The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

July, 1921 occurring is probably somewhat greater than the number recorded The limited area of the storm within the tropics (the diameter of the area of violent winds is here frequently less than 100 miles) and the scarcity of observing vessels in the region throughout which the storms manifest their greatest activity make it probable that a considerable percent- age escapes observation. The occurrence during the Ill-year period, 1890-1900. according to.the records of the United States hydrographic office, was as shown in the table on the facing page. In south latitude the storm season is from September to May, February and March being -the worst months. It would thus appear that in both hemis- pheres the storm season corresponds to the tiine when the sun is approaching the equator on its return from the great- est declination north or south. During the 'season of tropical storms whatever interferes with the regularity of the diurnal oScillation of the barom- eter should be considered an indication of a change of weather. .The barom- meter is by fO means an_ infallible evide for warnings much in advance, but after the beginning of the storm it will more or less accurately indicate the rapidity of approach and distance from the center, and its indications should in no case be disregarded. One of the earliest indications of the approach of a tropical storm is the ap- pearance of the sky and general clear- of the atmosphere. Tropical cy- clonic storms are almost invariably pre- ceded by a day of unusual clearness, when distant objects not usually visible stand out with great distinctness. The temperature at such times is more than unusually oppressive. liess This is frequently accompanied by an unusually high barometer. Later it may be followed by a restless oscillating ot pumping of the mercury caused by the disturbed condition of the atmosphere. Then the sky becomes overcast and: re- mains so, at first with a delicate cirrus haze, which shows no disposition to clear away at sunset, but which later becomes gradually more and more dense until the dark mass of the true hurricane cloud appears upon the horizon. From the main body of this cloud portions are detached from time 'to time and drift. acrass the sky, their progress marked by squalls of rain and wind of creasing force. Rain, indeed forms one of the most prominent features of the storm. In the outer portions it is fine and mist like, with occasional showers, these latter increasing in fre- quency and jin copiousness. In_ the neighborhood of the center it falls in torrents, The rain area extends farther in advance of the storm than in the rear. A long swell from the direction of MARINE REVIEW 120° 307 40" ld 30° |\me Ic Ib! 130° TRACK OF ASIATIC TYPHOONS the storm frequently sets in before any other indications become marked. When the sky first becomes overcast with the characteristic veil of cirrus the storm center will most probably lie in the direction of the greatest density of the cloud. : When the hurricane cloud appears over the horizon it will be densest at the storm center. By this time the barometer will usually be showing unmistakable evi- dence of a fall, and one may confidently look for a storm and begin observations to determine the location of its center and the direction in which it is moving. Surrounding the actual storm: area is a territory of large extent throughout which the barometer reads a tenth of an inch or more below the average, the pressure diminishing toward the cen- tral area, but with no such rapidity as is noted within that area itself. Throughout the outer ring unsettled weather prevails. The sky is ordinarily covered with a light haze, which in-