by Paul B. Sears (1935)
On the basis of what is known of the various glacial limits and the present temperature zonation of plant life there has been considerable conjecture as to the pattern of vegetation at the times of maximum ice advance. The absence of barriers to the south and west in North America afforded a condition which does not obtain in Europe where the retreat of plants away from the ice was restricted. The general theory is that tundra, scrub, conifers and deciduous forest shifted as belts before the advancing ice, then followed back in its wake as it melted. Actually, of course, the ice [[p. 38]] lobes extended into the interior grasslands as well as into the forest regions. Moreover, the exact character of the change of climate which produced glaciation is not proved. Certainly the mere proximity of glacial ice is not enough in itself to prevent the growth of trees. Trees ferns are found within sight of the New Zealand glaciers. Spruce-fir forests develop almost at the edge of the retreating ice in Alaska and there is some evidence to show that in places the ice has advanced over standing, possibly living forests (11). In the north central states it has been difficult to find any trace of tundra plants at the base of the oldest living deposits examined, a fact which will be discussed later.
On the other hand, Piech (38) states definitely that the crest of the Riss (III) ice sheet was fronted by a strip of treeless tundra more than 100 km. wide and Firbas (19) reports similar conditions nearly 450 km. ahead of the Würm (IV) glaciation. In the Florida peat deposits Hanna finds diatoms which he considers to have been deposited during the climax of the last glaciation (27). The species are so definitely northern in character that he considers the evidence "overwhelmingly in favor of cooling." Many of the forms now occur no farther south than New England. This leads him to consider the warm Pleistocene flora found by Berry in North Carolina to be interglacial or at most intraglacial, rather than a proof that the influence of the ice did not extend far south of its limits.
The whole problem of displacement of vegetation at the maximum advance of the ice is so significant that it warrants a careful study of all of the scattered and discontinuous peat deposits south of the glacial limits. The task is enormous but it seems the best means of deciding whether glaciation was accompanied by a general and extensive lowering of temperature. Evidently snow limits were lower in tropical mountains during the Pleistocene than now and many of today's deserts were then humid. But we do not know how the latitudinal limits of the various zones beyond the ice edge were affected.
Passing from the question of glacial to that of interglacial climate, we are confronted with an embarrassing confusion as to the number, character and correlation of these interglacial stages. Opinion, even today, ranges from the belief that only one such [[p. 39]] interval can be dignified by the term interglacial, all other periods of recession being intraglacial, to the belief that five or more truly interglacial intervals existed during the Pleistocene.
An excellent idea of the situation may be obtained by realizing that the Toronto interglacial beds have at different times been attributed to at least four different intervals. Of the Aftonian interval Thwaites remarks, "strangely enough, although the evidence on which the Aftonian was originally defined is worthless there is, nevertheless, a real 'Aftonian' interval of great length and possibly constituting a genuine interglacial stage" (45).
There is a growing tendency to emphasize four major glaciations during the Pleistocene. Brückner (7), with Leverett, regards them as essentially contemporaneous in Europe and North America, just as students of lesser climatic fluctuations of recent times point to a certain correlative behavior throughout the northern hemisphere. Disregarding subglaciations the scheme is as follows:
Period North America Europe
As to subglaciations there is still much discussion. The Iowan substage is important because of its well marked interval, the Peorian. Leverett apparently still regards these as late Illinoian, but Kay in Iowa and Leighton in Illinois have recently agreed in placing them in the early Wisconsin, thus following rather than preceding the Sangamon interglacial (8). Similarly in Europe there is some disagreement as to the position of the Warthe sub-glacial as indicated by Firbas and Grahmann (21) and Premick and Piech (38).
Szafer (44) in 1925 presented an account of the Polish interglacial periods in which he ascribed prolonged warm conditions, partly oceanic, partly continental to the last interglacial (Riss-Würm, Masovien II). Jessen and Milthers (32), working in Jutland and northwest Germany, agree with him, describing the last or Riss-Würm interglacial as passing up to an atlantic (warm, humid) [[p. 40]] climax and back, with a brief temperate fluctuation during the cool closing period. The penultimate (Mindel-Riss?) is described as simpler, progressing up to a warm climax and then back to glacial conditions.
