is reached the development is different in different sub-groups. In Batrachospermum filaments arise from the carpogonium on all sides; in Chantransia and Scinaia on one side only; in Helminthora the filaments are enclosed in a dense mucilage; in Nemalion, prior to the formation of the filaments, a sterile segment is cut off below. In all these cases, however, the end- cells of the filaments each give rise to a carpospore, and the aggregate of such sporiferous filaments is a cystocarp. Again, in the family of the Gelidiacea, the single filament arising from the carpogonium grows back into the tissue and preys upon the cells of the axis and larger branches, after which the end-cells give rise to carpospores and a diffused eystocarp is formed. In the whole group of the Cryptonemiales the parasitism becomes more marked still. The filaments arising from the carpogonia grow into long thin tubes, which fuse with special cells rich in protoplasm contents; and from these points issue isolated tufts of sporogenous filaments, several of which may form the product of one fertilized female cell. In Naccaria, one of the Gelidiacea, it is observable that the ooblastema filamient, as the tube arising from the fertilized carpogonium has been called, fuses completely with a cell contiguous to the carpogonium before giving rise to This is also the case the foraging filaments already referred to. among Cryptonemiales. In a whole series of Red Alge, the existence of a highly-specialized auxiliary cell in the neighbour- hood of the carpogonium is a characteristic feature. In the Gigartinales it is already differentiated previous to fertilization; in Rhodymeniales it arises subsequent to fertilization. In the Gigartinales, the filaments which arise from the auxiliary cell may spread and give rise to isolated tufts of sporogenous filaments, as in the Cryptonemiales. In the Rhodymeniales a single tuft arises directly from the auxiliary cell. The carpo- spores are in all cases bright red naked masses of protoplasm when first discharged. They soon acquire a cell-wall, and germinate without a period of rest. When the cystocarps or segments of cystocarps are formed in the substance of a thallus, the site is marked merely by a swelling of the substance. When the cystocarp is produced externally, it may form a berry-like mass without an envelope, in which case it is known as a favella. In Rhodomelaccce there is a special urn-shaped envelope sur- rounding the sporogenous filaments. This is a ceramidium. The attachment of the cell of an ooblastema filament to a cell of the thallus may be effected by means of a minute pore, or the two cells may fuse their contents into one protoplasmic mass. In the latter case, and especially where the union is with a special auxiliary cell, it is of importance to know what happens to the nuclei of the fusing cells. Schmitz was of opinion that in the cases of open union there occurred a fusion of nuclei similar to that which occurs in the sexual union of two cells. He founded his generalization to a large extent upon the observation that in Glæosiphonia capillaris two cells completely fuse, and that only one nucleus can be detected in the fused mass.
Oltmanns has recently reinvestigated the pheno- mena in this plant, among others, and has shown that the nucleus of the cell which is being preyed upon recedes to the wall and gradually atrophies. The nucleus of the ooblastema filament domi- nates the mass, and from it all the nuclei of the carpospores are thus derived. There thus seems to be no justification for believing, as Schmitz taught, that a second sexual act occurs in the life-cycle of these Floridece. The Bangiales are a relatively small group of Red Alge, to which much of the description now given does not apply. Structurally they are either a plate of cells, as in Porphyra, or fila- ments, as in Bangia. There is no exclusive apical growth, and the cells divide in all directions. The characteristic pit is also absent. Sexual and asexual reproduction prevail. The male cell is a spermatium, but the female cell bears no such receptive trichogyne as occurs in other Rhodo- phycece. After fertilization the equivalent of the oospore divides directly to form a group of carpospores. There is thus a certain resemblance to Eufloridece, but sufficient difference to necessitate their being grouped apart. Fertilization by means of non-motile spermatia and a trichogyne are known among the Fungi in the families Collemaceae and Laboulbeniacea. After this survey of the four groups comprised under Algæ, it is easier to indicate the variations in the limits
Limits of the Algæ. of the class as defined by different authorities. To consider the Cyanophyceae first, either the marked contrast in the method of nutrition of the generally colourless Bacteriaceae to that of the blue- green Cyanophycea is regarded as sufficient ground
for excluding Bacteriacea from Algæ altogether, notwith- standing their acknowledged morphological affinity with Cyanophyceae, or, in recognition of the incongruity of effecting such a separation, the whole group of the Schizophyta-that is to say, the Cyanophyceae in the narrow sense, together with Bacteriaceae, is included or excluded together. Again, while Conjugato may be shut out from Chlorophyeece as an independent group co-ordinate with them in rank, the Characece constitute so aberrant a group that it has even been proposed to raise them as Charophyta to the dignity of a main division co-ordinate with Thallophyta. Similarly, while Diatomacea may be excluded from among Pheophycea, though retained among Algæ, the Cryptomonadaceae and Peridiniaceae, like Euglena and other Chlorophycea, may be excluded from Thallophyta and ranged among the Flagel- late Protozoa. (See PROTOZOA.) It is doubtful, however, whether the conventional distinction between plants and animals will continue to be urged; and the suggestion of Haeckel that a class Protista should be established to receive the forms exhibiting both animal and plant affinities has much to recommend it on phylogenetic grounds. To adopt a figure, it is probable that the sources from which the two streams of life-animal and vegetable-spring may not be separable by a well-defined watershed at all, but consist of a great level upland, in which the water- ways anastomose. Finally, while Chlorophycece and Phao- phyceae exhibit important affinities, the Rhodophyceae are so distinct that the term Algæ cannot be made to include them, except when used in its widest sense. It has been well said that the attempt to classify plants according to their natural affinities is an attempt to con- struct for them the genealogical tree by which Phylogeny. their relationships can be traced. Algæ are, however, so heterogeneous a class, of which the constituent groups are so inadequately known, that it is at present futile to endeavour thus to exhibit their pedigree. A synoptical representation of the present state of know- ledge would be expressed by a network rather than by a The following table is an adaptation of a scheme devised by Klebs, and indicates the inter-relationships of tree.
the various constituent groups. The area included in the thick boundary line represents Alga in the widest sense in which the term is used, and the four included areas the four main subdivisions. A continuous line indicates a close affinity, and a dotted line a doubtful relationship. In comparing Alge with the great Archegoniate series which has doubtless sprung from them, it is natural to Alternation of generations. inquire to what extent, if any, they present evi- dence of the existence of the marked alternation of generations which dominates the life-history of the higher plants. Turning first to the Rhodophycea, both on account of the high place which
