Pre-Pleistocene Glaciation on Earth" Implications for Climatic History of Mars I

ICARUS 50, 423--443 (1982)

NICHOLAS CHRISTIE-BLIC K

Exxon Production Research Company, P. O. Box 2189, Houston, Texas 77001

Received August 18, 1981: revised March 29, 1982

The history of ice ages on Earth, extending back more than 2 Ga ( 10 a years), has been established by the recognition in strata of many ages of the assemblage of erosional features and deposits generated by glacial activity. The most useful indicators are widespread diamictite, striated and faceted clasts, polished and striated pavements, and laminites containing "dropstones'" inferred to be ice-rafted. The chronology of ancient glaciation is limited by our ability to recognize the results of glacial activity in an incomplete geologic record and by the age resolution attainable with available dating methods. Ice ages are currently known from the Early Proterozoic (2.5-2.1 Ga ago), several intervals in the Middle to Late Proterozoic (1.0-0.57 Ga), the Late Ordovician to Late Silurian (440-415 Ma, or 10e years), possibly latest Devonian (360-345 Ma), Permo-Carboniferous (335- 245 Ma), and late Cenozoic (26 Ma ago to the present). However, the timing of pre-Phanerozoic ice ages, occurring before 0.57 Ga ago, is known only approximately, and glacial events would have been shorter than is suggested by the bracketing ages. A generally warm climate seems to have prevailed for the remainder of Earth history, although Mesozoic-Cenozoic seismic stratigraphic evidence suggests that a small ice sheet may have been a persistent feature near the south pole throughout the last 500 Ma. Terrestrial climatic fluctuations occur on several time scales. The occurrence of ice ages on a time scale of 108-I0 a years, and possibly higher-frequency fluctuations (10e-10 ~ years), appears to be controlled by changes in solar output and atmospheric composition, and by the rearrangement of continents and oceans as a result of the motion of lithospheric plates. Climatic changes on a time scale of 104-10 ~ years are driven by perturbations of the Earth's orbital parameters. The principal evidence suggesting that significant climate changes have occurred on Mars consists of widespread dry channels (3.5-0.5 Ga?) and of geologically young dust-ice layered terrains in the polar regions (< 10 Ma?). By analogy with the Earth, polar layered terrains have probably existed episodically on Mars for a considerable time although there is little direct evidence for this. Factors affecting a potential sedimentary record are the degree to which ice sublimed or melted, the effectiveness of eolian destruction of layering, and, in contrast to the present polar terrains, the tendency for ice to flow and slide over its substrate. As on Earth, climate changes on Mars have probably been induced by the long-term evolution of the atmosphere, by changes in solar luminosity, and by variations in orbital parameters. Major tectonic events may have affected Martian climate through changes in planetary obliquity.

INTRODUCTION investigation. Recent summaries are by Crowell and Frakes (1970), Steiner and The history of ice ages on Earth extends Grillmair (1973), Keller (1973), Harland and back more than 2 Ga 2 (Fig. 1), and has been Herod (1975), Crowell (1975, 1978), established by over a century of geologic Edwards (1978), Frakes (1979), Pollack (1979), Harland (198 la), and Hambrey and 1 Presented at the "Workshop on Quasi-Periodic Harland (1981), whose compilation is the Climatic Changes on Earth and Mars" held at NASA's most comprehensive currently available. Ames Research Center, Moffett Field, California, Feb- ruary 24-26, 1981. Evidence for climate changes on Mars has 1 Ga = 109 years. 1 Ma = 108 years. 1 Ka = 103 been recognized comparatively recently years. following the observation of dry channels 423 O019-1035/82/050423-21 $02.00/0 Copyright (c~ 1982 by Academic Press. Inc. All rights of reproduction in any form reserved. 424 NICHOLAS CHRISTIE-BLIC K

Ga Go Ma ...... O ~////" ~ Cenozoic PHANEROZOIC

,///Z/////A • 5 7- 777-/77777.-'7 Late

Middle 245 7-~ Perrno- Carboniferous --I.6-- 7~36 07-~ Devonian PROTEROZOIC ~,415 7~I Ordovician - 2.1 Early 40 TM Silurian

