measure only about 2 cm. in length and 4-5 mm. in width. The female plants (Fig. 9, B) are about 10 cm. in length and 7 mm. wide.

The wings of the thallus are relatively very wide, and as usual but one cell thick. They are waved slightly on the margin, which is entire except for an occasional cell which projects slightly so as to form an inconspicuous tooth. Anteriorly the lamina of the antheridial plant narrows abruptly to the apex of the shoot, which is indented. The midrib projects strongly on the ventral side. The rhizoids are much less numerous than in P. radiculosa, and quite different in color, being a rather light brown, instead of the deep purple-red found in P. radiculosa. The conducting strand of the midrib is much like that of P. radiculosa.

From the sides of the midrib in the male plant extend a series of scales which form a more or less continuous shelf-like structure. These scales have their free margins deeply lobed and toothed, and underneath the shelf formed by these confluent scales the antheridia occur, either singly or in groups of several together (Fig. 9, A). The antheridia arise in acropetal succession, and are arranged in a somewhat broken row along each side of the midrib. They are more completely covered by the subtending scales than is the case in P. radiculosa, and are not so evidently divided into groups separated by sterile areas.

The antheridia, to judge from a somewhat cursory examination, appear to agree in all essential details of structure with those of P. radiculosa.

The apical cell of the thallus is not so deeply placed as in P. radiculosa, but a study of horizontal and vertical sections shows that it has the same form. On the ventral surface of the midrib in P. Levieri, as in P. radiculosa, there may occasionally be found small groups of meristematic cells, which appear dormant, and do not show a definite apical cell. The smaller groups are slightly sunken; the larger ones, perhaps having resumed activity, form hemispherical protuberances. The origin and development of these groups of cells is apparently the same as the similar ones in P. Zollingeri, and probably under favorable conditions these give rise to normal branches.

The archegonial plants of P. Levieri (Fig. 9, B), besides being much larger than the male plants, show also a very different appearance at the apex of the shoot. There is no indentation at this point, but the thallus is prolonged ino a nearly cylindrical process of some length, in which the wings are quite suppressed. It is possible that an examination of a larger number of individuals might show that this difference in apices of the male and female plants is not constant. The rhizoids of the female plants

of P. Levieri were much longer than those of either of the other species that were studied, sometimes reaching a length of a centimeter or more The archegonial receptacles are smaller than those of P. radiculosa, and are more like those of P. Zollingeri.

The Antheridium

Of the three species examined, P. radiculosa was the best for the study of the young antheridium, as most of the stages of development were found in the material.

Fig. 10. Pallavicinia radiculosa.

Early stages in the development of the antheridium.

Longitudinal sections. D, F, are cut in a plane at right angles to E, G.

The young antheridium arises very near the apex of the thallus, as a single large cell, projecting from the side of the midrib (Fig. 10, A). The first division wall is transverse, and separates a basal cell which takes no part in the development of the antheridium itself, and an outer cell, which is the real mother-cell of the antheridium (Fig. 10, B). At about the same time that this transverse division is formed in the young antheridium, certain neighboring superficial cells of the midrib become evident, which later form special structures accompanying the antheridia. Some of these "companion cells" secrete the mucilage which bathes the young antheridia; while others, not always readily distinguishable from the earliest stages of the antheridia themselves, finally develop into the characteristic scales covering the older antheridia, somewhat as described by Campbell [1] for Aneura.

The divisions in the antheridium show some variation. Of the two cells formed by the first transverse wall, the inner one divides by a vertical wall into two cells, which remain sunk in the midrib and usually divide no further (Fig. 10, C-E), and may usually be recognized at the base of the stalk in the fully developed antheridium.

The first wall in the antheridium itself is also transverse. Of the two cells thus formed, the lower gives rise to the stalk of the antheridium, and also to the layer of cells separating the sperm-cells from the stalk of the antheridium. The outermost of the two original cells divides first by a median vertical wall, and each of these cells is next divided by a nearly periclinal wall into two very unequal cells. This wall intersects both the outside wall of the antheridium and the median wall, and is quickly followed by a second similar wall which meets the first one and also intersects the median wall. A cross-section of the antheridium at this stage (Fig. II, A) shows two triangular central cells surrounded by four narrow peripheral ones. The young antheridium at this stage closely resembles that of Porella Bolanderi (Campbell [1], Fig. 52), and differs from the usual type of the Jungermanniales, where, according to Leitgeb ([1], II, p. 44) these two peripheral cells do not extend to the top of the antheridium, and a third peripheral cell is cut off before the separation of the central cell is complete. The appearance of longitudinal sections of the young antheridium, cut respectively in the plane of the first median wall, and at right angles to it, are shown in Fig. 10, D, G, E, H. Cross-sections of similar stages are shown in Fig. 11, A, C. The next divisions take place in the peripheral cells, and in the stalk, the two primary spermatogenous cells remaining undivided until the stalk is well developed and

the four original peripheral cells of the antheridium have each divided at least once (Fig. 10, G). The first division in the central cells is a transverse one, quickly followed by a vertical division, so that whether in longitudinal- or cross-section, the central cells are arranged quadrant-wise (Figs. 10, H, II, B). The subsequent divisions follow rapidly, but without any

definite succession being evident. There is but little displacement of the original division-walls, so that up to the last divison of the sperm-cells the limits of the earlier divisions can be plainly traced, and the spermatocytes are in irregular blocks marking the early divisions. While within these blocks of cells the nuclei are usually in the same stage of mitosis, each segment of the antheridium may show a different stage of nuclear division. This was also noted by Humphrey in his study of Fossombronia longiseta [1].

Spermatogenesis

Pallavicinia Zollingeri proved the best species for a study of spermatogenesis, and the investigation of this subject was mainly devoted to that species. In the earliest stages procurable the final mitosis had taken place, and the two resulting nuclei had begun to assume the elongated form found in the completed spermatozoid (Fig. 12, A, B). The final mitosis is accompanied by the formation of a delicate but perfectly evident division wall separating the pair of spermatocytes. The spermatocytes at this stage closely resemble those of Calycularia radiculosa (Campbell [2], Fig. 7). The length of the young spermatocyte in P. Zollingeri is about 7μ. Fig. 12, A, shows the pair of spermatocytes at this stage. The blepharoplast (bl.) now has the form of a somewhat curved rod at the end of the elongated nucleus. Sometimes the blepharoplasts of the pair of spermatocytes are at the same end, sometimes at opposite ends. With the development of the spermatozoid, the blepharoplast, as usual, elongates rapidly, and becomes a slender curved rod, hooked at the free apex, and following the curve of the elongating nucleus which becomes crescent-shaped, with the anterior end more or less conspicuously attenuated (Fig. 12, D-K). In some of the preparations there was present between the blepharoplast and the anterior part of the sperm-nucleus a thick, rod-shaped body, which perhaps represents the "Nebenkörper" described by Ikeno [1] in Marchantia. The cilia arise from the blepharoplast a short distance back of the apex, and become finally about as long as the body of the free spermatozoid, which is about 16μ in length.

The development of the spermatozoid in P. radiculosa was found to be so much like that of P. Zollingeri that no attempt was made to follow it in detail. However, some of the later stages in the last mitosis of the spermatocyte, which were not seen in P. Zollingeri, were secured in P. radiculosa. The nuclei are so small that difficulty was experienced in determining the number of chromosomes, which is probably eight (see Fig. 12, N-Q). The spermatocytes are separated, as in P. Zollingeri, by a delicate membrane, which is more difficult to demonstrate