GENUS CYCLOCLYPEUS. 295 



however, are very minute, their diameter being not above 1 - 1 0,000th of an inch. They are 

 wanting in certain parts of the shell [a, a), which there presents a transparence that contrasts 

 strikingly with the semi-opacity produced by the tubular perforations. By the comparison 

 of vertical with horizontal sections taken in difterent planes, it appears that these non-tubular 

 portions of the shell have a conical form, the base of each being on the surface of the shell, 

 and its apex pointing to one of the angles at the outer margin of a chamber (fig. 5, c c, d d). 

 Their gradual widening towards the surface causes the diameter of their bases to increase 

 with every addition to the thickness of the shell ; and thus it is on the older portion of the 

 shell, and especially on its central protuberance, that they become most conspicuous as 

 rounded spots, sometimes rising into tubercular elevations (1 494) . In horizontal sections of 

 the superficial lamellas, they form a large proportion of the area ; whilst in similar sections 

 near the chambered plane they become blended with angular projections of the annular 

 partitions, whicli fill up the spaces left between the proper walls of the chambers by the 

 rounding-off of their angles (as shown in fig. 4). The lamellated structure is seen in these 

 conical pillars (fig. 3, a, a), the lamellae being continuous with those of the tubular part of 

 the shell ; so that at each increase in thickness a tubular and a non-tubular portion must be 

 superimposed upon the corresponding parts of the preceding lamella. Both in the tubular 

 structure of the shell, and in the presence of these non-tubular columns, there is an exact 

 conformity to the structure of Nummulina, OpercuJijia, &c. 



498. I have now to speak of the system of iidersej}fal ciuiah, which is traceable 

 throughout the whole of the solid skeleton, though less regularly distributed than in some 

 of the preceding types. It may be conveniently described as consisting of radial, vertical, 

 and annular canals. The radial canals are seen both in vertical and in horizontal sections 

 (fig. 5, h, h ; fig. 6, y, //,) excavated in the shelly substance which is interposed in the 

 radiating partitions between the proper walls of adjacent chambers of the same annulus. 

 When the canal reaches the end of the radial septum, it usually subdivides into two, which 

 diverge at a considerable angle from each other, so as, by traversing the annular septum, to 

 reach the two alternating radial partitions of the next annulus ; and as each branch, before 

 entering the partition (fig. 6, i, /') towards which it runs, unites with another branch that inclines 

 towards it from the radial canal next adjacent. It follows that just as every chamber com- 

 municates (normally) with the two alternating chambers in the annuli internal and external 

 to it, so do the interseptal canals of every radiating partition communicate with those of the 

 partitions alternating with it in the internal and external annuli. In each radial partition 

 there are at least two, and very commonly three tiers of such canals, and sometimes they 

 are yet more numerous (fig. 5, i). Short transverse branches, apparently communicating 

 with the cavity of the chambers (fig. 6), are sometimes seen to proceed from the longitudinal 

 canals ; in regard to these communications I would not speak with confidence from what I 

 have seen in Cijdoclypeas ; but that they exist in other organisms is unquestionable. The 

 horizontal radiating canals communicate (as at g ■, g, fig. 6) with vertical canals, which ^^ass 

 directly towards the two surfaces of the disk, whereon they open ; these canals are best seen 

 in horizontal sections taken near the upper or under surface of the chambers (fig. 4), in 

 which they present themselves in regular rows, c, c, corresponding to the radial partitions ; 

 whilst in similar sections taken nearer the surface they are seen to be less regularly disposed, 



