1849.] CARPENTER ON THE STRUCTURE OF NUMMULINA. 25 



distance from each other about 1-1 5,000th. In a thin vertical sec- 

 tion of the shell (fig. 15) they are seen to run parallel to each other, 

 and to be free from sinuosities or interruptions. The whole of this 

 portion of the shell, therefore, is minutely porous. The structure 

 in question can seldom be clearly distinguished in those Numraulites 

 which have had their texture altered by calcareous or siliceous infil- 

 tration ; but as the appearances which these present correspond 

 closely with those exhibited by specimens of N. Iceviyata which have 

 been subjected to the same change, I have no doubt that the tubular 

 structure in question is common to the whole group. Very fre- 

 quently the metamorphosis is such as to give the appearance of a 

 minute prismatic arrangement to the shell-substance (as shown trans- 

 versely in fig. 12, 6, 6), by which I was myself deceived until I had 

 examined specimens in which less alteration had taken place. All 

 the Nummulites which I have examined present a remarkable de- 

 parture from this structure, in that portion of the shell which forms 

 the margin of each whorl. Here, instead of an assemblage of minute, 

 closely-set, parallel tubuli, we have a much coarser arrangement, the 

 solid substance being perforated with a smaller number of tubes of 

 two or three times the diameter of those last mentioned, which pass 

 in a radiating manner from the inner to the outer surface. Some 

 indications of this difference are seen in fig. 4 ; but it is much more 

 clearly displayed at 6, 5, fig. 15, which represents a portion of a very 

 thin section taken in the same direction, and viewed by transmitted 

 light. The openings of these tubes on the outer margin of the whorl 

 are not readily discernible, partly in consequence of the somewhat 

 oblique direction of the orifices, and partly through these being 

 usually covered with a calcareous incrustation. When this has been 

 removed by the application of dilute acid, they are easily seen when 

 properly looked for, as was first pointed out to me by Mr. J. Morris. 

 Each successive whorl of the Nummulite, as is well known, not 

 merely surrounds the preceding whorl, but completely invests it ; so 

 that, in a vertical section, each chamber of the medial plane is seen 

 to be covered with as many layers of shell above and below its own 

 roof and floor, as there are chambers intervening between it and the 

 nearest margin of the section. Thus in fig. 4, the chamber d has 

 three chambers on its exterior, and is invested by three layers of shell 

 above and below, in addition to those by which it is itself inclosed. 

 On the other hand, the chamber e, which is the tenth from the mar- 

 gin, — that is, which has nine whorls on its exterior, — is invested by 

 nine layers above and below. In some species of Nummulites, as the 

 N. complanata, these investing layers are closely applied to each 

 other, over the whole of each surface ; the chambers only occupying 

 the margin of the whorl, as shown in fig. 17. But in by far the greater 

 number of species, these successive layers are not in contact with each 

 other, being separated by prolongations of the marginal chambers, 

 which extend over the entire surface of the disk. These chambers 

 are still divided by prolongations of the marginal septa, which are 

 continued between the investing portions of the contiguous whorls, 

 and form a series of vertical partitions, upon which the successive 



