A Critique of Phanerozoic Climatic Models Involving Changes in The

Home , Anthony Hallam, Evolution of fungi, Mycotroph, Paleoatmosphere, Rhizocarpon geographicum

Earth-Science Reviews 56Ž. 2001 1–159 www.elsevier.comrlocaterearscirev

A critique of Phanerozoic climatic models involving changes in

the CO2 content of the atmosphere A.J. Boucot a,), Jane Gray b,1 a Department of Zoology, Oregon State UniÕersity, CorÕallis, OR 97331, USA b Department of Biology, UniÕersity of Oregon, Eugene, OR 97403, USA Received 28 April 1998; accepted 19 April 2001

Abstract

Critical consideration of varied Phanerozoic climatic models, and comparison of them against Phanerozoic global climatic gradients revealed by a compilation of Cambrian through Miocene climatically sensitive sedimentsŽ evaporites, coals, tillites, lateritic soils, bauxites, calcretes, etc.. suggests that the previously postulated climatic models do not satisfactorily account for the geological information. Nor do many climatic conclusions based on botanical data stand up very well when examined critically. Although this account does not deal directly with global biogeographic information, another powerful source of climatic information, we have tried to incorporate such data into our thinking wherever possible, particularly in the earlier Paleozoic. In view of the excellent correlation between CO2 present in Antarctic ice cores, going back some hundreds of thousands of years, and global climatic gradient, one wonders whether or not the commonly postulated Phanerozoic connection between atmospheric CO2 and global climatic gradient is more coincidence than cause and effect. Many models have been proposed that attempt to determine atmospheric composition and global temperature through geological time, particularly for the Phanerozoic or significant portions of it. Many models assume a positive correlation between atmospheric CO2 and surface temperature, thus viewing changes in atmospheric CO2 as playing the critical role in r regulating climate temperature, but none agree on the levels of atmospheric CO2 through time. Prior to the relatively recent interval of time in which atmospheric CO2 is directly measurable, a variety of biological and geological proxies have been r proposed to correlate with atmospheric CO22 level or with pCO temperature. Atmospheric models may be constructed for the Pre-Cenozoic but the difficulties of assessing variables in their construction are many and complex. None of the modelers have gathered enough biological and geological data to impart reliability to the model constructs.

Most modelers focus almost exclusively on one or a few variables as proxy to measure atmospheric CO2 , nor consider the many other variables involved, nor agree on what these variables are or how to measure them. In this paper, it is the reliability of the present data bases used in these atmospheric models that we wish to consider. We focus most attention on the Berner models, such as GEOCARB I, II and BLAG, because of the basic role they attribute to tracheophytes in regulating atmospheric CO2 and our own interest in pre-tracheophytic land plants and the atmospheric composition of the pre-tracheophytic Paleozoic. We survey the presence of symbiotic mycorrhizae and question the assumption that all tracheophytes are obligately associated with them.

) Corresponding author. Fax: q1-541-737-0501. E-mail address: [email protected]Ž. A.J. Boucot . 1 Deceased.

Although pre-tracheophytic embryophytes, cyanobacteria, and possibly other organisms preceded tracheophytes on land by millions of years, Berner’s models do not consider a significant role for them in affecting pre-Devonian climatertemperature and atmospheric composition. In effect, Berner assumes that pre-tracheophytic life inhabited a world governed largely by abiotic physical and chemical reactions. We consider uncertainties raised by minimizing possible roles for pre- tracheophytic oxygenic and heterotrophic microorganisms analogous to those speculated to be unique to tracheophytes both with regard to an active role in biodeterioration of rock and soil mineral substrates and in the sequestration of organic carbon. Additionally, Berner does not consider marine productivity, which might have been high in the Precambrian and Early Paleozoic and possibly consequent organic carbon sequestration, even in the possible absence of terrestrial organisms, or even in the absence of a significant preserved biomass of terrestrial and marine organisms. The important roles played by cyanobacteria, for example, are briefly reviewed by Giller and Malmqvist in lakes and rivers as regards both planktonic and benthic taxa, and it is not safe to assume that these organisms were absent or of no potential significance in the pre-embryophytic, i.e. earlier Ordovician and well back into the Precambrian.

Berner’s models have met with a large measure of consensus about CO2 balance during the Phanerozoic, about the role played by tracheophytes, and have been used to test or evaluate other data. After reviewing the biological and geological assumptions and estimates on which these Models are based, we conclude that they do not provide reliable information about atmospheric CO2 composition through Phanerozoic time, particularly in the Early Phanerozoic. We compare many atmospheric CO2 models, while considering the numerous proxies on which they are based and conclude that the competing models are inadequate for atmospheric CO2 estimation. Many possibilities not considered in present models must either be included or eliminated based on reliable evidence. We suggest that assessing Phanerozoic climatertemperature based on the available geologicalrclimatic proxies would appear to provide a more reliable method of estimating variations in CO222 , and hence atmospheric CO :O balance, than most proxy constructs on which atmospheric models are presently based, because of the critical role postulated for atmospheric CO2 in regulating Earth’s surface temperature. We present our own Phanerozoic climate estimate, based on r readily available geological climatic data, for comparison with postulated coeval atmospheric CO2 levels as a test of the postulated correlation. q 2001 Elsevier Science B.V. All rights reserved.

Keywords: global climatic gradient; atmospheric CO2 ; Phanerozoic; stomates; mycorrhizae; paleogeography; weathering

Contents

1. Introduction ...... 5

2. Phanerozoic climate and atmospheric composition ...... 7 2.1 Introduction ...... 7 2.2 Carbon and CO2 sources and sinks ...... 9 2.2.1 Carbon and CO2 sources ...... 9 2.2.2 Carbon and CO2 sinks...... 10 2.3 Criticisms ...... 10 2.4 An alternative ...... 11

3. Atmospheric CO2 proxies ...... 11

4. Proxies without independent atmospheric CO2 curves ...... 12 4.1 Stomates ...... 12 4.1.1 Physical and biotic variables affecting stomatal densities and occurrences ...... 13 4.1.2 Application of stomatal data to atmospheric CO2 estimates ...... 15 4.2 Tracheophyte productivity and standing crop...... 17 4.3 Proxies with atmospheric CO2 curve models ...... 17 4.3.1 Berner models ...... 17 a. Introduction ...... 17 b. Atmospheric CO2 fluctuations in the Berner models ...... 19 c. Berner’s data bases ...... 19 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 3

I. Estimates and assumptions...... 19 a. Sequestered organic carbonŽ. continental ...... 19 b. Sequestered inorganic carbonŽ. marine ...... 21 II. The modern autotrophic weathering basis of Berner’s 1984–2001 models...... 21 a. 1991 Model ...... 21 b. 1992, 1994, 1995, 1997 models ...... 22 4.3.2 Comment and criticism of the Berner models ...... 26 I. Marine phytoplankton ...... 26 II. Pretracheophytic autotrophs ...... 26 III. Phytogeography and evolution ...... 27 IV. Climatically sensitive rocks ...... 27 V. Need to collect adequate geological data...... 27 VI. Is tracheophyte weathering evolutionarily progressive? ...... 27 VII. Sequestered organic carbon ...... 29 VIII. Natural hydrocarbons ...... 30 IX. Deep, subsurface bacterial carbon ...... 30 X. Were Early and Middle Devonian floras confined to the lowlands? ...... 30

5. Organic carbon ...... 30 5.1 Oceanic ...... 31 5.2 Epicontinental ...... 32 5.3 Continental...... 33 5.4 Organic carbon in soil ...... 33 5.5 Inorganic soil carbon ...... 34 5.6 Nonsoil carbonate rock ...... 35

6. Use of strontium isotopes as a measure of weathering ...... 36

7. Berner’s use of modern river water data bases for assessing weathering ...... 36 7.1 Is flux a credible proxy for weathering rate? ...... 38

8. Carbon isotopes from calcrete as a measure of atmospheric CO2 ...... 41

9. Models inspired by T.C. Chamberlin ...... 44

10. Worsley et al. model...... 51

11. Francois et al. model ...... 52

12. Tappan model based on phytoplankton periodicity ...... 53

13. Freeman and Hayes model ...... 54

14. Budyko et al. model ...... 54

15. Chemical weathering: evaluating some of the factors in a multifaceted problem ...... 56

16. The relief and elevation effects ...... 59

17. The paleogeographic effect ...... 61

18. Tracheophytes and the soil effect ...... 62 18.1 What is soil? ...... 62 18.1.1 The role of the rhizosphere ...... 64 a. What is the rhizosphere? ...... 64 b. Bulk soil versus rhizosphere ...... 66 I. Mineral weathering in the rhizosphere ...... 68 4 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

c. The role of free-living fungirmicrobiota in soil weathering...... 70 d. The role of tracheophyte roots, their symbiotic fungi and other microorganisms in weathering ...... 70 e. Climatic correlates with biomes in varied climates ...... 73 18.1.2 The role of temperature in chemical weathering ...... 74 18.1.3 The role of CO22Ž. pCO -atmospheric and CO 2 partial pressure in chemical weathering ...... 74

19. The surface bedrock and vadose zone effect ...... 76 19.1 Bacteria and cyanobacteria: chemical and physical weathering of surface bedrock and regolith ...... 79 19.2 Bacteria and cyanobacteria: chemical and physical weathering of subsurface sediments and bedrock ...... 83 19.3 Lichens...... 86 19.4 Fungi ...... 92 19.5 Algae...... 94 19.6 Bryophytes...... 95

20. Evolutionary history of the rhizosphere: conventional and alternative hypotheses ...... 96 20.1 Summary...... 97

21. Are tracheophytes ultimate brokers of increased mineral solubility in the rhizosphere?: some evolutionary afterthoughts ..... 100 21.1 Extant VAM–tracheophyte relations ...... 104 21.2 Extant endophytic fungal–angiosperm relations ...... 105 21.3 Extant endophytic fungal–gymnosperm relations ...... 107 21.4 Extant endophytic fungal–pteridophyte relations ...... 109 21.5 Extant endophytic fungal–bryophyte relations ...... 110 21.6 The Rhynie Chert and a molecular benchmark for mycorrhizal–tracheophyte relations ...... 115 21.7 Conclusions ...... 118

22. Where is the pre-tracheopytic fossil record? ...... 118 22.1 Lichens...... 118 22.2 Fungi ...... 119 22.3 Mycorrhizae ...... 119

23. Autotrophic productivity and sequestered organic carbon, or, where is the pre-tracheophytic biomass? ...... 120 23.1 Precambrian regoliths ...... 121 23.2 Non-tracheophyte productivity ...... 122 23.3 Summary...... 127

24. A climatic test for atmospheric models ...... 129 24.1 A geologic climatertemperature model...... 129 24.2 The Ordovician glaciation as an example of the problem ...... 131 24.3 A Permo–Triassic boundary example...... 133 24.4 A Cretaceous example...... 133 24.5 Eocene examples ...... 134 24.6 A Pliocene example ...... 134

25. Summary and conclusions...... 135

Acknowledgements...... 137

References ...... 138 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 5

‘‘He had been Eight Years upon a Project for cles of warming and cooling. Raymo et al.Ž. 1998 extracting Sun Beams out of Cucumbers, which provide additional evidence from the marine sedi- were to be put into Vials hermetically sealed, and mentary record consistent with associated climatic

let out to warm the Air in raw inclement Sum- changes and atmospheric CO2 going back into the mers.’’Ž. Swift, 1726, A Voyage to Laputa, etc. . earlier Pleistocene. These correlations do not necessarily reveal cause and effect and it is clear that understanding the biotic feedbacks in the global 1. Introduction climatic system is complicated by many poorly In general, Earth’s atmosphere is predicted to constrained variablesŽ Trabalka and Reichle, 1986; have changed over the past 3 billion years from a Woodwell and Mackenzie, 1995; Shine and Forster, hot, CO2 -rich, low-oxygen atmosphere to the com- 1999, their Fig. 2. . paratively low CO2 , comparatively oxygen-rich at- Additionally, there is the apparently paradoxical mosphere and more hospitable climate of the present situationŽ. Schlesinger, 1991; Robinson et al., 1998 Earth. In considering Earth’s atmospheric evolution that during the past century or so, despite good through geologic time, there is general agreement evidence for significant increase in atmospheric CO2 , that oxygenic photosynthetic organisms have af- some have not seen a correspondingly agreed on fected atmospheric oxygen concentration beginning increase in global temperature. According to Robin- some time in the Precambrian. Even though there is son et al.Ž. 1998 , global average temperatures have no general consensus on when oxygen reached its actually cooled slightly, although this is not the present atmospheric level, wide variations in atmo- majority view at this timeŽ Sigman and Boyle, 2000, spheric oxygen are not anticipated for the Phanero- as example.Ž. . Broecker 1999 suggests that the zoic. But organic and other controls on changes in evidence of apparent lack of correlation between atmospheric CO222 and the atmospheric CO :O bal- atmospheric CO2 and global warming of the imme- ance through geologic time have been the subject of diate past results from a Anatural coolingB eventŽ the considerable discussion and sometimes considerable Little Ice Age. that temporarily countered the ex- difference of opinion, with wide fluctuations in at- pected warming. But, after reviewing historical data mospheric CO2 that differ conspicuously from to- of the past few centuries he concludes that it is still day’s atmospheric CO2 values. The actual course of too soon to be certain about the correlation between A B these changes and the causes of potential CO2 increasing atmospheric CO2 and increasing tempera- fluctuations are the subject of many models. Most ture, although he is inclined to favor this correlation. such models attribute different relative importance to He has also suggestedŽ. Broecker, 1997 that climatic abiotic geological, geochemical and biological pro- gradient changes during the later Quaternary have cesses in modelling atmospheric CO2 and surface been very rapid, and that this results from interac- temperature over time. All suggest that Phanerozoic tions involving oceanic circulation changes and addi-

CO2 fluctuations are then correlated with extreme tions and subtractions of water vapor, mainly from variations in global surface temperatures. the tropics; hence atmospheric CO2 might not have For the immediate- and long-range future, the been the most important greenhouse gas during the prospect of global warming, associated with high recent past. atmospheric CO2 , has generated increasing interest Others have suggested a variety of complicating in atmospheric composition and atmospheric compo- factors during the immediate past, equivalent to the sitional change, especially anthropogenically induced Little Ice Age in terms of temporary effects, which changes related to release of sequestered carbon might complicate the potential, short-term correlation stores into the atmosphereŽ for example, Gates et al., between temperature and atmospheric CO2 . 1983; Trabalka and Reichle, 1986; Schlesinger, 1991; Smith and ShugartŽ. 1993, p. 523 note that changes Woodwell and Mackenzie, 1995; Shine and Forster, in vegetation and soil type that result in a net release

1999. . For at least the past 200,000 years, ice-core of CO2 could be more rapid than changes that result records indicate a high correlation between atmo- in a net increase in terrestrial carbon storage. A B spheric CO24 , CH and N 2 O concentrations and cy- In a perturbed climate, they conclude that such a 6 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

AtransientB climate change would mean that the ter- mote past, the issue is additionally complicated by restrial vegetationrsoil system could be a net, signif- physical variables for which we have limited direct r A icant source of carbon CO2 to the atmosphere in knowledge such as, for example, a theoretical ca. the first 50–100 years following a climate warming, 30% increase in solar luminosity over Earth’s history increasing the atmospheric CO2 concentration by up Ž.Walker et al., 1981 . to a third of the present level.B Likewise, Oechel et If a dependency between silicate weathering, tem- al.Ž. 1993 note that climate change over the past perature and atmospheric CO2 is assumedŽ Brady century may be only a part of normal climate varia- and Zachara, 1996, and their Fig. 10; Caldeira and tion, yet it has significantly affected the current Kasting, 1992. , a decrease in the sun’s luminosity carbon flux from the Arctic tundra ecosystem, turn- and a potentially corresponding decrease in mean ing the latter from net CO2 sinks to major atmo- global temperature in early Earth history can be spheric CO2 sources because of the large amount of compensated for by a corresponding decrease in dead organic matter stored below ground. Nadelhof- silicate weathering rate which means more CO2 in fer et al.Ž. 1999; see also Raymo and Horowitz, 1996 the atmosphere to offset the temperature decrease emphasize that atmospherically available nitrogen due to lower luminosity. In this way, the feedback may have a significant effect on North Temperate mechanism between silicate weathering and atmo- forest growth that in turn acts as a carbon sink, an spheric CO2 , which is either consumed or increased additional complicating factor. Indermuhle et al. depending on the level of chemical weathering, re- Ž.1999 suggest, after studying ice-contained gas from sists both large increases and large decreases in an Antarctic core, that A . . . atmospheric variations global temperature and helps to maintain over the during the Holocene were driven by a combination long-term, stable global surface temperatures. of growthŽ. 11–7 kyr BP and decay of terrestrial Among geophysicists, geologists and paleontolo- biomass, and by an increase in sea surface tempera- gists, a number of models have been proposed that ; 8 tureŽ. SST of 0.5 C between 9 and 6 kyr BP. In attempt to trace variations in atmospheric CO2 addition, the slow re-equilibration between the ocean through time and determine controls. Our purpose and sediment systems in the wake of the glacial–in- here is to discuss the reliability of Phanerozoic atmo- B terglacial transition may also have contributed. spheric CO2 estimates, the climatic implications Raich and SchlesingerŽ. 1992 point out that soil based on several of these proposed models, and the respiration rates of CO2 should also affect levels of reliability of the proxies on which these models are atmospheric CO2 with a potentially positive correla- based. We focus on models proposed by Berner and tion between higher rates of soil respiration and co-workers over the past decadeŽ herein the Berner enhanced accumulation of greenhouse gases and model. because they are widely cited and accepted. global temperature. They note a close positive corre- Additionally, these models are most interesting to us lation between mean annual soil respiration rates and because of their implications about global climate mean annual net primary productivity but point out and atmospheric composition in a pre-tracheophytic that the rate at which CO2 moves from soil to world. For comparison we review a number of other atmosphere is influenced by a variety of physical models and the proxies on which they are based, factors while there is considerable variance in soil pointing out differences in predicted atmospheric respiration rates among major vegetation biomes to- CO2 and temperature that distinguish them from the day, and presumably in the past. Stephens and Keel- Berner models. ingŽ. 2000 suggest that waxing and waining of These models and proxies are considered against

Antarctic sea ice may also affect atmospheric CO2 an isotopic background in which the ratio of inor- levels. ganic:organic carbon isotopes, C12rC13, does not If such known AminorB perturbations can affect change significantlyŽ. Schidlowski, 1988 from the atmospheric CO2 , how much more difficult is it to Late Archaeozoic to the presentŽ remaining ca. 99% predict varied factors back into geological time, and carbon-12 and 1% carbon-13. . The light AorganicB the potential controls that present many more biotic carbon-12 isotope is carbon sequestered by au- and geologic imponderables? During the more re- totrophic and heterotrophic organisms, respectively A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 7 through photosynthesis and the consumption of pho- escape of terrestrial thermal infrared radiation and tosynthetic organisms, and present in sedimentary this radiative energy is reradiated back to Earth, rocks as reduced carbon. The heavy AinorganicB heating the surface. carbon-13 is carbon present as oxidized carbon in Only in the most recent time interval are direct carbonate rocks, that result ultimately from Ca and measurements of atmospheric CO2 possible, from Mg derived from the weathering of Ca- and Mg-rich atmospheric gas trapped in reliably dated glacial ice. silicates combined with atmospheric CO2 dissolved During the past 160,000 years, for example, climatic in waterŽ. bicarbonate and precipitated out as lime- warming is commonly associated with an increase in stone and dolomite. Stability in the organic:inorganic CO24 and CH levels in the atmosphereŽ Stauffer et carbon isotope ratio in, respectively, disseminated al., 1998. ; climatic cooling with a lowering of CO2 organic matter, kerogen, of sedimentary organic and CH4 . WoodwellŽ. 1995, p. 3 and Woodwell et rocks, and in sedimentary carbonates from the Late al.Ž. 1995, p. 395 note that during this time interval a Archaeozoic to the present, implies that there have 18C change in temperature produced an approxi- been similar major feedback mechanisms in place mately 7.5-ppm change in CO2 Ž. Woodwell, 1995 or from the beginning with limited change in the qual- a change in the atmospheric burden of about 15 ity and quantity of photosynthesis for approximately petagramsŽ 1015 . carbon. Oeschger and Stauffer 3.5 billion years or in Schidlowski’sŽ. 1988 words Ž.1986 note a number of technical problems in the Aan extreme degree of evolutionary conservatism in ice core methodology, as well as some AnaturalB the biochemistry of photoautotrophy . . . B ŽSchidlow- problems, but overall they conclude that the ap- ski and Aharon, 1992; see also Schidlowski, 1986, proach has real merit. They suggest that because of

1988. . This implies both the existence of, and a potential problems in measuring atmospheric CO2 in similar flourishing of photosynthetic life in available ice cores, climatic data preserved in the geological ecosystems through time, although it is possible that record related to a variety of physical and biological the absolute amounts of carbon have changed in one proxies provide the best means of investigating way or another. changes in atmospheric CO2 concentration over time and its role in controlling global climate. Sigman and BoyleŽ. 2000 provide an excellent summary of the

2. Phanerozoic climate and atmospheric composi- ice core CO2 data, while considering the difficulties tion in trying to understand the global carbon cycle. If the greenhouse hypothesis is correct, there 2.1. Introduction should be a positive correlation between changes in

atmospheric CO2 and climate during the last 600 The Phanerozoic history of the Earth’s atmo- m.y. as there appears to be in the last 160,000 years. spheric composition is regarded as essential for pre- There are no techniques for directly measuring atmo- dicting surface temperature and climate over time. spheric CO2 except for the past few hundred thou- The tenor of atmospheric CO2 is especially impor- sand years; all such measurements are based on a tant because of the role attributed to accumulation of variety of proxies, discussed below. Estimating past atmospheric CO2 as a major factor in global warm- atmospheric CO2 levels for time intervals more dis- ing, e.g. as the major greenhouse gas that helps to tant than a few hundred thousand years, when CO2 control Earth’s surface temperature. According to the is directly measurable, involves many variables, hypothesis of global warming, when atmospheric many approaches and many problems.

CO2 is high, surface temperatures are predicted to be Methane is another greenhouse gas. Today it is a high and when atmospheric CO2 is low, surface minor atmospheric component, but recent concern temperatures are predicted to be lowŽ but see Robin- with the potential release of methane hydrates from son et al., 1998; Schlesinger, 1991. . The predicted the ocean floor has given it attention. Schutz et al. correlation between atmospheric CO2 and surface Ž.1990 reviewed the role of methane in the modern temperature is based on the fact that high atmo- environment. Methane is chiefly produced by the spheric CO2 absorbs and prevents or decreases the anaerobic destruction of organic matter in relatively 8 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 anaerobic environments provided by swamps and indicators for the pre-Pliocene Phanerozoic. Their peats today, as well as by a minor contribution from compilation emphasizes that widespread coals are the AexcretionsB of individual organisms ranging present in the Carboniferous, Permian, Late Triassic, from termites to ourselves and our domestic animals. Jurassic, Cretaceous and Cenozoic, whereas wide- A minor amount occurs in volcanic gases. Methane spread arid climate indicators characterize the Cam- is consumed in arid to semi-arid regions within the brian, Ordovician, Silurian, Devonian, plus the Early soil systemŽ. Striegl et al., 1992 , i.e. this is probably and Middle Triassic, with somewhat less widespread an important methane sink, particularly during inter- arid conditions in the Late Triassic through Jurassic. vals of widespread aridity. Schutz et al.Ž. 1990 sug- However, the correlation overall of these time inter- gest that methane accounts for approximately 20% of vals with coal and arid climate occurrences does not the greenhouse global warming during the past cen- agree very well with the summed evidence for tury. However, methane, like CO2 , levels cannot be Phanerozoic global climatic gradients. Inspection of directly measured by means of samples obtained our Fig. 1 indicates numerous disparities between from older time intervals. Still, one can suggest that overall global climatic gradients, high to low, and those past time intervals characterized by extensive the distribution of coal deposits and arid regions. areas of swamps, coal swamps in particular, might This being the case one can only conclude that if the have produced more methane than is the case for methane generating and lacking evidence cited above intervals of widespread aridity. Other things being is anywhere near correct, the correlation between equal one might infer that intervals of widespread methane and global climatic gradients is low. How- coal formation should have been overall warmer than ever, some have suggested the possibility that solid the reverse. Boucot et al.Ž. submitted have summa- methane hydrates released from continental shelf rized the distribution of coals and of arid climate sediments might be responsible for massive methane

Fig. 1. Times when warm temperate indicators are present at high latitudes are taken as indicators of intervals of low climatic gradient. Times when temperate indicators are present at high latitudes will be taken as indicating moderate climatic gradients, and times when there are cool to cold climatic indicators that suggest freezing winters will be taken to indicate Amoderately highB climatic gradients. Continental glaciation at one pole will be taken as an interval of high climatic gradient. The only bipolar continental glaciation recognized, the Quaternary, will be taken as a Avery highB climatic gradient. Based on data from Boucot, Chen Xu and ScoteseŽ. submitted . A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 9 pulses that could have added considerable organic they operate as feedback mechanisms that potentially carbon to the atmosphere very suddenly and have control atmospheric CO2 through geologic time and acted as a greenhouse gas well before any human tend to remain in balance on time scales of roughly interventionŽ see Norris and Rohl, 1999, for a Late greater than 105 yearsŽ. Volk, 1989 . In general, there Paleocene–Early Eocene possibility; Cowling, 1999, is no consensus on which sources and sinks are most for a mid-Cenozoic warming possibility; Kennett et important, and in some cases what are the driving al., 2000, for Quaternary evidence. . mechanisms that add or subtract carbon, such as the relative importance of silicate mineral weathering.

2.2. Carbon and CO2 sources and sinks Models attempt to look at the many variables that affect this cycle and the time scale on which these The Ca and Mg released by the weathering of Ca transfers operate through geologic time. Some of and Mg silicate rocks combines with bicarbonate, these variables bear on potential sources of atmo- from the solution of carbonate rocks, to precipitate as spheric CO2 ; others bear on potential sinks of atmo- limestone and dolomite in the marine environment, spheric CO2 . Sources and sinks operate on both long as well as with atmospheric CO2 dissolved in rain and short time scales that need to be evaluated water that is also transported as bicarbonate in sur- individually. We note below a number of what have face runoff and ground water. This completed cycle been deemed the most important AsourcesB and A B represents the transfer of CO2 from the atmosphere sinks that have been used in model construction. to the lithosphereŽ. Drever, 1994, citing Urey, 1952 . As Brady and CarrollŽ. 1994, p. 1855 note, accurate These oceanic carbonates constitute a major long- identification of carbon sources and sinks is essential term CO22 sink. From this lithospheric sink, CO to predicting variations in atmospheric CO2 over the may be recycled back into the atmosphere over Earth’s life time. We also offer some critical com- geologic time due to erosion and metamorphism. ments related to these sources and sinks.

Thus, models that calculate atmospheric CO2 levels These sources and sinks are played out against a and mean global surface temperatures rely heavily background over geologic time of varied paleogeo- on estimates of CO2 consumed by chemical weather- graphic information with different land–sea distribu- ing at the Earth’s surface just as they commonly rely tion and different sizes and heights of continental on the volume of sequestered carbonate rocks. Brady land massesŽ. cf. Otto-Bliesner, 1995 . Ž.1991, pp. 18, 101 notes that whereas Ca and Mg silicates are the primary atmospheric CO2 sink in weathering, alkali feldspar dissolution rates or car- 2.2.1. Carbon and CO2 sources bonate aquifer data that are generally used in models CO2 sources provide positive feedback in the cause calculated CO2 levels and temperatures to tend carbon cycle, i.e. additional gaseous CO2 . Advo- to be anomalously highŽ. cf. Berner et al., 1983 cated sources that introduce or recycle CO2 back indicating a looser coupling between chemical into the atmosphere include volcanic gases; surface weathering and climate than weathering rates calcu- oxidation of organic matter in soils and sedimentary lated from Ca and Mg silicates. Gwiazda and rocks; oxidation of contemporaneous plantŽ includ- BroeckerŽ. 1994 point out that opinions are divided ing both lower autotrophs and tracheophytes. and on a mechanistic link between levels of CO2 and its animal remains; solution of limestone and dolomite sinks. in marine and non-marine waters; metamorphism

The transfer of carbonŽ as CO2 and as organic and magmatic conversion of carbonate rocksŽ meta- and inorganic carbon. between biosphere, litho- morphic decarbonation, Volk, 1989, p. 107. where sphere, oceans and atmosphere is termed the carbon the CO2 reaches the surface dissolved in water under cycleŽ see Sigman and Boyle, 2000, for helpful pressure; and metamorphism of organic material discussion. . The carbon cycle depends on feedback which produces CO24 as well as CH and CO. between a number of sources and sinks, and operates Petsch et al.Ž. 2001 provide a microbial mecha- on both short- and long-term time scales. The rela- nism for oxidizing kerogen back into the atmo- tive importance of carbon sources and sinks is that spheric CO2 . 10 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

2.2.2. Carbon and CO2 sinks weathering phenomena since it is the Ca and Mg CO2 sinks provide negative feedback in the mod- derived from weathering of silicate rocks that pro- els, i.e. remove gaseous CO2 . Continental weather- vides the necessary cations for precipitation of these ing of Ca and Mg silicate minerals, and transfer of carbonate rocks. Increased Ca–Mg silicate mineral the resulting Ca and Mg ions via continental runoff dissolution rates lead to an increase in Ca and Mg into oceans and lakes with subsequent surface pre- ions in surface runoff. This in turn leads to an cipitation of Ca and Mg carbonates as limestone and increase in precipitation of Ca–Mg carbonates which dolomite from CO2 dissolved in water is a major lowers atmospheric CO2 by removing that part of the sink. This sink transfers CO2 from the atmosphere– atmospheric CO2 used in the precipitation. Lowered ocean to carbonate rocks. Photosynthesis by marine atmospheric CO2 leads to lowered global surface and non-marine autotrophic organisms with conse- temperatures. quent sequestering of CO2 ; sequestering of organic Others have also examined the role of low-tem- matter in sediment and sedimentary rocks and soils, perature seafloor weathering as a major Ca source the latter often ignored by modelers; and the poten- and as a major sink for carbon in the oceans and tial of oceanic weathering of oceanic ridge basalts to atmosphereŽ Caldeira, 1995; Francois and Walker, yield Ca ions which then reprecipitate as carbonate, 1992; Brady and Gislason, 1997. and have con- thus forming a potentially large carbon sink; all cluded that it may or may not be as important as, or merit consideration. more important than, weathering of continental rocks. Many models see a compelling correlation be- The basic problem is that laboratory experiments tween global climatic change through geological time dealing with basalt dissolution at different tempera- and chemical weathering of silicate minerals in con- tures combined with assumptions about seafloor tem- tinental rocks because atmospheric CO2 is ‘con- peratures do not provide unique solutions. There are sumed’ during silicate dissolution removing atmo- few data compiled about the weathered or non- spheric CO2 that in turn affects global surface weathered condition of oceanic basalts from JOIDES climate. Thus, the rate of silicate weathering plays an cores, nor about how much higher temperature important role in ’dictating’ increases and decreases weathering is done at the oceanic spreading centers. in atmospheric CO2 over time and in global climatic Until such information becomes available, the rela- change, or to put it another way, the rate of silicate tive importance of seafloor weathering vis-a-vis con- weathering plays a major role in buffering the amount tinental weathering will remain uncertain. of atmospheric CO2 . Weathering of continental rocks according to some 2.3. Criticisms may be the major sink of atmospheric CO2 and thus A B A B one of the major controls of atmospheric CO2 and Few of these sources and sinks can be mea- climate. Dorn and BradyŽ. 1995, p. 2847 state that sured or estimated in any meaningful way. There is A . . . the response of silicate weathering to global no accurate estimate, for example, of the amount of environmental change is one of the primary determi- CO2 released during Phanerozoic volcanism, al- nants of atmospheric CO2 levels, and through the though if there were a reliable synthesis through time Greenhouse Effect, global climate.B Consequently, of Phanerozoic volcanics of varying types such an any physical or biotic factor that alters the amount or estimate is potentially possible. There is currently no rate of silicate weathering, such as size, relief and effective method for estimating the amount of CO2 elevation of land masses, total runoff, humidity, released during metamorphism through time. temperature and vegetation will ultimately affect the Nor is there, unfortunately, an accurate synthesis deposition of marine and non-marine Ca and Mg of the amount of carbonate rock either deposited or carbonates, and the significance of the organic and dissolved through the Phanerozoic, although in prin- inorganic carbon sink that figures so prominently in ciple estimates could be made after the massive, most model reconstructions. unsynthesized geological data are brought together. The volume of Ca and Mg carbonates, an impor- There have been no attempts by any modelers to tant CO2 sink, would provide some estimate of estimate differing rates, or actual amounts, of rock A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 11 weathering during the Phanerozoic, using data from tracheophytes have played a major weathering role the rocks themselves. All weathering estimates are and only in soil environments. Involved here is the based on proxies, such as solute fluxes in river problem of trying to estimate whether or not the Late waters, of indirect value rather than rates of uplift Silurian–Devonian advent of tracheophytes resulted through time and the volumes of the varied rock in far higher levels of chemical and physical weath- types subject to weatheringŽ. see below . Bluth and ering than occurred in the pre-tracheophytic world. Kump’sŽ. 1991 effort involves too many dubious Assessing the absolute and relative importance of assumptions to be useful, since they use the out- these sources and sinks and the relative time scales moded Ronov maps and gross time intervals, ignore on which they operate through geological time is the effects of relief, and do not effectively estimate critical to accurately estimating the amount of atmo- r rock volumes time interval. spheric CO2 and global temperatures through time. Few agree on the ultimate controls on weathering This has not been done. or the correlation between weathering rates and climatic variability. Some models stress biotic factors 2.4. An alternatiÕe and the effect on weathering rates of different plant ecosystems over evolutionary time. Other models An alternative to the sources and sinks approach, ignore the effects of biotic weathering in favor of which is difficult and time consuming to evaluate, is physical and geochemical variablesŽ. see below . to estimate the ever changing Phanerozoic global For example, estimating the volume of Ca- and temperature gradient from the distribution of climati- Mg-rich rocks being weathered through the Phanero- cally sensitive depositsŽ evaporites, coals, calcretes, zoic in principle would indicate the variability of bauxites, laterites, kaolins, tillites, dropstones, glen- weathering for varied time intervals that is poten- donites. and temperature-sensitive animals and plants tially capable of sequestering atmospheric CO2 . Any with living representatives. For example, the pres- abiotic or biotic factor that can potentially increase ence of high latitude evidence of warmth indicates a the rate of weathering therefore assumes importance. low global climatic gradient and evidence for high For example, the Berner models see tracheophytes as latitude cool to cold climates indicates a high gradi- the ultimate cause of variability in weathering rate, ent. It is possible to provide some reliable estimates while acknowledging the importance of numerous of atmospheric CO2 on the assumption that the other variables, such as temperature, climate, relief, changing global climatic gradient indicated by the etc., without evaluating them with reliably compiled distribution of climatically sensitive deposits and geological data. organisms is a direct function of atmospheric CO2 . It Just as there is no accurate estimate of most is possible to model the temperature effects of topog- sources, neither is there an accurate synthesis of the raphy and of the global distribution of the land varying amounts of Ca and Mg silicate minerals masses, in combination with the climatically sensi- weathered during the Phanerozoic, although unsyn- tive data for a better estimate of climatic gradient thesized paleogeological data might provide an esti- and climate distribution. mate. It is also not yet possible to provide global estimates of the rock and soil mineral-weathering capabilities of tracheophytes, and their commonly 3. Atmospheric CO2 proxies associated symbiotic fungi plus their associated root microbiota, as contrasted with lower autotrophsŽ pho- In the absence of reliable syntheses of the geolog- tolithotrophs and chemolithotrophs. , heterotrophs ical data bearing on climate, referred to above, others Ž.Žchemoorganotrophs cyanobacteria, eubacteria, have used varied proxies to estimate the magnitude actinomycetes, algae, fungi.Ž and lichens fungal–al- of individual sources and sinks from evidence avail- gal AsymbiosesB.Ž. Some AmodelsB see Schwartzman able in the geological record. Below we provide et al., 1993. attach a great deal of importance to accounts of previously proposed proxies for atmo- biotic weathering by varied microbiota. But some spheric CO2 through time from which independent that integrate biotic information conclude that only atmospheric CO2 curves have not been prepared as 12 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 well as proxies for which independent atmospheric too is discussed at length below. The relative stabil-

CO2 curves have been prepared. ity of the organic carbon:inorganic carbon isotope During the past 20 years, a number of atmo- ratio, and its significance for these problems, is also spheric CO2 models have been proposed. Since at- considered below. mospheric CO2 cannot be measured in the distant Independent proxies for which Phanerozoic CO2 past directly a number of proxies have been pro- curves have not been proposed are: stomatal infor- posed. We will review some of these models below. mation and higher land plant Cretaceous biomass. Before doing so, however, it is clear that most of the Below we consider eight models with their prox- disagreement over the relative importance of the ies that have been proposed to estimate Phanerozoic varied proxies is needless if one pays attention to the atmospheric CO2 , to explain changes in the Earth’s basics of Phanerozoic bedrock geology. Many of the atmospheric composition through the Phanerozoic. arguments concern just what is the control, or con- These demonstrate the widely disparate proxies used trols, over silicate mineral, bedrock weathering that and widely different conclusions determined by dif- in principle provides Ca and Mg ions that combine ferent approachesŽ see also Sundquist, 1986, who with bicarbonate ions in the oceans to form carbon- discusses additional models. . The significance and ates. For much of the Phanerozoic, very large conti- importance attached to different proxies differs in nental areas have been covered by epicontinental different models. For example, in some models, seas, thus removing them from the possibility of changes in the isotopic composition of the strontium bedrock weathering. Non-marine continental areas preserved in carbonate rocks of varying age is criti- during much of the Phanerozoic are largely low calŽ. see discussion below ; in others, changes in the elevation, low relief ApeneplanedB areas, that pro- carbon isotopic composition of pedogenic calcretes vide little possibility for silicate mineral weathering through time assume major importance; still in oth- under either warm humid or warm arid climatic ers, estimates of surface runoff through time com- regimesŽ. see discussions below . One is therefore bined with estimates of modern river water composi- restricted for significant silicate mineral weathering tion that is extrapolated back through time are most to areas of higher relief that involve tectonism andror critical. All of these possibilities involve complica- vulcanism. When such tectonically uplifted or vol- tions that deprive them of reliability in estimating canic areas occur in cool to cold regions, mostly at atmospheric CO2 through time or past climates. We higher latitudes, there will be little chemical weather- provide an evaluation of each proxy discussed. ing. Contrariwise, in lower latitude regions subject to We will use the following format: discuss proxies warm, humid conditions one should expect signifi- unaccompanied by models and their problems. Next cant silicate mineral weathering. Also involved is the we will summarize each model, and its proxies, question of whether radiogenic:non-radiogenic stron- presenting the rationale on which it is based by the tium isotopes derived from marine carbonate rocks authorŽ. s . We follow this with a discussion of some may be used as proxies for weathering intensity. This of the problems presented by each model and its possibility is negated by the fact that large areas proxies. subject to weathering commonly include significant areas of older carbonate rocks, marbles, limestones and dolomites, as well as carbonate containing 4. Proxies without independent atmospheric CO2 greenstones, that may provide significant older radio- curves genic strontium that negates the simplistic possibility. There is the additional problem of whether most 4.1. Stomates silicate mineral weathering takes place under the control of the co-evolved mycorrhizal:tracheophyte A biotic source of potentially independent, direct relation or not; we discuss this possibility and its estimates of atmospheric CO2 levels and fluctuations problems at length below. There is the allied prob- since the Early Devonian is stomatal anatomy and lem of whether or not lower autotrophs have any stomatal frequency indicesŽ stomatal density and significant silicate mineral weathering capability; this stomatal index defined as density of stomates ex- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 13 pressed as a percentage of the density of epidermal variety of sampling problems related to taphonomic cells plus stomates by Poole et al., 1996; and as Athe variables that could skew stomatal data, including proportion of stomata to leaf epidermal cellsB by such things as numbers and kinds of leaves and their Beerling and Woodward, 1996, p. 183. from leaf representiveness in the local and regional environ- cuticle of modern and fossil tracheophytesŽ for ex- ment. Such problems may be significantly exa- ample, Woodward, 1987; van der Burgh et al., 1993; cerbated in the Paleozoic and especially in the Ferris and Taylor, 1994; Beerling and Woodward, pre-Carboniferous where fossil plants are both un- 1996; Kurschner et al., 1996; McElwain and Cha- common and commonly represent transport assem- loner, 1995, 1996; McElwain, 1998; Poole et al., blages subject to preservational biasŽ durability for 1996; Royer, 2001; Royer et al., 2001. . It has been example.Ž. . Poole et al. 1996, p. 710 emphasize the suggested that stomates, providing diffusion path- need for large sample size if the data Aare not to be B ways for CO2 and water vapor into and out of the susceptible to random error because of the natural leaf during photosynthesis and transpiration, respec- variability of stomates on the leaf surface. tively, are an anatomical proxy of CO2 levels in the atmosphere in which they grew expressed by an 4.1.1. Physical and biotic Õariables affecting stom- inverse relationship between stomatal frequency atal densities and occurrences Ž.stomatal density and stomatal index and ambient A large number of physical variables, related to atmospheric CO2 concentration. One example, study growth conditions, have been implicated in stomatal of herbarium specimens of certain tree leaves, indi- characters, such as stomatal density, beside CO2 cates that stomatal density and index have both levels. These include light, temperature, elevation, decreased with increase in atmospheric CO2 concen- water and nutrient levels of the soils in which the trations since the Industrial Revolution in the mid- plants are grown; these differences may be reflected 19th CenturyŽ Woodward, 1987; additional herbar- at different growth stages during life of the plant ium studies are cited by Beerling and Woodward, Ž.Ž.references in Poole et al., 1996 . Bakker 1991 has 1996. . If stomatal density responses to atmospheric shown experimentally that stomatal density and size

CO2 changes over the past few centuries can be vary with humidity. Poole et al.Ž. 1996 also empha- established, it could be argued that the relationship size biological variables such as Aleaf insertion posi- between stomatal density and ancient atmospheric tionB and developmental stageŽ. ontogeny and postu-

CO2 through geologic time, could be demonstrated lated that there might be differences in stomatal in fossil leaves to provide not only a method of development and distribution in different growth determining changes in the atmospheric pCO2 , but forms, herb as opposed to tree, for example, as well an important proxy to paleoclimates in the geological as during ontogenetic development. Beerling and record. Use of stomatal data requires, however, for WoodwardŽ. 1996 have also discussed the potential pre-Pliocene plants, but especially for Paleozoic differential effects of geographic distribution and plants, that the climatic signals generated in response differential photosynthetic activity in different vege- to variations in CO2 , can be correctly interpreted for tation typesŽ highest stomatal densities occur in the plants with no living taxonomic, morphological or most photosynthetically active vegetation for exam- ecological counterparts. In the Early Devonian, for ple. . The literature dealing with the correlation be- example, low stomatal frequencies have been corre- tween stomatal density and atmospheric CO2 indi- lated with the high CO2 levels postulated in some cates that the relationship is complex and commonly models. But can it necessarily be concluded that this ambiguousŽ. Poole et al., 1996 . is a cause and effect relationship and not related to A large number of identified physical variables other environmental variablesŽ cf. Edwards et al., related to growth conditions bearing on stomatal 1996; Edwards and Axe, 1992.Ž. ? Berner 1998 uses response exist not only for single plants but for stomatal indices to reinforce his conclusions but at different taxa, even living within the same commu- the same time the stomatal workers use the Berner nity, and response cannot necessarily be assumed or models conclusion’s to reinforce or support their demonstrated to be related to atmospheric CO2 con- data. Additionally, it is possible that there are a centration. For example, Kurschner et al.Ž 1997, 14 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 p. 26. note that leaf anatomical alterations may there need not be a 1:1 correlation between stomatal A B involve profound changes in stomatal density that densities and atmospheric CO2 concentration. Some are closely affected by amounts of incident light have suggested that the AresponseB may be species- available during growth. In this regard, Wagner et al. specificŽ. Malone et al., 1993 , others phenotypic. Ž.1996 studying a single Betula xus tree, and Poole Wagner et al.Ž. 1996 emphasize the potential for et al.Ž. 1996 studying a single Alnus glutinosa tree, considerable phenotypic variation in the annual and both demonstrate significant differences in stomatal semi-annual responses even though there was some density and index of abaxial surfaces of AsunB and long-term response in the long-term reaction of a AshadeB leaves as well as major stomatal variations single B. pendula over some 40 years, as well as a Ž.intra-leaf variation within single leaves that do not long-term reduction in stomatal density correlated occur in a consistent pattern on the leaf surface. Both with increasing atmospheric CO2 , i.e. leaf samples studies demonstrate a high level of variation both of a limited kind may well provide misleading data. within and between leaves of a single tree. This is a potentially significant sampling problem. Light intensity, only one of many factors in the Woodward et al.Ž. 1991, p. 30 suggest significant environment of a plant, has such profound effects on genetic variations in the mechanisms of CO2 re- variations of stomatal density on sun and shade leaf sponses that may affect leaf area expansion, among morphotypes, that Athe potential errors for palaeoat- other things. If genotypes within a species differ in mospheric CO2 reconstructions if natural variation response to CO2 concentrations, there may be genet- and sampling bias are not taken into accountB ŽPoole ically controlled within- and between-population dif- et al., 1996. are profound. ferences in stomatal CO2 response, seen as marked Even the response to changes in CO2 levels may and unpredictable responses of different species. be unpredictable. For example, under experimentally Ferris and TaylorŽ. 1994, p. 448 comment: ASince increased CO2 , stomatal response may be highly stomatal density can also be affected by a number of variable, sometimes decreasing, sometimes increas- environmental factors including altitude, light and ing, and sometimes being unchanged; moreover it shade, temperature and water availabilityŽ Schoch et can be demonstrated that the response may be differ- al., 1980; Yegappan et al., 1982; Ferris, 1991; Rahim ent on different surfaces of the same leafŽ Weyers and Fordham, 1991. it is possible that in many of the and Lawson, 1997, pp. 335, 339; Ferris and Taylor, historical studies the effects of elevated CO2 were 1994; Poole et al., 1996.Ž . Ferris and Taylor 1994, p. confounded by other environmental variables affect- 452. studied stomatal characteristics of four herbs ing the expansion of leaf areaŽ Beerling and Chaloner, native to the chalk grassland community of southern 1993a.Ž. .B Wagner et al. 1996 emphasize the high Britain and commented on the amount of variability: level phenotypic variation in stomate density, i.e. a AThe results of this study show that there are large serious sampling problem if one deals with a small effects of elevated CO2 on the stomatal character- number of fossil leaves. istics of native herbs commonly found in chalk Direct use of stomatal information to determine grasslands. Perhaps most significant was the discov- CO2 levels for extended geological time, that is to A B ery that elevated CO2 caused both increases and model atmospheric CO2 changes, would have to decreases in stomatal indexŽ. SI , stomatal density, fulfill a number of requirements: that the effects of epidermal cell density and size, depending on evolution, taxonomy, habitat and other environmen- Species.B With the single parameter, stomatal den- tal variables on stomatal frequency and anatomy be sity, they point out that the parameter is Aextremely disentangled from atmospheric effects, that well-pre- variableB between species and habitats for example, served cuticular remains for which it is possible to and can be affected by environmental factors such as see epidermal cell outlines and number of stomates, altitude, light and shade, temperature and water are common and continuously available in the strati- availability. graphic record, and, additionally, that taphonomic Weyers and LawsonŽ.Ž 1997 and also Malone et effects related to preservation and occurrence of al., 1993; Ferris and Taylor, 1994; Rosenzweig and leaves and other stomatal-bearing remains can be Hillel, 1998; Poole et al., 1996. make it clear that disentangled. One might predict that fulfilling all A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 15 these requirements will not be simple for modern In contrast, the NLE concept assumes that one plants and fossil plants with living relatives, let alone can deduce the environmental requirements of an in wholly extinct plants of the Paleozoic! Poole et al. extinct taxon by employing an enÕironmentally simi- Ž.1996, p. 705 caution AUntil the variability in extant lar extant taxon, rather than a taxonomically similar material has been fully assessed, conclusions drawn one. McElwainŽ. 1998 defines NLE species as Aplant solely from counts of stomata from fossil material species from the present day that show ecological are of limited value and must be treated with ex- andror structural similarity to the fossil plants with treme caution.B which they are being comparedŽ in terms of their stomatal parameters. ,B and later states that A . . . mor- 4.1.2. Application of stomatal data to atmospheric phological matches . . . B and Asimilar ecological B CO2 estimates habitats . . . are the two requirements for the selec- Despite the demonstrated variables that affect tion of a suitable NLE species. Unfortunately, McEl- modern material based on herbarium studies and wain concludes that similar structuresŽ morpholo- experiments involving modern taxa raised in atmo- gies.Ž imply similar environments p. 87, AAll the spheres of elevated CO2 , attempts have been made Recent species chosen as NLE species comply with to use stomatal information secured from fossil ma- both of these ecological wxitalics ours criteria.B.. terials as CO2 proxies. Some studies have employed Consideration of modern higher land plant structures single species of Cenozoic age. Others have ex- does not support a 1:1 correlation between plant tended this information to taxa without close living structure and environment. In essence, McElwain is relatives. concluding that morphological similarities are far To cope with the problem of fossil taxa with no more environmentally significant than taxonomic close living relatives, McElwain and ChalonerŽ 1995, similarity in the study of pre-Late Tertiary plants. 1996. introduced the concept of Anearest living But, if the taxonomic relation is of no consequence equivalentŽ. NLE BA. . . as contrasted with the near- why does she regard the comparison between Ctenis est living relativeŽ. NLR B approach of the paleoecol- and Zamia with any concern? AOn the basis of these ogist. The traditional, uniformitarian NLR concept uncertainties regarding the Mesozoic cycads, the points out that one may estimate the environmental comparison of Ctenis and Z. furfuracea stomatal requirements of a fossil taxon by concluding that it data must be viewed with some caution.B Unfortu- should fall somewhere within the environmental re- nately, the determination of ancient environments is quirements of its closest living relatives, which leads not always a straightforward, unambiguous business. to the conclusion that a fossil assigned to an extant McElwain and ChalonerŽ. 1995 , for example, have species should probably have environmental require- taken stomatal density data available from two ex- ments essentially identical to those of the living tinct phylum-level taxa from the Early Devonian, members of the species, that a fossil assigned to an concluded that they existed in a Apoorly drained extant genus should probably have environmental environment,B and used a modern Awetland habitatB requirements no broader than those enjoyed by the monocotyledon as the comparator. Taxonomically it living species of the genus, and that a fossil assigned is clearly risky to compare stomatal densities, or to an extant family should have environmental re- anything else, from taxa belonging to different phyla. quirements probably no broader than those enjoyed Additionally, their conclusion is suspect because the by the living genera and species assigned to the Early Devonian environment from which the two family. The NLR approach appears to work very extinct plants were taken was a semi-arid region, well when there are a number of extant representa- with widespread calcretesŽ. Boucot et al., submitted tives, but becomes more and more tenuous as the in which the plants were preserved in a siliceous hot number of extant taxa decreases markedly. An ex- spring depositŽ. Trewin, 1994; Rice et al., 1995 tinct genus and its species may have had environ- earlier misidentified as a bog. Thus, the comparison mental requirements very distinct from a single, or of stomatal densities between the modern and the even a few living genera assigned to the same fam- Early Devonian example has no validity taxonomi- ily. cally or environmentally. 16 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

