134 • CLIMATIC CONDITIONS 



This law is also connected with the changes of direction experienced by the 

 general upper current in the tropical regions. In fact, during the whole year, in 

 Habana, if we except the hurricane season, the upper currents come from the west. 

 In the first half of June and particularly in the second half of October, the cur- 

 rents of the cirrus clouds incline to the south and southwest which is precisely 

 where the cyclones come from at that time of the year, for they reach us as they 

 are about to recurve, or just after recurving. From the end of June to the begin- 

 ning of October, the upper current sets from the east, and this is (without any 

 exception whatever in all the observations that I have made so far) the only time 

 of the year when the cirrus clouds come from that quarter. This fact, when taken 

 in connection with the tracks of the cyclones, is very significant: for, precisely at 

 that time, if a cyclone advances toward Habana, it has necessarily to come from 

 the east, since it must recurve to the north of the tropic and consequently we must 

 receive it in the first branch of its track. On the contrary, excepting in the season 

 just mentioned, in all the rest of the year, the cyclones that pass by Habana, or 

 in its vicinity, are all more or less from the west and never from the east. In 

 fact the cyclones of the end of October all come from the third quadrant, having 

 previously recurved. Those of November, December and January all pass to the 

 north of Habana in a northeast direction, as cyclones moving along the second 

 branch of their track. In all the rest of the year they pass to the north of and at a 

 greater or less distance from Habana more or less in the direction mentioned above. 

 The facts that I have brought forward seem to indicate that the cyclones are 

 directed along their tracks by the upper currents, which to my mind seems highly 

 probable. 



The law of recurving has also intimate connection with the greater or less 

 cyclonic activity in the West Indian seas during the different months. The maxi- 

 mum and minimum latitudes reached by the recurving point correspond respectively 

 to the maximum and minimum of cyclonic activity. 



In the second fortnight of August the hurricanes are, in general, more num- 

 erous and more violent; they move along their tracks with greater velocity, reach 

 greater altitudes, and the parabola which they describe is very wide, so that what- 

 ever be the force that, projectile-like, impels the cyclone, its reach and amplitude 

 are greater, and so consequently must be its impulsive energy. Besides, if the 

 general currents direct cyclones in their courses, this fact would denote that at 

 this time of the year these currents attain their maximum activity and reach 

 higher latitudes. The second fortnight of August and the beginning of September 

 are moreover the epochs for simultaneous or twin cyclones; so much so that in 

 1886, during the last decade of August, there were at one time four cyclones around 

 Habana: One in the island to the east-southeast, one to the northeast, and two 

 more in the Gulf of Mexico. Excepting at this season I only know of three cases 

 of simultaneous cyclones near Habana, one occurring in September and two in 

 October. 



In July and September the cyclones are less numerous, generally less violent, 

 they move along their tracks with less velocity, are more inclined to the west and 

 describe narrower parabolas. 



Finally, in June so few cyclones are observed that they are scarcely sufficient 

 to establish a law. In October they are somewhat more numerous, but still few: 

 some are quite intense; they move with but little velocity along the first branch 

 of their track and while recurving. 



The relative position of the seas and continents appears to have some influ- 

 ence on the recurving of hurricanes, for any one may observe that a great number 

 of the cyclones of August recurve in the Gulf of Charleston without extending to 



