/\!0,\a)3'.'2>O PORTO RICO AGRICULTURAL EXPERIMENT STATION MAYAGUEZ, PORTO RICO - Under the supervision of the UNITED STATES DEPARTMENT OF AGRICULTURE BULLETIN No. 30 COFFEE VARIETIES IN PORTO RICO BY T. B. McCLELLAND, Horticulturist Issued June, 1924 ulf DEPOSITORY tfN VS»Y — ~i •r) WASHINGTON GOVERNMENT PRINTING OFFICE 1924 PORTO RICO AGRICULTURAL EXPERIMENT STATION. tUnder the supervision of the Office cf Experiment Stations,, United States Department of Agriculture. ] E. W. Allen, Chief, Office of Experiment Stations. Walteb H. Evans, Chief, Division of Insular Stations, Office of Experiment Stations. STATION STAFF. D. W. May, Agronomist in Charge. T. B. McClelland, Horticulturist. W. V. Towee, Entomologist. T. Beeggee, Plant Breeder. H. C. Heneicksen, Specialist in Farm Management. J. O. Caeeeeo, Assistant Chemist. W. P. Snydee, Assistant in Plant Breeding. J. A. Saldana, Assistant Horticulturist. C Alemae, Jr., Clerk. PORTO RICO AGRICULTURAL EXPERIMENT STATION MAYAGUEZ, P. R. Under the supervision of the UNITED STATES DEPARTMENT OF AGRICULTURE BULLETIN No. 30 Washington, D. C. June, 1924 COFFEE VARIETIES IN PORTO RICO. By T. B. McClelland, Horticulturist. CONTENTS. Page. Kinds of coffees 1 Arabian group 2 Liberian group 15 Page. Kinds of coffees — Continued. Robustoid group 20 Summary L'5 During the 20-year period 1901 to 1920, the three principal articles of export from Porto Rico were, in the order named, sugar, coffee, and cigars. While sugar held first place each year, coffee held second place for 12 of the 20 years and for the period as a whole. The average annual exports aggregated 35,000,000 pounds, with a maximum annual exportation of over 51.000,000 pounds. In 1920, 32.776.754 pounds of coffee, valued at more than $9,000,000, were exported. This valuation is the highest since the American occupa- tion. The coffee crop thus holds a very prominent place in the agriculture of Porto Rico, is one for which another crop can not readily be substituted, and is peculiarly adapted to the mountainous and often rather inaccessible regions in which it is grown. The Porto Rico Experiment Station, since its establishment, has for these reasons been doing everything possible to develop the industry by promoting better agricultural methods among local growers, and introducing and testing numerous varieties and species not formerly grown on the island to ascertain which are best suited to local conditions. This bulletin gives the results of some of the tests in order that growers contemplating planting new varieties may know with what success thev have been grown at the station. 1 KINDS OF COFFEES. The coffees of commercial importance fall into three groups, the Arabian, the Liberian, and the robustoid. To the Arabian group belong the coffees known as Porto Riean, Padang, Bourbon, Pointed Bourbon, Erecta, Columnaris, Maragogipe, San Ramon, 1 Varieties and species having little or no economic value on the island are not discussed in the bulletin. 83926— 24 1 1 2 BULLETIN 30, PORTO RICO EXPERIMENT STATION. Mocha, and Murta ; to the Liberian group belong Liberica, Excelsa, Dewevrei, Abeokutse, and Dybowskii: and to the robustoid group belong Robusta, Canephora, Laurentii, Quillou, Coffea Uganda hybrid, and C. congensis hybrids. ARABIAN GROUP. The Arabian is the largest' of the three groups and includes the greater part of the coffees of South and Central America, the West Indies, that known as Java, 2 and in short what the public generally knows as " coffee " where no qualifying adjective is ap- plied. In this group are found the best cup coffees. The leaves of the Arabian group are smaller and thinner than those of the Liberian group, and the trees 3 are smaller than those of either the Liberian or the robustoid group. The pulp is much thinner than that of many members of the Liberian group, and the ripe fruit drops more readily, necessitating picking at shorter intervals. Porto Rican Coffee. According to historians, coffee of Arabian stock was introduced from the Far East into Java and other places, whence it spread to Martinique toward the close of the first quarter of the eighteenth century. Within the next quarter of the century it found its way to the other islands of the West Indies, and was taken by emigrants from Haiti to Porto Rico, where it is said to have produced a crop of some 700,000 pounds in 1770.* A century later the annual ex- portation from Porto Rico ranged from 15,000,000 to 25,000,000 pounds, and by 1896 had reached the maximum, 57,961,291 pounds. with the second highest exportation, 51,125,620 pounds, in 1915. According to the Federal census, 159,860 acres of coffee were harvested on the island in 1919. 5 Porto Rican coffee is botanicaily known as Coffea ardbica. The shrub is a small evergreen which attains a height of 10 to 15 feet. When young it has a single, very erect trunk or stem which later bends under the weight of its crop. Other stems of similar form and habit then develop along it, giving to the whole a somewhat bushy appearance. The lateral branches are opposite in arrange- ment, horizontal, and in pairs. Occasionally they occur in whorls of three. The leaves, which are dark, glossy green on the upper surface and light green underneath, opposite, and in pairs, are generally 1| to ^2 inches broad by 4 to 7 inches long, the length being 2-| to 3 times the breadth. They are elliptical in form, with acuminate tip and attenuate base. (PI. I, fig. 1.) The flowers are pure white with five or six petals, and are borne on short pedicels in one to four axillary clusters of one to four flowers each. The blossoming season is irregular, and flowering may take place at any time from December to May. At Mayaguez the principal blossoming occurs usually in February and March, and sometimes in January and April, the season being earlier on 3 At present the Arabian coffee forms a very small percentage of the Javan crop. 3 The term " tree" is arbitrarily used in this bulletin to cover all forms of Coffea spp. * Informe de la cafe. National Coffee Growers' Association, Ponce, P. R., 1910, p. 128. 8 Ann. Rpt. Governor of Porto Rico. 1920, p. 319. COFFEE VARIETIES IX PORTO RICO. 3 the coast than in the hills. Generally there are two or three major blossomings, followed by several minor ones. Upon fading the flowers are replaced by small berries, which when ripe resemble carmine cherries, hence the term " coffee-cherry." The fruit requires from six and one-half to seven months to mature. At Mayaguez the main crop is gathered during September, October, and November at intervals usually of about two weeks. Table 1 gives the ripening season for Porto Rican coffee, as shown by the harvest over a period of seven years from experimental plats located 7 miles east of Mayaguez, and for other varieties grown at the station. Table 1. — Ripening season and percentage of crop harvested each month for the espective varieties. Variety. Monthly quantity harvested for the whole period. Aug. Sept. Oct. Nov. P.ct. 6 14 11 27 13 22 21 37 40 13 15 13 6 1 Dec. Jan. Feb. Mar. Apr. P.ct. 'slay. June. July. Mocha P.ct. 5 15 G 14 3 1 2 P.ct. 37 31 40 32 39 31 12 t 5 P.ct. 52 37 38 24 40 41 28 34 27 1 1 3 1 2 P.ct. P.ct. P.ct. P.ct. P.ct. P.ct. P.ct. Bourbon . . 3 5 3 •j 5 26 18 22 29 27 20 19 26 9 Murta Porto Rican, Las Vegas Erecta Padang San Bamon 11 4 6 44 45 43 46 28 41 5 Maraeoeipe Columnaris Canephora 10 9 13 18 20 26 5 3 2 8 9 9 16 28 Ouillou 1 Congensis hybrids Robusta Dewevrei 4 4 22 1 1 3 Liberica 2 E xoelsa 15 7 In 10 samples of Porto Rican coffee gathered in the vicinity of Mayaguez the number of cherries per liter 6 ranged from 330 to 415, with an average of 357. In Porto Rico coffee is picked by the almud. 7 a measure of 20 liters' capacity. The average weight of an almud of coffee cherries is 28 pounds and 3 ounces as determined by measuring and weighing over 2\ tons of coffee cherries during a period of more than five years. Table 2 gives some idea of the size of the cherries of various coffees, based upon the average number per liter. Table 2. — Number of cherries per liter of different varieties of coffee. Varietv. Liberica Maragogioe Dewevrei" (general 1 Dewevrei (tree No. 2510). Padang Erecta Murta Porto Rican (Mayaguez). Average number ' of cherries per liter.** ' Variety. 139 i 209 | 263 268 328 352 355 357 Columnaris Excelsa San Ramon Mocha. Robusta Congensis hybrids. Canephora.". Quillou Average number of cherries per liter.a 388 410 515 543 564 638 654 681 a Computed from the count of cherries in 4 to 21 liter measures, excepting C dewevrti tree No, 2510, 6 In estimating size for other varieties this number is. for purposes of comparison, arbi- trarily assumed to be representative of size of Porto Rican coffee cherries as grown in this vicinity. 