286 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



without forming pointed spines. This is the case only in those species in which 

 the cirri are short and the possible scope of intersegmental motion is very limited, 

 the action of the cirrus as a whole being largely localized in the basal segments; 

 there has therefore been no opportunity for the formation of everted distal dorsal 

 edges to the segments, though the sharpening of the median dorsal line has taken 

 place as usual. 



In most species with very long cirri, as evidenced particularly by species of 

 Perometrinse (fig. 387, p. 307) and Thalassometridse (figs. 363-368, p. 297), there 

 has been, in the distal portion of the cirri, a combination of these processes; dorsal 

 spines have been acquired through metamorphosis from a primitive transverse 

 ridge; but in the outer segments there has been, due to the shortening of these 

 segments and the progressively more and more ventral position occupied by the 

 transverse fulcral ridge, a considerable tendency toward an excess of the dorsal 

 deposit of stereorn, so that the spines are more or less masked by the resultant high 

 carination, which as a rule reaches to their apices, and the dorsal processes assume 

 a form resembling that of the teeth of Serrasalmo. 



Typically the cirri may be said to consist of from 15 to 20 segments with 

 longitudinally straight sides and meeting end to end without overlap, the first 

 two segments short, the third about as long as broad, the following three slightly 

 longer than broad, then gradually becoming slightly broader than long; as the 

 segments begin to decrease in length their distal dorsal edges thicken and gradually 

 come to project, especially in the median dorsal line; the cirri are at first broadly 

 oval, often nearly circular, in cross section, but soon become somewhat flattened, 

 though still regularly oval, and after the first appearance of the distal dorsal 

 processes more flattened, and in cross section somewhat pointed dorsally. 



This typical or average type of cirrus, which careful study has indicated as the 

 primitive comatulid type of cirrus, differing but slightly from the generalized 

 pentacrinite type as found in Teliocrinus (fig. 127, p. 197) or in Ilypalocrinus, 

 does not occur in any known form, though in certain of the genera both of the 

 Oligophreata and of the Macrophreata the cirri of some species approach very 

 closely to it. Among the oligophreate genera most of the species belonging to the 

 family Charitometridse (fig. 369, p. 299), as well as those of the genus Catoptometra 

 (fig. 334, p. 283) and certain species of Comanthus (as for instance Comanihus 

 parvicirra) (fig. 331, p. 281), possess cirri close to the primitive' type, while the 

 same is true of some of the species of Antedon (fig. 312, p. 271) and of Mastigometra 

 among the macrophreate forms; but in all of these genera there is more or less 

 deviation in various directions. It is somewhat remarkable that these six genera, 

 all of which are highly specialized, and so widely different that they must be placed 

 in two distinct suborders and four families, should have departed so slightly from 

 the primitive cirrus structure as deduced not only from a critical comparative 

 study of mature cirri, but from a study of the ontogeny of the cirri in all the groups. 

 Their cirri might be supposed to have converged from entirely different types 

 toward a common central type as a result of similar requirements; but if this were 

 so we should expect the cirri of the young, or immature, or regenerating cirri, to 

 recapitulate these ancestral forms before reaching the mature form, but nothing 



