if I Reply to Mr. Brooke. 



405 



II. On the Evidence of the Doctrine of Isomorphism. 



It has already been said that analyses of the same mineral 

 often give results so widely different that it is hopeless to at- 

 tempt to discover in them the same chemical compound of 

 the same ingredients. This being so, we find that a law 

 nevertheless does prevail in the analyses of the same species, 

 connecting them together in a definite and certain manner, 

 and bringing them under a common formula. This law is 

 given by the doctrine of isomorphism. The evidence of the 

 doctrine must depend upon the accuracy with which it repre- 

 sents the facts; and that the reader may judge of this, I will 

 take the analyses of eight varieties of garnet, given by Leon- 

 hard in his account of that species [Handbuch, p. 489). No- 

 thing can apparently at first sight be more anomalous than 

 these analyses. The alumina varies from to 22 per cent., the 

 lime from to 34, the magnesia from to 13, the protoxide 

 of iron from to 34, the protoxide of manganese from to 7. 

 If we can find any law which approximately includes these 

 eight analyses, we shall have strong reason to believe that it 

 has some foundation in nature. 



The following is the constitution of each of the eight speci- 

 mens, expressed in atoms of the ingredients, and re- 

 duced to such a scale that the atoms of silica are 4.* 



If 



* I have in this paper used the notation which I have elsewhere recom- 

 mended (.Journal of the Royal Institution, No. 111). In this, S represents 

 silica, A alumina, C lime, M magnesia, K potassa, Fe and Fes the prot- 

 oxide and peroxide of iron, in which the (iroportions of oxygen are as 

 2:. '5; Mn the protoxide of manganese; as, arsenic, s sulphur, c' carbonic 

 acid. I have used the atomic numbers of Berzelius. 



In the Hiil. Mag. and Annals for August, Mr. Prideaux has dissented from 

 some of the views concerning chemical notation to which I have just referred. 



I will 



