The Effect of Polygyny on the Sex Ratio of Mice (Mus Musculus)

In 1924–25 while studying the possible linkage of certain factors in the house mouse, and in order to test as many characters as possible at one time, a back cross was made to a five-point recessive. The number of males available was limited, necessitating the mating of each male to a large number of females. Again, during the latter part of 1928 and first of 1929, another series of experiments involving linkage required the use of a limited number of double recessive males, due to the difficulty of obtaining fertile ones. Each of these experiments gave an excellent opportunity to test the effect of polygyny and excessive sexual activity of the male mouse on the sex ratio of his offspring.

In the mouse, as in other mammals where the sex chromosome complex is of the x−y type, under normal conditions, the determination of sex depends on the type of sperm fertilising the egg. Thus if sexual activity is to have any influence on sex ratio it would show through the male rather than the female. Parkes (1925) comes to the conclusion “that in this mammal (mouse) a rise in the percentage of males follows polygynous mating.”

Parkes used seven matings of single males with from 8 to 12 females and obtained a total of 62 litters containing 234 males and 161 females. The percentage of males is thus 59·2 ± 1·67; compared to a total of 1701 mice obtained from monogynous matings containing 905 males and 796 females and a percentage of males of 53·2± -8·1. He states “The difference between the two percentages is 6·0 and the error on this difference will be ⁠. Hence the difference 6·0 ± 1·85 is more than three times its probable error, and is therefore statistically significant.”

In the first series of experiments, the author was testing the possible linkage of the waltzing factor with other factors in the house mouse (Gates, 1926). The males used were homozygous for the recessive factors of non-agouti, dilute brown, piebald, and waltzing, and heterozygous for pink eye. The females were heterozygous for all six of the above factors. In the second series, the males were albino and were mated to females heterozygous for albinism and shaker (Lord and Gates, in press). In both series the females were confined in several pens, but were restricted to three or four at the most in each pen. The males were changed from pen to pen every second or third day.

As soon as a female was observed to be pregnant, she was removed and placed in a pen by herself, where she was permitted to litter without being disturbed. In this way postnatal mortality was reduced to a minimum. Only after the characters of the mits had been recorded and checked was she again returned to the breeding pen. The number and sex of all young were recorded at birth. The sex was checked again later when the other characters were recorded, which in the first series was on the sixth and eighth days and in the second series on the fourteenth day or later. In the case of depleted litters where the young died between first recording and checking, the first determination of sex is assumed to be correct. The error due to this is probably negligible as the mortality between recording and checking was insignificantly small in the first series and relatively low in the second. No differential mortality as to sex has been observed in the author’s colony. Experience likewise shows that the error of determining sex at birth is insignificant, especially in such numbers as are involved.

The time during which the males were continuously active in the first series varied from 13 months in the case of males 1157, 2011 and 2018, to 6 weeks in the case of males 2014 and 2504. In the second series all males were mated on November 6th, 1928, and unmated on June 24th, 1929. In the first series, each male was mated to certain females only, being rotated from pen to pen, but only to such pens as were assigned to that particular male. In the second series, no definite pens were assigned to any one male, but each and all males were rotated through each and all the female pens regularly and in order. The females in the first series varied somewhat from time to time as older individuals were replaced by younger. In the second series the females remained constant except for death ; no new females were added.

The results of the first series are given in Table I. This table shows a total of 1437 young from the eight males, of which 724 were females and 713 males, a sex ratio of 98·48 males to 100 females ± 3·48. The record of some of the individual males is interesting. Male 1157 was mated to a total of 55 different females during the 13 months; not all of which however, as stated above, were in pens at any one time. As the older females showed signs of declining reproductive activity, they were replaced by younger individuals. Probably at no one time were there over 25 females in pens assigned to this particular male. The total number of litters sired by this male was 97 which is an average of one litter every 4·07 days over a consecutive period of 13 months. The total number of mits produced was 570, 274 females and 296 males, a sex ratio of 108·03 ± 5·61. The average size of each litter was 5·87 which was slightly above the average for the colony. The reproductive activity of this male, considering the number of litters sired, and the length of time of his activity, is at least unusual, if not rather remarkable.

The other two males of this series that served equally as long as 1157 also show a high reproductive activity. Considering the fewer number of females to which they were mated, their production is about equal to that of 1157. This is especially true of male 2011. Taking the average number of females that were mated to male 1157 at one time as an example, we find that each of them produced 3·88 litters apiece during the 13 months, or an average of one litter every 3·35 months.

The results of the second series are given in Table II. Here, as stated above, no definite number of females were mated to any one particular male, and no attempt was made to identify the sire of any litter. It is only possible therefore to give averages. On account of lack of fertility of the albino shaker male, only six tested males were used. These were very kindly supplied by Dr E. C. MacDowell of the Department of Genetics of the Carnegie Institution of Washington. The 146 females to which these six males were mated were distributed through 52 pens, and the males changed every two or three days.

