2 ANNALS NEW YORK ACADEMY OF SCIENCES 



in another species. In I'xihiinmus. the nmst t'i-c(|ii('iit vnrialion in the 

 radial series was (hie to the absence of tiie first costal, or rarely, to the 

 presence of an extra distichal below the axillary one. A greater amount 

 of variation occurs in the number of interbrachials, as these are simply 

 space fillers, and iIhmi- number depends mainly on the relative height and 

 width of the cu]). The most marked variatixjn^ was found in the anal area. 

 Of the seventy-live specimens of Halu<rinus studied, Idiii- had two plates 

 in the second row of the anal series, as in the Ai'linocrinida'. and one of 

 twenty-five specimens of Cactocrinus had three plates in the second row, 

 after the manner of the Ratocrinida". As will be readily seen, this amount 

 of variation furnishes little evidence which can be used for working out 

 phylogeny, except in the most general way. 



The ornament' on the surface of closely related species differs in degree 

 rather than in kind, and hence does not show changes sufficiently definite 

 to furnish satisfactoiy results. 



The stems of crinoids often show very definite and well-marked changes 

 from the proximal to the distal portion, but it is so comparatively rare to 

 find the crinoid stems intact for any considerable distance from the calyx 

 that they are only occasionally helpful. 



A study of crinoid arms has shown that in some genera, at least, more 

 satisfactory results may be obtained. It has long been recognized that 

 the uniserial condition at the base of many arms which later become 

 biserial is reminiscent of ancestral species whose arms were uniserial 

 throughout. Applying the same principle to other characters, it is be- 

 lieved that changes in the form and ornament of the arm are indications 

 of changes through which the ancestors of a species have passed and that, 

 taken in connection with other characters, they may bo used to determine 

 the phylogeny of the group to which the species belongs. This method of 

 study will obviously be most useful in genera having highly modified 

 arms, and the genus Cactocrinus has furnished the material upon which 

 the present paper is based. No single character can be used alone in 

 working out relationship, and in this study constant reference has been 

 made to the characters of the calyx as well as of the arms. The columns, 

 in nearly all the specimens available for study, were not presented- 



In the descriptions of species of Cactocrinus which follow, a reference 

 is given to the original description and to a later full descri])tion and fig- 

 ures. Only such facts are added here as have a bearing upon the present 

 investigation, except in the case of new species and Cactocrinus prohosci- 

 (JaJis. which may be used, as a standard of comparison for other species 

 and is fully described. 



