Nov. 28, 1878] 



NATURE 



89 



showed that phenomena feeble and contradictory when caused 

 by radiation external to the bulb, became \-igorous and uniform 

 when the radiation was applied internally by the agency of an 

 electrically-heated wire. It was hoped that some of the more 

 obscure phenomena shown by the deep cups with movable 

 screens in front (referred to above) might be intensified if set in 

 action by a hot wire. Several inds of apparatus and experi- 

 ments with them are described, but the results are too compli- 

 cated to be given in abstract. One experiment proves that the 

 direction of pressure is not wholly normal to the surface on which 

 it is generated, but that some of it is tangential. 



The author then describes the turbine radiometer, early speci- 

 mens of which were exhibited before the Royal Society on April 

 5, 1876. In the ordinary form of radiometer the number of 

 disks constituting the fly is limited to six or eight, a greater 

 number causing interference one with the other and obstruction 

 of the incident light. In the turbine form of fly there is no 

 such difficulty, the number of vanes may be considerably increased 

 without overcrowding, and with correspondii^ advantage. In 

 the earlier turbine radiometers the flies were made of mica 

 blacked on both sides, and inclined at an angle like the sails of 

 a windmill, instead of being in a vertical plane. This form of 

 instrument is not sensitive to horizontal radiation, but moves 

 readily in one or other direction to a candle held above or below. 

 A vertical light falling on the fly gives the strongest action, but 

 rotation takes place, whatever be the incident angle, provided 

 the light is caught by one siu-face more than by the other. Ether 

 dropped on the top of the bulb to chill it causes rapid negative 

 rotation. If the turbine radiometer is floated in a vessel of ice- 

 cold water, and the upper portion exposed to the air of a warm 

 room, it rotates rapidly in the positive direction, acting as a heat 

 engine, and continuing so to act xmtil the rotating fly has 

 equalised the temperature of the upper and lower portions of the 

 bulb. By reversing the circle of operations — by floating the 

 turbine radiometer in hot water and cooling the upper portion 

 of the bulb — the fly instantly rotates in the negative direction. 



After describing experiments in which the same fly was made 

 to rotate first in a large bulb and then in a small one at the same 

 degree of exhaustion, the author proceeds to discuss the in- 

 fluence exerted by the inner side of the glass case of the radio- 

 meter as a reacting surface. A flat metal band was put equa- 

 torially inside a radiometer, and lamp-blacked, so that the 

 molecular pressure generated under the influence of light should 

 react between the fly and the black band, instead of between the 

 fly and the glass side of the bulb. It w as found that the maxi- 

 mum speed with the band present was 40 revolutions a minute, 

 against 8J revolutions when the band was absent. 



The rotation of the case of a radiometer, the fly being held 

 immovable by magnetism, is next described. A preliminary 

 note on this subject having already appeared in the Proceedings} 

 it need not be again described in detail. Many different forms 

 of instnunent for effecting this rotation are described, and their 

 mode of action explained. 



The reacting inner siu-face of the envelope being thus proved 

 to be essential to the rotation of the fly, other instnmients were 

 luade in which this necessary reaction is obtained in a more 

 direct manner. In one, the radiometer is furnished with a fly 

 carrjing four flat aluminium vanes, polished on both sides. 

 Three vertical partitions of thin clear mica are fixed in the bidb, 

 with their planes not passing through the axis of rotation, but 

 inclined to it, thus throwing the obliquity off" the fly on to the 

 case, and giving three fixed planes for the reaction to take place 

 against. Candles arranged sj-mmetrically round the bulb make 

 the fly rotate rapidly against the edges of the inclined planes. 

 Breathing gently on the bulb gives negative rotation. A hot 

 glass shade inverted over the instrument causes strong negative 

 rotation, changing to positive on cooling. When thelfly it fur- 

 nished with clear mica or with silver flake mica vanes, the same 

 results are obtained as when aluminium vanes are employed. 

 The principal action is produced by dark heat warming the bulb, 

 screens, and vanes. 



The otheoscope is the next subject treated on in the paper. This 

 has already been given in abstract,* and need not be again re- 

 ferred to. Many different varieties of otheoscope are°fieured 

 and described. 



It was suggested by Prof. Stokes that a disk might be made 

 to revolve on its axis, and the author describes an instrument in 

 wtuch-this suggestion is carried out. The disk is horizontal, 



' Proc. Roy. Soc.. No. 168. March 30. 1876. 