2.5- Witwatersrand diarnictite --L Stillwater 570 diamictite

ARCHEAN

Oldest rocks 4.0--

Age of Earth 4.65-- FIG. 1. Timing of ice ages on Earth (diagonal ruling). Ice ages indicated may include significant intervals of globally warm climate. Small ice sheets may have existed at times not normally considered as ice ages. The timing of Proterozoic glaciation is not well established. and polar layered deposits by the Mariner 9 and significance of periodic and episodic and Viking spacecraft (Pollack, 1979). It is climatic fluctuations on Earth; and (3) to therefore reasonable to ask whether any of consider ways in which the terrestrial rec- the terrestrial experience bears on the in- ord may bear on the ancient climatic his- vestigation of Martian climate, and espe- tory of Mars. Details of relatively recent cially to consider the significance of peri- climatic fluctuations on both planets (104- odic and episodic fluctuations in the climate 10 r years) are discussed elsewhere in this of Earth in view of the rhythmic appear- volume. ance of Martian layered deposits. The purpose of this paper is threefold: (1) RECOGNITION OF ANCIENT GLACIATION ON to present a brief overview of the history of EARTH pre-Pleistocene glaciation, emphasizing the A notable feature of terrestrial ice sheets uncertainties inherent in the recognition and glaciers (in contrast to Martian ice and dating of ice ages in an incomplete geo- caps) is their tendency to flow plastically logic record; (2) to review the evidence for and, where basal temperatures are at the PRE-PLEISTOCENE GLACIATION ON EARTH 425 melting point, to slide over their beds [sum- ages and especially in determining whether marized by Boulton (1975)]. Debris is incor- terrestrial glaciation has occurred periodi- porated in the moving ice by basal erosion cally, episodically, or more or less continu- and following the gravitational emplace- ously during some intervals. However, the ment onto the ice of mass-wasted material chronology of ancient glaciation (Fig. 1) is from nunataks or valley sides. Subse- limited by our ability to recognize the quently, this debris is deposited directly results of glacial activity in an incomplete from the ice, after gravitational flow or after geologic record and by the age resolution partial sorting by water or wind (Boulton, attainable with available dating methods. 1972, 1975; Anderson et al., 1981). The Continental ice tends to be concentrated principal way of recognizing ancient glacia- at high latitude, and even during a major ice tion on Earth is therefore to identify the age, glacial debris is usually distributed assemblage of erosional features and de- over only a small part of the Earth's conti- posits generated by glacial activity. nental shelves and interiors. During the An important product of glaciation is Pleistocene, for example, the Scandinavian diamicton [or diamictite, where lithified; ice sheet spread southward nearly to lati- Flint et al. (1960a,b)], a poorly sorted sedi- tude 51°N (Andersen, 1981) and the Lau- ment in which clasts may range in size from rentide sheet reached latitude 40°N (May- boulders to clay-size material. Such diamic- ewski et al., 1981), but only about 29% of ton (diamictite) is generally termed till (til- the world's land area, including Antarctica, lite) where a glacial origin is well estab- was covered by ice (Flint, 1971). This is in lished [see Boulton (1976, 1980) for a recent spite of the fact that land is currently con- summary of terminology]. One difficulty is centrated at mid to high northern latitudes that not all diamictite is glaciogenic, and and near the south pole (Barron et al., 1981, superficially similar rocks may on close Table 2). About 48.3% of the Earth's land scrutiny contain no definite glacial imprint occurs between latitudes 35 and 90° (37.5%, (e.g., Crowell, 1957, 1964; Dott, 1961; in the northern hemisphere and 10.8% in Schermerhorn, 1974). the southern hemisphere), whereas the total Criteria for establishing the occurrence of area between these latitudes constitutes glaciation have been discussed by Crowell only 42.6% of the surface of the globe. It is (1964), Schwarzbach (1964), Harland et al. clear that if ice sheets have formed prefer- (1966), Spencer (1971), Schermerhorn entially at high latitude for most or all of (1974), and Flint (1975). Without reviewing Earth history, the continental glacial record all the arguments, the most useful of many should be sparse even if the Earth were gla- indicators are: (1) the occurrence of diamic- ciated continuously. tite over a large area; (2) the presence in In addition to limitations in the glacial rec- diamictite or associated conglomerate of ord imposed by this latitude factor, early- hard, (glacially) striated and faceted clasts, deposited till is removed or modified by the some of which are far-traveled; (3) the oc- advancing ice sheet, and the final complex currence beneath diamictite of a (glacially) assemblage of till, outwash, and lacustrine polished and striated pavement; and (4) in sediment is subject to postglacial erosion. subaqueous deposits, (ice-rafted) ~'drop- Glaciogenic sediment tends to accumulate stones" in finely laminated fine-grained unevenly, depending on factors such as rock matrix. subglacial topography and the temperature, velocity, and thickness of the ice (Boulton DATING ANCIENT GLACIAL EVENTS ON et al., 1977). If preserved in the interior of a EARTH continent, numerous unconformities are The dating of glacial events on Earth is present, and it may be possible to decipher crucial to the elucidation of the causes of ice only the youngest history in any detail (e.g., 426 NICHOLAS CHRISTIE-BLIC K the Pleistocene tills of North America and domes of the northern hemisphere are northern Europe). Where deposited on cra- thought to have originated as ice shelves tonic continental crust, glacial sections are over inland seas or embayments [but see typically only a few tens to hundreds of me- Boulton (1979) for an opposing viewpoint]. ters thick (Flint, 1971 ; Crowell and Frakes, Cenozoic climatic fluctuations have been !970). Tillites of continental interiors are inferred in great detail from the examina- therefore more likely to be eroded than tion of the distribution of fossil taxa in oce- temporally equivalent thick sections in ad- anic deposits and from oxygen isotopic jacent subsiding marine basins (Carey and analysis of foraminifera and molluscs [sum- Ahmad, 1961). For example, mostly marine marized in Savin (1977) and Frakes (1979)]. basinal glaciogenic rocks of Late Protero- During late Cenozoic time, glacial detritus zoic age occur at many localities in western was rafted and released by melting icebergs North America (Crittenden et al., 1972: over vast areas of the ocean basins. A com- Christie-Blick et al., 1980), but equivalent parable climatic record is not available for platform rocks, themselves in part marine, most of Earth history because virtually all are known only from northwestern Utah, oceanic crust and sediment cover of pre- where they overlie glacially sculptured to- Jurassic age has been subducted as a result pography (Crittenden et al., 1952; Blick, of large-scale movements of lithospheric 1979; Ojakangas and Matsch, 1980). Else- plates. Only scraps of this old oceanic crust where on this continental platform any rec- are preserved as ophiolites in orogenic belts ord of glaciation has been eroded at a re- (Dewey and Bird, 1970). The pre-Jurassic gionally extensive sub-Cambrian uncon- record of ancient glaciation is therefore lim- formity. There are of course some excellent ited exclusively to the continental crust. examples elsewhere of well-preserved, The chronology of glaciation prior to the nonmarine deposits, particularly in the Pa- Cenozoic has been established principally leozoic. For example, continental tillites by a combination of paleontology and iso- occur in both the Proterozoic and Ordovi- topic dating. However, fossils are generally clan sections of the Taoudeni Basin in west rare in glaciogenic beds, and it is often diffi- Africa (Deynoux and Trompette, 1976; cult to relate nonmarine sections displaying Deynoux et al., 1972). the best evidence for glaciation to the Thick marine sections tend to contain the worldwide time scale (Crowell, 1978). The most complete history of advancing and re- problem becomes more acute in pre- treating ice sheets, but generally do not re- Phanerozoic rocks. In spite of remarkable cord the beginning and end of a glacial advances in the study of stromatolites (Hof- event. An ice sheet may already have been mann, 1973: Walter, 1976; Awramik et al., in existence for several million years before 1979), their use in biostratigraphy is limited its margin reaches sea level (e.g., the Ceno- by their long age ranges, as well as by their zoic ice sheet in Antarctica), and an ice cap environmental sensitivity. Microfossils may may survive for some time after the last prove useful in dating latest Proterozoic recognizable marine glaciogenic sediments glaciations, although Knoll et al. ( 198 I) sug- are deposited (Crowell, 1978). This can be a gested that one supposed Vendian (650-570 significant problem in dating Phanerozoic Ma) form, Bavlinella faveolata, may be an ice ages, although it is less important in the indicator of glacially induced ecological Proterozoic, for which less satisfactory age stress. For Paleozoic beds, fossil ages are resolution is possible. In addition, it is pos- related to the isotopic time scale by direct sible that not all ice sheets originate in con- correlation in many sequences (e.g., tinental interiors, according to Denton and Harland et al., 1964; Harland and Francis, Hughes (1981), who proposed a controver- 1971; Van Eysinga, 1975). For pre-Phanero- sial model in which the late Pleistocene ice zoic rocks isotopic ages must come from PRE-PLEISTOCENE GLACIATION ON EARTH 427 the sequences in which the glacial strata are leoslope and ice rafting is not required to observed. In these cases, well-defined ages, explain the lonestones. The glacial interpre- such as those on basement, commonly pro- tation of these beds must therefore be re- vide only the coarsest bracketing. Ages on garded as possible but not proven. rocks within a particular sequence (e.g., K- Diamictite of the Witwatersrand Super- Ar and Rb-Sr determinations on volcanic group [2.7-2.6 Ga old; Van Njekirk and rocks, shale, or sedimentary minerals such Burger (1978) was interpreted as tillite by as glauconite) are generally tentative and Wiebols (1955)] and Fuller (1958), and a gla- subject to significant errors. cial origin was tentatively accepted by In conclusion, the timing of ancient glaci- Frakes (1979) and Harland ( 198 la, b). How- ation is uncertain as a result of an incom- ever, recent work by W. E. L. Minter (K. plete geologic record and the limited age A. Eriksson, personal communication, resolution attainable. Although an overall July, 1981) suggests that the diamictite may chronology has been established, the times represent nonglacial mudflows. A pa- of initiation and termination of ice ages are leomagnetic study of the Ventersdorp lavas mostly poorly known, and climatic fluctua- (2.6 Ga old), which overlie the Witwa- tions on the scale of glacial and interglacial tersrand Supergroup, suggests that the intervals generally cannot be dated directly diamictite was deposited at high latitude or in pre-Cenozoic rocks. prior to an interval of rapid apparent polar wander [see Fig. 1 of Piper (1976)]. How- SUMMARY OF KNOWN ICE AGES ever, paleomagnetic results from such ancient rocks are difficult to interpret, and the a. Archean (4.65-2.5 Ga) climatic significance of the Witwatersrand Archean terranes of low metamorphic diamictite remains in doubt. grade (the greenstone belts) occur on all continents (Windley, 1977), but alleged til- b. Proterozoic (2.5-0.57 Ga) lites of this age are rare. Two reasonably 1. Early Proterozoic (2.5-1.6 GaL a Well- well-dated examples of diamictites for established tillites are known from two in- which a glacial interpretation was suggested tervals in the Proterozoic. The older inter- recently occur in Montana and South Af- val, best represented in North America, rica, but in neither case is a glacial origin includes the Huronian Supergroup [2.5-2. I convincingly established. Ga; summarized by Roscoe (1973)] of On- Metadiamictite occurs in the wall-rocks tario and correlative sequences in Quebec, of the Stillwater Complex in southwestern the Northwest Territories, Michigan, and Montana, and has been dated as at least Wyoming (Young, 1970, 1973; Long, 1974; 2.75 Ga old and possibly older than 3.14 Ga Karlstrom and Houston, 1979). Huronian [U-Pb determination on zircon; Nunes and tillites occur in three groups separated by Tilton (1971); Page (1977)]. The diamictite unconformities, and both terrestrial and is locally as thick as 100 m and extends dis- marine facies are represented. Morris continuously along strike for more than 20 (1977a) obtained a paleolatitude of 35 ° for km. The principal evidence for glacial activ- the uppermost group of the Huronian (mean ity cited by Page and Koski (1973) is inclination of magnetization direction A = the occurrence of supposed dropstones. I 54°). Higher paleolatitudes of 50-87 ° [Sy- briefly examined the rocks in 1980, and ob- mons (1975), and references in Steiner and served that the diamictite locally contains Grillmair (1973)] probably correspond to a abundant deformed sedimentary fragments and that lonestones (isolated clasts) are as- a Lower, Middle, and Upper Proterozoic are used in sociated with soft sediment deformation. the sense of Harrison and Peterman (1980) for intervals Both features suggest emplacement on a pa- limited by ages of 2.5, 1.6, 0.9, and 0.57 Ga. 428 NICHOLAS CHRISTIE-BLIC K secondary magnetization or direction B of glacial component and the need for sub- Morris (1977a). stantial ice sheets [e.g., Blick (1979); Lower Proterozoic tillites also occur in Christie-Blick et al. (1980); and summa- the Griqualand West and Transvaal Basins rized in Hambrey and Harland (1981)]. One of South Africa, where they have been enigmatic feature of the tillites is their close dated as older than 2.2 Ga [Rb-Sr isochron association with carbonate rocks, including on andesite; Burger and Coertze (1973- dolomite, normally considered an indicator 1974)] and younger than the Ventersdorp la- of warm-water sedimentation (e.g., Spen- vas (Visser, 1971, 1981; De Villiers and Vis- cer, 1971). Two explanations have been ser, 1977). The tillites appear to have been proposed. The first is that carbonates may deposited at high paleolatitude (Piper, 1976, be deposited in cold water (Carey and Fig. 1). Ahmad, 1961; Bjcrlykke et al., 1978), al- Diamictites recognized in the Bijawar though the timing and conditions for dolo- Group of central India (Mathur, 1981) and mitization have not generally been estab- in the Hamersley Basin of Australia (Tren- lished. A second explanation calls for rapid dall, 1976) are of questionable glacial origin and extreme climatic fluctuations (Spencer, and are only approximately dated as Early 1971; Williams, 1975a). Unfortunately, at- Proterozoic. tempts at measuring paleotemperatures us- 2. Middle to Late Proterozoic (1.6-0.57 ing stable-isotope geochemistry have pro- Ga). The younger interval of the Protero- duced equivocal results (Williams, 1979, zoic with documented tillites extends from 1981; Donnelly, 198l). about 1 Ga ago to Cambrian time (570 Ma). Perhaps the most problematic feature of Glacial beds of this age are known from all the Middle to Upper Proterozoic tillites is continents with the possible exception of the paleomagnetic evidence suggesting Antarctica [but see Neethling (1970), equatorial glaciation (Piper, 1973: Tarling, quoted by Crowell (1975)], and they occur 1974; Morris, 1977b: McWilliams and mainly but not exclusively in marine ba- McElhinny, 1980). Four explanations have sinal sections. Available age determinations been suggested. The first is that the time- appear to cluster into three groups: -1000- averaged position of the virtual geomag- 850 Ma, -820-740 Ma, and -655-575 Ma netic pole does not approximate the geo- (Williams, 1975a), and many sequences graphic pole, although this is contrary to contain two major tillite units. However, considerable Phanerozoic evidence (Run- the climatic significance of these wide- corn, 1964; Briden and Irving, 1964; spread glacial beds has been very contro- McElhinny, 1973; Drewry et al., 1974). A versial, and it is appropriate in an overview second explanation is that for these sedi- of Earth's glacial history to discuss some of mentary rocks, the remnant magnetization the disagreements. measured significantly postdates the time In many marine sections, it is clear that of sedimentation, so that with sufficiently some diamictite is probably nonglacial, a rapid plate motion (e.g., McElhinny et al., product of widespread continental rifting in 1974), the tillites could have been deposited Middle to Late Proterozoic time, and at higher latitudes (Crowell, 1975). How- Schermerhorn (1974, 1975) has argued ever, this idea does not account for poles forcefully that the diamictite is mostly of determined on igneous rocks, or for con- tectonic rather than glacial origin. How- sistently low latitudes of thick sedimentary ever, reappraisal of many of the alleged sections (e.g., McWilliams and McEIhinny, tillites during the last decade, including 1980), and it does not appear to be rooted in those deposited in continental interiors any firm evidence. A third possibility is that (Deynoux and Trompette, 1976), confirms there were two or three discrete, simultane- most earlier interpretations of a significant ous, prolonged ice ages sufficiently extreme PRE-PLEISTOCENE GLACIATION ON EARTH 429 for ice sheets to advance into the tropics Silurian tillites occur principally in South (Harland, 1964a,b). However, there is no America, where they range in age from evidence for the very large sea-level fluctu- Early Llandoverian in the Amazon Basin ations (several hundreds of meters) that (Caputo et al., 1972; Carozzi, 1979) to should have accompanied glacierization of Wenlockian in Bolivia, Peru, and Argentina most of the continents. In addition, such (Crowell et al., 1980). large ice sheets would have increased Evidence for Devonian glaciation is lim- global albedo so much that it is debatable ited, but is best displayed in the Amazon whether the ice would have melted again. A Basin, where probable tillites are of Fras- fourth explanation of the paleomagnetic nian-Strunian age (Caputo eta/., 1972: results is that during late Proterozoic time, Carozzi, 1979). the Earth experienced an obliquity pertur- The Late Paleozoic ice age is recorded on bation great enough (~ > 54°) for low and all the continents of Gondwana, namely, middle latitudes to be glaciated in prefer- South America and the Falkland Islands, ence to the poles (Williams, 1975a). The Africa and Madagascar, India, Australia, main difficulty with this hypothesis is the and Antarctica (Crowell and Frakes, 1975). lack of firm evidence for obliquity radically The earliest evidence of this glaciation is different from the present at other times (al- found in Visean (Upper Mississippian) though see Williams, 1972, 1973; Williams rocks of South America (Crowell, 1978). and Austin, 1973). During Pennsylvanian time, ice centers de- In summary, there is evidence for wide- veloped in many parts of the then intact spread glaciation during a span of about 400 supercontinent, but by the Early Permian, Ma in Middle to Late Proterozoic time, glaciation had virtually ceased in South with ice sheets extending to sea level, but America. In contrast, Australian tillites are the details of timing and the significance of most widespread in the Early Permian the problematic paleomagnetic observa- (Crowell and Frakes, 1971), and on that tions are yet to be worked out. continent there is evidence for ice rafting in rocks as young Kazanian (early Late Per- c. Paleozoic (570-230 Ma) 4 mian). The overall picture is one of many Evidence for glaciation is known princi- ice centers waxing and waning during a pally from three intervals in the Paleozoic: time span of about 90 Ma, but with the ice Late Ordovician to Late Silurian (440-415 age lasting somewhat less than this at any Ma), latest Devonian (360-345 Ma), and particular locality. Permo-Carboniferous (335-245 Ma). Or- Independent paleomagnetic data for the dovician tillites are best preserved in north Ordovician-Permian interval have been in- and west Africa, where they are mainly of terpreted in several somewhat different Ashgillian age (Beuf et a/., 1971; Des- ways (McElhinny eta/., 1974; Schmidt and tombes, 1968; Deynoux et al., 1972; Dey- Morris, 1977; Morel and Irving, 1978), but noux and Trompette, 1981; Tucker and indicate that ice centers formed preferen- Reid, 1973). The ice sheet was located be- tially at high paleolatitude [mostly within tween the Saharan region and the Guinean 30-40 ° of the south pole; Scotese eta/. Gulf. Upper Ordovician tillite also occurs in (1979)]. The progression in ages of tillites South Africa, with an ice sheet coming from Ordovician in North Africa to Per- from the north and west (Rust, 1981). Late mian in Australia roughly corresponds to Ordovician glaciation in Africa may have the apparent wander path of the virtual geo- persisted into the Early Silurian. However, magnetic pole. d. Cenozoic (65-0 Ma) 4 Absolute ages for the Phanerozoic are mainlyfrom Van Eysinga(1975). The earliest direct evidence of Cenozoic 430 NICHOLAS CHRISTIE-BLIC K glaciation in the southern hemisphere con- to locate Permo-Carboniferous ice sheets sists of Upper Oligocene(?) (26 Ma) gla- with some success by Crowell and Frakes ciomarine sediments in the Ross Sea [(1975), and in a succession of papers deal- (Hayes and Frakes, 1975; Barrett, 1975) ing with each of the Gondwana continents]. and in the northern hemisphere of Middle However, the size of ice sheets is difficult Miocene sediment in the Gulf of Alaska to estimate even with widely distributed de- (Plafker and Addicott, 1976). Geitzenauer posits. With worse age resolution and an et al. (1968) and Margolis and Kennett even more fragmentary record, typical of (1971) interpreted sand grains in South Pa- the Proterozoic ice ages, it becomes virtu- cific sediments, thought to be of Early Eo- ally impossible to reconstruct ice sheets at cene age, as ice-rafted, but this suggestion any particular time. Similarly, the lengths is not corroborated by apparently low tem- of ice ages are not well known in pre- perature gradients near Antarctica [discus- Permo-Carboniferous rocks, and it is sion in Frakes (1979)]. From analysis of likely, for example, that several distinct D.S.D.P. samples in the southern ocean, Proterozoic events are currently not re- Lazarus and Hayes (1981) concluded that solved by the few dates available. More- ice rafting was insignificant before the Plio- over, both the lengths of ice ages and the cene. Presumably, however, a substantial areal distribution of deposits are likely to be ice sheet already existed in Antarctica for measures of the paleolatitudinal position of several million years prior to the arrival of continents rather than purely a function of ice at sea level around much of the conti- global climate. The thickness of deposits in- nent. Corroboration for this, in addition to dicates subsidence and sedimentation rates the Ross Sea evidence already mentioned, rather than the scale of glaciation. Sedimen- is the rapid sea-level fall, indicated by the tation rates can vary by several orders of marked Late Oligocene (29 Ma) seaward magnitude from a few meters per million shift in coastal onlap observed in seismic years in some nonmarine deposits to 1000 profiles from many parts of the world (Vail m/Ma in the Gulf of Alaska, for example and Hardenbol, 1979; and Fig. 2). Details of (Plafker and Addicott, 1976). Isolated pa- climatic fluctuations during the late Ceno- leomagnetic determinations of tillite pa- zoic Ice Age are reviewed elsewhere in this leolatitude do not constitute a good indica- volume (Imbrie, 1982). tor of the scale of glaciation because the growth of ice sheets is very much influ- RELATIVE SIGNIFICANCE OF ICE AGES enced by geographic factors affecting pre- It would be desirable in an evaluation of cipitation, such as the distribution of land Earth's climatic history to assess the rela- and sea, patterns of atmospheric circula- tive significance of the ice ages, but, unfor- tion, and topography. Furthermore, it is not tunately, a quantitative comparison is diffi- clear that the enigmatic low paleomagnetic cult in part for reasons already discussed in latitudes associated with some Proterozoic the section on dating glacial events. Possi- tillites necessarily reflect global climates ble criteria that might be used to compare colder than those during other ice ages. Fi- ice ages are the areal extent of former ice nally, glacioeustatic changes in sea level sheets inferred from the paleogeographic are distinguished with difficulty from other reconstruction of glacial deposits, the nonglacial eustatic controls or from more lengths of ice ages, the thickness of de- local changes in the rates of subsidence and posits, paleolatitudes of tillites, and the sedimentation that affect relative sea level. magnitudes of glacioeustatic changes in sea In this regard, seismic stratigraphy (Vail et level. However, none of these criteria is al., 1977a) has revolutionized the analysis particularly reliable. of eustatic sea-level changes, especially for Paleogeographic reconstruction was used Mesozoic and Cenozoic time. PRE-PLEISTOCENE GLACIATION ON EARTH 431

