22 B. Silliman on the relation between the intensity of 
theorem, namely, 3 cubic feet, while No. 3, which represents — 
exactly this limit, brings the result within the range of experi- 
mental error—it being impossible to make two series of 15 
photometric observations which will accord more closely than 
1 fish-tail burner consuming 5 cubic feet. gave 132-94 candle power. 
2 “ “ 1°5 oe “cc 12°89 “ 
Computing the second observation we have: 
: By Farmer’s theorem for No, 2 14322 candle power 
“direct ratio ° eeee * 2 
This is an extreme case in which the volume of gas con- 
sumed in the second observation is far too low, but it is clear 4 
that by the old rule the result coming from the consumption 
of so small a volume of gas is perfectly worthless, while by 
Farmer’s theorem the difference of 10-28 candles is within 7-7 — 
_ per cent, while if the true intensity of this remarkable gas is 
le 1ere is good reason to believe it should be, at 142 — 
HACE, a8 
candles, the agreement in the two observations is absolute. 
_ Every photometric observer can confirm the results here given 
by reference to his own records of former observations, or by 
direct experiment designed to test the accuracy of the theorem 
here announced, 
In Sugg’s “Gas Manipulation ” (London, 1867), page 64, is 
atabular statement of the results of an experiment designed 
e f=) rm , ( og 
a form of argand) to develop the highest intensity of 
perhour, By this statement the burner in question produced 
_ Yeducod to 4°5 cubic 
as 
ee 
Mes ea 