In practice, the NLE probably can only be suc- caused by a significant increase in global tempera- cessfully applied within the later Cenozoic since one ture, using Yapp and Poths’Ž. 1996 isotopic data Ž see can be reasonably certain about the overall nature of below. . Their conclusions, as well as Yapp and Neogene floras and environments in terms of extant Poths’, are in disagreement with the geological data floras and environments, where the NLR approach Ž.Fig. 1 indicating no difference in global climatic would probably be just as effective if not more gradient across this boundary. effective in most cases. It is also clear that for the It seems clear that some attempts to make use of NLE approach one needs to pay attention to the stomatal anatomy and frequency for the Paleozoic climate associated with the fossils, and that compar- are not strictly based on independent interpretations ing materials, living and fossil, taken from different of the stomatal dataŽ see McElwain and Chaloner, modern and ancient latitudes may require significant 1995, 1996; Edwards et al., 1996; McElwain, 1998. adjustments. but are predicated on what appears to be a reason- McElwain and ChalonerŽ. 1996 employ the NLE able correlation between morphology and model- while trying to use stomatal evidence from the Juras- based CO2 levels assumed to be correct. In other sic. Here again, there is the problem about whether words, the morphological data are being AtestedB modern plant samples taken from environments very against a hypothetical model assumed to be valid and distinct from those of the pastŽ United Kingdom CO2 models in turn use stomatal evidence to bolster modern examples from a temperate climate com- the validity of the CO2 estimatesŽ Berner, 1997, pared with Jurassic examples from a warm temperate 1998, 2001.Ž. . McElwain and Chaloner 1995, p. 393 environment. and modern taxa very distinct from the admit that a variety of environmental factors, ‘novel fossils may be safely compared as environmentally physiology,’ and stomatal ‘position’ on the plant or taxonomically comparable. In this case the Ginkgo may contribute to observed stomatal differences in might be reliable but the conifers are questionable mid-Paleozoic plants without biological counterparts, owing to the taxonomic closeness and distance, re- but as they say Adespite these potential effects, we spectively of the taxa. prefer wxour italics to attribute . . . stomatal density McElwainŽ. 1998 provides additional data com- and index differences to major changes in atmo- B paring stomatal data for a variety of Tertiary and spheric CO2 . Even for fossil plants with living Jurassic taxa, some without modern relatives and relatives, can it safely be assumed that stomatal whose living environments may have been different frequency data developed for living species can be than their modern AcomparatorsB. For example, she extrapolated to extinct genera and species or that compares a modern cycadŽ a species of Zamia, living species do not have different stomatal frequen- which today is tropical. with a Middle Jurassic taxon cies when responding to the same CO2 levels or that may not have been a cycadean, and may have responding at all? The use of presumed ecological included deciduous and frost-tolerant forms. Her correlates or Anearest living equivalentsB such as the Ž.1998 comparison of Middle Jurassic extinct conifer match between the Early Devonian Aglaophyton and genera with modern Norfolk Island Pine is flawed in the extant angiosperm Juncus, is obviously open to a terms of their very different latitudes as well as the variety of criticisms and introduces a variety of usual taxonomic distinctions, despite the apparent potential problems. similarity of the warm climate, coastal environments. RetallackŽ. 2001 provides the most extensive Despite these taxonomic and environmental prob- summary of stomatal index data from the fossil lems, which she points out, she concludes that record, and attempts to use it for estimating changing A . . . standardized fossil stomatal ratio data are in levels of atmospheric CO2 from the Permian to the good agreement with both carbon isotopic data from present. His stomatal index data are take from the terrestrial and marine sources and long-term carbon literature and his own new information. He has cycle modelling estimates . . . B. restricted himself to consideration of fossil and mod- McElwain et al.Ž. 1999 , using stomatal data from ern Gingko plus three extinct Permian and Triassic Late Triassic and Early Jurassic leaves conclude that genera of the Peltaspermales that some consider the Triassic extinction of many land plant taxa was related to the Gingkophytes. Be that as it may taxo- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 17 nomically, the genera he employs are derived from nian that made possible the evolution of leaves; the areas with widely differing climatesŽ Early Permian suggestion is falsified by the data indicating that r Rhachiphyllum from cool cold South Africa; Late atmospheric CO2 during the time prior to the appear- Permian Tatarina from the arid Russian Platform; ance of true leaves was relatively low rather than Late Triassic Lepidopterus from arid to humid South high. Africa; Ginkgo species from the Mesozoic and 4.2. Tracheophyte productiÕity and standing crop Cenozoic involving different latitude warm temperate environments. . Retallack assumes that the mod- KrassilovŽ. 1999 has used another approach to ern data for Gingko biloba stomatal index controls estimating atmospheric CO2 in the Cretaceous. He will be no different than that for older examples of has taken data for standing crop of modern tracheo- the same species, for other Mesozoic and older phytes in each climate-correlated vegetational belt Cenozoic species assigned to Gingko, or for extinct from poles to tropical regions, and used it to estimate genera belonging to a separate family Peltaspemales; its equivalent amount of CO2 . For selected parts of these assumptions are highly questionable. The only the Cretaceous he has outlined analogous vegeta- extant species, Gingko biloba, from the warm tem- tional belts, estimated the areas involved by employ- perate of South China, but now widespread in the ing available paleogeographies, and then derived an temperate as an exotic, has had stomatal index data estimate of atmospheric CO2 appropriate for the studied using temperate exotics under conditions of standing crop. He has assumed that analogous mod- A B both normal and elevated CO2 conditions. How- ern and Cretaceous vegetational belts are character- ever, the study of modern leaves has not begun to ized by similar standing crop biomass. He states take into consideration all of the environmental vari- Ž.personal communication, Dec., 1997 that his over- ables that we have pointed out above as significantly all conclusions about atmospheric CO2 level do not effecting the stomatal index. There is no information correspond with Berner’s estimates. They do corre- from the present about the varied taphonomic factors spond well with climatic gradients determined from that may affect Gingko biloba leaves, nor of the geological criteriaŽ global climatic gradients through fossil leaves considered by Retallack. In view of the time summarized in Fig. 1. . above, Retallack’sŽ. 2001 conclusions concerning Upchurch et al.Ž. 1998 make clear that climatic levels of atmospheric CO2 over geological time are models taking into consideration the presence of in strong disagreement with those recently provided forests during the Maastrichtian are more realistic by Royer et al.Ž. 2001 . Royer et al. Ž. 2001 provide in than ones which ignore them. depth laboratory and herbarium studies of modern 4.3. Proxies with atmospheric CO curÕe models Gingko and Metasequoia that supports their own 2 conclusions for the mid-Paleocene-earlier Eocene and 4.3.1. Berner models mid-Miocene that are derived from work with appropriate fossil leaves belonging to the same genera; 4.3.1.1. Introduction. Since 1983, Berner has pro- they provide the first work using stomates from duced a number of stimulating and widely used fossil leaves that appears to be reliable. models that examine changes in atmospheric CO2 A great deal more work on living taxa is neces- through the PhanerozoicŽ see Berner, 1991b, 1993b, sary before stomatal morphology and frequencies 1994, 1995, 1997, 1998, 2001 for recent examples, can be employed in the interpretation of paleontolog- justification and conclusions. . ical data as a reliable proxy for levels of atmospheric The Berner CO2 models are discussed below CO2 in the past. Finally, the NLE approach is fraught together with the premises on which they are based. with so much uncertainty that we seriously question These models can be said to be Atracheophyte based.B its utility. They depend on the critical role that he and others

Allied to the estimation of atmospheric CO2 using Žfor example, Beerbower, 1985; Knoll and James, stomatal data is Beerling et al.’sŽ.Ž 2001 see also 1987; Drever, 1994. assign to tracheophytes in di- Kenrick, 2001. suggestion that it was the alleged rectly or indirectly promoting silicate mineral weath- decline in atmospheric CO2 during the later Devo- ering during the Devonian and younger Phanerozoic. 18 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

All Berner models consider somewhat different vari- regions, and that laboratory rates are higher than the ables, adding some and discarding some from model limited field data, i.e. this is limited information with to model. We refer to these collectively as the Berner which to make sweeping, global generalizations. models. Tracheophytes are implicated from their first It is important to consider both the weathering appearance, as the primary proxy, or AcausalB organ- role through time of tracheophytes and the geologi- isms because of what are concluded to be their cal data used to provide estimates for climate models enhanced effect on chemical rock weathering. such as BLAG and GEOCARB. BernerŽ.Ž 2001 and earlier references . attributes Summing up his perceptions of the role associated silicate mineral weathering effects largely to tracheo- with the presence of tracheophytes, BernerŽ 1995, p. phytes and their commonly associated fungal sym- 574.Ž and other papers . lists: secretion of organic bionts in the rhizosphere and minimizes other possi- acids and chelates; organic litter decomposition to bilities. BernerŽ.Ž 1998 see also Berner and Berner, H23 CO and organic acids; increased rainfall and 1997. notes the wide array of other variables affect- thereby increased waterrmineral contact time; soil ing atmospheric CO2 , including topography, temper- anchorage against erosion that permits retention of ature and rainfall that potentially exercise some con- water and increases weathering time. The implication trol over silicate mineral weathering. The major role is that pre-tracheophytic carbon-producing pho- is attributed to the appearance and development of totrophs, including early embryophytes, had no tracheophytes in accelerating silicate decomposition greater effect on atmospheric CO2 and climate than and in providing the organic carbon for burial as one might expect from chemical and physical weath- resistant, plant-derived organic remains. ering in an abiotic worldŽ. Berner, 1992 . The poten- BernerŽ. 1993b, p. 374 suggests that questioning tial effects of pre-tracheophytic embryophytes and the importance of chemical weathering by tracheo- other autotrophs and non-autotrophic fungi and bac- phytes, based on silicate mineral solubilities in water teria on organic carbon production and in organic in the laboratory andror field studies of Ca and Mg carbon sequestration he considers trivial. The role of ions in natural waters, misses the influence of the sequestered, non-coal basin organic carbon, either important mineralizing microenvironment created by marine or non-marine is not considered. tracheophyte roots. He also suggests that it is a Especially critical is the advent, and assumed combination of root-associated microorganisms in rapid dispersal, of rooted tracheophytes. The CO2 conjunction with the subsurface tracheophyte root fluctuations in his models are causally related in the network that creates a significant microenvironment Devonian and post-Devonian to the appearance and for increased weathering effectiveness over Aprimi- diversification of the tracheophytes. He concludes tive organisms such as lichens, or by abiotic pro- that organic AexudatesB from plant roots and their cesses . . . B Still the question remains, as he himself commonly symbiotic fungi are of prime importance asks, Ahow much less effective . . . ?B DreverŽ 1994, in promoting mineral dissolution. BernerŽ 1992, 1995, pp. 2325, 2330. likewise writes that Adirect chemical p. 576. assumes that tracheophytes became Aeffec- effects of land plants on weathering rates of silicate tive as weathering agentsB during the Late Devo- rocks appears to be relatively minor . . . no more than nian–Early Carboniferous, when they Afirst became a factor of 2B except perhaps Ain microenvironments extensive in well-drained upland areas . . . B. This adjacent to roots and fungal hyphae.B Drever and evolutionary event, assumed to be correlated with a VanceŽ. 1994, pp. 147–148, 157 discuss the rela- globally simultaneous rapid increase in silicate min- tively greater concentrations of organic acids in rhi- eral weathering processes in soilsŽ. the rhizosphere , zosphere and mycosphere microenvironments sur- is made to correlate with an alleged, spectacular rounding plant roots, as compared with the amounts drawdown in atmospheric CO2 in the Late Paleozoic found in the soil solution away from such environ- when he concludes that silicate mineral weathering A ments, and suggest that weathering rates may be first became a significant CO2 sink. greatly accelerated in these microenvironments.B Dr- Recently, Elick et al.Ž. 1998 demonstrated that ever and ClowŽ. 1995 make it clear that the data some late Early DevonianŽ. late Emsian tracheo- concerning weathering rates are only from temperate phytes had roots penetrating to about a 1-m depth, A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 19 with an above ground plant height of 2–3 m. If such ValdesŽ. 1997 emphasize that without consideration plants were globally widespread and had a signifi- of past cloud cover we may be attributing its effects cant influence on atmospheric CO2 , they might well to other variables; this is apparently a major defect in have been capable of lowering the atmospheric CO2 many modern attempts to model past climates. Thus, level in the Early Devonian, significantly earlier in we are faced with the possibility that the short-term, time than assumed by BernerŽ. 1997 . It is pertinent relatively recent temperature correlations with atmo- m here to point out that roots as small as 100 m may spheric CO2 indicated by the Antarctic ice core data be present in bedrock fissuresŽ Zwieniecki and New- have no parallel in the longer term climatic changes ton, 1995. , which provides plants with this capability evident in the pre-Quaternary Phanerozoic. Still, of tapping deeper water supplies in arid regions the rapid, short-term Late Quaternary temperature during dry seasons; such minute roots would not be changes might not be recognizable with the less detected paleontologically. detailed data available for older time intervals.

4.3.1.2. Atmospheric CO2 fluctuations in the Berner 4.3.1.3. Berner’s data bases models. During the pre-Devonian Phanerozoic, in the Estimates and assumptions. The principal parame- absence of tracheophytes, BernerŽ. Fig. 3 postulates ters used in his Phanerozoic Models are:Ž. a se- that CO2 concentrations in the atmosphere were 15 questered organic carbon in coal basin sedimentary to 20 times the present level. Considered by him in rock, which is equated with tracheophyte productiv- this interval relative to CO2 levels are physical ity;Ž. b sequestered inorganic carbon present in lime- variables such as energy output from the sun, conti- stone and dolomite estimated from C12rC13 ratios, nental runoff and paleogeography. During the which are equated with silicate rock weathering;Ž. c Permo-Carboniferous he postulates that the atmo- weathering rates, based on amount of runoff from the spheric CO2 concentration was similar to that of the modern continents and its ionic composition in terms present related to carbon sequestered in coal basin of tracheophyte cover for which strontium isotopic rocks. His atmospheric CO2 values for the Mesozoic data are used as a proxy in some models; andŽ. d and Cenozoic are intermediate between those of the input of primordial CO2 from volcanism, particularly Cambrian–earlier Devonian and the present, with that associated with mid-ocean ridges. We will dis- postulated highs in the Triassic and Jurassic. Ceno- cuss all but the last of these, as well as some of the r zoic CO2 perturbations might be reasonable since information needed to evaluate rectify the presently there are proven changes in the concentrations of inconclusive information used in the models.

CO24 and CH in the Late Quaternary, based on Sequestered organic carbon() continental . Berner direct measurements of these atmospheric gases, Ž.1991b, 1993a, 1994, 1997 suggests that atmo- within dated Antarctic ice coresŽ Barnola et al., spheric CO2 concentration is related to the appear- 1987; Jouzel et al., 1993; Petit et al., 1997; see ance of the tracheophytes in the post-Silurian Sigman and Boyle, 2000, for a good summary. , and Phanerozoic. He suggests that only tracheophytes this verifiable record has been extended back to have significantly affected atmospheric CO2 fluctua- more than 400,000 yearsŽ. Petit et al., 1999 . The tions. atmospheric CO24 and CH concentrations correlate The use of tracheophytes to estimate CO2 in the very well with temperature in these cores, as does N2 atmosphere and potential atmospheric fluctuations in r Ž.Fluckiger et al., 1999 . However, in the pre- CO2 requires a variety of largely untested and or Quaternary Phanerozoic there is limited or no evi- untestable assumptions and hypotheses based on the denceŽ Kerr, 1999; Pagani et al., 1999; Flower, 1999 effects of tracheophytes and tracheophyte-generated review the many alternative possibilities. suggesting biomass in the ecosphere. Key among them are the that reliable correlations exist between atmospheric following:

CO2 and temperature, rather than a variety of factors Ž.1 A pronounced increase in organic carbon pro- such as changing paleogeography, surface oceanic ductivity. circulation patterns, and tectonic events resulting in Ž.2 A pronounced increase in the amount of significant topographic changes. Sellwood and recalcitrant organic carbon Atemporarily lostB or se- 20 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 questeredŽ and thus unavailable for primary produc- regions, combined with overall greater productivity tivity. , in sedimentary rocks as the result of tracheo- on the shelf and adjacent slope regions. They also phyte biomass. Most of this, in his models, ends up note, however, that the presence of tracheophyte chiefly in coal basin deposits with some deposited on lipids and abundance of n-alkanes derived from ter- the continental shelf as discussed below. Estimates restrial plant waxes indicates a Asignificant input of of sequestered biomass are necessarily correlated terrigenous organic matter to the open oceanB in with the concept of increased productivity that ac- some circumstances. They note ca. 20% of total companied the change to a tracheophyte-dominated organic carbon is introduced into some deep-sea terrestrial flora. Estimates of sequestered biomass are sediments in the central equatorial Pacific by long- also correlated with recalcitrantŽ. to oxidation ligni- range aeolian transport from the continents, demon- fied tissues of tracheophytes. Both biomass and re- strating that terrigenous organic matter may be more calcitrant lignified tissues contribute to large vol- important in some modern deep-sea sediments than umes of sequestered organic carbon in sediments and previously thought. Nevertheless, they conclude that sedimentary rocks. marine organic matter is derived overwhelmingly Berner and CanfieldŽ.Ž 1989, Table 1 and Berner, from plankton within the upper 100 m of the sea 1991b. assume that the bulk of global sequestered surface but that the settling flux decreases precipi- organic carbon is locked up in Permo-Carboniferous, tously with water depth. Cretaceous and other post-Permian Acoal basin sedi- The fact that modern oceanic sediment contains mentsB using Ronov’s dataŽ Ronov, 1976, Table 1; relatively little organic carbon provides support for Ronov et al., 1980, Table 1. . In the Ronov papers, Berner’s conclusion about modern oceans, but can- sequestered organic carbon is only tabulated in Acoal not necessarily be extrapolated to justify that similar basin sedimentsB; there are no pre-Late Devonian circumstances have pertained throughout geological coal basins. Berner assumes that no other carbon time and that information from oceanic sediment of producing organisms are capable of providing signif- the past is irrelevant. icant sequesterable carbon, while saying nothing Ž.3 A significant, direct role for tracheophytes in about significant amounts of non-coal basin kerogen enhancing weathering rates of silicate soil minerals commonly thought to be the major repository of which affect atmospheric CO2 and therefore climate. organic carbon.Ž Schidlowski, 1983, comments on In 1992, Berner discussed various speculative the presence of widespread Precambrian kerogen. . Aweathering scenariosB relevant to the effects of BernerŽ.Ž 1982 see Schlunz and Schneider, 2000 rock and mineral weathering in a Abarren landscapeB for significant modifications of Berner’s conclu- Ž.abiotic world , an Aalgal-covered landscapeB Žpoten- sions. also concludes that most organic carbon trans- tially including Acyanobacteria, lichens andror al- ported by rivers to the sea is deposited in sediments gaeB., and in a tracheophyte-inhabited world. He on the continental shelf, but he only sampled mod- concluded that there is limited or no evidence to ern continental shelf and ocean basin sediments. support the possibility of significant, biologically Mueller and SuessŽ. 1979 as well as Hedges and based chemical weathering before the advent of the KeilŽ. 1995 also conclude that most organic carbon tracheophytes. Citing a then unpublished study by is deposited on the continental shelf. Hedges and Cochran and himself, he stated, based on a perceived Keil are really using Berner’s model predictions. absence of weathering beneath a single lichen species Calvert and PedersenŽ. 1992, and literature therein on a Hawaiian basalt, that there were no greater Žsee Schlunz and Schneider, 2000, for significant weathering effects for lichensŽ or any microorgan- modification of Calvert and Pedersen’s conclusions. isms. , on Aeither a decadal or thousand-year time conclude that oceanic plankton is the chief source of scaleB than those promulgated by Asimple water–rock oceanic organic carbon, whereas lignin is the chief reaction.B He concludedŽ. 1992 , based on inspection source of continental shelf and slope carbon. They of 200-year-old lichen-encrusted arkosic sandstone conclude that most organic carbon is stored on the gravestones in Connecticut, that Aphysical weather- continental slope owing to the combined effects of ingŽ. freeze–thaw in a temperate climate is much dissolution with depth, far greater for the vast oceanic more effective than lichen-induced chemical weath- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 21 ering in breaking down feldspathic rocks,B and that enters the ocean, is critical to inferring the quantity breakup by abiotic physical weathering combined of carbonate rocks deposited. These carbonates con- with simple mineral dissolution by water is more stitute a major CO2 sink, referred to as sequestered effective than lichen induced weatheringŽ Berner, inorganic carbon. The CO2 , Mg and Ca flux may be 1992, p. 3228. . The alleged absence of lichen-in- measured from river water composition and volume duced chemical weathering is contradicted by the and from the rate at which this flux is transported extensive evidence for lichen weathering discussed downstream. Direct estimates of the amounts of below. Based largely on the result of this unpub- weathered minerals through time are currently un- lished study, BernerŽ. 1992, p. 3230 discounted available. The problem of using river water studies weathering by non-tracheophytic organisms, and as- as a basis for estimating rock weathering is further sumes as well Athat weathering by more primitive discussed later, where it is pointed out that the vascular plants was less effectiveB than that of the absence of information about the amount of weather- angiosperms. ing affecting suspended and bed load, as well as the Sequestered inorganic carbon() marine . In addi- dissolved ion content of rainfall are serious deficien- tion to sequestered continental organic carbon, Berner cies in the present data. Ž.1994 also evaluated sequestered inorganic carbon BernerŽ. 1994 later used strontium isotopic stud- in order to estimate atmospheric CO2 through the ies and modern ionic runoff information as indirect Phanerozoic. proxies for continental rock weathering throughout To evaluate this inorganic carbon, he used data the Phanerozoic. Strontium isotopic information from provided by Bluth and KumpŽ. 1991 who in turn carbonate rocks was considered a suitable proxy to used data from Ronov and collaboratorsŽ Ronov, the uplift and weathering of old radiogenic rocks in 1976; Ronov et al., 1980, 1974, for the Cambrian; mountain belts. This information was extrapolated Ronov et al., 1976, for the Ordovician; Khain et al., from a few modern AweatheringB studies based on 1977, for the Silurian; Ronov and Khain, 1954, for ionic runoff to support the role of tracheophytes in the Devonian; Ronov and Khain, 1955, for the Car- weathering. Strontium isotope data adds another boniferous; Ronov and Khain, 1956, for the Permian; variable that is difficult to evaluate; most workers Ronov and Khain, 1961, for the Triassic; Ronov and now conclude that strontium isotopic data as a proxy Khain, 1962, for the Jurassic; Khain et al., 1975, for for weathering is invalidŽ. see above and below . the Cretaceous; Ronov et al., 1978, for the Paleo- The modern autotrophic weathering basis of gene. . Berner’s 1984–2001 models. Berner’s extrapolations The inorganic carbon is mostly derived from the of weathering estimates are based on some modern weathering of silicate rocks that provide Ca and Mg field studies of the dissolved ions in river waters and cations that then combine with atmospheric CO2 the bedrock weathering effects of varied autotrophs. dissolved in river and marine waters to form lime- These are not rhizosphere studies in the sense that stone and dolomite overwhelmingly precipitated in they provide direct measurements of silicate weather- the marine environment. ing in soils. They were used to support conclusions Berner uses modern river water runoff data to about the preeminent role of rhizosphere tracheo- arrive at an estimate of weathering by tracheophytes phytes in silicate mineral weathering. One or more of and what he interprets as the limited impact of these supporting studies depend on the assessment of abiotic weathering through the Phanerozoic. Modern microbial weathering of basalt vis-a-vis tracheophyte river water is his proxy for the estimated amounts of weathering in soilsŽ. see below . We will discuss weathered Ca and Mg minerals throughout the these studies in chronological sequence. Phanerozoic. These are, of course, the ions that 1991 Model. In his 1991 model, BernerŽ. 1991b y would combine with HCO3 in stream and marine cited a published studyŽ. Cawley et al., 1969 which waters to form limestone and dolomite. The products he claimed supported his conclusion that vascular of carbonate and silicate rock weathering and plants played an instrumental role in mineral weath- y HCO3 are transported downstream in solution. The ering. Cawley et al.Ž. 1969 estimated chemical quantity and rate of this ionic flux, most of which weathering rates from bicarbonate concentrations in 22 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 runoff waters draining Aolder vegetated volcanicsB weathering in areas with plant cover as compared and Ayoung unvegetated volcanicsB in cool temper- with those that are apparently barren as assessed by ate Iceland. Berner cites this study as the only work bicarbonate concentrations. Thus, they report that the that addresses the question of how much tracheo- bicarbonate concentration can Abuild up to one-half phytes accelerate weathering compared with an the maximum observed value in the absence of AabioticB environment. SubsequentlyŽ Moulton et al., higher land plants. This in turn indicates that the 2000; Moulton and Berner, 1998. made a far more influence of higher land plants on rates of chemical detailed study of stream waters from Iceland drain- weathering in this area is not overwhelming.B They ing terrains that are inhabited only by lichen and conclude that Arates of chemical weathering in the moss, birch or a conifer, with results suggesting that absence of higher land plants in the pre-Silurian need far more calcium and magnesium was weathered not have been very different from present day ratesB from the basaltic substrate associated with the tra- Ž.p. 392 . cheophytes. However, this study has a number of DreverŽ. 1994, p. 2329 criticized the Iceland study, deficiencies in that the stream waters were sampled pointing out that Cawley et al.Ž. 1969 could not A from streams of varied lengths, with some draining distinguish between the effects of CO2 pressureŽ or far more than the area occupied by a particular organic acids. and the physical effect of longer vegetative type, no account was provided about the contact time or increased exposure of the mineral amounts of calcite one would expect to be present in surface area.B He also suggests that results of this basaltic rocks in both the matrix and veinlets, nor the study were complicated by presence of loess and ash degree to which the weathered bedrocks in each case falls. are comparable or not. All of this again indicates the BernerŽ. 1991a, p. 860 acknowledged from the need for carefully controlled greenhouse experiments Iceland study that Aestimates of quantitative effects in order to solve this question once and for all. of land biota on weathering in the geological past Berner cited the then unpublished study by will be speculative in nature . . . B HeŽ Berner, 1992, Cochran and Berner on lichen weathering of a p. 3229. pointed out that there are possibilities that Hawaiian basalt. He concluded, based on what he might Acause differences in ion concentrations that perceived as absence of weathering beneath a single are unrelated to plant activities,B such as differences lichen species, no greater weathering effects for in hydrology or microclimate. lichensŽ. or any microorganisms than those promul- 1992, 1994, 1995, 1997 models. Several models gated by Asimple water–rock reaction.B Largely published from 1992 to 1997 are supported by a based on this study, weathering by other organisms number of the same, largely anecdotal accounts about was discounted and it was assumed that weathering the effects of biotic weathering today. These models by more primitive tracheophytes was less effective all assume that tracheophytes became Aeffective as than that of angiosperms. Berner concluded that weathering agentsB during the Late Devonian–Early there is limited or no evidence to support the possi- Carboniferous, when they Afirst became extensive in bility of significant organism-mediated weathering well-drained upland areas . . . B ŽBerner, 1992, 1995, before the advent of the tracheophytes. He assumes a p. 576. . Berner assumes that rooted tracheophytes major role for tracheophytes, since bicarbonate con- were confined to wet lowland environments, until the centration in waters draining from the vegetated development of seeds and the amelioration of cli- Iceland basalt flows was about 3 times higher than in mates in the Late Devonian–Carboniferous enabled that from Aunvegetated volcanics.B their spread into drier upland environments. However, Cawley et al.Ž.Ž 1969 see also discus- The Drever and ZobristŽ. 1992 study cited by sion of this information in Holland, 1984. were BernerŽ. 1992, 1994, 1995 , attempted to estimate unimpressed with the evidence that Ahigher plantsB weathering effects along an elevational transect on greatly accelerate chemical weathering and relatively the southern side of the Alps extending from forested more impressed by the limited increase in weather- areasŽ. 220 m to unforested areas Ž 2400 m . by ing associated with vascular plants, pointing out the measuring the concentration of dissolved cations in Arelatively small differenceB in the rate of chemical waters draining the AsameB rock type. Drever and A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 23

ZobristŽ. 1992, p. 3210 saw systematic changes in gion in Colorado as evidence for tracheophyte chemical weathering as a function of increased ele- weathering capabilities. Berner assumed that analy- vation, Awhere elevation is a surrogate for . . . vari- ses of soil waters from the forested and Adowned ables such as physical erosion rate, soil thickness, forestedB regions, as compared with stream waters vegetation type and temperature.B They conclude from a largely unforested region, implied greater that so many factors important to the weathering weathering effects by tracheophytes. AThey found process, such as temperature, soil thickness and veg- that the cationic denudation rate per unit area of the etation vary along the more than 2 km elevation forested area was 3.5 times higher than that for the gradient that Ait is impossible to isolate the effects of surrounding barren areaB Ž.Berner, 1995, p. 577 . any one factor.B Although the study demonstrated a Like the Alpine study noted above, Arthur and Fahey correlation between chemical weathering and eleva- expressed concern about the contribution of hydro- tion, Drever and Zobrist reached no specific conclu- gen ion regionally from acid rain. Their study was sions about the major cause of weathering. The study not designed to assess weathering effects. did not take into consideration the contribution of In connection with elevational transect studies in hydrogen ion from acid rain to this regional effect. general, VelbelŽ.Ž. 1993, p. 1060 discussed below BernerŽ. 1992, p. 3231 suggested that a difference questions the apparent correlation between decline in in bicarbonate concentration with elevationŽ data de- silicate mineral weathering rates and increased eleva- rived from Table 1, Drever and Zobrist, 1992. might tion. VelbelŽ. 1993 concludes that there is a high be Aascribed to vegetation, and not elevation or correlation between temperature and solution rates of temperature . . . B Ž.Žor soil thickness and Drever 1994, mineralsŽ according to Gwiazda and Broecker, 1994, p. 2329. on reconsideration also concluded that the p. 153, Velbel sees this only as the effect of tempera- results of this study indicate thatŽ italics those of ture on mineral dissolution, although variables tied to A Drever. at this particular location at this particular mean temperature like soil CO2 and productivity can time in its geomorphic eÕolution the effect of vegeta- also vary. . tion on weathering rate appears to be a factor of In a brief study intended to prove a limited role about 7–8.B DreverŽ. 1994, p. 2329 states that Dr- for lichen weathering to counter the work of Jackson ever and ZobristŽ. 1992 Aattributed the difference in and KellerŽ.Ž 1970 discussed below . and to support weathering rate to two effects: the effect of tempera- the role of tracheophytes in weathering, Cochran and ture which they estimated to be a factor of about BernerŽ. 1993a,b provided an account for Hawaiian 3 . . . , and the presence of thicker soils at lower basalt flowsŽ bare, or occupied by fruticose lichens elevations, which could be attributed to the binding or vascular plants. . They stated that the common effect of vegetation.B These conclusions were not fruticose lichen Stereocaulon Õulcani, the taxon discussed in the original paper and as noted above, studied by Jackson and KellerŽ. 1970 , has no greater no conclusions were drawn about the major weather- effect upon its substrate than does Aabiotic water– ing factor. rock interactionB Žwhich in fact may not be entirely BernerŽ. 1995, p. 577 , citing the Drever and trivial: see Robinson and Williams, 1994; Webley et ZobristŽ. 1992 paper, noted that the Adata Ž from the al., 1963. . Cochran and Berner claimed minimal southern Alps.Ž. suggests a large effect 8= of vas- chemical weathering and crumbling related to expan- cular plants on the rate of silicate weatheringB if the sion and contraction of lichen holdfasts upon wet- y stream flux of dissolved HCO3 is due mainly to the ting. They provided no information on the age of the presence of trees at the low elevation. He stated that lichenŽ. had it been present for 1 year or hundreds? this Aconclusion is tentative because of the possibil- that would enable assessment of the information. ity that differences in hydrology andror microclima- Certain of this information was amended in a later tology between . . . elevations could explain the re- studyŽ. Cochran and Berner, 1996 discussed below. sults.B The weathering capabilities of lichens in generat- BernerŽ. 1995, p. 577, and submitted cited the ing a lateritic-type AproductB on well-dated Hawai- study by Arthur and FaheyŽ. 1993 , a study from a ian basalt flows is shown by information provided by high elevationŽ. 3100–4000 m Rocky Mountain re- Jackson and KellerŽ. 1970 and Jackson Ž. 1993 . Jack- 24 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 son and KellerŽ. 1970 concluded that there is an the flow surface. On these flows they concluded that acceleration of chemical weathering by an order of chemical denudation rates beneath higher plants are magnitude of at least 10 to 100 times for lichen- Aat a minimum 10 times greater than those with only covered rock relative to bare rock. Brady et al. lichens or microbiotaB while on younger flows, they Ž.1999 provide good evidence that lichen covered could detect comparatively little AweatheringB be- Hawaiian basalt is significantly more rapidly weath- neath lichen crusts. They suggest that on newly ered than the same basalts under purely abiotic con- exposed silicate rocks, vascular plants are much ditions providing support for Jackson and Keller’s more effective in promoting chemical weathering conclusions. than microbiota alone and that the appearance and Cochran and BernerŽ.Ž 1993b see also Berner, spread of vascular plants would have had a major

1992. regard information presented by Jackson and negative effect on atmospheric CO2 . But on the KellerŽ. 1970 as ambiguous and unsubstantiated evi- geologic time scale short-term weathering would dence of weathering. JacksonŽ. 1993 , Schwartzman not show up. Rather it would be with long-term wea- Ž.1993 , and Cochran and Berner Ž 1993b . engaged in thering that lichens begin to be effective in dis- a spirited exchange regarding the weathering effects solution. of Hawaiian lichen species. JacksonŽ.Ž 1993 based Following Schwartzman and VolkŽ. 1989 , Co- on the detailed study of Jackson and Keller, 1970. , chran and BernerŽ. 1993a also state that the process claimed that in zones of high rainfall, lichens greatly of lichen dust trapping Amight have allowed primi- intensified and accelerated the chemical weathering tive land biota to restrain atmospheric CO2 from of Hawaiian basalt by Aorders of magnitude.B reaching excessive values before the rise of vascular Cochran and BernerŽ. 1993b largely dismissed plants . . . B but suggestŽ. Cochran and Berner, 1993a , Jackson’s evidence for an Aintensely leached weath- based on the Hawaiian example, that lichen-mediated ering crustB as Awind-blown dustB accumulated weathering does not approach that adjacent to the around the lichen growths, while discounting sub- Afine roots of vascular plantsB on the Hawaiian strate interactions for lichens. BernerŽ. 1995, p. 576 flows. BernerŽ.Ž 1995, pp. 575–576 citing Cochran, states that the material described by Jackson and 1995. , in support of the effectiveness of tracheo- KellerŽ. 1970 as alteration of the underlying basalt phyte-induced chemical weathering, illustratesŽ his by lichens is Alargely dust trapped by the sticky, Fig. 3. plagioclase phenocrysts from a young Hawai- lipid-rich lichen thalli and altered in situ.B Cochran ian basalt Acompletely dissolved away by biological and BernerŽ. 1993a, p. 213 note that lichen capacity exudates, leaving molds of the phenocrysts in to trap and leach aeolian mineral matter, indepen- the immediate vicinity of plant rootlets . . . B dently of its effect on its substrate, might result in but indicating that away from the rootlets there is no the weathering of large quantities of rocky material evidence for such intense weathering. In general, that might otherwise be transported and buried with- higher plants do not invade and establish on such out chemical weathering. Lichens do act as binding hard, pavement-like surfaces except in microsites agents for mineral materialsŽ. Walton, 1993 which such as cracks filled with fine mineral material and may have led to some of the differences in interpre- which concentrate water and nutrients from micro- tation generated by the studies of Jackson and Keller watersheds on the flow surface or where ash was Ž.1970 vis-a-vis Cochran and Berner Ž 1993a,b . , but incorporated with the flowsŽ Aplet et al., 1998; this capacity does not negate a role for them in Kitayama and Mueller-Dombois, 1995, also empha- chemical dissolution as Cochran and Berner claim. size that in a wet, tropical region, nutrients are The Cochran and BernerŽ. 1996 study, involving quickly depleted during weathering, after rising to a the same fruticose lichen, Stereocaulon Õulcani, weathering peak early on; see also Kitayama et al., amended some aspects of their earlier study and 1995. . Therefore, it is hardly surprising that the conceded that on older flows, exceeding 3000–5000 surface area subjected to weathering by vascular years in age, Azones of intense dissolution exist plants should be greatly constrained. beneath lichensB with depth of weathering a function BernerŽ. 1995, p. 576 concluded that even if of local lichen abundance and microtopography of lichens and other microbial organisms were effective A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 25 weathering agents prior to the advent of the tracheo- bris plus micaceous schist debris rather than Mg and phytes, their slow growth rate and the limited interfa- Ca-rich substrates which deprives the conclusions of cial area between the lichen and rock surface com- much utility for answering the questions we are pared with the growth rate and interfacial area concerned with. between roots and soils must mean that Arooted The Hubbard Brook experiments were carried out higher plants are much more effective in attacking in a largely uncontrolled manner in a cool temperate rocks and weathering them . . . B region, not an AaverageB environment from which to Regardless of which side one prefers vis-a-vis the obtain results of global significance. This under- Hawaiian basalt question, there is no doubt that scores the necessity for biologists to carry out the lichens play a significant role in chemical dissolution greenhouse work necessary to evaluate the problems of rock substrates. However, conclusions based on a involved. single species, as in the Hawaiian basalt issue, are an To summarize, in order to reliably evaluate the inadequate basis for global generalization. Aplet et relative importance of tracheophytic, including an- al.Ž. 1998 studied biomass and diversity on the slopes giospermous, soil and rock weathering capabilities of Mauna Loa, showing that a number of variables, versus that of the lower autotrophs and fungi one including flow age, moisture, temperature, elevation, needs to combine several classes of information. and lava texture were important; their work, although First, one needs to have reliable greenhouse data on not dealing with lichensŽ they dealt only with vascu- varied soil and rock type weathering capabilities of lar plants. makes it clear that the superficial studies the various groups, with far more than a single taxon done by all parties to this Hawaiian controversy form for each major group, under different temperature an inadequate basis for reliable conclusions about the and moisture regimes. We need ratervolume of soil relative weathering capabilities globally of lichens as and rock weathering data, with different rock and contrasted with vascular plants on bedrock and re- soil types involved. We presently lack such infor- golith surfaces. mation. Then one needs to take this information, In his 1997 model, Berner relies heavily on the globally, and combine it with the best available AHubbard Brook Sandbox ExperimentB carried out paleogeographic and climatic information for the in southern New HampshireŽ. Cochran et al., 1994 . Phanerozoic, time interval by time interval, in order In this study, three basins were filled with sand- to estimate which of the higher taxa will dominate in size granite ŽAfeldspar-rich glacial sandB in Berner, how large an area for each climatic regime. This 1998. . One was planted with pine: size and number information will then provide a far more reliable of the propagulesŽ. whether saplings or mature trees basis for estimating the actual weathering capabili- unspecified. The second with two grass species: ties of the varied autotrophic and fungal taxa through number of adult plants and planting density unspeci- the Phanerozoic. fied. The third was left to accumulate autotrophic Most of our present data has come from temperate bacteria, algal spores and lichens from the environ- climatic regimes and vegetation types, mostly North ment. Analysis of the basin ArunoffB indicated that Temperate, which is not an AaverageB sample for the pine was far more effective at weathering than grass. Phanerozoic. Most of this temperate information ap- Grass, however, was less effective in weathering plies to soils, with very little for bedrock of different than the lower autotrophs. It is puzzling that a gran- types. Thus, most of the conclusions based on these ite sandŽ mineralogy and chemical composition un- temperate region data are scarcely applicable to ei- specified. was used rather than a basic rock rich in ther the present or the past globally, with emphasis calcium and magnesium that might have provided on both humid tropical and warm, arid regions. more information about the possibilities for seques- Without the information noted in these two para- tering atmospheric CO2 as carbonate. Bormann et al. graphs it is clear that estimates of the relative rock Ž.1998 carried out a better experiment involving the and soil weathering capabilities of the tracheophytes Hubbard Brook area, but using only pine seedlings compared with other weathering organisms, includ- versus a plot lacking megascopic vegetation, in a ing cyanobacteria, lichens and fungi can only be cool temperate environment, employing granitic de- regarded as speculative. 26 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

4.3.2. Comment and criticism of the Berner models assumed that lower autotrophs had no significant A number of GEOCARB assumptions and data effect on rock or soil mineral weathering. base choices require comment and criticism. There Organic acids produced by bacteria and the lower are two categories of information:Ž. 1 items that autotrophs play an important but often neglected role were not considered; andŽ. 2 items that were consid- in the weathering of rock-forming silicatesŽ Ehrlich, ered inadequately. We begin with the first. 1995, 1998; Barker et al., 1997, 1998; Pittman and Lewan, 1994.Ž. . Ellwood et al. 1980 provide a useful 4.3.2.1. Marine phytoplankton. Berner and Canfield review of overall microbial ecology, including mate- Ž.1989, p. 341 assumed Athat prior to the advent of rial relevant here. vascular land plants, the source of organic matter OthersŽ. see below have very differently assessed was restricted to marine organisms.B Thus, while the the role played by lichens in rock weathering, includ- presence of marine phytoplankton is not ignored, ing the lichen weathering of Hawaiian basalt. their buried biomass is implicitly assumed to be too In tracheophyte-inhabited soils, chemical weather- small to have seriously lowered the atmospheric CO2 ing can also be driven by organic acids produced by based on the very high Early Paleozoic atmospheric tracheophyte roots andror by the biomass of micro-

CO2 curve. The high curve for the Middle Cambrian bial organisms whose presence is enhanced by abun- through earlier Middle Devonian is inconsistent with dant tracheophyte-derived organic matter. In the first the geologicalrclimatological dataŽ see Fig. 1, sum- scenario, if root and root mycorrhizal AexudatesB are mary diagram. . Both pieces of information are in- the principal stimulus to weathering, chemical consistent with the conclusion that it was the pres- weathering is necessarily dependent on tracheo- ence of tracheophytes in the later Devonian that first phytes. If, on the other hand, tracheophytes con- significantly affected and depressed atmospheric CO2 tribute chiefly to soil weathering through their sup- levels. GaoŽ. 1993 provides Early Devonian oxygen port of microorganisms that could be supported isotopic data suggesting that contemporary ocean by any organic material including that from a water temperatures may have been similar to those non-tracheophytic or pretracheophytic source long of the present, and that similar data leads to the same before the advent of the tracheophytes, they are conclusion for the Ordovician and Silurian, in marked clearly not seminal to the effects of soil weather- contrast to GEOCARB predictions. ing. The potential contribution of marine phytoplank- How effective are tracheophyte-generated organic ton in the pre-Devonian might have made up for the secretions independent of other important factors? proposed deficiency in organic carbon production Do other biological andror physical factors signifi- that is claimed for non-marine autotrophic life. This cantly control root exudates? Could these other fac- assumes that non-marine lower autotrophs had no tors relegate organic exudates of the tracheophytic effect on pre-Devonian atmospheric CO2 levels, and rhizosphere to a subordinate, relatively unimportant implies an unverified major drop in marine phyto- position? plankton in the later Devonian in order to compen- In a tracheophytic world, the question seems to be sate for the major increase made possible by the the relative importance of chemical weathering that advent of the tracheophytes. Freeman and Hayes can be attributed either directly to the tracheophytes Ž.1992 consider the role of marine phytoplankton Žroot AexudatesB andror the acids produced by their Ž.we discuss their model below . mycorrhizal symbionts. , or that can be attributed indirectly to them because of the biomass of tracheo- 4.3.2.2. Pretracheophytic autotrophs. Berner as- phyte-generated organic carbon that provides nutri- sumes that CO2 utilizing non-marine organisms such ents for soil microbiota that can be viewed, rather as cyanobacteria, algae, lichens and pre-tracheophytic than tracheophytes, as the principal chemical weath- embryophytes had no measurable effect on atmo- erers in soils. If soil microbiota, such as bacteria and spheric CO2 because, in the absence of lignified fungi, are the chief chemical weatherers, then we tissue, they were incapable of making a significant would argue that any source of surface organic car- contribution to sequestered organic carbon. It is also bon could have provided nutrients to support regolith A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 27 weathering before the appearance of the tracheo- through time, that in turn could provide information phytes. about changing levels of atmospheric CO2 .

