A CASE OF MENDELIAN INHERITANCE COMPLICATED BY HETEROGAMETISM AND MUTATION IN OENOTHERA PRAT IN COL A BY FRIEDA COBB A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIRE- MENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY AT THE UNIVERSITY OF MICHIGAN 1921 A CASE OF MENDELIAN INHERITANCE COMPLICATED BY HETEROGAMETISM AND MUTATION IN OENOTHERA PRATINCOLA BY FRIEDA COBB A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIRE- MENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY AT THE UNIVERSITY OF MICHIGAN 1921 BIOLOG1R R A CASE OF MENDELIAN INHERITANCE COMPLICATED BY HETEROGAMETISM AND MUTATION IN OENOTHERA PRATINCOLA 1 FRIEDA COBB University of Michigan, Ann Arbor, Michigan Received June 21, 1920 TABLE OF CONTENTS INTRODUCTION 1 Hypothesis of heterogametism 2 Equivalence, within strain C, of pollen of Oenothera pratincola f. typica and mut. latifolia. . 4 The F 3 and F 4 generations of the cross mut. formosa from strain E X mut. latifolia from strain C 5 The FI and F 2 generations of the cross f . typica strain C X mut. formosa strain E 6 The FI and F2 generations of the cross mut. formosa strain E X f . typica strain E 7 The FI and F 2 generations of various crosses between f . typica M (the new Mendelian strain) and f . typica of strain C, f . typica of strain E, and mut. formosa 8 Various pedigrees of plants used in crosses 9 DISCUSSION 10 SUMMARY 13 LITERATURE CITED 15 APPENDIX TABLES 15 INTRODUCTION In the study of heredity in the genus Oenothera simple Mendelian results are very rarely obtained. It is generally conceded that the unusual results are due to the production, by a single morphologically uniform strain, of gametes of more than one type. The explanations have been various. The mutation hypothesis of DE VRIES is of course well known. It has not been generally accepted, however, as originally proposed. MORGAN (1918) and MULLER (1918) have suggested point mutation, followed by crossing over, an explanation suggested by their work on Drosophila. MULLER has worked out experimentally a balanced lethal stock of Drosophila in which certain seemingly aberrant phenomena of the Oenotheras find a parallel. To what extent the parallel affords an explanation we can not judge until the genetic analysis of the Oenotheras has been carried further. 1 Papers from the Department of Botany of the UNIVERSITY OF MICHIGAN, No. 181. GFNETICS 6:1 Ja 1921 458726 2 FRIEDA COBB The present paper is a contribution to the genetic analysis of Oenothera pratincola. It deals exhaustively with the first case in which simple Mendelian inheritance has been recognized and conclusively demonstrated when complicated by phenomena peculiar to the Oenotheras. The strains of Oenothera pratincola used in this investigation were those of which the history has already been published (BARTLETT 1915 a, b; COBB and BARTLETT 1919). Although morphologically alike, one of them, desig- nated as strain E, is genetically different from the other seven, of which strain C, the strain used in the experiments recorded in this paper, is a typical example. Strain C produces in every generation a small number of mutations of several kinds (BARTLETT 1915 a). Some of these kinds appear also in strain E, but much more conspicuous in strain E are numerous mutations, of a strikingly distinct series, which do not occur in the other strains. These mutations occur in such numbers as to merit the term mass muta- tion (BARTLETT 1915 b) as a designation of the phenomenon. The series consists of four distinct types, all alike in having narrow, strongly revolute leaves, and in producing nothing butrevolute-leaved plants in their progenies. Of these revolute-leaved mutations, mut. formosa (BARTLETT 1915 b), the strongest and most fertile of the series, was crossed with f . typica of strain E, and with strain C. In a former paper (COBB and BARTLETT 1919) it has been stated that in reciprocal crosses between mut. formosa and f . typica E, from which mut. formosa arises, inheritance is matroclinic. Strain C pollinated by mut. formosa gives a matroclinic progeny; but the reciprocal cross, mut. formosa pollinated by strain C, gives in the FI generation only f. typica, in the F 2 generation a Mendelian segregation of 3 f. typica: 1 mut. formosa. HYPOTHESIS OF HETEROGAMETISM The hypothesis of heterogametism offered (COBB and BARTLETT 1919) in explanation of these phenomena, assumes that two types of gametes occur in Oenothera pratincola, a gametes (usually female) and gametes (usually male), the a gametes carrying some factors not represented in the gametes. Each zygote is formed by the union of an a and a /3 gamete, and so gets (except in rare cases of metacliny) the a determiners of its pistillate parent and the determiners of its staminate parent. It, in turn, produces a (female) and (male) gametes. In the case of a cross, the zygote is quite unaffected by the nature of the a of its staminate parent and the /3 of its pistillate parent. MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 3 Besides its characteristic a or ft part, each gamete carries a group of factors common to both a and ft gametes. The characteristic a and ft portions of the gametes may consist of a single chromosome or of a group of chromosomes, but probably the latter, for very few characters have been found which are not connected with the a or ft portions of the gametes. At reduction, a and ft behave as units; that is, there is no interchange of factors or chromosomes between the characteristic a and ft portions, each passing into the gametes (the a into the female, the