ON UNITS. 169 



of the hydrogen thermometer, to which the name of normal thermometric 

 scale should be given. ^ 



Third Part. 



Certain proposals still awaiting their final form and considerations ou 

 certain points : — 



The Calorie. 



It has been pi-oposed on various occasions not to adopt an independent 

 unit of heat-energy. In the equation 



, aMcd=i 



we have — 



M = unit of mass of water ; 



c = specific heat of water ; 



6 = the interval of temperature corresponding to one degree ; 



j = either the C.Gr.S. unit or the practical unit of energy ; 



a = the heat-equivalent of this energy. 



The proposal in question is equivalent to putting a = 1, and, as we 

 cannot modify M, to fixing c and Q so as to satisfy the equation. 



Messrs. Preece and Forbes proposed in 1889 to modify the temperature 

 system, while Prof. Ostwald prefers to attribute to water a specific heat 

 difiering from unity (equal to 4*2, when M = 1 gramme, 6 = 1° C, and 

 j = 1 joale). I am also of opinion that it would be best to agree to this 

 last alternative finally ; but I have pointed out previously why I think it 

 preferable for the moment to adhere to the status quo, and to consider this 

 reform one of those to be accomplished ia a more or less distant future. 

 For the present it is, above all things, necessary to define the calorie — 

 that is to say, the temperature at which the specific heat of water shall 

 be considered equal to unity — but I do not think this decision can be 

 taken before the question has again been submitted to a very accurate 

 experimental investigation. The temperature measurements would, of 

 course, be necessarily referred to the normal scale. 



Radiation. . 

 No special unit has hitherto been employed for radiation in general, 

 and the only unit which has been used in a particular case is irrational ; 

 for in solar radiation all authors have taken for their unit the calorie yer 

 minute, a unit which conforms to no system at all. No radiation has 

 hitherto been measured to an approximation equal to that with which we 

 know the mechanical equivalent of the calorie, hence there would be no 

 inconvenience in using the watt as the unit of intensity of radiation. 

 This unit would yield very acceptable numbers for the majority of cases 

 of total radiation which we have to measure. We might even push our 

 analysis still further, and divide the spectrum of an incandescent body into 

 bands O'l /^ (/x=micron) in breadth ; the unit would then be the watt in each 

 of these bands. The spectram would thus be completely defined, and 

 moreover the sums of the partial intensities would be numerically equal to 

 the total intensity of the radiation. 



' The differences between the indications of various kinds of mercury thermo- 

 meter and the hydrogen thermometer have been determined by very accurate 

 experiments. The comparisons carried out by M. P. Chappuis with mercury thermo- 

 meters of hard French glass are the most complete ones that have hitherto been 

 made. We also possess reduction tables for thermometers made of hard Jena glass, 

 of French lead glass, English lead glass, &c. It is much to be desired that only 

 samples of hard glass should be used in the construction of thermometers. 



