NATURAL swam mmwssr me smamxm GENE ma YELLOW 309v coma m gr 020mm. {QELANQGASTER \ Thesis for We Degree of M. S. MSCH‘iGAN STATE COLLEGi H. Lee Meyers, 11'. 1953 This is to certify that the thesis entitled Natural Selection Against the Sex-Linked Gene for Yellow Body chlor' 1n DroeoPhila melanogaster presented BI] Harvey Lee Meyers has been accepted towards fulfillment of the requirements for "0 S 0 degree in ___Ml° MAM Major professoxr Date Augufit 7’ 19530 O~169 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE 03": 4 W ll 3’” 3:53 3 ll MSU Is An Affirmative Action/Equal Opportunity Institution c:\circ\daedm.pm3-p.1 NATURAL SELECTION AGAINST THE SEX-LINKED GENE FOR YELLOW BODY COLOR IN DROSOPHILA MELANOGASTER By H. Lee ‘ngyers, Jr. A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Zoology 1953 {‘w -' p; THESIS AC KF’W'TLEDC .CD'TEN TF3 The author is greatly indebted to Dr. H. R. Hunt, whose insight and logic led the wav to improved methods and prevent- ed many serious mistakes. Deep gratitude is also extended to Mr. Fav Pinckney, whose reliable work unfailingly provided the author with equipment and medium whenever it was needed. The author also gratefully thanks Dr. K. A. Stiles and Dr. R. F. Keller, whose careful and unbiased advice allowed independent decisions. . Without the timely prompting and kind concern of Mrs. Dale Henderson (Miss Mac), this thesis would never have been written. The author also tenders apologies and gratitude to Sally, to Ed, and to all those friends and parents whose patient ears were never closed to constant vocal dissertations concerning fruit flies and evolution. TABLES OF CONTENTS INTRODL‘CTI‘jprJO0.00.0.0O.00.000.00.000...OOOOOOOOOOOOOOO.0.. REVIEYN OF LITERATIJREOOOOOOOOOOOOOOOOOOOOOOOOOO00.0.0.0... methOdS 0f InveStigationOOOOO0.0000000000000000000000 1 1., Experimental Results with Equilibrium and Selection.. h 5 B‘IETH()I)OLOGYCOCOOOOO0.0.0.0000....00...O....\.O...O........O 8 8 platerialSOOOODOC...0..OO0.0.0....OOOOOIOOOOUOCOOOOOOO ProcedureSOOOOOOOOOOOOOO0......OOOOOOOOOOOOOOOOOO.0.. 9 TeChniqu.SOOOOOOO0.00.00...O0....00.00.000.000000000011 DATA...00.0.0.0....0O.0.00.0000...00.00.000.000...0...000.13 Calculation of Expected Frequencies..................13 POPUlatj-On cag.3 O I O O C O O O O O O O O O O O O O O O 0 O O O O O O O O O O O O O O O .16 Assortative Mating. O O O O O O O O O O O O O O O O O O O O C O O O O O O O O O O O O .21 Viability.0.0....O..0.00.0.00...00.00.000.0000000000025 DonSitYOOOOOOOOOOOOOOOOOO..0.00.0.0000...0.0.0.00000026 Rat. Of Development 0 O O O O O O O O O O O O O O O O 0 O O O O 0 O O O O O O O I O O .28 DISCUSSIONOOOOOO00.0.0000...O00.00.0000...COOOOOOOOOOOOOOOAS Types of Selection and Deve10pmental Stages..........45 mortalitYOOOOOCOOO00.0.0000...C...0.0.00.000000000000h6 Differential Gamete Contribution.....................h7 FertilitYOOCOOOOOOOOOCOOOOO0.000....00.0000000000000049 Rate or Development. 0 O O O O 0 O O O O O O O O O O O O O O O O O C O O O O O O O O 0&9 COflClUSionSQ O O O O O O O O O O O o O O O O O O O O O O O O O O O O O O 0 O O O O O O. O O .50 SmmRYOOOOOOOOOOO.OOOO0.0...OOOOOOOOOOOOOOOOOOO0.0.0.0....52 BIBLIOGRAIJHYOOOOOOOOOO.0O.O.OOOOOOOOOOOOOOOOOOOOO000......53 TABLE 1. 2. 3. LIST OF TABLES PA Sample Case of Sex-linked Equilibrium Frequencies.. Expected Phenotypic Frequencies, No Selection...... Expected Phenotypic Frequencies in "Scarcity" With NO SelectionOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO PrOportion of Color and Sex PhenotYpes among Total Flies in "Abundance"......................... PrOportion of Color and Sex Phenotypes among Total Flies in'"Scarcity".......................... Proportion of Yellow Flies Among Total Flies, "Abundance" and "Scarcity"......................... Proportion of.Yellow Flies in Each Sex, 'tAbundance"...0.00.0000...000.000.00.000...00...... PrOportion of Yellow Flies in Each Sex, "Scaer-ty".OOOOOOOOOOO.0O...OOOOOOOOOOOOOOOOOOOOOOI Assortative Mating: Female Preference.............. Assortative Mating: Male Preference................ Expected Frequencies of Genotypes: "Abundance", from Merrell's Coefficients........................ Viability ReSUJ-t’si0......00.0.0.00000000000000.00... 30 31 32 33 3h 35 36 36 37 38 LIST OF FIGURES FIGURE PAGE 1. Frequency of Yellow Females Among Total Flies, "Abundance" and "scarCityuoooo00000000000000.0000. 39 .2. ‘Frequency of Yellow Males Among Total Flies, fiAbundance" and "Scarcity"........................ 40 3. Proportion of Yellow Flies in Each Sex, "Abundance"‘.0.........Q.....Q...‘................ #1 h. Proportion of Yellow Flies in Each Sex, t'Scarc1ty'tOOOOOOOOOOOOOOOOOOO..0...OOOOOOOOOOOOOOO 1‘2 5. Proportion of Yellow Flies Among Total Flies, "Abundance" and "Scarcity"........................ LB 6. Selected Frequency Curves: Total Yellow Among Total FlieSCOOOOOOOOOOOOOO00.0.0000000000000 ha 1.‘ ‘n‘ l, r lllil [.1 fl. 1 I l INTRODUCTION The experimental study of selection is necessary for an understanding of the process of evolution. Certain factors may modify the effect of selection against a recessive gene. Replacements of,the loss of this gene may occur by migration or mutation. Gradual acquisition of modifier genes may less- en the intensity of selection. Superiority of heteroZygous forms can offset selection pressure against homozygotes. If these phenomena occur at constant rates, then a state of equilibrium may occur, in which the gene frequency is fixed. Equilibrium for autosomal factors is reached in the first generation following random breeding, providing there is no selection pressure (Pearson, 1904). Refinements of this principle by Hardy (1908) assumed equal fertility, random mating, and independence from sex. A sex-linked gene is not independent from sex, however, since the female possesses a heterozygote class and the male does not. But Hogben (1946) and Li (1948) have shown that Hardy's Law of gene frequencies (p2-+ 2pq‘+ q2=:l, Hardy, 1908) still applies at equilibrium in the case of sex-linked genes in the female. The proportion of hemizvgous males showing the sex- linked recessive trait is theoretically the square root of the proportion cf females with the same trait: Sex male female Genotype Y y YY YY YY 2 2 Frequency p q .p 2pq q This equilibrium for sex-linked traits is not reached in one generation of rendom breeding. If the proportion of dominant males in the initial population is equal to the proportion of dominant gametes among female gametes then under random breeding these proportions are fixed (Robbins, 1918). In mathematical terms, equilibrium occurs in the first genera- tion after random breeding only if 2 P = P ‘* Pq, 2 01‘ q ‘+ PQo .o n For all caseslin which the gene frequency in the males is not equal to the gene frequency in the females, equilib- rium is approached in an oscillatory manner (Li, 1948). Ac- cording to Li, approximate equilibrium is attained in about fifteen generations. The difference,(d), between the gene frequency in each sex is halved in each generation, n d = (-1/2) d n 0 (Li, 1948). Thus, the difference between the frequencies of a recessive sex-linked gene in the males and females of any generation (dn) is a function of that difference in the initial genera- tion (do)' In Hogbens' terminology, an = (RYO + R‘IlH-l/Z)“ (Hogben, 19h6) That is, the difference between the frequencies of a recessive sex-linked gene as.between the males and females of any gener- ation (dn) is a function of the frequencies of this gene in the first and second generation males. After three or four generations, the mean values of any two consecutive frequen- cies are so close to their limiting values that equilibrium can be assumed (Hogben, l9h6). All these mathematical ideas assume no selection. It was suggested by