1907.] AND SUPPOSED SPECIES IN CORALS. 545 



By the end of a liundred days after the receipt of injury the 

 nninjnred centi-al end of a branch is roughened all over its surface 

 by the projections of lateral zooids, and those of them that will 

 form new branches have already begun to develop secondary buds 

 from their sides. This process cannot be spoken of with absolute 

 accuracy as an actual repair, for the part injured surely dies, and 

 its substance is never regenerated ; but it is a process by which 

 the sympathetic activity of the remaining zooids in the community 

 is called into action, and the potential power of branch formation 

 inherently lodged in the dominant lateral zooids is turned into a 

 real power. The value of this process in the life-saving of a 

 species is obvious, for the growth tendency of a Madrepore colony 

 is to steadily develop upwards, and the limit of this upward 

 growth is either at the water-level, or at a level at which pro- 

 tection from rough water and moving fragments afforded by the 

 shelter of surrounding rocks, is lost. ISTow a colony that grows 

 up beyond its upward limit of safety will sooner or later have its 

 terminal ramifications killed by exposure, or broken by waves or 

 moving fragments ; and the injuiy that destroys the power of the 

 apical zooid causes the lateral buds to branch out at angles to the 

 parent stem, and spread fresh zooid-bearing surfaces far and wide 

 in the area of safety. It is this process that is the great deter- 

 mining cause of the general growth tendency of Madrepore 

 colonies, and the one that causes the deep-water and shallow- 

 water forms of the same species to vary in their vegetative forms 

 of growth. 



3. If in a colony the injury is such that the apical zooid is 

 neither injured nor destroyed, but the damage is limited to the 

 surfaces of a branch, then the repair takes place as in colonies in 

 which every zooid is of equal value. 



This repair is well seen after experimentally inflicted injuries, 

 ii,nd the process of carrying out the regeneration of the destroyed 

 area is very like that previously described as occui-ring in those 

 corals that grow like the massive forms of Porites. 



The first step is the active marginal growth and the formation 

 of an excessive quantity of new material, which, in the form ex- 

 perimented on {Madrepora jjulchra), is at first of a light-blue 

 colour, and is semi-transparent. In this new material the mouths 

 of corallites soon appear, and the edges become covered by a host of 

 unifoi-m zooids, which soon spread over the entire area destroyed 

 by the injury, provided that the area is not too large, and that 

 no alga settles on it in the meanwhile. It may be stated here 

 that in experimental injury, many experiments fail for the reason 

 that the destroyed or injured area commonly becomes a focus 

 for the invasion of boring parasites, worms, molluscs, sponges, 

 and algfe, and the pure results of injury and repair become 

 complicated. 



Where no such complication existed, areas of 5 by 5 millimetres, 

 20 by 8 mm., 20 by 10 mm., and 25 by 12 mm., were completely 

 covered by new material with a multitude of new zooids in the 



37* 



