November 18, 1921] 



SCIENCE 



491 



open sea, dense, saline and cold, to the Baltic. 

 For this reason Mr. Huntsman suggests that 

 the density of the surface water in which the 

 eggs develop may be a decisive element. 



In this connection, I may add a few addi- 

 tional data. In the group of Eock Cod or 

 Rose-fish (Selastinw), the northern genera 

 iSehastes, Sebastolohus) have twenty-nine to 

 thirty-one vertebrae, the tropical forms nearest 

 related twenty-four, and the intermediate 

 g^roup of many species on both sides of the 

 Pacific (Selastodes and its allies) were sup- 

 posed to have twenty-seven. 



In verifying this statement I find that four 

 of the more primitive of these forms (Sehas- 

 todes paucispinis, S. goodei, Eosicola pinniger 

 and R. miniatus, have but twenty-five vertebrae, 

 while all the others examined have twenty- 

 seven as supposed, and the metameres in the 

 very young are also twenty-seven. 



Hitherto the extinct species of this tribe 

 have remained unknown. I have, however, 

 lately discovered three Miocene species, which 

 ought to throw light on the problem. At any 

 rate they show that the variation is of long 

 standing. 



Two fossil species with thirteen dorsal 

 species, Bixator porteousi and R. ineziae, re- 

 lated to Seiastodes goodei, have, like the 

 latter species, twenty-four vertebrce, besides 

 the last one which supports the hypural. 

 This is evidence so far as it goes that the 

 smaller number (with greater individual de- 

 velopment of the bones) is very ancient. 

 Nearly all the spiny-rayed shore fishes of the 

 present day have twenty-five. 



But another fish of this type — also Miocene 

 iSehastavus vertebralis), has thirty-two verte- 

 brae. The relation of this species to existing 

 forms is not close, nor is it well made out. 

 All three of these Miocene species are found 

 in deposits made in shallow, sheltered bays, 

 in a temperate climate. As Mr. Huntsman 

 observes, "A fruitful field for investigation 

 is open in this direction." It should appar- 

 ently involve both embryology and paleontol- 

 ogy, as well as the study of adult fishes and 

 their distribution. 



David Starr Jordan 



ABSTRACTS AND TITLES OF SCIENTIFIC 



ARTICLES FROM THE LIBRARIAN'S 



STANDPOINT 



To THE Editor of Science: In his article 

 on " Scientific Abstracting " in Science for 

 ■ ber 30, Mr. Fulcher emphasizes the 

 point that the time of research men should 

 be conserved for their actual research by 

 facilitating for them in every way the secur- 

 ing of the scientific information already pub- 

 lished. No one would dispute this statement, 

 and its truth is becoming increasingly strik- 

 ing as the mass of literature yearly accumu- 

 lates, but it is suggested that from his list 

 of the agencies contributing to this end as 

 a part of what he calls " our scientific in- 

 formation service " Mr. Fulcher has omitted 

 a very necessary and important agency, 

 namely, the scientific library. A library of 

 a scientific institution has no other purpose 

 than to collect and make available the litera- 

 ture on the subjects of interest to that insti- 

 tution, and anything which facilitates this 

 work is ultimately of benefit to the investiga- 

 tors. There is no one to whom abstracts 

 such as those pled for by Mr. Fulcher would 

 be of greater help than to the scientific libra- 

 rian or bibliographer. As he points out, it is 

 imjwssible to rely on titles alone to show the 

 variety of information contained in an arti- 

 cle, so that it is necessary for a librarian 

 compiling a subject catalogue to glance through 

 each article so that he may be sure it is 

 entered under all the subjects of which it 

 treats. Abstracts in the form described, with 

 the italicized paragraph headings and sub- 

 titles would suggest at a glance possible sub- 

 ject headings, and in the case of articles in 

 highly specialized subjects would frequently 

 suggest headings which, without the abstract, 

 only the specialist would recognize as being 

 desirable. 



Speaking of this, the present writer has 

 thought for a long time that it would be well 

 for persons interested in increased economy 

 and efficiency in the recording of scientific 

 data to give the form of titles for periodical 

 articles careful consideration. No title can, 

 of course, describe all the contents of an 

 article, but many could easily be more de- 



