the narrowing of the trochlea at the posterior end. According to Volkov’s deter- minations this narrowing seems to belong in a higher degree to primitive races. Also among new-born Europeans it is strongly marked. If, according to Volkov’s method (1903, p. 695), we take the anterior width as roo, the relative width of the posterior end is among— New-born Europeans...:......esse0+. OL ay cers OVAIOS F016 2. vale s key 6 elses =e 79.0 Wa panese, MUalES Are sielasiaetlereels siete le AMS ao ly MesiAtsmiAleS mcrae sister es)e'e's cheers) = 79.4 Negritos; malesacccs acne tm e Orne ects Aen Tel USKAIN OSIM ALES Heres, «aise iene slolene.es = © 6 80.3 ANIStralians ptrlales erties a eierateters) oetste eas Ae a ES ODCRNS EMALCS i.e vis fo is vias «Wess Sra3 Melanesians stialessan we fisctcie iets: Gee MIE OLOES SINGLE ercia's. ch ons a6 0s «tas 60,85 81.9 Weddas = males eects sritewn sarectcuste sts ots hee 460 HUMAN REMAINS FROM THE NORTH KURGAN. The same index for the trochlea of our Anau talus on both sides amounts to 79.4. The narrowing is therefore somewhat more pronounced than in adult Europeans, but does not attain any exceptional degree. On the other hand the Anau talus differs in another characteristic considerably, not only from Europeans, but also from all other races concerning which we have the results of investigations, viz, in the angle by which the axis of the neck and head differs from the longitudinal axis of the trochlea (plate 96, fig. 2). For this angle Volkov (1903, p. 706) gives the following values: Negroes, sales Fateecr owe wee nree ter aye Hskimos;-males (ere as dees au ier -) he Melanesians males #2... eee ne cole 23.4°) Weddas, males... waiacceitine senetetnts 20° Nepritos, malesac omelet rents ceo 23° Patagonians) imales seeac cet oieini ie ceils 20° Eitesians? malesey arene ere eee ee 237 Europeans, males. bie dees 0s vaean at iy ice In Anau I the deviation of the neck for both sides is 31°; it is therefore higher than the average of any of the races enumerated. But even individual values as high as this were found by Volkov only in new-born children, where he measured angles up to 35°. The values given for Japanese up to 47° are clearly wrong, since the mean is stated as 19° and the minimum as 28°. Still, Adachi found among Japanese a mean of 19°, but individual variations up to 32°. Thus the Anau talus takes an extreme position in this clearly primitive characteristic, only Japanese occasionally showing as high a divergence. The torsion of the head of the talus is important, because the transversal arching of the foot is in part dependent on it. According to Volkov’s investiga- tions (1904, p. 320) its values form an increasing series: New-born Europeans...........+.++%. 16..5°) (Negroes... 5). <0) dante oie oe 36° Neégtitos tt etka oe ene eee 34° Japanése ye ise ts aie eens cae eens 39° Melanesiansimiie. «caches ree ee Sno EMO peas sc... s'90.4n pind tae veel oe 49° In the Anau talus the torsion is less than in any of the races investigated by Volkov, 30° to 33°, and is therefore a primitive characteristic of high importance. The calcaneus also shows peculiarities which differ from that of the European. Its sustentaculum tali is very strongly developed; as Volkov has shown ‘the lower races’’ form in this respect “‘a real transition between the foot of the anthro- poids and the foot of the European.”’ Plate 96, fig. 3, shows this difference. Still another characteristic is to be seen in this figure, one that points to a slight height of the arch of the foot. The joint-surface of the calcaneus intended for the cuboid is more wide than high, and, looked at from below, is less visible than in the European, because it stands more vertical in relation to the longitu- dinal axis of the calcaneus. The wide, low form of this joint-surface is especially striking in Anau I. We find in this a breadth-height index for both sides of 62.5, while 5 Europeans gave a mean value of 92.0. The anterior inner joint-surface for the talus on the left is divided into two separate facets, while on the right these are connected by a narrow isthmus. The right naviculare is very thick on its medial edge; on the lateral, on the other hand, itis narrow. An index which we calculate, taking the thickness at the inner edge as 100, amounts to 47.4. It is interesting to compare this with the figures given by Volkov (1904, p. 38) as shown at top of next page. HUMAN REMAINS FROM THE NORTH KURGAN. 461 Weddas smyales 2. oe 0isn/cn vans ons aes AAT Ome apanese sinales were vee he i 5am PROPRIA AUAICS sere San a de Coons 45 INGPTILOS witaleseer npr te een oc see we 54.5 Negroes, maleseym orth tec to Aceon meAtistrallanswmalesan) se ee eee. 54.8 Polyn€siaus, ‘males .2 06 o..2.259 ess enjea ea Srie Maroneats- 7 SOMO NESAticn IMAlCS crea srs shore 6 es once ob 86.9 PESKANIOS SUSICS eS oo oe (ae an Oates es 8 os 80.0 This index also approaches the primitive forms. In cuneiforme I we are especially interested in the difference in its height at its proximal and distal ends. The last measures according to Volkov (1904, p- 204) in percentage of the proximal among: BO UTO PEARS ATIAIES are tei, ele eaters lotare tore av D2 To aN CRTICOS WIMAlCS treats. Sacco, sits veces ec 6 2's 139.5 Neoroesoialesmae ye seein. ences 3 > L2GHOM MASI Alans sMtales sate seve sce 'sie renee 140.9 Melanesians mmales ses etter Cees 6 027/57 We find for this index a value of 160.0, which, if inserted in this table, would range beyond the most primitive forms. It would perhaps be more appropriate to bring the two joint-surfaces into relation with each other; the proximal would then measure 56.7 per cent of the distal. The slender form of the metatarsus (plate 96, fig. 4) shows itself most clearly in the relation of the epiphysis width to the length. Since for comparison we have to use the investigations of Volkov, we must, in calculating the indices, choose the length taken by him, which reaches from the middle of the upper edge of the proximal joint-surface to the posterior edge of the first phalanx. Though we do not have this, it is not difficult to say how far it may have extended, and we shall not be far out of the way in assuming a measurement of 61 mm. for the right and 60 mm. for the left. The index for the basal width would then be for the right 31.1. Volkov (1904, p. 238) finds for: Melanesianss mialesia. a.laccs secs es cle PITS eta PANESE, BIOS medial s gos cia tale evs dha dba bce en 28 ma metiorizouta, pandradien,” medial). ose es fe cade pews es. 35 Saar Mem INGexTOlateralewOONdy Adlets «aie 6 «eettie seis tie aids ou oe sles S75 87.5 Aree IG eRNOL MCCA MeO WOVUALLIE « mevnle viele isrsiaie viel vie eis see ¢0ine fv 80 ee a8 46. Joint circumference of condylus lateralis................++: 115 114 47. Joint circumference of condylus medialis.................+: 104 106 PESTER OSL IMIG Ve WICC oie) ccce dete cies anlk ae cee A ely = xe eae ae Sn 82 | 81 Mom ELIS MteOl CONG VANS ALCLALIShp a rnianiaic Sele cule a> ie 69 66 52. Index of diaphysis curvature, after Manouvrier (No, 51: No.2). T Set ae LARS ere 53. Strongest diaphysis curvature (reciprocal value of radius)..... 4? 4 Be AME ATL CLEZOT LOL SION Mere aenctereriNeysict chats. costeve © are, sal Shela cue lerapelies a) weer 262 28° eta ee OLN) “Clb VIS BUS LO irre Ate CEC? pip wd eae 5 2 oaks 6s 5 4d od ees 135 138 133 Ow ORE CICADAS ANON... cm tu shad vas ats ch cals s drae nee he 9° nigra yume Oe stort nn ie. OF COLUM: Git a ON inle ovis Aves « sedan Ne oa gwd = S30 33° ca. 30° i Semrt OLeTOb ited ODT Ui ee tee ar cceters ie aoe cia, wusicl ose ol aa seo eiecel dm. 19° T7o 20° 466 HUMAN REMAINS FROM THE NORTH KURGAN. Table of dimensions in millimeters —Continued, ; if Il Pelvis. Lett. Left. ti Vertical diameter ot meetabulisoiians tee eee eer rare tens 57 Ret 2. From posterior edge of joint-socket, to incis. ith... sn. ce ene t see on 40 % 3. Thickness Of tuber 2155 < bcaies e's > vie aevn kobe © poeccate onan yee aeons len eee 500. Fic. 11, Ash-skeleton of long epidermis-cells in potsherd from Anau, 500. a“ £ - : e; a. 4% - OWE, +4 Lag al . > ea St ; pee Ss: : ol ‘oe ~ - ~*~ a se oe eee PART X. “4 by STONE IMPLEMENTS AND SKELETONS EXCAVATED IN ANAU By LANGDON WARNER. _ [CHAPTER XXIV. ] CHAPTER XXIV. REPORT ON THE LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. Of the larger stone implements found in the northern kurgan at Anau and not treated of by Dr. Schmidt in his report, there are few types, and those are roughly made, common utensils. Commonest and most important of all are the flat or gently curved mealing- stones on which the household supply of flour was evidently ground. These occur at intervals all through the culture-strata, and are sometimes accompanied by the muller, a more or less smooth, cylindrical kind of stone, which was rolled or rubbed over the flat surface of the larger ones. All the mealing-stones found in this kurgan—and they were many—were composed of a fine-grained quartzite conglomerate. In form they were the exact counterparts of the ‘‘metates”’ used to-day by