344 PROCEEDINGS OF THE AMERICAN ACADEMY 



pointed to the aperture in tlie screen could be opened or closed at 

 pleasure, and the resistance of the strip measured, as it varied through 

 the eflPect of the i-adiant heat. In this way were examined various 

 metals such as gold-foil, platinum-foil and various grades of platinum- 

 wire, including some x^^jxr *^'^- "^ thickness ; gold-leaf gummed on 

 glass ; extremely thin sheet-iron, both blackened with camphor-smoke 

 and without such treatment, etc. The lamp-black augmented the heat 

 registered, but, if too thick, produced anomalies of its own, due to its 

 hygroscopic properties, which doubtless exist when it is used on the 

 thermopile, but are not so obvious there. For example, the warm 

 breath on such a lamp-blacked strip gave the indication of cold at the 

 first moment, possibly owing to the decreased resistance from absorbed 

 moisture. 



Metals deposited on films of glass are found not to answer our pur- 

 pose, because of the great amount of heat conducted away by the 

 glass, however thin. 



The requirements include, as was seen both from these preliminary 

 trials and from obvious theoretical considerations, considerable elec- 

 tric resistance, great change of that resistance by temperature, lamina- 

 bility, sufficient tenacity in the thin metal to enable it to support its 

 own weight, and freedom from oxidation. 



Iron would fulfil these conditions ver}^ well except the last, but it is 

 liable to rust. This tendency can be partly overcome by the applica- 

 tion of a thin coat of oil. Gold-leaf produced by the ordinary gold- 

 beater's process lacks continuity, being filled with minute rents, and 

 other metals are disqualified by other objections, such for instance as 

 low melting-points. That the teinperatui'e of metallic strips of the 

 thickness used may be very high, in spite of their great radiating 

 surface and even when the battery is feeble, is seen from such an 

 example as the following : — 



An iron strip 7 mm. long, 0.088 mm. broad, 0.003 mm. thick, hav- 

 ing the resistance of about 2^ ohms, was subjected to a current of 

 about 0.6 Weber which had before produced a uniform cherry-red glow 

 throughout the same length of platinum wire ^^^> cm. thick. The iron 

 glowed more brightly, but only for about 2 mm. at the centre, and was 

 melted at that point in about five seconds. 



A number of experiments were tried to determine the proper excess 

 of temperature of the strips used in the Bolometer over that of the 

 surrounding case, for this excess (due to the heating by the battery 

 current) must always exist ; and the amount to give the best effect 

 depends on many circumstances, and can only be determined by trial. 



