EXPLANATION OF PLATE I.

FIG. 1. Longitudinal section of a corn root, grown in a glass tube: a, X45; b, two hair-producing cells lapping over the other epidermal cells, X220.

FIG. 2. a, Longitudinal section of a corn root, grown in air; the section shows more than one line of epidermal cells with long and short hairs; ×75. b, rounded surface of a living root of corn, grown in redistilled water; the outlines of the epidermal cells were very indistinct; the only case observed where the difference in size was so great; X45.

FIG. 3. Roots of wheat plants which had been cut from the seeds shortly after sprouting; water culture; X.

FIG. 4. Longitudinal section of a root of corn, grown in air, showing the origin of hairs from the region where the cells are still short; X220.

FIG. 5. Longitudinal section of a root of corn grown in water, in the same experiment with the root shown in fig. 6; X45.

FIG. 6. Longitudinal section of a root of corn grown in air, in the same experiment with the root shown in fig. 5; X45.

FIG. 7. Longitudinal section of a root of corn grown in air, showing the beginning of the hairs; X220.

FIG. 8. Longitudinal section of a corn root curving on the surface of water; X34.

FIG. 9. Longitudinal section of a corn root grown in air, showing one of the large nuclei of the hypodermal cells; X220.

A CONTRIBUTION TO THE LIFE HISTORY OF

APOCYNUM ANDROSAEMIFOLIUM.

THEODORE C. FRYE and ELEANOR B. BLODGETT.

(WITH PLATE II)

It appears that not one of the Apocynaceae, a family of about 1000 species, has ever been studied carefully in reference to the minute morphology of the flower. Considering this in connection with the fact that the family stands near the Asclepiadaceae, with their peculiar pollen and stigma, it was believed that it deserved investigation. Buds and flowers of Apocynum androsaemifolium L. were collected in various stages, and the results of their investigation are herewith described.

The order of appearance of the floral whorls is centripetal. The calyx shows no peculiarities other than a ruffling of the epidermis on the abaxial surface near the base, suggesting a mechanism for the folding of the sepals.

Each petal of the campanulate corolla has on its inner surface near the base a ridge (fig. 1, r) running from the midrib diagonally outward and toward its base. It is highest at the midrib, and undoubtedly functions as an aid in compelling cross-pollination. The ridge arises from the more superficial cells of the leaf, and does not affect the course of the veins. Its meristematic crest forms the cells for its enlargement.

The stamens are peculiar in form, adjusting themselves neatly in a ring rather closely applied to the stigmatic head (fig. 1, s). At the base of each are two long auriculate appendages (figs. 1, 2, 3, ap) extending downwards dorsal to the filament. The sporangia are above the insertion of the filament, and do not extend into the appendages. The loculi open on their inner surfaces, somewhat laterally, by longitudinal slits, and immediately beneath them is a beard of epidermal hairs extending transversely across the faces of the anthers, forming a ring around the stigmatic head (figs. 1, 2, b). These hairs meet similar ones from a ring around the head, thus preventing pollen from rolling into the base of the flower.

In the development of the stamen the enlargement of the tip, foreshadowing the formation of sporogenous tissue, occurs just about the time the carpels appear. The hypodermal layer gives rise to a primary parietal layer, and another homologizing with what is ordinarily the primary sporogenous layer (fig. 4). The former divides once; the latter also divides, forming the tapetum and primary sporogenous cells (fig. 5). This has been observed in a few plants only (1), the tapetum usually arising from the primary parietal layer. The primary sporogenous cells elongate as they do in Asclepias (fig. 6), but divide into a mass of mother cells, thus reinforcing the presumption that in Ascelpias this stage is simply omitted. The pollen is in the mother cell stage when the ovules appear. The rounded mother cells do not always divide simultaneously. Division of the two daughter cells is simultaneous or nearly so (fig. 8), and almost so in all the daughter cells of a sporangium; but not in different stamens of the same flower. Sometimes one daughter cell fails to divide, and three microspores instead of a tetrad is the result (fig. 9). Occasionally some of the pollen grains near the tip of the sporangium disintegrate after tetrad division, probably serving as nourishment for the others. The whole tapetum also disintegrates soon after tetrad division.

The microspores remain in tetrads in maturity, and their arrangement with relation to each other is various. In fact, one can find all stages grading from the bilateral to the tetrahedral arrangement. Fig. 10 evidently resulted from the spindles in the second division being somewhat at right angles in the same plane, and is like a grouping found by WILLE (2) in Orchis mascula. Usually the four spores are in the same plane, but their arrangement with regard to each other varies; in fig. 11 four pollen grains meet at a point on each side of the group; in other cases there are four on one side and three on the other; in still others only three meet at a point on either side. The pollen grains in a tetrad are often plainly unequal in size. In such a group as is represented in fig. 11, suppose the upper half (cells a and b) were revolved on the lower half 90°, with xy as an axis; if then they adjusted their form to fit each other the result would be the prevailing dicotyl grouping-tetrahedral. Figs. 11, 12, 13 are three members of a series grading from the bilateral to the tetra