Dec. 27, 1877] 



NATURE 



173 



5. But if the tumbler be removed when the fly has come to 

 rest, it remains at rest, or nearly so. 



6. If the tumbler be more strongly heated, positive rotation 

 begins as promptly as with light. In this ca:-e the tumbler must 

 not le leU long over the radiometer, for fear the vacuum should 

 be spoiled by the evolution of gas from the pith. 



7. If the tumbltr be heated by holding it over the spout of a 

 kettle from which steam is issuing, and held there till the con- 

 densation of waler has approximately ceased, and be then 

 inverted ovtr the pith radiometer, the bulb is immediately 

 bedewed, and a w^'^'aftVir rotation is almost immediately setup, 

 though sometimes, just at the veiy first moment, there is a trace 

 of positive T( tation. The negative rotation is lively, but not 

 Ipstirg ; and after fifteen seconds or so, is exchanged for a posi- 

 tive rotation, which is not lively, but lasts longer. 



8. If the tumbler be lifted when the negative rotation has 

 ceased, and the dewed surface be strongly blown upon, a lively 

 but biicf positive rotation is set up. 



9. To produce positive rotation by blowing it is not essential 

 that tl e bulb be wet. If it be merely warm, and the circum- 

 stances aie such that the fly is at re^t for the moment, or nearly 

 so, blowing produces positive rotation, though much less strongly 

 than when the bulb is wet. 



10. If the tumbler be heated as in § 7, and inverted over the mica 

 radiometer, the rotation is positive, as when the tumbler is dry. 



11. If the tumbltr or a cup be smoked inside (to facilitate 

 radiatii n), heattd to a little beyond the temperature of boi ing 

 water, and inverted over the pith radiometer, a positive rotation 

 is produced ; and if, when this has ceased, which takes place in 

 a couple of minutes or so, the heated vessel be removed, a 

 negative, though not lively, rotation is produced as the apparatus 

 cools. 



12. These results do not seem difficult to co-ordinate so far as 

 to reduce them to their proximate cause. 



As regards the small quantity, if any, of heat radiated directly 

 across the g'ass of the bulb, the action of which was experi- 

 mentally distinguishable by its promptitude, both radiometers 

 behaved in the ordinary way. 



13. As regards the mica radiometer, when the bulb gets heated 

 and radiates towards the fly the fly is impelled in the negative 

 direction as if\\\t. white pearly mica were black and the lamp- 

 black weie white. And there is nothing opposed to what we 

 know in supposing that such is really their relative order of 

 darkness as regards the heat of low refrangibility absorbed and 

 radiated by the glass ; for the researches of Melloni and others 

 have shown that lampblack is, if not absolutely white, at any 

 rate very far from black as regards heat of low refrangibility. On 

 the other band, g!ass and mica are both silicates, not so very dis- 

 similar in chemical composition, and it would not therefore be 

 very wonderful, but rather the reverse, if there were a (general 

 similarity in their mode of absorption of radiant heat, so that the 

 heat most freely radiated by glass and accordingly abounding in 

 the nidiation from thin glass such as that of the bulb, were 

 greedily absorbed by mica. The explanation of the reversal of 

 the action when heat and cold were interchanged is too well 

 known to require mention. 



14. Wiih the pith radiometer, when the bulb as a whole is 

 heated, and radiates towards the fly, the impulse is positive, 

 though less strong than in the case of the mica (§ 4) ; and when 

 the bulb as a whole is cooler than the fly the impulse is negative 



{§")• 



But to explain all the phenomena we must dissect the total 

 radiation from or towards the bulb. When I first noticed the 

 negative rotation produced by a heated wet tumbler, I was dis- 

 posed to attribute it to radiation fiom the water, which possibly 

 the glass of the bulb might be thin enough to let pass ; but when 

 I found that hot water in a gla.ss vessel outside, even though the 

 g'ass of it were thin, produced no sensible effect, and that 

 blowirg on the heated bulb when it was dry produced a similar 

 effect to blowing on it when dewed, though of much less amount, 

 I perceived that the moisture acted, not by direct radiation from 

 if, and in consequence of a difference of quality between the 

 radiations from glass and water, but by causing a rapid super- 

 ficial heating oi the bulb; and, similarly, the blowing on the 

 dewed surface acted by causing a rapid superficial cooling. 

 When the dry tumbler radiates to the bulb, the radiation is 

 absorbed at various depths ; the absorption is most copious, it is 

 true, at the outer strata, but still the change of temperature is 

 not by any means so much confined to the immediate surface as 

 when we have to deal with the latent heat of vapour condensed 

 on it, or obtained from it by rapid evaporation. 



