STRUCTURE OF CORONA OF ARCHu^OCIDARIS WORTHENI. 215 



corona. The fact that it has been resorbed is shown by the position of 

 the tubercle, which lies nearer to the ventral edge, and also the fact that 

 a line drawn through the right and left apices of the hexagon does not 

 divide the plate equally, as in superjacent plates, but falls nearer the 

 ventral than the dorsal border. On account of the alternation of position 

 of adjoining columns the ventral plate of column 3 is a small plate, d, rep- 

 resenting less than the upper half of a hexagon, just as plate c has less than 

 half of the lower half of a hexagon. Plate d has no primary spine boss, 

 it having been resorbed with the lower portion of the plate. The ventral 

 plate of column 4 is again a large, three-quarters plate e, like plate c ; its 

 ventral border has been considerably resorbed and the spine boss lies 

 near the ventral edge. The lowest plate,/, of column 2 is wanting, but 

 a vacancy in the specimen allows for its legitimate reconstruction, as in- 

 dicated by the dotted lines. The relations of this resorption of ventral 

 plates of the corona is shown especially clearly in Ctdaris florigemma 

 (plate 8, figure 47), where, on account of the size and ornamentation of 

 the plates, their removal by resorption is strikingly apparent. One 

 striking difference in these two types exists. In Cidaris (plate 8, figures 

 47 and 48), the smaller plates on the border of the corona have a large, 

 central spine boss, and secondary bosses much as in complete plates 

 farther dorsally. It is evident that here there has been a readjustment 

 of parts, so that though the original ornamentation was more or less re- 

 sorbed with the ventral border of the plate, a new series of spine bosses 

 have grown up to take the place of and occupy the same relative posi- 

 tion as the original tubercles. In a young specimen of recent Cidaris 

 papillata Leske (plate 9, figure 55) the spine tubercles are being resorbed 

 together with the ventral border of the plates, and no readjustment has 

 as yet taken place. In the specimen of Archxocidaris studied there was 

 no evidence of a readjustment of tubercles in the ventral plates. 



In ambulacrum B^ of Archseo cidaris (plate 8, figure 43), as in other 

 ambulacral areas, the plates consist of two columns, a b, of small, low 

 plates, each having two pores. The sides of the plates are parallel and 

 the form very regular, more so than in species of the Melonitidsd. 



In interambulacrum C we find in the ventral row of plates a condition 

 like that described in area A. Plate c of column 1 is wanting, and the 

 plate lying on its dorsal border is twisted dorso-ventrally, but the empty 

 space for c remains, and it can properly be reconstructed, as indicated by 

 the dotted lines. The ventral plate d of column 3 is present, and con- 

 sists of less than the upper half of a hexagon, as in plate d, area A. It 

 has no spine boss. The ventral plate e of column 4 is hexagonal, with its 

 ventral border partially resorbed, as in the same plate of area A. The 

 ventral plate/ of column 2 is less than the upper half of a hexagon, like 



