314 PERSONAL PREPARATION FOR RESEARCH. 



6 The accuracy of a determination often depends much 

 more upon the skill o the operator than upon the con- 

 struction of the instrument used ; and thus Cavendish, 

 with nitric oxide as his reagent and water as the confining 

 liquid, made many hundred analyses of air, collected in 

 various localities, in 1781, and found the percentage of 

 oxygen to be invariably 20*83, a number nearly identical 

 with those obtained by Bunsen and Eegnault, with much 

 more perfect means. But the average chemist of that 

 day obtained the most discordant results with the same 

 apparatus and materials, and would doubtless also do so 

 at the present day. By improved apparatus and methods, 

 the work of the average chemist is made to equal, or 

 nearly so, that of the most skilled.' 1 



As every different substance possesses different pro- 

 perties, and every different apparatus produces different 

 effects, or is for a more or less different purpose ; and as 

 the number of known substances is extremely great, and 

 the kinds of apparatus extremely varied, it is evident that 

 the modes of manipulation are equally diverse. Some 

 of the substances employed are very explosive, others 

 are deadly poisonous ; some are in the highest degree 

 volatile, and require enormous pressures to liquefy them ; 

 some take fire on contact with air, others by contact with 

 water ; some explode in the presence of light, others by 

 a slight rise of temperature ; some require to be made 

 liquid by the very highest temperature, others by the 

 application of the greatest degree of cold, and the most 

 powerful pressure, &c. Many of the pieces of apparatus 

 also are extremely fragile, and require the most delicate 

 management, and in many of the experiments the effects 

 produced are so excessively minute, that only by the aid 



1 Address by Dr. Frankland. Conferences : Loan Collection, London, 

 1876, vol. ii. p. 11. 



