October 1, 1915] 



SCIENCE 



455 



is nothing new in this method, for it has been 

 used at rare intervals for a number of years. 



The following illustrations taken at random 

 from a number of measurements, show the 

 state of the ease as regards chemical journals, 

 but the proposition is equally true for the 

 journals of other sciences. The measurements 

 have been made on the supposition that the 

 same sized type will be used in both ways of 

 printing the references. A space of five milli- 

 meters is allowed between the period at the 

 end of one reference and the beginning of the 

 number of the next, and extra space is allowed 

 for in the case of articles containing over ten 

 references, for these will require the use of 

 numbers with two figures. Thus, an article 

 with sixty separate references printed in the 

 customary way will have, as a rule, from two 

 to eight references on the page, and rarely re- 

 quire the use of reference numbers with two 

 figures, while the shorter way requires the use 

 of numbers up to sixty both in the text and at 

 the end of the article. 



A consideration of those illustrations in 

 which the number of pages with no references 

 is given, will make two facts obvious; that the 

 estimate of the amount of space saved is only 

 a rough approximation; and that it might 

 easily be possible by deliberate selection of 

 one hundred pages to get results widely diflier- 

 ent from those given here. 



In the leading English journal, the Journal 

 of the Chemical Society, all the references 

 are printed in the text, so that there is no 

 saving of space. In the Journal of the Society 

 of Chemical Industry, Volume 32, 457-995 

 (1913), containing 100 pages of text of articles, 

 about one eighth of a page could be saved. 

 This journal has two columns to the page. 

 Several of the articles have all the references 

 at the end of the article. 



In the Bulletin de la societe chimique de 

 France, Volume 13, 320-45Y (1913), contain- 

 ing 100 pages of text of articles, over one half 

 a page could be saved. There are 62 pages 

 with no references. Some of the articles have 

 the references printed two to the liae, with 

 extra space between the two. 



In the Berichte der deutschen chemischen 



Gesellschaft, Volume 45, 403-503 (1912), about 

 one half a page could be saved. The references 

 in this journal at present are greatly abbre- 

 viated, thus, B. stands for the title of Berichte, 

 etc. If space allows, instead of printing the 

 references one to a line, they are always 

 printed two to a line, with a space of 5 to 15 

 millimeters between the two. This method (a 

 step in the right direction) naturally limits 

 the saving possible. There are 61 pages with 

 no references. 



In the Gazzetta chimica italiana, Volume 

 42, I., 316-416 (1912), nearly three quarters 

 of a page could be saved. There are 61 pages 

 with no references. 



In the Journal of the American Chemical 

 Society, Volume 34, 1631-1731 (1912), one 

 page can be saved. There are 54 pages with 

 no references. In the Physical Review an 

 article always begins at the top of a page. 

 The effect of this is that there is sufficient 

 space at the end of most articles to print aU 

 the references, often on the basis of a line for 

 each reference. 



In the case of journals printed with two 

 columns to a page, there is an incredible num- 

 bers of variations in the way in which num- 

 bers indicating references are used, as, for 

 instance, giving numbers consecutively for the 

 references in one article regardless of the 

 number of columns used for the articles, or 

 using numbers consecutively for one column 

 regardless of the number of articles in that 

 column, etc. These variations doubtless re- 

 lieve the monotony of the work of editors, 

 printers and proof-readers. 



There is one great advantage in the method 

 recommended here. It does away with " loe. 

 cit.," the abbreviation so easy to write, so sav- 

 ing of time and space in printing, so wasteful 

 of time to one who has to go back an indefinite 

 number of pages and read an indefinite num- 

 ber of references, in order to find where the 

 authority for a given statement may be found. 

 When many articles by one author, or many 

 patents by one man are given, this " loc. cit." 

 may be so indefinite that it wiU be necessary 

 to look through a number of the originals in 

 order to get at the one desired. In a series of 



