TRANSACTIONS OP SUCTION A. 547 



connection between radiation and the advective region. Of course it is known 

 experimentally that the absorption of the atmosphere for different wave-lengths does 

 vary very much, but it seemed possible that an average value would be sufficiently 

 accurate for practical applications. This proves now not to be the case, and calcula- 

 tions based on the assumption that it is permissible may be subject to large corrections. 



There must be a considerable portion of the spectrum for which the atmosphere 

 is nearly transparent, and it is this fact which makes possible a lower temperature 

 in the advective region near the equator. If this were not the case, the temperature 

 at the earth's surface and in the convective region would rise until the balance with 

 the absorbed solar radiation was established, which could only be after the tempera- 

 ture of the advective region had risen above its value for temperate latitudes. 



It has been suggested that the cause of the variation of T c with latitude is a greater 

 proportion of ozone in temperate than in equatorial regions in the upper atmosphere. 

 This is in itself insufficient, because T c depends not on the absorbing power of the 

 advective region but upon the 'effective' temperature of the incident radiation. 

 If the latter i3 unchanged, an increase in absorption in the advective region is accom- 

 panied by an increase in radiation from it and there is no increase in T c . The dis- 

 tribution of ozone may modify the conditions in proportion to its relative importance 

 as an absorbing constituent, compared with water vapour. It could affect T c in the 

 manner observed only (1) if water vapour were transparent to ozone radiation; 



(2) if the intermediate atmospheric layers were nearly transparent to ozone radiation ; 



(3) if ozone predominated in the advective region in temperate latitudes and water 

 vapour in the same region near the equator. 



If, as is generally accepted, the vertical motion near the equator is upwards only, 

 then for all heights above the condensation level, 2 km. say, the air must be saturated 

 with water vapour. Thus if the vertical temperature gradient is the same as for a 

 place in the temperature zone in latitude <p, the amount of water vapour between the 

 isothermal surfaces at temperatures t , t c must be greater near the equator than 

 in latitude <j>, t being the surface temperature and t c that of the advective region in 

 latitude (p. But the outward radiation across the surface, t ', is greater at the equator 

 than in latitude <p, owing to the higher temperature of the air beneath the surface. 

 There must then be more of this radiation absorbed unless the value of T c at the 

 equator is to exceed t c ; and this greater absorption is in accordance with the deduction 

 regarding the water vapour made above. We may say therefore that the variation 

 in the distribution of water vapour, combined with the presence of spectral regions 

 for which the atmosphere is nearly transparent, is alone sufficient to explain the 

 variation of T c and H„ with latitude. 



7. A Sensitive Bifilar Seismograph ivith some Records. 

 By Professor F. G. Baily, M.A. 



WEDNESDAY, SEPTEMBER 7. 

 The following Papers and Reports were read :— 

 1. Stars as Furnaces. By Sir Norman Lockyer, K.C.B., F.R.S. 



2. On the Evolutions of a Vortex. By Professor W. M. Hicks, F.R.S. 



3. Report of the Seismological Committee. — See Reports, p. 44. 



4. On the Rate of Propagation of Magnetic Disturbances. 

 By C. Chree, D.Sc, F.R.S. 



This paper was mainly devoted to a recent view of Dr. L. A. Bauer, head 

 of the Department of Terrestrial Magnetism in the Carnegie Institution of 