Dr. Hunt that the effect of selection on a sex- linked trait could be tested in actual populations of Drosoph- ila. Thus in this experiment with Drosophila melanogaster, the wild-type fly was introduced into each of two cages con- taining pOpulations of recessive flies. Since the recessive allele is usually less advantageous to the species than is the dominant (Boyd, 1953), it was suspected that the frequency of recessive flies would decrease in each generation as a conse- quence of selection. Therefore, a differential food supply was employed in which one cage received approximately half as much food as the other cage. The generations were kept sepa- rate in order to measure the generation-to—generation frequen- cy changes. REVIEW OF LITERATURE Equilibrium has been demonstrated in Drosophila by many researches. Investigations of equilibrium conditions have shown that sometimes the heterozygous form is superior to the homozygotes. Ingenious experiments (Dobzhansky, 19L?) with chromosome inversions in Q. pseudoobscura have demonstrated that, although the types of eggs laid are in prOportions agreeing with Hardy's Law, the inversion heterozygotes are more frequent than expected. The adaptive superiority of the heterozygote thus protects both types of chromosomes from se- lective pressure and results in a balanced polymorphism. In Drosgphila, nearly all inversion heterozygotes are adaptively superior to their corresponding homozygotes (Dobzhansky, 1951). Thus equilibrium conditions have been attained in many labor- atory experiments. ' ’ An analysis of the color pattern in Q. polymorpha (da Cunha, 19h9) showed equilibrium due to a differential mortal- ity between the egg and larval stages which favored the heter- ozygote. Dubinin and Tiniakov (1946) found heterozygote su- periority in Q. funebris to be the main cause of equilibrium. Similar results were reported by L'Heritier and Teissier (1933) and L'Heritier (1937) in D. melanogaster, by Reed and Reed (19h8) in D. melanogaster, by Freire-Maia (l9h9) in 2. montium, by da Cunha, Burla, and Dobzhansky (1950) in Q. fiillistoni, and by Shell (1952) in 2. melanogaster. Quite often such heterozygote superiority is accompanied by stronger selection pressures against one of the alleles than against the other. Thus elimination of such an unfavored gene may occur. Drosophila laboratories are constantly beset by contamination of mutant stocks by their wild alleles. Such contamination often leads to elimination of the mutant gene. The gene for vestigial wings, for example, may disappear rap- idly when competing against its wild allele for normal wings (Spencer, 1932). ‘ A sex-linked recessive gene would enjoy even less protec- tion by the heterozygous state, since the males are never het- erozygous. The low incidence of mutant genes in the sex chromosomes of wild-caught Drosophila, as compared with those in the autosomes, is evidence of this (Huxley, 19h3). Thus the sex-linked gene, yellow, was strongly selected against in populations ony. melanogaster under optimum food conditions (Ludwin, 1951). Reed and Reed (1950) found that the sex—linked gene for white eyes in 2, melanogaster decreased from a fre- quency of .500 to zero in twenty-five generations. Their ex- cellent work showed selective mating to be the primary cause. Equilibrium was not attained due to failure of white males to breed as often as wild males. Neither of these latter two experiments with sex-linked genes involved a generation-by- generation frequency count. The methods used by these investigators have been various. Most approaches were attempts to simulate nature, wherein generations interbreed. These approaches utilized population containers which supplied fresh medium for the flies a desired time intervals. Pioneer work on cages of this tvne was per- formed by L'Heritier and Teissier (1933) and L'Heritier (1937). Later modifications (Dobzhansky, 19h?) improved the efficiency of these cages. Dobzhansky's cage featured a detachable glass top on a wooden box with glass or screen sides. Fifteen holes in the bottom served to furnish a continuous supply of fresh food and to remove larvae and pupae for examination. A dif- ferent technique was employed by Reed and Reed (19h8, 1950). Two interconnected bottles formed the population unit. Exam- ination was accomplished by etherization, removal, classifica- tion, and replacement of the "groggy" flies. Although these small bottle-units yielded fewer flies, the use of many units permitted random sampling. Variations of this two-bottle unit were used by da Cunha (l9h9) and Ludwin (1951). Cages and units of these types may allow approximation to free-liv- ing populations, but certain disadvantages also exist. Fly excreta may accumulate, mites may infest the population, or excessive dryness may occur (Wright and Dobzhansky, l9h6). For generation-to—generation frequency counts, a differ- ent type of cage is desirable. Fly populations in such a cage are killed and classified after laying the eggs which would develop into the flies of the next generation. These eggs are removed in order to clean the cages. Then the eggs and fresh food are replaced. This technique removes possibility of mite infestation and excess excreta accumulation. Shell (1952) used glass-sided cages of this type for generation-to-gener- ation studies of the gene for sepia eye in Q, melanoggster. Similar cages were employed in this experiment. METHODOLOGY Materials All operations were undertaken in the constant tempera- ture room of the Natural Science Building at Michigan State College. The temperature was kept at 79°: 1°F.(26.C.). Two cages with identical dimensions were employed to obtain popu- lation data. Each cage measured thirteen inches high, eight- een inches wide, and twenty-six inches long. Glass panes held by a frame of one-inch wood stripping formed the sides and top of each cage. Access to the interior was provided by sliding the front pane up its grooved stripping. Each cage had a plywood base, three-fourths inches thick, thirty-two inches long, and twenty-four inches wide. The food of the flies was contained in Petrie dishes. The food consisted of a medium with the following ingredients: 10 liters of water 110 grams of agar 8 grams of Moldex 1/2 lb. baker's yeast (dissolved in water) 350 cc. of unsulphured molasses 350 cc. of Karo 1000 grams of corn meal This medium was stored in Petrie dishes in the refrigeration room at Ad’F. A strip of sterile paper toweling was laid on the surface of the medium in each Petrie dish to provide a place for larvae to pupate. Twenty-four hours prior to actual insertion in a cage, the Petrie dishes were removed from re- frigeration. This procedure permitted initiation of yeast growth and brought the medium to room temperature. The sex-linked gene for yellow body color was chosen for use in this experiment. This gene has occurred spontaneously in Q. melanogaster numerous times (Morgan, Bridges, and Sturt- evant, 1925). It is at locus 0 of the X chromosome. A simi- lar gene exists in the X chromosome in Q. simulans, Q. virilis, Q. obscura, and Q. willistoni. In 2. melanogaster, the yellow phenotype is easily distinguished from the wild phenotype. The body color is rich yellow, while the hairs and bristles are brown with yellow tips. The hairs and veins of the wings are yellow. The phenotype is also classifiable in the larval stage. Larval setae and mouth parts are yellow to brown in color (Bridges and Brehme, 19th). Procedures The line of flies using relatively more food has been termed "Abundance"; the line using relatively less, "Scarcity". The only differences in the procedure between the two experi- ments were a slight variation between initial frequencies of the yellow gene and the number