76 TEMPEEATURE. 



During the warm months of the year there is a similar tendency, but much less 

 marked. For the months November to February the correlation coefficients between the 

 reduced range and the minimum and maximum temperatures are - 4 and -0 respectively. 

 Thus we see that during the warm months neither the maximum nor the minimum tempera- 

 ture plays a large part in determining the unperiodic temperature change,-. 



In order to understand these results, the temperature conditions in McMurdo Sound must 

 be kept in mind. We will first confine our attention to the cold months May to August. 

 There is little difference in mean temperature during this period, hence they may all be consi- 

 dered together as having similar general temperature conditions. We have already shown 

 that during the cold weather the direction from which the wind blows has little effect on the 

 temperature, both northerly and southerly winds having higher temperatures than hold during 

 calm weather. We have also shown that low temperatures ar^ associated mth a cold layer 

 of air just above the surface, while high temperatures during the winter are caused by the 

 removal of this cold layer and the establishment of a normal temperature gradient through- 

 out the lower atmosphere. A little consideration will show that minimum temperatures are 

 much more largely governed by the cold layer than maximum temperatures, while maximum 

 temperatures are more affected than minimum temperatures by the relative warmth produced 

 by the winds. For every time a cold layer forms even for a few hours it is recorded by 

 the minimum thermometer, but it must exist undisturbed throughout the whole 24 hours 

 if it is to affect the maximum temperature. Similarly a wind which only lasts a few 

 minutes may raise the temperature and give a high maximum temperature, but if it does 

 not last 24 hours it may not affect the ininimum temperature. 



A month during which cold layers frequently form will have a low average minimum 

 temperature. But the cold layer is very unstable, for at a comparatively small distance above 

 it, there is warm air ; hence during a period when the minimum temperature every day is 

 low, there may be a large proportion of days with a relatively high maximum temperature. 

 On the other hand when windy conditions exist both the maximum and minimum temperatures 

 are high for the cessation of the wind for a few minutes does not at once lead to a cold 

 surface layer which takes time to form. Thus frequently when maximum temperatures are 

 high, minimum temperatures are high, but much less frequently are maximum temperatures low when 

 minimum temperatures are low. In other words where the temperature is largely governed by the 

 formation and removal of cold layers, average maximum temperatures cannot undergo such 

 large variations as average minimum temperatures. That these conditions are fulfilled in 

 McMurdo Sound during the winter is clearly seen from figure 27 in which the points for the 

 minimum temperatures are spread over a much larger range than those for the maximum 

 temperatures. In fact the average deviatio)i of the maximum temperatures from their common 

 mean is only a little more than half that of the minimum temperatures, being 3-8°r. and 

 6-2°F. respectively. 



The tendency then during this period is for temperature changes to take place between 

 two limits, the lower being the temperature of the cold layers, which vary largely from month 

 to month both in intensity and duration, and the upper being the approximately uniform 

 temperature which exists when the normal temperature gradient has been established in the 

 lower atmosphere. It is now clear why the non-periodic changes during the winter are much 

 more nearly related to the average minimum than the average maximum temperature, for the 

 latter is an indication of the prevalence and intensity of the cold layer, the formation and 

 sweeping away of which give rise to the majority of the temperature changes. 



In the summer months the cold siu-face layers do not form, while the temperature is now 

 affected by the wind direction. Hence we see that not only are the temperature variations 



