It seems at first sight a probable view that aggregation is due to the glands being excited to secrete more copiously, so that sufficient fluid is not left in their cells, and in the cells of the pedicels, to hold the protoplasm in solution.  In favour of this view is the fact that aggregation follows the inflection of the tentacles, and during the movement the glands generally, or, as I believe, always, secrete more copiously than they did before.  Again, during the re-expansion [page 56] of the tentacles, the glands secrete less freely, or quite cease to secrete, and the aggregated masses of protoplasm are then redissolved.  Moreover, when leaves are immersed in dense vegetable solutions, or in glycerine, the fluid within the gland-cells passes outwards, and there is aggregation; and when the leaves are afterwards immersed in water, or in an innocuous fluid of less specific gravity than water, the protoplasm is redissolved, and this, no doubt, is due to endosmose.

Opposed to this view, that aggregation is caused by the outward passage of fluid from the cells, are the following facts.  There seems no close relation between the degree of increased secretion and that of aggregation.  Thus a particle of sugar added to the secretion round a gland causes a much greater increase of secretion, and much less aggregation, than does a particle of carbonate of ammonia given in the same manner.  It does not appear probable that pure water would cause much exosmose, and yet aggregation often follows from an immersion in water of between 16 hrs. and 24 hrs., and always after from 24 hrs. to 48 hrs.  Still less probable is it that water at a temperature of from 125o to 130o Fahr. (51o.6 to 54o.4 Cent.) should cause fluid to pass, not only from the glands, but from all the cells of the tentacles down to their bases, so quickly that aggregation is induced within 2 m. or 3 m.  Another strong argument against this view is, that, after complete aggregation, the spheres and oval masses of protoplasm float about in an abundant supply of thin colourless fluid; so that at least the latter stages of the process cannot be due to the want of fluid to hold the protoplasm in solution.  There is still stronger evidence that aggregation is independent of secretion; for the papillae, described in the first chapter, with which the [page 57] leaves are studded are not glandular, and do not secrete, yet they rapidly absorb carbonate of ammonia or an infusion of raw meat, and their contents then quickly undergo aggregation, which afterwards spreads into the cells of the surrounding tissues.  We shall hereafter see that the purple fluid within the sensitive filaments of Dionaea, which do not secrete, likewise undergoes aggregation from the action of a weak solution of carbonate of ammonia.