52 THE ROYAL SOCIETY OF CANADA 
other hand, in 813, 45° it is lower than both, and in 813, 60° and 75° all 
three are practically the same. When we compare these differences 
with the probable error of measurement of the plates, less than one 
quarter of the numbers given at the bottom of the tables varying from 
.002 to .015, we are forced to the conclusion that they are systematic 
and personal in nature, but are at a loss to account for their cause. 
It is unfortunate that Mr. Adams was unable to measure more 
than one spectrum, but the close agreement of his result with Plaskett’s 
and the generally higher values of Miss Lasby and lower of De Lury 
would naturally, from the law of averages, lead to the acceptance of 
Adams’ and of Plaskett’s measures as probably being nearest to the true 
values. If such a conclusion be accepted then it would be necessary 
to apply a positive correction to De Lury’s measures in the 45600 region, 
which, when all the comparisons are taken into account, should be 
about 0.040 km. at the equator and possibly slightly less at the higher 
latitudes. A further evidence that this is probably the proper course 
is given by the practical agreement of Plaskett’s and De Lury’s measures 
in Series III at 4 4250. As the velocities of rotation obtained by 
Plaskett from the measures of Series I, II, and III are all practically 
the same, while those obtained by De Lury are about 3 percent. lower for 
Series I and II, but the same for Series III, the inference is that, in the 
poorer quality lines in the yellow green, some personal effect causes the 
difference and that this disappears when the lines become better defined, 
as is the case in the violet. On the other hand, if there be no systematic 
differences in the measuring of the line displacements by De Lury at 
the two regions À 4250 and À 5600, this would imply a difference in 
the rates of rotation as determined from lines of different wave length, 
a thing which though in itself not impossible is perhaps not very prob- 
able. 
ABSOLUTE VALUE OF VELOCITY. 
VARIATION OF VELOCITY WITH LATITUDE. 
18. The above discussion and comparison of measures have shown 
that it is hardly possible to state exactly the absolute velocity of the 
rotation of the sun and furthermore if, as seems likely, earlier deter- 
minations were affected in the same way, they are also uncertain to 
to the same extent, that of the “personal equation” of measurement. 
19. In order to place the preceding summaries of measures in a 
more convenient form for discussion and comparison, the following 
tables containing the observed mean linear velocities at the mean 
latitudes have been compiled. From these linear velocities, the ob- 
served angular velocities have been directly computed, while the other 
columns will be explained below:— 