ft into the male) just as it entered the zygote from the parent. There is, however, the usual free segregation among the remainder of the chromosomes, each member of the homologous pairs accompanying with equal frequency the a and the ft portions. Thus factors entering a zygote in the characteristic a or ft chromosome (or chromosomes) occur in the a or ft gametes, respectively, which it produces ; but those factors which entered the zygote in the remain- ing, freely segregating group of chromosomes, occur as frequently in the a as in the ft gametes produced. Characters whose factors belong to the a or to the ft portion are inherited matroclinically or patroclinically, respec- tively; those whose factors belong to homologous and freely segregating chromosomes common to both a and ft gametes are inherited in a Men- delian manner. Mut. formosa arose from i. typical by modification of a factor for flatness in the a portion of the a gamete, that is, by change of a factor which has no counterpart in the ft portion of the ft gamete. Change being in the a (female) gamete only, inheritance in crosses between mut. formosa and f . typica E is, therefore matroclinic. If mut. formosa is used as the pistillate parent the a gamete received by the progeny is the mutated a of mut. formosa (designated hereafter as <*'), and the progeny is therefore mut. formosa; if f. typica is used as the pistillate parent the a gamete received by the progeny is the normal a gamete of Oenothera pratincola, and the progeny is f . typica. Strain C differs from strain E in having, in addition to the factor for flatness in the a portion of the a gamete, a freely segregating (Mendelian) factor for flatness (F) present in both a and ft gametes, of which the reces- sive allelomorph (/) is carried by strain E. Thus the constitution of strain C is aftFF, and the gametes which it produces are aF (female) and ftF (male). The constitution of strain E is aftff, and the gametes which it produces are af (female) and ftf (male). The constitution of mut. formosa is a'ftff, and the gametes which it produces are a'f (female) and ftf (male). Therefore strain C (aftFF) pollinated by mut. formosa (a'ftff) gives a flat-leaved progeny (a&Ff) which breeds true, for the a gamete GENETICS 6 : Ja 1921 4 FRIEDA COBB concerned in the cross is normal; but mut. formosa (aftff) pollinated by strain C (afiFF) gives a flat-leaved progeny (a'pFf) (flat-leaved, notwith- standing the mutated condition of a, because of the presence of one Men- delian factor for flatness, inherited from the pistillate parent) which shows in the next generation a Mendelian segregation of flat-leaved plants (a$FF and a.'&Ff) and revolute-leaved plants (a$ff). Thee/ gamete concerned in the cross has lost the factor for flatness, and so those F 2 individuals that are recessive for the Mendelian factors for flatness show the revolute- ness determined by a '. This new synthetic f. typica (a&FF or afiFf), differing in genetic composition from both f . typica C and f. typica E, carry- ing the factor for revoluteness masked by at least one of the Mendelian pair of factors for flatness, will be called f. typica M (homozygous or hete- rozygous, as the case may be). The hypothesis of the genetic constitution of the plants, and the results of crossing, may be stated in brief as follows: Strain C, apFF, flat, and, with respect to this char- acter, immutable. Strain E, aftff, flat, mutable. Mut. formosa, a 'Pff, revolute-leaved. Strain E X formosa, aflff, flat, mutable. Formosa X strain E, a'pff revolute, breeding true with respect to this character. Strain C X formosa, a(3Ff flat, segregating with respect to muta- bility. 1 a0FF flat, immutable, breeding true. 2 a0Ff flat, continuing the segregation of the Fi generation. 1 apff flat, mutable, otherwise breeding true. Formosa X strain C, a'&Ff flat, segregating with respect to revo- luteness. 1 a&FF flat, non-segregating. (Formosa X strain C) FJ 2 'W^< continuing the segregation of the FI generation. 1 o'jSjfjT revolute, breeding true. EQUIVALENCE, WITHIN STRAIN C, OF POLLEN OF OENOTHERA PRATINCOLA F. TYPICA AND MUT. LATIFOLIA The former paper (COBB and BARTLETT 1919) gave data of the F, and F 2 generations of the cross mut. formosa E X mut. latifolia C. At the time the original crosses between mut. formosa and strain C were made, (Strain C X formosa) F 2 MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 5 the cross mut. formosa X f . typica C was unsuccessful. But the pollen of mut. latifolia had been shown to be equivalent to the pollen of the f . typica from which it arises (the mutation to latifolia being concerned with the a gamete, the /? gamete remaining as in f. typica), and therefore, in order to have a more complete series of crosses to work upon, the cross mut. formosa X mut. latifolia from strain C was used in place of the missing cross, mut. formosa X f . typica C. The substitution has been justified by later work, recorded in tables 5 and 6 of the present paper, which give anal- yses of the Fi and F 2 generations of the cross mut. formosa X f . typica C. This cross was successfully made three years after the original crosses upon which much of the work recorded here is based. The Fi generation (see table 5) consisted of 100 plants from six different crosses, all of them flat-leaved. The F 2 generation (see table 6) consisted of 3274 plants, 2399 flat-leaved and 875 revolute-leaved; that is, a ratio of 2.74:1, in sufficiently good accord with the 3:1 of the Mendelian mono- hybrid ratio. The cross mut. formosa X mut. latifolia C, as