212 Permanence of the Ratherfurd Photographic Plates. 



In order to understand the above, it is necessary to bear in 

 mind that it was Mr. E-utherfurd's invariable custom to make 

 two impressions on his plates. The}' differ a little in right ascen- 

 sion, but not at all in declination. It is, therefore, clear that by 

 reading the western impression with the micrometer changed by 

 half a revolution, as explained in the above instructions, the mean 

 result from the two impressions will be almost entirely free from 

 periodic errors of the micrometer screw. Moreover, it will be 

 noticed that the principle of repeating all observations in the re- 

 verse order is rigidly adhered to in the above instructions. 



While this careful use of the above principle may reasonably 

 be expected to remove the effects of any changes of temperature 

 that may take place during a single set of observations, it is 

 otherwise with changes that may take place at longer intervals of 

 time. Measures made on one day are not immediately compara- 

 ble with those made on another day. For the scale may have a 

 different temperature on the second day, so that we shall really 

 be measuring with a different standard of length on the two days. 

 It has therefore been necessary to standardize the scale every 

 day. This was done in the following way. 



A pair of stars were selected differing very greatly in both 

 coordinates, and therefore situated near opposite corners of the 

 plate. These were carefully compared with the scale each day, 

 and all the measures of that day were multiplied by a constant 

 factor of such magnitude as would make the distance between the 

 standards an absolute constant. Moreover, a special series of 

 inter-comparisons of the distance between the standards in the 

 a;-coordinate and the y-coordinate was made, so as to be sure that 

 the unit of measurement should be the same for both coordinates, 

 we shall return to this matter later, as it is desirable first to give 

 the further instructions furnished to the observers. These were 

 as follows : 



" To standardize the measures it is necessary to select a pair of 

 standard stars having a very wide difference of both coordinates. 

 These stars we will call A and B. The successive operations will 

 now be as follows : 



" (a) Compare the standard stars with the scale in the order 

 A, B, B, A. 



" (b) Compare the central star twice. 



" (c) Compare a number of other stars with the scale, and then 



