52 On the Methods of obtaining Mean Temperatures. 



mation to the truth, or obtain harmonious results from elements 

 the relations of which are so variable as those of the temperatures 

 of particular hours. 



The method which I have adopted, and had in use about three 

 years, I have found experimentally to give results, for this cli- 

 mate, almost in precise accordance with those derived from calcu- 

 lations based upon hourly observations. I have constantly kept 

 daily observations at 6 and 7 a. m., 12 m., 2 p. m., sunset, and 9 

 p. m., so that I have been able to give my observation any of the 

 common tabular forms, and have always obtained the general re- 

 sults according to each of the methods, and the results have veri- 

 fied my previous conclusions by varying from each other as fol- 

 lows : the mean annual temperature of this place, by my method, 

 is 64.24° Fahr. ; as derived from the sunrise, 2 p. m. and sunset 

 observations. 66.30°; and during the same period, the mean ob- 

 tained from the 7 a. m., 2 p. m. and 9 p. m. observations is 65.62°. 

 In this way I have compared results by each method with each 

 other, and with those obtained from hourly observations. I have 

 found that the following method gives for this place results more 

 accurate than any other I have been able to devise. The mean 

 temperature of the diurnal portion of the twenty four hours is 

 derived as the mean of the highest and lowest temperatures of 

 that portion, that is, of the sunrise observation and that of 2 p. m. 

 The nocturnal mean will in like manner result as the mean of 

 the highest and lowest temperatures of that portion, that is, of 

 the sunset and sunrise observations. The average of these two 

 means will give the mean for the twenty four hours. By this 

 method, the sunrise temperature being the lowest for the twenty 

 four hours, will belong to both the diurnal and nocturnal portions, 

 and will enter twice into the calculations, while the other two 

 observations will represent the maxima of the portions of the 

 twenty four hours to which they respectively belong. 



If it is as I suspect then, that the results of observations and 

 calculations are, particularly in hot countries, frequently too high, 

 a little attention to the subjects above hinted at may show that 

 there is really no difference between the mean temperature of cli- 

 mates and of water derived from such depths as not to be affected 

 by the change of seasons. In my tables the most exact agree- 

 ment is shown between these temperatures at this place. The 

 same appears to be the case in the island of Cuba, notwithstand- 