of dishes of medium used for each generation. The "Abundance" line was started in half-pint milk bot- tles and transferred to its cage after the P1 parents were removed. This P1 generation comprised five wild males, forty- 10 seven yellow males, and forty-nine yellow females placed in a half-pint milk bottle and allowed to breed. After the ap- pearance of the F1 pupae, these P parents were removed. They l were classified as to body color and sex, then counted. The bottle containing the F1 pupae was placed in the "Abundance" cage with three Petrie dishes. The food dishes were placed equidistant from the pupae bottle and from each other. A folded paper barrier was placed between the larva-pupa bottle to prevent larvae from crawling to the food dishes. The e- merging F1 flies bred and fed in the three food dishes. Some breeding occurred in the larva-pupa dish, also. F1 imagoes were permitted breed for seven davs. They were then removed in order to prevent intermixture of two successive generations. Two days before the F generation emerged, one 2 dish of F2 larvae and pupae was removed to be used in producing the F2 imagoes. The larvae and pupae in the remaining dishes and the bottle, along with the F parents were etherized and removed. These dishes and the bittle were discarded. The F1 parents were classified and counted. The cage was cleaned and aired, then the one F2 larva-pupa dish was replaced. Three Petrie dishes containing fresh medirm were added as above. Thus the parents of each generation were classified apart from their offspring. This cage technique was repeated three gener- ations after the appearance of the last yellow fly. The "Scarcity" procedure was carried out in similar fashion. The F1 generation consisted of four wild males, one wild female, one hundred and twenty-five yellow males, and ninety yellow fe- ll males mated in a half-pint milk bottle as with "Abundance". The inclusion of the wild female was due to an error of clas- sification, and occasioned the future use of the binocular scope for rechecking classes. Tha experiment was not invali- dated, however, since the initial gene composition was known in both lines. After the appearance of the F1 pupae, these P1 parents were removed and classified. The larva-pupa bottle was placed in the "Scarcity" cage. Only one dish of food was added. F1 flies emerging from the bottle bred and fed in this dish. After seven days, this F2 larva-pupa dish was removed to be used in producing the F2 generation. The F1 parents were etherized and classified, and the cage was cleaned. The F2 larva-pupa dish was returned along with one dish of food. This procedure was repeated until after three generations fol- lowing the appearance of the last yellow fly. ,Techniques Removal of flies and dishes was accomplished in the fol- lowing manner. First the case's front pane was lifted several inches. One side offshe opening was covered with a piece of cardboard to prevent escapes, while a dish of ether was quick- ly inserted. One Petrie dish containing the larvae and pupae of the next generation was removed. The pane was closed and sealed with tape to increase etherization efficiency.‘ When the flies were completely etherized, the remaining dishes were removed and examined for imbedded flies. The flies on the 12 paper-covered floor of the care were then swept onto a paper and removed. Thev were placed in a one half~pint bottle until examination. The cage was then cleaned with Bon Ami and 70 percent al- cohol, and a new floor of sterilized paper was provided. The previously-removed Petrie dish containing the eggs and larvae producing the next generation was replaced. Dishes of fresh food were added. Special care was given to prevent contamina- tion at this point by stray flies in the room. The pane was carefully closed, and the cage was placed in its rack.' After both cages had been emptied and restocked in this manner, the dead flies were classified and counted. For classification the dead flies were removed from thm container bottles and separated into classes on white paper in natural light. Classification should be performed immedi- ately, since delay brings ahout shrinkage and deformation of the-flies due to dessication. The yellow body color may be difficult to distinguish in immature flies, but the yellow wings are easily seen in all stages after eclosion. The flies were separated by the naked eye according to sex and body col- or. Then all four classes were re-examined by passing them under a binocular scope. Each class was then counted twice, and the numbers and percentages were recorded. 13 DATA The unique gametic relations necessitated by sex-linkage can be analyzed by examination of the formulas developed by Li (1948). Let p : dominant male frequency. q = recessive male frequency. r a homozygous dominant female frequency. 23 = heterozygous female frequency. t = homozygous recessive female frequency. Then, 2 rn : pn(pn-l)ooooooop This means that the frequency of the dominant homozygous fe- males of a given generation (rn) is the product of the fre- quencies of the dominant males of that generation (pn) and the previous one (pn_1). Further, at equilibrium the frequen- cy of the homozygous dominant females will be the square of the dominant male frequency (p2). The homozygous recessive female frequency of a given gen- eration (tn) is equal to the product of the frequencies of the recessive males of that generation (qn) and the previous one (Qn-l)’ When equilibrium is reached the frequency of the re- cessive females is the square of the recessive male frequency (qz). Thus, 2 tn = qn(qn-l)oooooooq lb The frequency of the heterozygous females of a given gen- euation l2sn) is equal to the sum of the product of the domi— nant male frequency of that generation and the recessive male frequency of the previous generation plus the product of the frequencies oftshe recessive males of that generation and the dominant males of the previous generation: an g pn(qn_1)«+ qn(pn_1)....2pq In actual practice, the frequency of heterozygotes is found by calculating the frequency of both homozygotes and subtrac- ting that from unity: 23D = l - (rn + tn) The male frequencies show the typical oscillatory pattern that is masked by heterozygosity in the females. The frequen- cy of either the dominant or the recessive males of a given generation is equal to the sum of half the frequencv of the heterozygous females of the previous genenation and the fre— quency of the respective homozygous females of the previous generation: pn = rn-l + Sn-l qn = tn-l * 3n-l 15 The male frequencies can also be calculated by other formulas: pn = l/2(pn--l + pn-Z) (Hogben, 19h6) The oscillations in the frequencies of recessive males in suc- cessive generations are shown by Hogben's formula; on = 1/3cao + 2q ) + 2/3(qo - q1)(-1/2)n 1 Table 1 shows an example in which the initial frequencies of the males have been.arbitrarily selected. TABLE 1 SAEPLE CASE OF SEX-LINKED EQUILIBRIUK FREQUENCIES Generation Frenuengy .0 .I {7 .2 .h .5 .6 .7 .8 .9 1.0 qo 1.0 q]. ’O.2 q2 ' 0.6 q . O. 3 h qh 0.5 05 ' Col-[>5 16 The dominant male frequencv is always unity minus the re- cessive male frequency. Both frequencies approach equilibrium values at the same rate then, and the equilibrium frequency for males can easily be calculated: 3? = q0 +2/3 (so+ to) - qo (Li,191.8) The equilibrium frequency is equal to the sum of the original male frequency and two-thirds of the difference between the female and the male gene frequencies in the 0 generation. In the case of Table l, ’3" = 1.000 + 2/3 (.200) - 1.000 = .257, Application of this formula to this experiment shows that the equilibrium frequency of the yellow gene in the case of no mutation, no migration, and n0 selection would have been 0.968 in "Abundance" and 0.982 in "Scarcity". Table 2 shows the expected frequencies in "Abundance"; Table 3 shows those of "Scarcity". The actual gene frequencies, however, must be inferred from the observed phenotype frequencies. These are r- considered next. Population Cage Data, Information from the ponulation experiments has been com- pletely presented in the form of tables and graphs for both cages and separately. Both "Abundance" and "Scarcity" showed frequency changes similar in direction but slightly different in intensity. 