Ubc TllntversttB o( dbtcago 
 
 FOUNDED BY JOHN D. ROCKEFELLER 
 
 ) 1 
 
 ■> > > 1 
 
 > ■) > 
 
 •) j j ^ ^ ’ 
 
 4 '> 
 
 o 
 
 1 ® > ’ 
 
 » ^ * 
 
 ^ i ^ >» 
 
 » ) > 1 
 
 ) T y 
 
 
 o y 
 
 •) y 
 ^ yy y 
 
 A STUDY OF REDUCTION IN OENOTHERA 
 
 RUBRINERVIS 
 
 A DISSERTATION 
 
 SUBMITTED TO THE FACULTY OF THE OGDEN GRADUATE SCHOOL OF 
 SCIENCE IN CANDIDACY FOR THE DEGREE OF 
 DOCTOR OF PHILOSOPHY 
 
 (department of botany) 
 
 BY 
 
 REGINALD RUGGLES GATES 
 
 Reprinted from 
 
 Botanical Gazette, Vol. XLVI, No. i 
 Chicago, 1908 
 

 
 G- '7.'^ S 
 
 • • 
 • • • • 
 
 ^ # t. 
 
 • » 
 
 •• • 
 • • 
 
 »* ** • • c 
 
 ► • • • 
 • • * • » 
 
 ‘ « • 
 
 « « 
 • t 
 • • 
 
 ‘ • • 
 • • 
 
 • •• 
 
 
 Composed and Printed By 
 The University of Chicago Press 
 Chicago, Illinois, U. S. A. 
 

 VOLUME XLVI 
 
 NUMBER I 
 
 
 J- 
 
 j; 
 
 
 Botanical Gazette 
 
 u o ® i> o 
 >. ) 
 » 
 
 0 > 
 
 JULY jooA 
 
 ^ O » 
 
 > O 3 
 
 A STUDY OF REDUCTION'.IN .OEiVO?T?BI?T 
 
 RUBRINERVIS- 
 
 CONTRIBUTIONS FROM THE HULL BOTANICAL LABORATORY III 
 
 Reginald Ruggles Gates 
 
 (with plates i-iii) 
 
 The present contribution is a statement of some of the results 
 obtained in the cytological study of Oenothera Lamarckiana and its 
 mutants. Since these results have a more or less direct bearing on a 
 wide range of fact and theory in various fields, their full discussion is 
 reserved for a future time. The facts presented in this communica¬ 
 tion will be taken almost wholly from the study of O. rubrinervis, one 
 of the most vigorous of the mutants. Other papers will be presented 
 later, giving the further evidence upon which the conclusions of this 
 paper rest, and attempting to indicate their bearing on the general 
 problems of cytology and variation involved. 
 
 Material 
 
 The plants from which the material for these studies was obtained 
 were grown from pedigreed seeds of DeVries, the purity of these 
 cultures being further proven, in some cases, by carrying on the pedi¬ 
 gree for another generation before collections were made. The 
 results of these cultures, which are still being carried forward to later 
 generations, will be published at another time in connection with an 
 account of other studies on variation and hybridization in Oenothera. 
 In this way it is hoped, if possible, to correlate the cytological data 
 with the work in hybridization and variation. It seems to the writer 
 that only by thus combining cytological with experimental studies is an 
 explanation of the peculiar and remarkable phenomena of variation 
 exhibited by the Oenotheras to be reached. 
 
 I 
 
 I 
 
2 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 The cytological studies presented here will be confined chiefly to 
 the phenomena of synapsis and reduction in the pollen mother cell. 
 Various forms have been studied, a complete series of stages being 
 obtained ir^ some forms and a partial series or only a few stages being 
 examined in j6thei,;s:. / riie forms investigated include (i) O. rubri- 
 nerviSj (/>) ( 9 ;. Lamqrckianay(^) O. gigas^ (4) O. nanella^ (5) O. biennis 
 cruciate, ai^ariety of the Eufoijcan O. biennis, (6) both O. lata (see 12) 
 and O. Lamorekiana from the Fj of O. lataX O. Lamarckiana, and (7) 
 plants resembling Q'. ‘pgas, from the Fj of O. lataXO. gigas. Pre¬ 
 liminary reports have already been made upon some of these studies, 
 in various connections (ii, 12, 13, 14, 15). Reference will be made 
 to some of these results later. 
 
 The material from each individual was collected separately in 
 nearly all cases, in order to observe possible individual differences in 
 the same race^ either in the number of chromosomes or in other 
 cytological features. I am indebted to Mr. C. H. Shattuck for 
 making a number of these collections. The material for the study of 
 O. rubrinervis was obtained from a number of individuals grown in 
 two different seasons and representing several strains derived from the 
 same original pedigree. Sections were cut from six of these, and it 
 may be stated here that in O. rubrinervis no individual differences 
 were discovered, either in the number of chromosomes, which was 
 14 in all cases, or in any other features. In some of the other mutants, 
 also, a number of individuals were examined. It was found necessary 
 to reserve the account of O. gigas, which presents several features of 
 special interest, for a separate paper. A preliminary report on this 
 form has already been made (14, 15). 
 
 For various reasons, O. rubrinervis was chosen as the most favor¬ 
 able form for a thorough study of synapsis and reduction. The 
 nuclei and chromosomes of Oenothera are small, and for this reason 
 the selection of the most favorable type for study is a matter of some 
 importance. In O. rubrinervis the pollen mother cells, although 
 they vary much in size, are usually considerably larger than in 
 O. Lamarckiana, the nuclei being also proportionately larger. The 
 reason for this will be explained later. The chromosome number 
 being low in most of the forms (2^ = 14, x=j), they can be counted 
 without any difficulty, notwithstanding their small size. Another 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 3 
 
 notable advantage in comparing this with other studies in reduction 
 is in the shape of the chromosomes, which are globular or somewhat 
 oblong or cylindrical in most stages of mitosis, and are never greatly 
 elongated or looped. For this reason it is a comparatively easy 
 matter to obtain accurate counts of the chromosomes in the pollen 
 mother cells of any of the forms. This shape is also found to be very 
 advantageous in a study of the events of reduction following synapsis. 
 The appearances are clear and easily interpreted, in striking contrast 
 to the forms with long twisted chromosomes, such as have been made 
 the basis of many of the studies on reduction in plants. 
 
 On the other hand, the somatic nuclei and chromosomes are very 
 much smaller, and in metaphase the latter are elongated and looped, 
 making it impossible to count them with the same degree of accuracy. 
 Some of these appearances have already been described elsewhere 
 (12, p. 19). Thus while it was found that the chromosomes could 
 be counted almost equally well in pollen mother cells of all the 
 forms studied, O. ruhrinervis was found to be especially favorable 
 for the investigation of reduction phenomena, especially the events 
 of synapsis and the prophases of the heterotypic mitosis. The 
 account given here will refer throughout to O. ruhrinervis, with 
 occasional comparisons with other forms. Later papers will take 
 up these other forms in detail, in so far as this is necessary after the 
 account presented here. Special attention will be given at that time, 
 in particular, to the later stages, beginning with the telophase of 
 the heterotypic mitosis, and also to the interesting conditions in some 
 of the hybrids. The detailed account in O. ruhrinervis will not be 
 carried farther than the metaphase of the heterotypic mitosis, at 
 which time the essential events have already taken place. 
 
 Methods 
 
 The usual methods of cytological technique were employed, 
 various chrom-acetic and chrom-osmo-acetic solutions being tried 
 until satisfactory fixation was obtained. The thickness of the sec¬ 
 tions- varied from 4 to 10 /-t. The latter thickness was found most 
 favorable for counting chromosomes, because it is somewhat greater 
 than the diameter of the nuclei, many of which in such sections were 
 therefore uncut. It is possible to determine easily whether a nucleus 
 has been cut by the knife by examining it in low and high focus. The 
 
4 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 chromosomes in such uncut nuclei can then be counted with absolute 
 accuracy, either in the prophase of the heterotypic mitosis before the 
 disappearance of the nuclear membrane, or in the telophase after the 
 walls of the daughter nuclei are formed. In nearly every individual 
 examined, large numbers of such cases, all yielding the same result, 
 were counted before the number was finally determined upon. The 
 chromosomes could also be counted in certain positions on the 
 spindle, particularly in anaphases, but in metaphase they were usually 
 too closely grouped to allow of satisfactory counting. 
 
 In the second division, particularly in the forms having seven 
 chromosomes as the gametophytic number, the chromosomes could 
 be counted with certainty in almost any stage of mitosis. The thinner 
 sections were used chiefly in the study of spirem and synapsis stages, 
 although here also the comparatively short length of the thickened 
 spirem frequently made it advantageous to study uncut nuclei in 
 which the spirem could be followed throughout its length. 
 
 Of the various stains Heidenhain’s iron-hematoxylin was found 
 to be superior for chromosome counting and for clear differentiation 
 of chromatic structures in nearly all stages of synapsis and reduction, 
 safranin-gentian being used occasionally for comparison or for 
 differentiating particular cytoplasmic structures. Orange G was 
 also used with the iron-alum stain for bringing out clearly certain 
 special features, particularly the protoplasmic connections between 
 mother cells, which will be described later. 
 
 Description 
 
 EARLY STAGES 
 
 Some of the very early stages of the anthers, previous to the forma¬ 
 tion of mother cells, have been studied particularly with the purpose 
 of tracing the origin of the bodies which were called heterochromo¬ 
 somes in my first paper. The provisional use of the name was 
 based on the very close resemblance of these bodies to the chromo¬ 
 somes in appearance, and their frequent presence close by, or in some 
 cases apparently attached to, the heterotypic spindle. They were 
 not stated to pass undivided into one of the daughter nuclei, as 
 misquoted by Tischler (32), but to remain outside in the cytoplasm 
 where they gradually disappear. The study of their early history 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 5 
 
 shows that no line of distinction can be drawn between them and the 
 large body readily recognized as the nucleolus. They are then 
 smaller nucleoli, not differing essentially in origin from the single 
 larger body which is almost constantly present in the mother cell 
 during synapsis and prophase, but diverging from the latter some¬ 
 what in their later history. 
 
 In the earliest stages studied, the young meristematic cells of the 
 anther primordia are very small {figs. /, 2), and the tissues are wholly 
 undifferentiated, except the epidermal layer. Usually several smaller 
 nucleoli are present in each nucleus of the meristematic cells, in addi¬ 
 tion to the larger nucleolus. Compared with the cells of the anther 
 wall when they are no longer meristematic, the smaller nucleoli of 
 the former are about the size of the nucleoli of the latter, which are 
 approximately equal in size. There is nothing in the latter corre¬ 
 sponding to the larger nucleolus of the meristematic cells. Probably 
 afterward one of these nucleoli enlarges as the cell increases in size, 
 or it is possible that the nuclei of meristematic cells are always derived 
 from previous ones which already possess a large nucleolus. 
 
