FAMILY OYNIPIDJ1. 149 



apical ventral segment is much larger than the others, 

 and in the gall-making species is frequently plough- 

 share-shaped, the apex being usually produced into a 

 sharp point, which is frequently hairy. 



As regards the life-history of the family, we find 

 three well-marked modes of existence, gall-making 

 species, inquiline species, and parasitic species. In 

 the first group the females deposit their eggs in the 

 living tissue of plants in such a way that they are 

 brought into contact with the cambium layer. This 

 contact causes a more or less complicated structure 

 or gall to form round the egg. On the juicy substance 

 found in the gall immediately surrounding it the larva 

 lives, feeding only during the period that the gall is 

 succulent, necessarily a short time. The part with 

 which it is surrounded becomes hardened, forming a 

 round cell, in which it passes the quiescent larval con- 

 dition as well as the pupa state, and finally emerges 

 from it as a perfect insect. They are thus purely 

 vegetable feeders, in which respect the inquilines 

 agree with them, only they do not form or give 

 origin to galls themselves. They deposit their eggs 

 in galls raised by the true gall-flies for their own pro- 

 geny when the galls are still soft and juicy, and before 

 the gall-fly larva has reached any size. Being of a 

 more rapid and vigorous habit than it, the inquiline 

 larva eats up the available food, and crushes the legi- 

 timate tenant out of existence. This does not always 

 happen in large galls where there is a sufficient supply 

 of nourishment for both, but it may be said to be in- 

 variably the case in the smaller galls. Sometimes 

 there may be only one inquiline larva in a gall if the 

 latter be small, but in the larger galls there may be 

 as many as five or six. The presence of the inquiline 

 larvae in a gall causes it to become more or less dis- 

 torted, and frequently enlarged. They also cause 

 monothalamous galls to become many-celled. Thus, if 

 the galls of Rhodites eglanteriae were left in the posses- 

 sion of the Rhodites larva, it would be thin-walled and 



