50 



find also, that cold is generated, by the melting, & heat, by the hardening, of 

 spermaceti. The cold produced by the melting of spermaceti is sufficient to 

 cool a quantity of water equal to it in weight above 70 degrees, & nearly the 

 same degree of heat is produced by the hardening of spermaceti. It was tried 

 by putting cold spermaceti into hot water, & hot spermaceti into cold water. 



P. 46. Some tin & lead were melted separately in a crucible, & a thermo- 

 meter put into them, & suffered to remain there till they were cold. The 

 thermometer cooled pretty fast, till the metal began to harden round the edges 

 of the pot : it then remained perfectly stationary, till it was all congealed : it 

 then began to sink again. On heating the metal, with the thermometer in it, 

 as soon as the metal began to melt round the sides, the thermometer became 

 stationary as near as I could tell at the same point that it did in cooling, and 

 remained so till it was entirely melted. On putting a thermometer into 

 melted bismuth, the phenomena were the same, except that the thermometer 

 did not become stationary till a good deal of the metal was hardened, unless 

 I took care to keep the thermometer constantly stirring about. It then re- 

 mained stationary till it was almost hardened. I do not know what this dif- 

 ference between bismuth & the 2 other metals should be owing to, except to 

 its not transmitting heat so fast as them. I forbear to use the word conduct- 

 ing, as I know you have an aversion to it ; but perhaps you will say the word 

 I use is as bad as that I forbear. 



P. 48. All the following mixtures, except the first, differ considerably from 

 the 3 simple metals in the manner in which they harden in cooling, as they 

 begin to abate of their fluidity in a heat considerably greater than that in which 

 they grow hard, whereas in the simple metals I could not perceive any differ- 

 ence between the heat in which they ceased to be perfectly fluid, & that in 

 which they hardened. 



P. 49. I think it seems likely that the reason why these mixtures begm to 

 abate of their fluidity in a greater heat than that in which they harden is, that 

 the metals of which the mixture is composed begin to separate as soon as the 

 heat is not sufficient to keep the mixture quite fluid. This is confirmed by the 

 following experiment. The mixture of equal quantities of lead & tin was 

 melted over again & suffered to remain quiet till cold ; it was then cut in two. 

 The specific gravity of the upper piece was 8*001, & that of the lower 9'031 ; 

 so that the upper piece seems to contain much less lead than the lower. 



P. 50. Thoughts concerning the above-mentioned phenomena. There are 

 several of the above-mentioned experiments which at first seemed to me very 

 difficult to reconcile with Newton's theory of heat ; but on further considera- 

 tion they seem by no means to be so : but to understand this you must read 

 the following proposition. — 



EXPERIMENTS ON ARSENIC. 

 [That the paper containing these experiments, like the former, was addressed 

 to some individual who pursued the same studies, appears from the following 

 expressions : " As I have spun these papers out to a great length, I will not 

 repeat the particulars of this experiment, ivhich I showed you before :" and 

 again, "thus mercury may be revived without the addition of any matter 



