ee ve : ; - 
Aug. 7, 1873| 
commencement of the following month similar observations of 
the horizontal magnetic force. For these elements of terrestrial 
magnetism two-hourly, as well as daily mean, curves are traced 
for each month. 
In the general table that closes the report, we notice that the 
prevailing wind never deviates, in any season, more than 13° 31’ 
from the east, and in spring it is only 3° 36’N. of E. The 
rainfall for the seasons given in millemetres was in winter, 71'1 ; 
in spring, 181°0; in summer, 480’0 ; and in autumn, 547°2 ; the 
number of rainy days being respectively 13, 15, 33, and 39. 
The coincidence of magnetic disturbances with local storms, 
with hurricanes in Florida and St. Thomas, with Auroras visible 
in distant lands, and with similar magnetic perturbations in 
England, was remarked in August, September, and November. 
The frequent disturbances of the needle noted in October cer- 
tainly do not agree with photograghic records in England, this 
month having been remarkably free from perturbations of this 
nature, 
The second book contains the results of acontinued series of 
barometric observations between the years 1858 and 1871. The 
reliance we may place on the accuracy of the work can he esti- 
mated from the fact, that the correction of 1‘07mm, for sea level 
was the result of 2,000 comparisons. 
A very regular double period is apparent in the daily range, 
which may be represented by the expression A= sin (a+?) 
+#' sin (6+ 2¢) ; but the range for the day hours is somewhat in 
excess of that of the night. The minima occur at 2—4 AM. and 
3—4 P.M., andthe maxima at9—I0 A.M, and 10 P.M., the times 
varying slightly with the seasons, 
In December, January, and February, the amplitude of the 
daily range is greatest, and then gradually decreasing it attains 
its minimum in June and July. This confirms the law of 
Ramond, cited by Kaemtz, that the amplitude of the baro- 
metric range within the tropics is least in the rainy season. This 
annual variation-of the daily range is, remarks our author, the 
more worthy of note, as it is directly opposed to what has been 
observed in Europe, where the range is greatest in summer. 
This remark appears to me to require some modification, for on 
turning to the monthly mean range observed, forinstance, at 
Stonyhurst, during the last quarter of a century, I find a perfect 
agreement with the annual variation for Cuba. The mean 
values for the several months at Stonyhurst are 1°448, 1°415, 
1378, 1167, 0°970, 0°896, 0°869, 0°927, 1'217, 1°323, 1°45], 
1°449. These means are, it is true, obtained from the extreme 
monthly maxima and minima, but our author informs us that the 
amplitude of the extraordinary oscillations, if we eliminate the 
four greatest which were due to hurricanes, resembles the mean 
annual variation of the range, being greatest in January and 
least in July. The mean values of the extraordinary oscillations 
being almost identical in November, December, February, 
and March, give this annual daily range curve at Cuba a sin- 
gular symmetry. The periodic recurrence of summer storms 
at fixed hours will account for the diminution of the range in that 
season. 
The mean values of the Daily Range have been deduced by 
several methods : 1. From the absolute maxima and minima, by 
which the irregular oscillations are not sufficiently eliminated. 
2. By Humboldt’s method, from the maxima and minima at 
fixed hours. 3. By Kaemtz’s method, from the arithmetical 
means of the maxima and minima, 4. From Bravais formula, 
R=/d*+d?,+..d",;d, d,... being the differences between 
n 
the monthly mean and those of certain fixed hours. There is a 
striking agreement in the results from all these methods, but the 
second shows in certain cases signs of a suspicious irregularity. 
Besides the Daily Range, and an annual variation of this 
range, there exists a yearly variation of the mean value analo- 
gous to the diurnal, having its principal maximum and mini- 
mum in January and October, and secondary ones in July and 
May. This double inflexion of the mean annual curve is peculiar 
to Cuba, since there is generally in the same latitude only a 
single maximum in January, and a minimum in July. 
The abnormal inflexion occurs during the month of June, 
July, August, and September. Kaemtz, in his ‘‘ Météorologie,” 
who is followed by Marie Davy, fixed the principal minimum in 
August, but this and other lesser differences arise probably from 
not eliminating extraordinary perturbations, and from the con- 
fessed imperfection of his series of observations. 
