A method to enable ready ext raction of the dates of occurrence of cold spells using only the afternoon temperature readings for the Hoba rt Divisional Office is devised and checked on 8 years' records (1951-1958). The se c riteria a re used to ext ract date s of cold spells over the 20 year period 1939-1958. These occasions are then subdivided according to the extent of snow on Mt. Wellington and fur the r subdivided according to the time elapsed between the onset of cold conditions and a marked change, if any, to milder conditions. Some conclusions are drawn as to the probability of widespread snow during each month of the yea r and also the probable duration of cold spells.
C. Pierrehumbert (Fri,) studied this question.