abstract (sometimes, never). The abstracts in 

 general give (too much, too little, about enough) 

 information. I think they should be continued 

 (yes, no). 

 I have the following suggestions to make: 



Replies were received from 805 readers, 

 including 83 readers of the Astro physical Jour- 

 nal, 502 readers of the Physical Eevieiv, and 

 220 readers of 'both. Although the answers 

 from the three groups have been tabulated sep- 

 arately, the results are so nearly the same for 

 each that only the results for all readers, that 

 is, the percentages of all readers answering 

 each question in each alternative way, will be 

 given here. Since some replies are incom- 

 plete, the sum of the percentages is usually 

 less than 100. 



1. How frequently are the abstracts read be- 

 fore the articles? 



always by 41.9 per cent, of readers, 



usually by 45.9 per cent, of readers, 



sometimes by 9.9 iper cent, of readers, 



seldom by 2.1 per cent, of readers, 



never by 0.2 per cent, ef readers. 



100.0 per cent. 



2. How many abstracts are read instead of 

 the articles? 



many by 44.5 per cent, of readers, 



some by 39.7 per cent, of readers, 



few by 9.5 per cent, of readers, 



none by 2.6 per cent, of readers. 



96.3 per cent. 



S. How many abstracts helped in understand- 

 in (j the articles? 



some helped 58.1 .per cent, of readers. 



few helped 17.7 per cent, of readers, 



none helped 11.8 per cent, of readers. 



87.6 per cent. 

 2 



