304 Clarke and Schneider — Experiments upon the 



character in such a way as to shed much light upon their inner 

 chemical structure. In general, the methods which have been 

 employed by us are not new, except in their application to the 

 silicate problem ; but in their analogies to the processes used 

 for the elucidation of organic compounds, we believe that they 

 will be found interesting. 



In brief, our mode of procedure has been as follows : First, 

 each mineral, selected and purified with great care, has been 

 completely analyzed ; and in each case enough material was 

 ground to a uniform sample to suffice for all subsequent exper- 

 iments. Secondly, each mineral has been subjected to the 

 action of dry hydrochloric acid gas, under quantitative condi- 

 tions. For this purpose a quantity of material was weighed 

 out in a platinum boat, which was then placed in a glass tube 

 and heated in a stream of the dry gas until, after repeated 

 weighings, constant weight was attained. The sample was then 

 treated with water and a drop or two of nitric acid, and the 

 soluble constituents of the mass were determined by the usual 

 methods of analysis. In every instance a sample was thus 

 treated at a temperature between 383° and 412° C, but in 

 some cases other temperatures were studied also. For heating 

 the tubes an ordinary combustion furnace was used ; and the 

 temperature was kept within the indicated range by placing 

 on one side of the platinum boat a sealed capillary tube 

 containing lead iodide, melting point 383°, and on the 

 other side a tube containing zinc, which fuses at 412°. The 

 flames of the furnace were then so adjusted that the iodide 

 fused, while the zinc did not. For a higher range of temper- 

 atures, which was employed in some cases, the indicators simi- 

 larly used were lead chloride, m. p 498°, and silver iodide, m. 

 p. 527°. These melting points are given according to Carnel- 

 ley. As a rule, in each series of experiments the sample under 

 treatment was weighed every two hours ; and before reheating 

 it was stirred with a fine platinum wire in order to expose a 

 fresh surface to the action of the acid. By this mode of treat- 

 ment different minerals are very differently affected ; there 

 being almost no action in some cases, and very notable action 

 in others. With this action of gaseous acid, the action of 

 aqueous hydrochloric acid upon each mineral was carefully 

 compared, and some points of great importance were thus 

 brought out. As a rule, one gram of mineral was treated 

 with 75 cc. of fuming hydrochloric acid on the water bath, and 

 the mixture was evaporated to dryness. When decomposition 

 was not complete, the mineral was further treated with hydro- 

 chloric acid of sp. gr. 1-12 for three days or even longer, and 

 the amount of action was ascertained and recorded. In several 

 instances this treatment with aqueous acid was repeated after 

 strong ignition of a mineral, and it was sometimes found that a 



