H,‘ :mmawm ammmwczas m LARGE, ammmmemm mammmm as: $223. gmmwggag, wags $358K}! 6M3?” 3349;; 535N253 @033 mm AWE? ' 35m 26% mam ‘E’hasfzs far v‘rim $392333 :33 33% S. Mfiififififi $13}! mum: 39am Eiéxafiafirh 33625; E“? "” ‘3; WW' '7'..- lo- 'K IllllllmlflllllMlllllllllllllllllllfllllllllllllllfllllll j 1293 10375 6981 I This is to certify that the thesis entitled Mosaic frequencies in large, random- breeung populations of w glances»: mob carry the genes for wild and sepia. eye color. presented by Summon has been accepted towards fulfillment of the requirements for AL. degree mm I Major professor Date ”t 261 152. 0-169 _..7.'. C s‘ MSU LIBRARIES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES w111 be charged if book is ‘ returned after the date stamped be1ow. “ .. PHENOTYPIC FREQUENCIES IN LARGE, RANDOM—JREEDING POPULATIONS 0F DROSOPHILA MELANOGASTQEEKHICH CARRY TBA GENES FOH.NILD RED SLEIA EYE COLOR Sara Elizabeth hell 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 IENCE Department of Z 0010 gy < \A‘ ll\.K theoretically expected from Hardy's Law. APPARATUS AND METHODS APPARATUS AND METHODS The population cages used in this experiment were of very simple construction. They consisted of a wooden platform upon which was built a smaller, box-like cage. The extension of the platform beyond the dimensions of the cage made the cage easier to handle. This box-like cage was completely of glass except for the wooden floor and the wooden frame which supported the glass sides and top. The cage was Opened by sliding upward the sheet of glass which formed one end of the cage. Four such cages were used in the entire experiment. The first two, designated A and B, were 9.5 inches wide, 15.5 inches long, and 10.0 inches in height. The other two cages, C and D, were 16.0 inches wide, 2h.0 inches long and 12.0 inches high. These cages were kept at all times in a constant temperature room at 26.!“0c. (i 1.00 c). Tie shades i-.- the room were kept drtwn so that the room was in a semi-dark condition during the day and completely dark at. night. Throughout the experiment, the medium provided fer the flies was of the cornmeal-molasses-agar type. The formula for the meditm was as follows: 10 liters of water 110 gr ms of agar 8 grams Moldex 1/2 pound baker's yeast (dissolved in water) 350 cc. unsulphurcd molasses 350 cc. Karo 1000 grams corn meal The stocks of Drosophila melanogaster used here were obtained from the Biological Supply House in Chicago. The genes selected for study were those for sepia eye color and its wild type allele. Sepia is com- pletely recessive and may be described as a deep, translucent pink eye color in freshly hatched flies which darkens as the fly ages to a very dark purplish black. It is a highly useful gene for experimentation because there is little fluctuation in the character and it is quite easily distinguished from the wild type eye color oven in newly hatched flies. It is located on the third chromosome at a locus of 26.0.6 The mutation of the wild type of sepia is quite rare although it has been reported a few times since its original discovery. In the first series of experiments with cages A and B, half pint milk bottles filled to a depth of 30mm with the medium were used in the cages. A strip of paper toweling was placed in the medium to provide a place for the larva to crawl when they were ready to pupate. The initial crosses between wild and sepia-eyed flies were made in the half-pint bottles outside the cages. Then the F1 pupae appeared, the parents were removed and five bottles of this type were placed in each cage together with ten bottles containing fresh medium. All the bottles in the cages were without steppers. The F15 were allowed to hatch out in the cage and to lay their eggs. Eben the F2 pupae appeared in the newer medium bottles, the adult insects and the five original bottles were removed from the cage. The F1 flies were then discarded. This removal was accomplished in the following manner. The cage{ was placed in a completely darkened room and a goose-neck lamp, with its light directed over the top of the cage at the end farthest from the opening, was placed near the cage. In about ten minutes time, most of the flies had gathered at the top of the cage in the area where the light was the strongest, since these insects are positively phototropic m4 and negatively geotropic. The cage was then Opened by raising the end glass piece just enough to reach in and remove the bottles, each of which was plugged with a sterile cotton plug as it was lifted from the cage. This method of removal proved fairly efficient. Some flies were lost and although the exact number is unknown, I would estimate that 5% of the flies in the cage was the maximum number lost at any time. The average loss would be much less than that. After all the bottles had been removed, cotton wadding, soaked in ether, was placed in the cage which was then tightly closed again. Uhen the insects had succumbed to the ether, they were gently swept out with a soft brush. The cage was then carefully cleaned with a detergent and dried. / The bottles from which the Fl's hatched were discarded. Half of the bottles (5) containing the F2 pupae were replaced in the cage along with ten bottles of fresh medig‘and the plugs were withdrawn. The other five F2 bottles were kept for counting purposes. After two days the F2 would begin to hatch; those in the cage were allowed to deposit their eggs and those in the remaining bottles were counted and the number of sepia individuals carefully noted. The two groups were assumed to have the same gene freguencies. This cycle was repeated through the F7 generation for the two separate cages A and B. The resultant data for experiments I and II are recorded in Tables I and II. The results obtained by the above method were too erratic to analyze. Since the percentages of sepia varied greatly from bottle to bottle in the same generation and in the same cage, it was assumed that the sampling technique was at fault. This variation was greater by far than could be accounted for by the law of change alone. To test this hypothesis, the adults in the cages A and B of the F7 generation, when they had laid their eggs were removed as usual but this time they were counted. This cage count was compared with the bottle count for the same cage in the same generation and a discrepancy of h.75% and of 12.27% in the two counts was found for cages A and B respectively. As a result, the method was modified so that no bottles were removed for counting purposes. Rather, the adults were all allowed to hatch in the cage, allowed to lay their eggs and were then removed by the method described above and all of them were counted. At the same time, the bottles (10) from which the adults had hatched were discarded, the bottles (10) in which they had laid their eggs for the next generation were left in the cage and were arranged alternately with the ten new bottles of medium added at this time. This method of separating the generations, removing and counting the flies, and of supplying new medialwas used in all the