TRANSACTIONS OF SECTION A. 545 



author of a paper has had to submit it to criticism, but, when it has been ap- 

 proved by competent jud^'es, it has been published without ado and without 

 ■expense to himself. This is as it should be. It is lijrht that due care should be 

 exercised to prune away all unnecessary matter, to reduce as far as mav be the 

 necessary cost. It will, however, be a great misfortune if judgment as"to what 

 curtailment is necessary is in future passed, not with the object of removing what 

 is really superfluous, but in obedience to the iron rule of poverty. Apart from 

 all other disadvantages, such a course would add to the barriers which are 

 dividing the students of different .sciences. A few lines and a rough diagram may 

 suffice to show to experts what has been attempted and what achieved, but 

 there is no paper so difficult to master as that which assumes that the reader 

 starts from the point of vantage which montbsor years of study have enabled 

 the author to attain. Undue pruning will not make the tree of knowledge more 

 fruitl'u^, and will certainly make it harder to climb. 



Connected also with the vast increase of scientific literature is a growing 

 necessity for the publication of volumes of abs^.racts, in which the main results of 

 recent investigations are presented in a concentrated form. English chemists have 

 long been supplied with these by the Chemical Society. The Physical Society, 

 though far less wealthy than its elder sister, has determined to undertake a 

 similar task. We are compelled to begin cautiously, but in January next the first 

 number of a monthly pamphlet will be issued containing abstracts of all the papers 

 •which appear in the principal foreign journals of Physics. In this venture the 

 Society will incur gi-ave responsibilities, and I avail myself of this opportunity to 

 appeal to all Biitish physicists to support us in a work, the scope of which will 

 be rapidly extended if our first efforts succeed. 



From this brief glance at what has been or is about to be done to promote 

 the study of Physics, I must now turn to the discussion of narrower but more 

 definite problems, and I presume that I shall be most likely to deserve your 

 attention if I select a subject in which I am myself especially interested. 



During the last ten years my friend Dr. Thorpe and I have been engaged 

 upon a minute magnetic survey of the United Kingdom. The main conclusions 

 at which we have arrived are about to be published, and I do not propose to 

 recount them now. It is, however, impossible to give so long a time to a single 

 research without having one's attention drawn to a number of points whi'ch 

 require further investigation, and I shall perhaps be making the best use of this 

 opportunity if I bring to your notice some matters in the practical and theoretical 

 study of terrestrial magnetism which deserve a fuller consideration than has yet 

 been given to them. 



In the first place, then, there is little doubt that the instruments at present 

 used for measuring Declination and Horizontal Force are affected with errors far 

 greater than the error of observation. 



We employed four magnetometers by Elliott Brothers, which were frequently 

 compared with the standard instrument at Kew. These measurements proved that 

 the_ instrumental differences which afl'ect the accuracy of the declination and 

 horizontal force measurements are from five to ten times as great as the error of 

 a single field observation. The dip circle which two generations ago whs so 

 untrustworthy is, in our experience, the most satisfactory of the absolute instru- 

 ments. 



In most cases these comparisons extended over several days, but the 

 Astronomer Royal has described in his recent report observations made at Green- 

 ■wich for two years and a half with two horizontal force instruments. These 

 differ between themselves, and the discrepancy is of the same order of magnitude 

 as those we have detected. 



If such difiereaces exist between instruments of the Kew pattern, it is 

 probable that they will be still greater when the magnetometers under inv'estio-a- 

 tion are of different types. " " 



This point has been investigated by Dr. Van Rijckevorsel, who five years ago 

 visited Kew, Pare St. Maur, Wilhelmshaven, and Utrecht, and, using his own 

 instruments at each place, compared the values of the magnetic elements deter- 

 1894. ^ ^ 



