common plant in wet ditches and marshes throughout the whole of Southern Britain, represents the very earliest petal-bearing type in this line of development; indeed, save that its petals are now pinky-white, while those of the original ancestor were almost certainly yellow, we might almost say that the marsh-weed in question was really the earliest petal-hearing plant of which we are in search. It closely resembles in appearance, and in the arrangement of its parts, the buttercups, which are the earliest existing members of the other or quinary division of flowering plants; and in both we seem to get a survival of a still earlier common ancestor, only that in the one the parts are arranged in rows of three, while in the other they are arranged in rows of five; and concomitantly with this distinction go the two or three other distinctions which mark off the two main classes from one another—namely, that the one has leaves with parallel veins, only one seed-leaf to the embryo, and an endogenous stem, while the other has leaves with netted veins, two seed-leaves to the embryo, and an exogenous stem. Nevertheless, in spite of such fundamental differences, 
Fig. 1.—a. ovaries; b, stamens, inner whorl; c, stamens, outer whorl; d, petals; e, calyx-pieces. we may say that the alismas and the buttercups really stand very close to one another in the order of development. When the two main branches of flowering plants first diverged from one another, the earliest petal-bearing form they produced on one divergent branch was the alisma, or something very like it; the earliest petal-bearing form they produced on the other divergent branch was the buttercup, or something very like it. Hence, whenever we have to deal with the pedigree of either great line, the fixed historical point from which we must needs set out must always be the typical alismas or the typical buttercups. The accompanying diagram will show at once the relation of parts in the simplest trinary flowers, and will serve for comparison at a later stage of our argument with the arrangement of their degraded descendants, the wheats and grasses.
Our own smaller alisma has a number of ovaries loosely scattered about in its center, as in the buttercups, with two rows of three stamens outside them, and then a single row of three petals, followed by the calyx or inclosing cup of three green pieces. Its close ally the water-plantain, however, shows signs of some advance toward the typical lily form in the arrangement of its ovaries in a single ring, often loosely divisible into three sets. And in the pretty pink flowering rush (not of course a rush at all in the scientific sense) the advance
