1915] Cocoons Among Ants 325 
species as many Dolichoderine (Leptomyrmex, Tapinoma, 
Bothriomyrmex, etc.) dispense with it? And why is it retained 
by the fierce Formica sanguinea and so often absent in F. fusca 
and its varieties, the gentle slaves of this ant? That the 
presence of the cocoon is primitive and its suppression a second- 
ary phenomenon must be granted, and it is evident.that the loss 
of the structure must be due primarily either to a loss of the 
spinning instincts on the part of the larve or to a change in the 
instinct of the adult worker ants in so far as these instincts 
are concerned with the treatment of the larve. The latter 
change might conceivably consist in a lapse of the habit of 
burying the mature larve in the earth or in covering them with 
particles of earth or refuse, since ant larve are known to be 
quite unable to spin their elliptical cocoons unless thus tempor- 
arily enveloped with foreign particles to which they can attach 
the thread from their sericteries. 
The frequent appearance of nude pupez in the typical 
Formica fusca and some of its varieties may be attributed to 
the fact that these forms are peculiar to high altitudes and 
latitudes, where the reproductive season of the colony is con- 
siderably abbreviated. The omission of the cocoon would 
seem, therefore, to be due to “‘tachygenesis,’’ or acceleration 
of ontogenetic development, since the spinning of this envelope 
not only requires a number of hours and an expenditure of 
energy and material, and therefore delays pupation, but the 
envelope itself is probably in some degree a nonconductor and 
would therefore tend to prolong development under circum- 
stances that demand a very rapid utilization of the sun’s heat 
by the pupa if it is to reach maturity at the proper time. The 
loss of the pupal covering in the Myrmicine, Dolichoderinz 
and Dorylinz in the remote past may have ‘been due to the 
same or similar causes. 
That the behavior of the workers of the ant-colony in some 
species does actually influence the spinning activities of the 
larve is indicated by the following facts. In the great majority 
of larve which use their spinning glands, these organs are called 
into activity only at the end of the larval stage and for the 
purpose of making the cocoon, but within recent years several 
observers have called attention to a series of tropical ants belong- 
ing to at least three very different genera of Camponotine, 
