THYSANOPTERA AND ORTHOPTERA. 185 



ing the factors determining the habitat of thrips. therefore, two answers 

 suggest themselves: (1) The character of the food, and (2) the amount 

 of protection afforded. If either of these factors predominates over 

 the other in the case of an individual species, the effect of this pre- 

 dominance should be evident in the local distribution. (1) If char- 

 acter of food is exclusively, or even chiefly. responsil)le for the habitat, 

 almost any part of a plant, provided it is tender enough to be pierced 

 or rasped by the mouth parts of thrips, should be infested. At least 

 we should expect the whole plant to be infested unless the enemies of 

 and checks to Thysanoptera were powerful enough to remove them 

 from the exposed portions of the plant after these were once infested. 

 Furthermore, nearly related plants should be infested by the same 

 species of thrips, whether or not they afford equal protection. Xo test 

 of the influence of the character of food on the choice of food plants can 

 be made by examining plants which afford abundant protection but 

 have unpalatable juices, hke the dandelion (Taraxacum officinale), or 

 plants that have palatable juices but afford no protection. For what 

 may seem to us highly unpalatable may be quite agreeable to tlu-ips. 

 (2) If, on the other hand, the protection afforded by a plant is the chief 

 factor in determining whether it is likely to be inhabited by thrips, the 

 exposed portions should not be occupied in equal degree with the con- 

 cealed portions; and nearly related species of plants, which afford un- 

 equal degrees of protection, should be unequally infested. 



The large number of plants on which Thrips tabaci and Euthrips 

 tritici have been taken, in nearly every case from the flowers, close tufts 

 of leaves, or some other equally concealed situation, suggests at once 

 that for these two species, at least, the protection is the determining 

 factor. It is interesting to note that on white clover {Trifolium repens) 

 Euthrips tritici was often abundant, and on red clover (Trifolium pra- 

 tense) not uncommon; while on the closely related yellow, and white, 

 sweet clovers (Melilotus officinalis and M. alba respectiveU'), on which 

 the flowers are much more widely separated than in the species of 

 Trifolium, and hence offer little protection, the same species of thrips 

 was rare. The specimens of Melilotus examined were in localities in 

 which these thrips were fairly abundant on other plants, so that op- 

 portunity was present for them to infest the sweet clovers also. . From 

 these observations, and the large number of unrelated or distantly re- 

 lated plants on which Thrips tabaci and Euthrips tritici may be found, 

 I conclude that the local distribution of these two species is determined 

 largely, and perhaps almost exclusively, by the protection afforded by 

 the inhabited plants. 



With two of the poephilous species. Anaphothrips striatus and Aptino- 

 thrips rufus, the 'case seems to be different. The former species was 



