I.— PHYSIOLOGY. 159 



inability to alter them. In those cases where further study provides 

 methods of more fully understanding and therefore more adequately 

 controlling these circumstances, valuable results follow almost at once. 

 For instance, certain of the obscure causes of different behaviour under 

 particular conditions are inborn, and can be controlled by the use of 

 inbred strains of animals such as those of the standard inbred white rats ; 

 or again, one may mention the far-reaching results of the observation by 

 Pavlov that the utmost care must be exercised when studying the con- 

 ditioned reflexes to exclude all stimuli however trivial they may appear, 

 except the one under consideration. 



Under the most favourable conditions, however, it has up to the 

 present been usual to find a considerable imavoidable margin of 

 variation in the results of many physiological experiments. By regarding 

 these provisionally as ' chance ' variations, considerable help may be 

 obtained by the application of the theory of errors, based on the theory of 

 probability. In reality this is an empirical method of which Poiucare 

 has said that ' everybody firmly believes in it, because mathematicians 

 imagine that it is a fact of observation, and observers that it is a theorem 

 of mathematics,' but nevertheless, although it cannot, as seems sometimes 

 to be assumed, be used to replace accurate observation, it does enable a 

 result to be brought out which might otherwise be obscured by small 

 variations beyond our control. Research by such statistical methods 

 provides a useful method of investigation, as, for instance, in the study 

 of the toxic or other action of drugs, the data of the oestrus cycle, &c. 

 An elementary deduction which can be drawn from the consideration 

 of these facts is that, where only a few experiments of any kind are 

 performed, important conclusions cannot be drawn unless it can 

 be shown that the conditions are so controlled, and the accuracy of the 

 actual observations so high that the sum of the individual ' chance ' 

 variations must be small. Observation of this precaution would, in my 

 opinion, reduce the bulk of contemporary physiological literature very 

 materially, with a corresponding improvement in its quality. 



Lastly, as a means for evolving generalisations out of experimental 

 data, and of bringing these into relation with the generalisations of other 

 branches of science, the use of mathematics is incontestable. One need 

 only mention as examples the fresh outlook which has been provided for 

 further investigation by the exact study of the data relative to the segrega- 

 tion and recombination of hereditary factors, the beautiful investigations 

 of L. J. Henderson on the equilibria in the blood, the theoretical study 

 of the phenomena of excitation, the employment of thermo-dynamics 

 and the numerous other applications of physico-chemical theory. 



Certain applications of physics to physiology are quite clear-cut and 

 need no further comment ; but in many respects conventional physics 

 has for our purposes serious limitations, which the physiologist must try 

 to make good by his own investigations. For instance, many hydro- 

 dynamical problems of a specialised kind are connected with the study of 

 the circulation. The physical theory of the flow of homogeneous liquids 

 in wide, rigid, unbranched tubes is fairly well established, though, I under- 

 stand, somewhat abstruse. But when we come to study the physical 

 aspects of a pulsatile flow of a heterogeneous mixture like blood along 



