20 



TEMPERATURE. 



as soon the observations arc plotted. Tliis has been done in the upjier curve in fi'mre 5 

 Bowers' observations being sliown ])y a dot in a circle and those of Meaies by a circle. 



Fig. 5. Tempeiatuies on Barrier. 



It will be seen that with one or two exceptions the observations lie on a regular curve. 



Another set of observations made in the same region at the same time proves the 

 correctness of the temperature curve arrived at by combining these two sets of observations 

 and shows also the imperative necessity of taking the minimum temperature into account 

 in calculating the mean temperature on the Barrier. A small party had gone ahead of the 

 main party to a pre-arranged position where they waited until the main party came up. The 

 junction was effected on the 21st November when, unfoi-tunately, the detached party ceased 

 to record temperatures. This party had been taking observations three times a day and had 

 used a minimum thei'mometer under a sledge to obtain a nigbt temperature. The results are 

 plotted in the lower curve of figure 5, and it will at once be seen that the second curve 



