Chap, i.] Phenomenon at Mofco'W. \399 



air infmuates itfelf ; but this has been refuted by water 

 being frozen under an exhaufted receiver, and the fame 

 cavities being found in the ice. The ice, in fad, in- 

 ftead of being heavier, was found to be lighter. The 

 expanfion of ice, indeed, is evidently owing to the 

 cryftaliization of the water, and the particles afiuming 

 a different arrangement, and not being in clofe con- 

 tact. If this is admitted, as I apprehend it muft, we 

 cannot fay, with propriety, that the folid particles of 

 water expand. The Ipiculas, of which ice confifts, cut 

 each other at angles of fixty or one hundred and twenty 

 degrees. 



When it fnows at Mofcow, and the air is not tot} 

 dry, it is obferved to be loaded with beautiful cryf- 

 tallizations, regularly flattened, and as thin as a leaf 

 of paper. They confiit of an union of fibres, which 

 (hoot from the center to form fix principal rays, - 

 which are themfelves divided' into extremely fmall 

 blades. 



It appears, however, that the air previoufly con- 

 tained in water is fet at liberty on the congelation of 

 that fluid, and may thus partly contribute to the fwel- 

 ling of ice, and occafion ibme of the cavities obfervable 

 in it. 



When water freezes flowly, its furface preferves its 

 tranfparency to Tome depth, owing to the air which is 

 feparated paffing downwards as the freezing proceeds. 

 The air bubbles, however, foon begin to colled, and 

 produce confiderable inequalities, .which increafe as 

 they approach the center. When, on the contrary, 

 water is frozen with great rapidity, the bubbles of air 

 difperfe themfelves pretty equally through the mafs, 

 which, in confequence, becomes opake through its 

 tyjbiole fubftance. 



Reaumur obferves that call iron melted, in patting 



from 



