24 Prof. J. W. Gregory on the 



wliich were growing side by side, not 18 inches apart, on 

 precisely the same sea-floor, rising to exactly the same level, 

 and equally exposed to wave and current. The position was 

 sheltered in the extreme, for it was on the shore of a small 

 land-locked lay in Parham Sound on the lee side of Bird 

 Island ; and the bay was further protected by a shoal across 

 its mouth. In an adjacent patch of reef .1/. palmata and 

 M. cervicornis were growing interlocked, but each species was 

 perfectly distinct. In another bay in deeper water there were 

 circular patches of M. j^cilmata and il/. cervicornis, forming 

 flat-topped tabular masses from 10 to 15 feet in diameter. 

 They were growing under identical conditions. 



The statements therefore that M. palmata and M. cervi- 

 cornis are dimorphic forms of one species and that they have 

 acquired diflferent habits owing to their occurrence at different 

 situations on a reef are not in accordance with their distri- 

 bution on the coasts of Antigua. Indeed, the fact that where 

 M. palmata grows in association with M. cervicornis the 

 former is represented by a digitate or branched variety is fatal 

 to the assumption of their specific identity ; for the M. alces of 

 Dana, and not the J7. cervicornis oi Lamarck, is the branched 

 variety of M. imlmata. 



The Evidence op Intermediate Specimens. 



Brook supported his argument by the existence of corals 

 intermediate between M, palmata and J/, cervicornis. I care- 

 fully looked out for such in all the reefs I had the chance of 

 examining, but the search was unsuccessful. Brook stated 

 that the " intermediate forms occur chiefly in the collection of 

 the British Museum." He enumerates them on p. 29 of his 

 monograph. The specimens are four in number, and, thanks 

 to the kindness of Prof. Bell, 1 have had the opportunity of 

 examining them. 



The first specimen was collected by the * Challenger ' expe- 

 dition at St. Thomas. Its registration number is 86. 12. 9. 274. 

 The specimen is 200 millim. long, and consists of a central 

 stem which gives off" a series of cylindrical branches. On one 

 side there is, a little above the base, an imperfectly separated 

 branch which subdivides into two and shows the proximal 

 ends of six cylindrical branches. On the other side there are 

 eighteen branches or branchlets. 1 fail to see any ajjproach 

 to M. palmata in this specimen. If the specimen were 

 palmate we should expect it to be so at the base. But at the 

 bottom the central stem measures 30 millim. wide and 

 20 millim. thick. There is nothing palmate in that. Tiie 



