6 BULLETIN 1076, U. S. DEPARTMENT OF AGRICULTURE. 
that the larve marooned on this leaf, unable to seek their usual pu- — 
pation quarters, were forced to enter the petiole. 
From laboratory experiments it was found that when placed in 
water the larve, especially those nearly mature, were sustained on 
the surface film and were able to make considerable progress by lat- 
eral contortions. When once fully submerged, however, they were 
unable to regain the surface. Larve of Pyrausta ainsliei under the 
same conditions behaved similarly but did not advance as rapidly 
when swimming. 
FEEDING HABITS ON THE PODS. 
The larvee of the first generation apparently utilize the green, 
growing pod for food as well as for pupation. After the floral parts 
fall away, the pod normally droops (Pl. I, B, 6) until its flat, seed- 
bearing face is vertical, becoming erect again when nearly mature 
(Pl. I, B, c). Until maturity the seeds remain tightly embedded in 
their sockets, and only as the pods ripen, turn brown, and dry do 
they become loose, ultimately to be shaken out and sink in the water 
and mud around the plants. 
In entering the young pod the larva usually selects a point just 
below the rim and on the underside of the pod as it hangs horizon- 
tally (Pl. IT, A); less often it cuts in between the seeds on the flat 
face of the pod (Pl. II, C). As the larva feeds within, soft brown 
frass is pushed out of the entrance hole in considerable quantity, 
eventually drying and falling away or being washed off by rain. 
The interior of the fruit or pod is filled with parenchymatous tissue 
through which run the vascular bundles nourishing the developing 
seeds. After entering, the larva eats out more or less of a cavity in 
this soft tissue and often cuts into or through two or three or more 
of the seeds. Whenever a seed is injured even slightly or the 
vascular bundle beneath it is cut, it turns brown and soon shrivels 
to a mere remnant. These empty or partially empty sockets (PI. 
II, B) are very conspicuous and almost invariably indicate the pres- 
ence or work of the larve. Such injured pods are also frequently 
much distorted (Pl. I, B, d) and very unlike the ornamental, perfect 
specimens. Although in a few cases the eocoon was found near the 
face of the pod and, in fact, lying partially in and through some of 
the injured seeds, the larva usually makes its way well toward the 
_ base of the pod before cocooning. 
In the pods collected at Kimberlin Heights July 19 the normal 
number of seeds per pod varied from 10 to 25, with an average of 
17. The work of the larve resulted in the destruction of 5.9 seeds 
on the average in each infested pod, or 34.7 per cent of the total 
number in the infested pods and 5.88 per cent of those in the entire 
plantation. The pods developing from the scattering flowers which 
