134 MADKEPORAKIA. 



We begin our series, then, from an initial stock, which best carries on the series begun by 

 the explanate and massive forms. That would be one, from the surface of which there arose in 

 the centre or round the edges, more than one eminence or lobe. Such lobes would grow on 

 and fork, and eventually form a tuft of short thick branches ; several perhaps rise from one and 

 the same base. 



It is not necessary to repeat here what was said above, see p. 18, on the origin and forma- 

 tion of knee-bends. It has, however, to be kept in mind in all analyses of the growth-forms. 



a. With thick encrusting base from which more than one branching lobe arises. 



P. Cura^oa 1. The original figure (reproduced, PI. XVII., fig. 1), suggests that the 

 branches grew from the edges of a thick encrusting base. The perspective of 

 the figure, however, may be wrong. The regular repetition of the forking 

 suggests a more normal start. 



P. West Indies x. 15 (PI. XVII. fig. 20, 7). As this form grows round a Gorgonid 

 skeleton, it is difficult to say whether it is a normal growth or not. 



P. West Indies x. ^4 (PI- XVII. fig. 10). On the problems relating to its growth- 

 form, see text, p. 101. 



The following forms may belong liere or to the next group. The cause of the doubt is 

 that small stout lobes or branches may perhaps fuse together to form a solid mass, which is 

 not the true original base, yet looks like it. This problem of fusion deserves special study for 

 we shall find it occurring at one time as if it were normal, while at another the branches 

 seem to go out of their way to avoid it. In the following cases, though the short branches 

 toss about in all directions, they do not show sufficient signs of fusing to justify our con- 

 cluding that their bases are due to fusions. 



P. AntigvM 1 (PI. X. fig. 1). This stock having grown upon the side of an over- 

 turned earlier stock looks in the figure as if it belonged to this group, but see 

 next page, c ; ii. o. 



P. Barbuda 1 (PL X. fig. 3). The fact that this coral seems also to start from 

 more than a single lobe of a massive base, serves as another link to unite it with 

 P. West Indies x. 24.. 



P. Bermvda 1 (PI. XII. fig. 4). There are more signs of fusion in this specimen. 

 Yet a close examination of it suggests its having started from a stout ridge-like 

 basal form which broke up into short lobee, three to four in number. 



P. West Indies x. 11 (PL XV. fig. 2). This again looks as if it might have started as 

 an irregularly bluntly angular ridge, which gradually divided into elongating 

 and forking angular processes, every one of which retains the flattened appear- 

 ance of the ridge. 



Pontes x. 7 (PL XV. fig. 1). I provisionally place this in this group. It has all the 

 appearance of being a massive base, from which branching lobes spring up ; and 

 further, it has much more the character of a West Indian, than of an Indo- 

 Pacific, form. 



h. Single short thick knobs or lobes rise, and divide, and form somewhat massive tufts 

 with only slight traces of free branching. In other words, the forkings of the lobes follow 

 one another very rapidly and at very short distances apart. 



