TREE-RING INDICES OF RAINFALL, TEMPERATURE, AND RIVER FLOW 
appreciated by a careful study of such comparisons 
as those in Fig. 2. Differences in growth response in the 
Gila area of southern Arizona and New Mexico, sub- 
ject to flash floods, as compared with the conservative 
Green River area of Wyoming, may be specially noted. 
The Douglas fir index for the Colorado River Basin 
HEE Ee) 1870 Hehe) 1890 He plo) Hee 1930 1940 
Foleo eke RIVER BeolN ABOVE LEE'S 
OCT-JUNE 
RAINFALL 
Heo) 
100+ STATIONS 
uGLas FIR 150 
Se 100 
PONDEROSA AND j50- 
LIMBER PINES 
TaN 
1860 1870 1930 1940 
Fic. 3.—Regional tree-ring indices in the Colorado River 
Basin compared with gage data. Rainfall is expressed in per 
cent of the mean; tree growth in per cent of the mean trend; 
runoff in acre-feet X 10°. The runoff of the Colorado River pre- 
ceding 1895 is an approximation by H. C. LaRue (U.S.G8. 
Water-Supply Paper 395, 1916) on the basis of Salt Lake levels. 
(From University of Arizona Bulletin.) 
above Hoover (Boulder) Dam is seen (Fig. 3) to pro- 
vide a fair first approximation to the winter rainfall 
chronology of that region and to the river runoff. 
Independent 342-year indices for two types of pines 
1027 
mits a sensitivity to rainfall fluctuations which is at 
best on a much lower level than that of drought-site trees. 
The big-cone spruce of southern California has, how- 
ever, been found to show a sensitivity rivalling that of 
the best trees of the Colorado Plateau; a 560-year record 
has been constructed. Its value as an index of rainfall 
and runoff is indicated in Fig. 4. 
1850 1860 1870 1880; 1890 
1900 1910 1920 1930 1940 % 
200 
| | N 50 
RAINFALL | 4 I Al Aant [MAN i 
pA Pe Nh My AVY 90 
prance i) WAN YW 
50 
RUNOFF SAN GABRIEL RIVER 
(NEAR AZUSA) 
A A , , a| /150 
SPRUCE |/% ee ‘! J 
SRT a IAT tT a pat SA 
Paty 4 IA aT’ 
RANGE WIV Wo VY -30- WA A V A} 89 
\ h I\M y are 
RUNOFF KINGS RIVER (AT PIEDRA) — 2/9 af I\ fe A nN WA 
Sia | Ny YY %, 
10%a.f, MA = 
\ A 7 
ROU ACECRAINE WADA AA Aheg tig . 100 
! A : AN 
50 
Fig. 4.—Regional tree-ring indices for southern California 
compared with gage data. Rainfall is expressed in per cent of 
the mean; tree growth in per cent of the mean trend; runoff in 
acre-feet X 10°. (From University of Arizona Bulletin.) 
The third regional survey of broad character, that 
of Keen [20] in the ponderosa pines of semiarid eastern 
Oregon, makes possible some centuries-long compari- 
sons with southern California and the Colorado River 
Basin. The correlation coefficients of the Oregon index 
with rainfall and runoff are seen in Table II to be of 
the same order of magnitude as those in the other two 
regions. 
The analysis by Antevs [3] in the general area studied 
by Keen is particularly useful in its detailed discussion 
of the complex of factors influencing the observed 
erowth departures in recent decades. Among other 
analyses, the study by Hardman and Reil [16] in the 
Truckee River Basin provides extensive data. 
Tasie IJ. CorrRELATION Corrrictents BETWEEN GAGE RECORDS AND REGIONAL TREE-RinG INDICES 
Tree index Correlated with 
Colorado River runoff 
Eastern Oregon rain 
Columbia River runoff 
Southern California Coastal Rain 
“ec “ce “ce “ 
Colorado River Basin 
Eastern Oregon 
“cc “ce 
King’s River runoff 
San Gabriel River runoff 
Period Interval (yr) r unsmoothed r smoothed* 
Oct.—Sept. 49 +0 .66 -+0.81 
Sept.—Aug. 66 +0.50 = 
Jan.—Dec. 57 +0.56 = 
July—June 94 +0.65 +0.75 
“s fs 49 +0.71 +0.85 
Oct.—Sept. 49 +0.52 +0.64 
Oct.-Sept. 49 +0.61 +0.86 
a+ 2b+c 
* By formula b’ = 7 
support in general the main index, but show detailed 
differences in mean growth, which are probably only 
partly traceable to the relatively small number of 
component specimens in the pines. Correlation co- 
efficients are shown in Table II. 
The giant sequoia of California has yielded by far the 
longest ring sequences, four of the trees analyzed by 
Douglass being over 3000 years of age [5]. Unfortu- 
nately, the relatively moist habitat of this species per- 
Temperature Indices in the Arctic. Tree-ring chro- 
nologies in Scandinavia exceed in volume those for any 
other region outside of the southwestern United States. 
Two factors seem largely responsible: the major eco- 
nomic importance of forests, and the stimulus of the work 
begun in 1879 by de Geer [4] on the annual clay layers 
or varves. 
Problems associated with the recognition and dating 
of false annual or locally absent rings, so important in 
