xii EXPLANATIONS AND REMARKS. 



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

 rough approximations. Heights near tide-vpater may be considered to be reliable. 



Unless otherwise stated, the mean tabular values of the temperature, always 

 expressed in degrees of the Fahrenheit scale, refer to the observing Jiours noted, 

 and are consequently uncorrected for daily variation. In all cases where the 

 observing hours were variable or were changed during the series, the results were 

 referred either to those observing hours maintained for the longest period or to 

 those susceptible of the greater accuracy, or else all were corrected for daily fluctu- 

 ation. The means for correcting observed values, taken at stated epochs of the 

 day and for any month, were furnished by the discussion of the daily variation, but 

 the stations available for such discussions are comparatively so very few in number, 

 and are almost wanting for the western part of the United States, that but a small 

 portion of our results could be so corrected. If we had better and more com- 

 plete materials for daily variation, it would undoubtedly have been preferable to 

 correct all tabular results for this inequality, but in their absence it was deemed 

 advisable to attempt no more than to present the results in any one series for a 

 imiform set of hours of observation, correcting as stated in all cases where the 

 observer has changed his times of observation; this gives us the advantage of 

 eff'ecting hereafter a more satisfactory reduction to the mean of twenty-four hours 

 whenever we come into possession of new and, it is to be hoped, automatic registers. 



Respecting the results obtained under the University System of the State of New 

 York, the daily mean was directed^ to be found by adding to the morning observa- 

 tion twice the afternoon observation, and twice the evening observation to that of 

 next morning, and dividing their sum by six. This may be symbolically expressed 

 by i^Oj. -|- 3a bis -[- (Oa + l*") '''^ + Orj; the morning observation was to be taken a 

 little before sunrise. The means given in the table were made out in accordance 

 with this rule.^ 



With respect to the Smithsonian system of meteorological observations, the result 

 of the three hours 7 A. M. 2 and 9 P. M. was found to approximate less closely to 

 the true daily mean than the result obtained by adding twice the reading at 9 P. M. 

 to the readings at 7 A. M. and 2 P. M. and dividing this sum by four. The latter 

 rule was therefore adopted, and is symbolically indicated by | j 7„, -[" ^a + 9^ ^is \ . 

 In the column headed observing hours the symbols Or and 0^ stand for sunrise and 

 sunset; the affixes m. and a. to any given hour indicate morning and afternoon 

 respectively ; N. and Mdt. stand for noon and midnight ; M. and E. for morning 

 and evening ; Max. and Min. for mean from maximum and minimum readings ; • 



' P. B. Hough, p. iv of the introduction to the results of meteorological observations made in 

 obedience to instructions from the Regents of the University at sundry Academies in the State of 

 New York, Albany, 1855. 



2 It should also be mentioned that for these Academy stations the monthly means are made up 

 from the half-monthly means, there is therefore a slight inconsistency in the results for the months 

 having an odd number of days (the first 15 days having been united into a mean for all months, 

 excepting February). The October mean is most affected, less so May and March ; the amount 

 generally less than 0°.l is small enough to be neglected. 



