368 On Salt Solutions and Attached Water. 



the slope of ascent thence to zero. Indeed every tangent to 

 the left-hand branch of each curve makes a more acute angle 

 with the zero axis than any tangent to the right-hand branch 

 of the same curve. If this be true throughout, the unknown 

 right-hand branch of the HC1 curve must be very nearly at 

 right angles to the axis of 0° C. There is no evidence to show 

 that the right-hand branch is continuous with the left-hand. 

 In the case of the halogen salts this almost appears to be di- 

 stinctly shown not to be the case. If the branches be conti- 

 nuous, it appears that the colder the cryohydrate the less the 

 curvature there. If the branches be discontinuous, the colder 

 the cryohydrate the more acute is the angle between the tan- 

 gents at the intersection of the curves. This rule holds good 

 in all cases ; but between NaCl and KI there is considerable 

 difference of sharpness at the cryohydrate, attending identity of 

 temperature. It is not the case that the deeper the valley the 

 more precipitous the ascent, travelling from left to right ; nor 

 is it true that the deepest valley has the widest mouth. That 

 is, salts the most soluble at 0° do not have the coldest cryohy- 

 drates, and therefore do not form the most powerful cryogens. 



In the shape of the bottoms of the valleys (that is, the cur- 

 vature about the cryohydrates) the curves of the oxygen salts 

 show a likeness to one another as distinguished from the halo- 

 gen salts. Considering the former by themselves, we might 

 be inclined to view their two branches as being in all cases 

 continuous. To settle the question decisively whether these 

 curves are continuous, hundredths of degrees would have to 

 be read with precision, and the initial formation of ice noted 

 within that range. If we judge by the aspect of the halogen 

 curves and the general continuity in properties between the 

 halogen and the oxygen salts, I am inclined to consider the 

 two branches discontinuous in all cases. 



The tracings assume, of course, somewhat different positions 

 if they are referred to molecular ratio instead of to absolute- 

 weight ratio. Without determining each tracing for all its 

 intermediate values, we may join by straight lines the points 

 representing the molecular solubility at 0° with the molecular 

 value of the cryohydrate. While each of these lines is, of 

 course, parallel to the corresponding line derived from fig. 1, 

 being got by determining its extremities by division of the 

 percentage by the molecular weight, such division necessarily 

 brings those lines relatively further towards the left whose 

 salts have heavy molecules. The extreme points of the lines 

 are got by dividing the percentage values of fig. 1 by the 

 molecular weights and multiplying by 100. I know no pro- 

 perty of the salts here under examination which would have 