recorded in the previous paper, gave in the FI generation 209 plants, all of them flat- leaved, and in the F 2 generation from normal f. typica plants of the FI generation, 6392 plants, 4759 flat-leaved and 1633 revolute-leaved; that is, a ratio of 2.91 : 1. This shows that the results in the F 2 generation are the same whether mut. formosa is pollinated by f . typica C or by mut. latifolia C ; and this, in connection with the previous evidence (CoBB and B ARTLETT 1919) of the equivalence of the pollen from the two sources, gives ample justification for the substitution of f. typica M descended from mut. lati- folia instead of the identical form descended from f. typica C. THE F 3 AND F 4 GENERATIONS OF THE CROSS MUT. FORMOSA FROM STRAIN E X MUT. LATIFOLIA FROM STRAIN C The F 3 and F 4 generations of the cross mut. formosa X mut. latifolia C show a continuance of the Mendelian behavior of the F 2 generation (COBB and BARTLETT 1919). Self-pollination of normal f. typica plants of the F 2 generation gave in the F 3 generation (see table 2) 22 progenies consisting entirely of flat-leaved plants, showing the presence of homozygous dom- inants (a'(3FF) in the F 2 generation, and 41 progenies in which there were both flat-leaved and revolute-leaved plants, showing the presence of heterozygous dominants (a'pFf) in the F 2 generation. The ratio 22 uni- form cultures to 41 segregating cultures very closely approximates the expected ratio of 1 homozygous dominant to 2 heterozygous dominants in the F 2 generation. In the segregating progenies of the F 3 generation GENETICS 6 : Ja 1921 6 FRIEDA COBB the ratio of flat-leaved to revolute-leaved plants is 3.08:1, in very close agreement with the expected 3 : 1 of the Mendelian monohybrid segregation. All progenies of fewer than 20 plants were omitted from table 2 as being unreliable. A very small culture of flat-leaved plants might, had it been larger, have included some revolute-leaved plants, and thus a heterozygous dominant of the F 2 generation might be recorded as homozygous. If these cultures were included in the table, the ratio of non-segregating to segre- gating progenies would become 33:56 instead of 22:41, and the ratio of flat-leaved to revolute-leaved plants in the segregating progenies would become 2.94: 1 instead of 3.08: 1. Self-pollination of 16 plants of f . typica M (a'&FF) belonging to the non- segregating cultures of the F 3 generation gave in the F 4 generation (see table 3) 1114 plants, all of them flat-leaved. From these plants an F 5 generation has been grown, consisting of 695 plants belonging to 7 prog- enies, no progeny consisting of fewer than 29 plants. All were flat-leaved. Self-pollination of 3 plants chosen at random among the f . typica plants of the segregating cultures of the F 3 generation gave in the F 4 generation (see table 4) 2 segregating progenies and 1 non-segregating, showing a continu- ance of the Mendelian splitting to the fourth filial generation. It seems unnecessary to carry the line further. The behavior of the recessives, mut.formosa (a'Pff), of the F 2 generation of the cross mut.formosa X mut. latifolia C. was also in accord with expec- tation. The F 3 generation consisted of 69 plants belonging to progenies of 4 mut. formosa plants of the F 2 generation. All were revolute-leaved, and the 62 grown to maturity all proved to be mut. formosa. Also, 2388 plants belonging to the F 3 progenies from mut. formosa plants of the F 2 generation, of the. cross mut.formosa X CD hyb. viscida were all revolute- leaved. Hyb. viscida is the form resulting from the cross Oenothera pratin- cola f. typica C X Oenothera numismatica (BARTLETT 1915 a). It is like f. typica C in all respects, except that, in addition to the pubescence nor- mally occurring on the flowers of Oenothera pratincola, it has the viscid pubescence of Oenothera numismatica. THE FI AND F 2 GENERATIONS OF THE CROSS F. TYPICA, STRAIN C X MUT. FORMOSA, STRAIN E j The Mendelian behavior following the cross mut. formosa X strain C has been demonstrated at length. The reciprocal cross, f. typica C X mut. formosa (afiFF X a'pff), is just as Mendelian in its segregation of the free factors for flatness, but the Mendelian segregation of factors finds no MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 7 chance to express itself in the zygote because of the ever-present a factor for flatness inherited from the pistillate parent. The only way for a revo- lute-leaved plant to occur in the F 2 or following generations of the cross is by an independent mutation from a to a in the presence of the recessive condition of the Mendelian factors for flatness. Only one-fourth of the plants of the F 2 generation, those with the constitution apff, are capable of becoming revolute-leaved by mutation. Apparently there is nothing to hinder mutation from a to a in strain C. But in pure strain C the change would not be indicated by outward sign, for the strain is homo- zygous for the Mendelian factors for flatness. That this change does sometimes occur is shown by the few revolute-leaved plants which occur in the F 2 generation of the cross f . typica C X mut. formosa (see table 8, and COBB and BARTLETT 1919, table 6). In table 8 there are 26 revolute- leaved plants in a total of 1654, or 16 per 1000. If the mutation to a' should take place in a plant of the FI generation (a&Ff), a 3: 1 ratio would occur in the F 2 generation. In the hope that this may sometime happen in the experiment garden, the cross f. typica C X mut. formosa has been repeated many times, and FI and F 2 progenies are