4.3.2.3. Phytogeography and eÕolution. If one main- 4.3.2.5. Need to collect adequate geological data.It tains that tracheophytes are the principal weatherers, is necessary to sample, compile and analyse the then one has to assess their biomass and area of results in order to provide the geological data needed occurrence on a global basis; area of occurrence to construct reliable atmospheric CO2 models. includes consideration of areas subject to the spec- Geological data gathered for other purposes are trum of global climatic regimes. This remains un- commonly inadequate for helping to understand done. atmospheric CO2 levels through time. Below we Phytogeographic changes, such as major vegeta- discuss several of the more important deficiencies in tional and climatic beltŽ. humid, arid, warm, cold , the data he used and the purposes for which they change over time, and potential evolutionary effects were originally gathered. of changes over time in major plant biomes are likely to be important. An example of phytogeo- 4.3.2.6. Is tracheophyte weathering eÕolutionarily graphic change is the area occupied by arid regions progressiÕe?. Knoll and JamesŽ. 1987 also attribute versus humid regions during varied Phanerozoic in- the significant role in accelerating the weathering of tervals. Evolutionary change would include the soil minerals to Avascular plants and associated myc- change from a pteridophyte-dominated to a gym- orrhizal fungi . . . ,B and increased production of soil nosperm-dominated world and the change from a organic acids, prior to the advent of angiosperms in gymnosperm-dominated world to an angiosperm- the Cenozoic, suggesting that weathering by lichens dominated world, which correspond to the Palaeo- and fungi alone would be negligible because lack of phytic, Mesophytic and Cenophytic eras of the pale- Aroot systems in fungi and lichensB and Aabsence of obotanist as contrasted with the Paleozoic, Mesozoic myccorhizzaeŽ. sic precludes any meaningful re- and Cenozoic of the paleozoologist. To these one moval of ions from soil water.B Knoll and James needs to add Gray’sŽ. 1993 pre-Eutracheophytic, up- Ž.1987, p. 1099 see tracheophytes as playing a dual per unit of her Palaeophytic, EotracheophyticŽ Late role in mineral weathering: reducing the Astability of Silurian and earliest Devonian. , Eoembryophytic soil minerals through a net export of ions from soil Ž.Middle Ordovician–Early Silurian and Phycomy- waters and through the release of complexing or- cophyticŽ. Precambrian through Early Ordovician . ganic acids by root mycorrhizae.B They writeŽ p. Gray’s units listed here deal with the pretracheo- 1100. : ARegardless of the nutrient cycling capacity phytic floras that need to be considered by all con- of plants, the presence of vascular land plants and cerned with autotrophic relations during the Precam- associated root mycorrhizae serves to enhance min- brian through the Early Silurian as they may have eral weathering,B; thus they conclude that Apost- affected atmospheric CO2 composition. Silurian soil minerals should be more weathered than Failure to plot paleogeography, to analyse the pre-Silurian soil mineralsB Ž.p. 1101 , a largely positions and areas of climatic belts, using available untestable hypothesis because of the inability to dis- paleogeologic evidence, limits the reliability of exist- criminate pre- and post-Silurian diagenetic effects. ing CO2 -weathering models. They did not provide any quantitative data. While Berner emphasizes the role of tracheo- 4.3.2.4. Climatically sensitiÕe rocks. The distribution phytes, his comment related to weathering effective- and significance of climatically sensitive sediments ness of angiosperms compared with Amore primitive that indicate the changing positions through time of vascular plantsB Ž.Berner, 1992 echoed papers by varied climatic belts and their relative widthsrareas Knoll and JamesŽ. 1987 and Volk Ž. 1989 that shift might be used to improve CO2 -weathering models. focus in soil weathering rates in the Late Cretaceous These climatic belts, when considered against what and Cenozoic, from emphasis on root systems to is known of their respective floras through time have litter fall or to what VolkŽ. 1989 refers to as Ahigh- the potential for indicating organic productivity weathering ecosystemsB. 28 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Of these, the Arole of biological activityB in rock tems developed by VolkŽ. 1989 produce formula- weathering rates is Aespecially important to the re- tions for global weathering which combine ecosys- sultsB of his modelsŽ. Berner, 1991b, p. 347 . At the tems that differ in their fractional global coverage same time, the role of biological activity provides and intrinsic rates of weathering. If however, an- Aone of the weakest aspects of my modelB because it giosperm advent should have increased weathering relies Aon assumptions about the effect of plant rates compared with pre-Cretaceous tracheophytes, evolution on the weathering of silicate rocksB and one is faced with the paradox that the predominant A because little current data can be applied to the Phanerozoic decrease in atmospheric CO2 should ancient world . . . B Ž.Berner, 1991a, p. 860 . have occurred in the mid-Paleozoic. However, the

Knoll and JamesŽ. 1987 , Volk Ž. 1989 and Robin- predicted drop in CO2 in the Berner models is many sonŽ. 1990b, 1991 have argued for differential times that shown for the CretaceousŽ. Robinson, 1991 weathering effects of different groups of tracheo- when the Earth is speculated to have been covered phytes, thus attempting to take into account one of by a high-weathering ecosystemŽ Knoll and James, the weaknesses of the Berner model—the effects of 1987; Volk, 1989. . different evolutionary groups of tracheophytes on The rationale for the emphasis on an increased silicate mineral weathering rates. Those attributing role for angiosperms stated by Berner is unstated; the weathering largely to organic acids produced by Knoll and James and Volk papers both claim that tracheophytes, have argued for an accelerating in- mineral weathering rates would be greater in soils crease in weathering through time with some corre- dominated by a deciduous vegetation than in ones sponding major perturbations in atmospheric oxygen, dominated by an evergreen vegetation and accord- related first, to the advent and spread of the tracheo- ingly lead to a decrease in atmospheric CO2 and phytes in the PaleozoicŽ. e.g. Berner , second, to the global cooling. A . . . high weathering rates of an- advent and world-wide diversification of the decidu- giosperm–deciduous systems were a result of the ous angiosperms in the Cretaceous and Early Ceno- discontinuous nature of leaf loss with the related zoicŽ. Knoll and James, 1987; Volk, 1989 . Thus, enhanced chance of severe ion loss in the runoff, in even in a tracheophyte dominated world some have comparison with the conifer–evergreens . . . B ŽVolk, attached significantly different weathering roles to 1989, p. 109. ASignificant cooling would have re- different ecosystems, involving differences in weath- sulted from the evolutionary origin and spread of ering rates to different ecosystems with the potential Ž.deciduous ecosystems with high rates of mineral that angiospermous–deciduous ecosystems, e.g. weatheringB Ž.Volk, 1989, p. 110 . ecosystems dominated by seasonal climatic cycles, Although Knoll and JamesŽ. 1987 and Volk Ž. 1989 lose soil K, Mg and Ca at a rate 3 to 4 times higher emphasize deciduousness as only an angiosperm than evergreen–coniferous ecosystems of cold and habit, RobinsonŽ. 1991, p. 860 points out that there wet tropical ecosystems with limited seasonality are a variety of deciduous gymnosperms in the Late Ž.Knoll and James, 1987; Volk, 1989 . Those arguing Paleozoic and Mesozoic, so effectively if litter fall for the significance of leaf litter have argued for a amounts influence weathering rates this can be seem major role in weathering related to the advent and as a post-Devonian phenomenon. One might argue, worldwide distribution of deciduous angiosperms in however, that the Northern Hemisphere Temperate the Cretaceous. However, TrappeŽ written communi- Zone is the only major ecosystem where essentially cation, 2000. points out that ADeciduous and ever- all the principal components are deciduous, as op- green systems certainly differ but after juvenile stages posed to the occasionally deciduous gymnosperms not in leaf fall. Witness the thick litter accumulation Ž.gingko, Taxodiaceae, etc. . She also argues that in many conifer stands, e.g. ponderosa pine or weathering rates are not necessarily higher in an- spruce-hemlock. Most evergreen leaves drop off in giosperm-dominated ecosystems than gymnosperm- 3–5 years, whether conifers or broad-leaved.B dominated ecosystems. Generalizations are difficult Trappe’s comments, based on the obvious, falsify to make, considering the complex of variables. How- this supposed distinction between coniferous and ever, as we have noted below, based on one study broad-leaved angiospermous floras. Thus, model sys- involving soil-weathering rates in acid soils in sepa- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 29 rate northern hardwood, pine and mixed hardwood- tal environment. In the continental, Berner only con- pine forests in a single geographic region, no simple sidered sequestered organic carbon of coal basin correlation between weathering rates and vegetation sediments. He did not consider sequestered organic types is possible without a consideration of the phys- carbon generated by marine plankton as contrasted ical and chemical characteristics of the soils them- with non-marine autotrophically produced organic selvesŽ. Cronan, 1985 . These unresolved problems carbon. Berner has assumed that preserved se- raise significant questions related to the evolution of questered organic carbon, i.e. standing crop, is a tracheophytes and the significance of the Earth’s direct function of productivity; there is no evidence surface through time covered by these so-called to support that assumption. Ahigh-weathering ecosystemsB Ž.Volk, 1989, p. 109 Can we assume, as Berner does, that the modern although they cannot be considered within a vacuum marine sediment sample, which is all that he consid- unrelated to other variables. ered, provides a reliable sample of Phanerozoic se- RobinsonŽ. 1990a,b makes the potentially impor- questered oceanic carbon? Can we assume that the tant point, based on a number of speculative argu- modern sample is necessarily AaverageB or Arepre- ments, that over Phanerozoic time the terrigenous sentativeB of the sequestered carbon content of conti- organic carbon pool has become increasingly suscep- nental shelf and ocean basin sediments? Without tible to biological decay due to a significant decline reliable estimates of this oceanic organic carbon in lignification in tracheophytes, and a major radia- through time, a major bias has been introduced into tion of lignin-degrading fungi. This would mean a Berner’s carbon sequestering estimates. potentially significant decrease in sequestration of Berner has implicitly assumed that direct and organic carbon over Phanerozoic time despite the indirect estimates of sequestered organic carbon in potential for increased productivity of organic Phanerozoic sedimentary rocks provide a measure of biomass. According to these arguments, rates of autotrophic productivity in the past. However, most carbon sequestration for the Paleozoic may have CO2 assimilated by autotrophic plants is destroyed been accomplished by a productivity ŽAby a bio- by heterotrophic organisms and returned to the atmo- sphereB. that was small by modern standards. This sphere. He did not consider sequestered organic car- argument is based on a number of speculations and bon present in regionally metamorphosed oceanic or evolutionary assumptionsŽ about the time and radia- continental rocks, or the kerogenŽ refractory organic tion of lignin-degrading fungi, for example, for which compounds insoluble in organic solvents. present in Robinson provides no information. but would pro- regionally unmetamorphosed rocks. vide an additional complication in assessing the sig- Berner did not estimate the amount of Phanero- nificance of the organic carbon sink and its impact zoic organic carbon sequestered as graphite in re- on atmospheric CO2 . gionally metamorphosed metasedimentary rocks. Un- One problem here is that we have no globally like carbonate crusts and soil carbon, graphite is a applicable, quantitative or even semi-quantitative data potentially very large volume of organic carbon. For through time on a global basis concerning the al- example, as noted below, the estimated organic car- legedly overwhelming importance of varied tracheo- bon present as graphite in the Precambrian Grenville phyte groups as contrasted with lower embryophytes Province alone is greater than that in all the Car- and non-embryophytic autotrophs. boniferous coals of the worldŽ Sozinov and Gor- bachev, 1987, p. 38.Ž. ! Volkert et al. 2000 provide 4.3.2.7. Sequestered organic carbon. In order to isotopic information from Mesoproterozoic graphite consider global organic carbon sequestration in New Jersey metasediments consistent with an throughout the Phanerozoic, it is necessary to con- organic origin. This question of the amount of or- sider both oceanicŽ. abyssal , marine epicontinental ganic carbon present in metamorphic rocks is further and continentalŽ. freshwater and terrestrial organic complicated by the discoveryŽ. Ahn et al., 1999 of carbon. In the marine environment, Berner consid- organic carbon interlayered with the clay mineral ered only sequestered organic carbon in modern illite, an organic carbon source that was previously marine deposits. He did not consider the epicontinen- unsuspected. 30 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

4.3.2.8. Natural hydrocarbons. Although minor, one ŽKazakhstan, South China, tracheophytes with cal- should add in what is known about the organic cretes, Boucot et al., 1982. . The assumption that the carbon locked up in natural gas, petroleum, bitumen presence of the seed producing capability permitted and oil shaleŽ Ulmishek and Klemme, 1990; Bois et tracheophytes to move into the uplands for the first al., 1982. , with consideration given to normalizing time is speculative. There is no positive evidence this data for losses through time. supporting their thesis that the uplands could not have been colonized by spore producing plants early 4.3.2.9. Deep, subsurface bacterial carbon. Nobody on as contrasted with seed producers. Spores are has been concerned with deep, subsurface, bacterial ideally designed to take advantage of brief intervals carbon except GoldŽ. 1992 . Gold made a number of of moist conditions. assumptions, many of which may be questioned, about bacterial magnitude, leading to the conclusion that subsurface bacterial activity is responsible for an 5. Organic carbon amount of organic carbon globally more than equivalent to the amount of surface organic carbonŽ a layer To assure reliability and deal with the multitude 1.6 m thick!. . This bacterial carbon might become a of variables involved in organic carbon estimates the source when the deep, subsurface rocks are exposed following are needed:Ž. 1 a synthesis of global strati- to erosion. graphic geology in sufficient detail to approximate the areal extent and volume of organic carbon con- 4.3.2.10. Were Early and Middle DeÕonian floras taining rocks for each time interval;Ž. 2 a determina- confined to the lowlands?. BernerŽ 1992, 1995, p. tion of the concentration of organic carbon in these 576. assumes that tracheophytes became Aeffective rocks;Ž. 3 some knowledge of the areal extent and as weathering agentsB during the Late Devonian– volume of rocks lost through erosion in each time Early Carboniferous, when they Afirst became exten- interval, thus enabling an estimate of the carbon sive in well-drained upland areas . . . B. Algeo et al. content lost or recycled during each time interval;Ž. 4 Ž.1995, 2001 repeat the statement that tracheophytes isopachs of the rock units of interest needed to only colonized upland regions of the world begin- calculate their volumes as well as their organic car- ning in the Late Devonian. bon content;Ž. 5 construction of the appropriate pale- There is no positive evidence for this position. ogeologic maps for each time interval; andŽ. 6 an Using this judgment, they then go on to further estimate of organic carbon biomass from isotopic conclude that soils heavily influenced by tracheo- information. These estimates are essential if reliabil- phytes only appear in the Late Devonian. They make ity is to be imparted to the assumptions underlying other assumptions that need to be queried. For exam- atmospheric CO2 models. Brasseur et al.Ž. 1999 ple, the presence of thicker upland soils in the Late provide a good account of the modern organic car- Devonian is based on limited data from eastern New bon budget, with the many difficulties involved in YorkŽ. Retallack, 1986 , which at that time was a varied estimates; estimates made for the past are humid regionŽ evidenced by abundant kaolin; see presumably no less complex and difficult to make Friend, 1966. , while earlier Devonian paleosols have with assurance. not been carefully studied from either arid or humid An estimate of organic carbon biomass through regions elsewhere in the world, nor have Late Devo- time can be made from the isotopic ratios of C12 and nian arid region paleosols been carefully studied. C13 if combined with estimates of the volumes of Their argument about deeper root penetration begin- inorganic carbon locked up in carbonate rocks. The ning only in the Late Devonian disagrees with the isotopic data assembled by Schidlowski and Aharon information provided by Elick et al.Ž. 1998 . Their Ž.Ž1992 see also Schidlowski, 1986, 1988 . indicate conclusion that Late Silurian–Early Devonian tra- that the relative amounts of organic to inorganic cheophytes were restricted to moist lowland habitats carbonŽ C12rC 13. have remained relatively constant is speculative, since there are arid region sites known since the Late Archaean owing to the relative stabil- from the earlier Devonian at higher elevations ity of the isotopic ratios. Since there is a relatively A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 31 large volume of Proterozoic and Early Paleozoic strata, now represented by widespread black, or- carbonate rocks, the volume of organic carbon must ganic-rich shales. Erbacher and ThurowŽ.Ž 1997 see have been adequate to maintain this relatively stable also Erbacher et al., 1996. review the presence of isotopic ratio. A large proportion of this organic these anoxic events in the earlier Cretaceous through carbon must have been of autotrophic origin, whether the Cenomanian. Erbacher et al.Ž. 2001 provide ex- photoautotrophic or chemoautotrophic, in the ab- tensive data concerning sequestered organic carbon sence of evidence for significant heterotrophic in the western Tethys and North Atlantic of Albian biomass. While much organic carbon may have had age. Hochuli et al.Ž. 1999 review early Aptian in the a marine phytoplanktonic origin, a non-marine origin Alpine Tethys, while Kuypers et al.Ž. 1999 discuss for some of it cannot be ruled out. Schwartzman and more Cenomanian–Turonian data, and Wignall VolkŽ. 1989 make a strong case for the presence of Ž.1994, Fig. 8.1 plots the known global distribution significant autotrophy since at least the later Ar- of the Cenomanian–Turonian anoxic beds distribu- chaean. tion. Arthur et al.Ž. 1985a,b note that a high propor- The question might be whether the relatively small tion of this sequestered carbon, in some intervals, perturbations in this C12rC13 ratio can be translated was land-derived. Hofmann et al.Ž. 2000 show that into atmospheric CO2 concentrations when com- in the western Atlantic during the late Albian, a large bined with appropriate assumptions about the seques- percentage of the organic carbon is of terrestrial tering of organic carbon in sedimentary rocks, inor- origin. However, Kuypers et al.Ž. 2001 provide geo- ganic carbon in preserved limestone and dolomite, chemical data suggesting that at least 80% of the and weathering rates. Albian deep sea organic matter is of marine rather than terrestrial origin. Hofmann et al.Ž. 1999 provide 5.1. Oceanic additional data concerning North Atlantic Ocean region organic-rich Cretaceous strata; Villamil et al. Berner has not considered the changing levels of Ž.1999 provide similar data for northern South Amer- sequestered, oceanic, organic carbon available in ica and Davis et al.Ž. 1999 for the Maracaibo Basin. Cretaceous and Tertiary oceanic sediment known The Late CretaceousŽ. earlier Turonian global anoxic from JOIDES cores, nor has he considered the possi- event recognized in oceanic and continental shelf bility that oceanic carbon sinks similar to those in environmentsŽ see Kerr, 1998, his Fig. 2, for an parts of the Cretaceous, discussed below, might have excellent diagram showing the event’s distribution. been present during the pre-Cretaceous. There are obviously had a major effect on organic carbon many intrinsic difficulties in attempting to estimate sequestration and probably had profound effects on changing levels of sequestered oceanic organic car- global atmospheric CO2 . Freeman and HayesŽ 1992, bon during the pre-Cretaceous. It is impossible to their Fig. 7. point out that the isotopic evidence from directly sample oceanic organic carbon lost to sub- sequestered organic, marine carbon at the Cenoma- duction, nor is there much capability of recognizing nian–Turonian boundary indicates significant atmo- non-subducted ocean basin remnants in the pre- spheric CO2 drawdown. Kuypers et al.Ž. 1999 pro- Cretaceous. The volume of both oceanic and conti- vide isotopic data also consistent with a global nental organic carbon lost during uplift and erosion drawdown in atmospheric CO2 correlated with or- during the pre-Cretaceous can be partly estimated by ganic carbon oceanic sequestration, and they claim a considering the paleogeology of the appropriate time 40–80% Cretaceous decrease in pCO2 related to a intervals. transition from C3 to C4 type land plants and marine Extrapolating only from modern oceanic-con- phytoplankton, both of which indicate global cooling tinental margin sediment analyses is false. Many Žthey also cite marine microfossil data consistent post-Cretaceous and some Cretaceous deep sea ma- with their isotopic data.Ž . Hart 1999, p. 254 . regards rine shales from JOIDES cores are far higher in the Cenomanian–Turonian extinction event, employ- organic carbon than the modern ocean basin samples. ing planktonic foraminiferal evidence, as global in For example, Arthur et al.Ž. 1985a,b note the high extent rather than being a series of localized events, burial rate of organic carbon in Cretaceous deep sea with the implication that the black shales are also 32 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 global in extent. JenkynsŽ. 1999 summarizes infor- been much richer in organic carbon than modern mation strongly supporting anoxic events in the early oceanic sediments. For example, Berry and Wilde Toarcian, early Aptian and Cenomanian–Turonian Ž.1978 proposed, based on AgraptoliticB black shales boundary, and provides an explanation suggesting of epicontinental platforms and basins, that it is that pulses of volcanogenic carbon dioxide may have plausible that Early Paleozoic deep oceans may have led to increased global temperatures, followed by been traps for organic carbon. Ulmishek and Klemme increased weathering that provided increased nutri- Ž.1990 cite a number of additional, mostly marine, ents that in turn led to increased oceanic productivity pre-Cretaceous, black, organic-rich shales deposited and subsequent oceanic organic carbon sequestration, within the epicontinental and non-marine environ- followed by global cooling. The information pro- ments. vided by the previous papers suggests varied, signifi- Berner and Canfield’s assumption about the ab- cant reductions in pCO2 following anoxic events sence of organic-rich strata in the Early Paleozoic is that is at variance with the partial pressureŽ between in disagreement with widespread organic-rich, black 3 and 12 times modern, pre-industrial values. sug- shales such as the agnostid–olenid facies of the gested by BernerŽ. 1992 . Cambrian, the graptolitic shale facies of the Ordovi- SliterŽ. 1989 describes Early and Late Aptian, cian and Silurian, and their well-known equivalents Pacific Ocean organic rich horizons, commenting in the Devonian. They ignore the kerogen present in that they contain more organic carbon than all the the even more widespread normal marine sediments petroleum and coal deposits of the world! Ryan and of the Early Paleozoic. By not considering this data, CitaŽ. 1977 detail something of both Early and Late their anomalously high, Early Paleozoic atmospheric

Cretaceous organic rich horizons in the Atlantic, CO2 is probably an artifact, particularly when com- Pacific and Indian Oceans. Follmi¨ et al.Ž. 1993 re- bined with what is known of the moderately high view intervals of high organic carbon deposition in global climatic gradientŽ. Fig. 1 from the Middle the Cretaceous and Cenozoic, recognizing the high Cambrian through the Eifelian. intervals of others. All of these studies, noted above, Despite the difficulty of recognizing possible emphasize that the amount of oceanic organic carbon oceanic TOC traps through geologic time, their in the Cretaceous and Cenozoic is not trivial. recognition is critical to any interpretation of global Schouten et al.Ž.Ž 2000 see also Weissert, 2000; sequestered organic carbon. In addition to recogniz- Hesselbo et al., 2000. provide evidence suggesting ing the carbon-rich intervals, it is equally important that there was a high organic carbon event during the to compile TOC data from intervals between or- Early JurassicŽ. early Toarcian employing a variety ganic-rich intervals. For the Cretaceous and Ceno- of geological and isotopic evidence. HallamŽ written zoic these data could be obtained from JOIDES communication, 2000. strongly supports this conclu- cores but is not easily available for the pre-Creta- sion. Their approach, gives promise for finding out a ceous. Finally, work in discriminating biochemically lot more about older, Paleozoic organic carbon-rich between continental carbon of humic type from ma- events. rine carbon of sapropelic type needs to be carried out All of the organic carbon-rich intervals noted on varied marine organic carbon samples through above were geologically brief and their effect on the timeŽ see the older paper by Breger and Brown,

CO2 levels will also have been brief, but measurable 1962, as an example of what might be accomplished, as demonstrated by Freeman and HayesŽ. 1992 . Al- as well as the Calvert and Pedersen, 1992 approach though many of these organic carbon-rich intervals employing organic biogeochemistry. . are seen as representing anoxic condition on the ocean floor, Pedersen and CalvertŽ. 1990 argue that 5.2. Epicontinental these carbon-rich strata may underlie areas of high productivity. The Berner models do not consider organic car- The presence of Cretaceous and Cenozoic deep- bon present in epicontinental environments. This lack sea carbon sinks raises the possibility that many is a major bias in the models. For example, Ronov unknown pre-Cretaceous oceanic shales may have Ž.1958 emphasized that the largest Phanerozoic or- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 33 ganic carbon peak for the Russian Platform is in the Ronov, 1958. . Overall, coals and peats are relatively Ordovician! Pagani et al.Ž. 1999 call attention to the minor organic carbon traps through time. many marine Miocene basins on the Pacific and Atlantic margins that are high in organic carbon, 5.4. Organic carbon in soil such as that in the California Monterey Formation. Hunt’sŽ. 1972 data supports the conclusion that There are many variables in the non-marine conti- epicontinental shales as well as oceanic shales, se- nental environment that may present the same kinds quester significant organic carbon as kerogenŽ for of limitations as failure to consider the possibility of additional data on kerogen distribution in epiconti- pre-Cretaceous oceanic carbon sinks. Gates et al. nental and oceanic sediments, see Welte, 1970; Hunt, Ž.1983 point out that great uncertainties exist con- 1979; Tissot and Welte, 1984; Whelan and Farring- cerning amounts of carbon stored in the world’s ton, 1992. . terrestrial ecosystems, with the greatest error being estimates of the amount of carbon stored in humus that comprises the recalcitrant, or stabilized, fraction 5.3. Continental of soil organic matter. The soil ecosystem, for example, is an extremely significant carbon sink that In the Ronov papersŽ Ronov, 1976, Table 1; should be considered in the overall equation. It has Ronov et al., 1980, Table 1. used by Berner to been noted that approximately three-quarters of biot- estimate sequestered organic carbon only Acoal basin ically controlled carbon stored on land is the organic sedimentsB were tabulated. The tables used in these matter of soilsŽ Schimel et al., 1994 suggest that soil papers did not include non-coal basin organic-rich carbon, a major component of the global carbon pre-Carboniferous or post-Devonian organic carbon- inventory, comprises approximately 2r3 of terres- bearing sediments, and were not intended as an trial carbon storage. constituting Aone of the largest inclusive statement about Phanerozoic sequestered near-surface stores of carbon on EarthB ŽSchlesinger, organic carbon. Because there are no pre-Late Devo- 1977 concludes that the global carbon content of nian coal basins, Berner assumed no significant pre- detrital carbon is about 1456=109 t. . Schlesinger Late Devonian sequestered organic carbon. Ž.1985 , while considering the formation of caliche To use coal basin sediments as the only volume of soils of the Mojave Desert, concludes that soil car- sequestered carbon does not provide an effective bonates contain about as much carbon as the atmo- estimate of sequestered continental organic carbon sphere. An extremely minor carbon sink in today’s through the Phanerozoic because of the variety of environment is that generated by cryptoendolithic other sources of organic carbon. microbial and lichen communities in hot and cold BernerŽ. 1991b, his Fig. 19 is impressed by the deserts and other environments, discussed below; great volume of coal-basin beds in the Permo- one cannot assume that prior to tracheophyte advent Carboniferous and Cretaceous, but provides outdated varied humid environments were devoid of these evidenceŽ. Berner and Canfield, 1989, Table 1 for lower autotrophs in abundance. Such specialized concluding that the Late Triassic, Jurassic, and much ecosystems may have been more widespread during of the Cenozoic are not also just about as rich in some intervals of the Phanerozoic and should be no coal-basin sediments as the Cretaceous. It is a mis- more ignored than some of the comparatively minor take to rely on the older coal-basin sediment sum- inorganic carbon sinks discussed below. maries which do not take account of more recent Soil organic carbon and carbonate crusts are data regarding the huge coal-basin sediment com- doubtless extremely labile compared with the mass plexes of AsiaŽ. former Soviet Asia plus China for of marine carbonates, even though soil organic and the Late Triassic, Jurassic and Cretaceous, as well as inorganic carbon reserves Aspan the realm between for the Cenozoic. Moreover, most sequestered or- the geological and biospheric carbon poolsB ŽSchle- ganic carbon is not preserved as coal, but as fine- singer, 1986, p. 196.Ž. . Schimel et al. 1994 indicate grained kerogen in marine and non-marine shales that rates of total soil carbon turnover times vary unassociated with coal basinsŽ Hunt, 1972; see also widely on a global scale, being significantly shorter 34 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 in tropical than in high latitude soils with variations It is not possible to directly measure total, global within specific detrital and soil organic matter frac- marine and non-marine sequestered organic carbon tions, being even more dramatic. Raich and for the Phanerozoic, nor is it possible to make esti- SchlesingerŽ. 1992, Table 3, p. 86 present compara- mates of the pre-Cretaceous oceanic carbon. Prior to ble results indicating turnover rates of ca. 500 years the Cretaceous when oceanic sequestered organic in tundra and swamp-marsh sediments to 10 years in carbon may be estimated for the first time, the tropical grasslands, with a mean residence time for oceanic moiety is unknown, and may have been all vegetation types of 32 years. Residence time important during many intervals. Even for the Creta- tends to increase with depth, however, and O’Brien ceous and earlier Cenozoic, when oceanic data first and StoutŽ. 1978 conclude that soil carbon is a became available, the problem of eroded organic mixture of AoldB and Amodern,B with carbon that carbon on the continents, plus subducted oceanic may be 1000s of years old being distributed almost organic carbon, makes estimates very difficult, and uniformly with depth in the soil profile. Their esti- increasingly unreliable back in time. However, it is mates suggest that at least 16% of the organic matter possible to provide estimates for preserved se- in the lower soil profile of a New Zealand pasture questered, epicontinental and continental organic car- was at least 5700 years old and that part of the old bon, in Phanerozoic sedimentary rocks, by working carbon could be at least 15,000 years old. Paul and out the paleogeology of the continents, and estimat- VoroneyŽ. 1980 likewise note that stabilized or ing the preserved volumes of organic carbon contain- recalcitrant organic materials may persist for long ing rocks plus estimating the amounts of organic periods of time in terrestrial soils—they note that carbon removed during erosive intervals. 40–60% of the non-acid hydrolyzable carbon of 5.5. Inorganic soil carbon terrestrial soils as compared with the readily decom- posable fraction, has persisted for at least 1000 In arid and semi-arid regions, soils store inorganic years. Interestingly, VestalŽ. 1993, Table 2 reports carbon as calcium carbonate in amounts exceeding that Antarctic cryptoendolithic microbial communi- the organic carbon storage of desert soilsŽ Schle- ties with the lowest primary carbon production of singer, 1986, 1995.Ž. . Schlesinger 1986 estimates, any present global ecosystem also have the longest based on carbon stored as carbonate in Arizona carbon turnover timeŽ. 576 to 23,520 years of any desert soils, that the amount of carbon stored in present global ecosystem. Although there is little caliche in the worlds deserts at the present time is possibility of obtaining detailed information from the 800=1015 g. An increasing number of soils are geological past, for any of these potential carbon being identified throughout the Phanerozoic whose sinks at the atmospheric–lithosphere boundary, esti- stored carbon, both as calcrete and as organic car- mates are essential. We do not currently possess data bon, should be considered. Additional to the more adequate to evaluate the global significance of this or less horizontal caliche stores in soils, Mc- carbon sink. Connaughey and WhelanŽ. 1997 point out extended In conclusion, sequestered organic carbon pre- vertical root calcification in many soil types, where served in the continental and epicontinental environ- calcification follows root penetration, an additional ments, as contrasted with the oceanic realm, is pre- source of stored carbon potentially significant in sent:Ž. a as coal basin sediments; Ž. b as kerogen in some areas. both epicontinental marine and non-marine continen- Ferris et al.Ž. 1994 track potential for carbonate tal deposits;Ž. c as coaly material and kerogen in precipitation among cyanobacteria in a variety of epicontinental, marine deposits; andŽ. d as coaly fresh to saline–alkaline lakes in Canada and in min- material and kerogen in non-marine, non-coal basin eralized crusts on weathered basalt in Iceland. They deposits. suggest that the weathering of silicate minerals in Epicontinental, continental and soil organic car- bedrock is biogeochemically coupled to phototrophic bon sources are not necessarily the major sources microbial carbonate mineral deposition. They con- compared with oceanic sources. Berner has only clude that this coupled process may provide a signif- considered the continental coal basin organic carbon. icant sink for atmospheric CO2 in the terrestrial A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 35 environment. Thus, as they note, estimated non-mac- carbon. Estimates of the volumes of carbonate rocks rophytic primary production in terrestrial wetland removed by erosion during any time interval are also and lakes amounts to ;1.4 petagrams CryearŽ cit- essential. ing Schlesinger, 1991. Athere is potential for photo- There are a number of problems with the studies synthetic microorganisms to precipitate upward of a Ž.Ronov and others used by Berner to estimate se- billion tonnes of carbonate carbon on a yearly basisB questered inorganic carbon, and additionally, there essentially equivalent to the present imbalance in the are problems with omitted data not available when global atmospheric CO2 budget. Additionally, they the Ronov and others studies were made: note that alkalinity changes and carbonate precipita- Ž.1 Overall, they provide conclusions about areas tion can be initiated by microbes other than photo- occupied by carbonate rocks without providing the trophic surface-dwelling cyanobacteria in a coupled data points on which these conclusions are based. reaction involving deep subsurface weathering of Without the data points the reliability of the informa- silicate minerals and secondary carbonate precipitation cannot be evaluated. For example, improve- tion. Anaerobic sulfate reduction, related to the activ- ments in stratigraphic correlation during the past ities of anaerobic sulfate-reducing bacteria, may re- thirty to 40 years might well shift areas with signifi- lease considerable amounts of sulfide, a strong base cant carbonate from one time interval to another. capable to taking up protons from bicarbonate, and Ž.2 They provide no information regarding the promoting carbonate precipitation. Likewise, micro- original, pre-erosional volumes of carbonate rocks as bial activities that increase alkalinity such as denitri- opposed to their present areal extent per time inter- fication, manganese oxide reduction and iron oxide val. reduction can also contribute to carbonate mineral Ž.3 They do not estimate volumes of potentially precipitation Ferris et al.Ž. 1994 . Obviously, micro- subducted rocks, since most of the studies were bial carbonate precipitation provides for potentially published prior to the advent and acceptance of plate important sinks that will vary geographically and tectonics. through geologic time; their potential cannot be ig- Ž.4 They seriously underestimate the pre-erosional nored. extent of carbonate rock units because they have DornŽ. 1998 has noted the variety of carbonate treated the zero isopach as the shoreline and basin crusts, in addition to subsurface pedogenic caliches margin. The zero isopach is not the basin margin or or calcretes, readily deposited in freshwater, littoral shoreline in many instances. marine, soil and subaerial contexts, pointing out that Ž.5 They are obsolete relative to the amount of the mass of CO2 stored in them has never been information recently accumulated about widespread calculated nor has this CO2 source ever been incor- carbonate rocks in many regions. The Ronov et al. porated into carbon models. Nor are any estimates data are based on relatively secondary sources for available about how rapidly such crusts are either significant parts of the world. being deposited or dissolved, although they may only Ž.6 They conclude that AgeosynclinalB rocks con- represent a temporary storage of CO2 . tain a far greater volume of carbonate rocks than Other possible calcium carbonate sinks are a vari- platforms, but provide no basis for this conclusion. ety of secondary carbonate deposits variously identi- Ž.7 They do not deal with the extensive, com- fied as tufas, travertines, sinter, and calc-sinter found monly poorly dated, regionally metamorphosed car- in springs, waterfalls, lakes and alluvial fans in bonate rocks present in many mountain belts of the limestone regionsŽ. Viles and Goudie, 1990 . As Viles world, nor have they attempted to estimate the vary- and GoudieŽ. 1990 note, these vary hugely in size ing amounts of tectonic thickening and thinning that but are potentially important in regions with tropical, commonly affects the relatively plastic marbles. semi-arid climate. Ž.8 They do not estimate carbonate cement present in many non-carbonate clastic rocks. 5.6. Non-soil carbonate rock Ž.9 They could not include the volumes of cal- Knowledge of the volumes of preserved carbonate cium carbonate oceanic ooze for each time interval rocks are critical for obtaining estimates of inorganic in the post-Triassic, not available when they made 36 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 their compilations. Volumes of carbonate oceanic continental, chiefly basement complex rocks, richer ooze could now be estimated for the Cretaceous and in87 SrŽ see Edmond, 1992; Blum, 1997; McCauley Cenozoic from JOIDES data, but we know of no and DePaolo, 1997, for review of the principles and work in this area. It would be interesting, for exam- problems underlying the use of strontium ratios. . The ple, to see what effects the carbonate oceanic oozes assumption is that marine limestone and dolomite, of the Mid-Mesozoic had on carbonate sequestration. the major sink for strontium through time, will pro- Ž.10 Another potential source of sequestered inor- vide strontium isotope values indicative of relative ganic carbon not considered by Berner is that pro- amounts of contemporary uplift and silicate mineral duced during the alteration of oceanic basalts. Brady weathering, since radiogenic 87Sr is mostly derived and GislasonŽ. 1997 state AAlteration of surficial from silicate mineral weathering of continental rocks. marine basalts at low temperaturesŽ. 408C is a poten- Berner did not consider the sometimes abundant tially important sink for atmospheric CO2 over geo- carbonate rocks present in many basement complex logical time.B They emphasize that knowledge of regions that may contain radiogenic strontium de- Phanerozoic rates of seafloor spreading and of the rived from older sources, which largely invalidate extent of oceanic basalts is necessary for evaluating the results. this variable. Brady and Gislason note that the net Quade et al.Ž. 1997 point out, employing data influx of Ca from low temperature weathering of from the Himalayan region, that trying to estimate oceanic crust is imprecisely known but that it could continental weathering from strontium isotope data be Asomewhat similarB to recent estimates of the from carbonate rocks is unreliable; that no direct contribution of ACa derived from the continental relationship exists between amount of weathering weathering of Ca-silicatesB Ž.p. 965 . Uncertainty and strontium isotope data. Blum et al.Ž. 1998 , also about this information is one of the major obstacles in the Himalayan drainage basin, found that most of to modeling the long-term C cycle and its role in the strontium was derived from dissolved calcite of climatic evolution. Because the atmospheric CO2 metamorphic origin rather than from silicate mineral reservoir is small, they suggest that minimal imbal- weathering. This means that there may be little ances between sources and sinks, occasioned by the correlation between level of silicate weathering and lack of information about CO2 consumption during contemporary atmospheric CO2 level, because the seafloor weathering, could cause rapid fluctuations in strontium in the metamorphic calcite may well be of atmospheric CO2 levels and global mean tempera- widely different ages. BlumŽ. 1997, p. 260 , Edmond tures. As they emphasize, quantification of Aweather- and HuhŽ.Ž. 1997, p. 349 , Bickel 1998 and Broecker ing fluid fluxes, temperatures, and spreading ratesB and SanyalŽ. 1998 all concur in being skeptical about is critical to understanding the history of CO2 and strontium isotopes from marine carbonate rocks as a temperature over time. proxy for continental silicate weathering. The points discussed above make it difficult to compile sequestered inorganic carbon globally, but without these data Phanerozoic estimates of atmo- 7. Berner’s use of modern river water data bases spheric CO2 are potentially misleading and seriously for assessing weathering in error. Berner bases his estimates for weathering on a number of modern runoff studies mostly in the tem- 6. Use of strontium isotopes as a measure of perate zone and angiospermous ecosystems rather weathering than direct data, such as estimated volumes of the key rock types, relevant to Phanerozoic weathering BernerŽ. 1994 tried to extrapolate weathering in- phenomena overall. The assumption that modern formation from 87Srr86 Sr in marine carbonate rocks. continental runoff can be extrapolated back through 87SrrSr86 preserved in Phanerozoic marine carbon- the Cambrian is speculative. He discusses several A B ates was used as a proxy for atmospheric CO2 . This field studiesŽ. discussed elsewhere . None provide ratio is assumed to be an index for the weathering of direct measurements of organic soil acid production A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 37 but depend on information derived from ionic con- ing as measures of bedrock weathering. Involved y centrations of dissolved bicarbonateŽ HCO3 . in con- here is the common absence of analyses and mea- tinental runoff, which is subject to a number of surements of suspended load materials, as well as of limitations. bedload material. Also missing is an assessment of Berner uses modern river water runoff data to rainwater input, since Allegre emphasized that rain- arrive at an estimate of weathering by tracheophytes water is commonly AdirtyB with ionic input from and what he interprets as the impact of abiotic ocean water and dust among other things. All of this weathering. The quantity and rate of this ionic flux, suggests that river water analyses, no matter how most of which enters the ocean, are critical to infer- carefully done, from both large and small drainages, ring the quantity of carbonate rocks deposited. To may well be misleading for providing bedrock provide cation flux data it is necessary to estimate weathering estimates. Also involved here for previ- the amount of limestone and dolomite plus Ca and ous time intervals is the problem of estimating the Mg silicate minerals subject to weathering in the amount of carbonate bedrock, evaporite bedrock, drainage basin, and then translate the water content acidic and basic bedrock, and so forth, that has been into precipitated carbonates downstream. There are exposed to surface weathering globally during each extensive, global studies that deal with the solute time interval; such work has not been carried out. chemistry of larger river basins for the present There are serious problems with the bicarbonate Ž.Meybeck, 1979 , as well as isolated studies for method for estimating the amount and rate of weath- relatively small drainagesŽ Drever, 1994; Velbel, ering. The estimation of weathered Ca and Mg ions 1993. . Smaller scale watersheds and catchments, Ž.from both silicates and carbonates requires reliable with better defined hydrology and solute budgets, are knowledge of the proportions of the appropriate min- less encumbered by regional meteorologic, geologic, erals in rocks subject to weathering for each time biologic and anthropogenic complexities and might interval in each region and the climate of each prove more useful in the interpretation of weathering region. This requires careful paleogeological analy- processesŽ. White and Blum, 1995 . sis, time interval by time interval, necessary for an MeybeckŽ. 1979 classified all the world’s rivers estimate of the areas of Ca and Mg silicate minerals over a certain size on the basis of water runoff, and carbonate rocks available for weathering and the average temperature of watershed with allowance for rates of uplift which affect the volume of weathered relief, and then determined average concentrations rocks. The importance of rates of uplift, in addition and rate of transport of a number of dissolved con- to relief, elevation and glaciation for understanding stituents as a function of a Amorphoclimatic classi- the amounts of Ca and Mg ion available for seques- B fication. Meybeck related chemical fluxes and tering atmospheric CO2 cannot be ignoredŽ Francois denudation rates to both temperature and runoff. et al., 1993. . To obtain volume information requires According to MeybeckŽ. 1979, table 5, pp. 222–223 , knowledge of areas of uplift and uplift rates capable riverine Ca and Mg are not indicative of silicate of exposing fresh minerals to weathering, plus weathering because they are derived overwhelmingly knowledge of climatic regimes in areas of weather- from dissolution of carbonate minerals. Berner ex- ing. It is possible to make potentially valid runoff trapolated this modern Mg and Ca data into the past, estimates of weathered Ca and Mg mineralsŽ silicates but did not alternatively plot the paleogeological and carbonates. through geologic time, but this redistribution of the rocks subject to weathering during quires thorough familiarity with Phanerozoic paleo- any time interval. If this had been done, it would be geography, paleoclimatology and paleogeology. Such possible to draw conclusions based on their areal information is poorly known for the pre-Permian and extent and volume plus prevailing climates to give especially for the pre-Devonian interval. Critical in- some idea of amounts of the cations made available. formation on runoff requires reliable information Gaillardet et al.Ž.Ž 2000 Allegre, 2000, oral pre- about the global distribution of climates and paleo- sentation of the Gaillardet, 2000, paper at the 31st geography. This synthesis remains undone. Berner’s International Geological Congress. make clear that theoretical atmospheric models are based on infor- river water analyses by themselves can be mislead- mation about the role of weathering in modern tem- 38 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 perate soils, briefly summarized below, rather than terrtime interval. of base cations such as Ca and Mg on a reliable knowledge of past geological data. in surface waters leaving a catchment regionŽ Stal- To estimate Phanerozoic ionic runoff, Berner lard, 1985; Meybeck, 1979, 1980. . These ions are Ž.1991b, 1994 has used ionic runoff data Ž Tardy et presumed to have been released by the weathering of al., 1989; Otto-Bliesner, 1993, unpublished until primary silicates and thought to reflect local weather- 1995. , which are seriously deficient in one or more ing reactions. The major flux of Ca and Mg in river of the regards discussed above. The runoff data water solution combines that derived from silicate provided by Tardy et al. assumes that modern runoff mineral weathering with that representing carbonate data for present day latitudes can be extrapolated mineral solution. back in time, while using outmoded paleogeogra- Local information is then extrapolated to the phies; runoff data provided by Otto-BliesnerŽ 1993, global scale. Thus, DreverŽ. 1994, p. 2325 writes: unpublished until 1995. was based on paleogeogra- AThe global weathering rate is simply the weighted phies and climatic distribution patterns provided by average of the rates in all individual catchments.B C. Scotese and J. GolonkaŽ Otto-Bliesner, 1995, p. Drever and others assume that Ain the context of 11,539.Ž that are now superseded Boucot et al., weathering of silicate rocksB this flux of base cations A submitted. . corresponds closely to the flux of CO2 from the It is clear that the available continental runoff data atmosphere into bicarbonate in the oceans.B Modern and its ionic composition are inadequate for mod- flux studies are used in the Berner model. Such elling purposes. In any case, how important is ionic modern flux from limited areas is used as a proxy for runoff? France-Lanord and DerryŽ. 1997 , employing estimating global weathering through the Phanero- cores from the Neogene Himalayan–Bengal Fan, zoic, with the implicit assumption that weathering indicate that buried organic carbon is 2–3= more rates, exposure of silicate rocks rich in Ca and Mg, A B effective in CO2 consumption than silicate weath- topography, and other variables, of the past were ering. about the same as those of the present. There is no BradyŽ. 1991 and Velbel Ž. 1993 suggest that basis for this assumption. The more reliable alterna-

Berner’s older estimates of atmospheric CO2 levels tive is a compilation of the necessary geological throughout the Phanerozoic were overly high owing data. to misunderstandings about silicate mineral dissolu- Since such studies rarely measure soil CO2 or soil tion temperature dependencies; these have since been organic acids, it is, as Gwiazda and BroeckerŽ. 1994 updated. More realistic silicate mineral rate depen- state, Adifficult to relate the weathering production to dencies cause Berner’s anomalously high atmo- any forcing factor.B Having such data available might spheric CO2 levels vis-a-vis the climatic data pre- help in an overall evaluation of the weathering prob- sented below to drop significantly. Nugent et al. lem. Additionally it is seldom possible in most basin Ž.1998 provide more information about the contrast studies to determine that the alkalinity flux out of the between laboratory and field data on feldspar weath- basin, is a Asteady state valueB and not a transient ering. White and BrantleyŽ. 1995 provide entry to factor resulting from changes in land use and input the literature on laboratory work on silicate mineral of anthropogenc acidsŽ. acid rain or wet, dry season- solubilities. Hochella and BanfieldŽ. 1995 review the ality. Thus, Gwiazda and Broecker note that the disparities between field and laboratory mineral adjustment time of soils to natural variations in weathered rates; this area is clearly important to any environmental conditions is on a century to millen- understanding of mineral weathering under natural nial time scale, casting doubt on surface runoff or conditions. catchment studies for providing reliable information for extrapolating to generalizations of the type cited 7.1. Is flux a credible proxy for weathering rate? for demonstrating the weathering effects of modern basin studies to climates of the remote geological Geochemists and others have addressed Aweather- past. Thus, as they sayŽ. p. 153 , the long time ingB primarily by charting the fluxŽ the concentration required for a soil to adjust to new conditions has and volume of specific ions dissolved in river wa- important implications regarding the selection of a A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 39 field strategy to study the influence of different concentrations in surface runoff are Astrongly corre- factors on silicate weathering. StallardŽ 1985, pp. lated with proportional ETwx evapotranspiration loss 298, 312, 1995. notes that while dissolved river load and evaporative concentration makes stream solute chemistry is thought to best reflect weathering pro- concentrations an inappropriate surrogate for chemi- cesses, that weathering products undergo various cal weatheringB Ž.p. 1729 . Additional quotations reactions as they move through fluvial systems or are support their conclusion: Asolute concentrations are stored on flood plains that affect river chemistry in not accurate indicators of relative weathering intensi- numerous ways not predictable from looking at local ties.BAIncreasing trends in solute concentrations can erosion. Thus, it may not always be possible to relate result both from increases in weathering rates with riverine transported weathering products to geology, temperature, as is assumed in the models, as well as landforms and soils. from concentration effects produced by ET, which No field studies have directly measured acids or also increases with temperatureB Ž.p. 1739 . This is a pCO2 in soils but have depended more ambiguously very important conclusion in that they correlate cli- on measurements of ionic runoffŽ Gwiazda and matic gradient with solute concentration, i.e. the raw Broecker, 1994. attempting to assess weathering in solute concentration data should not be employed terms of bicarbonate, silicate and in some cases until it has been normalized for climate, including strontium isotope measurements. Berner’s assess- precipitation. They report that chemical fluxes on the ment of silicate weathering rates has cited a number other hand are unaffected by evapotranspiration and of such studies. Ionic runoff studies, as discussed, SiO2 and Na weathering fluxes show systematic are subject to a number of limitations and constraints increase with precipitation, runoff and temperature, that have not been carefully estimated or considered while K, Ca and Mg fluxes exhibit no climatic and that can seriously skew potential evidence in correlation. support of weathering effects. White and BlumŽ. 1995 conclude that Berner’s There have been no prior attempts to estimate 1994 model, in which trends in solute concentrations volumesrtime interval of weathered Ca- and Mg-rich were assumed to result from the coupled effects of silicate minerals, or of dissolved carbonate dilution and solute residence time in the soil, could rocksrtime interval from paleogeological data. As a significantly underestimate watershed weathering substitute, a proxy, for this deficiency there have rates in wet tropical areas. As they suggest from their been attempts to employ the flux of dissolved cations studiesŽ based on 68 watersheds underlain by grani- and bicarbonate in modern river waters, which is toid rock types with global distribution. , moisture then extrapolated into the past. Such river water has a more significant effect on weathering rates studies are clearly subject to many constraints that than assumed in the Berner model, since weathering have not been carefully estimated or considered. rates are accelerated in wet, warm climates meaning Particularly important are the large changes in sea- that a larger proportion of global weathering will sonal concentrations of dissolved ions; unless these occur in tropical regions. However, theyŽ 1995, p. have been carefully evaluated, there is real danger 1739. point out that while solute discharge concen- for misinterpretation. trations can be correlated with an increase in weath- White and BlumŽ. 1995, p. 1745 conclude that ering rates with temperature, evapotranspiration, stream solute concentration studies, commonly the which also increases with temperature, will have the focus of studies attempting to determine weathering same effect on solute concentrations of major cations rates by correlating variations in solute concentra- in watershed streams. Thus, they argue since the tions and fluxes with temperature, precipitation, proportion of precipitation lost by evapotranspiration runoff and evapotranspiration, are not an effective increases with lower rainfall, that the effects of method of comparing chemical weathering rates be- precipitation and temperature must be considered tween watersheds with widely different climates be- simultaneously rather than as independent variables cause climatic conditions play a significant role in when modelling feedback between climate and determining solute chemistry and weathering rates in chemical weathering rates. White and BlumŽ 1995, watersheds. Their study demonstrates that solute p. 1745. conclude that solute concentration studies, 40 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 commonly the focus of attempts to determine weath- silicate mineral weathering. Walker et al.Ž. 1981 ering rates, are not an effective method of comparing building on Meybeck’sŽ. 1979 work, note that Mey- weathering rates between watersheds with widely beck showed that silicate weathering increased different climates. markedly with increase in either runoff or tempera- It should also be notedŽ cf. Brady, 1991; Brady tureŽ. p. 9779 . The significance of this is that Mey- and Zachara, 1996, p. 329 et seq.; Bowen, 1980, beck concluded that in the tropical regions, where who also reviews many of the other rhizosphere the bulk of the weathering occurs, it is the interaction problems. that most models that attempt to assess the of runoff Ž.AhumidityB and temperature that is im- weathering variables and their relative importance in portant, since the two factors interact with each weathering rely on laboratoryŽ see White and Brant- other. But Walker et al.Ž. 1981 say that the apparent ley, 1995. rather than field studies. Attempts to temperature and runoff dependence suggested by isolate and compare direct effects of single factors in Meybeck is AcontaminatedB because silicate weath- mineral weathering under laboratory circumstances ering may be further enhanced by partial pressure of are seldom satisfactory because they tend to approxi- CO2 in pore space of soils that is increased by mate rather than duplicate effects of mineral weath- respiration and decay of organic matter and further ering in field circumstances for a variety of reasons promoted by organic acids. Walker et al.Ž. 1981 Ž.cf. Brady and Zachara, 1996 . For example, as deduce from work of others that cation concentra- Brady and ZacharaŽ. 1996 note, field weathering tions in river water are Aapproximately independent rates of primary silicates are found to be 100–1000 of runoff but depend on temperature and possibly B times less than those measured in the laboratory. also on CO2 partial pressure. Few field studies of the type provided by April and Velbel’sŽ. 1993 field study investigated the effect NewtonŽ. 1992 have examined the direct mechanical of elevation-dependent factors such as temperature, and chemical effects on mineral grains within the soil thickness and vegetation on silicate weathering soil. As Brady and ZacharaŽ. 1996 and Gwiazda and rates in two catchment areas from the Appalachian BroeckerŽ. 1994, p. 142 report, most field studies Blue Ridge that differed only in temperature as a attempt to estimate the rate of silicate weathering, by function of elevation, being closely matched in other measuring the alkalinity fluxŽ. bicarbonate content or variables including vegetation type and weathering Si concentrations of surface water runoff from spe- profiles. Because of a decline in silicate mineral cific continental basins and from this information weathering rates with increasing elevation, Velbel attempt to evaluate the reactions that determine the Ž.1993, p. 1060 concludes that there is a high corre- water composition. lation between temperature and solution rate of min- Berner’s oversimplification of a complex problem erals, or as he wrote, when variables in the two assumesŽ. 1 an abundance of fresh, reactive or catchments are minimized, this leaves Atemperature weatherable minerals in soils andŽ. 2 that the rate of alone as the dominant factor controlling differences mineral weathering is similar in climatically different in weathering rates.B In turn, this means that the regions; and ignoresŽ. 3 the subsequent reaction or consumption of CO2 by weathering and its removal fixation of base cations that may cause precipitation from the atmosphere is closely dependent not only of new or secondary mineral species, such as clays in on the global mean temperature but local variabilities soilsŽ April and Newton, 1992, pp. 399, and subse- in temperature like those evident in elevational tran- quent; Drever, 1994.Ž. and 4 subsequent decrease in sects. weathering rates with increase in thickness and age Gwiazda and BroeckerŽ. 1994 also explore the of soils.Ž. 5 Berner’s theoretical atmospheric models question of temperature in controlling rate of mineral are built on information about the role of weathering dissolution which they attempt to evaluate via a in modern temperate soils; this is a gross oversimpli- numerical model based on processes measured and fication of the soil weathering problem. observed in the field Ain an average soil of warm Walker et al.Ž 1981; Brady and Carroll, 1994; as temperate climate.B Gwiazda and Broecker examine noted by Gwiazda and Broecker, 1994. also at- relationship between temperature, soil pCO2 Ž partial tributed a significant role to surface temperature in pressure of soil CO2 . and tracheophyte-fungal gener- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 41 ated organic acidity to alkalinity production and the 1969, p. 392. , e.g. to account for weathering effects mechanisms and response time of the soil to changes that were but little different from those in the tra- in the forcing factor of silicate weathering. After cheophytic world. When they examined the rate of attempting to assess the relationship between these removal of dissolved salts and bicarbonate in streams several factors potentially implicated in enhanced as a measure of the acceleration of chemical weath- silicate weathering rate, they concludeŽ. p. 153 that ering by tracheophytes, they found that there was Asoil temperature is by far the most important factor relatively little difference between areas in central that regulates mineral dissolution . . .Ž see Walker et Iceland without visible megascopic plant cover as al., 1981. , whereas the influences of soil pCO2 and compared with those inhabited by vascular plants. organic acid concentration are minor.B Thus, the bicarbonate concentration can Abuild up to However, White and BlumŽ. 1995, p. 1739 argue one-half the maximum observed value in the absence that since the proportion of precipitation lost by of higher land plants. This in turn indicates that the evapotranspiration increases with lower rainfall that influence of higher land plants on rates of chemical the effects of precipitation and temperature must be weathering in this area is not overwhelming.B They considered simultaneously rather than as independent conclude that rates of chemical weathering Ain the variables when modelling feedback between climate absence of higher land plants in pre-Silurian times and chemical weathering rates. White and Blum were not very different from present day ratesB Ž.1995, p. 1730 indicate that the 68 compiled water- Ž.Ž.Cawley et al., 1969, p. 392 . Cawley et al. 1969 shed studies, for which they examined the effects of were impressed by the limited increase in weathering climate on chemical weathering were heavily biased associated with vascular plants pointing out the Arel- toward alpine and temperate forested vegetation in atively small differenceB in the rate of chemical Europe and North America, and indicate no clear weathering in areas with plant cover as compared trends that Asuggests that vegetation systematically with those that are apparently barren as assessed by influences chemical weathering ratesB, although be- bicarbonate concentrations. cause vegetation type varies with precipitation and temperature, vegetational effects may be incorporated into climatic trends. 8. Carbon isotopes from calcrete as a measure of

Cawley et al.Ž. 1969 tested the hypothesis about atmospheric CO2 the weathering significance of carbonic acid in central Iceland. They argued, based on analysis of field A potentially independent AtestB of atmospheric 12r 13 studies, that the rate of chemical weathering pro- CO2 content might be the C C ratio of soil vided an analogue to pre-tracheophytic weathering. carbonates and associated organic carbon. The recent While it has commonly been assumed that the car- studies by Driese and MoraŽ. 1993 and Mora et al. bonic acid content of soils would be greatly en- Ž.1996 attempted to estimate the CO2 of the Late hanced by microbial respiration associated with Paleozoic atmosphere using the C12rC13 ratio of increased-tracheophyte based organic carbonŽ the ancient pedogenic calcretes and coexisting organic partial pressure of CO2 in soils below tracheophytes carbon. Ekart et al.Ž. 1999 have carefully reviewed is normally 10–100 times greater than in the present the many physical and biological factors involved atmosphere. , that this biogenically generated CO2 with trying to use calcretes for ancient atmospheric could be offset by a relatively small increase in CO2 estimation, and discussed the many factors that atmospheric CO2 to maintain very similar weather- may influence the ultimate isotopic measurements; ing rates throughout geologic time. Thus, they sug- this is clearly a complex system that needs to be gested that increased atmospheric CO2 pressures no studied very carefully, sample by sample. Ekart et al. greater than 5 times the present values, were all that Ž.Ž.1999, their Fig. 12 our Fig. 2 provide a Devonian A would have been needed to produce present-day to present atmospheric CO2 curve. mean rates of chemical weathering under comparable This isotopic methodology is not without difficul- climatic and geographic conditions before the devel- ties and requires that a variety of rather specific opment of higher land plants . . . B ŽCawley et al., conditions be fulfilledŽ Driese and Mora, 1993; 42 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Fig. 2. Ekart et al.’sŽ. 1999 Fig. 12 Ž modified . showing their calcrete generated atmospheric CO2 estimate Ž gray band . for the Carboniferous to present. Berner’sŽ. 1994 model is shown with a solid line bounded by dot-dash lines showing the levels of uncertainty. The dashed line is from Worsley et al.Ž. 1994 . The hollow triangles are Yapp and Poth’s Ž. 1996 soil goethite measurements. The open diamonds are Freeman and Hayes’Ž. 1992 estimates.