17 The frequency of yellow phenotypes in both sexes in "Abundance" dropped significantly (See Table a.) in the gen- e ation 2. The corresponding increase in wild flies was ac- complished solely by the reproductive ability of the hetero- zygous females in generation 1. Such a reproductive superior— ity indicated that the yellow gene might be eliminated in sev- eral generations. However, the yellow flies became more nu- merous each generation until they almost regained their orig- inal frequency by generation 6. Thereafter, a steady decrease of yellow occurred, until generation 18, when no yellow flies appeared. Since the three subsequent generations contained no yellow flies, it was assumed that the gene had been eliminated. The initial erratic behavior of the yellow frequency may be due to certain uncontrolled environmental factors. Humidity changes occurred which may have affected the mutant genes. Also, food control varies due to different-sized populations living on the food (Reed and Reed, 19h8). Such factors prob- ‘ably affected the flies showing the mutant trait more than the wild-types, since characters determined by mutant genes are more modifiable by environmental changes (Plunkett, 1932). Table 5 shows a similar elimination of the yellow gene in "Scarcity". However, the loss of yellow was accomplished much more quickly than in "Abundance". The sharp decrease in yel- low frequencies in both sexes in the second generation appar- .ently occurred through lack of matings between yellow males and yellow females. This is indicated by the relative scarcity of yellow females in generation 2 (only 2.6 percent of the 18 total flies). As in "Abundance" however, a period of recovery occurred. The recovery peak was only half that of "Abundance". Generation 10 shows an abrupt elimination of yellow phenotypes in both sexes. Such abrupt elimination of detrimental alleles is one aspect of the "Hagedoorn effect". This phenomenon per- tains to the behavior of detrimental allelomorphs in the low frequency ranges. Increased random deviations either eliminate the factor or throw it up into a higher frequency range where stronger selection pressures exist (Kemp, 1929). Thus the in- creased randomness of frequency changes in yellow phenotypes at low frequencies indicated a lower selection pressure in "Abundance" than in "Scarcity". "Scarcity" yellow phenotypes, however, were completely eliminated in both sexes, and from a .higher frequency. This pointed to a differential selection pressure in the two populations. Graphical comparison of the frequencies in "Abundance" and "Scarcity" illustrates the process of differential elimi- nation of yellow phenotypes in each sex. Figure 1 compares the proportion of yellow females in each generation in both popu- lations. It can be seen that similar changes took place in each generation until the sixth generation. At this point the "Scarcity" yellow female proportion dr0pped considerably while the "Abundance" increased to its highest peak. The yellow male pr0portions in both populations, as shown by Figure 2, show even more concordance up until generation 5. Such con- cordance may occur through extraneous environmental factors 19 simultaneously affecting both populations. For this reason the generations were staggered; the second generation in "A- bundance" coincides in time with the first in "Scarcity". The initial dip and recovery in both populations may represent some factor inherent in gene competition. We are unable to explain the abrupt rise in the frequen- cies of yellow flies from the fourth to the sixth generations of the "Abundance" experiment,\and a similar rise from the third to the fourth generations of the "Scarcity" series. Therefore, to estimate the rates of selection pressure these initial increases in the proportions of yellow flies will not be considered. Consider the pr0portion of yellow flies among the total flies as shown in Figure 5 and tabulated in Table 6. A com- parison of elimination rates beginning at phenotype frequen- cies of similar magnitude should give a general picture of any differential selection pressure. Elimination in "Scarcity" occurred during five generations following the generation 5 frequency of yellow, 38.6 percent, while in "Abundance" com- plete elimination did not occur until nine generations after a yellow frequency of 3h.6 percent in generation 9. In other words, the mean elimination rate from the initial selected frequency level in "Scarcity" was -7.72 percent per generation. The mean rate from a similar frequency level in "Abundance" 'was only -3.8bpercent. Figure 5 illustrates the changes in the proportions of yellow flies in both p0pu1ations. The-right hand column in Table 6 gives the t values for the differences 20 in the percentages of yellow flies in the two populations. All t values over 1.96 can be considered statistically signif- icant (Dixon and Massey, 1951). In sex-linked‘equilibrium the square of the frequency of yellow males should theoretically equal the frequency of yel- low females in the same generation. In Tables 7 and 8 the right-hand column compares these generation values. At low frequencies, the yellow female frequencies are consistently lower than expected. This occurred in both "Abundance" and "Scarcity". Differential mortality in the yellow females or mating discrimination against the yellow males could cause such a phenomenon. Graphical representations of the frequencies in Tables 7 and 8 illustrate generations in which selection may have "relaxed".' Generation frequency changes which are opposite in direction in each sex are typical of the approach towards equilibrium with non-adaptive sex-linked genes. .Such sex- linked fluctuations should be more evident where selection had "relaxed". Perhaps this is the case in generations 3, h, 5, and 1h in "Abundance" (Figure 3) and in generations h, S, and 8 in "Scarcity" (Figure 4). Food deficiencies in "Scarcity" may have prevented more such unmaskings of sex-linked fluctu- ations. 21 Assortative Mating The following data strongly suggest the presence of as- sortative mating. This phenomenon has been reported in yel- low flies by Tan (l9h6), Sturtevant (1929), Rendel (l9h5), and Merrell (l9h9). All these authors have reported a def- inite discrimination against yellow males on the part of both types of females. Matings were devised to test these findings.. Two types of tests were employed; male preference and female preference. In the male preference tests, five males of one type (yellow or wild) were placed with five yellow females and five wild females. The females were all virgins and about twenty-four hours old. All of the wild females in the experiment were heterozygous. After five hours together, the fifteen flies were removed, and each female was placed in a separate food bottle. A record was kept of the offspring of each fertilized female. Fertilized yellow females produced all yellow males, and fertilized heterozygous wild females produced half yellow and half wild males. Twenty-one successful matings were ob- tained in these male preference tests. Table 10 shows the results. The second type of experiment tested female preference. Five virgin