January 24, 1895] 



NA TURE 



293 



If this were so, by using broad bands instead of lines, one 

 would expect to get a coloured line at each border, while the 

 central portion remained uncoloured. Prof. Liveing's theory 

 would lead one to expect a uniformly coloured hand. 



Accordingly, we prepared a disc, one half of which was 

 blackenel and the other half left white, as in Mr. Benham's 

 top, and on the white half were described three black circular 

 bands about one centimetre broad, with radii of about 47 and 

 10 centimetres respectively, and each with an arc of 45". The 

 outer band was described in an opposite sense to the two inner 

 bands, that is to say, in such a manner that if the disc was 

 routing so that the order of succession for the two inner bands 

 was black field, bands, while field, for the outer band it was 

 white, band.s, black. The disc was rotated by attaching it to 

 the spindle of an electromotor, the speed of which could be 

 graduated by friction against the spindle. 



On rotating the disc at a slow speed, in such a direction that 

 the order of succession for the two inner bands was black, bands, 

 white, the following results were obtained : — The two inner 

 bands were each bounded on both inner and outer edges by a 

 bright red line, fading off towards the centre in a dark, some- 

 what reddish, ground, which became less coloured and darker as 

 it approached the centre. The outer band presented a marked 

 contrast to these two ; the whole e.xtent of the band itself was 

 black, entirely free from colour, but on the white ground on both 

 borders of the band, and apparently outside it, there appeared a 

 brilliant coloured band, varying in colour from blue to green. On 

 reversing the direction of rotation, the appearances w-re exactly 

 reversed ; the inner bands now became black bordered by green, 

 and the outer band reddish black lined by bright red. 



We do not contend that this experiment proves that the efl'ect 

 is one of irradiation, for it might easily be supposed that the 

 heightened effect at the border is one of contrast with the 

 surrounding whiter coloured field. Our next experiment, on 

 the efl'ect of change in the speed of rotation, also tends to 

 negative the idea that the colours are due to irradiation, as there 

 is no reason to think that if irradiation were the cause, there 

 would be a change in colour with a change in the speed of 

 rotation, as was found to be the case. 



To tot the effect of change in the speed of rotation, a disc of 

 one of Mr. Benham's tops was detached, fixed to the spindle of 

 the electromotor, and rotated in such a direction as to cause the 

 three cenlial bands to appear red with a slow speed of rotation. 

 On gradually increasing the speed, a remarkable series of 

 changes in appearance presented itself. The bright blood-red 

 of the three inner lines gradually grew darker and duller, and 

 then passing rapidly through a transition, the shades of which 

 we were unable to observe, gave place to a most vivid green, 

 which in turn, with still increasing speed, passed through 

 another transition stage into blue, deepening into a full violet 

 at the greatest speed we could obtain. On causing the motor to 

 slow down, the same changes in an inverse order from violet to 

 red were observed. These changes in colour unth the santc 

 direction of rotation are very remarkable, and seem to us to be 

 in direct opposition to Mr. Benham's explanation supplied with 

 the top ; for if the colours are due to a certain percentage of the 

 etherial vibrations being cut off, this percentage will remain the 

 same for all speeds, and there is no reason apparent why there 

 should be a change in colour with difference in speed. Neither 

 are they easily explainable on Prof. Liveing's theory that red is 

 the first colour to appear, and blue the last to disappear ; also, 

 the green we got at the intermediate rate was certainly not a 

 neutral grey or green, but a pure vivid green. Probably other 

 observers have not used a greater speed than that of the first 

 transition stage from red to green, which has a kind of neutral 

 green tint. 



An experiment was next made with a disc constructed 

 similarly to that of Mr. Benham, but having white lines drawn 

 on the black semicircle instead of black lines on the while part. 

 On rotating this disc so that white lines on black ground suc- 

 ceeded the black surface, with slow rotation the colour obtained 

 was red, but a different kind of red to the deep blood-red given 

 by Mr. Benham's top in the case of i/ac/' lines on a s«/nVt' ground 

 also following a black surface, viz. a very bright red, evidently 

 not saturated. With higher speeds there followed a light green 

 and light blue, both evidently containing white. The diflerence 

 in hue of the two series of colours seems obviou-Iy that in the 

 case of the white-lined disc the colours are mixed with white, 

 and in that of the black-lined are mixed with hlack. 



