The general rule is,--that each larva spins for itself a separate cocoon; but amongst those of _Arctia chrysorhea_ and others which live in society, two or three sometimes begin their operations so close together that they are under the necessity of forming one common cocoon, which serves for a covering to the whole number. The same thing happens to silk-worms, the double or treble cocoons of which are called _Dupions_ by the breeders. The larvæ of some Ichneumons, besides forming each its separate cocoon, spin a joint cottony covering for the whole[508], which is effected thus:--After they leave the caterpillar they have devoured, they fix themselves side by side at a little distance from it, and begin to spin each a cocoon; and in order to defend its end and side that is not covered by others, they spin further an envelope of loose silk, and thus this exterior covering is formed.
The size, figure, colour, substance, and texture of silken cocoons are extremely various. Their _size_ indeed is usually proportioned to that of the included larva or pupa; yet it is by no means always so. Some large caterpillars spin cocoons so small, that the observer can hardly conceive how they can be contained in so narrow a compass: _Eriogaster Catax_ is a moth of this description[509]. And others smaller in size lodge themselves in apartments apparently much more spacious than necessary. The transparent hammock-like cocoons of _Hepialus Humuli_ and _Arctia villica_, two other moths, would contain several of their pupæ. I possess one in which the pupa is suspended in the centre, that is ten times its size, and not very short in dimensions of that of _Attacus Paphia_, a giant silk-moth. The largest cocoon I ever read or heard of, is that thus described by Mr. Hobhouse in his _Travels_: "Depending," says he, "from the boughs of the pines, near the Attic mountain Parnes, and stretching across from tree to tree so as to obstruct our passage, were the pods, _thrice as big as a turkey's egg!_ and the thick webs of a chrysalis, whose moth must be far larger than any of those in our country."[510] If this statement is correct, and I am not aware that there is any reason for doubting it, the cocoon must be vastly larger than the pupa, or the moth it produced would far exceed in size any yet known. Perhaps, however, as this gentleman is probably no entomologist, what he took for a cocoon might be a nidus, in which many larvæ were associated, of the nature of those formerly described[511].