many Indians of Arizona, New Mexico, and Mexico, to grind their wokus and parched corn. It is probable that these were made by cracking apart a small quartzite bowlder and rubbing the two new surfaces together till they became flat. Then, if the mealing-stone would not stand true, it was either chocked up with pebbles under the curve, or in some cases chipped flatter with a few blows of a stone maul. I am convinced that when first made, these ‘‘metates’’ were perfectly flat, and took on the shallow curve we found in so many only after a considerable amount of use; and also that the muller was held fast in the hand, and not rolled, for the only two I found, one of which lay on its mealing-stone, showed worn facets running longitudinally, which intersected the original curve of the cylinder (see fig. 496). There were in this kurgan no true mortars for grinding grain or roots, although in the southern hill we found almost as many of them as of the flat mealing-stones. The single possible exception to this was a barrel-shaped stone 9 inches high, with a round depression in one end 2.5 inches deep and 4 inches in diameter (see No. 4, fig. 497). My reason for not classing this utensil with the flour mortars of the South Kurgan is that, unlike them, the inner edge is sharp and delicate and would not allow for the gyroscopic wobbling motion of the pestle which gives the grinding crunch so necessary with kernels. It may have been used to powder earths or charcoal for colors, or even for ceremonial uses as a mere receptacle. This stone was also grooved longitudinally, perhaps for after-use as a weight. In this respect it is obviously to be classed with No. 43, fig. 498, found in the South Kurgan. Figs. 499 to 504 represent a number of small quartzite dishes or saucers of indeterminable use. For the most part they were fairly well finished both inside and out, and suggest ceremonial intention. It is hardly possible that the oval form would be used for crushing anything, so they must be classed as receptacles, 477 478 LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. Catalogue No. 39 A.N.K. can best be described as a “doughnut-shaped’”’ stone of fine-grain quartzite conglomerates» Its diameter is 3 inches and that of the hole 0.7 inch (fig. 505). Its use can be only conjectural—perhaps a ceremonial mace-head, perhaps a chucking-stone in some game. Fig. 506, from +30 feet in terrace 11 of North Kurgan, I have classed with figs. 507 and 508 and also with the stones found in the South Kurgan (figs. 509 and 510). It is a large stone 10 inches by 17 inches by 2.5 inches, weighing 33,3, Russian pfund (=13.59 kilos). It has a hole cut near the top to form a handle, which is much worn by use. I know of no analogous form among stone implements 4 '‘ 1 | A 3 H Ne) 4 1 i>) ; H ' if Vv A . <-- 6m. ----> i} = v 2 2M sg y — Art * = Zz CN yy Z ae : 4, 1% 1 Zh Bae 1 Ly oe 1 ! : A ! l ! I ae y ' t j 1 t <---54in- --> en Ala eet: «<--- /0 In. ---—> 502 493 Figs. 496-502.—Stone Implements from the Anau Kurgans. Fig. 498 from the South Kurgan, all others from the North Kurgan. of other countries. Professor Pumpelly suggested the possibility of its use as a standard of weight. This seemed the more likely when we came upon other smaller stones, all broken, but showing the same form. From the loose wash earth of the South Kurgan the workmen took a millstone-shaped stone (fig. 511),12 inches in diameter, with a 2.5-inch hole in the center. The edges were much scarred and chipped in a manner that could not have come from horizontal use against a similar stone; nevertheless, I took it to be a more or less modern implement from a Persian mill, of which there were several on the little watercourses flowing from the mountains south of our work. LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. 479 Fig. 498 (A.S.K. 43) shows an oval stone of great weight, grooved longitudin- ally. One can only conjecture its use, but stones not unlike this are used in other parts of the world for anchoring the hide roofs of huts, for straightening the green wood to be made into bows and lances, and for ceremonial purposes. Fig. 512 (A.S.K. 13) shows a fragment of a round stone disk, with a shallow, saucer-like depression on one side, across which runs a smooth groove, apparently made afterward. This groove on a stone, if found in North America, might be thought to be the straightener and polisher for arrow-shafts. Fig. 513 (A.S.K. 34) is a shallow saucer, broken, but once evidently oblong in shape, with rounded corners. 508 503 WH z i. Wy WH y wt PE a ae <--- 4gin.--> <---8in,---» i 4 <----- £0 Figs. 503-508.—Stone Implements from the North Kurgan. Fig. 514 (Spec. Finds Cat. S.K. 325, plate 48, fig. 11) shows three views of another disk with the saucer-like depression, and a deep groove running across the bottom, not quite intersecting the center. Fig. 515 (A.S.K. 42) is the ‘‘door-stone’’ found in place with the rest of the threshold in terrace B over skeleton No. 27 (see “ Report on Burials, South Kurgan, Anau’’). It was a rough, unshaped piece of fine-grain quartzite conglomerate, with the well-defined marks of a swinging pivot that had left ridges in the hole. (Cf. Dr. Schmidt’s report.) Several others of this sort were found scattered through this digging. 480 LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. Fig. 516 shows a stone cupped not unlike a door-hinge stone, but lacking its characteristic ridges; these for want of a better term I call “‘cup-mortars.”’ The pestles and mullers of the South Kurgan show many more diversities in form than the simple cylinder of the northern. Figs. 517 to 524 show the shapes found. Fig. 524 (Spec. Finds Cat. $.K. 196, plate 48, fig. 8) from +21 feet in Terrace C, South Kurgan, is included among them because of its evident after-use for pounding, but it is doubtful if it was intended originally for that purpose. The : A ' ' \ ( : AS N ; % : ! = a y <----0/gin.---> «---10Yin.----> A PRL eg : = S z % Se ee Vv <---6in. --> A i : bs es v 613 * | t ayWiy, \\ +, ' N 2 at. Bea ’ :F N Z : ; MW, we | Vv <3fin. > LIL $090 rae 516 Figs. 509-516.—Stone Implements from the South Kurgan. implement shown in figs. 525-527 (A.S.K. 119) owes its shape only partly to design, being a curiously weathered stone adapted by man (see plate 48, fig. 7). The same may be said of fig. 528 (A.S.K. 220, plate 48, fig. 2). Fig. 529 (A.S.K. 53) shows a cylinder broken at both ends, made of micaceous schist. Fig. 521 shows a broken pestle of the cylindrical type found on the North Kurgan. It is similar to that shown in fig. 496, except that it lacks signs of lateral after-use as a muller. LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. 481 By far the best worked of the larger stone implements taken from either kur- gan, if we except the mortar shown in fig. 532, is shown in fig. 530. The shat- tered end is roughly squared; the other has a circular cross-section and is carved with beautifully smoothed flutings, converging to a point at the center. The ridges between the flutings are sharp and their bottoms are even and round. If the ancient inhabitants of this kurgan had other use for it than ceremonial, it must have been on soft material, for the repeated, and seemingly intended, blows which have shattered one end have had no corresponding effect on the other. Fig. 531 shows a large mealing-stone found in the South Kurgan, but typical of those of both, after they had been worn down in the middle by continued friction with a muller such as is seen in fig. 496. ' A A - / ‘ s ‘ § § ¥ ' on) Ny v 1 y i] v a : a La] e v ee Ci iioce A ‘ y 1 A : : ‘ : : XS) . it : ' ‘ ' . Y ¥ <-f A I.-> «--2/2 in.--> 520 524 523 Figs. 517-524.—Pestles and Mullers from the South Kurgan. Figs. 532 and 533 represent a highly developed form of stone mortar, which is all the more interesting from its close resemblance to that still in use. Fig. 532 is about 5 inches high. It is from the IV or Iron Culture at the top of the South Kurgan. The information we can get concerning the daily life of these people is but slightly augmented by examination of their larger stone implements. We find that by far their commonest household tool that has come down to us (not counting the spinning weight described in Dr. Schmidt’s report) is the mealing-stone and 482 LARGER STONE IMPLEMENTS OF THE KURGANS AT ANAU. muller with which they ground their flour. But whether it was a grain, wild or cultivated, that they used, or whether they were acquainted with some tuberous root, there is nothing to show. What, then, brought in the mortar form in the younger kurgan, where it exists side by side with the flat, matate-like stones? Did they find the cup easier for pounding cereals, or did they, perhaps, with the new culture, learn the use of a new vegetable from which they extracted juice? Did the more elaborate pestles of the younger kurgan mean different uses, or a budding artistic sense? Nothing but more extended excavation and better oppor- tunities for studying the great climatic changes could ever tell us. 526 J, ‘ | \| SN ang nee ee eee Figs. 525-528.