328 Annals Entomological Society of America [Vol. VIII, 
whole outer surface was covered with hundreds of ants, all 
forming living sutures along the contiguous edges of the leaves. 
In most cases these sutures consisted of a single row of ants, 
side by side, with their mandibles grasping the edge of one 
leaf and the large claws of all their backwardly directed feet 
fixed into the edge of another adjacent leaf. Where the gap 
between the leaves of the outer layer was too wide to be spanned 
by single ants, there were parallel chains each consisting of 
two to seven ants, each ant holding the petiole of the ant in 
front in its mandibles and being grasped in the same place 
by the ant behind, exactly as described by Dodd for CE. virescens 
and figured by Bugnion for the true smaragdina of Ceylon.* 
The longer chains often ran diagonally across the shorter. 
The nest presented a startling appearance, with these hundreds 
of green ants immovably fixed on its outer surface and accentu- 
ating the more or less contiguous borders of the leaves. Then | 
I noticed that there was great activity on the branch leading 
to the nest. Files of workers were running along it to and 
from the nest and could be traced to two much smaller nests 
consisting of dead leaves on an adjacent tree about 30 feet 
away. The colony had evidently outgrown these nests and 
was in the act of building the more commodious domicile I had 
been observing. Closer inspection showed that about one in a 
dozen of the workers coming to the new nest was carrying in its 
mandibles a minute milk-white larva, and as the larve accumu- 
lated, I was able to observe all the stages in the spinning of the 
silken film. Many of the ants worked from the inside, but quite 
a number stationed themselves on the outside of the nest where 
I could see them very clearly under my pocket lens, while they 
moved their larve back and forth as living shuttles from the 
edge of one leaf to that of another, pausing only while the 
larve attached their extremely delicate threads to the surface 
of the leaf. A single ant would sometimes work in the same 
spot for 10 to 20 minutes, moving its larva so nearly through 
the same arc as to produce a stout silken band or cord from the 
*The question naturally suggests itself as to whether the greatly elongated 
petiole of Gtcophylla is not an adaptation to this peculiar use. I have recently 
(1914) called attention to the fact that the fossil species of Gicophylla (brevinodis 
Wheeler and brischkei Mayr of the Baltic amber and sicula Emery of the Sicilian 
amber) have a shorter petiole than the recent smaragdina. This indicates, per- 
haps, that the species of the Lower Oligocene and Miocene had not yet acquired or 
were merely in process of acquiring the habit of forming chain sutures. 
