PHonxntAPHY IN ASTRONOMICAL HKSKAKCH. 175 



AvhethiM" a dcHiiito conclusion would c\-cr luivc been reached at all. 

 15y the sheer facility of the new method of work we have advanced 

 by leaps and bounds wjiere we could only ci-aw 1 before. 



Thirdly, there has been a oreat j>ain in accuracy from the intro- 

 duction of photoii-raphy ; and it is this <|uality which is above all of 

 value in the science of astronomy." 'I'he wondeid'id exactness of the 

 photojTfraphic i-ecord may perhaps best be characterized by sayinir 

 that it has revealed the deficiencies of all our other astronomical 

 apparatus — object glasses and prisms, clocks, even the observer him- 

 self. 



It has almost been forg'otten that in the early days the accuracy 

 of a photograph was doubted. Even now it can scarcely be said 

 that we know definitely the stage of refinement at which we must 

 begin to expect irregular displacements of the images from distor- 

 tion of the photographic film; but Ave have learned that they do 

 not occur in a gross degree, and that other ai)paratus nnist be im- 

 proved before we need turn our attention seriously to errors aris- 

 ing from such a cause. Consider, for instance, what photography 

 has told us about our optical apparatus, which we regard as having 

 reached a high state of perfection. AVe are accustomed to think of 

 properly made optical apparatus as being sufficiently similar in all its 

 parts; it is tacitly assumed in the principle of the heliometer. for 

 example, that one half of the object glass is sufficiently similar to 

 the other. But a stock adjustment recently adopted in photograph- 

 ing a spectrum for accurate measurement exhibits clearly the errors 

 of this assumption. Photographs are taken of the spectrum through 

 the two halves of the objective; and if they were properly similar 

 the lines in the two halves of the spectrum should fit exactly. A 

 mere glance is usually sufficient to show^ discordances. It is true that 

 one of the photographs is taken through the thick half of the prism 

 and the other through the thin, so that errors of the prism are in- 

 cluded; but these, again, are optical errors. They are, however, 

 not the only sources of .error which at present mask photographic 

 imperfections. Glass plates are not fiat, and this want of flatness 

 introduces sensible errors. Even with the great improvements in 

 our driving clocks which were called for immediately photographs 

 were to be taken — with electrical control and careful watching on 

 the part of the observer— there is apt to creep in a " driving. error " 

 which gives bright stars a s])uri()us displacement relatively too faint. 

 We must get flatter plates, better driving clocks, and watch more 



a Two tliiiiijs limy be iiionsurea on a iihoto^a-apliic plate — the position of an 

 object, or tlie density of the iniai^'e ; tlie foriiiei- lieinj? an indication of itH posi- 

 tion in tlie lieavens, and tlie latter of its bri^litness. With the latter topic I 

 ■ do not propose to deal, for the reason that it is in the hands of a nuicli abler 

 and more experienced exixmeiit ; bnt tlie former alone will provide enonsii food 

 for I'eflectiou. 



