BACILLARIALES or DIATOMS. The diatoms are micro scopic unicellular or colonial plants belonging to the Algae, and are specially distinguished by the complex structure of their cell walls which are usually strongly impregnated with silica. The majority are exceedingly minute, and one with a length of in. (o• I 2 5mm.) is well beyond the medium size. The first forms were discovered by O. F. Muller towards the close of the 18th century. With the perfection of modern microscopes and micro scopic methods, the study of these minute plants has been prose cuted with great vigour and over 15,000 species of diatoms have been described, while about 1,200 species are known to occur in the fresh waters and on the coasts of Great Britain and Ireland.
Many diatoms are free-floating, but some (Gomplionema) are attached to foreign bodies by simple or branched gelatinous stalks. The frustules of some are connected to form ribbon-shaped col onies (Fragilaria) or zigzag chains (fig. 3 D). In a few genera numerous individuals are enclosed in a tube-like gelatinous envelope.
The valves are thin and transparent and generally ornamented with a variety of markings. The latter often appear as fine lines Reproduction.—The ordinary mode of multiplication is by cell-division, which usually occurs at night. The protoplasm divides into two masses and the halves of the cell commence to separate; thereupon fresh valves, which are at first very delicate, are secreted on the surfaces of the new protoplasmic masses op posite to the valves of the parent. The new valves are situated within the girdle of the original frustule, but as their own con necting bands develop, those of the parent separate. Each of the new individuals thus has one valve derived from the parent, and one that is newly formed and more or less parallel to it. This process of division is repeated at frequent intervals, and it has been calculated that from a single frustule i,000,000,000 new indi viduals may arise in the course of a month.
The individual diatom, owing to the rigidity of the siliceous wall, is ordinarily incapable of any increase in length and, since (striae) on the face of the valve, and the best lenses show them in many cases to consist of a series of separate dots (punctae) which correspond to small cavities within the siliceous wall. The valves of certain marine genera exhibit a beautiful areolated structure due to the presence of larger chambers within the wall (fig. I). A good many diatoms, especially those of the navicu loid type, have thickenings at the centre and at each extremity of the valve, known as nodules (fig. 2) . These possess a com plex structure and are generally connected by a long narrow cleft, the raphe (fig. 2), which appears as a straight or slightly un dulating line in the valve-view. Some diatoms, while possessing nodules, lack a raphe, although in these cases the narrow area be tween the two rows of striae often simulates one (pseudoraphe).
The contents of diatom cells are very similar to those of other Algae. Beneath the wall there is a lining protoplasmic layer en closing the cell-sap, and connected either by two broad bands or by a number of anastomosing strands with the central proto plasm in which the nucleus is embedded. The chromatophores are either one or two of large size, or numerous small lobed discs (fig. I). The former often include a variable number of glis tening pyrenoids. The characteristic brown colour is due to dia tomin, a pigment resembling that of the brown Algae and masking the chlorophyll. The cells often contain abundant oil.
Many diatoms, but only those which possess a true raphe in their valves (Navicula, etc.), are able to move through the water, sometimes with considerable rapidity. The movements are doubt less in all cases related to the presence of the raphe, which in all probability contains streaming protoplasm connected with that in the interior of the cell by a complex system of canals within the nodules. Some believe that the movements depend upon an extrusion of mucilage from the region of the raphe. In Cocconeis, whose disc-shaped frustules are found attached in enormous num bers to Cladopliora and other freshwater plants, only the valve in contact with the substratum possesses a raphe.
the new valves are always formed within the girdle of the old ones, it would follow that one individual of every succeeding gen eration is reduced in length by the thickness of the girdle. This is not, however, strictly true, as daughter cells are sometimes formed which are larger than the parent cell, so that the reduc tion in size is not always proportional to the number of divisions.
It seems that often, when the diminution in size has reached a limit, the maximum size is regained by the formation of auxo spores, but other conditions than this no doubt also lead to auxo spore production. Auxospores in many cases arise as a result of the fusion of two individuals; prior to this conjugation, the proto plasmic contents of each may divide into two parts. The two halves of each frustule then separate, and the liberated contents fuse to form one (fig. 4 B) or two (fig. 4 A) auxospores, which are surrounded by the empty valves of the original individuals. In other cases, however, auxospores are formed merely from the contents of a single individual without any process of fusion (fig. 4 C) , and this is invariably their mode of formation in cen tric diatoms (fig. I). The auxospores usually undergo enlarge ment and then, after secreting new valves, constitute a new indi vidual.
In some marine centric diatoms the protoplasmic contents at times divide to form numerous microspores (fig. 4 D), which are liberated as naked cells, in some cases provided with cilia like the sexual cells (gametes) of Algae. It is not impossible that they represent the sexual cells of these, but the evidence is inconclusive.
The system of classification of diatoms usually followed is one put' forward by Schutt (1896). He separates them into two pri mary divisions, the Centricae and the Pennatae. The former in clude the diatoms in which the valve view possesses a radial sym metry around a central point, and which are destitute of a raphe (or pseudoraphe) (fig. I) . The Pennatae comprise those in which the valve-view is boat- or needle-shaped, with the markings ar ranged on either side of a median line (raphe or pseudoraphe) (fig. 3) . The Centricae are much more abundant in the sea than in fresh waters. It is possible that these two groups are not as nearly related to one another as has hitherto been supposed. In particular the fact that the sexual reproductive process appears quite different in the two cases makes a close affinity doubtful.
Vast deposits of diatomaceous earths have been discovered in various parts of the world, some of freshwater, others of marine origin. That at Richmond in Virginia (U.S.A.), extends for many miles and is in places at least 4of t. deep, whilst in the western states of America beds of 3oof t. thickness have been discovered. Other, though less extensive, deposits are known at Dolgelly (Wales) and at Bilin in Bohemia. It is a remarkable fact that the fossil genera and species are very similar to and in part quite identical with the living representatives of the class.
(F. E. F.)