Throughout the series of experiments we have tried to eliminate 



NO. I317. VOL. 51] 



psychical errors as much as possible, by experimenting on 

 persons unacquainted with the expected results. 



Belfast, January 15. J. M. Finnegan. 



B. Moore. 



The Kinetic Theory of Gases. 



The difficulty of reconciling line spectra with the kinetic 

 theory of gases, has been referred to by Prof. Fitzgerald 

 (N.^TUKE, January 3, p. 221). The following considerations 

 show that it is possible under certain suppositions to have a 

 number of spectral rays with a very restricted number of degrees 

 of freedom. Most of us, I believe, now accept a definite 

 atomic charge of electricity, and if each charge is imagined to 

 be capable of moving along the surface of an atom, it would 

 represent two degrees of freedom. If a molecule is capable of 

 sending out a homogeneous vibration, it means that there must 

 be a definite position of equilibrium of the "electron." Jf 

 there are several such positions, the vibrations may take place 

 in several periods. Any one molecule may perform for a cer- 

 tain time a simple periodic oscillation about one position of 

 equilibrium, and owing to some impact the electron may be 

 knocked over into a new position. The vibrations under these 

 circumstances would not be quite homogeneous, but if the 

 electron oscillates about any one position sufficiently long to 

 perform a few thousand oscillations, we should hardly notice 

 the want of homogeneity. Each electron at a given time 

 would only send out vibrations which in our instruments would 

 appear as homogeneous. Each molecule could thus suc- 

 cessively give rise to a number of spectral rays, and at any one 

 time the electron in the different molecules would, by the laws 

 of probability, be distributed over all passible positions of 

 equilibrium, so that we should always see all the vibrations 

 which any one molecule of the gas is capable of sending out. 

 The probability of ao electron oscillating about one of its 

 positions of equilibrium need not be the same in all cases. 

 Hence a line may be weak not because the vibration has a 

 smaller amplitude, but because fewer molecules give rise to it. 

 The fact that the vibrations of a gas are not quite homogeneous, 

 is borne out by experiment. If impacts become more frequent 

 by increased pressure, we should expect from the above views 

 that the time during which an electron performs a certain 

 oscillation is shortened ; hence the line should widen, which is 

 the case. I have spoken, for the sake of simplicity, as if an 

 electron vibrating about one position of equilibrium could only 

 do so in one period. If the forces calleil into play, by a dis- 

 placement, depend on the direction of the displacement, there 

 would be two possible frequencies. If the surface is nearly 

 symmetrical, we ^hould have double lines. 



The only weight I attach to these speculations lies in 

 illustration it affords that a number of spectral lines does 

 necessarily mean an equal number of degrees of freedom, 

 the existence of the "electron" I firmly believe; and 

 necessarily implies a very restricted number of variables 



Arthur Schuster 



the 



not 



In 



this 



"Acquired Characters." 

 It would appear that Prof. Lankester has not thought it worth 

 while to read all the letters that have appeared in Nature on 

 the question raised by Sir Edward Fry, unless it is to be inferred 

 from his remarks that he confines himself to the consideration 

 of ihe arguments of those who have a place on the scientific 

 Olympus of the Royal Society. In my letter, published 

 December 6, I defended Lamarck's laws against the accusation 

 that they were reciprocally destructive. Prof. Lankester 

 reiterates his accusation without any further support. But this 

 is not the whole question. In his last letter he suggests that 

 acquired characters corresponding to Mr. Gallon's definition 

 should be taken, and an investigation made as to whether they 

 are inherited or not in later generations. But in his former 

 lelter (November 29) he suggested very distinctly and deliberately 

 that such an investigation was unnecessary, because the question 

 was already settled. He has already condemned the heretic, 

 and now consents to his trial. His words were — " Since the old 

 character had not become fixed and congenital alter many 

 thousands of successive generations of individuals had developed 

 it in response to environment, but gave place to a new character 

 when new conditions' operated on an individual, why should 

 we suppose that the new character is likely to become fixed after 

 a much shorter time of responsive existence?" To apply this 