—Stone Implements from the South Kurgan. SUMMARY OF WORK DONE IN TERRACE IJ], NORTH KURGAN. On March 28 I was given charge of terrace 1, which had then been in process of excavation three days and been sunk to a level of +26.5 feet above datum. The digging extended north from the end of terrace 1 and measured 20 feet on the sides and 9 feet and 7 feet respectively on the south and north ends (see fig. 31). At +26.5 feet we came to a layer of fire-hardened earth extending out from the edges of a circular hole which was filled to the mouth with ashes. This hole SUMMARY OF WORK DONE IN TERRACE II, NORTH KURGAN. 483 proved to be a sort of oven in the burnt earth, 5 inches deep and 6.5 inches in diameter at the rim. The ashes it contained were fine and white. It occurred 3 feet down the terrace from the south end and 4 feet from the east wall. From 6 inches below the level of this hearth and some feet to the northwest of it, I took the child’s skeleton No. 6 (see Report on Burials of North Kurgan). Below the body as it lay was a layer of ashes and burnt earth in which was em- bedded a part of the base and side of a huge thick jar containing a greenish-yellow slag of fused ashes, and what probably had been bone. March 30, I took from the floor of the terrace at the same level (+26 feet) skeieton No. 7 (see Report on Burials of North Kurgan), which lay contracted on its right side and directly over a large stone with a small cup-mortar in one end (see Report on Larger Stone Implements, North Kurgan, fig. 497), the lip of which stood at +25 feet. This, however, was not taken out until later. 529 530 « A ' ' : ' s i * § v x 3s ' — OC + «------------- — lGinths - —-----—------- > Figs. 529, 530.—Stone Implements from the South Kurgan. 531.—Type of Mealing-stone in both Kurgans. 532, 533.—Stone Mortars from the South Kurgan, Culture IV. On April 1 we came upon a pithos in situ 12 feet south of the north end of the terrace, at a level of +26 feet. It was of heavy, undecorated clay. It contained earth and some few streaks and layers of white ashes. At the same level in the northwest corner of the terrace appeared another of the same sort, but much more badly wrecked than the first. Six inches lower down and some feet away lay another pithos, which proved like the others to be full of earth and ashes. At +25 feet above datum, and directly below the first hearth and oven-like hole, occurred another of the same sort. The earth was baked red for a foot or so about the edge of the hole and the same fine white ashes were found inside. Three 484 SKELETONS EXCAVATED IN NORTH KURGAN. inches below the level of this hearth and out in the middle of the terrace was a plain undecorated pot like the others found above. Near it lay a half-burnt clay brick measuring 9 inches by 5.5 inches by 2.5 inches. This lay in situ, but was found to be broken into many fragments. At this same level and directly below the site of skeleton No. 7, but lving to the west of the stone cup-mortar also found below it, were two pots in situ, showing remains of a decorative design done in black. On April 4, 5, and 6, I excavated the four skeletons numbered 11, 12, 13, and 14 (see Report on Burials of North Kurgan), all at the level of +22.5 feet. They all occurred over layers of ashes mixed with bits of charcoal, and they all had near them burial objects, such as white stone beads (N.K.114,144; plate 4o, fig. 2), carnelian beads (N.K. 114), lead tubes and copper corkscrew spirals (N.K. 185; plate 36, fig. 1). With numbers 12 and 14 air-dried bricks were laid to form a partial sepulture. No. 12 was found beneath a layer of ashes as well as above one, though in all probability the top layer belonged to a later period. The last thing excavated by me in this terrace was a large pithos. It lay in the northwest corner of the digging and contained besides fine earth a mealing- stone of gently sloping surface (see Report on Larger Stone Implements, North Kurgan), on which lay a muller (see fig. 496) of cylindrical form, the battered end of which showed former use as a pestle. This was the only case in either kurgan where mealing-stone and muller were found together. Dr. Schmidt in his report has treated the occurrence of the objects excavated in this terrace so fully that there is little to add. As it is just here that the transi- tion period between the older and younger cultures of the kurgan occurs, his results and classifications of the finds in one or the other group is of the greatest importance. REPORT ON SKELETONS EXCAVATED IN ANAU. NORTH KURGAN. When it became my duty to expose the skeletons in the Anau kurgans as they were come upon by the workmen, there had already been several fragments of human remains brought to light. One skeleton, more or less complete, was taken up in the north digging and reburied, as its position and orientation showed it to be the more or less modern grave of a native Turkoman. During the first few days there were also taken from the first terrace fragments of children’s skulls belonging to at least five individuals. When it had become clear to Professor Pumpelly and to Dr. Schmidt that human remains were likely to occur in more than a fragmentary, and haphazard way, I was given the work of excavating them and noting their occurrence. In every case where it was practicable, a large circle was drawn around the remains as soon as they were discovered and the men carried on their work outside its limits. In this way the floor of a terrace would be carried down 3 or 4 feet, leaving the skeleton untouched on a pedestal. This method was found to be of great help when the actual clearing of the bones was begun. SKELETONS EXCAVATED IN NORTH KURGAN. 485 The material of the kurgan was so closely packed as to make fine work extremely difficult, and nearly all the skeletons were so delicate that exposure to wind and sun destroyed them. I found that a soft brush and a fine knife- blade were often none too nice for the work. When air-dried bricks appeared, often the only way of detecting them was by the faint outlined contours in the cutting, for they were as easy to work in as the material of the hill itself. The layers of ashes and charcoal that so often occurred under the skeletons as well as in other places came as a pleasant relief to the workers. The human remains were uncovered from above and drawn and photographed while still embedded enough to be supported. When this was done the upper bones were removed and the position of the hidden portions carefully noted. Then the earth for some distance around was thoroughly searched for burial objects. In taking out the minute beads found with some of the skeletons, notably No. 9, it was found that even a very close sieve could not be trusted and that the fingers were more sensitive and sure. In this way many square feet of earth passed through my hands, and though the work was delayed it proved well worth while, for from that burial alone we took 1,066 drilled beads, each scarcely larger than a pin-head (see plate 4o, fig. 5). The three highest skeletons were excavated by Dr. Schmidt :* 7 at +36 feet in terrace IV; d at +34 feet in terrace v, and ¢ in terrace v, at +30 feet 2 inches above datum. Skeleton No. 1 (¢).—For the description of the latter, I quote from his notes: This child’s skeleton is 62 cm. long, with its head to the south lying on the right side, but disturbed and partly destroyed by the pick. The trunk is on its back, inclined to the right,the left arm on the left side with that hand by the pelvis. The right arm is destroyed. The pelvis bones are spread to the left and right. The right leg, bent at the knee, is on the left side. The right leg is ill preserved, with the upper part drawn up and to the right; the lower bones of it are lacking. Skeleton No. 5 (a4).—The next skeleton in order of altitude was one I have called ‘‘No. 5’’ (see fig. 534). It lay at an altitude of +30 feet in terrace 1. The bones were those of a young child, for the cranial sutures gaped wide and the epiphyses separated off the long bones. The body was placed on its right side, lying in a contracted position, with the head toward the southeast. The left arm * Das beruht auf Missverstandnissen. Ich finde in.meinem Tagebuch folgende Notizen: *“Sonnabend d.2. April. * * * “Terrasse IV. Es wird wieder das Skelett (7. e., y) eines Kindes in der Lage der liegenden Hocker auf der rechten Seite von Herrn Warner freigelegt und photographiert. Beigaben fehlen. “Terrasse V. * * * Neben dem Topf (z.¢., +34.5 feet) etwas unter dem Niveau, auf dem er steht, werden Skelette aufgedeckt und unter diesen die Spuren eines tieferen Topfes (7. e., +33 feet) * * * Herr Warner legt die Skelette frei; es ist ein unentwirrbarer Haufen von mehreren Kinderskeletten (7. e., 0), der von ihm beseitigt wird. * **Dienstag den 5. April. * * * ““Terrasse V, ca. 20 cm. unter dem Niveau der Kalkschicht, siidwestlich davon, neben der aufgedeck- ten Mauer, die Reste eines zerst6rten Kinderskelettes (7.e., ¢). * * * Das Kinderskelett ist 62cm.lang. Kopf nach Stiden auf der rechten Seite, aber verschoben und durch die Hacke z. T. zerstért. Rumpf auf dem Riicken mit einer Wendung nach rechts, linker Arm auf der linken Seite, linke Hand in Beckenhohe, rechter Arm zerstért; linkes Bein an der linken Seite und im Knie gebogen, rechtes Bein nur unvollstandig erhalten, rechter Oberschenkel nach rechts und nach oben gezogen, rechter Unterschenkel fehlt.”’ Also nur Skelett ¢ habe ich herausgenommen und zwar weil Herr Warner gerade in Terrasse II be- schafligt war. Dr. HUBERT SCHMIDT. BERLIN, d. 29. Marz, 1908. 486 SKELETONS EXCAVATED IN NORTH KURGAN. was bent in such a position as to bring the hand opposite the face and on a level with it, while the right arm was extended down at an angle of 40° with the trunk. In front of this body and parallel with its main axis lay two air-dried bricks, 7.5 inches long by 2.5 inches wide, set on edge. At a right angle with these, 1.5 inches from the top of the skull, I came upon another brick of indeterminable length, but the same thickness. The three were laid in so deliberate a fashion as to suggest an attempt at sepulture. Between the heels and the end of the spine four lapis-lazuli beads (see N.K. 50, plate 4o, fig. 6) came to light, drilled from both flat surfaces so that the hole was roughly double-conical. Four inches back of the neck appeared a smooth, clay, plummet-shaped object, possibly also a burial gift. Skeleton, bricks, and gifts all lay on an even bed of ashes mixed with small pieces of charcoal, resting on a layer of hard-burnt earth. Skeleton No. to.—In terrace vil on the southerly slope of the kurgan, the workmen came upon traces of human remains at +29.5 feet. These proved to be the jumbled bones of a child, evidently hauled about and dislocated by an animal, for I found a burrowextending straight down through the middle of it all. Two feet below this level, in what seemed to have been the bottom of the burrow, the missing bones that belonged above were found, together with two crania of small rat-like rodents. ‘The loose earth that had filled the hole from the hillside above had allowed so much dampness to enter that all the bones were in an extremely fragile condition. Skeleton No. 15.—In the same terrace vil, and at the same level, + 29.5 feet, though nearer the outer edge of the hill, were found the cranium and a few vertebrze and long bones of a young child. The position in which the body had lain was indeterminable, and the bones fell to pieces as soon as they had dried. Skeleton No. 9.—On April 7, in terrace v, we came upon a child burial at a height of 29 feet. No traces of the cranium could be found, though four teeth lay in an orderly row as if there had been no disturbance. The main axis of the body was approximately southwest to northeast. In the softer earth about the pelvis and lower limb-bones, I took out 1,066 minute white beads, apparently of stone (N.K.222, plate 40, fig. 5). They were cylindrical, about ;,-inch in diameter and length, and so delicately bored that a very fine needle was required to thread them. Their presence on the leg-bones and pelvis, and their absence on the upper parts of the body suggest that they might have been sewn to a kirtle or other garment, and not used in strings, as were the larger beads we found later. Skeleton No. 2 (3).—The next burial in order of altitude appeared in terrace I at +28 feet. It was the skeleton of a young child lying on its right side in a con- tracted position. From the top of the cranium to the end of the spine measured but 13 inches and the knees were so drawn up that the greatest width of the body in position was 8 inches. ‘The main trend of the body was southwest and north- east. The right arm bones lay parallel with and behind the vertebre, the left arm bent to bring the hand palm down in front of the face. (See fig. 535.) When the bones were removed they were found to have been laid on a hori- zontal layer of wood-ashes and charcoal varying from 2 to 3 inches in depth and SKELETONS EXCAVATED IN NORTH KURGAN. 487 extending a foot or so beyond the body on all sides. The weight of the earth above had broken in the skull along the sutures and also warped it out of shape, so that no measurements could be made, but as it lay in position the effect of its shape seemed markedly brachycephalic. Skeleton No. 3.—In this same terrace I, we came upon two other skeletons at the same level (+ 28 feet). One, very small, was left unexcavated in the east wall of the terrace, but the other, also in the east wall, I laid bare. The main axis of the body was southwest and northeast as far as could be determined by the few bones preserved. The body had evidently lain on its back, with the skull propped up in such a way that it now was set squarely on its base and lower jaw fitted over the first two vertebre so as almost to suggest a dislocation of the neck. The left leg was extended straight down the main axis northeast from the skull, but the right knee was slightly bent and thrown over to the right above it. Only six of the ribs remained, all on the left side, and the top of the cranium had broken down, leaving only the side-walls and lower portions in place. (See fig. 536.) Fig. 534.—No. 5, Terrace I. Fig. 535.—No. 2, Terrace I. Fig. 536.—No. 3, Terrace I. Skeleton No. 4.—The next burial was found at a height of + 27 feet in the third terrace and more nearly approached adult size than any laid bare up to this time. The cranial sutures were not, however, closed and the ossification of the epiphyses was not complete in the limb-bones. The body lay contracted on its right side with the main axis southeast and northwest. Both arms were drawn up as if to bring the hands (which were lacking) in front of the face, the left above the right. From the top of the badly broken cranium to the end of the spine measured 23 inches and from the back of the spine across to the ends of the leg bones was but 13 inches. (See fig. 537.) Skeleton No. 6.—At a height of + 26 feet in the second terrace the workmen broke with their picks into the skull of a child’s body that on examination gave 488 SKELETONS EXCAVATED IN NORTH KURGAN. fewer evidences of deliberate burial than any others excavated up to that time. In general the trend of the body was southwest and northeast. The broken skull lay on its face, with a slight lean to the right, the knees were so bent that the lower leg bones stuck up to a height that would have brought the feet, had they been present, above the level of the top of the skull. It seems as if the body must have fallen in such soft material that it was partially buried at once and a support thus given to the feet and lower legs. Just beneath the skull was a large fragment of the bottom and side of a great earthen jar. Filling the cavity of this inverted fragment was a greenish-white slag partly fused with ashes; and under both pot- sherd and skeleton was a layer of rough charcoal and wood ashes. (See fig. 538.) Fig. 537.—No. 4, Terrace III. Fig. 538.—No. 6, Terrace II. Fig. 539.—No. 7, Terrace II. Skeleton No. 7.—In terrace 1 also, and on a level with the skeleton just described, I uncovered the first adult remains we had seen. The skeleton lay contracted on the right side, with the knees drawn up to a right angle with the main axis, which was southeast and northwest. The left arm lay extended down along the body, but the right was bent enough to bring the hand opposite the pelvis. (See fig. 539.) Although the bones were too fragile to admit removal without elaborate gluing, and this was not thought advisable, the cranium showed. the sutures well closed, and the teeth were worn flat and dull. The extreme length of the skeleton in position was 53 inches and the extreme width 15.5 inches. Skeleton No. 8.—The next burial we came upon was in terrace VIII at + 25.5 feet above the established datum. The bones were those of a young child and, though much lacking, it was possible to determine the main axis of the body as southwest and northeast. It was lying contracted on the right side with SKELETONS EXCAVATED IN NORTH KURGAN. 489 the left upper arm (all that remained of that member) stretched back of the body at an angle of about 40° with it. The right upper arm bones were absent, but the lower arm was traceable, extending along the main axis. The little that remained of the cranium showed the sutures still open and the walls of almost paper-like thinness. (See fig. 540.) Skeleton No. 11.—From the last skeleton which occurred on the south slope of the kurgan I was summoned by the workmen, who had come upon remains again in terrace II at +22.5 feet. This burial, which I have numbered 11, proved to be that of a child lying contracted on the right side, with the main axis running southwest and northeast. I found no traces of the right leg or right arm; but uJ i, 1 (Semen —— sins 4) | i — ~~ a a Fig. 540.—No. 8, Terrace VIII. Fig. 541.—No. 11, Terrace II. Fig. 542.