at the Roycd Institution, 1 900-1 907. 389 



molecules and it is to this circumstance, it may be supposed, that ice 

 owes its peculiar viscosity and plasticity. As the temperature is 

 lowered, the liquid molecules gradually give place to sohd molecules 

 and the ice becomes more and more nearly rigid ; consequently the 

 rate at which the volume alters diminishes as the temperature is 

 lowered. Sir James Dewar's observations show that the density of ice 

 at - 185^ is 0" 1)2999, so that the mean cubical coefficient of change 

 in volume between this temperature and 0° is 0' 00008099. This 

 mean value is only about half that observed between 0° and - 20°, 

 namely 0' 0001551. Although clear pieces of ordinary ice dropped 

 into liquid air crack in all directions, owing to the sudden cooling, 

 when clear pieces tljat have been cooled slowly in liquid air are in- 

 troduced into liquid hydrogen they do not crack — this behaviour is 

 again proof that the expansibility diminishes at very low temperatures. 



As the formation of ice from liquid water involves expansion, the 

 freezing-point is necessarily lowered by pressure ; in other words, 

 the molecules of the solid are decomposed by pressure and forced to 

 take on some other contexture. Various calculations have been made 

 as to what would be the behaviour of water at very low temperatures 

 and pressures, assuming that its properties change in the manner 

 observed under more ordinary conditions ; the conclusions arrived at, 

 however, have all been upset by Tamman's remarkable observation* 

 that under high pressure two solid forms of water exist which are 

 probably both denser than water; one of these melts at -15 '8^ 

 under a pressure of 5040 atmospheres. This conclusion is completely 

 confirmed by Sir James Dewar's observation, that if water be frozen 

 in a steel cyhnder in successive portions and lead shot be included in 

 the middle and upper portions, on cooling such ice to - 80° and sub- 

 jecting it to pressure, the lead shot do not fall through the ice even 

 under a pressure of 100 tons per square inch ; it is clear that pressure- 

 does not operate to the extent formerly supposed in depressing the 

 freezing-point of water. 



It may therefore be regarded as established that chemical affinity 

 can operate in more than one direction between the units of the water 

 complexes and that, provided the temperature be low enough, pressure 

 alone may be effective in curbing the tendency of affinity to act in 

 one particular direction or set of directions and even of compelling it 

 or enabling it to act in some other direction which is less preferred 

 under conditions of greater freedom. The complete investigation of 

 the optical properties of the solid forms of water at low temperatures- 

 and under various pressures would be of great interest. There if* 

 some reason to suppose that colour is more highly developed in ice 

 than in water l)ut I am not aware that the comparison has been made 

 in any satisfactorv manner. The molecules of hydrone per se should 

 be colourless, if the explanation of the origin of colour previously 



* Ann. der Physik., 1900, ii. 1. 



