o8H Prof. H. E. Armstronij on Low-Tem^Jerature Research 



simple molecules but are characteristic of particular structural 

 arrangements of such simple molecules; in fact the remarkable 

 variation in properties met with among the iron alloys may well 

 be traceable, at least in large part, to structural differences. Un- 

 fortunately such problems cannot be dealt with at present except 

 inferentially ; on this account the insight afforded by the experi- 

 ments carried on at low temperatures is of very special value and 

 interest. 



Ice at Low Temperatures. — Water is undoubtedly one of if not 

 the most remarkable of known substances, both in its physical and 

 its chemical behaviour. On cooling the liquid, it ceases to contract 

 at 4° C. and expands slightly until the freezing-point (0° C.) is 

 reached, when it becomes ice ; the conversion of the liquid into 

 solid is attended by a great increase in bulk, the density of ice at 

 0° C. being only 0' 9151)9 grams per cubic centimetre. There can 

 be no doubt that ice molecules are present in solution in the liquid, 

 it may almost be said, long before solidification sets in, the tempera- 

 ture at which the density of water is at a maximum (4° 0.) being 

 that at which the contraction which the liquid undergoes on 

 cooling is just balanced by the expansion consequent on a certain 

 proportion of the molecules becoming arranged in the manner in 

 which they are present in ice. 



Liquid water doubtless consists only to a limited extent of the 

 simple molecules which constitute vaporous water, that is to say, of 

 molecules of the composition represented by the chemical formula 

 OHo (hydrone) ; in addition to these, it probably contains complexes 

 of several kinds formed by the association of the simple molecules of 

 hydrone. Ice presumably is formed from some one kind of the 

 complex molecules. In the case of oxygen as in that of carbon, 

 there is every reason to suppose that the force of chemical affinity 

 is exerted in certain specific directions and that when the affinities 

 come into operation the molecules necessarily tend to assume certain 

 relative positions. The oxygen in hydrone has a considerable 

 amount of residual affinity, which is the cause of its activity : when 

 owing to the reduction of temperature the vibrations of the molecules 

 become sufficiently damped to allow the residual affinity to act, the 

 affinities become, as it were, interlocked in certain directions and 

 crystallisation is the consequence. Water, near to the ice point, 

 may be likened to a pile of bricks placed one upon the other in 

 close contact and therefore occupying minimum volume : ice may 

 be likened to the same bricks arranged in open triangular form ; so 

 placed they enclose a hollow space and — inclusive of this space — 

 occupy a greater volume than when placed directly one upon the 

 other. But the formation of ice is an incomplete process of solidi- 

 fication ; apparently ice is not a homogeneous substance but an 

 equilibrated mixture, in certain proportions, of soHd and Hquid 



