954 
outweighs the effect of scanty precipitation. In his 
view, precipitation is everywhere sufficient to meet 
the very limited demands of plant growth throughout 
the Boreal and Arctic regions. 
This common assumption rests upon very little 
serious research or experiment. The Scandinavian 
school of ecologists appear to have been pioneers in 
this field. Several workers have studied the effects of 
climate upon radial and vertical growth of common 
coniferous species, such as the Scotch pine (Pinus 
sylvestris L.) (Hustich [29, 30]) and European spruce 
(Picea abies (L.) Karst.) (Hidem, quoted by Hustich 
[30]). Though the climatic relations of individual species 
are not strictly comparable with the relations of a 
vegetation formation, they offer a valuable indication 
of the mechanism of climatic control near the pole- 
ward limits of tree growth, and are hence of outstand- 
ing interest. 
The most recent review of this work comes from 
Hustich [80], who has carried out important field sur- 
veys in Labrador and Arctic Scandinavia, and who has 
thus had an opportunity to study the problem in 
differmg environments. He found that both vertical 
and radial growth in northern conifers near their north- 
ward limit were closely dependent on midsummer tem- 
perature. His results are of such interest that they de- 
serve a comprehensive summary. In a stand of Scotch 
pine at Utsjoki (Lapland, lat. 69°30’N) he found the 
following correlations: 
1. Radial growth (z.e., increase in diameter due to 
addition of annual ring of woody tissue) was highly 
correlated with July mean temperature for the same 
year at nearby Inari (68°57’N, 26°49’E, 153 m (502 
ft) above sea level). 
2. Vertical growth (7.e., increase in height) was 
correlated with July mean temperature for the previous 
year, though less clearly than in the case of radial 
growth. 
He also showed that variations in radial growth had 
followed quite closely secular variation of July tempera- 
ture since 1890. In another series near Sodankyla [29] 
he obtained a correlation coefficient of +0.54 + 0.12 
between July temperature and radial growth. In this 
series the corresponding coefficient between July rain- 
fall and radial growth was —0.24 + 0.16. All other 
attempts to show correlation between rainfall and 
growth proved equally abortive. Accordingly, he con- 
cluded that “‘.. . in the northern part of the temperate 
zone the correlation between temperature and growth 
is stronger than the correlation between precipitation 
and growth.” He ascribes this independence of growth 
on summer rainfall to the maintenance of soil moisture 
content by melting snow. 
Erlandsson [12] obtained even more convincing evi- 
dence from Karesuanda in northern Sweden. He was 
able to correlate radial growth in Scotch pine with 
mean temperatures over the period 1879-1931. His 
results are presented in Table I. 
This table emphasises the importance of July tem- 
perature, rather than that of the summer as a whole. 
Apparently the critical season is extremely short in 
POLAR METEOROLOGY 
these high latitudes. Like Hustich, Erlandsson con- 
cluded that precipitation was a negligible factor in 
determining growth. 
TaBLe I. CoRRELATION COEFFICIENTS BETWEEN CLIMATIC 
Factors AnD RaptaL Growts or ScorcH PINE 
AT KARESUANDA, SWEDEN, 1879-1931 
(After Erlandsson [12]) 
Climatic factor Correlation coefficient 
+0.29 + 0.13 
June, mean temperature 
July, mean temperature +0.70 + 0.07 
August, mean temperature +0.24 + 0.13 
Rainfall, June-September —0.19 + 0.13 
These important results depend upon the well-estab- 
lished technique of tree-ring analysis; the width of the 
annual ring of xylem is carefully measured, variations 
from year to year in thickness being assumed to indi- 
cate variations in climatic stimulus to-growth. 
Hustich [30] points out that the northern conifers, 
though they grow in a region of great climatic hazard, 
have a remarkable freedom from “‘suppressed” or “‘miss- 
ing”’ annual rings, such as are constantly referred to 
by the Arizona school of tree-ring workers. In lower 
latitudes, where temperatures are high enough to sup- 
port growth for a large part of the year, summer 
drought becomes an all-important factor. Along the 
forest-steppe margin, occasional drought years may 
virtually suppress radial growth, and the tree trunk 
therefore contains only a stunted, incomplete, or vesti- 
gial “ring” as a record. This is true of all parts of the 
world in which occasional droughts occur, even, for 
example, in equable southern England [83]. Precipita- 
tion is a notoriously variable element; hence in any 
region in which the maintenance of soil moisture 
through a long growing season depends on summer 
rainfall, the annual-ring spectra of trees will contain 
marked variations and “suppressed” rings. 
If precipitation is unimportant as a factor influenc- 
ing growth in northern regions, the annual-ring spectra 
should show only small variations in annual growth, 
for July temperature, the apparent prime control, is 
nowhere very variable. Hustich’s observation that 
suppressed rings are absent from northern Scandina- 
vian conifers confirms this deduction. Hustich’s finding 
is in turn confirmed by several investigations in North 
America. Thus Giddings, accustomed to work in the 
dry western part of the United States, apparently did 
not notice suppressed rings in the Alaskan [14] and 
Mackenzie regions [15]. Marr, in an important mono- 
graph on the forest-tundra ecotone of western Labrador- 
Ungava [38, pp. 142, 143], was even more specific. 
Referring to white spruce (Picea glauca (Moench) Voss), 
he writes “‘... annual growth has been so uniform at 
Gulf Hazard during the last 230 years, that only 41 
growth-rings are sufficiently below average width to 
permit visual detection of the deviation.” 
There is thus ample support for the view that sum- — 
mer temperatures, and especially July temperatures, 
are the major factor determining radial growth (and 
probably vertical growth and volume increment too) 
