Conjugation, the formation of auxospores, and the actual process of cell division are seldom seen, as they occur during the night or at least in darkness. It is advisable in order to observe reduplication to obtain the material about midnight and place it in very dilute alcohol. In filamentous forms, however, the cell division is easily observed at any time in its various stages. By immersing in picric acid (saturated solution), transferring to very dilute alcohol which is gradually increased in strength, and then passing through oil of cloves and finally to the mounting medium, excellent preparations can be made. By staining with gold chloride alone the nucleus is made apparent without further treatment. EVOLUTION OF FORMS It may be assumed that diatoms originated in the sea; to deny this requires evidence of the existence of fresh-water species previous to the Miocene period which is entirely marine. In those subject to fluctuations of the waves, as pelagic diatoms, their existence appears to be contingent upon the methods by which the separate frustules can cohere. Various devices, including hooks, spiral bundles, horns and processes exuding threads of plasma, exist for holding together the frustules. When marine forms are found in quiet waters some of these devices, being no longer of any value, cease to grow, although free swimming diatoms are rare. They either occur in long chains or are stipitate or sessile. If it is further assumed that the fresh-water diatoms are found in greater abundance in later periods, the action of running streams makes necessary the provision of some means by which the species may continue to colonize. This may be recognized in the occurrence of linear forms chiefly in streams. Circular forms, such as Cyclotella which have no raphe, are found in quiet waters, such as pools or ditches, and never exist living in running streams. Those forms only would be able to live in water having a more or less swift current under one of three conditions: they must, as in Gomphonema, be adherent to surrounding objects by a stipe; or be enclosed in a gelatinous tube, as in Homoeocladia; or have an independent motion powerful enough to overcome the influence of the current. It is true that many forms with a raphe have no apparent motion. In the case of Mastogloia provision is made in a gelatinous cushion in which the frustules are preserved. In Cocconeis, with a true raphe in one valve only, in Epithemia, with a partial raphe, or in certain Eunotiæ with a trace of one, we find species evidently degenerate and parasitic. The long Synedræ, having only a median line, live in running streams, since they are attached at one end to other algae. Forms with a true raphe appear to be more highly developed, since they are able to seek locations favorable to growth. Given, therefore, the structure of the valve, the habitat may be inferred. {12}THE MOTION OF DIATOMS