44^ 



NATURE 



\Sept. 21, 1876 



something to be said for this method of treatment, but 

 the question is whether the difficulties it entails are 

 not greater than those it is intended to avoid. The 

 arenaceous and porcellanous types have one important 

 character in common, in the imperforate condition of the 

 test ; and there is yet another peculiarity which some of 

 the members of both groups possess not shared by the 

 Perforata, namely, the tendency exhibited by such forms 

 as tolerate brackish water {Mihola and Trochammmd) 

 to assume a more or less chitinous or membranous invest- 

 ment in proportion to the decreased amount of mineral con- 

 stituents held in solution ; whilst under similar conditions 

 the shells of the more highly organised perforate forms 

 {Polystomella zxid. Nonionhia) become thinner and more 

 delicate, but never change their essentially calcareous 

 nature. On the other hand, it can be easily shown that 

 the Globigerinida have more points of connection with 

 the arenaceous group than the Miliolida have, for whilst 

 the latter furnishes but few examples with any approach 

 to sandy shell-texture, the former has a number of 

 types which might with good reason be associated with 

 some which have hitherto been classed amongst the 

 Litiiolida, to form an intermediate group, calcareous and 

 perforate under certain circumstances, arenaceous and 

 imperforate under others. 



The truth of the matter is that the variations of the 

 Foraminifera are too multiform and the connection of the 

 members of the Order one with another is too close to 

 be well adapted for divisional classification, but they lend 

 themselves readily and naturally to arrangement in linear 

 series, "fhus, in the first family of the English classifi- 

 cation, Miliolida, we find a large assemblage of forms 

 of various degrees of complexity, but having, with 

 trifling exception, compact, non-porous shells. In such 

 forms as Qtiinqueloctilma agglutinans and its fellows, the 

 " porcellanous " overlaps the " arenaceous " series, ren- 

 dering complete separation on the basis of shell-texture 

 impossible. Next in order come a number of types 

 essentially and invariably arenaceous — a series by no 

 means uniform in the structure of the investment or even 

 in the materials of which it is composed, but all compo- 

 site, imperforate, and opaque. These give place to another 

 intermediate set, partaking more or less of the characters of 

 the " arenaceous " and the " perforate " groups, comprising 

 such genera as Endothyra, Valvulma, Textularia, Buli- 

 miiia, and the like, that are, it may be, clear-shelled and 

 imperforate, sandy and imperforate, sandy externally but 

 with a perforate shell as basis, or even hyaline and perforate, 

 the mere size of the specimen having apparently much to 

 do with the nature of the test. These supply any required 

 number of transitional steps to the uniformly " perforate '' 

 types which constitute the highest group. We need not 

 dwell further on this subject. To the systematist it is 

 one of considerable difficulty, from whatever point it is 

 viewed, and unless some better basis for classification 

 than the minute structure of the shells of these little 

 animals can be suggested, it may be a question whether 

 an increase in the number of families by the recognition 

 of an intermediate group, or possibly of more than one, 

 would not be the course open to the fewest objections. 



The second point that demands notice is the recon- 

 struction of the important family NUMMULINIDA ; for 

 practically the characters assigned to it in the work before 



us would result in nothing less than reconstruction, if 

 literally read. It is not needful to reprint the entire para- 

 graph relating to the subject, for its essential element may 

 be stated in few words, viz., the reliance on a complicated 

 interseptal canal-system, as the characteristic feature of the 

 Nummuline group. As one consequence of this limitation, 

 and it is only one out of many that must ensue if consist- 

 ently carried out, the genera Amphistegina and ArcJice- 

 discus are placed amongst the Globigerinida. That there is 

 sorat primd facie ground for the change maybe taken for 

 granted, or it would not have found favour with so com- 

 petent an observer as the author of the Handbook, but the 

 more it is investigated the more we think it will appear 

 that reliance on a single character of this sort is suited to 

 the exigencies of an artificial system, rather than to the 

 exposition of natural relationship. As the point in ques- 

 tion is one of great importance, and involves the principles 

 on which accepted methods of classification are based, it 

 may be worth while to illustrate its general bearing by 

 one or two instances of the results that would follow the 

 adoption of a hard and fast definition of the nature pro- 

 posed. Take for example the well-known genera Polysto- 

 mella and Nonionina — types so closely related that the 

 latter is often treated as a mere sub-genus of the former, 

 and is perhaps best so regarded. In its higher modifica- 

 tions Polystomella has a very complicated canal-system, 

 whilst no trace of such organisation has ever been traced 

 in Nonionina. On the other hand, turning to the Glo- 

 bigerinida, we find that Rotalia (proper) in its highest 

 modifications has also a well-defined and complex canal- 

 system, and the same, moreover, is easily recognised in 

 Calcarinaj so that this character, even according to Dr. 

 Zittel's arrangement of genera, is not an exclusive feature of 

 the NUMMULINIDA. Prof. W. K. Parker, than whom few 

 have better/ight to be heard on such a point, regards A tnpht- 

 stegina (though a true generic type) as bearing a relation to 

 Nunwiulina, similar in kind if not in degree to that which 

 exists between Nonionina and Polystomella. It is true 

 that neither in Amphistegina nor in Archcedtscus has any 

 true canal-system been demonstrated, but it must be 

 recollected^of the latter, type that it has no septa, and it 

 is possible that the double tubulation occasionally observ- 

 able in its supplementary skeleton may represent this 

 special organisation in a rudimentary condition. That 

 Polystomella is a more highly organised type than Rotalia, 

 and Nimmiulitia presents a distinct advance upon either, 

 and that in general terms the fact may be demonstrated 

 by the relative complexity of the structure of the test, is 

 hardly open to question. What is here contended for 

 is this — that throughout the Foraminifera in each group 

 comprising the modifications of a single central type, or 

 of two or more closely allied types, there may be traced a 

 regular series of subordinate forms gradually increasing 

 in complexity of organisation, and that these cannot be 

 separated in a system of classification without doing 

 violence to the order of nature. In the types to which 

 reference has already been made, sach a sequence is easily 

 found. In Rotalia, the minute .hin-shelled, brackish- 

 water R. nitida presents the very simplest morphological 

 characters ; R. Beccarii with its double septal walls marks 

 a distinct advance, and, omitting a multitude of inter- 

 mediates, R. Schroeteriana exhibits the highest develop- 

 ment with a complete interseptal canal-system. In Poly- 