However, all of the sites except one which were described by Szafer as Riss-Würm (Masovien II) have now been correlated by Premick and Piech (38) with Mindel-Riss (Masovien I). In a detailed analysis they show the latter to have been of more marked character than the former, or last interglacial. Four beds of Mindel-Riss age show gradual progress from pine-birch through oak, then oak-linden-hornbeam to fir, fir-spruce, and back to pine. During this final pine stage there appears to have been a brief amelioration marked by temporary increase in oak and other hardwoods. These quickly vanished before pine, arctic birch and willow, heralds of the returning ice. A similar analysis of the last interglacial agrees in its general import but differs in detail, presenting a predominantly cool and moderately humid character. It appears to have been mainly coniferous with briefer and weaker expressions of the mixed oak phase.
Dokturowsky (14) has found similar conditions for the last interglacial in Russia. Following the Riss-ice was a cold period with pine, birch and willow. This was succeeded by a warm period with hazel, alder and mixed oak forest, while another cold period with spruce and fir preceded the Würm glaciation. Dokturowsky is unable to confirm Szafer's report of a prolonged and intricate character for the last interglacial. With respect to the preceding, interglacial information from Russia is scanty, remains of yew, beech, larch and Trapa appearing to comprise the finds reported. Brookes, with others, considers the penultimate or Mindel-Riss as of much greater length and more marked character than the other interglacial periods, and states that it should probably be correlated with the prolonged Yarmouth stage in North America.
Firbas and Grahmann (21) report that deposits following the Warthe subglaciation in south Brandenberg show fluctuation between birch and pine, the latter accompanied by traces of deciduous forest. The period appears to have been predominantly cool. Spruce appears, although afterwards completely absent from the late glacial and early postglacial stages in the same vicinity.
[[p. 41]] In addition to these more comprehensive recent studies there is a considerable number having to do with isolated and fragmentary profiles, correlated with varying degrees of certainty. From all of the work to date, incomplete as it is, certain principles appear clearly enough. There is a general and significant, even if not detailed, parallel between the events of postglacial time and those of the interglacial intervals, so far as the latter are known. In the second place there was a sufficient variety of mild climates during one or more of the intervals to have permitted extensive migrations of plants, including some which had been thought of as tertiary rather than as interpleistocene relicts (44).
For North America, the contents of the Toronto beds show that during whatever interglacial period they were laid down, vegetation flourished, which suggests a milder climate than that of Toronto today (10). Toronto is within but near the present northern range limit of the linden and papaw, while the natural range of the Osage orange is distinctly restricted to the southwestern Mississippi Valley.
Evidence of a marked interglacial climate during what Kay considers the Aftonian, or first interval, has been secured by Lane working with the Iowa Geological Survey. His results have not yet been published but the writer has examined, and in some cases checked, his material. Acknowledgment is here made for courteous permission to refer to this and other data in advance of publication. Lane finds evidence in the Denison beds of a prolonged steppe or grassland period which passes through a brief deciduous forest phase into coniferous forest.
Two independent studies have been made on the Sangamon, an unpublished one by Lane on the Wapello beds in Iowa, the other by Voss (47) in Illinois. The results of both studies are in substantial agreement. Voss reports vegetation indicating conditions in Illinois similar to those of the conifer belt of the northern states and southern Canada. Lane finds conifers throughout. Pine with grass, oak and spruce later yields to a cover of pine and spruce. The Wapello beds are considered late Sangamon.
The Peorian, now considered intraglacial, whether within the late Illinoian or early Wisconsin, has also been studied by Voss, with findings of a conifer climate similar to that described by him for the Sangamon. The climatic resemblance of the Sangamon to [[p. 42]] the Peorian is much more consistent with the recent reports from Poland regarding the Riss-Würm than with the earlier report of Szafer or the findings of Jessen and Milthers in Jutland and northwest Germany.
In addition to the report of Voss on the Peorian, we have reports on intraglacial beds of late Wisconsin age from Wilson (49) in Manitowoc County, Wis., and from Cooper and Foote (12) in Minneapolis. Both agree that the organisms found indicate conditions now prevalent to the north of the sites investigated, thus confirming their intraglacial character.
