EXTERNAL FORM AS A SPECIFIC CRITERION 377 



significance for the systematist, and it is only necessary to count the number of varieties, 

 subspecies and species described in the Hterature devoted to Antarctic sponges, and 

 founded on spicule size alone, to realize this. 



Variation in the shape of spicules 



The shape, too, of spicules is not so constant as is commonly supposed. It will vary 

 more in one species than another, it is true, and from one stage in the life of an individual 

 to another (cf. Burton, 193 1), but in all species allowance must be made for these minor 

 differences, instead, as has been the practice, of investing them with specific importance. 

 A good example of this is given under lopJion proximiim (Ridley) (see pp. 296-30 1 ), but the 

 most striking case, which I hope to deal with in a forthcoming report, concerns Clathria 

 toxipraedita, Topsent, in which the ornamentation of the acanthostyli, styli and dermal 

 subtylostyli is very variable, while the general shape of these spicules also varies con- 

 siderably. 



The arrangement of the spicules in the skeleton is more constant than either their 

 shape or size, but even here variations are found which can be very misleading. In 

 appreciating the unreliability of the skeleton for systematic study, we are driven therefore 

 to accept the external form of the sponge as the final and deciding criterion in its 

 identification. 



The suppression of microscleres 



The dropping out of categories of spicules is a common phenomenon, but one which 

 is not sufficiently appreciated. This seems to be particularly evident among the Mycaleae, 

 but all types of sponges are liable to it. It is probable that sufficiently long searching 

 would show that the apparently absent spicules will be found to exist, even though in 

 minute quantities, in some parts of a sponge. That may or may not be true, but in any 

 case the practical difficulty still remains. Apropos this point, there is a specimen of 

 Mycale magellanica in the present collection in which no spicules other than tylostyli 

 could be found ; yet the external form showed it clearly to belong to that species. I have 

 myself described (1930, p. 333) as Axiiiosia incrustans, sp.n., a specimen which proved 

 eventually to be an Amphilectus fucoriim (Johnston) with very rare chelae. Most of the 

 species of Retiiera described by Dendy (1921) from the Indian Ocean contain rare 

 sigmata, which may be found only after prolonged search. Numerous examples of this 

 could be related (see also Mycale magellonica, p. 288), yet to suggest that micro- 

 scleres are an unreliable criterion for taxonomic purposes is to make an unwelcome 

 proposal. The fact is that they are secondary in importance to the structure of the main 

 skeleton. 



Differentiation within a single category of spicules 



Thiele (1898, pp. 13-14) describes two species of Stelletta from Japan, S. validissima 

 and S. orientalis, the former having oxea, dichotriaenes, protriaenes and two sorts of 

 anatriaenes, the latter having oxea, dichotriaenes, protriaenes and only one sort of 

 anatriaene. Both are from adjoining localities, have the same external appearance, and 



