PRINCIPLES OF CLASSIFICATION. 51 



portions of that substance remaining in their original condition, so as to maintain a communica- 

 tion between the contents of the chambers and the parts of the calcareous skeleton most 

 removed from them, analogous to that which the Haversian canals afford in the case of latninte of 

 bone not in the immediate vicinity of a vascular surface. As, therefore, the development of the 

 Haversian system is related to the thickness of the bone-substance to be nourished, so does that 

 of the canal-system in Foraminifera seem to be related to the amount of the consolidating 

 substance which constitutes the supplemental skeleton. There is good reason to believe that 

 these canals are occupied in the living state by prolongations of the sarcode-body, which pass 

 from the chambers into the portions of the system in nearest relation to them, and proceed to 

 its peripheral extensions ; and they are largest and most numerous where nutriment has thus 

 to be conveyed to parts of the supplemental skeleton, which (like the outgrowths of Calcarina) 

 are very far removed from the segments of the ordinary sarcode-body. Now, although it is 

 only in the largest and most developed types of the hyaline series, that we meet cither with a 

 distinct " canal-system" or witli any considerable amount of that intermediate deposit which 

 it nourishes, yet the jDresence of these two peculiar features most strongly differentiates those 

 types from such of the porcellanous series as most nearly resemble tliem in general plan of 

 growth, and to which, according to any classification essentially founded on that character, 

 they would be most nearly approximated, — as, for example, the hyaline Ojjerculina from the 

 porcellanous Pciieroplls, the hyaline Cydodypeas from the porcellanous Orbitolites. 



64. Another strongly marked difference, that seems no less obviously related to the 

 physiological condition of the animal than the perforation or non-perforation of the shell, is 

 observable in the degree of separation that exists between the segments of the sarcode-body 

 in the two series respectively ; this body even in the most complex types of the imperforate 

 shells being an aggregate of mutually related parts, whilst even in the simplest types of 

 the perforated these parts acquire a much higher degree of independence, so as to live much 

 more /';/• and by themselves alone. The key to this difference is furnished by the relative 

 size of the aperture, which indicates in the unilocular types the degree of readiness with 

 which the animal can extend itself into the medium it inhabits, whilst in the raultilocular it 

 indicates not only this, but also the relative amount of connection between the several segments 

 of the composite structure. Thus, when we compare the unilocular Gromia (which, though 

 possessing only a membranous " test," may be considered as physiologically representing the 

 unilocular type of the imperforate Foraminifera) with the unilocular Layena, we are at 

 once struck with the extreme narrowness of the aperture of the latter, as compared with 

 the wide, open mouth of the former (Plate II, fig. 2) ; and this difi'erence will be readily 

 understood, when it is borne in mind that in one case the principal aperture is supple- 

 mented (so to speak) by the' pores distributed through the whole of the globular casing that 

 encloses the body of the animal, every one of which allows the passage of a pseudopodium, 

 whilst in the other the sarcode-body, shut up within its " test," has no other means of com- 

 munication with the external world than that afforded by its single orifice. So, if we compare 

 the apertures of Niibecularia, Vertebralina, and Miliola with those of the Nodosarian series, we 

 find a no less striking contrast between the mere constrictions that mark out the segmentation 

 of the sarcode-body in the former, and the almost complete separation that exists between the 

 successive chambers in the latter. In Pmeroplis, again, although the apertural pores are 