—No. 14, Terrace II. the left knee was drawn up at a right angle, and the left upper arm ran parallel with the vertebre, the elbow being bent so that the hand lay out at the level of the pelvis. (See fig. 541.) From near the collar-bone I took out 58 small white stone beads (N.K. 114), many of them double-conical. One larger white stone bead of a cylindrical shape (N.K. 114, plate 4o, fig. 2) and 11 flat beads of red carnelian (N.K. 114, plate 40, fig. 2). To reach this skeleton it was necessary to remove a hearth-like layer of hard- burnt earth topped by a layer of ashes 1 to 3 inches thick; and after the bones had been lifted I came upon a similar hearth beneath, extending over about 4 feet square. Skeleton No. 12.—In the same terrace (11) and at the same height, 22.5 feet, but about 5 feet north of No. 11, we uncovered a child’s bones lying in the position now so familiar—contracted on the right side, with the main axis southeast and 490 SKELETONS EXCAVATED IN NORTH KURGAN. northwest. The right knee lay under the left, but doubled to a slightly sharper angle. The left arm was bent so as to bring the hand in front of the face, while the right lay along parallel with the vertebre, the fingers underneath the pelvis. From between the lower jaw and the collar-bone, I took 67 small white beads (N.K. 144, plate 4o, fig. 8), like those found with skeleton No. 11. Along the back and beyond the head, at a right angle to the main axis, were traces of air- dried bricks as in skeleton No. 5, and, as in that case, the whole lay upon a layer of charcoal and ashes. Skeleton No. 13.—Again, at the same altitude in terrace 1 I came upon traces of a child’s skeleton in the midst of the caved earth of an animal’s burrow. The cranium and many other parts of the skeleton were entirely lacking. Among the jumbled bones, however, lay 2 small white beads of stone and 3 spirally wound cylinders of lead, possibly beads (N.K. 143, plate 40, fig. 3). Beneath the body was a layer of fine white ashes, below which the earth was burnt hard and red. Fig. 543.—No. 16, East Gallery. Fig. 544.—No. 17, East Gallery. Fig. 545.—No. 18, North Digging I. Skeleton No. r4.—The next remains in terrace 1 lay also at + 22.5 feet. They were those of a child and lay contracted on the right side with the left knee drawn up slightly higher than the right. The left upper arm lay parallel with the main axis (southeast and northwest) with elbow bent so as to bring that hand on a level with the pelvis. The right arm lay extended under the body and parallel to it. As in nearly every other case, the cranium was crushed by the weight of the earth. (See fig. 542.) In connection with this burial were taken out three bits of spirally wound tubes of copper and two plain lead tubes and one flint edge. Air-dried bricks were traceable along the main axis before and behind the body, and the whole was on a layer of ashes and charcoal with fire-reddened earth beneath. Skeleton No. 16.—Skeleton 16 next came to light, far below any human remains we had yet found. It was in the east gallery off from General Komorof’s trench, at a level of 8 feet below datum on the plain. It was that of a young child and SKELETONS EXCAVATED IN NORTH KURGAN. 491 lay on its face, with the top of the cranium crushed in. The arm bones were jumbled together in an indistinguishable mass. The left leg was straight from the trunk, but the right knee was pulled up at an angle. Five vertebre were found inside the skull. With the body I found two clay spin-whorls and two flint edges. Below was a layer of ashes and bits of charred wood over fire-hardened earth. (See fig. 543.) Skeleton No. 17.—The next burial was at the same level (—8 feet) and, like the other, in the east gallery. As before, the bones were those of a child, but in this case when they were laid bare they were found to be in the familiar contracted position. The main axis was southwest and northeast. The body had lain on its left side with the right leg drawn up to a right angle with the main axis and the left leg seemingly disturbed. The arms were bent to bring the two hands, one | over the other, in front of the face, much as in the predynastic Egyptian burials. (See fig. 544.) From the earth between the knees and the lower jaw I took 6 flat bean-shaped turquoise beads and two pierced snail shells (plate 40, fig. 7). When the bones were removed a layer of ashes and charcoal over fire-hardened earth came to light. Skeleton No. 18.—The lowest of all the human remains which we found in the North Kurgan occurred in the north digging No. 1 at a depth of —11 feet. As usual the bones were immature and the position was contracted, with the main axis southwest and northeast. The body lay on its left side, with the hands in front of the face, and beneath it I found ashes and charcoal mixed with occasional bits of burnt bone. (See fig. 545.) . The results to be superficially gained from the human burials in the North Kurgan may be thus briefly summed up: Out of 18 different remains laid bare, 10 were in the contracted position, 4 indeterminable, and 4 definitely not contracted. Of these all but 3 headed to the southwest, with the trunk running northeast. These 3 varied to the southeast. Ten were found on hearths or layers of ashes topping burnt earth; 8 (6 of the lower culture and 2 above) were found with small objects in obvious connection with the burials, and 3 had not only hearths beneath them, but traces of bricks carefully placed in reference to the bodies. These facts, taken in connection with the floors and walls found by Dr. Schmidt (see Dr. Schmidt’s report), seem to prove that the children (for we came upon but two adults) were buried in or near the houses and often on hearths that were especially prepared for their reception, or, what seems more likely, the domestic hearths within or just outside the dwellings. SOUTH KURGAN. The burials which occurred in the South Kurgan were excavated in the same way we had attacked those in the north, that is, by carrying on the work around them till they were left isolated on pedestals at a convenient height, and then personally exposing them. Nine out of the 10 came to light in terrace B on the southern slope of the hill, and the altitudes at which they were found only included a range of 6.5 feet (+27 feet to + 20.5 feet). 492 SKELETONS EXCAVATED IN SOUTH KURGAN. Skeleton No. 19 (a).—The first one of the skeletons found in terrace B was at +27 feet above datum. It proved to be the remains of a child lying contracted, with the main axis running northeast and southwest. Lying on the right side, the hands were spread in front of the face and the knees drawn up to an angle of 90° with the trunk. From among the bones of the left wrist was taken one small, cylindrical, yellowish-white stone bead (S.K. 244). (See fig. 546.) Skeleton No. 21 (f).—Skeleton No. 21, a few feet south and west of No. 19, was again that of a child lying contracted on the right side. It was also at an altitude of +27 feet. The left hand was beneath the chin, with that elbow close in to the trunk. The right arm was slightly deranged. The position of the leg bones was traceable as contracted, but they were for the most part mere channels in the hard-packed earth. Below the body was a thin horizontal layer of ashes and charcoal. Add << > i\ As \ | Fig. 546.—No. 19, Terrace B. Fig. 547.—No. 23, Terrace B. Fig. 548.—No. 28, Terrace C. moe H th © \\ oy 3 ... . 1. 10 Contracted: onvleft side wesc eet elas 5 Found without burial gifts.................. 18 Indeterminable position’. -—.— sae see eee ae 5 Not: contracted. position. .7%.cccumera eis ne 5 Total « + «jiscaai= Sida ee eae ee 28 Total igi 30 235 0suene ct eve see 28 Adult remains: .92.30<5 Viste. rcicts 2.Gaelees sn et cae 460 AGES Se aati pS Gene CORSE eae Her IIR, IN Siar 414, 415 Adobe roof (with or without wood), relation of, to growth of Kurgans....... 299, ff. PAUL OSEA Peet LUtiaS OL a sept oe ece ec ociece © eke 0) see eFsie.s Str ERMC ca eres re Warinh xe me atte SNE E E RSS OV ALC Ya oe spd ive wlSe or ntge. cars ds 282 BNL Ser PeRe ss cos ie hoi sis. nails: sj 5.5, Sa re’eue siege nian’ 456 PSHM Set teciere sterrsieye. © hes wa ae wes ee 283 lil GCS (a (BOS Be aS Sona eae Ge Cee e 287 PAIateMOUMCATTIS Ae. cs series 264, 267, 275, 277, 289 Pte Guth Cath Ce Cava mieieeies lanier. chai 264 eLOsiouscyCles iN. e eale ene 287 POR VERE trie S Aele sc ieee thh aikecaien « 259, ff., 260* IA VSIG ELA DI Yel aoe steeper ete eo ahem ore = 259-264 route through, from Bactra to Kashgar 307-310, 309% 3 LOLTACCSHINGE ices tieyaparare: 6 sys Se eae Ee 263* LN LOVED TS Fy Coes OVERS SRE OSI On Ane 2EE 252 PAT GTITATITI Ts MME A PNG Perec rel eiges te oot wie © wlele ace.’ 453 PASEO ARC OCL i iterates o, ora. e fbi eashega.< Afi ten oiaue's 332 Alluvial activity as a barrier to the formation GlCUTeS Wa eh sr macro. gatoves 244 Alluvial deposition, man’s control of, at Anau. 324, ff. Alltryialsdeposition at) Merv..22. 00: ......5. B32) /)- PREVA WATE LT AGCES wer, 0 epee cain in Sitges ahs 2 tis 263 Alluvial zone in an ideal desert basin... ...247—248 Tits hemhatitnn Basin vemene: peat nota hi. 283-286 mEthewmergatia, Basin. .0 9. Anite cies en 290 in the Aralo-Caspian Basin....... 