268 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



It is interesting to note that the species lacking cirri when adult are confined 

 to the East Indian region (extending westward to Ceylon) and to northern Australia, 

 while within this circumscribed area this feature is much more developed along the 

 Australian coasts then elsewhere. This is a fact of very great importance, as will 

 be explained under General Conclusions. 



The proportionate length of the cirri varies enormously; in some species of 

 Oligometra, m Antedon (&gs. 103, p. 165, 104, p. 167, 105, p. 169, and 106, p. 171), and 

 in Mastigometra, as well as in Comactinia (fig. 76, p. 129), they attain to only a very 

 small percentage of the arm length; they are here, however, stout and well adapted 

 for firmly fixing the comparatively slender and attenuated animals; in such genera 

 as Ptilometra (fig. 93, p. 153), Pterometra, and Asterometra (fig. 94, p. 155) and their 

 near relatives they attain a most extraordinary size, in Asterometra macropoda (fig. 

 94, p. 155) and in A. magnipeda being longer than the arms, sometimes as much as 

 one-fifth longer, and very stout. Generally speaking, the cirri are, on the average, 

 one-fourth or one-fifth of the arm length, as in the closely related stalked species 

 of the family Pentacrinitidse. 



The number of component ossicles in the cirri varies as much as the length; 

 while there may be in certain species not more than 6 or 8 (fig. 76, p. 129) , and very 

 often not more than 15, in Asterometra macropoda and in A. magnipeda there may 

 be as many as 120 or even more (fig. 94, p. 155). 



Fundamentally the cirri are simply somatic outgrowths from the body wall 

 normally ( as is indicated by their occurrence singly on nodal columnals) one to each 

 somatic division of the body. They are strictly comparable to the lateral somatic 

 outgrowths along the sides of the body in the arthropods, though they have become 

 so altered as to have lost almost all resemblance to the ancestral type. 



In the arthropods these lateral body processes occur normally in a lateral or 

 ventrolateral line, and are commonly double, arranged in two series, one above the 

 other; they occur in the mid-line of each segment. 



In the crinoids the cirri are dorsal, arranged in a circle of small diameter about 

 the extreme dorsal apex of the animal, and are normally single, though they may 

 be doubled or still further reduplicated. In the so-called monocyclic forms, con- 

 fined to the earlier horizons, they are interradial or midsomatic; in the comatulids 

 and in the pentacrinites they are always intersomatic, occurring in the radial areas 

 of the dorsal apex of the body. 



The anomalous position of the crinoid cirri, which are confined to the dorsal 

 apex of the animal, is easily accounted for. The cirri represent the dorsal row of 

 lateral processes in the articulates, while the coronal plates, as previously explained, 

 represent the ventral; the crinoid arms originated from a third row of similar body 

 processes which was essentially a duplication of the second, while the orals represent 

 a fourth, which again was a duplication of the third. 



If the crinoid cirri are true somatic processes they would naturally be expected 

 always to be interradial or midsomatic in position. But such is the case only in 

 the fossil so-called monocyclic forms. In all the recent types in which cirri occur 

 they are radial or intersomatic in position. 



