EEPOET ON THE RADIOLARIA. 483 



The second subfamilj^ of tlie Porodiscida are the Trematodiscida, which are derived 

 from the preceding Ai'chidiscida by concentric growth in the equatorial plane. In the 

 same way in which the simple chambered ring of Archidiscus is connected by radial beams 

 with the central chamber, so in Porodiscus a variable number of concentric rings is 

 connected with that first ring. The number of these concentric rings varies between two 

 and ten or more, but commonly amounts to between three and five. The radial beams con- 

 necting them are either piercing or interrupted ; theii- number increases gradually from 

 the centre towards the periphery. The chambers between them are sometimes more 

 regular, at other times more irregular in size and form. Their upper and lower wall 

 is formed by the two covering " sieve-plates," or the porous cover-plates, which are 

 continued from the central disk to the margin. If these two sieve-plates continue being 

 parallel, the disk becomes medal-shaped or a short cylinder ; if the two plates become 

 more or less concavely vaulted one against the other, the disk becomes biconvex 

 lenticular, the middle part thicker than the marginal part. Earely the contrary is the 

 case, the margin thicker than the centre, and then the disk Ijiconcave. 



In my Monograph I had separated as two different subfamilies the true Tremato- 

 discida (with cii-cular concentric rings) from the Discospirida (with spirally convoluted 

 rings). But the enormous mass of specimens, which I afterwards examined in the 

 Challenger collection, has convinced me that this separation was not natural. For in 

 one and the same genus of most nearly allied forms we find on one hand quite regular 

 concentric cii-cular forms (Trematodiscits), on the other hand spirally convoluted forms 

 (Discospira), and connecting between them such forms as are in the central part 

 concentric, in the marginal part spiral (Ferisjnra) — or conversely, these in the centre 

 spiral, on the margin concentric {Centrospiva) — and frequently also more or less 

 irregular forms with interrupted rings (Atactodiscus) ; therefore, all those genera 

 (Prodromus, 1881, Nos. 448-452) have only the value of subgenera of Porodiscus. 

 But a distinct genus is Perichlamydium, in which the two sieve-plates run on the 

 margin of the lens and form a broad hyaline porous or solid girdle. More important 

 is the distinction of the Ommatodiscida, in which the margin of the disk exhibits one 

 larger osculum, armed with a corona of spines {Ommatodisci(s), or two oscula, opposite on 

 the poles of one axis [Stomatodisciis). Whilst in many Porodiscida all chambers of the 

 concentric rings lie in one and the same (equatorial) plane, in many others with further 

 growth they become stratified in floors, and the whole disk is therefore composed of two 

 to four or more parallel disks, each with a system of concentric chambered rings or 

 girdles, c|uite as in the majority of the Coccodiscida (p. 457). Often the central part of 

 the lenticular disk becomes thickened by apposition of such floors or strata, whilst the 

 marginal part remains simple, with one single stratum. The communication between the 

 chambers of the different strata seems to be the same as in the similar Coccodiscida. 



Also the margin of the disk exhibits in the Porodiscida the same characteristif 



