Simpson — A Revision of lite Gorgonellidub. 271 



The different variations which occur in these types will be considered in 

 detail in describing the different species ; but certain generalizations must be 

 briefly referred to here. 



Tt is very important when describing spicules from any colony to state 

 precisely from what part of the colony the preparation has been made. 

 New species have been established on slight differences in the size and shape 

 of spicules, and also on the preponderance of one type of spicule over 

 another. 



With a view to testing the degree of certainty with which this procedure 

 might be justified, we have made different preparations under different 

 conditions from the same colony ; and we now give the results derived 

 from over 500 preparations. 



(1) The different types of spicules retain theii- own distinct characteristics, 

 no matter from what level of the colony they may be taken. 



(2) Spicules from ditfei-ent levels of the same colony or from colonies 

 of different ages show marked deviations in absolute size, but not in 

 projjoriionate size. 



(3) Spicules in the coenenchyma alone differ from those in the verrucae 

 alone, e.g. in Scirpearia furcata, the double-club type, with hemispherical 

 ends, is coiifiued to the coenenchyma, whereas the elongated double-club is 

 restricted to the verrucae. This obtains in all specimens examined. 



It therefore foUows that when examining spicules for specific deter- 

 mination the factor of primary importance is the character of the spicules. 

 Next comes the average size of these spicules; while of no importance whatever 

 is the proportionate numbers of each type, as this depends on the proportion 

 of coenenchyma and verrucae taken for the preparation. 



If, then, a single preparation be made from a certain part of a colony, 

 and no criterion be given as to the exact age of this portion, subsequent 

 workers will experience great difficulty in making preparations from a 

 similar part. To obviate this difficulty another method may be employed, 

 namely, to take coenenchyma and verrucae from different levels for the 

 single preparation, and so obtain a representative sample of the spicules of 

 the specimen. This method has been found to be of great service in 

 identification, and is the one employed in the preparation of this memoir. 



Now it has been seen that the disposition of the verrucae is not a 

 constant even in a single specimen, and that its inclusion as a generic 

 character is imtenable. If therefore the separation of the specimens of 

 this Juncellid- group of the Gorgonellidae into genera is to be accomplished, 

 it must be based on the character of the spiculation. 



K,I.\. I'ROC, VOL. XXVIII., SECT, B. [2 O] 



