EXPLANATION OF TABLE (A). 



The headings of the cliflf'erent columns of the table, in general, sufficiently 

 explain their contents, but a few additional remarks may be made, principally 

 relating to the positions of the stations and to the continuity and combination of 

 the records. 



The geographical positions are given to the nearest minute of arc, as far as 

 known, the longitudes are counted as usual, west from Greenwich. The positions 

 which became known through the operations of the United States Coast and Greodetic 

 Survey are reliable, as well as those given upon the authorities of the United States 

 Lake Survey, directors of astronomical observatories, and in general, all those posi- 

 tions which have been determined by direct astronomical observations and those 

 connected with the General Land Office. Positions given on the authority of the 

 observer, and these are by far the most numerous, are less trustworthy, since most of 

 them were taken from Sta,te or county maps having no adequate astronomical basis. 

 The results for longitude depending on the electric telegraph are of such recent date 

 that but few maps have as yet incorporated them, but longitudes of stations in their 

 vicinity have been improved by referring to them differentially. Althoiigh no pains 

 have been spared to render these geographical positions as trustworthy as possible, 

 they are, in general, when taken from maps, evidently affected in the given latitudes 

 with a probable uncertainty of from 3' to 5', and in the given longitudes with a 

 probable uncertainty of from 5' to 8'. Fortunately for the immediate wants of the 

 discussion of the rain-fall a moderate approximation to the true position suffices. 



The elevations of the stations above the ocean are given on the authority of the 

 observers, unless corrected from other, and apparently more reliable sources. These 

 data are known in many instances to be quite defective, as but few depend upon direct 

 hypsometric measures, or on measured differences of level from known railroad or 

 canal levels ; those depending on barometric observations can only be regarded as 

 rough approximations. Heights near tide-water may be considered reliable. 



The tabular means given for any one month are derived from all the monthly 

 amounts in the different years, and, unless the whole series of observations is 

 unbroken, may differ in weight, since they may be derived from a greater or less 

 number of years ; and on account of discontinuity in the observations the numbers 



(xix) 



