THE NATURE OF THE NEBULAE — HUBBLE 145 



scuration in the two directions, we find that the Cepheids in M31 

 average about 4.3 magnitudes fainter than Cepheids with correspond- 

 ing periods in the Cloud. Therefore, the relative distances are about 

 7.2 to 1. Since the current value for the distance of the Cloud is 

 95,000 light-years, the distance of the spiral is about 700,000 light- 

 years, subject to the previously mentioned uncertainty. 



NORMAL NOVAE 



About 120 normal novae have been discovered in M31 during the 

 past 21 years. The record is by no means complete, for an analysis 

 of the well-observed seasons suggests that the frequency is of the order 

 of 25 to 30 per year. These novae form a very homogeneous group; 

 their light curves are more or less similar over the regions that can be 

 observed, and the dispersion about the mean luminosities at corres- 

 ponding stages is remarkably small. 



A discussion of the data some 10 years ago suggested a mean 

 luminosity at maximum of the order of M= — 5.7 with a dispersion of 

 about 0.5 magnitude. The mean light-curve for the first month after 

 maximum appeared to be approximately linear with a slope of the 

 order of 0.05 magnitude per day. Spectra of two novae, obtained by 

 Humason in 1932, exhibited the familiar features of normal novae in 

 the galactic system. 



The general similarity to galactic novae is pronounced, but a detailed 

 comparison of maximum luminosities leads to discrepancies. Galactic 

 novae, as a class, are systematically brighter by more than a magni- 

 tude. Moreover, several cases are known — e. g., Nova Aquilae (1918), 

 Nova Persei (1901), T Scorpii (1860) — where the maxima were of the 

 order of M=— 9, which would be highly improbable on the basis of 

 the small dispersion about the mean maximum derived in M31. 



The study of galactic novae, as a class, is subject to effects of se- 

 lection which favor the brighter objects, and to difficulties of measur- 

 ing individual distances. For these reasons, it was hoped that the 

 group characteristics could be determined in the neighboring spiral, 

 and the results used in the interpretation of observations within the 

 galactic system. The procedure cannot be followed unless the two 

 groups of novae are known to be strictly comparable. Thus the ap- 

 parent discrepancies assumed unusual importance; they indicated 

 either intrinsic differences between the groups or errors in the inter- 

 pretation of the available data, and no progress was possible until the 

 question was settled. 



Quite recently, my colleague, Dr. Baade, has found a solution in the 

 latter alternative. Observations of M31 with large reflectors were 

 necessarily made only a few days in each month, and most of the 

 novae were discovered at some unknown interval after maximum. 

 The mean light-curvefwas*'constructed on the simple assumption that 



