550 
1 1 
ge SON nae 
lim So) 
Sufticient attention has not always been paid to this relation. For 
instance in Circular 35 of the Bureau of Standards 24 edition 1915 
side by side with the melting point of palladium 1549° the value 
c, = 14500 is found. This would give L = 2,895, a value which 
is 0,25 °/, too high, whereas the experimental determinations do not 
differ from each other by more than 0,08 °/,. 
I. LANGMUIR*) assumes c, = 14392. M. Prrani?) c, = 14400. 
In consequence of this greater uncertainties arise than are necessary 
in view of the good agreement of the most recent measurements. 
2. The above discussion naturally leads to a simple method of 
standardizing optical pyrometers, provided with colour-filters. The 
ratio v is measured of the intensities transmitted by the filter at the 
melting points of palladium and of gold. The effective wave-length 
may then be derived from 
ee LM 1.2542 
a logv oe log v 
As an instance, if the red filter N°. F 4512 of Scnorr and Gen. 
is taken about 5,8 mm. thick, and the effective wavelength between 
the two points is determined, the effective wave-length for other 
ranges of temperature may be derived from Hypr’s calculations. In 
this manner a very simple method of standardising is obtained. 
Summary : 
To a given value of the melting point of palladium a definite 
value for c, corresponds. 
If the melting point of palladium is taken as 1549° (seale of 
Day and Sosman) c, must be taken equal to 14465 + 5. If on the 
other hand c, is taken 14800 (P.T.R. scale), it follows that the 
melting point of palladium is 1557°. 
Physical Laboratory of the N.V. Philips- 
Incandescent-lamp- factories. 
1) Phys. Rev. (7) 153, 1915. 
2) Verh. D. phys. Ges. (17) 226, 1915. 
