distinct nutritive and circulating organs to the free moving ovigerous individual from the rooted polype, and prolong its existence, and it would then cease to have the ancillary character of a nurse to the ova of the fixed individuals, and would assume that of the perfected form of the species; and such, in fact, is the case with the larger gelatinous Radiaries, called Medusæ." Now, in so far as perfection means elaborateness of organisation, it is not, of course, to be denied that the medusa is in advance of the polype; but as regards the selection of the phase to be taken as the typical form of the species, I do not see how we can avoid these conclusions:1st, That the bare-eyed medusoids are really homologues of the parts of reproduction, inasmuch as they pass by a continuous gradation into generative organs of the simplest kind and, 2d, that in so natural an order, the relative position assumed for the puny bare-eyed Medusæ must hold also for their portly brethren of the hood-eyed kind.

My limits prevent me giving anything like a detailed view of the declension referred to from free medusoids to simple tunicated ova attached to the body of the parent; but the following may be noticed as observable links in the series.t In Campanularia dichotoma, the medusoids are no longer free as in C. geniculata; they have also more of the polypeform, and remain during their brief period of life attached to the edge of the horny "ovigerous capsule," characteristic of the zoophyte, and there emit the ova or spermatozoa with which they are charged; after which they wither away like blossoms, to be succeeded by a new expansion. In Campanularia lacerata, the ovarian sac advances to the mouth of the capsule; but, instead of a bell-shaped envelope, becomes invested merely by a thick gelatinous coat. In Sertularia generally, the appendages, with somewhat of the medusoid conformation, mature and discharge their contents while still within the "ovigerous capsule." In Cordylophora, the only medusoid features presented either by the spermatic or the ovarian cyst, are the presence of a central tongue or colu* Parthenogenesis, p. 12.

↑ See several papers by Dr T. S. Wright in previous numbers of this Journal.

mella to represent the proboscidiform mouth, and the existence of phlebenteric canals in its wall. In Hydractinia, we have the columella without the canals; and the cysts of some species of Plumularia and Eudendrium are, if possible, of still simpler structure, the latter containing but a single ovum. The progress of degradation reaches its maximum in the common Hydra, in which the large Medusa of the "Hydra tuba" are represented only by spermatic and ovarian cysts, of the most rudimentary organisation, attached to the exterior of the polype. Closely allied species sometimes differ remarkably in this respect; and even in the same species there may be as great a diversity in the opposite sexes; thus in Laomedea geniculata, the ova are formed in free swimming medusoids; the spermatozoa in simple cysts permanently attached. Variations of the same kind occur also in the allied order of Physograda. Such variations, though perplexing to the systematic zoologist, are especially valuable to the physiologist, as indicating the true relations of the forms which occur in dimorphous species; and I think they are fully sufficient to bear us out in the conclusion, that both the bare-eyed and the hood-eyed Medusa are to be considered as gamomorphic zooids, and the polype stock from which they sprang as the typical form in each case. In the one, the orthomorphic form is, as usual, the most conspicuous phase of the species, while in the other it is quite eclipsed by the resulting gamomorphic zooid, which is really a part of itself— a detached and overgrown organ of its own system.

As parallel cases, I would refer to the relation subsisting between the solitary and catenated Salpa, which, as described first by Chamisso, may be regarded as the original basis of the doctrine of alternation, and of which Mr Huxley has since given us a most lucid and philosophical account-to the detachment of reproductive zooids, made up of caudal segments, budded off from some Annelida, as described by MM. Edwards and Quatrefages-and to the derivation from the "Tania-head" of the proglottides, or cucurbitiform segments of the "body" of the tape-worm.

