380 Clarke and Schneider — Experiments upon, the 



discussed in our earlier paper, from which 6 '29 per cent of 

 magnesia was removable ; but more than double that amount 

 is required to represent the MgOH which must be present 

 under any recognized theory. 



These results were so unsatisfactory that further work 

 seemed to be desirable, but the tediousness of the process was 

 discouraging. As a possible abbreviation of the method a 

 new device was tried, which, although as yet unconclusive in 

 its outcome, is nevertheless interesting. Ammonium chloride, 

 as is well known, dissociates upon vaporization into ammonia 

 and hydrochloric acid. Will it then, dissociating in presence 

 of a silicate, act in any way by virtue of the acid so liberated ? 

 This question is still under investigation ; but some prelimi- 

 nary data may here be given. In each experiment the mineral 

 was intimately mixed with ten times its weight of sal ammoniac, 

 and then heated in a platinum crucible, well below redness, 

 until all the ammonium salt had been vaporized and driven off. 

 The residue was then leached out with water, and the soluble 

 magnesia determined. Upon olivine and waluewite there was 

 very little action, as in the case of the treatment with gaseous 

 hydrochloric acid ; but upon the chlorites and serpentine there 

 was a vigorous attack. The magnesia thus rendered soluble 

 was as follows : 



Ciinochlore. Leuchtenbergite. Serpentine. 

 Percent MgO, 2-12 3-98 4'93 



In a second series of experiments the minerals were heated 

 as before with ammonium chloride until the latter was driven 

 off. Then a fresh charge of chloride was added to the contents 

 of the crucible and expelled like the first. Finally a third 

 charge was added and likewise expelled. By this means more 

 magnesia was converted into chloride than in the first experi- 

 ments, as the subjoined figures show : 



Ciinochlore. Serpentine. 



3 heatings. 2 heatings. 3 heatings. 



Soluble MgO, 3-80 14-30 10-63 



The irregularities in all these data remain to be explained. 

 So far the results have only qualitative value. In those min- 

 erals which must contain the group MgOH, some action takes 

 place, but in variable amount. But the magnesium silicates 

 which cannot contain MgOH are practically unattacked, and 

 so far all the evidence is harmonious and clear. Contrary to 

 our earlier expectations, however, we cannot trust the dry 

 hydrochloric acid reaction quantitatively, at least until its con- 

 ditions and its nature are more fully understood. We can 

 determine when the group MgOH is present as an essential 



