MELONECHINUS. 391 



the genus repeat very fully the adult character of reprorntative lower species and genera in 

 its series. Dorsal localized stages could not be ascertained in Melonechinus giganteus, but 

 in the diagram (text-fig. 237, p. 231) such stages are supplied fnun Melonechinus keepingi, 

 which at the mid-zone has twelve columns of ambulacra! plates, and therefore in this respect 

 represents the same structural differentiation as yiganteus. 



The interambulacra of Melonechinus giganteus are slightly wider than the ambulacra, 

 and are arched up in strong melon-like ribs which outwardly are gently curved and laterally 

 dip steeply down to the adradial sutures (Plate GO, fig. 3). There are eleven columns of plates 

 in an area at or above the mid-zone (Plate 59, fig. 14; Plate 61, fig. 8). This is the highest 

 number known in the genus and family, and therefore, as giganleus has as great a number of 

 ambulacral plates as are known, it is considered the most highly evolved species of the genus 

 and family yet discovered. The interambulacral plates bear many secondary tubercles, tin-re 

 being about twenty-five on the larger plates (Plate 59, fig. 15). The spines are not preserved. 



The structure of the interambulacra of the type was described in detail and the point of 

 origin of columns tabulated in my earlier paper (1896, pp. 174-182). Eleven columns wen- 

 ascertained in four of the five areas, and it was estimated that there arc about 1,550 interam- 

 bulacral plates in the specimen. It has not only more columns of interambulacral plates than 

 any other species of the genus, but, as a natural corollary, it has a more accelerated development 

 than lower species of the genus. Considering here only one area, that figured (Plate 59, fig. 14), 

 it is seen that plates of the basicoronal row are wanting, but there were doubtless two, as restored 

 in Plate 59, fig. 13. Assuming these two as a basal row, there are three plates in the second 

 row and four in the third as usual in the whole family. The fifth column originates in the 

 fifth row (very early), the sixth column in the eighth row, the seventh column in the tenth, 

 and the eighth column in the fourteenth row, all very early as compared with the zone at which 

 these columns originate in other species of the genus (compare text-fig. 2I(>. p. 382; Plate .~>7. 

 fig. 1). The ninth column originates in the sixteenth row in a small tetragonal plate, in which 

 the two sides wanting are compensated for by the two heptagons, H' and H", on its left dorsal 

 and right ventral borders. The hexagon A lies on its left ventral border. In the seventeenth 

 row there are only eight plates, but in the eighteenth row the ninth column starts again with a 

 pentagonal plate bearing a heptagon, H, on its right ventral border as usual. The same 

 anomaly occurs in the origination of the ninth column in three other areas in this superb speci- 

 men, showing a strong tendency to a parallel radial variation. The tenth column originates 

 in the twenty-second row, just below the mid-zone, about the same zone in which the eighth 

 column ordinarily originates in M. multiporus (Plate 57, fig. 1). The eleventh column origi- 

 nates in this area (Plate 59, fig. 14) in the twenty-ninth row, or in other areas a trifle earlier. 

 Dorsally, as a senescent character, it drops out before reaching the apical disc. In the speci- 

 men in Vanderbilt University Collection 222, the eleventh interambulacral column has already 



