up the whole seed, and the peg necessarily rubbed against both tips, but did not hold either down.  The result was, that the cotyledons of five out of the nine seeds thus placed were raised above the ground still enclosed within their seed-coats.  Four seeds were buried with the end from which the radicle protrudes pointing vertically downwards, and owing to the peg being always developed in the same position, its apex alone came into contact with, and rubbed against the tip on one side; the result was, that the cotyledons of all four emerged still within their seed-coats.  These cases show us how the peg acts in co-ordination with the position which the flat, thin, broad seeds would almost always occupy when naturally sown.  When the tip of the lower half of the seed-coats was cut off, Flahault found (as we did likewise) that the peg could not act, since it had nothing to press on, and the cotyledons were raised above the ground with their seed-coats not cast off.  Lastly, nature shows us the use of the peg; for in the one Cucurbitaceous genus known to us, in which the cotyledons are hypogean and do not cast their seed-coats, namely, Megarrhiza, there is no vestige of a peg.  This structure seems to be present in most of the other genera in the family, judging from Flahault’s statements’ we found it well-developed and properly acting in Trichosanthes anguina, in which we hardly expected to find it, as the cotyledons are somewhat thick and fleshy.  Few cases can be advanced of a structure better adapted for a special purpose than the present one. [page 105]

With Mimosa pudica the radicle protrudes from a small hole in the sharp edge of the seed; and on its summit, where united with the hypocotyl, a transverse ridge is developed at an early age, which clearly aids in splitting the tough seed-coats; but it does not aid in casting them off, as this is subsequently effected by the swelling of the cotyledons after they have been raised above the ground.  The ridge or heel therefore acts rather differently from that of Cucurbita.  Its lower surface and the edges were coloured brown by the permanganate of potassium, but not the upper surface.  It is a singular fact that after the ridge has done its work and has escaped from the seed-coats, it is developed into a frill all round the summit of the radicle.*

At the base of the enlarged hypocotyl of Abronia umbellata, where it blends into the radicle, there is a projection or heel which varies in shape, but its outline is too angular in our former figure (Fig. 61).  The radicle first protrudes from a small hole at one end of the tough, leathery, winged fruit.  At this period the upper part of the radicle is packed within the fruit parallel to the hypocotyl, and the single cotyledon is doubled back parallel to the latter.  The swelling of these three parts, and especially the rapid development of the thick heel between the hypocotyl and radicle at the point where they are doubled, ruptures the tough fruit at the upper end and allows the arched hypocotyl to emerge; and this seems to be the function of the heel.  A seed was cut out of the fruit and