452 I'ROCICICDINGS OF TUIC NATIONAL MUSEUM. 



The new Alabama form resembles, as stated above, the jjenus Cyclo- 

 (ksmnH rather than the others tlius far mentioned. Jint from (Jyclodes- 

 mus, as well as from all other Mcro(;lieta, it dillcrs conspienously in tlie 

 possession on each se}4nient of a dei^p <;avity located at the base of the 

 carina in front. From the al(;oholic material available I have not been 

 able to dcMionstrate in connection with these pits tlie existence of any 

 repu^^iiatorial i)ores, and there are two facts that seem to negative such 

 an exidanation of their nature and function. 



The first is that the (cavities are not lo(;ated in ai)osition corres])ond- 

 ing to tiiat occu[)ied by i)ores in any other member of the order. In 

 general the pores of Mero(;heta are near the margin, but in the few 

 instances where they are far removed from the lateral edge the pore is 

 still distin(;tly a part of the carina, and not an excavation in tlu; body 

 cylinder as in tlie i)resent case, to say nothing of the anomaly of find- 

 ing a repiigiiatonal pore at the bottom of a deep (lavity. 



The se(;ond uni(iue circumstance is that in no diploi)oda are repug- 

 natorial pores known to occur in front of the fiftli segment, when? they 

 normally begin, all exceptions being in the sui)pression of the i)ores of 

 that segment. The peculiar cavities of the new tyi)e are, however, 

 api)arent on the fourth and third segments! The only other paiied 

 cavities affecting the dorsal surface of the segments of dii)loi)oda are 

 the so-called "scobiiia" of the Auocheta, but these are found near the 

 median liix^, are located near oi' at the anterior edge of the segment, 

 and oiler no similarity of form or structure which would give ground 

 lor asserting a homology with the pits discovered on the Alabama 

 specimens under dis(!Ussion. 



It is also noteworthy that these cavities form a continuous series 

 occurring on all segments from the third to the i)«nultinuite. Without 

 excei)tion the pore series is at least once interrupted in all known 

 Merocheta, so t»hat if the cavities were in form and function normal 

 pores, they would still be uni(jue in position, number, and distribution. 

 It accordingly seems desirable to hold the newly recognized type as 

 distinct from Gyclodesmus, notwithstanding the general resemblance iu 

 habit. 



Keferences to the- descriptions and i)lates of other American onisci- 

 form genera are added, together with descriptions of an interesting new 

 species of <)niscod(S)ui(s. A systematic arrangement is also proposed, 

 and a key to the familicis is given, but this must be looked upon as 

 artificial, the forms included not composing a natural group. 



ANALYIlCAl, KKY TO THK A.MKKICAN KAMIMKS OK ONISCIFORM MEKOCIIETA. 



.Sognieiits with a iK)Ht(»ri<)r bonier coinixjHod of a transverse row of convex, rectan- 

 gnlar areas; penultimate segment with its carina- as broad or broader than the small, 

 rounded lastHcgmont: Family ()niscoi>k.smii),i;. 



•Segments witliont a posterior aroate border; last segment snbqnadrat(^ miirh 

 broader than the carina- of segment 19 