following phases of this experiment, whether the media’was supplied in bottles or in petri dishes. The results of generations seven through thirteen for experiments III and IV are summed up in Table III. At this time, the two cages A and B became contaminated with white-eyed flies and the experiment was discontinued. There seemed to be a congregation of flies of one of the two types around the mouths of some of the bottles. In an attempt to remedy this, petri dishes were substituted for the bottles and used in all subseQuent phases of the experiment with the cages. Eight petri dishes, 100mm in diameter, were used in each of the two cages A and B in experiments we shall term V and VI. The petri dishes were filled to a depth of about 10mm with the medium and a strip of the paper toweling was added to each dish. The four dishes which had already produced flies were removed in each generation in the same manner as the bottles. Experiments V and VI did not begin as a straight F2 cross, but rather, different percents of sepia and wild type flies were introduced into the cages to start the populations. The data obtained from these crosses, generation one through fourteen in Cage A (experiment f&) and one through eight in cage B (experiment Tgl) are recorded in Tables IV and V. The larger cages C and D were begun with pOpulations of true Fls. These cages each held ten petri dishes 150mm in diameter, five of which were removed in each generation in the manner described previously. These dishes were filled with medium to a depth of about 15mm and a paper towel strip added to the top. The five plates containing the larva and pupa of the next generation were arranged between the five dishes con- taining the fresh medium. The data for the fourteen generations raised in each of these cages (experiments VII and VIII) will be found in Tables VI and VII. Throughout this series of experiments, the cages were carefully cleaned between each generation in order to prevent the growth of molds. The bottles were plugged and the petri dishes covered during these ex- changes to avoid contamination by other flies or by molds. As the experiment progressed, it appeared that selection was modifying the expected ratios. It occurred to me that this might be due to the selection of mates. Three tests were planned to determine this. The virgin females and the males used throughout this experiment were all to hours (I 2 hours) of age. In the first case, five virgin sepia females were placed in a bottle with five sepia males and five wild males for a period of four hours. The females were then removed and put in separate bottles so that the type of offspring she produced would show whether she had been fertilized by a sepia male, a wild male, both, or in the case of no offspring, neither. This was done with a total of seventy-five sepia females and an egual number each of the two types of males. Possible mate selection for the opposite sex was also studied. In the second case, five virgin sepia females and five virgin wild females were placed in a bottle with five wild males for a period of four hours. The females were then separated and records kept of which type of female was most often fertilized. A total of eighty females was used here, forty of each type. Lastly, the experiment was repeated with the two types of females and five sepia males, and again involved the use of eighty females. The results of these crosses appear in Table VIII. The relative reproductive ability of each of the stocks and various crosses between them when not in.competition with other kinds was also studied in relation to this problem. The males and virgin females used 10 here were also hé hours (I 2 hours) of ag , In each case, five males and five females were placed together in a half-pint bottle containing medium. Each type of cross involved the use of 100 males and 100 females or twenty bottles. A total of five types of crosses were made, wild X‘Wild, sepia x sepia, wild male x.sepia female, sepia male x wild female, and heterozygote x heterozygote. The parents were removed from the bottle when the pupa of the next generation began to appear. The average total number of offspring produced in the five days after the appearance of the first adults for each of the types of crosses is tabulated in Table VIII. DATA AND OBSERVATIONS ll DATA AND OBSERVATIONS The percentages of sepia in each generation in the eight populations studied is the basic material in this research. The population number is the actual number of flies counted and consists probably of a minimum of 95% of the actual pOpulation and in most cases a greater percentage than that. Male and female differences in reSpect to the freQuency of the sepia phenotype have been recorded in all except the first two experi- ments. Tables I and II give the percentages of sepia flies in the first seven generations in experiments I and II in which half the bottles con- taining the next generation of insects were removed from the cages and the hatching offspring counted. This sample was assumed to be like that of the bottles left in the cage to produce the succeeding generation. The number of flies and the percentages of sepia are given for each of the five bottles in each generation, and, in addition, the total per- centage for that generation. The numbers of flies in these bottles averaged over 200, so that the populations in each generation usually numbered over a thousand. The percentages varied so much from bottle to bottle that the erratic results obtained over the seven generations are probably due to the random selection of the bottles which were re- moved from the cage for purposes of counting. The over-all data ob- tained here were useless in the application of Hardy's Law, but the widely varying bottle counts were of value later in the explanation of 12 the phenomena occurring in later populations. As has been explained, these results were responsible for the discovery of the faulty sampling technique, which was then discarded. Table III provic es the percentages of sepia obtained in experiments III and IV, which were continuations of the two preceding experiments after the flies were counted upon their removal from the cage as was 3 explained in the foregoing section. In cage A there was a decline from about lh.§% in the seventh and eigllth generations to about 7 .7% in the last three generations. In cage B the decline was more severe, falling from a high of h3. 32% in the seventh generation to 7.18% in the thirteenth generation. The environmental conditions of the two cages were similar and the experiments were run simultaneously. Hence, the data obtained in the two may well be considered tOgecher. Some type of equilibrium seems to be re ac hcd in expe r iment III at about 7.5p and this sane percentage was reac}ed in experiment IV. The results of the crosses begun.with smaller preportions of the sepia insects than would be obtained from a dir: act F2 generation are recorded in Tables IV and V. These two experiments, V and VI, showed a rise from lO.L6p and 16 .205 in the second generation to 19.3;fi and 23.12% in the third gene ration for cages A an-i 8 re we tively. In cage A generations three throngh thirteen show an average of 18.72; with the eatest deviations being 16. 