1921] Wheeler: Some Social Beetles 107 



and the fact that they devote more attention to the posterior than 

 to the anterior end of the larger coccids, implies a delicate 

 discrimination, because both ends of the coccid's body are so very 

 much alike. Furthermore, the beetle's antennae and mouthparts 

 seem to be so clearly adapted to dealing with the coccids as to 

 indicate that the trophobiotic relations between the two species 

 have been in existence for a very long time. The beetle's extraor- 

 dinary perseverance in stroking individual coccids after they 

 have been exhausted by repeated emissions of honey-dew might 

 be interpreted either as a very thorough and hence highly 

 adaptive method of exploiting the coccids, or as due to a very 

 imperfect discernment of their physiological peculiarities. 



Obviously the most remarkable item of behavior in 

 Coccidotrophus and Eunausibius is the stroking of the coccids 

 by the larv^ of all stages, as well as by the adult beetles. No 

 one has even considered the possibility of a similar performance 

 by the larvje of ants, wasps or butterflies, since it is difficult 

 to imagine creatures more unfitted for such behavior. Neverthe- 

 less, Mr. W. F. Fiske informs me that while he was investigating 

 certain injurious insects in British East Africa, he saw small 

 worker ants climbing a tree with their larvae and holding them 

 to the posterior ends of aphids, so that they could feed on the 

 honey-dew voided in response to the antennal solicitations of their 

 nurses ! Unfortunately no specimens of the ants were preserved, 

 and from Mr. Fiske's description I am unable to determine even 

 the subfamily to which they belonged, but I have no reason to 

 doubt the statement of an entomologist so competent and so 

 keenly observant. I surmise that the ant must have been some 

 Myrmicine which is unable to feed its larvae by regurgitation, as 

 otherwise such behavior would be superfluous. 



So specialized a habit as the coccidophily of the two genera 

 of social Silvanids calls for some consideration of its possible 

 phylogenetic origin. Under existing conditions, the beetles either 

 find the coccids already established in petioles that have been 

 previously inhabited by other beetle colonies or by other insects, 

 or the coccids enter the young petioles just after they have been 

 perforated by the beetles, for insects with sucking mouthparts 

 cannot, of course, gain access to the cavities in any other way. 

 That they migrate into the petioles as very young individuals is 



