328 PART V. WORKING STAGE. 



It was natural to extend from (1) a sample of hydrogen at 

 a particular time to (2) that same sample at other times; from 

 this again to (3) other samples and (4) to the particular gas in 

 general. It was equally natural to extend the observation to 

 (5) some other gases less and less closely related to hydrogen, 

 and then (6) to all gases. Now since there are three states of 

 matter gaseous, liquid, and solid (including viscous), it is also 

 feasible to extend the generalisation tentatively to (7) some 

 liquid, then (8) to more liquids, thence (9) to all liquids, and 

 supposing that we ascertained that Boyle's law held good of all 

 liquids, we might, finally, extend the conclusion to (10) the 

 lightest solid body, (11) to a number of solid bodies more and 

 more heavy, and (12) to all solid bodies. I should then, granting 

 that the above-mentioned law also held good of all solids, state 

 the most general and universal law, namely, that the volume 

 of any given substance at any given temperature is inversely 

 proportional to the pressure to which it is subjected. This we 

 might name a universal generalisation. 



However, as a matter of fact, Boyle's law 1 applies neither 

 to liquids nor to solids, and therefore we should be compelled 

 to rest content with the more restricted compass of our gene- 

 ralisation, and even be obliged to allow for deviations assumed 

 to be .due to the attraction of the molecules for each other and 

 to the volume occupied by the molecules. 2 Of course, had the 

 question been the problem of the indestructibility or the gravi- 

 tational force of matter, we should have been in a position to 

 extend the first intimation derived from the study of the sample 

 of one body which proved indestructible or affected the scales, 

 to all matter gaseous, liquid, and solid. 



However, as in the science which deals with religious pheno- 

 mena, and in other new and intricate sciences, trustworthy 

 classifications may as yet not exist. In such an instance they 

 have to be created, and here the obstacles in the way of gene- 

 ralising may be most formidable. Or, what may be as bad 

 nearly, classifications may exist in abundance, as in proverbs 

 or in astrology, and these classifications may be unscientific 

 and misleading. In such circumstances the whole ground has 

 to be mapped out afresh, and the investigator needs to make 

 accurate studies of groups of facts, alertly seeking for significant 

 data and for multitudes of conclusions which should by degrees 

 enable him to formulate a classificatory scheme. 



"Boyle himself only proved his law in the case of atmospheric' air; but 

 the observation was subsequently (1676) generalised by Marriotte." (Thorpe, 

 op. cit., vol. 1, p. 139.) 



"Bodies which can support a longitudinal pressure however small, without 

 eing supported by a lateral pressure, are called solids. A liquid differs 

 essentially from a solid in being destitute of the power of sustaining pressure 

 unless it is supported laterally in every direction. The distinguishing feature 

 of a gas is that of indefinite expansion." (A. H. Hiorns, Principles of Metal- 

 lurgy, 1914, p. 8.) 