~ ! A6ES - RISINGIIJ,[ , I , , , [ IF--GI " !~ i ' Q_ 0 RIO, MESSINIAN' '=' --- ...... Td - TORTONIAN '~, TM3,1 __ SERRAVALLIAN _~ %=_ LANGHIAN ~ .-11~..F_ Tc BURDIGALIAN '~-~ ._..__.._...~_ TM12

- :-T0 z)-- -

RUPELIAN ~ TOt _ PRIABQNIAN ~___ IT.3 TO 8ARTONIAN _.~ LUTETIAN TEZl YPBESIAN -~ R~ W THANETIAN ...... -'__~-- --~'~TPZ2 Ta - --TF'LT" ==. DANIAN ~ TP1 -- MAASTRICHTIAN I ~ ~ ' - c~~-z CAMPANIAN ~ 100 -- z SANTONIAN CONIACIAN TURONIAN CENOMANIAN

~ BARREMIAN ~ LIEVB. HAUTERIVIAN c) VALANGINIAN ~._ I(1.1 ~' BERRIASIAN PORTLANDIAN I TITHON- 302 KIMME RI DGIANL~U~u.-- ..... -- OXFORDIAN ,13.1 CALLOVIAN ~_ J BATHONIAN J~ -- BAJOCIAN ~ ~1 __ AALENIAN .~ ---- - TOARCIAN PLEINSBACHIAN J1 -- SINEMURIAN HETTANGIAN RHAETIAN NORIAN

FiG. 2. Chart of global relative changes of sea level since the Triassic as indicated by relative changes of coastal onlap. Cretaceous cycles (hatchured area) have not been released for publication (Vail et a/., 1977b). 432 NICHOLAS CHRISTIE-BLIC K

