^WB 



^ TRSORIBS OP CRVSTAttOGRAfHY* 



l«wi of the ilftTy parallel to each other. We will proceed on two fimllar 

 wftalT^and cryftals of the fame fubftance and equal in folidit)', dividing 

 theory thence the firll into one fpecies, the other into a different fpecies, of 

 jtfulting. parallelopipedons, equal in folidity but not in furface; and let 



the diviilon of each be pulhed to its lalt term. But as we are 

 come by fmooth fedliohs to parallelopipedons of different fpe* 

 cies, thofe fedions have alfo produced their differences : but 

 by fuppofltion thefe parallelopipedons are the refult of the lall 

 poflible term of divifion without defiroying the chemical com* 

 portion, and being equal in folidity, though not in furface, 

 they cannot contain each other ; therefore if their differences 

 are not integrant parts of both, thefe differences muft ceafe to 

 be homogeneous, and we come to a fort of chemical decom- 

 pofition. It is true we cannot execute this exceflive divifion, 

 but we can form a very corretSl idea of if» If the little paral* 

 lelopipedons contain two forts of elements, their differences 

 will alfo, but alfo in different proportions; and, fir, if you 

 will turn to Berthollet's Bejhmxhes on the Lwjjs qf Affinities^ 

 you will fee him in all his experiments proving, that however 

 perfedly a chemical decompofition may have been made, the 

 refults will always contain a certain portion of thofe fubflances 

 from which it was the objed of the operation to feparate them» 

 If thefe reflexions, Sir, are well grounded, do they not give 

 us hopes, and perhaps fliow the poffibility, of defcending from 

 the integrant particles to the conftituent particles ? This fe- 

 cond refearch is of the fame nature as the firfl. It is more 

 than probable that the conflituent particles themfelves are di* 

 vifible, having no determined figure, but are aggregations, 

 fubje£t to the fame laws as the integrant particles. The objeft 

 of the natural philofppher is not to difcover the forms of the 

 ultimate particles, but to determine their refpedlive pofitions ; 

 which, if ever they could be determined in the integrant par^ 

 tides and their component parts, the grand problem of chemi- 

 cal affinities would be fully folved; and fliould fuch ever be 

 the cafe, to the Abbe Hauy's theory would be due the merit. 

 The Encydop<£dia Britannica, under the article Chemijliy, in 

 the Supplement, p. 396, fays : 



" This theory, to fay no more of it, is, in point of ing6» 

 nuity, inferior to few ; and the mathematical fkill and induflry 

 of its author are entitled to the greatefl applaufe. 



♦' But 



