"Strictures on Prof. Dove's Essay on the Law of Storms." (In reply to Robert Hare's "Objections to Mr. Redfield's Theory of Storms, and Strictures on Prof. Dove's Essay on the Law of Storms.")''On the First Hurricane of September 1853 in the Atlantic, with a Chart, and Notices of other Storms.'-' On Three several Hurricanes of the Atlantic, and their Relations to the Northern of Mexico and Central America; with Notices of other Storms. New Haven, 1846, 118 pp.'-Observations relative to the Cyclones of the Western Pacific, 1857.'

A complete list of Redfield's works may be found in Olmsted's Address on the Scientific Life and Labours of William C. Redfield.' New Haven, 1857, 8vo.

*

The materials which were collected with such care by Mr. Redfield have been considerably increased by the publication of a magnificent work on the subject by the late Major-General Reid, who was at the time Governor of Bermuda. This author has arrived at exactly the same results as Mr. Redfield. They have come to these conclusions, entirely independently of my previous publications, as I have ascertained by personal communication with them. The works of Redfield and Reid, in addition to the further confirmation of the theory of rotatory motion, contain some very important observations, and the credit of establishing these is due to them alone. The most important of these discoveries, that of the motion of the cyclone within the torrid zone from SE. towards NW., before its change of path at the outside limit of the Trade-wind, is due to Mr. Redfield

*An Attempt to develope the Law of Storms by means of Facts arranged according to Place and Time, and hence to point out a Cause for the Variable Winds, with the view to practical use in Navigation.' Illustrated by Charts and Woodcuts. London, 1838, 8vo.—‘William Reid. The Progress of the Developement of the Law of Storms and of the Variable Winds, with the practical application of the subject to Navigation.' London, 1849, 8vo. 424 PP.

alone. In fact, the position of an observer in America is particularly favourable for the solution of this part of the problem, inasmuch as the hurricanes which sweep the coasts of the United States traverse, in the tropical portion of their course, the West Indian Islands, where their extraordinary character has given rise to the title, "West India Hurricanes.' In the case of the cyclones which are felt in Central Europe, it is rarely possible to trace the tropical portion of the course. This gives us a satisfactory proof that we shall avoid one-sided views of natural phenomena, the greater be the area over which our observations are extended. As, in the science of astronomy, the geographical latitude of an observatory exercises an important influence on the problems to be solved there,

in the science of meteorology, it is equally important to know the position of the station at which the solution of a problem has been undertaken. I shall now seek to compare the observations of Redfield and Reid with the theory of a rotatory motion. When I published my first papers on the winds, I was disposed to refer the Law of Gyration as well as the rotatory motion of storms to the mutual interference of two currents of air which alternately displaced each other in a lateral direction. A closer examination of the phenomena showed me that the Law of Gyration depended on more general principles, and that it was a simple and necessary consequence of the motion of the earth on its axis.* This generalization of the principle of Hadley's Trade-wind theory explained perfectly all the rules which had been ascertained for the non-periodical fluctuations of the instruments, and enabled us to foretell what they would be in the southern hemi

* Ueber den Einfluss der Drehung der Erde auf die Strömungen ihrer Atmosphäre. (On the Influence of the Rotation of the Earth on the Currents of its Atmosphere.) Pogg. Ann. xxxvi. p. 321, 1835.

sphere. It did not, however, explain the rotatory motion. of storms, to which idea I had originally allowed too great an extension, in the introduction to my first investigations (Pogg. Ann. xiii. p. 597), where I said, "That, in general, all storms are rotatory is an ascertained fact which will be confirmed by every seaman.' I fell into an error, which few of those who have investigated storms have escaped, which consisted in referring all the rotations of the vane which may be observed, to cyclonical movements. I was, accordingly, compelled to retain my former theoretical explanation, in my Meteorological Investigations' (1837), in which work I combined into a whole all the papers which I had published up to that time. The reason of this was, that the empirical facts had received entire confirmation, while their connexion with the principle of the general theory was by no means so evident. It was not until later* that I was able to supply this deficiency, by proving that a cyclonical movement was produced whenever the interposition of any obstacle interfered with the regular change in the direction of the wind, which is due to the rotation of the earth, and consequently interfered with the regular rotation of the vane at any station. From the investigations of Redfield and Reid we ascertain the following facts.

1. The storms which arise in the torrid zone, preserve the original direction of their progression (viz., from SE. to NW.) unchanged, as long as they continue within that zone; but as soon as they enter the temperate zone they turn and move from SW. to NE., almost at right angles to their former path. The corresponding storms in the southern hemisphere, which move from NE. to SW. between the tropics, undergo a change of direction, analo

* Reports of the Royal Academy of Sciences of Berlin,' 1840, p. 232. Pogg. Ann. lii. p. 1.

gous to that above described, on their entrance into the temperate zone, and move from NW. to SE.

2. The diameter of the cyclone, which increases very slowly within the torrid zone, increases to a remarkable extent as soon as the change of direction takes place.

In order to exhibit the course of the storms by means of numerous examples, we have added to these a chart constructed by Redfield, which gives the course of several of these storms (Chart IV.).

Two of these storms were confined to the torrid zone, and consequently preserved a straight course. One of these was on June 23, 1831, and advanced from Trinidad, by Tobago, Grenada, and the middle of Yucatan, to the neighbourhood of Vera Cruz.

The other was on August 12, 1835, and passed from Antigua, by Nevis, St. Thomas, St. Croix, Porto Rico, Hayti, and Matanzas in Cuba, to Texas.

Eight of the storms travelled outside the limits of the torrid zone, and their course was as follows:

That which devastated Barbadoes on the night of August 10, 1831, reached Porto Rico on the 12th, Aux Cayes and Santiago de Cuba on the 13th, Matanzas on the 14th, the Tortugas on the 15th, the Gulf of Mexico on the 16th, and lastly, Mobile, Pensacola, and New Orleans on the 17th; so that it traversed a space of 2,000 nautical miles in about 150 hours, and consequently travelled with a velocity of 13 miles per hour. Its direction between the tropics was N. 64° W.

The storm which commenced in the neighbourhood of Martinique on August 17, 1827, reached St. Martin and St. Thomas on the 18th, passed to the north-east of Hayti on the 19th, reached Turk Islands on the 20th, the Bahamas on the 21st and 22nd, the coast of Florida and South Carolina on the 23rd and 24th, Cape Hatteras on the 25th, Delaware on the 26th, Nantucket on the 27th, Sable Island and Porpoise bank on the 28th: and thus travelled 3,000 nautical miles in eleven days. Its direction within the tropics was N. 61° W.; and, on the other hand, in lat. 40°, N. 58° E.