1 835 . ] Native Sulphate of A lamina and Iron . 61 



1. 4/S. + Al. S2 + 36 Aq. 



2. 3J/S'. + Al. S. + 341 Aq. 



3. 3/'S. + Al. Sii+27i Aq. 



And, as expressing the mean, we may adopt 

 3i/S'. + Al. SU 4- 32 Aq. 



Another specimen which had been preserved in a phial 

 for some years was also analyzed, and yielded. 



Sulphuric acid . . 35-600 - 2-37 atoms. 

 Protoxide of iron . . 13-500 - 1* ,, 



Alumina .... 7-127 - 1-05 „ 

 Water 43-773 - 12-9 



100-000 

 Which may be considered equivalent to/ S. + Al. S. + 

 13 Aq. with a great excess of acid. The salt had a strongly 

 acid taste. If we take the mean of this formula with those 

 which precede, we obtain nearly 



2/S. + Al. S. + 20 Aq. 

 which is quite different from the result of Phillips. 



Of the three analyses contained in the table, the third, 

 perhaps, approaches most nearly the mean composition of 

 this substance, as it corresponds with the first so far as 

 regards the acid and iron, and the water is identical with 

 the result obtained by direct experiment. 



From these facts, then, it appears that the hair salt of 

 the coal strata varies in its composition. But this deduc- 

 tion is what we should have been inclined to draw, from 

 the consideration of various analyses by different chemists, 

 of specimens of similar salts from other localities, which 

 affect the same form of crystallization, although consist- 

 ing of totally different constituents. Thus sulphate of 

 magnesia, sulphate of manganese, as well as sulphate of 

 alumina and iron are found, we have also seen in capillary 

 crystals. 



Upon what circumstances this remarkable asbestus form 

 of soluble salt depends, it is not easy to determine, because 

 they are indifferently met with in various species of rocks. 

 This form, however, in insoluble minerals, as has been 

 observed, is connected with serpentine rocks.* 



Thomson's Inorganic Chemistry, i. 161. 