The observations of fourteen years are insfficient to determine 
any certain law respecting the years of hurricanes; but an in- 
NATURE 
295 
spection of the yearly curves shows that 1865 and 1870 are dis- 
tinguished from the rest by the almost identity of the means for 
February, March, and April, followed by a rapid rise from May 
to July, a fall from July to October, and a still more marked rise 
from October to January. 
The third pamphlet gives a very interesting and detailed ac- 
count of the terrible hurricanes that caused such wide-spread 
desolation in the October of 1870. The first storm occurred ol 
the 7th and 8th, the second on the 19th and 2oth. 
The author adopts the theory of cyclones first enunciated by 
Redfield in 1831, and since accepted and modified by many emi- 
nent meteorologists, Situated N. of the Equator he considers 
the storm to be rotating in the direction from E. to W. through 
N. and the centre to be at the same time moving N.W. within 
the tropics, and N.E. in higher latitudes. The resultant path 
he finds to be a spiral wrapped round a parabola, the folds of 
the spiral being closest at the apex of the curve. The position 
of the centre or vortex is given at any moment by the height 
of the barometer and the direction of the wind. The barometer 
being lowest at the centre, the reading of the mercurial column, 
corrected for daily range and for the particular wind, furnishes 
data for determining the distance of the centre, whilst the 
angular bearing of the latter is known from its being at right angles 
to the direction of the wind, and to the rigtht hand of an observer 
facing the wind. This follows necessarily from the circular 
motion of the cyclone, and falls, asa particular case, under the 
general law of Buys Ballot, since we know that the barometer 
is lowest at the vortex. The latteris thought to move in a 
cycloidal curve with loops at the cusps, which just fall on the 
parabolic trajectory. The vortex is thus almost always to the 
E. of the parabola. The double motion of translation and 
rotation causes the effects of the hurricane to be much more 
disastrous in the N. hemisphere to the E, of the parabolic path 
than on the W, side, and the velocity of the wind ata given 
distance from the vortex for any points of the compass may be 
found from the formula V =\/7* + 7? + 2¢r cos @, where ¢ and 7 are 
the velocities of translation and rotation, and @ is measured from 
the E. point when the storm is moving N. The calm at the 
centre of the cyclone gives a ready means of estimating the 
velocity of translation. The storm of the 7th was felt from the 
4th to the 13th, the maximum humidity lasting till the 12th. The 
rate at which the vortex crossed the island was only four miles 
an hour, and this remained almost constant throughout, The 
second storm was much more sudden and rapid, and the in- 
creasing rate, from 93 to 20 miles an hour, at which the vortex 
‘was travelling, showed that the second branch of the parabola 
had been reached before passing Cuba. 
Equal heights of the barometer combined with the directions 
of the wind enable the meteorologist to lay down the parabolic 
trajectory with considerable accuracy, either from observations 
at a single station, or at several, Thus on the 7th at 2 P.M, 
the corrected barometer read the same at Havannah and at 
Cienfuegos, the wind being S. by W. at the latter, and N.E. at 
the former station, the vortex was therefore at that time S.E. of 
Havannah, and a few degrees N. of W. from Cienfuegos, but 
equally distant from the two places. The more rapid changes 
and greater fall of the barometer, together with the increase in 
the velocity of the wind, show that the storm passed more cen- 
trally over Havannah than over Cienfuegos. ‘The discharges of 
electric fluid were very intense, and at Cardenas an appearance 
similar to the aurora borealis was visible for ten minutes. The 
magnetic needles were much disturbed. The inundations from 
the rising of the sea were very destructive, and on the 7th 
the existing wind favoured the rise. This rise under the centre 
of the cyclone seems to follow from the removal of pressure, 
and the inrush of air of different temperatures fully accounts for 
the heavy rainfall. The diminution of atmospheric pressure is 
also offered as a probable explanation of the slight shocks of 
earthquake, due perhaps to the violent expansion of certain gases 
confined within the cavities that abound in the island. 
A careful consideration of the accounts published in the local 
papers, and a personal inspection of the localities, tended 
strongly to confirm the results of theory. p 
Cuba, from its situation just within the Tropic of Cancer, and 
at the entrance to the Gulf of Mexico, is admirably placed for 
the study of these cyclonic storms, and eight of those which 
have been best observed are traced on a map appended to the 
pamphlet, showing that in most cases the apex of the parabolic 
curve is not far from the island. It is a subject of congratulation 
that an observatory so well conducted, and so situated, has, by 
