k would be subject to
considerable variation, according to seasons, latitudes and atmospherical
conditions. Having noted these positions, I next combined them separately,
obtaining the remarkable results given in fig. 4. The combined midnight
positions of the Ursa Major or Minor, at the four divisions of the year,
yielded symmetrical swastikas, the forms of which were identical with the
different types of swastika or cross-symbols (the normal, ogee and volute,
etc.), which have come down to us from remote antiquity and are reproduced
here for comparison (pl. II, _a-f_). Reflection showed me that such
composite pictures of the Ursa constellations constituted an exact record
of their annual rotation, and afforded a perfect sign for the period of a
year. I moreover perceived how the association of rotatory motion with the
advance of time, and its division into fixed periods or cycles, would be
the natural outcome of the recognition of the annual rotation of the
star-groups.

The Calendar-Swastika, or cross of ancient Mexico (pl. II, _g_)
constitutes an absolute proof of the native association of the
cross-symbol with the ideas of rotatory motion and the progress of time,
and furnishes an indication that, in an analogous manner, the swastika may
have been primarily and generally employed by primitive races, as a sign
for a year or cycle. A close scrutiny of the respective forms of the
crosses yielded by Ursae Major and Minor shows that the normal swastika and
suavastika may be explained as the separate representations of the two
constellations--the angular break in the outline of Ursa Major suggesting
the direction of the bend to the right of the arms of the normal swastika,
whilst the form of Ursa Minor obviously suggests the bend to the left
which is characteristic of the suavastika.

[Illustration.]

Figure 5

My growing conviction that the Bear constellations had furnished the
archetype of the different forms of swastika and cross-symbols, found
subsequent support when I referred to the map showing the geographical
distribution of the ancient symbol published by Prof. Thomas Wilson in his
valuable and comprehensive monograph on the subject,(1) to which I am
indebted for much information and several illustrations (pl. II, _a-f_,
etc.). The map, reproduced here (fig. 5), proves that, with two
exceptions, which can be attributed to a migration southward, th