612 



degenerate as to lose all distinctive characters while the male retains 

 them. For this reason the classifications of Steenstrup and Lütken 

 and Canu cannot be accepted. The former is also open to the further 

 objection that it includes only the true parasites and leaves out of account 

 the free and semi-parasitic species. 



We have thus by elimination practically brought our choice down 

 to two systems, the one originally proposed by Thorell, and afterward 

 adopted and corrected by Glaus and Gerstaecker, and the other 

 put forward by Giesbrecht and subsequently modified by Sars. And 

 we may propose a second practical question. 



II. Is there anything aside from inherent excellence and the 

 authority and experience of the propounder that will guide us in select- 

 ing between these two schemes? 



a. Thor ell' s well known division received much credit when 

 proposed, but w^e have already shown that in the very next year it was 

 found that many of the genera in the second division really did possess 

 mandibles. Glaus and Gerstaecker tried to correct this and other 

 errors, and still preserve Thorell' s scheme. But the original basis 

 having proved false, the corrected schemes are necessarily patched up 

 affairs, and for that reason not satisfactory. In the final correction by 

 Glaus, for example, the first suborder is called Gnathostoma from the 

 structure of the mouth-parts, while the second sub-order is known as 

 the Parasita from the habits of its members. The Parasita are then 

 divided into those with a beak or sipho (Siphonostoma) and those which 

 have none. The first suborder thus corresponds to one of the divisions 

 under the second one. 



b. This scheme does not readily include all the known forms, either 

 free or parasitic. Gerstaecker acknowledged (Tierreich, pp. 716 bis 

 730) seven genera of doubtful position, and there are fifty others which 

 he enumerated under the several families but could not j^lace definitely. 

 This is more than 2h% of all in the entire group, far too large a pro- 

 portion to be omitted in any valid classification. 



c. Finally this scheme mixes up the genera badly, so that forms 

 otherwise closely allied are widely separated by some habit or anatomical 

 detail. Giesbrecht has well shown that forms with biting (Gnatho- 

 stoma) and others with suctorial (Siphonostoma) mouth-parts may occur 

 even in the same family. Again the females of the genus Ergasiliis are 

 true parasites while the males are free-swimmers all their lives. To 

 separate the members of such famihes or the sexes of such genera in a 

 scheme of classification is manifestly as unscientific as it is unnatural. 



After such an accumulation of adverse criticism of the one system 

 a third question naturally suggests itself. 



