PRINCIPLES OF CLASSIFICATION. Gl 



surface remains to be mentioned ; namel)% that hispid character which is given by the pro- 

 jection of conical spines from every part of it. Of this we have numerous examples among 

 the straight and slightly curved NodosaruB (Plate XII, fig. 2) ; whilst among spiral shells 

 we meet with it in the most pronounced degree in a varietal form frequently presented by the 

 young of Calcarina (Plate XIV, figs. 6, 7). The careful study which I have made of this last type 

 enables me to aflarra with confidence, that these conical spines are formed by an excessive 

 growth of the tubercles of non-tubular shell-substance already described, and that they usually 

 disappear with the advance of age, by an increase in the thickness of tlic ordinary shell 

 substance which fills up the spaces that intervene between them. And among the Nodusaria 

 they certainly have no more vahic as differential characters tlian the ridges already noticed. 



75. Taking our stand, then, upon the characters by which the Order Reticularia is 

 difi"erentiated from otlier Rhizopods, — viz., the minute subdivision and the free inosculation 

 of the pseudopodia, the imperfect differentiation of the endosarc and the ectosarc, and the 

 absence both of nucleus and contractile vesicle, — ^we have finally to inquire how the group 

 thus constituted can be most naturally subdivided in accordance with the principles that have 

 now been laid down. At first sight, it would appear as if the groups of Gromida and 

 FoRAMiNiFERA Were so strong!)^ differentiated by the deficiency in the former of that 

 calcareous envelope which is the special characteristic of the latter, that they should constitute 

 two sections or sub-orders of corresponding rank ; and such a view has been adopted by 

 MM. Claparede and Lachmann (xxv, p. 34). If, however, we attach a greater value to 

 the characters furnislied by the animal than to those afforded by the material of its envelope 

 (and this appears to me the more natural method), we find that the affinity of the Gromida to 

 those Foraminifera whose shells, being imperforate, do not give passage to pseudopodia, is even 

 closer than is that of the Foraminifera having imperforate shells to those of which the shells 

 are perforated; whilst the systematic value of the difference in the material of the envelope 

 is lowered by the circumstance, that among the true Foraminifera we occasionally meet with 

 instances in which the only part of the shell that is really formed by an exudation from the 

 animal is the cement that holds together the particles of sand from which it derives its 

 solidity. Following out this principle, tlie whole Order Reticularia may be subdivided into 

 two primary groups, according as the envelope (whether membranous or shelly) is imperforate 

 or is perforated ; the pseudopodia in the former case issuing only from the single or multiple 

 aperture, whilst in the latter they proceed from the general surface of the body. The imper- 

 forate sub-order may be divided into three very natural groups, according as the nature of the 

 envelope is membranous, jjorcellanous, or arenaceous ; and thus we have the families Gromida, 

 MiLiOLiDA, and Lituolida. Throughout the perforated sub-order, on the other hand, the 

 texture of the shell is hyaline or vitreous, save in the few instances in which the ordinary shell- 

 substance is partially replaced by particles of sand ; and there seems no other l)asis for a 

 division of that sub-order into families, than that which is afforded by tiic mutual affinities 

 of its generic t3'pes. 



The results of our inquiry up to this point, thei*efore, may be summed up as follows 



.« 



* Tlie nearest approach to the above principle of classification -wliich I fiud among preceding 

 systematists, is that hinted at, though not actually adopted, by j\I. Dujardiu (xxxvi). As the original 