being grown. THE FI AND F 2 GENERATIONS OF THE CROSS MUT. FORMOSA STRAIN E X F. TYPICA STRAIN E In the cross mut. formosa X f . typica E (afiff X aQff) the mechanism for Mendelian inheritance operates just as certainly as in the correspond- ing cross with f. typica C, but the two parents happen to be alike in the Mendelian factors under consideration, both being pure recessives, so the only type of inheritance which manifests itself is matrocliny, depending on the difference in factorial composition of the characteristic portions of the a and )8 gametes. This cross has been repeated successfully nine times, giving 305 plants in the FI generation (see table 9) and 628 plants, from four plants of the FI generation, in the F 2 generation (see table 10). All of the plants of both generations were revolute-leaved. The inheritance here is matroclinic, in contrast with the Mendelian inheritance in the cor- responding cross with strain C. The reciprocal cross, f. typica E X mut. formosa (afiff X otfiff), has not been successfully repeated since the publication (COBB and BARTLETT 1919) of the fact that this cross is also matroclinic, the number of revolute- leaved plants occurring in the progeny being no greater than might be expected from self-pollination of f. typica plants of strain E, the strain which regularly produces some revolute-leaved plants in every generation. GENETICS 6 : Ja 1921 8 FRIEDA COBB THE FI AND F 2 GENERATIONS OF VARIOUS CROSSES BETWEEN F. TYPICA M (THE NEW, MENDELIAN STRAIN) AND F. TYPICA OF STRAIN c, F. TYPICA OF STRAIN E, AND MUT. FORMOSA The data recorded in this paper concerning the FI and F 2 generations of crosses of eighteen different kinds between f . typica M (the new Men- delian strain from the cross mut. formosa X strain C) as one parent, and f. typica C, f. typica E, or mut. formosa, as the other parent, confirm the hypotheses of non-equivalent gametes and the presence of a pair of inde- pendent Mendelian factors in Oenothera pratincola. All f. typica M plants used in the crosses were self -pollinated to deter- mine whether they were homozygous or heterozygous. All of the flat-leaved types other than f. typica which occurred in the progenies of crosses are mutations regularly thrown by f. typica C, and some of them by f. typica E also. All of the revolute-leaved types which occurred, are regularly thrown by f . typica E and by mut. formosa. The cross mut. formosa X f . typica M (homozygous) (afrff X a$FF) gave in the Fi generation (see table 11) only flat-leaved plants (a'(3Ff), and in the F 2 generation (see table 12) a segregation of 3 flat-leaved plants (a'QFF and a'QFf) to 1 revolute-leaved plant (a'fff). The reciprocal cross, f . typica M (homozygous) X mut. formosa (a' X afiff), gave the same results (see tables 13 and 14). The cross mut. formosa X f . typica M (heterozygous) (aftff X a gave in the Fi generation (see table 15) progenies consisting of flat-leaved plants (a&Ff) and revolute-leaved plants (aftff) in approximately equal numbers, and in the F 2 generation from flat-leaved plants (see table 16) a segregation of 3 flat-leaved plants (a'pFF and a'flFf) to 1 revolute-leaved plant (a'fff). The reciprocal cross, f. typica M (heterozygous) X mut. formosa (a'ftFJ X a r pff) gave the same results (see tables 17 and 18). The cross f . typica E X f . typica M (homozygous) (a/3// X a'QFF) gave in both the Fi and the F 2 generations (see tables 19 and 20) only flat-leaved plants. It is known that the a of strain E frequently mutates to a', and a few revolute-leaved plants would therefore be expected in the F 2 progenies, by a combination of a and the recessive Mendelian factors. One-fourth of the plants of the F 2 generation, those with the constitution aftff, would be expected to become revolute-leaved by mutation with the frequency of mutation in pure strain E. But this did not occur. The only explana- tion that can be suggested is that because the germination percentage was much higher than was expected, the seedlings were very much crowded in MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 9 the seed-pans, and possibly revolute-leaved plants, which do not hold their own in a dense stand, died before the seedlings were counted off. This, however, does not seem likely, and the matter will be further investigated. The reciprocal cross, f . typica M (homozygous) X f. typica E (a'fiFF X aftff), gave in the FI generation (see table 21) only flat-leaved plants (a'fiFf) and in the F 2 generation (see table 22) a segregation of 3 flat-leaved plants (a'fiFF and cf&Ff) to 1 revolute-leaved plant (aftffl. The cross f . typica M (heterozygous) X f . typica E (a'fiFf X a$f) gave in the FI generation (see table 23) flat-leaved plants (a0Ff) and revolute- leaved plants (a'pff) in about equal numbers, and in the F 2 generation from flat-leaved plants (see table 24) a segregation of 3 flat-leaved plants (a'pFF and a'pFf) to 1 revolute-leaved plant (a'fff). The reciprocal cross is missing from the series. The cross f. typica M (homozygous) X f. typica C (a'&FF X a@FF) gave in both the FI and F 2 generations (see tables 25 and 26) only flat- leaved plants (a'pFF in both generations) . The reciprocal cross is missing from the series. The cross f. typica M (heterozygous) X f. typica C (a'QFf X aQFF) gave in the FI generation (see table 27) only flat-leaved plants (afiFF and a'pFf) and in the F 2 generation (see table 28) 11 progenies consisting entirely of flat-leaved plants (a'pFF) and 7 progenies showing a segrega- tion of 3 flat-leaved plants (a'pFF and a'pFf) to 1 revolute-leaved