INTERVALS WITHOUT IMPORTANT NH3 or CO2-H20 cloud (Sagan and Mullen, GLACIATION 1972; Sagan, 1977; Walker, 1976, 1977; There appear to be three intervals in Hart, 1978; Owen et al., 1979; and Pollack, Earth history for which there is minimal or 1979). However, a considerable range of no evidence for glaciation. These are: (1) climates may be envisaged for the early from the age of the oldest known rocks at Earth, and it is possible that evidence for about 4 Ga to around 2.5 Ga ago; (2) from Archean glaciation has simply been mostly 2.1 Ga to approximately ! Ga ago; and (3) or completely eradicated. from 245 Ma to 26 Ma ago or slightly ear- b. Earl3, to Middle Proterozoic Interval lier. (2.1-1.0 Ga) The interval from 2.1 to 1.0 Ga is charac- a. Archean Interval (4-2.5 Ga) terized by the presence of abundant carbon- There are few climatic indicators in sedi- ate rocks (Ronov, 1964), suggestive of mentary rocks older than 2.5 Ga and few warm-water marine sedimentation, and magnetic data with which to determine pa- widespread evaporites, consistent with lo- leolatitudes. Knauth and Epstein (1976) cally hot dry climates. Chert geothermome- used stable-isotope geochemistry to infer try suggests paleotemperatures in the range pre-Phanerozoic paleotemperatures as high 20-52°C (Knauth and Epstein, 1976). as 70°C for 3-Ga-old cherts in South Africa. Available paleomagnetic data indicate that However, it is not clear to what extent mea- several continents experienced large latitu- sured isotopic ratios reflect Archean oce- dinal shifts and at different times occupied anic temperature, the Archean isotopic both polar and equatorial positions (Irving composition of sea water, or the effects of and McGlynn, 1976; Piper, 1976; Morris, later diagenesis, hydrothermal alteration, 1977a). The absence of evidence for glacia- metamorphism, and isotopic exchange with tion during this long interval therefore sug- meteoric water over geologic time (Perry gests a period of persistently warm climate and Tan, 1972; Knauth and Epstein, 1976). on a global scale. On the other hand, a paleotemperature of 70°C is not incompatible with the probable c. Late Permian to Mid-Cenozoic Interval temperature tolerances of Archean living (245-26 Ma) organisms (Knauth and Epstein, 1976). The distribution of climatically sensitive From a theoretical standpoint, Archean sedimentary rocks on paleomagnetically re- climate would have been strongly con- positioned continents (Drewry et al., 1974), trolled by solar luminosity and by atmo- the absence of tillites, the wide latitudinal spheric composition. In spite of questions ranges of invertebrate and plant fossils, and raised by the low neutrino flux observed stable-isotope geothermometry suggest a from the Sun, Newkirk (1980) regards the globally equable climate for the entire inter- theory of stellar evolution--which predicts val from Late Permian to early Cenozoic a solar luminosity 4.6 Ga ago of about 70% ]summarized by Frakes (1979)]. Isolated its contemporary value--as robust. For stones, possibly dropstones, occur in Juras- current albedos and atmospheric composi- sic and Cretaceous deposits of northern Si- tion, the global mean temperature of the beria, in the Cretaceous of Svalbard, Earth would have been less than the freez- Alaska, and New Zealand, and in Pa- ing point of water about 2 Ga ago. The ab- leogene sediments of Svalbard, the sence of evidence for widespread Archean U.S.S.R., and New Zealand, all these re- glaciation is probably best explained by gions located at high paleolatitude [sum- considering the greenhouse effect of a prim- marized in Hambrey and Harland (1981) itive atmosphere composed of CH4 and and Collins (1961)]. Some of the stones may PRE-PLEISTOCENE GLACIATION ON EARTH 433 have been rafted by sea ice, but there is no 1973; Williams, 1975b; Mitchell, 1976). For direct evidence for significant glaciation. example, Steiner (1967) proposed a model The mid-Jurassic so-called Mawson Tillite for producing variable solar luminosity of Antarctica (Warren, 1962) is probably a based on the idea that the universal gravita- nonglacial volcanic breccia (Borns and tional "constant" (G) is a function of time Hall, 1969). and space. According to the model, solar In conflict with the impressive evidence luminosity varies because the Sun experi- favoring generally warm climates during the ences a variable gravitational function in its Mesozoic and Paleogene are some globally elliptical orbit. Ice ages supposedly occur synchronous and rapid shifts in coastal on- once per cosmic year [280 Ma in Steiner lap, determined by seismic stratigraphy, (1967)], when the Sun is at the perigalacti- that suggest sea level fluctuations that are in cure, or closest to the galactic center (ap- part glacially controlled, because the rates proximately its current location). Another involved appear to be too rapid for tectonic model, which predicts ice ages with the mechanisms alone (Vail et al., 1977b; Vail same periodicity, invokes temporary varia- and Hardenbol, 1979; Kerr, 1980; and Fig. tion in the Sun's radiation during passage of 2). Most of the ice responsible must have the solar system through interstellar clouds been situated in Antarctica, consistently in the spiral arms of the galaxy (e.g., Mc- positioned at high latitude, because this is Crea, 1975; Talbot et a/., 1976). Williams the only large continent where the evidence (1975b) suggested that ice ages have oc- can remain effectively hidden. Two impor- curred with a period of about 155 Ma since tant limitations on the size of a possible ice the Late Proterozoic, or twice per cosmic sheet are that it did not spread onto still year (estimated as ....ana+65 51 Ma). In his model, attached fragments of Gondwana and that it solar luminosity is reduced twice per orbit probably did not reach sea level. The latter as the Sun passes through diametrically op- is indicated by relatively warm oceanic bot- posite galactic flexures (Table I). Steiner tom water (Savin, 1977) and by the appar- and Grillmair (1973) obtained a betterfit be- ent absence of rafted detritus in oceanic tween the timing of ice ages and the cosmic sediments of Mesozoic age. The present year by cunningly abandoning strict period- Antarctic ice sheet is equivalent to about 55 icity. In their model, the Sun spirals inward m of water spread over the world ocean in such a way that the duration of the cos- (Denton eta/., 1971), suggesting that a Me- mic year decreases from 400 Ma in the Ar- sozoic ice sheet would have been equiva- chean to 280 Ma today. lent to somewhat less than 50 m of water. A The inferred timing of ice ages and times significant aspect of the seismic evidence is predicted by the models of Williams (1975b) that the presence of even a small ice cap and Steiner and Grillmair (1973) are shown during the demonstrably warm Mesozoic in Table I. It is clear from both Table I and Era suggests that an ice sheet may have Fig. 1 that the occurrence of ice ages is not been a relatively persistent feature near the strictly periodic, and even focusing only on south pole throughout the Phanerozoic. the last 1 Ga, Williams (1975b) incorrectly predicts significant glaciation at 145 Ma ago PERIODIC AND EPISODIC CLIMATIC (latest Jurassic). Steiner and Grillmair FLUCTUATIONS ON EARTH (1973) account for the glacial history of the Several authors have suggested that ter- last 1 Ga, but incorrectly predict ice ages at restrial ice ages occur periodically either 1119, 1489, 1873, 2660, and 3060 Ma for once or twice per cosmic year, the time re- which there is no evidence. On the other quired for the Sun to execute one orbit hand, major ice ages do appear to occur around the center of the galaxy (Umbgrove, episodically with a charactertistic time of 1947; Steiner, 1967; Steiner and Grillmair, 108-109 years, requiring overall controls on 434 NICHOLAS CHRISTIE-BLIC K

TABLE I

COMPARISON OF TIMING OF MAJOR ICE AGES AND QUASI-PERIODIC SOLAR ORBITAL PARAMETERS SUGGESTED BY WILLIAMS (1975b) AND STEINER AND GRILLMAIR (1973)

Timing of ice ages (Ma) Timing of Timing of solar orbital solar This paper Williams (1975b) Steiner and parameter (Ma) apsides (Ma) Grillmair (1973) Williams (1975b) Steiner and Grillmair (1973)

13_+ 13 -10 ± 30 7 ± 7 -lO -8_+4 145 290 ± 45 295 ± 55 288 ± 55 300 280 352 ± 7 427 ± 12 445 ± 5 440 ± 40 455 437 615 ± 40 615 _+ 40 616 ± 30 610 595 780 ± 40 770~]g 777 ± 40 765 766 925 ± 75 940~ 950 ± 50 920 937 1119 1489 1873 2300_+ 200 2288 _+ 87 2265 2660 3060

the Earth's climate that are effective on the al., 1981). Examination of the Mesozoic same time scale (Mitchell, 1976). Among and Cenozoic chart of global relative these are possible periodic and merely epi- changes in sea level (Fig. 2; Vail et al., sodic variations in solar output, long-term 1977b) suggests episodic, abrupt glacioeu- changes in atmospheric composition, and static lowering of sea level with a character- the rearrangement of continents and oceans istic time of 1-10 Ma. Among possible pri- brought about by the movement of litho- mary controls for these climatic variations spheric plates (Crowell and Frakes, 1970; are changes in atmospheric composition Pollack, 1979). Since several controls are (Fischer, 1982a), tectonism, and variations significant on this time scale, and since the in solar output, although effective feedback controls are probably largely stochastic, it mechanisms are definitely required to ex- is hardly surprising that the resulting cli- plain the apparently rapid growth and disin- matic variation is not periodic. tegration of small ice sheets in the generally Climatic variations with characteristic warm Mesozoic. times of 10e-107 years are indicated by sev- An explanation for the waxing and wan- eral lines of evidence. Dorman (1968) pro- ing of Pleistocene ice sheets on a time scale posed a 30-Ma periodicity in global temper- of 104-105 years has long been sought in a atures based on oxygen isotope analyses of variety of factors, such as variations in so- Cenozoic molluscs. Fischer and Arthur lar output, the absorption of solar energy by (1977) suggested that the expansion and de- interstellar dust, in the seasonal and latitu- cline of diversity in Mesozoic and Cenozoic dinal distribution of incident radiation due pelagic marine organisms, on a time scale of to changes in the Earth's orbital parame- 32 Ma, was in some way related to climate ters, atmospheric dust, and in the Earth's change. Equivalent but less regular events magnetic field [summarized by Hays et al. appear to have characterized the Paleozoic (1976)]. The application of time series anal- seas on a time scale of 4-46 Ma (Leggett et ysis to the climatic record in deep sea cores PRE-PLEISTOCENE GLACIATION ON EARTH 435 has now demonstrated that glacial and in- well established, but analysis of cratering terglacial stages can be largely, although statistics suggests that most channels not completely, explained by orbital pertur- formed over an extended period some time bations (Hays et al., 1976; Imbrie and Im- between 3.5 and 0.5 Ga ago (Masursky et brie, 1980; and Imbrie, 1982).The most im- al., 1977). The smallest channels, termed portant periodicities are 413, 100, 41, 23, gullies by Pieri (1976), are common over and 19 Ka. Observations in many well- much of Mars, but appear to be absent on dated rhythmically bedded sedimentary se- the youngest geologic features, such as quences, ranging in age from latest Protero- Tharsis. Morphologic contrasts suggest that zoic to Cretaceous, suggest that periodic channels are of several different origins orbital variations may have affected climate (Sharp and Malin, 1975), but most explana- for at least the entire Phanerozoic (Arthur, tions require that past climatic conditions 1979; Fischer, 1980, 1982a, b). However, on Mars were quite unlike those of today. the relative importance of the various or- Morphologic features such as dendritic bital parameters in controlling climate has patterns, sinuosity, and braiding suggest probably varied through geologic time that some of the channels were cut by run- (Fischer, 1982a" Ward, 1982). ning water, although liquid water is not now stable on the Martian surface (Sagan et al., GEOLOGY OF MARS AND EVIDENCE FOR 1973; Sharp and Malin, 1975; Masursky et CLIMATE CHANGE al., 1977). Gullies constitute the best evi- Mars is divided into two morphologically dence for rainfall or atmospherically in- distinct hemispheres by a great circle inter- duced melting of permafrost (Pollack, secting the equator at roughly 35° (Mutch et 1979). They are ubiquitous in both crater- al., 1976). The southern hemisphere is dom- free terrain and on crater rims in areas lack- inated by cratered terrain and basins, prob- ing much evidence for volcanic activity, ably older than 3.8 Ga (Head and Solomon, and are thus not a result of the melting of 1981), although there is abundant evidence permafrost by volcanism or meteoritic im- for continued volcanic modification of the pact. surface following the period of heavy bom- The origin of the largest channels, termed bardment early in Martian history. Much of outflow channels by Sharp and Malin the northern hemisphere consists of plains (1975), is controversial (Cutts and Blasius, of lower elevation, underlain by volcanic 1981). One possibility is that they were cut lava, but punctuated by giant shield volca- by catastrophic floods resulting from the noes. The most prominent of these are con- sudden release of subsurface water (e.g., centrated in the Tharsis plateau, 4000 km in Masursky et al., 1977; Carr, 1979). How- diameter and rising more than 10 km above ever, Nummedal (1978) suggested that the Mars datum (Head and Solomon, 1981). outflow channels were eroded by debris The principal evidence suggesting that sig- flows, and Cutts and Blasius (1981) pro- nificant climate changes have occurred on posed an eolian model. Whatever the pro- Mars consists of widespread dry channels cess, it does not seem to be operative on and of geologically young rhythmically lay- Mars now. ered terrains in the polar regions (Pollack, 1979). b. Polar Layered Terrains The polar regions of Mars are underlain a. Channels by extensive layered terrains consisting of Channels observed on Mars range in wind-blown dust and water-ice (Cutts, scale from on the order of l0 s km long and 1973). These deposits are among the young- 102 km wide to 10 km long and 1 km wide est features of the Martian surface, because (Sharp and Malin, 1975). Their ages are not fresh impact craters are absent (Cutts et al., 436 NICHOLAS CHRISTIE-BLIC K