yellow females were placed with five yellow and five wild males. Also, five wild females(heterozygous) were placed with five yellow and five wild males. In the case of the yellow female preference tests, the-phenotypes of the F1 22 females indicated the successful male. If all the F1 females were yellow, then the successful male was yellow; if all the F1 females were wild, then the wild male was the successful fly; and if there occurred both wild and yellow F1 females, then both male types had fertilized the female. Twenty—three yellow females were fertilized, and the wild males were suc- cessful in twenty-one. In the case of the heterozygous wild female preference tests, the successful male was again indicated by the pheno- types of the F1 females. If the F1 females were all wild, then the wild male was the successful male; if half the F1 females were yellow and half were wild, then the yellow male was successful. All twenty-four fertilized heterozygous fe- males produced all wild Fl females, which indicates that no yellow males had ever been successful., The results are shown in Table 9. Similar results were obtained by Ferrell (19h9) with tests of yellow and wild preferences in Q. melanogaster. Test re- sults can be used to determine mate selection coefficients. Such coefficients in turn can be applied to equations which calculate expected gene frequencies. Mating coefficients have been worked out by Merrell (1950). The equations are as fol- lows: number of yellow male-9 yellow female matings. number of yellow male-9 wild female matings. number of wild male-o wild female matings. number of wild male-9 yellow female matings. a: 03 u Illn' relative frequency of matings by females. - relative frequency of matings by males. 53»? I H 23 The coefficient of female mating is a ratio of fertilized yel- low females to fertilized wild females, MngA+BAo BE + AB The coefficient of male mating is a ratio of successful yel- low males to successful wild males. MmzAA'I‘AB. m Applying the coefficients to the genotype equations, (pages 12-14) qn ; Sn-l* Mf(tn:£1 (l'tn-l) Mf(tn-l) T3 :3 II |-‘ I .0 :3 rn = (pn-l) (pn) Pn-l 4' 741mm,“)? tn : Mm(qn_1)(qn) 1311-1 1’ Mm( qn_1) 23H 3 1 - (rn + tn) 24 When these equations are in turn applied to population data, comparisons can be drawn concerning the selection pres- sure in the actual populations. The material in Tables 9 and 10 yield the following assortative mating coefficient values: M = 1.1 M 3 0.08 However, Merrell's work with the same gene involoved A}? fertilized females compared to only 68 in our experiment. Since the tests in this experiment do not contradict Merrell's results, the relative mating coefficients used for frequency calculations were derived from his tests: Mf l.h 0.1 (Merrell, 19h9) Both Merrell's coefficients have been incorporated into the formulas shown on the preceding page. For comparative purposes these mating coefficients must be applied to generations which show little or no selection abatement. Thus, the last thirteen generations of "Abundance" have been chosen for comparison. When Merrell's coefficients are applied to the probable genotype frequencies of generations 25 6 to 18 in "Abundance", the resultant expected frequencies are as shown in Table 11. The irregular changes in "Scarcity" do not lend themselves to ready comparison with Merrell's coefficient-curve (Figure 6), but the general trend is sim- ilar. Thus assortative mating is one possible cause for the elimination of yellow genes, but the difference between "Abun- dance" and "Scarcity" is probably due to other factors relat- ing to the food deficiency in "Scarcity". In view of this, tests were conducted in regards to differential viability or O \/ Vigor between the egg and imago stages. Viability When virgin heterozygous females are mated with yellow males, four classes of offSpring are expected. These four classes, grey males, yellow males, grey females, and yellow females, are expected in equal proportions. Any deviations from this l:1:l:l ratio might indicate a differential selec- tion in the form of reduced vigor, increased mortality, or lengthened rate of development. Tests were made in half-pint milk bottles. The number of parents was varied in order to discover whether increased population pressure was a factor in the selection. No significant deviation from expected sex ratios was noted. Therefore, both sexes of yellow flies were included in a single class with an expected frequency of .500. Table 12 shows that in progeny from one female up to ten fe- w 26 males no significant selection differential exists. The in- formation concerning the number of progeny per female per day is similar to Pearl's findings (Pearl and Parker, 1922) con- cerning population density. Pearl concluded that fewer eggs were laid in dense populations (Pearl, 1927). If the density effect on egg production were differential, then selection could operate. Density The food deficiency in "Scarcity" was due to the fact that all the eggs of each parent generation were laid in one Petrie dish while the eggs of each "Abundance" parent genera- tion were laid in three Petrie dishes of food, one of which was selected for the next generation. Thus the food dish in "Scarcity" fed more flies than the dish in "Abundance". Since all the Petrie dishes used in both cages contained approxi- mately the same amount of food, then there was less food per fly in "Scarcity". To check the effect of density, the total pOpulation was correlated with the decline of yellow frequency in each gen- . eration. The coefficient of correlation calculated from 28 generations was 0.531. This value exceeds the normal coeffi- cient of correlation value for two independent variables when n (number of generations) = 28 at the 1% level of signifi- cance (p.16h, Dixon and Massey, 1951). From this it can be 27 said that the frequency decline of yellow phenotypes is NOT independent of the density of the population. Therefore, population density is a factor in the elimination of the yellow gene. Thus the higher densities of the "Scarcity" generations correspond with the sharper decreases of the yellow frequency. The less dense populations of "Abundance" produced generations with relatively smaller decreases in the frequency of yellow. The mean population per generation for "Scarcity" was 1,438 flies, compared to 1,036 flies for "Abundance". The rate of increase of the total population falls off as the density increases (Pearl, 1927). Kemp (1929) suggested that the decrease in offspring in dense pOpulations is due to a smaller proportion of the population engaging in reproduc- tion. The diminished breeding unit could have contained rel- atively more wild flies in the "Scarcity" generations of our experiment. However, Pearl held that decreased egg production caused this decrease in the number of progeny. L'Heritier (1937) explained the phenomenon by larval mortality. If Pearl's explanation is the answer, then egg production could occur differentially. If L'Heritier's is the answer, then larval differential mortality could take place. This latter mode was not detected in the tests in our experiment, although these tests involved less than eleven pairs of parents in the largest mating. Density-produced phenotypic alterations in size, however, may be indicative of 28 corresponding physiological changes. The gamete production of smaller yellow flies may have dropped more than that of smaller wild flies. Mating discrimination against yellow males may have become more pronounced with the smaller yellow males. Rate of Development Although yellow larvae were noticed to be more sluggish than wild larvae, no observable differences in the rates of deve10pment occurred. 