169 
Kn.—H.17 and the calculated absolute rate of H. 17 the absolute 
rate of the Kn. clock was derived. 
By this method the observed transits of the fundamental stars 
give only the mean clock correction at a mean sidereal time and 
did not give any data for finding the rate of the clock. 
The question now arose, whether it would not be possible, excluding 
the clock H. 17, to derive from the registered transits of the fund. 
stars not only the mean clock correction, but also the rate, both 
expressed in time of the Kn. clock. 
In order to arrive at a solution of this question | here give the 
postulates underlying each of the methods. 
I: (by means of H. 17). 
1. The times of transits of the fund. stars must be sufficiently 
accurate to give a reliable clock correction by their mean. 
2. The relative rate Ku.—H 17 obtained must be accurate, from 
which it follows that: 
a. The rate of the Kn. clock must be constant during the obser- 
vations. 
b. The registering of the beats of H.17 must be accurate. 
c. There must not be a systematic difference between the registration 
of these beats before and after three or four hours of observations. 
3. The rate of the H.17 clock must conform to the formula given 
above, not only in the mean for several days, but must do so 
without any retardation, in other words it must react immediately 
to every change of temperature and barometer. 
Il: (without the H.17 clock). 
1. The transits of the fundamental stars must have sufficient 
accuracy to give by their mean a reliable clock correction and at 
the same time a reliable clock rate. 
2. The rate of the Kn. clock must be constant during the obser- 
vations. 
A comparison of the two groups of postulates shows that II, 2 
and J, 2a are the same. 
The only difference between II,1 and I,1 is that in the second 
method the observations of the transits of fund. stars require a 
higher degree of refinement and the question is reduced therefore 
to the following: 
What conditions must be laid down for observations to make 
them suitable for the deduction of a clock rate? 
In the first place it is desirable that the stars of the observation 
programme should all be included between fundamental stars, so 
