300 ANIMAL AGGREGATIONS 



observations led to experiments designed to determine the dose of 

 hairworm eggs that would be fatal to the host; for the present pur- 

 pose it may be said that for young grasshoppers in the second instar 

 it averages well under 50 eggs. These later experiments led to the 

 discovery that when a slightly sublethal dose of eggs is given, the 

 resulting parasites are all males. In one case where 20 hairworm 

 eggs were fed to a grasshopper first freed from this parasite, 19 para- 

 sites were recovered, all males. Many tests and observations were 

 made, and all led to the same conclusion : Feeding a very few eggs 

 resulted in female parasites; feeding a large number, but not quite 

 enough to kill, resulted in male parasites. 



Similar experiments in treating the larvae of a midge {Chirono- 

 mus) with the hairworm Pseudo?nermis zykoffi gave similar results. 

 High parasitism resulted in the parasites being all males. Similar 

 observations were made on the hairworm Agamermis paradecaudata 

 infesting the tea bug Helopeltis theivora. No artificial infestation was 

 made in this case; the observations were confined to the parasite as 

 it existed in nature. Here, again, when there were a large number 

 of parasites all of them were males. 



Observations on the mirmithid Allomermis sp. infesting the 

 common dooryard ant Lasius niger americanus showed that when 

 parasitism is high the parasites are all or nearly all males, and that 

 when it is low the parasites are all females. 



The phenomenon is the same in hairworms of four different gene- 

 ra, found in host insects of four different orders. Between the ex- 

 tremes of parasitism in all these cases there is a gradient with mix- 

 tures of males and females, the proportion of the males increasing 

 with the severity of the infestation. 



In the case of the hairworm Mermis subnigrescens, Cobb and his 

 associates think (Christie, 1929) that by appropriate experiments 

 they have excluded other possible explanations of these results; 

 that they were not brought about by using the eggs from an individ- 

 ual hairworm, nor from a definite part of the uterus having a uniform 

 sex-potentiality, nor by death or other selective elimination. The 

 most convincing data at hand on these points are those collected by 

 Christie (1929). Table XXXIV shows the results from one set of 