Hence, thin as is the glass of the bulb (about 002 in. thick), 

 we must still, in imagination, divide it into an outer and inner 

 stratum, and examine the effects of these separately. The heat 

 radiated by either stratum depends only on its temperature, but 

 the radiation from the outer, on its way to the fly, is sifted by 

 passing through the inner, and the portion for which gla's is 

 most excessively opaque is in great part stopped. It apj ears 

 from the observed results that the residue acts decidedly nega- 

 tively, while when the bulb is pretty uniformly heated there is 

 positive action. We may in'er that if it were possible to heat the 

 inner stratum alone it would manifest a very decided positive 

 action. 



15. In the struggle between the opposing actions of the outer 

 ard inner strata we see the explanation of the strange behaviour 

 of the pith radiometer. In the experiment of § 7 the outer 

 stratum at first shows its negative action, but quickly the inner 

 also gets heated, partly by conduction from the outer, partly by 

 direct radiation from the tumbler, and then the inner prevails. 

 In the experiment of § 5 the whole bulb cools, partly by radia- 

 tion, partly by convection, while the fly remains warmer ; and 

 the slightly greater coolness of the outer thanot the inner stratum 

 makes up for the superiority of the inner when the two are equally 

 cool, so that the antagonistic actions nearly balance, and slight 

 causes, iuch as greater or less agitation of ihe air, su'tceto make 

 the balance incline one way or other. That the inner stratum 

 would prevail if the two were about equa ly cooled may be 

 inferred from the behaviour of the radiometer when the bulb is 

 pretty uniformly heated (§§ 4, ll), or shown more directly by 

 cooling the bulb with snow, when a negative rotation may be 

 obtained. 



16. The complete definition of a radiation would involve the 

 expression of the intensity of each component of it as a function 

 of some quantity serving to define the quality of the component, 

 such as its refractive index in a standard medium, or its wave- 

 length, or the squared reciprocal of the wave-length. ^ The 

 experimental determination of the character, as thus defined, of 

 a radiation consisting of invisible heat-rays is beset with difficul- 

 ties, at least in the case of heat of extremely low refrangibility ; 

 and in general we can do little more than speak in a rough way 

 of the radiation as being of such or such a kind. It is obvious 

 that the behaviour of radiometers by itself alone affords no indi- 

 cation of the refrangibilities of the kinds of heat with which we 

 have to deal ; neveitheless, by combining what we know of the 

 behaviour of bodies in respect to radiations in general (especially 

 luminous radiations, which are the most easily studied) with 

 what we observe as to the motions of radiometers, we may arrive 

 at some probable conclusions. 



17. We may evidently conceive a series of ethereal vibrations 

 of any periodic time, however great, to be incident on a homo- 

 geneous medium such as glass, and inquire in what manner the 

 rate of absorption would change with the period ; though 

 whether we can actually produce ethereal vibrations of a very 

 long period is another question, seeing that we can only act on 

 the ether by the intervention of matter, and are limited to such 

 periods of vibration as matter can assume when vibrating mole- 

 cularly, in a manner communicable to the ether, and not as a 

 continuous mass, in the manner of the vibrations which produce 

 sound. We may inquire whether, on continually increasing the 

 period of vibration, the glass (or other medium) would ultimately 

 beconre and remain very opaque, or whether, after passing 

 through a range of opacity, it would become transparent again, 

 on still further increasing the period of the incident vibrations. 



18. This is a question the experimental answer to which, as 

 it seems to me, could only be given, in so far as it could be 

 given at all, as a result of a long series of experiments, of a 

 kind that Melloni has barely touched on. A variety of consider- 

 ations, which I could not explain in short compass, lead me to 

 regard the second alternative as the more probable, namely, 

 that, on increasing the periodic time, homogeneous substances 

 in general (perhaps even metals, though this is doubtful) become 

 at last transparent, or at least comparatively so. The limit of 

 opacity, in all probability, varies from one substance to another ; 

 and the lower it is, the lower would be the lowest refrangibility 

 of the radiation which the same substance is capable of emitting. 



19. In what immediately follows I shall suppose accordingly 

 that glass is strongly absorbing through a certain range of low 



' A map of the spectrum, constructed with the squared reciprocals of the 

 wave-lengths for aDscissas, would be referred to a natural standard, no less 

 than that of Angstrom, wliich is constructed according to wave-lengths ; 

 while it would have the great advantage of admitting of ready comparison 

 with refraction spectra, the kind almost always used. 