plant (a* ftff) . The ratio 11:7 does not approach as closely as would be expected the ratio of one dominant factor to one recessive factor in the gametes of the heterozygous pistillate parent of the cross. In this table there seems to be a shortage both of segregating progenies and of revolute-leaved plants in the segregating progenies, indicating that the pistillate parent and its progeny produce either fewer or weaker gametes bearing the recessive factor. The reciprocal cross is missing from the series. A summary of tables 5 to 28, inclusive, is given as table 29. VARIOUS PEDIGREES OF PLANTS USED IN CROSSES A record of the parentage of all of the plants used in this work is given as table 1. All plants not otherwise designated were f. typica. It may be noticed that, though different strains of Oenothera pratincola behave differently as to the mutations that they throw when self -pollinated, and in the way that they behave in crosses, all f. typica plants within a strain, no matter how complicated, by crossing or mutation, their pedi- GENETICS 6 : Ja 1921 10 FRIEDA COBB grees may be, act in the same way. For instance, an f. typica from mut. grisella from f. typica C behaves exactly as f. typica C with no mutations in its direct ancestry, at least during the period in which the strain has been carried in the garden. It may also be noted that all plants of mut. formosa appear the same and behave the same genetically regardless of extraction. Those used in crosses were of five different types of extraction: (1) directly from f. typica E, by self-pollination; (2) from f. typica E, first by mutation to angusti- folia, then to mut. nitidissima and finally to mut. formosa, all by self- pollination; (3) from the cross mut. formosa X f. typica E, by matroclinic inheritance; (4) from the cross mut. formosa X strain C, by segregation; (5) from crosses with both strain C and strain E, by segregation and matro- clinic inheritance, e.g., ( (mut. formosa X mut. latifolia C} mut. formosa X f . typica E) mut. formosa. In many thousand offspring, mut. formosa has produced nothing but revolute-leaved plants. DISCUSSION It may seem to those who are used to working with organisms in which clear Mendelian inheritance is the usual thing, that this case in Oenothera pratincola has been worked out with unnecessary elaboration. But several considerations should be borne in mind: first, that evident Mendelian inheritance is so rare in Oenothera that only two indisputable cases have been recorded, that of mut. brevistylis (DE VRIES 1901, p. 223; 1903, pp. 151-179, 429) and that of the dwarf mutation from mut. gigas (DE VRIES 1915 b) and that all instances deserve therefore to be thoroughly examined; second, that in this case the Mendelian inheritance is apparently modified by inheritance of another kind, working simultaneously with and inde- pendently of the Mendelian inheritance; and third, that the hypothesis of heterogametism put forth to explain this other type of inheritance needs further testing. If the explanations offered here are correct, we have, in addition to Mendelian inheritance masked by heterogametism, mutation masked by Mendelian factors. That the a of strain C can undergo mutation to a.' is shown by the presence of a few revolute-leaved plants in the F 2 generation of the cross f . typica C X mut. formosa. These plants derive their a from strain C, but have the dominant Mendelian factors which are present in strain C replaced by their recessive allelomorphs (see table 8). There seems no reason to doubt that this mutation in a occurs just as frequently in the presence of the Mendelian factors for flatness (i.e., in pure strain C) MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 11 as in their absence. If such is the case, Mendelian factors hide the muta- tion until a suitable cross occurs to remove the factors. This may have some bearing on the question whether crossing induces mutation; it may be that crossing merely makes possible an external expression, by removing concealing Mendelian factors, of changes which occurred long since in the germ-plasm and have been passed on from generation to generation giving no visible sign of their presence. In the same way a single mutation, the loss of one concealing factor, in a single chromosome, might bring to light a whole series of new forms. (The several revolute-leaved types that have occurred in the experiment garden can all be permanently concealed by the single pair of Mendelian factors for flatness.) This may explain in part the apparent periodicity of mutability. Perhaps the organism does not have increased tendency to change, but hoards actual changes until chance brings them to light, "gruppenweise," by the removal of inhibiting factors through a mutation. Though the condition of the Mendelian factors in Oenothera pratincola strains other than strains C and E has not been investigated, it seems probable, since they have given no revolute-leaved plants, that the other six strains are homozygous with regard to the factor for flatness. If strain E is the only one of the eight strains carrying the recessive factors, it might seem likely that it arose from one of the other strains by loss of a dominant factor, and consequent Mendelian segregation, rather than that the reverse change took place, were it not for the fact that strain E has produced in the experiment garden a mutation (mut. nitidissima) which, as shown by its behavior in crosses with mut. formosa, is a homozygous dominant in regard to the Mendelian factors for flatness. Whether both