In the vegetable kingdom we have also a very parallel case in

the reproduction of the ferns. In these plants, it is well known the sexual elements are not formed in connection with the conspicuous vegetative stem, but in minute derivative phytoids, termed prothallia, which are produced by the germination of the spores. The prothallia bear antheridia and archegonia; and the embryo, formed on impregnation from the central cell of one of the latter structures, grows up from the prothallium, which comes to have very much the appearance of the seed-leaf of the young shoot. The prothallia I regard as gamomorphic phytoids, parallel to the medusiform reproductive zooids of the Polypifera. Hence I feel obliged to dissent from the parallelism which Hoffmeister would establish between the reproductive process in ferns and mosses. This great authority regards as equivalent structures the prothallium of the former and the leafy axis of the latter, on the ground of their being the parts which bear the sexual organ; and argues from this a corresponding relation between the frondiferous stem of the fern and the seta and capsule of the moss, as the immediate products of impregnation in the two cases respectively. A comparison of objects of such primâ facie diversity-objects more unlike than even the large Medusa and the ovarian cysts of the Hydra-ought not, I conceive, to be adopted, except on the most convincing evidence. But there is no such cogency in this case, on the admission of the general view which has now been advanced; for we have a readier solution of the difficulty, in assuming that the interpolation of an intermediate form occurs at a later stage of the genetic cycle in ferns than in mosses. For this view we have ample warrant in the analogy of the animal kingdom, where we find corresponding differences between the Cestoid and Trematode Entozoa, and between the latter and the Polypifera, among which, indeed, even nearly allied species differ in this matter of the interpolation of gemmation. Such a view, I submit, is a less tax on our powers of conception, than to regard the minute and fugitive capsule of the moss as the equivalent of the perennial and towering stem of the tree-fern. And it is to be borne in mind that the difference here is not merely one of a primá facie kind. In some respects it increases the more we con

template it; for it is clear, as Mr Jenner observes,* that the persistent character of the leafy axis of the moss, and its yielding, in perennial species, many successive sets of sporiferous capsules, assimilates it, quite independently of structural features, rather to the stem of the fern than to its prothallium, which is an organ even more evanescent than the capsule of the moss, its existence terminating when the embryo formed in it has begun to germinate.

§ 6. Orthomorphic Alternation.

In the intermediate period of the life-history of the species, that here termed orthomorphic, which intervenes between the appearance of the general typical character of the family and the maturation of sexual organs, gemmation, though perhaps a more frequent character than either in the incipient or terminal stages, rarely comes before us as a case of alternation of generations, in consequence of the gemmæ commonly remaining in adhesion to each other, so that their separate individuality is lost, and the whole aggregation passes as a single plant or animal. This is especially the characteristic arrangement in the vegetable kingdom, and in Polypifera and Polyzoa among animals. Where the gemmæ do become detached, however, the case may assume the aspect of a form of alternation, as we see strikingly exhibited in the propagation of the Aphides.

I may here briefly explain why I am disposed to refer the alternation of the Aphides to the commencement of this stage, rather than to the protomorphic. It is because the organisation has already acquired that partially advanced development characteristic of the larvæ of other insects, before the process of gemmation comes into play. We cannot say here that the primary product of impregnation buds off a set of embryos of a higher organisation; it is rather a larva-that is, a

* Edin. New Phil. Jour., New Series, Vol. III., p. 269. The occasional conversion of the fruit of the moss into a leafy shoot has been thought to indicate its analogy to the stem of the fern. Is it not rather a viviparous inflorescence, such as occurs at times in the higher plants?

naked embryo-already so far organised on the insect type, that buds off a series of similar larvæ, the last only of which become perfect insects.

§ 7. Resumé of the Varieties of Alternation.

On the grounds above stated, it becomes necessary to distinguish these three varieties of the so-called alternation of generations, that is, of the alternation of gemmation with sexual reproduction:

1. That in which the gemmation occurs in the protomorphic or germinal stage, prior to the appearance of the typical organisation;

2. That in which it occurs in the gamomorphic or later stage of the life-history,—that is, in connection with the maturation of the reproductive organs; and,

3. That in which it occurs in the orthomorphic or intermediate stage,—that is, during the manifestation of a more fully developed condition of the typical organisation, but prior to the maturation of the sexual organs.

The contrast lies principally between the two former varieties. They cannot, indeed, be identified or confounded, as they are by many authors, without losing sight of two important points of difference ;

1. In the budding-stock, which in gamomorphic alternation has both a higher organisation and a greater permanence of life than are possessed by the protomorphic zooid or germparent of the typical form;

2. In the off-sets or concluding links of the respective series, which have really nothing in common but the single point of sexual completeness, the medusoids, prothallia, and other gamomorphic forms being generally of the most rudimentary structure.

Hence, in the one case, we speak of the typical organism and its germ-like matrix; in the other, of the typical organism and its sexual offset. Both the matrix and the offset may assume, indeed, the form of independent beings, but their life is always transitory and provisional, having reference to