154 



NA TURE 



[December 15, 1904 



supply, and are evidently inspired by a determination 

 to give something like concrete value for public money. 



While awaiting the verdicts of science and the 

 deliberations of legislators, it is useful to have to hand 

 a work such as this, which gives a concise statement 

 and accurate picture of the present condition of the 

 great sea-fishing industry. 



The book is abundantly supplied with interesting 

 photographs. There is also a sea-fisheries map, in 

 which, however, is one glaring defect. From this 

 map it would appear that Yarmouth and Lowestoft 

 arc given over entirely to the drift-net fishing, and 

 that neither of these places has any connection by rail 

 with the metropolis. This is inconsistent with what 

 is stated in the text, and is opposed to common 

 knowledge. 



THE ELEVENTH EROS CIRCULARS 

 T"* HE appearance of this volume brings us definitely 

 ■*• face to face with a new situation in the derivation 

 of accurate positions of the heavenly bodies from photo- 

 graphs. It will be remembered that in the winter of 

 1900-1 the recently discovered small planet Eros made 

 a very near approach to the earth, and a large number 

 of photographs were taken with the view of determin- 

 ing the distance of the planet, from a knowledge of 

 which that of the sun, and the dimensions of the solar 

 system generally, could be inferred with (it was hoped) 

 considerably improved accuracy. The measurement of 

 the plates involves enormous labour, and has only been 

 partially accomplished in the intervening four years ; 

 and the discussion of the measures has necessarily pro- 

 ceeded even more slowly. But the present publication 

 of more than 400 quarto pages represents a notable 

 addition to the tabular statement of measures, and 

 contains an important contribution to the discussion. 



It appears that the plates taken at different observ- 

 atories are liable to disagreement in a serious manner. 

 Putting aside the planet itself for a moment, when the 

 positions of the stars found from plates taken at the 

 Algiers Obser\'atory are compared with those found 

 from plates taken at Paris, there is a difference vary- 

 ing with the brightness of the individual stars. Such 

 ■a difference is not altogether new in astronomy ; it was 

 pointed out by Sir David Gill a dozen years ago or 

 more that eye observations of stellar positions made 

 by different observers were likely to differ system- 

 atically in this manner ; but this was attributed to 

 human defects in the observer, and it was hoped that 

 photography would free us from the embarrassment. 

 So it probably will when rightly used ; but we have 

 apparently not yet completely realised the necessary 

 precautions. The instruments for taking the photo- 

 graphs at Algiers and at Paris are as precisely similar 

 as the constructor could make them ; they were used 

 in the same way ; the plates were measured similarly 

 and with careful attention to certain known sources 

 of error, and yet the resulting star places show the 

 following differences in seconds of arc in the mean of 

 5 groups of 87 stars each : — 



Mean Difference 

 magnitude ^^ 



8"8 -0-27 



94 — 0-42 



lo'4 -0-57 



il'2 -072 



II-6 -0-83 



There is a range of more than half a second, and 

 we want to measure the hundredth of a second ! This 

 is probably an exceptional case ; but what may occur 

 once may occur again, and in view of this fa'ct it is 



1 Conference astrophotographique Internationale de Juillet, igoo Cir- 

 ■culaire No. II. (Paris : Gauthier Villars, 1904 ) 



NO. 1833, VOL. 71] 



not too much to say that a very serious addition has 

 been made to the labour of determining the quantity 

 sought — the solar parallax — by this revelation. 



It is disappointing to find no satisfactory sugges- 

 tion of the cause of error in the paper which gives an 

 account of it. A suggestion is indeed made, viz. that 

 in measuring a plate the presence of an adjacent image 

 (for the exposure is repeated on the same plate so as 

 to show all the images more than once) may disturb 

 the eye of the measurer. All our experience hitherto 

 is against such a possibility. It seems more likely to 

 the writer that the cause may be sought in the object 

 glass of the photographic telescope, and, to be more 

 precise, in an error of centreing of the crown lens re- 

 latively to the flint. Such an error is well known to 

 opticians, and is easily detected in a visual telescope 

 by the fringe of colour on one side of a star image 

 when slightly out of focus. But the images formed 

 by a photographic telescope are not examined by the 

 eye in the regular course of work, and such an error 

 might therefore escape detection until revealed by such 

 a comparison of measures as is given above. The 

 stray light on one side of the image would not be 

 strong enough to affect the sensitive film in the case 

 of faint stars, but for a bright star it would spread the 

 image in that direction, and so introduce a spurious 

 displacement of the centre. If this explanation be 

 correct, the error can be both detected and eliminated 

 by turning the object glass through iSo° (with most 

 forms of telescope mounting it is only necessary to 

 turn the telescope to the other side of the pier), and this 

 can easily be done. Indeed, it ought to have been done 

 before now, under the admirable maxim for physical 

 work, " reverse everything that can be reversed," but, 

 so far as is known to the writer, the point has hitherto 

 escaped notice. 



If on examination this explanation will not fit the 

 facts, some other must be found. A few additional 

 details in the volume before us would have made it 

 possible to test this hypothesis; if, for instance, it had 

 been specified which plates were taken on one side of 

 the pier and which on the other, a comparison of the 

 two sets would have given very definite information. 

 Mr. Hinks has already given cogent reasons (see 

 Observatory for September, 1903) for regretting the 

 lack of information as to the identity of the individual 

 plates, and we have now to add this further reason. 

 For the systematic difference described is not confined 

 to .A-lgiers-Paris. If we turn to the paper following 

 that in which M. Trepied gives the figures above 

 quoted and arrange the differences found at the Good- 

 sell Observatory (Carleton College, Minnesota) accord- 

 ing to stellar magnitude, we find a well marked effect 

 in R..^. and a smaller one in dec. ; and probably other 

 cases, when duly examined, will give similar results, 

 though it does not seem to have occurred to astro- 

 nomers generally to make a properly searching inquiry. 

 For instance, at the end of the volume ^I. Loewy 

 tabulates a series of differences between two lists of 

 star places prepared with great care by himself and 

 by Prof. Tucker, of the Lick Observatory, and he com- 

 ments with satisfaction on the close accordance of the 

 two lists. But a very slight examination suffices to 

 show that the difl'erences are affected with " magni- 

 tude-equation," though in this instance the effect may 

 be due to the visual observations. 



In fact, while duly admiring the energy and dili- 

 gence with which this vast mass of material has been 

 collected and published, a result due in great part to 

 the powers of organisation of M. Lcewy, the director 

 of the Paris Observator)-, we may well feel some doubts 

 whether it will turn out to be, as he hopes, a " collec- 

 tion of homogeneous material, susceptible of being 

 immediately used without the necessity of undertaking. 



II 



