KEPOKT ON THE RADIOLARIA. 51 



spherical form is quite constantly preserved in the central capsule, and commonly also 

 in the fenestrated shell enveloping the latter, although in many forms the sphere is 

 more or less modified ; very frequently it is an " eudospherical polyhedron," i.e., a 

 polyhedron all the angles (or the nodes of the network) of which lie upon the 

 surface of a geometrical sphere ; more rarely the spherical form is more or less modified 

 and irregular. In the great majority of Sphseroidea there is no external indication 

 of the three dimensive axes ; but in many forms they are indicated by the regular 

 position of certain external radial spines or internal radial beams. However, in no 

 case are those three axes expressed in the form of the shell itself and of the enclosed 

 spherical central capsule ; this is the main character by which the Sphseroidea 

 difi'er from the following sections: — Prunoidea, Discoidea, Larcoidea, all 

 three of which arise from them. 



The section S p h se r o i d e a, in the sense here restricted, was founded by me in my 

 Protistenreich (1878, p. 103) and adopted by Hertwig (1879) in his Organismus der 

 Radiolarien (p. 39). The difierent groups appertaining to this large section were 

 characterised more accurately in my Prodromus (1881, pp. 448-456); there I gave 

 the characters of six subfamilies with thirty tribes, containing ninety-three genera. 

 Formerly, in my Monograph (1862), the S p h a3 r o i d e a were disposed in five difi"erent 

 families : — Ethmosphasrida, Cladococcida, Ommatida, Spongosphgerida, CoUosphserida. 

 At that time I could not separate them sufficiently from some Acantharia and 

 Ph^odaeia, which have a similar spherical lattice-shell. 



As the number of different genera and species in the S p h as r o i d e a is much 

 greater than in all other sections of Spumellaria, many forms were already described 

 by former authors. In the oldest system of Ehrenberg (1847, ha. cit., p. 53) they 

 represent one part of his Haliommatina (with four genera, Haliomma, Chilomma, 

 Stylosphcera, Spongosplueva). Most species, however, of these genera are 

 Discoidea. Also in the latest system of Ehrenberg (1875, loc. cit., p. 157) his 

 Haliommatina are a confused conglomeration of difierent Spumellaria (S p h se r o i d e a, 

 Discoidea, and P r u n o i d e a). 



The section S p h as r o i d e a is the largest division of S p h a3 r e 1 1 a r i a, comprising 

 not less than one hundred and seven genera and six hundred and fifty species. This 

 enormous number (easily to Ije augmented by further investigations) requires a careful 

 disposition in difi"erent families and subfamilies. For this disposition two different 

 principles only can be employed : firstly, the number and disposition of the radial 

 spines ; secondly, the number of the concentric latticed .^p>heres, which are connected by 

 radial beams. I give here the preference to the first principle, whilst in my Prodromus 

 (1881) I had preferred the second. The question, which of the two principles is more 

 important for the classification of S p h ae r o i d e a, is very diflicult to answer ; probabl}- 

 in many cases the former, in man)' the latter is more important for their phylogeny. 



