DEPTHS IN THK NORTH ATLANTIC OCEAN. 207 



the granules of the granule-heaps. The largest discoliths 

 are commonly free, but the smaller and smallest are very 

 generally found imbedded among the granules. 



The second kind of coccolith (fig. 4 a — 7)i), when full 

 grown, has an oval contour, convex upon one face, and flat 

 or concave upon the other. Left to themselves, they lie 

 upon one or other of these faces, and in that aspect appear 

 to be composed of two concenaic zones (fig. 4 d, 2, 3) 

 surrounding a central corpuscle (fig. 4: d, 1). The central 

 corpuscle is oval, and has thick walls ; in its centre is a clear 

 and transparent space. Immediatelj' surrounding this cor- 

 puscle is a broad zone {2), which often appears more or 

 less distinctly granulated, and sometimes has an almost 

 moniliform margin. Beyond this appears a narrower zone 

 (3), which is generally clear, transparent, and structureless, 

 but sometimes exhibits well-marked striae, which follow the 

 direction of radii from the centre. Strong pressure occasion- 

 ally causes this zone to break up into fragments bounded by 

 radial lines. 



Sometimes, as Dr. Wallich has already observed, the clear 

 space is divided into two (fig. 1 e). This appears to occur 

 only in the largest of these bodies, but I have never observed 

 any further subdivision of the clear centre, nor any tendency 

 to divide on the part of the body itself. 



A lateral view of any of these bodies (fig. 4/ — i) shows that 

 it is by no means the concentrically laminated concretion it at 

 first appears to be, but that it has a very singular and, so far as 

 I know, unique structure. Supposing it to rest upon its con- 

 vex surface, it consists of a lower plate, shaped like a deep 

 saucer or watch-glass ; of an upper plate, which is sometimes 

 flat, sometimes more or less watch-glass-shaped ; of the oval, 

 thick-walled, flattened corpuscle, which connects the centres 

 of these two plates ; and of an intermediate substance, which 

 is closely connected with the under surface of the upper plate, 

 or more or less fills up the interval between the two plates, 

 and often has a coarsely granular margin. The upper plate 

 always has a less diameter than the lower, and is not wider 

 than the intermediate substance. It is this last which gives 

 rise to the broad gi'anular zone in the face view. 



Suppose a couple of watch-glasses, one rather smaller and 

 much flatter than the other; turn the convex side of the former 

 to the concave side of the latter, interpose betAveen the centre 

 of the two a hollow spheroid of wax, and press them tog-ether 

 — these will represent the upper and lower plates and the 

 central corpuscle. Then pour some plaster of Paris into the 

 interval left between the watch-glasses, and that will take the 



