1904] HEMILEUCA BUD LEVI. II9 



died. Two in a tin would eat, but not heartily, and six or eight were needed to 

 make a success of life, at this stage. 



The fourth moult came in nine days. The caterpillars were as before except 

 that the long spines on the thoracic and anal segments were of a dull buff color, 

 much longer and more branching than the others. Few larvae ate their skins after 

 this moult. A few left the group and fed on separate leaves, but most of them fed 

 in close rows as before, lying as near each other as their spines would allow. 



Before each moult the yellow color disappeared from the spines, leaving them 

 like little glass tubes. This time the caterpillars spun the leaf to the twig, then 

 spun a mat on the leaf and lay side by side to moult. 



The fifth moult took place ten days after the fourth, bringing no change of 

 color except a white dot at each end of the spiracles, which hitherto had been diffi- 

 cult to find, even with a lens, they were so black, but had now a whitish line at the 

 opening, as well as a dot at each end. The larvae ate most of their cast skins. 

 The eversible sacs were still more noticeable, but no odor was perceptible. Ten 

 days later there was a faint, pleasant odor when the boxes were opened. 



The caterpillars ate voraciously, and seemed to have no instinct in choosing 

 their leaves. I found that certain saplings had specially succulent-looking leaves 

 and gave them to several sets of larvae. In every case the larvae had the symptoms 

 of severe colic but would not refuse the leaves. By giving them older leaves from 

 a tree, not a sapling, I could cure the sick, but they would not choose the older 

 leaves themselves, taking whichever kind was nearer. They were now a little over 

 four inches long, and the most superb caterpillars I ever saw. 



(hi August 28th, twelve days from the last moult, they emptied the intestines 

 and, two days later, spun dark brown cocoons, drawing leaves together around 

 them, and sometimes spinning to a twig. Each cocoon had one end open, and, 

 about half an inch within this open end, a grating of silk across the cocoon, admit- 

 ting air. 



The pupae were not formed — or did not cast the larva-skin — until September 

 1 2th, sixty-seven days from the egg. 



Part of a second brood omitted the fifth moult, pupating instead of moulting, 

 while the rest of the brood moulted five times. This second brood was sixty-nine 

 days in passing from egg to pupa. 



The spines of bndleyi urticated as much as those of io, rather more in the last 

 stage, but with no lasting effect. 



My larvae preferred Prunus virginiana^ and would not eat P. seroiina after the 

 second moult, but those I sent to Miss Eliot, half of each brood, fed wholly on 

 P. serotina. 



When the caterpillars reached the length of four inches, or nearly that, I had 

 but two in each large tin to insure their having plenty of space and food supply. 



