August 4, 1923] 



NATURE 



16 



a shifting of the earth's axis, a change in the shape 

 of the earth's orbit, or an increase of solar radiation. 



It is well known that perpetual ground frost to 

 within 10 or 15 inches from the surface does not inter- 

 fere with the prosperity of a black-spruce forest. At 

 Fort Macpherson, N.W.T., Canada, for example, we 

 have trees a hundred feet high growing straight and 

 close together, and yet I have observed in midsummer 

 that the perpetual frost around their roots was less 

 than a foot below the surface. 



As stated above, I do not offer this explanation 

 of certain of the coal measures in connexion with 

 any allegation that the Arctic was once an extensive 

 low land, but merely as an hypothesis which can be 

 called upon in case other evidence shows that extensive 

 low land may once have existed there. 



Coal has been found in the Antarctic no less than 

 in the Arctic. The Antarctic is at present in large 

 part an extremely high continent, but it is at least 

 worth considering whether it may not have been a 

 low land at the time when the coal was formed there. 



It has been abundantly shown that permanent 

 snow on land in the polar regions depends on altitude 

 and precipitation rather than latitude. Nansen has 

 said that on the low land of northern Siberia no 

 permanent snow has been found, and that he feels 

 certain none can be found. Many travellers, includ- 

 ing myself, have reported from northern Canada, 

 northern Alaska, and from the islands to the north 

 of Canada, the total absence of bodies of permanent 

 snow large enough to be called glaciers, though there 

 are small snow-drifts at the end of summer in the 

 shadowed bottoms of deep ravines in some of the 

 Canadian islands. Greenland is 90 per cent, covered 

 with ice, but the largest ice-free area in Greenland 

 is near its northern end, showing that altitude and 

 precipitation rather than latitude are the controlling 

 factors. The smaller glaciers of Franz Josef, Spits- 

 bergen, Ellesmere, Heiberg, North Devon, and the 

 one or two small glaciers of Baffin Island, depend 

 similarly on altitude and precipitation. A mere 

 change of altitude without change of area might 

 therefore remove the whole ice-cap of Antarctica — 

 or certainly it could be removed by a reduction to 

 a general level below 2000 ft. and perhaps a slight 

 increase in area. With the ice once gone, only the 

 Antarctic shores would be kept cool in summer by 

 the sea, the interior promptly adopting the extremely 

 hot June and July weather now found in the Arctic 

 lowlands, thus bringing conditions suitable for spruce 

 forests and the development from them of beds of 



coal. ViLHJALMUR StEFANSSON. 



The Trichromatic Theory of Colour Vision. 



The history of the spread of knowledge regarding 

 the Young-Helmholtz theory of colour vision is a 

 very curious one. As in the case of all other great 

 theories, its range of possible application far exceeds 

 the demands made upon it for the explanation of 

 actual facts. Limitations have to be imposed upon 

 it here and there in answer to inquiry as to which 

 choice out of several has been the one adopted by 

 Nature. This process is in accordance with the 

 development of all great theories. In the earlier 

 stages powerful restrictions are adopted in order that 

 advancement may be made. When these are found 

 to be too restrictive a wider postulate is made so as 

 to include a wider group of facts within the scope 

 of the theory ; and, the whole development being 

 simple and direct, the theory at last stands forth as 

 no longer a theory but a fact greater and wider than 

 any of the groups of facts which are contained within 

 its bounds. Thus the electron theory is now, apart 



NO. 2805, VOL. I 12] 



from certain tentative developments, a fact standing 

 upon as wide a basis of experience as any so-called 

 fact of which we are cognisant. 



This statement also holds in the case of the kinetic 

 theory in general. But, if the great developments 

 by Clausius, Maxwell, and others more recently, were 

 unknown ; if nothing were known beyond the 

 results of the early restrictive postulate of perfectly 

 hard, spherical, smooth, and elastic atoms ; com- 

 mentators of to-day might readily be found con- 

 demning the theory, and asserting that it could not 

 explain the facts which the recent workers have 

 shown to be direct and simple consequences of its 

 naturally developed postulates. This, or rather 

 worse, is exactly the position with regard to many 

 present-day criticisms of the trichromatic theory of 

 colour vision. These are evidently made in entire 

 obliviousness of developments actually made by 

 Helmholtz himself. 



A still more curious condition which subsists is 

 that the commentators are not entirely worthy of 

 blame. For the later developments have never 

 become common scientific property in Britain, while 

 the early developments became widely known. 



As examples of the criticisms I give some statements 

 taken from Dr. Edridge-Green's book on colour vision. 

 In doing so I desire to make it clear that I am making 

 no attack upon his valuable and interesting work ; 

 I am only replying to his strictures upon the Young- 

 Helmholtz theory, in which he, in my view, in- 

 advertently draws quite undeserved and wrong con- 

 clusions. In chap. XXX. he gives six arguments 

 which he holds to be conclusive against the Young- 

 Helmholtz theory of colour vision, and eight against 

 the Young-Helmholtz theory of colour blindness. I 

 assert, on the contrary, that the theory gives a simple 

 and direct account of the phenomena in each case ; 

 and I give the mode of deduction in five cases. 



" The theory does not explain why there should 

 be a defect in hue perception in those who have lost 

 one of their sensations." Now, actually, the theory 

 explains it beautifully. Thus in any stretch of wave- 

 lengths in which two of the three sensation curves 

 have opposite slopes, hue discrimination is corre- 

 spondingly strong. Therefore annulment of one of 

 these curves diminishes it. 



" The theory does not explain why many dichromics 

 have a luminosity curve similar to the normal." 

 This is an example of overlooking the later develop- 

 ments of the theory. If the dichromasy arises from 

 fusion of two of the sensation curves, the distribution 

 of luminosity may be unaltered. 



" There are not two or three definite varieties of 

 colour blindness, as there should be according to the 

 theory." Here again there is oversight. In the 

 hard-smooth-elastic-spherical-atom stage of the theory 

 this might have been asserted. Actually, according 

 to the theory as left by Helmholtz, there may be a 

 doubly infinite variety of cases of colour blindness. 



" How could the loss of half of a hypothetical 

 green sensation cause dichromatism ? " The answer 

 is simple. Given one sensation curve intersecting the 

 other two, if lessening of its ordinates by one half 

 makes it fall entirely within the others, dichromasy 

 is present. 



On p. 210, and also in the Phil. Mag., Nov. 1922, 

 Dr. Edridge-Green describes another case. " A man 

 with shortening of the red end of the spectrum and 

 normal colour discrimination will put together as 

 exactly alike a pink and a blue or violet much 

 darker. If, however, the pink and blue be viewed 

 by a normal sighted person through a blue-green 

 glass which cuts off the red encf of the spectrum, 

 both will appear identical in hue and colour. This 



