436 PROCEEDINGS OF THE NATIONAL MUSEUM vol.89 



said to be from the Red Sea — which is too improbable for belief. 

 Verrill's species has never been rediscovered and probably rests on 

 a young Encope. It is odd that Lambert and Thiery, after reviving 

 the genus Leodla for species with six lunules, put caroUniana, 

 eryfhrea, and pacifica in Mellita, which they distinctly say is 

 characterized by "cinq lunules." 



The genus Mellita^ restricted thus to those forms having five 

 lunules, four ambulacral and one interambulacral, contains but few 

 species, of which the type, quinquiesperforata (Leske), also well 

 known as pentapora (Gmelin) and testudinata Klein (or, more cor- 

 rectly, Agassiz), is much the best known. Lambert and Thiery list 

 five others, but as already pointed out three of these are really Leo- 

 dias. Of the other two, ampla Holmes, a fossil species from South 

 Carolina, is undoubtedly a synonym of qimnquies perforata^ but longi- 

 -fiHsa appears to be a valid species confined to the western coast of 

 Central America and Mexico. 



Hitherto the name qumqmesperforaia has been used for all the 

 5-lunuled Mellitas found on the eastern coasts of America from 

 Nantucket to southern Brazil, but the material from the National 

 Museum shows that several quite distinct forms have been included 

 under that long name. After critical study of this material, and of 

 the large series in the Museum of Comparative Zoology, it seems 

 best to recognize two new species and a well-marked variety of quln- 

 quiesp>erforata. Mellita thus becomes a genus of four species and a 

 variet}'^, which may be distinguished from each other by the following 

 key. It must be constantly borne in mind, however, that the younger 

 the specimen, the less well marked are its characters. Individuals 

 less than 40 mm. in diameter cannot always be positively identified. 



KEY TO THE SPECIES AND VARIETIES IN THE GENUS MELLITA 



a '. Test more or less circular or pentagonal, its length (100-150 mm. 

 in large adults) nearly equal to its width, often less, but 

 rarely less than 90 percent thereof ; apex more or less central, 

 but rarely definitely in front of abactinal system; anterior 

 half of test not mai'kedly thicker than posterior; periproct 

 usually longer than wide, often markedly so ; unpaired 

 luuule not very long, about 20 percent of test length. 

 6\ Test more or less pentagonal, rather stout, the apes tending 

 toward the anterior ; periproct little or moderately 



elongated quinquiesperforata 



1)1 Test nearly or quite circular, light and thin, the apex tending 

 to be posterior ; periproct very long and narrow. 



quinquiesperforata var. tenuis 

 a'. Test more or less elliptical, the width greatly exceeding the 

 length (70-98 mm. in large adults) ; length only 80-90 percent 

 of width; apex anterior, usually very evidently so; anterior 