01% in generation eleven and 21 hlb in generation seven. The fourteex 1th generation showi1527.515 of the sepia phenotype is not consistent with the remaining portions of the data, and any uncontrolled conditions which might have caused such an increase is 13 unknown. In cage B, the average of generations three through eight is 20.76%, with a high of 23.70fl and a low of 17.72%. In both cages a con- dition resembling eQuilibrium is evident at values that are close for the two cages and which is higher than the initial freguencies for the populations. The largest populations which were observed were contained in cages C and D in experiments VII and VIII. The data which .re recorded in Tables VI and VII show an initial rise in the F3 generation. Thereafter, there is a decline in cage C to 11.36; and a subsequent rise to an average value of l7.hh% in the last three generations. In cage D there is also a decline following the initial rise, which is followed by another rise. The last seven generations again suggest an equilibrium at an average of 22.68% with the greatest deviations being 2h.hl% and 21.30%, which is a relatively low range of variability. In most of the populations, a condition resembling some type of equilibrium was established. Hith the use of bottled medium, this equilibrium was reached at about 7.5% for the experiments III-ané—IV. In experiments V, VI, VII, VIII, this value varied from l7.hh% in cage C, 18.72% in cage.A, 20.76% in cage B, to 22.68% in cage D. In the pepu- lations of cages A and B this value was above the initial value of the population, below the initial value in cage C, and about equal to the initial value in cage D. In all except experiments I and II, separate counts were kept for the males and for the females. In all, sixty-one generations from six experiments were counted with the relation between the sexes in mind. 1h x2 was significant (at the 5% level) in four cases and highly significant (at the 15 level) in four other cases. Since the percentage of recessives favored the males in exactly half of these eight cases and the females in half, these data as a whole were not assumed to have great significance, and calculations were based on the total percents obtained by combining the males and females. One important difference was noted between the cages in which bottles of medifi'were used in comparison with those in.which the petri dishes were employed. The amount of moisture which condensed on the sides of the cages containing the petri dishes was much greater than those contain- ing bottles. During the time in which the medium was being autoclaved, the bottles were stOppered with cotton plugs which prevented the entrance of much moisture. The petri dishes were covered with their glass covers and the condensation of moisture inside these dishes was considerable during the cooling period following autoclaving. The media supplied in the two cases differ in moisture content and conseQuently the humidity within the cages was effected. Ehen the bottles were used, no moisture was evident on the sides of the cages, whereas, the use of the wetter medium in the petri dishes caused a clouding of the glass portions of the cage. In each case, the experiments were begun with a fairly large number of flies. Nevertheless, the growth from this point when considered generation by generation typically follows the latter portion of Pearl'sh8 population curve. That is, it grew at an accelerating rate until a maximum was reached in each generation which was dependent upon the 15 density of population which could be supported in the available environ- ment. Thereafter, the growth between generations was ever less until it was no longer perceptible. In such pOpulations of Drosophila as these, if the temperature is constant, the available food supply has been shown to be the most important limiting factor of the population.3h This would appear to be so here also, since the competition in the larval stages was very severe. If the maximum pOpulation number which occurs continuously for a few generations is roughly estimated in round numbers and the average area and volume of available medifi'are calculated the following correlation table may be set up: Estimated Maximum. Area of’Media Volume of Media Experiment Size of Population sq. cm. cu. cm. A B c III 1800 565 1693 IV 2100 565 1693 V 1300 628 628 VI 1800 628 628 VII L500 1767 2651 VIII . h700 1767 2651 . 6 AB , ,. , By u51ng the formula rAB ==-Erfifer-' , r38 18 equal to .9776 and rAC 18 .8999. These correlations are so large that the maximum pOpulation size must be dependent to a large degree on the available food material, which is what would be expected. Table VIII combines the results of the two collateral experiments. The first involves the testing of mate selection. The data are limited but they indicate that the two types of males are about equally competent in mating and that the sepia female shows little preference for either 16 type of male under the conditions tested. The females listed under the section unknown were either lost in the last transfer or became stuck in the medium before eggs had been laid. The last portion of the table lists the average number of offspring produced by ten parents (five males and five females) for the twenty bottles tested in each of the five crosses. These figures are a rough estimate of the reproductive abilities of these types of crosses under more optimal conditions than those found in the cages, since competition during the larva stage was much less in the bottles. 17 TABLE I EKPEhIMLNT I: PERCENTAGES OF SEPIA FLIES FROM BOTTLE COUNTS IN CAGE A Number of Percentages Total Total Total Generation Offspring Of Sepias in Number Number of Percentages Each Bottle Each Bottle Offspring Sepias Of Sepias F2 289 25.61 11b? 258 22.1h 211 19.h3 27h 19.3h 139 2h.h6 23h 22.22 F3 180 h2.86 965 313 32.hh 187 h3.85 192 30.73 223 29.15 223 21.08 F4 186 b.8h 1179 108 8.82 330 6.6? 