294-295, 298 Alluviation of the Kizil Su gorge in Karategin. 274 Gfethe: Zetalsnan GOLrge. smn eicis acle «st 279-279 Alluviation, glacial, see Glacial. Alluvium, wind-blown derivatives of... . 246, 283, ff. “natural sediments” and their relation to the archeology of Anau..... 3290, ff. Of; Mery A pvehe tts ote cateds eres soe. B35.16 RATA LMEE Ree 5 Fase e eke tance Senet real onthe bys 387, 388 em rettad Vit (LYRIS) rn gisiel fe hs x Oe See A 292, 295 Anau, animal remains from excavations at. . 341-442 estimation of amount of erosion on North Karrvan ate ac siverki arse caters 305 human skulls from excavations at..... 445-446 PUVSIOSTAD UY OL achtc se ea. ain cisiplsie «ae 320-330 peculiarities of, as an oasis type I (b).. . 320, ff. relative percentage of different animals bal JoWatey uu Lfvete ll Mba A oe aikc Meee 341-342 some human remains from excavations at449—-468 stone implements from the excavations 2H Ga RA tere 9 5b OES Gener Are 477-484 Skeletons excavated atau. ccacrs 6 are - 484-494 wheat and barley from North Kurgan at 471-473 Anau-li, irrigation deposits left by......... 326, ff. JMrer ue Sbsh te etenetorts oan atinascn bob aoe 325 SLOSLOMULCV.CIES) Ol ty eA rites texcteycreea viet 322 PANGHILNELUTML ane trees Wise ets eek oy stee 424 FATICIATTIANCSE Petes orrerthe ht dee ciutaeke oa eG 456 RST a ES OT eye mee Pe gee 308 PAGE. ZATIIATI OA CoRR RAE or, ne AS co Rens cys 459 PECL MEY EOE coca Fn Sc ad ai ER coe 342 PRIS UB OV PORE Re cencg 4c iss, 6 FAS AS ee 0 ke aM at 382 LALA oiry Calf pea OS, ee, a ale ar 407 PLN eR Mc Fm artic D sine Cele s Reak 431 Pbte Ol EV Dipores brelc hid the kp cia d «2s tie 359, 364, 365 NT ACLS AMEE ce Soto ise Wise" wit Ge 362 Aralo-Caspian Basin, physiography of..... 291-298 Arctic overflow of the Aralo-Caspian....... 292 DN oT dee oh cae Os, ee tae en a 427, 430 Aridity of desert basins reintensified at the beginning of each erosion cycle 248 Aridity of the Northern Pamir........... 251-253 of the Aralo-Caspian basin during the beginning of the second erosion CYCIOR Rl tate cits 1 292-293, 208 of the Aralo-Caspian basin at present, the beginning of the fourth erosion CY ClOMee ew ieni nd dors ee i=. 205; 298 Hed COS hy Fase 5.n, oh eee ne ne OEE Oe ee ee 253 See Deflation of Pamir, Aryan dialect in the Zerafshan Valley..... 265, 312 PASTE Nair (all oe ween eters re ac leve acts « 361 Asia, the continent from a dynamic point of NGA he, Sale, co RIS SIO ONCE RENEE 243 LASEMUS 1 CIVECARUS Heater enc le aie ies coe © «as 403 PASkha badane nie weiner re een epee. CAS cy'sr3 302 BEROSIAC WEA Bite Cas? ns eae ss «7s 249 ASS ope Ome Oe ren 403, 407, 408, 409, 413, 414, 415 SS VUIAN emit paetop ihe edt ens Seeebisisis snd 306.» 351, 361 ALIStPAlIANS ans. as ey A See 459, 461 Auvernier...... 353, 397, 403, 404, 409, 410, 417, 426 Babyloniaere saeco see o- 351, 360, 361, 440 Bactria, Alai Valley route from, to China. .307—310 Baran AL ery ete cee Mec se eat eneevey merc erence a2 a 333 Baplvarshy. peers tate isns tote Gani eteia altel Sisters 52 453, 456 SEN ab ety ai cr GEREN elect POO oe ae a 302 Bal Kuwi, physiographic shaft at......... 327-328 Bal kari: Gitte een wale ence a. - 5 pi 296-297 Balyakandosewmrever aeease ccs aa sa cce ss B07, Barley and wheat, casts of, in potsherds TLOMMAT AU) lew wa fo eo 471-473 HBT 8 0b wh ye DAL OCLC: Oe CATEGORIE AR nn eae 345 Birch tyellow semen cette oo + Seaiacns.s 273 Blanrvilles (Gems ie aresdanc Sass ss x = 404 BlaAStitseperper sewers cre etre rece aiac soe ois y's) Ges 370 Bohemiaameetrervr a tise a vletors tei suse fanuscie soe 345 BOKNaTA atte Miter teasers aes fecha dee sisignere 265, 276 at the mercy of Samarkand........... 319 TOULE LOM mtOLEISSAT anemia eae reyes ee 310 BO Speeeree ites Aner eas et as ore a, rae 5 2s 341-342 namadicus Falconer & Cautley..... 359, ff» 435 promigenius Bojanus................. 359, ff. WMUDGLOCEYOS MUeTStcns ci. Cree 383-384 Camelusicvon Gaia cae Cee eee ae ee 342-383 CIUENISES oe arom rat per tena ra ae 384 DACELANUS.. eionir teases oi ce. o, teste tee 384 Rnobloche. s aerstecn eo ete coe ieee tee tel 384 Canidae: 5 AR ae cae ee ee aces 345, 354 Vulpes montana Pearson.............. 345 CONT Se Soh ar eae a aero Ne ree 341, 342, 345-354 familiaris matris optime Jeitteles. . 348, ff. 430, 435, 438, 439, 441 hodo phylaxis: sea eee ee 346 inostranzewt...... 247" 348, 1.353. ao lupus Tignes on eee ee 346-347 MOLOSSUS: Beas lets she te ee le 350 pallipes Sykes........... 346,347,352, fe PALUSULUSS ieee erica oe sted ee 350 poutratint . . 347, 348, 349, 350, 353, ff, 435 Cuon alpinus Pallas. 220; 3)... te-ee 347 Camhing-town F20% te ieee ee ee 374 COPIA) Sore eee cate ae eee ee ee 342 hircus riitumeyert Duerst......... 378, 380-381 CPE PRE eh iris CE Cs Se oe ye 380 CYELENSUS eo pe ane eee 284-286 recent oscillations of the alluvial in Fergana 35.202 220. ee, See 290 INDEX. III PAGE. Deposition zones—Continued. the lacustrian in Aralo-Caspia........ 292-294 recent developments in the alluvial and flying sands of Aralo-Caspia. . 294, 295 i qi geeb aha dela. indie aot iminh Ot one Ais 303-307 the building of a tilting delta on....... aon Hides the; cOmroloL Mai wate icss. .- 324-326 interbedding of the alluvial and flying SAUCS iA teATIA Ue ete a aera the Murg-ab delta and man’s control of alluviation..... 330, 331, » 332, 33301 Deposits from ice, water, and wind. . 244-247 See Deposition zones, Oiiiary (Cultire;Gebiis)ice sure ee cee « 299, ff. WNgALOn Mepesits ne 4s eee se 324-327 of man, water, and wind at Anau..... 328-329 of man, water, and wind at Merv...... 8345 He of the Zerafshan River (glacial origin of LOGSS) Peart, LN MER ettiee aie are 281-282 of ancient dune-sand interbedded with the Anau delta clays........ 327-328 in the Aralo-Caspian Basin, variations ity aIMOUnt Ola se Ce ae 292 “1 Sp eae a ot Pliocene ASia sania as oe 243 of Gobi (uptilted borders of).......... 243 the Northern Pamir an extreme....... 251 GhGreatykK ara Wile cements aerate: 251-252 Gehlation ol Kara kolirs. cen cca eee 253 preservation of ancient topography on PATI tee ete eos Pte cree es 254 valley of Eastern Karategin........... 266 tz WaT en gave Matters! ences Lae 271 sDaAtatiey pees Cee A es sitar. ee 282, ff. transgression of the Kara Kum....... 293-294 Deserts, paystography: Ofte. a) sha len ee ees 243-298 Desert basin as an organic whole......... 243-250 essential ‘character Of axis. ican. ses 243, ff. three agencies of erosion and five deposi- AIOMIZONES 300 Be rayer aceite 244-246 interlapping of deposition zones ina. 247 the cyclical development of an ideal. 247-2 50 Desiccation (see Climatic) of Tarim basin. 286 PIE ENZATIO okie aiches sii cis asi pie sv) oe ee 418, 427, 430 DAES ere Ree tere: ospoee sae a a 352 Displacement, orogenic evolutions in the CeSert DASUUKt wire sets ccletere ete es 249 Distributaries, unstability of the Murg-ab, and their effect upon man.. . 331333 BISIEL ESE A TRL ola an tate a APR oe as 3% 390, 391 Dillan-oott Darya sii ci rantdnatitete vis eo '0s 318-320 Dog ancient Meyptian 1... 26a eed /. 0k. 352 Assyrian and Babylonian............. 352 PATEST TERNS ATE SE Pe Seal he. 348,. i). brought from Inner Asia to Europe. . . 354, 430 GANS O are So tarp ts 5 iy Se eee 348, ff., 435 OLPA MAIL D sx nee ser cheeses ncratn spelen Hee 348-354 TEIACTONSIN) Olirerter-ratetedera)a\ tar eta 350 tables of comparison with dogs of Eurasia and Africa...... 348, 349, 353 SLOSH ES Ak ee oe erie do wee yun wate adds a53 of Bohemia, pariah.....0..02.% 3. 349, ff. shepherd got ace tase pkeees eae gee OR 349, ff. CC ALY 2 Face aee nhc a ee ee Role ee 39 Domestic animals, census of, in Turkestan IME TQOF srs Hels Oe ee eae 433 relative percentage of, in Cultures I and Liat tAna ars aetna tars 438 POMARIVEL catenin tetris) herka ehut alate levee ta 292, 297 SOLID Olan eh tee Alore rc rey Nix ea e's See. 8 388 DON US =tatlin tere ee rete erate a meePohe fe qcie ce ester: 288 DD ORGO QING esctate toy te lairy oe o/s oe M lnk re oar eles 403 PAGE Drainage, recent decrease in surface....... 322-324 EMCeUSE POLS Py Liat ma ietatels Cay 3 dee toe ee 449 description of animal remains from exca- WationsiateA nat. syeeteee ens 341-442 SES A ue aoe Mm iasteers Los eens 405, 406, 408 Dunes, dune-sand, etc., see Sands. LuShetraibe lives sete «7 ee Cree 2 ee 270, 459 Dust.cyelonie-stormis ‘of; oa) oe Oe ad ks 244 precepitated as loess...) 20a Pook. Cane 244 Coe seteling cround of: 2.2.65 Sse eee 243 absence of dust deposits on Pamir..... 253 RMA INTEC ia ey in Ae OA RAN 318 |G Ae8) 5 AS go ae a 302, 334, 440 RACHEL eee en eh ei oso, ee ok 416 SE PRINS oS 2 ans oS obs ON 9 hak rw 295, 296 Bees OREM Pere Yoana Perot oe ah ane e 273 Equus OSTTUUS rae et Mees eS 403, 416, 417, 424 CUOGMUSM Rai go's oe as dnd Aah +s 390, 395 PRGIDOt PONV 22 oes i eine os 416, 421, 422 caballus domesticus «oo. i ie cece cus 360 COOGTTES JOSST GA ee Dyin Oe a oes 402 caballus fossilis Riitimeyer, the ances- tor of the oriental and occiden- Leh CES aeet ee wertorirste & /e-spércte > 431 caballus fossilis robustus seu germani- GIES IN ELIIT Sion ncaa te, 42 Af chaser a 28 CADGHUS PEFMANICUS® 2% va. ween ss 398 caballus germanicus seu robustus...... 