374 Count de Bournon on Bardiglione. 



still however retaining a slighter cohesion between its particles than 

 existed between those of the gypsum previous to its calcination, and 

 without receiving any degree of transparency. In the transition of 

 Plaster to this state of solidity, the gypsum has hitherto been 

 said to resume its water of cr^^stallization ; but what is this water 

 of crystallization ? what idea ought this expression to convey to the 

 mind ? 



If, in order to settle my opinion on the subject, I consult those 

 who have written on chemistry, or on mineralogy, the majority 

 conceive, that this water is that which a large quantity of salts and 

 earthy substances retain when they crystallize ; and that this water 

 is necessary, in these cases, to the ci-ystallization of the substances, 

 but constitutes no part of their essence. But how can the water, 

 absorbed by the plaster, whicli is evidently very different from water 

 of crystallization, be necessary to the process ? What part does the 

 water in crystallized gypsum act in the crystallization of it ? a satis- 

 factory answer to these two questions, is, I conceive, requisite in all 

 the cases In which they occur. 



Some of these authors think, that, in several instances, the water 

 enters into their composition as an essential ingredient ; and upon 

 this subject I refer to a passage in the first volume of the Mineralogy 

 of M. Brongniart, page 96; as also to a very judicious doubt ex- 

 pressed by the Abbe Haiiy, in his Mineralogy, vol. IV. p. 351. I 

 confess, that I had long ago adopted this opinion. But is all the 

 water, that may be included in these mineral substances, to be so 

 considered ? Certainly not. In many of them the water is foreign 

 to their substance, and has entered merely in consequence of the 

 attraction (to which I have given the name of attraction by approxi- 

 mation) more or less powerful, exerted upon it by their integrant 

 molecules ; in this case it is only imbibed and interposed between 



