The specific characters of the coral Stylaster roseus \ 37 



natural size (Fig. 10), and the terminal part of a branch, 4 times enlarged (Fig. 11). 

 The figures show that the tendency for a lateral arrangement of the cyclosystems on 

 the branches is much more obvious than in the specimens from Curasao, that the 

 branches are much less pronouncedly tapering, and that the cyclosystems definitely 

 extend over the surface of the branchlets, the smaller branchlets thereby becoming 

 distinctly zigzag-shaped, in contradistinction to the corresponding parts of the 

 specimens from Curasao. The cyclosystems of Stylaster erubescens (diameter 1 -2- 

 1 -5 mm) are about twice as wide as those of the specimens from Curasao (diameter 

 0-5-0-7 mm), and distinctly wider than in the specimens of Stylaster roseus examined 

 by Milne Edwards and Haime (diameter nearly I mm). The most important 

 difference of the two forms is that of the entirely different shape of the gastroslyle, 

 which in Stylaster erubescens is rounded (almost spherical), in the specimens from 

 Curasao conical, more than twice as high as broad (Fig. I). 



The Stylasteridae of the North Atlantic region remained imperfectly known till 

 1914, when Broch's important paper appeared dealing with the specific characters 

 of Pliobothrus symmetricus Pourtales, Allopora norvegica (Gunnerus), Stylaster gem- 

 mascens (Esper), and a species named by broch Stylaster roseus (Pallas). The material 

 of the last named came from depths between 263 and 1,400 metres; Broch notes that 

 the colonies display a marked difference between small branches, main branches, and 

 stem, that the cyclosystems are placed laterally and alternately on the small branches, 

 that the cyclosystems show from 8 to 17, generally 9 to 1 1 dactylopores, and that the 

 gastrostyle is almost spherical, with equal height and breadth. The characters here 

 cited from Broch's description are almost exactly those of the species Stylaster 

 erubescens as described and figured by Pourtales (1871). Broch's figures of colonies 

 in natural size (1914, PI. 1, Figs. 8 and 9, PI. 2, Figs. 10 and 1 1) represent corals with 

 an entirely similar form of growth as the specimen of Stylaster erubescens of Pourtales 

 (1871, PL 4, Fig. 10); moreover, the figures of enlarged terminal branches (Pourtales, 

 1871, PI. 4, Fig. 11 ; Broch, 1914, PI. 2, Fig. 17) are strikingly similar. 



Unfortunately Broch gave the name Stylaster roseus to the corals here dealt with, 

 while placing the name Stylaster erubescens in the synonymy of the species. In Broch's 

 paper there is an instructive figure of a longitudinal section (1914, PI. 3, Fig. 22) 

 showing in two of the cyclosystems the spherical gastrostyles, which, judging by this 

 figure, have a height and a breadth of about 0-3 mm. This figure has been copied in 

 other publications (Broch, 1924; Kuhn, 1939) as a longitudinal section of a branchlel 

 of Stylaster roseus (Pallas). 



Broch (1914, p. 15) remarks " As the species is the commonest Stylasterid in the 

 Atlantic north of the equator, it is probably the same form that served as a basis for 

 Pallas' description of Madrepora rosea "; this is right as far as the specific identity 

 of the North Atlantic corals with Stylaster erubescens is concerned, but it was an error 

 to identify this species with Stylaster roseus. Pourtales (1868, p. 136, footnote; 

 1871, p. 83) examined the two forms and referred to them as separate species, the one 

 from shallow water, the other from deeper water only; it is to be regretted that he 

 did not give a description of the specific characters of Stylaster roseus. 



The conclusion of the data dealt with above is that Stylaster roseus (Pallas), the 

 type species of its genus, has a rather complicated history. The descriptions and 

 figures of the various eighteenth-century authors contain some characters which may 

 be considered typical of the species. In later years these older data were overlooked 



