1006 
two. There seems little doubt that the main glaciations 
were contemporaneous on both sides of the North 
Atlantic. In most other parts of the world only two 
or three glaciations have hitherto been recognised, but 
even on the equator in Hast Africa three separate ex- 
tensions of the mountain glaciers can plausibly be 
equated to the Mindel, Riss, and Wurm glaciations. 
In Kashmir four distinct glaciations have been recog- 
nised; the fourth includes four glacial substages. Here 
the second interglacial period was the longest and the 
third the shortest. 
Penck and Briickner [27] made the first approach 
to a time scale for the Quaternary glaciation. Their 
methods were crude but effective, and their results have 
never been seriously challenged on geological grounds. 
Their dates, in thousands of years before the present, 
are shown in the left-hand column of Table I. It may 
be remarked that de Geer* estimated that the Wurm 
ice sheet left the coast of Germany about 18,000 B.c. 
The durations of the various glacial periods naturally 
varied from place to place, being longest near the centres 
of the ice sheets and shortest near the peripheries. 
Thus in Iowa and Ohio the latest glaciation ended 
roughly 25,000 years ago, whereas in Sweden the dura- 
tion of postglacial time is taken as only 8500 years. 
Recent estimates quoted by Flint [11] suggest that the 
durations of the interglacial periods may have been 
somewhat longer than the figures given by Penck and 
Brickner. 
In the last few years striking evidence of the succes- 
sion of glacial and interglacial periods has been recoy- 
ered from the floor of the Atlantic Ocean by the Swedish 
Deep-sea Expedition under the leadership of Professor 
Hans Pettersson, using a special ‘“‘corer’’ capable of ex- 
tracting cores up to 50 ft long from great depths. These 
cores are estimated to include the deposits formed during 
a period of at least a million years. An account of the 
results of this mvestigation is given by Ovey [26]. 
Glacial periods are represented by beds rich in erratic 
debris, showing that icebergs drifted at least as far 
south as 30°N. Interglacial periods are represented by 
foraminiferal oozes between the beds of debris. The 
species of foraminifera indicate the temperature of the 
surface water very clearly; m the tropics the alterna- 
tion of glacial and interglacial periods is shown by the 
alternation of layers with cool and warm species. 
A core obtained in the Caribbean Sea shows a succes- 
sion of four “glacial” periods which have been provi- 
sionally correlated withthe Nebraskan, Kansan, Ilimoian, 
and Wisconsin. The Nebraskan had three substages, 
the Kansan and Illmoian two each, and the Wisconsin 
four or five, but the most interesting point is that the 
second interglacial (Yarmouth stage) was about twice 
as long as the first (Aftonian) and three times as long 
as the third (Sangamon). 
Plwial Periods. Outside the borders of the glaciated 
regions, the rainfall during glacial periods was in most 
places appreciably greater than at present. In the mid- 
3. For a general summary of de Geer’s numerous papers 
see Geogr. Ann. Stockh., 16: 1-52 (1934). 
CLIMATOLOGY 
dle latitudes of the Northern Hemisphere this was 
no doubt due to the fact that ice sheets deflected the 
storm tracks southwards, so that the Mediterranean 
regions, for example, had a rainfall two or three times 
that at present, distributed fairly evenly through the 
year. The Mediterranean storms continued into south- 
west Asia; v. Ficker [10] calculated that at the maxi- 
mum glaciation the rainfall m northwest Pamir was 
four or five times greater than that at the present time. 
The change was most notable on the northern margins 
of the subtropical deserts: wandering storms penetrated 
into the Sahara and this area, now desert, was one of 
the main centres of population. Similar conditions pre- 
vailed south of the main ice masses of North America, 
the lakes of the Great Basin spreading to form large 
inland seas, the best known of which are Lakes Lahon- 
tan and Bonneville [32]. 
The most remarkable development of the pluvial 
periods took place in Hast Africa, near the equator, 
where a succession of great lakes grew from the union 
of a number of existing small lakes, left their deposits, 
and disappeared. The pluvial periods in nonglaciated 
regions coincided with the advances of the mountain 
glaciers, and almost certainly represent the glacial peri- 
ods of Europe and North America. A very early lake 
left deposits, now fragmentary, termed Kafuan; Way- 
land [33] thinks that this lake had two maxima sep- 
arated by a period of earth movements, and he pro- 
visionally equates the lake to the Gunz and Mindel 
glaciations. This was followed, after a long dry interval, 
by the ‘Great Pluvial” im which the very large Lake 
Kamasia was formed; Nilsson [25] equates this pluvial 
to the Riss glaciation. Lake Kamasia then dried up 
completely, and at the same time the mountain glaciers 
disappeared. Then followed a period of renewed lake 
formation, the Gamblian pluvial, less mtense than the 
Kamasian, since the lakes did not overflow their sepa- 
rate basins. Nilsson distinguishes four successive Gam- 
blian lake systems, probably corresponding to the three 
maxima of the Wurm glaciation and a late-glacial 
readvance. Following the close of the Quaternary Ice 
Age there have been a number of minor fluctuations 
of lake levels, which appear to correspond with minor 
advances and retreats of the mountain glaciers, and 
these fluctuations present a succession very similar to 
the postglacial history of northern Europe, though the 
exact correlation is not yet defined. 
Closed lake basins are very sensitive to variations 
of rainfall. From botanical evidence Moreau [23] es- 
timates the average rainfall in the last of the Gamblian 
(Wurm) stages as 44-50 inches, while during the post- 
glacial dry period, when the lakes dried up completely, 
it cannot have been as low as 27 inches. The present 
rainfall in this region averages 3714 inches. On this 
scale even the Great Pluvial period may have had 
a rainfall less than twice the present average. Never- 
theless these changes indicate climatic disturbances 
of large magnitude, which are important for the theory 
of climatic oscillations. 
Earlier Ice Ages. Great ice ages have recurred at 
intervals of roughly 250 million years—the Late Pal- 
