VOL. LXVII.] PHILOSOPHICAL TRANSACTIONS. J5y 



without being at all able to form a probable conjecture at the cause of my suc- 

 cess or disappointment. The pores were so very small, that they were not dis- 

 coverable when the metal had received a good face and figure on the hones, nor 

 till the last and highest polish had been given ; and then it frequently appeared 

 as if dusted over with millions of microscopic pores, which were exceedingly 

 prejudicial in 2 respects; for first, they became in time a lodgment for a mois- 

 ture which tarnished the surface ; and 2dly, on polishing the speculum, the putty 

 necessarily rounded oft' the edges of the pores, so as to spoil a great part of the 

 metal, by the loss of as much light and sharpness in the image as there were 

 defective points of reflection in the metal. Besides the trouble of a great num- 

 ber of experiments, in order to get rid of this mischief, and to ascertain the 

 cause to which it was owing, there was this additional inconvenience attending 

 it, that the fault was not discovered, as before observed, till a great deal of 

 trouble had been taken in grinding and even polishing the metal, the whole of 

 which was rendered useless by the mortifying discovery of this defect. 



I was extricated at last from this difficulty, and in some measure by accident. 

 Having one day made a great number of experiments, and having melted down 

 all the good copper I had or could procure ; though puzzled and fatigued, yet 

 not caring to give it up, I recollected that I had some metal which was reserved 

 out of curiosity, and was a part of one of the bells of St. Andrew's which had 

 been re-cast. Expecting however very little from this gross and uncertain com- 

 position, I was nevertheless determined to see what could be made of it by en- 

 riching the composition with a little fresh tin. Accordingly casting a metal with 

 it, it turned out perfectly free from pores, and in every respect as fine a metal as 

 ever I saw. I could not at first conceive to what this success was owing ; but at 

 last I hit upon the real cause of that defect which had given me so much 

 embarrassment and trouble, during a course of near 100 experiments, and in 

 consequence fell upon a method which ever after prevented it. 



I had hitherto always melted the copper first, and when it was sufficiently 

 fused, I used to add the proportional quantity of tin ; and as soon as the two 

 were mixed, and the scoria taken ofl^, the metal was poured into the moulds. I 

 began to consider that putty was calcined tin, and strongly suspected, that the 

 excessive heat which the copper necessarily undergoes before fusion, was suf- 

 ficient to reduce part of the tin to this state of calcination, which therefore 

 might fly off" from the composition in the form of putty, at the time the metal 

 was poured into the flasks. On this idea, after I had furnished myself with some 

 more Swedish copper and grain-tin, both which I had always before used, I 

 melted the copper, and having added the tin as usual to it, cast the whole into 

 an ingot ; this was, as I expected, porous. I then melted it again, and as in 

 this mixed state it did not acquire half the heat which was before necessary to 