7 One almud of cherries yields about 5 pounds of marketable coffee. One pound of cherries equals about 0.22 pound of marketable coffee, and 1 liter of cherries about 0.:) pound of marketable coffee. Approximately 300 pounds of marketable coffee is produced per acre in Porto Rico per annum. 4 BULLETIN 30, x^ORTO RICO EXPERIMENT STATION. Each cherry usually contains two oval seeds, termed "beans' 1 in commerce. (PI. I, fig. 2.) These are greenish or grayish olive in color, elliptical in outline, rounded on one side and flat- tened on the other, with a slightly curved suture furrowing the flat side. They are placed in the cherry with their flat sides to- gether, and when one becomes abortive the other rounds out on what was the flat side and develops into an oval bean which is called a " peaberry." Being rather attractive in appearance, the peaberry enjoys an unmerited reputation for superiority over the flat bean and retails at a slightly higher price. The coffees of the Arabian group contain a smaller percentage of peaberries than do those of the other two groups, and Porto Rican samples of the group when examined at the station showed a low content of peaberries. The beans are inclosed in a cartilaginous membrane known as the parch- ment, which is underlain by a semitransparent membrane termed the " silver skin." Representative beans measure 10 to 11 millimeters in length and 6 to 8 millimeters in breadth. The number of beans in a pound of Porto Rican marketable coffee varies with the altitude and condi- tions under which the crop is produced, and may range from about 2,400 to 3,000 or more. Judging from several counts, it is thought that about 2,800 beans may be considered a representative number for this vicinity, where a smaller cherry is produced than at higher elevations. Table 3 shows the number of beans to a pound of sun-dried, cleaned coffee, the proportion of peaberries to a pound of coffee, and the weight of moisture-free beans of the different varieties. Table 3. — Comparative size of beans, proportion of peaberries in pound. weight of different varieties. Variety. Number of beans to pound of sun- dried, cleaned coffee. 1 Propor- tion of pea- berries in pound. Weight of 1,000 dry beans. Liberica M aragogipe Dewe vrei (tree No. 2510) Dewevrei (general) Padang Excelsa Porto Rican, Las Vegas Columnaris Erecta Robusta Canephora Porto Rican (Mayaguez) Quillou Bourbon Congensis hybrids (planting No. 1) , Congensis hybrids (planting No. 2) . San Ram6n Mocha 1,612 1,674 2.144 2,164 2,494 2,707 2,725 2,749 2,742 2. 726 2,791 2, 857 2,903 2,973 3,285 3,352 5,021 Per cent. 14 6 10 23 11 29 6 10 17 24 41 8 23 11 29 2 18 6 14 Grams. 242 .4 235.9 189.2 186.6 157. 3 154.1 146.9 146.4 144.7 144.4 143.1 138.9 136.9 130.6 122.5 109.3 119.9 78.2 1 The moisture content of 11 of these was determined. For eight, it ranged from 12.2 to 13.8 per cent, a variation of 1.6 per cent, while for the three others it was 11.8, 10.8, and 8.4 per cent, respectively, a total variation of 5.4 per cent. This would indicate that a small allowance should be made for variation in. moisture content in these comparisons of sun-dried coffee. 2 From a count of 1,000 beans rather than from a pound. COFFEE VARIETIES IN PORTO RICO. 5 In preparing coffee for market the skin and pulp are removed in the pulping machine, then the beans are fermented, washed, and dried, and finally the parchment and the silver skin are removed. Table 4 shows the loss in weight of each variety in the process of preparing the beans for market. Table 4. — Reduction in weight of coffee occurring during preparation for market. \ Varieties listed according to ratio of weight of coffee cherry to weight of cleaned bean.] Variety. Quillou Canephora Congensis hybrids Robusta San Ram6n Porto Rican (Las Vegas) 2 Maragogipe Mocha Columnaris Erecta Bourbon Padang Excelsa .Dewevrei (tree No. 2510) . Dewevrei (general) Liberica Weight Weight weight of 1 almud of coffee cherries. of sun- dried, of coffee cherries cleaned beans from almud. required per unit of cleaned beans. Pounds. Pounds. 31| 8 3.9 29§ 7| 4.0 29^ 7 4.2 29& n 4.4 2m 5H 5.1 29i <*T5 5.4 29| 5Sr 5.5 28& 5"T5 5.5 2S& 5 5.7 28H *i 6.3 28} ^ 6.3 29-& *?* 6.6 28H- 4 7.2 28A 3i 7.5 29£ 3f 7.9 28^ 2f 10.2 Loss in weight from dry coffee in parch- ment to cleaned beans. 1 Per cent. 12 12 13 14 15 17 17 20 18 19 16 18 26 24 27 33 1 Computed for the first four from 500 grams each of parchment coffee rather than from the almud. A sample of Porto Rican coffee as check for these ran 17 per cent. 2 These data were obtained from a coffee plantation about 7 miles east of Mayaguez, where the land is fairly representative of the coffee region. In the sample of Porto Rican (Las Vegas) coffee reported in Table 4 an almud of cherries gave 5 pounds 7 ounces of cleaned beans, the reduction from cherry coffee to cleaned bean being in the ratio of 5.4 to 1, a very favorable ratio, surpassed in this group by only one variety. CHARACTERISTICS OF PORTO RICAN COFFEE. Porto Rican coffee has been popular in Europe but is little known in the States. The following are some opinions regarding it : For some years past France has offered the best market in Europe for Porto Rican coffee and ranks next to Cuba, which takes the bulk of the crop. * * * The current retail price for a good grade of Porto Rican coffee is 37.7 cents per pound. For purposes of comparison it may be stated that the retail prices per pound for other coffees are now as follows, in cents : St Marc, 32.2 ; Mala- bar, 35 ; Santos, 31.3 ; Rio, 31.3 ; Gonaives, 31.3 ; Mocha, 36.8 ; Java, 36.8 ; Bour- bon, 37.7. Porto Rican coffee is highly regarded in France and is frequently blended with Mocha in equal proportions.* * * * The superior grades of coffee from Porto Rico now rank among the best growths produced anywhere. 8 The coffee market of Havre, by Consul J. B. Osborne, Havre, France. Dailv Cons, and Trade Rpts., Nov. 23, 1912, p. 1006. 6 BULLETIN 30, PORTO RICO EXPERIMENT STATION. The bean is large, uniform, and very attractive in appearance, the color rang- ing from a light gray-blue to a dark greenish blue. The coffee makes a par- ticularly good roasted product, and has as heavy a body as the fanciest Col- ombians or Mexicans. It is not as rich in the cup as these two coffees, but it blends very satisfactorily with both of them, as well as with other growths, and in consequence is being used very much more extensively than heretofore. 8 In a conversation with the writer, Mr. Harris, quoted above, said that Porto Rican coffee has a higher color value than any other kind. From various published reports it would appear that the amount of caffein in coffee may range from to 2.95 per cent. * * * The great majority of commercial coffees, however, have about 1.5 per cent caffein, more or less. ♦ * * Certain coffees, as the Porto Rican, are known as mild coffees from the fact that they can be easily borne by persons who say they can not drink other coffee. As a rule (doubtless subject to exceptions), Porto Rican coffees contain less than 1 per cent of caffein, and therefore are naturally "caffein- poor." 10 Determinations of caffein in three samples of Porto Rican coffee were supplied the author by W. J. McGee, chief, Food and Drug Inspection Station, United States Department of Agriculture, San Juan. The results are given in Table 5. Table 5. — Caffein content of three samples of Porto Rican coffee. Condition of coffee. Moisture. Caffein. Caffein (dry basis). Green Per cent. 11.03 2.55 2.63 Per cent. 0.98 1.31 1.30 Per cent. 1.12 Roasted 1.35 Do 1.34 From the data given in the above table it can readily be seen that Porto Rican coffee is comparatively low in caffein. This, coupled with its pronounced aroma and high color value, makes it a very desirable article of commerce where these characters are appreciated. COFFEES BOTANICALLY SIMILAR TO PORTO RICAN. Such coffees as Padang, Preanger, Menado, Kamerun, Philippine, Kona, Surinam. Guadeloupe, and Jamaican Blue Mountain so closely resemble Porto Rican in tree and fruit as to be indistinguishable from it in the field. The degree to which the quality of these coffees has been changed by change in location is not known. In 1913 samples of Porto Rican grown Padang, Pantjoer, Preanger, Guade- loupe, Blue Mountain, Erecta, Columnaris, Maragogipe, Mocha, and Porto Rican coffees were sent to S. B. Morison, of New York, who had them tested and reported on them as follows : I had the samples tested by some of the best tasters in tlie city. On some things they agreed. One of them was that there is very little, if any, of the character of the seed left in the Mocha, Java, etc.. that you had grown in Porto Rico. They all had no trouble in picking out the genuine Mocha and Java, which I had drawn with your samples. In all cases, as is customary in tests of this kind, we had "blinded" the cups. As to the various samples of Porto Rican coffee, the highest-grown coffees 11 were, almost without exception, the Coffee characteristics, No. 7. — Porto Rican coffee, by W. B. Harris, coffee expert, U. S, Dept. Agr. In Tea and Coffee Trade Jour., vol. 27, No. 5 (1914), p. 423. 