72 Zoologica: N. Y. Zoological Society [HI; 3 



observation, just after the beetle had tossed the larva to a dis- 

 tance of about four millimeters from the coccid by an unusually 

 well-aimed blow. 



Scenes of this description were so frequently enacted that 

 they could be readily observed in almost any of the colonies after 

 they had been kept for several days and the beetles, larvae and 

 coccids had all grown very thirsty. In such colonies the bodies 

 of the coccids and larvse become visibly attenuated and some- 

 what shrivelled as a result of the loss of water from the tis- 

 sues of the Tachigalia petioles. All the insects now become rest- 

 less. The beetles leave the petioles, wander about on the walls 

 of the tubes and finally collect about the plugs of cotton in an 

 endeavor to escape to the outside. The coccids, too, withdraw 

 their beaks from the parenchyma in the floors of the food- 

 grooves and wander aimlessly about, vainly seeking more favor- 

 able pastures. But before this stage of demoralization is reached, 

 both the beetles and the larvse become cannibalistic and one 

 may often see them, singly or in groups voraciously devouring 

 partly dismembered larvse or immature beetles. Within a few 

 days all the larvse and immature beetles are consumed, but tht 

 coccids, immune from attack, still wander about till they die 

 of starvation. 



I believe that Coccidotrophus is rarely or never cannibal- 

 istic under normal conditions. It is, as already stated, almost 

 impossible to split a freshly gathered Tachigalia petiole con- 

 taining one of the beetle colonies, without cutting some of the 

 insects iu two, and such disabled individuals are soon devoured 

 by their fellows, but both in such cases and in the cannibalism 

 that supervenes in dried petioles, I believe that thirst or the need 

 of water and not a veritable carnivorous instinct, such as seems 

 to be manifested by some species of Cucujid beetles and their 

 larvse, must be regarded as the true explanation. I am con- 

 firmed in this view by Heins' recent investigations (1920) on 

 meal-worms (Tenebrio molitor). He found that when the larvse 

 of this beetle are reared in dry meal as many as 24.2% of 

 them may be devoured by their fellows, but that if wet slices 

 of rusk or of vegetables are placed in the breeding jars the 

 mortality from cannibalism is reduced to 7.5%. In this connec- 

 tion Bodine's observations (1921) on grasshoppers are also of 



