DISTRIBUTION IN THE ATMOSPHERE OVER ENGLAND. 269 



will be air that has recently risen and conversely, for the air changes its temperature 

 slowly under the influence of radiation, but instantly under the influence of change 

 of pressure. Also, the magnitude of the temperature departure from the mean 

 should be an index of the intensity of the ascending or descending motion. 



For convection to occur it would seem necessary that the lower layer should be 

 potentially at least as warm as the air above it, that is to say, that if the lower layer 

 rises it shall, after cooling adiabatically, still be as warm as the air on its fresh 

 level. For unsaturated air the adiabatic gradient amounts to 10 C. per kilometre, 

 and this is the practical limit to the gradients met with in the individual ascents, but 

 for saturated air at ordinary temperatures the value is much less, the exact value 

 being dependent on the temperature. A fair average for the year in the first few 

 kilometres in England is about 6 C. per kilometre, and this is somewhat over the 

 mean gradient up to 3 km. As the height increases so also does the adiabatic 

 gradient, because the colder air contains less moisture to be condensed, and con- 

 sequently the latent heat set free is less. The observed gradient also increases, but 

 never becomes as great as the adiabatic gradient for saturated air at the temperature 

 that usually prevails at the corresponding height. The lower layers are potentially 

 the colder by about 1 C. per kilometre, even for saturated air, and convective currents 

 should not occur. One would have expected to find the adiabatic gradient up to the 

 point to which convection extends, or at least up to the limit of the clouds, 7 or 8 km. 



Probably owing to the general disturbance caused by winds and the convection 

 currents of the lower strata, a certain amount of forced mixing occurs, and since the 

 lower layers are potentially colder than the upper, the mixing will cool the upper 

 strata. Mixing means the existence of vertical currents, and therefore raises the 

 height of the isothermal region, which begins at the point where vertical currents 

 cease. Could we by any means thoroughly mix up the vertical strata, just after the 

 mixing we should have the same potential temperature throughout, that is to say, 

 the surface layers would be very much warmer and the upper layers very much colder 

 than they are now, there would be no isothermal region, but the adiabatic gradient 

 would prevail to the very top. 



It seems possible that the strong convection currents that prevail in the equatorial 

 regions may produce this forced mixing up to a considerable height, and thereby raise 

 the height of the isothermal and lower the temperature. Or the explanation 

 subsequently given as to the conditions that prevail in cyclones and anticyclones may 

 hold (see p. 277). The general rule is that the greater the height H e the lower the 

 \alue of Tc. This holds for change of latitude and for the change from cyclone to 

 anticyclone, but it fails in the case of the annual variation, where apparently the 

 values of VL e and T c should follow those produced by change of latitude. 



To have met with the lowest natural temperatures ever recorded in the equatorial 

 region is not what one would have expected, but is what has happened. The 

 isothermal region over the equator receives at least as much solar radiation as it 



