MELIT.EA I. 



The larvffi were immediately placed upon a plant of Clielone, and in obtain- 

 ing this I noticed many little webs already constructed. One colony which I 

 encountered was in a state of great agitation, its members running about wildly 

 and throwing their heads and two thirds the Ijody in a jerking way from right 

 to left, all in same manner and like so many aut(jmata. The cause of tlie alarm 

 seemed to be a small crimson ichneumon fly which was hovering about, and which 

 alighted on one of the leaves which protruded from the web. My presence 

 fi-ightened it away, however, so that I failed to see its mode of attack. 



Placing the young larvfe on the leaves, some at the top and some well down 

 the stem, each lot began at once the construction of a web ; in case of the ter- 

 minal leaves, drawing them together. While part were busy at this, othei's were 

 eatiitg the pulp, and even gnawing the midi-ib, by which the leaves easily yielded 

 and were soon folded over and incorporated in the web. Those on one of the 

 lower leaves in same way contrived to double the leaf over, but I noticed that 

 in a few hours all those which began life low down the stem had climbed to the 

 upper leaves and joined the colony there. As the larvae grew, leaf after leaf was 

 inclosed, a detachment working at the next pair of leaves below, which were 

 taken possession of by the colony as soon as the upper ones were consumed. 

 These first webs^were slight, and quite transparent, the warp composed of long, 

 regular and colorless threads, which ran from the stem to the middle of the next 

 leaves below, and were bound together by innumerable ci'oss-threads which took 

 no regularity whatever. Two or more holes were left for egress, and the eat- 

 ing away of the leaf made other openings. The web at this stage does not resist 

 rain verv well, and in a long continued storm more or less destruction of the 

 larva? follows. But as soon as the skies clear, the survivors set themselves at 

 repairing or rebuilding altogether. 



Six days after leaving the egg, the larv^ began to pass the first moult, which 

 process continued for two days before all the colon}' had changed, and forthwith 

 they eat voraciousl}', stripping the leaves much faster than they could form a 

 web to cover them. Manv of the larva? during this staj^e remained outside in 

 small clusters, and fed on separate leaves, even on separate stems ; but as the 

 time for the second moult drew near, they came together and constructed a loose 

 web, within which the moult took place. 



About the middle of the next stage, on a bright day, I noticed that each of my 

 colonies was suddenly active in spinning, and it at once occurred to me that they 

 anticipated a storm and were providing against it. Before night much more 

 substantial and larger webs than had been hitherto constructed were ready for 

 occupation. During the night a steadj^ rain set in and continued for twenty-four 

 hours, but it found each coloixy safely housed. The next day was clear, and 