 ‘ Chromatic staining bodies are also found closely appressed to the 
 nuclear membrane in all the meristematic cells {figs, i, 2). This 
 tendency for chromatic material to accumulate on the nuclear walls 
 gives these nuclei a characteristic appearance. These bodies often 
 appear like a thickening of the membrane itself. 
 
 At the next stage studied many cell divisions have taken place, 
 and the sporogenous, tapetal, and wall tissues have been differentiated. 
 The sporogenous cells have increased enormously in size, and form 
 a single row in longitudinal section down the center of the anther, the 
 walls of'these cells being especially thickened and distinct {fg. fi). 
 The cells of the surrounding tapetal layer have also increased greatly 
 in size and are distinctly marked off from the sporogenous row. In 
 the sporogenous cells the nuclei {fg. 4), though much increased in 
 size, have not increased in proportion to the cytoplasm. The large 
 nucleolus, much larger than in the earlier stage, is now a conspicuous 
 object in the nucleus. Smaller nucleolar bodies are almost invariably 
 present, but masses are no longer found attached to the nuclear 
 membrane. (The characteristic masses, however, may remain for 
 some time attached to the nuclear walls of the tapetal cells). 
 
6 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 Figs, j-io are from drawings of other nuclei at this stage of develop¬ 
 ment. In the majority of cases one or two smaller nucleoli occur 
 in addition to a single large one, but rarely {fig. 6) two large nucleoli 
 of equal size may be found; and very frequently the number of small 
 bodies, of equal or unequal size, may-be greater, reaching as many as 
 five or six. Figs. 5, y, 8, g show these in various stages of fusion with 
 each other and with the large nucleolus.^ They are thus not in any 
 sense autonomous bodies. It appears that usually these fusions 
 take place until only one large nucleolus and one or two smaller ones 
 are present during synapsis and diakinesis. But occasionally the 
 fusions do not take place, and several of these bodies may then be 
 present in the later stages. The number of these nucleoli finally 
 present depends, then, largely upon the amount of fusion which has 
 previously taken place between them. In the later stages one large 
 nucleolus is almost invariably present and usually a smaller one 
 bearing a certain proportion to the larger in size, though the latter 
 may vary in size and number as already stated. There is usually 
 a clear area around the large nucleolus, as in the earlier stage, and 
 threads of the reticulum may or may not cross this and appear to be 
 attached to the nucleolus {fig. 4). The reticulum of the cytoplasm 
 usually stains rather more deeply at this time than that of the nucleus. 
 It may as well be stated at this time that in the resting nuclei of the 
 pollen tetrad and in the nuclei of the nearly mature pollen grains of 
 Oenothera one finds {fig. ii) the same condition of the nucleoli as 
 in the mother cells, namely, usually one large and one small nucleolus 
 bearing a rather definite size relation to each other, with sometimes 
 additional small ones. 
 
 The sporogenous rows are differentiated from the tapetum by 
 the greater growth of the cells, nuclei, and nucleoli of the former. 
 At the same time they are distinctly marked off by the formation of a 
 continuous thickened wall between tapetum and archesporium 
 {fig. j). It is obvious that as the cells and nuclei increase in size, 
 the nucleolus grows also. Up to the time of synapsis the mother cells 
 usually form a compact tissue, but about this time the cells begin to 
 
 I Miss Nichols (21) figures what are in all probability stages of fusion of large 
 and small nucleoli in Sarracenia pollen mother cells, but interprets them as a budding- 
 off of small bodies from the nucleolus. The budlike attachments to the nucleolus, 
 frequently observed by other authors are doubtless to be explained in like manner. 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 7 
 
 break apart at the corners where they meet, and before diakinesis is 
 reached they are completely rounded off and independent, or they 
 frequently remain connected with other mother cells only at the ends. 
 In the meantime the cavity of the loculus grows rapidly, so that the 
 mother cells, in normal development, usually lie loose in the cavity. 
 
 The events of synapsis and reduction usually go forward simul¬ 
 taneously throughout a flower, with comparatively little variation in 
 the different parts of the same loculus or in the different anthers of a 
 flower. In one flower, however, wide variation was found in different 
 anthers, but comparative constancy in the loculus. One anther of 
 this flower was in synapsis, another in diakinesis, another in meta¬ 
 phase of the heterotypic mitosis, and in still another some of the 
 mother cells had completed the second mitosis. No abnormalities 
 in the cytological condition of this flower were observed. 
 
 SYNAPSIS 
 
 After the stage described in fig. 4, the nucleus increases greatly in 
 size, but without an appreciable increase in the size of the cell. The 
 single row of sporogenous cells divides, so that there are usually two 
 rows of pollen mother cells. Occasionally three or more mother 
 cells are found in the cross-section of a loculus. In general there are 
 fewer divisions than in the other forms, and this is at least one of the 
 reasons why the mother cells are on the average larger than, for 
 example, in O. Lamarckiana. 
 
 The resting nucleus of the pollen mother cell increases in size and 
 begins to show signs of approaching synapsis. Figs. 12, ij, 14 
 show stages in the beginning of this process. A number of these 
 stages were found—although they are uncommon—^in the same 
 sections with regular synapsis stages. In some cases they occurred 
 side by side with mother cells in which the synaptic knot had already 
 been formed. A complete series of stages may be found in the same 
 section, from the beginning of contraction to the formation of a close 
 synaptic ball. The cytoplasm in these cells shows no contraction 
 whatever, but is perfectly fixed. For this and other reasons there 
 can be no doubt that this is a real contraction stage, leading to synap¬ 
 sis, and not a result of imperfect fixation, as one might judge on first 
 examination. 
 
 
8 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 That these nuclei are going into synapsis and not coming out is 
 shown by several features: (i) the extremely delicate character of the 
 threads, like those of the resting nucleus; (2) the fact that the 
 periphery of the reticulum as it contracts frequently preserves perfectly 
 the curved outline of the nuclear wall {jig. 12 ); (3) immediately after 
 synapsis the thread is somewhat shorter and thicker than previously 
 and appears to be continuous, while in the earlier contraction stages 
 we still have the appearance of a reticulum {jig. ij).^ As the con¬ 
 traction progresses, the threads are gradually rearranged from an 
 anastomosing reticulum to a very long and continuous delicate thread. 
 The exact manner of this rearrangement could not be observed, but 
 one finds many transitions {jig. 14) from the anastomosing reticulum 
 of the resting nucleus to the closely coiled and apparently continuous 
 spirem of the synaptic knot {jig. 15). The contraction may take 
 place from one side of the nucleus, leaving the reticulum attached for a 
 time to the nuclear membrane at one point {jig. ij), or it may take 
 place simultaneously from all sides {fig. 12). A few threads of the 
 reticulum usually remain attached for a time to the nuclear membrane 
 while the contraction is going on. These are drawn in finally as the 
 synaptic ball becomes more compact. 
 
 The small number of these intermediate stages found indicates 
 that they are passed through rather rapidly, the frequency of the occur¬ 
 rence of synapsis stages indicating, on the other hand, that this con¬ 
 dition is of considerable duration. 
 
 No indication of a doubling or pairing of the threads during these 
 intermediate contraction stages could be observed, though they were 
 carefully searched for. Moreover, in the earliest stages of the synaptic 
 ball the thread appears to be as thin and delicate as in the reticulum, 
 which does not favor the view that a pairing has taken place. The 
 evidence, then, £0 far as it goes, is decidedly not in favor of a pairing. 
 
 During these stages the nuclear membrane is often indistinct, 
 making it difficult to define accurately the limits of the nucleus. The 
 
 2 This explanation assumes, of course, that the synapsis stages themselves are 
 normal and not due to artifact, as I presume all cytologists will now agree, although 
 ScHAFFNER (29) apparently still entertains some doubt on the subject. The regularly 
 coiled arrangement of the thread in the synaptic ball appears to me to be one of the 
 best arguments against this stage being an artifact. Evidently a rearrangement of 
 the threads is going on as contraction proceeds. 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 9 
 
 same condition is observed during synapsis, which is found in the 
 same sections. In places the nuclear membrane has either disap¬ 
 peared or is too delicate to be observed. The cytoplasm, however, 
 retains the original outline of the nucleus. Mottier (20) has 
 apparently observed similar conditions of the nuclear membrane at 
 this time. In some cases it is ruptured and a portion of it is actually 
 carried inward with the nuclear reticulum at the beginning of the 
 contraction {-jig. 12). One is tempted to explain this as an artifact; 
 but that this is not the explanation is shown by the considerations 
 already mentioned. The explanation appears to be that as contrac¬ 
 tion proceeds a portion of the nuclear membrane may be torn away 
 and carried inward attached to the threads. Frequently in these 
 stages one finds the nuclear membrane present on one side of the 
 nucleus but invisible elsewhere. This is the case in jig. 12, although 
 the membrane was drawn as though complete. Observations of other 
 nuclei bear out this interpretation, the nuclear membrane being clearly 
 visible in some cases attached to portions of the reticulum which have 
 contracted far away from the original position of the nuclear wall. In 
 the late prophase, when the definitive chromosomes are formed, a 
 distinct and perfect nuclear membrane is invariably present, so it 
 would appear that in such cases as those just described a new mem¬ 
 brane is afterward formed. 
 
 Mention must now be made of the chromatic staining material of 
 the nucleus during these stages. The nucleolus is frequently, though 
 not always, included within the synaptic knot. There is a tendency 
 for other dark-staining bodies to accumulate near the periphery of 
 the nucleus {jigs. 12, 14 ); as contraction proceeds these are swept in 
 by the reticulum. The exact relation they bear to the threads is not 
 known. In some cases they appear, in the later stages of contraction 
 at least, to form a part of the threads themselves, in other cases they 
 appear to be merely inclusions in its coils. These bodies show no 
 constancy in number, size, or shape. As the spirem takes on the ap¬ 
 pearance of the synaptic knot, they are still found in its meshes, and 
 portions of the thread itself may also stain darkly, suggesting a 
 solution of a part of their substance and its transfer into the thread. 
 Even when the greater part of the spirem is completely decolorized 
 certain portions of it retain the stain. This appears to be partly 
 
lO 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 due to the denser aggregation of the spirem in these regions, but 
 in some cases it is evidently due to the presence of bodies which retain 
 the stain and appear to be giving up the stainable part of their sub¬ 
 stance to the spirem. These bodies are evidently not the prochromo¬ 
 somes found by Overton (22) in certain dicotyledons, nor are they 
 the gamosomes of Strasburger (30, 31). 
 
