Heterogynis paradoxa. 729 



a rather improbable attribution of instinct to the Pimpla 

 larva. Another explanation offers itself. The ITcierogpiis 

 larva must be not unfrequently stung by both parasites as 

 they are so common; what will happen in this case? Since 

 no larva produced two parasites (we may assume that each 

 species only stings a larva once, recognizing and avoiding 

 one already stung by its own species), either they grow 

 2Mri passic, and both perish for want of pabulum, or one 

 destroys the other; if it does this by treating the other as 

 part of its host's tissues and devouring it, we have again 

 no outward indication that two parasites were at any time 

 present. But suppose the Casinaria larva already well 

 grown at the time the Pimpla stings its victim, then 

 using its piercing ovipositor, it, quite by accident, as it 

 were, pierces also and lays her egg in the Casinaria larva, 

 the result would be what I observed, the Casinaria would 

 be so far ahead of the Fimp)Ia as to make its cocoon. The 

 condition would be one of ordinary hyperparasitism so far 

 as procedure goes, but as regards intention and instinct 

 would be purely accidental. It presents, however, no 

 matter how we explain it, a very interesting accident as 

 affording material which might be elaborated by natural 

 selection into true hyperjmrasitism. 



Explanation of Plate XXVIII. 



Larva of Heterojyms paradoxa, Kainli. (First stage.) 



Fig. 1. Dorsal view x 20, showing tubercles and pseudo-tubercles. 



Fig. 2. Portion of same x 250. Below 1 is a tubercle ; Iselow 2 and 

 3, its two attendant pseudo-tubercles. 



December .30, 1902 



