Changes in the Sun’s apparent Diameter. 269 
That is, if with Wagner an observer assigns a weight to each 
observed transit of the sun, this weight expressing the goodness 
of the image as to steadiness and definition, it will be found 
that each class of observations so defined will give a diameter 
peculiar to itself, and differing in a constant way from the 
diameters deduced from the other classes. Dr. Gyldén has 
ound the same thing to be true of observations of the sun’s 
vertical diameter made with the Pulkowa vertical circle, and 
Dr. Becker of Neuchatel corroborates Wagner's results for hori- 
ner’s statement, and to show that the observations of both 
diameters made at Washington in the years 1866 to 1870 en- 
tirely confirm it. : 
e great importance of the conclusions drawn by Secchi 
from his observations has induced us to test them by a method 
different from those of Auwers and Wagner. The difficulty 
which besets this entire subject is to distinguish between actual 
variations of the sun’s diameter and errors of observations. 
at two different observatories, so that each observation of the 
ies is accompanied by a simultaneous one of the other 
° 
i=] 
than the average, the probability of finding a positive correc: 
tion will be more than } at each observatory, and hence the 
probability of an agreement of sign will be greater than 4. - 
the probabilit in each case be $+4, it is easy to see that the 
probability ef an agreement will be $+2a?. 
