34 GENERAL HISTORY OF THE INFUSORIA. 



the latter into a right and a left with reference to the frustule when lying on 

 a primary side. The left side is often concave, and the right convex ; mostly, 

 however, the two are alike. As a general rule the primary sides correspond 

 with the lateral surfaces, and the secondary with the dorsum and venter in 

 the terminology of Ehi'enberg. Mr. Ralfs, in his papers in the Ann. Nat. 

 Hist., used the simple terms '/>'on< view^ and ^ side vieiu,' corresponding 

 respectively T^ith Kiitzing's names primary and secondary sides. The Rev. 

 W. Smith adopts this nomenclature as the most convenient for the English 

 student, and uses the term 'front vieiu' to denote the aspect of the frustule 

 when the valvular suture (connecting membrane), or the line along which 

 self-division takes place, is turned towards the observer, and the term * side 

 vieiu^ when the centre of one valve is dii-ected to the eye. He adds — 

 " Even these terms will requii'e modification when applied to some of the 

 more complex and irregular forms ; but in general theii' meaning will be 

 sufficiently obvious." 



It must happen, therefore, from this terminology, that at times a front 

 view cannot be said to exist, viz. when the connecting membrane is obsolete 

 and the opposed valves are closely applied to each other, a suture alone in- 

 dicating the line of junction. 



In size the Diatomea) vary very greatly ; some individual frustules are 

 cognizable by the naked eye, whilst others require the highest powers of 

 the microscope to display them. Even among specimens of the same species 

 great diversity of size prevails, — a peculiarity much determined by the cii'- 

 cumstances surroimding a frustule at the period of its development, and 

 afterwards perpetuated through a long series of individuals multiplied by 

 self-fission. 



The Diatomese exist under three chief forms, as — 1. Single isolated free 

 frustules. 2. Frustules attached by a stalk, stij^es, or pedicle. 3. Frustules 

 coherent in chains, or aggregated together in ramose tufts by an interposed 

 gelatinous substance. The third form is the consequence of incomplete fission, 

 or of imperfect separation after fission. Incomplete fission and consequent 

 concatenation are observed in Bacillaria, Meridion, Himantidi^mi, Melosira, 

 Odontidium, Striatella, Fragilaria, &c. (IX. 131, 167, 171, 175, 176, 177) ; 

 and the form of the chain or filament produced -^-ill be determined by the 

 figiu^e of the indi^-idual frustules composing it. For instance, if these be 

 rectangular, then the resulting chain (IX. 171, 172, 176 ; XIV. 2, 4, 6, la) 

 is straight, but if wedge-shaped, it is cun-ed or spiral (IX. 177, 179 ; XII. 

 21). The extent and degree of attachment of the adjoining frustules difier 

 in difierent genera ; thus, in Bacillaria it is very shght, and readily yields, 

 allowing one segment to glide on another, or to separate from it, except at 

 one point, yet at the same time possessing the power of recovering itself. In 

 Odontidium, Himantidium, Denticida, and Meridion (IX. 177), the mutual 

 adhesion of the several segments is stronger ; and after the opposed sui-faces 

 have been separated, futiu^e adhesion is not effected. In Fragilaria, the ad- 

 herence is more tenacious. In Diatoma, Tahellaria, Ghr(mmatoj)7iora, Am- 

 phitetras, &e. (II. 46 ; XI. 22, 52 ; XIV. 23), the frustules hang together by 

 a sort of hinge inserted between adjoining angles in a zigzag fashion. In 

 IstJimia, this hinge or connecting link attains a greater magnitude, and, in 

 fact, is double. In Podosira (II. 45) and in some species of Melosira a 

 junction-process is developed from the centre of each frustule in the chain. 

 In other Melosirce and in OrtJiosira, the junction-siufaces are toothed (den- 

 tate), and thus hold the adjoining frustules in firm union. In the instance 

 of Biddidpliia (II. 46, 48), the siufaces in union are curiously elongated at 

 the angles into rounded or horn-like processes, whilst theii' convexity is 



