484 0« the Corundum SUfte/rotn AJia, 



diftinft cryftal of which was 3,950 ; of a fragment of ruby-coloured corundum, 3,959 ; and 

 of a fragment of corundum with vitreous luftre, 33954. 



It may be objefted to me, that Bergman has ftated the variety of fpecific gravity in gem« 

 to be fo great, as to leave no certain rule of judging thereby of the fpecies. He obferved, 

 that the topaz generally prevails in weight, being from 3,460 to 4,560 ; 0ie ruby from 3,180 

 to 4,240 ; then the fapphire from 3,650 to 3,940*. But, in the preceding page, he had 

 faid, " Analyfi cryftallorum tarn ejufdem quam diverfx figurse multum lucis fcientia cx- 

 peftat. nise quarum antea compoGtionem explorare licuit naturali forma per artem privatae 

 erant." It is not, therefore, an hypothefis unworthy of examination wliich I advance, that 

 gems, derived from the reftangled oftoedra, whofe fpecific gravity is above 3,300 to 3,800, 

 will be found to be diamonds, or oiSloedral rubies ; and thefe will be eafily diftinguilhed 

 from each other by their luftre and .hardnefs. Diamonds, whether red, yellow, blue, or 

 white, being hardeft, though their fpecific gravity will be lefe j viz. from 3,356 to 3,471, 

 as I found among different diamonds in my collection : whereas the oftoedral ruby was 

 from 3,571 to 3,625, and inferior in hardnefs not only to the diamoitd, but to the corun- 

 dum ; the fpecific gravity of which, in its different appearance of form and colour, I found 

 to vary from 3,876 to 4,166 ; and I fuppofe it to be fubjeft to a variation from 3,300 to 

 4,300 : after which the jargon will come with a fpecific gravity of 4,600 ; eafily diftin- " 

 guiihed alfo by its cryftallization from the above-mentioned gems. The above fpecific gra- 

 vities Mr. Hatcliett very obligingly aflifted me in taking with his accurate fcales, in the tem- 

 perature of 60°. It will not be underftood that I depend entirely on the fpecific gravity \ 

 on the contrary, I connedt this quality with cryftallization : hardnefs is the next criterion \ 

 and analyfis muft feparate the component parts, and demonftrate the analogy, or identity, 

 of fubftances, or of compounds. The improvements of Mr. Klaproth's procefs are evi- 

 dent by the comparifon of his firft analyfis, and in his laft analyfis of corundum. 

 In the firft it confifted of 



Corundum earth - - « - 6% o 



Siliceous earth -- » • « 3^5° 



Iron and nickel - » u ^ » o 50 



By the laft analyfis of Mr. Klaproth, the corundum of the peninfula of India con- 

 fifted of 



Argillaceous earth , - - » 



Siliceous earth « - _ 



Oxide of iron _ . . - - 



Lofs - - » » • • 



» De Terra Gemmaium. Berg. Opufc. vol. II. p. 104. 



The 



