46 CANADIAN FOSSILS. 



The upper termination of the cells (fig. 6c) again expands, so as 

 to form broad and closely placed tesserae, with only small cavities 

 between them. These were of course, if the fossil be a Foraminifer, 

 filled up with solid calcareous matter, now removed. 



One character in particular which serves to connect the fossil 

 with its living analogue is the very considerable space occupied by 

 the animal matter — now solid spaces filled with silex — both on the 

 upper and lower surfaces. 



It forms, as above noted, rhomboidal plates on the lower surfaces, 

 which plates are somewhat imbricated ; and the concavities within 

 them (not distinct cells, as in compound Orbitolites,) communicate 

 with each other at one or more of the angles (figure 7) very much in 

 the way shewn by a section of Orbitolites (see Carpenter, 1. c. } 

 plate 6, figs. 1, 2), while the intervening calcareous walls are linear 

 and thin. (Sometimes, as in R. australis, figs. S, 9, these basal plates 

 are lobed.) 



If, instead of comparing Receptaculites with the Orbitolites, we 

 should suppose it related to any of the millepore corals, or still more 

 probably to such a form as that of the purple organ-coral {Tubipora 

 musica), the reverse of all this would be the case. There should be 

 a calcareous plate or epitheca on the lower surface, from which the 

 tubular corallites would spring, and the walls of these latter, however 

 thin, ought to be visible in the transverse section, which we do not 

 find to be the case in slices of the columns viewed by transmitted 

 light. Again, the upper extremity of the tubes should be open, not 

 closed by convex plates as in the fossil, since such could only be the 

 case when the walls of the coral-cells were so greatly thickened as 

 nearly to close the mouth, while we have seen that in the fossil the 

 corresponding part expands, and is covered over by a definite and 

 often lobed plate, to all appearance continuous with the walls of 

 the cells. 



If any analogy be suspected between Receptaculites and such 

 Palaeozoic corals as Halysites (the chain-coral), or with Syringopora? 

 there should, besides the characters above mentioned, be indications 

 of transverse plates which have never yet appeared. 



I believe all this applies equally to the Receptaculites Neytuni of 

 the Devonian rocks of Belgium, but there is some appearance of a 

 thin investment of the columns in transverse sections of that species 

 which requires further investigation. And as it is of a deeply 

 infundibuliform shape, there is of course a possibility that it and the 

 other Receptaculites may be very regularly formed sponges ; but I 



