158 TABLES OF DIFFERENCES OP MEAN TEMPERATURES. 



they have in common, as shown in diagrams C and D. In the first diagram the 

 curves for the northern stations appear single-crested, in the second the curve of 

 the middle latitude stations is double-crested and that of the Gulf stations exhibits 

 three or more elevations and depressions, all ill-defined. The marked feature of 

 the low^ latitude range is its great uniformity throughout the year. In the Arctic 

 regions, with the returning day, the range suddenly rises to its maximum in March; 

 in Canada and northern New York the range is greatest in July or about the time 

 of greatest heat; along the coast from Massachusetts to the District of Columbia 

 the range attains two maxima, one early in June the other late in September, with 

 an intermediate depression of range during the hottest season. As our observations 

 become more extended, other features in the march of the daily temperature will 

 undoubtedly make their appearance, and those already recognized will become 

 better defined. At San Francisco especially, it would be interesting to have a series 

 of hourly observations, extending at least over one year, this locality being other- 

 wise noted for anomalous temperature relations. According to Dr. Gibbons the 

 coldest and warmest periods of the day are not far from sunrise and noon, and by 

 taking the differences of the mean monthly temperatures at these times, as given 

 in the Smithsonian report for 1854, p. 231 and foil. For the years 1851 to 1854, 

 I obtain the following table of daily range of temperature at San Francisco. 



January .... 



12°. 



February . . . 



14.0 



March .... 



• • • iS-3 



April 



. . . 16.S 



May 



. . . 14.9 



June 



. . . 16.2 



July is°-7 



August 12.8 



September 14.9 



October 16. i 



November 13.7 



December 11.5 



These numbers are approximations only, yet they indicate a comparatively large 

 range, a minimum range in December and two maxima — one in spring, the other 

 in autumn. 



The modification which the daily fluctuation undergoes in the course of a year 

 can be advantageously brought out by a comparison of its value in December when 

 near the least, with its value in June when not far from its greatest development. 



The fluctuations observed at Van Rensselaer, Port Foulke, Port Kennedy, and 

 Boothia Felix were united into a mean, those at Thunder Bay Island, Toronto, 

 Mohawk, Amherst, and Philadelphia into another, and those at Fort Morgan, Key 

 West, and Galveston into a third ; these localities are designated, Arctic stations. 

 Temperate stations, and Gulf stations respectively. 



Before taking means, the record for Galveston, Texas, was made complete by 

 interpolation. 



