Reactions and Peoducts in Inteespecific Ceosses 139 



principles of the Mendelian reaction and the general evolutionary conceptions 

 of the factorial theory. These in most respects are duplicates in a broad way of 

 the differences in reaction found in the cross of L. dwersa and L. signaticollis, 

 the differences in both series being the product of the conditions of the medium 

 in which the cross was made. 



I first found, in 1907, that when L. signaticollis with the Ac value at 38 to 40 

 was crossed with undecimlineata with the usual Ac value of about 63 under the 

 following complex, that the results in P^ were further upset and productive of 

 much confusion. 



Table 16b. 



This arrangement of the complex and its rapid extreme changes in the experi- 

 ment, when brought to bear upon the crossing at the time of mating, maturing 

 of the gametes, and during fertilization and early development produced a most 

 distressing array in P^, especially so when the behavior of other crosses of the 

 same materials is known not to give anything but the most orderly and typical 

 reactions. The actual result of this cross under the conditions is shown in 

 plate 15 which shows that in P^ there appear the three types of adults — ^visible 

 undecimlineata^ mid-type, and signaticollis types — in about a monohybrid ratio 

 of 1:2:1. 



The Pi larvae split in the second stage into two groups of about equal numbers 

 of yellow and white, and in the third stage are still, in equal numbers, yellow and 

 white, both without the meristic-spot series, so that the L. undecimlineata was 

 the dominant one in this respect. The white larvae always gave types that were 

 undecimlineata, that in all respects are duplicates of the Fj appearing undecim- 

 lineata types of other series, slow breeding, with difficulty dissociated, in fact 

 masked heterozygotes. The yellow larvse gave two types of adults, on the aver- 

 age of two mid-types to one signaticollis type. The mid-type gave only the 

 normal reaction, and were in no observed respect different from the other Pi 

 heterozygotes that have been seen and tested. 



The signaticollis type that appeared in Pi, however, was pure-breeding, at 

 least to the extent that they came true to type for an undetermined number of 

 generations. They were a fixed type arising in Fi that bred true to type and 

 had all of the characters of the signaticollis normal type. They bred at once, 

 had about the same length and type of rhythm in reproduction as the mid-types, 

 but were true to the Pi condition in subsequent generations. These were 

 another type of heterozygote, as could be shown quite readily by applying the 

 test reaction used in the cross of diversa and signaticollis, which showed the 

 recovery in F^ following the test of the undecimlineata type. Biometric tests 

 of the pure breed Pj also showed in two fraternities that I had tested the typical 