Driese et al., 1993; Mora et al., 1991, 1996; Ekart et provided carbon isotopic data interpreted as indicat- = al., 1999.Ž. . Driese and Mora 1993, p. 26 , Driese et ing a 10 above modern atmospheric CO2 level, al.Ž. 1993, pp. 63–65 , Rossinsky and Swart Ž. 1993 , thus, significantly overestimating atmospheric CO2 and Cerling and QuadeŽ. 1993 note the numerous Ž.Mora et al., 1996 . They explain this anomaly and parameters involved in calcrete formation, all of correct for it by assuming that the minute amounts of which influence carbonate chemistry and hence the organic carbon extracted from their nodular calcrete stable isotope composition of pedogenic carbonate: for carbon isotope study might have represented latitude, elevation, temperature, vegetation type, iso- plants that grew in a wet soil. But the presence of topic composition of meteoric and ground waters, calcretes indicates a semi-arid regional environment, evaporation, variable soil productivity, depth of for- which means that the plants could only have grown mation, pedogenic processes such as soil cracking in a seasonally wet environment. Mora et al.Ž. 1996 and drying, and diagenetic alteration of original iso- have no information about which autotrophsŽ for topic composition. Driese et al.Ž. 1993, p. 64 caution example, bacteria, lichens, bryophytes, extinct higher that the use of carbon isotope composition of pedo- tracheophyte taxa. are represented by the minute genic carbonates without regard to their environment amounts of carbon they analysed, or whether or not of precipitation commonly results in atmospheric the physiologyŽ. ies of these unknown autotrophs re-

CO2 levels that are too high because of atmospheric quired a wet soil. mixing. For example, it can be demonstrated that Unless physical and chemical parameters can be rhizolith calcrete formed at depth provides the most determined to be uniform during paleosol formation, reliable data, compared with near surface nodular it is not surprising that calcrete isotopic data may be calcrete formed within the zone of Asoil crackingB inconsistent in many instances with climatic data of and enriched by greater atmospheric CO2 input. a more classical type involving the global distribu- Their Late Silurian near surface nodular calcrete tion of such things as coals and evaporitesŽ Fig. 1 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 43

summarizes these for the Phanerozoic. , or that differ- concurrence, with his theoretical atmospheric CO2 ent calcrete isotope studies should lead to conflicting curve. Tanner et al.Ž. 2001 discuss the disparity in results about atmospheric CO2 tenor. The conflicting estimated atmospheric CO2 suggested by some to be information provided in these varied studies indi- present across the Triassic-Jurassic boundary, with cates the difficulty of relying on pre-Cenozoic cal- their own results indicating atmospheric levels little crete isotopic data for atmospheric CO2 estimation. different than the modern. The most discordant results are those for the Late Worsley et al.Ž. 1994 point out that Berner’s error Silurian where geological criteria suggest a moder- level is 300%, thus capable of including just about ately high global climatic gradient, whereas the any data generated by any set of assumptions. For conclusions from calcrete data suggest very high example, Berner suggests that a 17-fold decline in atmospheric CO2 consistent with a very low global atmospheric CO2 occurred in his theoretical models climatic gradient. from the Silurian to Late Carboniferous–Permian

We illustrate some discrepancies inherent in this when CO2 levels were close to the present. Berner’s methodology in nearly age-concordant calcretes with post-1994 curves are not significantly different from a few examples. At the 150 million year point on his pre-1994 curves as regards the error estimate. Berner’s curveŽ. 1997, p. 545 , Yapp and Poths’ Mora et al.Ž. 1996 supported Berner when stating A Ž.1992 calcrete data disagrees with Cerling’s Ž. 1991 . that their results of atmospheric CO2 decline closely The calcrete data for the Early Cretaceous provided followedB the decline indicated in Berner’sŽ 1993a, by Freeman and HayesŽ. 1992 are close to Yapp and 1994. theoretical models. However, the decline indi- Poths’ but not to Cerling’s. There is also conflicting cated by their isotope workŽ Mora et al., 1996, their information between the calcrete isotopic data and Fig. 4. was only 10-fold and occurred specifically

Berner’s atmospheric CO2 informationŽ. Fig. 3 , said between the Late Silurian and the Late Devonian, by him to be A . . . in rough semi-quantitative agree- whereas Berner’s AdropB occurred within the Missis- B ment, as can be seen from the figure ŽBerner, 1997, sippian long after the CO2 decline indicated by the Figure, p. 545; Berner, 1998, his Fig. 5. . The data calcrete data. Within the estimated limits of pre- points indicate both disparityŽ. our Fig. 3 , as well as dicted CO2 from the isotope workŽ Mora et al.,

Fig. 3. Berner’sŽ. 1997, p. 545, modified model-based atmospheric CO2 curve Ž the broad band is the area of uncertainty . on which he has indicated varied calcrete-based atmospheric CO2 levelsŽ. vertical bars taken from the literature. 44 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

1996.Ž. , the Late Devonian Famennian , Carbonifer- Freeman and HayesŽ. 1992 agree with Andrews et ous and earlier Permian appeared to remain more or al.Ž. 1995 that Early Cretaceous atmospheric CO2 less stable during the time of significant fall based levels were higher than those present in the Late on the theoretical models. In Berner’sŽ. 1997 model, Cretaceous. CerlingŽ. 1991, 1992 and Sinha and the curve has now been altered to accommodate the StottŽ. 1994 in agreement with Cerling Ž 1991, 1992 . , previously discordant Late Devonian calcrete data also concluded that Early Cretaceous atmospheric provided by Mora et al.Ž. 1996 , so that the CO2 CO2 levels were higher than those of the Late AdropB now occurs in the Late Devonian rather than Cretaceous and Early Eocene. All of these conclu- in the Mississippian, eradicating disagreement at 370 sions conflict with the geologically deduced climatic Ma between the calcrete data and the theoretical data. Driese et al.Ž. 1993, p. 64 argue that overesti- curve. mation of atmospheric CO2 in the Cerling models is Other discrepancies exist between Berner’sŽ 1997, a Adepth effectB that exposed the calcrete to a more p. 545.Ž. CO2 curve Fig. 3 and the calcrete data. significant component of atmospheric CO2 . Between 100 and 200 Ma, Cerling’sŽ. 1991 calcrete The Ekart et al.Ž.Ž. 1999, their Fig. 12 our Fig. 2 data indicate consistently higher CO2 values than the curve also indicates relatively low Tertiary atmo- theoretical curve. Yapp and PothsŽ. 1992, 1996 mea- spheric CO2 , as contrasted with much higher levels surements are concordant at about 420 and 90 Ma, during the Jurassic and most of the Cretaceous. This somewhat discordant at about 440 Ma and at 150 anomaly also disagrees with the geological data indi- Ma, and discordant at about 360 Ma. Early Eocene cating that the Paleogene, particularly the Late Pale- calcrete data provided by CerlingŽ. 1991 and Sinha ocene–Early Eocene was a time of very low global and StottŽ. 1994 indicate significantly lower levels climatic gradient, and also the problem of bipolar than the theoretical curve. Andrews et al.’sŽ. 1995 glendonite and dropstone distributions during a good Late Cretaceous dataŽ plotted too low by Berner, part of the earlier CretaceousŽ. see our Fig. 1 . The 1997, fig. on p. 545. agrees with Ghosh et al.’s Silurian isotope data supplied by Mora et al.Ž. 1996

Ž.1995 data, and both are higher than Berner’s CO2 is not in accord with the moderate global climatic curve. gradient indicated by the geologicalrclimatic infor- Some calcrete data are also discordant with mation. This discrepancy might be due to their de- climatic information deduced from varied geolo- pendence on nodular calcretes because of the ab- gicalrclimatic proxiesŽ. Fig. 1 . Yapp and Poths’ sence of rhizoliths in the SilurianŽ Mora et al., 1991, Ž.1996 conclusion from studies of soil profile goethite 1996. . with occluded hydrated iron carbonate and associated calcretes led them to infer that the Early Juras- = sic atmospheric CO2 level was as much as 18 the 9. Models inspired by T.C. Chamberlin modern concentration, as contrasted with a more or less modern level for the Late Triassic, whereas the The models produced by Raymo et al.Ž. 1988 , geological dataŽ. Fig. 1 indicates a similar, low RaymoŽ. 1991 and Francois et al. Ž.Ž 1993 see below . global climatic gradient for both. Andrews et al. Ž.our Fig. 4 follow dicta presented by Chamberlin Ž.1995 conclude that Early Cretaceous calcretes indi- Ž.1899 a century ago, who posited that atmospheric cate high atmospheric CO2 in disagreement with CO2 levels are correlated with the prevalence of high latitude evidence for freezing winter conditions uplifted areas, times of orogeny, and high levels of Ž.dropstones and glendonites in both hemispheres. continental erosion. Uplift increased the landmass

They conclude that Maastrichtian atmospheric CO2 area subject to erosion, and weathering in turn in- = levels were 2–3 those of the present, in accord creased atmospheric CO2 loss, which in turn reduced with Berner’s theoretical curve. However, the data the atmospheric CO2 concentration. Correlations be- they present indicates atmospheric CO2 values that tween the effects of tectonic uplift and climate could be anywhere from 2.3–9= higher than mod- change, within the context proposed by Chamberlin, ern values, not 2–3= higher. are fully explored in RuddimanŽ. 1997 . A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 45

Fig. 4. Francois et al.’sŽ. 1993, modified atmospheric CO2 curve based largely on topographic effects on climate. Run a1 is for no seafloor weathering combined with silicate and carbonate weathering; Run a2 is for seafloor weathering combined with silicate and carbonate weathering; Run a3 is for seafloor weathering combined with Sr 87rSr 86 estimates of silicate and carbonate weathering; Run a4 is for seafloor weathering combined with strontium isotope estimates of silicate weathering, and carbonate weathering proportional to land area.

Those favoring the role of tectonically driven ated with tectonic events, give rise to high levels of, increase in geochemical, abiotic weathering conclude or enhanced, continental chemical rock weathering, that increased weathering rates caused a drawdown that in turn consumes atmospheric CO2 , leading to of atmospheric CO2 that leads to global cooling. drawdowns of atmospheric CO2 , which leads in turn They conclude that uplift is the primary cause of to important intervals of continental glaciation, i.e. changes in atmospheric CO2 during the Phanerozoic, climatic change. and that strontium isotope dataŽ. as discussed above Some students have used strontium isotope data to provide a measure of montane uplift and increased infer tectonism and glaciation and put more empha- weathering. RaymoŽ. 1991 used the strontium iso- sis on the strontium data than others. According to tope data as a proxy for topographic relief. these models, mountain building, not tracheophytes, Raymo, Ruddiman and others conclude, as Cham- is the Aprimary factorB determining the rate of chem- berlin did, that highly elevated landmasses, associ- ical weathering. 46 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

The RaymoŽ. 1991 model suggests that the record Ma. In her modelŽ. our Fig. 5 the more positive r of tectonism and glaciation can be read from the 87 86Sr values, assumed to indicate enhanced conti- signature left in the history of the strontium isotope nental erosion, are associated with decreased CO2 ratio. Thus, Raymo et al.Ž. 1988 suggest that the levels and globally colder climates. More negative r relation between chemical erosion, climate and 87 86Sr values, indicating enhanced sea-floor spread- orogeny can be evaluated in terms of the marine ing and hydrothermal activity, are associated with 87r86 carbonate Sr record of the Phanerozoic as dis- increased CO2 levels. A cussed above under the use of strontium iso- Intervals of high CO2 in Raymo’sŽ 1991, her Fig. topes . . . B, tacitly assuming that most of the stron- 1. model are: mid- to late-Ordovician; mid-De- tium was present in contemporary ocean water, and vonian; mid-Mississippian; latest Permian; Late then coprecipitated at the time when Ca and Mg Jurassic and minor increases in the mid-Cretaceous ions, derived from contemporary silicate rock weath- Ž.see our Fig. 5 . Another problem with this model is ering, was precipitated as carbonate in the oceans of that it attributes the drawdown of atmospheric CO2 the time, rather than heavily involving significantly to the weathering of Ca and Mg silicates, but ignores older weathered carbonate rocks with their loads of the addition of CO2 to the atmosphere due to oxida- much older radiogenic strontium, as is now known to tion of organic carbon sequestered in the rocks sub- commonly be the case. ject to weathering. This model disagrees with the RaymoŽ.Ž 1991 our Fig. 5 . used strontium iso- others discussed in this review. topes in marine carbonates as indices of atmospheric Moreover, her strontium isotope curve, plotted in

CO2 and global climatic conditions, such as comparison with intervals of widespread continental Phanerozoic glaciations, and of intervals of increased glaciationŽ. summarized from Crowell, 1983 , fails to orogenic activity, all of which she assumed are support some of her climatic inferences andror oro- correlated. From these data, RaymoŽ. 1991 generated genic episodesŽ. our Fig. 5 . Her strontium isotope a climatic model based on variations in strontium curve fails to show widespread orogenic activity in isotope ratios in marine limestones over the past 500 the Middle Devonian, AAcadianB Orogeny, glacia-

87r 86 Fig. 5. Raymo’sŽ. 1991, Fig. 1, modified atmospheric CO2 curve based on an assumed high correlation between Sr Sr ratios and atmospheric CO2 levels. The horizontally ruled band covers 93% of the strontium data; the black bars are the inferred times of continental glaciation, including erroneous Cambrian data. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 47 tion in the Late Ordovician and Early Silurian, ab- which Ashould be reconsideredŽ following Chamber- sence of glaciation or increased orogenic activity lin.Ž as an important if not the most important . Ž.our Figs. 1 and 5 at the Silurian–Devonian bound- influence on global temperatures both in the Late ary, and indicates both a decreased atmospheric CO Cenozoic . . . and during the PaleozoicB ŽRaymo, r 2 and a highly positive 87 86Sr value in the mid- 1991, p. 346. . Cambrian, where there is no evidence for glaciation. It needs to be emphasized here that these ap- Raymo attributes evidence for a mid-Cambrian proaches, apart from the RaymoŽ. 1991 model, em- glaciation to CrowellŽ. 1983 , but he notes only Ž pp. phasize data from the Cenozoic, particularly the 247–248. the ApossibleB evidence for a Cambro- development of the major montane regions Ordovician glaciation in central Bolivia indicated by Ž.Himalayan–Tibetan, Andean . Their conclusions boulders and blocks in the Limbo Formation, and emphasize that regional uplift induced by tectonic states that Ait is uncertain whether glaciation is processes is largely responsible for the greatly in- needed to explain these beds . . . B Discussions and creased weathering that releases large volumes of data in Hambrey and HarlandŽ. 1981 make clear that calcium and magnesium later precipitated as marine there is no indubitable geological evidence for carbonates. This weathering of silicate minerals re- widespread Cambrian glaciation, or for any localized lated to tectonic activity involving uplift, suggests a Cambrian glaciation. Nor are there any well-zoned correlation between purely physical erosion rates, Cambrian fossils occurring below or lateral to any measured by rates of mechanical erosion, and rates allegedly Cambrian glacial deposits. The dating by of chemical weathering that contribute to solute in- means of fossils for a later, approximately Mid- flux. This approach concludes that continental Cambrian phase of the lengthy Pan-African Orogeny weathering wasris principally a function of rates of that extends well back into the later Proterozoic is uplift and glaciation that exposed fresh mineral sur- not well agreed on, nor is the Mid-Cambrian geo- faces. This process would remove large amounts of graphic distribution of this orogeny. Increases in atmospheric CO2 , induce glaciation in at least some 87 r86 seawater Sr Sr in the Early Triassic and the Late cases in areas of high relief and act as a major CO2 JurassicrEarly Cretaceous indicated on her curve are sink. not associated with widespread glaciation. There is A number of workers have considered questions also widespread earlier Cenozoic evidence for cir- that bear on the significance of the tectonic model, cum-Pacific and Alpine–Himalayan Belt orogeny that and the impact of weathering and the related poten- is not associated with continental glaciation. These tial for climatic effects. The spirited exchange be- inconsistencies can be juxtaposed against some con- tween Raymo and RuddimanŽ. 1993a,b , Volk et al. sistencies in her isotopic curve with co-occurring Ž.1993 and Caldeira et al. Ž. 1993 illustrates the diffi- orogenic activity and continental glaciation, such as culties involved in evaluating the complex relation- those for the Late Paleozoic and the Late Cenozoic. ships between weathering rates and other parameters. As detailed above, there is reason to be skeptical Whipple et al.Ž. 1999 Alay to restB the concept that marine carbonate strontium isotopes provide a that there is a 1:1, reciprocal relation between cli- useful proxy for silicate weathering. The assump- mate and tectonism, one of the principal conclusions tions on which the methodology is based do not of the Raymo-Ruddiman et al. models. They con- accord with the facts, and the Aisotopic systematics,B clude that tectonism is actually independent of cli- in Edmond’sŽ. 1992 words, are more complex mate and that climatically induced erosion does not than originally envisaged. Raymo and Ruddiman automatically result in isostatic rebound that in turn Ž.1993a,b now indicate reasons why the87 Srr 86Sr gives rise to tectonism. ratios cannot necessarily be accepted as direct prox- White and BlumŽ. 1995 also provide a dissenting ies or correlatives of atmospheric CO2 values and voice. They correlated variations in solute concentra- tectonism. tions and ionic fluxes in surface runoff with tempera- In the Raymo and others models, variations in ture, precipitation, runoff and evapotranspiration for atmospheric CO2 throughout the last 600 Ma are 68 globally distributed watersheds, finding no strong controlled chiefly by episodic uplift and erosion, relationship between steep topography and high 48 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

chemical weathering rates and concludedŽ. p. 1729 : mospheric CO2 drawdown. Low elevation, high dif- ANo correlation exists between chemical fluxes and ferential relief areas subject to tropical weathering topographic relief or the extent of recent glaciation, will provide large amounts of silicate weathering implying that physical erosion rates do not have a products, whereas low elevation, high differential critical influence on chemical weathering rates.B, relief areas subject to coolrcold climate weathering while also statingŽ. p. 1745 : Athis study indicates will not. In contrast, low relief, low elevation regions there is not a strong relationship between steep to- are the most areally widespread during the Phanero- pography and high chemical weathering rates, and zoic. therefore, does not lend support to the concept of ATectonic hypothesesB depend ultimately on the tectonic control of global climateB as implied in the continuous exposure to erosion of fresh mineral sur- models of Raymo, Raymo and Ruddiman and Ed- faces with various levels of uplift andror glaciation mund. They concludeŽ. pp. 1729, 1743, 1745 that the as the principal factor in the acceleration or inhibi- effects of uplift and erosion cannot be considered tion of silicate weatheringŽ Francois et al., 1993; independently from the effects of both temperature Raymo et al., 1988; Raymo, 1991; Raymo and Rud- and precipitation. They considered mostly temperate diman, 1993a, 1992; see Ruddiman, 1997, for addi- areas, with little data from the tropics or polar re- tional papers dealing with this question. . There are gions, while concluding that precipitation and tem- potential complications to the concept of continuous perature are important correlatives of weathering. In weathering and erosion. Dorn and BradyŽ. 1995 , temperate area watershed studies, they found no Brady and ZacharaŽ. 1996 , Stallard Ž 1985, 1992, correlation between fluvial chemical fluxes and topo- 1995.Ž. and Stallard and Edmund 1983 discuss the graphic relief or extent of glaciation, implying that paleoclimatic significance of Abare-rockB weathering physical erosion alone does not have a critical influ- and Aregolith coveredB weathering. In the latter soil ence on chemical weathering rates. Thus, they write formation is rapid enough to bury primary minerals that lack of any correlation between chemical fluxes, whose dissolution is then limited by Athe transport of e.g. high chemical weathering rate, topographic re- the reactants into, and the products out of soil pro- lief or extent of recent glaciation, does not lend files.B They refer to the former as Aweathering lim- support to the concept of tectonic control of global ited landscapesB in the sense that physical erosion climate or support the concept that high rates of exposes unaltered minerals at a faster rate than they physical erosion Ahave a critical influence on chemi- can be chemically weathered and colonized by lichens cal weathering rates.B Despite the relevance of their and associated microbiota. In an environment of study, emphasizing the importance of temperature chemical weathering, where physical erosion is less and precipitation, the absence of significant data intense, referred to as Atransport-limited landscapesB from tropical and polar regions leaves open the all weatherable minerals contribute to the solute possibility that there may be latitudinally correlated load. They suggest that weathering limited environ- distinctions. To this should be added the effects of ments are likely to be highly important in CO2 different global climatic gradients. consumption, independent of biota, because when Another factor tectonic models should consider is erosion dominates, Ca- and Mg-silicates are continu- the duration of high relief areas, since short duration ously exposed to weathering while transport-limited relief will not provide large CO2 perturbations, soils will possess a smaller proportion of cation-rich whereas the reverse is the case in areas of long-term minerals because hydrological factors control solute reliefŽ. Himalayas, Alps and Andes . It is also critical fluxes in and out of soils are more important than the to know whether high relief areas are at high or low weathering susceptability of the silicates. They note elevations. The critical point in the tectonic uplift that it is essential to establish how the consumption models is that areas of bare rock not subject to of atmospheric CO2 by silicate weathering is ef- extensive biotic chemical weathering are largely fected during the transition from bare-rock to re- limited to high montane regions which may be golith weathering. Chemical weathering in soils short-lived, contributing very little to long-term at- destabilizes the surface that is then removed, and A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 49 exposing AnewB, unweathered surfaces to fresh physical erosion rates, and soil ages are of secondary weathering and new soil build up and the cycle is importance relative to precipitation and temperature repeated. in controlling weathering rates.B StallardŽ Stallard, 1985, 1995; Stallard and Ed- Others have concluded that limited relationship mund, 1983; Drever, 1994, who bases his discussion exists between significantŽ. or high chemical weath- on Stallard. notes the significance of topographic ering rates and topographyŽ. or steep topography . relief and continental denudation in the global geo- Berner and BernerŽ. 1997 conclude, like Holland chemical cycle drawing, however, a major distinc- Ž.1984, p. 181 , that Ahigh topographic relief in conti- tion between Atransport-limitedB Žslight slopes with nental areas does not, per se, lead to high rates of thick soils.Ž and Aweathering-limitedB steep slopes chemical weathering . . . B, rather they argue that a with limited soils. erosional processes in contribut- number of other variables—such as rainfall and ing to bedload denudation and river load and river temperature and most significantly vegetation—will chemistry. also play a significant role and cannot be ignored in In areas of steep topography, if transport pro- comparison with sole emphasis on topographic up- cesses are temporarily decreased, rooted plants will lift. increase chemical weathering rates through their ef- The tectonic uplift model leads to conclusions in fects in binding and retaining soil particlesŽ and conflict with other atmospheric CO2 models. For thereby increasing the surface area of minerals ex- example, according to Raymo and Ruddiman posed to weathering and the residence time of water Ž.1993a,b, p. 119 , Berner-type models predict a prin- A B and bioacids which increase reaction time. . Ulti- cipal decline in atmospheric CO2 and global temper- mately, this chemical weathering leads to soil insta- atures between 100 and 50 MaŽ mid-Cretaceous– bility and erosion exposing AnewB unweathered sur- Paleocene.Ž , while geological evidence uplift of face to fresh weathering, followed in time by a new the Tibetan Plateau, Himalaya, Andes, Alps, Ameri- cycle of soil accumulation and renewed chemical can Cordillera. suggests AGreater climate cool- weathering and another destabilized surface. In sub- ing . . . primarily between 50 Ma and the present dued topographyŽ. e.g. low relief areas , on the other Ž.post-Eocene .B By the same token it is essential to hand, by anchoring soil and allowing very thick soil determine whether weathering might not be inhibited accumulations, the rate of chemical weathering may rather than promoted in some topographic and cli- actually be decreased by binding secondary products matic settings by surfaces stabilized by an effective and isolating bedrock and unweathered minerals from root system, in addition to determining what effects meteoric water, chemical weathering thus being neg- roots, independently of their associated biota, actu- atively correlated with soil thicknessŽ Drever, 1994; ally have in the weathering process. Stallard, 1985. and soil age. Soils over time thus Pagani et al.Ž. 1999 conclude, employing evi- show a significant decrease in chemical weathering dence of carbon isotope changes in oceanic plank- rates as reactive minerals are removed. As a conse- tonic foraminifera and diunsaturated alkenones, that quence, watersheds with younger soils and large the highest later Cenozoic pCO2 was in the latest areas of exposed bedrock may be expected to have Oligocene and decreased rapidly into the Miocene. higher weathering rates than areas with accumula- The Early and Middle Miocene was characterized by tions of thick soils. low pCO2 , which corresponds to inferred organic However, White and BlumŽ 1995, p. 1744, their carbon burial and glacial episodes; lowest concentra- Fig. 13. tested this hypothesis by comparing weath- tions occurred during the Middle Miocene. They ering fluxes for glaciated and non-glaciated water- conclude that the strontium isotope data generated by sheds and found no clear differentiation of SiO2 and RaymoŽ. 1994a,b are inconsistent with their atmo- Na fluxes from watersheds that had experienced spheric CO2 data developed from carbon isotopic recent alpine glaciations, those that experienced information. Pleistocene glaciation, and those that had not been StallardŽ 1985; also Drever, 1994, who bases his glaciated for at least 250 ka. They concluded from discussion on Stallard. notes that in areas of steep these data that the Aeffects of watershed topography, topography, if transport processes are temporarily 50 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 decreased, rooted plants will increase chemical more geographically restricted high relief tropical weathering rates through their effects in binding and situations, white water rivers, there is high nutrient retaining soil particlesŽ and thereby increasing the availability. GibbsŽ. 1967 provides an excellent ex- surface area of minerals exposed to weathering and ample, the Amazon Basin, showing that the bulk of the residence time of water and AbioacidsB which the chemical weathering occurs in the non-tropical increase reaction time. . Ultimately, this chemical montane hinterlands to the west of the lowland tropi- weathering leads to instability and erosion, followed cal region, while Reynolds and JohnsonŽ. 1972 indi- in time by a new cycle of soil accumulation and cate an extreme of chemical weathering activity in renewed chemical weathering. In subdued topogra- the high Northern Cascade Mountains in what they phy, on the other hand, by anchoring soil and allow- define as a Atemperate glacial environment,B al- ing very thick soil accumulations, the rate of chemi- though it should be noted that the microclimates cal weathering may actually be decreased by binding within and upon the rock surfaces at higher elevation secondary products and isolating bedrock and un- commonly are warm enough to almost qualify as weathered minerals from meteoric water, chemical AtropicalB. They note the global correlation for any weathering thus being negatively correlated with soil given region, between amount of runoff and amount thicknessŽ. Drever, 1994; Stallard, 1985 and age of of concurrent chemical denudationŽ citing Liv- soil. ingston, 1963.Ž . Strakhov 1967, pp. 4–9 . on the Decrease in the rate of chemical weathering is other hand, stressed that increased relief accelerates related to soil thickness. The last is particularly the physical weathering and suppresses chemical weath- case in tropical regions of low relief subject to ering, but his conclusions are based largely on higher lateritic type weatheringŽ Edmond and Huh, 1997, latitude non-tropical data plus information from the Tables 1 and 2. where there is essentially no rock lower reaches of a few tropical rivers rather than weathering in regions with thick mantles of laterite their upper reaches where chemical weathering is Ž.Berner and Berner, 1997 . important. Gorham et al.Ž. 1979, p. 60 conclude that Discussing the question of plant stabilized soils, if high rainfall and high relief are associated the BernerŽ. 1992, p.3229 states AOverall, I feel that removal of weathered minerals likely takes place vascular plants must bring about a net acceleration of under conditions that favor chemical weathering of chemical weathering, and not a deceleration of primary minerals. Based on varied disparate interpre- weathering due to the stabilization of soils, if for no tations, this is clearly a question deserving of study other reason than that they cannot recycle nutrients and quantification. Edmond and HuhŽ. 1997 express with perfect efficiency and, thus, require prolonged concern about the effects of weathering in humid rock decomposition to provide the nutrients neces- tropical regions of low relief, but similar concern has sary for continued growth and existence.B This state- not been shown for low relief arid regions, ment is in error. It is common knowledge that the widespread during some past geological intervals. widespread, low differential relief, low elevation We would conclude that low relief, arid region pale- forests of the wet tropics and subtropics, growing on osols would armor unweathered bedrock from addi- nutrient-poor, lateritic soils, such as oxisols, are very tional weathering in a manner similar to that effected efficient at recycling nutrients. Lucas et al.Ž. 1993 by lateritic cover. Note that White and Blum’sŽ. 1995 make a good case for silica recycling by plants in the conclusions about there being no correlation between Amazon region, which also accounts for the common high relief and chemical weathering are not in agree- occurrence of a kaolinite-rich layer immediately be- ment with the above. neath the trees, and above the alumina-rich lower During the Phanerozoic low differential relief, layers. . There is a rich and varied literature on this low elevation regions were far more widespread than topicŽ for examples see, Herrera et al., 1978; Jordan, those with high differential relief and elevation mak- 1982; Jordan and Herrera, 1981; Longman and Jenik, ing it possible to question the amount of global 1987, pp. 58–65, 211–218; Sanchez, 1976; Proctor, weathering that has taken place. In low relief, low 1989.Ž. . Herrera et al. 1978 point out that in the far elevation regions of tropical lateritic type weather- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 51 ing, once the geologically brief initial period of such estimates, realistic evaluation of the contribu- intense silicate mineral weathering has been com- tion of biotic, chemical rock weathering to atmo- pleted, mineral solution takes place at a very low rate spheric CO2 is impossible. because the thick, underlying lateritic deposits virtu- In view of the above, the papers referred to above ally lack undissolved silicate minerals. These low-re- are only partially relevant to the problems they at- lief, lateritic regions will contribute very little to CO2 tempt to address because they ignore one or more of drawdown due to weathering. These regions essen- the key variables involved through geological time, tially correspond to Stallard’sŽ. 1985 Atransport- as well as the present. limitedB regimes. In contrast with humid tropical regions, low relief 10. Worsley et al. model regions subject to arid climate regimes present a problem because we have little understanding of how Worsley et al.Ž. 1994 provide another model their soil biota, during the pre-tracheophytic or tra- prompted by the early work of Chamberlin. In addi- cheophytic time intervals, affect the undissolved sili- tion to the importance of tectonic uplift and glacia- cate minerals in soils. Arid region soils are rich in tion, they stress, unlike the Raymo, Ruddiman et al. smectite, indicating Mg retention, rather than loss; models, the effect of different global paleogeogra- their pedogenic calcretes indicate Ca retention fol- phies on the distribution of climatic belts, which in lowing silicate mineral rock weathering. The large turn will effect the distribution of weathering envi- regions of low relief arid conditions present during ronments. Their preliminary results clearly support the Phanerozoic suggest that, as with the low relief the premise that the changing positions of major land lateritic regions, there is little, long-term silicate areas through time will result in globally large cli- mineral weathering involved. matic changes. Phanerozoic atmospheric CO2 levels Complicating the use of calcretes as a potential generated by the Worsley et al.Ž. 1994 model Ž our measure of silicate weathering to provide a measure Fig. 6. disagree significantly with the CO2 curves of released calcium that might combine with atmo- generated by the BernerŽ. 1993a, 1991b Models, spheric CO2 is Capo and Chadwick’sŽ. 1999 work RaymoŽ. 1991 , and the Cerling Ž. 1991 and Mora et indicating that the source of calcium in some Pleis- al.Ž. 1991 calcrete data. tocene calcrete from the U.S. Southwest is atmo- The paleogeographic approach combined with the spheric dust transported from outside with isotopic tectonic approach would provide an effective model evidence indicating the source to have been lime- for global climate and atmospheric composition stones and evaporites, i.e. not contributing to net through time, which might be close to the results silicate weathering at all. obtained from considering climatically sensitive geo- It is clear that high relief, lower elevation regions logical deposits through timeŽ. our Fig. 1 . Important will be where most silicate rock chemical weathering impetus has been given to the paleogeographic ap- takes place, with temperature and precipitation con- proach by Pearson and PalmerŽ.Ž 1999 see also Kerr, trolling its extent, high under tropical conditions, 1999. who point out, employing boron isotopic data, moderate under temperate conditions, and low under that the Middle Eocene globally relatively warm cold polar conditions. High relief, higher elevation interval can be most effectively explained as due to regions just about everywhere will be dominantly the potential of surface ocean currents carrying warm cold and not very conducive to extensive silicate waters to higher latitudes as contrasted with hypo- rock chemical weathering. thetical increases in atmospheric CO2 . Pearson and In attempting to evaluate the global role of tra- PalmerŽ. 1999 suggest that the different paleogeogra- cheophytes in rock weathering over geological time, phy of the Middle Eocene may have been crucial in an evaluation of the relative areas of high and low setting the oceanographic stage! They comment A relief, high and low elevation subject to tropical . . . atmospheric pCO2 was probably similar to weatheringŽ. with dense vegetation or arid region modern concentrations or slightly higher.B weatheringŽ. with minimal tracheophytic vegetation , Walker et al.Ž. 1981 comment that the changing time interval by time interval, is essential. Without areas on the continents subject to silicate weathering 52 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

A B Fig. 6. Worsley et al.’sŽ. 1994 Fig. 11 Ž modified . , indicating disagreement between varied Phanerozoic atmospheric CO2 levels. a is A B A B Berner’sŽ 1990, 1991a,b . CO2 curve; b is Cerling’s Ž. 1991 and Mora et al.’s Ž. 1991 curve; c is Worsley et al.’s Ž. 1994 curve based on global latitudinal and land area data. would affect the amounts of calcium and magnesium flawed owing to its reliance on unreliable assump- released for precipitation as carbonate, thus reducing tions about the significance of the strontium isotopes, the amount of atmospheric CO2 , but go into no the seafloor spreading rates based on Vail’s sealevel detail about this effect through geological time. curve through time conclusions and the Berner assumptions for several of the models. It disagrees with our Fig. 1 that summarizes the geological data 11. Francois et al. model on Phanerozoic global climatic gradients. Francois et al.Ž. 1993 provide four models Ž our All their models indicate a Middle Cambrian to

Fig. 4. for atmospheric CO2 in which varied parame- earlier Devonian CO2 high that disagrees with the ters are adjusted to conform in part to certain Berner geologic approachŽ. our Fig. 1 , fail to show a low for and Raymo models. None of the Francois et al. the Late Ordovician glaciation, or for the latest Ž.1993 models agree with the global climatic data Devonian glaciation, and do not indicate the Late Ž.our Fig. 1 . Francois et al. Ž. 1993 rely on the Jurassic–Early Cretaceous interval of cool, high lati- strontium isotope curve derived from carbonates tude winters with freezing conditionsŽ. our Fig. 1 . through time, which has been found to be defective Francois et al.Ž. 1993 noted some inconsistencies for Ž.see above . They conclude that the strontium curve many geological variables, including such things as correlates with glaciation and mountain building, but the different compositions of rocks subject to weath- not with CO2 or global temperature. The Gaffin ering in different time intervals to mention one. Ž.1987 conclusions about spreading rates are derived Francois and WalkerŽ. 1992 suggest that seafloor from assumptions about the Vail curveŽ see Payton, basalt alteration and not terrestrial silicate rock 1977. for Phanerozoic sealevel through time. The weathering feedback is the most important factor

Vail curve appears to be relatively reliable for the determining long-term steady state atmospheric CO2 Jurassic to present, where there is available seafloor concentrations. They test a number of Amodel possi- spreading data, but is unreliable for the older rocks. bilitiesB against the strontium isotope record. They In summary, the Francois et al.Ž. 1993 effort is assume the seawater 87Srr86 Sr record as a AgivenB A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 53 and then try to make continental weathering, vol- that it is not possible to discount the effects and that canism and surface lithology conform in order to there may be variations in the fluxes of CO2 into and explain the strontium data. A . . . a history of atmo- out of the oceans and atmosphere related to alter- spheric carbon dioxide and climate during Phanero- ation in oceanic basement that Amay influence atmo- B zoic time, consistent with the strontium isotopic data, spheric pCO2 , without regulating it ŽCaldeira, 1995, is reconstructed with the model and is shown to be p. 1103.Ž . Brady written communication, 2001 . com- compatible with paleoclimatic indicators, such as the ments ANote though that CaldeiraŽ. 1995 neglected timing of glaciation and the estimates of Cretaceous any temperature dependence of basalt weathering, paleotemperatures.B the specific feature that gives rise to the continental Brady and GislasonŽ. 1997 argue that weathering weathering sensitivity.B Thus, he concludes that Aoc- of bedrock in seawater is also temperature-sensitive ean basement carbonate accumulation should be in- and that deep ocean temperatures are Aparticularly cluded in global carbon budgets and may impact B sensitive to atmospheric CO2 levels. They define atmospheric CO2 content by affecting the degassing B seafloor weathering as the oxidative and non-oxida- of CO2 during subduction zone metamorphism Žp. tive alteration that occurs when seawater is in con- 1077. . None of the available seafloor weathering tact with the upper several hundred meters of the data is truly amenable to reliable estimation. oceanic crust as well as alteration by seawater of exposed ocean floor. As in continental weathering, they are especially interested in Ca leaching and in 12. Tappan model based on phytoplankton peri- A B CO22 consumption, here marine CO . They con- odicity clude that weathering of surficial marine basalts at temperatures less than 408C is a potentially impor- TappanŽ.Ž 1968 our Fig. 7 . has argued that the tant sink for atmospheric CO2 over geologic timeŽ p. atmospheric CO22 :O balance since the early Precam- 965. , although the absolute size of the seafloor brian has been controlled by oceanic phytoplankton weathering sink for CO2 remains unclear as does the photosynthesis rather than varied geochemical pro- thermal regime that drives it. They state from results cesses. Tappan hypothesized that atmospheric CO2 of petrographic analysis of laboratory-generated would decrease significantly when atmospheric O2 basalt weathering in seawater-like solutions that a increased as a result of higher phytoplankton produc- significant feedback exists between seafloor weather- tivity. Thus, a CO22 :O imbalance would result during and climate. They also hypothesize that their ing periods of minimal phytoplankton productivity. results Asupport the notion that seafloor weathering Phytoplankton productivity thus became the proxy to A might have moderated climate in the continent-free atmospheric O2 composition. According to the phy- world of the Archean . . . B Ž.p. 971 . toplankton periodicity modelB Ž.our Fig. 7 intervals

Thus, some have argued that the long time control of decreased atmospheric O2 coincided with a crisis of atmospheric CO2 is related to seafloor alteration in phytoplankton productivity that began in the latest rather more than it is to continental silicate rock Devonian and lasted into the Triassic, followed by weathering. CaldeiraŽ. 1995 has concluded, however, briefer intervals of low productivity at the close of on the basis of extensions of the Francois and Walker the Mesozoic, and several low productivity levels in model and additional models and calculations, that the Cenozoic. Tappan’s model depicts equivalently silicate rock weathering on land, rather than low- high O22 , low CO for the Ordovician, Silurian, temperature seafloor basalt alteration, is the primary Devonian, Cretaceous and much of the Cenozoic. control on long-term atmospheric CO2 concentra- TappanŽ. 1968 and Tappan and Loeblich Ž. 1970 tion. While he concludes that seafloor basalt dissolu- assumed that atmospheric O22 and CO were both a tion is unlikely to be a major or dominant factor function of phytoplankton productivity. It can be regulating long-term atmospheric CO2 because con- assumed from context that they considered high phy- tinental silicate rock weathering is more sensitive toplankton biomass to be a function of taxic diversity Aby several orders of magnitudeB to changes in through time, but this is not discussed. There is no ocean–atmospheric CO2 content, Caldeira concedes evidence for correlating phytoplankton biomass or 54 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Fig. 7. Tappan’sŽ. 1968, Fig. 1, modified atmospheric O2 curve, based on fossil phytoplankton abundance derived from the assumption that changes in taxonomic diversity, high to low, are paralleled by phytoplankton biomass changes from high to low. APALB refers to the Present Atmospheric Level of oxygen, with correspondingly high and lower levels shown on either side, which presumably correspond to numbers of phytoplankton taxa.

productivity and phytoplankton taxic diversity. No peratures they probably would achieve results in data exist on changing phytoplankton biomass accord with the geological evidence. through time. Their atmospheric composition curve has no counterpart in the other modelsŽ. our Fig. 7 and in some cases suggests inverse relationships 14. Budyko et al. model compared with the other models. Nor do Tappan’s conclusions agree with the geological dataŽ our Fig. 1. used to estimate global climate gradients. The Budyko et al.Ž.Ž 1987 Budyko, 1986, his Fig. 7.1.Ž. model Fig. 5 is causally related to globally changing levels and intensity of volcanic activity through geologic time since they assume that most 13. Freeman and Hayes model A atmospheric CO2 is ejected during volcanic erup- tionsB and related to Adegassing of the upper mantle Freeman and HayesŽ. 1992 consider the carbon and the higher layers of the crustB ŽBudyko et al., isotope fractionation caused by marine phytoplank- 1987, pp. 11–14.Ž . Budyko 1986, his Fig. 7.1 . calcu- ton, modern and ancient, as a basis for estimating lated variations in CO2 determined from Ronov’s past levels of atmospheric CO2 . Their methodology Žsee Ronov, 1976, for undocumented estimates of depends in no small part on reliable estimates of past organic carbon in Phanerozoic sedimentary rocks. surface water temperatures, which they have based investigations of organic carbon in Phanerozoic con- largely on data supplied by others. When their continental and Cretaceous and younger oceanic sedi- clusions are compared with the geologically deter- ments, and igneous rocks in the PhanerozoicŽ see our mined global climatic gradientsŽ. our Fig. 1 it is earlier comments regarding the Ronov data. . Budyko clear that these temperature estimatesŽ. their Fig. 6 suggests that to understand causation in the fluctua- for Late Jurassic and Early Cretaceous strata are far tions in atmospheric CO2 content, it is necessary to too high, whereas their estimates for the earlier study igneous rock formation per time interval, con- Eocene are far too low. If they recalculated their sidered to correspond to volcanic activity fluctua- data, employing more realistic surface region tem- tions and intensity. Thus, he concludes that great A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 55 changes in volcanic activity took place during the any of the other models. It is especially interesting

Phanerozoic, and that volcanic activity varied rhyth- that the drawdown in atmospheric CO2 concentration mically with highest peaks occurring every 100 mil- posited by HollandŽ. 1984 and later by Berner for lion years. BudykoŽ. 1986 and Budyko et al. Ž. 1987 the Devonian related to the spread of higher land used RonovŽ. 1976 estimates of changing global plants is not corroborated in this model, where atmo- levels of volcanism, which are provided in tabular spheric CO2 concentration is increased rather than form without any information about their sources or decreased in the Devonian and Early Carboniferous. the reliability of those sources. One of the basic The atmospheric CO2 models discussed above are deficiencies of the Ronov data base is the absence of based on poorly constrained data. None of these the documentation necessary to evaluate the summa- models agree. Suggested CO2 fluctuations in the rized data. Well documented global assessments of Berner models, bear no relation to changes in the differing Phanerozoic levels of volcanism, including CO22 :O balance suggested by TappanŽ. 1968 , to the some idea of the relative amounts of different types, curve produced by Budyko et al.Ž. 1987 , or to the have yet to be made. Raymo-typeŽ. 1991 curves. All employ different

Their model has six intervals of maximum CO2 proxies to arrive at their disparate conclusions. None concentrations during the Phanerozoic: Early Or- agree with the climatic informationŽ. our Fig. 1 dovician; A . . . the second started in the Devonian, based purely on geological data. They all disagree in and in the Early Carboniferous the carbon diox- some major features and agree in others. It is clear ide concentration was highest for the entire Pha- that all of the variablesrproxies considered in the nerozoic . . . B; Early PermianŽ. second greatest ; Mid- different models are involved to one extent or an- dle Triassic; Late Jurassic; and Late Cretaceous other in affecting climate, rock and soil weathering

Ž.Ž.Budyko et al., 1987, pp. 80–81 our Fig. 8 . The and atmospheric CO2 . Global paleogeography and last two maxima Aoverlap a noticeable increase in climatic distribution need to be evaluated through the level of carbon dioxide concentration that em- time, as does the distribution of high elevation, high braced most of the Jurassic and the entire Creta- relief and low elevation, low relief regions, plus the ceousB Ž.Budyko et al., 1987, p. 81 . The atmospheric distribution through time of vegetation belts, includ-

CO2 curve of this model does not agree with that of ing arid regions poor in tracheophytes, and the rela-

Fig. 8. Budyko et al.’sŽ. 1987, Fig. 24, modified atmospheric CO2 curve indicating their conclusion about there being a high correlation between the abundance of volcanic rocks and atmospheric CO2c . V is the rate of formation of volcanic rocks; M , solid line is the 1987 calculation of atmospheric CO2c , whereas M , dashed line, is an earlier, less precise calculation. Nowhere is there any indication that the volumes of volcanic rocks through time were directly measured, which deprives Budyko’s conclusions of much reliability. 56 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 tive abundances of Ca- and Mg-rich silicate and None of the modelers agree about the major carbonate rocks subject to weathering. This calls for factors that ultimately play a significant role in the intensive time-sequence compilation of the available chemical weathering of silicate minerals, and their geological data rather than for making models based relative importance, or the principal proxies neces- on single or a few proxies, since there are no model- sary to assess the significance of weathering. Each ing procedures that can reliably deduce the geologi- modeler tends to emphasize one factor or one proxy cal data. at the expense of others, suggesting that one is far more important than others. The factors range from the biotic to the abiotic. 15. Chemical weathering: evaluating some of the Broecker and SanyalŽ. 1998 emphasize that there factors in a multifaceted problem needs to have been an overall equilibrium between

Rock and soil weathering involve physical and atmospheric CO2 and dissolved calciumŽ and magne- chemical processes that may be biotic andror abi- sium. provided by weathering of silicate minerals, otic. In chemical weathering, both relatively stable and dissolved CO2 and dissolved carbonate rocks. secondary minerals and relatively soluble substances Therefore, whether one favors tracheophyte evolu- are produced. While there is general agreement that tion, tectonically induced uplift, changing paleogeog- soils are major sites of weathering, the significance raphy, global climatic gradient, or something else as of bedrock weathering tends to be overlooked. being the AkeyB factor one must ultimately appeal to