200 5.00 288 15.32 215 11.63 F5 295 5.76 1288 188 11.21 310 13.55 258 13.57 228 11.8h 193 11.92 F6 222 30.18 11h9 260 22.63 173 26.59 238 15.13 213 28.68 303 16.50 F7 291 25.77 l3h8 257 19.07 2&2 9.50 286 19.58 282 21.07 287 18.12 18 TABLE II EXPERIMENT II: PERCENTAGES OF SEPIA FLIES FROM BOTTLE COUNTS IN CAGE B Number of Percentages Total Total Total Generation Offspring Of Sepias in Number Number of Percentages Each Bottle Each Bottle Offspring~ Sepias Of Sepias F2 261 28.52 1088 222 21.26 220 16.82 210 16.67 186 23.12 167 25.75 F3 235 27.23 1091 290 26.58 198 20.10 ‘ 220 27.73 187 38.76 255 23.92 F4 266 52.26 1059 879 85.23 163 81.72 128 9.68 302 86.36 208 58.82 F5 328 9.57 1087 205 19.58 229 88.58 188 18.88 133 16.58 177 9.08 Fe 282 28.10 1058 287 27.23 278 21.53 170 15.88 173 89.71 195 28.10 F7 275 32.36 1280 385 31.05 210 80.88 260 16.58 228 23.25 267 83.07 19 TABLE III Mean-mas III 151-11) IV: PLJLC J‘ETAGES OF SEPIA FLIES FROM CAGE COUNTS IN CALiES A AND B Percentage Percentage Total Total Total Per- Generation or Sepia Of Sepia x2 of Sex No. of No. of centages Males Females Differences Flies Sepias of sqgias CMEA F7 18.86 18.08 .020 789 113 18.32 F8 18.97 18.85 .008 1828 213 18.92 F9 13.83 9.07 8.880 1600 185 11.56 F10 8.82 8.37 .128 2078 179 8.63 F11 5.75 6.00 .033 1176 69 5.87 F12 6 .02 9 .37 5 .501 1382 105 7 .60 F13 7.86 8.70 .525 1080 83 7.98 F14 7.97 7.03 .813 1323 100 7.56 F7 81.82 85.08 1.262 1175 509 83.32 Fa 31.87 30.93 .068 1875 555 31.20 F9 22.19 20.67 .681 1871 801 21.83 F10 17.95 15.62 2.120 2183 367 16.81 Fll 13.59 15.52 .988 1265 188 18.55 F12 8.77 12.12 6.898 2151 221 10.27 F13 7.16 7.19 .000 1867 137 7.18 20 TABLE IV EXPERIMENT v: Pziczuiioss OF SEPIA FLIES FROM JAGE COUNTS IN ones A Total Percentage Percentage Total Total Per— Generation Of Sepia Of Sepia x2 of Sex No. of No. of centagcs Males Females Differences Flies Sepjas of Sepias 2 12 .76 7 .63 .663 526 55 10 .86 3 21.11 18.70 6.886 128 28 19.35 8 21.05 16.33 .829 125 28 19.20 5 17.55 25.76 2.079 258 50 19.68 6 17 20 17 28 .000 302 52 17.22 7 20.68 22.76 .210 381 73 21.81 8 17.88 15.79 .886 911 152 16.68 9 19 .98 18 .31 .519 1233 237 19 .22 10 18.89 20.51 567 887 172 19.39 11 15.82 16.76 .260 787 126 16.01 12 18.05 19.23 .785 913 170 18.62 13 19.33 19.02 .229 1297 289 19.20 18 28 .57 26.33 .887 1352 372 27 .51 EXPERIMENT VI: TABLE V ERCENTAGES OF SEPIA FLIES FROM CAGE COUNTS IN 21 ACE B _‘ __ PerbEEtage PErcentage Total Total Total Per- Generation Of Sepia 0f Sepia x3 of Sex N0. of No. of centages Males Females Differences Flies Sepias 0f Sepias 2 16.39 16.01 .068 2352 381 16.20 3 25.58 21.25 3.250 1276 295 23.12 8 25.15 22.39 1.500 1826 338 23.70 5 17.88 21.58 3.831 1828 365 19.97 6 16.75 18.38 .789 1732 307 17.72 7 17.22 22.11 7.232 1908 392 19.90 8 22.77 17.81 8.008 1792 360 20.09 22 111B LE VI EXPERIMENT VII: PERCENTAGES OF SEPIA FLIES FROM CAGE COUNTS IN CAGE C Percentage Percentage Total Total Total Por- Generation Of Sepia 0f sepia x3 of Sex No. of No. of centages Males Females Differences Flies Sepias 0f Sepias F; 25.58 19.79 5.352 1136 257 22.62 F3 30.03 28.36 .088 1887 823 29.23 F4 18.36 18.00 .036 1677 305 18.19 F5 16.88 16.98 .003 1660 281 16.93 F6 13.23 12.80 .063 1608 210 13.06 F7 11.85 11.27 .053 3979 852 11.36 Fe 15.86 17.58 1.777 2257 372 16.88 Fe 13 .6? 13 .87 .030 3556 890 13 .78 Flo 13.09 18.82 1.671 8876 616 13.76 Fn 16.37 18.99 1.622 8516 706 15.63 F12 16.23 19.56 8.357 8858 801 17.98 F13 16.98 17.18 .030 8225 722 17.09 F14 18.21 16.31 2.260 3589 612 17.28 23 TABLE VII EHERD’UNT VIII: PdtCETITAOES 0F SEPIA FLIES FROM CILGE COUNTS IN CAGE D Percentage Percentage Total Total Total Per- Generation Of Sepia 0f Sepia x3 of Sex No. of No. of centages Males Females Differences Flies Sepias of Sepias F2 23.79 20.75 1.181 886 197 22.23 F3 23.35 26.88 2.253 1810 352 28.96 F4 20.20 20.63 .031 1079 220 20.39 F5 16.56 19.16 1.785 1577 278 17.63 Fe 15.37 15.83 .102 2511 392 15.61 F7 12.65 12.62 .000 2557 323 12.63 F8 22.59 22.68 .001 3011 681 22.62 F9 21.80 28.75 3.822 3119 725 23.28 F10 20.60 22.06 1.086 3829 732 21.35 F11 21.58 21.05 .183 8832 988 21.30 F12 28.80 28.00 .802 8708 1188 28.38 F13 21.28 21.68 .117 8256 918 21.88 F14 25.08 23.81 1.027 1153 28.81 8723 TABLE VIII PART I: MATE SELECTION 28 Sepia ‘Nild Females Females Cross: Sepia female x (sepia male and wild male) Fertilized by sepia male 27 Fertilized by wild male 23 Fertilized by both males 13 Fertilized by neither 10 Unknown 2 Total E Cross: (Sepia female and wild female) x sepia male Fertilized by sepia male 28 25 Not fertilized 15 18 'Unknown. 1 1 Totals E0 E0 Cross: (Sepia female and wild female) X'Wild male Fertilized by wild male 26 25 Not fertilized 13 13 Unknown 1 2 Totals 80 80 PART II: REPRODUCTIVE ABILITY Average Number Offspring For Twenty Bottles Cross: Heterozygote x heterozygote Ifild X‘Wild Sepia x sepia 'Wild male x sepia female Sepia male X'Wild female 288 283 258 261 256 DISCU SSION 25 DISCUSSION The main objective of this research was to test the validity of Hardy's Law in controlled populations of Drosophila melanggaster. Even as the study was just beginning, however, there was evidence that the ratios expected by Hardy's Lax were being modified in some way. The numbers were Quite large and theoretically, deviations from the expected ratios should not have been great. However, in experiments I, II, VII, and VIII, each of which was begun as a simple F1 cross, the F2 ratios in each case were less than the 25% theoretically expected according to this law. If the Fz's of these experiments are taken together, x2 is egual to 19.155 which is a deviation from the expected which is highly significant. This evidence did not, in any sense, mean that Hardy's Law was erroneous and that the failure of these laboratory pepulations to follow it in any way disproved the law. Rather, it would appear that the expectations were too naive and that one or more of the basic conditions imposed upon Hardy's Law to make it valid were not being met. A natural population is not a static entity, but a dynamic one which undergoes constant changes,2 and the proportions expected by Hardy's Law are constantly deviating due to certain pressures. 'Which type or types of pressures then might be acting on these more carefully buffered laboratory populations? There are several types of nonrecurrent change which might be termed accidents.57 These might be useful in explaining a single or at most a few erratic ratios, but not an entire series of percentages such as was obtained here which deviate from the equilibrium values of Hardy's Law. In addition, there are systematic pressures which bear constantly upon a natural p0pulation, namely, mutation, mi- 55 gration and selection. In this particular problem the effect of mutation should be negligible and that of migration, zero. The only remaining pressure then is selection, which is the most complex in many ways because so many variables are concerned with it. There are numerous ways in which it may effect the equilibrium ratios of Hardy's Law, and many of these will be discussed here in an attempt to find an explanation for the results obtained in this research. It is important to remember that in any phase of the discussion of selection, and especially in the application of mathematical methods to this pressure, biological facts must be reduced to a mere abstract of their real complexity.S That is, only one phase will be discussed at any one time, but perhaps not one variable, but a combination of several are acting simultaneously to produce the selection phenomenon as it is observed here. Possible Modes Of Action Of The Selection Pressure Natural selection alters the frequencies of the genes in the next generation due to the diSproportionate contribution of the carriers of the different genotypes in the preceding generation. Some of the geno- types are adaptively incompetent and may be eliminated or reduced by natural selection. Other genotypes possess Optimal fitness in certain (D nvironments, and if these environments recur freauently, hese types may becor e lasting