 Just the relation these bodies sustain to the spirem is not easy to 
 determine. From figs. 12 and 14 it is evident that they are at first 
 small “nucleoli” caught in the contracting reticulum, but quite 
 independent of it. Later they appear to give up a portion at least of 
 their material to the spirem, finally disappearing as independent 
 bodies. Usually, however, at least one of these bodies remains inde¬ 
 pendent, and appears in synapsis and diakinesis as a small nucleolus 
 bearing a definite relation to the size of the large nucleolus, being 
 about the size of a chromosome. These bodies are usually free in 
 the nuclear cavity {fig. ij). A certain depth of stain is required for 
 demonstrating them during synapsis, for they usually decolorize more 
 quickly than the large nucleolus. With a favorable stain they are 
 found to be of strikingly uniform occurrence at this time. A plasma 
 stain such as orange G may be used with advantage to demonstrate 
 their presence. The uniformity in their occurrence is so great that 
 for some time they were thought to be constant in size and number. 
 With the demonstration of their inconstancy and their origin we have 
 chosen to call them merely small nucleoli, as there appears to be no 
 sufficient reason for another name. The (large) nucleolus disappears 
 with great promptness immediately after the nuclear membrane breaks 
 down, only persisting for a time in a few rare instances. In no case 
 has fragmentation of the nucleolus, previous to its disappearance, been 
 observed, although the presence of deeply staining globular bodies 
 occasionally found near the periphery of the cytoplasm might be 
 accounted for in this way. The mass of the latter, however, is some- • 
 times greater than that of the nucleoli. The smaller nucleoli persist 
 and are frequently found close by the heterotypic spindle. They 
 may also be found on the homotypic spindle {fig. 41). Apparently 
 they never reenter a nucleus, but remain in the cytoplasm until they 
 finally disappear. These bodies have been found showing the same 
 behavior in all the forms studied. 
 
 % 
 
 4 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 II 
 
 POST-SYNAPTIC STAGES 
 
 Synapsis lasts for a comparatively long time, as shown by the fre¬ 
 quency of its occurrence in the material sectioned. During this time 
 the spirem shortens and thickens and then begins to arrange itself 
 more loosely in the nuclear cavity. This shortening and thickening 
 is progressive {jigs. 16-18) and apparently continues for some time. 
 During these stages the thickness of the spirem may be nearly uniform 
 throughout, or it may vary greatly, giving a moniliform appearance, or 
 the spirem may appear irregularly constricted at varying intervals. 
 In other cases, with a certain depth of stain it is seen to be composed 
 of lighter and darker areas more or less regularly alternating. Por¬ 
 tions of the thread may appear homogeneous or may show the lighter 
 and darker areas, according to the depth of stain {jig. if). In more 
 deeply stained nuclei, such as jig. 16, the thread appears homogeneous 
 throughout. These darker areas are the chromatin discs or chromo- 
 meres of various authors; and they give the thread a very character¬ 
 istic appearance. During this well-defined stage the greatly thickened 
 spirem is loosely distributed in the nuclear cavity. Deeply staining 
 bodies still appear attached to or enmeshed in the coils of the thread. 
 
 At this time one^ finds undoubted indications of parallel threads. 
 When represented by camera drawings in one plane the evidence for 
 this is not so satisfactory as in the original preparation, but there is no 
 doubt of their occurrence. As already stated, in the earlier stages 
 previous to and during synapsis, parallel threads could not be observed, 
 and it has not been determined whether they were really absent or 
 whether the failure to observe them was due to their extreme delicacy. 
 Hence it cannot now be stated whether they have arisen through an 
 approximation of spirems at an earlier period, or through a split in the 
 single continuous spirem. This matter will be discussed later. 
 
 Following this stage a second well-marked contraction takes place 
 {jigs. 18, 20, 21), apparently quite as typical and constant in its occur¬ 
 rence as the first contraction stage, which is ordinarily identified as 
 synapsis. This contraction is of much shorter duration, however, 
 and entirely different in appearance, owing to changes which the 
 thread has undergone since synapsis, resulting in a great amount of 
 shortening and thickening of the spirem. Mottier (20) has recog¬ 
 nized this second contraction stage in Podophyllum, Lilium, and 
 
12 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 Tradescantia, though he formerly thought it resulted from bad fixa¬ 
 tion ; and it appears to have been observed also by Farmer and Shove 
 (lo). Mottier states that in these forms there is but little shorten¬ 
 ing of the spirem between synapsis and segmentation into chromo¬ 
 somes. In Oenothera, on the contrary, as is evident from a 
 comparison of 'jigs. 15 or id with 22, a very considerable amount of 
 shortening as well as thickening of the spirem takes place during this 
 interval. During the second contraction the paired threads apparently 
 fuse, and further shortening of the (from now single) thread results in 
 an enormous amount of thickening of the spirem, so that when it uncoils 
 from this second contraction it has approximately the thickness of a 
 chromosome and exhibits only a few loops. It can then frequently 
 be traced through nearly its whole length. At this time there is a 
 great amount of variation in the thickness of different parts of the 
 spirem, as seen in jigs. 22 and 2j. Fig. ig is a portion of the spirem 
 at this period, drawn with a higher magnification. It shows the 
 chromatic bodies, which vary in size, imbedded in the linin substra¬ 
 tum. As to how far two different substances are represented, I am 
 at present unprepared to say. 
 
 DIAKINESIS 
 
 The single thick thread now segments transversely into 14 chromo¬ 
 somes, the sporophyte or 2X number. At this time there is no indica¬ 
 tion whatever of a longitudinal split in the thread.- Even when greatly 
 washed out, the material of the chromosomes appears perfectly 
 homogeneous, or if a granular structure is observable there is in its 
 arrangement no indication of the previous split. At the time of this 
 second contraction a pair of chromosomes is frequently observed 
 separated from the spirem and apparently always lying with their 
 long axes parallel and connected at one end {figs. 20, 22). This 
 condition occurs very commonly, although in other cases the spirem 
 is continuous throughout (jig. 21). In no case has more than one 
 pair of chromosomes been observed to be thus precociously cut off in 
 O. ruhrinervis, though two such pairs have been observed in O. lata 
 (see 11, jig. ig ). In no case has a single chromosome been observed 
 to be cut off in this manner, and apparently they are invariably cut 
 off in pairs, that is, bivalvent chromosomes are detached. 
 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 13 
 
 What significance this early separation of chromosome pairs may 
 have is not known, but it appears that the later history of these pairs 
 on the spindle can be traced. In the paper just cited (ii), the 
 writer wrongly identified them with the smaller nucleoli which persist 
 by the heterotypic spindle. These chromosome pairs are frequently 
 so closely approximated at the end opposite the end of actual connec¬ 
 tion as to give the appearance of a ring. It was thought that these 
 rings by condensation (which actually takes place) were reduced to 
 the size of these nucleolar bodies. The latter had the size and shape 
 of chromosomes, and with a certain depth of stain invariably appeared 
 hollow. These pairs are not condensed to rings, however, but to 
 chromosome pairs of the ordinary Oenothera type. 
 
 The spirem at this time varies greatly in thickness in different 
 parts, exhibiting constrictions and dilatations which indicate more 
 or less clearly where segmentation into chromosomes will take place. 
 This segmentation may happen while the spirem is still in the con¬ 
 tracted condition {jig. 25), or after it has again uncoiled and distrib¬ 
 uted itself in the nuclear cavity {jigs. 24, 26, 28), or before this 
 uncoiling is completed. The segmentation appears to be in some 
 cases nearly simultaneous {jig. 24 ); in other cases the segmentation is 
 successive, as in jig. 2j, where the spirem is clearly divided into three 
 portions and the constrictions for the formation of the chromosomes 
 are so far advanced that the number of chromosomes to be formed 
 by each segment can already be foretold with practical certainty. 
 The segmentation at this time is into 14 chromosomes, the sporophyte 
 number. • A large number of counts made at this time demonstrate 
 the absolute constancy of this number in all the individuals of O. 
 ruhrinervis examined. It is possible, however, that individuals of this 
 race may be found whose chromosome number differs from this 
 number by one. This matter will be discussed later. 
 
 In every single case where the count could be determined with 
 certainty it was shown to be 14. These counts were all made from 
 sections 10 //- thick, and from nuclei which were uncut by the knife. 
 The less numerous counts made in the multipolar stage of the hetero¬ 
 typic spindle gave invariably the same number. In this case all in a 
 given cell were obtained by examining the adjacent sections. In all, 
 hundreds of counts were made. In such nuclei as jigs. 26, 2g, jo, ji 
 
14 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 there can be no possible doubt of the number of chromosomes 
 present. 
 
 As already shown {fig. 20), one or in some cases more pairs of 
 chromosomes may be cut off from the spirem before it undergoes 
 segmentation, and frequently while it is still in the second contraction 
 period. The exact method of origin of these pairs has not been 
 observed, but they invariably, so far as observed, lie with their long 
 axes parallel and connected at one end, from which it would appear 
 that they were successive chromosomes on the spirem. In later 
 stages, when the spirem has constricted into a chain of chromosomes 
 arranged near the periphery of the nucleus, one or more pairs of 
 chromosomes are found separated from the rest. Some of these have 
 doubtless had the origin shown in fig. 20. Others appear to have 
 originated later, as indicated in some of the figures, by successive 
 chromosomes on the chain swinging around parallel to each other and 
 thus pairing. Usually in diakinesis one or two such pairs are found, 
 though occasionally there is no evidence of pairing. The highest 
 number of pairs observed at this stage was five, with indications of 
 pairing among the others {fig. 2g ); which is unusual. Later, in the 
 multipolar spindle stage two distinct pairs are usually found in vary¬ 
 ing stages of conjugation {figs, jy, jd). A single case was observed 
 {fg. J7) in which the fourteen chromosomes were all paired. 
 
 As the figures indicate, constriction of the spirem at regular inter¬ 
 vals proceeds progressively until a chain of chromosomes is formed. 
 When this has taken place, the chromosomes are at first several 
 times longer than broad, and their margins have a very irregular, 
 sinuous outline, like that of the spirem just previous to segmentation. 
 They are not so long, however, that they can be twisted and looped 
 in the confusing manner of many heterotypic chromosomes of plants. 
 This is very gratifying in the study of these stages, since it permits 
 a clearness of interpretation which would otherwise be unattainable. 
 Figs. 22 and show the beginning of contraction, which has pro¬ 
 ceeded farther in fg. 24, leaving only the so-called linin connection 
 between the chromosomes. The constrictions are all equivalent and 
 the spirem thus segments into the sporophyte number of chromosomes 
 and not into the reduced number of chromosome pairs. If successive 
 chromosomes on the spirem are really the members of a pair, there is 
 
1908} 
 
 GATES—REDUCTION IN OENOTHERA 
 
 15 
 
 nothing in the manner of segmentation of the spirem to indicate this. 
 However, it is clear enough that one chromosome frequently swings 
 around, as already mentioned, and pairs with its neighbor on the 
 spirem. We do not really have, then, a transverse division of chromo¬ 
 some bivalents, but a separation of whole (somatic) chromosomes. 
 Nothing has been found in the earlier stages which would correspond 
 to the gamosomes and zygosomes of Strasburger, and even should a 
 pairing of parallel threads during synapsis occur (a possibility which 
 will be discussed later), the final pairing is between chromosome 
 bodies which were lying end to end on a single spirem thread. 
 