10 Scientific analvses of coffee, bv O. W. Wilcox and M. J. Rentschler. Tea and Coffee Trade Jour. (1911), vol. 20, No. 1, p. 32, 33. 11 Coffees grown at highest altitude. COFFEE VARIETIES IN PORTO RICO. 7 best. There were one or two samples that were supposed to be high grown i did not show up the snap they ought to. As a rule, the coffees seemed to lack what we call acid or snap. That was, I should think, the one most consistent criticism of the various tasters. * * * The consensus of opinion about the samples of the coffees that you have imported from other countries to grow in Porto Rico was that generally they retain little, if any, of the quality of the original. The different tasters do not agree as to the amount retained. Some say nothing and others a distinguishable amount. Coffees of the Arabian group at the station are growing in stiff clay soil on rather steep slopes. The soil is no better than that of the average coffee plantation, and is poorer than on many planta- tions; and the grove is located on an elevation just above sea level, where conditions are less favorable for Arabian coffee than is the case at higher altitudes. Padang Coffee. Padang coffee is highly thought of in the world's markets. In November, 1921, it was quoted at 22 to 24 cents with Porto Rican at 114 to 15-J cents a pound. Padang coffee is a Sumatran variety, the seed of which was received at the experiment station from the Botanic Gardens, Buitenzorg, Java. (PI. II, fig. 1.) In general appearance both tree and bean are similar to the Porto Rican. More than 250 trees of Padang coffee have been planted at the station. Seed was planted in Decem- ber. 1908, and the seedlings were transplanted to the field to an area of one-fifth acre the following August. Thev were 4. 6,* 9, and 10 to 11 feet high, respectively, in March of 1911. 1912. 1913, and* 1914. Plantings were made on both fertilized and unfertilized ground, and after the small 1911 crop. The fertilized plat for the 11-year period 1912-1922 annually yielded an average of about 3.1 liters of cherries per tree, and the unfertilized plat, approximately 2.7 liters with a maximum yield of 5.1 liters. 12 Among the Arabian varieties the Padang bean ranks next largest after Maragogipe, considerably surpassing Porto Rican samples. (See Table 3.) Padang loses considerably in weight from cherry to cleaned bean* when prepared for market, there being much less cleaned coffee per almud than in case of the Porto Rican variety. (See Table 4.) Bourbon. De Wildeman 13 says that Bourbon coffee, owing to its fine aroma, has always been one of the most highly esteemed sorts. (PI. V, fig. 1.) According to Lalie.re. 14 it is grown in the deepest and rich- est soils on one-fifth of the plantations of the State of Sao Paulo, Brazil, and it is more exacting in its requirements than the ordi- nal* Brazilian coffee. In April, 1915, seed of this variety was received from J. Hill, the owner of a large coffee plantation at Santa Ana, Republic of Salvador, Central America, who wrote regarding it: I am staking my whole future upon Bourbon. 1 have Ifaurbon out in sun at 2,000 and at 3.000 feet, and in hoth places it looks well, and at two years from planting out the trees have a good crop on them. 12 The approximate yield of marketable coffee per nee. or per acre, arbitrarily ;t*snimnsr 700 trees to the acre, can easily be estimated, since there are about 4 liters of 'cherries of the local or similar coffee to a pound of cleaned coffee beans. 13 Les plantes tropicales de grande culture. E. de Wildeman. 1902, p. 99. a4 Le cafe (Bresilt, p. 39. Amour Laliere. 1909, p. 39. 8 BULLETIN 30, PORTO RICO EXPERIMENT STATION. In September, 1916, 60 young trees were set in the field, where, having been fertilized, they are making good growth. In general appearance they do not differ greatly from the Porto Rican. Bour- bon is an early variety, nearly half of the crop having been har- vested by the end of September. The reduction in weight from cherry to cleaned bean is greater for Bourbon than for Porto Rican (see Table 4), and the bean is smaller (see Table 3), both of which points are market disad- vantages. Bourbon may be mixed with Porto Rican in marketing. The production is given in Table 6. Pointed Bourbon. According to Lecomte, 13 Pointed Bourbon was discovered by Le- roy in Reunion, and its aroma is not so good as that of the oval-seed type of Bourbon, known in Reunion as " cafe du pays." Lecomte also states that this coffee is classified by some as C. laurina. Pointed Bourbon is mentioned by de Wildeman 16 as one of the two varieties of C. arabica grown in Reunion, where it is known as Cafe Leroy. It differs from Bourbon proper in tree, foliage, and fruit. The tree grows very slowly, 8-year-old trees at the station being about 6 feet high. Older, fully developed trees, however, are 8 to 10 feet in height. In form, the tree is conical with a rather stiff, erect trunk. The primary lateral branches turn upward when young, the secondary laterals are numerous, and the internodes are short. Some of the internodes on the laterals may measure l'J inches, and the average internodal lengths for the whole branch may be only -J to f of an inch. The leaves are elliptical, tapering to an acute base and apex, and are three times as long as broad, representative specimens ranging from 2 to 3^ inches in length. The cherry is comparatively small, as many as 414 being counted in a liter. The bean is pointed in shape and should not be mixed with typical Arabian coffees for marketing since it differs from them in shape. The tree, being small, yields a small crop. In 1922 8-year-old trees at the station yielded, on the average, 0.9 of a liter of cherries each. The variety is not recommended for planting. Brecta. The variety Erecta was received from the Botanic Gardens, Buitenzorg, Java. (PL II, fig. 2.) Dr. P. J. S. Cramer, of the Department of Agriculture, Dutch East Indies, in his correspond- ence with the experiment station, writes, regarding this coffee: C. erecta is found from time to time in our plantations of C. arabica typica. It is probably a mutation repeating itself. * * * I consider it well fit for spots exposed to heavy winds. Productivity, satisfactory; growth, some- what more vigorous than the type. Not planted on a commercial scale. In June, 1909, 236 small Erecta trees were transferred from the nursery to their permanent place in the field. During six years, beginning with 1911, the trees attained in height approximately 15 Le caf6 : Culture, manipulation, production. Henri Lecomte. 1899, p. 298. 16 Les plantes tropicales de grande culture. E. de Wildeman. Bui. 30, Porto Rico Agr. Expt. Station. Plate I. Fig. I. — Coffee Foliage: I, Porto Rican; 2, Padang; 3, Columnaris; 4, ERECTA; 5, BOURBON; 6, POINTED BOURBON; 7. SAN RAMON; 8, MURTA; 9, MOCHA; 10, MARAGOGIPE; II, C. QUILLOU: 12, C. DYBOWSKII; 13, C. LIBERICA; 14, C. DEWEVREI; 15, C. EXCELSA; 16, C. CANEPHORA; 17, C. ROBUSTA. §*••••• •••• • SCTA # • • • • • | C. -:'. ' • #§•##•§•• C. POv . 001 m Fig. 2. — Coffee Beans. Bui. 30, Porto Rico Agr. Expt. Station. Plate II. LU HI _ o O o LU CC LU 6 X ^- o O a z D o >- LU LU Lu Ll_ O O a z < a < Q. a Ll Bui. 30, Porto Rico Agr. Expt. Station. FLATE III. lU > r ■ ' ■ jyij; ^a*k^2 ^SrL" I--*-* ^ "-HlfciJS&O ^^^^■^^PPEiil ^^^R^WKKiaK wmS^I ' ^^^*- ^BBg^^jK.^Lfcirv"'^^w?^^HH ^^^~^^^Ey ^ ** 2- -'•^ v • -, «j2 * ^ K. '* ■ . <• jS . & • ' § ~~1 ^HE^I - ■ * m 'GIlflDtillMA- ^"^ f* : • -r \ * 4»- jj ■ >^ 2^ '' ^ **£.-.-- ^"V» . *■"* '-■- v ■■ '•». .***—» ■ . * I **''£>• _£ ''■ *i£2 mtm-.i ■t. *r id^CSfetSftLt ^m£». i"— «^ ^ • •S B^Sfij v - rfti ■ UJ UJ LL u. o o CO X < z o o o Ll Bui. 30, Porto Rico Agr. Expt. Station. Plate IV. UJ 111 Ll Ll o O z < CC z < CO LU UJ cc cc < lu >- I Oi o *~ Pi. v. a> W pa < c a. > 60 o Eh 8 © CM CM © © oo © © © CO © CM © "3 03 a .2 © 'C 03 P c>j co t-3 »h co *A co cm — *i CO CO 00 00 ■>* t> "5 r+ 'i-4cM ' »H — '< CM rt ^!CONNCC»NMN®0> fc©^©i-5^r^coCMt^cOcNoo ~CNCO — CO !N CM — — NOSCflN'*^ KONONtOiiMN © oc cm © cm © io> oocor^©cococco0'<* cm i^oonv r-i OKCOOOO^CO r^. © co >c © cm © oc oc ^ c^© — icascMiOTf — -^cm©© ^ io cc cm e-(i-( "5 O ■* tC » t^ ' CM" — •& CM i-< CM O0 CM CO i-H CO " i-J io co <-! CO CM © — -11(5 rtOO ce9 oo oo Tt< r- r>- io ic r- t- cm r— © n©ncec-h-< -^ ■* >o ■* m Tf io "5 io © O i—l O O O C © O — © © — — © © i— ' —< i-H HHHrtrtr«HHH ©©©©©03©0^©©©OS © © © © © CTi © CT>Cfc©©©C2©©© o.S M _5 4-: C p. a o — *■£ pV§.8 8 ^ w 3 g bCt*c _, « «S b o S o T t; - a,' © ce c« cs c •- in 2 — a, cq W W 85 C fl 8^' of ce o O c 'tic C C3 ^d 03 S DO 5 c« C, '- C OS a 03 — CM CO — CM-HCMr-lCM cccoooooo trt£MitjCWM;txCfcCbc .5 .S .S .