OAl-OARKOUS SKKI.ETON OK TllK ANTUO/OA. 533 



" cells " ill the positions (leseril)e(l by Kowler ; aiul, as far as 1 

 could determine, the "cells" in (,uestion were not neat 

 (;olnninur cells with deliuite cell boundaries, as figured lu 

 Fowler's diajifram, but resembled those of Mussa (vide fig. 29), 

 —being, in fact, long amoeboid, vacuolated, and partially fused 

 masses of protoplasm containing nuclei. The cellular cha- 

 racter of the calicoblasts is, however, more evident in l>oi)ho. 

 helia and in Mussa than in any other forms which 1 hive 



examined. 



Fig. 21) is a representation of the elongated and (piasi- 

 columnar calici.blasts of Mussa distans, drawn under a high 

 magnificatu)n. The section was stained by Ileidenhain's iron 

 hixjmatoxylin method. Under a lower power (120 diameters) 

 the calicoblasts appear to form a layer of simple columnar 

 epithelium, but a higher power (lOOO diameters) l)rint;s out the 

 characters shown in the figure. Thecellsare eonlluent, both at 

 their bases and at their free ends, especially the latter. The 

 alveolar structure of the cytoplasm is well shown, and besides 

 the minute alveoli there are a number of small vacuoles, each o( 

 which contains a microsome. The presence of vacuoles con- 

 taining mierosomata is characteristic of well-preserved calico- 

 blasts. A comparison with fig. 10 shows that there is no 

 essential dilleieiuH! between the calicoblasts of Mussa and 

 Caryophyllia, though the elements are much smaller in the 

 latter species. Fig. 21) does not show the presence! of an 

 external limiting membrane, but there are indications of it in 

 fig. 27, which was drawn from the same slide. 1 have reason 

 to believe that there is always an external limiting membrane 

 to the calicoblasts in Madreporaria, and that when it is absent 

 in sections it has been destroyed in the course of decalciliea- 

 tion. As has already been stated, a similar membrane occurs 

 in llcliopora. The limiting membrane is interposed between 

 the calicoblasts and the corallum, and may be regarded as a 

 sheath of the latter. It is clear that the carbonate of lime 

 secreted by the calicoblasts must pass through the membrane, 

 just as it has to pass through the spicule sheath in Aleyo- 

 Tluoughout my researches 1 have looked most 



uarians. 



