MELONECHINUS. 



The interambulacra are a little wider than the ambulacra and are gently rounded up in 

 moderate melon-like ribs, laterally sloping gradually down to (lie ;idr;i<li;il suture- I'hn 

 figs. 1-3; Plate 60, figs. 1, 2). The ribs are quite distinct 1mm tho-e of (lie nearly allied M. 

 liratus (Plate 54, fig. 4), in which the ribs are high and much more steeply rounded in tlie am- 

 bulacra, and also are more sharply elevated laterally in the interambulacra. The interamlnil- 

 acra of mulliporus have typically eight or nine columns of plates in each area, very rarely only 

 seven columns. Typically, there are seven columns of plates at the mid-zone, the eighth column 

 coming in a little later (text-fig. 245, p. 382; Plate 57, fig. 1). This is the character in tin- 

 specimens shown in Plate 55, figs. 1, 2, and Plate GO, fig. 1. In addition, there may be a ninth 

 column appearing dorsally in one or more to all the areas, and tin- >tnicture of having nine 

 columns is about as frequent as only eight columns in an area. In the specimen. Plate 54, 

 fig. 5, there are nine columns in area A, but eight columns in all other areas. In the splendid 

 great specimen, Plate 55, fig. 3, there are nine columns in areas C, E, and G, the only ones that 

 can be determined with certainty. There are also nine columns shown in two areas in the speci- 

 men figured in Plate 56, fig. 11. In the very complete specimen, Plate 60, fig. 2, there are nine 

 columns of plates dorsally in all five areas. As worked out by Jackson and Jaggar (1896, 

 p. 162), it was found that eight columns are a little more frequent than nine in this specie-. 

 Rarely there are only seven columns in an area (p. 50). In a specimen figured by .lacksc m and Jag- 

 gar (1896, Plate 4, fig. 18) there are evidently only seven columns in one area, but this area has 

 a very retarded development, the fourth column not coming in until much later than in any other 

 specimen known in the family; in another area in this specimen, however, the fourth column 

 originates in the next row after the third as usual (p. 439). When seven columns only 

 are seen in an area, it is usually due to the fact that it is imperfect dorsally, -o that that 

 portion where the additional one or two columns appear is wanting. Miss Klein (1904, p. 43) 

 says that in a collection of several hundred specimens one can find any number of specimens 

 with ten, eleven, and even twelve columns in one or more interambulacral areas. I have 

 a good many specimens, but none with more than nine columns in an area. The adradial 

 plates are pentagonal as usual in the genus and family, and bevel under the ambulacrals on 

 the adradial sutures. All the plates of median columns are hexagonal or the mechanical 

 equivalent of the same. Interambulacral plates, like the ambulacrals, bear secondary 

 tubercles, similar to those shown in M. giganteus (Plate 60, fig. 3; Plate 61, fig. 8). Spines 

 thickly cover the plates in exceptional specimens; the spines are about 3 mm. in length (Plate 

 56, fig. 13), swollen at the base, and taper gradually to the tip, apparently with slight constric- 

 tions. Such constrictions have not been noticed in spines of other Palaeozoic specie-. 



Ventrally, the interambulacra, when perfect, have two plates in the basicoronal row ' 



1 Miss Klem (1904, p. 2) says that there are three plates ventrally when the six-ciiiien is perfect. She i> certainly mis- 

 taken, as discussed on pages 66, 67. 



