dialyser and allowed to diffuse into water, the latter being occasion- 

 ally changed. After the lapse of five days seven-eighths of the 

 original silicic acid was found to remain liquid upon the septum, and 

 to be so free from hydrochloric acid and chloride of sodium as not to 

 give a precipitate with acid nitrate of silver. The true hydrated 

 alumina, and also Mr. Cram's metalumina, were obtained soluble by 

 dialysing solutions of these oxides in the chloride and acetate of the 

 same metal. So also the hydrated peroxide of iron, in addition to 

 the hydrated metaperoxide of iron of M. Pean de Saint Gilles, and 

 the soluble hydrated chromic oxide. The varieties of prussian blue 

 are obtained soluble by dialysing their solution in oxalate of ammonia, 

 the latter salt diffusing away. Stannic and titanic acids appear as 

 insoluble gelatinous hydrates. 



A solution of gum-arabic (gummate of lime), dialysed after an addi- 

 tion of hydrochloric acid, gave at once the pure gummic acid of 

 Fremy. Soluble albumen is obtained in a state of purity by dia- 

 lysing that substance with an addition of acetic acid. 



Caramel of sugar, purified by repeated precipitation by alcohol and 

 afterwards by dialysis, contains more carbon than any of the cara- 

 melic bodies of Gelis ; it forms a tremulous jelly when concentrated, 

 and appears decidedly colloidal. Caramel, like all other colloids, has 

 a soluble and an insoluble modification. The latter has its solubility 

 restored by the action of alkali, followed by that of acetic acid and 

 subsequent dialysis. 



Dialysis proves highly useful in separating arsenious acid arid 

 metallic poisons from organic fluids. Defibrinated blood, milk, and 

 other organic fluids charged with a few milligrammes of arsenious 

 acid, and placed upon the dialyser, were found to impart the greater 

 proportion of the arsenious acid to the external water in the course 

 of twenty-four hours. The diffusate was so free from organic matter, 

 that the metal could be readily precipitated by sulphuretted hydro- 

 gen, and the quantity weighed. 



Ice at or near its melting-point appears to be a colloidal substance, 

 and exhibits a resemblance to a firm jelly in elasticity, the tendency 

 to rend, and to redintegrate on contact. 



The consideration of the properties of gelatinous colloids appears 

 to show that osmose is principally an affair of the dehydration of the 

 gelatinous septum under influences having a catalytic character, and 