1976). The layering appears to be climati- might be preserved under conditions of low cally induced and related to pronounced atmospheric pressure would depend on variations in the orbital parameters of Mars whether dust liberated from the ice was on a time scale of 105-106 years (Ward, subsequently recycled by eolian activity or 1974, 1979; Pollack, 1979; Toon et al., effectively buried. Toon et al. (1982) have 1982). If this is true, the layered terrains shown that water-ice could melt in wind- could have accumulated in only a few mil- sheltered regions during times of high or- lion years. bital obliquity if atmospheric pressure were Several models have been proposed to substantially increased. This suggests that account for the stratigraphic, topographic, another possible product of the disintegra- and geomorphic characteristics of the polar tion of polar layered terrains would be anal- deposits. These include the suggestion that ogous to outwash on Earth, but much finer dust was deposited during a single event grained. Howard (1982) has interpreted nar- (Cutts, 1973), models in which dust accu- row braided ridges conspicuous in the south mulates contemporaneously with ice abla- polar region of Mars as possible eskers, that tion (Howard, 1978; Squyres, 1979; Toon et is, fluvial sediments resulting from the basal al., 1982; Howard et al., 1982), and models melting of a thick ice cover, now ablated. involving episodic accumulation of dust and In contrast to ice sheets on Earth, Mar- ice (Cutts et al., 1976, 1979; Cutts and tian ice caps do not appear to flow, and it Lewis, 1982). However, none of the hy- has been assumed in the above discussion potheses yet explains all the observed char- that ancient polar terrains were similar to acteristics of the layered terrains (Blasius et present ones. It is not clear whether earlier al., 1982). layered terrains could have flowed and slid over their substrates, but such behavior DISCUSSION: IMPLICATIONS OF could have led to basal erosion and pro- TERRESTRIAL RECORD OF ICE AGES TO duced glacial deposits more similar to ter- ANCIENT CLIMATIC HISTORY OF MARS restrial ones. Possible controls on the ten- The geologic record and controls of cli- dency of polar terrains to flow are: (1) matic change on Earth are now reasonably composition (e.g., the relative proportions well known, and some progress can be of dust, water-ice, and CO2-ice): (2) layer- made toward understanding ancient Mar- ing characteristics; (3) the total thickness of tian climatic history by means of a compari- deposits [inferred to be currently 1-2 km in son of the two planets (e.g., Pollack, 1979). the south and 4-6 km in the north: Dzurisin Let us consider, for example, the possi- and Blasius, (1975)]; (4) lateral extent [cur- ble characteristics of fossil polar layered rently about 10° of latitude in the northern terrains on Mars and the likelihood that hemisphere; Cutts et al. (1979)]: (5) surface such deposits would be preserved. The rec- temperature [estimated to be 205°K in late ognition of ancient glaciation on Earth is summer near the north pole; Kieffer et al. based on the identification of erosional fea- (1976)]; and (6) gravitational acceleration tures and deposits generated by glacial ac- [371 gals on Mars, in comparison with 980 tivity. The sedimentary product of the cli- gals for the Earth; Mutch et al. ( 1976)]. matically induced disintegration of a The terrestrial lithosphere has been rela- Martian polar terrain would be quite unlike tively mobile for much of Earth history and its terrestrial counterpart and potential evidence for high-latitude glaciation is dis- characteristics would depend largely on the tributed over all the continents, even those mechanism of "deglaciation." Under the now in equatorial positions. In comparison, present atmospheric pressure of only a few large-scale lateral movements between millibars, ice sublimes rather than melts. lithospheric blocks do not seem to have oc- Hence the degree to which primary layering curred on Mars (Head and Solomon, 1981). PRE-PLEISTOCENE GLACIATION ON EARTH 437

However, two lines of evidence suggest sition (Pollack, 1979). The time at which that polar wandering may have taken place polar ice caps first formed on Mars may relative to the whole lithosphere. The first have been quite different from the time at is the observation of unique arcuate struc- which this event occurred on Earth. How- tures near the north pole of Mars (Murray ever, in view of Earth's long history of epi- and Malin, 1973). The second is the obser- sodic glaciation, it would be surprising if vation of layered terrains near the Martian polar ice caps had existed on Mars for only equator that are strikingly similar to polar a few million years. deposits, and may indicate former pole positions (Schultz and Lutz-Garihan, 1981). If SUMMARY polar wandering has occurred on Mars, an- Ice ages have occurred episodically on cient polar deposits should be sought on all Earth during at least the latter 50% of its parts of the planet, not merely near the history, and a small ice sheet may have present poles. been a persistent feature near the south On Earth, vertical tectonics affects cli- pole throughout the last 500 Ma. Major gla- mate by influencing atmospheric circula- cial events are known from the Early Pro- tion. Mars is sufficiently small that large- terozoic, several intervals of the Middle to scale vertical tectonics, such as that Late Proterozoic, the Ordovician-Silurian, associated with the Tharsis region, may possibly latest Devonian, Permo-Carbonif- have distorted the lithosphere sufficiently to erous, and late Cenozoic. A generally warm affect axial obliquity (Ward et al., 1979). It climate seems to have prevailed during the is therefore possible that changes in Mar- Archean, much of the Early to Middle Pro- tian climate could have been triggered by terozoic, and the Mesozoic to early Ceno- major tectonic events, since climate seems zoic. to be strongly affected by obliquity. On Earth, climatic fluctuations occur on The link between periodic changes in or- several time scales, but only those driven bital parameters and quasi-periodic climatic by perturbations of the Earth's orbital pa- fluctuations on Earth is now so convinc- rameters on a time scale of 1@-105 years ingly established by evidence from the are demonstrably periodic. The occurrence ocean floor (Imbrie, 1982) that an analogous of ice ages with a characteristic time of 108- explanation for layering in polar terrains on 109 years is explicable in terms of predicted Mars seems very credible. This is perhaps variations in solar output, inferred changes the most important implication of the ter- in atmospheric composition, and the rear- restrial record of ice ages to the climatic rangement of continents and oceans result- history of Mars. ing from the movement of lithospheric Finally, I will mention the importance of plates. Controls for climatic variations on a possible long-term changes in planetary at- time scale of 108-107 years are less well un- mospheres and solar luminosity to the evo- derstood, but may also be related to atmo- lution of climate. The initiation of glaciation spheric composition, tectonism, and varia- on Earth between 2.5 and 2.1 Ga ago ap- tions in solar output. pears to have been due, at least in part, to a An appreciation of the geological record gradual shift in the balance between a pro- and possible causes of ancient ice ages on gressive decrease in the effect of the atmo- Earth suggests nonuniformitarian as well as spheric greenhouse and increasing solar lu- uniformitarian ways of thinking about an- minosity. The same controls were probably cient climates on Mars. Fossil polar layered operative on Mars, although the Martian terrains could be preserved if the ice com- greenhouse effect was probably reduced by ponent sublimed and layering within the the lowering of atmospheric pressure, as dust was not destroyed by eolian activity. well as by a change in atmospheric compo- Melting of polar terrains under conditions 438 NICHOLAS CHRISTIE-BLIC K of greater atmospheric pressure would pro- Sheets (G. H. Denton and T. J. Hughes, Eds.), pp. duce outwash deposits analogous to those 1-65. Wiley, New York. of Earth. Martian ice caps do not appear to ANDERSON, J. B., D. D. KURTZ, E. W. DOMACK, AND K. M. BALSHAW (1981). Glacial and glacial marine flow or slide now, but it is possible that they sediments of the Antarctic continental shelf. J. did in the past, and this could have pro- Geol. 88, 399-414. duced glacial deposits still more similar to ARTHUR, M. A. (1979). Sedimentologic and Geochem- those of Earth. The Martian lithosphere is ical Studies of Cretaceous and Pah, ogene Pelagic relatively stable, but if polar wander has Sedimentao' Rocks: The Gubbio Sequence. Ph.D. occurred, ancient polar deposits should be dissertation, Part I, Princeton University, Princeton, N.J. sought on all parts of the planet, not merely AWRAMIK, S. M., A. HAUPT, H. J. HOFMANN, AND near the present poles. M. R. WALTER (1979). Stromatolite bibliography 2. Abrupt changes in Martian climate may Precambrian Res. 9, 105-166. have been triggered by major tectonic BARRETT, P. J. (1975). Textural characteristics of Ce- events through changes induced in plane- nozoic preglacial and glacial sediments at site 270, Ross Sea, Antarctica. In Initial Reports ~t"the Deep tary obliquity. Periodic climate changes, Sect Drilling Project 28, pp. 757-766. U.S. Govt. such as suggested by layering in the polar Printing Office, Washington, D.C. terrains, seem to be due to more regular BARRON, E. J., S. L. THOMPSON, AND S. H. fluctuations in the orbital parameters. Polar SCHNEIDER (1981). An ice-free Cretaceous? Re- ice caps probably first formed on Mars as a suits from climate model simulations. Science 212, 501-508. result of a gradual shift in the balance be- BEUF, S., B. BIJU-DUVAL, O. DE CHARPAL, P. tween a progressive decrease in the effect ROGNON, 0. GARIEL, AND A. BENNACEF (1971). of the atmospheric greenhouse and increas- Les grbs da Paldozoique ln]brieur au Sahara. L'In- ing solar luminosity. By analogy with the stitute Frangais du P6trole, Paris. Earth, polar layered terrains have probably BJORLYKKE, K., B. BUE, AND A. ELVERHOI (1978). Quaternary sediments in the northwestern part of existed episodically on Mars for a consider- the Barents Sea and their relation to the underlying able time, but observational evidence sup- Mesozoic bedrock. Sedimentology 25, 227-246. porting this idea still needs to be assembled. BLAS1US, K. R., J. A. CUTTS, AND A. D. HOWARD (1982). Topography and stratigraphy of Martian po- ACKNOWLEDGMENTS lar layered deposits, h'arns 50, 139-159. BLICK, N. H. (1979). I thank K. T. Biddle, M. V. Caputo, J. C. Crowell, J. Stratigraphic, Structural and Pa- A. Cutts, W. B. Harland, J. M. G. Miller, and P. D. leogeographic Interpretation ~f Upper Proterozoic Glaciogenic Rocks in the Sevier Orogenic Belt, Snavely III for critical reviews of this paper. J. A. Northwestern Utah. Ph.D. dissertation, University Cutts was especially helpful with his comments on of California, Santa Barbara. Martian climate and in supplying several pertinent ref- BORNS, H. W., JR., AND B. A. HALL (1969). Mawson erences. Exxon Production Research Company ap- proved publication and allowed me to attend the "Tillite'" in Antarctica: Preliminary report of a vol- NASA-sponsored "'Workshop on Quasi-Periodic Cli- canic deposit of Jurassic age. Science 166, 870-872. matic Changes on Earth and Mars," which lead to the BOULTON, G. S. (1972). Modern Arctic glaciers as depositional models for former ice sheets. paper being written. Ideas presented are in part the J. Geol. 128, 361-393. product of doctoral and postdoctoral studies com- Soc. London BOULTON, G. S. (1975). Processes and patterns of sub- pleted at the University of California, Santa Barbara, glacial sedimentation: A theoretical approach. I n supported by NSF Grants ATM 74-24201, EAR 77- Ice 06008, and EAR 78-15194 to J. C. Crowell. N. J. Page Ages: Ancient and Modern (A. E. Wright and F. Moseley, Eds.), pp. 7-42. Seel House Press, Liver- kindly pointed out in correspondence the best expo- pool. sures for examining diamictite in the wall rocks of the Stillwater Complex, which I visited in July 1980. 1 BOULTON, G. S. (1976). A genetic classification of tills thank A. G. Fischer for a preprint of his paper on long- and criteria for distinguishing tills of different origin. term climatic oscillations. Geografia 12, 65-80. BOULTON, G. S. (1979). A model of Weichselian gla- cier variation in the North Atlantic region. Boreas 8, REFERENCES 373-395. ANDERSEN, B. G. (1981). Late Weichselian ice sheets BOULTON, G. S. (1980). Till species--Terminology and in Eurasia and Greenland. In The Last Great Ice criteria for recognition. Quat. Newsl. No. 31. PRE-PLEISTOCENE GLACIATION ON EARTH 439