 The linin connections during diakinesis appear to be merely the 
 more finely drawn out portion of the spirem between the chromosomes. 
 As condensation and contraction of the chromosomes progress, 
 these linin connections become longer and more delicate {jigs, ji, jj). 
 The chromosomes become more dense and compact, being at first 
 oblong-cylindrical {figs. 24, 26) and then more nearly globular or 
 pear-shaped {fig. ji). Certain chromosomes sometimes undergo this 
 contraction more quickly than others, as in fig. 2g, and the different 
 stages of this condensation may occasionally all be found in the same 
 nucleus. In other cases the globular appearance is due to the 
 position in which certain chromosomes happen to be lying {fig. J4). 
 
 HETEROTYPIC MITOSIS 
 
 During the prophase stages last outlined the cytoplasm usually pos¬ 
 sesses a more or less obscurely radiate appearance. A felt-work 
 of fibrillae finally appears around the nuclear membrane. Later 
 these fibrillae come to run tangentially to the latter, terminating in the 
 
 cytoplasm, and by their aggregation in certain regions the multipolar 
 
 _ • 
 
 spind'e is formed. From this stage the fibers are rearranged to form 
 the bipolar spindle, passing through conditions in which the spindle 
 appears quadripolar or tripolar in section. In the meantime the 
 nuclear membrane has dissolved and the chromosomes are found 
 at first in a cavity surrounded by fibers which preserve the outline of 
 the nuclear wall. Later they come in and become attached to the 
 chromosomes. Usually the large nucleolus has vanished before this 
 time, but occasionally it may still be seen {fig. J5). In fig. jy the 
 small nucleolus is shown, which can very frequently be seen at this 
 
i6 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 time. Figs. j6 and j/ are merely sketches of the spindle fibers to 
 indicate their general direction. Fig. J5 is an unusual case. A cone 
 of fibers appears to have been formed on one side only of the nucleus. 
 The fibers are coming in and finding attachment to the chromo¬ 
 somes. The large nucleolus is still present, as well as two smaller ones. 
 
 The most critical stages of reduction have now been described and 
 the remaining stages will be taken up with less detail at this time, but 
 will be presented in full in a later paper. The chromosomes are at 
 first irregularly arranged on the heterotypic spindle. As already 
 seen, during spindle formation many of the chromosomes are fre¬ 
 quently separate and unpaired. The attraction between the chromo¬ 
 somes which leads to pairing is evidently weak, so that it is doubtful 
 if any pairing takes place at metaphase between chromosomes 
 which had not previously paired. On the other hand, chromosomes 
 which have once paired, no matter how early, appear to remain 
 together until their separation in the metaphase of the heterotypic 
 mitosis. Hence probably in many cases the chromosomes pass to the 
 poles of the heterotypic spindle without having previously paired with 
 each other, that is, they were merely lying loosely in the equatorial 
 region of the spindle in metaphase, so that it was largely a matter of 
 chance which pole any particular chromosome went to. This is 
 believed to be a matter of prime importance in determining the final 
 result of the reduction divisions in Oenothera, and the nature of the 
 distribution of chromatin elements which takes place. Its possible 
 significance will be pointed out in the discussion. Fig. shows the 
 chromosomes just being drawn into the equatorial plate of the hetero¬ 
 typic spindle. In the examination of thousands of spindles in about 
 this stage, one usually finds the chromosomes spread out in several 
 planes along the long axis of the spindle. Of course some of these are 
 early anaphase stages in which the chromosomes have begun their 
 journey to the poles, but the condition is seldom found where the 
 chromosomes are arranged regularly in pairs on the spindle. The 
 daughter chromosomes seldom advance toward the pole in a single 
 plane, as is the case in so many forms, but are more or less irregularly 
 strung out along the spindle in their passage to the poles. This is in 
 striking contrast with their behavior in the homotypic mitosis. 
 
 Usually in the early anaphase of the heterotypic mitosis a longi- 
 
 ( 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 17 
 
 tudinal split appears in the daughter chromosomes. This split does 
 not stop short of one end, giving a V-shaped body as>n niany plant 
 chromosomes, but usually passes right jthro'agt.L.JoriTiing two inde- 
 pendent bodies, which, however, remain paired’m fhc’telophase, and 
 occupy a great variety of positions in regard, ,to -^ach other. The 
 homotypic chromosomes thus assume rnany ot the., characteristic 
 shapes which are usually observed in the heterpvypic chromosomes of 
 other forms, such as X, Y, V, H, etc. The failure of the heterotypic 
 bivalents to form these shapes is due partly to the weaker attraction 
 between the members of a pair, but largely to a difference in their 
 shape, each member of a pair being usually more rounded in the 
 heterotypic and more elongated and rodlike during the stages between 
 the two mitoses. 
 
 The telophase of the heterotypic mitosis is one of the best stages for 
 counting the chromosomes, as they are distributed at equal intervals 
 around the periphery of the nucleus, no two ever being in contact and 
 the halves of each (bivalent) chromosome rarely separating. The 
 chromosomes now evidently repel each other, while the halves of 
 each chromosome attract each other rather strongly. The halves of 
 these bivalent chromosomes are usually short rods, but they may be 
 dumb-bell or hour-glass shaped, or nearly globular, as previously 
 mentioned (12). Sometimes, however, this split fails completely to 
 occur in the anaphase, the daughter chromosomes remaining single 
 and globular or somewhat elongated {fig. jp). These telophase 
 stages and the prophases of the homotypic mitosis will be taken up 
 in detail in a paper dealing with different forms. These results, 
 therefore, will not be duplicated here, but a brief statement of the 
 events of the second mitosis will be given. 
 
 HOMOTYPIC MITOSIS 
 
 In the telophase of the heterotypic mitosis the nuclei never pass 
 into the resting condition and the chromosomes never lose their 
 identity completely, though they spread out and anastomose with 
 each other more or less. Nucleoli are formed, as previously described 
 (ii). These stages between the two mitoses last for some time, but 
 the events of the second mitosis are passed through very quickly. The 
 two homotypic spindles are formed simultaneously and their axes are 
 
i8 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 at various angles to each other. Spindle formation is the same as for 
 the heterotypic mitosis, except that the spindles, are smaller. In 
 regard to the chi^omatln,’Suffice it at present to say that the chromo¬ 
 somes-of ^he hOmotypic' prophase show the same general types and 
 are often identical in' appearance with those of the heterotypic telo¬ 
 phase. ' ■ ''Ki'e’recan. be n'o'dOubt that the bivalent bodies which appear 
 on the homotypic. spindle" are the same bodies that were present in the 
 telophase of the heterotypic. Fig. 41 shows an early anaphase of the 
 second mitosis, the members of each pair having just separated. One 
 of the small nucleoli appears by one of the spindles. 
 
 IRREGULARITIES 
 
 In fig. jg spindle fibers are seen in the cytoplasm by the side of the 
 spindle in anaphase. This may be connected with a condition which 
 is illustrated in fig. 40. Six such cases were observed in which a 
 regular spindle occurred at the side of the mother cell instead of 
 between the daughter nuclei, after the partial or complete disappear¬ 
 ance of the heterotypic spindle. Some of these cases were in the telo¬ 
 phase of the heterotypic spindle {fig. 40 ); others were in the prophase 
 of the homotypic. In these cases the spindles were regularly formed 
 and rather sharp-pointed and occupied the same position at the side 
 of the cell; of course they contained no chromosomes. The method 
 of their origin is unknown, but it seems probable that they are con¬ 
 nected with the condition observed in fig. jg. Mother cells which 
 probably indicate an intermediate condition, in which irregularly 
 arranged fibers were found at the side of the cell, were occasionally 
 observed. They may merely indicate a persistence of the kinoplasm 
 of the heterotypic spindle after its function has ceased, but their 
 structure appeared remarkably definite in most of the cases observed. 
 
 A single case of extra nuclei in the pollen tetrad was observed in 
 O. rubrinervis. These have been previously described in O. lata 
 (ii), where they are common occurrences in connection with pollen 
 degeneration. The single case observed in O. rubrinervis is sketched 
 in fig. 42. ' Two small nuclei are present in addition to the four 
 larger ones composing the tetrad. The nuclei had passed too far 
 into the resting condition to count the chromosomes in each 
 nucleus. 
 
1908] 
 
 CjATES—REDUCTION IN OENOTHERA 
 
 19 
 
 POLLEN DEGENERATION 
 
 The general question of pollen degeneration in Oenothera is an 
 interesting one. It reaches its extreme expression in O. lata, which 
 is usually completely sterile in this regard, and in which I have 
 already shown (ii) that irregularities occur during the reduction 
 divisions similar to those found in sterile hybrids. The question of 
 sterility is evidently, as Tischler (32) suggests, a relative one. 
 
 In O. rubrinervis one is led from a gross examination to judge 
 that the pollen production is copious and probably equal to that of 
 O. Lamarckiana itself, but in reality many of the pollen mother cells 
 fail to complete their divisions. From an examination of sections of 
 anthers of O. rubrinervis it is found that in some loculi a large number 
 or perhaps nearly all the mother cells may be degenerating in the 
 synapsis stage. Frequently the cells are flattened and distorted, 
 appearing pressed together for lack of space in the loculus. The 
 chromatic contents of such cells often form a dense irregular mass, or 
 their nuclei may be in normal synapsis or mitosis, notwithstanding the 
 distorted shape of the cell; while still other cells of the same loculus 
 may be entirely normal. Even earlier, in the archesporial stage, the 
 tapetal cells in many sections were found to be breaking down, as in 
 O. lata (ii). No indications of degeneration have yet been observed 
 in mother cells of O. Lamarckiana, and very few in the tapetum. 
 
 The percentage of mother cells which thus degenerate in O. 
 rubrinervis was not determined. Tischler (32) suggests that the 
 causes of sterility in mutants are the same as those in hybrids and in 
 plants under cultivation. This general cause he designates as a 
 disturbance or derangement of the constitution of the idioplasm, which 
 he thinks has taken place in the production of mutants as well as in 
 hybrids and under the conditions of cultivation. 
 
 PROTOPLASMIC CONNECTIONS 
 
 It is an interesting fact that large and rather conspicuous proto¬ 
 plasmic connections are found between the pollen mother cells in 
 O. rubrinervis. They are usually quite easily seen and it is probable 
 that they are always present. They consist of delicate strings or 
 threads of cytoplasm connecting adjacent mother cells. In size they 
 vary greatly, from the delicacy of a spindle fiber to a coarse thread or 
 
20 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 strand connecting the cells {figs. 45, 46). When the cytoplasm has 
 shrunken slightly away from the cell wall they are particularly clearly 
 observable. These connections appear to be in all cases between 
 mother cells, and in no case have they been observed between the 
 mother cells and the tapetum. Generally one such strand is seen 
 connecting two cells, but not infrequently there are two or three or 
 occasionally even more. There is no constriction or change in the 
 nature of the connective as it passes through the cell wall. These 
 connections are even larger and more conspicuous in O. gigas, where 
 the mother cells are also much larger. They have not been observed 
 in O. Lamarckiana or the other forms, but they doubtless occur in all, 
 being probably smaller and more inconspicuous in some. 
 
