INBREEDING AND SELECTION IN DROSOPHILA AMPELOPHILA 147 
b. Strain 207 (low female ratio). From pair 207 with a ratio 
of 1:0.89 (84 «: 75 2) it was hoped to develop by selection a 
strain showing a low female ratio. Seven matings from the first 
generation produced 536 « and 579 9¢, or a ratio of 1:1.08. 
The range of ratios of the individual pairs was from 1:1.22 (99 ¢: 
121 ¢) to 1:0.86 (79 «: 68 ¢). This selection was continued 
for four generations, the matings being made from broods with a 
low female ratio. The ratios of all the offspring in thesuccessive 
generations were 1:1.06 (220 7: 228 ¢) 1:1.10 (581 2 : 640 
Sheed: 104 (142 & : 14749)s 117, (18h@ + 607 2) for; the 
second, third, fourth and fifth generations respectively (See Table 
9). This low female ratio showed itself rather uniformly in all the 
individual matings, a notable exception occurring in the fifth gener- 
tion (see Table 9, pair 3.) with a ratio of 1:2.53 (45 #: 144 9). 
On the other hand no pairs threw a great preponderance of males, 
the most notable among those from which a large number of 
progeny was obtained being pair 2 in the third generation in 
which the ratio was 1:0.87 (115 #: 101 ¢). For the sixth 
generation two sets of matings were made as in the fifth genera- 
tion of the strain 206. One of these was made from a brood with a 
ratio of 1:2.53 (45 o« : 144 ¢) and the other from a brood with a 
relatively low female ratio, 1:1.36 (72 @: 98 9). From the 
former the total progeny of eight matings gave a ratio of 1:1.42 
(461 =: 654 9?) and from the latter the ratio of eleven matings 
was 1.1.05 (944 7:997 9). 
c. Discussion. It seems from the above experiment that the sex- 
ratio in this creature is a strongly transmissible character. Start- 
ing with a pair that throws an offspring showing either high or a 
low female ratio it was possible to maintain, by selection, a strain 
maintaining the respective ratios. The offspring from a given 
pair, when mated in pairs, show a considerable variation in the 
sex-ratio of their children. It is thus possible to develop a strain 
with a low female ratio from one with a high female ratio, or the 
reverse, as is shown in the fifth and sixth generation of experi- 
ment 206 and 207 respectively (tables 8 and 9). The sex-ratio is 
clearly amenable to selection like any other character. 
