903 
May 15,1925 Possibility of Sex Control by Artificial Insemination 
already mentioned, observation has 
apparently shown that there are slight 
but consistent deviations which vary 
for different species {S3). Among 
most mammals it seems that there is a 
slight excess of males, the normal sex 
ratio being usually somewhere be¬ 
tween 103 and 110 males per 100 fe¬ 
males. This may very easily be due 
to some difference in the motility or 
viability of the two classes of sperma¬ 
tozoa. The presumption is that in the 
rabbit also there will normally be a 
slight excess of males, and the data 
from this flock for the litters produced 
by a natural service without any treat¬ 
ment afterward support that assump¬ 
tion, for there were 421 males and 394 
females, a ratio of 106.9 c? S : 100 9 9 • 
However, no report of large num¬ 
bers has been found for the rabbit. 
The numbers available are too small 
for calculations to be based upon ratios 
drawn from them for a standard, and 
therefore the calculations in these ta¬ 
bles are based upon exact equality as a 
rigid standard to be given up only 
when enough data are available to es¬ 
tablish definitely how far the normal 
sex ratio in the rabbit does deviate 
from equality. The probability that 
the normal sex ratio is somewhere be¬ 
tween 105 and 108 males per 100 fe¬ 
males should be kept in mind, how¬ 
ever, in considering the results given in 
these tables. 
The common method of stating sex 
ratios in terms of the number of males 
per 100 females is open to criticism in 
that a change in the number of males 
produced is not given equal numerical 
value with an equal change in the num¬ 
ber of females produced. Thus in a 
population where the sexes are equal, if 
a change were made which reduced by 
one-fourth the number of males born 
the sex ratio would be given at 75 c? cf: 
100 9 9 , whereas in the same popula¬ 
tion if the change had reduced the num¬ 
ber of females produced by one-fourth 
the sex ratio would have been stated as 
1333^ cf cf : 100 9 9 - In the one case 
the ratio would be described as 25 
points below the normal and in the 
other as 33h£ points above the normal, 
although the magnitude of the changes 
had been the same fundamentally, but 
took place in different sexes in the two 
cases. This seems to be a minor point, 
but it tends to obscure what may be a 
very vital relation. For example, in 
her work on selection for abnormal sex 
ratios in the rat, King {23) states that 
in the high male line after the seventh 
generation the sex ratio was 122.3 cf cf : 
100 9 9 , while in the low male line for 
the same period it was 81.8 cf cf: 
100 9 9 . Now it is not readily appa¬ 
rent without calculation that a ratio of 
81.8 cf cf : 100 9 9 is the same as a 
ratio of 100 cf cf : 122.25 9 9 , or almost 
exactly the same as the ratio in the high 
male line except that the sexes are re¬ 
versed. If these ratios had been ex¬ 
pressed in percentages, that fact would 
have been apparent at once and, taken 
in connection with the fact that selec¬ 
tion appeared on the whole to have 
very little effect after the twelfth gen¬ 
eration, would have brought home a 
conclusion not emphasized in that 
paper—namely, that in the albino rat 
selection can alter the sex ratio in 
either direction with equal ease but 
can only carry it the same distance on 
either side of equality. That fits in 
very well with the hypothesis that the 
results which she secured were due to 
sex-linked lethal factors or factors 
whose lethal action was confined to 
one sex, and suggests that the ratio 
which she adopts as a normal, 105 cf cf : 
100 9 9 , is really the result of some 
such factor as a difference in motility 
of the two types of spermatozoa. There 
seems to be no good reason for express¬ 
ing sex ratios in terms of the number 
of males per 100 females except that it 
is customary and magnifies the differ¬ 
ences between ratios. 
The attempt to develop a practical 
method of sex control by means used in 
these experiments was not successful. 
In Table I, where the different methods 
of breeding are compared, the only sex 
ratio which deviates far enough from 
equality and concerns numbers enough 
to approach statistical significance is 
that of the young produced by a natu¬ 
ral service where all the fluid possible 
was withdrawn with a catheter. If the 
normal sex ratio for the rabbit be taken 
as 106.9 S c? : 100 9 9 , as it was found to 
be with the litters produced in the nor¬ 
mal way in this flock, then the deviation 
of this class from expectation becomes 
more than 3.64 times the probable 
error, and, looking at it from a purely 
statistical standpoint, it may be said 
that the deviation is probably signifi¬ 
cant. However, when it is asked what 
factor has changed the sex ratio and 
how it has operated to produce this 
result, the answer is not apparent. The 
only readily apparent change which has 
been made in this type of breeding is a 
reduction in the number of spermatozoa 
which have been left in the uterus and 
vagina, but just how this could influ¬ 
ence the sex ratio is not apparent. A 
reduction in the number of spermatozoa 
present might cause an already abnor¬ 
mal sex ratio to approach nearer equal¬ 
ity in a way analogous to the differential 
