294 WILLIAM LAWRENCE TOWER. 



The extracted forms of the signaticollis type, when inbred, con- 

 tinued to breed true in F3; the extracted forms Hke the male 

 diver sa type hkewise came true in Fg, F^ and F,; and the mid 

 types in generations F3 and F4 gave in all cases close approxi- 

 mations to the ratio 1:2:1, The behavior observed in* experi- 

 ments of this class is shown in Plane I., a digest of Exp. No. H 409. 

 This behavior would suggest that either one or the other of 

 the parents used was a "heterozygous form." When, however, 

 identical material, brothers and sisters of the first, were reared 

 under other conditions, a different result was obtained. 



L. signaticollis 9 X cP L. diver sa. 

 Exp. No. H 410. 



In Plate II. are represented the results obtained in crossing 

 a female L. signaticollis and a male L. diversa of exactly the same 

 stocks as described in Exp. No. H 409, but under the following 

 conditions : 



Food : Normal — uniform. 



T. R. H. 



Day Av. 75° F. ± 5°. Day Av. so per cent. ± 10 per cent. 



Night Av. 50° F. ± 5°. Night A v. 80 per cent. ±15 per cent. 



In the Fi generation of this experiment there was one class of 

 adults, invariably of the mid type, and these, when inbred, gave 

 in the Fg generation a typical Mendelian grouping of types, in 

 the majority of cases in as nearly perfect ratio of 1:2:1 as could 

 be desired. These two crosses were repeated eleven times under 

 the same conditions with the same results, the only variation 

 being a slight oscillation in the 1:2:1 ratio of the F^ generation 

 and a slight variation in the mid types. The number of pairs 

 that have been used would in itself be sufficient evidence that 

 the results were not due to the chance selection of individuals 

 which would give these results, but that they were due entirely 

 to the conditions surrounding the germinal materials at the time 

 of their combination and during their early development. To 

 still further test the point, a more crucial experiment was carried 

 out as follows: 



