W. E. Agar 183 



was however recorded, and these form the material available for the 

 table. One example of how the table is to be read will explain it at 

 once. Take generation 10' for example. In this generation ten first 

 broods were tested for sex — i.e. the first broods from ten individuals of 

 the preceding generation. By "tested for sex" I mean that one or more 

 of the young were brought to maturity and the sex recorded. In this 

 case altogether 26 individuals were taken fnjm the ten broods. In the 

 same generation no individuals belonging to second broods were kept 

 to maturity, but one third brood was tested by the isolati(3n of two 

 members of it. As the members of a brood in Swiocephalus (as in 

 Daphnia) generally all, or mainly all, belong to the same sex, it is 

 important to state how many broods were tested, and not merely the 

 number of individuals without reference to the number of families from 

 which they were taken. 



The re.sult of the experiment was as follows : — No sexual (ephippial) 

 females appeared during the whole of the experiment. Males appeared 

 in three generations only, where the asterisks are placed in the table. 

 In generation 6 there was one male from one of the first broods, in 

 generation 7 two of the first broods furnished between them five males, 

 and in generation 9 there were three males from one of the first broods. 

 All the other broods and all the other generations afforded partheiio- 

 genetic females only. 



Thus so far from a progressive increase in sexuality, the very few 

 sexual fonns that did appear were in the early generations of the 

 experiment. 



These 46 generations were passed through in twenty-one months, 

 at room temperature. The average duration of each generation 

 would however have been far greater at the mean annual temperature 

 in their natural ponds. Papanicolau estimates the maximum number 

 of parthenogenetic generations in a natural " cycle " at twelve, and this 

 agi'ees roughly with my estimate of the greatest number that could be 

 produced in a year under the climatic conditions of Glasgow. Even 

 during the time in which the line was bred in my experiment, partheno- 

 genesis lasted therefore for nearly four times as many generations as 

 may be expected in nature (though of course the natural " cycle " is by 

 no means a rigid one), and at the end of the experiment no trace of loss 

 of vigour was detectable. 



According to Papanicolau, not only should sexual forms appear with 



1 The original parthenogenetic female from which the Une was started being counted 

 as generation 1. 



