MICROSCOPIC FOJRMS OF VEGETABLE LIFE. 293 



as iii many other Diatoms; most assuredly they are not orifices, as sup- 

 posed by Prof. Ehrenberg. Of this type, again, specimens are found 

 presenting 6, 8, 10, or 12 radial divisions, but in other respects exactly 

 similar; on the other hand, two specimens agreeing in their Dumber of 

 divisions may exhibit minute differences of other kinds; in fact, it is 

 rare to find two that are precisely alike. It seems probable, then, that 

 we must allow a considerable latitude of variation in these forms, before 

 attempting to separate any of them as distinct species. Another verv 

 beautiful discoidal Diatom, which occurs in Guano, and is also found 

 attached to Sea-weeds from different parts of the world (especially to a 

 species employed by the Japanese in. making soup), is the Arachnoidis- 

 cus (Plate XL), so named from the resemblance which the beautiful 

 markings on its disk cause it to bear to a Spider's web. According to 

 Mr. Shadbolt, 1 who first carefully examined its structure, each valve con- 

 sists of two layers; the outer one, a thin flexible horny membrane, inde- 

 structible by boiling in nitric acid; the inner one, siliceous. It is the 

 former which has upon it the peculiar spider's web-like markings: whilst 

 it is the latter that forms the supporting frame-work, which bears a very 

 strong resemblance to that of a circular Gothic window. The two can 

 occasionally be separated entire, by first boiling the disks for a consider- 

 able time in nitric acid, and then carefully washing them in distilled 

 water. Even without such separation, however, the distinctness of the 

 two layers can be made out by focussing for each separately under a l-4th or 

 l-5th inch Objective; or by looking at a valve as an opaque object (either 

 by the Parabolic Illuminator, or by the Lieberkuhn, or by a side light) 

 witha4-10ths inch Objective, first from one side, and then from the 

 other. * This family is connected with the succeeding by the small 



froup of Eupodiscece, the members of which agree with the Coscino- 

 iycecB in the general character of their discoid frustules, and with the 

 Biddulphiece in having tubercular processes on their lateral surfaces. In 

 the beautiful Atilacodiscus (Plate I., Fig. 5) these tubercles are situated 

 near the margin, and are connected with bands radiating from the 

 centre; the surface also is frequently inflated in a manner that reminds 

 us of Actinoptychus. These forms are for the most part obtained from 

 Guano. 



292. The members of the next Family Biddulpliiece differ greatly in 

 their general form from the preceding; being remarkable for the great 

 development of the lateral valves, which, instead of being nearly flat or 

 discoidal, so as only to present a thin edge in front view, are so convex 

 or inflated as always to enter largely into the front view, causing the cen- 

 tral zone to appear like a band between them. This band is very narrow 

 when the ne\v frustules are first produced by self-division ( 278); but it 

 increases gradually in breadth, until the ncwfrustule is fully formed and 

 is itself undergoing the same duplicative change. In Biddulphia (Fig. 

 167) the frustules have a quadrilateral form, and x remain coherent by 

 their alternate angles (which are elongated into toothlike projections), so 

 as to form a zigzag chain. They are marked externally by ribbings 

 which seem to be indicative of internal costw partially subdividing the 

 cavity. Nearly allied to this is the beautiful genus Istlimia (Fig. 181), 

 in which the frustules have a trapezoidal form owing to the oblique pro- 



1 " Transact, of Microsc. Society," First Series, Vol. iii , p. 49. 



2 These valves afford admirable objects for showing the * conversion of relief 

 in Nachet's Stereo-Pseudoscopic Microscope (g 38). 