BOULTON, G. S., A. S. JONES, K. M. CLAYTON, AND CROWELL, J. C., AND L. A. FRAKES (1970). Phanero- M. J. KENNING (1977). A British ice-sheet model zoic glaciation and the causes of ice ages. Amer. J. and patterns of glacial erosion and deposition in Sci. 268, 193-224. Britain. In British Quaternary Studies: Recent Ad- CROWELL, J. C., AND L. A. FRAKES (1971). Late Pa- vances (F. W. Shotton, Ed.), pp. 231-246. Oxford leozoic glaciation: Part IV, Australia. Geol. Soc. Univ. Press (Clarendon), London/New York. Amer. Bull. 82, 2515-2540. BRIDEN, J. C., AND E. IRVING (1964). Palaeolatitude CROWELL, J. C., AND L. A. FRAKES (1975). The Late spectra of sedimentary palaeoclimatic indicators. In Paleozoic glaciation. In Gondwana Geology ( K. S. Problems in Palaeoclimatology (A. E. M. Nairn, W. Campbell, Ed.), pp. 313-331. Australian Na- Ed.), pp. 199-224. Interscience, New York. tional Univ. Press, Canberra. BURGER, A. J., AND F. J. COERTZE (1973--1974). Age CROWELL, J. C., A. C. ROCHA-CAMPOS, AND determinations, April 1972 to March 1974. Annu. R. SUAREZ-SORuCO (1980). Silurian glaciation in Geol. Surv. S. Aft'. 10, 135-141. central South America. In Gondwuna Five (M. CAPUTO, M. V., R. RODRIGUES, AND D. N. N. DE M. Cresswell and P. Vella, Eds.), pp. 105-110. A. VASCONCELOS (1972). Nomenclatura estratigrfifica A. Balkema, Rotterdam. da bacia do Amazonas hist6rico e atualiza9~to. An CUTTS, J. A. (1973). Nature and origin of layered de- XXVI Congr. Bras. Geol. Soc. Bras. Geol. 3, 35-46. posits of the Martian polar regions. J. Geophys. CAREY, S. W., AND N. AHMAD (1961). Glacial marine Res. 78, 4231-4249. sedimentation. In Geology of the Arctic (G. O. CUTTS, J. A., AND K. R. BLASIUS (1981). Origin of Raasch, Ed.), Vol. 2, pp. 865-894. Univ. Toronto Martian outflow channels: The eolian hypothesis. J. Press, Toronto. Geophys. Res. 86, 5075-5102. CAROZZl, A. V. (1979). Petroleum geology in the Pa- CUTTS, J. A., AND B. H. LEWIS (1982). Models of leozoic clastics of the Middle Amazon Basin, Brazil. climate cycles recorded in Martian polar layered de- J. Pet. Geol. 2, 55-74. posits, h'arus 50, 216-244. CARR, M. H. (1979). Formation of Martian flood fea- CUTTS, J. A., K. R. BLASIUS, G. A. BRIGGS, M. H. tures by release of water from confined aquifers. J. CARR, R. GREELEY, AND H. MASURSKY (1976). Geophys. Res. 84, 2995-3007. North polar region of Mars: Imaging results from CHRISTIE-BLICK, N., P. K. LINK, J. M. G. MILLER, G. Viking 2. Science 194, 1329-1337. M. YOUNG, AND J. C. CROWELL (1980). Regional CUTTS, J. A., K. R. BLASIUS, AND W. J. ROBERTS geologic events inferred from upper Proterozoic (1979). Evolution of Martian polar landscapes: In- rocks of the North American Cordillera. Geol. Soc. terplay of long-term variations in perennial ice cover Amer. Abstr. Progr. 12, No. 7, 402. and dust storm intensity. J. Geophys. Res. 84, 2975- COLLINS, F. R. (1961). Core tests and test wells, Bar- 2994. row area, Alaska. U.S. Geol. Surv. Prof. Pap. 305- DE VILLIERS, P. R., AND J. N. J. VISSER (1977). The K, 569-641. glacial beds of the Griqualand West Supergroup as CRITTENDEN, M. D., B. J. SHARP, AND F. C. CALKINS revealed by four deep boreholes between Postmas- (1952). Geology of the Wasatch Mountains east of burg and Sishen. Trans. Geol. Soc. S. Aft. 80, 1-8. Salt Lake City, Parleys Canyon to Traverse Range. DENTON, G. H., AND T. J. HUGHES (1981). The Arctic In Geology of the Central Wasatch Mountains (R. Ice Sheet: An outrageous hypothesis, in The Last E. Marsell, Ed.), pp. 1-37. Utah Geol. Min. Surv. Great Ice Sheets (G. H. Denton and T. J. Hughes, Guidebook No. 8. Eds.), pp. 437-467. Wiley, New York. CRITTENDEN, M. D., JR., J. H. STEWART, AND C. A. DENTON, G. H., R. L. ARMSTRONG, AND M. WALLACE (1972). Regional correlation of upper Pre- STUIVER (1971). The late Cenozoic glacial history cambrian strata in western North America. 24th Int. of Antarctica. In The Late Cenozoic Glacial Ages Geol. Congr. Rep. Sect. 1, 334-341. (K. K. Turekian, Ed.), pp. 267-306. Yale Univ. CROWELL, J. C. (1957). Origin of pebbly mudstones. Press, New Haven. Geol. Soc. Amer. Bull. 68, 993-1010. DESTOMBES, J. (1968). Sur la nature glaciaire des s6di- CROWELL, J. C. (1964). Climatic significance of sediments du groupe du 2 bme Bani, Ashgill sup6rieur de mentary deposits containing dispersed megaclasts. l'Anti-Atlas, Maroc. C.R. hebd. Seances Acad. Sci. In Problems in Palaeoclimatology (A. E. M. Nairn, Paris Set. D 267, 684-686. Ed.), pp. 86-99. Interscience, New York. DEWEY, J. F., AND J. M. BIRD (1970). Mountain belts CROWELL, J. C. (1975). The significance of glaciations and the new global tectonics. J. Geophys. Res. 75, in Precambrian correlation. In Symposium on the 2625-2647. Correlation of the Precambrian. Int. Geol. Corr. DEYNOUX, M., AND R. TROMPETTE (1976). Late Pre- Progr., Moscow. cambrian mixtites: Glacial and/or nonglacial? Deal- CROWELL, J. C. (1978). Gondwanan glaciation, cyclo- ing especially with the mixtites of west Africa. thems, continental positioning, and climate change. Amer. J. Sci. 276, 1302-1324. Amer. J. Sci. 278, 1345-1372. DEYNOUX, M., AND R. TROMPETTE (1981). Late Or- 440 NICHOLAS CHRISTIE-BLIC K

