SPIKAL NEBULAE — -CURTIS. 131 
in the Milky Wa}' may well be ascribed, at least in part, to such 
occulting matter. 
Though Ave thereby run the risk of arguing in a circle, the fact 
that no spirals can be detected in our galactic plane, a natural result 
of such a ring of occulting matter, Avould in itself appear to lend 
some probability to the hypothesis. The peculiar distribution of the 
spiral nebulae would then be explained as due, not to an actual asym- 
metrical and improbable distribution in space, but to a cause within 
our own galaxy, assumed to be a spiral with a peripheral ring of 
occulting matter similar to that observed in a large proportion of the 
spirals. The argument that the spirals must be an integral feature 
of our own galaxy, based on a relationship of avoidance, would then 
lose its force. The explanation appears to be a possibility, even a 
strong probability, on the island universe theory, and T knoAV of no 
other explanation, on any theory, for the observed phenomenon of 
nebular distribution about our galactic poles. 
SUMMARY. 
The spiral nebulae as island universes. 
1. On this theory it is unnecessary to attempt to coordinate the tre- 
mendous space velocities of the spirals with the thirtyfold smaller 
values found for the stars. Very high velocities have been found for 
the Magellanic Clouds, which may possibly be A'ery irregular spirals, 
relatively close to our galaxy. 
2. There is some evidence for a spiral structure in our OAvn galaxy. 
3. The spectrum of the majority of the spirals is practically identi- 
cal Avith that given b}^ a star cluster ; a spectrum of this general type 
is such as Avould be expected from a vast congeries of stars. 
4. If the spirals are separate universes, similar to our galaxy 
in extent and in number of component stars, we should observe 
many new stars in the spirals, closely resembling in their life history 
the 27 novae Avhich have appeared in our own gsilaxy. Over a dozen 
such novae in spirals have been found, and it is probable that a 
systematic program of repetition of nebular photographs Avill add 
greatly to this number. A comparison of the average magnitudes of 
the novae in spirals with those of our own galaxy indicates a distance 
of the order of 10,000,000 light-j-ears for the spirals. Our own galaxy 
at this distance would appear 10' in diameter, the size of the larger 
spirals. 
5. A considerable proportion of the spirals show a peripheral 
equatorial ring of occulting matter. So many instances of this have 
been found that it appears to be a general though not universal char- 
acteristic of the spirals ; the existence of such an outer ring of occult- 
