TRANSACTIONS OF SECTION G. 683 
in different parts of the world, these large bulb standard photometric lamps have 
proved very useful, and also they have proved of utility in obtaining the proper 
coefficients for the correction of the candle power of flame standards for atmo- 
spheric moisture and temperature. 
7. Some Investigations on the Ten-candle Power Harcourt Pentane Lamp 
made at the National Physical Laboratory.!. By CuirrorD C. Paterson. 
This paper gave briefly the results of investigations carried out at the National 
Physical Laboratory in order to determine changes in the illuminating power of 
the ten-candle Harcourt lamp due to variation 
(i.) In barometric pressure. 
(ii.) In the quantity of water vapour present in the air, 
When photometric comparisons are made between the pentane standard and 
a source of light unaffected by atmospheric conditions, as, for instance, an electric 
glow lamp, errors of the order of 5 per cent. may be introduced into candle-power 
measurements if corrections are not made for the hygrometric and barometric con- 
ditions existing at the time. 
In order to ascertain the amount of variation, photometric comparisons were 
made against two large bulb Fleming-Ediswan electric standard glow lamps. The 
double comparison method only was employed, and the two electric lamps used 
to standardise a comparison glow lamp anew for each experiment, so that it was 
only necessary to burn the standards for five or ten minutes at a time. 
Tt is found convenient to state humidity volumetrically, as the number of litres 
of water vapour to a cubic metre of pure dry air, at the barometric pressure 
existing at the time ; so that if 
b = Reading of barometer in mm., 
e = Aqueous pressure, 
e, = Vapour tension of carbon dioxide present in the atmosphere, 
the litres of water vapour per cubic metre of pure air 
e 
ee LOO, 
b—e-e, a 
Upwards of sixty observations have been made on different days over a range 
of humidity varying from five to twenty litres per cubic metre, which are about 
the limits obtained under ordinary conditions in a well-ventilated photometer 
room ; eighty per cent. of these, when corrected by means of the formula, fall within 
+ or — } per cent. of ten-candle power. 
The barometric pressure in these observations has varied from 739 to 780 mm. 
of mercury, so that, applying the method of least squares for these two variables, 
the following formula has been obtained for correcting the candle-power of the 
lamp to the standard atmospheric conditions of 760 mm. of mercury and ten litres 
of water vapour per cubic metre of pure dry air. 
Candle power = 10 + 0:066 (10 — <)—0-008(760 —8), 
where ¢ is the humidity as explained above and 6 the height of the barometer 
in mm. 
From this it will be seen that a variation of one litre per cubic metre in the 
moisture causes a variation in candle-power of about 0:7 per cent., and that 
10 mm. change in barometric pressure brings about an alteration of 0:8 per cent. 
in the illuminating power of the lamp. 
The standard humidity of ten litres per 1,000 has been fixed upon as being the 
mean value for three years found at Kew Observatory. The figure is also borne 
out by observations made at the Meteorological Office in Victoria Street, London. 
1 Published in the Hlectrician, August 26, 1904. 
