18C6] 263 t*^^-'''®- 



crease in temperature, under a vertical sun at noon, of 180° X .00067 

 -r.367r=0°.82S6, and a like decrease at midnight. This amount 

 of disturbance is entirely independent of the heat-waves which are 

 radiated from the sun, and it would, therefore, remain unchanged, 

 even if the sun were darkened or an adiathermanous screen were 

 interposed. 



If the hypothesis of a constant ratio between the solar attraction 

 and the solar radiation is correct, the average temperature at dif- 

 ferent hours of the day, like the weight-disturbance, should vary 

 nearly as the co.sine of the hour-angle. In Table I, the stations are 

 arranged, in each quarter of the globe, in the order of latitude, and 

 one hour is allowed for the lagging which is due to atmospheric 

 inertia,* the differences being taken between l** and 13'', 2'" and 12'', 

 &c., instead of between 0'' and 12'', l** and ll*", &c. The striking 

 theoretical correspondence in the average of the joint observations 

 at Gottingen and Halle, is specially interesting on account of the 

 position of the two stations, south of the Hartz jMountains, and near 

 the opposite extremities of the range. The New Haven ratios, 

 which are deduced from the discussion, by Profs. Loomis and New- 

 ton, of 8G years' observations, approximate more nearly to the cosine 

 ratios than those which are based on a shorter series of observations 

 at most other American stations. The Amherst ratios, however, 

 representing a single year's observations under the direction of Prof. 

 Snell, are very noticeable for their close coincidence with theory. 

 At nearly all the stations the abnormal differences are greatest when 

 the sun is near the horizon, which may be partly owing to the 

 greater extent of air traversed by the solar rays. 



The force of gravity at the poles is P — (§gg)^= y|^ greater than 

 at the equator. The amount of heat required to expand air j|„ of 

 its volume is 180° X ^ i^ -~ .367 = 3°.27. According to Johnston's 

 Phy.'^ical Atlas, the mean minimum temperature of the globe is 

 1°.75, the maximum 85°. 3, which indicates an average difference 

 of 83". 55 between the polar and equatorial temperatures. Daguin 

 (Traits de Physique, v. 2, p. 115) fixes the range at 46°. 2 C., or 

 83°.16 F. 



Table II exhibits the average daily range, at all the stations of 

 hourly observation which are embraced in the third edition of 

 Guyot's Tables. 



* The approximate average lagging, as shown by a great variety of barometric 

 observations, is about 60 minutes. 



VOL. X. — 2K 



