Clarke and Schneider — Experiments, etc. 407 



ble, and it is not practicable to determine exactly what changes 

 have taken place The group — Fe — OH might behave like 

 MgOH, and yet subsequent alteration might prevent any esti- 

 mate of the extent of the reaction. 



By digestion with strong, aqueous hydrochloric acid, both 

 ripidolite and prochlorite were completely decomposed. Leuch- 

 tenbergite, on the other hand, left an insoluble residue, resem- 

 bling garnet, which was originally present as an inclusion in 

 the mineral. All of these minerals decompose with aqueous 

 acid more slowly than the serpentines. 



By sharp ignition, ripidolite and prochlorite give up in the 

 free state small quantities of silica, which are determinable by 

 extraction with soda solution. The percentages were as follows : 



Ripidolite. Prochlorite. 



Si0 2 liberated 2-98 2-45 



These quantities represent only one-tenth of the total silica 

 in the minerals, and have no evident significance in a discussion 

 of the chemical structure. 



Although ripidolite is readily decomposable by aqueous 

 hydrochloric acid, it appears to be split up by prolonged igni- 

 tion into a soluble and an insoluble part. A weighed quantity 

 of the mineral was heated for nine hours over the blast-lamp, 

 and then digested for three days with hydrochloric acid of sp. 

 gr. 1'12. The residue amounted to 4847 per cent of insoluble 

 matter, from which boiling with sodium carbonate solution ex- 

 tracted 28*73 of silica belonging to the decomposed silicates. 

 The final undissolved residue, 19*74 per cent, was analyzed; 

 and, treated as an independent substance, gave as follows : 



Si0 2 6-32 



Sesquioxides 67*81 



MsO 25-67 



99*80 



If the small quantity of silica here found, only 1*25 per cent of 

 the original material, be neglected as non-essential, the remain- 

 der, 18*49 per cent of the ripidolite, has exactly the composi- 

 tion of spinel. Like spinel it is quite insoluble, and in all rea- 

 sonable probability it may be regarded as that compound. The 

 formation of such a magnesian aluminate, MgAl 2 4 , is peculiarly 

 suggestive when we come to consider the structure of the chlo- 

 rites. 



Similar experiments with the prochlorite gave similar but 

 not identical results. After long ignition, six hours, and three 

 days' digestion with hydrochloric acid, 35*61 per cent of residue 

 remained, of which 18*16 per cent was insoluble in carbonate 



Am. Jour. Sci. — Third Series, Yol. XL, No. 239. — Nov., 1890. 

 26 



