877 



///jO in flus way, if free fVom pei-oxide. AUliouo-li the Icinperalure was 

 increased up to 'i6"25° (J. llii^ lime, we were not able to tiiid any 

 heat-effect. The crucible was intUited like a balloon, but did not 

 crack. After being opened, we found the powder baked together, 

 however ??(>/ molten. An unimportant trace of peroxide only was formed. 

 Therefore we must conclude, that as Troost already pointed out, the 

 melting-point lies very high, — higher than can be determined by 

 means of our i)latinumresistance-furnaces and thermoelements. Probably 

 the meltingpoint will be in the neighbourhood of 1700° C. 



Tiiirdly we determined the meltingpoint of the IJ^ Si O^ quite in 

 the same way, and found it (without any correction), in several 

 determinations; at an E.M.F. of 12640 M.V., if the thick thermo- 

 element (0,6 m.ra.) wTis used. As the cori-ection of it was — 35 ?*I.V. 

 foj- this temperature, the true meltingpoint is 1256'" C. (G. Th.), thus 

 not ditfering appreciably from the formerly obtained value. The 

 heat-eflfect was not large; but if the thermoelement is fixed in a 

 suitable way at the bottom, which must be bent a little inwardly, 

 the lieat-efFect is clearly localised on the heating-curves. On cooling, 

 we found a cr^^stallization at 1249° C. 



It is of course possible, that this temperature of 1256° C. indeed 

 corresponds to a steep and not very extended maximiuu in the cnrve. 

 However we made some more experiments, to find out the form of 

 the curve for mixtures, which are still richer in Li J), by means 

 of the described new method. With a mixture, corresponding to 54 

 weight-procents Li^O (or 71 Mol. proc. Li^O), we found a good 

 observable heat-effect at 1405° C. ; then the bulb cracked by the 

 enormous vapour-tension. Thus it can be, judging from this, that the 

 temperature of 1256° C. must be considered rather as a temperature 

 of "transformation", at which the orthosilicate melts under dissocia- 

 tion into Li^O, and into a liquid, whose composition is very close 

 to that of the pure compound. 



§ 13. Microscopical investigations. All preparations were micros- 

 copically investigated, and the results were in every case compared 

 with those of the thermical determinations. It is an agreeable task to us, 

 to express our thanks once more to Dr. F. E. Wright in Washington 

 for his kindness and readiness to look over our preparations again, 

 and for the information he has given us in some doubtfnl cases. 



In general we can say, that the results of these investigations by 

 means of the microscopical method agree fairly well in all respects, 

 with the conclusions drawn from our thermical work. The prepara- 