£ .5 .9 .S -S .£ '43 +j '43 '■£> "-P "+i '-^ +j «a pcccdaccc C C Cv C C CC bC nS — = > ■c gg 0J 3 * CO o DO O Q o Oh c X! Ch o 03 O c? x) a « b£ a o •o a S3 CD o Xi y- +» *a > u S3 Si 03 1 O e — 03 £ >, c *i ~eb to CO 03 C3 o X > 03 05 CD u CD 'ro P o o a o CD X 03 © 03 T3 a _© 03 O C ■s o d CD CO oc © CM © t~ •H S© 03 CO >> CD X CC X ° "3 r/- ? 03 cct a CO ■a C 1 co •— s © s 3 o lo — to "£ co © © * 3 &.'^3 _ 3 03* s & n CD 60 X'C +J o co iF cp 3 E^2 © T3 co © CD 73 © *> 2^ CD 6C © 3 2 * co - i^- 1 r/l •CO — r^ ^ e8"3 i|*^o! ^ 03 © C2 2"E-£ «-© — ®.S2 £ - -cxx S- 5^** BIS 'OcJp W 3 3 CD S3 03 g ^•3 6cj; © co co 3 2*S T' £ ^ £ § 3! © ^ 3 2c3g£o Ci fc © 3 ■si I ^ &o ^^2.£>»§ -|g^S£ V- "* +3 p,© O -3 3 3 — ^ *- © O 2 03 £_ d 2 © tr © "3 E £ co -r so S3 O'+aX o C« 3 «- on ij V- 3 k!"*"l3 co © ©5- ^5 « o 3 +-~ * ° ra^^n o3 c^ S o£3^?J »?" fl u o t- - 1/ ©O feP -5 © c .,© -|2 2 ^ ce « W -*»3 3S <=>x - X - o © - 3 3* O «3 'cc © r^ CO X O c- O© s3>2g^ 3 3 S« f_i <, f 1 ^* i-i — g w © 03 r 6 £ 5 — ,_ w CO DO COFFEE VARIETIES IX PORTO RICO. 11 Although the cherry was found to be slightly smaller than that of the Porto Rican in the samples compared ( Table 1 ) . the bean is large, approximating in size the Porto Rican grown at higher elevations (Table 3). An almud of cherries, reduced in weight in processing in the ratio of 5.7 to 1, gave 5 pounds of marketable beans. (Table 4.) This variety may be mixed with the Porto Rican. from winch it is in- distinguishable in appearance. Its cup flavor is excellent. Columnaris is recommended for planting because of its vigorous growth, fair productivity, relatively large bean, and cup flavor. Maragogipe. Keable 18 states that Maragogipe was discovered in 1870 near the town of the same name in Bahia. Brazil. He also says that the variety is commercially important and has been rather extensively grown in Guatemala. Seed of this variety was received by the station from Guatemala and from Java. In September. 1909, 153 trees, pre- sumably 20 months from seed, were set in the field. Root disease present in the soil has slowly and steadily reduced the number of trees, as in the Columnaris planting. Maragogipe. differs from typical Arabian in both foliage and in- ternodal length., (PL III. fig. 2.) The internodes attain such a length as to limit their production on a given area of branch and cause the tree to present a straggling appearance. The longer internodes on the uprights may range from 7 inches to over a foot in length. As is true of the other varieties, the internodal length varies constantly. The internodal lengths measured, from base to tip. 6. 51, 3i. If. 2. 4i. li. and 2i inches on a representative lateral branch. immediately above a 9-inch internode. The leaves of the variety are larger than those of other varieties of Arabian coffee, some of them ranging from 7 to 10 inches in length and 3 to 4 inches in breadth. The shape also differs, being lanceolate with an upper convex surface. The ripening season of Maragogipe is almost as late as that of Columnaris. (See Table 1.) The record of yield for the 11-year period 1912-1922 is shown in Table 6. In the last four crops' are included some younger trees, more than 5 years old, which were set to replace dead trees. Xot only was the average annual yield for the entire period low. being 1.6 liters of coffee cherries to a tree, but the maximum average annual yield was only 2.4 liters. The cherries and beans of Maragogipe are the largest among the Arabian varieties. In a count of the cherries in 10 liters of each it was found that only 59 per cent as many Maragogipe cherries were required as of Porto Rican to fill the measure. The 1.000-bean weight was 70 per cent heavier for Maragogipe than for the Mayaguez sample of Porto Rican coffee. (See Table 3.) From an almud of cherries 5 pounds 7 ounces of marketable coffee was obtained, and the ratio of reduction in weight. 5.5 to 1. was very favorable. The cup flavor seems much more distinctive than is true of the other coffee varieties tested at the station. Doctor Cramer 19 says that it ; > "the finest coffee known : it has a hisrhlv developed, splendid flavor. ^Coffee. B. B. Keable. London. 1909, p. 16. -• Coffees of the. Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jou*-., vol. 35, No. 4 (1918), p. 321. 12 BULLETIN 30, PORTO RICO EXPERIMENT STATION. Maragogipe is, however, a light yielder." The latter statement seems to voice the general experience. Maragogipe is not recommended for general planting because of its low production. San Ramon. Seed of San Ramon, received from J. Hill, Santa Ana, Salvador, was planted in the nursery in February, 1913, and 321 seedlings were transplanted to the field in August, 1914. Regarding the variety, Mr. Hill, in his correspondence with the station, writes as follows: Some call this variety "San Lorenzo," but the right name is San Ramon, and it originated in one of these Central American Republics. These trees seem made by nature to stand a very exposed windy outlook as they are strong and stiff and hardly bend under a high wind. * * * This San Ram6n does well where the winds make the cultivation of Arabica or Maragogipe difficult, and it seems to do well in other places where the Arabica has been growing for years and has got played out. The San Ramon, being a small tree, is planted closer than the Arabica, say 2 by 2 yards, and its pro- duction per acre is about 600 pounds clean bean. The quality in the cup is all right. The slow growth and dwarf habit of San Ramon were shown by measurements of average height taken annually at the station. At time of setting the trees averaged 19 inches, and thereafter annually 27, 39, 48, 53, 61, and 68 inches, which was an increase in average height of never more than a foot a year. (PI. IV, fig. 1.) Although there is a rather wide range in variation in the station plantings, representative trees are dwarf and bear short internodes, permitting of the production of a large number of cherries on a comparatively short length of branch. (PL IV, fig. 2.) A 15-inch branch may carry 15 nodes or even a much larger number. The leaves are broadly elliptical in outline, very dark green in color, and are comparatively flat, being only slightly wavy. The base and apex are obtuse, except for the acute tip of the former. Representative leaves range in length from 2J to 5 inches, and in breadth from 1 to 2J inches, 4 by 1J being a typical size. This variety comes into bearing very young. In August, 1915, one year from setting and two and one-half years from planting the seed, 17 per cent of the trees bore fruit. This percentage is increased as much again if only half of the planting receiving greater care in transplanting is considered. 20 In 1916, 90 per cent of the trees were producing. San Ramon is one of the three late-season varieties of the Arabian group, more than one-third of the crop maturing in December and January. In addition to ripening late, this variety has a long bearing season, 14 per cent of the yield having been picked in August and September. Table 6 shows the average production per tree, which, during seven years, amounted to 0.9 of a liter of coffee cherries. Although this yield is small in comparison with that of other varieties, it is large in proportion to the size of the tree. The dwarf form is much easier to harvest than is i lie typical Arabian. Owing to its stocky form and short internodes. the trees may be set 4 to 5 by 6 feet apart. The uneven coloring of the ripening cherry is characteristic of San Ramon, the surface above the line of division between the two beans 30 Porto Rico Sta. Bui. 22. Effect of different methods of transplanting coffee, p. 8. COFFEE VARIETIES IX PORTO RICO. 13 turning deep red while that over the convex face of the bean is still ^Bv 9 . "it* § — .1 ■■* ■ ^v '1 » 1*. HUM * a **L mm -. * Wh *< JL>.---. .