697 
element has only little influence on the observed temperature. Both 
the quantity of substance that is to be heated (metal of which the 
thermo-element consists) and the quantity of substance which directly 
causes the heating (salol), are about proportional to the width of the 
thermo-strip. The rise of the temperature will, therefore, not depend 
on this width. 
The width can, however, have influence on the rise of the 
temperature of the joint, because the heat which the thermo-strip 
leads off through the glass wall, is proportional to the width. As 
an approximate calculation teaches, this quantity of heat is, however, 
very small, and will, therefore, not appreciably influence the result 
of the measurement. 
Observations with thermo-elements of different widths (75 to 
200 u) have proved that the rise of temperature decreases slightly, 
when the width of the element increases. This influence of the 
breadth is very probably owing to a slight obliqueness of the 
thermo-strip. If this is exactly at right angles to the axis of the 
tube, it indicates the temperature in a plane that is at right angles 
to this axis, and in which the temperature depends but little on the 
place. If, however, the strip is placed obliquely, it indicates a mean 
temperature of different perpendicular sections of the tube. The 
thickness of the layer over which the temperature is averaged, is 
proportional to the width of the strip. In consequence of this the 
top of the curve, which strictly speaking, ought to be perfeetly 
sharp *), is rounded, and the more so, as the thermo-element is 
broader with the same obliqueness. [f the thermo-element is adjusted 
with great care, this error cannot reach a great amount either. 
3. The influence of the thickness of the thermo-element is more 
difficult to estimate a priori, than that of the breadth. It is self- 
evident that the rise of temperature measured with thick material, 
is smaller than that which is determined with thinner thermo- 
elements. Whether, however, the real temperature is sufficiently closely 
approximated when the thickness amounts to 5 u, as was the case 
in my first measurements, can only be decided when the influence 
of the thickness on the rise of temperature, is examined. For this 
reason I have repeated the measurement with the same tube and 
at the same temperature of the thermostat, with thermo-elements of 
different thicknesses, varying from 1.4 to 13 u. As was to be expected 
the rise of temperature decreased with increasing thickness of the 
thermo-element. The difference in temperature, however, amounted 
!) i. e. where the differential quotient of the temperature as function of the 
time, would have to be discontinuous. 
