May 19, 1905.] 



SCIENCE. 



793 



They are characterized by their reduced max- 

 illary, the absent symplectic and subopercu- 

 lum, the union of supraoccipital and parietals 

 to form a single parieto-occipital bone, the 

 absence of parapophyses, the sessile ribs and 

 the compressed caudal vertebrae. All but the 

 Argiinae are covered, in part at least, by bony 

 plates. 



Mr. Eegan's paper is based on the material 

 in the British Museum and the Paris Mu- 

 seums and the types in the Harvard collec- 

 tions. A total of 189 species are recognized, 

 thirty-four of which are described as new. 



The genera and sometimes the species are 

 found widely distributed; two of the genera, 

 Chwtostomus and Arges, are confined to the 

 Andes of Peru. A table gives the known geo- 

 graphical distribution of all the species. In 

 this table the following systems are considered 

 as units : (1) Western coast-streams of Peru 

 and Ecuador; (2) Kio Magdalena system 

 (with Panama); (3) Venezuela and Guiana; 



(4) Upper Amazon (the Amazon and its trib- 

 utaries above its junction with the Yapura) ; 



(5) middle and lower Amazon; (6) Eio Par- 

 anahyba and Rio San Francisco and their 

 tributaries, and smaller coast streams in their 

 neighborhood; (7) Rio Parahyba, Rio Grande 

 do Sul and other coast streams of southeastern 

 Brazil; (8) Rio de La Plata and its tribu- 

 taries. 



Little need be said of this classification ex- 

 cept that it has long been known that Rio 

 Grande do Sul should be classed with the Rio 

 de La Plata and that the coast streams from 

 Rio de Janeiro to Bahia form a natural unit 

 distinct from the rivers to the north or south. 

 About fifty genera that are found both in the 

 Amazon on the north and the La Plata on the 

 south are absent from this area. The Rio 

 Paranahyba should probably be classed with 

 the Lower Amazon. The Pacific slope of 

 Panama should be classed with western Peru 

 and Ecuador. The number of species in these 

 systems is, respectively, 15, 18, 35, 64, 42, 17, 

 29, 32. Approximately one third of all the 

 species are found in the upper Amazon, while 

 but half as many are found in the vast La 

 Plata Basin. 



Brief chapters on sexual differences and 



changes during growth introduce the syste- 

 matic portion of the paper. The systematic 

 paper makes a distinct advance over the last 

 revision of this family by Eigenmann and 

 Eigenmann* in so far as Regan takes into 

 consideration the details of the skeleton of 

 the various groups, and inasmuch as he had 

 a much more abundant material, especially 

 of the species grouped by Eigenmann under 

 the generic names Hemiancistrus and Choe- 

 tostomus. 



Considerable difference exists between the 

 two papers on account of the estimate placed 

 on the importance of some characters. Regan 

 accepts fewer genera. 



For museum purposes and for purposes bf 

 general classification larger genera and fewer 

 names are preferable, but for all more inti- 

 mate discussions of variation, geographical 

 distribution and genetic origin of faunas 

 smaller units are vastly preferable. Regan's 

 species are also museum species with little 

 recognition of the biological significance of 

 varieties. For instance, Eigenmann and 

 Eigenmann accepted the Plecostomus affinis 

 of Steindachner as a variety of commersoni 

 (not of Ya\.) = punctatus. The parent, or 

 type form, is from Rio Janeiro, Santa Cruz, 

 Macacos, Itabapuana, Rio Parahyba, the va- 

 riety affinis from Rio Janeiro, Mendez, the 

 rios Mucuri, Parahyba, Muriahe, Doce and 

 San Antonio. The variety affinis is more 

 abundant than punctatus, ' over 50 specimens 

 having been examined ranging from .13-.26 

 m.' A second variety was described, three 

 specimens .30-35 m., from Sao Matheos. Of 

 these varieties Regan says : 



These varieties [based in one ease on over fifty 

 specimens] scarcely seem wortli recognition; in all 

 yoimg specimens the spots are large and the lower 

 surface of the head and abdomen partially naked, 

 and the persistence of these features in the adult 

 must be regarded as cases of individual variation 

 only. 



Of Plecostomus lima atropinnis Regan 

 says : Eigenmann has given the name P. lima 

 atropinnis to a specimen (presumably of this 

 species) from Goyaz, with the fins uniformly 



* ' A Revision of the South American Nema- 

 tognathi,' Occasional papers, Cal. Acad. Sci., I., 

 pp. 1-508, 1890. 



