18 MADREPOEARIA. 



be almost as clear throughout as it seems to be in these "West Indian forms. I say " almost " 

 because there is an admitted stiffness of growth shown in the group here dealt with, wliich 

 contrasts with the plasticity of the Indo- Pacific forms.* 



Eeferring the reader to § IV., p. 20, for further discussion on this elementary growth- 

 principle we turn to a brief review of some of the variations shown in the forms of the 

 branchings tocks. 



The possible variations in method of forking are, of course, very great, even assuming that 

 the division is always into two, whereas, in passing, it may be noted that here and there (e.g. 

 P. Florida 1, PI. XII. fig. 2) there are traces of division into more than two, traces which 

 require further investigation. The vast majority of the forms noted in this Volume certainly 

 divide typically into two prongs, and only two. 



There may be endless differences of height, thickness, and roundness of the initial or 

 basal stem. If thin and long, their repetitions result in thin open stocks, branching only at 

 considerable distances apart. If the forking is at a small angle, the stock towers upwards ; if 

 the angle is wide, the stock is divaricate. If the initial stem is very short, forking almost 

 immediately it rises above the ground, stunted almost nodular stocks may be produced ; and 

 so on. While this repetition is the essential principle underlying the growth-forms of the 

 West Indian branching Pontes t and supplies us with taxonomic characters for their comparative 

 descriptions, certain interesting modifications have to be noted, modifications which may be seen 

 to be quite natural on the principle above laid down. The most important is the habit of 

 aborting one prong. The one which survives is, as a rule, that which is nearest to the vertical, 

 and into it the vegetative energies of the colony are concentrated, thereby insuring the successful 

 upward growth of the stock. A reason for this abortion, at least in certain cases, may perhaps 

 be found in the fact that a position far out of the vertical is not favourable to the growth of the 

 prong, treated as a new unit and trying to repeat the initial or basal colony ; hence its inability 

 to grow. These aborted prongs may be found in all stages from knobs, small mammillate pro- 

 cesses, or as mere bends in an apparently smooth stem. We would explain the mammillate 

 processes on such an extreme form as P. Barbados 6, PI. IX. fig. 5, as due to the abortion of 

 lateral prongs, with the result that the stem seems to shoot up as a single, tall, thin, tapering rod. 



The abortions of prongs which lead to bends in the smooth stem, bends which may become 

 very characteristic of certain forms, are always referred to in the descriptive text as " knee- 

 bends." P. Porto Bico 4^ shows many examples of them. In connection with such bends 

 and distinct from them, although at times they may be difficult to distinguish, are the natural 

 phototropic upward curvings of prongs which have been turned outwards too far from the 

 vertical This is very pronounced, for instance, in the type specimen of Lamarck's P. furcata, 

 preserved in the Paris Museum, see PI. XII. fig. 1. 



It is true that it will be long before we shall have discovered the growth-principles of all 



• There is certainly one Goniopora («. A., p. 161) described in the Supplementary Catalogue of 

 that genus given in this Volume, which shows the same phenomenon in almost as striking a way as 

 does the young Porites {Wed Indies x. 14) above described. 



f See Dr. Vaughan's excellent illustration, Bull. U.S. Fish Commission, ii. (1900) pi. xxx.- 



