38 



I'KKSIJJKXTIAL A])])RK.SS SKCTIOX A. 



only by reference to the meridian catalogue positions of 

 whatever stars there Arere on the plates Avliose position had 

 been determined by meridian telescopes at different epochs. 

 The detailed investig-ation of the jDlates is carried on at the 

 Union Observatory with the helj) of a stereo-comjDarator and 

 blink-microscope. This is a wonderful instrument, Avhich 

 enables corresponding i:»ortions of the two jslates to be com- 

 pared in rapid succession, and shows at a glance relative 

 disj^lacements amongst the stars of a small region. The same 

 result would be obtained, but with an enormous expense of 

 time and labour, by measuring the co-ordinates of each star 

 image on the two plates and comi^aring the measures for the 

 corresponding stars. Discrejjancies here would also show 

 relative displacements amongst the stars on the plate. But 

 whereas in this method every star must be measured to find 

 the few showing motion, the blink-microscope selects just 

 those few and practically ignores the rest, with the resiilt 

 that not only is the labour very much less, but also the 

 liability of error. In actual practice, something like one in 

 a hundred stars will show a small displacement with reference 

 to its immediate surroundings. It remains to inteiiDret 

 actually what this displacement means. Theoretically, if the 

 majority of the stars on the plate are at an infinite distance, 

 so that it can be assumed that their relative configuration 

 remains unaltered as seen from the extremities of a base line 

 as long as that over which the Sun would move in twenty-five 

 j-ears, i.e., 100 units (assuming twenty-five years to be the 

 interval of time between the taking of the two photographs), 

 then the displacement seen for any one star would indicate 

 its shift against the infinitely distant background. This shift 

 would be due to either or both of two causes — (1) the change 

 of viewpoint, i.e., the motion of the Sun, or (2) actual motion 

 in a fairly near star. Thus from either point of view the 

 blink-microscope would pick out the nearer members of the 

 stellar universe. 



A larg-e number of plates taken with astrographic 

 telescopes for the purposes of the Carte-du-Ciel has been 

 examined at the Union Observatory with the blink-microscope. 

 Mr. Innes has examined pairs of plates of the Greenwich, 

 Melbourne and Sydney Zones, and, with myself, is examining 

 the whole of the' Cape Astrographic Zone. I have re-photo- 

 graphed practically the whole of the Cape Zone to secure new 

 plates to compare with those taken originally. The time 

 interval between the members of each pair of plates is about 

 one-quarter of a century. The work of examining the plates 

 is not yet completed, and there is the further great task of 

 analysing the results to be done. Already it is evident from 

 a preliminary survey of the portion of the work done that the 

 moving stars so picked out by the blink-microscope divide 

 themselves into two main classes. These are Kapteijn's 

 Drifts I and II, and the directions of motions of these 

 groups coincide fairly nearly with the directions predicted 

 from the generallv accepted positions of the apices of the 

 drifts. Further, the division into two classes is so sharp and 



