36 THE METHOD OF SCIENCE 



thinking exactly. 1 Mathematics, which is not an experimental 

 science, is exact for similar reasons. The study of heredity is not, 

 and cannot be, exact in the same sense, for no two individuals, 

 nor even two of their cells, are ever precisely alike. When we 

 unite a given quantity of oxygen with a given quantity of hydrogen, 

 the result, which can always be predicted with absolute confidence, 

 is ever a definite quantity of water of a definite composition. But, 

 when two germ-cells are united, it is impossible to foresee the 

 exact nature of the product. Doubtless, when dealing, not with 

 individuals, but with populations, biometry, the statistical method 

 of investigating the phenomena of life, does in some cases enable 

 us to forecast the future with considerable accuracy ; but this 

 accuracy is infinitely less comprehensive than that obtained in 

 physics and chemistry where predictions are true, not only as 

 regards averages, but also as regards particular instances. It is 

 conceivable, though, owing to the complexity of the conditions, not 

 at all probable, that biologists will one day be able to predict 

 with the same accuracy and confidence as physicists and chemists ; 

 but it is hard to understand how this goal will be the sooner 

 attained by ignoring everything that is patent to our senses. 

 Physicists and chemists, as rational beings, would be happy if 

 their facts were patent, and they were thus spared some of the 

 labours of the laboratory. It is their good fortune, not their merit, 

 that they are able to measure and test exactly. 



65. In brief terms, sciences are experimental or non-experi- 

 mental in proportion as the facts on which they are founded are, 

 or are not, obscured. They are exact or inexact in proportion as 

 they are, or are not, capable of being founded on precise measure- 

 ments. Some sciences are both experimental and exact ; but the 

 two circumstances have no necessary relation. Dogs run and wag 

 their tails. But they do not run because they wag their tails. 

 Neither are sciences exact because they are experimental. In 

 science, as in the ordinary affairs of life, a source of error as 

 common as or more common than the use of inaccurate data, is a 

 neglect to use the whole of the facts available. Certainly the use 

 of all the evidence is, through the exposure of discrepancies, one 

 of the most effective means of ascertaining whether any of it is 

 inaccurate or has been misinterpreted. 



66. Science is founded on facts. But collections of facts do 

 not by themselves constitute science. They are only the raw 

 materials out of which it is manufactured. Science consists in the 



1 See 824. 