0n the other hand, Neel (19A1) found that in starved larvae, eclosion took place at about the same time as in well-fed larvae and even slightly earlier at times. For our experiment, developmental time in the wild stock was compared with that of the mutant stock. The times for devel- Opment for wild and yellow flies in competition with each other were also noted. Significant differences were not observed. TABLE 2 EXPECTED PHENOTYPIC FREQUENCIES IN "ABUNDANCE", NO SELECTION MALE % FEMALE_% 0 9.6 90.4 - . 100.0 1 - 100.0 9.6 ‘90.4 2 4.8 95.2 4.8 95.2 3 2.4 97.6 7.1 92.9 h 3.6 96.4 5.9 94.1 5 3.0 97.0 6.5 93.5 6 3.3 96.7 6.2 93.8 7 3.1 96.9 6.3 93.7 8 3.2 96.8 6.2 93.8 9 3.2 96.8 6.3 93.7 n 3.2 96.8 6.3 ‘93.? TABLE 3 EXPECTED PHENOTYPIC FREQUENCIES IN "SCARCITY", NO SELECTION #» MALE % FEMALE % 933. @113 Yellow flild Yellow 0 3.1 96.9 1.1 98.9 1 1.1 98.9 h.2 95.8 2 2.1 97.9 3.2 96.8 - 3 1.6 98.h 3.8 96.2 h 1.8 98.2 3.5 96.5 5 1.7‘ 98.3 3.6 96.h 6 1.8 98.2 3.5 96.5 7 1.8 98.2 3.6 96.4 n 1.8 98.2 3.6 96.4 Gen. \oooxnoxmrwmp P‘ +4 F‘ .4 b‘ 14 F4 ,. h‘ mxloxmk-wmpo 96'.4 00 NUMBER AND PROPORTION OF COLOR AND SEX PHENOTYPES AMONG TOTAL FLIES IN "ABUNDANCE" TABLE 4 Ma 9 Wild Yellow ._fl__ __%__ __E__ _.%_. 5 .050 47 .465 0 .000 .132 .426 262 .228 290 .252 160 .307 134 .257 144 .113 491 .384 103 .129 294 .367 30 .022 650 .476 88 .089 384 .387 167 .186 309 .344 277 .220 349 .277 228 .256 210 .236 368 .340 173 .160- 756 .422 211 .118 596 .454 73 .056 715 .435 107 .065 269 .440 8 .013 706 .461 22 .014 526 .468 3 .003 734 .498 0 .000 694 .503 0 .000 502 .505 #0 .000 101 310 1149 522 1279 801 1366 991 897 1261 891 1981 1793_ 1314 1643 611 1533 .1124 1475 1381 994 Fema e Wild Yellow ' N _Z__ .__E. ___2 12221 O .000 49 .485 39 .126 139 .448 336 .292 261 .227 116 .222 112 .215 460 .360 184 .144 170 .212 234 .292 118 .086 568 .416 146 .147 373 .376 227 .253 194 .216 543 .435 87 .069 420 .471 33 .037 510 .472 30 .028 822 .458 4 .002 640 .487 5 .004 821 .500 0 .000 334 .547 O .000 805 .525 0 .000 595 .529 0 .000 741 .502 0 .000 687 .497 0 .000 492 .495 0 .000 TABLE 5 NUMBER AND PROPORTION OF COLOR AND SEX PHENOTYPES AMONG TOTAL FLIES IN "SCARCITY" 32 Male Wild Yellow __N_ Prop. __N_ Prop. 4 .018 ‘125 .568 31 .022 641 .464 149 .259 .139 .234 279 .267 278 .266 235 .138 639 .374 198 .242 202 .247 351 .311 186 .165 245 .228 271 .253 406 .244 374 .224 1154 .347 466 .140 645 .488 0 .000 921 .510 0 .000 1167 .531 0 .000 1364 .489 0 .000 Female Wild Yellowv _§: Prop. N Prop. 1 .005 90 .409 32 .023 676 .490 277 .481 15 .026 305 .292 184 .176 724 .424 109 .064 304 .372 113 .138 534 .473 59 .052 496 .462 61 .057 726 .436 160 .096 1626 .489 79 .024 677 .512 0 .000 885 .490 0 .000 1031 .469 O .000 1425 .511 0 .000 220 1380 576 1046 1707 817 1130 1073 1666 Total 3325_ 1322 1806 2198 2789 FLIES: TABLE 6 {RQPORTION OF YELLON FLIES ANONG TOTAL "ABUNDANCE" AND "SCARCITY" C) \ooe-QOUIJ-‘wmr-Joig 0 A) +4 Fl #4 94 #4 rd P‘ r» I4 h‘ OxomxloxmthI-do "Abundance" "Scarcity" .950 .874 .480 .471 .528 .659 .892 .764 .561 .346 .273 .188 .120 .059 .065 .013 .014 .003 .000 .000 .000 .977 .954 .260 .442 .438 .386 .217 .309 .321 .164 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 t Value 1.1 4.0 9.8 1.1 5.0 11.4 45.0 23.3 12.0 13.3 18.2 15.7 15.8 9.1 10.8 2.6 4.7 1.88 .000 .000 .000 33 PROPORTION OF YELLOW FLIES IN EACH SEX: TABLE 7 "ABUNDANCE" O Id #0 FJ +4 Ia F‘ *4 rd 14 (n «q Ox v1 ¢~ \» A) +4 <3 19 20 Male Proportion Wild .096 .000 .475 .544 .227 .259 .044 .186 .351 .442 .521 .680 .782 ..891 .870 .971 .970 .994 1.000 1.000 1.000 Yellow .904 1.000 .525 .456 .773 .741 .956 .814 .649 .558 .479 .320 .218 .109 .130 .029 .030 .006 .000 .000 .000 Female Proportion Wi1d .000 .219 .563 .509 .714 .421 .172 .281 .539 .863 .927 .944 .995 .992 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Yellow 1.000 .781 .437 .491 .286 .579 .828 .719 .461 .137 .073 .056 .005 .008 .000 .000 .000 .000 .000 .000 .000 34 TABLE 8 PROPORTION OF YELLOW FLIES IN EACH SEX: "SCARCITY" C) O 5 j: l ‘ W L 1 Male Proportion Female PrOportion \omxzoxmrwmon HHH NHO H C» 1.000 1.000 1213 Yellow Wild Yellow .031 .969 .011 .989 ' .046 .954 .045 .955 .525 .475 .949 .051 .501 .499 .624 .376 .269 .731 .869 .131 .495 .505 .729 .271 .654 .346 .901 .099 .475 .525 .890 .110 .521 .479 .819 .181 .712 .288 .954 .046 1.000 .000 1.000 .000 1.000 .000 1.000 .000 1.000 .000 1.000 .000 .000 .000 \h) C\ TABLE 9 ASSORTATIVE RATING: FEMALE PREFERENCE Successful Males Female' Total Genotype Females Yellow 2 Wild Fertilized N 1% N % 2 8. 21 l. yy 23 7 9 3 Yy ‘24 0 0.0 24 100.0 TABLE 10 ASSORTATIVE NATING: NALE PREFERENCE Male Total Fertilized Females Genotype Males . Successful Yellow Wlld ‘ N % N 4 y o 5 55.6 4 44.4 Y 12 6 50.0 6 50.0 37 TABLE 11 EXPECTED FREQUENCIES OF GENOTYPES: "ABUNDANCE", FROM MERRELL'S COEFFICIENTS Male Female. wild Yellow Wild Heterozygous Yellow Total Gen. (p) (q) (r) (23) (t) Yellow 6_ .044 .956 .002 .170 .828 .892 7 .065 .935 .020 .344 .636 .778 8 .153- .847 .063 .441 .496 .682 9 .236 .764 .151 .576 .273 .516 10 .396 .604 .298 .555 .147 .373 11 .543 .457 .471 .470 .059 .258 12 .689 .311 .635 .341 .024 .179 13 .797‘ .203 .763 .229 .008 .107 14 .874 .126 .853 .144 .003 .065 15 .924 .076 .911 .088 .001 .034 16 .955 .045 .940 .060 .000 .022 17 .970 .030 ‘.965 .035 .000 .014 18 .018 .979 .021 .000 .009 .982 38 TABLE 12 VIABILITY RESULTS Number of Seventh Day Percentage Imagoes Female Yellow of Yellow Female Parents Ofggpging Offspring X2 Day 1 144 46.2 0.289 20.6 1 160 53.5 0.245 22.9 2 185 49.0 0.200 13.2 3 178 54.4 0.387 8.5 4 224 50.4 0.003 8.0 5 268 43.8 0.769 7.7 6 227 42.7 1.066 5.4 7 169 42.4 1.155 3.4 ‘8 309 49.4 0.007 5.5 9 278 47.3 0.146 4.4 10 304 50.0 0.000 4.3 :_*a 9 3 unoHpmpecoc Tma mfiwflmm am am HHO il .144 1:14 . coo. (DH 0 .. OOH. t CON. T 00m. . .0090... aepfionmom= ..._.. 004. O 00%.. . 00m. zoocmccsnf I :owuuomoum 9.3.: 4.290% 462024. mm «Emu 21.....qumw mo Mozamcmmm H $5on 40 unowumhunvo - 8886826488116mammwo . . . 11---.---.:.-a . . . . .. . n . . 1% 000. m :.ooa. ... .00. 9’. on o. 00. . nIOONo '0 .0 O ..o. o co. ... ... , .. . .. .n - ..OOM. v . .looe. =hpwuhmum=233§ ... . a ...vofiwvcfifldzolle . . ... 00m. 000. soapy..opd maHaa 42909 82028 mega: 2044a» mo ZOHBmomoma N mapmHe nGOHpmuoneo om 8H ma ea 6H ma 8H ma 6H ma 9“ 6 m a w o . - ooo. apm «pd, arm N H l .....h...‘.'.. L. 0...... 0'00 j . .ooH. ..oom. O A u . .00m. 0 L . .oo:. .OOm. A .005. L ... e- 000. At 000 e uonEom 30:3. 0...... .e.. . r00m. munE 30:0»! .. 8.000. .oo.H nofipuoaopm :m02¢023m<: .xmm mo¢m 2H mmHAh BOAQWM mo onPmomomm m mmDUHm 42 unawpmnecmm m m o 6.... . 1 1+ 4 w m m w m 000. ... . ... 00 O e 00 e 0 e... .... OOH. O .0 O . O O O. .0 00 O. O. O. u .. .. . . . ..oom. O I O O I . . . .. 0. 0 e 1' 00m. to 0 . C . ..OOE. .. 0 1.00m. 0 O O Iv com. o o o .7005. o o . 1.000. noamfivm BOHHGM eeeeee . monE zeaaew . 1.000. ee 0 coauuwmmum zweHomqom: .xmm mo¢m 2H mqum Hoqqmw m0 onemomomm s mmaoHe 3 cm .H 3 fl ma mm 4 mm- mm uncapmuocow 22.4211481_ 7:». - p111. . . «111-44-11.41. 4 d 1 q . 2 u 4 . OOO , .53. f. 0.0 0. Become ... CO. 000 00.0. . 00.0000. .00 co. co. 1' COM. 0 e o ..:. .. . .403. 0000‘. 0 ... .. ooh. .. 1.86. ..005. a. . 136.289. 2:... oow zoocwccanf IIIII ... fooo. ..:: 000.H coauuoaoum mmHAm 44909 0203< mmHAm hoqqmw m0 onemomomm m mmDOHm 44 ncowpmuvccu NH Ha 0H 0 m u m 4 m N I I. . 4 u- .7 4 61 1 4 1 1 w Amoco. ll... 0 , o. III. 1100‘“. e0 .11.! I so. fl 0 1. 00m. . /.h . . 0 .0 ' 0 O o 00 . [“0 0 0 Lv COM 3.0... 5.. . u 0 0 o ,fl0 .. “ 1l8+~e 00 . IIeOee .00 w I . 1.00m. . O I o I Do , o I u 2.08 0 /I. « ooa. . . nucofioflmeoo II o. 1 . 