from the glands; so that, when an insect comes into contact with a drop, it is able to crawl away, but soon touches other drops, and then, smothered by the secretion, sinks down on the sessile glands and dies.  Another difference is, that the secretion from the tall glands, before they have been in any way excited, is strongly acid, and perhaps contains a small quantity of the proper ferment.  Again, these glands do not secrete more copiously from being excited by the absorption of nitrogenous matter; on the contrary, they then absorb their own secretion with extraordinary quickness.  In a short time they begin to secrete again.  All these circumstances are probably connected with the fact that insects do not commonly adhere to the glands with which they first come into contact, though this does sometimes occur; and that it is chiefly the secretion from the sessile glands which dissolves animal matter out of their bodies.

Roridula.

Roridula dentata.—­This plant, a native of the western parts of the Cape of Good Hope, was sent to me in a dried state from Kew.  It has an almost woody stem and branches, and apparently grows to a height of some feet.  The leaves are linear, with their summits much attenuated.  Their upper and lower surfaces are concave, with a ridge in the middle, and both are covered with tentacles, which differ greatly in length; some being very long, especially those on the tips of the leaves, and some very short.  The glands also differ much in size and are somewhat elongated.  They are supported on multicellular pedicels.

This plant, therefore, agrees in several respects with [page 343] Drosophyllum, but differs in the following points.  I could detect no sessile glands; nor would these have been of any use, as the upper surface of the leaves is thickly clothed with pointed, unicellular hairs directed upwards.  The pedicels of the tentacles do not include spiral vessels; nor are there any spiral cells within the glands.  The leaves often arise in tufts and are pinnatifid, the divisions projecting at right angles to the main linear blade.  These lateral divisions are often very short and bear only a single terminal tentacle, with one or two short ones on the sides.  No distinct line of demarcation can be drawn between the pedicels of the long terminal tentacles and the much attenuated summits of the leaves.  We may, indeed, arbitrarily fix on the point to which the spiral vessels proceeding from the blade extend; but there is no other distinction.

It was evident from the many particles of dirt sticking to the glands that they secrete much viscid matter.  A large number of insects of many kinds also adhered to the leaves.  I could nowhere discover any signs of the tentacles having been inflected over the captured insects; and this probably would have been seen even in the dried specimens, had they possessed the power of movement.  Hence, in this negative character, Roridula resembles its northern representative, Drosophyllum.