34 DR. W, RAE SHERRIFPS ON EVOLUTION 



relatively little variability of architecture, though there are, of 

 course, quantitative differences in size of spicules, strength of 

 supporting bundle, and the like, which are merely growth 

 characters. I must emphasize the saving clause in most cases, 

 in view of the occurrence of a number of very variable species, 

 such as D. gigantea, where there is great variability of antho- 

 codial armature in one and the same colony. In the second part 

 of this paper particular attention is paid to the interesting form 

 referred to ; and it does not stand alone. 



§ 3. The multiplicity of form detail within a narrow range 

 a.dmits of va.iious interj)retations : — 



I. («). It may be urged that the observed differences are 

 indivichial, not specific, characters. If the observed differences 

 are purely individual, they may be either extrinsic modifications 

 or of the nature of intrinsic variations. And before considering 

 these alternatives separately, I may note the general answer 

 that in many cases a considerable number of specimens from the 

 same or difterent localities agree together. Thus Kitkenthal 

 speaks of twenty specimens of D. savignyi, twelve of B. gigantea, 

 ten of D. rubra, seven of D. pumilio ; while I have myself got 

 twelve of D. ehrenbergi, ten of D. gigantea, ten of D. clavata, eight 

 of J), habereri, and the same number of I), hyalina. 



Therefore the general suggestion that observed differences are 

 purely individual does not apply in every case. At the same 

 time, the unsatisfactoriness of basing a new species on a single 

 specimen when the specific characters are of such a tenuous 

 description is plain. 



(6) It may be said that the observed' differences are of the 

 nature of exogenous modifications correlated with environmental 

 peculiarities. Ag-ainst a too ready acceptance of this easy 

 solution is the occurrence of the same form in widely separated 

 localities. For example, D. robttsta is reported from Ternate, 

 Borneo, Zanzibar, and Madagascar ; D. florida from Hong Kong, 

 the Philippines, and Port Jackson (Australia). 



It may also be pointed out that certain factors in the 

 immediate environment in different parts of a large colony must 

 show considerable diversity, cf. shelter ; and yet, as already 

 pointed out, there is rarely diversity in a colony except in colour 

 and the like. 



(c) It may be, however, that the observed differences are 

 expressions of individual intrinsic vai'iability, not correlated with 

 any peculiarities of environment. Thei-e is no way of testing this 

 interpretation until numerous similar specimens from the same 

 locality and of the same age and size are examined by some 

 investigator with abundance of living material at his disposal. 



II. It may be that the multiplicity of distinct forms within a 

 nan^ow range is due to cross-fertilization between allied species. 