■■■'■; W--$ ^1 Is |yg^ «m - y i < O o Z UJ UJ tr Ul UJ CO - o UJ (r DC U. > < UJ UJ < UJ u. u. o o o Ll UJ UJ u. Ll o O cc D u. . OS o LU CO LUg r~CQ UJ cc D I- < a ll Bui. 30, Porto Rico Agr. Expt. Station. Plate VII, Ll < CO -I UJ o X u < UJ Li. _ o O 6 Bui. 30, Porto Rico Agr. Expt. Station. PLATE VIM, FIG. I. — COFFEA EXCELSA AT A LITTLE MORE THAN 6 YEARS FROM PLANT- ING Seed. FIG. 2. — COFFEA ROBUSTA AT 3i YEARS FROM PLANTING SEED. COFFEE VARIETIES IN PORTO RICO. 17 favorable ratio shown by any variety tested. A representative bean is 12 to 13 millimeters long by 8 to 9 millimeters broad, fairly plump, with the concave face furrowed with a broadly open suture. One thousand Liberica beans were found to weigh more than an equal number of Maragogipe, the largest of the Arabian group. The bean is of a greenish straw color, and is wrapped in silver skin of glisten- ing clay color. Two samples of marketable coffee were found to contain 1,612 and 1,786 beans per pound, respectively. When the very heavy reduction in weight from cherry to market- able bean is considered in connection with the individual yield, and also the wider spacing required for trees of this species, it will be seen that the harvesting is costly and the yield per acre very light. Considering further the difficulty of harvesting, due to height, if the trees are untopped, and of pulping, there is little to warrant the planting of this variety. EXCELSA. Excelsa is said 30 to have been discovered in 1905 by Aug. Chevalier in West Africa, in the region of the Chari River, not far from Lake Tchad. In writing of it in Java, Doctor Cramer says that it is not particular as to soil, and is suited to the same climate as is Liberian. but will do very well also in higher altitudes. He states that at Bangelan the best field produces 760 to 1,500 pounds per acre. Doctor Cramer pronounces its quality good and its growth vigorous, and asserts that it is one of the best kinds to grow for produce. The first Excelsa coffee in Porto Rico was received by the station from the French Government in 1906. Two inarched plants, severed from this tree in December, 1912, and set in the field in May, 1913, were 22J and 23^ feet high in June, 1922. The trees began to bear in 1917 and made an average yield of 13 liters of cherries each in the season of 1921. Fiftv-four trees, the seed of which was received from Bangelan, Java, in November and December, 1915, were set m the field in August and September, 1917. These have made very vigorous growth (PI. VIII, fig. 1), attaining an average height of 10.9 feet in June, 1922, and having an average spread of 9.1 feet. At this time the height of the tallest tree was 18-J feet and its spread 1Q| feet. The spread of the broadest tree was 14 feet with a corre- sponding height of llf feet. The only pruning given these trees up to this time consisted in the removal of suckers. Doctor Cramer, however, who has had extensive experience with Excelsa coffee in Java, where hundreds of acres are planted to it, in his correspondence with the office, advises topping at 3 or 4 feet when young in order to develop the lower primary laterals. When these have attained con- siderable size an upright is allowed to develop and continue the growth of the tree to a height of 12 feet. It is again topped and thereafter kept at this height to permit of picking from ladders. Doctor Cramer also advises the removal of all suckers, not only from the main trunk, but likewise from the primary laterals. Topping is necessary for all members of the Liberian group to facilitate picking. 30 Coffees of the Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jouv., TOl. 36, No. 1. p. 25. (19.1JU 83926 — *>a ° 18 BULLETIN 30, PORTO RICO EXPERIMENT STATION. In general appearance Excelsa somewhat resembles Liberian, but there are very decided differences in foliage, flower, and fruit. The young tree is much more spreading and the leaves are larger, oval, and more rounded than those of Liberian, though of similar stiff texture. Representative leaves are 8 to 12 inches in length and 4 to 6 inches in breadth, a typical specimen measuring about 9J by 5 inches. On laterals representative internodal lengths are 2 to 5 inches. The large, white flowers have only four to six petals, like the Arabian, but are unlike it in blossoming season, heavy blossom- ings appearing in summer or autumn when Arabian coffee is not in flower, and the trees at times carrying simultaneously buds, blossoms, and immature and mature fruit. The fruits are thickly clustered, as many as 35 and 40 appearing at a single node. (PL VII, fig. 2.) They are much smaller than those of Liberian coffee, and appear solid colored. Upon examination, however, there is found a fine striping in nopal red to maroon, and in some specimens, alternate pale yellow and red. The disk, whicji is very variable in size and often only slightly russeted, is not so prominent as is that of Liberian coffee, but it is more prominent than is that of Arabian. Of the 54 trees set in 1917, 1 died, 2 fruited in 1920, 41 in 1921, and 50 in 1922. The major part of the crop has ripened in the three months of spring. Since the Arabian crop ripens in the fall, the production of both on the same plantation would offer certain ad- vantages in regard to labor. Another advantage in planting Excelsa is that the cherries, in common with the other members of the Liber- ian group, do not fall on ripening. Pickings at the station have been made at monthly intervals. The ripening season is shown in Table 1, and the yield in Table 6. Seven samples were found to contain between 369 and 455 cherries, with an average of 410 cherries per liter, showing them to be ma- terially smaller than those of the Porto Rican. The pulp is firmer than that of Arabian coffee, but unlike Liberian in that it is rather thin and therefore pulped with greater ease. Five almuds of freshly picked cherries ranged in weight from 28 pounds 3 ounces to 28 pounds 11 ounces. Two almuds of cherries picked at different times in the season checked very closely, weighing when fresh 28 pounds 10 ounces and 28 pounds 11 ounces, respectively; and giving, after pulping and drying,, 4 pounds of cleaned beans each. Each lot showed a reduction in weight of 26 per cent on removal of parch- ment and a total reduction in weight from cherry to marketable bean in the ratio of 7.2 to 1. One thousand moisture-free beans from each almud weighed 153 and 154.1 grams, respectively. In weight, the bean does not differ greatly from the Porto Rican, there being both a greater and a less number per pound than was found in the two samples of Porto Rican given in Table 3. A rep- resentative size is 9 or 10 millimeters long by 7 or 8 millimeters broad. The silver skin is of a pale brownish color and the bean is straw-colored or yellow. A considerable portion of the silver skin adheres to the beans, giving them a rough and uneven appearance, unless they are rapidly dried by artificial heat, and a fragment of the parchment is likely to be held in the deep sutures when being broken off. Since the appearance of the bean is totally different from that of the Porto Rican, the two should not be mixed for marketing. COFFEE VARIETIES IN PORTO RICO. 1$ Excelsa is the most promising of the Liberian group of coffees so far tested at the station. Its good cup quality, vigorous growth, comparatively low (for the Liberian group) ratio of reduction in weight, indications of productivity, and resistance to the leaf miner recommend it for trial, especially in localities which suffer from ravages of this pest. Dewevkei. This species is indigenous in African Congo, whence seed was received by the experiment station in September, 1911. The sta- tion planting shows an extremely wide variation in type, and sug- gests a heterogeneous collection of several varieties. Individual trees differ in size of foliage ; in color of blossoms, some being white and others pale pink; in number of petals, flowers having five, six r seven, or as many as eight petals; in size, shape, and color of cherry; in size, protrusion, or flatness of navel; in thickness of pulp; in color of silver skin ; and in shape of bean. In August, 1913, 56 trees were transplanted from the nursery to the field, some of the trees then being 3 feet high. Two years later the trees had attained an average height of 5-| feet, and annually thereafter for five years they were 7f , 9^, 11J, and 14| feet high, re- spectively. In another two years the average height was 17^ feet, and the tallest trees were about 25 feet high. (PI. VI, fig. 2.) Fully developed specimens are said to attain a height of 50 feet. Some unpruned trees make a tall columnar growth and develop rather frail lateral branches, while others spread into pyramidal form and develop very heavy, thick laterals. In plantation practice it is ad- visable to limit the height to 12 feet in order to facilitate picking. These trees, like the others of the Liberian group, have dark green, leathery leaves ranging from 9 to 14 inches in length by 3£ to 6 inches in breadth, the length usually being 2-J, and in some instances 3 times the breadth. In general, the leaf shape is subspatulate with acuminate base and acute to obtuse apex, and has a slightly wavy surface. The foliage is dense, the tree forming a close screen. On most of the trees the cherries are more closely clustered and their pulp is thinner than is the case with Liberian. Counts of 11 samples of cherries, picked