368 



NATURE 



[February i6, 189; 



kinds of light not liable to be absorbed in notable degree 

 by the yellow pigment. On the whole, it is accordingly 

 found that this pigmentation is subjectively influential 

 only in cases in which the rays in the neighbourhood of 

 the line F play a prominent part, as, for example, in a 

 certain mixture of this blue with red (mentioned on page 

 354) which, if it looks white when our eyes are directly 

 fixed upon it, will show blue predominant when we look 

 in a slightly different direction. 



As far as hitherto-known facts go, it appears very im- 

 probable that Hering's theory of dichromasy can be 

 carried through. Nevertheless, further observations in 

 this direction are very desirable. The influence which 

 the colouration of the yellow spot has in individual eyes 

 can be estimated by comparing the appearances of colour- 

 mixtures in the centre of the field of view with their ap- 

 pearances very near the centre. Such comparisons will 

 show with certainty where such influence is present and 

 where it is absent. 



The following is a summary, by Prof. Everett, of 

 two passages from the new edition of Helmholtz's 

 " Physiological Optics," which are important as supple- 

 menting the foregoing critique of Hering's theory : — 



In discussing the results of experiments for determining 

 the exact positions of the three elementary sensations 

 with respect to actual colours, in Newton's diagram or in 

 Lambert's pyramid, Helmholtz represents the results by 

 a triangle with the three elementary sensations at its 

 corners, and with the colours of the spectrum plotted 

 along a curve which lies entirely in the central portion of 

 the triangle. He says, p. 457 : — 



" This curve shows that every simple colour excites 

 simultaneously in the trichromic eye the three nerve- 

 elements which are sensitive to light, and excites them 

 with only moderate differences of intensity. If we then 

 hypothetically refer all these excitations to the presence 

 of three photo-chemically alterable substances in the 

 retina, we must conclude that all three of these must have 

 nearly the same limits of sensibility to light, and must 

 show, in the rates of their photo-chemical actions for the 

 different wave-lengths, only secondary variatioi\s of 

 moderate amount. Similar variations, arising from the 

 presence of foreign substances, from substitutions of 

 analogous atom-groups, and so on, occur also in other 

 photo-chemically alterable substances as used in photo- 

 graphy ; for example, in the different haloid salts of 

 silver." 



In a mathematical discussion of colour-blindness, 

 commencing at p. 458, he points out that in dichromic 

 vision there must be a linear relation between the three 

 independent elements of trichromic vision, and in Lam- 

 bert's colour pyramid there must be a certain line through 

 the vertex, such that any plane drawn through it is a plane 

 of uniform colour. Newton's diagram of colour may be re- 

 garded as contained in any plane which cuts the axis of the 

 pyramid ; and it is very important to determine the point 

 in which the above-mentioned line cuts such a plane ; for 

 any line in Newton's diagram that passes through this 

 point is a line of uniform colour to the dichromic vision 

 in question. Experiment shows that it always lies out- 

 side the triangle of actual presentable colours. 



Addendum. 



Prof. Everett adds the following remarks of his own 

 on the present position of the problem of colour-vision : — 



On the one hand, it is established, as a fact of experi- 

 ment, that the excitation of colour-sensation in the normal 

 eye depends upon only three variables, and that their 

 effects are superposable, so astoadmit of being expressed 

 by equations of the first degree, otherwise called linear 

 equations. The simplest choice of three variables is that 

 adopted in Young's theory, because it only requires posi- 

 tive values of the variables. 



NO. I 2 16. VOL. 47] 



On the other hand, the various colours regarded as 

 subjective appearances do not naturally class themselves 

 under a threefold heading. Yellow does not look as if 

 it consisted of red and green. Colour-sensations as 

 known to us in consciousness are not threefold but 

 manifold. 



The two facts taken together seem to imply two succes- 

 sive operations intervening between the incidence of 

 light and the perception of colour. The first operation is 

 threefold, and may consist (as above suggested by Helm- 

 holtz) of the photo-chemical decomposition of three 

 different substances. The second operation consists in 

 the effects of the first operation upon a complex organism, 

 and the distinctions of colours as we see them arise out 

 of the nature of this organism. 



The number of independent variables required for 

 specifying the condition of a system is a very different 

 thing from the number of well-distinguished states in 

 which the system can exist. For example, the state of 

 a given mass of water-substance is completely determined 

 if its volume and temperature are given, and therefore 

 depends on only two variables. But the number of its 

 well-distinguished states is three. In like manner colour 

 depends on three variables, but the number of well-dis- 

 tinguished colours, besides white, may be said to be seven, 

 namely the six principal colours of the spectrum and 

 purple. 



What differences of condition in the organism cor- 

 respond to these eight distinct appearances in the field of 

 view, and how these different conditions are produced by 

 the three primary excitations, are problems awaiting 

 solution. 



AUTOMATIC MERCURIAL AIR-PUMPS. 



OF late years, and more especially during the last 

 decade, men of science have devoted much thought 

 and ceaseless energy to the invention of an apparatus 

 which should admit of the automatic working of mer- 

 curial air-pumps. Of the numerous inventions brought 

 forward, the ingenious apparatus of SchuUer and Steam 

 are especially deserving of mention. 



But notwithstanding the present extensive employment 

 of the mercurial air-pump in science as well as in tech- 

 nics these appliances are neither much known, nor have 

 they been used to any great extent, although they are of 

 great importance, and would probably be very advantage- 

 ous. This may be explained by the fact that they are 

 wanting in the necessary simplicity and trustworthiness, 

 without which the advantages of automatically working 

 mercurial air-pumps are somewhat doubtful. 



We shall describe now an apparatus for the per- 

 fectly trustworthy and automatic working of mercurial 

 air-pumps, as well as the shape of the glass pump used 

 in connection with it, which, while possessing the greatest 

 possible simplicity, admits of the highest rarefactions 

 hitherto known. 



The figure shows the automatic apparatus in connection 

 with an improved Toepler mercurial air-pump. The glass 

 ball H is connected on the one hand by flexible tubes 

 with the pump Q, on the other hand by the tube L with 

 the accumulator M. The water-pipe K runs into the 

 bottom of the accumulator, and by means of a specially- 

 constructed three-way cock K can either be connected 

 with the hydrostatic pressure-pipe Ki or the discharge- 

 pipe Ko. 



If water under pressure is admitted through the tubes 

 ^j, K and k into M, the air contained in M is compressed. 

 This air again exerts a pressure through the tube L on 

 the mercury contained in H, and drives it into the pump 

 Q. As soon as the mercury has risen sufficiently high and 

 the cock K is reversed, the compressed air forces the 

 water out again through k, K, and k.^, and the mercury 



