368 TRANSACTIONS OF SECTION B. 



obtain a pure hydrosol, free from other substances, it is best to boil the glucoside 

 with water in a quartz vessel. 



The aqueous solutions of saponarin, obtained by any of the above methods, 

 are seen to be colloidal when examined ultra-microscopically. The alkaline 

 solutions appear to contain particles of various sizes ; neutralisation is marked by 

 the appearance of a large number of small particles. The author is indebted to 

 Dr. W. M. Bayliss, F.K.S., for these observations. 



As in the case of starch, the addition of a pure aqueous iodine solution to the 

 hydrosol does not bring about a blue coloration which, however, appears on the 

 addition of potassium iodide. It is found that in the present case many other 

 salts (ammonium chloride, aluminium sulphate) are also effective. The blue 

 colour disappears on the addition of organic solvents or of sodium thiosulphate ; 

 also (temporarily) on heating. 



If, together with the iodine, but little salt is added, the blue adsorption com- 

 pound remains dissolved, but the further addition of electrolytes causes floccu- 

 lation and the separation of a dark blue gel. The concentration of salt required 

 depends primarily on the cation, and is in accordance with Sehulze's law. Thua 

 in a given case for a number of K, Na, NH 4 salts this concentration was 0"053 

 gram equivalents per litre, for Ba, Ca, Sr about 0-0047, and for Al salts 0-0006. 

 If electrolytes are excluded as far as possible, the blue hydrosol does not travel 

 appreciably with a current of 110 volts. 



The conductivity of the blue hydrosol is less than that of the potassium iodide 

 contained in it, so that some of the salt is probably adsorbed, and this salt, or 

 its cation, favours the adsorption of iodine by saponarin, just as cations facili- 

 tate the absorption of congo red by filter paper. 3 Owing to experimental diffi- 

 culties it has not yet been possible to determine the effect of the valency of the 

 cation. 



The iodine content of the blue hydrogel depends on the concentration of tho 

 iodine left in aqueous solution, as was shown for starch by Raster,' 1 and for basic 

 lanthanum acetate by W. Biltz. 



Preliminary experiments yielded the following results. 



C„ 

 0-0000533 



0-000171 

 0-000376 

 0000606 

 0-000792 



Here (J represents the free iodine (in grams) remaining in 1 c.o. of water and 

 C„ the iodine precipitated with 1 gram of Saponarin. The third column is calcu- 

 lated according to the formula C, = a[CJ", where a :- 0-47 and n = 0'166. The 

 value for n is much smaller than is generally the case with solid adsorbents {e.g., 

 charcoal). It is apparently somewhat higher in very dilute solution. The 

 formula does not apply to concentrated solutions of iodine, for which n falls off 

 much below 0'166. 



(iv) The Colloid Theory of Cements. By Dr. C. H. Desch. 



The explanation of the setting of calcareous cements, as caused by the 

 crystallisation of the products of hydrolysis from a supersaturated solution, 

 fails to account for the great mechanical strength of such cements. The colloid 

 hypothesis proposed by Michaelis attributes the setting to the formation of a 

 gel of calcium silicate, which subsequently hardens by loss of water and 

 adsorption of lime. Microscopical examination confirms Michaelis' view. 

 The only constituent of the cement, which is acted on is the alite. The hydro- 

 lysis of the complex substances contained in the alite first sets free calcium 

 aluminate, which separates in the form of crystals. This constitutes the initial 

 set. The calcium silicate is more slowly hydrolysed, and the calcium mono- 

 silicate produced, being extremely insoluble, separates as a colloidal gel. A part 



a Bavliss, Biochem. Journ., 1906, 1, 175. 



i Annalen, 1894, '283, 360. « Bcr., 1904, 37, 719. 