dominant factors were present in the original plant of mut. nitidissima and its FI generation is a question which cannot be answered, but it is clear that they are now present in the strain as it is carried on in the garden. That a pure dominant strain might arise from a pure recessive strain, or the reverse, leaving no heterozygous plants to tell the tale, seems pos- sible, for the heterozygotes, as they appear in the experiment garden, tend to have poor and irregular leaf development as young rosettes, and might easily be eliminated by natural selection. The modification of a factor for flatness in a single chromosome of a homozygous dominant, or the reverse modification in a homozygous recessive, would then be the only change in the germ-plasm necessary to produce one homozygous strain from the other. Some explanation should be made of the unusually low germination per- centages occurring in the cultures recorded in this investigation. As the GENETICS 6: Ja 1921 12 FRIEDA COBB character flatness vs. revoluteness is one which is evident in even very young seedlings, many sowings were made merely for the sake of recording the nature of the seedlings, and the seedlings were then discarded. In order to give all possible space and attention to the cultures intended for the summer garden at the time which is most favorable for planting, the sowings for seedling counts were made early and got out of the way. This meant that the seeds did not have as long a rest period as they apparently need before germination. Had they been planted a month or two later (in February and March instead of December and January) the percent- ages of germination would have been very much higher. In problems connected with comparative fertility and sterility it is of course quite necessary to know that all viable seeds are forced to ger- minate. The method worked out by DE VRIES (1915 a) and applied by DAVIS (1915) for forcing germination to completion, is especially suited to such problems. It will also throw light on the types that may be lost through selective mortality and selective germination rate when less thorough methods are used. This information is most valuable. But that the low germination percentage has no significant effect on the results of the experiments recorded here is shown by table 30. Here the ratios of flat-leaved to revolute-leaved plants are assembled from all of those cultures in which the expected ratio is 3:1, and arranged according to the germination percentage. It will be seen that three-fourths of the cultures have a germination percentage under 25 percent. But it may also be seen that the average of the ratios of cultures in which the germination is from 50 to 81 percent is no nearer the expected 3:1 ratio than are the averages from cultures with poorer germination. Even when less than 5 percent of the seeds germinated, the average is as close to the theoretical ratio as is that of any one of the five cultures with a germination over 50 percent, or as the average of these five cultures. That selective germination, in connection with the types with which this problem is concerned, occurs to any significant extent seems impossible. The only evidence of such selection is a slight excess of revolute-leaved individuals, especially noticeable when the germination is poorer. There are four tables which show in the total an excess of flat-leaved plants, ten which show an excess of revolute-leaved plants. In work with this species it has been noticed that when the percentage of germination is very low the percentage of mutations is very high (BARTLETT 1915 a). General observations lead to the conclusion that poor germination tends to bring the unusual types into prominence rather than to conceal them. MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 13 This paper records the case of a single unit character of the zygote, revoluteness, determined by a complicated set of phenomena: an allelo- morphic pair of factors (F and /), the dominance and recessiveness of which produce no effect on the zygote except when the particular muta- tional change from a to a has taken place; a mutation (a to a) occurring repeatedly, but concealed, as long as self-pollination continues, by the Mendelian factors FF, and Mendelian segregation concealed by matro- clinic inheritance dependent on heterogametism (a and /3 gametes). It is hoped that the case may help to throw light upon the seemingly peculiar behavior of the Oenotheras. SUMMARY 1. The male and female gametes of Oenothera pratincola are not alike. Each zygote is formed by the union of an a (female) gamete and a (male) gamete (except in rare cases which it is needless to mention here) and has the constitution . It, in turn, produces a (female) and /3 (male) gametes. 2. The a (female) gametes may undergo such mutation that, unless certain factors for flatness are present, the resulting plants are revolute- leaved. Such mutated gametes are designated a. 3. The |8 (male) gametes have no such possibility of producing revolute- leaved plants. 4. Strain C carries in both male and female gametes, but not in the characteristic a and |8 portions of the gametes, a freely segregating factor for flatness (F), in the presence of which revoluteness can not occur, even though the mutation to a has occurred. The constitution of f. typica strain C is afiFF. 5. Strain E does not carry in either male or female gametes the inde- pendent factor for flatness which occurs in strain C, but its allelomorph (/). The constitution of f. typica strain E is afiff. 