 » Proc. Koy. Soc., No. xSo, April 26, 1877. 



mounted like the fly of a radiometer, and for lightness' sake is 

 of mica, blacked above. Fixed to the bulb above the disk are 

 four flat pieces of clear mica ; each extends from the side of the 

 bulb to near the centre, and ends below in a straight horizontal 

 edge, leaving just space enough for the disk to revolve without 

 risk of scraping. The edge is in a radial direction, and the 

 plane of the plates is inclined about 45'' to the horizon in the 

 same direction for them all. Exposed to the light of a candle 

 the rotation is against the edge. By slightly modifying this 

 form the instrument becomes much more sensitive. 



Whilst experimenting with the otheoscope it was found that, 

 for a given exhaustion, the nearer the reacting surfaces were 

 together the greater was the speed obtained. In the Proceed- 

 ings of the Royal Society for November, 1876,* the author 

 described an apparatus by which he was able to measiu-e the 

 thickness of the layer of molecular pressiu^e generated when 

 radiation impinged on a blackened surface inclosed in an 

 atmosphere the rarefaction of which could be varied at will. 



It was found that in this apparatus repulsion could be ob- 

 tained at ordinary atmospheric pressures. Observations are 

 given at normal pressure and at various d^jrees of rarefaction, 

 with the driving and moving surfaces separated i, 2, 3, 4, 6, 8, 

 and 12 millims. ; and diagrams of the resulting curves are 

 shown when the atmospheric tension and the force of repul- 

 sion are used as abscissae and ordinates. The tables and curves 

 show that the law of increase of the force w ith the diminution 

 of the distance between the disks does not remain uniform at 

 all rarefactions. At the lowest exhaustions the mean path of 

 the molecules of the attenuated gas is less than i millim., as 

 rendered evident by the force of repulsion diminishing rapidly 

 as the distance increases. At exhaustions higher than 9 millims. 

 this condition alters, and as the gauge approaches barometric 

 height the molecular pressure tends to become uniform through 

 considerable distances, the mean path of the molecules now 

 being comparable with the greatest distance separating the 

 surfaces between which they act. 



A similar apparatus to the one in which the last experi- 

 ments were tried was used to measure the action at pressures 

 at and approaching atmospheric. At pressiu-es between atmo- 

 spheric and 210 millims. the first action is ver)' faint repul- 

 sion, immediately followed by strong attraction. The attrac- 

 tion then begins to decline, until, at 1 5 millims. pressure, it dis- 

 appears. At the same time the repulsion, which begins to be 

 apparent at 250 millims., increases as the attraction diminishes. 

 The author considers that the attraction is the result of air- 

 currents, caused by the permanent heating of the surface in fi-ont 

 of the movable disk. 



The paper concludes with experiments undertaken to measure 

 the amount of repulsion, using a horizontal torsion balance,* on 

 the principle of Ritchie's, in which the force of repulsion is 

 balanced by the torsion of a fine glass fibre. The pan of the 

 balance is a clear mica disk, and a similar disk is fastened to the 

 tube in which the beam oscillates. This fixed disk is lamp- 

 blacked on the upper side, and beneath is a spiral of platinum 

 wire, connected with terminals sealed through the side of the 

 tube. When the spiral is ignited by a constant electric current 

 the blacked mica disk fixed above it becomes heated, and the 

 molecular pressure thereby generated between it and the mica 

 pan causes the latter to rise. The glass thread attached to the 

 beam is thas twisted, and by means of a graduated circle the 

 nxmiber of degrees through which the thread has to be turned ia 

 order to bring the beam back to equilibrium is noted. This gives 

 a measurement of the pressure exerted, in torsional degrees, and 

 these are converted into grains by ascertaining how many tor- 

 sional degrees correspond to a known weight. A ray of light 

 reflected from a mirror in the centre of the beam is used as an 

 index, being brought back to zero at each experiment. The 

 author gives in a table, and also shows in the form of a curve, 

 the results obtained with this apparatus, giv'ing the force of 

 molecular pressure in grains weight at exhaustions varying 

 between 2,237 and 07 millionths of an atmosphere. 



Mathematical Society, November 14. — Lord Rayleigh, 

 F.R.S., in the chair. — The Treasurer's and Secretaries' reports 

 having been read and adopted. Prof. W. G. Adams, F.R.S., 

 consented to act as auditor. — The scrutators declared the follow- 

 ing gentlemen elected as the Council for the ensuing session, 

 viz., Mr. C. W. Merrifield, F.R.S., President; Prof. Cayley, 



' Proc. Roy. Soc., No. 175, voL xxv. p. 310. 



' For a description of this form of torsion balance, see the author's paper, 

 Phil. Trans., 1876, vol. clivL pt. 2, p. 371. 



