254 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



have been; specimens of species of Comasteridae or Thalassometridse, as well a3 of 

 Himerometridse or Colobometridas, maj r be recovered from a mass of laval or coral 

 detritus which has been turned over and over in the dredge, and yet have practically 

 all the cirri intact. This is the more remarkable in the Thalassometridse, in which 

 family the sometimes enormously long cirri are often very slender. But in the 

 Macrophreata the cirri are deciduous and, besides, very brittle, so that it is very 

 difficult in many cases to recover any of them at all. This is the more unfortunate, 

 as the presence of the smaller apical cirri is such anomalous genera as Psathyro- 

 metra, Zenometra, Atopocrinus and Atelecrinus would give us a valuable clue to 

 their systematic affinities. 



There is a great difference in the facility with which cirri are lost in different 

 genera, and this is always correlated with a corresponding facility of fracture in 

 the brachial syzygies. As a general rule the genera in which there is the most 

 resorption of the dorsal pole and the most proportionate increase in the thickness 

 of the walls of the centrodorsal have the most tenacious cirri; but this is to be 

 expected, since these genera, by these very characters, show the greatest approach 

 to the Oligophreata. Large species are less likely to lose their cirri easily than 

 smaller ones in the same genera, and in the same species large specimens are usually 

 more nearly perfect than smaller ones; but here again the large species or the large 

 specimens take on certain oligophreate characters. The very small species, again, 

 are less liable to lose the cirri than the others on account of the immunity conferred 

 by their size. 



Of all the macrophreate comatulids the species belonging to the subfamily 

 Antedoninse are the least liable to loss of cirri, with the species of Perometrinse a 

 more or less close second. The species of Bathymetrinse usually have at least some 

 of the cirri present, although they are quite unknown in one of the species of Bathy- 

 metra. In the species of Heliometrinse cirri are rarely found in place; so far as I 

 have seen, when taken under ordinary conditions, not more than one in five or six 

 hundred specimens of the species of Solanome.tr a, Heliometra, or of Hathrometra have 

 any cirri at all, and I have never seen a single specimen of any species of any one of the 

 three genera with the cirri perfect, although I have examined probably at least 50,000 ; 

 Promachocrinus agrees with Heliometra in this respect, as would be expected, but 

 the cirri of Isometra and of Trichometra appear to be somewhat more tenacious, 

 though the cirri of several species of the latter genus are as yet unknown. This 

 apparent tenacity may, however, be due in part to the fact that these genera com- 

 monly inhabit softer bottom. The Thysanometrinse are, as a whole, like the Helio- 

 metrinse, though none of them retain the cirri so well as Trichometra; the cirri of 

 one of the species of Iridometra are unknown. In the Zenornetrinse, Atelecrinidse and 

 Pentametrocrinidse specimens retaining even the basal segments of the cirri are 

 very rare, so that we are quite ignorant of their structure in half of all the known 

 species, including three entire genera. Of the seven genera, in only two, containing 

 two species each, are the cirri adequately understood. 



It is a fortunate circumstance that in two of these three groups with very 

 deciduous cirri the centrodorsal is of the highest systematic value, presenting much 

 more important characters than the cirri. 