components of t11e population As we have observed in this research also, the visible effect of selection is upon phenotype. In this case, it is upon the sepia pheno- (I) type. The de re 0 of corr68p oncence bet"ee neenotype and phenotype is fundamental in determining the permanent reSponses of the population to 83 the selection pressure. Nor can we even assume that the selection rate for a single gene such as se 6pia is constant. There are conditions under which one gene has a selective advantage onl1y until a certain gene-ratio is esta oli shed while for higher ratios it is at a disa Mv tags. In such cas es the gene ratio will be most stable at this limiting value for the selection and this value will tend to be restored wh6never it is disturbed from either direction.3 Neither is the adaptive value of a type necessarily proportional to its survival value at all the developmental stages of the organism. In some cases it has eon shown that some types which show relatively higher mortalities than other types between the egg and adult stages proved nevertheless to be adaptively superior to the latter 1 in 6 6.6617. 61136.20 (D Selection may be complete. That is, it may involv ve al6 thal gone, so that one or the other of the cenotypes is completely eliminated every )1 . generation.4 This could not have occurred in this experiment, howevez, :3 because all genotypes were present 1 the population and an e5111‘er11m was reached at the approximate poi n +s alre a.Cy noted 0r selection may be partial, so that only a pro1ortion of certair1 genotypes are eliminated L in each generation. This type of selection is mos tcommonly against the homoz35oas recessives, althOU‘J not no cessaril3 so. If selection is D partial and against the recessive gene in the homoz3omus state onl3, the frequenc c3 of this gene in the next generation is where s eguals the percent of re cessives r6 ject6 d and r equals the fre- guency of the recessive gene. This type of selection is characterized by a constant decline in the recessive phenotype in every generation so that after an infinite number of generations the gene should be elimi- na6 ed, or nearl3 so.LL Adverse selection against a recessive rone is most effective when the homozygous recessives make up a fairl3 large proportion of the population. As the prOportion of homozygous recessives decreases, the effectiveness of the adverse selection also decreases but at a much slower rate. A decline of this sort has been observed by several workers. L'Héritier and Teissier)S found that in a laoora tory pOpulation of D. aelanogaster, the gene "‘oar" with an initial frequency of .999 was almost completely replaced in 15 O da3s b3 the wild 5e r1e when a wild fly accide entl3 contaminate d the "0 ar" population. In later experiments, at the end of approximately 600 days, the frequency of the "bar" game had been reduced to .0037 and .0105 in each of two such pOp ulations .36 These q 0 0 same two workers aided by Neefs)7 found that the gene for vestigal wing was selected against in similar populations. If a breeze blew through 29 the cage, however, the vesti al gene was favored. Although these 8 experiments were not carried on long enough to determine whether com- plete elimination would take place, these geneticists believe that neither "bar"nor Vestigal would be eliminated entirely from the pcpu— ’ lation, but that an eduilibrium condition might be reached for a rather 31 low freQuency of the mutant gene. Gordon32 released a pepulation of 36,000 individuals of Q. melanoiaster carrying the gene for ebony with a freqaency of .5. This species is not endemic in England. After 120 days, the frequency of the ebony gene had been reduced to .1. This experiment could not be continued because cold weather wiped out the entire population. If the type of selection just described was acting in our sepia- wild populations, a constant decrease in the frequency of the sepia phenotype should have occurred. Such was not the case. A single factor may be in stable eguilibrium under selection if the heterozygote has a selective advantage over both homozygotes. An inSpection of the equi- librium values which have been obtained in the sepia-wild pepulatiens indicates that such is probably the type of eauilibrium which has occurred in this problem. Actually, a type of equilibrium will be obtained if the heterozygote is either better or worse adapted than the two homozygotes. Fisher3 believes that the two cases will not exist equally freguently, however. If the heterozygote is at a disadvantage, the equilibrium is unstable, for if Aq equals 0, then 3 (at the equilibrium condition) is one-half, A . . . . zero or one. At q equals one-half, complete or partial elimination of 30 the heterozygote will not change the gene frequency in the population. Therefore, random mating will restore the original zygotic proportions in the succeeding generation. But if‘g eguals any value less than one- half, selection diminishes the frequency of the rarer allele in each generation until q becomes zero. And the eQuilibrium at one-half is always unstable and if lost, cannot be restored.5 However, if the heterozygote is better adapted, the eQuilibrium is more stable and will tend to persist until this stability is upset. Equilibrium will be restored once it is disturbed. At this equilibrium position, the adaptiveness of the entire pepulation is enhanced at the price of the production of some less well adapted individuals. In this case of balanced polymorphism the average fitness of the individual in the population will be greatest when the equilibrium condition is reached.2 If the adaptive value of the heterozygote is taken an unity then those of the homozygotes are (l-s) and (l-S) respectively where s is the selection coefficient against the homozygous recessives and 3 against the home— zygous dominants. The frequency of the recessive gene, q, at which equilibrium is established is A S q = (s + 85 and the frequency of the dominate gene, p, is S A _ 19 p is + 55 . If p and q are the frequencies of the dominant and recessive genes in any generation, then the frequency of the recessive gene in the next 31 generation becomes 9(1 - 8a) l - sz - sq2 and the amount of change in q per generation is 29(Sp - ss) 4 = q l - Sp2 - sq2 . - o 01 o - o ,.. o A At equilibrium Aq is equal to zero and the gene frequency ratio, u, at this point is vi A 11:2: 9 Uflm It should be noted that in the case of the superior adaptive value of the heterogygote, the equilibrium values of a gene frequency are independent of the initial gene frequency of the population and are com- pletely determined by the selection coefficients of the two homOZygous genotypes. This equilibrium condition is a stable one and regardless of the initial frequency, it will be eventually reached since these equilibrium values of the gene frequency give the maximum adaptive value of the population as a'whole.5 This phenomenon may be observed in the sepia-wild pepulations in this study from experiments V, VI, VII, VIII which may be considered together since the equilibrium values of them all are very close. In experiments V and VI, the equilibrium values, about 19% and 21% reSpectively, are above the initial values. But in experiment VII it is below the initial value and about equal to the beginning value in VIII. This condition lends considerable evidence to the supposition that the pOpulations in Question show a superior adaptive value of the heterozygous types over the homozygotes. 