Madreporarian Genus Turbinaria. 505 



radial growth, diverged sufficiently from their next neighbours 

 to admit of buds appearing between them, these appear, while 

 the parent polyps bend sharply upward towards the axis of 

 the cup. As soon as this bending is effected, a fresh bud or 

 fresh buds grow out close to the bend {cf. PI. XIX. figs. 3, 

 2th). This new generation (or incomplete ring) of buds may 

 remain for a time hidden in a ridge of the coenenchyma, which 

 then forms the edge of the cup. 



Without the bending up of the polyps as each series ceases 

 to form the growing edge, it is clear tliat the cup-shape could 

 not be maintained, the corallum would droop and curl under 

 on all sides. It was this more or less sudden bending up of 

 each polyp-cavity, with the bud starting from the bend, which 

 led Ehrenberg to describe it as stolonformatlon, and not 

 gemmation, and Dana to assert that the gemmation was 

 basal. The more or less sudden bendlngs upward of the 

 polyps were very naturally mistaken for the bases of the 

 corallites. This, however, is obviously not the case, if the 

 process be followed up from its starting point, viz. the budding 

 of the axial polyp of the stalk of the young cup. 



In this way, then, by the continual addition of a fresh 

 series of polyps outside the one last formed, the edge of tlie 

 corallum grows outwards into an ever-expanding cup or disk. 

 The Flowing of the Coenenchyma and the Thickening of the 

 Stalk. — It is obvious that increase of size of the cup or disk 

 requires a stouter stalk and walls — that is, the basal region of 

 the cup has to be thickened. In Madrepora, as already de- 

 scribed, the basal thickening of each upright branch (consisting 

 of an axial polyp surrounded by irregular tiers of daughter- 

 polyps) can often be seen to submerge the lower earlier-formed 

 buds (fig. 3 a). The downward streaming of the fluids can be 

 gathered from the longitudinal channels between the costte 

 and from the gradually increasing density of the coenen- 

 chyma. The same is the case with Turbinaria ; while the 

 coenenchyma of the growing edge of the corallum is spongy 

 both inside and outside of the cup, a short distance from 

 the edge it is furrowed by a system of channels runnino- 

 downwards. The channels are separated by ridges which 

 are, as we have seen, the most distal edges of the costse. In 

 nearly all cases the gradual thickening of these costaj can be 

 followed from thin echlnulate ridges, not thicker than septa 

 near the growing margin, into dense masses in the stalk. 

 This downward flow of matter, however It is to be explained 

 physiologically, is a very striking feature in the Turblnarlans. 

 Within the cup it frequently fills up the bottom, often com- 

 pletely submerging all the polyps which formed the cup at 

 Ann. & Mag. N. Hist. Ser. 6. Vol. xv. 35 



