FACTITIOUS PUZZOLANA. 337 



<! repeated thefe experiments many times, and. had for a 

 medium refult from an hundred pounds, chemical weight, 

 50 parts of filex ; 

 31 — of iron ; 



16 of alumen ; 



3 of manganefe, and lots. 



ioo 



Ifthisanalyfis be compared with that of the puzzolana of*^^ 1 ^™ 

 Italy, which contains in 100 parts — ■ 

 50 of filex ; 

 25 of alumen ; 

 1 6 of iron ; 

 3 of lime ; 

 6 of lofs; 



.100 

 their refpe&ive properties may be appreciated according to 

 the proportions of their integrant parts. 



The excefs of alumen caufes the plaifters made from the Excefs of aiu- 

 Italian puzzolana to crack and chap, in the open air: this ^lia^kia^ t^ 

 fault arifes from their great oxidation. I have been able to crack, 

 replace in them thofe principles which they loft by decom- 

 pofilion. 



Ana!j/fit in the dry Way. 



I endeavoured to obtain a regulus from thefe oxides of iron Analyfis in the 

 by uting a violent heat. I followed the. procefs ©f r Kirwan jJJ r ^[. 8 7 *' 

 for the fufion of filiceous and argillaceous ores of iron ; yet I cefles; 

 never obtained a tingle metallic button ; and only found at the afforded no 

 bottom pf the crucible a vitrified mafs of an opaque black, or meta ' 

 a fcoria in the ftate of crude caft iron. 



Defirous to know if I could procure a malleable button The blowpipe 

 by ufing the blowpipe, taking borax for the flux and fuppott- JJJ|L un ucce *" 

 ing the oxide on charcoal ; I ftill could only obtain a fpongy 

 ingot refembling crude caft iron, and breaking both when hot 

 and when cold. 



Being placed on a fupport of glafs (according to my method The o*ide fufed 

 puplifhed in the Journal de Phyfique, Tome XXXI. pages on g!afs ' 

 116 and 139), the oxide fufed at die fecond attempt, the fup- 



Vol. XII. — Supplement. Z port 



