EEPOE.T ON THE CRINOIDEA. 261 



(PI. XLVIII. fig. 3). But in other cases they show a considerable alteration both in the 

 size and in the character of their component joints, as seen on the left side of fig. 1 on 

 PI. XXXVI. 



The double row of lateral projections on the joints of these proximal pinnules is 

 developed in rather a singular manner. Their basal joints are somewhat flattened 

 against the arm, and the upper edge of their broad dorsal surface is sharpened, and more 

 or less carinate, while its distal end is marked by a median process of variable prominence, 

 as is well seen in PL XXXVI. figs. 4-6. As the following joints lose their flattened 

 appearance, and become more rounded, the carination of the upper edge develops into a 

 strong blunt process at the distal end of the joint on its inner side ; while the medio- 

 dorsal prominence passes into a corresponding process on the outer side (PI. XLIX. 

 fig. 2). There is much variation, however, in the exact nature and mode of development 

 of these processes. 



The frecpiency of the ray-divisions of this species, and therefore the number of arms, 

 is subject to great fluctuations. A second post-radial axillary only occurs in the single 

 specimens which I named Antedon variipinna and Antedon crenidata respectively, and 

 sometimes also in the form which was described by Bell as Antedon irregularis. A 

 large number of individuals were obtained by the " Alert," and the majority of them 

 have two or more palmar series, though in others, as in the Challenger specimen (PI. 

 XLIX. fig. 1), palmars are entirely absent. Bell gives the number of arms as ranging 

 from eleven to twenty-two, but seems to have overlooked one example in which there 

 are twenty-five. The occurrence of an individual with only eleven arms makes it cpuite 

 possible that a ten-armed variety of this protean type may be eventually discovered. 

 In fact, the two individuals which I formerly called Antedon dubia are not improbably 

 of this nature. The one has two distichal series, and the other only one. But in each 

 case they result from regeneration of the arm at the syzygy in the third joint above the 

 radial axillary. This may perhaps have originally supported a distichal axillary ; or it 

 may have given rise to one arm only, which was replaced by two after fracture, as is 

 so often the case, an excellent instance of it having been described by Dr. Carpenter 

 in Antedon rosacea. 1 Under these circumstances I have therefore thought it safer 

 to assign Antedon variipinna a place among the ten-armed species, to which it can 

 definitely be referred if ever an individual is found in which distichal series are entirely 

 absent. No harm will be done if it never turns up, and should it do so, it will run less 

 risk of being baptised as a new species, having undergone that process too frequently 

 already. 



There is one point relating to the extremely variable characters of this species, which 

 seems to me to be of special importance. The variations which I have noticed above 

 are not altogether due to difference of locality. Varieties Nos. 5 and 6 were found 



1 Phil. Trans., 1866, p. 725, pi xxxviii. fig. 8, B. 



