34 REFLECTING TELESCOfE. 



small ladle, out of the crucible, and poured it on a cold flag; 

 and observed the following appearances : 

 Fiom a cast First. If the metal assumed, in cooling, a lively blue, or 



pccimen. parple colour, commonly intermixed with clouds, or shades of 

 green or yellow; and if, when broken, the face of the fracture 

 exhibited a silvery whiteness', as bright and glistening as quick- 

 silver, without any appearance of gfain^ or inequality of tex- 

 ture ; then the degree of saturation o f the metal, with the tin-, 

 was complete and perfect. 

 And the frac- Secondly. If the surface of the metal became of a dun oc 

 ture and co- mouse colour, and especially if of a brown or red; and, when 

 broken, the fracture exhibited a more yellow, or tawny hue, 

 than that of quicksilver ; then the quantity of tin in the com- 

 position was deficient, and it was necessary to add more.* 



Thirdly. If the colour was an uniform dull blue, like lead, 

 where broken, discovered a dull colour, with a coarse gi"ain, 

 like facetts ; the due saturation v/as exceeded, and there was 

 an over proportion of tin in the metal. 

 Explanation of These colours would be more distinct, if a small quantity 

 the colours. ^f ^}^g metal were cast in a flask, which had been previously 

 smoaked, by a candle, made of resin mixed with tallow ; in 

 which way I used to prepare the moulds. I attribute the 

 formation of the colours to this: that, as the calx of every 

 metal has its own peculiar colour, so, the heat of the melted 

 mass, calcining some of the particles on its surface, which are 

 in contact with the air, these display the colour of the calces of 

 those ingredients which prevail in the composition. Whence, 

 it may be expected, that, if the copper is the redundant metal, 

 the mass will exhibit a reddish tinge, which is appropriate to 

 the calx of copper; and, if the tin be prevalent, a blueish die 

 ought to appear. Either of these colours, therefore, appear- 

 ing unmixed, shews the redundance of that metal, to which 

 each belongs. And, as brass, when cast alone, has alway^s a 



* This can always be done by degrees, and without any trou- 

 ble, till the point of saturation is found; whereas, if too much 

 tin were added at first, there would be a necessity for meltiu'g 

 more copper separately', and repeating the whole process : and 

 different specimens of copper will require different proportions 

 of tin ; so tliat the due quantity can never be known, a priori, but 

 on trial only. 



