126 THE GROWTH OF GROUPS 



Three of the former and one of the latter appeared as 

 offspring of multitaniata. Both kinds were isolated and 

 were found to breed true, each to its own kind, for several 

 generations, without any tendency to revert to the type 

 of the parent species. A pure line of melanothorax was 

 raised from a single male specimen, by crossing with the 

 normal multitaniata, the second generation contained a 

 certain number of the kind melanothorax of both sexes, 

 in accordance with the Mendelian expectation. 



The line of rubicunda, which was raised directly, both 

 male and female sports of this kind being obtained, was 

 found to be " almost completely sterile " with the parent 

 species, but quite fertile with specimens of rubicunda 

 found at Toluca in natural circumstances. 



In another series of experiments, a number of multi- 

 tczniata were subjected to unusual heat and moisture 

 during the formation of some of the eggs, and to normal 

 conditions during the formation of the remainder. As 

 in the similar experiments with decemlineata, the first 

 batch of eggs gave rise to many sports, a few of the kind 

 rubicunda and many of the kind melanothorax. The 

 second batch of eggs gave rise to offspring of the parental 

 kind. It is then a property of multit&niata to give rise 

 to rubicunda and melanothorax, just as it is a property of 

 decemlineata to produce the forms pallida, melanicum, etc. 

 We may expect such events to occur, just as we may 

 expect any of the well-known reactions between chemical 

 substances. There is, of course, no similarity between 

 animal mutation and chemical reaction. But it is evident 

 that mutation like chemical reaction is a phenomenon 

 which has been observed. 



At present these two classes of events are not looked 



