28 WRITINGS OF JOSEPH HENRY. [1855- 



the temperature of a given place every hour in the day, add 

 all the observations into one sum for a year, and divide by 

 the number of hours in a year, v^^e shall get the mean 

 annual temperature. By this method of observation we 

 shall ascertain the warmest and the coolest hours of each day, 

 and by repeating the same process for a number of years, 

 we shall learn the temperature of each hour, eliminated 

 from all perturbations, and in this way arrive at truths which 

 could not be obtained by any other means. If we examine 

 the individual records we shall find the warmest time to 

 recur, on different days, at different hours. We know how- 

 ever that if there were no perturbing influences the warmest 

 period of the day would be that at which the heat received 

 from the sun is just equal to the cooling of the earth by 

 radiation into space. At every instant from the rising of 

 the sun previous to this the earth would be receiving more 

 heat than it gave off, and hence the temperature would con- 

 stantly increase until the heating and cooling were equal. 

 After this the earth would give off more heat than it would 

 receive, and the temperature would begin to descend. On 

 individual days however, clouds may intervene, or winds of 

 varying temperatures and velocities may prevail, so as to 

 change the hour of maximum heat ; but as these are not 

 periodical or governed by recurring laws, the probability 

 is that they will act in opposite directions ; that is, on some 

 days hasten the maximum period, and on other days retard 

 it, and thus in the course of a year, or several years, neutral- 

 ize each other. The method of averages therefore enables 

 us to separate the effects produced by irregular variations 

 from those which are due to permanent causes. The latter 

 are called periodic variations, while to the former has been 

 given the name of non-periodic. By continuing the obser- 

 vations for a number of years, in ascertaining the tempera- 

 ture at a given place, we find by the method we have ex- 

 plained a result from which that of the individual years 

 will oscillate on either side within certain limits, while for 

 two separate decades of years it will scarcely differ at all ; 

 and this is the mean temperature of the place. The same 