32 If the proportions of recessives, heterOZygotes and dominants are R, H, and D respectively, then (1 - s)rn2 ’ H = 2rn(l - rn) P n P 9 Rn: <1 - Sm - rnf and Dn = P where r equals the frequency of the recessive gene, S and s are the 2 h selection coefficients, and P = l - er 2 - S + 2Srn - Srn . n The explanation of the superiority of the heterozygote seems most adequately to explain the results obtained in our sepia-wild populations, since in each case an equilibrium was established which seemed to be independent of the initial values. The results obtained from the various populations may be subdivided into three groups. The first group would be composed of experiments I and II in which the counts were made from bottles removed from the cages rather than directly from the cages them- selves. This first group which proved only to reveal the weakness in the sampling technique need not be considered further in this analysis. The second group would include experiments III and IV since they were carried on simultaneously and in both cases medium was provided in half- pint bottles. The equilibria reached in III and perhaps approached in IV were at the same level, although the initial frequencies differed quite markedly from one another. Experiments V, VI, VII, VIII compose the last group. Although the population size in the first two cases differs from that in the last two, it would seem that the similar results would warrant their consideration 33 as a whole. The equilibrium points of them all vary within rather small limits. Relations to the initial values have been discussed. In this particular analysis it has been found that the formulas just mentioned in connection with heterozygote superiority are practic- ally useless since they all involve knowledge of the actual gene frequencies, p and q. Sepia is completely recessive, so the frequencies of the homozygous wild and heterozygotes cannot be determined. If the populations had not shown selection, but had been in the equilibrium described by Hardy's Lax, then the square root of the fre- quency of the homozygous recessive genotype would equal the frequency of the recessive gene. In our F2 populations, the theoretical distri- bution of the genotypes according to this law, would be .25 aa, .50 ha, and .25 AA. The genotypes AA and aa are apparently selected against. The preportions of AA and ea in the population would, therefore, decline and the frequency of the Aa genotype would increase preportionately. Obviously, the square root of the frequency of ea in this case would be less than .5, and the same would be true for the square root of the frequency of AA. Therefore, the sums of the square roots of the per- centages of AA and aa individuals would be less than one. Consequently, since this reasoning is true in all generations, it is evident that the square root of the frequency of the aa genotype is less than q, which is the actual value of the gene a. Since p and q are unknown and cannot be calculated, the selection coefficients and adaptive values of the three genotypes cannot be determined. Moreover, we cannot even be certain that S and stare stable values. It is quite probable that they 3h might vary with the pepulation density and with the frequency of the recessive gene in the total population. In addition, it is impossible even to state whether the homozygous wild or the homozygous recessives had the higher adaptive value. There is, then, only one relationship that is known, and that is that the equilibrium value of q is greater than the value of the square root of the frequency of the sepia phenotype at the equilibrium condition. If the square root of the average frequencies of the sepia phenotype at this stable condition as they have already been stated are taken, a is greater than the approximate values of .27h for experiment III, .L32 for v, .Lss for VI, .hlB for VII, and .h76 in VIII. The type of equilibrium observed here has been found rather fre- quently by a number of investigators. The superiority of the hetero- zygote has been shown in the experimental laboratory pepulations of 50 Reed and Reed, who studied pepulations of D. melancgaster in which the homozygotic condition for an inversion of the X chromosome was almost completely lethal when in severe competition with the wild type. However, the inversion flourished in the :eterozygous condition so tlat it was not eliminated from the pepulation. An equilibrium was quickly establishai with 235 homozyrous females, 28) hemizvgous normal males, L2£ heterozygous females and 2S hemizygtus inversion males. This shows that the heterozygous female had a strong selective advantage over the 20 . c 01 o ' _‘. ‘I - homozygous types. Dobzhansky found an equiliorium of a third chrono- '\ do bscura at 793 and 533 3.4 some inversion in two populations of D. Lee- reSpectively for the inversion in which the original breeders were obtained from two different natural pepulations. This same type of IO esuilibrium was found by Cunha inr regard to the trait for light and dark abdomen color in Brazilian populations of D. polxmorpha with sur- 0 .v’ '1 1'1 1a? 9 Vival values of .50 for Eh, 1.00 for he and .23 for so. Freire-Mai also observed a stable value for dark and light abdomen color in Brazilian populations of D. montium. Additional details of this type of equilibrium were obtained by L . . Kalmus‘r0 in populations of D. melanOgaste r With reSpect to the trait ebcn‘. He found that at a hi3 ier temperature and highs? humidity, the wild type was adaptively superior to ebony. But the ebony gene was found to be superior under the opposite conditiozs However, in each of the cultures, eventually an equilibrium was approached even though the point of stability differed with the varying conditions. L'EM eritier . 