EEPORT OF THE SECRETARY. 51 



servations. Although the primary causes of the change of the 

 weather are, on the one hand, the alternating inclination of the sur- 

 face of the earth to the rays of the sun, by which its different parts 

 are unequally heated in summer and in winter, and, on the other, the 

 moisture which is elevated from the ocean in the warmer and precip- 

 itated upon the colder portions of the globe; yet the effects of these 

 are so modified by the revolution of the earth on its axis, the condi- 

 tion and character of the different portions of its surface, and the 

 topography of each country, that to strictly calculate the perturba- 

 tions or predict the results of the simple laws of atmospheric 

 equilibrium with that precision which is attainable in astronomy, will 

 probably ever transcend the sagacity of the wisest, even when as- 

 sisted by the highest mathematical analysis. But although such 

 precision cannot be looked for, approximations may still be obtained 

 of great importance in their practical bearing on the every-da}^ 

 business of life. 



The greater part of all the observations which have been recorded 

 until within a few years past has been without system or co-ordina- 

 tion. It is true that the peculiar climate of a given place may be 

 determined by a long series of isolated observations, but such obser- 

 vations, however long continued, or industriously and accurately made, 

 can give no adequate idea of the climate of a wide region, of the 

 progress of atmospheric changes, nor can they furnish an approxi- 

 mation to the general laws of the recurrence of phenomena. For 

 this purpiose a system of observation must be established over widely 

 extended regions within which simultaneous records are made and 

 periodically transmitted to a central position, where, by proper 

 reduction and discussion, such general conclusions may be reached as 

 the materials are capable of yielding. 



In discussing the records, the empirical method does not suffice. It 

 is necessary that a pn'oH assumptions should be provisionally adopted,, 

 not, however, at random, but chosen in strict accordance with well- 

 established physical principles, and that these be finally adopted, re- 

 jected, or modified, as they are found to agree or disagree with the 

 records. It is only by this method that the different causes which co- 

 operate in the production of a series of complex phenomena can be dis- 

 covered, as is illustrated in the history of astronomy, which, previous to 

 the investigations of Kepler, consisted of an unintelligible mass of 

 records of observations. But even with the application of the best 

 possible process of discussion, the labor necessary to be expended 

 on such large masses of figures, in order to deduxe simple results, is 



