Abstract
Two years of upperβair data from the four daily radiosonde ascents from five stations in the British Isles are analysed for the pressure levels 150 mb, 100 mb and 80 mb β some 14,000 ascents in all.
The apparent diurnal temperature changes are discussed and the need is shown for further experimental evidence concerning the reality of the measured highβlevel daylight temperatures. The observed daylight heating is too great to be caused by the absorption of radiation and must either be dynamical in origin or due to radiation error of the thermometer element in the radiosonde. From 1948, the times of ascent are suitable for computing the nocturnal cooling throughout the year, independent of radiation error. The temperature falls during the night by about 0.05Β°C hr^β1^ at these levels, the sort of change to be expected if radiative processes predominate.
The apparent rate of change of temperature closely following sunrise and just preceding sunset is about ten times more rapid than the nocturnal cooling.
It is concluded that the amplitude and rate of the apparent variations of temperature measured in daylight ascents should be treated with reserve until further experimental evidence as to their reality is available.