Eeactions and Peoducts in Intbkspecifio Ceossbs 



97 



the extracted, types gave, no matter what the point of origin nor what the 

 history, the same uniform results obtained in the crossing of the original stocks. 



That there might be no personal equation involved in the series presented 

 after the initial combinations were made, the censuses of the fraternities were 

 made by a technical assistant, who was not instructed what to look for, but to 

 separate the fraternities into the different groups that he could recognize, and 

 to isolate any that were questionable as to where they belonged. After this was 

 done I took the materials and checked the counts and determinations of the 

 array; any that we did not agree upon where used in breeding to test their 

 character. Likewise, matings of the different portions of the fraternities were 

 random, the separated classes of a fraternity being put into glass jars and the 

 pairs drawn at random for the matings. The results presented are as free from 

 the errors of personal equation as is possible. 



From the inspection of the Fj and F^ fraternities that came from the crossing 

 of these two species, as well as the extracted types in Fj and ia later generations, 

 the impression was forced upon the observer that the total type had remained 

 as a unit, segregating out in the gametogenesis of the heterozygous forms. There 



Ky a 



Fig. 2. 



was no evidence of the interchange of any characters between the species dis- 

 coverable by inspection. To test this point, several of the fraternities were 

 investigated, as also the parent species of the stocks used in experiment and 

 from nature. 



For this purpose I studied the form-index and its behavior in the materials, 

 in that it is a good indicator of the persistence and integrity of the gametic 

 type With each type are associated many minor characteristics, and these, 

 together with the form differences, make for a distinct difference between the 

 two species in general aspect as well as in special characteristics. 



This form-index is determined in these materials from measurements of cer- 

 tain parts, not liable to change in drying in preserved specimens or by changes 

 incidental to reproduction, or other ontogenetic processes. In figure 3 I have 

 shown the measurements that were made of the adults, and the index as used 

 was determined by the summation of the values 1, 2, and 3, divided by the 

 value 4 These readings were made with special vernier calipers reading to 

 tenths of a millimeter. Experience has shown that this index is the most 

 reliable and least variable, as well as the least liable to distortion by ontogenetic 

 or post-mortem changes in the specimens examined. 