399 COUGH AES IGUAL S ile] tlehalo Se os cle «26 « 384 caballus nehringt.......-..+..: 398, 399, 420 thetoGrestitiyoe arr cderom cz. vere: 430, 431 caballus pumpellit..... 397, 398, 412, 413, 417 418, 419, 422, 423, 424, 426, 429, 430 MGyEY Pe Of TNE COSELT xe 6c a. win es 431 caballus robustus Nehring, the type of CHEISHEN PES ay mk mie eos Asie e's 431 OGRA Ti incur he OS Oe Cee OE 398 hemtonus . . . .387, 388, 389, 392, 393, 395, 396 hemionus (var. kiang)......... 387, 390, 391 VU LTONLCUS eee Pe eA 3h OE oy dis 402 eNO Geen Git Gd Oe Oper RR 428 ONDZEY Me ees 387, 388, 390, 391, 394, 396 YT TNS SE Bons Cech) ge Canes ERO ee DLPo ae 401 PUSCUNENSES Danan tet en ates Sante st ann Maks 428 PILGULCH SON ELE port ast ern alae oie 28 PU ZC AIS RUD date ecae oe 387, 388, 390-398, 401 411-415, 419-426, 428, 429, 431 GUAT LOCI Sen oy ot eisieploteey olan aheie eine 402, 428 quaggovdes affinis Woldrich.......... 428 POOUSTUSER Nec Mone A tact as Gunes wer «oo 401, 417 SUVOMIO WSIS Loe etc se Rett Es ee 428 SDELBUS OR Sto ets sie Hale eisbs eM gare Abn 402, 428 SLENONUS] oes. «as hoi 395, 390, 402; 417,418 420, 421, 422, 423, 428 stenonis affinis W slarich aces. cts orate 428 PERSO SS eos ie sctnha Lae ONG EO tes 416 yi Le ae eo tiim rig elon lores Qos an, ees 334 Erosion, the agencies of. .............0.4. 244 manasa geologic factor in............ 299 in the obliteration of kurgans........ 303-307 Erosion cycles of an ideal desert basin..... 247-250 ChHemeAamiicn were shits aweaelels Bae ev ctytens 259 PHOEA ATEN AG Vemcaaratkattr res cohen > sakes 264 general discussion of the duration and Memmi Ue Of A cc.0s heats as sae 267-268 Katateoin andy bissatine. occ. em ctae a « 274 the Zeraisnan Valleyeceme: dee ates 277-282 AWa\auarit, Ac tteo.o ccecto tee Gua boae ee eICnCR to he 286 EerA TateMOuntai 1S nese a iaeente torres ets 287 the Begonia Basins. ota: 2 wh an Sea sree 290 1V INDEX. PAGE. Erosion cycles—Continued. the Aralo-Caspian) Basin’ ao ane 298 the Anau. Sa... 2s ewan ee eee te ee 322 growth and oscillations of the Anau delta during the various nic.n 1s = 330 HskimoS acs: scbik bre cro crenata 459, 460, 461 Euphrates, cick eicAeed Boa pee eee 437 Evolution of a continent.,..sstme trent 243 desert: basin sui akknge ht 22k eerie 247-250 EX 00f; POLY 25.6 en wae ern eee a 414 Faiza bad - ic smpeeteeie omen ceeley Orns Raters 272 Fault-scarp in the loess plain of Hissar.270, 272*, 274 Fergana, abandoned oases of............. 317-320 Fergana Basin, physiography of.......... 287-290 Fischer, Hos fin sss sWrw ck 2 ee ee 452 Kish traps of the Kazil Sus. eee eee 271% Flower, Wi. Harkins hess eet area 459 Horsyth Ma jore|ia Corwen 386, 394, 402, 405 Fowyents Aut cok ke cee ielee ee eae 403, 419, 429 Fox fromsAnauts so

=e sone ee ke 312 Galchas. . . 2097, 2707, 314", 315%, 316* Galcha Beg with his hunting eagle’ ei g cee 266* Garm 2's. sadn ale coe ee Se ress 269 Gaudry,; Alber tiie sa setae ats cena ene 433 Gazella subgutturosa Gueldenstedt......... 382, 435 Genghis Khatsi.sia sls 2s yassceeee eae 332 Gera .octs diohen ley oottene Ae mPa ee x eee na 414 Gervais) Patil aos 05 ce Soe. curso eres 260 during the glacial period.............. 298 brackish water, and wire-,............ 252 areas shrinking in Fergana............ 290 Gtisonsis. ainsi Saree ere piste eee 415 Gross Czernosek........ 367, 381, 411, 419, 425, 426 Grotte dw Pontil 4... .2-5 7a. eee ee 375 Gruim-Grshimatloz .. Se. 2. ape eee eee 402, 429 Guanchass 2707.25: ee cee orotate 456 Gulbrandsdals iki cs oe: ote ee 415 Gulcha foie 20s ip ae ce ace ee oa 289 Gulcha River :o2 5. 0.2204 ebee ae eee 289 Gyaours River vanishing: ....°...c.. eee B28 Ha wee. cide som « Sng Sn eae ee ene ee 269 Hallstadts..¢.. S252. seas decease green ee 410 Hensel Or ..0 5 250s op ep bw eter ee ee 421 Hepbtirn 24 05..t ye cea ee eiceire eee 451 Heschelermocncci bree 392, 393, 394, 416, 426 Hickory s00. vv 5 ee etek eects ae 273 High valley oases... .... 4.60 sos oe 301, ffs of the Zeraishan suc. acc sme eee 311-313 Hindu-Kisie ovis eats ere einen tere 251 Hep partots. a. tdci wil aera 395, 397, 422, 424 Hissar, ancient route frome. aac. cre eres 310 Gitadel Of 024 Ue ee a 316, ff., 320* {NOUNEAITIS ©...) toe eee ee 272, 2746 Hissar Valley, preservation ofkurgansinthe. 306 physiograpity, Ola, whe as eee 270-274 Hodja Mussaii Ashar}... i050>503 <1: eee 314 Hommel F 5 ige he ee eee 351 Homo Primigentus ne oe ses odaie sg cine 454 Hordeum distichum, chaff of, in pottery of J Gat- 5 la! Arena crareten RA inspira nr 472 Horse of Anau, Equus caballus pumpellir Diterst’s Si, skaters cee arene 387-397 tables of measurements comparing with other subfossil equids, pre- molars and molars... 387 dental indices. 5 2 heat 388 JAWS F506 fo 2 Fees eo ee 390 indices of metacarpus tertius.......... 391 Scapiilacsites clots cage: tetas trans 391 humerus, radius, and metacarpi....... 392 fenivr aid tibia h ata sa, yenlace oo 393 metatarsi and phalanx Iand II........ 394 phalasux IT], astragn lis. 2 ooh cas anee 395 type and-relations of cic. caucuses ee 396 meta tarsus tertulus a sacs es aware ea ec 396 height of withers sian sheers tee 396, 397 characteristics of Anau horse summar- TZOO a teh ee ete teens 397 Horse of the Quaternary period of Europe.. 402 prehistoric, Giese. vive es = heen cee 1 4038 the neolithieiager tween. ee aneriate 403 INDEX. Vi PAGE, PAGE. Horse—Continued. Horses, fossil, subfossil, and recent—Continued. tHe COPPEMiaAl Cam a wtertaccusee culersiancee ie 410 LB TE hel sinners o os cin ean ten ete eam 412 La Dene times) signee < ces sccsiens dng Sisal 410 European diluvial....... 391, 394, 395, 416 Halistadtitime ss. beans eee racket: 410 Exim oonmponyienua sha te ain 414 PROMIAN 411 Ret AGORA 2) tien d lad cuttns Heaneem 412, 415 the skull and its proportions.......... 414 GrossiCzerlosekie rer 4II, 419, 425, 426 PORES CV PO naa cs afeissl sho letele: aig oeisteS 399, 431 SHAG saal .,o5...00 ecaemaceane we 415 GESERE TVD Cor meukes iaicl> sarees FinpeLe Gates 399, 431 EHelveto-Gallic) (iron) 74... .6- «6 410 SLEP Pe Ly DE rs aeipesic i ceetewusye tir ees ves chet 399, 431 RIOIST CR 42 bh oi 0b ss oc eae Mak ero es 412 STAC OMIA cen uesars ore. spas eid tbc. omaveney asa 434 Hostomitz (iron, La Téne)...390, 411, 412 thew egtiatvs Ol MH UrOp es dancin ess te bere 28 414, 415, 417, 418, 419, 420, 421, 425 only one type of, on Eurasiatic continent PEORIA TN Soper sb ye ipiene iene ewinrarshe wis: a. 415 in Pliocene and Pleistocene ig ciiaitipa cere it rere. Scrat RA Be co: 390, 412 time, of which Equus przewal- Kal nny Keer eta a teen ent 394 skit is last wild representative.. 428 PB UUGEIIT ie coach Goa tm alos 415 regional racial differentiation of, conse- Kesslerloch (diluvial)....391, 392, 393, 394 quent upon regional changes in 395, 412, 413, 417, 421, 423, 425, 429 climate and vegetation in post- K6nigsfelden (alemannic)411, 412, 414, 417 PIACIA, EINES 4's) ao as ie aes os 429 PSOE DER rs. occ cee aiaiecieiee «| wt 418 the horse of the bronze age and early iron Kutterschitz (neolithic).414, 415, 416, 417 age of Europe was the Anau JE ATIO IAG CS Gre MeN cael, (tees yannoie a hae 425 horse, Equus caballus pumpellit. 430 GAs enes(iron) tanner b 301, 392; 3930415 this latter used in ennobling Roman 418, 425, 426, 428 IOLSES ee ee eg er cea es 430 Leitmeritz (neolithic) . . .391, 392, 419, 429 Horses, oriental group (broad-fronted)...... 386 Benaskiimersrasmciey ewe os scton « 391, 393, 394 occidental group (narrow-fronted)..... 386 ie bs alisenanertacrrcie avers cess vis as A4It cause of characteristic differences be- TGV OUIILZ skeet tate ckexcyerenera) 6 aut ccd Shred 411, 426 tween the oriental and occi- WG ttrd eSerwav enters crates eis eMail 392 dentalforoupseriea ce eee 2 or 398-399 Lottwerné (neolithic), .’..4....... 392, 419 fossil and subfossil, and characteristics.. 402 ISITESCLICE Za eyaed el siacheh ys Stas sty eh as 415 broad and narrow-fronted............. 416 Djachown Island eetieer ctec occ 's 2s 394 Cuilinyial or CrOrmaan ye mye ecto tts sh. ce 416 IMadelainesmemar ya rc. cottons +. 392 of Solutré and Kesslerloch...........- 416 Moehringenieeysa rei: crs oolaels i's 410 index of basilar and palatalregionof.... 416 IMoossecdOrim wet ee cuscn tc. nsle sic ok 415 comparative table of cranial dimensions INussdort (dilitvialjion. ssc. ce «54 5 414 of, in Equus caballus pumpellir Occidentalieeris sciences: 401, 412, 430 and other subfossil horses..... 417 ot Arabs and Berbers... .. -). +. ...399, 431 comparative tables and discussions of the Oriental eee dtm ¢ 401, 412, 417, 430 bones of the extremities. .... 417-426 Petersinsel tan pace stevie w lated eke a 425 calculated sizes of diluvial and prehistoric 426 PinZoalion, sos. eee sn « 412, 415, 417, 418 INCL WSI OLS: seep ey Os imaa shears) Secheueesy es og 427-433 Quedlinburg (diluvial)............ 