1881. J NATURAL SCIENCES OF PHILADELPHIA. 185 



row of raclials. The formation of higher orders of radials takes 

 place in a similar manner, only one branch generally bifurcating, 

 and one remaining single. In species with but two orders of 

 I'adials the number of primary arms is limited to two in each 

 ray, which, however, sometimes branch after they become free. 

 Species with tertiary radials may have either three or four arms to 

 the ray, depending upon whether one or both divisions bifurcate. 

 Upon the same principle species with four orders of radials may 

 have five to eight arms, and those with a fifth order may have 

 from eight to twelve and even sixteen. The first number can onl}- 

 occur when all the branches remain simx^le ; the latter when thej- 

 all divide. Hence the number of primar}- arms is dependent upon 

 the number of orders of radials represented in the species, and 

 whether part of them bifurcate again or remain simple. It is 

 important to note that simple arms are always given off from 

 opposite sides alternately. 



The number of arms is most frequently only of specific import- 

 ance, but in cases where certain rays throughout a number of 

 species are distinguished bj^ a smaller or greater number, the arm 

 formula may become almost a generic character, as for instance 

 in Batocrinus^ where the posterior rays are generally more 

 developed. In Eretmocrinus the anterior and two posterior rays 

 are less developed than the antero-lateral rays. In some genera 

 the arms appear to be limited to a certain number, in others 

 the}^ vary. In the typical Actinocrinus there are species with 

 four, five, six, seven and eight arms to the ray, but the normal 

 number is uniform in the different rays. In all these cases, how- 

 ever, as well as not unfrequently within the limits of species, 

 there are exceptions. 



3. Interradials and Anals. 



The interradial and anal plates occupy the intermediate spaces 

 between the five radial divisions or raj'S in the bod}', and their 

 number and shape depends altogether upon the number, position 

 and. proportions of the radial plates. They vary from a single 

 plate to twent}' or more, but are represented even in the young 

 crinoid by at least one plate. In species in which the secondary 

 radials of adjacent rays abut laterally, the number of interradials 

 is naturally small, but when higher orders of radials are present, 

 and especially when the raj'S are widel}- separated, the number is 



