night to a very variable amount, rarely more than 45o; and in one lot of seedlings (purchased under the name of O. tropaeoloides, but certainly belonging to the above variety) they rose only from 5o to 15o above the horizon.  The pulvinus is developed imperfectly and to an extremely variable degree, so that apparently it is tending towards abortion.  No such case has hitherto, we believe, been described.  It is coloured green from its cells containing chlorophyll; and it is seated nearly in the middle of the petiole, instead of at the upper end as in all the other species.  The nocturnal movement is effected partly by its aid, and partly by the growth of the upper part of the petiole as in the case of plants destitute of a pulvinus.  From these several reasons and from our having partially traced the development of the pulvinus from an early age, the case seems worth describing in some detail. [page 119]

[When the cotyledons of O. corniculata were dissected out of a seed from which they would soon have naturally emerged, no trace of a pulvinus could be detected; and all the cells forming the short petiole, 7 in number in a longitudinal row, were of nearly equal size.  In seedlings one or two days old, the pulvinus was so indistinct that we thought at first that it did not exist; but in the middle of the petiole an ill-defined transverse zone of cells could be seen, which were much shorter than those both above and below, although of the same breadth with them.  They presented the appearance of having been just formed by the transverse division of longer cells; and there can be little doubt that this had occurred, for the cells in the petiole which had

Fig. 64.  Oxalis corniculata:  A and B the almost rudimentary pulvini of the cotyledons of two rather old seedlings, viewed as transparent objects.  Magnified 50 times.

been dissected out of the seed averaged in length 7 divisions of the micrometer (each division equalling .003 mm.), and were a little longer than those forming a well-developed pulvinus, which varied between 4 and 6 of these same divisions.  After a few additional days the ill-defined zone of cells becomes distinct, and although it does not extend across the whole width of the petiole, and although the cells are of a green colour from containing chlorophyll, yet they certainly constitute a pulvinus, which as we shall presently see, acts as one.  These small cells were arranged in longitudinal rows, and varied from 4 to 7 in number; and the cells themselves varied in length in different parts of the [page 120] same pulvinus and in different individuals.  In the accompanying figures, A and B (Fig. 64), we have views of the epidermis* in the middle part of the petioles of two seedlings, in which the pulvinus was for this species well developed.  They offer a striking contrast with the pulvinus of O. rosea (see former Fig. 63), or of O. Valdiviana.  With the seedlings, falsely called O. tropaeoloides, the cotyledons of which rise very little at night,