684 II. KIRKPATRICK. 



vertical sections of the calcocytes on the surface of a crypt; 

 here the nucleus is pressed outwards or away from the 

 skeletal face of the cell to form a little mound on the opposite 

 face. The most difficult part of this investigation was the 

 discovery of the precise nature of the surface of the masses of 

 crypt-tissue. This difficulty was partly due to the slowness 

 with which fi.^ing reagents ])enetr;ited into the nearly inacces- 

 sible crypts, and partly, I think, to actnal variation in the 

 condition of the surface-cells owing to functional causes. 

 PI. 34, fig. 6, shows a well-fixed Fleunning preparation, and 

 PI. 36, fig. 1, a badly fixed osniic-picro-carmine one. Some- 

 times flattened surface-calcocytes are very well shown; at 

 other times one can make out only a diffused layer of granules, 

 the boundaries of cells being invisiijle. 



There is a possibility that the large inner cells may be 

 trophic stores, but in my opinion the surface crypt-cells are 

 amoebocytes of^ the same nature as the inner ones, and also I 

 think they are calcigenous, and not l)y secretion, but by 

 actual transformation. 



Fig. C on PI. 36 shows the calcocytes forming an epi- 

 thelium-like layer closely moulded to one of the calcareous 

 bars of a polygonal mesh at the growing edge of a crust of 

 Merlia. 



The theory that the calcocytes become wholly transformed 

 into calcified lumps oi- scales is confronted Avitli one difficult}^, 

 viz. to account for the fan-like, fibrillar structure of the 

 skeleton. I can only suggest that possibly the fibrillation is 

 a secondary change. A decalcified section of a tubercle (PI. 

 34, fig. 3h) shows radiating, fibrillar structure apparently 

 corresponding with that of the hard skeleton. 



A propos of the calcocytes, it only remains to say that the 

 cylindrical, moniliforin masses of these cells are simply basal 

 prolongations of the s[)onge body in which they have accumu- 

 lated. The cells are imbedded in the finely fibrillar maltha, 

 just as are the rest of the tissues of the sponge. Occasionally 

 the maltha in the crypts contains coUencytes, scleroblasts, 

 and spicules, and, veiy raiely, undeveloped wedge-shajied 



