THE GROWTH OF GROUPS 121 



become common knowledge, for the evidence is clear 

 and direct, more so even than the theory of descent in 

 general. The beetles illustrated herein are each of 

 different kinds in the usual sense. As a rule each kind 

 breeds true to its own kind, but occasionally some kinds 

 give birth to others. Similar events seem to occur among 

 the rats, and doubtless among the animal kingdom at 

 large. 



Let us now continue our summary of Tower's researches. 

 The sports were found singly among swarms of the normal 

 or common decemlineata, sometimes they were found in 

 twos or threes, and on one occasion six of the kind 

 pallida were found in the same neighbourhood. These 

 six were kept for experiment : they hibernated as usual. 

 One female and two males survived the winter and 

 became sexually mature in the following spring. From 

 a pair of them, descendants were raised for seven genera- 

 tions, in nearly every generation some hundreds of beetles 

 were raised, large numbers had to be destroyed in order 

 to keep the experiment within bounds. All, without 

 exception, were of the pallida type. No mention is made 

 of any ill effects arising from inbreeding. This experiment 

 shows conclusively that if a pair of pallida were to be 

 isolated physiologically or geographically in a suitable 

 place, they would give rise to a species. 



A second experiment shows, that pallida may persist 

 when mingled with decemlineata, in spite of the fact that 

 the sport is not sterile with decemlineata and sometimes 

 breeds with it in nature. Pallida does not then become 

 swamped by crossing. In this experiment twenty-two 

 pallida and thirty decemlineata were confined together and 

 allowed to breed freely, in each group about half were 



