958 
a thermally controlled divide in the same position 
climatically in both Labrador and Russia. 
The Southern Limit of the Boreal Forest. This zone 1s 
less clearly defined climatically and is variable in eco- 
logical character. In Alberta, Saskatchewan, and Mani- 
toba, the forest passes south into prairie (or steppe) 
lands through the so-called parklands, or forest-steppe 
ecotone. A similar forest-steppe boundary is observed 
in Siberia from the Urals to the Yenisei. In eastern 
North America the Boreal forest passes southwards 
into a mixed hardwood-softwood formation, the Great 
Lakes—St. Lawrence forest of Halliday [18], or the Lake 
forest of Weaver and Clements [64, pp. 496-500]. A 
similar passage is observed in European Russia along 
a line extending from Estonia through Yaroslavl to 
the Urals near Sverdlovsk. Both in America and Rus- 
sia, the mixed forest is distinct floristically from the 
Boreal forest to the north and the deciduous forest to 
the south, though it has some elements in common 
with both. 
In eastern North America, certain species typical 
of the Great Lakes-St. Lawrence forest penetrate the 
true Boreal forest, reaching an east-west line with po- 
tential evapotranspiration about 48 cm. South of the 
isopleth H = 51 cm, the dominants are everywhere 
the white and red pine (Pinus Strobus L., Pinus resinosa 
Ait.), yellow birch (Betula lutea Michx. f.), and various 
maples; the spruces (Picea spp.), balsam fir (Abzes 
balsamea (L.) Mill.), and Jack pine (Pinus Banksiana 
Lamb.), typical of the Boreal forest, are subsidiary 
elements. Here again we are faced with a thermally- 
correlated boundary; the southern limit of the Boreal 
forest follows an isopleth of Thornthwaite’s thermal 
efficiency function. 
Once more the same value appears, on a casual in- 
spection, to hold for the Russian forests. Moscow, 150 
mi south of the Boreal-mixed forest line at Yaroslavl, 
has a value of # = 54.2 em, which in North America 
would put it well into the mixed belt. On the other 
hand, Perm (Molotov), 75 mi north of the line, has 
E = 50.8 cm. These values are obviously very close 
to those applicable in eastern North America. What 
is even more surprising is that the inner (forest) edge 
of the forest steppe both in the Canadian Prairies and 
im western Siberia closely corresponds to the same value 
of potential evapotranspiration. 
Gathering all these threads of evidence, we can draw 
these conclusions, all of which remain tentative because 
of the fragmentary evidence employed: 
1. Through most parts of the Boreal region of the 
earth, the temperature of midsummer is the main 
factor determining variations in growth of the common 
tree species. 
2. Variations in summer precipitation have little or 
no effect on growth, except near the forest-steppe mar- 
gin. 
3. The Boreal-forest formation shows a broad divi- 
sion into latitudimal zones: a forest-tundra ecotone, 
next a woodland belt, then a close-forest region. The 
arctic tree line and all the interzonal boundaries are 
highly correlated with Thornthwaite’s thermal efficiency 
POLAR METEOROLOGY 
function, the potential evapotranspiration (#). The 
values of H applying along these lines are much the 
same in the Soviet Union west of the Yenisei River. 
Obviously these conclusions need to be checked and 
rechecked. There are probable sources of error at every 
stage of the analysis. For the ecologist, there is the 
task of more adequately mapping the zonal boundaries 
of the Boreal forest. For the climatologist, there is 
the responsibility of collecting and analysing more and 
more climatic data. Thornthwaite’s stimulating cli- 
matic classification, used above to illustrate the de- 
pendence of the Boreal forest on thermal efficiency, 
needs to be scrupulously examined as to its applicabil- 
ity to the cold climates. Sanderson’s work in the Mac- 
kenzie Valley [51] is an important contribution in this 
particular. Above all, the type of analysis here at- 
tempted needs to be applied to the mountainous lands 
of Alaska and the Yukon, and of Siberia east of the 
Yenisei. Here the montane effect must vastly compli- 
cate the correlations quite easily established in Labra- 
dor and the western Soviet Union. 
Permafrost Distribution. A widespread condition in 
the cold lands is that of permanently frozen soil, chris- 
tened “permafrost” by Muller [43]. This condition is 
achieved when the annual wave of summer heating 
fails to descend to the base of the layer of frozen soil. 
The mechanical properties of permafrost are the busi- 
ness of the engineer, but its distribution is of interest 
to climatologists. Figure 1 shows its approximate extent 
in the Northern Hemisphere. The data used in its 
preparation come from studies by Muller [43] and 
Jenness [32]. 
Few attempts have been made to correlate perma- 
frost with climate. Sumgin (quoted by Muller [48, p. 
4]) states that Russian estimates give OC (82F) annual 
mean temperature as the outer limit of patchy perma- 
frost, and that geographically continuous permafrost 
occurs north of the annual mean isotherms of —3C to 
—6C (26-21F). Thomson (quoted by Jenness [32]) 
suggests that the —5C (23F) isotherm comes closest 
in Canada. All these are approximate estimates, and 
it is quite obvious that such quantities as depth and 
duration of snow cover must also affect the distribu- 
tion. In detail, the occurrence of permafrost is affected 
by soil drainage, slope, exposure, the presence of large 
bodies of water, and the geological history of the region. 
The climatological correlations of permafrost urgently 
await study. 
THE DISTRIBUTION OF SEA ICE 
It was earlier suggested (p. 952) that the term “Arc- 
tic”? might well refer at sea to those areas perennially 
or annually affected by sea ice, either landfast or pack. 
The presence or absence of sea ice is of prime impor- 
tance to both the meteorologist and the climatologist, 
who are confronted with questions of two sorts: (1) 
the effect of ice formation on air-mass structure, and 
hence upon climate; and (2) the climatological require- 
ments for the formation of ice, and for its drift mto 
other waters. The literature on these questions is abun- 
dant but scattered; moreover much of it comes from 
