376 On JLLANITE, a new 



5. From the preceding analyfis, fuppofing it accurate, it fol- 

 lowed, that the mineral was compofed of 



Silica, - - - 37* 10 " 



Lime, - - - 9 t2 3 



Alumina, - - - 4. 14 



Oxide of iron, - 42.40 



Volatile matter, - - 3,98 



96.91 

 Lofs, -■"-..■■- 3*09 



100.00 



But the appearance of the fuppofed oxide of iron, induced me 

 to fufpecT:, that it did not confift wholly of that metal. I 

 thought it even conceivable, that the yttria which the mineral 

 contained, might have been rendered infoluble by the applica- 

 tion of too much heat, and might have been concealed by the 

 iron with which it was mixed. A number of experiments, 

 which it is needlefs to fpecify, foon convinced me, that, befides 

 iron, there was likewife another fubftance prefent, which pof- 

 ferled properties different from any that I had been in the ha- 

 bit of examining. It poflefled one property at leaft in com- 

 mon with yttria ; its folution in acids had a fweet tafte ; but few 

 of its other properties had any refemblance to thofe whica the 

 chemifts to whom we are indebted for our knowledge of yttria, 

 have particularifed. But as I had never myfelf made any ex- 

 periments on yttria, I was rather at a lofs what conclufion to 

 dra^v. From this uncertainty, I was relieved by Mr Allan, 

 who had the goodnefs to give me a fmall fragment of gadoli- 

 nite, which had been received directly from Mr Ekeberg. 

 From this I extracted about 1 o grains of yttria 5 and upon com- 

 paring its properties with thofe of the fubftance in queftion, I 



found 



