The Great Shapley-Curtis Debate, 1920

Harlow Shapley (Mt. Wilson Observatory)	Heber D. Curtis (Lick Observatory)
Basic Views The galaxy is approximately 300,000 light years in diameter and the sun is located far from the center. The spiral nebulae are associated with the galaxy, although outside the main body. The nature of the spirals is not known, but is probably some combination of gas and faint stars. The Milky Way and its “halo” of globular clusters and spiral nebulae is all there is to the universe.	Basic Views: The galaxy is approximately 30,000 light years in diameter and the sun is located near the center. The spiral nebulae are “island universes”, i.e., other galaxies comparable in size to the Milky Way. The universe contains a large, indeterminate, number of galaxies spread out over a large, indeterminate, volume of space.
Dimensions of the Milky Way	Dimensions of the Milky Way
The dimensions of the galaxy are deduced from the distribution of globular clusters. The distance to the brightest globular clusters using Cepheid variables and the bluest stars range from 20,000 to 45,000 years, distances greater than Curtis assigns to the dimensions of the entire galaxy. Assuming that all globular clusters are roughly the same size, the dimmest clusters must be hundreds of thousands of light years away. Since the galaxy is inside this “halo” of clusters, its maximum dimensions must be comparable to the maximum globular distance. This yields a galactic diameter of approximately 300,000 light years. The globulars “outline” the Milky Way, but they are not evenly distributed. They are concentrated in an area centered in the constellation Sagittarius. This happens to be the brightest part of the Milky Way. Also, from Newton’s universal law of gravity, we can be sure that any physical body will always be attracted to the greatest concentration of mass. Thus, the center of the galaxy must be in the direction of Sagittarius, not in the vicinity of the sun.	Shapley’s estimate of the Milky Way’s size is wrong because his method for measuring distance is inadequate. He concentrates on two types of unusual stars, namely bright blue stars and Cepheid variables. The absolute magnitude magnitudes of these types of stars are both poorly known. The possible range of magnitudes for bright blue stars can be almost an order of magnitude, while there is no parallax measurement available for any Cepheid. The method Shapley uses to derive the absolute magnitude vs. period for Cepheids is too uncertain to be trusted. A more accurate measurement of distance is to discard these rare bright stars and use the average magnitude of all the stars in the clusters. In our neighborhood, most stars are types F through K with an average absolute magnitude of +4.5. Shapley’s average distance to the globular clusters (on the order of 100,000 light years) would imply that the average absolute magnitude in these clusters is +0.6, some 40 times brighter than the stars in our neighborhood. Since there is no reason to believe that the stars in clusters are any different than nearby stars, this difference makes no sense. If we assume an average distance to the globular clusters to be on the order of 10,000 light years rather than 100,000 light years, the average absolute magnitude of cluster stars drops to +5.6, a very reasonable figure.
The Nature of the Spiral Nebulae	The Nature of the Spiral Nebulae
There are three basic reasons for rejecting the theory that the spiral nebulae are “island universes” (we would say “other galaxies”): The spiral nebulae are distributed uniformly about the Milky Way with a very distinct “zone of avoidance” containing no spiral nebulae running exactly along the galactic equator. Such a distribution strongly suggests an association with the Milky Way (Why would randomly distributed galaxies at vast distances always avoid only this specific region of the sky?) Spectroscopic studies have shown conclusively that many nebulae are gaseous (they have emission spectra). It is true that spiral nebulae have absorption spectra, like stars, but it is also true that all nebulae have some stellar association. For example, the Orion nebula (M42) appears mostly gaseous, but in a telescope we see stars embedded in it. On the other hand, the Pleiades appear to be simply a group of stars to the naked eye, but long exposure photographs reveals that it, too, is embedded in a gaseous nebula. It would appear that gaseous and non-gaseous nebulae are not distinct classes of objects, but form a continuum, going from almost all gas to almost all stars. Finally, the enormous size of the Milky Way argues against the island universe theory. If the spiral nebulae are really other galaxies, their sizes must be on the order of the Milky Way. To appear as small as they do, this implies that the distance to even the nearest is enormous – millions of light years. Two facts about spiral nebulae argue that they cannot be that big and that far away: First there are the novae, stars that brighten, then fade. In 1885 there was a nova in M31 (the Andromeda galaxy) that almost reached naked eye visibility. If it is really a million light years away, to appear that bright, the peak luminosity of this single star would have to equal the combined luminosity of billions of suns! Then there is rotation. In 1915 A. van Maanen published a paper on the rotation of the spiral nebula M101. His data indicated that M101 rotates with a period of approximately 85,000 years. Now if this spiral were millions of light years away and comparable to the size of the Milky Way, it is easy to show that a point on the edge of this galaxy would be traveling at a speed greater than the speed of light, a physical impossibility.	The spiral nebulae are island universes comparable to the Milky Way because: Since the Milky Way is only one tenth the diameter in this model than it is in Shapley’s, it follows that the spirals are correspondently smaller and, therefore, closer. Assuming a diameter of around 20,000 light years for M31, for example, gives a distance of about 500,000 light years. This effectively eliminates the problem of the super bright novae and super fast rotation. The novae are still bright and the rotation is still fast, but both are possible within the realm on known physics. The spectra of the spirals, while not conclusive, are very consistent with the island universe theory. The integrated spectra are of the F-K type, just like the integrated of the typical stars in the Milky Way, and just the sort of spectra we would expect from a large galaxy of stars. If we assume the spirals are galaxies like the Milky Way, we can use their observed properties as analogies for our own galaxy. Many edge-on spirals, for example, show dark bands of obscuring material along the major axis of their disks. We see similar regions of obscuring matter in the Milky Way. This offers an explanation for why spirals are never seen in the Milky Way (the zone of avoidance) – they are blocked from view. A large number of novae have been identified in spiral nebulae. This makes sense if spiral nebulae are island universes, but hard to explain if spirals are part of the Milky Way. Why should this region of the Milky Way have such a high concentration of novae? If the spirals are as close as Shapely claims (20,000 light years for M31), then the large number of novae observed there have very high absolute magnitudes (meaning very low absolute luminosity – the larger the magnitude, the dimmer the star), much higher than observed in the Milky Way. Why should these objects and these objects alone have extraordinarily dim novae? On the other hand, if the spiral distance is more like 5 times Shapley’s estimate, then the absolute magnitudes correlate well with the absolute magnitudes of local Milky Way novae. If the spirals are other galaxies, it makes sense from the standpoint of stellar evolution. All other stellar phenomena are concentrated around the Milky Way – stars, gaseous nebulae, clusters. Spiral nebulae are not. Where do they fit in, if they are not external to the system?