In other cases the carpels are entirely absent and their place is supplied by a leafy shoot as in a species of Verbascum, which came under my own observation. In this case the petals were virescent, and the stamens and pistils were entirely absent, hence in truth, the so-called flower more nearly resembled a branch. In a flower of a May Duke cherry, for which I am indebted to Mr. Salter, there was a gradual change from the floral to the foliar condition; thus there were five distinct lanceolate sepals, the arrangement of whose veins betokened that they were leaf-sheaths rather than perfect leaves, ten petals partly foliaceous and sheath-like as to their venation, one of them funnel-shaped, but whether from dilatation or cohesion of the margins could not be determined. The stamens were eight or ten in number, their connectives prolonged into foliaceous or petaloid appendages, so that the filament represented the stalk of the leaf. The pistil was entirely absent and its place was supplied by a branch with numerous perfectly formed stipulate leaves.

Some flowers of Anagallis arvensis described by Dr. Marchand[119] are so interesting and show so well the gradual stages by which this malformation is arrived at, that it is desirable to cite the summary of Dr. Marchand's researches as given in the 'Gardeners' Chronicle' by Mr. Berkeley, taking that instance first in which the parts of the flower departed least from the normal condition, and then the others in their proper order. In all the parts there was a greater or less tendency to assume a green tint; in some they were entirely green, in others the brighter colours were confined to the more recently developed parts.

"1. In the first case then, the sepals and petals were in their normal position, though rather more dilated than usual; the anthers were fertile, the principal change existing in the ovary, the upper part of which was wanting, so that the ovules were exposed seated on the central placenta.

2. In the next step the calyx, more developed than usual, was separated from the corolla by a long peduncle, and the ovary, which was ovate, contained instead of a placenta a sort of plumule or young shoot.

3. In this case the corolla and calyx were distant from each other; there was no trace of stamens, but the axis was continued from the centre of the corolla, and ended in a leaf-bud.

4. The calyx and corolla nearly as before, but instead of stamens a whorl of little leaves was developed, in the centre of which the axis was continued, bearing at its tip two whorls of leaflets, alternately three and three.

5. In this case two out of the five stamens were normal, the other three changed into leaves, showing clearly the origin of the leaflets, in the last case, which took the place of the stamens.

6. The ovary varied in different flowers. In some the placenta was crowned with ovules; in others the ovules were replaced by a single whorl of leaflets; in others there was every shade of change from ordinary ovules to perfect leaflets; while in others, again, every ovule was converted into a leaf with a long petiole.

7. In these flowers shoots were developed in the axils of the sepals, or on the face of the petals between the point of their insertion and that of the stamens, and, what is most curious, in the interior of the ovaries round the foot of the placenta.