Chemical weathering rates of soil and continental some kind of equilibrium involving atmospheric CO2 bedrock are complex functions of temperature, hu- and calciumŽ. plus magnesium supplied by weather- midity, organic activity, exposure, relief and eleva- ing as the ultimate control leading to equilibrium. tion and sometimes ageŽ cf. Brady and Gislason, In order to accommodate this diversity of opin- 1997. which cannot be assessed by one proxy. Fac- ions about the critical factors and proxies for tors important in seafloor weathering are not only Phanerozoic chemical weathering, we have divided less clear, but have been comparatively little studied this discussion as follows: Relief and Elevation Ef- Ž.Brady and Gislason, 1997 . Brady and Gislason fect, covering the potential weathering effects, chiefly Ž.1997, p. 971 examine the linkage between seafloor abiotic, of relief and elevation; Paleogeographic Ef- weathering and climate and note that while the re- fect; The Soil Effect, covering the soil weathering sponse of seafloor weathering to changes in atmo- capabilities of tracheophytes and their mycorrhizae; spheric CO2 is less than continental weathering, it is and Bedrock and Vadose Zone Effect, which covers not Aorders of magnitude . . . smallerB. bedrock weathering capabilities of bacteria, lichens, Aside from the Tappan and Budyko et al. models, fungi, algae and bryophytes. other atmospheric models discussed above agree that Below we discuss some of the principal biotic and chemical weathering of Ca- and Mg-rich silicate abiotic effects that geologists have seen as impli- rocks is a first order control over atmospheric CO2 cated in mineral weathering rates as they affect levels and global surface temperature through time, atmospheric composition over geological time. In in addition to the oxidation of organic carbon that assessing the role of weathering in relation to model- adds CO2 to the atmosphere. Silicate mineral weath- ing, the C-cycle and climates one or more of these ering of calcium- and magnesium-rich species is effects has commonly been emphasized. In the A B viewed by many as the primary sink by which CO2 tectonic effect , for example, chemical weathering is removed from the atmosphere to counterbalance in soils and bedrock is viewed chiefly as a function r additions of atmospheric CO2 from organic matter of erosion related to elevation and or continental oxidation, volcanism and metamorphism. Dissolution glaciation. In the Apaleogeographic effectB, positions of Ca and Mg silicates is seen as particularly impor- of the continents relative to the climatic belts are tant because of the role that these cations play in viewed as the chief control on mineral weathering. In precipitation of oceanic carbonates; the dissolution both of the above AeffectsB, mineral weathering is of Na and K feldspars is less criticalŽ Walker et al., chiefly viewed as due to abiotic chemical and physi- 1981; Brady, 1991. . cal phenomena. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 57

In the Asoil effect,B predominant roles have been for submarine bedrock weathering, as well as its attributed to various organic and inorganic chemicals importance relative to terrestrial weathering in the produced by tracheophyte roots, symbiotic fungi ultimate removal of CO2 from the atmosphere. Giles Ž.mycorrhizae and associated free-living bacteria and et al.Ž. 1994 indicate that subsurface organic acids do fungi. Soil weathering is commonly assumed to be a not generate much porosity. McKinley et al.Ž. 2000 biological phenomenon. No significant, biologically even suggest that there are possible subsurface bacte- mediated, chemical, mineral weathering is assumed rial fossils, i.e. a capability for recognizing fossilized to have taken place before there were tracheophyte- bacterial weathering, subsurface activity. There is inhabited soils. In this view soils are a phenomenon agreement that Ca and Mg silicate minerals are associated with the advent of the rooted tracheo- inherently unstable at the Earth’s surfaceŽ April and phytes. Although some, like DreverŽ 1994; see Jack- Newton, 1992. . son, 1996, for a contradiction of Drever’s views. , We have divided the discussion below into min- attribute soil weathering rates to Aland plantsB inter- eral weathering in soils and mineral weathering of preted to mean the Adirect effect of higher plants and surface and vadose zone bedrock. It seems clear that the effects of associated bacteria and fungi,B microbiotic surface bedrock weathering was an important bial organisms and pre-tracheophytic embryophytes phenomenon before the advent of the tracheophytes, are seen to play no significant, independent role as it is today in some tracheophyte-deficient regions, either in non-tracheophytic soils or in bedrock in addition to the obvious importance of tracheophyt- weathering. BurnsŽ. 1986 considers the possible role ically induced weathering. The literature on mineral of microbially generated enzymes in the soil system. weathering in soils is extensive; the literature on Others have stressed the potential significance of bedrock weathering is growing in volume. Seldom is temperature and the role of CO2 Ž atmospheric and there considered to be any commonality in the pro- soil CO2 partial pressure. in soil mineral weathering. cesses between them. There have been only limited r This CO2 effect may be biotic and or abiotic. It attempts to cross compare mineral weathering effects involves the potential for significant CO2 weathering between exposed bedrock in climatically varied envi- of soils prior to the advent of land plants in the ronments and between bedrock and soils derived scenario advocated by HollandŽ. 1984 who stressed from exposed bedrock in climatically varied environ- that increased CO2 can occur in soils occupied by mentsŽ humid tropics, warm arid, cool temperate, organisms as a result of microbial respiration, andror polar. . Since soils are derived from bedrock there are as the result of dissolved atmospheric CO2 . obviously many transitions between fresh bedrock Lastly, we consider the Abedrock and vadose zone and mature soil. effect,B in which we advocate a potentially signifi- There is agreement that physical and chemical cant role for microbial weathering on bedrock sur- weathering affect minerals at the Earth’s surface. faces and in pre-tracheophytic soilsrregoliths and There also seems to be general agreement that min- suggest that weathering rates of silicate minerals erals in soils rather than minerals in bare bedrock are perpetrated by these organisms may have had mea- particularly susceptible to chemical weathering be- A surable effects on the global CO2 budget and global cause of the confluence in soils of atmospheric, temperatures long before the changes, advocated by hydrological, biological, and geological processesB some, that are believed associated only with the Ž.April and Newton, 1992, p. 379 and the fact that evolution of the tracheophytes and development of the largest store of Ca and Mg is within minerals rhizosphere soils. In the vadose zone effect, we constituting forest soils at least in North American consider the potential of subsurface chemotrophic and European temperate forestsŽ April and Newton, bacteria, which could have been present far back in 1992, pp. 378–379.Ž . April and Newton 1992, pp. time. Some evidence also exists for microbially me- 378–379. state that minerals in soils are particularly diated bedrock weathering in the vadose zone below susceptible to chemical weathering in the Atransi- soils in deep aquifers which can contribute to tional boundary between Earth and its atmosphereB groundwaters and to dissolved ions in runoff. Far where weathering is driven Aby the confluence of less well understood are the mechanisms responsible atmospheric, hydrologic, biological and geological 58 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 processes.B However, they do not consider tropical appearing in the earlier Devonian, today lies in the and arid region soils. effects of mycorrhizal fungi, both ecto- and endo-, so April and NewtonŽ. 1992 sum up the complex of intimately associated with manyrmost tracheo- variables in any given region which influence the phytes. The weathering effects of these rhizosphere rate and extent of soil mineral weathering to include Žthe tracheophyte root system plus its associated Ž.p. 379 Afactors such as climate, annual precipita- symbiotic fungi and their mycelium extending into tion, soil temperature, soil water content and chem- the soil, ectomycorrhizal extensions of the roots. istry, soil pH, vegetation, mineral grain size.BAThese generated soil components are clearly important ad- include mineralogical factors such as mean annual ditions to the weathering process. However, it should temperature and rainfall; hydrologic factors such as be emphasized that the important contribution of hydraulic conductivity and permeability of the soil; tracheophytes to weathering is made chiefly in hu- physiographic factors such as slope, aspect, relief, mid regions where tracheophytes are abundant, less and elevation; and chemical factors such as precipita- so in arid regions, and very little in highly arid tion and throughfall chemistry, soil pH, and the regions. This was probably also true during the nature and quantity of organic compounds generated Devonian to present. It has been assumed that pre- by the decomposition of organic matter. The interac- tracheophytic regolithsŽ definable as the raw, largely tions among these variables ultimately determine the organic-free layer of rock and rock fragments pro- weathering rate within a defined settingB Ž.p. 405 . In duced by bedrock weathering. or whatever term one order to understand soil weathering it is essential to prefers, were significantly lower in organic carbon as consider the entire complex of variables. This has are modern desert crusts in arid regions and other not yet been done. When it is accomplished we will environments lacking in tracheophytes, whereas it is have a far better understanding of the synergistic only with the incoming of tracheophytes that organic effects of uplift, paleogeography and ever changing carbon in soilsŽ definably AsoilsB are basically com- climatic distribution patterns on Phanerozoic atmo- binations of disintegrated and decomposed rock and spheric composition and climates. Bare bedrock sur- decayed remains of organisms and living organisms. faces also fall into the narrow transitional boundary becomes volumetrically significant. Nash et al.Ž 1977; between AEarth and its atmosphereB just as soils do. see also Moore, 1998. provide some data on lichen The impact on bedrock of the variables affecting soil abundances in desert crust from the North American weathering are not commonly considered to be the southwest, suggesting that they are commonly signif- same. Biotic factors that influence the rate and extent icant although minor components. This assumption, of weathering in soils are commonly considered to the large scale absence of organic carbon, is sup- have minimal or no effects on bare bedrockŽ April ported by the absence of pre-Late Devonian coals, and Newton, 1992, p. 379. . even from presumably humid regions, based on the Today there is a distinct succession leading from evidence that massive accumulations of desert crust- unweathered bedrock to bedrock covered partly or like or other organic autotrophic andror fungal accu- entirely by a soil in which tracheophytes thrive. The mulations, in the pre-Devonian, capable of generat- succession involves varied lower autotrophs, such as ing coaly layers did not exist. However, the absence cyanobacteria, lichens, algae, plus fungi, that have of coals, or massive organic carbon accumulations in the capability of making bedrock and soil nutrients the pre-Devonian, potentially reflects as much the available to the later arriving tracheophytes. Al- absence of suitable conditions for accumulation as though today some angiosperms may be early rock anything else. WestŽ. 1990 makes it clear that today surface colonizers of isolated soil pockets generated there is significant organic carbon produced by lower by these earlier succession organisms, there is no autotrophs in semi-arid and arid regions, but that evidence that any Paleozoic tracheophytes now there are no reliable global summaries about rate of known were able to fulfill the role of these earlier production and magnitude of this organic carbon; successional organisms. However, the unique charac- until this situation is remedied it would be unwise to ter of tracheophyte-generated soils, presumably first dismiss lower autotrophs as a significant source of A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 59 organic carbon either today or in the past, particu- All these workers conclude that continental weather- larly owing to the widespread existence of dry re- ing wasris principally a function of rates of uplift, gions now and in the past. as well as implied high relief which most of them do Ultimately, bedrock weathering is the first step in not discuss, and glaciation that exposed fresh mineral soil formation; this is a role that we argue cannot be surfaces. assumed by tracheophytes with limited capacity for Many workers have advocated relief andror ele- bare bedrock colonization and physical bedrock de- vation as major factors in physical andror chemical struction. weathering and as central components in the global geochemical budget and climatic change. But generalizations about chemical and physical weathering cannot be made purely on the basis of differential 16. The relief and elevation effects relief and elevation alone without consideration of other interacting factors, including vegetation cover, Tectonic hypotheses depend ultimately on the and humidity and temperature and their duration continuous exposure to erosion of fresh mineral sur- over time for each set of climatic and physical faces related to various levels of uplift. Edmond et Ž.relief, elevation conditions. al.Ž. 1995, p. 3322 stress that this can only come One weathering item that may be relatively inde- about in Atectonically active areas of continually pendent of climate, topography or elevation concerns renewed reliefB not in low relief areas, regardless of sulfide oxidation. Sulfides, primarily pyrite, when elevation andror glaciationŽ Francois et al., 1993; exposed to moist air invariably oxidize, providing Raymo et al., 1988; Raymo, 1991; Raymo and Rud- sulfate ions that actively attack silicate minerals diman, 1993a,b; see Ruddiman, 1997, for additional among others. The presence of efflorescences of papers dealing with this question. as the principal varied sulfates, AalumsB beneath rock overhangs is factor in the acceleration and inhibition of silicate universal. Thus, a certain amount of Ca and Mg is weathering. Raymo and Ruddiman conclude that liberated from silicate minerals by this process, but mountain building and the accompanying exposure its overall importance is probably small.Ž see Alpers of fresh mineral surfaces has stronger effects than et al., 2000, for discussion of sulfide mineral weath- temperature on enhancing the alkalinity flux to the ering, including the commn biotic involvement, to oceans. Edmond et al.Ž. 1995 argue that random, but sulfates. . specific tectonic events are the major drivers of As an example of the importance of relief andror global weathering evidences, similar to the argu- elevation to weathering in different latitudes, the ments of Raymo and Ruddiman whom they cite. following scenarios consider the relation between SomeŽ. particularly Raymo and Ruddiman have chemical weathering and different relief, elevation, specifically raised the possibility of enhanced chemi- temperature and humidity: cal weathering of silicate minerals related to tectonic Ž.1 At low latitudes, evidence suggests that under activity involving uplift, suggesting a correlation be- conditions of high humidity and temperature in high tween purely physical erosion rates, measured by relief regions at low elevations there will be intense rates of mechanical erosion, and rates of chemical chemical weathering, leading to a deeply laterized weathering that contribute to solute influx. These soil profile. workers have based their conclusions largely on Ž.2 At low latitudes, under conditions of high Cenozoic examples, and on the Tibetan Plateau in humidity, high temperature, and low relief and low particular. elevation, chemical weathering will be far weaker, Uplift effects depend on pedogenic weathering of owing to the presence of a laterized soil profile. rock-forming minerals in the virtual absence of or- Ž.3 At low latitudes, under conditions of high ganic acids produced by living organisms or follow- elevation and low relief, regardless of humidity, but ing from the decomposition of organic matter. Impli- with lower temperatures owing to the higher eleva- cated in these effects is significant geochemical tion, there will be far lower levels of chemical weathering related to inorganic geological processes. weathering than in the higher temperature regions. 60 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Ž.4 At low latitudes, under conditions of high with little net consumption of atmospheric CO2 and elevation and high relief combined with high humid- little effect on the global CO2 budget. ity, but still lower temperature, there will be an even However, White and BlumŽ 1995, pp. 1729, 1743, lower level of chemical weathering. 1745. correlated variations in solute concentrations Ž.5 At temperate latitudes, under conditions of and ionic fluxes with temperature, precipitation, high relief at lower elevations, but low temperatures runoff and evapotranspiration for 68 globally dis- and high humidity, there will still be lower levels of tributed watersheds, and found no strong relationship chemical weathering, while at higher elevations between steep topography and high chemical weath- chemical weathering will be much reduced. ering rates ŽAimplied physical erosion rates do not Ž.6 At polar latitudes, at all elevations with either appear to have a dominant impact on chemical high or low relief, regardless of elevation, under weathering ratesB.. They considered mostly temper- conditions of low temperature but high humidity, ate areas, with little data from the tropics or polar there will be a low level of chemical weathering. regions, and concluded that precipitation and temper- To estimate the levels of chemical weathering one ature are important correlatives of weathering. In might expect through time, this information from the temperate area watershed studies, they found no present indicates the necessity of employing all the correlation between fluvial chemical fluxes and topo- resources of paleogeography, paleogeology and the graphic relief or extent of glaciation, implying that changing positions of ancient climate belts. It is not physical erosion alone does not have a critical influ- possible to extrapolate information about modern ence on chemical weathering rates. Thus, they write river water compositions, from the present back that lack of any correlation between chemical fluxes, through time and expect reliable weathering results. e.g. high chemical weathering rate, topographic re- Elevation and relief, by themselves are an inadequate lief or extent of recent glaciation, does not lend basis for understanding levels of chemical weather- support to the concept of tectonic control of global ing in the past, without any understanding of past climate or support the concept that high rates of global climatic belts and their ever changing posi- physical erosion Ahave a critical influence on chemi- tions. cal weathering rates.B Despite the relevance of their Edmond et al.Ž. 1995 provide an elegant example study, emphasizing the importance of temperature from the Guyana Shield, Venezuela, Colombia and and precipitation, the absence of significant data Brazil, demonstrating that relatively low relief, low from tropical and polar regions leaves open the elevation conditions in a tropical region do not yield possibility that there may be latitudinally correlated high rates of chemical or physical erosion. Thus, as distinctions. they discussŽ. p. 3321 the mix of environmental These conclusions are similar to those reached by conditions, tropical temperatures at low and interme- StrakhovŽ. 1967, p. 4 who argued that in regions of diate altitudes, high and strongly seasonal rainfall, rugged montane relief and rapid denudation, chemi- and extensive cover of dense forest appear to be cal weathering is almost entirely suppressed in com- strongly conducive to rapid denudation, but an esti- parison with physical weathering, even in very moist mation of weathering processes cannot be deter- climates. Thus, he argues that surface layers of rock mined from river composition, e.g. the relationship must not be removed so rapidly that change in the between elevation and Aweathering yieldB on this eluvium from an alkaline to an acid state which tectonically quiescent shield is not pronounced. Thus, promotes removal of soluble derivativesŽ chemical they argue that elevation per se is not the determin- weathering.Ž. , can occur. Strakhov concludes p. 4 : ing variable in weathering yield, but whether mecha- ARelief by itself is not . . . a significant factor; it nisms of elevation also generate relief. Only tecton- depends on epeirogenic movements. Rapid epeiro- ism that generates high local relief and steep slopes genic uplift gives rise to high, mountainous relief, leads to continual availability of fresh surfaces and suppressing chemical weathering; when epeirogeny rapid removal of weathered material while low relief is sluggish, scarcely perceptible, the landscape con- areas, regardless of elevation, have a deeply weath- sists of plains, which favor the most active chemical ered mantle and contribute little to fluvial runoff weathering.B Even where there is a highly favorable A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 61 combination of climatic factors, especially tempera- pedogenic calcretes account for much of the Ca ture and rainfall, chemical weathering can be entirely released during silicate mineral weathering. The large suppressed or fostered in different tectonic regimes. regions of low relief arid conditions present during StrakhovŽ. 1967, p.9 cautions against the transfer of the Phanerozoic suggest that, as with the low relief data from present day weathering to the geologic lateritic regions, there is little, long-term silicate past, but argues that strong epeirogeny has sup- mineral weathering involved. This suggests that most pressed chemical weathering in all earlier geological erosion takes place in high relief regions, whereas intervals. However, most of Strakhov’s data are from most mineral weathering takes place in both high higher latitude regions or from the lower reaches of relief and adjacent low relief regions. tropical rivers where the intense effects of higher elevation weathering under tropical conditions have been diluted by downstream waters. The overall generalization is that, as with so many factors, there 17. The paleogeographic effect is an overall latitudinal chemical weathering gradient at higher elevations, from the relatively high rates in tropical humid, warm regions to the much lower Worsley et al.Ž. 1994 have stressed the signifi- rates in the much cooler temperate to polar regions cance of different paleogeographies in affecting the of similar elevation. distribution of climate types globally, in part through Others have criticized these efforts because they the alteration of oceanic circulation, which in turn did not emphasize the importance of different levels will affect the distribution of weathering environ- of temperature and humidity, i.e. the weathering ments. Changing paleogeographies also involve difference between low latitude, tropical and high changing areas subject to weathering under differing latitude polar environments. To this, one should add climatic regimes, thus giving rise to different cation the effects of different global climatic gradients. fluxes in surface runoff. Those working only in the Another critical point in the tectonic uplift models Cenozoic need not be concerned with major changes is that high relief, montane areas of bare rock subject in global landmass positions, although oceanic gate- to weathering, lacking any evidence for lichen-based, ways need consideration. tracheophyte or other megascopic autotrophs, are A major implication for global weathering- largely limited to high montane regions which are climatic models that consider the implications of either very short-lived, or only moderately long-lived silicate weathering on atmospheric CO2 and temper- Ž.Himalayas, Alps, Andes , contributing very little to ature is that there is necessarily a relation between long-term atmospheric CO2 drawdown. silicate weathering and the area of landmass in tropi- Low relief regions, mostly at low elevations, sub- cal latitudes as contrasted with higher latitudesŽ White ject to tropical lateritic type weathering, once the and Blum, 1995, p. 1745. as implied by Otto-Bliesner geologically brief initial period of intense silicate Ž.1995 . Otto-Bliesner Ž. 1995 argues that changes mineral weathering has been completed, result in during the Phanerozoic in the size, elevation and mineral solution taking place at a very low rate global positions of landmasses affect continental owing to the very thick, underlying lateritic deposits weathering and changes in total runoff. Based on virtually lacking in undissolved silicate mineralsŽ see simulations using paleogeographic reconstructions for Edmond et al., 1995, for an excellent example. . the Phanerozoic she sees a feedback in weathering These low-relief, lateritic regions will contribute very rate between size and distribution of continents little to CO2 drawdown due to weathering. through time and surface temperatures, precipitation, Again, low relief arid climate regions present a evaporation, and runoff. Thus, she states that the problem because we have little understanding of how weathering feedback between temperature and runoff their soil biota, today or in the past, affect undis- may be largely dependent on land–ocean configura- solved silicate minerals as contrasted with the tropi- tion. In essence she agrees with Worsley et al. cal environments. Arid region soils are rich in smec- Ž.1994 about the importance of global paleogeogra- tite, indicating Mg retention, rather than loss, and phy in constraining weathering. 62 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

18. Tracheophytes and the soil effect dent and interacting variables as temperature and soil CO . On the other hand, without some attempt it is 18.1. What is soil? 2 difficult to know which one, or which ones, are most Soil is more than merely crushed, finer grained significant in mineral solubilization. bedrock. Disaggregated bedrock and partially disag- Our view, repeated again below, is that the first gregated bedrock, even those that have been subject step, presumably in the earlier Precambrian, from an to a certain amount of abiotic weathering are not inorganic regolith, Araw soilB, is the addition of soil. The term regolith, although commonly avoided organic matter contributed from lower autotrophs. in many but not all soil science texts, involves such The biotically influenced soil part of the ecosystem disaggregated bedrock materials that may have been came into existence by at least the Late Archaeozoic, also subjected to a low level of biotic activity, with a with the advent of cyanobacteria, and possibly other small content of organic material. The concept of micro-organisms. In support of this view, Watanabe soil involves the interaction of bedrock grains of et al.’sŽ. 2000 geochemical evidence for an ancient varying sizes with biotic materials autotrophic and Ž.2.3–2.7 Ga, Early Proterozoic organic carbon-rich heterotrophic. From this perspective, there is good paleosol is notable. By the earlier Middle Ordovi- evidence that soils have been generated by organ- cian, there were hepatic-like embryophytes that isms from well back in the Precambrian to the should have been capable of adding organic carbon present. For some, however, the concept of soil-gen- to the soils with their preexisting micro-organisms, erating organisms necessarily requires tracheophytes. possibly including free-living fungi. Much later in Geological evidence makes it clear that such a re- time, possibly by the mid-Silurian, the soil ecosys- strictive definition requiring tracheophyte involve- tem also had a rootless, possibly tracheophytic addi- ment is in errorŽ see Gutzmer and Beukes, 1998, for tion, possibly with free-living soil fungi that might a Precambrian bauxite based example. . have been intimately associated with possible tra- To determine the relative chemical weathering cheophytesŽ a potential first step in the ultimate effects that may have been imposed by the advent of establishment of a closer relationship between fungi tracheophytes and their effects on atmospheric com- and tracheophytes—the mycorrhizal relationship— position, among questions that need to be answered may have begun with the establishment of fungal are: colonies on plant surfaces. . Still later, possibly in the Ž.1 Are tracheophytes the principal AbrokersB of earlier Devonian, root-bearing tracheophytes were weathering, that is do they AmediateB by their pres- added. This is ultimately followed by the appearance ence, almost all biotically induced weathering? of tracheophytes, with their mycorrhizae that to- Ž.2 What is the relative importance of biotic and gether make up the rhizosphere, with its much high- abiotic chemical weathering? er organic carbon content. The evidence for my- Ž.3 Are soils the only or main site of chemical corrhizae before the Carboniferous is contentious weathering? Are there global differences in soil ŽPirozynski, 1981; Pirozynski and Dalpe, 1989; weathering related to differences in temperature and Pirozynski and Hawksworth, 1988a; Pirozynski and moisture, i.e. climatic control? Malloch, 1975. . If the rhizosphere is defined to mean Ž.4 Was there sufficient ground surface life to roots only, then it appears in the earlier Devonian have promoted biologically mediated chemical with the advent of true rooted tracheophytes, but if it weathering in surface regoliths analogous to that must involve symbiotic mycorrhizae, the rhizosphere seen in tracheophytic soils before the advent and may not have appeared until the Carboniferous. Later spread of tracheophytes? yet in the Devonian or Carboniferous, there is the Ž.5 What is the potential importance of abiotic definite addition of mycorrhizae to the soil ecosys- carbonic acid, that someŽ. cf. Holland, 1984, p. 181 tem, coevolved with tracheophytes. In one sense the have called the Amajor weathering acidB during rhizosphere did not exist prior to the Devonian ad- Amost, if not allB of Earth history? vent of roots; in the second sense the rhizosphere did It would seem difficult, if not impossible, to not exist until the advent of symbiotic mycorrhizae. independently evaluate such obviously interdepen- The supposed Early Devonian Rhynie Chert symbi- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 63 otic fungi are present in stems, not in rhizomes or Others have arrived at similar conclusionsŽ Keller true roots, although many have assumed that they and Wood, 1993. as we discuss below. were physiologically and functionally no different The Asoil effectB refers to mineral weathering, than the fungal symbionts of true roots. including that of silicate rocks, in the soil zone, It seems to be commonly assumed that most above bedrock. It is conventional to divide modern pre-tracheophytic rock and mineral weathering was soil into rhizosphere and bulk, or non-rhizosphere related entirely to mechanical stresses and abrasion, soils, which we define and discuss below. Most while chemical mineral weathering, in soilsrre- pedologists discussing soil weathering seldom make goliths now regarded as largely biological, was clear the distinction between rhizosphere and bulk largely abiotic. Although the rate of chemical weath- soils doubtless because as Campbell and Greaves ering for soils and bedrock may be different, we Ž.1990, p.24 note, there usually is no distinct bound- argue that bedrock weathering cannot be ignored as a ary between the two. Therefore, conclusions about factor in atmospheric models, pre-Devonian and oth- soil weathering are lumped under a generic soil erwise, particularly for the pre-Carboniferous where category. it seems unlikely that soils closely approximated Under natural, field conditions, it is virtually im- modern soils or were necessarily widely developed possible to consider rhizosphere and bulk soil weath- on a global scale. ering separately, even though many of the factors Thus, it is conceivable that there was a complete assumed in rhizosphere weathering are thought to evolutionary continuum from bedrock and raw Ain- have minimal effects in bulk weathering. They can organicB soil immediately derived from them, re- only be evaluated in combination with an attempt to golith to rhizosphere type soils. Additionally, it is determine the principal forcing agent on which the also clear that non-rhizosphere, lower autotrophic other variables may be ultimately dependent or their soils continue to the present in appropriate environ- relatiÕe importance in silicate weathering in the soil ments, chiefly arid paleosols and regoliths like those environment. that preceded the advent of the tracheophytes. The The rhizosphere is defined in terms of tracheo- physical weathering continuum has presumably been phyte roots and their symbiotic fungi, and therefore present since the earlier Precambrian, but as we will is necessarily concerned with both tracheophyte and stress and demonstrate below, there was also a con- fungal evolution, and the development of their coe- temporary biotic weathering continuum. volved symbiotic relations. Bulk soils are soils, even An increasing number of regoliths and immature in forest situations, that are outside the narrowly AsoilsB are now identified in the Precambrian and circumscribed zone that encompasses tracheophyte Early Paleozoic that show the effects of weathering roots and their mycorrhizal extensions. This implies such as laterites, including bauxites. We argue that that most soils will be bulk soil if roots are widely the weathering effects of pre-tracheophytic organ- spaced, and mycorrhizal extensions are limitedŽ cf. isms that inhabit bedrock surfacesŽ. epiliths as well April and Newton, 1992. . as organisms that live below the surface of rocks In the tracheophytic world, most chemical weath- Ž.such as endolithic lichens and cyanobacteria could ering of minerals in Anatural systemsB is assumed to have contributed sufficient organic material to these take place in soils, where solution pH is controlled regoliths to sustain populations of autotrophic, by pCO2 and the presence of biotically generated heterotrophic, and chemotrophic bacteria and hetero- organic and inorganic acids and other substances trophic fungi adequate to instigate weathering of the ŽKeller and Wood, 1993; Gwiazda and Broecker, type that we now find in soils where mineral solubi- 1994, p. 142. . lization is as much accomplished by the organic Weathering in the rhizosphere, the narrow zone of acids produced by these free-living organisms as it is soils subject to the influence of living rootsŽ April by root and mycorrhizal Aexudates.B All that is and Newton, 1992. is extremely complex. Brady needed to sustain these microorganisms, independent Ž.1991 and Gwiazda and Broecker Ž. 1994 note some of tracheophytes, is a source of organic carbon. of the principal factors implicated in the control and 64 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 inhibition of the silicate weathering rate in the tra- zosphere environment and its role, it is important to cheophytic rhizosphere. While various models have understand the complexities of this microenviron- stressed the differential importance of three or four ment and its organic components in attempting to factors, all of them ultimately are either directly evaluate its significance through time and in differ- tracheophyte-dependent or are related to other organ- ent biogeographic and climatic regions of the world. isms, whose biomass in the rhizosphere is ultimately tracheophyte-dependent, except temperature. Factors 18.1.1.1. What is the rhizosphere?. By definition, the that have claimed attention in silicate dissolution rhizosphere is that microhabitat adjacent to and in- include, among others, temperature, soil CO2 , soil volving living tracheophyte roots plus their associ- pH, soil moisture, and organic acidity generated by ated symbiotic fungi and free-living fungi and mi- root AexudatesB and soil microbiotaŽ Brady, 1991; croorganisms dependent on the readily available and Gwiazda and Broecker, 1994. . Only the last need continuous supply of organic carbon compounds de- directly implicate the tracheophytes. rived in part from roots. It also includes fungal or Below we discuss several of these possible Afor- mycorrhizal extensions out from the roots into the cing factorsB ŽBrady, 1991; Gwiazda and Broecker, soil that increase the surface area effective in ion 1994. that have been implicated in mineral dissolu- uptake from the soil. Typical discussions are pro- tion in soils and the overall significance attributed to vided by LynchŽ. 1990 and Curl and Truelove Ž. 1986 . each by different people. The first is biotic and Mycorrhizae, literally Afungus rootsB, is the name includes organic secretions from tracheophyte roots applied to Acomposite organsB formed by roots and and their mutualistic, symbiotic fungi emphasized by symbiotic fungi which have been suggested to exist Berner and others. The second, is soil temperature, in mutualist association with roots, although Smith potentially a totally abiotic factor. The third is CO2 and ReadŽ.Ž. 1997, p. 106 also Reid, 1990 caution in which biota may or may not play a role. Finally, that it is difficult to demonstrate that the mycorrhizal we discuss the role of organic secretions from free- symbiosis is mutualistic in the sense of benefiting living fungi and microbiota in pre-tracheophytic soils both organisms especially for VA mycorrhizae that which potentially may play a role in either tracheo- utilize a high proportion of the photosynthate of the phyte inhabited or bulk soils, which have been gen- host with some cost to the host. erally discredited or ignored by others. Obviously, In modern soils, mycorrhizae consist principally organic secretions from free-living fungi and micro- of VA mycorrhizaeŽ VAMs, sometimes generally biota may also play a role in tracheophyte-inhabited referred to as endomycorrhizae. and ectomycor- soils and in bulk soils. StotzkyŽ. 1986 has reviewed rhizae, together with several smaller categories not what is known about the complex relations between further discussed hereŽ see Harley and Smith, 1983; bacteria and virus with soil and soil components; this Reid, 1990; Smith and Read, 1997; Read, 1999, their is clearly a very complex and little understood area Fig. 1. . In VA mycorrhizae, the fungi are obligate owing in no small part to the basic complexity of biotrophs that cannot be cultured outside of the host soils and the difficulty of applying the results of root. They depend on the host since they lack free overly simplified laboratory experiments to the natu- saprophytic ability and cannot exploit as carbon ral world. BouwmanŽ. 1990 emphasizes how poorly source organic detritus of the soil. Caldwell et al. constrained are the many factors involved in modern Ž.2000 show that some dark-septate, root endophytes soils in terms of greenhouse gases; there is no reason can utilize many detrital and litter sources of soil to think that things were any simpler in the past. carbon, nitrogen and phosphorus. Trappe points outŽ. written communication, 2000 that AOf the several types of mycorrhizae, the vesicu- 18.1.1. The role of the rhizosphere lar–arbuscular type is the common one on herba- In the Berner models, atmospheric change can be ceous and many woody plants today. As it is possi- seen as a rhizosphere phenomenon since he regards bly the only type known from the Devonian fossil it as the site of all mineral weathering sufficient to record, it is the only one considered in the present have environmental impact. To comprehend the rhi- discussion. The fungi involved are in the Order A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 65

Glomales and related orders and are obligate host of free-living autotrophic and heterotrophic or- biotrophs: they cannot access carbon from dead organisms, varied animals and protists found at the ganic matter. The fungal mycelium enters receptive root surfaces and within a generally small Aradial feeder rootlets, where it grows between and into root distanceB of roots. High concentrationsŽ perhaps cortical cells, forming nutrient exchange structures highest. of microorganisms also occur in the organic termed arbuscules and storage cells termed vesicles. carbon-rich surface litter. Many of the free-living In this association, the tracheophyte host provides organisms depend nutritionally on organic com- photosynthates and growth regulators to the fungus. pounds released from roots and from decomposition The fungal mycelium, growing into the soil from the of sloughed off root cells and mucilage, a poly- root, in turn takes up minerals from the soil and saccharide, secreted by epidermal cellsŽ Campbell translocates them to the host root tissue. The fungal and Greaves, 1990. . Thus, the organic material de- mycelium thus acts as a geometrically and physio- rived from roots mycorrhizal fungi influences num- logically adept extension of the tracheophyte root bers, kinds and activities of virtually all microorgan- system.BAMineralB in the biologist’s sense refers to isms at the root surface and their biological activity solubilized ions derived from varied mineral species within a limited sphere immediately adjacent to roots. present as grains in the soil. This nutrient-rich, microbial-rich environment of the In this AmutualismB, roots supply photosynthate rhizosphere is a limited, circumscribed zone within a to support fungi, and the fungi absorb mineral nutri- few mm of the root surface and decreases rapidly ents from the soil and transport them by hyphae with distance. extending into the surrounding soil to tracheophyte The mycorrhizosphere is occupied by varied mi- host cellsŽ Wilcox, 1996; Allen, 1991; Safir, 1987; croorganisms including heterotrophic and au- Smith and Read, 1997; Waisel et al., 1996. . totrophic bacteria, microalgae, protozoans, micro- With the development of symbiotic root fungi, faunaŽ. nematodes and mites and free-living fungi. which act as root extensions into the substrate and Concentration in the rhizosphere depends on and is increase the potential area for action of fungal-gener- enhanced by the nutritional benefits derived from a ated enzymes, organic and inorganic acids in com- continuous supply of organic carbon and inorganic parison with tracheophyte roots, the soil rhizosphere compounds derived from living rootsŽ. exudates , came into being. TrappeŽ written communication, from surface litter, from sloughed off epidermal hairs 2000. points our that AThe rhizosphere of mycor- and cortical cells and so on, at the root surface rhizal plants is actually a Amycorrhizosphere,B influ- ŽSilverman and Munoz, 1970; Hattori, 1973; Bag- enced by exudates of both host and fungus.B This yaraj, 1984; Bennett and Casey, 1994; Curl and microhabitat is complex but narrowly circumscribed: Truelove, 1986, pp. 2 and 4; Campbell and Greaves, The rhizosphere represents a narrow mm Aradial 1990.Ž . But as noted by Smith and Read 1997, p. zoneB of the soil subject to the influence of living 240. , the ability of free-living rhizosphere fungi to roots and their mycorrhizal extensions; the rhizo- use organic carbon sources in the soil, such as lignin plane represents the root surface. Thus, narrowly or cellulose, has little relevance to nutritional re- circumscribed, the bulk of even forest soils, Alies quirements of symbiotic VAM fungi, which depend outside of the rhizosphereB ŽApril and Newton, 1992, upon carbon provided by current or recent photo- pp. 414–415.Ž . Smith and Read 1997, pp. 66–67, synthesis by the host plant, rather than from litter fall 253. indicate that it may be difficult in soils to or dead organisms. Smith and ReadŽ. 1997, p. 239 . determine the actual lengths of hyphae associated argue that the great mass of the soil does not contain with mycorrhizal fungi and their biomass from fun- abundant supplies of readily available carbon com- gal populations of soil saprophytes and root pounds since, as noted above, most of this is used up pathogens but suggest C flow to the soil via hyphae by the microbial populations—ca. 75% or more. well beyond the zone normally designated as rhizo- TheyŽ. 1997, pp. 105, 125, 239 et seq. explain that sphere. glomalean fungi have limited saprophytic ability and In addition to tracheophytes and their fungal sym- limited capacity to produce mycelia in the non- bionts, this complex environment is inhabited by a mycorrhizal state when they have no autotrophic 66 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 partner, thus they are dependent on the photosynthate According to Bazin et al.Ž. 1990, their Fig. 2 , produced by their host as much as 20% of which can different colonizing populations may also be in- be transferred to the fungus; in the mycorrhizal state, volved. In addition to obvious differences in pres- the mycelia network assures access to soil-derived ence or absence of rooted plants in rhizosphere and nutrients, some of which are passed to the host. non-rhizosphere soils, there is also a difference in Thus, roots influence the number, kinds and activ- their microbiotasŽ non-symbiotic fungi, bacteria, al- ities of virtually all groups of microflora common in gae, actinomycetes. , principally related to biomass of soil near the root surfaceŽ Curl and Truelove, 1986, these organisms in proximity to roots and litter, p. 6. . But the diversity and abundance of microor- rather than their diversity or presence or absence. ganisms in the rhizosphere also depends on the plant Species diversity as opposed to biomass concentra- type, soil nutrientŽ. humus contents among other tion, may actually be greatest in the root-free zone properties, and a variety of environmental factors Ž.Bazin et al., 1990, p. 104 . Curl and Truelove including temperature, pH and water among others Ž.1986, p. 94 say that it is also possible to have ŽBlock et al., 1994, p. 674; Hattori, 1973, p. 395, higher population numbers of microorganisms in their Figs. 8–18, shows variation in humic contents root-free soil. and bacterial numbers in various soils that clearly Bazin et al.Ž. 1990, their Fig. 1 say that decrease indicate differences in bacterial biomass; Curl and in population density comes commonly within a Truelove, 1986, p. 113 et seq.. . radial distance of 5 mm from the root with the actual distance ultimately dependent on the amount of car- 18.1.1.2. Bulk soil Õersus rhizosphere. Specifically in bonaceous material lost through the roots and dis- discussing soils, a distinction is invariably drawn tance traversed by root-derived diffusable metabo- between so-called AbulkB or non-rhizosphere soils, litesŽ. carbonates . Root-derived organic compounds sometimes referred to as the edaphosphere, and rhi- are classified according to their soil mobility. Only zosphere soils, the soil habitat in the immediate diffusible–volatile compounds move either rapidly vicinity of roots and their symbiotic fungi. The or even centimeter distances into the soil and accord- rhizosphere is sometimes defined as the volume of ing to Curl and TrueloveŽ. 1986, p. 52 may even soil influenced by rootsŽ Campbell and Greaves, affect the activities of microorganisms lying outside 1990. , sometimes as the narrow zone of the soil the rhizosphere; so-call diffusible, water-soluble subject to the influence of living roots and their compoundsŽ amino acids, organic acids, hormones, symbiotic fungi. The extent of the rhizosphere from flavonoids, enzymes and so on. , whose diffusion is the root surface varies from a few millimeters or at enhanced by soil–water contentŽ Curl and Truelove, most a few centimeters dependent on soil type, plant 1986, p. 127. , are metabolized by bacteria at the root species, age and many other factorsŽ Curl and Tru- surface and within a few millimeters into the rhizo- elove, 1986; Campbell and Greaves, 1990; Bazin et sphere. Thus, the outer limits of the rhizosphere will al., 1990. . be defined by the diffusible–volatile compounds. As a rule, there is a rapid fall-off in the popula- Bazin et al.Ž. 1990; cf. their Fig. 2 defined the outer tion density of heterotrophic microbes, especially limit of the rhizosphere as the point at which concen- bacteria, away from roots and mycorrhizal exten- trations of all microorganisms in the environment sions into the rhizosphere, as one approaches so- surrounding the root have returned to the level found called bulk or non-mycorrhizal root-free soilsŽ Hat- in an identical, root-free environment. tori, 1973, their Figs. 7–11, Tables 7-2 and 7-4; Increase in microbial biomass is correlated with Marschner and Romheld, 1996, p. 570; Curl and the increased source of organic carbon available in Truelove, 1986 Table 4.1, p. 103 and Table 4.3. p. the rhizosphere close to and at root surfaces making 105; Bagyaraj, 1984; Bazin et al., 1990, their Fig. 2. , the few mm of soil around roots Aa virtual cultural although most information on the Arhizosphere ef- medium for microbial growthB ŽCurl and Truelove, fectB does not provide information on biomass or 1986, p. 76. , and from litter fall and Acanopy dripB, weight of living microorganisms under these circum- the organic substances exuded from leaves. In ex- stancesŽ. Curl and Truelove, 1986, p. 95 . press comparisons between the rhizosphere and A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 67 non-rhizosphere, it is clear among other things that portant role in distribution of mycorrhizal symbioses sharp dropoff in microbial populations within micron and may selectively promote the establishment of or millimeter distance from roots provides some idea such symbioses. Leyval and BerthelinŽ. 1991 , of the diffusible distance that can be traveled by through in vitro experiments, also demonstrated that these acids. Only soil algae and cyanobacteria, typi- bacterial inoculation significantly promoted mycor- cally photoautotrophic, that appear not to be affected rhizal colonization in pine roots. by root exudatesŽ Curl and Truelove, 1986, p. 106, Ultimately, because the rhizosphere is so specifi- Table 4.3. run counter to this trend, with higher cally circumscribed to a very narrow zone subject to population abundanceŽ. densities in the non-mycor- the influence of living roots, even the bulk of forest rhizal, root-free, bulk soils than in mycorrhizal soils. soils lie outside the rhizosphereŽ April and Newton, Although soil microalgae may affect soil chemical 1992, p. 414. . and physical processes, there is only limited Acir- Estimates of mycorrhizal hyphae extent from the cumstantialB evidence that they play an active role in root into the substrate vary. But spread of arbuscular solubilization of mineralsŽ. Metting, 1990, p. 355 . mycorrhizaeŽ. AM from the root into the soil adja- Based on such information, it is assumed that the cent to the roots can form an extensive, highly rhizosphere is the site of most biochemical weather- branched network that penetrates for up to several ing of minerals and formation and precipitation of meters into the zone immediately above the water new minerals and mineral reduction. table, and for 4.6 m or deeper in arid soils where Thus, in tracheophyte-inhabited soils, the immedi- mycorrhizal hyphae by increasing the root length ate surfaces of roots and the microhabitats in the have greater exploring ability than the roots, thus immediate vicinity of roots are biologically active increasing significantly the potential area of mineral habitats occupied by metabolically active and abun- exposure to oxalic and other acids derived from the dant free-living microbial populationsŽ Pittman and fungusŽ. Varma, 1995, pp. 562–563, 566, 567, 571 . Lewan, 1994; Harley and Smith, 1983; Varma and ReidŽ. 1990, p. 290 notes that hyphae of ectomycor- Hock, 1995, p. 394; Drever and Vance, 1994. that rhizas in anthracite coal mine spoil can extend for include actinomycetes, fungi, algae and both hetero- meters. Cromack et al.Ž.Ž 1979 see also Graustein et trophic and autotrophic bacteria. Mycorrhizae have al., 1977; Griffiths et al., 1994. show something the effect of concentrating nutrients in the rooting about the oxalate generating capability of the mycor- regions as the result of activities during active stages rhizal fungi present in a temperate Douglas-fir forest. of life and from the local release of nutrients as they In most habitats, mycorrhizae fungal activity de- senesce and decomposeŽ Harley and Smith, 1983, p. creases significantly with root penetration into deeper 393. . soilsŽ. Trappe, oral communication, 2001 . Cronan Roots and mycorrhizae thus can be said to influ- Ž.1985 in his study indicates that weathering rates ence numbers, kinds and activities of virtually all are much higher in near surface soil horizons than in microorganisms at the rhizoplaneŽ. root surface and deeper layers. within a limited zone adjacent to the root in the As the outer limits of the rhizosphere, or the soil nearby rhizosphere so that the numbers of these zone influenced by living roots, are approached, the organisms are Ainvariably higherB than in the concentrations of all microorganisms are highly di- edaphosphere or Abulk soilB terms applied to soils minished compared with their numbers at the root beyond the influence of living rootsŽ Curl and Tru- surface and adjacent to the rootŽ Bazin et al., 1990, elove, 1986 pp. 4 and 6; April and Newton, 1992. . their Fig. 2; Curl and Truelove, 1986, pp. 4 and 6; At the same time, GarbayeŽ. 1994a,b has Paul and Clark, 1996.Ž . Campbell and Greaves 1990, proposedrdemonstrated that rhizospheric bacterial p. 141. say that micro-organisms tend to congregate populationsŽ. rhizobacteria play a seminal role in around root exudates or multiply during decomposi- mycorrhizal establishment by interacting with the tion of fresh organic matter. According to Harley and fungi during their free-living stage of development SmithŽ. 1983, pp. 5, 387 et seq. , both the rhizosphere before they find receptive entry points in the root. He population and most mycorrhizal fungiŽ specifically argues that such Ahelper bacteriaB may play an im- VAM. depend upon carbon provided by current or 68 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 recent photosynthesis by the host plant, rather than have to seriously rethink the whole topic of the carbonaceous detritus from litter fall or dead roots. effect on atmospheric CO2 played by soil weathering Biochemical or physiochemical differences in the organisms. rhizosphere compared with bulk soils include pH EhrlichŽ. 1998 emphasizes the importance of Ž.acidityralkalinity , moisture and nutrient status, free-living bacteria in mineral, including soil, weath- electrical conductivity and redox potentialŽ Lynch, eringŽ. see discussion below . 1990.Ž. . Campbell and Greaves 1990, p. 129 and Mineral weathering in the rhizosphere. But as the Marschner and RomheldŽ. 1996, p. 557 show that above discussion makes clear, the effective weather- the pH in the rhizosphere may differ by 1–2 units ing effects of the root microhabitat are for the most from that of root-free, bulk soil with a consequent part highly circumscribed and interrelated to a num- significant effect on bacteria, actinomycetes and some ber of variables that are either poorly known or fungi. which have not been effectively measured in most Moreover, because of the overlap in production of modern circumstances even within the modern inorganic acids and organic compounds of many biomes. To extrapolate modern weathering scenarios rhizosphere-inhabiting organisms, including tracheo- into the geologic past, into different vegetation types, phytes, it is impossible to determine how much phytogeographic and climatic circumstances is obvi- effective weathering is being perpetrated by any of ously fraught with uncertainties. these organisms independent of one another. The Organic materials potentially effective in control- above scenario may implicate tracheophytes princi- ling the mineral weathering rate in the rhizosphere pally as a source of organic carbon, certainly an are derived from a number of sources: tracheophyte effect not to be discounted if it accounts for the rootsŽ variously classified as excretions, exudates, increased biomass of other organisms, but does not secretions, leakages, mucilages and others. , canopy mean that they are necessarily the principal organ- drip, associated mycorrhiza and from associated mi- isms involved in mineral weathering. By implication, crobial organisms. . Some are the metabolic byprod- however, it largely relegates the potential for mineral ucts of living organisms; others are decomposition weathering by non-rhizosphere organisms to an in- products from decaying organic material or from significant independent role if significant weathering microbial activities on decomposing organic mate- activities can only take place within the microhabitat rial. created by tracheophyte roots in the rhizosphere. It Literature cited below discusses the biogeochemi- implies that non-rhizosphere organisms played no cal effects of various organic materials generated active, important role prior to the advent of the directly by tracheophytes and the possibly synergis- tracheophytes; it also implies that they play no ac- tic effects of their mycorrhizal fungi and a host of tive, important role at the present time. autotrophic and heterotrophic organisms in the rhizo- Under natural conditions, the relative importance sphere. The high biomass of microbial organisms at in soil weathering of free-living fungi and other and near the root surfaces serves to concentrate these microbiota have never been disentangled from the organic substances, depending on whether they are effects of roots and their mycorrhizae. Numerous diffusible or not, within the rhizosphere in close laboratory experiments demonstrate that bacteria and proximity to the roots. Low molecular-mass acids in other microbiota are independently effective in soils are rapidly produced and consumed by microor- weathering minerals, including silicates. Until the ganisms at or near roots so that there may be limited relative importance in soil weathering of these two penetration into adjacent substrata depending on categories are disentangled in the field, we will be whether they are diffusible or not; such acids are unable to understand their potential importance back highest in organic layers at the top of the soil profile into the geological record. It is possible, for exam- and decrease with depthŽ. Lewan and Pittman, 1994 . ple, that free-living microbiota were important in soil Mineral weathering in the rhizosphere is complex, weathering well before the advent of the tracheo- with a large number of interrelated variables that phytes and the mycorrhizae with which many of bear on weathering and weathering rates in different them are coevolved. If this were the case, one would soils. CronanŽ. 1985 points out, for example, that A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 69 weathering rates can vary 5-fold between acid forest soluble exudatesŽ e.g. sugars, amino acids, organic soils with different physical–chemical properties, in acids such as acetic, oxalic, fumaric, oxaloacetic, the same general geographic region in northeast North malic, humic and fulvic acids, phenolics, hormones AmericaŽ. New York and New Hampshire , with and vitamins that leak from roots without involve- different rates inversely correlated with mean soil ment of metabolic energy; secretions, or synthates, particle size and positively correlated with total ex- such as polymeric carbohydrates and enzymes that changeable bases in the soil profile. Weather- depend on metabolic processes; lysates released from ingrmineralization rates may also be distinctly dif- cell walls, sloughed whole cell when they autolyse; ferent between separate soil horizons within single and gases such as ethylene and CO2 and others. , are soil profilesŽ. Cronan, 1985, p. 185 . Plant- and mi- reviewed by WhippsŽ. 1990, p. 60 , Curl and Tru- crobially mediated processes such as the release of eloveŽ. 1986, pp. 52–92, Table 3.2 , Marschner and organic acids were found to be Asecondary to other RomheldŽ. 1996 , Bar-Yosef Ž. 1996 , Lewan and factors in controlling weatheringB Ž.p. 191 . Vegeta- PittmanŽ. 1994 , Stevenson Ž. 1994 , McKeague et al. tion types associated with these soils are oak-pitch Ž.1986 and Tan Ž. 1986 among others. pine, northern hardwood and white pine. It is instruc- Paul and ClarkŽ. 1996 argue that since microor- tive that the highest weathering rates occur in the ganisms make immediate use of these carbon sources, white pine forest, and the least in the oak-pitch pine quantification of losses must also include consequent forest. microbial biomass and microbial generated respira-

Soil and inter-soil variables, with potential effects tory CO2 . on weathering rates, include such things as biota, soil Because the types of chemical components and general rock types and source mineralogy, soil AleakedB or lost from rootsŽ variously referred to as textureŽ. grain size, etc. , structure and thickness, AexudatesB, excretions, leakage, secretions, among permeability-aeration, compaction, moisture and other terms: Curl and Truelove, 1986, p. 55. , and moisture residence time, root mass, deep or shallow metabolites synthesized or generated through mycorrhizal hyphal length, density and biomass in metabolic activities of roots, symbiotic fungi and soil, extent and intensity of colonization of the root non-symbiotic microbiota from root-generated mate- system by mycorrhizaeŽ see Smith and Read, 1997, rialsŽ sloughed off cellular material, senescing cells, pp. 68 and 71. , etc., biological release of organic etc.. and litter-fall, overlap in kind and amount con- acids; microbial nitrification, distance away from taining many of the same compounds, it is impossi- roots where these acids, etc., are found, soil pH, soil ble in field circumstances to differentiate between pCO2 and so on and a variety of broad physical mineral dissolution effects of tracheophytes, their variables such as climatic region, elevation, relief, symbiotic mycorrhizae, and the variety of non-sym- temperature, vegetation type, hydrology, etc. Gorham biotic organismsŽ bacteria, actinomycetes, fungi, al- et al.Ž. 1979, p. 59 also note the importance of the gae. that are commonly found in profusion in the parent rock material, whether composed of easily rhizosphere or even the source organisms from which weathered minerals or differentially weatherable organic acids might have been derivedŽ cf. Curl and minerals where chemical weathering may first accel- Truelove, 1986. . erate and then decline as resistance accumulates at How many of these organic substances can be the expense of more soluble minerals, and so on. attributed directly to roots rather than their mycor- Gorham et al.Ž. 1979, pp. 64 and 74 also note that rhizae and associated microbiota, all of which are the availability of weathering products to the local known to produce similar substances, is unclear, system will depend on the depth of unstable minerals particularly in natural field experiments where it is in the soil profile in relation to root distribution in virtually impossible to disentangle root exudates— the same profile thus both root mass and root pene- root organic matter—from microbial metabolites tration into the soil profile are important considera- which contain many of the same compounds found tions. in AexudatesB, and determine potentially superim- The types of organic and inorganic substances, posed, synergistic effects of the associated microor- e.g. rhizodeposition lost from roots, include water- ganisms on root growth and physiologyŽ Curl and 70 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Truelove, 1986, pp. 54, 55, 87. . Most information TrueloveŽ. 1986, p. 142 state that higher concentra- about root exudates comes from studies of plants tions of carbonic acid Aalong with microbially in- grown in nutrient solutions under sterile conditions duced organic acids and nitric or sulfuric acid, in- Ž.Curl and Truelove, 1986, p. 53 . Culturing experi- creases the solubilization of phosphates, ferric oxide, ments indicate that commonly similar substances are and manganese dioxide.B derived from roots, mycorrhizae and associated, In considering the comparative roles of ion uptake free-living microbial organisms. Even in laboratory from soil minerals by plant roots and fungal hyphae situations, because endomycorrhizaeŽ glomalean in mycorrhizae, CooperŽ. 1984, pp. 159–169 notes VAMs. cannot be cultivated independently of the that external fungal hyphae provide a more extensive host organisms, unlike ectomycorrhizal fungi, it is and better distributed absorbing surface than roots impossible to determine the source of the organic suggesting that mycorrhizal roots are more efficient acids involved in weathering effects. in weathering effects than roots alone. Insoluble ions like those in crystal lattices are not available to 18.1.1.3. The role of free-liÕing fungirmicrobiota in non-mycorrhizal roots, but there is also some ques- soil weathering. Non-symbiotic populations of the tion whether VAM can mobilize ions from these rhizosphere and rhizoplane that appear to be related same sources of insoluble ions. either directly or indirectly to the organic material Among these plant-generated substances are Aex- derived from roots are composed of bacteria, actino- udates,B organic and inorganic substances that AleakB mycetes, fungi and algae as well as a heterogeneous from tracheophyte roots without involvement of group of protists and animals that are primarily metabolic energy. Estimates of photosynthetic prod- bacterial and fungal feedersŽ see Curl and Truelove, ucts or organic carbon loss via root exudates and 1986. . These are typically fluctuating populations sloughed cells into the rhizosphere can run 40% or that diminish with distance from the root surface and higher and depending on the nature of the material that are in part responsive to growth-stimulating Ždiffusible–volatile, diffusible, water-soluble and AexudatesB derived from roots and in part to growth non-diffusible. affects its availability to microbiota factors synthesized by other organisms in the root ŽBazin et al., 1990, p. 104; Campbell and Greaves, zone. As Curl and TrueloveŽ. 1986, p. 113 state 1990. . Most of this is utilized by microorganisms; A . . . microbial populations are affected by many of only small amounts remain in the soil. the same factors that affect root exudation: age and The amounts and proportions of exudates vary developmental state of plant, soil type and treatment, considerably under different circumstances and can environmental factorsŽ light, temperature, moisture, be enhanced or increased by microorganisms and by pH. , foliar application of chemical, and microbial various forms of AstressB ŽCurl and Truelove, 1986, interactions.B p. 53; Marschner and Romheld, 1996, p. 567. . Root exudates vary in kind and amount dependent on such 18.1.1.4. The role of tracheophyte roots, their symbi- variables as plant ontogeny, plant species and otic fungi and other microorganisms in weathering. physiological condition and with varied abiotic con- Biochemical weathering in the rhizosphere appears ditions including temperature, soil structure, soil aer- to be due to varied organic substances produced ation and water contentŽ Curl and Truelove, 1986; during the decomposition of soil organisms and to Campbell and Greaves, 1990. . reactions with acids and organic compounds secreted The difficulty of assessing the nature and quantity by soil organisms. Metabolic respiration by soil mi- of exudates in the natural soil environment can be croorganisms also contributes to increased CO2 con- attributed to the absorption and utilization of exudate centration in the rhizosphere with the increased compounds by microorganisms, and the synthesis of potential for enhanced mineral solubilization by car- microbial metabolites in the rhizosphereŽ Curl and bonic acid among other organic and inorganic acids. TrueloveŽ. 1986, p. 88 . Curl and Truelove Ž 1986, pp.