from eight trees individually, showed between 208 and 329 specimens per liter, with an average of 263 cherries. (PL VII, fig. 1.) In color scheme the cherries are reddish- yellow, producing in some a solid effect, or yellow striped with red. The individuals differ widely in their season of ripening, and ma- ture cherries may be found on some trees during almost any month of the year. The major part of the crop has ripened in the winter months at the station. The trees came into bearing in the winter of 1915-16. The aver- age yield for the planting as a whole and the individual yield of four of the best trees for that period, and for subsequent years, is given in Table 6. On various occasions more than an almud of cherries has been picked from a single tree at one picking. The crops of 1920, 1921, and 1922, taken together, showed an average annual yield of over half an almud per tree for 38 per cent of the trees. The average bean is intermediate in size between that of the Mara- gogipe and the Padang. The ratio of reduction in weight for the 20 BULLETIN 30, PORTO RICO EXPERIMENT STATION. two almuds on which data are given in Table 4 was less favorable than for Excelsa, the two almuds of Dewevrei each giving 3 pounds 12 ounces of marketable coffee. In this planting there are some worthless trees, together with a number of notably vigorous and productive trees of a much more desirable type than is the ordinary Liber i an coffee. Where such variation is exhibited among trees it would seem ad- visable to practice shield budding of the more desirable strains for further propagation. This is a simple process but is not practiced locally. The trees should be set at planting distances of about 12 by 12 feet and be topped at a height of 12 feet. Other Members of the Liberian Group. Coffea abeokutce. — In the station plantings there are a few trees of this species which came from St. Lucia, West Indies. These are vigorous and in foliage, fruit, and manner of growth are similar to C. lib erica. The cherries are large, 99 having been found in a liter, and the pulp is thick. Being very inferior to C. excelsa and to many individuals in the planting of C. dewevrei, this species is not recom- mended for trial. Coffea dybowskii. — Seed from a tree of this species that was grow- ing surrounded by trees of C. excelsa, and thought to contain possi- bly some hybrids of the two, was received from the Java Experiment Station in the years 1916-1918. Trees of the earlier importation are just coming into bearing. They look promising, but it is too soon yet to predict their merits. In growth they are very vigorous, and in general appearance are suggestive of a magnified 0. excelsa. with their longer internodes, larger leaves, and larger closely clustered cherries. The internodes on laterals range from 3 to 9 or more inches in length, 4 to 7 being a representative length. The leaves are oval, of heavy texture, and the largest of any found in the station plantings, measuring 6 by 14 to 9 by 20 inches. The flowers are five- petaled and If inches across. The cherries are very large, averag- ing 185 per liter in two lots taken from one tree, and they have a rather thick pulp. ROBUSTOID GROUP. Robusta coffee, with its related species, has come into prominence only within recent years. Being considered poor in cup quality, this coffee has been rather reluctantly purchased by American deal- ers. It has been rather extensively planted in the East Indies be- cause of its heavy production. In growth the tree is intermediate between that of the other two groups. Its ripening season in Porto Rico is distinct from that of the Arabian coffee, the crop beginning to mature toward the end of the Arabian harvest. The pulp is very thin and an almud of Robusta cherries will therefore contain more marketable coffee than is true of either Arabian or Liberian. The ripe cherries do not drop readily, and pickings may be made every four weeks instead of every two, .as is the local practice. The grow- ing of this coffee for export is not recommended, but it may serve to furnish a low-priced article to that part of the local trade which pre- fers a cheap rather than a high-priced coffee of good quality. COFFEE VARIETIES IX PORTO RICO. 21 Kobusta coffee was discovered in the Congo by Emil Laurent in 1898. 31 It was put on the market under the name of Coffea robusta by a horticultural firm in Brussels. Doctor Cramer states S2 that young seedlings were imported in 1900 from Brussels into Java, and that what is known as Robusta coffee is probably a group of coffees of different varieties rather than a pure species. The rapidity with which Robusta coffee has been taken up by the Javan planters has been phenomenal. Within 20 years of its introduction. 298,774 acres, or 83.5 per cent, of the total area in coffee in the Dutch East Indies were planted to Robusta, while only 19,780 acres, or 5.5 per cent, were in Arabian coffee. 88 Of Robusta coffee in Java, Doctor Cramer writes as follows : 34 Robusta is marked by rapid growth, early fruiting and high productivity, * * * likes wet climates, especially those with regularly divided rains. It is grown in Java from sea level up to more than 3.000 feet and in Sumatra at even higher altitudes. The general altitude for it may be put at 1,000 to 3,000 feet above sea level. * * * Robusta does not like wind, and suffers especially from continuously dry winds : it requires shade, at least in higher altitudes. * * * Under favorable conditions the crop may be 1,520 pounds per acre. The average may be estimated at 750 to 1,150 pounds per acre. * * * The crop comes in all the year round in climates with regularly distributed rainfall ; in regions with a sharply marked dry season the crop lasts three to four months, according to the rains. * * * Robusta and allied species are not so appreciated as other kinds. They rank a little under good average Santos. To clean the beans from the silver skin artificial drying is necessary. The skin may be removed by drying in the sun and washing, but. then, second quality is obtained. Robusta is a bulk produce, not, as Liberian and Java coffee, a fancy quality. .At the station robustoid coffees have been received under such names as Coffea robusta. Coffea quillou. Coffea canephora. Coffea laurentii. and Coffea ungandce hybrid. Judging from the plantings at the station, a number of specific names have been applied to a group of closely related coffees which were propagated by seed and will likely continue to cross. In some instances differences can be seen: in others, the difference in name is the only dissimilarity. All. however, would be classed in the group known as Robusta coffee. Here, also, are included the station plantings from seed 2'eceived as Coffea congensis. as explained later under that heading. Robusta. Especially selected seed of the best trees of Robusta coffee was received in October. 1914, and November, 1915, from the experiment station at Buitenzorg, Java Two plantings of the first lot of seed were set in September, 1915, one of 114 trees and the other of 28 trees, and a third planting of 60 trees of the second lot of seed was set in August. 1917. A few trees of each lot were lost within the period reported. At a year from setting the tallest trees in the first planting were 50 inches high, but the average height was only 30 inches. By the 81 Coffee robusta. W. J. Gallagher. Dept. Agr. Federated Malav States Bui. 7, p. 1. (1910.) 82 Coffees of the Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jour. (1918). vol. 35, No. 5, p. 417. 88 Coffee in the East Indies. J. S. Fowler. Tea and Coffee Trade Jour. (1920), vol. 39, No. 3, p. 299. 84 Coffees of the Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jour. (1918), voL 35, No. 5. p. 417. 