formed b})- the elastic and invisible vapor which is present everywhere in the atmosphere; 

 the precipitation of aqueous vapor is clearly owing to the cold produced by the radiation 

 of bodies towards a clear sky. Looking at the question in this way, leaves, wood, glass, 

 varnish, lampblack, become covered with dew, because they emit heat easily, and are con- 

 siderably cooled under a clear sky. Metals, on the other hand, remain dry, in conse- 

 quence of the difficulty they have in radiating their heat to the upper regions of the at- 

 mosphere; and, in fact, a great difference is observed between the indications of a thermo- 

 scope, when a vessel of polished metal, full of boiling water, and an exactly similar one 

 coated with lampblack, are succcssivel}^ presented to it; the action of the second being 

 much greater than that of the first. The deduction is correct; but it must be allowed 

 that it does not necessarily appear so to everybody. Indeed, Prevost, and Saussure be- 

 fore him, attributed the absence of dew on metals to an electric force; Leslie explained 

 the same phenomenon by a particular repulsion, which, he said, existed between metallic 

 surfaces and watery vapor; and those who maintained that dew arose from the earth, ex- 

 plained the same thing by the heat and electricity disengaged by the chemical action of 

 metals upon the particles of this same vapor, at the moment of their passage to the liquid 

 state. To show that these hypothesis are untenable, I first take three thermometers 

 with graduated stems, and on each tube I fix a small cork about five or six millimetres 

 above the bulb. This cork helps to support the metallic cases in which the thermometers 

 for experiments on nocturnal cooling, are inclosed. The first case consists of a small, thin, 

 polished silver or copper cup, like a common thimble, and large enough to contain the 

 bulb of the thermometer; the second is a tin cylinder, open at one end and closed at the 

 other; this serves as an envelope for the graduated tube. The two metallic pieces (which 

 can easily be put off or on,) are kept in their places by the elasticity of the coik. 



In the next place I procured three tin cups, each having a lateral opening near its bot- 

 tom, through which the bulbs of the prepared thermometers can be passed, while the 

 stems with their envelopes remain horizontally on the outside. These cups are supported 

 by fine metallic tubes, provided with covers of the same nature, and the whole were ex- 

 posed to the air on a calm fine night. One of the thermometer cases was blackened, and 

 the other two were in their natural state, and the cups were sometimes open and some- 

 times shut. Such was the apparatus with which I compared the nocturnal radiation of 

 silver with that of lampblack. Suppose the cups to be first shut, the three thermometers 

 then mark the same temperature. Then by opening two of the cups, and leaving the third, 

 containing one of the bright thermometers, shut, it will be seen that the metallic thermo- 

 meter which is now exposed to the air, falls so little that hardly any change can be obser- 

 ved, except with the finest instruments; while the thermometer coated with lampblack, 

 falls very visibly, and after a few minutes it will mark three or four degrees less than the 

 thermometer in the closed cup — an evident proof that this difference is owing to the heat 

 radiated by the lampblack, and not at all to the contact of the exterior air, which 

 equally surrounded the polished metallic casing of the other exi^osed thermometer. My 

 results confirm, in a striking manner, the assertion of MM. La Provostaye and Desains, 

 viz: that the emissive power of metals is much less than the experiments of Leslie, Du- 

 long and Petit, led people to suppose. 



The radiating power of lampblack being 100, that of laminated silver I found to be 

 3.026. MM. LaProvostaye and Desains find 5.37 for silver chemically precipitated on 

 copper, and 2.1 when the silver is polished. According to the last mentioned gcntle- 

 the emissive power of recently laminated silver is 2.94; while 2.38 is that of 

 silver burnished. 



