1195 
The remaining deviations Obs.—Cale. have been placed in the 
last column. They yield for the mean error of a normal place 
ENC 
Nd 
Fig. 2. 
according to the mean of the two series, 0.21 (if we adopt this 
same value. for both series, then each maximum has a mean error 
of 04.11), from which we find 0.84 as mean error of one obser- 
vation, while 0.7 had been found from the differences between 
the separate results and the adopted normal places. The deviations 
of the normal places from the sinusoid, it is true, show a systematic 
character, in the sense that the maximum is very sharp, the mini- 
mum very flat, hence that a term with 2p is indicated, the positive 
maximum of which falls together with the maximum of the principal 
term. Since, however, nothing of this kind is to be observed in the 
light-curves of HeRTzsPRUNG and STEBBINS, no further attention has 
been paid to this phenomenon. Thus my observations yield as epoch 
of the maximum, after reduction to Greenwich-time: 
1894 Aug. 4.81 Gr. M.T. = J.D. 2413045.81 + 04.08. 
The interval between my normal-epoch and that of Hertzspruna 
J. D. 2418985.86 is 5940.05 days = 1497 periods of 3.9680 days. 
In order to reduce the brightness of maximum and minimum to 
the same photometric scale, the catalogues of Potsdam and Harvard 
were used. For the reduction of the magnitudes given there to the 
homogeneous scale that has been derived and adopted in my dissertation 
“Untersuchungen über den Lichtwechsel Algols” (p. 146—158) first 
a correction was added to the values of Harvard 44, in order to 
reduce them to Harvard 14. This was derived from the differences 
between the two catalogues, calculated by Mürrur and Kempr and 
communicated in their ’Generalkatalog der photometrischen Durch- 
