Makch 25, 1904.] 



SCIENCE. 



509 



comparison of the generic treatment in the 

 two works has not been attempted for all the 

 groups. The following, however, will show 

 that this difference in numbers is not a mere 

 multiplication of genera by Saccardo. Of the 

 64 genera of the Agaricacese given by Hen- 

 nings in Engler and Prantl, 15, or nearly 28 

 per cent., are not recognized by Saccardo, 

 many of them not appearing even as synonyms, 

 though his work is two years the later. If 

 we reverse the comparison, the showing for 

 uniformity in modern usage is much worse, 

 since of the 82 genera given by Saccardo only 

 41, or exactly 50 per cent., are recognized in 

 Engler and Prantl. This is certainly a case 

 where it will be difficult to say what is the 

 'prevailing usage.' Taking the Hymenomy- 

 cetes as a whole, Engler and Prantl give 147 

 genera, 25 of which, or 17 per cent., are not 

 recognized by Saccardo. If these glaring dif- 

 ferences cause us to investigate as to which 

 of these works is based on the more logical 

 and consistent usage, and, therefore, which is 

 the safer nomenclatorial guide, we are forced 

 to the conclusion that neither of them follows 

 any recognizable or consistent rule of nomen- 

 clature. The case of each genus seems to 

 have been settled on an independent basis and 

 according to the whim of the moment. 

 Doubtless the claim would be made for each 

 work that the names were selected on the basis 

 of priority, but priority has been flagrantly 

 and repeatedly violated in both of them. 



Again is the ' prevailing usage ' furnishing 

 us at the present time with a safe rule for the 

 establishment of new genera on a sure and 

 stable basis? That this is no idle qiiestion is 

 shown by the vast increase of over 2,000 

 genera since 1829 and of 663 since 1889, if 

 we count on the basis of the ' Sylloge,' and 

 the tendency is for the still more rapid multi- 

 plication of genera in the near future. Every 

 revision of a large genus in these days re- 

 sults in breaking it up into smaller generic 

 groups. It is vitally important that this shall 

 be done on some basis that will prove stable. 

 What is really being done is illustrated by a 

 recent revision of Bavenelia. After an ex- 

 haustive and critical study of the species the 

 author very properly decides to break up the 



genus. He leaves the majority of the species 

 under the old generic name and proposes new 

 names for the smaller segregations. Now it 

 happens that Bavenelia was founded on a 

 single species, B. glandulosa. In the pro- 

 posed revision this species falls in one of the 

 smaller groups and is no longer called a 

 Bavenelia, while that name is applied to a 

 group of species none of which were included 

 under it by the author of the genus. If 

 priority is to be more than an empty name 

 such practices can certainly not be allowed 

 to stand unchallenged, yet the author could 

 point to hundreds and hundreds of precedents 

 to justify his usage. In fact, we must admit 

 that this usage has been the prevailing one 

 ever since the time of Linnseus. The chaotic 

 condition that must inevitably be produced by 

 following this so-called ' method of residues ' 

 is well illustrated by the following figures 

 taken from my work as a member of the 

 nomenclatorial committee in finding the types 

 of the older genera of fungi. I have listed 

 485 names that were proposed between 1753, 

 the first edition of Linnseus's ' Species 

 Plantarum,' and 1821, including the first 

 volume of Eries's ' Systema Mycologicum.' Of 

 these, 242, or one half, are to be rejected for 

 various reasons. Some are hyponyms, never 

 having been associated with a recognizable 

 binomial species ; some are typonyms, being 

 based on species already used as the types of 

 other genera; some were only proposed as 

 subgenera, and some were based on sterile 

 mycelia, monstrosities, insect work or plants 

 that are not fungi. The remaining 243 

 names are available for use at the present 

 time. The types of these have been deter- 

 mined according to the code proposed at the 

 Washington meeting. One hundred of them, 

 or 41 per cent., were monotypic, being based 

 on a single species. In 135 of them, or 55 per 

 cent., the type was determined by page 

 priority. Nine are historic types taken from 

 pre-Linnajan authors and 4 were inferred from 

 the form of the specific name. Of these 243 

 available names, 118 are used by Saccardo in 

 their proper historic sense, being still asso- 

 ciated with their original type species. In 

 the other 125 cases the names are either not 



