The same question rises to the lips of every observer: How can they possibly have been brought into such a situation? The marvel does not grow less when we know that, instead of being closely compacted, the stars of the cluster are probably separated by millions of miles; for we know that their distances apart are slight as compared with their remoteness from the Earth. Sir William Herschel estimated their number to be about fourteen thousand, but in fact they are uncountable. If we could view them from a point just within the edge of the assemblage, they would offer the appearance of a hollow hemisphere emblazoned with stars of astonishing brilliancy; the near-by ones unparalleled in splendor by any celestial object known to us, while the more distant ones would resemble ordinary stars. An inhabitant of the cluster would not know, except by a process of ratiocination, that he was dwelling in a globular assemblage of suns; only from a point far outside would their spherical arrangement become evident to the eye. Imagine fourteen-thousand fire-balloons with an approach to regularity in a spherical space—say, ten miles in diameter; there would be an average of less than thirty in every cubic mile, and it would be necessary to go to a considerable distance in order to see them as a globular aggregation; yet from a point sufficiently far away they would blend into a glowing ball. [Illustration: Star-cluster in Hercules (Photographed with a two-foot reflector)] Photographs show even better than the best telescopic views that the great cluster is surrounded with a multitude of dispersed stars, suggestively arrayed in more or less curving lines, which radiate from the principle mass, with which their connection is manifest. These stars, situated outside the central sphere, look somewhat like vagrant bees buzzing round a dense swarm where the queen bee is sitting. Yet while there is so much to suggest the operation of central forces, bringing and keeping the members of the cluster together, the attentive observer is also impressed with the idea that the whole wonderful phenomenon may be _the result of explosion._ As soon as this thought seizes the mind, confirmation of it seems to be found in the appearance of the outlying stars, which could be as readily explained by the supposition that they have been blown apart as that they have flocked together toward a center. The probable fact that the stars constituting the cluster are very much smaller than our sun might be regarded as favoring the hypothesis of an explosion. Of their real size we know nothing, but, on the basis of an uncertain estimate of their parallax, it has been calculated that they may average forty-five thousand miles in diameter—something more than half the diameter of the planet Jupiter. Assuming the same mean density, fourteen thousand such stars might have been formed by the explosion of a body about twice the size of the sun.