236 



Meteorology of New Haven. 



ature of the year and the highest isothermal ; the two variables being 

 the hour of the day, and the day of the year. The hours of the day 

 are indicated by equidistant vertical lines which are numbered at 

 both top and bottom ; Avhile the beginning and middle of each month 

 are indicated by the horizontal lines. The smallest curve on this 

 map shows at what hours and upon what months the mean tempera- 

 ture of New Haven is 78°. The next curve shows at what hours, 

 and upon what months the mean temperature is 76°, and so on. 

 These curves show an unexpected degree of symmetry, which is 

 probably due to the length of the series of observations, by which 

 the influence of irregular causes is mostly eliminated. 



Plate III exhibits the isothermals which are comprised between 

 the mean temperature of the year, and the lotcest isothermal. The 

 smallest curve on this map shows at what hours, and upon what 

 months the mean temperature of New Haven is 22°. The next curve 

 shows at what hours and upon Avhat months the mean temperature is 

 24°, and so on. These curves are less symmetrical than the preced- 

 ing, being crowded much more closely together in the morning than 

 in the evening hours; indicating that in winter the temperature 

 changes much more rapidly in the morning than in the evening. An 

 effect of the same kind is noticeable in summer, but is somewhat less 

 conspicuous than in winter. 



In order to determine whether the mean temperature of New Haven 

 has changed since the time of the earliest recorded observations, we 

 have divided the entire series of observations into two groups, the 

 first embracing the observations down to 1820, forming a series of 41 

 years; the second embracing the observations since 1820, forming a 

 series of 45 years. The mean temperature of each month according 

 to the two series of observations is shown in the following table ; the 

 mean of the observations being reduced to the true mean temperature 

 by applying the correction from the table on page 232. 



Months. 



January, 



February, 



March, 



April, 



May, 



June, 



These differences are generally small, with repeated changes of 

 sio-n; which seems to indicate that they are mainly due to those 

 irregular causes which render the mean temperature of a given 



