26 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 13, No. 2 
rimeter, so the final loss is less than given in the table. On the other 
hand, the wire will not take the calorimeter temperature just at 
the surface, and the point of mean effective contact may be 1 or 2 
em. within. This will make the air loss as great as if part of the extra 
length were added to the wire in air, so it will be safer, in applying table 
1, to take 2X as 2 cm. more than the free length of the wire. 
TABLE 1.—Approximate Loss To THE AiR, IN CALORIES, IN 3 MinuTEs FROM HALF 
or 2 Leap Wires or LENGTH 2X BACH. 
DIAMETER, Inm. 
paths Oe 
1 0.25 
l ampere 1 0.00016 0.023 
2 0.0012 0.14 
3 0.0039 0.35 
+ 0.0086 
5 0.016 
10 
14 0.14 
5 amperes 1 0.57 
2.5 
6.5 
5 amperes 0.5 0.14 
1.5 
3.5 
Of more importance may be the effect of the uncertain contact on 
the flow of the heat that does not go into the air. If the contacts are 
symmetrical no error results; if one is effectively 1 em. deeper than the 
other, that is equivalent to a displacement of 0.5 em. in the potential 
lead attachment. This, for 1 ampere, 110 volts, gives a systematic 
error of about 8 per million of the total heat with No. 24, and of only - 
30 per million even with No. 30 leads. 
Thus with a little care even No. 30 wire will usually make satisfac- 
tory leads for a heater that can raise a liter calorimeter 5° in an ap- 
propriately short time. Coiled leads of larger wire are safer, but less 
convenient. Direct leads of larger wire increase various undesirable 
leakages unnecessarily.° If 5 amperes is used the heat produced 
in the wires is 25 times as great, but at 110 volts the allowable leakage 
® For a precaution necessary with No. 18, or even with No. 24 leads see Walter P. 
White, op. cit., p. 1884. 
