166 CORRELATION OF PHYSICAL FOECES. 



oxygen, phosphorus, and sulphur. In this case we must ei- 

 ther divide the atoms of phosphorus and sulphur so as to re- 

 duce them to small fractions, or multiply the atoms of the 

 other substances by extravagant numbers ; thus to preserve 

 the unit of one of the constituents of this substance, chemists 

 say it is composed of 400 atoms of carbon, 310 of hydrogen, 

 120 of oxygen, 50 of nitrogen, 2 of sulphur, and 1 of phos- 

 phorus. This is a somewhat extreme case, but similar diffi- 

 culties will be found in different degrees to prevail among or- 

 ganic compounds ; in very many no constituent can be taken 

 as a unit to which simple multiples of any of the others will 

 give their relative proportions. By the mode of notation 

 adopted, if any conceivable substance be selected, it could^ 

 whatever be the proportions of its constituents, be termed 

 atomic. A solution of an ounce of sugar in a pound of wa- 

 ter, in a pound and a half, in a pound and a quarter, in a 

 pound and a tenth, might be expressed in an atomic form, if 

 we select arbitrarily a multiplier or divisor. 



It is true that in the case of solution, different proportions 

 can be united up to the point of saturation without any dif- 

 ference in the character of the compound, though the same 

 may be predicated to some extent of an acid and an alkali ; 

 but even where the steps are sudden, and compounds only 

 exist with definite proportions, they cannot, in a multitude of 

 cases, be reconciled with the true idea of an atomic combina- 

 tion, i. e. one to one, one to two, &c. 



Although, therefore, nature presents us with facts which 

 show that there is some restrictive law of combination which 

 in numerous cases limits the ratios in which substances will 

 combine, nay, further, shows many instances of a proportion 

 between the combining weights of one compound and those 

 of another ; although she shows also a remarkable simplicity 

 in the combining volumes of numerous gases, she also gives 

 numerous cases to which the doctrine of atomic combinations 

 cannot fairly be applied. 



