BUTTERFLIES AND ANTS. 165 
butthey take the greatest care of the larvae, defend them from 
their enemies to the best of their ability, and when they are about 
to turn to pupx, conduct them to a safe place, where they may per- 
form their transformations, and allow the newly-emerged and at 
first helpless butterflies to escape unmolested. 
On the penultimate or twelfth segment ofthese Lycenid larva on 
the upperside are two erectile tentacula, and on the cleventh segment 
on the upperside in the dorsal line is an oval opening from which 
exudes a sweet liquid, of which the ants are inordinately fond, and 
to obtain which their care and attention of the larve is entirely due. 
The two tentacula on the twelfth segment do not apparently give off 
any fluid, and I have a theory, whether correct or no I cannot prove, 
that they were originally developed in the larvee to drive away their 
enemies, probably Ichneumon-flics, much as the tentacula behind 
theheads of Papilio larvee are used. In the larve of the genus 
Curetis, which do not appear to be affected by ants, these tentacula 
are very large, and when the insect is touched, they are extended 
with lightning-like rapidity and whirled round with great velocity. 
On Plate No. 26 will be found two enlarged figures of the larva of 
QO. thetys, Drury, bred by mein Calcutta, also an enlarged figure of 
its eurious jelly-like pupa. In the genus Curetis the tentacula are 
very long, much longer than in any other Lyceenid larvee known to 
me, and furnished at the apex witha tuft of long hairs forming a 
rosette, very similar in appearance to the anal tufts which the males 
of butterflies of the sub-family Danaince can thrust out at will. 
Curetis larvee appear to have no honey-gland, so ants do not attend 
them, nay, would probably eat them with much relish if they came 
across them, and were not frightened away by the rapidly-whirling 
tentacula. In the larve with the honey-gland these tentacula are 
much smaller and have short hairs at the apex; in these larvee they 
are probably becoming aborted from want of use. In the larvee 
of the genus Virachola, which feed in the interior of certain fruits, 
I have been unable to discover that they possess either a honey- 
gland or tentacula. They are not therefore, I believe, attended by 
ants, though Mr. W. C. Taylor and his daughter, Mrs. Wylly, hold 
to the contrary opinion, from observations they have made on the 
habits of these larvae. 
M. Guenée seems to have been the first to notice the existence of 
these organsin the larve of Polyommatus beticus, Linnzeus, which 
occurs in France as well as in India and elsewhere, This was in 
