. . [ I.I ] 



lated powder from which it had been feparated. The phseno- 

 menon remained to be accounted for; this was effeded by an 

 accurate examination of powder in its damaged ftate, when with 

 the aiiiftance of a convex- lens I difcovered a new cryftalHzation 

 of the nitre (called by powder-makers the ftarting of the petre) 

 which having been partially diflblved, fhot its minute falts to 

 the furface of the grain, where they appeared like the fpiculas of 

 hoar-froft, 'till broken and detached by the attrition produced in 

 moving the powder, they were converted into that duft, which 

 confequently contained the effejice of the compofition. The 

 eprouvette experiments were corroborated by the lefs fallacious 

 teflimony of analyzation, and this erroneous pradice correded. 

 The foregoing obfervations muft, however, be applied to fuch 

 powder only as though injured, in part retains its grain. When 

 it is fo far damaged as to cake, the cryftallization of the nitre 

 being more compleat, and its Hiootings larger, they adhere more 

 tenacioufly to the lumps, or when broken off are prevented by 

 their magnitude from that intimate admixture with the fulphur 

 and charcoal duft which is elTential to forcible explofion ; all 

 attempts to renovate powder, when thus far decorapofed, are 

 nugatory, and can only be ditilated by ignorance or fraud ; it 

 fhould be immediately transferred to the extradling houfe. The 

 ftrength of new powder is not diminiftied by reducing it to duft, 

 but rather increafed, a fecret well underftood by powder mer- 

 chants, who mix duft in fmall quantities with that powder they 

 apprehend will not rife to proof. It was formerly the pradice of 

 government to manufa(3ure their powder as fmall in the grain 

 as that made at Dantzick or Battle is at prefent ; whether the 



large 



