BREEDING HABITS OF HETERONEREIS. 155 



through lunar tidal variations, rhythmical alterations of condi- 

 tions of nutrition. (Cf. Hempelmann, 1911.) This assumption 

 at least explains fairly well the conditions as we find them. 



2. The Runs. The collector receives very definite impression 

 concerning the size and composition of swarms and the general 

 features of the runs. But to give these quantitative expression 

 is not a simple matter, though it is essential in order to reveal 

 the order that exists. Moreover, any quantitative system of 

 representing the data must be based on some one feature to the 

 exclusion of others, and it is of course limited to this extent. 

 As a result of experience we finally decided in the first place to 

 limit the collections used in the tables to one station, and in the 

 second place to record only females caught. The latter provision 

 is due to the relatively enormous number of the males on certain 

 nights which makes accurate estimation out of the question. 

 The collecting was done each evening from the same station in 

 the same way as nearly as possible. Practically complete records 

 were kept for 1911 and 1912, and these have been tabulated in 

 the form of curves in which the ordinates represent the number 

 of females caught and the abscissae represent days from the 

 first quarter of the moon. 



Of these records of runs we may select that of August, 1912, 

 as being in some respects the most typical (see curve i). It 

 begins on the fifth day after the first quarter of the moon (July 

 25), and is quite atypical in this one respect, for no other run 

 began earlier than the eighth day. On the fifth day two feamles 

 were caught, the sixth day one, seventh day ten, eighth day 

 eight, then the catch gradually ascends to a maximum on the 

 twelfth day. The numbers in the successive swarms then rapidly 

 diminish to the I5th day, when only three females were caught, 

 but on the i6th day there were sixteen females, and on the ijth, 

 1 8th, I9th, 20th, and 2 1st, over thirty. The numbers then 

 gradually decreased to the 26th day, which was the last day of 

 the run. This run had been preceded by eight days in which 

 no Nereis occurred, and it was followed by a barren period of 

 eleven days It was therefore sharply separated from preceding 

 .and succeeding runs. 



The other runs may be compared with this one as follows: 



