HYPSELASTKK F KACILIS. W) 



to find. The former have valves about .20 mm. long, with the loop .04-.0S nun. 

 more, while in the latter, the valves are only about .10-. 12 nun. 



This interesting and well-characterized species is evidently closely Pel 

 to limicolus but aside from the differences in the pedicellariae, which are v<-ry 

 marked, the differences in the relative size of the petals and in the Dumber of 

 pore-pairs therein are important features. In rotundus petals I and V are just 

 one half as long as II and IV, while in limicolus I and V are nearly two thirds 

 as long as II and IV. In rotundus petals I and V have fewer than .65 as many 

 pore-pairs as petals II and IV while in limicolus the posterior petals have three 

 fourths as many pore-pairs as the anterior. The shape of the test readily dis- 

 tinguishes rotundus from any of the following species. 



Station 4946. Between Kagoshima and Kobe, Japan. Bott. temp. 68.7°. 

 39 fms. Br. s., brk. sh., p. 



One specimen. 



Hypselaster fragilis. 



Periaster fragilis A. Agassiz and Clark, 1907. Bull. M. C. Z., 51, p. 138. 



Plate 148, figs. 5-8. 



Length 16 mm.; width, where widest, back of abactinal system, 14 nun.; 

 height, back of abactinal system, 10.5 mm. Test highest just back of middle, 

 whence it slopes posteriorly and much more markedly anteriorly. The abac- 

 tinal system is anterior to the middle of test, the general form of which, with 

 the arrangement of its component plates will be most easily understood from 

 the figures (PI. 148, figs. 5-8). The anterior petals, especially III, are slightly 

 sunken and the interambulacra 2 and 4 are compressed near the abactinal 

 system to form low, rounded ridges. Petals I and V are only about 2 mm. long 

 and have only seven or eight pore-pairs on each side, while petals II and IN' 

 are over 5 mm. long and have 18-20 pore-pairs on each side. The periproct is 

 well up on the posterior end of the test and is visible from above; it is relatively 

 small, scarcely 2 mm. across. There are no genital pores and the outlines of 

 the genital plates cannot be clearly determined. Interambulacra 1 and 4 

 appear to be shut out from the peristome (PL 148, fig. 8) while dorsally they 

 seem to terminate each in a single plate (PL 148, fig. 7). Probably, however, 

 better material will show the usual pair of plates here. The peripetalous fasci- 

 ole is well marked and of nearly uniform width. About 3 mm. below the peri- 



