been appearing in the Astro physical Journal 

 and Physical Review are of the analytic type 

 developed by the National Research Council. 3 

 They aim to give a complete description and 

 adequate summary of the results reported in 

 each article. The replies to the question: 



6. Is the amount of information given, in 

 general, about enough, too little, or too much? 



about enough 79.8 per cent, of readers. 



too little 6.7 per cent, of readers. 



too much 4.8 per cent, of readers. 



91.3 per cent. 



these replies indicate that the abstracts, aver- 

 aging from 5 to 6 per cent, of the articles, have 

 been of about the right length, and it is not 

 believed they could be made much shorter with- 

 out considerably decreasing their value. 



A distinctive feature of the abstracts is that 

 they contain italicized subtitles wnich give 

 the various subjects involved more completely 

 and precisely than in general it is possible for 

 the author's titles to do. One purpose of these 

 subtitles is to enable a reader who is not inter- 

 ested in the subject indicated by the author's 

 title to determine by glancing through the sub- 

 titles whether something of interest to him has 

 not been incidentally included. Until readers 

 became accustomed to these subtitles, however, 

 it was to be expected that most readers would 

 consider them more of an annoyance than a 

 help; but the following replies show that a 

 surprising percentage of the readers have 

 already found the subtitles useful: 



7. Do you ever read the subtitles in the ab- 

 stracts first? 



sometimes 64.1 per cent, of readers. 



never 17.5 per cent, of readers. 



81.6 per cent. 



3 Described in ' ' Scientific Abstracting ' ' re- 

 ferred to a.bove. 



4 



