112 bulletin: museum of comparative zoology. 



r 



posed to distinguish Ophiopeza from Pectiuura (J., e. the absence of supplementary 

 oral plates) is not constant, as he found specimens of Ophiopeza which had one or 

 more such plates. 



In 1872 Ljungman adopted most of Liitken's suggestions but added two new 

 genera (for which, however, he unfortunately failed to designate types), one 

 (Ophiopezella) appears to be for his species, 0. spinosa, and the otlier (Opliiarach- 

 iiella) for 0. gorgonia (and possibly 0. infernalis and 0. septemspinosd) . Lyman 

 never accepted either of Ljungman's genera, or Grube's Opliiochasma, or Liitken's 

 Ophiopsammus, but he followed Liitken's lead as regards Ophiarachna's associa- 

 tion with Ophiocoma, and in recognizing the two genera Pectiuura and Ophiopeza. 

 He, however, refers several times to the unsatisfactory nature of the character 

 which was supposed to distinguish the two latter. Except for the recognition of 

 Ophiopezella by de Loriol aud Koehler, Lyman's classification has been used with 

 scarcely any modifications down to the present day. Several writers have re- 

 ferred to the inconstancy of the presence or absence of supplementary oral plates, 

 but no one has ventured to attempt a different grouping of the species. De 

 Loriol, some years ago, called attention to the close resemblance between 

 Pectiuura and Ophiarachna, but it has apparently been agreed that such resem- 

 blance was simply parallelism and not an indication of relationship. 



Taking up the question de now, I have been forced to the conclusion that 

 Liitken's separation of Ophiarachna from the Pectiuura group and its association 

 with Ophiocoma was unfortunate, and can only be rectified by a complete return 

 to the position of Miiller aud Troschel that 0. incrassafa is closely allied to 

 0. gorgonia and its allies. If we consider the morphological characters of the 

 three genera concerned we find that they all have the disc covered with scales, 

 which are concealed by a close grauulation that often covers the radial shields 

 and may even extend out a little way on the arms. In Ophiocoma and its allies 

 the arm-spines arise from a ridge occupying approximately the vertical mid-line 

 of the side arm-plate, and they stand out at nearly a right augle from the surface of 

 the plate. In Pectiuura aud its allies the arm-spiues arise from the distal margin 

 of the side arm-plate and are generally more or less flattened and appressed to the 

 arm ; not infrequently there are notclies in the proximal margin of the next plate 

 into which they may fit. At first sight the arm-spines in Ophiarachna appear to 

 be of the Ophiocoma type, for they are long and thick and often stand out at 

 a decided angle from the arm. It was mainly on the strength of this character 

 that Liitken and Lyman agreed on the association of Ophiarachna with Ophio- 

 coma, but a careful examination of the genera concerned has satisfied me that 

 even in its arm-spine arrangement Ophiarachna is nearly allied to Pectiuura. 

 While there is, in most specimens, a distinct ridge from which the spines arise, 

 it does not occupy the median part of the plate as in Ophiocoma, but is really 

 only the thickened distal margin of the plate ; aud furthermore, in many, if not 

 in all cases, the proximal margin of each side arm-plate shows notches like those 

 found in Pectiuura. It is clear then that in Ophiarachna we have a genus allied 

 to Pectiuura, in which the increase in size of the arm-spines has led to a super- 



