also frequently found on the surface of uncleaned shells.[32] From the few examples given, a faint idea only can be formed of the variety and beauty of the engravings on the diatoms. It had long been doubted whether those on the valves of Coscinodiscus, Triceratium and others, were elevations or depressions, but Professor Rood of New York, United States, has proved them to be depressions by an optical arrangement which will be useful for the investigation of microscopic forms. [Illustration: Fig. 14. Actinocyclus undulatus.] Diatoms increase by spontaneous bisection, by conjugation, and by the resolution of their endochrome into minute spores, called gonidia. When bisection is about to take place the cell elongates, the hoop increases in breadth, the endochrome divides into two equal parts, and the coating of the cell bends in between them, which gives the diatom the appearance of an hour-glass. At last they separate, and upon each of the new surfaces a new silicious half is formed, usually the exact counterpart of the old one, so that there are two diatoms instead of one; and the process may be continued indefinitely. In most cases, the new diatoms thus produced are free and independent. Sometimes, however, they adhere to one another by a fragment or connecting membrane, and if they happen to be slender and rectangular, and attached side by side, they form a slender filament, or if attached by alternate angles they form a zigzag chain, as in fig. 11. The Meridion circulare (fig. 15) is a diatom of exquisite beauty, millions and millions of which cover every submerged stone, twig, or blade of grass, and even form the mud at the bottom of the streams at West Point, in the United States of North America. Its frustule or single diatom is long, slender, and rectilinear, but being broader at one end than at the other, by continued bisection and adhering to one another they form a circular, spiral, or flattened helical screw of several turns. The individual frustules of some marine diatoms have a precisely similar form, being rectilinear and broader at one end than the other, but each frustule is attached by its narrow end to the extremity of branching cellulose stems fixed to sea-weeds or stones, and by a continuous subdivision of which the stem does not partake, they are spread out at their free ends like a fan. [Illustration: Fig. 15. Meridion circulare.]