CONTINUOUS ELECTRIC CALORIMETRY. 
125 
In working out these results, the standard value of the specific heat was taken 
as Js 0 = ’1400 joule per gramme degree, since the specific heat of mercury is 
approximately -g^th of that of water. If we take only the values of the difference D 
corresponding to the first and third flows, and calculate the values of h and J sd, 
according to the elementary theory, neglecting the correction term Id/dJsQ, we find 
J sd = (D' - D")/(Q' - Q") = - -0091/4-159 = - ‘00219, 
whence, h = "0568, J s = -1400 — "00219 = "13781 joule per gramme degree. 
As a verification, we may calculate the value of D for the intermediate flow (2). 
We find Ih = "0420, in place of the observed value "0416. The difference is only 
four parts in 10,000, on EC/c/d, and might well be attributed to errors of obser¬ 
vation in these preliminary experiments. 
Inserting the correction for the variation of the temperature gradient in the flow- 
tuhe, by subtracting li 2 J 3JsQ from each of the corresponding values of D, and then 
calculating as before, we find the corrected results given in the last column, which 
exceed those calculated on the elementary theory by nearly one part in 700. It 
should be remarked that this correction can be deduced with certainty from (1) 
and (3) without reference to (2). It cannot be calculated satisfactorily from three 
flows, as it depends on small differences. 
(36.) Correction of Results with Water Calorimeter. 
In applying the correction to the water calorimeter, it is necessary to take some 
account of the heat-loss from the outflow-tube round the thermometer bulb, as well 
as that from the fine flow-tube. This changes the numerical factors, which depend 
to some extent on the dimensions of the tubes, but the theory of the correction 
is otherwise unchanged. Assuming the heat-loss from the thermometer bulb to 
be two-fifths of the whole, which is sufficiently exact, since the whole correction is 
very small and a change of one-tenth in the ratio would not alter the result by more 
than 2 or 3 per cent, of itself, equation (10) becomes 
EC /tie - Js 0 Q = D = Js 0 Qd + h 0 (1 - add/ 10) + lif 1 l/25JsQ . . ( 12 ), 
and the correction to be added on this account to the value of the specific heat in 
joules, as calculated by Dr. Barnes, is given by the expression 
Correction for Variation of Gradient with Flow = -f- 1 l/) 0 2 /25J*-Q , Q ,/ . (13) 
In calculating this correction it is desirable to use the (corrected value of h 0 , which 
is obtained from that of h as given by Dr. Barnes by adding the terms 
Correction to value of h = + ahdO/10 — (11 A~/25-J.s) (1/Q' -f 1/Q") . (14). 