391 tabulation of points of agreement as basis Remagen (diluvial)......... A412, 413, 414 for determining relationship be- 416, 425, 426 tween the various diluvial and RixdOrt te seen ss ake oe oe ats 6 419 DLEUIS COT Comte ries plete peea i 427 DAIL A Vek sete es cers 411 fossil, subfossil, and recent: Schlossberg (iron).......391, 392, 393, 418 PMIDASGI ES en clerstctier sath pete ena bases) oes 414 419, 425, 426, 427 PASC ENN aie Seen ee ee 416 Inscriptions; Arm bic. is oe ac 2. + «ee eee 316 LOUL: 6.5 ah pheasants oe eee Oe or eee 512 lori Kargay ty is cco ee ae eee BIA, ary Irkeshtand oct dace oe eee 308 Tron, Door:2'¢). 5 cas ordi auene se 308 Irrigation terraces, abandoned..........326*, 327* Irrigation sediments, irrigation terraces, origin and descriptionvola 2 eee e 325, 326 tabulated with deposits from man, water, and wind ares aes 28 how they are recognized. ... soas...5-- Bo interbedded with ‘‘canal sediments”. . . 336 Iskander (Alexander)... cesses cs scene ae srr Iskandeér, Kala ences 05.0 ae te ee nae Japanese a ae ey eee 455, 459, 460, 461 Jaxartes Rivets wine. acces eee eee ee 310 Jettteles cncigt 2) sa Oia i ee ee ee 350, 353 Ke Grnigan 2% 5 ied hee, g ui Skee 270, 274 Rak Passi. casi Sued teins cee eee eee a Kanaka 2 20 or te nee nao akc ee 270 Kara BUgassosiy oes oe enh oe ae ee 245 Karakorum oe.ec oe i ere te ee 249, 251 Karta: Kul sete men cesarean 251-259, 251* Kara; Kin deserts.) eee eee ee 398 Scie Ahearn AL Oe Hees 297, 324, 330, 435 Kara Kum, ancient and modern dunes of.. . 327, ff. transgressed by the inland sea........ Karatagh:.< eee ee oe ee Karategin and Hissar, physiography of... . 265-274 Kata Tepe (Westera) .: las ie ee oa ole MIUSUIMON ON aa here easy Roce Aa ee 381 CDRA an eae are Pek en, cee ee sar orcentales, Gimelitien «+. aaaet soe 370-371 DIGIVUTE @LYPUNE. vow vioes oat cee B77 PUALY UT UUCHOVUCH Wye asa is eee sae 376-377 POE Greve te ees ayer ke bea Sy Mele tec 252 Oa 8 2) 4 8 Rane RIN ark oc FS) cae EUR igdl fic vignet arkal Lydekker....... 8701372) 3751435 tables of measurements and com- PATISONG cc hades dae eer Oe 371-372 GN EVES CVELICN OR Le © Soccaie vin as 3 Cpe ae ae 370 ammotragus tragelaphus............... B73 (OWyet RE eaciieten es Gs alntiectess 386, 401, 404, 405, 409 CORACUICH Cee ee OU Ce eee ate ithe: coher Soaeeet ca? 441 Ox, domestic, of Anau I and II...........364-370 tables comparing long-horned, of I, and short-horned, of II, with re- mains from Europe and with Egyptian long and __— short- HOrNEedyA PISs kisses sialon eke 366-369 Vill PAGE. Ox, wild (Asiatic urus) of Anaul......... 359-364 table of the measurements of different bovids.. goku ceme eReer ora 359 table of dimensions comparing with other Eurasiatic finds........ 362 in Babylonia and ancient Persia....... 360 Sumerian Character10r.. eee ate 360 was intermediate between Bos primi- genius and B. namadicus...... 364 Ox, remains of the wild, of Anau I, studied CONIPALAtIVely wes ae eee ee 359-364 long-horned domestic breed established Bt ADA nccc ich: eed ene 369 identity of this with Egyptian Apis and Babylonian long-horns....... 369 short-horned domestic breed of, in Ama ls xeric te ete re hee 369 census of cattle in Turkestan in 1903... 433 ORTIS TIVES science bon eee itereerneny atte see ee 265, 272 (Amu Darya), recent changes in the COUtSecOle neg coe eee 295-297 Packshittte ceosn gee cree nee meee tan 279 Patkent oats ate eas ent cern eae ee 276 ancient route LO Missate va-c se ee ee 310 Paladrtt- es oc oat oe See es B75 Paleotherum miedwimas,. ne oe ea 422 Paleothervuamen nus rie cm ee ee eee 424 PARI sie yok, cinta ois tena ey echs ee eee 277 physiograpliy oOlzi-.k v6. sem eet 251-259 PATSECS ac dhe tee kere tere ot gee Oh ee eee eg eee 442 Patavonians: <;..0kie. oan tie oe 460, 461 Pavilot, “Mime oot: deiner tne aa eee 405 Parson yee ciamarehe rcs oe a ee oie eee 345 Peking: 5.7. srg cesntoh tans Paap ee see mee 347 Peneplain iia oct deci ee ee ee SON OS Petitvians i.) set eee a Ce eet 461 Pétersinsel ea. deta eee ee ene 397, 425 Peter-the-Great Mountains................ 266 Pigs Ana Ieand: Dagens cept eee eee B55 identity of, with Torfschwein of pre- historic’ Hurope.eshce aaa 355 CUGDALYs. Ol SCUIOSSDELE. peer ee ee eee 355 a Renan eek eee et ene 355 neolithic Swiss lake-dwellings.......... 355 Hirrasiahie wild boarl.3:tssess eee 357 ot Angit. ae’ Ae set eee eee eee 355-358 Cur bary Raita «eee ee ke See 355, ff: MaOris 2.27 nic dnc boc ae aes 356 wild, of lake-dwellings................ 356 Pildona es Jnana eee peak ee Ee eee 269 PINZVAU - 6 nhs ase a CRE ee 412, 415, 417, 418 Pliocene Aralo-Caspian Sea......:....... 293, 294 Plomawild 225 <2 oa Oteren ee ee 273 Poephagus sruintens. ane eee ee 363 Polyiesians. Aes wee cee ge ee 459, 401 Pomel. 3. be ene ont eee 384 Poplar s\n. wuss, boo eee cca. tee ee en eee 273 Pottery, chaff of wheat and barley mixed With Clay.Gl..y c.0- ase 472 Precipitation, recent decrease in.......... 322-32 oscillations of, in an ideal desert basin. .247, 249 on the: Pamir ..,.20 5 con raeae. 34 0a SU at See Climatic. Prosperity Of Ogse9.... 41.9 se eee ee ee gar Przewalsle ere astra ate eer eee 383, 398 Pumpelly eee 256, 286, 291, 304, 305, 341, 345 351, 360, 375, 437, 440, 449, 478 Pumpelly, R. W., on physiography of Cen- tral-Asian deserts and oases. . 243-337 INDEX. PAGE Quaternary Aralo-Caspian Sea............. 293, ff. Quedlin burg 52 ota aor aa eee 391 Rainfall, recent decrease in..........+-6> 322-324 See Precipitation; Climatic. Ray Lancastets Fe iiss suic.s elt Apa eee 403 Reinach 904) juceia s+. 20m, ee eee 441 Remagen . . 397, 402, 412, 413, 414, 415, 425, 426, 428 Rivers of the Aralo-Caspian Basin......... 291 River-bank and flood-plain oases.......... BONG he of the lower Zerafshan.......... 310-311, 310" Rixdorf (dilrvial Torse) «0: «sons meee 419 Robern hase focuses gid eke a ee ee 367 ROBINSON aang, ce 4 aries oes ee ee 459 ROCVALs cre ncrartin eeuieick isis 5/0 ee 317 Rollestone. oatso., tia tateascec eae eee 355 ROMEttE Sekine ths etaie side Se oe ee 273 Roofs, thatched . rac.0- cn os eae ereeeee 273,075" adobe domes; flat adobe 0. eee 299, ff. Rosellint irish. acct utre se eae 352 ROE Y Chee sewas 342, 355, 360, 373, 374, 381, 386 403, 404, 405, 409, 410, 415, 421, 422, 423 Salenski, W.. . .407, 409, 411, 412, 414, 415, 428, 429 SAMarkand , si... chit = evel eee ee oa 276 surface horizon of antiquity preserved from burial by warping up of the steppe.g)-. ate eee 305-306 OASIS: Of fais onset e Sic ee eee 311 upwarped loess steppe of........ 281-282, 283* ancient route from, to’ Hissar,..... 20. 310 Sands (flying sands, dune-sands) : how derived by wind-work over alluvial Plaiws 0.5.42 aoe forming a great nucleus of dunes sur- rounded by the other deposi- tion! ZONES... ciety ee in the Alai Valley. 3. owues co ee 260-261 residual from the deflation of old alluvial deposits in Tarimt,. atc son sce 2 burying the ancient cities of Tarim..... invading the oases of Fergana......... ancient dunes buried below the present shore of the Caspian Sea...... origin of flying sands in the Kara Kum. . ancient and modern dunes of the Kara Kum and interbedding of the delta margins with them..... 327-328 lying second in the Merv series (loess, dune-sand, alluvium)........ 336-337 Sauson, A. «nin. sobs eee 401, 415, 416 SAALENAY sei ss Ves Pew ee cee ee ee Sardai-miona gorge, physiography of...... 272-274 Sargon of Accad bar Kamish , Sivcy ee aera ee 293, 296, 297 DCHABi Ss oh ciosus 5 soci ts aaa oe eee 346, 274 376 schellenberg, Prof, Fi. Creer anaes con ee 437 on wheat and barley from Anau I..... 471-473 von Schiagenweit, :S ® a y* é hes “a ' aa = i ‘ - 7 x INDEX. PAGE. d Vaksh Valley. .-... 03 ..0ss sess eee ence S65 1) Walekens, Bi Gw.ais,.0 06 -esin 386, 388, 397, a river, captured by the Kizil Su....... 272,274 | Wills, Bailey oo... 2.4 2.28 see ee ; Vain béry on ccnicouls ties sons ame atten oer BST yl WAM OWs civ Gorcdae sell eeees eon aoe nen ee Vanshan Darya iiie.cs ona eins ane 375 “| Wohontsch? soe -0..tetatepet ee ee 392, 41 Vindonissa..... 356, 394, 410, 417-421, 425-427, 430 | Woldrich) J. No... 0..58. 50s 40 Volga ‘river icadcachosals cae ee ee a 207, 402 + Wolf, trom Anat: (Po. 15.0 eee ets See, Volkov, Ye ..6 pace ee 459, 460, 461 table of dimensions of................ . Woodworth, J. B., fracture Sens. on the — Wadi-Halfa oii. clsn ocean eee eae 353 grander Scale. (iin Gee eee : Wrialthamstowieeaetr etter an nerene 362, 368, 374; 379 | Wirtemberg....:.«... 9:5. .os seer eee Walther, To.cccoe.varaseiteue eater eek cane ene 293 Warner, L., on stone implements from the Wana River Video hie cei 387, 393, 396, 507426 excavations at Anau........ 477-484 | Yarkand Darya... .........eceeeet ee eeee : on skeletons excavated at Anau...... 484-494] Yarkitely Glaciersé .< 420. .xue - ee on work done in Terrace II, North Yellow days caused by dust... ...705.-5 ae Kurcanse een era eee 482 Warped loess ce of Samarkand......... 283% | Zend-A neste. ooiaca bee sagen eee «eee } Warping of Karategin (274200 85-10 eer ee 269 .| Zemishan Glacter 2... ajnc