XX THE VOYAGE OF H.M.S. CHALLENGER. 



The whole group of chelse, with their almost endless modifications, may be divided 

 into two subgroups according to whether the two ends of the spicule are equal 

 (Fig. v., 5 ; Fig. VI., 2, 2a) or unequal (Fig. V., 6 ; Fig. VI., 1, la); to the former class 

 the name isochelse is applied, and to the latter the name anisochelw. 



The most complex form of chela as^yet known to us appears to be that of Melonanchora 

 elliptica, figured and descrilied by Mr. Carter.^ Not infrequently the anisochelse are 

 found in " rosettes" (PI. XVII. fig. 7), all adhering together by their small ends, which 

 are attached to a central, granular (?) mass; no satisfactory explanation has, so far as 

 we are aware, as yet been given of this state of things, but possibly the rosettes are com- 

 parable to the toxodragmata and sigmadragmata, which we have described above, and 

 orio-inate like them by the development of a number of spicules in one and the same 

 cell. Mr. Carter also records this phenomenon in the case of the isochelse of Desmacidon 

 tituhans, Schmidt.^ 



In the embryo of Esperella nnammiformis we have succeeded in tracing the 

 development of the chelse in a very interesting manner. This developmental history 

 throws an important light on the relationships of the chelae to other forms of micro- 

 sclera. In the earliest stage observed the sj)icule consists of a sim^jle, slender shaft, 

 very slightly curved in the same direction at each end, and also pointed at each end. It 

 appears that the two pointed extremities then curve sharply inwards so as to form each 

 an acute angle with the shaft, thereby giving to the spicule the appearance of a simple 

 sigma with short, sharply recurved hooks and almost straight shaft. At first, as already 

 noted by Carter,^ the two ends are equal, but this condition does not persist for long. 

 The teeth or palms are now developed ; the central or anterior palm is formed by a 

 delicate, oval, flattened outgrowth from the end of the hook (possibly also in part from 

 the sides of the hook), lying in a plane at right angles to the plane of the original, 

 simple, hooked shaft. The lateral palms are formed as lateral outgrowths from the two 

 ends of the straight portion of the shaft. The hooks of the original spicule form the 

 median falces and tubercles of the adult. Thus it appears that the chelate spicules 

 develop from sigma-like forms by the formation of flattened outgrowths or buds at the 

 two ends ; hence we are justified in grouping sigmata and chelse in the same category. 

 In Esperella mammiformis the lateral outgrowths of the shaft (lateral palms) remain 

 connected with the shaft even in the adult, the spicule (PI. XV. figs. 18, 18a) being palmate; 

 but in other species, as already noted, they often become cut away from the shaft 

 and form distinct teeth (e.g., Cladorhiza tridentata, woodcut. Fig. VI., 1, la ; PL XXI. 

 fig. 20); new lateral outgrowths may then be developed on the shaft, and these may 

 again become cut away as teeth {e.g., Meliiderma stipitatum, PI. XXI. fig. 14), or 



1 Ann. and Mag. Nat. Hist, ser. 4, vol. xiv. p. 212, pi. xiii. figs. 9, 10, 11, 12. 



2 Ann. and Mag. Nat. Hist., ser. 5, vol. ix. p. 298, pi. xii. fig. 24, h. In this paper will be found a good deal of in- 

 teresting information concerning the chelfe. 



^ Ann. and Mag. Nat. Hist., ser. 4, vol. xiv. p. 102, pi. x. fig. 13. 