This table shows a total of 268 litters sired by these six males in 224 days, an average of 1·19 litters per day or an average of one litter for each male every 5·01 days over the entire period of nearly seven and a half months. The average total number of litters sired by each male was 44·66. The average size of litter was 7·43 which is unusually high, but is accounted for by the fact that only well-matured females were used and the experiment terminated, before they had reached an age of declining fertility, and small litters. The fact that each female had to raise her entire litter up to an age of 14 to 16 days, or even 18 days in the case of doubtful individuals in the litter, meant that her own reproductive activity was retarded. Likewise the fact that each female only produced 1·83 litters during the 224 days of the experiment showed that her own reproductive capacity probably was not taxed to any great extent. The larger size litters, however, and the longer period of nursing undoubtedly tended to lengthen the time between parities.

In this series there were produced a total of 1993 mits, of which 1079 were females and 914 males, giving a sex ratio of 84·71 males to 100 females ± 2·48.

Combining the two sets of data we have a total of 3430 young produced by polygynous matings of 14 males. Of this number, 1803 were females and 1627 males; a sex ratio of 90·23 males to too females ± 2·18.

The average sex ratio of the colony exclusive of the experimental animals was obtained by using all the litters born during the entire period covered by the two series of experiments. The total number of such young born was 12,881 in 2508 litters, or an average of 5·13 mits per litter. Of this number 6672 were females and 6209 males or a sex ratio of 93·06 males to 100 females ± 1·56. These figures do not of course represent monogynous matings only, as it is the general practice in this colony to mate from one to three females to each male, depending on the number of young saved out of each litter, as brother-sister matings are the rule. It is customary, however, not to mate over three females to any one male unless absolutely necessary, on account of the small pens in which the mice are kept. Rotation of males among two or three pens of females was practised in a few instances. However, as it is customary to remove pregnant females from the breeding pen, and as most of such females reared their young, the conditions approaching monogynous matings are obtained.

Comparing the sex ratio of the two series with that of the average of the colony, we find that the first is higher and the latter lower. The ratio of the combined total of the two series is quite close to that of the control or the entire colony exclusive of experimental animals.

In two series of experiments it was necessary to mate a limited number of males to a large number of females. The females were kept in separate pens and the males were rotated from pen to pen every second or third day. The results show that these polygynous matings resulted in unusual sexual activity on the part of the males, one male in the first series producing 570 mits and the average of the six males in the second series 662·16. It would seem that such sexual activity over extended and continuous periods of time ought to be a sufficient test of the effect of polygyny and over stimulation of the sex organs on the sex ratio, if such activity is to have any effect at all. Allowing for some non-productive copulations, the rate at which these males mated probably must have been close to the maximum physiological functioning of the reproductive and associated glands.

In the first series of experiments there is a great variation in the sex ratio, some males producing a low, some a high ratio. The three males mated to the largest number of females produce a ratio varying from normal to considerably above. The less active males, with one exception, produced a ratio below normal. In the case of the exception, the numbers are not sufficiently large to warrant any conclusion. The sex ratio in the first series is not available month by month.

A study of the second series shows a marked lowering of the sex ratio among the offspring of sexually active males. The average 84·71 is considerably below the normal of the colony, namely 93·06. By itself this might demand explanation, but when compared with the results of the first series, it probably is due to individual variation. The sex ratio rises and falls from month to month in this series with no correlation as to season or breeding activity. The months of greatest activity may show a high or low ratio and vice versa. The food and environmental conditions are quite uniform from month to month and year to year, and could have little or no influence.

The difference between the average of the two series, 90·23 ± 2·18, and that of the colony, 93·06 ± 1·56, is but 2·83 ± 2·68. A difference only 1·06 times the probable error and therefore not significant. The size of the litters which average above the mean for the colony shows that the reproductive capacity, that is, the ability to produce virile, functioning sperm, of the males was not impaired on account of their extreme activity.

Parkes (1924) states that the sex ratio in 1031 monogynously mated mice was 118 males to 100 females. Other observers find that the number of males born is usually in excess of females. It is therefore interesting to note that in the author’s colony over a number of years the number of females born is in excess of the males. Whether or not the geographical environment affects the ratio has not been investigated. However, whatever environmental conditions might be operative to lower the average sex ratio, these must be identical in the case of both experimental and control animals.

The evidence, therefore, gathered from large enough numbers to be significant, shows that polygyny does not affect the sex ratio, but that variation from the normal probably is due to individual, natural or inherited qualities, possibly also to age and not to over stimulation and activity of the sex organs.