42 



SCIENCE 



[N. S. Vol. XXXI. No. li 



within certain limits— were regarded as 

 indefinite chemical compounds, or com- 

 pounds according to variable proportions. 

 So Robert Bunsen used to teach that we 

 may have compounds according to definite 

 proportions and also compounds according 

 to variable proportions, the latter com- 

 pounds being the group known as solu- 

 tions. 



The careful quantitative investigation 

 of solutions really dates from the time of 

 Lavoisier, who, as is well known, intro- 

 duced the balance into the chemical lab- 

 oratory. Before this the observations 

 made were generally only qualitative in 

 character; at any rate they were often 

 crude and faulty. The very fact that so- 

 lutions were regarded as chemical com- 

 pounds led to their study by much the 

 same methods adopted for the investiga- 

 tion of definite chemical compounds, i. e., 

 chemical compounds in the narrower sense 

 in which the term is at present commonly 

 used. So the qualitative composition and 

 the quantitative composition of solutions 

 were carefully studied. The density, the 

 color, the boiling point, the specific heat, 

 the optical activity, the thermal accom- 

 paniment of the formation of solutions 

 and of their reactions with other sub- 

 stances, as well as their other physical, 

 chemical and physiological properties, were 

 studied in much the same way that these 

 various properties were determined for 

 definite chemical compounds. And yet, 

 the fact that the composition of solutions 

 may be varied gradually and arbitrarily 

 within certain limits and that this can not 

 be done in the ease of definite chemical 

 compounds, has for nearly a whole cen- 

 tury been considered to be the vital differ- 

 ence between a solution and a definite 

 chemical compound, and this is quite 

 proper. 



To obtain a definite chemical compound 

 in the pure state usually requires a con- 



siderable amount of work. The usual 

 operations of purification as in vogue at 

 present are crystallization, solution and 

 precipitation, sublimation and distillation. 

 By means of the so-called purification 

 process a product is finally obtained whose 

 composition does not change further, 

 though the substance be subjected to 

 further similar treatment. As F. "Wald 

 states it, a chemical compound is a phase 

 whose composition remains constant though 

 temperature, pressure and contact with 

 other phases be varied within certain lim- 

 its inside of which the substance in ques- 

 tion is stable. In a sense then the so-called 

 definite chemical compounds are really ob- 

 tained in certain cases as the more resist- 

 ant cleavage pieces resulting when the 

 purification processes are applied. That 

 the latter processes after all frequently 

 represent rather violent treatment will 

 probably not be gainsaid by any one. 



The law of definite proportions was con- 

 sidered by Ostwald in his Faraday lecture, 

 which in turn was discussed by others, 

 among whom Benedicks voiced the senti- 

 ment that after all when closely scrutin- 

 ized it becomes evident that there is an 

 arbitrary element in judging as to when 

 Ave really have a pure, definite compound 

 before us, and that the matter of definite 

 proportions is to some extent one of defi- 

 nition. As to the law of multiple propor- 

 tions, this has been directly challenged by 

 P. Duhem as a tenet that can neither be 

 proved nor disproved, though I must 

 frankly confess my inability to agree com- 

 pletely with him in his argument. 



The year 1887 is noteworthy, for it 

 brought both the van't Hoff theory of di- 

 lute solutions and the theory of electro- 

 lytic dissociation of Arrhenius. These 

 theories really supplement each other, as 

 is well known. They may well be called 

 physical theories of solutions as distinct 

 from the chemical views of solutions al- 



