1193 
period of 3.968 days. Lack of time, because of my work at the 
observatory, prevented me from immediately reducing my observations 
by means of this value for the period and so testing the variability. 
The probability that @ Ursae minoris was indeed a short-period- 
variable of the type of d Cephei grew stronger, when I found in 
1906 ') that it showed the same peculiarity in its spectrum as those 
stars (c-character after Miss Maury) and has, as all stars of short 
period of this type, an extraordinary slight density. In a footnote 
attention was already drawn to these moments of probability. 
Starting from the consideration, that for all these short-period- 
variables the photographic amplitude is much larger than the visual 
one, Hirrzsprunc at Potsdam has thereupon (in 1910 and 1911) 
taken a great number of photographs (418 plates in 50 nights) of 
Polaris, and from this settled with absolute certainty a variability 
with an amplitude of 0.17 magnitude *). For the epoch of maximum 
light he found J. D. 2418985.86 + 0.08 Greenwich M. T. Subsequently 
J. SreBBINs has executed a number of photometric measurements with 
his exceedingly sensitive selenium-method in 1911—12; these also 
clearly show a variability with a visual amplitude of 0.07 magnitude *). 
The epoch of greatest brighness as found by him, viz. J. D. 2418985.94 
Gr. M. T. agrees very well with Hrrrzserune’s result. 
I have also reduced my observations of 1890—1900 with the aid 
of the periodic time 349681, as spectrographically found. In the 
second half of each year [| used for comparison the stars of Perseus 
and Andromeda, in the first half those of Ursa major. Thus the 
observations form two mutually independent series, partially over- 
lapping in wintertime. For the 1% series « Persei = 6.3, 8 Andromedae 
= 3.8, y Andromedae = 3.1, and exceptionally a Arietis = 5.4 and 
«a Andromedae = 2.3 were used as a scale of comparison-stars ; for 
the 2rd series served e Ursae maj. = 2.4, 4 Ursae maj. = 0.0, and, 
exceptionally, « Ursae maj. = 4.0. The observations were not corrected 
for atmospheric extinction, since this influence disappears in the 
mean of many observations and at the most can make the mean 
error seem too great. Taking all together, from 1890 up to 1899 
259 comparisons with the Perseus-Andromeda-stars were available 
and 251 comparisons with those of Ursa major. With the aid of 
the periodie time 3.968 all epochs of observation were reduced to 
') See A. PANNEKOEK, The luminosity of stars of different type of spectrum. 
Proceedings Acad. Amsterdam 9, 1906, p. 134, 
*) Astronomische Nachrichten 4518 (Bd. 189, 89). 
5) Astronomische Nachrichten 4596 (Bd. 192, S. 189). 
