BV COMPRESSION. 399 



he llluftrated ivhal he laid by a6lual trial. All iron, when ex- flee) at i higher» 



pofed to a certain heat^cruilies and crunables under theliammer; ^"^^-^Xr. wh?te^'^ 



but the temperature in which this happens, varies with every hear, a:id old 



different fpecies. Thus, as he Ibewed me, cafi iron cruOies f^'^;^ ^^ ^ ft'" 

 ' ' , . ,, Jiigher temper- 



in a dull red heat, or perhaps abJut 1.5" ol Wedgwood ; atuie. 



fteel, in a heat perhaps of 30°; Swedilli iron, in a bright 

 white heat, perhaps of 50° or 60'^ ; old fable itfe'f, likewife 

 yields, but in a much higher heat, perhaps of 100**. 1 merely 

 gueffed at thefe temperatures; but I am certain of this, that in a 

 heat iimilar to that in which Swediih iron crumbled under the 

 hammer, the old fable withflood a ftiong blow, and fecmed to 

 polfefs contiderable firmnefs. It is from a knowledge of this 

 quality, that the blackfmith, when he firft takes his iron from 

 the forge, and lays it on the anvil, begins by very gentle 

 blows, till the temperature has funk to the degree in which the 

 iron can bear the hammer. I obferved, as the ftrong heat of 

 the forge a6led on the Swedidi iron, that it began to boil at the 

 furface, clearly indicating the difcharge of fome gafeous mat- 

 ter j whereas, the old lable, in the fame circumflances, ac- 

 quired the fliining furface of a liquid, and melted away without 

 any eflfervefcence, I procured, at this time, a confulerable 

 number of bars of that iron, which fully anfwered my expec- 

 tations. 



By the experiments laft mentioned, a very important point The complete 

 was (j[ainetl in this inveHiealion ; the complete fufibilily of the^"1'"'"y °^ ^^^ 



,. n , , n 1 -A r Carbonate under 



carbonate under preUure bemg thereby eltablimed. But from preflure was af- 

 this very circumftance, a necedity arofe of adding fome new *^^"*"'=<^ '" 

 devices to thofe already defcnbed: for the carbonate, in fu- ^eis, ' " 

 fion, fpreading itfelf on the infide of the tube containing it, 

 and the two uniting firmly together, fo as to be quite infepa- 

 rable, it was impodible, after the experiment, to afcertaiti 

 the weight of the carbonate by any method previoufly ufed. 

 I therefore determined in future to adopt the following ar- 

 rangement. 



A fmall tube of porcelain (ik, Fig. 23.) was weighed by Arrangemer>t for 

 means of a counterpoifeof fand, or granulated tin; then tbe ^J'^'^^',";^/""™^ 

 carbonate was firmly rammed into the lube, and the whole which arofe from 

 weighed again: thus the weight of the carbonate, previous to ^''^^/"^*°"* ^'^^ 



" . ° ° ' f^ carbonate was 



the experiment was afcertained. After the experiment, the put into a fmall 

 tube, with its contents, was again weighed ; and the variation P'^f'^^'ai" tube 

 of weight obtained, independently of any mutual adion that had frcure^inT^' ^ 



taken •^''S^f' Thcfe 



