Morphological J^ariatio)i and Its Causes in A. tigrinum 57 



attacking each other, except the strong stimukis of food is in- 

 troduced (earthworms, meat), when they become violently ag- 

 gressive, biting each other or anything that conies to hand. Only 

 smaller larvae, or larvae from quite different locations, or, still 

 more, larvae dift'ering much in size, are incompatible as com- 

 panions. But not so these broad-mouths. Among other dispo- 

 sitions, I placed in each of four small aquaria, four larvae of 

 equal length. Two in each acjuarium were normal, two were 

 broad-mouths. I thought, by unlimited feeding upon the most 

 luscious of liver, to reform the cannibals, and to be able to watch 

 the development and metamorphosis of the two types. My hopes 

 were soon dissipated, however, for wuthin a few days but one of 

 the eight cannibals was left, and several of the others had been 

 killed. 



It was curious that in every instance where two or more of the 

 cannibals were placed at close quarters, even though other larvae 

 were present, the result was the destruction of one or both of the 

 cannibals, while the others frequently remained unharmed. This 

 result is not due to the natural enmity of competitors or to a wise 

 foresight with regard to a limited food supply, but purely to the 

 strongly modified reactions of the cannibals themselves. While 

 an ordinary larva instinctively avoids close contact with another, 

 and beats the most precipitate retreat at the merest touch of can- 

 nibalistic jaws, the possessors of these weapons themselves are 

 apparently wholly divested of this innate fear. Unless decidedly 

 hungry they lie sluggishly at the bottom, either ignoring the 

 chance contacts of other specimens or savagely nabbing the in- 

 truder. The violence and instantaneousness of their occasional 

 movements contrast strongly with their sluggish inactivity be- 

 tween whiles. Even complete satiety does not usually check their 

 savage attacks, provided that the proper stimulus is offered ; the 

 prey is then seized and held some time or half swallowed, to be 

 then as quickly rejected by a sudden jerk much like the one by 

 which it was seized. Thus it is that cannibals in close proximity 

 almost invariably prove each other's undoing, the swallower fre- 

 quently succumbing as well as the swallowed. Even when taken 

 in the reservoir, not a few of the broad-heads were sadlv bitten 



253 