6.2.44.2..IIII I... .. I o 1-000. zuhflfiohflcmz eeeeeeee I 0. zoocmvczp<=v _ =eonmncsn<= mo 0 cowpmuenvwllbu. .1000. zhpfioneomz no 0 coapmhecoc . mqum Adeoe 0202a Boqqmw A<909 o mmDOHm md>m00 Wozm00mmm OMBOMAmm 000.H GOHpLOQOhm 45 DISCUSSION In an article defending his theory of equilibrium, Hardy (1908) admitted that selective influences caused the entire situation to become "...greatly complicated...". Simpson (1951) warns that selection "...involves complex and delicate interplay with those genetic factors in populations...". Ad- vice more pertinent to this problem was offered by Wright (1939) when he cautioned that with sex-linkage selection ef- fects "...may be rather complicated because of the necessity of considering the sexes separately". Since no equilibrium was reached in either "Abundance" or "Scarcity", the manifold effects of selection had to be analyzed. ~¢ Types of Selection and Developmental Stages Basically, selection is differential reproduction (Simp- son, 1951). As such, selection mainly proceeds through a continuum of relative effectivity; from maximum potency to complete sterility, from highest viability to complete lethal- ity, or from random copulation to absolute assortative mating. These differentials exist in three general forms; mortality, gamete contribution,and fertility. They may occur at differ- ent stages of deVelopment in every organism. In Drosophila, five main developmental stages can be considered with respect to selection: the adult, the gamete, the egg, the larval, and the pupal stages. 46 Mortality The mortality of the adult fruit fly varies considerably. Spencer (1932) found that vestigial-winged flies lived longer than normal-winged flies when subjected to complete lack of food and water. Under optimum conditions the normal-winged flies lived longer. Differential adult mortality could not have been a factor in the elimination of yellow flies in this experiment. Less than 0.5 percent were found dead in the medium. In assortative mating tests, however, yellow males died significantly more often than any other type. This may have been caused by a differential viability with reapect to ether. V There is no available evidence concerning differential mortality in gametes. A differential mortality somewhere between the egg and larval stage was studied in Q. polymorpha (da Cunha, 1949). This differential favored the heterozygote and resulted in equilibrium. There is no evidence of such a mortality differ- ential with respect to the gene for yellow body color in Q. melanogaster. Crosses in Q. virilis between yellow males and heterozygous females yielded four classes in approximately equal proportions (Metz, 1916). The expected offspring of such a crossare males and females which are wild and yellow in a 1:1:1:1 ratio. The same results were obtained in this experiment (See Table 12) with Q. melanogaster. These results suggest that no differential mortality exists in the gamete, 47 egg, larval, 0r pupal stage. This is a startling piece of evidence, considering that Q. melanogaster lays ten times as many eggs as will develop (L'Heritier, 1937). Even in the absence of competition in optimum food conditions, only 43 percent of the eggs in one of L'Heritier's strains reached the adult stage. Apparently, such heavy selection pressure is random, in the light of the viability tests shown in Table 12. Perhaps the tests were not severe enough. A study should be done using the technique of J. V. Neel (1941). Neel studied crossing-over in starved larvae which had been fed only water- moistened Kleenex. Such severe starvation might introduce- differential selection pressures which would be absent in the milder tests such as were carried out in this experiment. Sturtevant (1929) mentioned that the viability of yellow phenotypes in D. simulans was "Usually good", and the gene for yellow in that species is allelomorphic with the same gene in D. melanogaster. Viability, as with other selective dif- ferentials, is remarkably.sensitive to environmental change. For example, Timofeeff-Ressovsky (1933), found that the mutant, eversae, was superior to the ‘wild fly at 24° - 25°C., but in- ferior at 15° - 16°C. and 280- 300 C. No evidence of reduced viability was observed in our experiments, however. Differential Gamete Contribution A numerical disparity in gamete contribution can be caused by various factors. Assortative mating is often the primary 48 causative agent. Wright (1921) described two types of assort- ative mating based on somatic resemblance: perfect and imper- fect. Had perfect assortative mating been the case here, then heterozygosity in the females would have rapidly diminished until only flies homozygous for yellow and for wild remained. The gene frequencies would have been unchanged, however. Imperfect assortative mating, however, may mean partial failure of gene transmission on the part of a discriminated class. Such partial discrimination has been demonstrated by mating tests with SGXrlinked genes in Q, pgeudoobsgura (Tan, 1946), Q. subobscura (Rendel, 1945), and Q. melanogastgr (Rendel, 1945; Merrell, 1949; Reed and Reed, 1950). Tan showed that yellow females'were preferred over wild females by both types of males. Merrell demonstrated the same phe- nomenon, but Rendel stated that yellow females mate normally. Tests in our experiment did not contradict the hypothesis that' yellow females are preferred. Rendel (1951) wrote that it is not always possible to distinguish between a specific mating preference and an indirect effect of general vigor. In fact, Sturtevant (1929) concluded that the weaker and more inactive female is more likely to be mated with. However, weakening experiments did not increase the frequency of "cross-mating". This condition results in a biological paradox in that less- vigorous flies enjoy a reproductive advantage due to lack Of vigor. But in our experiment any mating superiority enjoyed by yellow females is more than cancelled by the strong discrimi- 49 nation against yellow males by females of both phenotypes. All mating tests with yellow have shown this fact. Terrell (1949), for instance, found that yellow males of 2. melano- gaster in competition with wild males mated successfully with yellow females only 11 percent of the time and only 6 percent with wild females. Such breeding discrimination against ye1-' low males along with a preference for yellow females results in a relatively higher yellow male frequency in the offspring than expected. 0n the other hand, the class of yellow females is smaller than expected (See Tables 9 and 10). Fertility Hogben (1946) lists numerous criteria used to define fertility. The egg-laying capacity of the female, the viabil- ity of spermatozoa, the length of reproductive cycle, and the interval which elapses between two fertile periods are a few. The frequency of mating attempts would influence assortative mating tests. Our tests, however, reflect actual discrimina- tions and not merely a lessened frequency of mating attempts. This conclusion is based on actual observations of rejections of yellow males by females. As for the above criteria of fertility, they were not tested in our experiment. Rate of Deve10pment Although no mutant gene is known that affects only the larval or pupal stage of D. melanogaster (Morgan, Bridges, 50 and Sturtevant, 1925), the yellow larvae are certainly physi- ologically different from the wild larvae. Vhether their slower movements and limited ran:e reflect lessened viability or slower growth rates is not known. Many mutants have exces- sively slow rates of development (Bridges and Brehme, 1944). Curly-wing in Q. pseudoobscura, for instance, develops slower than the normal-wing (Dobzhansky and Spassky, 1944). No evi- dence for differential development rates was observed in our experiment, however. Conclusions As Sewall Wright (1921a) predicted, a differential re- productive rate among classes prevented equilibrium. Inves- tigation of the selection modes causing this differential re- production suggests that the main factor involved in the elimination of yellow was assortative mating. A search for the cause of the differential effect of density in "Scarcity" and "Abundance" was attempted. Adult mortality was ruled out. A survey of selection literature brought out several possible effects of population density: a decreased breeding unit, a decreased egg production, or an increased larval mortality. Several other possibilities can logically be considered: in- creased discrimination against yellow males or decreased de- velopmental rates. The fact that one successful courtship Stimulates other males to redoubled activity (Handel, 1951) may be a factor. Any of these variables may have operated 51 in a differential manner, thus causing a differential elimina- tion of yellow. These same factors may have varied with the frequency levels, with the humidity, or with any of numerous envoronmental effects. 