22 BULLETIN 30, PORTO RICO EXPERIMENT STATION. following September, or two years from setting, the maximum height attained was 119 inches, the average height was 59 inches,, and a number of trees were over 7 feet. (PL VIII, fig. 2.) Fur- ther measurements were not made as topping at about 8 feet was begun within the year. Unless they are topped the trees will soon make such vigorous growth as to render picking difficult. The lat- eral branches are rather delicate. Representative lengths are 2J to 4J feet, although some exceed this. The internodes on laterals range in length from 2 to 6 or more inches, with an average of 3 or 4 inches. The leaves of the various individuals vary somewhat in shape, some being oval with obtuse base and apex having acuminate tip, and others being elliptical with acute base and apex. In length they may range from 6 to 12 inches, but generally are 7 to 10 inches. The length ranges from less than twice to nearly three times the breadth. The leaf texture is not so heavy as in the Liberian group, and the leaves are subject to attacks by the leaf miner. The blossoms are white, five or six petaled, and from 1^ to 1£ inches across. They appear at irregular intervals, and may be seen on the trees simultaneously with green and mature cherries. (PL X, fig. 2.) The ripening season for the three plantings is given in Table ]L In common with the other members of this group the crop began to ripen in October, but ripened mainly in the winter months, nearly half maturing in January and the rest in the spring. Only a negligible quantity of cherries ripened in the fourth year from seed. The yields, which are shown in Table 6, have been surprisingly low for a coffee celebrated elsewhere for its productivity. The cherries are closely clustered, there being 15 to 30 or more at a node. (PL IX. fig. 2.) When ripe they are blood red, some being of solid color and others finely striped in dark and light red. Striping is less conspicuous when the cherries are very ripe. Occasionally the navel is flat, but more frequently it protrudes, and the pulp is very thin. The cherry is smaller than that of the Arabian coffee, representative specimens be- ing 14 millimeters long, 13 millimeters in the diameter at right angles to the flat faces of the beans, and 10 to 11 millimeters in the other diameter. Five samples showed between 528 and 630 cherries per liter with an average of 564. Two almuds of freshly picked Robusta cherries weighed 28 pounds 11 ounces and 29 pounds 9 ounces, respec- tively. The latter gave 6 pounds 12 ounces of cleaned marketable beans, a reduction in weight in the ratio of 4.4 to 1. Two samples were found to contain 2,726 and 2,736 beans per pound, respectively, showing the beans to approximate the Porto Rican coffee in weight, though it differs from the latter in its more rounded form. The silver skin is brownish and rough in appearance, and adheres closely to the bean. It does not clean off readily after slow drying in the sun, but is effectively removed when rapidly dried by artificial heat. The bean does not present an attractive appearance unless the silver skin is thoroughly removed. Canephoba. Considerable variation between individuals is found in the station plantings of Coffea ccmephora, The range is from the type showing narrower leaves or more flattened cherries than those of the typical C. robusta to others which are wholly indistinguishable from O. ro- busta in habit, foliage, and appearance of cherry. This is perhaps COFFEE VARIETIES IN PORTO RICO. 23 accounted for by the crossing of two closely related strains. Leaves may vary from 7£ or 9 to 10 or 12 inches in length, and from 2J or 3 to 4J and 6 inches in breadth. Usually the cherries are striped, but some appear solid in color, being a lighter red and slightly smaller than those of G. robusta. One tree, received from the French Government in 1906, has fruited heavily in some seasons, giving a maximum yield of 14 liters of cher- ries at a single picking. (PL XI, fig. 1.) Two plantings have been made of G. canephora, one of 24 trees from seed sown in the last quar- ter of 1914 and the other of 60 trees from seed sown a year later. Seed of the latter and part of the former came from the Java Experiment Station. The yield of the two plantings is shown in Table 6. An almud of cherries, weighing 29 pounds at picking, gave 6 pounds 6 ounces of marketable beans, or a reduction in weight in the ratio of 4.5 to 1. Another almud picked in a different season, weighing 29 pounds 8 ounces, gave 7 pounds 6 ounces of marketable beans, or a reduction in weight in the ratio of 4 to 1. Counts of six samples showed between 552 and 704 cherries per liter, with an average of 654. Twenty-five unselected beans from a large and representative lot of C. robusta and G. canephora coffee were measured for purposes of comparison. In length, breadth, and thickness, G. robusta averaged S.7, 7, and 4.4 millimeters respectively, and G. canephora, 8.2, 6.6, and 4.2 millimeters respectively, showing the latter to be the smaller in each of the three dimensions. As is true of Robusta, the silver skin adheres closely to the seed of G. canephora, but it is more uniformly brown in color than is the former. Latirentii. One tree of Goffea laurentii (S. P. I. No. 28080), which is culti- vated in the Congo Free State, Africa, and known also as G. robusta, grows at the station. (PL X, fig. 1.) It has not been observed to differ from the other Robusta coffees, and has fruited very heavily, often producing half an almud of cherries in a season. In one season it produced 19.4 liters (a little less than an almud) of cherries. Counts of three samples showed 581, 633, and 645 cherries per liter, respectively. Quillou. Goffea quillou was imported into Java from the French Congo by the Java Experiment Station, which later supplied the Porto Rico station with seed. Doctor Cramer 35 states that under favor- able circumstances G. quillou coffee produces more than other varieties, yielding in good years 2,500 pounds of merchantable coffee to the acre. Two plantings have been made at the Porto Rico station, the first of 47 trees from seed received and planted in October, 1914, and set in the field in September, 1915, and the second of 60 trees from seed planted in November and December, 1915, and set in the field in August, 1917. The development has been vigorous in good soil. The trees differ from those of Robusta in that they are more compact, * Coffees of the Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jour. i(l»18), TOd. 35, No. 5, ». 420. 24 BULLETIN 30, POETO RICO EXPERIMENT STATION. the leaves are narrower, and the internodes are shorter, giving to the whole a more densely foliated effect. (PL IX, fig. 1.) Representative leaves are 8-J to 9J inches long and from less than 3 to 3J inches wide, the length of some being more than three times the width. The average internodal length on laterals varies from 2 to 3 inches. The first planting produced a. few cherries in the fourth year from seed. The subsequent crops are given in Table 6 and the season of maturity in Table 1. C. quillou ripened slightly earlier than did G. robusta, but the fruiting habit of the two is similar. The cherries of G. quillou are a lighter red than those of 0. robust a, and some are solid colored while others are striped. They are also less rounded and slightly smaller. Counts of four samples showed between 640 and 726 cherries per liter. The skin is very thin and the yield of marketable coffee to the almud of cherries was the highest of any variety tested, both as to the amount of coffee and proportional reduction in weight. Two almuds of cherries weighed 29 pounds 9 ounces and 31 pounds 6 ounces, respectively. After being pulped and dried, the latter yielded 8 pounds of marketable coffee, a reduc- tion in weight in the ratio of 3.9 to 1. Two pounds were found to contain 3,128 and 3,159 beans, respectively. The bean is less rounded than that of G. robusta, being proportionally longer and in some cases rather pointed. Unless the closely adhering brown silver skin is thoroughly removed, the bean presents an unattractive appear- ance. Where the robustoid coffees are to be planted, this variety would seem worthy of trial. Coffea Uganda Hybrid. Seed of this hybrid was received from the Java Experiment Station, where this variety is considered the best of the robustoid coffees for the less favorable soils. From seed planted in November,. 1916, a small planting was set two years later. Both growth and production have been satisfactory, but the planting is small, and too young as yet to permit of the drawing of definite conclusions. The leaves are considerably smaller than those of Robusta, being 7 to 8 inches long by 3 to 3J inches wide. In shape they are generally oval, but there is some variation between individuals. The cherry is either striped or solid in color and is on a rather long pedicel* like that of the Arabian. The produce of one individual differs from that of another in both size and shape of bean. Coffea Congensis Hybrids. Doctor Cramer, who supplied the station with seed of G. congensis, does not include it in the Robusta group, stating in his letter of transmittal that " 0. canephora enters into the group of G. robusta. C. congensis and its varieties come nearer to G. arabica" He also says that " G. congensis var. chalotii is probably a hybrid of G. con- gensis with G. canephora." 3 * Inasmuch as the two plantings at the station, the first from seed received as G. congensis and the second as G. congensis var. chalotii, • ' i 1 ; — : ! 