REPORT ON THE CRINOIDEA. 49 



series present, so that the number of arms is reduced to thirteen, 1 and Bell has himself 

 examined individuals with less than twenty. 2 According to his rule, therefore, the 

 formula should be — A'. (3). (3). but in the formula which he actually gives the brackets 

 are altogether omitted. I should write it myself as — a. 3. (3), to indicate that whde some 

 distichal series are always present in every individual, palmar series may occasionally be 

 entirely absent. This appears to me to be the only possible way in which brackets can 

 be profitably employed. Bell, however, thinks otherwise, as is shown by the following 

 passage : 3 — 



" From the table of Antedon formulae some facts become at once apparent : — 



" (a) There are six examples among the more than ten-rayed forms in which the arms 

 are not a regular multiple of ten — that is, not 20, 40, or 80 ; this is clear from the sign 

 for the palmar or post-palmar being in these cases placed within brackets." 



The first line of this passage contains a repetition of an error in terminology which 

 was made by Bell in 1882, 4 and was afterwards corrected by myself. 5 He seems, how- 

 ever, to consider the point an unimportant one and continues to use the expression to 

 which I took exception. There are no ten-rayed forms of Antedon, though there are 

 plenty which are ten-armed. The arms were clearly distinguished from the rays by 

 Midler, who laid the foundation of the descriptive terminology now in use for the 

 Crinoids. But Bell persists in using the word rays when he only means arms. This is 

 unfortunate, as it leads to confusion between the five-rayed but ten-armed Antedon and 

 the truly ten-rayed Promachocrinus, a point to which I have before alluded. 



Bell has evidently made the generalisation quoted above on the basis of his formulae, 

 without special reference to the individuals he examined. He describes his single 

 specimen of Antedon gyges as having forty-one arms, and I find this to be due to the 

 presence of one post-palmar series, of which Bell's formula gives no hint. He is thus 

 able to include this type among those forms in which the arms are a regular multiple of 

 ten, i.e., forty. Then again he gives the formula of Antedon articulata as A.2.2. But 

 the exact number of forty arms which this expression denotes does not occur in his 

 specimen, which also has one post-palmar series ; while I have seen individuals with less 

 than forty arms. According to Bell's own system the formula of this type and perhaps 

 also that of Antedon gyges should be A. 2. (2). (2). We find then that not only on the six, 

 but in all the eight multibrachiate forms of Antedon for which he gives formulae the arms 

 are not a regular multiple of ten. But this is in no way a specially remarkable fact. The 

 singularity would be if the number of arms always were a regular multiple of ten, as is 

 generally though not always the case in Aetinometra paucicirra (PL LIV. figs. 1, 2). 

 But this is a most exceptional species. No one can examine any large collection of multi- 

 brachiate Comatulae without becoming immediately aware of the extreme irregularity in 



1 Trans. Linn. Soc. Land. (Zool.), 1879, ser. 2, vol. ii. pp. 51, 52, pi. ii. fig. 9. 2 " Alert " Report, p. 168. 



3 Ibid., p. 155. 4 Proc. Zool. Soc. Lond., 1882, p. 532, note. 6 Ibid., p. 732, note. 



(ZOOL. CHALL. EXP. — PAET LX. 1888.) OOO 7 



