492 



SCIENCE. 



[N. S. Vol. XXII. No. 564. 



they are not themselves concerned. The type 

 should be the best known species from the 

 standpoint of the author. Frequently the first 

 reviser (as of Esox and Syngnathus) selects 

 as type a species v?hich was by no means cen- 

 tral or typical in the estimation of the orig- 

 inal author. Still more frequently it is im- 

 possible to tell who is the first reviser, unless 

 that phrase itself receive accurate definition. 

 In the early days, many authors paid little 

 attention to earlier genera, and in their re- 

 views they encroached on the groups named 

 by their predecessors, without limiting them or 

 fixing their species. If the phrase is retained, 

 the first reviser should be the one who first- 

 consciously limited the range of the genus by 

 fixing the actual name to one of the actual 

 original species. This at least is tangible. 

 When a type is not fixed either by the original 

 author or by his ' first reviser,' the code makes 

 certain recommendations to the systematist. 

 These seem to be of the nature of advice, and 

 are void and of no effect when a type has been 

 previously fixed. Third among these comes 

 the method of elimination, a plausible process, 

 but one which has never been defined and 

 which in complex cases leads to as many dif- 

 ferent results as there are writers who attempt 

 to use it. 



In the code, these recommendations are made 

 subordinate to the rule of the ' first reviser.' 

 It is a question, however, whether the first and 

 second of these recommendations (using as 

 type the species suggesting the generic name 

 as Lutianus lutianus, and using the one per- 

 sonally best known to the original author, as 

 Esox lucius) should not have had precedence 

 over ' the first reviser rule.' The present 

 writer finds difiiculty as above stated with the 

 rule of the first reviser. In fishes, he finds 

 the method of elimination practically worth- 

 less, at least, unless some rigid definition of 

 it can be agreed upon. The arbitrary choice 

 as type of the first species named under each 

 new genus by its describer, is a rule which 

 could have been enforced without confusion 

 and which yet may be found necessary. It 

 is, perhaps, too late now to go back to it, al- 

 though several of the chief writers on fishes, 

 Bleeker and practically Lacepede and Cuvier 



have more or less consistently adopted it. It 

 is at least fair to apply this rule to these 

 particular authors and to others who begin 

 their account of each genus with the ' type ' 

 or ' chef de file.' 



A great deal can be said in favor of a prin- 

 ciple in nomenclature, which may be stated as 

 follows : The determination of the significance 

 of each name, generic or specific, must be made 

 on evidence furnished by the author framing 

 the name or on evidence existing at the time. 

 It is possible to give an exact type to every 

 genus or species on this basis, or in default 

 of this to follow the simple and just rule of 

 page precedence. This gives fixedness at 

 least, and we need demand nothing else. This 

 method would release zoology from the un- 

 welcome and profitless task of finding out 

 what an author means, by studying the effect 

 of his words on his successors. In other 

 words, our studies in this line would be limited 

 to the author himself and to those on whom he 

 may have relied. The adoption of the rule 

 that a specific name might be identical with 

 the name of a genus has saved us, in the ag- 

 gregate, years of investigation among useless 

 and forgotten synonyms. This same kind of 

 study is forced upon us by the rule of the first 

 reviser or the still more complex custom of 

 the application of the method of elimination. 



Dr. Stiles evidently appreciates the incom- 

 pleteness of article 30, for he supplements it 

 by twelve rules of his own, saying that ' No 

 existing code of nomenclature provides for 

 all cases that arise, so that authors make 

 supplemental rules for themselves.' But 

 these supplemental rules are necessarily parts 

 of a completed code. The final form of this 

 code should, therefore, contain or replace these 

 twelve. excellent rules of Dr. Stiles. Till this 

 is done, we may recommend that these sup- 

 plementary rules be favorably regarded by 

 naturalists, though in our judgment page- 

 precedence — as a remedy for taste or whim — 

 will ultimately be given a place higher up the 

 line than that assigned by Dr. Stiles, and 

 ' absolute tautonomy,' ' virtual tautonomy ' 

 and the Linnsean rule of using ' the best 

 known European or officinal species ' as type 



