For the most part it inhabits the main stem of the tree, its gallery always having an outlet, which is covered with a curtain of silk and refuse, and is spun exactly level with the surrounding bark, and very inconspicuous. These burrows usually run towards the ground, and are mostly two or three inches from the surface of the trunk. In some instances the larvæ inhabit branches, in which case, if they are small, the tunnels are made near the centre. Later on in its life, but probably some time before its transformation into the pupa, the caterpillar of this insect constructs a far more complicated burrow than the above. It consists of a spacious, irregular, but shallow cavity, just under the bark, having a very large opening to the air, which is entirely covered { 130}with a thin silken curtain, almost exactly the same shape and size as the numerous marks occurring at intervals on the trunks of many of the trees. Three large tunnels open into this shallow cavity: one in the centre, which runs into the middle of the stem, and one on each side, which run right and left just under the bark. These lateral tunnels are usually very short, but sometimes they extend half-way round the tree, and occasionally even join one another on the opposite side. The central tunnel has a slightly upward direction for a short distance inwards, which effectually prevents it from becoming flooded in wet weather; afterwards it pursues an almost horizontal course until it reaches the centre of the tree, when it appears to suddenly terminate. This, however, is not the case, for, if the gallery floor be carefully examined a short distance before its apparent termination, a round trap-door will be found, compactly constructed of very hard, smooth silk, and corresponding with the surrounding portion of the tunnel so exactly that it almost escapes detection. When this lid is lifted a long, perpendicular shaft is disclosed, which runs down the middle of the tree to a depth of 14 or 16 inches, and is about ½ inch in diameter. The upper end of this shaft is lined with silk, which forms a framework on which the trap-door rests when closed. The lid itself is of a larger size than the orifice which it covers, and this makes it extremely difficult, if not impossible, to force it open from the exterior, especially as it always fits down very closely as long as the insect remains in its burrow. The object of this contrivance is, no doubt, to prevent the ingress of enemies, large numbers of spiders, slugs, wood-lice, and various orthoptera being frequently found in both central and lateral tunnels, but they are quite unable to pass the trap-door. The galleries of individual larvæ are all wonderfully alike, the only differences observable being in the length of the perpendicular shaft, and in the direction of the horizontal burrow, which is sometimes curved. These variations are usually caused by the presence of other tunnels in the tree, which the larva appears to carefully avoid; at least I have never known an instance where a larva has allowed its tunnel to communicate with another one, whether inhabited or otherwise.