Before leaving the subject of climate during the glacial period, it seems advisable to mention the importance of the loess. This wind deposited soil is considered an excellent criterion of dry conditions. Several occurrences of it have been correlated as interglacial and so used to predicate marked continental conditions during those intervals. This assumption is strengthened, not only from the character and mode of origin of the loess itself, but from the character of the molluscan remains which it shows. Baker (4) has studied the latter in loess deposits of the Yarmouth, Sangamon, Peorian and two Wisconsin intraglacial periods. In addition to several extinct species he finds others whose present range is definitely western or northern. If practicable it would be of interest to see the same statistical methods applied to molluscan remains which have proved so useful in the case of fossil pollen.
Five distinct soil horizons, separated by loess, have been described from Bermuda by Sayles (40). The soils have been attributed to mild humid inter-glacial conditions, while the loess is considered to be due to the severe, windy periods of glacial advance. Harper (28) has found somewhat similar conditions in Oklahoma which he ascribes to the Pleistocene. The Rocky Mountain region and the Southwest abound in fossil soils or humified layers separated by layers of aeolian material or heavy alluvial wash suggesting arid intervals, but their correlation is necessarily a difficult matter, requiring great caution.
The climate and vegetation of postglacial times has been much more thoroughly studied than have Pleistocene conditions. Nearly 100 years ago in the Danish peat bogs, Steenstrup found successive strata of buried forests not showing evidence of a steadily warming climate since glacial times. As decades went by, the study of [[p. 43]] fossil remains in the peat and of the peat layers themselves suggested strongly that the climate of Europe had once been milder than today. This probability seemed to be increased by the presence of relict steppe plants in places having today a forest climate.
The solution of the problem was greatly complicated by historical, traditional and even stratigraphy evidence indicating climate fluctuations of considerable intensity but short duration. Moreover, the destruction of plant cover and humus following human occupation always tends to create an impression of greater aridity than the climate justified, except perhaps in areas having a marked oceanic climate. This, of course, confuses the issue and has doubtless been responsible for much of the prevalent notion that our present climate is becoming drier. That it is becoming cooler seems warranted by the cultural history of Iceland, the Esquimaux, and by the southward shift of the limits of the vine.
Hypothesis has passed through four important phases as follows:
First, the simple assumption that vegetation had moved south before the advancing ice and subsequently returned to a more or less stable equilibrium. This idea is still implicit in many quarters (24, 48).
Second, the Blytt-Sernander hypothesis, postulating a series of fluctuations in both moisture and temperature, and designating the following periods: pre-boreal, boreal, atlantic, sub-boreal, sub-atlantic. The atlantic or middle period was conceived of as a climatic optimum, warm and humid, between the very continental boreal and sub-boreal. The sub-atlantic represented a return of humidity. The stages so set forth were associated with a definite stratigraphy and archeology, and dated with great accuracy through the method of DeGeer (42). They thus afforded something very tangible, capable of being checked with most minute care throughout Europe. The value of this hypothesis as a scientific tool is not to be underestimated, even by those who have rejected it. As is often the case, it has probably suffered more from its friends than its enemies.
Third, G. Anderson (1), holding the view that temperature changes have been the predominant ones, conceded that glaciation was followed by a very dry and ultimately warm boreal period, succeeded in turn by an equally warm but much wetter atlantic. But he denied the reality of a second dry or sub-boreal period and [[p. 44]] predicated a gradual climatic deterioration from the atlantic (neolithic) down to the present. In this general interpretation Gross (25) and Rudolph (39) appear to concur.
Fourth, L. von Post (46) divides postglacial time into three periods: 1., a period of increasing warmth; 2., a period of maximum temperature; 3., a period of decreasing temperature. On this general frame-work, he admits, lesser fluctuations of the various climatic factors have had their place. Under suitable conditions the minor fluctuations have doubtless been extreme enough to produce a marked effect. Very soundly, he concludes that only by the most minute analysis between correlated horizons, on a widely coordinated plan, can we determine the extent, character and intensity of the actual fluctuations. Gams (22) appears to agree essentially with von Post but considers that the waning of the warm continental maximum has occurred in oscillating fashion, thus explaining whatever evidence may exist for a dry sub-boreal about 3000 years ago. He also considers the evidence of recent drying not to be wholly due to human interference.
