The few species of diatoms first discovered were included by Lyngbye, Dillwyn, and others in the genus Conferva. In 1824, the species, increased to forty-eight, were separated by Agardh into eight genera distinguished partly by their mode of growth. But little change was made until Heiberg, in 1863, advocated the division into symmetrical and asymmetrical forms. Without entering upon a general review of the later classifications, including Pfitzer's and Petit's divisions according to the number and location of the chromatophores, or the arrangement of Prof. H. L. Smith, because of the presence or absence of a raphe, or that of Mereschkowsky into motile and immotile forms, the modification of all of these methods by Schuett is here adopted, varied in accordance with certain monographs which appear to offer advantage.
It is customary, especially among writers who are familiar with other classes of plants, to decry any classification of diatoms according to the markings of their siliceous envelopes. As, however, one of the chief distinctions of the class is the possession of a more or less siliceous and indestructible frustule, and as the cell and its contents are never seen except within the valves, their variety forms the only available method of identification. The cell contents, owing to the difficulty of observing their living condition, their continued change, their lack of distinct variation and their entire absence in fossil forms, render their consideration as a complete method of classification an impossibility. If, however, the cell contents can be brought into relation with the markings of their siliceous envelope, it will be a consummation for which the future student of these complicated forms ought to be grateful. That this result is one to be expected may be inferred from the fact that the arrangement of protoplasmic masses in the interior of the cell is coincident in some cases with markings on the valve, and the character of the endochrome is assuming a certain value in accentuating the difference between such forms as Pleurosigma and Gyrosigma, or in the resemblance between Hantzschia and Nitzschia, or between Surirella and Campylodiscus. Mereschkowsky, however, states that it is necessary to be careful in "establishing the relationship between diatoms based on the resemblance of their chromatophores," and further observes that in Hantzschia amphioxys, Scoliotropis latestriata and Achnanthes brevipes, three widely separated forms, the chromatophores are essentially the same.
In one of the earliest classifications of diatoms, the individual cell received less consideration than the nature of the filament or thallus in which many species occur in the first stages of their growth. Those, however, which exist in colonies at first are, sooner or later, broken up into separate frustules, either before or at the time of their maturity or previous to conjugation, while very many species are never seen except in a free state. The union of frustules, therefore, is of secondary importance and the group must be considered as filamentous or unicellular algæ. Their relation to other algæ is not well determined. Among the Desmidiaceæ, a family of the order Conjugales, of the class Chlorophyceæ, the cells are in many forms divided by a constriction into symmetrical halves. The Conjugales are starch forming, with walls of cellulose. In the Diatomaceæ the starch is replaced by oil globules, while the walls of cellulose are more or less filled with a deposit of silica. The Conjugales, however, reproduce by zygospores and usually contain pyrenoids, as may be seen in the parietal chromatophores of Spirogyra. In the class Heterokontæ we have the reserve material in the form of oil, instead of starch, but there are no pyrenoids. To this class belongs the order Confervaceæ, in which the cells are unicellular or filamentous, and to which all of the Diatomaceæ were referred. While, therefore, Diatomaceæ have a close affinity to the Desmidiaceæ and to the Confervaceæ, the determination of their origin, one from another, or from a common ancestral type, appears to be a matter of conjecture.
MORPHOLOGY AND DEVELOPMENT
THE CELL
The cell membrane is composed of two usually equal parts, each of which consists of a valve and a girdle or zone formed of cellulose modified by silica deposited in an insoluble state from a very dilute aqueous solution. The valves are more siliceous and robust than the girdle. Both are in most species easily separable, or at least the bands of the girdle which may be more or less closely fastened to the valves have a motion over each other permitting the cell to enlarge at pleasure. The longitudinal diameter of the cell, or the distance between the centres of the two valves, will vary according to the convexity of the valve and the age of the frustule which may be often determined by the width or number of the girdle bands. These, owing to their diversity of form and arrangement, will be further described under the generic diagnoses.
The siliceous cell-wall is covered on the outside by a layer of protoplasm called the coleoderm. This layer may be quite thin and evident only when treated with fuchsin or Bismarck brown, or it may be of considerable thickness. The cell contains the cytoplasma, protoplasm, cell-sap, endochrome, pyrenoids, oil globules and nucleus, together with certain other less understood bodies.
The Cytoplasma is a thin skin of colorless plasma covering the entire inner surface of the cell. It is invisible in the living cell but is evident in plasmolysis. In long forms it is thickened at the ends and is condensed at the plasma bridge which frequently connects the two valves and divides the cell into two parts, each containing more or less protoplasm surrounding the vacuole in which are found the cell-sap and certain granules. In some forms, as Meloseira, the cytoplasma includes the entire mass of protoplasm.
The Endochrome is seen in the form of one or more bands or plates, of a yellowish or brownish color, on the inner side of the valves or connective zone, or in granules or irregular masses, more or less numerous, on the inner walls, or sometimes grouped near the centre. It consists of a mixture of chlorophyll and diatomine which differ in their relative solubility in alcohol and in their spectroscopic analyses. The color varies from green to a chocolate brown in proportion to the amount of diatomine. So far as the function of the endochrome is concerned it does not appear to differ from that of ordinary chlorophyll, absorbing, under the influence of light, the carbon, and disengaging the oxygen of the carbonic anhydride in the water. Diatoms do not live in absolutely pure or non-aërated water. The individual plates or granules of the endochrome are called chromatophores. Their number and significance will be referred to in the description of genera.
The Pyrenoids.—In the chromatophores of many species are found colorless, homogeneous bodies, strongly refractive, of various shapes, usually lenticular or fusiform, which are known as Pyrenoids (Schmitz). They are scarcely evident in the living cell, but are distinguished by the action of hæmatoxylin and other reagents. Flat forms occur in Surirella and Pleurosigma, lens forms in Pinnularia, Stauroneis, Synedra, Fragilaria and Nitzschia, while a spherical form is found in Cymbella cuspidata. The pyrenoids are always imbedded in the chromatophore. Their growth is by division. Schmitz considers them a part of the living chromatophore, and their substance as working material which in excess has become resolved into the nature of a crystal which its form sometimes resembles. Comparisons are made between them and crystalloids found in certain monocotyledons. The pyrenoid is evidently concerned in the formation of the chromatophore, or in its division. Much of the conjecture, however, is due to the behavior of pyrenoids in other plants.