 Discussion 
 
 The method of reduction described in this paper at once raises a 
 number of questions of prime importance from the cytological stand¬ 
 point, as well as from that of the relation subsisting between heredi¬ 
 tary and cytological phenomena. A discussion of all these features 
 will not be attempted at this time, the intention of the writer being 
 merely to indicate the general directions in which the facts point and 
 the possible bearing which these data may be found to have on the 
 problems connected with the phenomena of mutation in Oenothera. 
 A fuller discussion of these subjects is reserved for a future time, after 
 the presentation of further data. In the present communication 
 reference will be made only to the most recent papers on reduction in 
 plants, the purpose not being a review of the literature, or a dis¬ 
 cussion of present views, except in so far as they bear directly on the 
 matter in hand. 
 
 The recent accounts of reduction in plants, given by Berghs 
 ( 3 > 4 > 5 » 6), Gregoire (16), Strasburger (31), Allen (i, 2), 
 Miyake (18), Overton (22), Rosenberg (25), Yamanouchi (33), 
 and others, have agreed in so far as the following general course of 
 events is concerned: In synapsis a pairing of homologous maternal 
 and paternal elements occurs either in the form of gamosomes (Stras¬ 
 burger and Miyake), prochromosomes (Overton), or parallel 
 threads (Allen, Rosenberg, Gregoire, Berghs, Cardiff 7, and 
 Yamanouchi). In every case two parallel threads result, which unite 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 21 
 
 more or less intimately about the time of synapsis or later. After the 
 events of synapsis, a longitudinal split reappears in the thickened 
 spirem threads, this split representing the line of approximation of the 
 two original spirems. Transverse segmentation into pairs of chro¬ 
 mosomes, which are believed to be homologous somatic chromosomes 
 of maternal and paternal origin, then takes place. The halves of these 
 bivalent chromosomes, which lie side by side, are then distributed 
 in the heterotypic mitosis, which is thus a reduction division. In 
 the anaphase of the heterotypic mitosis a longitudinal split appears in 
 the daughter chromosomes, which is regarded as a premature split 
 for the homotypic mitosis, the latter being thus an equation division. 
 The persistency with which this general account has been given, not¬ 
 withstanding differences in detail, particularly preceding and during 
 synapsis, leads the writer to the belief that it is probably correct in its 
 main outlines, at least in many of the forms described. This being 
 judged to be the case, every effort was made to bring the account in 
 Oenothera into harmony with this general course of events but with¬ 
 out success, for Oenothera is found to deviate in some important 
 particulars, as is already evident from the description. 
 
 Another general account of reduction in plants, which was adhered 
 to by Strasburger as late as 1904 (30), and has been held notably 
 by Farmer and Moore (8, 9), Farmer and Shove (10), Schaffner 
 (28), Mottier (19, 20), and others, to mention only a few of the recent 
 papers, is in general as follows: The split in the spirem which occurs 
 at about the time of synapsis is a true split, such as may occur in the 
 prophase of somatic mitoses, and is not preceded by a pairing of 
 parallel threads, but the thread is single from the beginning. This 
 split afterward closes up as the thread shortens and thickens after 
 synapsis, and the single spirem so formed segments usually into the 
 reduced number of chromosomes, which are thus arranged successively 
 end to end. Each such bivalent chromosome thus consists of two 
 halves arranged end to end, not side by side, and the heterotypic 
 mitosis thus separates successive whole chromosomes on the spirem, 
 being therefore, as in the other account, a reduction division. The 
 split which appears in the anaphase of this mitosis is interpreted as a 
 reappearance of the earlier longitudinal split of the spirem. The 
 homotypic mitosis is therefore an equation or longitudinal division. 
 
 t 
 
22 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 There are of course minor differences in these accounts, Schaffner 
 (28) stating, for example, that in Lilium tigrinum there is a splitting 
 of granules in the spirem, but the linin thread remains single. Differ¬ 
 ences of opinion are also expressed regarding the arrangement of the 
 loops of the spirem before segmentation, and their relation to the 
 chromosomes formed. 
 
 These two general schemes agree that the heterotypic mitosis is 
 a reduction division separating whole somatic chromosomes, while 
 the second division is longitudinal. The essence of the distinction is 
 that the first view regards the chromosome bivalents as formed by a 
 side-by-side union of homologous chromosomes through the medium 
 of parallel threads, while the second view holds to an end-to-end 
 union. It will be seen that, omitting the points which are left undeter¬ 
 mined, the account in Oenothera corresponds more nearly with the 
 latter scheme than with the former, though differing in some respects 
 from both. Rosenberg (25), from a comparison of forms having 
 long and short chromosomes, has attempted to harmonize the latter 
 view with the former. He examined Listera, Tanacetum, Drosera, 
 and Arum, and found that, for example in Drosera, which has short 
 definitive chromosomes much like those of Oenothera, the spirem 
 first segmented into long twisted chromosomes lying in pairs with 
 their long axes parallel. Later, as they condensed into the short, 
 rounded definitive chromosomes, they frequently swung around end 
 to end, so that an observer seeing only the later stage would conclude 
 that they had been arranged tandem on the spirem at the time of their 
 origin. Similar conditions were sometimes observed in Listera. 
 I think my figs. 22-28 make it evident that this explanation will not 
 apply to Oenothera. The chromosomes in Oenothera do not undergo 
 any such great amount of condensation, but are already thick, heavy 
 bodies when first formed from segmentation of the spirem {fig. 24). 
 Their diameter at this time is about the same as that of the spirem 
 just previous to segmentation, as is shown by comparing figs. 22 and 
 
 with figs. 24 and 26. The fact that as many as eight or more 
 chromosomes may be found forming a single connected chain {fig. 26) 
 also renders this explanation impossible. 
 
 Miyake (18) finds that after the pairing of elements in synapsis 
 (the exact method of this pairing need not be entered into here) in 
 
1908] 
 
 GATES—REDUCTION IN OENOTHERA 
 
 23 
 
 Galtonia and Tradescantia, a longitudinal split appears in the thick¬ 
 ened thread, and the double spirem thus formed breaks transversely 
 into the reduced number of chromosome pairs. Later, in these 
 forms, a secondary union between the chromosomes is claimed to take 
 place, forming a single connected chain of chromosomes (as in Oeno¬ 
 thera). Sometimes a pair of chromosomes lies free by itself at this 
 time. Then by further shortening the chromosomes of Galtonia 
 again fall apart into pairs, though in Tradescantia they frequently 
 remain connected even after spindle formation. The apparent 
 similarity of the chromosome chain thus described by Miyake in 
 Galtonia to the condition in Oenothera, led the writer to make an 
 endeavor to harmonize the two accounts. But instead of this, all 
 the evidence obtained from a critical study of the stages concerned 
 shows that in Oenothera a single very thick spirem breaks transversely 
 into the sporophyte number of chromosomes. A critical examination 
 of jigs. 22-28 will make it clear, I think, that we are following the 
 progressive segmentation of a single spirem, and there is no room 
 for stages between, in which a double spirem breaks into two parallel 
 series of chromosomes. Moreover, it is hardly likely that secondary 
 fusions between chromosomes would take place to such an extent as is 
 shown in jigs. and 24. In nuclei such as 'jig. 20, in which a pair of 
 chromosomes is cut off prematurely from the spirem while still in the 
 second contraction, they are invariably connected at one end and 
 rarely, if ever, at the other (though sometimes the close approximation 
 of the latter ends may give the false appearance of a ring). This 
 would not be the case if they came from separate paired threads 
 merely lying side by side, so that this connection shows them to have 
 been really successive on the spirem. From this evidence the writer 
 cannot see how anything except a distortion of the facts can lead to the 
 assumption in Oenothera of two parallel threads breaking into chro¬ 
 mosomes. Hence the conclusion is that the double threads appear¬ 
 ing in the stage represented by jig. 17 have united to form a single 
 thread, which then breaks transversely into the sporophyte number 
 of chromosomes. 
 
 This corresponds fairly well with Strasburger’s 1904 (30) account 
 of the post-synaptic stages in Galtonia, and suggests to the writer 
 that perhaps after all the earlier account may be nearer the facts. 
 
24 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 SO far as the points here under discussion are concerned, than the 
 paper of 1905 (31). The close similarity of the conditions in Galtonia 
 and Tradescantia during diakinesis to those in Oenothera suggests 
 that they may be found finally to conform to Oenothera in these later 
 stages. Whether or not this will be found to be the case, we must 
 conclude that in Oenothera the longitudinal fission in the spirem 
 (however it originated) closes up, and that after the second contrac¬ 
 tion, or during it, the thick thread segments into the sporophyte 
 number of chromosomes. Since this diverges in important respects 
 from nearly all the recent accounts of reduction in plants, the con¬ 
 clusion is that reduction probably takes place differently in different 
 plants. Whether or not the results are different from the standpoint 
 of a qualitative distribution will not be discussed now. The writer 
 believes the above conclusions to be necessary, despite the fact that 
 authors have reached different conclusions in regard to the same 
 plant, particularly in such cases as Lilium and Podophyllum. 
 
 The next important point which requires discussion and which 
 was left undecided in the statement of observations, is in regard to 
 whether the double thread observed after synapsis arises from an 
 approximation of parallel filaments or through a primary split in the 
 thread. It may be well to examine the results which follow from 
 either assumption. The writer hopes later to determine more defi¬ 
 nitely this difficult matter. On the first assumption of a lateral 
 approximation in synapsis of two spirems representing respectively 
 the maternal and paternal chromosomes, we should expect the double 
 thread so formed to segment into the reduced number of chromosome 
 pairs, in order to conform to the current account in forms in which 
 there is a pairing of spirems, for example Allen (i), Gregoire (16), 
 and Yamanouchi (33). Instead, however, the spirem segments 
 into the unreduced number of bodies. We may still assume that 
 each of these bodies consists of maternal and paternal longitudinal 
 halves still closely held together and resulting from a previous approxi¬ 
 mation. According to this view the first mitosis would separate 
 bodies which were arranged successively on the spirem, while the 
 second mitosis would separate the maternal and paternal halves of 
 these bodies. The reason for such a result would be that the maternal 
 and paternal spirems remained closely fused after pairing, so that 
 
1908] 
 
 GATES—REDUCTION IN OF.NOTHERA 
 
 25 
 
 their elements were, separated in the second mitosis instead of the 
 first. This view is scarcely admissible for several reasons. In the 
 first place, on this hypothesis transverse segmentation of the spirem 
 must have taken place not only between the (bivalent) chromosomes 
 but also in the middle of each chromosome, in order to give a chain 
 of fourteen bodies. Such a segmentation seems unlikely. Another 
 possible explanation would be that the chromosomes have lost their 
 identity during synapsis, and that the bodies we are dealing with now 
 are new arrangements of the chromatic material, irrespective of the 
 somatic chromosomes. Many considerations, however, strongly sup¬ 
 port the belief that these bodies really represent the somatic chro¬ 
 mosomes. The facts so far educed in Oenothera, in the opinion of 
 the writer, all favor the hypothesis of the separate existence and 
 genetic continuity of the chromosomes from one generation to another. 
 In this connection may be cited certain plants from the Fi of O. lata 
 XO. gigcLS, which as stated elsewhere (14) have 21 chromosomes as 
 somatic number, 10 of which regularly go to one pole of the hetero¬ 
 typic spindle and ii to the other. Occasionally, however, the segre¬ 
 gated numbers of chromosomes are 12 and 9, one chromosome having 
 gone to the wrong pole of the spindle. In this hybrid 7 of the chro¬ 
 mosomes are maternal and 14 paternal. If in this case there were a 
 pairing of maternal and paternal spirems, it is difficult to see how it 
 could be accomplished and result in the distribution of chromosomes 
 in the heterotypic mitosis already stated. 
 
