ON THE FACTS AND THEORY OF EARTHQUAKE PHENOMENA. 53 
catalogue, the same relation of scale as before (figs. 1 and 2, Plate VIII.) 
being maintained between the northern and southern abscisse, we find 
some of the apparent anomalies disappear. In fig.1, Plate X. the curve of 
season for the northern hemisphere assumes a very regular form, and gives 
a decisive minimum for the summer season (in May and June), and an 
equally clear maximum for the winter season (in December and January). 
In fig. 2, Plate X. the corresponding curve for the southern hemisphere, 
however, still shows two maxima and two minima, the maximum at the 
commencement of winter, with second maximum at midsummer; the 
minima in spring and autumn assuming the months constituting the re- 
spective seasons reversed in the two hemispheres. It must be borne in view, 
however, that the base of induction for this hemisphere is from only 223 
observations, against 5879 in the northern; that if the southern curve had 
been drawn to the same vertical scale as the northern, it would have ap- 
peared to the eye as almost a straight line; so that very little weight is to 
be attached to the discordance it appears to present to the corresponding 
curve, its necessarily exaggerated scale falsely addressing the eye. 
In fig. 3, Plate X., the two curves preceding are combined, but to the 
same scale of vertical or of seismic abscissa; and the result shows how little 
in reality the data that we possess as yet for the southern hemisphere are- 
capable of modifying the facts we have for the northern. The southern 
curve, in fact, scarcely alters to the eye the preceding northern one ; and 
the new curve of season for both hemispheres presents still the winter maxi- 
mum and summer minimum. 
In fig. 5, Plate X., a curve has been obtained for the whole period of 
the catalogue and for both hemispheres, representing graphically all recorded 
earthquakes occurring near or at the equinoxes and solstices (the critical 
epochs of Perrey and others) within a limit of twenty days, ¢.e. ten days be- 
fore and ten days after each equinox and solstice. The base of induction is 
moderately large, the catalogue containing the following numbers :— 
Vernal equinox (March 10—30)........... . 310 
Summer solstice (June 11—July 1).......... 254 
Autumnal equinox (Sept. 13—Oct. 3) ...... 249 
Winter solstice (Dec. 11—31).............. 318. 
This we may call the equinoctial and solstitial curve of comparative seismic 
energy. It indicates a distinct maximum about the winter solstice, and an 
equally distinct minimum rather before the autumnal equinox. Taking the 
average of the whole year for any lengthened period, it may admit of much 
doubt, whether there is any real seismic paroxysm at the equinoxes and sol- 
stices, although a clear preponderance is shown by our catalogues at two out of 
the four annual epochs at which all are recorded ; yet, from the accordance 
of Perrey’s results with those given by this much larger base of induction, 
we cannot put aside the possibility that the fact may have a cosmical basis. 
The most direct connexion in such case that we should expect to find, 
with other ascertained, periodical phenomena, would be with the annual 
march of the barometer. In fig. 4, Plate X., the annual curves of mean 
mensual barometric pressure are laid down to the same scale of ordinate for 
time as the equinoctial and solstitial seismic curve below (fig. 5), giving the 
variation in atmospheric pressure for places in several and distant latitudes, 
Macao, Havanna, Calcutta, Benares; and in Europe, Halle, St. Petersburg, 
Berlin, Paris, and Strasburg,—the curves themselves having been reduced 
from those of MM. Buch, Dove, and Kaemtz. 
On comparing these barometric curves with the seismic one, an obvious 