dovician tillites of the Taoudeni Basin, west Africa. FRAKES, L. A. (1979). Climates Throughout Geologic In Earth's Pre-Pleistocene Glacial Record (M. J. Time. Elsevier, New York/Amsterdam. Hambrey and W. B. Harland, Eds.), pp. 89-96. FULLER, A. O. (1958). A contribution to the petrology Cambridge Univ. Press, London/New York. of the Witwatersrand System. Trans. Geol. Soc. S. DEYNOUX, M., O. DIA, J. SOUGY, AND R. TROMPETTE Afr. 61, 19-50. (1972). La glaciation "Fini-Ordovicienne" en Afri- GEITZENAUER, K. R., S. V. MARGOL1S, AND D. S. que de l'ouest. Soc. Geol. Mineral. Bretagne Bull. EDWARDS (1968). Evidence consistent with Eocene C. IV, 9-16. glaciation in a South Pacific deep sea sedimentary DONNELLY, T. H. (1981). Discussion: Sedimentology, core. Earth Planet. Sci. Lett. 4, 173-177. stable-isotope geochemistry and palaeoenvironment HAMBREY, M. J., AND W. B. HARLAND, Eds. of dolostones capping late Precambrian glacial se- (1981). Earth's Pre-Pleistocene Glacial Record. quences in Australia. J. Geol. Soc. Aust. 28, 99- Cambridge Univ. Press, London/New York. 101. HARLAND, W. B. (1964a). Evidence of late Precam- DORMAN, F. H. (1968). Some Australian oxygen iso- brian glaciation and its significance. In Problems in tope temperatures and a theory for a 30-million-year Palaeoclimatology (A. E. M. Nairn, Ed.), pp. 119- world-temperature cycle. J. Geol. 76, 297-313. 149. Interscience, New York. DOTT, R. H., JR. (1961). Squantum "tillite," Massa- HARLAND, W. B. (1964b). Critical evidence for a great chusetts-Evidence of glaciation or subaqueous infra-Cambrian glaciation. Geol. Rundsch. 54, 45- mass movements? Geol. Soc. Amer. Bull. 72, 1289- 61. 1306. HARLAND, W. B. (1981a). Chronology of Earth's gla- DREWRY, G. E., A. T. S. RAMSAY, AND A. G. SMITH cial and tectonic record. J. Geol. Soc. London 138, (1974). Climatically controlled sediments, the geo- 197-203. magnetic field, and trade wind belts in Phanerozoic HARLAND, W. B. (1981b). The late Archaean(?) time. J. Geol. 82, 531-553. Witwatersrand conglomerates, South Africa. In DZURISIN, D., AND K. R. BLASIUS (1975). Topography Earth's Pre-Pleistocene Glacial Record (M. J. Ham- of the polar layered deposits of Mars. J. Geophys. brey and W. B. Harland, Eds.), pp. 185-187. Cam- Res. 80, 3286-3306. bridge Univ. Press, Cambridge/New York. EDWARDS, M. B. (1978). Glacial environments. In HARLAND, W. B., AND E. H. FRANCIS, Eds. (1971). Sedimentary Environments and Facies (H. G. Read- The Phanerozoic Time-Scale. A Supplement. Geol. ing, Ed.), pp. 416-438. Elsevier, New York/Amster- Soc. Lond. Spec. Publ. 5. dam. HARLAND, W. B., AND K. N. HEROD (1975). Glacia- FISCHER, A. G. (1980). Gilbert-Bedding rhythms and tions through time. In Ice Ages: Ancient and Mod- geochronology. Geol. Soc. Amer. Spec. Pap. 183, ern (A. E. Wright and F. Moseley, Eds.), pp. 189- 93-104. 216. Seel House Press, Liverpool. FISCHER, A. G. (1982a). Long-term climatic oscilla- HARLAND, W. B., K. N. HEROD, AND D. H. tions recorded in stratigraphy. In Report on KRINSLEY (1966). The definition and identification of Pre-Pleistocene Climates. Nat. Acad. Sci., in tills and tillites. Earth-Sci. Rev. 2, 225-256. press. HARLAND, W. B., A. G. SMITH, AND B. WILCOCK, FISCHER, A. G. (1982b). The 21,000 year shale-lime- Eds. (1964). The Phanerozoic Time-Scale. Geol. stone rhythm. J. Geol. Soc. London 139, in press. Soc. Lond. Spec. Publ. I. FISCHER, A. G., AND M. A. ARTHUR (1977). Secular HARRISON, J. E., AND Z. E. PETERMAN (1980). North variations in the pelagic realm. In Deep-Water Car- American Commission on Stratigraphic Nomencla- bonate Environments (H. E. Cook and P. Enos, ture Note 52--A preliminary proposal for a chrono- Eds.), pp. 19-50. Soc. Econ. Paleont. Mineral. metric time scale for the Precambrian of the United Spec. Publ. 25. States and Mexico. Geol. Soc. Amer. Bull. 91, 377- FLINT, R. F. (1971). Glacial and Quaternary Geology. 380. Wiley, New York. HART, M. H. (1978). A history of the Earth's atmo- FLINT, R. F. (1975). Features other than diamicts as sphere. Icarus 33, 23-39. evidence of ancient glaciations. In Ice Ages: Ancient HAVES, D. E., AND L. A. FRAKES (1975). General syn- and Modern (A. E. Wright and F. Moseley, Eds.), thesis. In Initial Reports of the Deep Sea Drilling pp. 121-136. Seel House Press, Liverpool. Project 28, pp. 919-942. U.S. Govt. Printing Office, FLINT, R. F., J. E. SANDERS, AND J. RODGERS (1960a). Washington, D.C. Symmictite: A name for nonsorted terrigenous sedi- HAYS, J. D., J. IMBRIE, AND N. J. SHACKLETON mentary rocks that contain a wide range of particle (1976). Variations in the Earth's orbit: Pacemaker of sizes. Geol. Soc. Amer. Bull. 71, 507-510. the ice ages. Science 194, 1121-1132. FLINT, R. F., J. E. SANDERS, AND J. RODGERS HEAD, J. W., AND S. C. SOLOMON (1981). Tectonic (1960b). Diamictite: A substitute term for symmic- evolution of the terrestrial planets. Science 213, 62- tite. Geol. Soc. Amer. Bull. 71, 1809-1810. 76. PRE-PLEISTOCENE GLACIATION ON EARTH 441

HOFMANN, H. J. (1973). Stromatolites: Characteris- EMBLETON (1974). Palaeomagnetic results and late tics and utility. Earth-Sci. Rev. 9, 339-373. Precambrian glaciations. Nature (London) 248, HOWARD, A. D. (1978). Origin of the stepped topogra- 557-561. phy of the Martian poles. Icarus 34, 581-599. McWILLIAMS, M. O., AND M. W. MCELHINNY (1980). HOWARD, A. D. (1982). Etched plains and braided Late Precambrian paleomagnetism of Australia: The ridges of the south polar region of Mars: Features Adelaide Geosyncline. J. Geol. 88, 1-26. produced by basal melting of ground ice? In Planet. MARGOLIS, S. V., AND J. P. KENNETT (1971). Ceno- Geol. Princ. Inv. Meeting, NASA TM, in press. zoic paleoglacial history of Antarctica recorded in HOWARD, A. D., J. A. CUTTS, AND K. R. BLASIUS subantarctic deep-sea cores. Amer. J. Sci. 271, 1- (1982). Stratigraphic relationships within Martian 36. polar cap deposits. Icarus 50, 160-215. MASURSKY, H., J. M. BOYCE, A. L. DIAL, G. G. IMBRIE, J. (1982). Review of the astronomical theory SCHABER, AND M. E. STROBELL (1977). Classifica- of the Pleistocene ice ages. Icarus, in press. tion and time of formation of Martian channels IMBRIE, J., AND J. Z. IMBRIE (1980). Modeling the based on Viking data. J. Geophys. Res. 82, 4016- climatic response to orbital variations. Science 207, 4038. 943-953. MATHUR, S. M. (1981). The Middle Precambrian Gan- IRVING, E., AND J. C. McGLYNN (1976). Proterozoic gau Tillite, Bijawar Group, Central India. In Earth's magnetostratigraphy and the tectonic evolution of Pre-Pleistocene Glacial Record (M. J. Hambrey and Laurentia. Philos. Trans. Roy. Soc. London Ser. A W. B. Harland, Eds.), pp. 428-430. Cambridge 280, 433-468. Univ. Press, London/New York. KARLSTROM, K. E., AND R. S. HOUSTON (1979). MAYEWSKI, P. A., G. H. DENTON, AND T. J. HUGHES Stratigraphy of the Phantom Lake Metamorphic (1981). Late Wisconsin ice sheets of North America. Suite and Deep Lake Group and a review of the In The Last Great Ice Sheets (G. H. Denton and T. J. Precambrian tectonic history of the Medicine Bow Hughes, Eds.), pp. 67-178. Wiley, New York. Mountains. Contrib. Geol. Univ. VCyorning 17, 111- MITCHELL, J. M., JR. (1976). An overview of climatic 133. variability and its causal mechanisms. Quat. Res. 6, KELLER, B. M. (1973). Great glaciations in history of 481-493. the earth. Int. Geol. Rev. 15, 1067-1074. MOREL, P., AND E. IRVING (1978). Tentative paleocon- KERR, R. A. (1980). Changing global sea levels as a tinental maps for the Early Phanerozoic and Pro- geologic index. Science 209, 483-486. terozoic, J. Geol. 86, 535-561. KIEFFER, H. H., S. C. CHASE, JR., T. Z. MARTIN, E. MORRIS, W. A. (1977a). Paleomagnetism of the D. MINER, AND F. D. PALLUCONI (1976). Martian Gowganda and Chibougamau Formations: Evidence north pole summer temperatures: Dirty water ice. for 2,200-m.y.-old folding and re magnetization event Science 194, 1341-1344. of the southern province. Geology 5, 137-140. KNAUTH, L. P., AND S. EPSTEIN (1976). Hydrogen MORRIS, W. A. (1977b). Paleolatitude of glaciogenic and oxygen isotope ratios in nodular and bedded upper precambrian Rapitan Group and the use of cherts. Geochim. Cosmochim. Acta 40, 1095-1108. tillites as chronostratigraphic marker horizons. Ge- KNOLL, A. H., N. BLICK, AND S. M. AWRAMIK ology 5, 85-88. (1981). Stratigraphic and ecologic implications of MURRAY, B. C., AND M. C. MALIN (1973). Polar wan- late Precambrian microfossils from Utah. Amer. J. dering on Mars? Science 179, 997-1000. Sci. 281, 247-263. MUTCH, T. A., R. E. ARVIDSON, J. W. HEAD III, K. LAZARUS, D. B., AND J. D. HAYS (1981). Late Mio- L. JONES, AND R. S. SAUNDERS(1976). The Geology cene history of the southern ocean. Geol. Soc. of Mars. Princeton Univ. Press, Princeton, N.J. Amer. Abstr. Progr. 13, no. 7, 495. NEETHLING, D. C. (1970). Geology of the Ahlmann LEGGETT, J. K., W. S. MC KERROW, L. R. M. CocKs, Ridge, Western Queen Maud Land. Map sheet 7, AND R. B. RICHARDS (1981). Periodicity in the early plate VII, folio 12, Antarctic map folio series, Amer. Palaeozoic marine realm. J. Geol. Soc. London 138, Geogr. Soc. 167-176. NEWKIRK, G., JR. (1980). Solar variability on time LONG, D. G. F. (1974). Glacial and paraglacial genesis scales of 105 years to l09-e years. In Proc. Conf. of conglomeratic rocks of the Chibougamau Forma- Ancient Sun (R. O. Pepin, J. A. Eddy, and R. B. tion (Aphebian), Chibougamau, Quebec. Canad. J. MerTill, Eds.), pp. 293-320. Earth Sci. 11, 1236-1252. NUMMEDAL, D. (1978). The role of liquification in MCCREA, W. H. (1975). Ice ages and the galaxy. Na- channel development on Mars. Rep. Planet. Geol. ture (London) 255, 607-609. Prog., NASA TM 79729, 257-259. MCELH1NNY, M. W. (1973). Palaeomagnetism and NUNES, P. D., AND G. R. TILTON (1971). Uranium- Plate Tectonics. Cambridge Univ. Press, London/ lead ages of minerals from the Stillwater Igneous New York. Complex and associated rocks, Montana. Geol. Soc. MCELHINNY, M. W., J. W. GIDD1NGS, AND B. J. J. Amer. Bull. 82, 2231-2250. 442 NICHOLAS CHRISTIE-BLIC K

