A MONOGRAPH OF THE EXISTING CRINOIDS 87 



in Capillaster sentosa, and is usually between 12 and 25. The length and stoutness 

 of the cirri and the number of cirrus segments are in general proportionate to the 

 size of the species or the individual. While this is more or less true in all the comas- 

 terids, it is especially noticeable in the Capillnsterinae. The earlier cirrus segments 

 are always somewhat, and usually considerably, longer than the distal, which bear 

 dorsal spines or tubercles. 



The inclusion of the genera containing species with only 10 arms in this sub- 

 family rather than in the Comasterinae calls for a word of explanation. The cirri 

 of these species are quite different from those found in the species of Comactiniinae, 

 since they always have dorsal processes on the distal segments. In the multibra- 

 chiate species of Capillasterinae the cirri, in contrast to the arms, are fairly uniform 

 in structure and also in relative size; they are absent in only a single species. The 

 10-armed comasterids possess cirri which, except for the attenuate cirri of the very 

 small forms, resemble those of the multibrachiate Capillasterinae. A close approach 

 to this type of cirrus is found in certain species of Comanthus, such as C. trichoptera 

 and C. japonwa, but the cirri of the 10-armed forms do not resemble the cirri of 

 these species so closely as they do the cirri of the smaller of the multibrachiate 

 Capillasterinae. Besides this, in the 10-armed genera the length, stoutness, and 

 number of segments in the cirri are roughly proportionate to the size of the species, 

 as in the multibrachiate Capillasterinae. In the Comasterinae, on the other hand, 

 the cirri are highly diversified. The distal segments usually bear dorsal processes, 

 but these are very variable, and may be absent, as in Comanthina schlegelii and 

 Comanthus bennetti. While the cirri may be very large and stout, as in Comanthus 

 bennetti and C. pinguis, they are commonly weak and poorly developed, more or 

 less deficient, or entirely absent. So there seems to be no escape from the conclu- 

 sion that the 10-armed genera should be included in the subfamily Capillasterinae. 



The single species of the curious genus Comanthoides usually has 10 arms, but 

 sometimes more than 10. In the latter case the arm division may resemble that of 

 Capillaster, or of the genus Comanthus in the Comasterinae. The species of 

 Comanthoides, therefore, would seem to be a form uniting, in its multibrachiate in- 

 dividuals, the Capillasterinae and the Comasterinae. In its cirri and pinnules, 

 however, it agrees very closely with the species of Comissia, and I believe that it 

 should be regarded as an aberrant member of the Capillasterinae rather than as a 

 connecting link between the Capillasterinae and the Comasterinae. 



The tropical American genera of Capillasterinae which include 10-armed species 

 Leptonemaster, Comatonia, Comatilia, and Micromatula are especially interesting. 

 In Comatilia we find the only comasterid with deficient pinnulation, there being no 

 pinnules between the first and the fourth pairs. The single species of Microcomatula 

 is the smallest comatulid known, be>ng somewhat smaller than the smallest adults 

 of the minute Compsometra parviflora. It has an arm length of only 12 mm. The 

 cirri are greatly attenuated, and the comb on the oral pinnules has only three teeth. 

 Sacculi, otherwise absent in the Comasteridae, occur in Comatilia and in Comatonia. 



It is worthy of note that in many of the small species of Capillasterinae the 

 mouth is always central, while in some of the larger ones, as Capillaster multiradiata, 

 it is frequently nearly or even quite central. In the species of Comactiniinae and 

 Comasterinae the mouth is always marginal or submarginal. 



