CALCAREOUS SKKLETON OP THK ANTHOZOA. 535 



traces of tlie calcareous skeleton of Madreporaria arc wliat 

 von Koch described, sphseroids of calcium carbonate. I 

 shall show presently that similar sphpcroids may be observed 

 on the septal surface of the adult Caryophyllia. At the 

 same time the ''calcareous scale" figured by Mrs. Gordon 

 for Galaxea exists, and calls for explanation. I have said 

 enough to show that the scales in question are certainly not 

 calcified desmocytes. I have found these scales in preparations 

 of the vesicular endotheca of Galaxea laperonsiana made 

 for the purpose of verifying Mrs. Gordon's statements, but I 

 have found them best developed in preparations of Madrepora 

 subulata, M. rosacea, and M. hyacinthus, taken from 

 the collection of the late Mr. Geo. Brook. In tolerably thick 

 sections, longitudinal or transverse, of any of these species, 

 the calcareous scales arc very evident at the sides of the septa, 

 ncfincnchymal trabeculse, and costal spines included in the 

 section. The outer edges of the scales arc commonly coloured 

 bright crimson in specimens stained in borax carmine (see fig. 

 41), but this effect is not produced by aniline dyes, and after 

 some study I became convinced that there is no staining of an 

 organic residue, but a deposit of carmine particles on the 

 edges of the scales. But the whole structure is so striking 

 that it seems to afford positive proof of the statement that 

 " the growth-lamellae of Madrepora from first to last are 

 composed of coalesced calcified calicoblasts, which once repre- 

 sented living ectoderm '' (Mrs. Gordon, loc. cit., p. 213). One's 

 belief is shaken, however, by the study of sections, both of 

 hard and soft parts together and of decalcified specimens. 

 The layer of continuous vacuolated protoplasm shown in fig. 42 

 lends no support to the conception of successive shedding off 

 of calcified cells from a definite cell layer, nor does a side-by- 

 side comparison of as much of the calicoblasts as can be seen 

 in a ground-down section with the adjacent corallum. A 

 portion of such a section, magnified 420 times, is shown in 

 fig. 41. And here let me say that neither in this nor in any 

 other sections of several species of Madrepora could I find 

 any trace of a continuous layer of desmocytes. Here and 