6. Mut. formosa, a revolute-leaved mutation thrown by strain E, differs from f. typica strain E in that it contains a mutated a (a). The constitu- tion of mut. formosa is a'Qff. 7. There are two types of inheritance going on simultaneously and independently in Oenothera pratincola, matroclinic inheritance, connected with certain constant differences in factorial composition between male and female gametes, and Mendelian inheritance, connected with an inde- pendent segregation of factors carried by both gametes. 8. The cross mut. formosa X strain C (afiff X aftFF) produces only flat-leaved plants in the FI generation; in the F 2 generation there occurs a Mendelian segregation in the ratio of 3 flat-leaved plants to 1 revolute GENETICS 6 : Ja 1921 14 FRIEDA COBB leaved plant. This segregation has been followed to the F 4 generation. The f . typica plants descended from this cross have the constitution afiFF or a'0Ff, and are called f. typica M (homozygous) and f. typica M (hete- rozygous), respectively. 9. The reciprocal cross, strain C X mut. formosa (a$FF X a'pff) also gives only flat-leaved plants in the Fi generation; but in the F 2 generation there is no Mendelian segregation, though there occurs by mutation a small percentage (1.6 percent) of revolute-leaved plants. In this cross the inheritance appears to be matroclinic. 10. In reciprocal crosses between mut. formosa (a'pff) and f. typica E (apff) the inheritance is purely matroclinic, as the two parents are alike in regard to the Mendelian factors for flatness. 11. The results, recorded in the tables of this paper, of various crosses between f. typica M and f. typica C, f. typica E and mut. formosa are all such as to favor the hypotheses of heterogametism and the presence of a pair of Mendelian factors for flatness in Oenotkera pratincola. All the results obtained could be correctly predicted on the assumption that: f . typica C = a@FF f . typica E = afiff mut. formosa a'ftfj 12. The a of strain C may become mutated to a ', but in pure strain C this change can find no expression, because of the Mendelian factors for flatness for which this strain is homozygous. That the change does some- times occur here, as in strain E, is shown by the occurrence of a few revo- lute-leaved plants in the otherwise uniformly flat-leaved F 2 generation of the cross f . typica C X mut. formosa, plants in which the a portion of the constitution came from strain C and in which the Mendelian factors for flatness are replaced by their recessive allelomorphs from strain E. 13. The difference, with regard to Mendelian factors, between strains C and E is paralleled by the difference, with regard to the same factors, between mut. nitidissima, a type which has arisen in the experiment garden, and strain E from which it arose. Strain E is recessive; mut. nitidissima is a homozygous dominant. Evidently a homozygous dom- inant strain can arise from a homozygous recessive strain ; the reverse proc- ess has not as yet been known to take place in the garden. 14. It is concluded that mutation may be masked by Mendelian factors, and that the apparent induction of mutation by hybridization may be merely the first appearance of changes which occurred in the past and MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 15 were carried on unseen until their appearance in the zygote was made possible by the removal, through hybridization, of inhibiting Mendelian factors. LITERATURE CITED BARTLETT, H. H., 1915 a Additional evidence of mutation in Oenothera. Bot. Gaz. 59: 81- 123. 1915 b Mass mutation in Oenothera pratincola. Bot. Gaz. 60: 425-456. COBB, FRIEDA, and BARTLETT, H. H., 1919 On Mendelian inheritance in crosses between mass-mutating and non-mass-mutating strains of Oenothera pratincola. Jour. Wash- ington Acad. Sci. 9: 462-483. DAVIS, B. M., 1915 A method of obtaining complete germination of seeds in Oenothera and of recording the residue of sterile seed-like structures. Proc. Nation. Acad. Sci. 1: 360- 363. DE VRIES, HUGO, 1901, 1903 Die Mutations-Theorie. 2 vols., xii + 648 pp.; xiv + 752 pp. Leipzig: Veit & Co. 1915 a The coefficient of mutation in Oenothera biennis L. Bot. Gaz. 59: 169-196. 1915 b Oenothera gigas nanella, a Mendelian mutant. Bot. Gaz. 60: 337-345. MORGAN, T. H., 1918 Concerning the mutation theory. Sci. Monthly 6: 385-405. MULLER, H. J., 1918 Genetic variability, twin hybrids and constant hybrids, in a case of bal- anced lethal factors. Genetics 3: 422-499. APPENDIX-TABLES Explanations applying to all of the tables In table 1, all plants not otherwise designated were f. typica. F. typica M is the synthetic, Mendelian strain of Oenothera pratincola which arises from the cross mut. formosa X f. typica C (or other form with equivalent pollen) v The numbers in columns headed "Key number" refer to corresponding numbers in table 1. The numbers in columns headed "Parent plant" are the numbers of the individual plants in the progenies resulting from the crosses or self-pollinations recorded in table 1. * indicates that the seeds sown were from a single capsule. t indicates that seeds from two or more capsules were sown together. All of the flat-leaved types mentioned in analyses of cultures are regularly thrown by self-pollinated f. typica C, and some of them also by f. typica E. All of the revolute-leaved types mentioned are regularly thrown by f. typica E and by mut. formosa. Types other than f. typica, mut. latifolia, and mut. formosa mentioned in the following tables are: Mut. albicans (BARTLETT 1915 b, page 449). Mut. angustifolia (BARTLETT 1915 b, page 438). Mut. dimorpha, an undescribed