branch by the production of lateral swellings, which are elongated and then divided off by cellulose partitions. It seems reasonable to suppose that branching in this case simply means that one part of the cell- wall is weaker, less fitted to withstand internal tension, than the rest. If the cell- wall were uniform it would stretch uniformly, and the want of uniformity in the cell-wall may probably be traced to its being part of an aggregate. The string of cells is moored by one end, and any motion of the water will cause it to assume an oblique position. It is then unequally influenced by gravity and by light, and these two forces are known to be competent to modify growth in an unsymmetrical manner under such circumstances. Reasoning of this kind will at any rate suggest the way in which the specialization of individual cells would be a result of their aggregation. With this specialization they would lose an increasing amount of their capacity for in dependent existence. Specialization of function will, of course, be reflected in corresponding morphological modifi cations, and thus a mechanical aggregate of independent cell elements gradually passes into a physiological aggregate
of interdependent ones.One obvious penalty of specialization is the loss of the power of reproduction by individual cells when detached. Amongst plants, however, even the highest types preserve some measure of it. It is sufficient to adduce the common horticultural method of propagating Begonias from fragments of the leaves. Such a mode of reproduction in one of the highest plants is essentially the same as that which exists in Pleurococcus, and although reproduction from fragments of structures so specialized as a leaf is comparatively rare, there are in almost all plants provisions for agamogenesis, which depend upon the retention by fragments of the organism of this capacity for independent existence. And it may even happen that the perpetuation of the race for consider able periods may come to depend upon such a property. One of the most striking instances of this is the distri bution over the British Isles of the American water-weed (Anacharis), of which, nevertheless, only the female plant exists in this country. If we had the male, we should have a good instance (and others might be given) of gamo- genesis and agamogenesis proceeding side by side in in dividuals of the same generation, although probably not without the two processes reacting upon one another. Throughout the vegetable kingdom, however, we find gamo- genesis and agamogenesis occurring in separate generations, which are often extremely different. The two modes of reproduction are then subject to a cyclical arrangement, and the comparison of the forms under which this alter nation of generations occurs in different groups throws a good deal of light upon their taxonomic relations. It will be convenient to use the word "Sporophore " for the agamo- genetic generation, in which special cells (spores) are de tached from the parent to serve as a means of propagation ; while for the gamogenetic generation, in which conjuga tion takes place, or in which special cells (oospheres) are fertilized by antherozoids, and become oospores, "Oopkore" may be employed.
Our present knowledge has rather added to than dimi nished the difficulty of devising even a plausible phylogenetic classification of the vegetable kingdom. Greater success has, however, been attained in establishing the primary and larger secondary groups, so as to allow us to feel some confidence that they are really natural assemblages. The relation of these groups to one another is a problem, the more than approximate solution of which will probably have to be some time postponed. Many of them appear to represent the later developments of simpler types or aggre gates than anything that these groups at present contain. The Equisetacece, for example, are extremely ancient, yet we know of no form, living or extinct, which enables U3 to trace the connection of their very remarkable organiza tion with that of other groups. In the animal kingdom the "recapitulation theory" steps in, and obtains from the study of the development of the organism that kind of information which is wanting amongst plants as to the simpler ancestral forms of the different vegetable types. This information, however, is drawn in the animal king dom to a large extent from the phenomena presented by the differentiation of the sustentative organs of the embryo. For reasons already pointed out, there is nothing analogous to this amongst plants, where the rapid extension of sur face is usually the primary object to be attained. The earliest stages in the development of any plant are subject to conditions so simple that there is little room for special ization, and the economy of nutrition has probably gene rally led to the suppression of recapitulative structural details.
On the other hand, if we derive little help from re capitulation in studying the process of development in plants, from the unicellular stage of spore or oospore, it proves extremely suggestive, if we take in the whole cycle comprised between two processes of gamogenesis, and com pare the relations to one another of the gamogenetic and agamogenetic generations, or using the nomenclature intro duced above, of the Oophore and Sporophore.
Thallophyta.
Endlicher, in 183G, divided the vegetable kingdom into Tliallophyta and Cormophyta; and these divisions still hold good, though it is by no means easy to frame characters which will strictly limit them. With the exception of the absence in the one, and presence in the other, of the " opposition of stem and root," none of the distinctions which Endlicher pointed out are available now. Through out the greater part of the Thallophyta anything like a distinct segmentation of an axis furnished with lateral appendages is altogether wanting. As thus limited by the absence of a clear differentiation of root, stem, and lateral appendages, this sub-kingdom comprises an assemblage of plants, which were divided by Bishop Agardh, in 1821, into the three well-known groups of Algce, Fungi, and Lichens. And to these, for reasons which will be presently pointed out, must be added the hitherto problematical group of Characece.
It has long been seen that, with respect to the three former groups, it was impossible to assign morphological characters which would separate them strictly one from the other. Accordingly, Berkeley and Lindley were compelled to fall back on distinctions of a physiological kind. Algce were defined to be generally aquatic in their mode of life; Fungi and Lichens, on the other hand, were aerial, but the former drew their nutriment from the " substratum," while the latter obtained it from the air.
There are several grounds on which this arrangement appears to need reconsideration.
Contrasting, in the first place, Algoe and Fungi, we now know that the plants belonging to the former group in variably contain chlorophyll, while those belonging to the latter are equally devoid of it. In their morphological aspects, however, the two groups present a remarkable parallelism.
chlorophyll, and the difference in the mode of life which . results, would have greater weight for classificatory pur poses, were we not familiar with instances in other parts of the vegetable kingdom in which it proves to have no value at all. Amongst flowering plants we are acquainted
with many cases where plants closely allied in structure to