" Law " in Physical Optics. 



331 



Here it is evident that the value of ft remains practically 

 constant through a range of 60° C. The values given were 

 calculated from 0° as one of the temperatures in each case. 



If ft has the meaning here assigned to it, the above 

 measurements indicate that 1 gram of carbon disulphide 

 actually occupies 0*046 cubic centimetre. Moreover, if we 

 take the ratio ft/v, we find it to be 0"062 at 0° C, which is 

 again very near the average value yielded by the many liquids 

 already given. 



Water. 



Temperature, 



v. 



/XD. 



(3 calculated 

 from first. 



/3 calculated 

 for each pair. 



20-9 

 31-0 

 56-0 

 87-0 

 94-0 



1-00196 



1-0044 



1-0149 



1-0336 



10386 



1-33287 

 1-33177 

 1-32776 

 1-32107 

 1-31943 



0-204 

 0-150 

 0114 

 0-108 



] -204 

 1 -114 

 | -085 

 1 -067 



The result here is quite different from that shown by carbon 

 disulphide ; ft varies with every change in the temperature. 

 It is given in two columns, the first containing values obtained 

 by taking each observation in conjunction with that at the 

 lowest temperature. The next column contains values obtained 

 by taking each pair of observations. This last series is the 

 more suggestive. 



1st. It shows ft to be diminishing as the temperature rises, 

 and gives the value it has between 87° and 94°. At this 

 temperature, where the water is approaching its boiling-point, 

 the value of ft becomes 0*067 cubic centim., a close approxi- 

 mation to that given by the other liquids. 



2nd. The ratio ft/v at the high temperatures is 0*063, 

 almost exactly the ratio 0*064 given by the other liquids. 



These results are suggestive, and at once raise a question 

 as to the meaning of the diminution in ft as the temperature 

 rises towards boiling-point. Any one acquainted with recent 

 researches on the physical properties of liquids will at once 

 recall those of Ramsay and Shields on Molecular Complexity; 

 those of Thorpe and Rodger on Viscosity of Liquids, &c, as 

 bearing on this question. The investigations mentioned, 

 along with others, have rendered it almost certain that the 

 molecule of water is a complex of the form tiH 2 0, and that 

 the value of n in this expression diminishes as the temperature 



