Turning now to very young seedlings, the cotyledons of which do not rise at night and are only from 2 to 2½ mm. in length, their petioles do not exhibit any defined zone of small cells, destitute of chlorophyll and differing in shape exteriorly from the lower ones. Nevertheless, the cells at the place where a pulvinus will afterwards be developed are smaller (being on an average .015 mm. in length) than those in the lower parts of the same petiole, which gradually become larger in proceeding downwards, the largest being .030 mm. in length. At this early age the cells of the blade are about .027 mm. in length. We thus see that the pulvinus is formed by the cells in the uppermost part of the petiole, continuing for only a short time to increase in length, then being arrested in their growth, accompanied by the loss of their chlorophyll grains; whilst the cells in the lower part of the petiole continue for a long time to increase in length, those of the epidermis becoming more conical. The singular fact of the cotyledons of this plant not sleeping at first is therefore due to the pulvinus not being developed at an early age. We learn from these two cases of Lotus and Oxalis, that the development of a pulvinus follows from the growth of the cells over a small defined space of the petiole being almost arrested at an early age. With Lotus Jacobæus the cells at first increase a little in length; in Oxalis corniculata they decrease a little, owing to self-division. A mass of such small cells forming a pulvinus, might therefore be either acquired or lost without any special difficulty, by different species in the same natural genus: and we know that with seedlings of Trifolium, Lotus, and Oxalis some of the species have a well-developed pulvinus, and others have none, or one in a rudimentary condition. As the movements caused by the alternate turgescence of the cells in the two halves of a pulvinus, must be largely determined by the extensibility and subsequent contraction of their walls, we can perhaps understand why a large number of small cells will be more efficient than a small number of large cells occupying the same space. As a pulvinus is formed by the arrestment of the growth of its cells, movements dependent on their action may be long-continued without any increase in length of the part thus provided; and such long-continued movements seem to be one chief end gained by the development of a pulvinus. Long-continued movement would be impossible in any part, without an inordinate increase in its length, if the turgescence of the cells was always followed by growth.