November 5, 1891] 



NATURE 



co.iflictjng interpretations of the portion of the accounts of the 

 Koh-i-nur from Babar's time onward. There are still some 

 interesting questions of a difficult kind regarding its history 

 antecedent to the days of the Mogul Empire. But I believe I 

 have said now my last word regarding the later history, and 

 leave to my readers the decision as to the side in this little 

 controversy on which the truth is more likely to lie. 



^ N, Story Maskelyne. 



f Basset Down House, October 26. 



A Rare Phenomenon. 



Auroras were visible at Lyons, New York, on September 

 9, ID, and II. That on September 9 was very fine, flickering 

 streamers and arches forming at intervals from 8 o'clock to 10 

 o'clock- p.m. A peculiar feature of this aurora was an arch 

 similar to that described in Nature of September 17 (vol. xliv. 

 p. 475), as having been seen by Mr. Tuckwell at Loughrigg, 

 Ambleside, on September 11. The arch seen at Lyons on 

 September 9 was visible shortly after sunset, and remained in 

 the same position throughout the evening. It consisted of a 

 narrow band of light, which arose vertically from a point on the 

 horizon nearly due west, and passed through the constellations 

 of the Northern Crown and the Lyre, and just south of the 

 zenith down to the eastern horizon. When it was brightest, at 

 about 10 p.m., a few small streamers formed in connection with 

 it nearly in the zenith ; otherwise it consisted simply of a narrow 

 band of white light separated by a wide interval Irom the 

 auroral coruscations and streamers in the northern heavens. 

 This seems to have been very similar to the band seen by Mr. 

 Tuckwell. Other instances have been noted by the writer in 

 which some peculiarity of form or colour characteristic of an 

 outbreak of the aurora has attended its appearance in localities 

 remote from each other. M. A. Veeder. 



Lyons, N.Y., October 17. 



Two instances of the occurrence of the rare phenomenon 

 referred to in your issue of September 24 (vol. xliv. p. 494), 

 by Prof. R. Copeland and Mr. W. E. Wilson, will be found 

 recorded in the Transactions of the Nova Scotian Institute of 

 Natural Science, vol. vi. p. 100. The dates of these occur- 

 rences were July 31 and September 5, 1883. The general 

 appearance and position of the luminous arch were the same in 

 lx>th cases as in those described by Prof. Copeland and Mr. Wilson. 

 Two additional points were noted, however, which are worthy 

 of mention, viz. (i) that the arch of September 5 had a slightly 

 marked rayed structure, which, when first observed, was in the 

 direction of its length, but which gradually changed to a direc- 

 tion inclined about 45° to the longitudinal, and (2) that the 

 spectrum of this arch, as determined by one of Hilger's pocket 

 spectroscopes, consisted of two lines in the green, one quite 

 bright and the other faint. 



On Tuesday, September I of this year, I again observed the 

 same phenomenon at Halifax, N. S. I was unable to make 

 accurate observations, but noted the following facts : — The 

 luminous arch was quite bright when first observed, at 11.30 

 p.m., and extended from horizon to horizon. Fifteen minutes 

 later it had completely faded away. It was about 4" or 5° in 

 width throughout its whole length. It met the horizon at 

 points about 10° or 15^ to the north of the east and west points, 

 and passed through a point a few degrees south of the zenith. 

 When first observed, it was approximately uniformly bright 

 throughout, except at the edges, where its brightness diminished 

 rapidly outwards. To the eye its light seemed to be white, and 

 stars were visible through it. In fading away, the east and west 

 ends disappeared first, and the main body of the arch became 

 gradually fainter, wider, and more variable in width. The night 

 was bright and clear, and the temperature lower than it usually 

 is in the beginning of September, and there was no appearance 

 of aurora in other parts of the sky. 



Except on this occasion I have neither observed this pheno- 

 menon nor heard of its occurrence since 1883. But as it might 

 readily occur without my either seeing it or hearing of it, I cannot 

 say that I know it to be rare. J. G. MacGregor. 



Dalhousie College, Halifax, N.S., October 14. 



what indistinct, but at all events the luminous band extended 

 east and west almost through the zenith, and was preceded by 

 an auroral display. It occurred in August or September of 

 1883 or 18S4. 



