Notices respecting New Books. 305 



and \vhen converted into potash, they gain a grain and -jt-j 

 and as 7-1 : 8 : : 67 : 75-4 ; and as 1-6:8: :''l5 : 73, giv- 

 jng the number representing potai^sium as about 75. 



" It is easy to form a series of proporiionai numbers, by 

 taking | ot these numbers, on the supposition that water is 

 composed uFone proportion of hydrogen aTid one of oxvgen: 

 but in this case the number representing »he proportion in 

 which oxygen combines must contain a traction ; and the 

 calculations are much expedited, and the formula rendered 

 more simple, by considering the smallest proportion an in- 

 teger. 



" Mr. Higgins has supposed that water is composed of 

 one particle of oxygen and one of hydrogen, and Mr. Dal ton, 

 of an atom of each ; but in the doctrine of prof)ortions de- 

 rived from facts, it is not necessary to consider the combi- 

 ning bodies, either as composed of indivisible particles, or 

 even as aUvavs united, one and one, or one and two, or one 

 and three proportions. Cases will hereafter be pointed out, 

 in s^'hich the ratios are very different ; and at present, as we 

 have no means whatever of judging either of the relative 

 numbers, figures, or weights, of those particles of bodies 

 which are not in contact, our numerical expressions ought 

 to relate only to the results of experiments. 



" If it should liereafter be disc'overed, that any of those 

 substances now considered as undecompounded, consist of 

 Other elements, these elements must be represented by sonie 

 division of iheir numbers; and should even h\droo-en be 

 found a compoundtd body, it would merely be necessary to 

 multiply all the numbers representing the other elements, 

 by some common number which would admit of a division 

 into proportions, representing the elements of hydiogen ; 

 9o that no discovery concerning the composition of bodies 

 can interfere with the general law of the detinite nature of 

 the'r combinations." 



Such is the i^oh ssor's illustration of his theory of de- 

 fiime pro|)ortions in the chemical composition of bodies. 

 Its simplicity and comprehensiveness must render it a great 

 jehef to I he memory, enable the mind to methodize the 

 vast multitude of tacts which modern experiments have 

 developed, and contribute to furnish a collateral test of the 

 accuracy of analytical results in general. Facts seem to 

 prove " that two or more proportions of one bodv attract a 

 miigle proportion of another bodv with more energy than 

 cue proportion, and that two proportions or njore adhere 

 V) a .HiDL'le |)roport|on with less energv than one propor'ion ; 

 %r, at least, that a Bicond ora tbipd proporuuu adheres with 



V'yl. 40. No, 1 74. Ocf . 1 8 1 2. U less 