52 SUMMARY Selection prevented equilibrium between the sex-linked gene for yellow body color and its wild allele. The main selective pressure was possibly due to assorta- tive mating. Specifically, the major factor was probably severe discrimination against yellow males during court- ship. Frequency changes in yellow flies were also significantly correlated with population density. The greater popula- tion per generation in "Scarcity" indicated less food per fly than in "Abundance". Evidence was presented to show that differential viability in the egg-to-imago stages was probably not a factor.- Differential mortality in the adults was probably not a factor. Observations indicated that a differential rate of devel- opment was probably not a factor. 53 BIBLIOdRAPHY Books Boyd, W.C., Genetics and the Races of Man, Scranton, Pa.: Haddon Craftsmen Inc., 1953, #53 pp. Dixon, W.J. and bassey, F.J., Introduction to Statistical Anal sis, New York: McCraw-Hill Book Co., Inc., 1951, 375 pp- Dobzhansky, Th., "Mendelian Populations and Their Evolution", Chap. 25,Dunn, L.C., Editor, Genetics in the Twentieth Centur , New York: The NacMillan Co., 1951,—634 pp. Hogben, L., An Introduction to Nathematical Genetics, New York: W.W. Norton afid‘Co., Inc., 1956, 260 pp.—w Huxley, J., Evolution, the Modern Synthesis, New York and London: Harper and Fros., 19h3. 555 PP. Li, C.C., 5n Introduction to Population Genetics, Peiping, China: National Peking UniVersity Press, 1958, 321 pp. Simpson, G.G., The Meaning of Evolution, New York: The New American Library, 1951, 192 pp. Wright, Sewall, Exposes de Biometrie et de Statistioue Biolggiquc. XIII. StatiSticaI Genetics in Relation to Evolution, Paris:*Hermann and Co., 1939, 63 pp. Periodicals Bridges, C.B. and Brehme, K.S., "The Butants of Drosophila Melanogaster", Carnegie Institution of Washington, Pub. 552’ pp'zzq‘232g Ighhe da Cunha, A.B., "Genetic Analysis of the Polymorphism of Color Pattern in Drosophila polymorpha", Evolution, Vol. 3, pp.239-251, 1959. da Cunha, A.B. Burla, H., and Dobzhansky, Th., "Adaptive Chromosome Polymorphism in Drosophila willistoni", Evolution, Vol.h, pp.212~235, 1950. 5A Dobzhansky, Th., "Adaptive Changes Induced by Natural Selection in Wild Fonulations of Urosonhila", Evolution, Vol.1, pp. 1-16, 19a7. Gobzhansky, Th., and Spassky, B., "Genetics of Natural Popula- tions. XI. Manifestation of Genetic Variants in Drosophila pseudoobscura in Different Environments", Genetics, Vol. 29, pp. 270-290, lQAh. "'“"‘ Dubinin, N.P. and Tiniakov, G.G., "Inversion Gradients and Natural Selection in Ecological Races of Drosophila funebris", Genetics, Vol.31, pp.537-545, 19A6. Freire-Maia, N., "Balanced Polymorphism in Drosophila montium", Evolution, Vol.3, p.98, 1949. Hardy, G.H., "Mendelian Proportions in a Mixed Population", Science N.S., Vol.28, pp.A9-SO, 1908. Kemp, W.B., "Genetic Eouilibrium and Selection", Genetics, Vol.1h, p.85, 1929. L'Heritier, Ph., "Etude de Variations Ouantatative Au Sein d'une Espece Drosophila melanogaster”, Archives de Zoolggie Experimental et Generale, Vol.78, pp.255-356, 1937. L'Heritier, Ph. and Teissier, G., "Etude d'une Population de DrOSOphiles en Equilibre", C.B. Acad. Sci. Paris, Vol.197, pp.l755-1767, 1933. Ludwin, 1., "Natural Selection in Drosophila melanogaster under Laboratory Conditions", Evolution, Vol.5, pp.231- 2A2, 1951. Merrell, D.J., "Selective Mating in Drosgphila melanogaster”, Genetics, Vol.3h, pp.370-389, l9h9. ‘ Merrell, D.J., "Measurement of Sexual Isolation and Selective mating", Evolution, Vol.h, pp.326-331, 1950. Metz, C.W., "lMtations in Three Species of Drosophila", Genetics, Vol.1. pp.591-6O8, 1916. Morgan, T.H., Bridges, C.B. and Sturtevant, A.H., "The Genetics of Drosophila", Bibliographia Genetica, Vol.2, pp. 3-267, 192;. Neel, J.V., "A Relation Between Larval Nutrition and the Frequency of Crossing Over in the Third Chromosome of Drosophila melanogaster",Genetics, Vol.26, pp.h69-5A2, 1951. 55 Pearl, R., "The Growth of Populations", Quarterly Review of Biology, Vol. 2, pp. 532-548, 1927. Pearl, R. and Parker, 3. L., "On the Influence of Density of POpulation on the Rate of Reproduction in DrOSOphila", Proc. Nat. Acad. Sci., Vol. 8, pp. 212-218, 1922. Pearson, K., "On the Generalized Theory of Alternative Inheri- tance, with Special Reference to Mendel's Laws", Phil. Trans. Roy. Soc. (A), Vol. 203, pp. 53-86, 190A. Plunkett, C.C., "Temperature as a Tool of Research in Pheno- genetics: Methods and Results", Proc. VI Int. Congr. Genet. (Ithaca), Vol. 2, pp. 1583160, 1932. ' Reed, S. C. and Reed, E. W., "Natural Selection in Laboratory POpulations of Drosgphila", Evolutigg. Vol, 2, pp. 176- 186, 19h8. Reed, S. C. and Reed, E. W. "Natural Selection in Laboratory Populations of DrOSOphila" II. Competition Between a White-eyed gene and its Wild-type Allele"., Evolution , Vol. A, pp. 3h-h2, 1950. Rendel, J. M., "Failure of Mating Between Two Mutants of Drosthila subobscura", g. Genetics, Vol. A6, pp. 287- 3 g 1+ O ‘ Rendel, J. M., "Mating of Ebony Vestigial and Wild Type Drosophila melanogaster in Light and Dark", Evol., Vol. 5, pp. 226-230, 1951. Robbins, R. B., "Application of Mathematics to Breeding Prob- 1ems II." Genetics, Vol. 3, pp. 73-92, 1918. Shell, S. E., "Phenotypic Frequencies in Large, Random-breed- ing Ponulations of Drosophila Melanqgaster which Carry the Genes for Wild and Sepia Eye CdIor", Unpublished Michigan State College M.S. thesis., 1952. Spencer. W. P., "The Vermilion Mutant of Drosophila h dei Breeding in Nature", Amer. Nat., Vol._66, pp. A7 - ., 1932. . Sturtevant, A. B., "The Genetics of Drosophila Simulans", Carnegie Institution of Washington, Pub. 399, pp. 5-55, 1929. Tan, C. C., "Genetics of Sexual Isolation Between Drosophila pgeudoobscura and DrOSOphila pgrsimilis", Genetics, V01. 31’ Pp. Egg-573, 1614'60 56 zen- Timofeeff- Ressovsky, N. W., "Uber die Vitalitat einiger mutationen und ihrer Kombinationen in Drosophila funebris und ihre Abhanzigkeit vom "genotypischen‘i und vom auss~ren Zeitschrift fur Inductive Abstammungs-und Vererb- Milieu", nvslehre, V01. 66, pp. 319-3hh 1931 ystems of Mating.III. Assortative Matlng ' Vol. 6, pp. 1AA- Wright, Sewall, "S Ba5ed on Somatic Resemblance", Genet1cs "Systems of Mating. V. General Considerations", 161, 1021. 6, pp. 167-178,1921a. Bull. Wright, Wright, Sewall Genetics. Vol. Wright, Sewall, "Statistical Genetics and Evolution" Amer. Math. Soc., Vol. A8, pp. 223- 2A6, l9h2. S. and Dobzhansky, Th., "Genetics of Natural Popula- tions. XII. Experimental Reproduction of Some of the Chan- ces Caused by Natural Selection in Certain Populations of ’ O O Broso hila pseudoobscura", Genetics Vol 31 pp. 125- 157 19A5. ’ \x K K \ ‘\ Sex» . i . USE ONLY ---7J“§ 36‘“ ‘5‘/ 3° 30 '54 W2 5“ HICHIGQN STRTE UNIV. LIBRQRIES 31293008692737