Decomposition of organic matter also elevates CO2 89 and 91. note that it is difficult to distinguish and carbonic acid concentration. Thus, Curl and between exudate compounds derived from mycor- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 71 rhizal roots and volatilernon-volatile compounds enhanced biomass of microbes and is in fact affected synthesized and released by the fungal symbiont or by root and mycorrhizal exudates, then there should from the degradation products of sloughed root cells. be increased mineral solubilization closely associated In some literature, the lion’s share of mineral with the rhizosphere. Insofar as the biomass of or- solution effects are attributed to the associated free- ganisms affects various chemical reactions including living microbiota, rather than to the effects of root or silicate mineral weathering, such reactions will mycorrhizal-generated organics. clearly be faster in the rhizosphere, particularly in All of these organisms—tracheophyte roots, myc- the immediate vicinity of roots, and their mycor- orrhizae and microbiota—are also assumed to have rhizal extensions into the soil than in non-mycorr- some physical, e.g. mechanical non-biological effect, hizal soils. on minerals, such as breakage or fracturing, realign- Leyval and BerthelinŽ. 1991 note that while the ment, fragmentation, etc. Physical effects assume ability of plant roots and soil microorganisms in the importance because they expose fresh surfaces to rhizosphere to weather minerals is reported in a weathering andror increase the area of mineral sur- number of studies, few studies have attempted to face exposed to weathering but not always with distinguish plant from microbial effects or the com- expected resultsŽ. April and Newton, 1992, p. 416 . bined synergistic effects of mycorrhizae and bacteria Dissolution of silicate and carbonate minerals by in weathering. In vitro greenhouse studies of mica organic acids has been demonstrated in controlled Ž.phlogopite weathering by inoculated and non-inoc- experiments in laboratory settingsŽ Pittman and ulated pine roots and rhizospheric microorganisms Lewan, 1994; White and Brantley, 1995; Banfield over a 2-year period, showed that the total amounts and Nealson, 1997; April and Newton, 1992. have of acids were almost twice the level for non-inoc- shown that minerals show weathering effects with ulated plants and thus to confirm previous observa- organic acidsŽ. pitting, dissolution, etc. where root tionsŽ. cf. Robert and Berthelin, 1986 that mineral exudates, mycorrhiza-generated acids and microbial weathering occurs in microsites where bacteria are acids are segregatedŽ Barker and Banfield, 1996, associated intimately with minerals and related to 1998. . their increased production of organic acids. They, Gwiazda and BroeckerŽ. 1994 also conclude that and Fritz et al.Ž. 1994 , documented a spatial correla- Aif a more accurate evaluation is to be done about tion between rhizosphere bacteria around pine roots the role of organic acids as driving agents of chemi- and areas of highest K depletion in micas. cal weathering, more data isŽ. sic needed on low OchsŽ. 1996 reports, based on laboratory weather- molecular weight organic acid abundance next to ing experiments, that exudates from non-mycorrhizal mineral surfaces, on surface absorption of organic spruce roots and humic substances have no or lim- ligands, and on aluminum–organic anions complexa- ited affect on weathering rates compared with the tion in soils. At face value, the results of this study effects of low molecular weight organic ligands pro- indicate that organic acid variation plays a very duced by mycorrhizal spruce roots thus putting more minor role in enhancing or inhibiting the weathering burden on fungal than root exudates. rate. Despite an increment in the total amount of Harley and SmithŽ. 1983, pp. 5, 387 et seq. also organic acids in solution, the formation of organic– argue that the microbial population and most mycor- aluminum complexes prevents the increase in free rhizal fungi depend on carbon provided by current or organic anions, a necessary condition for the acceler- recent photosynthesis by the host plant, rather than ation of the organic ligand promoted dissolution rate carbonaceous detritus from litter fall or dead roots. and for the decrease of soil pH.B However, saprobic fungi and other organisms make Nevertheless, in the soil profile, there is some use of such dead material and litterŽ see Jennings, direct evidence for enhanced weathering associated 1995. . with these organic products and with roots particu- While it is clear that the concentration of organic larly in comparison with bulk, non-mycorrhizal soils acids is enhanced in the rhizosphere compared with and it seems safe to conclude that if the incidence of Abulk soils,B the role of organic acids in silicate acid production, etc., is closely correlated with the mineral dissolution in field studies is less clear in the 72 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 absence of compelling studies. Thus, Lewan and zosphere soils, with ambiguous results as noted be- PittmanŽ. 1994, p. 11 write: AThere is considerable low. controversy regarding whether organic acids at natu- VilesŽ. 1990, p. 8 notes the paradox that while it ral concentrations significantly accelerate the rate of is always assumed that plant roots have an impact on dissolution of primary silicate minerals. It is not weathering, there has actually been little serious likely that organic acids are important in the weath- study of whether roots do in fact instigate rock ering of granitic rocks, but they may be important in weathering and what processes might be involved. the weathering of more mafic rocks. Organic acid YatsuŽ. 1988 even questions how much we gen- concentrations may be significantly higher in mi- uinely know about the Awedging effectsB of growing croenvironments surrounding rootlets or fungal hy- plant roots. Although the physical effectiveness of phae than in bulk soil solutionsB. Lewan and Pittman roots in widening Arock cracksB and prizing apart Ž.1994 review the literature indicating that organic joints is always assumed, it is not always clear which acids have the potential to dissolve minerals in soil is the cause, Athe existence of the joint or the and that organic acids derived from kerogen in source growing plant root.B He argues that data Aon axial rocks can dissolve carbonate and silicate minerals in and radial pressures of growing roots of trees mea- subsurface sandstones and complex metals. sured at the apical meristem, regions of enlargement, April and NewtonŽ.Ž 1992 as cited in Gwiazda and regions of maturation of roots as well as at the and Broecker, 1994, p. 152. conclude that roots lower parts of stems, are urgently required to solve Ahave more than a passive role in mineral dissolu- the problem.B tionB as evidenced by differences in clay mineralogy It has been argued that roots tend to follow lines next to roots compared with in bulk soils; by appar- of least resistanceŽ. Drew, 1990, p. 39 or they tend ent mechanical effects of roots on mineral grains, to follow large gaps or channels between soil parti- that include fracturing and exposing fresh surfaces to clesŽ. Curl and Truelove, 1986, p. 50 pathways in weathering; fracturing and bending phyllosilicates, the soil of least resistance to growth—grow selec- and similarities in shapes of roots and adjoining tively along channels which are much wider than the mineral grains suggesting preferential dissolution; diameter of the rootŽ. p. 50 . There is no evidence sometimes there is a similarity between the root that roots penetrate smooth rock surfaces. Trappe shape and adjacent mineral shape suggesting prefer- Ž.written communication, 2000 notes that mycor- ential dissolution. rhizal mycelium can enter tiny cracks. RootsŽ as all While experimental evidence clearly demonstrates litter. have significant indirect effects on chemical dissolution of minerals related to organic acids, ex- weathering through their decomposition by bacteria, perimental evidence creates artificial circumstances. actinomycetesŽ. filamentous bacteria and fungi when Whether this information can be extrapolated to field acids are released or modified in the decomposition circumstances remains uncertain because of the vari- processŽ Yatsu, 1988; see also Boyle and Voigt, ety of unnatural, e.g. artificial circumstances neces- 1973.Ž. as Berner 1992 has also recognized. But the sarily created in the laboratory settingŽ White and important role here belongs to the associated mi- Brantley, 1995. . croorganisms, not to the roots. Berner’s emphasis on There are few direct field studies where both tracheophyte roots might give the impression that biological and physical effects have been measured microorganisms are wholly ineffective weathering and none where the effects of different organisms agents until they attach themselves to roots. can be sufficiently differentiated to determine whe- The real crux of the argument, vis-a-vis weather- ther the effects of these organisms are genuinely ing effectiveness of tracheophytes versus non- synergistic or whether the microbiota, for example, tracheophytic organisms, seems to be the specific are the most effective of these organisms in mineral microenvironmental circumstances created by tra- weathering. The April and NewtonŽ. 1992 study is cheophyte roots and their associated microbiota. Al- among the few studies attempting to discriminate though BernerŽ. 1992 promoted the direct activities comparative weathering effects within bulk and rhi- of organic acids from root AexudatesB in rock weath- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 73 ering, he also noted roles for Asecretions by mycor- for in response to the particular requirements of the rhizalŽ. sic symbiontsB and Amicrobial decay of plant biome. Thus, Smith and ReadŽ. 1997 note that plants litter,B entangling the weathering roles of varied in mineral soils in nival alpine zones and in high organisms. Arctic and Antarctic latitudes are non-mycorrhizal StevensonŽ. 1967 and Stevenson and Fitch Ž 1986, even though the same taxa may be mycorrhizal in p. 50.Ž note that a high proportion of bacteria A other biomes. TrappeŽ. 1987a,b, Table 4 notes the Gram-negative short rodsB. in soil are capable of sizable proportion of dicots and monocots in synthesizing 2-ketogluconic acids that are effective Arctic–Alpine habitats that are facultative or non- in dissolving phosphates and silicates. Stevenson mycorrhizal. Smith and ReadŽ. 1997 note the preva- Ž.1967 cites the high proportion of bacteria in soil as lence of ectomycorrhizal plants in boreal and tem- well as those living on rock surfaces that produce perate forest zones, although local communities on 2-ketogluconic acid. Bennett and CaseyŽ 1994, p. certain substrates may be VA mycorrhizal. APlant 188. state that ACitric, oxalic, and 2-ketogluconic communities . . . dominated by trees or herbaceous acids have been identified in soils and interstitial plants . . . colonized by VA mycorrhizal fungi pro- waters, and it is expected that high concentrations of gressively replace those with ectomycorrhizal roots these organic acids may build up around bacteria on a global scale as, with decreasing latitude, mean adhering to a mineral surface.B Several bacteria and annual temperatures and evapo-transpiration rates fungi produce siderophores to enhance Fe solubility increase . . . B Ž.Smith and Read, 1997, p. 432 . In Ž.in Drever and Vance, 1994, p. 144 . Webley et al. tropical biomes trees with VAM colonization pre- Ž.1963 and Duff et al. Ž. 1963 showed that organic dominate in wet and seasonally wet tropical forests acids produced by bacteria enhance mineral weather- although communities dominated by ectomycorrhizal ing in soils. species are characteristically restricted to the most Bennett and CaseyŽ 1994, pp. 188–195, and cit- nutrient-poor soils with surface accumulations of ing references. note that where a carbon substrate is litter and raw humusŽ p. 445, see Read, 1984, whom available for metabolic processes, bacteria rapidly they cite.Ž . Janos 1987, pp. 119–120 . notes the colonize fresh mineral surfaces and the organic acids prevalence of VA mycorrhizae in lowland, humid that result from degradation of this carbon source tropical ecosystems in forest trees, but less frequent can Adeeply weatherB quartz and feldspar grains in or low intensity or no infection, in seral and herba- as brief a time as 14 months, with greatest weather- ceous species from a number of global regions and ing associated with high bacterial density. AMicrobes no available information for forest understory vege- may be an important component of the organic-en- tation. However, non-mycorrhizal species in a forest hanced weathering scenario, as they may produce in Brazil included 25 species among 86 examined, deep chemical gradients near microbial colonies and accounting for almost one quarter of all species Aby high concentrations of dissolved organic acidB Žp. importance value.B In lowland dry tropical forests, 195. . VA mycorrhizae are also commonŽ. Janos, 1987 . VarmaŽ. 1999 notes the facultative mycorrhizal 18.1.1.5. Climatic correlates with biomes in Õaried status of most natural vegetation in desert, arid envi- climates. Smith and ReadŽ 1997; Read, 1993; Brun- ronmentsŽ Table 5, cactus, Table 6, grasses, some of drett, 1991. discuss variability in mycorrhizal popu- which are non-mycorrhizal.Ž . Varma 1999, p. 543 . lations in plants in different ecosystemsrbiomes. also notes that dominant pioneering plants are com- This variability has possible consequences in consid- monly facultative AM mycorrhizal in sand dune ering the presence of this relationship in the past and areas in western India, with low nutrients and or- potential for weathering in certain kinds of envi- ganic material. Coastal sand dunes and volcanic ronments that may have previously been more terrain pioneers are similarly facultativeŽ Allen and widespread than at present. Not just climate but Allen, 1990, as cited in Varma, 1999. while high common soil types play a major role so that in available nutrient levels and saturated soils Aoften different biomes distinctive symbioses fulfill multi- inhibit mycorrhizal symbiosisB ŽVarma, 1999, p. functional roles, some of which appear to be selected 544. . 74 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Clearly, not all tracheophytes are mycorrhizal and attributed to carbonic acidŽ H23 CO , an aqueous solu- some may form mycorrhizae in some environments tion of CO2 . . but not others. Extrapolating any generalizations However, the flux of soil CO2 and its absolute about mycorrhizae into the distant past is clearly amount vary considerably from biome to biome hazardous. Ž.Raich and Schlesinger, 1992 related both to climate and vegetation. Deserts are at the low end and wet tropics at the high end, with temperate regions in 18.1.2. The role of temperature in chemical weather- between, related to net primary productivity. Soil ing respiration also varies positively with temperature, Some emphasize the role of temperature as the precipitation, and soil water vapor, generally on a principal, or dominant factor in silicate mineral dis- seasonal basis. Soil respiration overall is derived solution, since changes in pH and temperature from all soil carbon sources, living and dead, plant strongly affect silicate dissolution ratesŽ Brady, 1991; and animal. One assumes that the weathering effect Brady and Carroll, 1994; Brady and Zachara, 1996. . of dissolved soil CO will be a function of its overall Thus, BradyŽ. 1991, p. 18,105 argues Aon the basis 2 abundance relative to accumulated labile organic of experimental weathering rates for a stronger matter, the availability of unweathered silicate min- weathering temperature dependency in global carbon erals, as well as of temperature and precipitation and cycle models.B Higher temperatures may lead as well that the carbonic acid flux will vary with exhaustion to higher organic productivity, which in turn could of labile organic matterŽ. Gorham et al., 1979 . Over Ž.or potentially could affect both soil pCO and soil 2 short time spans, soil CO is linked to organic organic acids, in addition to an acceleration of the 2 productivity; on a geologic time scale, soil CO will reaction rate and an acceleration of the global water 2 vary as well with atmospheric CO , though soil cycle that could lead to a higher alkalinity flux from 2 thickness and moisture and plant evolution will also the continents into the oceanŽ Brady and Carroll, play a partŽ. Brady and Zachara, 1996, p. 335 the 1994; Gwiazda and Broecker, 1994, and references latter related to differences in productivity and struc- they cite. . Thus, temperature could become the sig- ture. nificant forcing mechanism in silicate weathering in Moreover, considerable differences of opinion ex- the rhizosphere through a number of secondary ef- ists about the relative importance of the direct effects fects. However, these conclusions refer only to re- of CO on silicate weathering. Brady and House gions where the soil is high in unweathered silicate 2 Ž.Ž1996, their Fig. 24, p. 259 also Brady and Carroll, minerals, i.e. not in the widespread, high biomass, 1994. , for example, conclude from laboratory experi- low relief tropical climatic zones where mineral ments, that dissolution rates for augite and anorthite weathering is very lowŽ. the lateritic regions . Most exposed to high CO solutions dissolve at essentially of this weathering information about temperature and 2 the same rate as they dissolved in solutions relatively soil refers to work in the Temperate, particularly the depleted of CO . Brady and CarrollŽ. 1994 argue North Temperate Zone, which is not very widespread 2 that silicate mineral weathering is not directly af- compared to the tropical. fected by soil CO2 , but is extremely sensitive to temperature. They see the effects of soil CO2 on 18.1.3. The role of CO22( pCO -atmospheric and CO 2mineral weathering as indirectly related to its effects partial pressure) in chemical weathering on increased organic activity and the production of

In tracheophyte-inhabited soils, where the CO2 corrosive organic acids. Welch and UllmanŽ. 1993 content of pore space air generated through the note that although carbonic acid as well as other respiration of soil microorganisms, plant roots and inorganic acids such as nitric and sulphuric were at aerobic decomposition of all soil organic matterŽ Curl one time considered responsible for most mineral and Truelove, 1986, p. 4. is 10–100 times the atmo- weathering, organic acids are far more effective. spheric CO2 contentŽ Brady and Carroll, 1994, citing Still, one wonders whether inorganic acids such as Holland, 1978; Holland et al., 1986; Keller and nitric, sulphuric and phosphoric, might not be signif- Wood, 1993. , chemical mineral dissolution can be icant in some situations. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 75

Because of differences of opinion about the rela- and organic respiration, and thus closely related to a tive importance of carbonic and organic acidsŽ Tan, nutritional source provided by tracheophytesŽ Hol- 1986. in soil silicate mineral weathering, widely land et al., 1986; Keller and Wood, 1993. . different opinions also exist about the relative rates In modern soils, the biogenesis of CO2 is princi- through geologic time of silicate mineral dissolution pally dependent on respiring organisms that utilize and the rates of return of atmospheric CO2 to the nutrients such as organic carbonŽ the biogenesis of Earth’s surface and the ocean. SomeŽ cf. Holland et CO2 in soil requires respiring organisms and nutri- al., 1986, p. 32; Holland, 1984, as cited in Keller and ents such as organic carbon: Keller and Wood, 1993. . Wood, 1993; see also Cawley et al., 1969. attribut- Thus, according to this scenario, the increase in soil ing mineral dissolution largely to carbonic acid microbial activity that is believed to have accompa- formed from atmospheric CO2 dissolved in rainwater nied the appearance of the tracheophytes, related to that combines with soil CO2 , argue that silicate increased litter fall, is concluded to have been ac- mineral weathering has remained relatively constant companied by an increase in soil CO2 generation. during much of Earth history, with correspondingly According to this scenario, an increase of in situ relatively minor perturbations in atmospheric CO2 . biogenically generated soil CO2 means that micro- However, until this sweeping generalization is tested bially generated CO22 replaced atmospheric CO in against a global paleogeological analysis concerning soils as the primary agent of chemical weathering possible production rates of weatherable silicate min- ŽKeller and Wood, 1993, p. 223; Holland et al., erals one will be uncertain about its validity. 1986. but may not have greatly enhanced the silicate

Because pore water pH is controlled by pCO2 weathering rate or necessarily increased soil CO2 and organic acids, some see pCO2 in soils, where content. CO2 in aqueous solution forms a weak acidŽ carbonic However, Brady and CarrollŽ. 1994, p. 1854 em- A acid. , as the most important control on soil mineral phasize that CO2 does not directly control silicate weathering rates. Thus, it has been argued that CO2 weathering. Any acceleration of silicate weathering in soils controls the mineral weathering rate and that by soil CO22 must therefore arise through CO fertil- quantitatively, carbonic acid is now and always has ization of organic activity. Increased organic activity been the dominant weathering acid although organic should result in increased production of corrosive acids may play a significant part in weathering today organic acids and a higher overall rate of silicate Ž.Holland et al., 1986; also Holland, 1984; Tan, 1986 . weathering . . . B They also note, based on experimen-

CO22 derived from CO dissolved in rainwater is a tal evidence, that decrease in soil pH brought about weathering agent for soils and bedrock potentially by increased pCO2 would have minimal indirect free from the presence of surface vegetation. It has effect on organic free weathering in near neutral soil been suggested, that while the mechanisms of how solutions.

CO2 gets into soils have changed over geologic As Keller and WoodŽ. 1993 point out, the in- time, that an increase in atmospheric CO2 before the crease in microbial activity in the rhizosphere that is advent of higher plants could account for increased believed to have accompanied the rise of the tracheo-

CO2 in the soil that would partially offset the high phytes is also commonly believed to have increased pCO2 in vegetation-occupied soils, so that the over- CO2 generation in the soilŽ which as noted above is all silicate weathering rate need not have changed anywhere from 10–100 times that of the atmosphere. , appreciably since the mid-ProterozoicŽ Holland et so that soil CO22 replaced atmospheric CO as the al., 1986; Holland, 1984. . primary agent of chemical weathering. However, Prior to evidence for large-scale ground surface they show that the subsoil vadose zone, the aerated vegetation and significant associated soil microbial region above the water table, exerts a strong influ- activity, Holland et al.Ž. 1986 and Holland Ž. 1984 ence on the CO2 concentration to which runoff is argue that atmospheric CO2 entered regoliths dis- exposed as it percolates beneath the soil. In the solved in rainwater and through atmospheric gaseous vadose zone, microbial respiration is presently sus- diffusion. In contrast, soil CO2 is now microbially tained largely by nutrients generated at the land generated by bacterial oxidation of organic matter surface by tracheophytes, but as they suggest, pretra- 76 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 cheophytes could have generated a flux of organic bedrock and ArawB soils to the rhizosphere associ- material that would also have been available in the ated with tracheophytes and their symbiotic fungi. subsurface because pre-tracheophytic soils had high We have also made the distinction between the permeability and infiltration capacity related to thin- weathering of surface bedrock, i.e. potential soil ness and sandy and rocky character. Thus, they progenitor, and vadose zone bedrock weathering. It argue, as we have above, that even small amounts of seems to be commonly assumed that most pre- surface life may have been adequate to sustain mi- tracheophytic weathering of AsoilsB was related crobial life possibly supplemented by carbonaceous largelyrentirely to mechanical stresses and abrasion, material generated in marine waters that occasionally and inorganic rather than biotically mediated chemi- may have covered land surfaces. cal solution. Microbially mediated bedrock weather- In addition to microbially generated organic acids ing at the Earth’s surface is commonly relegated to a which would have enhanced weathering rates, this comparatively minor role, although the destructive infiltrated organic matter could also have supported role of epi- and endolithic bacteriaŽ. cyanobacteria , microbial colonization of mineral surfaces and Athe algae, fungi and lichens in the uppermost littoral development of microreaction zones in which silicate Ž.intertidal and supratidal zone of limestone coasts, dissolution has been observed to be concentrated in has long been regarded as significant and widespread aquifer sediments . . . B ŽKeller and Wood, 1993, p. Ž.Schneider, 1976 . 223.Ž. . As Keller and Wood 1993 point out, the Although the rate of biochemical weathering for increase in microbial activity in the rhizosphere that soils and bedrock and everything in between may be is believed to have accompanied the rise of the different, we argue that bedrock weathering cannot tracheophytes is also commonly believed to have be ignored as a factor in atmospheric models particu- increased CO2 generation in the soilŽ which as noted larly for the pre-Carboniferous where it seems un- above is anywhere from 10 to 100 times that of the likely that soils closely approximate modern soils. It atmospheric concentration; citing Holland, 1984. is also worth considering in view of the rock weath- without the necessity to invoke high atmospheric ering capability of the modern microbiota that they

CO2 to explain this apparent constancy. They argue may also be capable of considerable weathering of that even small vadose respiration rates Acould have soil mineral grains; all of this, of course, requires an charged early terrestrial runoff with carbonic acid available carbon source, with the exception of such and weathered mineral mass to levels similar to things as cyanobacteriaŽ. also including the lichens post-Silurian levels, even if atmospheric pCO2 val- which may have been the carbon source from the ues were as small as today’sB Ž.p. 224 . They argue very beginning. that a pre-Silurian vadose weathering agency is theo- April and NewtonŽ. 1992, pp. 378–379 state that retically possible and that it has some present day minerals in soils are particularly susceptible to chem- analoguesŽ. Schwartzman and Volk, 1989, p. 457 . ical weathering in the Atransitional boundary between Earth and its atmosphereB where weathering is driven Aby the confluence of atmospheric, hydro- 19. The surface bedrock and vadose zone effect logic, biological and geological processes.B This is It is common to disregard the effects of mineral no less true for bedrock weathering discussed below. weathering of bedrock and discount bedrock weath- Most geologists, including BernerŽ. 1991b, 1992 , ering as playing any role in significant chemical assume minimal or no rock weathering by micro- weathering that might provide a sink for atmospheric biota and do not discuss or mention the observational

CO2 . Bedrock weathering and the potential impact and experimental evidence cited in many publica- of bedrock weathering on atmospheric CO2 , as well tions that have appeared during past decades, on the as the weathering potential of non-tracheophytic soils, weathering effectiveness of varied microbial and are issues that we raise here. other organisms that support the importance of non- We have polarized this discussion by drawing a tracheophyte biodegradation of mineral substrates sharp distinction between AsoilsB and AbedrockB. and the precipitation and formation of new minerals There is, of course, a complete continuum from Žfor example, Krumbein, 1968; Syers and Iskandar, A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 77

1973; Ehrlich, 1995, 1998; Barker et al., 1997, 1998; cations in the substrate to form soluble organo- Berthelin, 1983, 1988; Eckhardt, 1985; Yatsu, 1988; metallic compounds. YatsuŽ. 1988, p. 330 , for exam- Leyval and Berthelin, 1991; Viles, 1990, 1995; Wal- ple, notes that the role played by prokaryotic and ton, 1993; Viles and Pentecost, 1994; Piervittori et eukaryotic microorganisms in alteration of rocks and al., 1994, 1996; Hirsch et al., 1995a,b; Brady et al., modification of minerals has not been Asufficiently 1999; Riding and Awramik, 2000; Urzi et al., 1993. . appreciatedB; VilesŽ. 1990, p. 11 states that lower Wessels and KappenŽ. 1993 , working in semi-arid plants and microorganisms Aare particularly impor- South Africa, make it clear that data on the atmo- tant to weathering processesB; and WaltonŽ. 1993 spheric carbon dioxide consuming abilities of varied remarks that Abiological weathering by cryptogams lichens may be obtained in the field, as well as in the is now an established fact, involving both biogeo- laboratory, while referring to similar information chemical and biogeophysical processes.B Hirsch et obtained elsewhere. al.Ž. 1995b point out the variety of rock types Ž sand- Most geologists speculating about Aweathering stone, granite, gneiss, amphibolite, limestone, basalt, scenarios,B arrive at the conclusion that non-trac- dolerite. as well as the variety of AartificialB stones heophytic organisms have no greater weathering ef- Ž.bricks, glass affected by microorganisms Ž fila- fect than abiotic physical and chemical reactions. mentous fungi, microcolonial fungi, yeasts, and aero- Although it is easy to make such an assumption, bic, facultative anaerobic heterotrophic bacteria: biodeterioration of church walls, gravestones and Ehrlich, 1995, p. 134; 1998. and phototrophic bacte- ancient monuments in Spain, Italy and Greece, and ria and algae, that live on or within rocks. Hirsch et elsewhere, by heterotrophic and phototrophic bacte- al.Ž. 1995b note that these organisms affect the ria, lichens, filamentous and microcolonial fungi, weathering process not only in their physiologically algae, and epilithic mosses are of sufficient concern active state but in a wide variety of ways: formation to conservationists, to suggest that rock weathering of chelating compounds; formation of mineral and effects by these organisms are not minimalŽ for organic acids; release of cations andror anions; example, Round, 1981; Eckhardt, 1985; Palmer and change of local pH; formation of polymers; forma- Hirsch, 1991; Gorbushina et al., 1993; Blazquez et tion of secondary minerals; formation of crustsŽ bio- al., 1995; Gomez-Alarcon et al., 1995a,b; Hirsch et films, patinas. and by active boring in limestones al., 1995b; Piervittori et al., 1994, 1996; Sterflinger and limestone matrix rocks. They point out that and Krumbein, 1997. and should not be discounted microbial degradation of rocks is not dependent on in the pre-tracheophytic or modern non-tracheophytic climate, location, or annual seasons and has been landscape and may have played a role in pre- documented in every environment from hot to cold tracheophytic raw soils or regolithsŽ see Keller and and wet to dry. Robert and BerthelinŽ. 1986, p. 455 Wood, 1993.Ž. . The comments by Berner 1992 about point to the role that strong adhesion ŽAoften virtu- New England gravestone weathering are falsified by ally glued to the mineral surfaceB. by bacteria, algae, the data cited above. fungal or lichen hyphae may play in dissolution For many concerned with rock surface weathering processes. They note that this process usually occurs processes in geochemical cycles and geomorphic at the microsites where living organisms adhere to phenomena, an active role for microorganisms is the minerals and can be observed with various sili- endorsed even if their relative impact on total weath- cates and with different living organisms. Ehrlich ering processes, particularly with regard to environ- Ž.1996, p. 8 notes that for microbes that live in mental conditions and specific rock characteristics, is biofilms on a mineral surface, the critical metabolic not known. products they produce Aencounter the silicate or Stevenson and FitchŽ. 1986, p. 49 report that aluminosilicate at relatively high concentrations and bacteria, fungi and lichens bring about the disintegra- thereby have a positive effect on the kinetics of the tion of rock surfaces to which they are attached solubilization of the mineral.B through the production of chelating agents; chelation, Thorseth et al.Ž. 1995, p. 141 note that microbiota a biochemical process, includes reactions by which Aselectively oxidize, reduce and chelate a large num- amino and organic acids combine with mineral ber of elements. . . . during microbial activity inor- 78 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 ganicrorganic acids and alkalies, resulting from mechanisms are discussed by EckhardtŽ. 1985 , metabolism will appear, and . . . cause changes in the BerthelinŽ.Ž. 1983, 1988 , Yatsu 1988 , Krumbein Eh and pH of the original environment. Chemo- Ž.1983 , Viles Ž 1990, 1995 . , Thorseth et al. Ž. 1995 organotrophic microbial communities may lower the and EhrlichŽ. 1995, 1998 . Excretion of inorganic and pH to 2–4, while phototrophic communities may organic acids is the most important mechanism by B A increase pH to above 10 due to CO2 utilization . . . which microrganisms attack minerals and thus the Biofilms, crusts, and patinas, the lithobiontic coat- structure of rocks and stonesB ŽEckhardt, 1985, p. ings of Golubic et al.Ž. 1981 , which are formed on 168.Ž. . Krumbein and Urzi 1991 review the whole and within rock surfaces by eubacteria and archae- spectrum of biological AdecayB affecting marble. bacteria, help to enhance the loss of surface material Webley et al.Ž. 1963 found that a high proportion and alter mineral crystalline networksŽ Gomez- of bacteria, actinomycetes and fungi isolated from Alarcon et al., 1995b, p. 231, citing Krumbein et al., ArawB soil in rock crevices rendered silicates soluble 1987; de la Torre et al., 1993b; see also Banfield and when tested in the laboratory and that their numbers Nealson, 1997; Riding and Awramik, 2000. . increase with colonization of lichens and mossesŽ as Although the significance of these weathering cited in Stevenson and Fitch, 1986, p. 50. . Others interactions has yet to be Aquantified at the ecosys- have shown that both bacteria and fungi preferen- tem level,B as WaltonŽ. 1993 , Hirsch et al. Ž 1995b . tially mobilize silica during mineral weathering reac- and BellŽ. 1993 note, it is obviously too early to tionsŽ Kutuzova, 1969; Avakyan et al., 1981, both as dismiss them in comparison with tracheophyte-medi- cited in Bennett and Casey, 1994, p. 188. . Both may ated weathering, for which ecosystem quantification also utilize silica supplied as quartz or silicate min- at a global scale is equally limited. Nor should their eral in metabolic processesŽ Heinen, 1967, 1968 and potential for generating organic carbon biomass be Krumbein and Werner, 1983, both as cited in Ben- ignored. EckhardtŽ. 1985 notes, for example, that nett and Casey, 1994. . within 3 days of exposure, freshly worked building Thus, despite proclamations about the signifi- stones are colonized by 10 3 colony forming units of cance of tracheophytes, comments and examples, heterotrophic bacteria and filamentous fungi per gram below, drawn from available literature and again of stone. considering some of the least hospitable environ- Literature synthesesŽ for example, Krumbein, ments for tracheophytes, support conclusions made 1968; Berthelin, 1983, 1988; Eckhardt, 1985; Yatsu, by many others that lower autotrophs and hetero- 1988; Viles, 1990, 1995; Walton, 1993; Gomez- trophicŽ. non-autotrophic microorganisms are capa- Alarcon et al., 1995a,b; Ortega-Calvo et al., 1995; ble of considerable physical and chemical rock Hirsch et al., 1995b; Ehrlich, 1995, 1998. and stud- weathering as well as generating organic carbon ies on the microbiology of rock surfaces and weath- biomass independently of tracheophytes and tracheo- ered stones in specific regions like those made by phyte-induced microhabitats. In a discipline that is Webley et al.Ž. 1963 , Adamo et al. Ž. 1993 and others now an active area for research, it is entirely possi- noted below, make it clear that microorganisms are ble, as VilesŽ. 1995, p. 31 points out, that an active able to actively attack minerals and rocks, physically role for biological influences will ultimately be re- and chemically, and that weathering by bacteria, vealed in areas where weathering is supposed to be fungi, algae and lichens is Aan important factorB— dominated solely by abiotic processes. even in Antarctic sandstones. The fact that these Schwartzman and ShoreŽ. 1996 argue that the organisms need no soil and commonly can grow on origin of prokaryotic and complex life forms has had and permeate substrates too hard to be penetrated by both geochemical and climatic effects on the Earth’s roots and can thus effect microscale biodegradation, surface that extends well back into the Archaean should not be forgotten in assessing their potential with the evolution of extreme thermophiles. They weathering effects, nor should the wide range of also argue that the progressive introduction of new environmentsŽ. aerobic, anaerobic, acidic, alkaline organismic groups led to lower surface temperatures and climates in which microbial weathering can correlated with increased weathering. Thus, as they occur. Various microbial weathering processes or suggest, enhancement of weathering by land biota A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 79 has occurred throughout the history of life Aby the processes. EhrlichŽ. 1996 notes that bacteria form stabilization of soil under the protective cover of the minerals or dissolve them to gain energy, carry out first microbesŽ. so-called cryptogamic soils . . . as respiration when oxygen is limiting or absent, and to well as other biotic effects including soil carbonic satisfy trace element requirements; that they are both and organic acid elevation . . . B Thus, as they argue ubiquitous even in environmental extremes so that Ž.p. 428 each new innovation in the microbial soil they occur on minerals everywhere in the biosphere. community resulted in greater biotic enhancement of Gorbushina and KrumbeinŽ. 2000 review the exten- weathering, culminating in the rhizosphere. sive evidence for the soil and rock weathering capa- Huang and KiangŽ.Ž 1972 see also Fox and Com- bilities of microbial mats. erford, 1992. demonstrated the rock weathering Pittman and LewanŽ. 1994 provide entry to the capabilities of organic acids produced by lower au- extensive literature dealing with the greatly increased totrophs on feldspars, i.e. the effectiveness of or- solubility of rock-forming silicate minerals exposed ganic acids as weathering agents which solubilize the to the organic acids produced by bacteria and the nutrients needed by autotrophs from silicate minerals lower autotrophs or by their decomposition or that of leaving behind a kaolinitic residue. Huang and other plant materialŽ see also Ehrlich, 1995, 1998; SchnitzerŽ. 1986 and Huang and Violante Ž. 1986 Barker et al., 1997, 1998. . The volume edited by provide more data on the interactions of organic Garg et al.Ž.Ž 1993 see also Riding and Awramik, acids and soil minerals, as well as with microbes, i.e. 2000, and Krumbein, 1979. contains a number of providing more information on the capabilities of papers that include information about the bacterial organic acids in silicate weathering. biodeterioration of stone, including Petushkova and Most of the above information is directed towards LyalikovaŽ. 1993 , Ortega-Calvo et al. Ž Table 1 has a better understanding of the organisms involved in references to almost 100 examples of substrates oc- the weathering of silicate minerals. However, McIn- cupied by epilithic and potentially Arock weatheringB tosh et al.Ž. 1997 summarize the extensive data on cyanobacteria; Kawano and Tomita, 2001, for bacte- the bacterial weathering of sulfide minerals, which rial weathering of modern volcanic ash and forma- makes it very clear that these widely scattered rock tion of biogenic minerals.Ž. , Tiano 1993 , Jain et al. minerals, although volumetrically minor, will cer- Ž.Ž1993 Tables 5 and 6 list acids produced, elements tainly be of material assistance in overall rock affects and general geochemical activity for bacteria, weathering under natural conditions. algae, fungi and lichens.Ž. and Martinez et al. 1993 ; these papers provide an extensive set of examples 19.1. Bacteria and cyanobacteria: chemical and but not much information on the details of the physical weathering of surface bedrock and regolith bacterial biodeterioration processes. Orial and Brunet Ž.1991 review some of the many bacteria involved in Here we include cyanobacteria as well as eubacte- the decay of marble. ria and archaeobacteria, e.g. both autotrophic and FriedmannŽ 1971, 1982, 1986; Friedmann and heterotrophic bacteria whether aerobic, facultative, Ocampo, 1976; Friedmann and Weed, 1987. dis- anaerobic, etc.Ž see Ehrlich, 1995, p. 134, Amany cusse the effects of biogenic rock weathering and heterotrophic bacteria whether aerobic, facultative, mineral mobilization in hot and cold deserts by anaerobic, etc.B, 1998. . Bacteria are among the so- bacteria, lichens and algae that live on rock surfaces called lithobiontic, rock inhabiting, organisms of and within the rock fabricŽ. cryptoendoliths and Golubic et al.Ž. 1981 that commonly live in synergis- within surface cracksŽ. chasmoendoliths of sand- tic associations with fungi, algae and lichens, which stones, limestones, granites and granodiorites, among may enhance the capacity of all these microorgan- others. isms for weathering rock surfaces. Barker et al. In addition to physical exfoliation, endolithic Ž.1998 point out that field data as well as in situ field cyanobacteria and lichens, commonly associated in experiments and in vitro experiments show that mi- distinctive microbial communities with heterotrophic croorganisms accelerate the degradation rates of both eubacteria, fungus, and green algae, appear to con- rocks and minerals by both physical and chemical tribute to rock weathering by mineral solubilization 80 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 and possibly by shrinking and swelling of gelatinous eroding capabilities of cyanobacteria, using Cayman cell sheaths with alternation of dry and humid condi- Island examples, which they termed phytokarst. Viles tionsŽ.Ž. Friedmann, 1971 . Vestal 1993 notes the Ž.1984 capably summarized the available informa- different chemical environments in each of these tion on phytokarstŽ she prefers the term biokarst, primitive microbial ecosystems that provide differ- defined to cover both depositional and erosional ences in pHŽ 5–6 in the lichen community; 8–9 in features. , with data indicating that the limestone the cyanobacterial community. . He suggests that this dissolving and depositing capabilities of organisms is probably related to the acidic organic compounds Ž.bacteria, lichens, fungi, algae are relatively minor produced by the lichen community and production of Žshe does not consider the overall effect on inland organic amines in the cyanobacterial community. limestone exposures of biologically induced solution Ortega-Calvo et al.Ž. 1993 specifically dispute since it is commonly not involved with karstic fea- any direct evidence for a chemical role for cyanobac- tures. . The limestone weathering data referred to teria, as opposed to heterotrophic bacteria, in biode- above is not, of course, directly applicable to silicate terioration of rockrstone surfaces, but indicate a role weathering but does indicate something about the in mechanical biodeterioration or physical break- weathering capabilities of the organisms involved. down. MettingŽ. 1990 as noted above, also ques- Field and laboratory studies such as those per- tioned the evidence for solubilization of mineral formed by Wagner and SchwartzŽ. 1967 , Marathe in soils by microalgae, including cyanobacteria. Or- and ChaudhariŽ. 1975 , Ortega-Calvo et al. Ž. 1991 , tega-Calvo et al.Ž.Ž 1991 see also Ortega-Calvo et and BakkerŽ.Ž 1970 cited in Dorn, 1998 . have do- al., 1995. found 15 species of cyanobacteria and cumented the effectiveness of bacterial coloniza- algae on building stones and from induced colonization and biodeterioration of mineral species. Still, tion of calcarenite in the laboratory by filamentous as Ortega-Calvo et al.Ž. 1991 note, the deleterious Microcoleus were able to demonstrate conclusive effects of cyanobacteria on building materials is a evidence for mechanical deterioration. In contradis- Amatter of controversy,B and AaggressiveB action tinction to the evidence presented by others, which of cyanobacteria, and algae, has been considered they note is often contradictory, Ortega-Calvo et al. Anegligible by someB ŽOrtega-Calvo et al., 1993, Ž.Ž1995, p. 339 see also Ortega-Calvo et al., 1993 .1995. . state, there Ais no direct evidence for a chemical role Pietrini and RicciŽ. 1993 , however, present con- of algaeŽ referring here to cyanobacteria and eukary- clusive evidence for the effectiveness of at least one otic algae. in stone decay.B They dispute, in contra- taxon of cave-dwelling calcareous cyanobacteria, diction to others, a chemical role as opposed to a Scytonema julianum in biodeterioration. They note physical role in biodeterioration for cyanobacteria. that this filamentous, subaerial AalgaB precipitates Gomez-Alarcon et al.Ž. 1995b note the presence extra-cellular calcium carbonate encrustations in the of heterotrophic bacteria isolated from crusts on form of calcite which it obtains by a Aprogressive granite, limestone and brick monuments in Spain and impoverishmentB or absorption of calcium carbonate discuss their biodeterioration potential. The role of from the calcarenite substratum. They also refer to cyanobacteria in rock weathering, whether entirely additional, similar examples of cyanobacterial weath- physical andror physical and biochemical, is contra- ering from elsewhere. DornŽ. 1998 and Krinsley dictory. Bull and LavertyŽ. 1982 provide solid evi- Ž.1998 discuss the role of bacteria, including cya- dence regarding the limestone weathering capabili- nobacteria, in the formation of some types of rock ties of cyanobacteria, using cave entrance region coatings. Although such rock skins are often consid- examples from MalayaŽ the rough, light-oriented, ered abiotic, KrinsleyŽ 1998, see also Krumbein and surface produced in the cave entrance region, is Jens, 1981. notes that Mn–Fe casts of bacteria are termed phytokarst.Ž. . Danin and Garty 1983 provide commonly preserved in rock varnish although the similar evidence for arid region limestones from the organisms themselves are seldom preserved. Such Negev, where lichens and cyanobacteria are in- rock coating are a Asecondary weathering productB volved. Folk et al.Ž. 1973 discuss the limestone that results in the formation of new mineral assem- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 81 blages although rock coatings also promote mechani- through the degradation of cyanobacteria, algae and cal weathering by enhancing preexisting rock weak- fungi leading to the production of recognizable nessesŽ. Dorn, 1998 . biokarstŽ as cited in Viles, 1987a,b, see Folk et al., Viles and GoudieŽ. 1990 suggest that endolithic 1973, whom she cites.Ž. . Schneider 1976 has stressed or partially endolithic cyanobacteria that produce the importance of the destructive capability of mi- boreholes, such as Schizothrix, Plectonema and croorganisms as compared with inorganic processes Phormidium, may contribute both to the precipitation Ž.inorganic chemical solution that some have stressed of a variety of secondary freshwater calcium carbon- relative to the destruction of limestone coastline ate depositsŽ. tufas, travertine , as well as to their rocks. active degradation, noting that cyanobacterial weath- VilesŽ. 1987a has reviewed the wide range of ering is occurring on both active and relict tufas at plausible biophysical and biochemical effects of the present time. cyanobacteria on limestone weathering in varied en- Marathe and ChaudhariŽ. 1975, p. 67 demon- vironmental situations. She specifically attempted to strated that at least one species of cyanobacteria examine the potential differential weathering effects Ž.Aulosira aenigmatica is capable of deeply weath- of epilithic, chasmolithic and endolithic bacteria as ering siliceous rock in monsoonal India in a short well as the synergistic effects of their interactions on time interval. They specifically note with regard to terrestrial limestone on Aldabra Atoll in the Indian the comparative effects of abiotic and biotic weather- Ocean. She observed penetration of endolithic taxa ing, that bare rocks show very little corrosion, to between 40 and 800 mm and for many samples an Awhereas below the algal community corrosion was altered layer ranging in thickness from 100 to 4400 considerable.B Wessels and BudelŽ. 1995 argue per- mm related to the activities of epiliths especially suasively for a dominant role for cryptoendolithic from dry terrestrial sites. The latter effect, however, cyanobacteria living between the quartz grains in the appeared to be constructive rather than destructive. porous Clarens Sandstone, Orange Free State, South While she concluded that microorganisms were ac- Africa, in exfoliative weathering that has led to cliff tive in weathering terrestrial limestones she also overhangs or hollowed out sandstone cliffsŽ holk- concluded that it was difficult to assess how quanti- ranses. . Thus, they posit that substrate alkalization to tatively important their effects are in relation to other a pH of almost 10 by the unicellular nanocyte processes in producing the overall rates of surface Chroococcidiopsis has increased the solubility of lowering recorded for the atoll.

