15 
The average daily mean temperature (based on twenty- 
four hourly observations daily) of 54 degrees for July shown 
in Table I places Churchill significantly south of the 50°F. 
(10°C.) July isotherm, believed by some biologists and geogra- 
phers to be a suitable indicator of the southern limit of the 
true Arctic. This isotherm, shown in Chart 1-5 of Thomas (1953), 
lies in most places considerably north of the polar limit of tree- 
like conifers shown in Figure 7 of Hustich (1953). A much closer 
correspondence with the northern tree line is attained by the 
“Nordenskjold line” shown in Figure 1 of Polunin (1951) and 
Figure 1 of Hare (1951). The formula upon which this iso- 
therm is based employs not only the factor of mean temper- 
ature of the warmest month but also that of mean temperature 
of the coldest month for the localities through which it is 
drawn. Hare notes that Tierra del Fuego, at the southern tip of 
South America, is forested in spite of the fact that the average 
temperature of the warmest month is only between 46— 50°F., 
presumably because mean daily temperatures throughout the 
winter remain above freezing. Apparently a truer reflection of 
the influence of temperature upon growth is given by the com- 
bined use of summer and winter temperatures than by the use 
of the former only. 
Sanderson (1948) has outlined the climates of Canada 
according to Thornthwaite's revised classification (1948), 
placing Churchill in the humid, cool microthermal zone, with 
an evapotranspiration value of 12.6 inches (32 cm.). Potential 
evapotranspiration is the combined evaporating power of the 
vegetation and a soil surface in which the supply of moisture 
is unlimited. It is expressed as a function of day length as 
well as of temperature and is used as a measure of the thermal 
efficiency of a region. It was Thornthwaite's belief that, in the 
cold climates, restriction of growth by cold far outweighs the 
effect of scanty precipitation, moisture normally not being a limit- 
ing factor for plant growth in arctic and subarctic regions. This 
belief is supported by tree-ring studies of Hustich (1949) in 
Scandinavia, Giddings (1941; 1947) in Alaska and District of 
Mackenzie, and Marr (1948) in the Richmond Gulf area on the 
east coast of Hudson Bay. A remarkable uniformity of ring width 
