APATURA II. 
larvae on the Plate, but the group in nature would be compact as I have before 
stated. I found that they fed principally at night, the leaf in the morning hav¬ 
ing been eaten at one spot as if all had fed at the same time. When finally 
there remained nothing but the patch on which they rested, they were forced to 
shift to a fresh leaf. From the earliest stage, the surface of the leaf about and 
beneath these larvae was kept thoroughly clean, but just outside the group was a 
mass of excrement in a pretty regular ridge and as if it had been voided at that 
place. It was some days before I discovered the explanation of this, by happen¬ 
ing to come upon the larvae one morning when they were engaged in a general 
cleansing. The colony was comparatively active, many of its members moving 
about, some crawling over the backs of the rest. When a bit of frass was encoun¬ 
tered by one of these last, who seemed especially deputed to act as scavengers, it 
was seized by the mandibles, and if very small, the head and thoracic segments 
were jerked back, and by a snap the frass was thrown some distance, at least 
two lengths of the caterpillar. If this happened near the edge of the colony the 
frass probably went over; if not, and it fell short, either one of the larvae on 
which it struck seized it, or it was met by one of the scavengers, and was again 
snapped off, until by repeating the process as often as was necessary, the obnox¬ 
ious thing was gotten rid of. When a considerable mass was encountered, it was 
broken by jaws and feet, or two or three of the larvae tugged at it till it was 
brought near the edge and toppled over. In some cases, as one of the larvae 
voided its frass, it turned and seized it, snapping it away. Presently the colony 
was cleansed and all its members resumed their usual attitude of rest. This 
sanitary work could only have been necessary when the larvae were in confine¬ 
ment, as in nature they would have rested on the under side of the leaf. 
Before the first moult took place, the larva remained immovable for about two 
days; the skin on the second segment became swollen and nearly covered the 
head. When the change occurred, this burst, disclosing the new head and ant¬ 
lers, and the skin was speedily shuffled off' and devoured. This process of moult¬ 
ing was by no means simultaneous throughout the colony, but was going on for 
two or three days before all were changed. And the same may be said of each 
successive moult. 
0 
Soon after the second moult, which occurred 1st August and days following, 
roost of the larvae began to change color, first to pink, then to brown. In other 
cases the change was very gradual, and the winter coat was not assumed before 
1st September. All these at last rested on a common bed of silk web which cov¬ 
ered the surface of the leaf, each with head bent under so that the face was in 
same plane with the lower side of the body, the back arched, and the last seg¬ 
ments appressed. But three of the larvae behaved differently, and after having 
