544 



SCIENCE. 



[N. S. Vol. VIII. No. 199. 



the West Indies and southern Florida, was 

 first taken at Woods Hole in 1894, and has 

 since been found on several occasions. The 

 species attains a weight of over 25 pounds, 

 but only small specimens (3 inches or less) 

 have up to this time been obtained here. 



One species of half-beak (Syporhamphus 

 roberti) is common at Woods Hole, and in 

 the current year another species (Hemi- 

 rhamphiis hrasiliensis) was found for the first 

 time. The latter is reported from Chesa- 

 peake Bay, but from no other localities 

 north of Florida. 



In August, 1898, there was taken a small 

 file-fish of the genus Alutera, which resem- 

 bles a fish known from Asiatic waters since 

 pre-Linnsean times and described by Os- 

 beck in 1757 as Batistes monoceros. It also 

 has some points of similarity to the Cuban 

 fish described and figured by Parra in 1787 

 under the vulgar name of ' lija barbuda,' 

 which was subsequently identified by Poey 

 and called by him Alutera guntheriana; the 

 latter is regarded by some recent authorities 

 as identical with A. monoceros, but the lack 

 of specimens has prevented a settlement of 

 the question. The Woods Hole fish differs 

 in a number of important features from the 

 foregoing, and apparently represents an un- 

 described species. 



Hugh M. Smith. 



TJ. S. Commission of 

 Fish and Fisheeies. 



NOTES ON INORGANIC CHEMISTRY. 

 Up to the present time there has been 

 little experimental evidence of the trivalence 

 of the so-called rare earths. They form no 

 volatile compounds in which the density 

 can be determined. The single j^roof of the 

 correctness of the formulae Ce^Og, LaPj, etc., 

 has been the determination of the specific 

 heat of the metals by Hillebrand, which 

 would give the atomic weights of the metals 

 as about 140. This has been generally ac- 

 cepted by chemists, but from time to time 



certain French chemists, notably Wyrou- 

 boff, have questioned the trivalence. This 

 has been largely on crystallographical 

 grounds. Wyroubofi' shows that the sili- 

 cotungstates of cerium, lanthanum and 

 didymium are isomorphous with that of 

 calcium, and argues from this that these 

 metals must be bivalent. It is also stated 

 that these metals in their compounds have 

 certain strong resemblances to the alkaline 

 earths. The whole subject is taken up by 

 W. Muthmann, of Munich, in the Berichte 

 and very fully discussed. He finds that 

 many of these supposed resemblances do not 

 exist in reality and that others do not sub- 

 stantiate the inference drawn from them. 

 Particularly he shows that the fact that 

 metals replace one another in isomorphous 

 salts by no means proves them to have the 

 same valence. By this argument a large 

 share of the metals would be made bivalent. 

 Especially in salts of high molecular weight, 

 as in the salts of complex acids, the nega- 

 tive complex is the dominating influence in 

 determining crystallographic form. This 

 important principle is well sustained by 

 Muthmann. To settle the matter of the 

 valence of these metals beyond controversy, 

 he has determined the valence, by the con- 

 ductivity of solutions of lanthanum nitrate, 

 sulfate and chlorid of diSerent strengths, 

 and the molecular weight of cerium chlorid 

 by the boiling-point method. In every case 

 a trivalent formula is obtained, and the 

 correctness of the usually-given formulae 

 for compounds of these metals may be con- 

 sidered as finally established. 



The work of Professor Jorgensen, of 

 Copenhagen, on the cobalt-ammonia bases is 

 well known. It has continued over many 

 years, following the work of Gibbs and 

 Genth, which was published in the Smith- 

 sonian Contributions to Knowledge in 1856. 

 In the last number of the Zeitschrift fiir 

 anorganisclie Chemie, Jorgensen gives a most 



