21 



Ihe iiol yel wliolly calcified |)arl of the Irontal iiiemltrane has loiined a compen- 

 sation sac by an invagination proximally to the operculum. 



The basal wall like the frontal may also be membranous, as in Membranipora mein- 

 branacea and Electru pilosn, and even in numerous, well-calcified, incrusting mem- 

 bers of the division Ascophora the basal wall is slightly calcified or jjarlly uncalcified, 

 sometimes even quite membranous. I may for instance mention the incrusting 

 species of the genera Eschaielbt. Escharina, Micropoielld, Hippothoa, etc. The ))asal 

 wall in tlie calcified state seems as a rule to be a Gymnocyst, and it is only in 

 very few cases that it is covered with a membrane in species appearing in free 

 colonies. Harmer' has for example shown that the free, one-layered colonies 

 of Euthyris dathratu and Eiith. obtecta (PI. XV, figs. 2c, 2d) are provided over 

 the whole surface with a covering membrane which is kept stretched by ])ro- 

 jections from the underlying Cryptocyst. A covering membrane over the whole 

 surface of the colony is also j)resent in Urceolipora nana (PI. XV, figs. 1 a — 1 e) 

 and it seems also to be found in species of the genus Cupiilaria. 



Under the names of Sleginoj)ora and Uisteginopora d'Orbigny- has described 

 a number of fossil species, which have possessed a double roof, of which the 

 lower except in .S7. irregularis seems to have quite the same structure as the roof in 

 Cribrilinidae and thus to be formed by spines connected with one another. On 

 first consideration of the drawings given we should be inclined to supjjose that 

 the upper roof is formed by calcification of such a projecting membrane as the 

 one we find in Eulhijris ohtecta; but according to Jullien's^ investigations the 

 upper roof is formed by a partial fusion together of very large spines, broadened 

 out and plate-like at the ends, and this view is undoubtedly right. That this roof 

 cannot be explained in the above-mentioned way appears clearly from the fact, 

 that the membrane which corresponds with the mentioned projecting cover in 

 Euthgris, has its place below the fused spines in Cribrilinidae. hi a number of 

 fossil (Iribrilina forms as well as in the one-layered Steginopora species we find 

 a varying number of robusj projections at the back of the single zocecia, which 

 Harmer' thinks have served as sup|)orts for a membranous cover, similar to 

 the one which is found in Eulhgris. Against this view speaks firstly Ihe circum- 

 stance, that while the mentioned sui)ports in Euthgris obteclu are slender, cy- 

 lindrical rods, the projections in the mentioned Cribrilina species, with which 

 Harmer compares them, have the form of tubercles, which are very differently 

 developed in number and size in Ihe different species, and their rounded end- 

 ])art does not seem to have been connected with a membrane. Further, these 



• 18, p. 16 and 19, pp 267, p. 277, 278: « 86, pp. 235, 498, 499; ' 44. p. 609; ' 18, p. H 