SiO2 nearly 4 times, which in turn has led to acceler- Thorseth et al.Ž. 1995 summarize the literature ated sandstone weathering and an increased flaking that discusses the weathering capability of bacteria rate. AOne might consequently speculate that the and their role in solubilization of varied silicate origin of holkranses in Clarens sandstone are largely minerals and volcanic glasses. From experimental due to increased weathering as a result of the alkaliz- work, Thorseth et al.Ž. 1995 demonstrate the effects ing activity of cryptoendolithic cyanobacteriaB Žp. of bacteria derived from naturally altered tuff in 224. . altering basaltic glass. They showed that over time BakkerŽ.Ž 1970 as cited in Dorn, 1998 . notes that the alteration rate is increased at least by a factor of cyanobacterial organic acids remove aluminum and 10 compared to that of the first 6 months of the iron from granodiorite leading to the formation of experiment possibly related to a newly dominant clay materials in humid–tropical Surinam. Golubic group of bacteria. The effects were both chemical Ž.1973 briefly reviews the role of intertidal en- and textural and included bacterial attachment and dolithic microorganisms, including cyanobacteria, in development of biofilms irreversibly attached to the actively dissolving and deeply boring carbonate sub- glass surface; corrosion textures such as etching pits strates noting that carbonate penetration Afollows a or hollows; chemical alteration; and accumulation of pattern indicative of chemical dissolution . . . B Žp. elements in bacteria from dissolution of glass, such 471.Ž. . Schneider 1976 has observed that the lime- as Si and Al, indicating that the bacteria acted as stone coast of the northeastern Adriatic, Yugoslavia, element sinks and might provide the site for subse- is being removed at the rate of 0.25–1 mmryear quent neomineralization. Thorseth et al.Ž. 1992, 1995 82 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 discuss cyanobacteria in volcanic glass relating to and contraction of cells, generation of polysaccha- Ahigh alkaline microenvironments.B rides which on swelling split apart mineral grains, EhrlichŽ. 1995, pp. 226–233; 1996, 1998 summa- mineral transformations and so on. rizes information related to the biomineralization The organic matter produced by phototrophic or- mode of attack: microbially induced acids that act as ganisms such as cyanobacteria on exposed rocks ligands that pull cations from the crystal lattice and supports chemoorganotrophic bacteria and fungi that facilitate subsequent breakage of framework bonds; derive chemical energy from oxidation of organic of cations ; microbially produced organic or inor- compounds that they also use as carbon sourceŽ Brock ganic acids; microbially produced alkaliŽ ammonia et al., 1994, p. 640. . The CO2 produced during or amines. or microbially produced extracellular respiration by chemoorganotrophs is converted to polysaccharides acting at acid pHŽ Glycocalyx, carbonic acid and many chemoorganotrophs also ex- Ehrlich, 1996. manifest in weathering of rock sili- crete inorganic and organic acids: both are effective cates, of silica and silicates in nature that is and in dissolution of rocks. Sulphur, ammonia and nitrite recycling silica in nature. EhrlichŽ. 1996 talks about oxidizing chemoorganotrophic bacteria through their bacterial slimesŽ. acid polysaccharide that form com- metabolic activities excrete nitric and sulphuric acids plexes with silicate that lead to silicate solubilization. effective in mineral transformation and dissolution EhrlichŽ.Ž 1995, 1998 see also Nesbitt, 1997 . has Ž.Brock et al., 1994; Viles, 1995 . Ortega-Calvo et al. summarized much of the literature about the rock Ž.1994 note that some cyanobacteria use gypsum to silicate solubilization by bacteria, with illustrations support their growth with the sulphate released dur- of varied etch patterns for some of the minerals. ing gypsum dissolution progressively incorporated StevensonŽ. 1967 discusses the high proportion of into the carbohydrate sheath and suggesting that bacteria in soil as well as those living on rock gypsum might play a role in colonization of sul- surfaces that produce 2-ketogluconic acid. phate-bearing substrates and microbial biomass. FolkŽ. 1994a,b, 1999 , Folk et al., 1994 and Silli- Barker et al.Ž. 1998 used populations of soil and toe et al.Ž. 1996 provide data suggesting that rock groundwater bacteria, otherwise unidentified, to test weathering by AnannobacteriaB is an important pro- for aluminosilicate weathering in dissolution experi- cess both on bedrock and in the subsurface. Al- ments. They were able to show that bacteria that though the nature of these tinyŽ. 20 to 150 nm attach directly to mineral surfaces by extracellular sub-bacteria-size organisms is controversial. Unwins polymers or that synthesize exopolymers to form gel et al.Ž. 1998 detect living colonies of Anano- layers or biofilms that often cover much larger areas organismsB Ž.nanobe growths in the size range 20 of mineral surface than the bacteria alone, accelerate nm to 1.0 mm, on Triassic and Jurassic sandstones mineral weathering reactions by producing organic and other substrates that potentially confirm that and inorganic acids, metal-complexing ligands, by nannobacteria are genuine biological organisms. changing redox conditions or mediating formation of VilesŽ. 1990, Tables 2.2 and 2.3 has catalogued secondary mineral phases. Their dissolution experi- examples of Amicro-scaleB rock and mineral weath- ments demonstrated that microbial colonization of ering by micro-organisms and Alower plants,B and surfaces greatly accelerates mineral weathering reac- statedŽ. p. 11 that the importance of organisms to tions and Areleasewx d up to two orders of magnitude weathering Adoes not directly relate to the size of the more material to solution that abiotic controlsB Žp. biomass.B YatsuŽ. 1988 , Schwartzman and Volk 1560. . they also demonstrated that mineral surfaces Ž.Ž.1991, p. 359 and Viles 1995 list varied mecha- beneath microbial extracellular polymers Awere often nisms by which bacteria, including groups formerly extensively etchedB Ž.p. 1561 while experimental called eubacteria and archaebacteria, that include weathering studies demonstrate that chemical inter- autotrophic and heterotrophic bacteria, affect both action between extracellular polymers and mineral physical and chemical rock and mineral weathering, surfaces and mineral ions, increase dissolution Aby among them, the production of both organic and up to several orders of magnitude . . . B Ž.p. 1562 . inorganic acids and chelating agents, microfracturing As discussed above, Ferris et al.Ž. 1994 biogeo- of mineral grains and exfoliation related to swelling chemically couple weathering of silicate minerals in A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 83 surface bedrock by autotrophic cyanobacteria and in al., 1987, 1988; Sargent and Fliermans, 1989; Sin- the subsurface vadose zone by anaerobic sulfate-re- clair and Ghiorse, 1989; Balkwill, 1989; Fredrickson ducing or denitrifying bacteria, with extensive micro- et al., 1989; Hicks and Fredrickson, 1989; Phelps et bial carbonate mineral deposition in lakes and on al., 1989a,b; Madsen and Bollag, 1989; Francis et weathered basalt. This precipitation is consistent with al., 1989; Jones et al., 1989; McMahon and Chapelle, extensive deposition of carbonate minerals in regions 1991; McMahon et al., 1992; Amy and Haldeman, where the chemical composition of surface and 1997; Pedersen, 1999; Spark et al., 2000; Machel groundwaters is determined mainly by the bedrock and Foght, 2000. . weathering of silicate minerals. Sinclair and GhiorseŽ. 1989 discuss distribution of aerobic bacteria, cyanobacteria, protozoa, algaeŽ in- 19.2. Bacteria and cyanobacteria: chemical and cluding rare phototrophs. , actinomycetes, and fungi physical weathering of subsurface sediments and to depths of over 216 m in three boreholes from the bedrock Upper Atlantic Coastal Plain, South Carolina; Balk- willŽ. 1989 the numbers and diversity of aerobic Ž or The above studies have stressed microbial bedrock facultatively anaerobic. , chemoheterotrophic bacte- weathering at the atmosphererlithosphere interface. ria; Fredrickson et al.Ž. 1989 chemoheterotrophic

There is also evidence for significant subsurface bacteria, N2 -fixing microaerophiles, and nitrifying, r sediment bedrock weathering 100–300 m and at sulfur-oxidizing, and H2 -oxidizing lithotrophic bac- even deeper levels, below soil horizons both in the teria and Jones et al.Ž. 1989 anaerobic microorgan- unsaturatedŽ. vadose zone between the ground sur- isms, methanogens, sulfate-reducing bacteria and face and the water table and in the phreatic or coliform bacteria, from the same deep subsurface saturation zone below the level of the water table, sediments. Phelps et al.Ž. 1994a provide entrance where all voids are water saturated. These subsurface into the literature that deals with the presence, abun- environments could have had a counterpart in Pre- dance and activity of microorganisms and microbial cambrian and Early Paleozoic regoliths and imma- communities in subsurface sediments including those ture AsoilsB before the advent of tracheophytes as of the vadose zone, aerobic and anaerobic water argued by Keller and WoodŽ. 1993 . Stevens Ž. 1997 saturated sediments and basalt interbeds. According similarly argues for the potential of a subsurface to Phelps et al.Ž. 1994b, with reference citations , cradle for the origin of life, advancing the hypothesis microbial trophic groups identified in subsurface sed- that the deep subsurface may contain remnants, or iments include aerobic heterotrophs, methanotrophs, analogs to, ecosystems once present at the surface. nitrifiers and denitrifers, fungi, protozoa, anaerobic AppenzellerŽ.Ž 1992 see also Pedersen, 1999 . pro- heterotrophs, sulfur or metal reducers, and me- vides a good introductory series of comments about thanogens. Phototrophs suggest influence of sur- the potential importance of widespread subsurface face microbial populations in the recent past rather rock weathering activities by varied bacteria. than indigenous populationsŽ Sinclair and Ghiorse, Although subsurface environments, especially 1989. . those hundreds to thousands of meters below the At shallow depthsŽ. Ehrlich, 1998; 10 to 50 m , surface, previously have been considered largely abi- viable bacteria, actinomycetes and microeukaryotes otic, there is now a burgeoning literature related to such as algae, fungi and protozoans, are active, but deep and shallow subsurface microbiology that cata- further down only bacteria, primarily unicellular het- logues the kinds and communities of organisms, erotrophic and chemoautotrophic bacteria are in- physicalŽ. diagenetic and chemical effects of mic- volved ŽABelow a depth of 250 m down to several robiota on the substrate and ground water bio- thousand meters, only viable bacteria have been chemistry, and the array of factorsŽ nutrient, water recovered to date . . . B Ehrlich, 1998, p. 54. although availability, sediment texture, pH, concentration of Bennett et al.Ž. 1996 cite literature indicating Aviable metallic cations, among others. that effect the pres- heterotrophic bacteria and fungus . . . to depths of ence, abundance and activities of microorganisms in 1600 m . . . B and Rogers et al.Ž. 1998 microbial subsurface environmentsŽ for example, Chapelle et populations to 1600 m at temperatures exceeding 84 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

1008C. Boone et al.Ž.Ž 1995 cited in Amy and Halde- silica and aluminum from the mineral surfaceŽ Hie- man, 1997, p. 1. claim to have recovered microbiota bert and Bennett, 1992, p. 278. . In subsequent exper- from subsurface depths greater than 9000 ftŽ 3000 iments with crushed quartz and feldspar, it was m.Ž. and Pedersen 1993 suggests ca. 4 km maximum demonstrated that bacteria rapidly colonize the sili- depths for subterranean microbial activity. cate surfaces with evidence of chemical weathering Chapelle et al.Ž. 1988 demonstrated that a signifi- after 14 monthsŽ Hiebert and Bennett, 1992; Bennett cant bacterial flora of predominantly facultatively and Casey, 1994, p. 189. . anaerobic heterotrophsŽ 1075 –10 cellsrgram sedi- Bennett et al.Ž. 1996 AseededB mineral samples ment. was present to ca. 40 m depth in core samples Ž.quartz, microcline, albite, anorthoclase and so on in from the Floridan aquifer, South Carolina and to microbially active groundwaterŽ maximum ca. 120 depths of 182 m in coastal plain sediments, Mary- m. for 12-month intervals and then recovered to landŽ. Chapelle et al., 1987 . Sinclair and Ghiorse study microbial colonization patterns and weathering Ž.1989, p. 19, their Figs. 1–3 found population den- features. This experiment demonstrated that quartz sities of viable bacteria in several aquifer samples at and some feldsparsŽ microcline, anorthoclase and more than 200 m depth to be as great as in surface oligoclase. were widely colonized and deeply weath- soil at the same site; BalkwillŽ. 1989 noted that the ered at colonization sites with secondary clays pre- diversity of viable, aerobic, chemoheterotrophic bac- cipitated on some uncolonized surfaces in anortho- teria did not decrease with depth, down to 265 m, clase, while other silicatesŽ. albite, olivine and others despite the presumably nutrient-limited environment were uncolonized and unweathered. They speculate and variable carbon sources; Fredrickson et al.Ž. 1989 that these colonizationrweathering differences may noted that total heterotrophs, lithotrophs and mi- be related to microbial nutrient requirements and croaerophylic N2 -fixing bacteria were not influenced mineral composition such as the presence of potas- by depth with some of the highest populations from sium, an essential nutrient. depths greater than 200 m, where the carbon source Stevens and McKinleyŽ. 1995 describe active, may be particulate matter or found in association anaerobic lithoautotrophic microbial ecosystems in with mineral surfaces. deep basalt aquifers where bacterial growth and sur- Relatively high bacteria biomassŽ cellsrcm3. of vival in low organic carbon environments is pro- aerobic chemotrophic bacteria in deep subsurface moted and sustained by a purely geochemical energy aquifersŽ. Balkwill, 1989 indicates the potential many source obtained from weathering reactions. Thus, the meters below the rhizosphere for weathering effects relatively high populations of anaerobic autotrophic similar to surface weathering effects. Hazen et al. and heterotrophic bacteria are sustained in igneous Ž.1991 and Holm et al. Ž. 1992 indicate that most rocks within aquifers below sedimentary interbeds, such chemotrophic bacteria are attached to mineral by autotrophic metabolism coupled to mineral surfaces where by etching mineral surfaces they weathering, totally independent of photosynthesis. release ions to solution and act as nucleation sites for StevensŽ. 1997 recognizes two principal microbial secondary mineral phasesŽ Barker et al., 1997 pro- ecosystems in the subsurface based on trophic poten- vide entrance to this literature. . tial: heterotrophy-based ecosystems, metabolically Hiebert and BennettŽ.Ž 1992 also Bennett and sustained by available organic carbon available in Casey, 1994; Bennett et al., 1996. discuss the effec- sedimentary rock formations; and lithotrophy based tiveness of heterotrophic eubacteria in the coloniza- ecosystems, metabolically sustained by inorganic tion and chemical weathering of quartz and feldspar geochemical nutrient sources. The latter, sustained grains in an oil-contaminated aquifer at rates consid- by chemolithoautotrophy, utilize inorganic carbon erably faster than theoretically predicted, with high- sourcesŽ. Stevens and McKinley, 1995 . est bacterial density associated with the greatest Similar evidence is provided for mineral weather- weathering. Etching occurred in a reaction zone at ing and geochemical cycling of elements in shallow the bacterial–mineral interface where high concen- aquifers when bacteria colonize feldspar and quartz trations of organic acids, formed by bacteria during surfacesŽ. see Barker et al., 1997, p. 393 by creating metabolism of hydrocarbon, selectively mobilized Amicro-reaction zones of organic acids and other A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 85 metabolites concentrated at the mineral surface.B deep drill cores hundreds of meters below the atmo- These subsurface microorganisms have the capacity sphererlithosphere interface might seem unlikely to to affect both the physical properties of the aquifer have a significant impact on atmospheric CO2 , over system as well as the chemistry of the ground water intervals of geologic time, subsurface weathering Ž.Chapelle and Bradley, 1997 . surfaces can be exposed by tectonic uplift and re- Vandevivere et al.Ž. 1994 found in laboratory moval of surficial cover, and waters derived from simulation studies, that subsurface bacterial isolates, bacterial weathering identified in subsurface aquifers some recovered from depths of 181–267 m substan- may come to the surface or, move upward to be tially increased dissolution of the feldspar bytownite, incorporated with and diluted by meteoric waters to as well as silicate minerals such as albite, quartz and form part of surface runoff and contribute to the kaolinite, at neutral pH by excreting the organic acid limestonerdolomite carbonate sink. Importantly, gluconate and much smaller amounts of other acids some such subsurface weathering zones indicate the such as lactate, formate, and unidentified 2-ketoacid, bacterial potential for colonization and weathering deemed of minor weathering consequence. By main- with or without any available organic carbon source taining the culture at near neutral pH, they elimi- in sedimentary rocks and even for being sustained by nated the potential that dissolution was pH-depen- a purely geochemical reaction source, lithoau- dent. totrophic microbial ecosystems that are independent Based on in situ field microcosm experiments of photosynthetic inputŽ. Stevens, 1997 . with aquifer sediments, Rogers et al.Ž. 1998 found Keller and WoodŽ. 1993 show that Ž for extant that feldspars containing phosphateŽ as apatite inclu- atmospheric CO2 concentrations. a low level of mi- sions. and ammonium in small amounts are preferen- crobial respiration in the vadose zone supports ap- tially colonized and deeply weatheredŽ preferentially preciable subsoil CO2 concentrations because of the destroyed. by bacteria, including fermenters, iron slow rate of CO2 loss through diffusion to the reducers and methanogens, accompanied by precipi- ground surface. Thus, they argue that even for small tation of secondary clay minerals. They hypothesize amounts of microbial soil respiration, CO2 concen- that phosphorus is released due to microbially pro- trations in the vadose zone might have been suffi- duced acidity and chelating ligands in the vicinity of cient to account for the apparent constancy in chemi- the attached microorganisms. In contrast, similar cal weathering since the mid-ProterozoicŽ citing feldspars without phosphateŽ. phosphorus were found Holland, 1984. without the necessity to involve high to be barren of attached microorganisms and com- atmospheric CO2 to explain this apparent constancy, pletely unweathered even after a year’s time. They which Keller and WoodŽ. 1993 state could be used note that the observed weathering rates often far to support Holland’s contention that the rate of exceeded those predicted from observed laboratory chemical weathering has shown no significant in- rates and were related directly to the colonization of crease in time since the mid-Proterozoic. A microorganisms and not simply secondary to Chapelle et al.Ž. 1988 also postulate that the CO2 macroscopic, aquifer-scale chemical dissolutionB Žp. produced by subsurface facultative anaerobic hetero- 1538. . They also note that release of a limiting trophic bacteria as they metabolize sedimentary or- nutrient necessary for survival, such as phosphorous ganic material could drive dissolution of silicates and or nitrogen, as a result of weathering of a silicate carbonates along the aquifer flow path to modify source of these elements promotes microbial growth, chemical and isotopic composition of ground waters which in turn acts to increase mineral weathering. and ground water geochemistry. Ullman et al.Ž. 1996

CO2 may also be produced in the subsurface as a demonstrated, by controlled laboratory studies using result of microbial respiration, as heterotrophic bac- strains of bacteria isolated from subsurface environ- teria metabolize sedimentary organic material with ments, the potential impact of bacteria and microbial consequent dissolution of silicate and carbonate min- metabolites on mineral dissolution in subsurface con- erals along the path of groundwater flowŽ Chapelle tinental environments. Their results indicate that or- et al., 1988; Phelps et al., 1994a,b. . Although sub- ganic acidsŽ. and sometimes polymers are excreted surface mineral weathering of the type recognized in by a number of these strains when sufficient carbon 86 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 substrate is present but that aerobic growth is inhib- possible the physical weathering; Paradise also pro- ited by lack of nutrients. In oxygen-depleted environ- vides references to additional rock weathering by ment, organic acids were produced by fermentation. lichens reported elsewhere. SeawardŽ. 1996 summa- In summary, it now appears necessary to consider rizes current views about the rock weathering capa- the amounts of organic carbon present today in the bilities of lichens. Further extending the knowledge subsurface, and to also consider the possibility that of both epi- and endolithic lichen rock weathering considerable organic carbon may have been se- and atmospheric carbon dioxide consumption is questered in the subsurface during the past. Wessels and Kappen’sŽ. 1993 work on varied taxa. Syers and IskandarŽ. 1973 , Wilson and Jones Ž.1983 , Jones and Wilson Ž. 1985 , Jones Ž. 1988 , Yatsu 19.3. Lichens Ž.1988 , Singh and Sinha Ž. 1993 , Longton Ž. 1992 , WaltonŽ. 1993 , Viles and Pentecost Ž. 1994 , Ehrlich Although mineral and rock weathering effects Ž1995, 1998 . , Viles Ž. 1995 , Dorn Ž. 1998 , Cooks and now widely attributed to lichens were once seen as FourieŽ. 1990 , Cooks and Otto Ž. 1990 , Barker and due largely to physical causes, evidence that the BanfieldŽ. 1996 , Barker et al. Ž. 1997 , Barker and lichen–rock interface is the site of considerable BanfieldŽ.Ž 1996, 1998 , Easton 1997, see his Table chemical and physical activity is no longer in doubt, 2. , among others, provide entrance to the extensive as evidenced by the compiled citations to the ex- literature dealing with lichen weathering of disparate tended literature that deals with lichen biodeteriora- substrates including rocks and inorganic building tion of Astone work,B natural and artificialŽ Piervit- materials, where lichen growth is of almost universal tori et al., 1994, 1996. . Additionally, as Wilson and occurrenceŽ. Jones and Wilson, 1985 . Saiz-Jimenez JonesŽ.Ž 1983 see also Lee, 1999 . point out, scanning and Garcia-RoweŽ. 1991 discuss some of the lichen electron microscopy and electron microprobe analy- affecting marble, limestone and other tesserae from sis that make it possible to study the rock–lichen Roman pavements in Spain. Anagnostidis et al. interface in greater detail than was possible with Ž.1991 discuss the many biological agents affecting only the optical microscope, have removed previous the Parthenon, including lichens, bacteria, actino- doubts about lichen capability for the physical disin- mycetes, other fungi, and chlorophytes. Seaward tegration and chemical decomposition of bare rock Ž.1988 discusses rapid lichen weathering of terra surfacesŽ. weathering and transformation of rock ma- cotta. terial into soilŽ. pedogenesis despite statements to Lichens, like lithophilous photoautotrophic cyano- the contraryŽ. Syers and Iskandar, 1973 . As the bacteria and chemolithotrophic bacteria, can develop careful work on lichen weathering of Scottish basalt on bare rockrstone surfaces totally lacking in or- by Jones et al.Ž. 1980 indicates, the physical and ganic material, thus they can be among the first to chemical effects of lichen weathering can no longer occupy bare mineral surfacesŽ see Jackson, 1971, for be denied once scanning electron microscopy makes an Hawaiian example. . possible careful examination of rock surfaces be- Nieboer et al.Ž. 1978 provide circumstantial evi- neath lichen-encrusted rock. Simulation experiments dence, data concerning elemental abundances in have also confirmed the interpretation of alteration lichens, that is consistent with their playing a signifi- of rock features as biogeochemical weathering pre- cant role in the biochemical weathering process. cipitated by lichensŽ. Wilson and Jones, 1983 . Jones Friedmann and GalunŽ. 1974 summarize information et al.Ž. 1981 point out the utility of SEM in studying about the presence of lichens in varied desert envi- the rock weathering capabilities of lichens. Paradise ronments, from soil to bare rock. The volume edited Ž.1997 provides evidence from Arizona that in the by Garg et al.Ž. 1993 provides examples of rock sandstone-weathering lichen Xanthoparmelia the in- weathering involving lichens, especially the papers ner part of the colony overlies a rhizine rich area of by Tiano, Jain et al. and Singh and Sinha. Singh and physically weathered sandstone, whereas the outer, SinhaŽ. 1993 provide a detailed summary from the younger portion overlies a chemically weathered re- literature of the various physical and chemical pro- gion, i.e. chemical weathering precedes and makes cesses of rock deterioration attributed to lichens. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 87

EdwardsŽ. 1997 describes lichen weathering of weathering effects result in a distinctive microtopog- calcareous substrata. Preto et al.Ž. 2000 describes raphy. lichen weathering of granite, while emphasizing that SchneiderŽ. 1976 provides information on the rock different lichen taxa have different weathering capa- weathering capability of lichens on the Istrian coast- bilities, a very important point. Gehrmann et al. line immediately above the supratidal zone. He Ž.1988 provide solid evidence of lichen weathering describes this Alichen zoneB as soil covered where of both limestone and sandstone, some from Temper- epi- and endolithic lichens occur as closed carpet ate North German tombstones that falsifies Berner’s replacing cyanobacteria. This is a limited zone of Ž.1992 earlier comments about the absence of lichen weathering globally, but the lichen weathering capa- weathering from New England tombstones. bilities by both surface corrosion and chemical solu- Gehrmann and KrumbeinŽ. 1994 provide excel- tion are worth noting. FryŽ. 1927 provides visually lent evidence regarding the biopitting, corrosion ca- dramatic evidence of the effects of lichens on varied pabilities of eleven different lichen genera and even rock types, concluding that the work is mostly me- more species on varied limestone and marbles. They chanical, although at this early date techniques for also have recognized the organic acids produced by understanding the chemical possibilities were not yet the varied lichens involved as well as some of the available, nor were there SEM possibilities. products of the acid-caused solution, including var- Effectiveness of lichens in basalt weathering and ied hydrated calcium oxalatesŽ mostly weddellite and mineral decomposition at the rock–lichen interface whewellite, plus some calcium oxalate trihydrate. , has been demonstrated by Jones et al.Ž. 1980 . They and traces of goethite, glushkinite and nesquehonite. also performed simulation experiments involving Whewellite was predominant in most of the epilithic fresh labradorite and the common oxalic acid-secret- lichens, whereas weddellite was in most of the en- ing soil fungus Aspergillus niger. Jones et al.Ž. 1980 doliths. This type of investigation makes it clear that demonstrate that lichen weathering of ferruginous similar studies of varied lichen species with varied clays and ferromagnesian silicates, associated with rock types should tell us a great deal about weather- the basalt, yield an ochreous crust of ferruginous ing capabilities, particularly if quantified under care- Ž.ferrihydrite and alumino-silicate materials similar fully controlled greenhouse conditions. Such studies to that reported by Jackson and KellerŽ. 1970 . would provide a solid basis for evaluating the rock- Thus, there is reason to rejuvenate the classical weathering capabilities, or not, of the lower au- concept of the role of lichens as pioneer organisms totrophs, as well as guiding us in our thinking about in transforming rocks into raw soils—in soil forma- the possibilities for significant organic rock weather- tion because of their capability of colonizing sur- ing prior to the advent of the tracheophytes in the faces prior to the capability of most other organisms Late Silurian–Early Devonian. to do so. Lichens are particularly significant in the VilesŽ. 1987b summarizes some earlier informa- colonization and physical and biochemical alteration tion about lichen weathering by saxicolousŽ rock-en- of plain, bare rock surfaces, where, unlike scree and crusting. lichens at the rockrthallus interface appar- other unconsolidated surfaces and creviced rocks, ently by both mechanical and biochemical means tracheophytes are unable to establish early succes- involving lichen metabolic by-products. Her own sional populationsŽ. Syers and Iskandar, 1973 . focus on limestones from Somerset in southern Eng- Lichens together with other microorganisms may be land covered by crustose and partially endolithic entirely responsible for the weathering of rock sur- lichens, indicates that lichen weathering of lime- faces and the earliest pedogenic action leading to the stones is an ongoing, significant process but appar- formation of soils, in comparison with mineral ently less well developed in humid than dry environ- weathering within already established, tracheophyte- ments, and difficult to relate specifically to rates of inhabited rhizospheres, stressed by many, that de- surface lowering versus a protective effect. In pends on a consortium of microorganisms whose soil Mediterranean climates, KlappaŽ. 1979 discusses the biomass is at least in part co-dependent on tracheo- role of lichens in weathering carbonate rocks and in phytes, and potentially on the tracheophytes them- the production of laminar calcretes indicating that selves, although their role independent of associated 88 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 microorganisms is less clear. However, there are few Some, perhaps especially epilithic, crustose, saxi- data on lichen biomass, particularly for terricolous colous species that are firmly attached to the sub- formsŽ. Wein and Speer, 1975 , which hampers our strate by a mass of hyphae, and crustose endolithic ability to extrapolate into the distant pass. It is species that grow within rocks, and associated interestingŽ. Bell, 1993 that within a desert environ- micro-organisms, such as bacteria and algae not ment cryptendolithic biomass is concluded to equal directly involved in the symbiosis, appear to play a to or exceed vascular biomass. variety of roles, biogeophysical and biogeochemical, Schwartzman and VolkŽ. 1989 , for example, have in rock weathering and pedogenesisŽ although there suggested that the acceleration of silicate weathering are inherent difficulties in measuring lichen weather- by lichens to a hundred to thousand fold increase ing rates. . Foliose saxicolous lichens, more loosely over abiotic conditions during the Precambrian may attached to substrates by distinct clusters of hyphae have been adequate to depress atmospheric CO2 and Ž.rhizinae and haptera Ž suckerlike sheaths . , may also consequently lower terrestrial surface temperatures. play a role, particularly in chemical weathering al- Thus, there are a variety of physical environments though this is more difficult to demonstrate because throughout geologic time, but perhaps especially in of the possibly rapid removal of dissolution products some of the more extreme, more widespread, more by rain waterŽ. Ascaso et al., 1976 . arid environments of the Early Paleozoic, where It is clear that there is considerable taxon variabil- rooted tracheophytes can have had only minimal ity in the weathering capability of lichens; such effect on weathering in comparison with lichens. variability is also related to rock type. Singh and Danin et al.Ž. 1983 , as cited in Longton Ž 1992, p. SinhaŽ. 1993 note that the effects on rock-forming 34. , for example, considered that lichen cover accel- minerals vary according to the chemical composition erated weathering in dry environments, but protects and crystallographic structure of the rock. Ascaso et rock surfaces in wet conditions. Lichens are among al.Ž. 1976 on the other hand, noted differential ef- the dominants in extreme environments and the dom- fects related to different lichens. They studied the inant cryptogam in extreme environments with cal- pedogenic effects of three lichens on granite and careous bedrock such as the Negev Desert Highland, gneiss in natural circumstances, as well as albite, Israel where Shachak et al.Ž. 1987 note that 70% of orthoclase, biotite, muscovite and quartz, obtained the ground area is covered byŽ. small rocks partially from outcrops, in in vitro experiments and found that covered by epilithic lichens and containing extensive the lichen least attached to the rock substrate gener- areas of endolithic lichens found from 1 to 7 mm ated no detectable new minerals, although two taxa within the rocks. In addition to the role of endolithic closely adpressedrattached to the substrate both did. lichens in exfoliation, in the Negev Desert these Cooks and Fourie’sŽ. 1990 study involving different endoliths are adequate to sustain populations of her- lichen taxa and different quartzite and gabbro pro- bivorous snails that browse at depths up to 3 mm and vided similar results. Ahave the unexpected and major impact of weather- The mycobiontŽ. fungal partner of the lichen that ing this desert’s rocks at a rate of 0.7 to 1.1 metric appears to be mainly responsible for the biodeteriora- tonsrharyear.B This type of weathering contributes tion ability of lichens, induces chemical weathering to the process of soil formation at a rate that is by the production of organic acids such as oxalic similar to wind-borne dust deposition which is a acid. Some crustose lichens can produce up to 50% major factor in arid zones. Even in the present global of their total weight of oxalic acid. , various ‘lichen environment, DornŽ.Ž 1998, p. 57 citing Koppes, acids’Ž polyphenolic compounds of limited solubility 1990. reports that lichens are Adominant on over 8% in water believed to form metal complexes with of the Earth’s surface . . . B It is not difficult to imag- substrate silicate minerals. and by changing the ine circumstances in the past when Earth’s climate as chemical microenvironment at the rock surface such a whole was more similar to that of the Negev as change in pH. CO2 produced during respiration is Desert, and when lichens may have been far more transformed within the thallus into carbonic acid that prevalent than they are on the Earth’s surface today. can be responsible for biodeterioration of stone even A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 89 in the absence of organic acidsŽ Tiano, 1993; Jain et know whether this might have been a factor in the al., 1993.Ž . Jain et al. 1993, Table 6 . list the variety Pre-Carboniferous. of acids attributed to lichens and Singh and Sinha Barker et al.Ž. 1997, p. 395, their Fig. 1 suggest Ž.1993, pp. 371–372 provide a partial list of lichen that the ability of microorganisms, particularly lichen substances that act as metal complexing agents. communities, to physically disaggregate rock sur- Barker and BanfieldŽ. 1996, p. 64; 1998, p. 229 faces, is Aperhaps the most strikingly obvious bio- and Barker et al.Ž. 1997, p. 402 question the role of logical process involved in silicate mineral weather- water-insoluble lichen acids in mineral dissolution ing . . . B Robert and BerthelinŽ 1986, p. 457, their because such substances are usually generated and Fig. 12-2. note the effectiveness of fungal or lichen concentrated in the upper thallusŽ the photosynthetic hyphae in penetrating inside rocks and minerals and zone of Barker and Banfield, 1998. rather than at the microdividing mica crystallites commonly accompa- lichenrmineral interface, but document clearly both nied by dissolution. They note that minerals scattered varied biochemical and biophysical weathering ef- among the mass of hyphae within the lichen thallus fects of lithobiotic lichen communities. Barker and are both physically microdivided and chemically dis- BanfieldŽ. 1996, 1998 suggest that lichens can pro- solved. Dissolution patterns on quartz have been mote and accelerate weathering in varied ways on clearly distinguished under the encrusting thallus of outcrop surfaces which they have colonized: in the a number of lichens. The precipitation and formation Aindirect biochemilithic zoneB by organic acids that of new minerals beneath lichens have also been penetrated cleavages and grain boundaries that are observedŽ. Robert and Berthelin, 1986 . Thus, as too small to permit microbial colonization; and in the DormaarŽ.Ž 1968, p. 223 who may be citing Jacks, Adirect biochemilithic zoneB of direct microbial ac- 1965. notes, as an overall assessment of lichen cess both primary and secondary mineral phases are weathering capability, lichens absorb mineral ele- coated by thick layers of Amicrobial exopolymers, ments from bare rocks, being among the first organ- primarily acidic mucopolysaccharidesB that serve as isms to gain a foothold on such bare surfaces, syn- templates for the formation of secondary minerals. thesize a variety of secondary clay minerals and Barker and BanfieldŽ. 1996, pp. 67–68 suggest that produce a clay–humus complex, an accumulation of these extracellular biopolymers, which may be syn- organic detritus mixed with the crumbling rock sur- thesized by a diverse community of bacteria, algae face that Acan be regarded as an extremely thin and fungi that inhabit the lichen–mineral interface, organic-mineral surface horizon . . . that allows such may play a more important role in mineral weather- plants as mosses to become established.B The latter ing reactions in lithobiotic communities that hereto- in turn produce additional organic matter that partici- fore realized. Lichens also accelerate physical weath- pates in the build up of the amount of AsoilB on the eringŽ exfoliation, separating mineral grains from the rock surface. Thus, it seems clear that lichens in substrate, cracking, increase in pore volumeŽ see conjunction with other organisms play a potentially Barker and Banfield, 1996, 1998; Lee, 1999, for critical role in early stages of soil formation before discussion and SEM illustrations. by hyphal and they are habitable by tracheophytes, and doubtless rhizineŽ. hyphae bundles penetration and potential played a similar role before the advent of tracheo- expansion and contraction, and by the expansion and phytes. contraction of partially endolithic crustose thalli and Several specific fieldrlaboratory studies on the epilithic foliose thalli, in response to hydration cy- pedogenic effects of lichens, among many that might cles, e.g. changes in water content and freeze–thaw be cited, illustrate something of these effects. Ascaso cyclesŽ Syers and Iskandar, 1973; Jones and Wilson, et al.Ž. 1976 examined the effects of different saxi- 1985; Robert and Berthelin, 1986; Longton, 1988, p. colous lichen thalli in physically or structurally alter- 81, 1992; Yatsu, 1988; Kerr and Zavada, 1989; ing minerals of acidic rocks such as granite and Walton, 1985, 1993; Viles and Pentecost, 1994, Table gneiss, and the affect of specific lichen acids on 7.1; Viles, 1995; Tiano, 1993. . As noted above, particle samples of albite, orthoclase, biotite, mus- herbivores that browse endolithic lichens have an covite and quartz from these rocks, in extracting additional impact on weathering. It is difficult to cationsŽ. calcium, aluminum, iron, magnesium from 90 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 silicates, and in generating new minerals both in lia. were involved in both biophysical and biochemi- nature and under controlled laboratory conditions. cal breakdown of both rocks with effects more pro- Ascaso et al.Ž. 1976 were able to demonstrate that nounced on the quartzite than the gabbro, where lichens induce chemical and morphological changes Buellia was more effective than Acarospora. Their in acidic rocks such as granite and gneiss, alter the study illustrates that lichen weathering can be taxon edges and surfaces of mica and feldspars in the rock specific, as well as rock specific. Cooks and Otto and their mineral content through the generation of Ž.1990 provide geochemical data and SEM pho- new minerals at the interface between the rock sur- tographs implicating the endolithic South African face and the closely appressed thallus. Of the three lichen, Lecidea aff. sarcogynoides, both in biochem- lichen taxa studied in this context, only the taxon ical weatheringŽ. utilizing oxalic and lichen acids with a loosely attached thallus not in close contact and biophysical weathering of the Magaliesburg with the rock substrate failed to generate new miner- Quartzite. They conclude that lichens in general play als in nature or alternatively that new minerals that a far more important role in weathering of the land- were generated, were removed quickly by rainwater scape than hitherto realized, with endolithic taxa because of the loosely attached thallus. especially effective because they act internally as Hallbauer and JahnsŽ. 1977 observed the attack of well as externally on the rock. lichen hyphae on undisturbed micron-thick thallusr SEM investigations of lichen activity on sand- rock interface on quartzites and quartzitic conglom- stone of the Cedarberg Mountains, South Africa erates, noting morphological developments that they Ž.Viles and Pentecost, 1994, p. 111 show that the correlated with chemical changes such as apparent thin cover of epilithic crustose lichens is etching the evidence of Achemical boringB evidenced by the quartz grains and removing cement from the rock, porous rock surface. WaltonŽ. 1985 documented the but is concentrated in a thin surface zone. A study by Aconsiderable mechanical disruption of rock sur- McCarroll and VilesŽ. 1995 demonstrated that the facesB caused by penetration of crustose lichens into weathering rates caused by a single taxon of euen- mica schist in Antarctica. Kerr and ZavadaŽ. 1989 dolithic lichenŽ lichens that actively bore into rocks provide data for lichen weathering of South African where the thallus is imbedded. on gabbroic boulders veld rocks that demonstrate that these symbionts in arctic, alpine Norway, are at minimum 25 to 50 have considerable mineralizing capability under nat- times that of abiotic weathering in the same region, ural conditions. Their work demonstrated that an and probably much higher. In an attempt to measure increase in pH associated with the microenviron- the extent and intensity of weathering over time, they ment, caused solubilization of silica and subse- studied an age sequence of boulders and found that quently its removal from the underlying substrate. lichens penetrated between 500 mm and 2 mm into Adamo et al.Ž. 1993 note the intense disaggregation the rock and that within 240 years essentially all and fragmentation of rock surfaces immediately be- boulders were affected with evidence of a clear low lichen crusts caused by six lichen species on increase over timeŽ including increase in affected serpentine, gabbro and dolerite. They suggest that surface area of each boulder as well as degree of weathering occurs as the result of surface adhesion weathering. , that surface damage was widespread, of the lichen and penetration of hyphae, by dissolu- and that the lichens were Aweakening gabbroic rock tion of existing minerals and by precipitation and quite spectacularlyB Ž.p. 200 . formation of new minerals. Sanders et al.Ž. 1994 studied the physical and Cooks and FourieŽ. 1990 studied the effects of chemical interaction perpetrated by rhizomorph-for- epilithic lichen weathering on Magaliesberg quartzite ming lichens colonizing a calcareously cemented and Bushveld gabbro in South Africa by comparing conglomerate and a siliceous schist in Spain. They thickness of biochemically weathered zones in rocks noted that both chemical and mechanical weathering with and without lichens, penetration depth of hy- occurred in the AcontactB zone of the substrate sur- phae, and chemical compounds present in weathered face, with endolithic lichens integrated within the and unweathered plagioclase and pyroxene. They substrate having an enhanced contact zone compared concluded that epilithic lichens Ž Acarospora, Buel- with epilithic lichens, but not necessarily increased A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 91 interaction with the substrate. In both cases rhizoidal Jackson and KellerŽ.Ž 1970 see also Brady et al., hyphae, rhizomorphs or Amycelial cordsB that differ 1999. claim that there is an acceleration of chemical from rhizines because of irregular form and indeter- weathering by an order of magnitude of at least 10 to minate growth, commonly reach several centimeters 100 times for lichen-covered rock relative to bare depth in rock substrates and produce an extensive rock. JacksonŽ. 1993 , based on the detailed study of Asubstratic networkB that interpenetrates and incor- Jackson and KellerŽ. 1970 , claimed that in zones of porates substrate particles and separates superficial high rainfall, lichens greatly intensified and acceler- substrate laminae to enhance the potential for biode- ated the chemical weathering of Hawaiian basalt by terioration. In calcareous matrices direct penetration Aorders of magnitudeB. Lichens do act as binding occurred, probably by chemical dissolution rather agents for mineral materialsŽ. Walton, 1993 but this than mechanical disruption. capacity does not negate a role for them in chemical On the other hand, LeeŽ. 1999 shows that for one dissolution. As the careful work on lichen weather- lichen species Rhizocarpon geographicum, the Amain ing of Scottish basalt by Jones et al.Ž. 1980 indicates, weathering effectB on two British granites, is purely the physical and chemical effects of lichen weather- mechanical rather than chemical, induced by the ing can no longer be denied once scanning electron action of the rhizoids, principally following lines of microscopy makes possible careful examination of weakness. rock surfaces beneath lichen encrusted rock. Tretiach and PecchiariŽ. 1995 studied calcicolous However, conclusions based on a single species, endolithic lichens from a variety of rocky habitats as in the Hawaiian basalt issue, are an inadequate from the Trieste Karst area of northeastern Italy basis for global generalization. Adamo et al.Ž. 1993 because, as they note, these lichens often colonize speak to this question, involving six lichen species building stones and natural rocky surfaces with on mafic rocks in a limited region, suggesting that Aserious consequences for the integrity of the sur- differences are likely related to physiology rather facesB related to dissolution of the substrate by than thallus morphology but noting that there is a metabolic products secreted by hyphae. While few paucity of literature dealing with such differences data are available on the physiology and primary and that investigation is much needed. VilesŽ 1995, productivity of calcicolous endoliths, they are viewed p. 24. calls attention to the fact that the weathering as one of the main causes of surface pitting of rock action of fungi has also been much debated, because surfaces. Calcicolous endoliths appear to be signifi- some species are effective weathering agentsŽ as cantly different in ecophysiology and growth forms indeed with each group of organisms. , while others from the so-called cryptoendolithic lichens of are not, and one tends to take a position depending Antarctica. on the taxon one is studying. Sanders et al.Ž. 1994 The weathering capabilities of lichens in generat- suggest, for example, that species capable of producing a lateritic-type AproductB on well-dated Hawai- ing rhizomorphs, as discussed above, need to be ian basalt flows is shown by information provided by taken into account when considering lichen growth Jackson and KellerŽ. 1970 and Jackson Ž. 1993 . Ef- form and relationship between it and biodeterioration fectiveness of lichens in basalt weathering and min- potential. As noted below, there are groups of fungi, eral decomposition has also been demonstrated by such as the chemoorganotrophic black fungi, black Jones et al.Ž. 1980 from examination of weathering yeastlike fungi of the Dematiaceae group, now rec- at the rock–lichen interface and through simulation ognized as major agents of rock decay and color experiments involving fresh labradorite and the com- changeŽ Gorbushina et al., 1993, cite them as playing mon oxalic acid secreting soil fungus A. niger. a major role in deterioration on antique marbles and Jones et al.Ž. 1980 demonstrate that lichen weather- historical limestone buildings under differing cli- ing of ferruginous clays and ferromagnesian silicates matic regimes throughout Europe and in Asia, Africa associated with the basalt yield an ochreous crust of and America; see also Soukharjevski et al., 1994; ferruginousŽ. ferrihydrite and alumino-silicate mate- Urzi et al., 1994; Diakumaku et al., 1994. , whose rials similar to that reported by Jackson and Keller rock penetrating capabilities in natural outcrops and Ž.1970 . on monuments have long been attributed to other 92 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 organisms and to physicochemical factorsŽ Sterf- weathering was strongly outpacing that of surfaces linger and Krumbein, 1997. . colonized by lichens but that there were limited or Obviously, both ignorance and taxon variability no obvious weathering effects beneath the thallus as suggested by Viles, makes generalizations about cover. the weathering role of microorganisms and lichens Without adequately controlled greenhouse experi- difficult. She also suggests, as do Gomez-Alarcon et ments and detailed observational data including SEM al.Ž. 1995a and Arino et al. Ž. 1995 , that many litho- work on field samples, these questions about lichen biontic communities are mixed assemblages of pho- mediated weathering cannot be solved, i.e. the evalu- toautotrophs and heterotrophic bacteria and fungi, ation of the contribution to rock weathering made by and that the synergistic effects of these mixed, bio- each organism. Field observations in which only logically more active communities on weathering is maximum ages of the varied plant cover types are often radically different from and considerably more assumedŽ. e.g. Cochran and Berner, 1996 , as in the potent than the effects of individual organisms. case of the Hawaiian lava flows where ages of cover Thus, studies limited to one group of microorgan- types are based on the ages of the flows, are an isms Acannot give a complete idea of the complex inadequate basis for making useful conclusions, processes involved in stone deteriorationsB ŽGomez- whether dealing with lichens or tracheophytes. Wes- Alarcon et al., 1995a, p. 253. . At the same time, it sels and KappenŽ. 1993 show have such work can be may not be possible to always correctly identify the carried out in the field as well as in the laboratory. effective agent in corrosive activity, when a mixed community is present. In an example discussed by 19.4. Fungi Ortega-Calvo et al.Ž. 1995 , the calcium oxalate dehy- drate patina on monuments in Italy originally at- The weathering potential of organic acids pro- tributed to endolithic cyanobacteria was on further duced by fungi in mycorrhizal relationship with tra- study attributed to endolithic lichen activity. Erdman cheophytes and associated with rhizobacteriaŽ see for et al.Ž. 1977 report relatively high amounts of cal- example, Leyval and Berthelin, 1991. is widely ac- cium oxalate in a terricolous lichen, while Whitney cepted where the effects seem to be obviously syner- and ArnottŽ. 1986 discuss the calcium oxalate pre- gistic. It is also clear that free-livingŽ non-mycorr- sent in the fungus Geastrum. Jones and Wilson hizal. soil fungi are effective mineral weatherers Ž.1985 and Arino et al. Ž. 1995 also call attention to where the carbon source is organic material con- the fact that lichen encrustations under some specific tributed by tracheophytes and other organisms. Jen- environmental circumstances might play a relatively ningsŽ. 1995 provides an overview of fungal nutri- more active role in bioprotection than biodeterioration that is useful in understanding their weathering tion even though deleterious effects are still occur- capabilities. Jongmans et al.Ž. 2000 provide details ring on the substrate beneath lichen colonization. about how individual hyphae may enter a mineral For example, in environments where abiotic phys- grain and use their organic acids to release mineral ical and chemical weathering is so aggressive that nutrients; this is the first really detailed evidence substrate instability makes active lichen colonization about the process. difficult, biodeterioration may be a slower process But it is equally clear, as discussed above, that beneath the lichen cover, where it is possible for non-rhizosphere, synergistic fungal relationships with such a cover to be established, than on unprotected cyanobacteria and algae, as lichens, can also lead to substrate, even though characteristic biodeterioration extensive weathering. There is also evidence that processes such as disaggregation, calcium oxalate lithobiotic rock-dwelling fungus, that coat and in- deposition and crystal etching can be observed at the vade rock surfaces, quickly covering fresh rock sur- lichen–substrate interfaceŽ. Arino et al., 1995 . In faces, can be equally effective weatherers in the such an environment, and with only casual observa- presence of an organic carbon sourceŽ see Dorn, tion of the substrate beneath the thallus, it would be 1998, p. 49, for additional literature; also Riding and easy to conclude not only that physico-chemical Awramik, 2000. . Thus, free-living, rock-surface A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 93 fungi, isolated from sandstone rock surfaces, such as effects of neither appear to have been extensively species of Penicillium and Cephalosporium, have studied. For example, experimental and observa- also been demonstrated to be effective chemical tional studies of the effectiveness of meristematic weathers of rockŽ. Williams and Rudolph, 1974 black fungi, Dematiaceae, a group of fungi with probably through the production of citric acid. yeast-like growth patterns, never previously consid- Williams and RudolphŽ. 1974 compared the bio- ered significant in rock weathering activities, are chemical weathering potential, as measured by iron extolled by Sterflinger and KrumbeinŽ 1995, p. 490, chelation, of a lichen with that of fungi isolated from 1997.Ž also Wollenzien et al., 1995 . and they are the same rock surface and found that the fungi now believed to Ahave a high destructive potential chelated iron in solution, probably related to the towards the inhabited rock surfacesB where they are production of citric acid, 2 to up to 6 times as much found. Sterflinger and KrumbeinŽ. 1995 specifically as the lichen acid, squamatic acid, isolated from the note that the DematiaceaeŽ Wollenzien et al., 1995; lichen Cladonia squamosa. Gorbushina et al., 1993. have long been underesti- SchneiderŽ. 1976 notes that fungi, especially en- mated as weathering agents, and that their rock- dolithic fungi, play a Amost significantB role in rock penetrating activities have been attributed to other weathering on the Istrian coastline, where they organisms and to physicochemical factors. Yeast-like AlichenizeB endolithic cyanobacteria as nutrient Dematiaceae have now been described as inhabitants source. Degradation is more intense with increasing of marble, limestone and granite surfaces in the moisture e.g. in floors of rock pools rather than drier Mediterranean region where they are mostly found in surfaces in the supratidal zone. But in supratidal connection with surface lesions such as cavities and habitats commonly characterized by blanketing colo- fissures. Gorbushina et al.Ž. 1993, p. 210 attributed nization of lichens, fungi may occupy depths up to their effects on rock surfaces primarily to physical 1.5 exceptionally 2 mm, where the thallus spreads attack, specifically physical boring, since they were out under the surface by dissolving the rock unable to identify organic acids associated with their Ž.Schneider, 1976, p. 61 . This is a limited zone of fungi. The volume edited by Garg et al.Ž. 1993 weathering globally, but the fungal weathering capa- includes several papers that involve significant fun- bilities are worth noting. Urzi et al.Ž. 1991 discuss gal rock biodeterioration data, including those by fungal weathering of marble, as well as that accom- TianoŽ. 1993 and by Jain et al. Ž. 1993 . plished by bacteria, cyanobacteria and algae. There are numerous publications on the important It is clear that any soil-dwelling microorganisms role played by the organic acids produced by tra- capable of physical and chemical weathering effects cheophyte coevolved mycorrhizal fungi. Experimen- maintain both capabilities in non-soil situations. The tal evidenceŽ. Boyle and Voigt, 1973 makes it clear only question is the organic carbon source and the that non-mycorrhizal, free-living fungi also are capa- effectiveness of the activity where the carbon source ble in rock weathering, which opens up the possibil- is lower than it is in many soils or rhizospheres. ity for important, pre-tracheophytic rock weathering Fungi that are most effective in dissolving sili- capabilities of fungi. cates produce citric or oxalic acids, which are strong AEnhanced weathering of quartz has been identi- chelating agentsŽ. Stevenson and Fitch, 1986, p. 50 . fied in fungal systems and metabolically produced Henderson and DuffŽ.Ž 1963 as cited in Stevenson citric acid was found to be the reactive component and Fitch, 1986, p. 50. indicated the wide variety of ŽAristovskaya and Aristovskaya, 1968; Silverman and silicates attacked by a variety of fungi. VilesŽ. 1995 Munoz, 1970. illustrated by the weathering action of provides entrance to the literature that assesses indi- lichensŽ. Silverman, 1979; Jones and Wilson, 1985 B vidual roles for fungi that are found widely on and Ž.Bennett and Casey, 1994, p. 188 . Silverman and within terrestrial and aquatic rock surfaces, and green MunozŽ. 1970 isolated Penicillium from weathering and other algae that may also form lithobiotic com- basalt and cultured it in the laboratory in a glucose– munities on and within rocks. Although it is clear mineral salts medium with specimens of basalt, gran- that fungi and algae may play an active role in both ite, granodiorite, rhyolite, andesite, peridotite, dunite physical and chemical weathering, the weathering and quartzite. They were able to demonstrate that 94 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 after 7 days, as much as 31% of the silica, 11% of absence. Gomez-Alarcon et al.Ž. 1995a have opined the aluminum, 64% of the iron and 59% of the that epilithic growths of algae and cyanobacteria manganese of some of the rocks was solubilized. may contribute directly to biodeterioration, but that Thus, it is clear that Penicillium-produced acids their main role may be as supporters of heterotrophic attack rocks, although not all rocks were equally populations of fungi and bacteria with Ahigher de- susceptible. Weed et al.Ž. 1969 discuss mica weath- structive potential . . . B ering by fungi. Lewin and CharolaŽ. 1981 provide Gomez-Alarcon et al.Ž. 1995b list fungal genera evidence of the rock weathering potential of fungi. found colonizing granite, limestone and brick build- Similarly, yeast-like and meristematic black fungi, ing materials in Spain, and noted the capability of known in northern Europe, Mediterranean basin, individual taxa for producing organic acidsŽ acido- Africa, North American and Asian deserts, can be genic fungi. , forming biofilms, and oxidizing man- demonstrated in the Mediterranean region, to etch ganese and iron, thus promoting biodeterioration that and penetrate rock surfaces that appear to provide progresses though cracks and fissures to a few mm their natural ecological nicheŽ Wollenzien et al., depth. De la Torre and Gomez-AlarconŽ. 1994 iso- 1995; Gorbushina et al., 1993. . These pigmented lated acid and non-acid generating filamentous fungi fungi, a common cause of the blackening of rocks Ž.see their Table 1 from weathered sandstone, lime- and architectural surfaces, are especially tolerant of stone and granite of Spanish cathedrals and assayed extreme environmental conditions, and based on field their ability to oxidize both manganese and iron as a observations, stereomicroscopy and SEM Aseems to means of contributing to biodeterioration and crust be particularly significant in rock degradationB ŽWol- exfoliation. From their culturing experiments they lenzien et al., 1995, p. 292.Ž . Dorn 1998, p. 50–55 .were able to conclude that many filamentous fungi provides evidence that black epilithic Amicrocolonial have the capacity to oxidize both manganese and fungiB in Arizona deserts bore into host rock varnish iron, important components in the Earth’s crustal and underlying rock, and also release organic acids rocks, by indirect or directŽ. enzymatic mechanisms. that dissolve rock varnish coating. Robert and Physical weathering is also accelerated by pene- BerthelinŽ. 1986, p. 457, their Fig. 12-2 note that tration of fungal hyphae and rhizomorphsŽ organized fungal or lichen hyphae penetrate inside rocks or root-like aggregates, discussed above in association minerals and between biotite layer packs, which with the fungal member of lichens. into rocks that results in the Asegregation and microdivision of min- permit water entry as well as providing Aaccess eralsB often an important phenomenon in crystalline channelsB for bacteria, fungi and algae potentially rocks. Commonly a deposit of calcium oxalate at the capable of biogeochemical attackŽ Longton, 1988, p. base of detached minerals indicates a concurrent 82, 1992; Walton, 1993, p. 47. . Cracks and crevices chemical dissolution that aids in detachment. in rocks, also provide potential AmicroenvironmentsB Hirsch et al.Ž. 1995a recovered microbial isolates, for microbial activity, analogous to those suggested including yeasts and filamentous fungi, from sand- by Berner for roots and their associated microbiota. stone and granite from Antarctica and from a variety Friedmann and GalunŽ. 1974 summarize data about of stone monuments. Although none of these endo- the soil fungi of deserts. or epilithic fungi were restricted to the rock environ- 19.5. Algae ment, having arrived at the rock surface from the aerobiota, they were assessed as probably contribut- The role of eukaryotic algae as biodeterioration ing substantially to rock deterioration in a variety of agents is still being assessed, as indeed are others of different waysŽ cation mobilization as result of the organisms discussed aboveŽ Viles, 1995; Ortega- chelating activity of excreted acids and through acid- Calvo et al., 1991, 1993, 1995. . Reviews by Viles ity itself; hyphal swelling; formation of secondary and Ortego-Calvo et al. provide entrance to the minerals.Ž. . Hirsch et al. 1995a,b note that fungal literature on green and other algae, including diatoms weathering will be exacerbated by the presence of and brown algae, that may form lithobiotic commu- primary producersŽ. algae and cyanobacteria as or- nities on and within rocks. Ortega-Calvo et al.Ž. 1993 ganic carbon source, but can still be shown in their Žin Garg et al., 1993, including Tiano, 1993; Jain et A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 95 al., 1993. also assess the role of epilithic and en- role for algae, similar to that reported for cyanobac- dolithic green algae in weathering processes of build- teria, related to mucilagenous sheath contraction and ing stones. Hall and OtteŽ.Ž 1990 as cited in Viles, expansion during cycles of drying and moistening. 1995, p. 25. suggest a role for cryptoendolithic green Jain et al.Ž. 1993 indicate that epilithic and chas- algae similar to that of cryptoendolithic cyanobac- molithic algae contribute to physical weathering by teria in exfoliation, although chemical attack may retaining water which can expand in freeze–thaw also play a role in detachment of rock flakes cycles. SchneiderŽ. 1976 provides information on the ŽTschermak-Woess and Friedmann, 1984, as cited in limited rock-weathering capability of endolithic al- Viles, 1995, p. 25. . The physical rock flaking pro- gae in the chlorophyceae and rhodophyceae on the moted by the expansion and contraction of algae Istrian coastline immediately above the supratidal living within rock cracks is one that is said to zone. He reports that both groups are insignificant drastically alter the appearance of rock surfaces in biodegraders in the inter- and supratidal zones com- cold-wetŽ.Ž Hall and Otte, 1990 , cold-dry Friedmann pared with cyanobacteria, lichens and fungi. et al., 1967.Ž and warm-wet climates Fiol et al., 1996; Dorn, 1998, p. 57. . Friedmann and Galun 19.6. Bryophytes Ž.1974 review the presence of different algal groups in and on different desert soils, as well as under Bryophytes and lichens show striking similarities rocks, and within the cracks and outer layer of desert in ecology and physiology that enable them to pio- rock exposures. neer the colonization of bare and newly exposed In comparison with phototrophic bacteriaŽ cyano- rock surfaces uninhabitable by tracheophytes, and bacteria. and heterotrophic microorganisms, some Apromote soil formation by accelerating physical and have claimed a role for algae primarily in the support chemical weathering, by trapping wind-blown or- of other, potentially more effective weathering organic and inorganic material, and by contributing ganismsŽ see Gomez-Alarcon et al., 1995a, p. 252; directly to undecomposed organic matterB in addi- Ortega-Calvo et al., 1993, 1995. because their growth tion to concentrating K, P, S and N, commonly represents a significant input of organic matter to deficient in immature soilsŽ. Longton, 1992 . How- rock surfaces that can be utilized by heterotrophic ever, the global coverage of bryophytes as compared bacteria and fungi that can give rise to organic acids. with other autotrophs is not well quantified. The role of opportunistic species of chlorophytes LongtonŽ. 1992 implies similarity in the physical Ž.green algae , found in soil and air, in colonization and chemical weathering effects by bryophytes and degradation of man-made stone structures is Ž.mosses and lichens, wholly unrelated organisms reviewed by Ortega-Calvo et al.Ž.Ž 1995 see also that nevertheless may show striking similarities in Ortega-Calvo et al., 1993, 1994. , although much of ecology and are often complexly associated in differ- their paper deals with cyanobacteria included by ent habitats. Both he and WaltonŽ. 1993 report that them, for purposes of discussion, among the algae. moss rhizoids penetrate rocks along cracks, some up Ortega-Calvo et al.Ž. 1993 dispute any direct evi- to a depth of 5 mm, to assist in physical weathering dence for a chemical role for algae in stone decay, and to provide access channels for bacteria, fungi although they note that the information is contradic- and algae. Like lichens, mosses also scavenge and tory and they report that microbial films may contain accumulate wind-blown mineral particles which may significant amounts of adsorbed inorganic materials be chemically attacked by moss exudates. Longton derived from the substrate such as quartz and cal- Ž.1992 suggests that because mosses grow more cium carbonate, although it is unclear whether they rapidly than lichens, they have an increased capacity are referring to cyanobacteria, algae or both. Jain et for trapping such material and accelerating the con- al.Ž. 1993 list the variety of acids and extracellular tribution of organic material to the substrate as well products produced by green algae and diatoms which as initiating succession. potentially could cause stone biodeterioration. LongtonŽ. 1992 also notes a potential role for Ortega-Calvo et al.Ž. 1993 , like Woess and Fried- bryophytes in retarding weathering, similar to that manŽ. above suggest a mechanical biodegradation reported for lichens, by protecting surfaces from 96 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 erosion and by absorbing precipitation reducing frost carbohydrate-rich material, i.e. this is a symbiotic, shattering. co-evolved relation. Mycorrhizae have a number of functions in their association with roots. Among the most important, is that together with root hairs fun- 20. Evolutionary history of the rhizosphere: con- gal mycelia act as root extensions into the substrate, ventional and alternative hypotheses thus potentially enlarging the root system and the As defined in relation to tracheophytes, the rhizo- source of available nutrients. sphere as an ecosystem could not have existed prior Having presented the conventional view of rhizo- to the advent of the tracheophytes and mycorrhizal sphere evolution we here present an alternative pos- fungi and the organic biomass contributed by their sibility. Our view is that the first step, presumably in presence and that of associated organisms. the earlier Precambrian, from an inorganic regolith, With the advent of tracheophytes near the begin- Araw soilB, is the addition of organic matter con- ning of the Devonian or the Carboniferous, depend- tributed from lower autotrophs. The biotically influ- ing on what evidence one accepts for the joint enced soil part of the ecosystem came into existence presence of tracheophytes with mycorrhizae, their by at least the Late Archaeozoic, with the advent of effects on the weathering of soil silicate minerals and cyanobacteria, and possibly other micro-organisms. its relationship to atmospheric CO2 levels must also By the earlier Middle Ordovician, there were hep- be taken into consideration. Through mineral weath- atic-like embryophytes that should have been capa- ering soils are the ultimate source of elemental nutri- ble of adding organic carbon to the soils with their entsŽ. Ca, Mg, K, Na, PO4 , etc. utilized by plants preexisting micro-organisms, possibly including and made available to them through roots and their free-living fungi. TrappeŽ written communication, mycorrhizae. 2000. points out that APresent day bryophytes have Tracheophytes provided the first group of plants Amycorrhizoids,B i.e. rhizoid fungal endophytes. with roots capable of penetrating unconsolidated These have been interpreted as mycorrhiza-like in substrate often to considerable depths below the land function but with little evidence other than that they surface. Evidence suggests that roots as opposed to do not seem to harm the host.B Much later in time, the rhizomesŽ underground stems that characterized possibly by the mid-Silurian, the soil ecosystem also the earliest tracheophytes and tracheophyte-like had a rootless, possible tracheophytic addition, possi- plants.Ž first appeared in the Early Devonian Elick et bly with free-living soil fungi that might have been al., 1998. , although the distribution of root-bearing intimately associated with possible tracheophytes. A plants is commonly not thought to have been exten- potential first step in the ultimate establishment of a sive until sometime in the Late Devonian–Early closer relationship between fungi and tracheophytes Carboniferous. —the mycorrhizal relationship—may have begun Moreover, tracheophytes are perceived to be the with the establishment of fungal colonies on root first terrestrial organisms to form extensive associa- surfaces. Still later, probably in the earlier Devonian, tions with fungi involving the roots in the elaborate root-bearing tracheophytes were added. This is ulti- fungus-root synergism, known as mycorrhizae. Some mately followed by the appearance of tracheophytes, perceive this to be what enabled tracheophytes to with their mycorrhizae that together make up the become established on landŽ Pirozynski and Malloch, rhizosphere, with its much higher organic carbon 1975. . Until recently, tracheophytes have been inter- content. The evidence for mycorrhizae before preted as the first embryophytic land plantsŽ Gray, the Carboniferous is contentiousŽ Pirozynski and 1985. : a concept now discredited. Pirozynski and Hawksworth, 1988b. . If the rhizosphere is defined to MallochŽ. 1975 concluded that tracheophytes by mean roots only, then it appears in the earlier Devo- themselves would have been unable to secure neces- nian with the advent of true rooted tracheophytes, sary mineral nutrients from ArawB mineral grains in but if it must involve symbiotic mycorrhizae, the the soil, whereas mycorrhizal fungi have this capabil- rhizosphere may not have appeared until the Car- ity. Mycorrhizal fungi AexchangeB these mineral nu- boniferous. Later yet in the Devonian or Carbonifer- trients with their tracheophyte hosts in exchange for ous there is the definite addition of mycorrhizae to A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 97 the soil ecosystem, coevolved with tracheophytes. In Ž.2 Volume of tracheophyte-inhabited soils versus one sense, the rhizosphere did not exist prior to the non-tracheophyte—estimates on a global scale of the Devonian advent of roots; in the second sense the landmass occupied by tracheophytes. See section on rhizosphere did not exist until the advent of symbi- the rhizosphere for details. otic mycorrhizae. The supposed Early Devonian Ž.3 Variability in weathering from soil to soil de- Rhynie Chert symbiotic fungus are present in stems, pending on mineralogy and even in different hori- not in rhizomes or true roots, although many have zons even in the same climatic region and in differ- assumed that they were physiologically and function- ent climatic belts. ally no different than the fungal symbionts of true Ž.4 Low elevation, low relief regions versus high roots. relief, montane regions. It seems to be commonly assumed that most Ž.5 Effectiveness of soluble versus diffusible or- pre-tracheophytic rock and mineral weathering was ganic materials in wet and dry soils. related entirely to mechanical stresses and abrasion, Ž.6 Incidence of soils through time Ž especially in while chemical mineral weathering, in soilsrre- the mid-Paleozoic, at the AdropoffB in Berner’s goliths now regarded as largely biological, was curve. . largely abiotic. Although the rate of chemical weath- Ž.7 Areas of minimal tracheophyte habitation. ering for soils and bedrock may be different, we Ž.8 Decrease in weathering in AoldB soils. argue that bedrock and soil weathering cannot be Ž.9 How to estimate on a geological time scale ignored as a factor in pre-Devonian atmospheric areas where lichens were dominant, in areas where models, particularly for the pre-Carboniferous where moisture is a problem for tracheophytesŽ see Dorn, it seems unlikely that soils closely approximated 1998, p. 61. . modern soils or were necessarily widely developed Ž.10 Brady Ž written communication, 2001 . is con- A on a global scale. Particularly in arid, hypersaline cerned about the direct or indirect effects of pCO2 regions, so widespread during many Phanerozoic and temperature.B intervals, it is clearŽ. Hunt and Durrell, 1966 from Ž.11 How to directly measure the weathering im- modern examples that cyanobacterial activity is high, pact of tracheophytes as noted by DornŽ. 1998, p. 60 with the implication that such activity probably ex- Ž.and citing Stallard, 1985, 1995 ; weathering infor- tended well back into the older Precambrian. mation is now angiosperm based—is it possible to Thus, it is conceivable that there was a complete extrapolate into the geological past to unfamiliar evolutionary continuum from bedrock and raw Ain- non-angiosperm vegetation types; some questions organicB soil immediately derived from them, re- about particle size effects on weathering in soils; golith to rhizosphere-type soils. Additionally, it is quantitative estimates of areas covered by soils and also clear that non-rhizosphere, lower autotrophic plants in the Paleozoic—where did these plants live soils continue to the present in appropriate environ- and how much land surface did they cover; are trees ments, chiefly very arid paleosols and regoliths like more important than herbs?; how to estimate the those that preceded the advent of the tracheophytes. quantitative effect of plants on the rate of weathering The physical weathering continuum has presumably Ž.as Berner, 1995, p. 577, asks . been present since the earlier Precambrian, but as we The studies mentioned above are among the in- will stress and demonstrate below, there was also a creasing number of recent investigations on the ef- contemporary biotic weathering continuum. fectiveness of microbes, fungi, and lower autotrophic organisms in physical and chemical bedrock weath- 20.1. Summary ering and organic carbon generation. They emphasize, as with the case of the dematiaceous fungi, that Among the variables that must be considered in their rock-decaying capabilities have long been un- assessing the actual quantitative effects of plants in known andror underestimated. Since these weather- the silicate weathering rate are the following: ing interactions have not been Aquantified at the Ž.1 Root mass and variations in root mass, highly ecosystem level,B as noted above, significant com- individualistic according to plant habitat and biome. parisons vis-a-vis the effectiveness of tracheophytes 98 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 and their associated microbiota in the microenviron- of their effects from a few climatic regimes and ments associated with tracheophyte root systems in regions of the world and involving only limited all ecosystems are not really possible. microorganism taxa. Silver et al.Ž. 1986 consider the There clearly are modern ecosystems where tra- varied microbial systems present and effective in soil cheophytes have little weathering impact, due to weathering. Until the effects of tracheophyte root their low biomass and the limited time when they are systems, independently of their associated micro- functional, related to temperature and moisture re- biota, have been critically examined, it may be pre- quirements. Some such ecosystems were doubtless mature to attribute significantly increased weathering more widespread in the past. BernerŽ. 1995, p. 576 rates to the appearance of arborescent tracheophytes suggests that even if lichens and other microbial in the Devonian. It is clear from the Sand Box organisms were effective weathering agents prior to experiments noted above, where the weathering ef- the advent of the tracheophytes, their slow growth fects in microbe inhabited environments were more rate and the limited interfacial area between the significant than those of grass-inhabited environ- lichen and rock surface compared with the growth mentsŽ. some grasses are mycorrhizal , that rock sur- rate and interfacial area between roots and soils must face weathering by microorganisms is as likely to be mean that Arooted higher plants are much more as effective as tracheophytes in many regions of the effective in attacking rocks and weathering them . . . B world in the presence of only shallowly or poorly There certainly may be environmental circumstances rooted tracheophytes. where this is a valid argumentŽ we have limited Until controlled experiments are done to assess information about soil bacteria. . There are doubtless the potential contributions of lower autotrophs as ecosystems, where lichens, bacteria and other au- well as fungi and bacteria under different tempera- totrophs dominate in the vegetation and cover exten- ture and moisture regimes, it would be premature to sive areas of rock surfaces so that even if they are arrive at any conclusions about the weathering capa- less effective weatherers, their overall effects cannot bilities of these organisms prior to the Devonian be regarded as negligible. There are limited data on arrival of the tracheophytes. In the absence of such the extent and importance of non-mycorrhizal soil information, estimates based on one or two studies weathering. are questionable. BernerŽ. 1992, p. 3225 is aware of The examples provided above obviously cannot the speculative nature of his conclusions supporting AproveB that pre-tracheophytic embryophytes and only a limited role for non-tracheophytic organisms, terrestrial phototrophic microorganisms had recog- stating A . . . much more work needs to be done on the nizable effects on pre-Devonian climate and atmo- quantitative effect of both primitive biota and vascu- sphere any more than it can be proven that they did lar plants on the rate of weatheringŽ and their re- not. Nor do they indicate that pre-tracheophytic or- sponses to changes in atmospheric CO2 . before any- ganisms can be dismissed from consideration in thing more definitive can be said.B Work such as that models of atmospheric CO2 concentration without done by Wessels and KappenŽ. 1993 indicates that far more observational and decisive experimental the atmospheric carbon dioxide consuming capabili- data that specifically address comparison between ties of the lower autotrophs may have been consider- tracheophyte-dominated and non-tracheophyte-domi- able in the past. nated habitats from different ecosystems and cli- It should be clear from the above discussion, that mates and take into consideration climatic variations there are a large number of poorly understood and through geologic time that might have favored non- complex variables, both with regard to organisms, tracheophyte-dominated habitats. even potentially at the taxon level, and physical Even if microbially mediated weathering rates environmental variables including microenvironmen- were lower without the specific microenvironment tal variables and rock characteristics that are poten- created by rooted vegetation, and are lower today in tially crucial and extremely difficult to estimate for some sparsely tracheophyte-inhabited regions, it is the geological past. Only poorly constrained approxi- premature to discount microbes, algae, fungi and mations are possible. One notes in discussions of bryophytes, based on only perfunctory examination these models, for example, limited consideration for A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 99 the effects and potential effects of tracheophyte evo- microbial activity in the Precambrian as the result of lution through timeŽ the large number of parameters such Asoil stabilizationB would provide for a 100- to introduced to justify them without any understanding 1000-fold increase in weathering over that in an of their feasibility in terms of tracheophyte evolution abiotic world and thus exercise a major control on through time. . Additionally, in order to assess the atmospheric CO2 and climate. magnitude of the claimed comparative superiority of DreverŽ. 1994 points out that a major unknown in tracheophytes and mycorrhizae in enhancing weath- estimating the initial effect of land plantsŽ e.g. tra- ering, it is essential to have information both about cheophytes.Ž on weathering rates, is the nature thick- the weathering rates of microbial and non-tra- ness, particle size distribution and permeability. of cheophytic autotrophs, and comparative information the pre-Silurian or we would say pre-Devonian, re- about rock and soil weathering rates caused by golith. We know even less about the nature and cyanobacteria, varied terrestrial algal groups, lichens, extent of Acryptogamic soils.B Without information fungi, liverworts, mosses and tracheophytes from all about the global extent and nature of pre-tra- global environments involving a diversity of taxa in cheophytic soils, it is difficult to assess both the all cases. It is essential to consider whether weather- initial effects of early tracheophytes or the potential ing rates might be radically different in different large scale effects envisioned by Berner in the Late climatic environments, even in the presence of tra- Devonian–Early Carboniferous for the spread of ar- cheophytes, and whether the comparative weathering borescent tracheophytes with extensive root systems. effects of rock and soil inhabiting, non-tracheophytic DreverŽ. 1994, p. 2331 concludes that the advent associations of microbial organisms may vary from of tracheophytes had little direct effect on the weath- region to region in a significant way. For example, ering of silicate rocks, Aprobably less than an order KrumbeinŽ.Ž 1968 see also Viles and Pentecost, 1994, of magnitudeB with the dominant effects being phys- who note divergent roles of deterioration and protec- ical, related to the binding and retaining of fine tion, and Viles, 1995. reports that the part played by particles, and the indirect effects of tracheophytes on microflora in the weathering process is dependent on the regional distribution of precipitation patterns on the microclimate and the environment and composi- continental land masses, that he suggests may be as tion of the rocks concerned. Krumbein notes that in critical a factor on weathering rates as their direct humid climates lichens commonly protect rocks from effects. For example, related to Drever’s comment weathering, but that lichens are Aone of the most about the nature of soils and weathering: small soil effective deteriorating agents under arid climatic particles have a much greater surface area per unit conditions.B A semi-permanent, lichen-microbial weight than do large particles, and those with large cover may retard inorganic weathering of rock and surface area per unit weight, dominate the ion ex- mineral surfaces by insolation and protection from change properties of soilsŽ. Nobel, 1999, p. 366 , erosionŽ Viles, 1990, p. 11, 1995; Longton, 1992, p. which means in effect that mineral weathering would 34. . potentially be more rapid in such soils than in soils Schwartzman and VolkŽ. 1991, pp. 359–360, 369 where particle size is mostly over 1 mm in diameter. suggest that stabilization provided by a Acryptogamic The question thus arises whether early em- soilB Ža AcrustB or soil stabilized by microbes—a bryophytes and varied oxygenic and other microor- microbial counterpart to soil stabilized by the subsur- ganisms that preceded them in the terrestrial habitat face root system of tracheophytes. actually acceler- had any measurable effect on pre-Devonian climate ates the rate of chemical weathering or decomposi- or atmospheric composition? Did they enhance tion, by increasing the surface areas of minerals weathering in any way? Are there indeed pro- exposed to weathering and reaction times, while nounced differences in productivity in all environ- retarding physical weathering or disintegration. ments including those only minimally inhabitable by Schwartzman and VolkŽ. 1989, 1991, p. 360 , extrap- tracheophytes? Are only tracheophytes capable of olating from the weathering data provided by Jack- generating large volumes of sequesterable organic son and KellerŽ.Ž 1970 discounted by Berner, 1992; carbon—sufficient to cause perturbations in atmo-

