142 I. P. Koch. 
In case we suppose this difference to be exclusively due to the re- 
fraction, the error of same as computed becomes 
2.93 | 
о 
where г is the value indicated by Bessel. 
It might thus seem as if the computed value of the refraction 
had been a good deal too great, and in so far the result of the 
investigation coincides with the view expressed by BORGEN and 
CoPELAND; but here also the agreement ceases, as they estimated the 
error of the computed refraction at — — г, ог in other words, three 
times the value deduced above. It would be absurd to deduce a 
term of accuracy of this value. The fact that it does not appear 
directly from the latitudes computed that the values of the northern 
culmination decrease, while those of the southern increase with the 
zenith distances, already shows, in a sufficiently clear manner, the 
unreliability of the correction deduced. In appraising the uncertainty 
I shall only add that by including the ten last-mentioned stars with 
zenith distances from76° to 86!/2°, the result would have been: 
During the period from the end of July till the beginning of 
October 1907 we performed a series of circummeridian zenith distance 
observations of the sun. Each group of observations consisted of 
six with “telescope right” and six with “telescope left”. Of these 
six sets of zenith distances three were measured to the lower limb, 
the other three to the upper limb. 
In the tables below the results of these groups of observations 
are given; besides I have included the result of a single group, 
measured on June 3rd 1908. 
In all we are dealing with 26 x 12 — 312 observations of the 
limb of the sun, of which 120 with zenith distances below 70°. 
Within each group the single observations harmonize very well. An 
investigation of the accuracy of the observations in the case of two 
groups gave the mean error of the single observation of a limb 
from 2” to 4” (28/VII and 24/VI1). 
The results are here, as in the case of the star observations, 
collected in two tables corresponding to the zenith distances below 
and over 70°. 
It is apparent that the value of the latitude which may be 
deduced from the ten determinations, the zenith distances of which 
are below 70°, coincides pretty well with the corresponding value 
