Daniell on Meteorological Observations. 99 



The form in which I have extracted the observations from my 

 register, the times of observing, and the division of the periods, 

 are the first things that require remark. I have constructed 

 three tables, which will be found to contain the greater part of 

 the results necessary for our purpose. The first is similar to this 

 one for the preceding year, published in the tenth volume of the 

 Journal, p. 1 3 1 . It contains the means of the different periods 

 in large figures, and extremes in smaller, shewing the range of 

 the several instruments in the respective intervals. The division 

 of the year is into halves, quarters, and half-quarters, which I 

 have found to be convenient and comprehensive. The grouping 

 of the quarters represents Autumn, Winter, Spring and Summer; 

 and agrees with that adopted by Mr. Howard, except that he 

 commences each quarter with the second, instead of the first, 

 week of the month. The reasons which he has assigned for de- 

 parture from the customary division, are conclusive; mine, which 

 was made before I had any knowledge of his work, approaches 

 sufficiently near to excuse me from the laborious drudgery (not 

 easily appreciated but by those who have undertaken such a 

 task,) of recalculating the results. The commencing with Au- 

 tumn was accidental. It was with this season that I began first 

 to observe regularly, and of course at this season the first year 

 ended. If this irregularity should be thought inconvenient, 

 from the nature of the division, it may be easily rectified. The 

 hours of the day at which the observations are made, are a con- 

 sideration of greater moment ; they are, nine in the morning, five 

 in the afternoon, and eleven at night; very rarely, indeed, vary- 

 ing an hour from the appointed time. The maximum and mini- 

 mum of temperature, by a register thermometer, may be added 

 to these. If the selection of three periods in the day had been 

 guided merely by meteorological considerations, I do not mean 

 to say that these would have been the most proper ; I think not: 

 but it is not a question that is easily determined. They are not, 

 however, destitute of peculiar advantages. The first object of 

 repeated meteorological observations is to obtain accurate mean 

 results. The most proper method of effecting this, even for the 

 temperature, where we are assisted by an instrument which 

 H 2 



