66 Proceedings of the Royal Physical Society. 



crystallographic work ; but this mode of division, thongh of 

 great importance in crystallographic work, is not needed in 

 biology. It is drawn by measuring off tangents on the metre 

 scale regarded as of the length of the radius. Its construc- 

 tion presents no difficulty. 



V. Observations upon the Bathymetrical Distribution of 

 Beef-huilding Corals. By J. G-. Goodchild, H.M. Geo- 

 logical Survey, F.G.S., F.Z.S. 



(Bead 16th April 1902.) 



One of the principal arguments upon which "Darwin's 

 Theory of the Origin of Coral Eeefs " is based, is that reef- 

 building species of corals cannot live below a certain depth 

 from the surface. This depth is variously stated: some 

 authors setting the figure at 25 feet, while others give it at 

 as many fathoms. A few writers on the subject admit a 

 possible downward limit of growth at 40 fathoms. All agree 

 that the depth, be it what it may, is comparatively small as 

 compared with the depth to which many other species of 

 corals, not reef-builders, may extend and thrive. 



It must strike any thoughtful reader of these statements, 

 regarding the downward limit to which reef-building corals 

 can flourish, that some of the figures are rather more of the 

 nature of vague estimates of what is likely, than statements 

 of facts actually observed. It would seem, indeed, as if no 

 actual proof has ever been advanced. The statement, again, 

 which is usually given in books, that reef-building corals 

 thrive best in sea-water at depths between 6 fathoms and 

 10, is also of too general a nature, and needs to be used with 

 caution, and with special reference only to the particular 

 locality where the observations were made upon which that 

 statement was founded. 



On the whole, therefore, it appears safe to conclude that 

 no really trustworthy statements relative to the depth, in 

 general, at which reef-building corals can live, have yet been 

 made — or, at any rate, none that can be applied to all the 

 cases likely to be met with. It will probably be found, 



