54 PHIL RAU AND NELLIE RAU 



sheltered from the weather and enemies, and their food is at 

 hand for them when they arrive, and with the two fierce struggles 

 removed (hunting for food and being hunted), one would expect 

 little or no loss of life, and it certainly would appear that, aside 

 from inherent tendencies, all had exactly equal chances of 

 survival. 



Hence we must admit that Natural Selection is vigorously at 

 work here, eliminating weaklings while the vigorous survive, 

 even in this secluded life where the individuals do not in any 

 way come into the open struggle for existence. It is precisely 

 this elimination in the potentially viable young in single families 

 that we shall now take up. 



It is difficult when so many elements are involved to get any 

 tangible proof that inherent characteristics of families have 

 anything to do with the mortality or viability of these young. 

 We may approach it however by calculating what the death- 

 rate would be in each of the five stages or combinations of 

 stages 21 if the elimination followed only chance. The entire 

 formula for all of the possible combinations of the five stages 

 is so long that we shall only take from it what we want to com- 

 pare with the actual occurrences in the preceding tables. It is 

 obviously necessary that all of the nests must have at least 

 five good cells to justly compare them with the distribution 

 of five chances, so we use in the following comparison only the 

 nests of that size or greater. Of these, we had 120. 



Now if no factor but mere chance controlled the distribution 

 of the fatalities, then there would be only one chance in 3,125 

 of all five of the young in a nest meeting death in any one stage, 

 while in fact we find this actually occurring at a much higher 

 rate, e. g., all died in the prepupal stage in 22 of the 120 large 

 nests, or at the rate of 572± in 3,125. Other comparisons are 

 as follows: 



21 We must make due allowance for the fact that in this calculation we have 

 assumed that the chances of death are equal in the five stages, while in fact some 

 stages are obviously of longer duration or more trying than others. 



