October 20, 1910] 



NATURE 



495 



rufous (or chestnut), but one is fawn-coloured and has 

 dark " patches at the fetlocks." Doubtless dun may con- 

 lain bay, brown, and black, but on one occasion I obtained 

 a stripid dun by crossing a black Shetland pony with a 

 striped bay Arab — an Arab which mated with a yellow- 

 dun Connemara mare produced a pure black. These may 

 only be exceptions that prove the rule. 



J. C. EWART. 



Tests for Colour-Blindtiess. 



I AM surprised to see in the review of my book in X.xture 

 of September i the statement : — 



" \Ve do not see that Dr. Edridge-Green has furnished 

 us with any increased security, or indeed that any better 

 security is needed, than is obtained from Holmgren's test 

 when this is employed in the precise manner directed by 

 its originator." 



In this issue of my book I have devoted nearly nine 

 pages to the detailed condemnation of the Holmgren test, 

 and this portion remains as it was in the 1891 edition. 

 The statements there have been confirmed by numerous 

 observers, amongst whom are some of the ablest scientific 

 men the world contains. In fact, at the recent Inter- 

 national Physiological Congress I did not meet with a 

 single man who was satisfied with the Holmgren test. 



I will only refer to the statements of Prof. Nagel, who 

 has done so much in connection with colour-blindness. 



I pointed out that normal-sighted persons were rejected 

 by this test, and this is abundantly evident by the number 

 of men rejected by the Board of Trade who get through 

 cm appeal. 



Prof. Nagel in iSqS found thirty-nine cases (2-75 per 

 cent. !) in 1420 e.\aminations in which typical dichromic 

 (red-green blind) mistakes were made with the Holmgren 

 test, and yet when examined by other and more trust- 

 worthy methods, as, for instance, the spectroscope, were 

 found not to be dichromics. 



I stated that the test green was not the best colour for 

 a first test. Nagel says the same thing. In the reports 

 of the Board of Trade it will be seen that many have 

 passed the green test and failed with the rose test. It 

 may be noted that the Board of Trade have never at any 

 time used the test in strict accordance with Holmgren's 

 instructions, because they have used all three test skeins, 

 whereas Holmgren stated that when the green test had 

 been passed the person might be regarded as normal 

 sighted. Nagel points out the varieties and number of 

 colour-blind persons who are passed by the Holmgren test, 

 and gives the reasons, which are similar to mine. 



F. W. Edridge-Grf.en. 



The Institute of Physiology, University College, 

 London, October 7. 



In iSoo or 1891 the Royal Society appointed a very 

 strong committee, of which Lord Rayleigh was chairman, 

 and it included, among other " able scientific men," 

 Lord Kelvin, Sir George Stokes, Sir William Abnev, and 

 Prof. Michael Foster, to report on the general subject of 

 colour-vision and on the tests proper to be used in con- 

 nection with it. Dr. Edridge-Green gave evidence before 

 this committee, stated fully his objections to Holmgren's 

 test, and displayed the methods which he recommended 

 in lieu thereof. His book was published before he gave 

 evidence; and, as his original objections to the Holmgren 

 test are reprinted verbatim in the 1909 edition, it is fair 

 to suppose that no fresh evidence in support of them has 

 been obtained during the intervening time. Besides hear- 

 ing many witnesses, the committee carried out an extended 

 series of practical investigations, and on April 28, 1S02, 

 it unanimously recommended the Holmgren test for 

 .idoption bv railwav companies, ship-owners, and the 

 Board of Trade. The committee pointed out that varia- 

 tions in the amount of deficiency in colour-perception are 

 numerous, and, " when small, are often difficult to 

 classify." No one claims for the Holmgren test that it 

 affords a sufficient basis for a minute classification, but it 

 does afford the surest and most convenient means of 

 excluding from certain industries the small number of 

 persons who could not engage in them without danger to 

 the communitv. The Reviewer. 



Water Vapour on Mars. 



