The cylindrical exit-way passes through the strata of wood along the shortest line, almost normally, after a slight bend which connects the vertical with the horizontal, a curve with a radius large enough to allow the stiff Buprestis to tack about without difficulty. It ends in a blind-alley, less than a twelfth of an inch from the surface of the wood. The eating away of the untouched sheet of wood and of the bark is all the labour that the grub leaves the insect to perform. Having made these preparations, the larva withdraws, strengthening the wooden screen, however, with a layer of fine sawdust; it reaches the end of the round gallery, which is prolonged by the completely choked flat gallery; and here, scorning a special chamber or any upholstery, it goes to sleep for the nymphosis, with its head towards the exit.
I find numbers of specimens of a black Buprestis (_B. octoguttata_) in the old stumps of pine-trees left standing in the ground, hard outside but soft within, where the wood is as pliable as tinder. In this yielding substance, which has a resinous aroma, the larvæ spend their life. For the metamorphosis they leave the unctuous regions of the centre and penetrate the hard wood, where they hollow out oval recesses, slightly flattened, measuring from twenty-five to thirty millimetres[1] in length. The major axis of these cells is always vertical. They are continued by a wide exit-path, sometimes straight, sometimes slightly curved, according as the tree is to be quitted through the section above or through the side. The exit-channel is nearly always bored completely; the window by which the insect escapes opens directly upon the outside world. At most, in a few rare instances, the grub leaves the Buprestis the trouble of piercing a leaf of wood so thin as to be translucent. But, if easy paths are necessary to the insect, protective ramparts are no less needed for the safety of the nymphosis; and the larva plugs the liberating channel with a fine paste of masticated wood, very different from the ordinary sawdust. A layer of the same paste divides the bottom of the chamber from the low-ceilinged gallery, the work of the grub's active life. Lastly, the magnifying-glass reveals upon the walls of the cell a tapestry of woody fibres, very finely divided, standing erect and closely shorn, so as to make a sort of velvet pile. This quilted lining, of which the Cerambyx of the Oak showed us the first example, is, it seems to me, pretty often employed by the wood-eaters, Buprestes as well as Longicorns.