hour, we have the following variations of configuration :—At the winter solstice the Milky Way passes nearly through the zenith, crossing the horizon towards the south-east and north-west. Its brightest part is low down towards the last named quarter, where the constellation Cygnus lies close to the horizon. One month later, the Milky Way crosses the horizon towards the south-south-east, and northnorth-west, and is bowed some 20° from the zenith towards the west-south-west. Cygnus is now half-set. Yet another month, and the Milky Way is found crossing the northern and southern points of the horizon, and bowed about 40° from the zenith towards the west-Cygnus more than halfset. At the Vernal Equinox, the Milky Way crosses the horizon towards the south-south-west, and north-north-east, and is bowed upwards of 50° from the zenith towards the west-north-west. One month later (that is about April 20), we find the Milky Way crossing the horizon towards the west somewhat southerly and towards the east somewhat northerly, and only raised about 20° above the northern horizon. It is now easy to follow the remaining changes without special comment. The eastern end travels southwards, the western northwards, along the horizon, the central part approaching the zenith, just as hitherto it has been seen to leave the zenith. The bright parts in Cygnus and Aquila are more and more favourably seen as they approach the zenith, being best seen in July and August (at 10 p.m.). About this time we see the southern portion of the Milky Way somewhat beyond Antares (the heart of Scorpio), whereas six months before the greatest range on the opposite side (including nearly the whole of the gap in Argo) had been visible. At the end of October the Milky Way (at 10 P.M.), is seen vertically overhead, and crossing the horizon towards the east and west quarters, the western half being the most conspicuous.

Galileo was the first to prove, though earlier astronomers had entertained the notion, that the Milky Way is composed of a vast number of stars, crowded closely together. But no attempt was made to offer a theory of its structure until, in 1754, Thomas Wright, in his Theory of the Universe,' propounded views closely according with those entertained later by Sir W. Herschel. Wright, having examined a portion of the galaxy with a reflecting telescope, only one foot in focal length, came to the conclusion that our sun is in the midst of a vast stratum of stars; that it is when we look along the direction in which this stratum extends, that we see the zone of light constituting the Milky Way; and that as the line of sight is inclined at a greater and greater angle to the mean plane of the stratum, the apparent density of the star-grouping gradually diminishes.

But it is to Sir W. Herschel, and the supplementary labours of Sir J. Herschel, that we owe the more definite views now commonly entertained respecting the Via Lactea. The elder Herschel, whose nobly speculative views of nature were accompanied by practical common sense, and a wonderful power of patient observation, applied to the heavens his celebrated method of gauging. He assumed as a first principle, to be modified by the results of observation, that there is a tolerable uniformity in the distribution of stars through space. Directing his twenty-feet reflector successively towards different parts of the heavens, he counted the number of stars which were visible at any single view. The field of view of this reflector was 15' in diameter, so that the portion of the sky included in any one view was less than one-fourth of that covered by the moon. He found the number of stars visible in different parts of the heavens, in a field of view of this size, to be very variable. Sometimes there were but two or three stars in the field; indeed, on one occasion he counted only three stars in four fields.

In other parts of the heavens the whole field was crowded with stars. In the richer parts of the galaxy as many as 400 or 500 stars would be visible at once, and on one occasion he saw as many as 588. He calculated that in one quarter of an hour 116,000 stars traversed the field of his telescope, when the richest part of the galaxy was under observation. Now, on the assumption above-named, the number of stars visible when the telescope was pointed in any given direction was a criterion of the depth of the bed of stars in that direction. Thus, by combining a large number of observations, a conception-rough, indeed, but instructive-might be formed of the figure of that stratum of stars within which our sun is situated.

One section of the galactic nebula, as determined from Herschel's observations, is given in fig. 1. The projections

FIG. 1.

extending to the left correspond to those portions of the particular great circle considered, which cross the double part of the Milky Way; the opposite projection represents the portion crossed by the single stream; while the comparative flattening of the central part indicates the gradual diminution of star-density in directions removed from the galactic zone. It is, of course, to be understood that Herschel was far from supposing that such a figure correctly represented the figure of the section. He looked upon it as affording but a rough indication of the true figure. He had indeed noticed, so early as 1785, that there is a tendency in the Milky Way to

cluster around definite regions of the heavens; and he saw that the fact of such clustering was sufficient to account for many irregularities of the figure, quite irrespectively of the absolute extent of the Milky Way in space. If we are looking from a height at the lights of a large town, we may fairly assume that a row of many lights very closely ranged, lies at a greater distance from us than another row containing lights more widely dispersed, if we have reason to suppose that throughout all the streets of the town the lights are separated by distances approximately equal. But if we have reason to suspect that there are some streets lighted more fully than others, the inference would be no longer valid. And again, Herschel suspected that there are stars so large as to bear a sort of sway among other stars by superior attractive influence. Here, then, was another element of difficulty, since it becomes clear (1) that the brilliancy of a star is no positive evidence of proximity; and (2) that there may be (besides the obvious clusterings already considered) laws of systematic distribution, which might largely modify the evidence afforded by star-ganging. For instance, returning to the illustration given above, if we have reason to suspect that there are many lights of superior brilliancy, in some parts of a town, and that further there are in some streets laws of arrangement among the lights, or that there are irregularities of surface-contour, which produce here and there a greater or less foreshortening than would result on a level ground, we should have to make allowance for these points in attempting to form an estimate of the distances at which the different parts of the town are removed from us.

Still, the results obtained by Sir W. Herschel have very properly been accepted as affording general evidence of high

value.

Sir J. Herschel, during his residence at the Cape of Good

Hope, carried out an extensive series of observations of the southern heavens. Applying his father's method of gauging, with a telescope of equal power, he obtained a result agreeing, in a most remarkable manner, with those obtained by Sir William Herschel. It appeared, however, that the southern hemisphere is somewhat richer in stars than the northern, a result which has been accepted as indicating that our system is probably somewhat nearer the southern than the northern part of the galactic nebula. Moreover, Sir J. Herschel was led to believe that the sidereal system forms a cloven flat ring rather than a disc.

Combining the results obtained by the two Herschels, we should assign to the stratum of stars a figure somewhat resembling that of the solid cloven disc exhibited in fig. 2. This, being a side view, gives the figure of the imaginary section which appears so often in works on astronomy. In fig. I the small circle near the centre indicates the probable

FIG. 2.

extent (on the theory in question) of the sphere within which stars down to the fifth and sixth magnitude may be supposed to be included.

The main difficulties in attempting to form an estimate of the real configuration of the galactic system are those which have been already mentioned. Have we evidence confirming or disproving (1) the tendency to clustering suggested by the elder Herschel, (2) the possible variability among starmagnitudes, and (3) the action of influences exerted by large stars in guiding or swaying others? It appears to me that there are indications of a very obvious and important character, which have been either altogether unnoticed, or much less