Madreporarian Genus Turbinaria. 513 



separated by chambers or fissures. These flat coralla, seldom 

 noduh^ted, are often of great thickness and strength, as indeed 

 their form requires. They appear very often to be semi- 

 circular, as if their shape were adapted to horizontal growth 

 from a more or less vertical substratum. Such a horizontal 

 growth requires far greater strength than does an erect frond. 

 Consequently a section through such a tabulate form shows 

 great thickness of the coenenchyma both above and below the 

 line of intercommunication between the polyp-cavities. The 

 texture of the coenenchyma is also very massive and dense. 



In this case and in the next it seems to me as if we have 

 periodical growths, or, at any rate, such a modification of the 

 usual progressive dying down that it practically amounts to 

 periodicity. 



Seventh Type of Growth: Glomerate. — I was for a long 

 time inclined to consider all glomerate forms as mere varieties 

 of other species, varieties which had become glomerate owing 

 to some accidental influence, perhaps of the form of the sub- 

 stratum ; and it is undeniable that the likeness between certain 

 glomerate forms and other Turbinarians found growing near 

 them is very great. But this resemblance admits of another 

 explanation, and will be referred to again. On the other 

 hand, if any classificatory value is to be placed upon methods 

 of growth, and I do not see how this can be disputed, we are 

 bound to look upon the glomerate type o£ growth as one of 

 the most marked and peculiar. 



The corallum expands very little superficially. The 

 coenenchyma is built up, as it were, in situ. It is therefore 

 always irregularly reticulate — that is, it shows none of the 

 regular channels which indicate streamings. The consequence 

 is that the corallites have to lengthen enormously to keep their 

 apertures at the surface of the ever-thickening coenenchyma. 

 The budding of the polyps is, however, of the usual Turbi- 

 narian type, as is well shown in a section revealed by a 

 broken specimen in the National Collection (PI. XIX. fig. 7). 



Here again it appears as if we have periodical growth. A 

 fine specimen in the collection (fig. 8) shows three successive 

 growths in vertical series. It appears as if each new growth 

 must have started from the highest point of the old (perhaps 

 from adventitious budding in the thickest part) and crept 

 slowly out in all directions, covering it up. 



We here have an excellent illustration of the great import- 

 ance of recognizing clearly the essential morphology of the 

 Turbinarians as shown in their method of budding. The 

 method of budding and of growth of these glomerate Turbina- 

 rians is quite definite and distinct, and, in spite of the occa- 



