TRANSACTIONS OF SECTION A, 4.67 
Under certain conditions (%) should be a minimum when we have hit on a real 
periodicity. The conditions are made obvious by supposing the periodicity perfect 
and the observations without error, when by choosing Ho properly, we can make 
all the residuals zero, and hence (4) will be zero. But if Ig. is not properly chosen, 
(b) will not be zero. Its greatest value will be mz? when all the observed maxima 
fall just midway between consecutive tabulated maxima. Hence the need for 
forming also the algebraic mean (@) which in the former case will be zero and in 
the latter wil] be mm. If (a) comes out sensibly different from zero, we must alter 
Ep (in a manner which need not here be explained) until (@) is small enough. 
When there is no periodicity near 2” the residuals will have indifferently any 
value from 0 to m, and (if 2 is an integral number of days) the sum of the squares 
of the residuals will approximate to : mn(2n +1) or say 5 mn”, When we get 
serious differences from this, it is worth while to investigate further. 
The®method is being applied to the case of U Geminorum, and some interesting 
results obtained. 
7. An Analytical Study of the Meteorological Observations made at the 
Glossop Moor Kite Station during the Session 1906-1907. By Man- 
caret Wuits, T. V. Prine, and J. E. Peraven, I’.R.S. 
In Part I. the change of temperature with height was considered. The rate of 
fall, or temperature gradient, is measured in degrees centigrade per 100 metres. 
The average gradient for the present year was given for the various English 
stations at each successive 500 metres above sea-level. The values for Glossop 
Moor are 1:24° C., 091° C., 0°61° C., and 0:33° C. per 100 metres at heights of 
400, 750, 1250, and 1750 metres, These results were compared with the obser- 
vations made in previous years in America at Blue Hill, and in Europe at Berlin, 
Oxshott, and Crinan. 
The temperature gradient varies: (1) with the wind direction, and is at all 
levels a maximum for a N.W. wind; (2) with the amount of clouds, and is at all 
levels 2 maximum on clear and fine days; (3) with the wind velocity, and is a 
maximum for low velocities. 
Diagrams illustrating these facts were shown. 
Part II. dealt with the question of wind velocity. The average velocity had 
been calculated for all stations at ground level, 500, 1000, and 1500 metres. For 
Glossop Moor the averages are 135, 26, 30, and 33 miles per hour respectively at 
these heights. Except at the ground level, where it is influenced by local condi- 
tions, the average wind velocity at each height differs but little from station to 
station. 
Comparing the velocities at various heights it is found that above a height of 
500 metres the average increase is 0°7 mile per hour per 100 metres difference of 
level. 
Curves showing the connection between the wind velocity and barometric 
gradient were exhibited. 
The effects of the absolute barometric height, of the season of the year, and of 
the wind direction were also considered. 
With regard to the alteration of wind direction with height, the results indicate 
the usual slight rotation in a clockwise direction, 7.e., a west wind tends to become 
more northerly, &c. 
A number of subsidiary questions, such as the conditions prevailing during tem- 
perature inversions and the occurrence of vertical air-currents were discussed. 
8, Siath Report on the Investigation of the Upper Atmosphere by means of 
Kites. See Reports, p. 99. 
HH 2 
