On the Coiirses of Hwricanes. 203 



he thinks he finds evidence of a northerly or variable course, in 

 the Barbadoes hurricane of 26th July, 1837 ; and also of a varia- 

 ble or northeastern course, in the Antigua hurricane of August 2d, 

 1837 ; or at least of a lateral movement or oscillation, in the course 

 of these hurricanes.* He also suggests that hurricanes may alter- 

 nately dilate and contract during their course. 



The inquiries of GoL Reid, aided by his excellent charts and 

 delineations, appear to have done much towards settling these 

 questions, so far at least as relates to the particular storms referred 

 to by your correspondent ,• and it may be observed, that in tracing 

 the course or track of a storm, we must be governed by its regular 

 geographical developments or progress, rather than by any induc- 

 tions from the directions and changes of the wind at a given place, 

 grounded on the known whirlwind character of these storms. It 

 is true that these inductions, if carefully made, will commonly 

 harmonize, with wonderful accuracy, with the actual course or 

 path of the storm ; but there are various sources of error, which 

 may at times mislead us in our deductions, when made from a lim- 

 ited number of observations ; some of which may here be curso- 

 rily noticed. 



1. The reported observations are not always correct as to the 

 point of compass from which the wind blows, and the changes 

 which it exhibits, during the storm. This is not unfrequently the 

 case with the reports of unpractised observers ; or with observa- 

 tions made in the gloom of night ; or in the tumultuous crisis of 

 the hurricane, when the whole energies of the seaman are directed 

 to his more immediate duties, and the preservation of his ship ; 

 and Avhen in the darkness and turmoil of the storm, the swinging 

 of the ship may sometimes be mistaken and reported for the irreg- 

 ular veering of the wind. Verbal or typographical errors, will also 

 have sometimes occurred in the reports which are under conside- 

 ration ; and in some localities, an important difference between 

 the magnetic and the true points of direction, is frequently con- 

 founded, or unnoticed. 



2. The inductions in question are usually made on the theory 

 of an exact circle in the course of the winds, which in large 

 storms, and for practical purposes, is, in most cases, sufficiently 

 accm-ate. But it sometimes happens, that the higher portions of 

 the storm, overrun the inferior portions, and reach the surface m' 



* See Nautical Magazine for January, 183S, pp. 35 — 40. 



