44 THE ROYAL SOCIETY OF CANADA 
niade directly following one another on the same plate, and under 
apparently identical conditions, such errors creep in, as for example 
in Table V in the equator plates of series III. In plates 860, 865, 867, 
869, the difference in the displacements of successive exposures are 
0.066, .074, .051, .051 km. per second, a greater difference than the one 
in question. Consequently although the method will be tried later it 
was not deemed desirable to delay further the publication of the ob- 
tained values, but to determine if possible, the probable corrections to 
be applied to the above given velocities. 
15. For this purpose all the equator spectra of Series I and 7 of 
Series III were measured by De Lury and all of Series II by Plaskett 
to determine systematic differences at the equator. In addition, to see 
how this difference varied with the latitude, 5 complete plates (7 
latitudes on each) of Series I were measured by De Lury and 5 com- 
plete pates of Series II by Plaskett. Two representative plates of Series 
I, Nos. 813 and 820 were sent to Mt. Wilson and were kindly measured 
by Mr. Adams and Miss Lasby in order to compare Ottawa and Mt. 
Wilson measures. All these comparisons are tabulated below and 
serve to show not only the difference in velocity obtained by different 
measures from the same plates which appear to be generally systematic 
in character, but indicate also the accidental errors of measurement 
to be looked for. The detailed measures for plates 813 and 820 show 
the great differences in accuracy of setting, for the probable error of 
setting on a single line (given below the means) varies on the average 
from 0.008 by Miss Lasby to 0.019 by Adams and Plaskett to 0.052 km. 
per sec. by De Lury equivalent, in linear values, to 0.0004, .001, and 
.003 mm. | | 