 It will be instructive to compare the chromosome history in this 
 cross with the often-quoted condition found by Rosenberg (23, 24) 
 in Drosera longijoliaXD. rotundifolia. D. rotundijolia has 10 chro¬ 
 mosomes and D. longifolia 20, as the gametophyte number. The 
 hybrid naturally has 30 chromosomes in its sporophyte tissues, 
 but in diakinesis 20 chromosome bodies appear, 10 of which are 
 double, consisting of a larger and a smaller half, while the remaining 
 10 are the unpaired (smaller) longifolia chromosomes. The larger 
 and smaller halves of the 10 bivalents separate and pass regularly 
 to the poles of the heterotypic spindle, but the unpaired chromosomes 
 are irregularly distributed or left out of the daughter nuclei. Later 
 the pollen deteriorates. This result is strikingly different from that 
 in the Oenothera hybrid, and, while perfectly in harmony with the 
 
26 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 idea of the pairing of threads in synapsis in Drosera, makes it highly 
 probable, and in fact necessary, that the method of reduction in the 
 Oenothera hybrid be different. This is a strong argument not 
 only against pairing of maternal and paternal spirems in Oenothera, 
 but in favor of the probability that reduction takes place in diverse 
 ways in the two genera. A considerable amount of time has already 
 been devoted to the study of reduction in this Oenothera hybrid, and 
 an account will be published later. So far as observed it shows no 
 differences in method from the account given here for the pure races. 
 
 The hypothesis of the pairing of parental spirems in synapsis in 
 Oenothera being thus rejected, the other alternative remains, namely, 
 that the double spirem results from a split; and this appears to satisfy 
 all the facts. The observations have already shown that the spirem 
 segments into a single chain of chromosomes. The description of 
 events in Oenothera from synapsis on thus agrees in outline with the 
 1904 account of Strasburger (30) in Galtonia, and in general also 
 with that of Farmer and Moore (9) in Lilium, Osmunda, Psilotum, 
 and Aneura, Farmer and Shove (10) in Tradescantia, and Mottier 
 (19, 20) in Lilium, Podophyllum, and Tradescantia. The belief of 
 the writer is that some of these forms will be found to correspond 
 more nearly with the account which involves a pairing of threads, and 
 some with the account involving only a split. 
 
 Another important matter which requires mention at this time 
 is the nature of the chromosome distribution which takes place on the 
 heterotypic spindle in Oenothera. As already observed, the chromo¬ 
 somes even during spindle formation are frequently unpaired. This 
 appears to be due to the weakness of the mutual attraction which 
 ordinarily leads to pairing. Granting that homologous maternal and 
 paternal chromosomes unite when pairing takes place, what are the 
 possibilities regarding the unpaired chromosomes ? Pairing insures 
 ordinarily that the members of the pair will proceed to opposite 
 poles of the spindle, and hence that the homologous maternal and 
 paternal elements will enter different nuclei. There is no such 
 certainty in the distribution of the unpaired chromosomes, so that it 
 might be expected that in certain cases both members of a pair would 
 enter the same daughter nucleus. It is important to note that this 
 result is entirely independent of the origin of these chromosome 
 
1908] 
 
 CiATES—REDUCTION IN OENOTHERA 
 
 27 
 
 pairs, whether from an end-to-end or side-by-side union of somatic 
 chromosomes, or in any other manner, so that this question holds no 
 necessary relation to the method of reduction. On the common 
 cytological assumption that the chromosomes are qualitatively 
 different (which has apparently been shown to be a fact in certain 
 well-known cases in animals, that need not* be cited), germ cells 
 would occasionally arise lacking both members of a pair^ and hence 
 lacking the possibility of developing certain qualities. In this manner 
 it is conceivable that a series of types might arise from the parent 
 O. Laniarckiana, each lacking the possibility of developing a certain 
 group of characters possessed by O. Lamarckiana. 
 
 On this view, which is suggested merely as a tentative hypothesis, 
 we would have in the mutations of O. Lamarckiana an analytical 
 process in which a series of types arises from the parent form, each 
 lacking in a different group of qualities or capacities which the parent 
 form possessed. This does not apply to O. gigas, however, which will 
 be taken up at another time. The further bearings of this hypothesis 
 on the mutation theory of DeVries will not be followed up in this 
 discussion, but it may be pointed out here that such a hypothesis 
 accounts for the absence of reversions of the mutants to O. Lamarcki¬ 
 ana, and it may also account for some of the peculiarities of 
 hybridization among the Oenothera mutants. I should therefore sug¬ 
 gest that there may be a relation between the type of reduction in any 
 organism and its variation and hybridization phenomena. 
 
 In Galtonia and probably also in Tradescantia there are apparently 
 the same possibilities that both chromosomes of a pair may occasion¬ 
 ally enter the same daughter nucleus. In other plant forms studied 
 the attraction between chromosomes seems to be strong enough to 
 keep the members of a pair together until their separation in the 
 anaphase of the heterotypic mitosis. The segregation of the members 
 of a pair into separate germ cells is thus insured. In cases where, 
 as in Oenothera, the members of a pair do not always remain in 
 contact, but are loosely arranged on the spindle, such a result as 
 already suggested seems certain to occur in certain instances. 
 
 It has already been mentioned that occasionally one chromosome 
 goes to the wrong pole of the heterotypic spindle. This is found to 
 be the case particularly in the hybrids, for example, in the O. Lamarcki- 
 
 t 
 
28 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 ana plants from the Fi of O. lataXO. Lamarckiana ( 13 ), in which 
 sometimes eight chromosomes pass to one pole and six to the other; 
 but it may also occur rarely in the pure races. This matter was 
 briefly discussed elsewhere ( 14 ). Assuming that the 14 chromosomes 
 are in two similar sets of 7 each, and that homologous members of 
 these sets conjugate except when there is a failure to pair, then when 
 8 chromosomes go to one pole and 6 to the other, both members of one 
 of the pairs must have gone to the same pole. This probably takes 
 place in cases where such members were unconjugated, for the 
 purpose, or at any rate, the result of the pairing is in ordinary cases 
 that one member of every pair shall be distributed to each pole. If, 
 while two members of one pair thus go to one pole, the second member 
 of another pair goes to the other pole, we should have an equal 
 numerical distribution of chromosomes, but one daughter group would 
 be lacking both members of one pair and the other would be lacking 
 both members of another pair. It is highly probable that such a 
 distribution occasionally takes place, though it would be less common 
 than the case, already proved, where the members of a single pair are 
 unilaterally distributed. It should be borne in mind that such cases 
 are most likely to occur, not when the members of a pair are con¬ 
 jugated, but when they lie separately in diakinesis and on the spindle. 
 
 Miss Lutz ( 17 ), from an examination of root tips, states that she 
 has observed several individuals belonging to different strains having 
 15 chromosomes instead of 14. This is to be anticipated from the 
 irregularities in chromosome distribution in reduction already men¬ 
 tioned. I have observed one such case in O. lataXO. gigas (14) — 
 a certain plant having 20 chromosomes instead of 21. All the plants 
 of O. lata (12) and O. nanella (13) thus far examined by me had 14 
 chromosomes, while Miss Lutz ( 17 ) finds in root tips some O. lata 
 plants with 14 and also some with 15 or she thinks possibly 16 chro¬ 
 mosomes. She reports finding two O. fianella plants with 14 chro¬ 
 mosomes and one with 15. Two O. albida seedlings are said to have 
 15 chromosomes and two O. oblonga plants 15, while a third has 14. 
 Disregarding the possibility that these results might be due to the 
 well-known variation in chromosome numbers in root tips, they are 
 such as would be likely to arise in different individuals from the 
 cytological irregularities I have already described. Whether there 
 
GATES—REDUCTION IN OENOTHERA 
 
 29 
 
 1908] 
 
 are external differences between the plants having 14 chromosomes 
 and those of the same race having 15, is as yet unknown. But it is 
 quite conceivable that no such differences will be found, for if the 
 sporophyte chromosomes consist of two complete sets (and for a 
 variety of reasons this seems the only tenable view at the present time 
 if we assume qualitative differences at all), the presence of an addi¬ 
 tional chromosome, which is already present in duplicate, would 
 scarcely be expected visibly to affect the plant. 
 
 Rosenberg (26) has found an analogous situation in Hieracium. 
 For example, H. excellens'KH. Pilosella gives hybrids with different 
 numbers of chromosomes. This he ascribes to the fact that the eggs 
 of H. excellens differ in their numbers of chromosomes, which he finds 
 is due to irregularities in chromosome distribution during the reduction 
 divisions. The writer has pointed out elsewhere (12) certain similari¬ 
 ties between the hybridization phenomena in Hieracium and Oeno¬ 
 thera, and this seems to be a further similarity between the two genera.' 
 
 Rosenberg (27) has since shown that H. excellens produces three 
 kinds of embryo sacs: (i) Normal embryo sacs which require fertili¬ 
 zation for their development. These are presumably the only ones 
 which can be hybridized. The egg cells in these sacs vary in their 
 number of chromosomes owing to the fact that some of the chromo¬ 
 somes, lacking in “affinity,” remain univalent (that is, fail to pair) 
 during the heterotypic mitosis and are irregularly distributed. It is 
 evident that this lack of affinity between chromosomes is similar to that 
 in Oenothera. (2) In rare cases apogamous embryo sacs are formed 
 after a single division of the megaspore mother cell, and without 
 reduction. (3) More frequently the condition occurs which Rosen¬ 
 berg calls apospory, in which tetrad formation takes place and then 
 an adjacent cell of the nucellus enlarges, displaces the tetrad, and 
 forms an embryo sac without reduction. 
 
 Summary 
 
 In conclusion a brief summary of the facts and considerations 
 here presented will be useful. 
 
