driven up to the blind end where it passes around the partition, back through the other chamber to the sinus of the parent. It is of course unnecessary to state that when the circulation of the parent is reversed that of the stolon changes also.

By the formation of the partition above described the tube is divided longitudinally into halves, and each half is destined to be converted into the series of zooids on one side of the chain. The outer wall of the tube, which has been shown to be part of the muscular tunic of the parent, becomes the muscular tunics of the zooids; the chambers, which are continuous with the sinus system of the parent, form the body cavities or sinus systems of the zooids, and the central tube, which is a prolongation of the pericardium of the parent, forms the nervous, digestive, and branchial organs of the zooids of the chain. It is probable that the cavity of this inner tube gives rise to lateral diverticula, which form the cavities of the digestive organs and branchial sac of the young, but this point could not be determined with certainty, nor could any connection between the cavity of this inner tube and any of the cavities of the parent be discovered.

Before the tube becomes differentiated into the organs of the zooids, in fact, before there are any indications that the tube is to give rise to the chain, two new organs are formed, one in each of the sinus chambers of the stolon. These new organs are long clubshaped masses of protoplasm, which are not at first attached to the tube, but are free within the chambers, and do not seem to be derived from any of the pre-existing parts of the solitary Salpa, but are formed directly from the blood. As the tube grows these organs lengthen as well, and soon a row of germinative vesicles is seen extending along each of them; they are the ovaries. (Fig. X., x.) At the time that the constrictions, which are the first indications of the zooids, make their appearance on the outer wall of the tube, each ovary is seen to be made up of a single row of eggs, equal in number to the constrictions which indicate the number of the future zooids, and as these latter are developed, and their sinus systems become separated from the common cavity of the tube, the chain of ova divides, so that a single egg passes into the sinus system of each zooid, and becomes suspended there by a gubernaculum, by means of which it is attached to the wall of the branchial sac, as already described.

Since the chain Salpa at birth always contains an unimpregnated ovum, organically connected with its body, and since this egg and the resulting embryo are nourished by the blood of the chain Salpa by means of a placenta, and since no reproductive organs have ever been observed within the body of the solitary Salpa, it seems most reasonable to accept the belief that the solitary Salpa is the asexual, and the chain Salpa the hermaphrodite sexual generation, and that

the developmental history of the genus presents a true example of "alternation of generations." When, however, we have traced backward the history of one of the zooids, which compose a chain, and find that the egg is present at all stages of growth, and is of exactly the same size and appearance as at the time of its impregnation; when we find one organ after another disappearing, until at last we have nothing but a faint trace of a constriction indicating upon the wall of the stolon the position of the future zooid, the conclusion seems to be irresistible that the animal, which has as yet no existence, cannot be the parent of the egg which is already fully formed.

The life history of Salpa may then be stated in outline as follows: The solitary Salpa is the female, and produces a chain of males by budding, and discharges an egg in the body of each of these before birth. These eggs are impregnated while the zooids of the chain are very small and sexually immature, and develop into females which give rise to other males in the same way.

After the foetus has been discharged from the body of the male the latter attains its full size, becomes sexually mature, and discharges its spermatic fluid into the water to gain access to the eggs carried by other immature chains.

The fact that impregnation takes place, not, as we might expect, within the body of the solitary, but within that of the chain Salpa, is no objection to this view, for the number of animals whose eggs are fertilized within the body of the female is quite small, and in at least one genus, Hippocampus, the eggs are received into a specialized. brood-sac in the male, and are there impregnated.

We can also find analogy for the singular fact that the eggs always develop females, while the males are formed by budding. The fertilized eggs of the bee always give rise to females, while the males are developed by the virgin bee, through what seems, as pointed out by Professor McCrady, to be most properly regarded as a process of internal gemmation; and we cannot fail to mark the very striking parallelism between the process of reproduction as manifested in Salpa and the bee.

The fertilization of the eggs within the bodies of zooids produced by budding from the body of that whose ovary gave rise to the eggs is not unusual among the Tunicata. The zooids of most of the Tunicata are hermaphrodite, and develop eggs of their own, but, at least in the case of Pyrosoma, Perophora, Didemnium, and Amauricium, the egg which undergoes impregnation and development within the body of the zooid is derived, not from its own ovary, but from that of the generation before, and the eggs produced in the body of the second generation must pass into the bodies of the zooids of the third generation before they can be fertilized. The essential difference between this process and that presented by Salpa, is that in

the

Salpa the sexes are distinct, and as the chain Salpa has no ovary process of budding stops with the second generation; while as the zooids of the other Tunicata are hermaphrodite the process may go on indefinitely.

The history of Salpa is of especial interest, as it throws a great deal of light upon the manner in which separation of the sexes may be brought about in forms which were originally hermaphrodite, and it is also interesting to note that the elæoblast, the history of the development of which shows it to be the homologue in the female of the testicle of the male, is concerned in reproduction, although it has lost all the characteristics of a sexual organ, and is simply a supply of food.

We cannot fail to notice the connection between the manner in which the male Salpa is produced, and the numerous cases, through the various groups of the animal kingdom, in which the male is, to some extent, parasitic upon, or supplemental to, the female.

The Cirrhipeds, Arachnids and the Argonaut, will at once suggest themselves, as familiar instances of the occurrence of such a relation between the sexes.

These interesting theoretical points are simply mentioned here, as a more exhaustive discussion of them is reserved for another place.-A Paper read before the Boston Society of Natural History.

EXPLANATION OF PLATE CXLIV.

The small letters have the same signification throughout.

k. Atrial aperture.

7. Cavity of atrial chamber.
m. Muscles.

n. Ganglion.

o. Nucleus.

p. Esophagus.

s. Stomach.

t. Intestines.

u. Elæoblast.

v. Pericardium.

w. Inner tube of stolon.
x. Ovary.

FIG. I.-Egg within the sinus system, and attached by a gubernaculum to wall of branchial sac, within the cavity of which a few spermatic filaments are

seen.

FIGS. II., III., IV., and V.-Successive stages of segmentation.

FIG. VI.-Gastrula within the brood-sac.

FIG. VII.-Embryo, soon after the primitive digestive cavity has become divided into the branchial and placental chambers.

FIG. VIII.-Embryo considerably advanced, showing the mid-atrium, and one of the lateral atria m, which has already begun to split and form the muscles. FIG. IX.-Embryo at about the time that the stolon appears.

FIG. X.-Stolon, at a very early stage, showing the ovaries x, x; [in this Figure the letters a and b were accidentally transposed, so that b represents the outer tunic, and a the branchial sac].