PRESENT STATE OF OUR KNOWLEDGE OF THE UPPER ATMOSPHERE. 111 
The following table gives the total number of days in each month on 
which the observations, used for the calculation, were made :— 
— Jan. | Feb. Mar. Apr. May June July | Aug. Sept. Oct. Nov. Dec. 
| = | | mY: | a ~ ==} 
Surface | 155 | 141 | 155 | 150 | 155 | 150 | 155 | 155 | 150 | 155 | 150 | 155 
500 | 155 | 139 | 154 | 150 | 155 | 150 | 155 | 155 | 150 155 | 150 | 154 
1,000 | 146 | 133 | 150 | 146 | 154 | 150 | 151 | 152 | 148 | 152 | 147, 150 
1,600 | 115) 98) 110 | 122 | 132 | 128 | 134 | 132 | 129 | 134 | 124 120 
2,000 | 79 73) 76) 89| 97100} 91/111] 109, 112 94 83 
2,500 | 49) 48| 48) 57| 67| 71| 66| 85) 85} 84| 77] 60 
3,000 | 24| 29 | 39} 55 | 50| 55 | 65| 60| 55/ 36 
made at 40 m. They have been reduced to 122 m. by taking for each 
of the 27 months the mean gradient observed in that month between 
40 m. and 500 m. 
The outstanding features of the results are a steady decrease in the 
amplitude and in the phase of the whole year term. The minimum tem- 
perature occurs near January 15 at the surface, while at 2,000 m. it does 
not occur until February 1. 
The results for 3,000 m. are unduly influenced by the fact that the 
observations at that height in January 1907 were made during a very cold 
period, and probably diminished the mean value for that month below its 
true value. The effect would be to increase the phase for that height. 
The amplitude at 2°5 km. is only 3 of its surface value, but it does 
not appear to decrease above this height. 
Hann! gives the following values for the amplitude of the yearly 
variation on mountain peaks and well-exposed places :— 
Sonnblick 
Place Altstitten | Trogen | Gibris | Rigikulm  Siintis 
| Height 460 m. 880m. | 1,250 m. 1,790 m. 2,500 m. | 3,100 m. 
| Range | 19°4 17°11 | LD Dre 14°4 / 14°-1 / 14°'5 C. 
These values show the same general decrease up to 2°5 km. as those 
found for the free atmosphere, but they are in all cases 2°-3° C. greater 
than for the free atmosphere. 
The four components have been given to show the comparative regu- 
larity of the third both in amplitude and phase, compared with the second 
and fourth, the latter of which appears to be merely the result of accidental 
irregularities. 
The result appears interesting and worth investigation, because there 
is a similar period in the velocity of the centres of anticyclones. 
The following expression for the annual variation of this velocity in 
m.p.s. in America has been deduced from results given by Herrmann :” 
V=11:441-30 sin (nt + 68°) + 0°12 sin (2nt + 138°) 
+ 0°31 sin (8nt +18°)+0°09 sin (4nt + 78°). 
Here the third component has an amplitude more than double that of 
the second, and treble that of the fourth, and its phase agrees very nearly 
with the corresponding term in the variation of temperature. This 
implies that when anticyclones are moving rapidly in America, the mean 
temperature over Berlin is above the normal value. 
1 Lehrbuch, p. 105. 2 Monthly Weather Review, April 1907. 