 I. In Oenothera the heterotypic mitosis is a reduction division, 
 separating whole chromosomes which lie successively on the spirem. 
 The homotypic mitosis is an equation division, separating the longi- 
 
30 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 tudinal halves of the daughter chromosomes of the heterotypic 
 mitosis. Whether an approximation of threads or a split in a single 
 thread occurs in synapsis was not determined with certainty from the 
 observations, but various considerations lead to the belief that in 
 Oenothera the doubling is due to a split which closes up later, rather 
 than to an approximation of separate spirems. 
 
 2. The conclusion that the method of reduction probably differs 
 in different genera is based on two considerations: (i) the fact that 
 in most of the recent accounts of synapsis and reduction in plants 
 a side-by-side pairing of chromosomes from maternal and paternal 
 spirems is described, while in Oenothera the members of a pair are 
 arranged end to end on a single spirem; and (2) on differences in 
 chromosome distribution during reduction in certain hybrids of 
 Drosera and of Oenothera (see p. 25). If reduction took place in 
 the same manner in both genera, the chromosome distribution during 
 reduction in these hybrids with reference to the parental chromosome 
 numbers should be the same in both, but this is not the case. 
 
 3. Pairing between the definitive chromosomes during diakinesis 
 and the prophase of the heterotypic mitosis does not always take place, 
 owing to a weak attraction between the chromosomes. This allows 
 irregularities of distribution in the heterotypic mitosis, so that both 
 (unpaired) chromosomes belonging to one pair will occasionally enter 
 the same daughter nucleus (see p. 26). Germ cells will thus arise, 
 from which both members of a given pair of chromosomes are 
 absent. 
 
 4. If we assume qualitative differences between the chromosomes 
 or parts of them, various types would be expected to originate in this 
 manner, each of them lacking the ability to develop certain qualities 
 possessed by the parent form. On this view the mutations of Oeno¬ 
 thera Lamarckiana are an instance of a process of analysis by which 
 from the parent form arises a series of types, each lacking in certain 
 characters or capacities possessed by the parent. This hypothesis 
 would account for the absence of reversions arnong Oenothera 
 mutants, and perhaps also for some of the peculiarities of hybridiza¬ 
 tion in Oenothera. This matter will be considered at another time. 
 This explanation does not apply to all the mutants, however; for 
 example, O. gigas. 
 
GATES—REDUCTION IN OENOTHERA 
 
 31 
 
 1908] 
 
 5. It is suggested that there is probably a direct relation between 
 the events of reduction in a given genus and its variation, as well as its 
 hybridization phenomena. 
 
 I desire to express my thanks to Professors John M. Coulter 
 and Charles R. Barnes for valuable suggestions and adequate 
 facilities in connection with this work. 
 
 The University oe Chicago 
 
 LITERATURE CITED 
 
 1. Allen, C. E., Nuclear division in the pollen mother cells of Lilium canadense. 
 Annals of Botany 19:189-258. pis. 6-g. 1905. 
 
 2. -, Das Verhalten der Kernsubstanzen wahrend der Synapsis in den 
 
 Pollenmutterzellen von Lilium canadense. Jahrb. Wiss. Bot. 42:72-82. 
 pi. 2. 1905. 
 
 3. Berghs, Jules, La formation des chromosomes heterotypiques dans murs la 
 sporogenese vegetale. I. Depuis le spireme jusqu’au chromosomes murs, 
 dans la microsporogenese d*Allium fistulosum et Lilium lancifolium {speci- 
 osum). La Cellule 21:173-189. pi. i. 1904. 
 
 4. -, II. Depuis la sporogonie jusqu’au spireme definitif, dans la micro¬ 
 
 sporogenese de I’d21:383-397. pi. I. 1904. 
 
 5. -, III. La microsporogenese de Convallaria maialis. Idem 22:43-50. 
 
 pi. I. 1905. 
 
 6. -, IV. La microsporogenese de Drosera rotundifolia, Narthecium 
 
 ossifragum, et Hellehorus foetidus. Idem 22:141-160. pis. 2. 1905. 
 
 7. Cardiff, Ira D., A study of synapsis and reduction. Bull. Torr. Bot. 
 Club 33:271-306. pis. 12-15. 1906. 
 
 8. Farmer, J. B., and Moore, J. E. S., New investigations into the reduction 
 phenomena of animals and plants. Proc. Roy. Soc. 72:104-108. figs. 6. 1903. 
 
 9. --- —, On the maiotic phase (reduction divisions) in animals and plants. 
 
 Quart. Jour. Micr. Sci. 48:489-557. pis. 34-41. 1905. 
 
 10. Farmer, J. B., and Shove, Dorothy, On the structure and development of 
 the somatic and heterotype chromosomes of Tradescantia virginica. Idem 
 
 48:559-569- pis. 42, 43’ 1905- 
 
 11. Gates, R. R., Pollen development in hybrids of Oenothera lata X O. 
 Lamarckiana, and its relation to mutation. Bot. Gazette 43:81-115. 
 pis. 2-4. 1907. 
 
 12. —-Hybridization and germ cells of Oenothera mutants. Idem 44:1-21. 
 
 figs. 3. 1907. 
 
 13. -, International Zool. Congress, Boston, Aug., 1907. 
 
 14. -, The chromosomes of Oenothera. Science N. S. 27:193-195. 1908. 
 
 15. -— , Further studies on the chromosomes of Oenothera. Idem 27:335. 
 
 1908. 
 
32 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 16. Gregoire, V., La reduction numerique des chromosomes et les cineses de 
 maturation. La Cellule 21:297-314. 1904. 
 
 17. Lutz, Anne M., Chromosomes of the somatic cells of the Oenotheras. 
 Science N. S. 27:335. 1908. 
 
 18. Miyake, Kiichi. Ueber Reduktionsteilung in den Pollenmutterzellen 
 einiger Monokotylen. Jahrb. Wiss. Bot. 42:83-120. pis. j-5. 1905. 
 
 19. Mother, D. M., The development of the heterotypic chromosomes in pollen 
 mother cells. Bot. Gazette 40: 171-177. 1905. 
 
 20. - , The development of the heterotypic chromosomes in pollen mother 
 
 cells. Annals of Botany 21:309-347. pis. 27, 28. 1907. 
 
 21. Nichols, M. Louise, The development of the pollen of Sarracenia. Bot. 
 Gazette 45:31-37. pi. 5 . 1908. 
 
 22. Overton, J. B., Ueber Reduktionsteilung in den Pollenmutterzellen einiger 
 Dikotylen. Jahrb. Wiss. Bot. 42:121-153. pis. 6, 7. 1905. 
 
 23. Rosenberg, O., Das Verhalten der chromosomen in einer hybriden Pflanze. 
 Ber. Deutsch. Bot. Gesells. 21:110-119. 1903. 
 
 24. -, Ueber die Tetradentheilung eines Drosera-Bastardes. Ber. Deutsch. 
 
 Bot. Gesells. 22:47-53. P^- 4- 1904- 
 
 25. -, Zur Kenntniss der Reduktionstheilung in Pflanzen. Bot. Notiser 
 
 1905: pp. 24. Jigs. 14. 
 
 26. -, Cytological investigations in plant hybrids. Report 3d Internat. 
 
 Conf. Genetics, pp. 289-291. 1906. 
 
 27. -, Cytological studies on the apogamy in Hieracium. Bot. Tidsskrift 
 
 28:143-170. Jigs. 13. pis. 7, 2. 1907. 
 
 28. ScHAEFNER, J. H., Chromosome reduction in the- microsporocytes of 
 Lilium tigrinum. Bot. Gazette 41:183-191. pis. 12, 13. 1906. 
 
 29. -, Synapsis and synizesis. Ohio Nat. 7:41-48. pi. 4. 1907. 
 
 30. Strasburger, E., Ueber Reduktionstheilung. Sitzungsber. k. k. Preuss. 
 Akad. Wiss. 18:587-614. Jigs. 13. 1904. 
 
 31. -, Typische und allotypische Kerntheilung. Jahrb. Wiss. Bot. 42:1- 
 
 71. pi. I. 1905. 
 
 32. Tischler, G., Zellstudien an sterilen Bastardpflanzen. Archiv Zellforschung 
 1:33-151- fig ^- 120 . 1908. 
 
 33. Yamanouchi, Shigeo, Sporogenesis in Nephrodium. Bot. Gazette 45: 
 1-30. pis. 1-4. 1908. 
 
 EXPLANATION OF PLATES I-III 
 
 The figures were drawn with the aid of a Bausch & Lomb camera lucida. All 
 except Jigs, i and ig were drawn under a Zeiss apochromatic objective 2™™ ap. 
 1.30, with a Zeiss compensating ocular 18. The figures are reduced one-fourth in 
 reproduction, giving a magnification of nearly 3000 diameters. Fig. i was drawn 
 under a 2™^ objective and compensating ocular 6; Jig. ig under B. & L. objective 
 iV N. A. 1.32 and Zeiss ocular 18. 
 
GATES—REDUCTION IN OENOTHERA 
 
 33 
 
 1908] 
 
 PLATE 1 
 
 Figs, i, 2.—Young meristematic cells of anther primordium showing one 
 large nucleolus and several smaller ones, and chromatic masses adherent to the 
 nuclear membrane. 
 
 Fig. 3. —Longitudinal section of anther, showing size relations of nucleoli in 
 sporogenous, tapetal, and wall cells. 
 
 Fig. 4. —One sporogenous cell from stage of fig. j, previous to synapsis; 
 cytoplasm somewhat vacuolate. 
 
 Figs. 5, 7-9.—Nuclei at same stage, showing fusions of nucleoli. 
 
 Fig. 6. —Two nucleoli of equal size; an unusual condition. 
 
 Fig. 10.—Several small nucleoli, and no indication of fusion. 
 
 Fig. II.—Nucleoli of young pollen grain nucleus. 
 
 Fig. 12. —Beginning of synaptic contraction; the reticulum has contracted 
 from the nuclear membrane on all sides, leaving several loops attached to the 
 membrane; on the side on which the reticulum retains the curved outline of the 
 nuclear membrane the latter has been drawn inward attached to the threads; on 
 the rest of the circumference, between the loops, the nuclear membrane remains 
 in situ; the cytoplasm is perfectly fixed. 
 
 Fig. 13.—Another contraction stage, showing loops attached to the nuclear 
 membrane, which is intact. 
 
 Fig. 14.—A slightly later stage of contraction, in which the rearrangement 
 of threads is taking place. 
 
 Fig. 15. —Synapsis; dark-staining bodies are still held in the meshes of the 
 spirem; a small nu-cleolus, usually about the size of a chromosome, is generally 
 present in addition to the large nucleolus. 
 
 Fig. 16. —After synapsis; the thread thicker and shorter and loosely coiled. 
 
 Fig. 17. —Slightly later stage than fig. 16, and less deeply stained; thread 
 shows the characteristic light and dark areas; indications of parallel threads in 
 two places; edge of thread may be even or moniliform.—5 ai. 
 