OJAKANGAS,R. W., AND C. L. MATSCH (1980). Upper SAGAN, C., O. B. TOON, AND P. J. GIERASCH (1973). Precambrian (Eocambrian) Mineral Fork Tillite of Climatic change on Mars. Science 181, 1045-1049. Utah: A continental glacial and glaciomarine se- SAVIN, S. M. (1977). The history of the Earth's surface quence. Geol. Soc. Amer. Bull. 91, 495-501. temperature during the past 100 million years. Annu. OWEN, T., R. D. CESS, AND V. RAMANATHAN (1979). Rev. Earth Planet. Sci. 5, 319-355. Infrared CO2 greenhouse to compensate for reduced SCHERMERHORN, L. J. G. (1974). Late Precambrian solar luminosity on early Earth. Nature (London) mixtites: Glacial and/or nonglacial? Amer. J. Sei. 277, 640-642. 274, 673-824. PAGE, N. J. (1977). Stillwater Complex, Montana: SCHERMERHORN, L. J. G. (1975). Tectonic framework Rock succession, metamorphism and structure of of Late Precambrian supposed glacials. In Ice Ages: the complex and adjacent rocks. U.S. Geol. Surv. Ancient and Modern CA. E. Wright and F. Moseley, Prof. Pap. 999. Eds.), pp. 241-274. Seel House Press, Liverpool. PAGE, N. J., AND R. A. KOSKI (1973). A Precambrian SCHMIDT, P. W., AND W. A. MORRIS (1977). An alter- diamictite below the base of the Stillwater Complex, native view of the Gondwana Paleozoic apparent southwestern Montana. J. Res. U.S. Geol. Surv. l, polar wander path. Canad. J. Earth Sci. 14, 2674- 403-414. 2678. PERRY, E. C., AND F. C. TAN (1972). Significance of SCHULTZ, P. H., AND A. B. LUTZ-GARIHAN (1981). oxygen and carbon isotope variations in Early Pre- Ancient polar locations on Mars: Evidence and im- cambrian cherts and carbonate rocks of southern plications. In 3rd Mars Internat. Colloqu. Abstrs., Africa. Geol. Soc. Amer. Bull. 83, 647-664. Pasadena, California, pp. 229-231. PIERI, D. (1976). Distribution of small channels on the SCrtWARZBACH, M. (1964). Criteria for the recognition Martian surface. Icarus 27, 25-50. of ancient glaciations. In Problems in Palaeoelima- PIPER, J. D. A. (1973). Latitudinal extent of Late Pre- tology CA. E. M. Nairn, Ed.), pp. 81-85. Inter- cambrian glaciations. Nature (London) 244, 342- science, New York. 344. SCOTESE, C. R., R. K. BAMBACH,C. BARTON, R. VAN PIPER, J. D. A. (1976). Palaeomagnetic evidence for a DER VOO, AND A. M. ZIEGLER (1979). Paleozoic Proterozoic supercontinent. Philos. Trans. Roy. base maps. J. Geol. 87, 217-277. Soc. London Set. A 280, 469-490. SHARP, R. P., AND M. C. MALIN (1975). Channels on PLAEKER, G., AND W. O. ADDICOTT (1976). Gla- Mars. Geol. Soe. Amer. Bull. 86, 593-609. ciomarine deposits of Miocene through Holocene SPENCER, A. M. ( 1971 ). Late Pre- Cambrian glaciation age in the Yakataga Formation along the Gulf of in Scotland. Geol. Soc. London. Mere. 6. Alaska margin. In Recent and Ancient Sedimentary SQUYRES, S. W. (1979). The evolution of dust deposits Environments in Alaska iT. P. Miller, Ed.), pp. 1- in the Martian north polar region. Icarus 40, 244- 23. Alaska Geol. Soc., Anchorage. 261. POLLACK, J. B. (1979). Climate change on the terres- STEINER, J. (1967). The sequence of geological events trial planets. Icarus 37, 479-553. and the dynamics of the Milky Way Galaxy. J. Geol. RONOV, A. B. (1964). Common tendencies in the Soc. Aust. 14, 99-132. chemical evolution of the Earth's crust, ocean and STEINER, J., AND E. GRILLMAIR (1973). Possible ga- atmosphere. Geochemistry 8, 715-743. lactic causes of periodic and episodic glaciations. ROSCOE, S. M. (1973). The Huronian Supergroup, a Geol. Soc. Amer. Bull. 84, 1003-1018. Paleoaphebian succession showing evidence of at- SYMONS, D. T. A. (1975). Huronion glaciation and po- mospheric evolution. In Huronian Stratigraphy and lar wander from the Gowganda Formation, Ontario. Sedimentation (G. M. Young, Ed.), pp. 31-47. Geol. Geology 3, 303-306. Assoc. Canada Spec. Paper 12. TALBOT, R. J., JR., D. M. BUTLER, AND M. J. NEW- RUNCORN, S. K. (1964). The connexion between pa- MAN (1976). Climatic effects during passage of the laeomagnetism and palaeoclimates. In Problems in solar system through interstellar clouds. Nature Palaeoclimatology CA. E. M. Nalrn, Ed.), pp. 191- (London) 262, 561-563. 199. Interscience, New York. TARL[NG, D. H. t1974). A palaeomagnetic study of RusT, I. C. (1981). Early Paleozoic Pakhuis Tillite, Eocambrian tillites in Scotland. J. Geol. Soc. Lon- South Africa. In Earth's Pre-Pleistocene Glacial don 130, 163-177. Record (M. J. Hambrey and W. B. Harland, Eds.), TOON, O. B., J. B. POLLACK, W. WARD, J. A. BURNS, pp. 113-117. Cambridge Univ. Press, London/ AND K. BILSK! (1982). The astronomical theory of New York. climatic change on Mars. h'arus, in press. SAGAN, C. (1977). Reducing greenhouses and the tem- TRENDALL, A, F. (1976). Striated and faceted boulders perature history of Earth and Mars. Nature (Lon- from the Turee Creek Formation--Evidence for a don) 269, 224-226. possible Huronian glaciation on the Australian conti- SAGAN, C., AND G. MULLEN (1972). Earth and nent. Geol. Surv. West. Aust. Annu. Rep., 88-92. Mars: Evolution of atmospheres and surface tem- TUCKER, M. E., AND P. C. REID ( 19731. The sedimen- peratures. Science 177, 52-56. tology and context of Late Ordovician glacial ma- PRE-PLEISTOCENE GLACIATION ON EARTH 443

fine sediments from Sierra Leone, West Africa. Pa- Mars: Fourth-order accuracy in orbital e and 1. J. laeogeogr. Palaeoclimatol. Palaeoecol. 13, 289- Geophys. Res. 84, 237-241. 307. WARD, W. R. (1982). Comments on the long-term sta- UMBGROVE, J. H. F. (1947). The Pulse of the Earth, bility of the Earth's obliquity. Icarus 50, 444-448. 2nd ed. Martinus Nijhoff, The Hague. WARD, W. R., J. A. BURNS, AND O. B. TOON (1979). VAIL, P. R., AND J. HARDENBOL (1979). Sea-level Past obliquity oscillations of Mars: The role of the changes during the Tertiary. Oceanus 22, 71-79. Tharsis Uplift. J. Geophys. Res. 84, 243-259. VAIL, P. R., R. M. MITCHUM, JR., AND S. THOMPSON WARREN, G. (1962). Mawson Tillite. New Zealand III (1977a). Seismic stratigraphy and global changes Geol. Surv. Bull. 71, 117-122. of sea level, part 3: Relative changes of sea level WIEBOLS, J. H. (1955). A suggested glacial origin for from coastal onlap. In Seismic Stratigraphy-Appli- the Witwatersrand conglomerates. Trans. Geol. cations to Hydrocarbon Exploration (C. E. Payton, Soc. S. Afr. 58, 367-387. Ed.), pp. 63-81. Am. Assoc. Petrol Geol. Mem. WILLIAMS, G. E. (1972). Geological evidence relating 26. to the origin and secular rotation of the solar system. VAIL, P. R., R. M. MITCHUM, JR., AND S. THOMPSON Mod. Geol. 3, 165-181. III (1977b). Seismic stratigraphy and global changes WILLIAMS, G. E. (1973). Geotectonic cycles, lunar of sea level, part 4: Global cycles of relative changes evolution, and the dynamics of the Earth-Moon sys- of sea level. In Seismic Stratigraphy-Applica- tem. Mod. Geol. 4, 159-183. tions to Hydrocarbon Exploration (C. E. Payton, WILLIAMS, G. E. (1975a). Late Precambrian glacial Ed.), pp. 83-97. Am. Assoc. Petrol. Geol. Mem. climate and the Earth's obliquity. Geol. Mag. 112, 26. 441-465. VAN EYSINGA, F. W. B., Compiler. (1975). Geological WILLIAMS, G. E. (1975b). Possible relation between Time Table. Elsevier, Amsterdam/New York. periodic glaciation and the flexure of the galaxy. VAN NJEKIRK, C. B., AND A. J. BURGER (1978). A new Earth Planet. Sci. Lett. 26, 361-369. age for the Ventersdorp acidic lavas. Trans. Geol. WILLIAMS, G. E. (1979). Sedimentology, stable-iso- Soc. S. Afr. 81, 155-163. tope geochemistry and palaeoenvironment of dolo- VISSER, J. N. J. (1971). The deposition of the Gri- stones capping late Precambrian glacial sequences quatown Glacial Member in the Transvaal Super- in Australia. J. Geol. Soc. Aust. 26, 377-386. group. Trans. Geol. Soc. S. Afr. 74, 187-199. WILLIAMS, G. E. (1981). Reply: Sedimentology, sta- VISSER. J. N. J. (1981). The Mid-Precambrian tillite in ble-isotope geochemistry and palaeoenvironment the Griqualand West and Transvaal Basins, South of dolostones capping late Precambrian glacial se- Africa. In Earth's Pre-Pleistocene Glacial Record quences in Australia. J. Geol. Soc. Aust. 28, 102- (M. J. Hambrey and W. B. Harland, Eds.), pp. 180- 105. 184. Cambridge Univ. Press, London/New York. WILLIAMS, G. E., AND P. M. AUSTIN (1973). Global WALKER, J. C. G. (1976). Implications for atmospheric tectonics and the earth's rotation. Mod. Geol. 4, evolution of the inhomogeneous accretion model of 185-199. the origin of the Earth. In The Early History of Earth WINDLEY, B. F. (1977). The Evoh,ing Continents. Wi- (B. F. Windley, Ed.), pp. 537-546. Wiley, New ley, New York. York. YOUNG, G. M. (1970). An extensive early Proterozoic WALKER, J. C, G. (1977). Evolution of the Atmo- glaciation in North America? Palaeogeogr. Palaeo- sphere. Macmillan Co., New York. climatol. Palaeoecol. 7, 85-101. WALTER, M. R., Ed. (1976). Stromatolites. Dev. Sedi- YOUNG, G. M. (1973). Tillites and aluminous quartz- mentol. 20. ites as possible time markers for middle Precam- WARD, W. R. (1974). Climatic variations on Mars. 1. brian (Aphebian) rocks of North America. In Huro- Astronomical theory of insolation. J. Geophys. Res. nian Stratigraphy and Sedimentation (G. M. Young, 79, 3375-3386. Ed.), pp. 97-127. Geol. Assoc. Canad. Spec. Paper WARD, W. R. (1979). Present obliquity oscillations of 12.