of Aldrovanda.  The spiral arms no doubt act as accessory traps.  Until fresh leaves are examined, it cannot be told whether the line of junction of the spirally wound lamina is a little open along its whole course, or only in parts, but a small creature which forced its way into the tube at any point, would be prevented from escaping by the incurved hairs, and would find an open path down [page 451] the tube into the neck, and so into the utricle.  If the creature perished within the spiral arms, its decaying remains would be absorbed and utilised by the bifid papillae.  We thus see that animals are captured by Genlisea, not by means of an elastic valve, as with the foregoing species, but by a contrivance resembling an eel-trap, though more complex.

Genlisea africana (South Africa).—­Fragments of the utriculiferous leaves of this species exhibited the same structure as those of Genlisea ornata.  A nearly perfect Acarus was found within the utricle or neck of one leaf, but in which of the two was not recorded.

Genlisea aurea (Brazil).—­A fragment of the neck of a utricle was lined with transverse rows of hairs, and was furnished with elongated papillae, exactly like those within the neck of Genlisea ornata.  It is probable, therefore, that the whole utricle is similarly constructed.

Genlisea filiformis (Bahia, Brazil).—­Many leaves were examined and none were found provided with utricles, whereas such leaves were found without difficulty in the three previous species.  On the other hand, the rhizomes bear bladders resembling in essential character those on the rhizomes of Utricularia.  These bladders are transparent, and very small, viz.  Only 1/100 of an inch (.254 mm.) in length.  The antennae are not united at their bases, and apparently bear some long hairs.  On the outside of the bladders there are only a few papillae, and internally very few quadrifid processes.  These latter, however, are of unusually large size, relatively to the bladder, with the four divergent arms of equal length.  No prey could be seen within these minute bladders.  As the rhizomes of this species were furnished with bladders, those of Genlisea africana, ornata, and aurea were carefully [page 452] examined, but none could be found.  What are we to infer from these facts?  Did the three species just named, like their close allies, the several species of Utricularia, aboriginally possess bladders on their rhizomes, which they afterwards lost, acquiring in their place utriculiferous leaves?  In support of this view it may be urged that the bladders of Genlisea filiformis appear from their small size and from the fewness of their quadrifid processes to be tending towards abortion; but why has not this species acquired utriculiferous leaves, like its congeners?

Conclusion.—­It has now been shown that many species of Utricularia and of two closely allied genera, inhabiting the most distant parts of the world—­Europe, Africa, India, the Malay Archipelago, Australia, North and South America—­are admirably adapted for capturing by two methods small aquatic or terrestrial animals, and that they absorb the products of their decay.