CARPENTER. — FEEDING REACTIONS OF THE ROSE CORAL- 159 



While the reaction in the zooid actually touched might be accounted 

 for by the direct stimulation of the endodermal muscles by mechanical 

 pressure exerted through the overlying tissues, this explanation will 

 not apply to the reaction in distant polyps. Nor can the movements 

 in the outer portions of the colony be the result of the strain exerted 

 by the contracting central polyp on contiguous polyps, and by these, 

 when they in turn contract, on zooids still more distant. For, as We 

 have seen, a central polyp in a colony, contracting because of the direct 

 application of a chemical stimulus, does so without affecting the re- 

 tractor muscles of neighboring polyps. And, finally, the wide-spread 

 reaction cannot be explained by assuming that the protoplasm of the 

 colony is remarkably sensitive to slight molecular vibrations set up in 

 it by touching one of the polyps. If this were so, the colonies would 

 constantly be responding by obvious muscular contractions to vibra- 

 tions in the surrounding medium or in the solid substratum on which 

 they rest. That such reactions do not occur is evident from repeated 

 observation. 



When we turn from these experimental evidences of nervoid trans- 

 mission to the histology of the Madreporaria, our attention is naturally 

 directed to the mesogloea, through which we infer the impulses must 

 pass, and in which we should, therefore, expect to discover at least 

 some trace of transmitting tissues. According to Duerden (•.02) "the 

 mesogloea of coral polyps has generally been described as a perfectly 

 structureless layer " ; but the following quotation from page 416 of his 

 monograph on West Indian Madreporaria shows that this statement 

 does not hold for the form with which we are concerned : 



" In large polyps, such as Isophyll'm dipsacea, and also in Maeandrina, 

 the mesogloea is rather thick, and minute connective-tissue cells occur 

 sparsely throughout. In sections the cells are circular or oval in 

 shape, with a central nucleus, and minute prolongations extend in all 

 directions ; many of these reach one or other of the surfaces of the 

 layer, and there come into contact with the ectodermal or endodermal 

 cells. In some instances the processes extend right across from one 

 layer to the other, but are mostly disposed in an irregular stellate 

 manner." 



These " connective-tissue cells " are probably homologous with those 

 forming the mesogloeal cell plexus in the related Alcyonaria. This 

 plexus was formerly considered to be nervous in function, but the 

 studies of Pratt (:05) on the digestive organs of the Alcyonaria indi- 

 cate that the amoeboid cells whose processes give rise to the network 

 are concerned with the ingestion and transportation of food. By means 

 of these wandering cells " nutriment may be conveyed from the diges- 



