SILICEOUS ENVELOPE OF DIATOMACE^. 279 



one to another, after the process of duplicative subdivision has 

 once been completed; nor being in any V7ay connected, either 

 by a stipes, or by a gelatinous investment. This is the case, for 

 example, with Triceratium (Fig. 79), Pleurosigma (Fig. 80), Acti- 

 nocydus (Figs. 84, 101, b b), Heliopelta (Fig. 85), Arachnoidiscus 

 (Fig. 86), Campylodiscus (Fig. 87), Surirella (Fig. 88), Coscinodiseus 

 (Fig. 101, a, a, a), and many others. The discoid forms, how- 

 ever, when obtained in their living state, are commonly found 

 cohering to the surface of seaweeds. 



174. We have now to examine more minutely into the curious 

 structure of the siliceous envelope, which constitutes the charac- 

 teristic feature of the Diatomacese, and the presence of which 

 imparts a peculiar interest to the group, not merely on account 

 of the elaborately marked pattern which it often exhibits, but 

 also through the perpetuation of the minutest details of that 

 pattern, in the specimens obtained from fossilized deposits (Figs. 

 101, 102). The siliceous envelope of every Diatomaceous "frus- 

 tule" or cell, consists of two valves or plates, usually of the most 

 perfect symmetry, closely applied to each other, like the two 

 valves of a Mussel or other bivalve shell, along a line of fracture 

 or suture; and each valve being more or less concavo-convex, a 

 cavity is left between the two, which is occupied by the cell-con- 

 tents. The form of this cavity, however, differs very greatly ; 

 for sometimes each valve is hemispherical, so that the cavity is 

 globular; sometimes it is a smaller segment of a sphere, re- 

 sembling a watch-glass, so that the cavity is lenticular ; some- 

 times the central portion is completely flattened, and the sides 

 abruptly turned up, so that the valve resembles the cover of a 

 pill-box, in which case the cavity will be cylindrical; and these 

 and other varieties may coexist with any modifications of the 

 contour of the valves, which may be square, triangular (Fig.' 79), 

 heart-shaped (Fig. 87), boat-shaped (Fig. 88, a), or very much 

 elongated (Fig. 80, a), and may be furnished (though this is rare 

 among the Diatomacese) with projecting outgrowths (Fig. 81). 

 In all instances, the frustule is considered to present its "front" 

 view, when its suture is turned towards the eye, as in Fig. 88, b, 

 c ; whilst its " side" view is seen, when the centre of either valve 

 is directly beneath the eye (a). Although the two valves meet 

 along the suture, in those newly formed frustules which have 

 been just produced by binary subdivision (as shown in Fig. 81, 

 A, e), yet as soon as they begin to undergo any increase, the 

 valves separate from one another, and the cell-membrane which 

 is thus left exposed, immediately becomes consolidated by silex, 

 and thus forms a sort of hoop that intervenes between the valves 

 (as seen at b) ; this hoop becomes broader and broader with the 

 increase of the cell in length ; and it sometimes attains a very 

 considerable width (Fig. 81, a, b). As growth and self-division 

 are continually going on when the frustules are in a healthy 

 vigorous condition, it is rare to find a specimen in which the 

 valves are not in some degree separated by the interposition of 