1 i i 88 Coffees of the Dutch East Indies. Dr. P. J. S. Cramer. Tea and Coffee Trade Jour. (1919), vol. 36, No. 2, p. 120. Bui. 30, Porto Rico Agr. Expt. Station. Plate IX. Fig. I. — Coffea quillou in Blossom, from Above. Fig. 2. — Coffea robusta in Fruit. Bui. 30, Porto Rico Agr. Expt. Station. PLATE X. 2 O CO CO O -l CO CO CO o < u LL Ll o o I oi o ll o oo o oo CN 6 z <2cr _ < - LU UJ rj- li UJ HI " d o £ o Ll Bui. 30, Porto Rico Agr. Expt. Station. Plate XI, s **^tffc SfifcKj ■ M0^i h - - -4 - f. , £ • p RV,^:,^> 1 V 1 / l PS • « v m . yf > •'■ $ 4 f • f. >• V. E2SI ■I - f. *jiS»: , >.''<> \ • V g * ■ 1 ' $ FIG. I. — COFFEA CANEPHORA BENT WITH HEAVY CROP. FIG. 2. — COFFEA CONGENSIS HYBRID IN FIRST PLANTING. COFFEE VABIETIES IN PORTO RICO. 25 have many points in common with the Robusta group, it is thought that they are made up chiefly of hybrids of C. congensis with mem- bers of the Robusta group, and as such they are included here. (PL XI, fig. 2.) Seed of both lots was received in December, 1915, and 30 trees of each were set in the field in August, 1917. In the first lot there is a wide variation between individuals. Detached leaves of some trees would readily pass as foliage of Arabian coffee, while those of others resemble the Robusta type more than the Arabian. In leaf dimensions, individuals may vary from 5 to 6 inches long by 2 to 2£ inches broad on branches having 2-inch internodes, to others 8 to 8-J inches long by 3 to 3^ inches broad at nodes more widely spaced. On the whole, the foliage may be con- sidered as intermediate in size between that of the two groups. The young branches point upward. The cherry is either striped or solid in color, flattened, and like the Arabian is borne on a longer pedicel than is true of Robusta. The second lot, C. congensis var. chalotii, resembles more closely the Robusta type, with larger leaves 7 to 9 inches long by 3 to 4 inches wide, and cherries that are borne on shorter pedicels. The yield of the two lots is given in Table 6. The first lot, in comparison with the general plantings of Robusta, has been fairly productive. C. congensis hybrids ripen in the same season as does the Robusta group, which they resemble in size and appearance of cherry, reduc- tion in weight, closely adhering brownish silver skin, and bean. Four counts showed between 612 and 666 cherries, with an average of 638 cherries, per liter. The skin of the cherries is thin, and the beans are the smallest of this group. An almud of fresh cherries weighed 29 pounds 1 ounce, which, after being dried, gave 7 pounds of cleaned coffee beans, a reduction in weight in the ratio of 4.2 to 1. SUMMARY. In both area under cultivation and value as an agricultural export, coffee holds second place in Porto Rico. Porto Rican coffee is of the Arabian type. It is highly esteemed in Europe, but is little known in the United States. It is com- paratively low in caffein content. No distinct varieties of coffee are recognized in Porto Rico, but that produced at the higher altitudes is considered better and sells at a higher price than does coffee grown nearer sea level. The experiment station has imported and tested numerous varieties from other coffee growing countries. Some of these are indistin- guishable from the Porto Rican, and the distinctive flavor for which they are known seems to be modified by the change in environment, while others differ essentially from the Porto Rican. The ripening season for the different varieties includes every month in the year. Some plantations suffer heavy losses through inability to secure sufficient coffee pickers at the season of heaviest ripening. The planting of varieties which do not mature their main crops simultaneously would help this labor shortage. If sun- drying is practiced, varieties maturing in the dry season offer an ad- vantage over those ripening in the season of heavy rains. Of the coffees tested, three members of the Arabian group — Columnaris, Maragogipe, and San Ram6n — mature considerably later than does 83926—24 3 26 BULLETIN 30, PORTO RICO EXPERIMENT STATION. the Porto Rican variety. These are followed by the robustoid group, and later still by the Liberian group, of which Excelsa appears to be the latest. The three groups in which the coffees of commercial interest have been placed were clearly divided as to cherry size, except in case of Maragogipe (large) and Excelsa (small). The Liberian group contains large cherries, and the robustoid group small cherries^ nearly twice as many of the latter as of Porto Rican being required to fill a given measure. The variation in weight of an almud of cherries was slight. The reduction in weight from cherry to marketable bean not only showed marked differences between the groups but a considerable range within the group. The Liberian group suffered the heaviest reduction, the ratio of weight of cherry to marketable bean ranging from 10.2 : 1 for Liberica to 7.2 : 1 for Excelsa, an almud of cherries giving 2| and 4 pounds, respectively, of marketable coffee. The ratio of reduction for the varieties in the Arabian group ranged from 6.6 : 1 for Padang to 5.1 : 1 for San Ramon, which was the only variety in this group showing as favorable a ratio as the Porto Rican. The robustoid group showed the least loss in weight and consequently the most favorable ratio of reduction of any of the groups, the ratio ranging from 4.4 : 1 for Robusta to 3.9 : 1 for Quillou, an almud of cherries of the latter giving 8 pounds of marketable coffee. The groups were not clearly divided as to bean weight. The Liberian group is made up of larger beans than the robustoid. In the Arabian group there is a wide range, from Mocha with the smallest bean to Maragogipe weighing three times as much, and nearly equaling in weight the bean of Liberica, which was th^ largest of those compared. In the Arabian group Maragogipe was first and Padang was second in bean size, with Columnaris and Erecta comparing favorably with the Porto Rican. The Arabian group ran lowest in peaberries, 6 to 17 per cent; the Liberian group was intermediate with 14 per cent for Liberica to 29 per cent for Excelsa ; and the robustoid group was highest, 18 to 41 per cent. As the coffee planter often sells his product in the parchment, the buyer making an arbitrary deduction of 18 to 20 per cent, according to its condition and appearance, in lieu of actual weight of parch- ment, a comparison of the varieties for reduction in weight on re- moval of the parchment was made. The loss for the Liberian group was very high, ranging in the samples examined from 24 to 33 per cent. The Arabian group showed a reduction of 15 to 20 per cent; the Porto Rican, 17 per cent; and the robustoid group showed the least reduction, 12 to 14 per cent. Of the Arabian coffees imported by the station, Padang, Column- .aris, Erecta, and Bourbon have shown themselves to be vigorous and productive under favorable conditions, and of good cup quality. In marketing, they could be mixed with the Porto Rican. A point against Bourbon, however, is the small size of its bean. The growth of Columnaris has been the most vigorous of any of the Arabian coffees. Maragogipe, notable for its large bean and distinctive flavor, has given too low a yield to warrant commercial plantings. Mocha, San Ramon, and Murta have given low individual yields, but as the trees of the last two varieties are small, the acre yield COFFEE VARIETIES IN PORTO RICO. 27 may be increased by closer planting than is advisable for the other coffees. While Mocha is of excellent cup quality and may be grown for home use, it should not be planted for commercial purposes. San Ramon is of very dwarf, stocky, and vigorous habit, and may be useful for planting in windy locations less suitable for varieties of taller, spindling growth. The bean is considerably smaller than that of Porto Rican. In sections suffering greatly from ravages of the leaf miner it may be advisable to plant coffees of the Liberian group, the foliage of which is less subject to leaf miner attacks, presumably on account of the leathery texture of the leaf. Excelsa seems to be the most promising of this group so far tested. Certain individuals of Dewevrei have shown themselves very prolific and give much promise. The Robust a and related varieties are planted in certain coffee- growing countries for their heavy production. While the average production for this station has not been high, the tests have been of shorter duration than for the Arabian varieties. The yield of cer- tain individuals has been high, indicating either heavy-yielding strains or heavy yields for plantings under favorable conditions. Notwithstanding the many points in its favor, Robusta coffee is inferior in cup quality, a fact which should be borne in mind by those contemplating planting it. ADDITIONAL COPIES OF THIS PUBLICATION MAY BE PROCURED FROM THE SUPERINTENDENT OF DOCUMENTS GOVERNMENT PRINTING OFFICE WASHINGTON, D. C. AT 10 CENTS PER COPY V /Vit; 7 I 0.