TRANSACTIONS OF THE SECTIONS. S3 



general truth of the principle of prediction is, first, that correctness in the result 

 depends on the correctness of all the elements of the formula, but their values are 

 only mean values and cannot be relied on in individual cases ; secondly, any error 

 m the theoretical expectation of the value of «j^^ is, on the whole, doubled in the 

 prediction of Ve, because the difference between what was expected of rj^^, and 

 what was fulfilled in Wj, is heaped on to v,., which has, therefore, to bear the entire 

 error of expectation of i)j+^. It was concluded from this, and from other previous 

 deductions from some years of Dublin observations, to which reference was made, 

 that the fame of the barometer is due to its success in predicting a type of storm 

 very rarely met with in the British Isles, but frequently in liuvricane-latitudes, 

 when the fall of the mercury far outstrips the increasing severity of the weather. 

 In ordmary gales, and much more in ordinary weather, the author considered the 

 barometer to be useless as a guide when consulted without a knowledge of what is 

 occurring at adjacent stations— in short, without such information as is supplied bv 

 the 'Daily Weather Eeport.' 



On the Temperature of the Air at A feet, 22 feet, and 50 feet above the 

 Ground. By James Glaisher, F.R.S., F.R.A.S. 



}\ *ii^ '"^Poi't to the British Association, for 1866, on the experiments made by means 

 of balloon, I stated that the law of decrease of temperature with increase of eleva- 

 tion was variable throughout the day, and variable m different seasons of the year, 

 that at about sunset the temperature was nearly the same up to 2000 feet, and that at 

 night (from the only two night-ascents) the temperature of the air increased from 

 the earth upwards. From this it was evident that, instead of only a few ascents 

 being necessary, a much larger number were required than it was possible for me to 

 ^*^ i „f o^'t'i^'^tely, in the second year of the balloon experiments, I planted at the 

 ■^"^o f r'^*'''^' G^reenwich, a dry- and a wet-bulb thermometer at the height 



+ vT ^ '^Dove the soil, readings of which have been taken daily since that time, 

 Bi, the hours of 9 a.m.. Noon, 3 p.m. and 9 p.m. Sometimes readings at the higher 

 pomt were above those at 4 feet from the gi-ound ; but no particular value was 

 attached to tins fact, until, on the observations made in M. Giffard's captive balloon 

 being reduced, the results proved that the decrease of temperature with increase of 

 elevation had a diurnal ran^e, and was different at different hours of the day, tlie 

 changes being greatest at about midday, and least at or about sunset (see lieport 

 for 1809), whilst sensible changes occurred within 30 feet of the earth. In conse- 

 quence, the observations made at the height of 22 feet were reduced by takiu"- the 

 differences between the readings of the two thermometers, and affixing the'^sigu 

 plus (-I-) to that difference when the temperature was higher at the higher eleva- 

 tion, and the sign minus (-) when vice versa. All the observations made in the 

 years 1867-70 were treated in this way, 



_ By selecting the gi-eatest number with a + sign, and the greatest with a - sign 

 in each month, it was found that in the winter months the temperature at 22 feet 

 height ranged from 2 to 4 degrees above, and from 1 to 2 degrees below, that at 4 

 feet, and in the summer months from 4 to 5 degrees above to 6 or deo-rees below 

 that at 4 feet height, as will be seen by the following Table :— ° 



1867-70, .January ti 1~9 " 



February 2o 15 



March 19 49 



April 4-2 5-,5 



May 4-7 61 



June 41 60 



July 4-3 5-5 



August 5-2 6-1 



1867-69, September 4-0 4-8 



October 3-5 33 



November 4-8 2-6 



December . . . ,. 2-3 1-2 



1870. 3 