Having thus outlined the trends of theory, let us examine the phenomenology upon which they are based. The difficulty of accepting an hypothesis of return to stable climatic equilibrium following glaciation in Scandinavia has already been noted. The pioneer work of Steenstrup was followed by much more detailed study of bog profiles, plant fossils and molluscan remains, much of which went counter to the idea of a stable postglacial equilibrium. Both in Europe and North America taxonomists and students of floristics found themselves confronted with problems to be explained only by assuming a warm continental climax some time during the postglacial (23). Geological analysis of glacial retreat showed that it had been marked by periods of re-advance. Furthermore, the demonstration of huge-scale fluctuations, such as the Permian3 and Pleistocene as well as the so-called "cycles" of small magnitude have, it must be confessed, placed scientists in a receptive mood toward the demonstration of fluctuations of an intermediate order, to be measured in tens of centuries perhaps.
[[p. 45]] Allowing for appropriate differences in alpine and other extreme habitats, the general vegetational history in Europe has been about as follows. Widespread and prolonged postglacial tundra conditions are indicated by remains of Dryas, birch and willow, far south of the ice front. It is generally found that this was succeeded by a period in which pine predominated, variously accompanied by birch and hazel. The pine was followed by oak, the hazel frequently persisting on into the oak period. In fact, hazel, considered to be an indicator of dry continental conditions, exhibits a singular behavior so far as its relative pollen frequency is concerned. Its maximum may precede, accompany or follow that of the oak. Perhaps as much as any other factor this contributes to the uncertainty as to whether conditions were dry during both the pine (boreal) and oak (sub-boreal) or only during the former. Erdtman (15), in a very penetrating analysis, points out certain contingencies which may affect preservation of the hazel pollen even when abundant, while Firbas (20) notes that weathering of peat during the hazel maximum may interfere with the study of its pollen content. In the words of Theophrastus, "this, then, is proper subject for further investigation."
The oak period, characterized by an admixture of other deciduous trees such as linden and elm, exhibits a variable length and intensity of expression. It is least marked at high altitudes and latitudes, more so in continental than oceanic areas. In a large part of Europe the oak appears to have been replaced eventually, either through an intermediate stage of beech or without it, by conifers, birch and alder, suggesting cooler and perhaps more humid conditions.
In addition to the changes of vegetation, as shown by pollen analysis, the changing character of the peat itself affords some evidence. In the late glacial stages peat formation seems to have been restricted and exceedingly slow which has prevented an adequate record of events and measure of time. With the establishment of a sub-arctic forest of birch and pine, heavy formation of sedge peat occurred. Later during the warm, dry (boreal) climax (hazel) this peat was subjected to heavy humification which often spoiled the record. During the mixed oak stage, formation was renewed (atlantic) but apparently often this was followed by another period of weathering (sub-boreal). On the latter point there is [[p. 46]] considerable controversy between those who believe in a dry sub-boreal, and their opponents. The weathered sub-boreal (?) layer in question is known as the Grenz-horizont. According to Gross (26) its origin and even existence is frequently a matter of doubt. Above the mixed oak strata is found a renewal of peat formation, rather generally attributed to "climatic deterioration"--either lowered temperature, increased moisture or both. Quite generally the surface layers of peat show disturbance, attributed by many to cultural effect, but thought by others to be due to a recurrence of dryness.
There is apparent a growing and hopeful effort to isolate all of the factors such as migration rate of trees, late glacial interstages, local plant succession, soil development and climatic change. The reader interested in further information should consult the recent, more general papers of Gams (22), Gross (25), Firbas (19), Keller (34), von Post (46), Rudolph (39) and others cited by Erdtman (17).
By comparison with Europe, North American studies are still meagre and tentative. The area is vast and diversified with its own geological history. The absence of marked barriers to plant migration and a decidedly richer forest flora than that of Europe introduce complications. Although there is a growing belief in the similarity of climatic trends, it is essential that the American problem be approached from a strictly objective basis.
As in Europe, the known peat record does not extend back continuously beyond a maximum of about 10,000 years. Such late-glacial records as are known appear to be isolated and to indicate general northerly conditions, one from the Connecticut valley being pronounced by Hollick (29) to be the American equivalent of the European Dryas flora. There is not much but conjecture as to conditions beyond the ice-front, although Hanna's (27) report of diatoms in northern Florida which now range from New England northward is significant.
