I had frequently noticed that tunnels brought from the tank on the under surface of _Limnanthemum_ leaves had a _Hydra_ fixed to them. This occurred in about a third of the occupied shelters examined. The _Hydra_ was always in a contracted condition and often more or less mutilated. By keeping a larva together with a free polyp in a glass of clean water, I have been able to observe the manner in which the polyp is captured and entangled. The larva settles down near the base of its column and commences to spin a tunnel. When this is partially completed, it passes a thread round the polyp's body to which it gives a sharp bite. This causes the polyp to bend down its tentacles, which the larva entangles with threads of silk, doing so by means of rapid, darting movements; for the nettle-cells would prove fatal should they be shot out against its body, which is soft. Its head is probably too thickly coated with chitin to excite their discharge. Indeed, small larvæ of this very species form no inconsiderable part of the food of the polyp, and, so far as my observations go, a larva is always attacked in the body and swallowed in a doubled-up position. When the _Hydra_ has been firmly built into the wall of the shelters and its tentacles fastened down by their bases on the roof, the larva proceeds, sometimes after an interval of some hours, to eat the body, which it does very rapidly, leaving the tentacles attached to its shelter. The meal only lasts for a few minutes; after it the larva enjoys several hours' repose, protected by remains of its victim, which retain a kind of vitality for some time. During this period it remains still, except for certain undulatory movements of the posterior part of the body which probably aid in respiration. Then it leaves the shelter and goes in search of further prey. Its food, even when living in a tunnel, does not consist entirely of _Hydra_. I have watched a larva building its shelter near a number of rotifers, some of which it devoured and some of which it plastered on to its tunnel. The tubular shelters occasionally found are very much stouter structures than the tunnels, but are apparently made fundamentally of the same materials; and structures intermediate between them and the tunnels are sometimes produced. The larva as a rule fastens to them branches detached from living colonies of Vorticellid protozoa such as _Epistylis_[AT]. [Footnote AT: Further particulars regarding the life-history of this larva will be found on pp. 114 and 115, J. Asiat. Soc. Bengal, ii (n. s.) 1906.]