PRESIDENTIAL ADDRESS—PRIESTLEY. 13 
The tendency towards realism, which is all too common 
in the scientific world, arises mainly from the concentra- 
tion of individual workers on their own fields at the 
expense of due consideration of general scientific develop- 
ment. The scientist should take the advice so often 
tendered to politicians and social reformers—Read history! 
A thoughtful study of the history of the growth of science 
reveals the arbitrary nature of our current views, which 
have been adopted for convenience, not of necessity. Since 
all the sciences are influenced by Physics, it is sufficient 
to notice the way in which physical science develops by 
mutual reaction between mathematical theory and physical 
observation. Observation suggests postulates for mathe- 
matical theory, and deductions from these postulates are 
tested by further observations. In order to carry out the 
tests it is necessary both to attach physical significance 
to the concepts of the mathematical theory, as Newton did 
when he identified a physical force of gravity with his 
mathematically defined central forces, and also to decide 
on canons of interpretation of the observations. The 
arbitrary element in the latter step is apt to be overlooked; 
in fact, great man as Helmholtz was, he failed entirely to 
recognize it when he undertook his experiment to deter- 
mine whether space was subject to the laws of Euclidean 
Geometry. It was left to Poincare to point out that Helm- 
holtz’s result was susceptible of two interpretations: space 
is Euclidean and a beam of light is straight, or space is 
non-Euclidean and a beam of light is curved. Further, 
it could not be upheld that the first interpretation must 
be taken on the grounds that Physics had already shown 
that light travelled in straight lines, for this fact was not 
a necessary conclusion from experience but merely a 
possible interpretation of experience. To return to the 
main question, when physical significance has been given 
to the mathematical theory, and when canons of interpre- 
tation of observations have been adopted, comparison can 
be made between theoretical and experimental results. 
Discrepancies between the two are reconciled by modifying 
the mathematical theory, the physical interpretation of 
the theory, or the interpretation of observations. By the 
constant interaction of theory and experiment a consistent 