There appears to be little or no record of tundra at the bottom of living peat profiles. Instead, the record begins with a marked abundance of fir and spruce pollen, not only in southeastern Canada (3) and the Lake States (41) but even in Iowa (35). This does not preclude the possibility of a long tundra stage. The basins suitable to peat formation may have been occupied by dead ice until the forest returned northward to surround them. Tundra [[p. 47]] may possibly have been followed by steppe conditions so dry as to preclude peat formation. Or, as in Alberta (16), the central states may have been covered with an aspen parkland and yet have shown a pollen profile in which only conifers were indicated. Antevs (2), in common with many others, believes that a tundra zone existed but that it may have been comparatively narrow. Further study may elucidate this point.
While the bottom of the record in eastern North America shows the general presence of spruce and fir, above that point important regional differences appear. In southeastern Canada (3), Minnesota and Northern Wisconsin (48) pine quickly replaces the other two coniferous genera and is predominant until the upper layers where spruce and fir once more increase at the expense of pine. Whatever indications of deciduous trees there may be, they are strongest during the middle of the pine period, not inconsistent with the idea of warmer conditions than now obtain. In Canada, hemlock shows a slight but definite tendency to appear before and after the pine, rather than with it. This suggests to Auer that the pine corresponds to the European sub-boreal, preceded and followed by more humid periods, the atlantic and sub-atlantic, respectively. On the other hand, Voss reads no such meaning into his own profiles from the north lake states although he concedes their similarity to those of Auer. He attributes the recent increase of spruce to local conditions without discussing their striking coincidence with local effects across half of this continent, not to mention northern Europe and Japan (33).
Profiles from central Wisconsin (48) and in less detail from Michigan (37) and northern Indiana (31, 36) resemble in their general character those just described save that there is a definite interval of deciduous predominance instead of a mere showing during the pine period. To varying degrees these profiles show indications of a recent increase of conifers at the expense of deciduous genera, again suggesting the passing of warmer conditions.
In Ohio (41), central Indiana (31) and northern Illinois (48), following the conifers, the deciduous interval is marked and prolonged through to the present. In northern Illinois, where it occupies the upper two thirds of the profile, Voss states that climate has been essentially uniform and attributes any fluctuations to edaphic and topographic causes. In Ohio the fir and spruce are first [[p. 48]] replaced by pine. The deciduous forest interval then begins with an appearance of white pine, beech and hemlock, later showing a predominance of oak-hickory with a recent and slight but definite increase in beech. Sears has interpreted the oak-hickory period as dryer, perhaps warmer than those which preceded and followed it. The transition from coniferous to deciduous he considers to be moister than the pine period which preceded it.
In Iowa (35) the coniferous stage passed through one of deciduous forest into grassland near the base of the profile. The grassland shows at least two periods of abundance of pollen, indicating semi-arid conditions. Sears has correlated these tentatively with the pine and oak-hickory periods in Ohio, and the latter with Auer's pine period in Canada.
It must be conceded that any correlations at the present time are highly provisional and that an immense amount of work needs to be done. The first task is to establish the broad outlines of major change. A simple approach should be to test whether von Post's hypothesis of increasing warmth, warm maximum, and decreasing warmth is justified as it seems to be by much of the work except that in Illinois. In a transitional region like Illinois it is essential that search be made for strata showing grass pollen and that species, as well as genera, having a complementary character be studied. In Ohio this was done by using beech and hickory. Hotchkiss and Ingersoll (30) report geothermal measurements from the Calumet district which agree with the requirements of von Post's outline. Incidentally, Dachnowski (13) reports petrographic sequences in his peat profiles such as would be expected if not only the general hypothesis of von Post but the moisture fluctuations postulated by Auer and Sears were a reality. Certainly these assumptions all need much more rigorous proof than they have at present. Complications may be expected because of human disturbance of recent vegetation and the possibility, however remote, that there has also been a recent swing towards drier climatic conditions. Furthermore, it is essential to reckon with the probability that climatic shifts, if any, "have not operated uniformly nor persisted equally in all types of habitat" (42).
In the meantime, there is a healthy enough divergence of viewpoint among American students to insure that any hypothesis whatsoever will be thoroughly tested, including the hypothesis that [[p. 49]] much of postglacial climate has been uniform (3, 6, 9, 13, 24, 42, 48).
1. ANDERSSON, G. Die Veränderungen des Klimas seit
dem Maximum der letzten Eiszeit. Cong. Geol. Internat. 11th Session 1:
Contribution No. 31, Bot. Lab., Univ. of Okla. [[on