JIYEMECOPHILOUS LYa^ENID CHRYSALIS. 



75 



caudal borders, respectively. A direct connection is thus indicated as 

 existing between the interior of the body of the pupa and the chitinous 

 crater. 



It is a great pity that the pupa sent to me was not suitable for 

 sectioning. Prom the striking size of the chitinous crater and the con- 

 nection with the inside of the pupa, we ma]^ conclude with comparative 

 safety that secretory glands are also present. This pupa must therefore 

 be considered mjTmecophilous, and certainly, as such, of a species which 

 actually supplies the ants with some kind of secretion. It is an ascer- 

 tained fact that the pupae of myrmecophilous Lycjenids are often found 

 in ant nests. So far as all the palearctic species are concerned, the 

 presence of lepidopterous pups in colonies of ants is not usual, but more 

 or less accidental. At the time of pupation, when the caterpillars leave 

 the food plant, they very often wander into the immediate vicinity or 

 even into the nests of ants which frequently have their colonies located at 

 the roots of the plant. That the pupte are permitted in the nests of ants 

 is easy to understand on the basis of the s3'mbiotic relationship existing 

 between the caterpillar and the ants. The ants experience the trans- 

 formation of -the caterpillar into the pupa in their very midst, as it 

 were. Very probably their toleration of the pupa is enhanced 'by tlie 

 existence of some pleasantly scented substance, secreted by glandular 

 cells situated in its epidermis; at least the action of the ants seems to 

 point in that direction. In none of the lyctenid pupss hitherto known 

 to me is the reason for this tolerance to be sought in the presentation of 

 some kind of gift by the pupa to the ants, because all pupffi of this family 

 so far laiown are without a secretoiv organ. 



Figs. 1, 2, 3. — Lateral, Ventral and Dorsal Aspects. 



Fig. 4. — Myrmecophilous Organ. 