3C . . . . and Teiss ier J found that in their pepulations of ebony and Wild, equilibrium was escablish ed at about 15% of the ebony flies. This same situation has been observed in natural pOpulations by Dobzhanshy and Levene.2h These two workers found that the eggs laid by D. pseudo obscwra are in conformity with the HardyJJeinberg Law in the proportions of homozygotes and heterom gotes for different types of the third chromosome. But a differential mortality takes place between the egg and adult stag-es which favors the heterozy cote. Thus far, we have considered that the wild-sepia pepulations studied have had the ratios exgected by Hardy's Law modified by the pressure of selection which favored the heterozygote. It would be interesting to attempt to exilain what environmental factor or factors n:ight have 3o contributed to this selection and in what stage of the life cycle it might act. Period Of The Life Cycle'Uhere Selection Exerts Its Pressure Selection may occur at any stage in the life cycle. If there is a differential fertility of the adult flies, it may be in the number of eggs or sperm produced or in the relative survival ability'cf the gametes. l, A smaller preportion of the eggs of one genotype may hatch as has been L7 shown to be the case in some of the other mutant stocks of Dresgphila. One type may be at a disadvantage in the competition in the larval st ges. Perhaps one is better adapted to survive the pupa stage. The time and length of reproductive activity may vary. In the adult stage it may occur in the selection of mates. Nor is any one of these constant, but is modified by numerous environmental conditions, such as temperature, food and population density. Mate selection has been observed in certain populations of Drosophila. (In, Reed and Reed)1 found that natural selection favored the wild gene in laboratory populations of this insect to the extent of eliminating the gene for white eye. They discovered that the ratio of red males which succeeded in mating compared to white males was 1.00 to .75. Thus they were able to conclude that selective mating was the most important factor contributing to the decrease in the white gene. This same conclusion was reached by another workerl2 in.p0pulations of wild and yellow-white. t 2 . . Rendel) observed the courtship pattern in D, ubscura of yellow males and found that it did not differ from that of normal males except that it 37 was longer. But the normal female resisted the advances of the yellow . L5 . male. herrell in his selective mating experiments also concluded that the occurrence of non-random mating was due primarily to the be- havior of the female in D. melanogaster, but that in time, practically all females of a pOpulation would be fertilized. Sturtevantgh has also stLdied the problem of sexual selection in this fly. His results indi- cate that actual choice is not involved. But that any female willing to mate will accept any ma e and a male ready to mate will do so with the first female which will let him. However, if the female is not willing, the male of the more vigorous stock will have an advan age. As a result, the weaker female is most often mated with. It was with this possibility in.mind that the experiment on mate selection with the wild a d seiia stocks was begun. However, the results are startlingly close for the small numbers worked with. The wild and sepia males seem to be egually successful in mating with both types of females and either of the two males seem to be acceptable to the females. The numbers are not large enough to be conclusive but they indicate that mate selection probably does not play a very important role in these sepia—wild populations, in producing the selection pressure. In addition, the numbers of offspring produced by the different types of wild-sepia crosses indicated in Table VIII show little vari- ability in their ability to reproduce at uncrowded, more optimal condi- tions. This information would cause me to believe that the reproductive capacities of the different types of crosses does not differ too much, and that under Optimal conditions, about the same number of offSpring 38 reach maturity. Moreover, differential mortality in the adult stage may be eliminated since never more than two or three dead flies were found in any cage during a generation's time. The most probable point of the action of the selection pressure is during the larval stage. The cages were supporting the maximum number of flies possible with the available food supply and the medium dishes were Quite crowded with larva and the, competition at this point was quite keen. Possible Agents of Selection Numerous environmental factors act as agents of natural selection, and many of them have been studied in Drosophila populations. Seasonal variations have been observed in the relative frequency of gene arrange- ments in the third and sex chromosomes in two of three populations of D. pseudoobscura obtained from localities near San Jacinto, California.16 hree Moscow pOpulations of Q. funebris,25 and a certain pepulation of 2. robustah2 from Virginia also showed seasonal variations. In the latter case, the changes were significant in males only. The relative freQuency of black hamsters has been reported in some regions of U.S.S.R. to under- go regular a d significant changes from season to season.31 All of these reports indicate that selection may act quickly in changing the freguencies of gene arrangements and of genes. Probably several environmental factors which change with the season cause these changes. 57 In 2, pseudoobscura,‘¥right and Dobzhansky have found in laboratory populations that certain of the third chromosomal arrangements are better 17 adapted at higher temperatures and others at lower temperatures. These 39 varying adaptive values to temperature were also found in certain chromo- somal variants of the second and fourth chromosome in this same Species.23 . 38 1" I ’- ' ° 1 Hovanitz found tuat tne frequency of white females in the butterfly, Colias chgysotheme are largely correlated with changes in climatic con- ditions. Large size seems to be similarly correlated in the English Sparrow.8 Extreme cold temperatures which caused individuals of D, funebris to hibernate was found to favor one inversion and to discriminate against another .2 8 Other environmental factors not clearly outlined serve as selection pressures on pepulations of Q. funebris observed by Dubinin and Tiniakov.26 These workers consistently observed higher frequencies of certain inver- sions in the populations of Moscow and other cities than in the rural districts nearby. Another factor which acts as an agent of natural selection is popu- lation density. Pearl and Parker149 have found in early studies of Drosophila that the reproductive rate per female declines as the pepu- lation becomes more dense but the decrease is at a decreasing rate at the highest densities. Crowding has been seen to play an active part in some laboratory populations. Moreeb'6 found that the mutant gene ebony in pOpulations of D. melanogaster was nearly as viable as wild if there were 1 little crowding. But the viability of ebony decreased as the crowning continued and as the competition became more intense. Other workers have observed that some inversions on the second and fourth chromosomes23 and in crosses involving three gene arrangements on the third chromosome17 of D. pseudoobscura there was a differential viability in accord with the ho pOpulation densities. ther gene arrangements of the third chromosome were little modified by these conditions. The incidence of black coat color in some hamster pepulations of the U.S.S.R. was found to be positively correlated with the population densities of that species.30 The additive effects of temperaturea and humidity on the freguency . . . . was . . . of eoony in Drosophila populations wee-mentioned very early in this Discussion. Khich of these factors then seem greatest in influencing the selec- tion observed in the wild-s epia populations of this research? Since temperature was carefully controlled, the changes were not due to any fluctuations here. However, the temperature was probably very important 1 i. aete -rmining the adaptive values of the three genotypes and therefore the actual oint of equilibrium. The humiditr was aniroximatelv equal in a i a 2 each cage throughout any one exq: eriment, since it seemed to (1ep311d on the wetness of the medium provided. Its probable import nce in produc in the widely varying equilibrium points at about 7.5S in III and—I¥ and -’ 1 p an' 22.18fl in experiments V-VIII has alread" been men- ) betw {3511 17.) '11 4.