 Fig. 18. —Later stage; thread much shortened and greatly thickened and 
 entering upon second contraction phase; nucleus uncut.—10 p. 
 
 Fig. 19. —Higher magnification of a portion of the thread in figs. 20 and 21. 
 
 PLA TE II 
 
 Fig. 20. —Second contraction stage; a pair of chromosomes cut off from 
 spirem; nucleus uncut.—10 /x. 
 
 Fig. 21.—Second contraction stage; nucleus uncut. 
 
 Fig. 22.—Uncoiling from second contraction stage; pair of chromosomes 
 detached; nucleus uncut. 
 
 Fig. 23.—Spirem segmented in three places, each segment showing constric¬ 
 tions which will form the chromosomes; certain chromosomes already detached; 
 nucleus uncut. 
 
 Fig. 24. —Constriction of spirem has proceeded farther, the chromosomes being 
 elongated bodies with irregular margins like the spirem, and connected by rather 
 
34 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 thick “linin’’ bands; pair of chromosomes detached earlier lies at side of nucleus; 
 n, small nucleolus; nucleus cut. 
 
 Fig. 25.—Spirem more or less completely segmented into chromosomes while 
 still in the second contraction stage; preparation considerably destained; 13 
 chromosomes in view. 
 
 Fig. 26.—Spirem segmented, showing chain of eight chromosomes and three 
 pairs; nucleus uncut. 
 
 Fig. 27.—Chain of six chromosomes, and probably four pairs; linin con¬ 
 nections between members of a pair not always visible; nucleus uncut. 
 
 Fig. 28.—Fourteen chromosomes; several small nucleoli; nucleus uncut. 
 
 Fig. 29.—Fourteen chromosomes, including five pairs more or less closely 
 associated; linin connections not visible; one pair of chromosomes has already 
 contracted into the globular shape. 
 
 Fig. 30.—Fourteen chromosomes, several in pairs; apparent inequalities in 
 size due to positions in which some of the chromosomes are lying. 
 
 Fig. 31.—Slightly later stage; the fourteen chromosomes contracted into the 
 globular or pear-shaped definitive form; linin connections longer and extremely 
 delicate; nucleus uncut. 
 
 Figs. 32-34.—Other groups in diakinesis, showing various peculiarities of 
 chromosomes. 
 
 Fig. 35.—Peculiar case of spindle formation; three nucleoli present and four¬ 
 teen chromosomes, including three or four pairs. 
 
 Fig. 36.—Multipolar stage of heterotypic spindle; two more or less closely 
 united pairs of chromosomes present. 
 
 PLA TE III 
 
 Fig. 37.—Same as fig. j6; an unusual case in which all the chromosomes 
 are closely joined in pairs; seven such pairs present and a small nucleolus. 
 
 Fig. 38.—Heterotypic spindle in metaphase; spindle has usually more mantle 
 fibers than in O. Lamarckiana; chromosomes usually loosely arranged in equatorial 
 region of spindle. 
 
 Fig. 39.—Late anaphase; an uncommon case; daughter chromosomes have 
 failed to divide, and fibrillae are scattered in cytoplasm at side of cell; chromatic 
 staining material also present. 
 
 Fig. 40.—Telophase of heterotypic mitosis; exceptional case, in which a 
 rather sharp pointed spindle is formed at side of cell; it probably originated from 
 the fibrillae shown in Jig. jg. 
 
 Fig. 41.—Early anaphase of homotypic mitosis; small nucleolus having the 
 characteristic appearance, present on one of the spindles. 
 
 Fig. 42.—The single case of extra nuclei observed in O. ruhrinervis pollen 
 mother cells. 
 
 Figs. 43, 44.—Nuclei from telophase of second mitosis, passing into resting 
 condition. 
 
 Figs. 45, 46.—Protoplasmic connections between mother cells. 
 
igo8] 
 
 CzATES—-REDUCTION IN OENOTHERA 
 
 33 
 
 PLA TE 1 
 
 Figs, i, 2.—Young meristematic cells of anther primordium showing one 
 large nucleolus and several smaller ones, and chromatic masses adherent to the 
 nuclear membrane. 
 
 Fig. 3.—Longitudinal section of anther, showing size relations of nucleoli in 
 sporogenous, tapetal, and wall cells. 
 
 Fig. 4.—One sporogenous cell from stage of fig. j, previous to synapsis; 
 cytoplasm somewhat vacuolate. 
 
 Figs, 5, 7-9.—Nuclei at same stage, showing fusions of nucleoli. 
 
 Fig. 6.—Two nucleoli of equal size; an unusual condition. 
 
 Fig. 10,—Several small nucleoli, and no indication of fusion. 
 
 Fig. II. —Nucleoli of young pollen grain nucleus. 
 
 Fig. 12.—Beginning of synaptic contraction; the reticulum has contracted 
 from the nuclear membrane on all sides, leaving several loops attached to the 
 membrane; on the side on which the reticulum retains the curved outline of the 
 nuclear membrane the latter has been drawn inward attached to the threads; on 
 the rest of the circumference, between the loops, the nuclear membrane remains 
 in situ; the cytoplasm is perfectly fixed. 
 
 Fig. 13.—Another contraction stage, showing loops attached to the nuclear 
 membrane, which is intact. 
 
 Fig. 14.—A slightly later stage of contraction, in which the rearrangement 
 of threads is taking place. 
 
 Fig. 15.—Synapsis; dark-staining bodies are still held in the meshes of the 
 spirem; a small nucleolus, usually about the size of a chromosome, is generally 
 present in addition to the large nucleolus. 
 
 Fig. 16.—After synapsis; the thread thicker and shorter and loosely coiled. 
 
 Fig. 17.—Slightly later stage than fig. 16, and less deeply stained; thread 
 shows the characteristic light and dark areas; indications of parallel threads in 
 two places; edge of thread may be even or moniliform.—5 u. 
 
 Fig. 18.—Later stage; thread much shortened and greatly thickened and 
 entering upon second contraction phase; nucleus uncut.—10 u. 
 
 Fig. 19.—Higher magnification of a portion of the thread in figs. 20 and 21. 
 
 PLATE II 
 
 Fig. 20.—Second contraction stage; a pair of chromosomes cut off from 
 spirem; nucleus uncut.—10 u. 
 
 Fig. 21.—Second contraction stage; nucleus uncut. 
 
 Fig. 22.—Uncoiling from second contraction stage; pair of chromosomes 
 detached; nucleus uncut. 
 
 Fig. 23.—Spirem segmented in three places, each segment showing constric¬ 
 tions which will form the chromosomes; certain chromosomes already detached; 
 nucleus uncut. 
 
 Fig. 24.—Constriction of spirem has proceeded farther, the chromosomes being 
 elongated bodies with irregular margins like the spirem, and connected by rather 
 
34 
 
 BOTANICAL GAZETTE 
 
 [JULY 
 
 thick “linin’’ bands; pair of chromosomes detached earlier lies at side of nucleus; 
 n, small nucleolus; nucleus cut. 
 
 Fig. 25.^—Spirem more or less completely segmented into chromosomes while 
 still in the second contraction stage; preparation considerably destained; 13 
 chromosomes in view. 
 
 Fig. 26.—Spirem segmented, showing chain of eight chromosomes and three 
 pairs; nucleus uncut. 
 
 Fig. 27.—Chain of six chromosomes, and probably four pairs; linin con¬ 
 nections between members of a pair not always visible; nucleus uncut. 
 
 Fig. 28.—Fourteen chromosomes; several small nucleoli; nucleus uncut. 
 
 Fig. 29.—Fourteen chromosomes, including five pairs more or less closely 
 associated; linin connections not visible; one pair of chromosomes has already 
 contracted into the globular shape. 
 
 Fig. 30.—Fourteen chromosomes, several in pairs; apparent inequalities in 
 size due to positions in which some of the chromosomes are lying. 
 
 Fig. 31.—Slightly later stage; the fourteen chromosomes contracted into the 
 globular or pear-shaped definitive form; linin connections longer and extremely 
 delicate; nucleus uncut. 
 
 Figs. 32-34.—Other groups in diakinesis, showing various peculiarities of 
 chromosomes. 
 
 Fig. 35.—Peculiar case of spindle formation; three nucleoli present and four¬ 
 teen chromosomes, including three or four pairs. 
 
 Fig. 36.—Multipolar stage of heterotypic spindle; two more or less closely 
 united pairs of chromosomes present. 
 
 PLA TE III 
 
 Fig. 37.—Same as jig. j6; an unusual case in which all the chromosomes 
 are closely joined in pairs; seven such pairs present and a small nucleolus. 
 
 Fig. 38.—Heterotypic spindle in metaphase; spindle has usually more mantle 
 fibers than in O. Lamarckiana; chromosomes usually loosely arranged in equatorial 
 region of spindle. 
 
 Fig. 39.—Late anaphase; an uncommon case; daughter chromosomes have 
 failed to divide, and fibrillae are scattered in cytoplasm at side of cell; chromatic 
 staining material also present. 
 
 Fig. 40.—Telophase of heterotypic mitosis; exceptional case, in which a 
 rather sharp pointed spindle is formed at side of cell; it probably originated from 
 the fibrillae shown in fig. jp. 
 
 Fig. 41.—Early anaphase of homotypic mitosis; small nucleolus having the 
 characteristic appearance, present on one of the spindles. 
 
 Fig. 42.—The single case of extra nuclei observed in O. ruhrinervis pollen 
 mother cells. 
 
 Figs. 43, 44.—Nuclei from telophase of second mitosis, passing into resting 
 condition. 
 
 Figs. 45, 46.—Protoplasmic connections between mother cells. 
 
nOT^[XICAL (iAZKTTK,XL VI 
 
 PLATE I 
 
 m 
 
 (#ATi<:s on RKDiTCTioN in OKxarnii:i?A 
 
nOTANICAL (iAZETTE, XJA I 
 
 ELATE II 
 
 (iATKS on IU:i)U(rnO.\ in OKNOTHERA 
 
■ 
 
 I*. N 
 
 f> ■ 
 
 t , 
 
 I .* 
 
 ^■0 
 
 ^ . 
 
 I-. • 
 
 -w 
 
 
 V 
 
 
 1 
 
 r*f; 
 
 x*' 
 
 3 
 
 •f' 
 
 T» >*• . % n 
 
 • "V 
 
 
 
 
 < 
 
 
 
 i 
 
 t, 
 
 
 
 I 
 
 V -t 
 
 . « 
 
 
 l''-“ 
 
 l'.'. 
 
 -r 
 
 \ 
 
 r/ 
 
 b<< 
 
 r 
 
 4 
 
 « 4 . » 
 
 
 iv'.’' " 
 
 ‘ ^ 1 
 
 *> 1-1 
 
 ♦ 
 
 
 
 
 
 
 X- 
 
n(n\ iA 76 : i/. (jA 7 .ettk, xl \y 
 
 PLA TE IH 
 
 OATI<:S on RHlDUCriOX in OEXOI'URRA