174 i'i.siii^G GiiouNDS OF ^■oIr^ll ameiiica. 



The isotherms of 75° ami 80° also bend abruptly between South Caroliua and the Florida 

 Reefs, that of 75° occurring at Fowey Rocks about sixty days in advance of Martin's Industry, 

 and th;it of 80° about twenty-six days in advance. 



RELATIONS OF THE AIR AND SURFACE ISOTUERMS. 



As stated in the explanation of Chart No. 32, there appears to be no constant relation between 

 the air and surface isotherms at any of the light-house stations. During the periods of both ris- 

 ing and falling temperature, the air temperature of any degree, as a rule, precedes the surface tem- 

 perature of the same degree, but the length of time intervening may vary from two or three days 

 to over a nionth. 



THE WIND RECORDS. 



Very complete records respecting the direction of the winds were kept at all the stations 

 excepting Thatcher's Island, the observations being taken twice each day, at the same time as 

 the temperatures. The wind records are even more complete thau are those for surface tem- 

 perature, having fewer breaks at any of the stations; and as temperature is greatly influenced by 

 the wind, it has been thought advisable to present in this connection a tabulation of the observa- 

 tions made. In this table (p. 170) the means of tive years' observations for each month are arranged 

 according to quadrants of the compass, beginning with the northeast quadrant. Northerly winds 

 have been included in the same quadrant with northwesterly winds, to which they appear to be 

 most nearly related in their eft'ects upon temperature. The extent of the several quadrants is, tliere- 

 fore, as follows : Northeast, from NNE. to E., inclusive ; southeast from ESB. to S., inclusive ; south- 

 irest from SSW. to W., inclusive; northvcst from WNW. to K, inclusive. 



The general arrangement of the table scarcely requires an explanation. The data respecting 

 the several stations are classitied by quadrants under each month, and that for each station extends 

 from left to right across the table. In the columns of figures, each one-tenth of a unit represents 

 one observation, ami each unit ten observations, extendiug through five years, the latter being 

 equivalent to one day's ob.servations for Qvc years. The time ratio for eacli qnndnmt is, therefore, 

 represented by days and fractious of a day. For example, at the Tortugas station in January, 

 which has thirty-one days, the northeast quadrant shows a mean record of 11.9 days ; the ;*outheast 

 of 7.2 days; the southwest of 2 days, and the northwest of ~j.'.) days, making a total of thirty 

 days. The discrepancy of one day results from calms and variable winds. The terms used by 

 the observers to express the velocity of the winds not being nnifonn at the diflerent stations, it 

 has been impossible to tabulate them. 



At the Florida Reef stations northeasterly and southeasterly winds prevailed tliroughout the 

 entire year, the records for the northwest and southwest quadrants being relatively small dnring 

 nearly every month. Northeasterly winds predominated at the Tortugas during every month 

 excepting June, in which the prevailing winds were southeasterly, and they afford an exceedingly 

 high record during October, November, and December. The records for Carysfort Reef and Fowey 

 Rocks, which are situated only about 23 miles apart, present some, though not considerable difter- 

 ences. At the former station northeasterly winds prevailed during February, May, September, 

 October, and November; at the latter, during September, October, November, and December; 

 the prevailing winds for the same stations during the other months being southeasterly. North- 

 westerly winds very rarely occur at the Florida Reel's during the summer months, and in only one 

 instance did they exceed a mean of eight days during the winter months, that being at Carysfort 

 Reef in December. ContiTiuons winds from tlie north and northwest cause a marked reduction in 



