Sept. 8.—The hurricane is central about the axis of the Gulf Stream, off Hatteras. The area of high barometer in advance is retarding its northward progress, and preventing it from recurving to the northeastward. The heavy surf and the increased height of the tides, due to the storm-wave of the hurricane, begin to attract general attention and to cause damage along the low-lying portions of the coast between Norfolk and Newport; warm, moist ocean air is being driven in over the cold inshore current, and dense fogs are encountered off the coast north of the 35th parallel. The storm is losing a little of its tropical intensity, but its area is increasing and winds of hurricane force are raging over a vast area between Hatteras and Bermuda.
Sept. 9.—The storm is still raging with great violence between Nantucket, Hatteras, and Bermuda. Tremendous seas and tides are driving in on the coast. It is blowing with hurricane force close in shore near Hatteras. The storm-center is still moving northward, but more slowly, and the great area of high barometer into which the hurricane has forced its way stands fast, the pressure rising to 30.30 over the Gulf of St. Lawrence and the isobar of 30.00 reaching south on either side well down toward the tropics. To the northwestward, northward, and northeastward, close to the outer limits of the great whirlwind, warm, sultry weather prevails, with calms or light, variable winds, hazy weather, and barometer above the normal.
Sept. 10.—The hurricane seems to be gathering all its strength for a last desperate struggle to force its way along its normal track to the northeast. The wind-arrows plotted on the Chart illustrate very graphically the marked circular character of the great whirlwind. The storm-wave, or general elevation of the oceanic surface caused by the in-rushing and whirling winds, and the reduced barometric pressure (acting as a partial vacuum), bank up the water in the bight of the coast about Sandy Hook, and cause the greatest floods on record at many places along the eastern coast of New Jersey and the southern shore of Long Island.
Sept. 11.—The barometric pressure at the storm-center has increased noticeably. The low area is filled up, and the barometric gradients are very much less steep. Very few winds of full hurricane force are reported. The storm-center has moved in toward shore, however, and stormy winds are still felt along the coast. The clouds are breaking away in places, and the cyclonic circulation is no longer so well marked. The warm waters of the Gulf Stream have been driven in toward the coast off Block Island by the long-continued and furious southeasterly gales on the right of the storm track, and, mingling with the cold in-shore current, cause dense fogs and squally, unsettled weather.
Sept. 12.—The great hurricane has blown itself out, and although a few reports still give a force of wind as high as 10 of Beaufort's scale, the storm has practically ended. The remnants drift inland during the 13th and 14th, with cloudy, rainy weather in eastern Virginia and Maryland. It leaves a mountainous cross-sea that lasts for several days, a coast line strewn with wreckage and already some twenty additions have been made to the list of derelicts and drifting wrecks whose positions are plotted on the Pilot Chart as a caution to navigators. How many vessels it wrecked at sea can never be fully known, but its entire track is marked by wrecks and wreckage.

The importance to navigators of a true appreciation of the law of storms—not the mere memorization of a set of rules, but an intelligent comprehension of the subject—is now perhaps clearly evident to the reader: at any rate, that is the object I have aimed at, rather than a mere formal statement of generally accepted principles and an abstruse discussion of isobars and gradients.

It will be seen that the probable bearing of the center, as indicated by the direction of the wind at a single station, is the great question, so far as the navigator is concerned. There are men who want and must have a hard-and-fast rule,—an 8-point, a 10-point, or a 12-point rule—something to act on without thought, while every nerve is strained to save the ship's spars, sails, boats, engines, and cargo, from damage or destruction. Under such circumstances, I think that perhaps the safest general plan is to use the old 8-point rule, but applied to the low clouds, instead of to the wind. This is equivalent, generally speaking, to a 10-point rule, applied to the wind. That any such rule, if intended for general application, is only roughly approximate, goes without saying, or ought to do so, at least. The angle of bearing changes in different parts of the storm, it varies with the quadrant, with the latitude, with different storms, and with various other conditions, too numerous to be mentioned or even wholly known. One good general rule is that in rear of a hurricane the wind blows somewhat decidedly toward it; and yet that there are marked exceptions is well illustrated by the chart of the hurricane of November 25, 1888, already referred to. As a good example of the wind circulation in a hurricane in the tropics the accompanying diagram is of interest. This represents two days (the 3d and 5th) of the great Cuban hurricane of September, 1888, the intervening day (September 4th) being omitted, for the sake of clearness. Its severity is sufficiently indicated by the fact that it caused the loss of fully a thousand lives in Cuba, and destroyed property of the estimated value of $5,000,000 in the single province of Sagua. Now take any point on any one of these spiral lines, and observe the bearing of the center: in rear of the storm, especially, the 8-point rule is hardly applicable, and action based upon it might result disastrously.

The Cuban Hurricane of September, 1888, illustrating the surface wind-circulation on September 3d and 5th, at noon, Greenwich mean time.

The next and last plate, entitled,

HURRICANES IN THE NORTH ATLANTIC.—TYPICAL CIRCULATION OF THE WIND, FROM ACTUAL OBSERVATION,