The statement attributed to Director Campbell on p. 317 

 of Nature for September 8, to the effect that the nights 

 in September, 1909, on which his spectrograms of Mars 

 were taken, *' were as perfect for the purpose as could 

 be wished," is open to question. Though the sky may 

 have been clear and the surface humidity low, this does 

 not prove that the aqueous vapour in the upper air was 

 small in amount. September is the month when the total 

 vapour-content of the atmosphere is a maximum, and 

 February is the month when the vapour-content is a 

 minimum, in north temperate latitudes. This is well 

 shown in the curves of energy in the infra-red solar spec- 

 trum for February 19, 1903, and September 14, 1903, in 

 the article on " The Absorption of Water Vapour in the 

 Infra-red Solar Spectrum," by F. E. Fowle, jun. (Smith- 

 sonian Miscellaneous Collections, quarterly issue, vol ii., 

 part i., p. I, 1904, Plate i.). The ratio of the intensities 

 of the bands of aqueous absorption ](.Mars-t-earth)/earthf 

 will be greatest when the total absorbent column of the 

 earth's atmosphere contains least water, that is, other 

 things being equal, the ratio may be expected to be 

 smallest in September and largest in February. Director 

 Campbell has chosen the worst month, and Dr. Slipher, 

 who observed in January and February, the best months 

 for making the experiment. 



The statement that " with a nearly evanescent a band, 

 the more water vapour one attributes to the terrestrial 

 atmosphere the less remains attributable to that of Mars 

 is, of course, true, and because the water vapour of Mars 

 is not great in amount it is not desirable to attempt to 

 observe it at a time when the feeble Martian absorption 

 band is swamped in a more powerful terrestrial band. 



The depths of the aqueous absorption bands in Mr. 

 Fowle 's figure (/or. cit.) is many times greater in Sep- 

 tember than in February ; but this does not express the 

 degree of unfavourableness of the September observation 

 adequately, for it is increasingly difficult to detect an 

 increment of absorption due to the addition of a constant 

 amount of vapour, as the total absorption grows greater, 

 and this for the reason that many of the absorption lines 

 have reached a maximum intensity already, so that any 

 further increase of the depth of the absorbent only affects 

 the feebler fines. It should be understood that, with the 

 low power employed, the band is not resolved into its 

 separate lines in the Martian spectrum. 



Through the favour of Dr. Percival Lowell I have been 

 permitted to measure the spectra of Mars and the moon 

 photographed by Dr. Slipher at the Lowell Observatory 

 in January and February, iqo8. The seasonal gain from 

 lower temperature and diminished moisture in the upper 

 air in winter at Flagstaff is more than an equivalent for 

 any gain in this respect to be obtained by even a double 

 altitude in summer. I have made quantitative measures 

 of the absolute intensity of the little a band in both 

 spectra. The ratio, a(Mars)/a(moon), obtained from six 

 different plates on as many nights, varies from night to 

 night as changes in the terrestrial atmospheric humidity 

 mav determine, but all of the plates unite in telling the 

 same story, and show that little a is stronger in the 

 spectrum of Mars. 



One of Dr. Slipher 's plates, which is, unfortunately, not 

 the best one photographically, although it is better in 

 this respect than any of Director Campbell's, was taken 

 under almost ideal conditions, the air at the surface having 

 a dew point of —14-8° C, and the exposure on Mars 

 being equally divided on either side of the lunar exposure 

 with both bodies at the same altitude (40°). The result 

 is conclusive, and shows that the Martian band on this 

 occasion was two and one-half times as intense as the 

 telluric one. Still larger ratios were obtained from other 

 plates. 



Similar consistent measures have also been made of the 

 oxygen Band, great B, showing that it is in like manner 

 stronger in the soectrum of Mars, although the measure- 

 ment is a difficult one, because the earth's atmosphere 

 is much denser than that of Mars, and the further small 

 addition of absorbent has but little effect. 



Frank W. Very. 



Westwood Astroohysical Observatory, West wood, 

 Massachusetts, October i. 



NO. 2138, VOL. 84] 