My attention was again directed to a similar appearance on 

 the evening of September 9 of the present year, while near 

 Toronto. The narrow band of light, as before, extended from 

 the eastern almost to the western horizon, passing through the 

 zenith, and was accompanied by an aurora. 



It is worthy of note that Isaw the phenomenon at Toronto on 

 the evening of September 9, not September II. 



R. N. Hudspeth. 



Bishop's College, Lennoxville, P.Q. 



It has twice been my good fortune to observe phenomena 

 similar to that described in Nature of September 24 (vol. 

 xliv. p. 494). My recollections of the first occasion are some- 



NO. U49, VOL. 45] 



Apparent Size of Objects near the Horizon. 



Some years ago there appeared an account of an investigation 

 into the cause of the sun and moon looking larger when low 

 down than when high up in the sky. The theory advanced as 

 the result of the investigation attributed the effect to a physio- 

 logical cause. One could not expect an explanation of this kind 

 to be applicable to all individuals, but rather that with different 

 persons there would be different results ; so I have made obser- 

 vations — 81 in number — to find out what law applies to my own 

 case. These observations were made by taking notice of two 

 .stars near the horizon, and then looking up near the zenith to see 

 what stars in that situation appeared to be the same distance 

 apart as those near the horizon. I took a great variety of 

 different cases, the length of the compared arcs varying from 

 i°"4 to 100°. I observed them also in various angles of position, 

 from horizontal to vertical ; and sometimes had the two arcs at 

 the same angle of position upon the retina, and at other times at 

 different angles. 



The result of this investigation is an unexpected one, showing 

 that the length of the observed arc greatly affects the result of 

 the estimation — short arcs appearing longer when near the 

 horizon than when high up, and long ones appearing shorter. 



The comparisons were made in either of two ways ; according 

 to one method, after I had carefully taken note of the ap- 

 parent length of ihe arc near the horizon, and had fixed an 

 idea of it in my mind, I then took a single glance at the stars 

 near the zenith and fixed in a moment upon an arc that appeared 

 to be of the same length ; whereas in the other plan I made as 

 deliberate and careful an estimation of the arc near the zenith as 

 of that near the horizon with which it was compared, looking to 

 and fro from one to the other till \ was satisfied as to their 

 apparent equality. 



One would naturally expect that the instantaneous estimations 

 would be less accurate than the careful ones, and this is found to 

 be the case. Taking all the observations, I find the average 

 deviation from the truth of a single estimation is 77 per cent, in 

 the case of careful comparisons, and 10 3 per cent, in the case of 

 the instantaneous ones. The following formula is based upon 

 the careful comparisons — 



L = /{ I -h ^^^-(-085 - -00321/) I , 

 I 74 ■' 



where / and L are the lengths (in degrees) of apparently equal 

 arcs at a, the lower altitude, and at A', the higher altitude, re- 

 spectively. According to this formula, an arc 26°*48 long appears 

 the same length at whatevtr altitude it is situated, but an arc 

 shorter than 26° '48 appears longer at the horizon than at 

 the zenith, and an arc in excess of 26° 48 would actually appear 

 longer near the zenith than near the horizon : an arc i"'-4 long 

 (the shortest in my observations), when at the horizon, would 

 appear equal to an arc in the zenith 109 85 per cent, of its 

 length ; while an arc 100° (about the longest in my observations) 

 at the horizon would appear equal to an arc of 71° 30 only in the 

 zenith [i.e. with its middle point in the zenith). When the 

 above formula is applied to all the observations, the average 

 deviation (of the observed lengths from the computed) is reduced 

 to 4 "2 per cent, in the case of the careful comparisons, and 7*0 

 per cent, in the case of the instantaneous ones. If this formula 

 can rightly be applied to obj;;cts of such small dimensions as the 

 sun and moon, it (as will be seen) allows only a small increase 

 for their apparent size near the horizon upon that when they are 

 seen at a considerable altitude. 



It would be easy to find a more complex formula which would 

 satisfy the observations still better, but these are not sufficiently 

 numerous to warrant the doing so. 