-. tioned. It is probably one important factor in determining this differ- once. Another factor already mentioned is that of population densities. (J This crowding was Quito severe during the larval stages, and intense competition was, no doubt, of utmost i;:iportanc as a selective agent. The exact importance of any one of these factors would have to be determined experimentally, but I believe that it is safe to as sane that as a group, t emf) era ture, humidity and crowding were of some importance in determining the equilibrium points in the wild-sepia populations. L1 The Importance Of Balariced Pol3morphism As Determined By Selection.Pressures \‘v (D f I In natural populations the b lance‘ polymorpiii ism observe I in th_ D-v‘ .4 pOpulations D‘u fers the Sfecies a ainst envirormental change ani at t 3 sxie time it does not consume or deplete the store of hereiitary vari- ability pres 2r nt in the population. The total adaptiv: ability of the ‘tulation is thus greatly enhanced. 18 A balanced polymorphism also preserves Certain of the new g;ne rranwe-ents and g-;n:3 mutations whioh a.ise from time to time in any t naintains these tao most im- f—lo population. Tiis is imlortant beiw use portant rcw materials of evolution,13 which are present mach mor: fie 1uently in Drosophila pepulations than would be expostzd b"r casual 11:39 T. J (I. d. PI; (0 :5 r5 acted upon by restriction 9;. observation. ese raw materia 1'- ~.' - .L, ' .— 3—1. AP 1.3“ '. L' ,. of pa mul Lion DlLU, he irzl s; lec ction and the umVfilurlelb of iso lacinp 1 rt.) 1‘" '0' t ' t‘~r3 . -n~~rw-1 yo 1' 1rn1~ "1."h I I'1~JC iuIIlquS O lnsllre 11L 1);. 'J ‘_',. 1v 3 O r‘ v _, *k». Lulu“ . m - ' ' 3' ' I . ‘ . “ . L- r5 ~ ‘ P ‘2 ".“I w -— A “ ‘( "' L . ' -h; plrdlooldfls in this Stacy were all fairly lalbe once b in the in “A .,r‘+—. ‘_,“ .v~ ._ “._ _f ‘3‘ 1' a" “I_'— h, ' ‘_‘_I r‘ _' 1 W h" “‘1 irst geno.acichs in ex;3riment J. If the else had contihved orail and ~-~ - , ‘ ~‘ 0 rar "\ 1H’L . r‘ «w V. r- /\' \ h -. ' ' ‘ L - "jfirw ‘\-. n a CUlASiI ‘Fjr‘aISle nwyml‘ol‘ OJ. Lilinbl Kn; Ulbllb 1.?- -s }D‘;L3.l.| wld—(J’J‘zki DU {J «4.00 a \.,'.L .LJ- H v” y n 1 y ‘3 . ‘V : W. . ‘ " ‘1 " '-' W . Z ‘V'.‘ 4‘ ‘1 0"" 'I " ’ in 3 lion the edllllbrlhm values would have b9'4fl exieot3l iiom chance 0 a 1: ective Size of the breeding pepulation *‘J (.1 ' v‘ 1 x 7“ .7 fl . 4- -~ I‘ alone. In such populations the ei r—I/ must be taken into accourt ” ani tie lADLizolP" coefficient must be 22 .- - ..- '0 ,- ... ..-..1..V...- uptkfmlhv. since there wlll be a corieseohcin 1.. ‘r‘ - ,.,-.-, n “-1— 3““ x -\Ir\.‘ ~' .r“ H q‘ n -\3 .1!‘ s V: \ cu: to these laooOFS. In these Smallur populations, oiancc piaJs a xloh - ‘I ~"‘ -\ 1 ' . ~ 1 ,~ ~ ’3 . . A -‘,1wrr-,- xvi h . 1.: .-u """y‘. r D ~ .. ”‘1‘ bitutpr role in pioducing iJlu ely .;3 an populations. Tile t3pe Ci uiiiu has been seen to play an impor ant part in the oificrentian of certain I!“ R) r~ 1 -' \a HAL. . . -_-- h ‘.- .: ~~ ,- V” . ‘0 local EVHUAJUlUHS in several bLUClSD of LF Shuhll? the apraur t of Ciffer— <, p. H I V_ "I " C 0 fl ' If? I) ention eugenaing upon the effective Size 01 the br;ee, 1»-:L; .--fs, anc Georgm 7cs Insects et Selection Naturclle, " C. h. D - 'n cram/1’- " -7 1 /\AH vaer'L'VLQ, L01. ()Lf, fjp. WY-m /~’/, 1;)! Hovantiz, Hilliam, "'l“ne Distribution of Gene Freq lencies in7fi d v--. - .,1- . 4--, nafi ._ ' L ,3 .“ r ‘Ji. «Dad-ti l‘dll b (Jf 0’ 1... ~ 5.; G .4- .~- KILS‘; 3 VIC-1 l. 2 9 ’ I'll . 31- :0 , 13L;4 . Ives, Philip Truman "The G;n:tic Structtre of.im3ricn.Icfu1a;igns cf Droso: hila Pelano*act~r " Genetics, Vol. 30, pp. le7—l93, 193'?) o Kalmus, H.,' 'l:1aptive anL Selective Responses of a Population of Drosollila Nolan;éeet;r co mt ining e and e+ to Differercie‘ in TLmierfmtwre, humidity, and to 3~lection for DeveIOpmental Speed," Jourr: 1 of' Gin wii cs, Vol. N7, pp. 58~C3, b925-L6. I Keller, P. C., "cnetics of Na tural Populations. III. Ger ne arrange- ments in Pepulatiors of DrosolhilgpPseudoohscura from Contitous Localiti2s,‘ Genet ,VOl. 2L, pp. 22-33, 1939. Levi+ an, Max, "Selective Differe neies Be éen Males and Females in Drosophila Rcbust3,' Lmerican Naturalist, Vol. 85 ,pp. 3c;— 3C8 ,1951. Mather, K., and B. J. Harrison, "The Manifold Affect of Selection," Heredity, Vol. 3, pp. 1-52, l9h9. chwen, RO7QPL Stanley, "The heactions to Li"ht and to Gra it‘ In DrQSophila and its Mutants," JOH”r°1 0f “ir~r‘mchb-‘ eeleio', Vol. 23, pp. 19-105, 1918. Nerrell, David J., "Se active “aiilg in Dros.shila Melanogast r," .1.— Genctics, Vol. 3h, pp 37 0-389, l9e9. ioree Ray, "Experiment tal Pleasurem ent of the hzlative Viability of the Mutant “bony in Drosoi hila Nelarovaster," American M"Ldrlll° ,VOl. L6, pp. L)-LU, 1932. Morgan, T. H C. B. Bridges and A. H. Sturtcvant, "The Genetics of -rosopnila," Bibll’"r_ph]1 Genetica, Vol. II, pp. 1-292, 1923. 19 Pearl 119.. ey'mont‘ , -, "The GrOUth Of Po12ulations, " Th-2 Quarterli Review of Biolod , Vol. 2, pp. 532-:L8, 1927. Pearl, H., and S. L. Parker, "On the Influence of De nsitw of Pepu- lation Upon Rate of Reproduction in Dre Msoplila, " Pr: e~o1n~s of q5t1oflql (csdor L of Science, Vol. 8, pp. 212- ~2l9, 1/22 . 1 Need, S. C. and m, J. Head, "Natural Selection in Laboratory Popu- lations of Drosophila, " Evolution, Vol. 2, pp. l7f~lté, lQLE. pLed, 3. C. nd 3"g. Reed, "Natural Selection in Laboratory Popu- la tions of Dr s,A1Jla II. Competition Between a “hite-eye Gone ano It5' 'ild Pype Allele," Evolution, Vol. h, pp. 3h-L2, 1950. ill Normal ana Selective hatirgs in Drosothila 51? Journal of Genetics ,cv l. hJ, pp. 2L7 -302, 19 \()V§Sfi a" " V11'fi c1, 7],}- Rentlell, J. H. ., "Genetics and Cvtoloov ofL D rcsool a Subobs cura.. II. " 1 44-1.) 0 Spencer, 'Narren P. ,"Genetic Drift in a Pepulation of Drosqphila Imr1"r9*° " Evoluticn, Vol.1, pp. lC3—llO, 19L7. Sturtevant, A. H., "Experiments on Sex;Recognition a.nd the Problem of Sexual Selection in Drosophila," Journal of Anina.l Eelav io r Vol. 5, pp. 351-350, 1915. Seuall, "Evolution in Mendelian Populations," Genetics, Vol. 16, pp. 97-139, 1931. Sewall, "Size of POpulation ano Breeding Structure in Relation to Evolution, " Sciance ,Vol. 87, pp. h30-L31, 193C. 7.'Nright, Sewall, and Th. Dobzhansky, "Genetics of Natural Populations. XII. Experimental Reproduction of Some of the Changes Caused by Natural Selection in Certain POpulations of Drosonhila Pseudoobscura," Genetics, Vol. 31, pp. 125-156, l9ho. ~' \ Pi. (LI. IPID)‘ HICHIGQN STQTE UNIV. LIBRQRIES 31293103756981