mutation. Mut. ericacea, an undescribed mutation. Mut. fallax, an undescribed mutation, as a seedling very much like mut. nummularia, and thrown by the same strains. GENETICS 6 : Ja 1921 16 FRIEDA COBB Mut. gigas (BARTLETT 1915 b, page 443). Mut. grisea, an undescribed mutation. Mut. grisella, an undescribed mutation. Mut. nitidissima (BARTLETT 1915 b, page 440, table IV). Mut. nummularia, a mutation commonly thrown by strain C but never by pure strain E. (BARTLETT 1915 a, page 97; COBB and BARTLETT 1919.) Mut. revoluta (BARTLETT 1915 b, page 450). Mut. setacea (BARTLETT 1915 b, page 450). Mut. sub-latifolia, an undescribed mutation. Hyb. mscida, a hybrid of Oenothera pratincola X Oenothera numismatica (BARTLETT 1915 a, page 86). This form is like Oe. pratincola f. typica with the addition of the viscid pubescence of Oe. numismatica. TABLE 1 Record of the parentage of all of the progenies recorded in the following tables with key numbers. All plants not otherwise designated are f. typica I 2 E - 5 - 199 formosa - 28 formosa] X [ 4 C - 22 - 13 latifolia - 87 latifolia} 2 E - 5 - 199 formosa - 28 formosa X C - 22 - 13 latifolia - 87 latifolia 190 3 E - 5 - 199 formosa - 28 formosa} X [ 162 C - 22 - 13 latifolia - 87 latifolia} 4 E - 5 - 199 formosa - 28 formosa} X f 162 - 164 formosa] C - 22 - 13 latifolia - 87 latifolia] X C - 52 - 6 grisella 3 - 31 grisella - 73 - 3J (= 3 No. 164 formosa X f. typica C) 5 E - 5 - 199 formosa - 28 formosa} X [ 162 - 153 formosa] C - 22 - 13 latifolia - 87 latifolia} X C - 52 - 6 grisella - 31 grisella - 73 - 3J (= 3 No. 153 formosa X f. typica C) 6 E - 5 - 199 formosa - 28 formosa] X [ 162 - 30 formosa] C - 22 - 13 latifolia - 87 latifolia} X C - 52 - 6 grisella - 31 grisella - 73 - 3J (=3 No. 30 formosa X f . typica C) 2 These numbers are designated "key numbers" in subsequent tables. * An undescribed mutation. MENDELIAN INHERITANCE IN OENOTHERA PR AT IN CO LA 17 TABLE 1 (continued) 7 E - 5 - 199 formosa - 28 formosa} X \ 162 - 138 formosa] C - 22 - 13 latifolia - 87 latifolia} X C - 52 - 6 grisdla - 31 grisdla - 73 - 3j (=3 No. 138 formosa X L typica C) 8 E - 5 - 199 formosa - 58 formosa - 15 formosa} X 1 C - 52 - 6 groeKa - 25 - 1 - 4lJ 9 E - 5 - 199 formosa - 28 /omosa] X [ 162 - 164 formes*} C - 22 - 13 latifolia - 87 latifolia] X F 2 formosa E - 43 - 89 - 5 - 19J X C- 22 -7- 40 -5-9-2 ( = 23 No. 2 formosa X f . typica C) 10 C - 52 - 6 gr//a - 25 - 1 - 43 - 2 E - 5 - 199 formosa - 58 formosa - 15 /omosa] X X \ 63 formosa} E - 43 - 89 - 5 - ij ( = f . typica C X 25 No. 63 formosa) 11 C - 52 - 6 groetfa - 25 - 1 - 43 - 19J E - 5 - 199 formosa - 58 formosa - 14 formosa} X f X I 27 formosa} E - 43 - 89 - 5 - l] (= f. ty#ca C X 25 No. 27 formosa} 12 c- 22 -7-40-5-9-1] X ? 27 formosa E - 5 - 199 formosa - 28 formosa - 62 formosa} 13 C - 22 - 7 - 40 - 5 - 9 - 1 E - 5 - 199 formosa - 28 formosa] v X [ 162 - 164 formosa} C - 22 - 13 latifolia - 87 latifolia] X ) 16 formosa E - 43 - 89 - 5 - 19j (=f. typica C X 23 No. 16 formosa) 14 C - 52 - 6 grweKa - 25 - 1 - 43 - 18] E - 5 - 199 formosa - 58 formosa - 15 /orwosa] X > X [ 63 formosa} E - 43 - 89 - 5 - IJ (= f. ty#ca C X 25 No. 63 formosa) 15 C - 22 - 7 - 40 - 5 - 9 - 2] E - 5 - 199 formosa - 58 formosa - 15 /0rw0sa] X rwosaj X E - 43 - 89 - 5 - IJ (= L typica C X 25 No. 63 formosa) 63 f GENETICS 6 : Ja 1921 18 FRIEDA COBB TABLE 1 (continued) 16 C - 22 - 7 - 40 - 5 - 9 - 2 E - 5 - 199 formosa - 28 formosa} x X [ 162 - 164 formosa} C - 22 - 13 latifolia - 87 latifolia} X [ 3 formosa E - 43 - 89 - 5 - 19j (= f. typica C X 23 No. 3 formosa) 17 E - 5 - 199 formosa - 58 formosa - 35 formosa} X E - 43 - 89 - 5 - 19J 18 E - 5 - 199 formosa - 28 formosa} X ^ 162 - 153/0ri E - 5 - 199 formosa - 28 formosa - 62 formosa) (= 1 No. 165 i.typica M (heterozygous) X formosa) 45 E - 5 - 199 formosa - 28 formosa} X [ 4 - 164b] C - 22 - 13 Jatf/oWa - 87 latifolia) X E - 5 - 199 formosa - 28 formosa - 31 formosa) (= 1 No. 164b f . fy^ica M (heterozygous) X formosa} 68 formosa X 6 viscida 68 formosa X 1 viscida X 46 E - 5 - 199 formosa - 28 formosa} X \ 4 - 163bl C - 22 - 13 latifolia - 87 latifolia} E - 5 - 199 formosa - 28 /cmosa] X f 4 - 33 formosa) C - 22 - 13 te//0#a - 87 latifolia) (= 1 No. 163b i.typica M (heterozygous) X 1 No. 33 formosa) E - 43 - 89 - 5 - 1 - 21 E - 5 - 199 formosa - 28 formosa} X X } 190 - 4 - 2j C - 22 - 13 /a/t/0/ta - 87 latifolia) (= i.typica E X 2 No. 4 - 2 i.typica M (homozygous)) 48 E - 5 - 199 formosa - 28 X 4-64-9] C - 22 - 13 /a/*/Wia - 87 latifolia} X E - 43 - 74 - 41 - 45 - 2J ( = 1 No. 64 - 9 f . ty/wa M (homozygous) X f . ty^t'ca E) 4 A hybrid of Oenothera pratincola X Oe. numismatica (BARTLETT 1915 a, p. 86). This form is like Oe. pratincola f. typica with the addition of the viscid pubescence of Oe. numismatica. MENDELIAN INHERITANCE IN OENOTHERA PRATINCOLA 23 TABLE 1 (continued) 49 E - 5 - 199 formosa - 28 formosa} X \ 190 - 4 - 6] C - 22 - 13 latifolia - 87 latifolia] X E - 43 - 74 - 21 - 3J ( = 2 No. 4 - 6 f . ty/^'ca M (homozygous) X f . typka E) 50 E - 5 - 199 formosa - 28 formosa} X }> 190 - 54 - 2] C - 22 - 13 latifolia - 87 latifolia} X E - 43 - 72 - 5 - 6 - 12J (= 2 No. 54 - 2 f. typica M (homozygous) X f. typica E) 51 E - 5 - 199 formosa - 28 formosa X 190 - 4 - 2] X C - 22 - 13 latifolia - 87 latifolia E - 43 - 89 - 5 - 1 - 6j ( = 2 No. 4 - 2 f . typica M (homozygous) X f . typica E) 52 E - 5 - 199 formosa - 28 /0jw. 739 747 793 772 ll 741 750 795 782 1 I 742 750 804 830 1 750 755 808 854 d* * ^ 765 766 825 914 > ^ *M 765 806 836 jo *o ja 766 815 *S 1 778 818 ~ 3 S 800 897 3 g 808 917 -|