82 HURRICANE from the station ; on board a vessel, the baro- metric fall is further complicated by the move- ment of the observer. The best idea of the barometric disturbance is given by a chart of synchronous observations on which isobaro- metric lines are drawn, these isobars will be found to be crowded together on one side (generally the advancing half) of the storm more than on the other, and to enclose a small oval or circular region of lowest pressure, al- most if not quite identical with that of the area of calms, though sometimes apparently in advance of it. In a general way it may be stated that the velocity of the wind increases with the crowding of the isobarometric lines. The exact relation between the two is quite complicated, and may be deduced from the formulas of the above mentioned treatise by Ferrel, combined with the considerations in- troduced by Peslin in 1867 and Reye in 1872. It is evident that the law above given for the rotation of the wind may be converted into a rule for finding the centre of calms, which will also hold good for finding the centre of lowest barometer ; this latter is generally spoken of as the storm centre or axis. Buys-Ballot has expressed this generalization in the form known as Buys-Ballot's rule, viz. : in the northern hemisphere stand with your back to the wind, and the lowest pressure will be on your left hand and somewhat in front thereof; a rule that applies especially to, and was apparently suggested by, the behavior of the winds of hurricanes and similar storms. The dimen- sions of hurricanes generally increase from day to day until the dissipation of the entire storm, while the intensity of the winds is believed on the average to diminish somewhat ; this will however depend upon the atmospheric condi- tions favoring the development or the deca- dence of the disturbance. Given a proper sup- ply of warm moist air, and it can be shown that the central depression with the attendant wind and rain must steadily increase up to a certain limit. These favorable circumstances are gen- erally found combined in a remarkable degree in the region of the Gulf stream, the Kuro Siwo, and similar ocean currents; accordingly, on reaching these the area of cloud and rain ex- pands, as also do the diameters of the isobaric curves. The dimensions of the central depres- sions vary quite irregularly, hut appear on the average to increase as the storm continues; while the actual height of the barometer at the centre changes much less, but is believed to diminish gradually so long as the intensity of the wind increases. If a curve, enclosing a region in which the winds attain the force ordinarily described as a moderate gale, be as- sumed as the limit of the storm, it will be found that in the earliest stages of the hurri- cane it has a diameter of from 50 to 200 m., which increases in the course of 5 or 10 days to from 400 to 1,200 m. ; thus a disturbance that may have been originally designated as small or local, increases so as to involve half the surface of the North Atlantic ocean. The track of the centre of the hurricane is a fair indication of the progress of the storm over the earth, and much labor has been bestowed upon such collations of logs of vessels as would elucidate this important branch of the subject. But notwithstanding the labor expended, there have as yet been very few hurricanes traced back to what appears to be very near their origin, and in not a single instance has unmis- takable evidence of their origin been adduced. The general position of hurricane tracks in the earlier parts of their course therefore remains obscure, although the immense accumulation of material by the labors of the various na- tional government weather bureaus is rapidly dissolving our ignorance on this point. So far as the known hurricane tracks are concerned, it may be stated that in the North Atlantic ocean each uniformly appears to be a segment of a parabola having its axis coincident with the parallels of 25 to 35 N. latitude, and the longitudes of whose apices fall between the meridians 40 and 100 west of Greenwich, but mostly between 65 and 85. At the southern extremity of the parabolic track, the branch passes either to the north of or over the Windward islands, while the northern branch passes to the south of or over Newfoundland. In a few cases the first portion of the track has been traced southeastward nearly to the coast of Senegambia, and the latter portion of the track northeastward to the ocean be- tween Iceland and Scotland ; some tracks that curve northeastward before reaching Ion. 40 may even strike England or France. The hur- ricanes of the southern hemisphere describe similar parabolic tracks, which lie at a corre- sponding distance south of the equatorial belt of calms, and are symmetrically disposed with reference thereto. Very few have been traced in the South Atlantic ocean, but in the south- ern Indian ocean the majority of the hurricanes pass from Sumatra and Java sonthwestward to within 500 m. of Madagascar, then south- ward and southeastward. In general, Mohn (1870) and Reye (1872) state that all cyclones (of which hurricanes are the grandest examples) move in the direction in which for the longest time the warmest and moistest air has been rising, and producing the heaviest cloud and rainfall. If we combine with this law the tendency of the whirlwind as a whole to move away from the equator, as proved by Ferrel, it seems to the writer that we have a very close approximation to the full statement of the reason for the parabolic form of their orbits. The rate of progression of the West Indian storm centres varies from 50 m. per hour in a few cases to 10 or 15 as the other extreme ; that of the storms of the southern Indian ocean varies from 1 to 20 m. The rate in gen- eral in the North Atlantic increases with the growth and northward movement of the hurri- cane, and, though sometimes quite variable, is not so much so as in the case of the similar