Cochran and Berner, 1993a,b. , claim that terrestrial spheric CO2 content—as these models claim? Are 100 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 models that fail to consider whether pre-tracheophytic growth season environments, that dominate in tem- embryophytes and other terrestrial autotrophs are perate northern and southern forests. The global deficient in all these regards in comparison with importance of this small group of plants Ais greatly tracheophytes, adequate to deduce atmospheric CO2 increased by their disproportionate occupancy of the concentration? Even if non-tracheophytic autotrophs terrestrial land surface . . . B ŽSmith and Read, 1997, are deficient in all these aspects can it be assumed p. 167.Ž. . Smith and Read 1997, p. 422 note that that they necessarily had negligible, non-measurable ectomycorrhizal trees are natural dominants in boreal effects on atmospheric composition before the rise of and temperate biomes of the world, particularly on the tracheophytes, or after the rise of the tracheo- acid soils, while base-rich soils, even in these re- phytes, in all areas of the globe? gions, may be dominated by conifers and hardwoods that are either obligate or facultatively VA mycorrhizal. In addition to north and south temperate 21. Are tracheophytes ultimate brokers of in- regions, the geographic occurrence of ectomycor- creased mineral solubility in the rhizosphere?: rhizal phanerogams, includes sub-arctic regions, tem- some evolutionary afterthoughts perate and subtropical Australasia Ž Eucalyptus and other Myrtaceae. and at least some taxa found in VA mycorrhizae or VAMsŽ vesicular–arbuscular monsoonal and moist tropical regions of Indo– mycorrhizae, now sometimes referred to as AM or Malaysia, regions of seasonal vegetative activityŽ see arbuscular mycorrhizal fungi because of the inability Harley and Smith, 1991, pp. 107–111; Wilcox, 1996, of many taxa to form root vesiculesŽ Smith and p. 690; Smith and Read, 1997, Table 6.1. . Ectomyc- Read, 1997; Varma and Hock, 1995. include six orrhizal taxa also include some shrubs and a few generaŽ. of aseptate endophytes in zygomycetous genera of annual or perennial herbsŽ in the Aster- order GlomalesŽ all formerly included in the Endog- aceae, etc., Cyperacae, Polygonaceae. . Ectomycor- onales, an order now exclusively comprised of rhizae are formed by a large number of fungal saprobes or ectomycorrhizal associates: Smith and species, predominantly basidiomycetes and many as- Read, 1997, p. 15; see Table 1, pp. 4–8 in Smith and comycetes and a few zygomycetesŽ Reid, 1990; Smith Read, 1997; for a classification of Mycorrhizae; and Read, 1997. . Alexopoulos et al., 1996. . VAM are universally dis- Ectomycorrhizae are most intensively developed tributed, but find their greatest development in moist in or immediately under the litter zone in surface soil temperate and tropical lowlands where VAM herbs, horizons, with litter production a single consistent shrubs and trees dominate in the vegetationŽ Harley requirement in some forest typesŽ Harley and Smith, and Smith, 1983, p. 401. . They predominate in 1983; Reid, 1990; Wilcox, 1996, p. 711; Smith and ecosystems where mineralization of organic matter is Read, 1997, p. 423. . In ectomycorrhizae, the fungi rapid enough that it does not accumulateŽ Wilcox, encase the root with a structure known as a mantle or 1996, p. 712. ; as a generalization they tend to be sheath from which hyphae grow inward between the more characteristics of plants growing on mineral epidermal and cortical cells of the root but intra- soilsŽ. Smith and Read, 1997 . cellular penetration is limited in comparison with VA mycorrhizae are referred to as endophytes VA mycorrhizae. In these it is extraradial mycelia, because they invade the cortical cells of the partner extending from the mantle into the rhizosphere or forming characteristic structuresŽ vesicles, arbus- soil substrate, even if only during the growing season cules. in addition to having an external hyphal net- of the host which may have limited function in work or mycelia extending from the mycorrhizal mycorrhizal systemsŽ. Smith and Read, 1997, p. 251 , roots into the soil. Some obligately endosymbiotic that increase the surface area of the colonized root in mycorrhizal fungus themselves harbor obligately in- contact with the soil. tracellular bacteriaŽ. Bianciotto et al., 1996 . The potential absorptionŽ. and weathering area By contrast, ectomycorrhizae are associated with gained by the root system through extraradial mycelia roots of a small but important group of predomi- is enormousŽ. Smith and Read, 1997, pp. 210–211 . nantly woody perennials of seasonal or intermittent As a broad generalization Smith and ReadŽ. 1997 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 101 suggest that ectomycorrhizae in comparison with VA minimal populations of shallow-rooted tracheo- mycorrhizae, are more characteristic of plants grow- phytes, and in the well-developed, long-established ing in relatively organic soils but note that some soil rhizospheres of the type known in the Cenozoic genera may have VA mycorrhizae in either mineral inhabited by angiosperms or mixed angiosperm– or nutrient-rich soils, while ectomycorrhizae predom- gymnosperm assemblages. inate in organic soilsŽ. p. 169 . Predicting weathering effects for the Paleozoic on Of the two principal mycorrhizal relationships, varied assumptions and speculations that may not be endophytes are considered more primitive or ances- applicable throughout geologic time and in non- tral in comparison with the ectophytes and faculta- angiospermous ecosystems is highly speculative. Dr- tive mycotrophy and non-mycotrophy areŽ com- ever raised this problem as noted in his quotation monly. considered the most advanced trophic types above, the one about regolith thickness and allied Ž.Trappe, 1987a,b . It is generally assumed that the questions. In determining tracheophyte-dependent non-VA mycorrhizal, non-mycorrhizal and coloniza- weathering rates, little or no account is taken of the tion in a single host of different types of coexisting many changes in vegetation through time, nor of the mycorrhizae, has evolved a number of times in effects of varied climatic changes on distributions of different taxonomic linesŽ Trappe, 1987a,b, p. 22; Earth’s vegetation during that time. Berner agrees Smith and Read, 1997, pp. 89–90. . that this is an unknown factorŽ. his quotation above As discussed below, it is basically Aectomycorr- but has not considered potential problems of the type hizal environments,B of the temperate zone that have discussed by RobinsonŽ. 1990a,b, 1991 . provided most information about soil weathering ef- Nor for the Paleozoic and Mesozoic do we have fects, based on Arun-offB studies, such as those in the other than a skewed knowledge of the kind of vege- Andrew’s Forest part of the Oregon CascadesŽ Martin tation and range of habitats occupied, based on and Harr, 1988.Ž , the Rocky Mountains Arthur and preservational artifacts. In the Paleozoic, for exam- Fahey, 1993.Ž , the Alps Drever and Zobrist, 1992 . , ple, we know a great deal about lowland coal swamp etc., and the prevalence of plants with ectomycor- vegetation, which we might speculate was largely rhizal roots. The vast majority of the worlds plants non-mycorrhizal in waterlogged largely anaerobic are endomycorrhizalŽ. VAM , and such environments environments, but comparatively little about any have provided few weathering studies. Significantly, other ecosystems either before the Late Carbonifer- information about the weathering role of tracheo- ousŽ. Pennsylvanian or after the Late Carboniferous. phytes depends almost entirely on what we know Gymnosperm–Pteridophyte-dominated ecosystems about forest nutrient dynamics and mineral weather- like those of the Mesozoic and Pteridophyte domi- ing in ecosystems dominated by deciduous an- nated ecosystems like those of the Paleozoic, where giosperms or mixed hardwoods and conifers, e.g. in angiosperms were wholly absent, have no modern the temperate zone, or where angiosperms are signif- vegetational or ecosystem analogues and many Pale- icant in the vegetation and dominantly ectomycor- ozoic and even some Mesozoic gymnosperms and rhizal arborescent populations. pteridophytes have no closely comparable taxonomic Attributing soil weathering effects to tracheo- analogues today. Also significantly, placing sole sig- phytes because of their synergistic relations with root nificance on tracheophyte mycorrhizae in determin- fungi in the rhizosphere is widely accepted in mod- ing mineral weathering and pedogenesis, presumes a ern ecosystems. But in considering the weathering long established, long-lasting coevolution between effects of tracheophytes particularly in the Paleozoic, fungal endophytes and tracheophytes and to a large and through the rest of geologic time, it seems extent ArequiresB that mycorrhizal fungi are obli- necessary to make a distinction between biogeo- gately associated with tracheophytes in all habitats chemical and physical weathering effects that neces- throughout time. sarily occur on bare, AcleanB rock surfaces where In VAMs, regarded as the oldest, most Aprimi- tracheophytes are unable to establish populations, tiveB and most widespread of the tracheophyte-fungal inorganic regoliths that may form shallow surface mutualistic partnerships, the endophytes are Aob- cover over bedrock surfaces able to support only ligateB biotrophs, but there is a remarkable lack of 102 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 hostrendophyte specificity when compared with not species or sub-species specific, especially VAMs, other obligate, co-evolutionary relationshipsŽ Nicol- although the concept of species specific relation- son, 1975; Harley and Smith, 1991, pp. 364–365, ships, pervades most theoretical thinking on coevolu- Table 80. . VAM endotrophs belong to a single tion. The often ArelaxedB nature of the relationship order the GlomalesŽ Phylum Zygomycota, Class Zy- indicates that it is highly facultative in many envi- gomycetes. , with six endomycorrhizal genera ronments as discussed below. It tends to be more Ž.Alexopoulos et al., 1996 , but the host plants in- AhabitatB correlated than taxon correlated such that clude an enormous variety and varied growth forms external environmental characteristics seem to play a over a wide geographic distribution. As Harley and major role in regulating mycorrhizal formationŽ see SmithŽ. 1983, p. 364 report, A . . . a vesicular–arbusc- Allen, 1991, p. 36. . There is seldom a one to one ular fungus isolated from one species of host plant relationship involving a single tracheophyte taxon can be expected . . . to infect any other species which and a single fungal taxon, indeed many woody tra- has been shown to be capable of forming vesicular– cheophytes can be endomycorrhizalŽ. VAM or ecto- arbuscular mycorrhizasB permitting a combination of mycorrhizal at the same time, even though the fungi wide host range and plant–fungal compatibility with- involved belong to entirely different classes. These out specificity between particular endomycorrhizal relations may not warrant the implication of an biotrophs and particular host plants e.g. there is no obligate relationship as the term AcoevolutionB im- 1:1 relationship, e.g. permanence of association be- plies, despite the fact that many VAM endophytes tween the mycorrhizial biotrophs and a particular are obligate. host plant. The fact that almost any mycotrophic Mutualism is commonly thought of as a relation- plant can be colonized by almost any species of AM ship that is mutually advantageous to the parties fungus suggests that these fungi are the ultimate involved, as in the zooxanthellae of many marine generalists. However, Molina et al.Ž. 1992 provide a taxa. The concept of coevolution for embryophytes is useful summary of the wide spectrum of fungal complicated by the implied time of evolution of specificity and generality in relation to their hostŽ. s ; certain of the fungal lineages which are presumed to this is not a simple relationship! have formed mycorrhizal relations with early tra- Among AadvancedB ectomycorrhizae formed by cheophytes. The concept of coevolution in tracheo- basidiomycetes, ascomycetes, and species of Endo- phytes is also complicated by limited evidence for gone in the zygomycetesŽ cf. Read, 1999; Debaud et endotrophic relations with other embryophytesŽ of al., 1999. , the circumstances are different. In these which modern bryophytes are preserved as lineal ectomycorrhizae, many families of Basidiomycetes descendent. which appear now to have preceded and a number of families of AscomycetesŽ see Har- tracheophytes into terrestrial habitats and which ap- ley and Smith, 1983, p. 111; Wilcox, 1996; Debaud pear ultimately to have given rise to tracheophytes et al., 1999. can infect a single host. A single Ž.see discussion in Allen, 1991, pp. 27–29, etc. . tracheophyte AhostB may support 100s or even 1000s Duckett and ReadŽ. 1991, Table 1 compile the lim- of hostsŽ in Douglas fir for example, there are as ited records of AsymbioticB fungus recovered from many as 2000 species of associated mycorrhizal hepatic rhizoids that indicates restriction to liver- fungi: Trappe, 1977; see also Trappe, 1962, and worts except for zygomycete occurrence in Phaeo- Harley and Smith, 1991, pp. 360–361 for a list of ceros laeÕis, a hornwortŽ. Anthocerotales but present fungi that form ectomycorrhizae on species of pine. . no evidence for the nature of the relationship. Schus- A given species of fungus may form similar mycor- slerŽ. 2000 describes a Amycorrhiza-likeB symbiotic rhizae with many hosts including both gymnosperms fungus, Glomus claroideum, with the hornwort An- and angiospermsŽ Harley and Smith, pp. 360–363, thoceros punctatus, that forms arbuscules. The fact Table 79; Smith and Read, 1997, Table 6.2; Debaud that ericoid mycorrhizae, a small group of endomyc- et al., 1999. but many fungi are specific enough with orrhizae largely restricted to genera of EricalesŽ see a single host genusŽ. Debaud et al., 1999 . Smith and Read, 1997; Read, 1999; Read and Ker- As Harley and SmithŽ.Ž 1983 see also Allen, ley, 1999. are found in rhizoids of 5 families of 1991, p. 141. note, mycorrhizal symbioses are often liverworts and share the same ascomycetous myco- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 103 biont Ž.Hymenoscyphus ericae invites speculation tiated from any paleobotanical or other information that some of the nutritional and protective benefits that the uplands were not already significantly colo- accrued to the heathland ericoid hosts may also nized by tracheophytes. Arguments that support co- benefit the liverwort AhostsB, but as Duckett and evolution are less tenable than they once were be- ReadŽ. 1995, p. 447 state, Athe functional signifi- cause of difficulties in recognizing a mycorrhizal cance of the association for these plants has not been relationship. investigated.B Questions arise: What is actually known about the Although implications of some of the statements prevalence of mycorrhizal relationship in tracheo- above will be met by objections, the point here is phytes in different habitats and ecospheres and that the synergistic relationship is not one that can be biomes, apart from the angiosperms, what do we assumed as Berner has in his weathering scenario actually know even for the angiosperms, and what do Ž.Fitter and Merryweather, 1992 . Still, it seems clear we know about the AgymnospermsB and Apteridos- that the weathering effectiveness of varied microor- permsB? ganisms is enhanced by the presence of organic Does this information about modern taxa provide matter in the rhizosphere, whether or not there was a sufficient information to extrapolate to vegetation of synergistic relationship in the Paleozoic with mycor- the past? In extant groups like the Pinaceae, whose rhizal fungi. They should also be enhanced by any members are ectomycorrhizal, what can we assume source of organic carbon. Can these rhizosphere and about the Mesozoic Pinaceae, if as some have stated non-rhizosphere effects be compared or are they there is no evidence for ectomycorrhizae until the negligible enough in any environment except the Cretaceous? rhizosphere to be totally discountedŽ organic acids Is there any evidence from extant VAM– from decomposing organic matter, leachates from tracheophyte relations or from the fossil record, that leaves, root exudates, fungal exudates, bacteria, etc.. ? supports the timing of mycorrhizal symbiosis coinci- Opportunistic relationship: The question can also dent with the appearance and radiation of the tra- be raised whether there is sufficient information to cheophytes that supports Berner’s assumption about evolutionarily link the appearance of endophytic the timing of increased rock silicate weathering and

VAM fungi with the appearance of tracheophytes—a the CO2 drawdown of the Devonian? Knoll and necessary prerequisite to Berner’s and Knoll and JamesŽ. 1987 have suggested that Afirst-order in- James’Ž. 1987 conclusions about the increase in creases in overall mineral weathering,B should also weathering effectiveness associated with this group have taken place in the early Tertiary in addition to of plants. Since it seems clear that it is the Amicroen- the mid-PaleozoicŽ. but as Robinson has noted, above B vironment associated with roots, rather than roots the CO2 drawdown in the Cretaceous is trifling as themselves where weathering effects are effective in compared with the supposed Paleozoic drawdown. the rhizosphere. Some have said that roots and myc- Even if weathering effectiveness is increased in orrhizal fungi have co-evolved or evolved Aco-later- modern angiosperm-dominated environments can it allyB Ž.Nicolson, 1967 . be assumed to have been true throughout geological Pirozynski and MallochŽ. 1975 , Malloch et al. time, with varied groups of plants for which we have Ž.1980 , Pirozynski Ž. 1981 and Pirozynski and Dalpe no or few extant members? In part to attempt to Ž.1989 suggest mycorrhizism with the first tracheo- answer that question, it is important to examine what phytes of the Devonian, lending credence to the we know about the occurrence of mycorrhizal rela- supposed spectacular Late Devonian drawdown in tions in extant plants, including those with links, if atmospheric CO2 in Berner models that he has tied any, to taxa that dominated during the Paleozoic and to an hypothesized dramatic increase in weathering Mesozoic. Unless we assume that the relationship is potential related to the adaptive radiation and habitat an obligate one that commenced with the earliest expansion of presumed mycorrhizal tracheophytes. tracheophytes and has been maintained in that lin- The increase in the area of colonization of tracheo- eage through time, as some have hypothesized, it phytes in the Devonian Ainto the uplandsB which is makes sense to ask whether figures for extant taxa part of this scenario, is speculative. It is not substan- can be extrapolated into the geological past. 104 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

21.1. Extant VAM–tracheophyte relations continent and all ecosystems, found relevant information on the mycorrhizal status for about 1% of Fitter and MerryweatherŽ. 1992 asked Awhy are the total number of Earth’s vascular plants or less some plants more mycorrhizal than others?B There than 3000 species. This information included an- are remarkably few comprehensive studies of the giosperms, gymnospermsŽ unspecified other than extent of mycorrhizal relations in the world’s tra- Pinaceae.Ž and Pteridophytes unspecified, except for cheophytes. There are numerous extrapolations based the fern families Hymenophyllaceae and Aspidi- on a meagre and commonly angiospermous only data aceae. . In any taxon the percentage of species with base, although the latter information is not always mycorrhizal statusŽ meaning that at a site at least made clear. These meagre compilations provide the some plant roots, percentage unspecified, were myc- source for various estimates of the prevalence of orrhizal, but not necessarily all. ranged from 100% VAM mycorrhizae in extant tracheophytes. These to 8% percentage of species with non-mycorrhizal estimates run the gamut from about 70–95% of all status from 87% to 2%, while a sizable proportion of plant speciesŽ Fitter and Merryweather, 1992; Trappe, speciesŽ. 28–2% in some families had species that 1987a,b. and permit statements that AMycorrhizae, were mycorrhizal at some sites and non-mycorrhizal the symbioses between fungi and plant roots, are at others. Their information also demonstrated that nearly universal in terrestrial plants . . . B ŽMalloch et taxa were clearly segregated into ectomycorrhizal or al., 1980, p. 2113.Ž . As Trappe 1987a,b, p. 6 . points vesicular–arbuscular mycorrhizal, and those capable out, such unsupported and potentially erroneous of forming both in addition to other smaller groups statements are commonly extrapolated from the very of mycorrhizal types. different estimate that Aabout 95% of the world’s TrappeŽ. 1987a,b presents data based on 3000 present species of vascular plants belong to families papers, theses and booksŽ Trappe, 2000, written that are characteristically mycorrhizal . . . B All that is communication, now about 6000, but with the same really known are figures based on broad compila- caveats. without respect to the types of limitations tions of the evidence for mycotrophy in families and specified by Newman and Reddell, that tabulates orders of tracheophytesŽ. Trappe, 1987a,b . More- mycorrhizal relations for a limited number of species over, some have argued that Amycotrophyrnon- Žca. 6500 species or about 3% of the worlds an- mycotrophic is an expression of ecological adap- giosperm species. in a limited number of subclasses tation and there may be no consistency within and orders of monocots and dicots. As he notes a designated families or even generaB ŽAllen, 1991, p. large proportion of the species examined have been 32. . Thus, whether some taxa are mycorrhizal or studied Aonly from a single specimen.B He also non-mycorrhizal may depend on the specific edaphic notes, that the more a species is studied, especially conditions where they are growing. for annuals, the Amore likely it will be found to be a It is important to understand the statistical base on facultative mycotroph in at least some kinds of habi- which these and similar figures rest and whether this tatsB Ž.1987, p. 12 . As he also makes clear there are data base is largely angiospermous or includes a many groups that have been disproportionally exam- representative sampling of other tracheophytes. It is ined. It is also clear that there are many ecosystems also important to know whether the relationships that have been disproportionately examined. occur in all tracheophytes with similar abundance These actual figures are clearly a minute sample and whether it is sufficiently obligate to justify the of tracheophytes. Although they might be seen to statement that the Apresence of vascular plants and support the belief that mycorrhizae are Anearly uni- associated mycorrhizal fungi have probably served to versalB in tracheophytes, the documented facultative accelerate the weathering of soil mineralsB through- nature of the symbiosis and the variability from site out all of geological timeŽ Knoll and James, 1987, p. to site and in different habitats, renders that an 1100. . untenable conclusion. Newman and ReddellŽ. 1987 The compilation by Newman and ReddellŽ. 1987 caution that the small sample provided by them is of post-1960 literature, excluding crop plants, tree inadequate for anything but cautious generalization seedling and pot-grown plants, and covering every about the extent and commonness of fungal symbio- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 105 sis in tracheophytes, including generalizations within ditions including water-logging that is commonly a family for assumptions about the mycorrhizal or characterized by non-mycorrhizal taxa including non-mycorrhizal status of a taxon. some that are normally mycorrhizal in other circum- In addition to the inadequacy of the sample for stances. considering the role of the tracheophyte–mycorrhizal 21.2. Extant endophytic fungal–angiosperm relations symbiosis in mineral-weathering effectiveness, it should be emphasized that in largest part much of Is angiosperm-based information, where the VAM the information available about organic acid produc- relationship is best known, necessarily applicable to tion and physiological function associated with myc- the vast majority of tracheophytes–pteridophytes and orrhizal symbiosis is extrapolated from angiosperm various groups loosely included in the gymnosperms, VAM relationships, and from comparative weather- which preceded them in geological time? Can it be ing values for angiosperm-dominated floras. assumed that the weathering effectiveness of the Smith and ReadŽ. 1997, pp. 28, 446–447 point angiosperms—based on the knownrapparent preva- out that there are a number of habitats that are lence of mycorrhizal relationships among them— characterized by a relatively high proportion of provide a valid basis for projecting the weathering Aconsistently non-mycorrhizal species or by species effectiveness of all tracheophytes that preceded them that sometimes occur in the non-colonized stateB, based on the necessarily speculative implication for a e.g. where the relationship is clearly facultative. mutualistic relationship? Indeed, according to Bon- Habitats that they enumerate in this category include: fante-FasoloŽ. 1984, p. 28 it is only in angiosperms very moist or arid; highly disturbed, particularly in that all morphological and functional criteria for the early stages of succession; and very nutrient-rich identifying VAM can be satisfied. However, Trappe soils. TrappeŽ. 1987a,b, Table 4, p. 22 notes that Ž.written communication, 2000 points out Athat VAM xerophytes tend to have a high incidence of both VA abounds in gymnospermsŽ Cupressaceae, Taxaceae, mycorrhizae and facultative or non-mycorrhizal taxa. Cycadaceae, etc.. . Often it is hard to find arbuscules, Primary successional habitats such as newly exposed because the fungi seem to form coils in these hosts glacial till in the Oregon Cascades are colonized by instead, but the same fungi do form arbuscules in ectomycorrhizal conifersŽ Cazares, 1997, found that other plants.B Bonfante-FasoloŽ. 1984 also states that the initial primary succession on a glacial forefront there is a high correlation between mycorrhizal colo- in the North Cascades was of non-mycorrhizal plants, nization and frequency and length of root hairs such followed soon by ectomycorrhizal willows and that plants with abundant fine rootlets and long root conifers. . At Krakatoa, following eruption, early hairs have the ability to grow without mycorrhizal colonists were non-mycorrhizal or facultatively myc- fungi. orrhizal species. In the primary succession on sand Within the handful of angiosperm families exam- dune ecosystems at the drift line non-mycorrhizal inedŽ. 23 in Newman and Reddell, 1987 , endotrophic species predominate. There is no doubt that on a fungal infection is widespread but not universal, and global scale, in the absence of disturbance, biome-re- VAM relationships are rare, absent or facultative in lated segregation of predominantly mycorrhizal types many important, primarily herbaceous, angiosperm can be seen even though a given type rarely, if ever, families, but also in the Southern Hemisphere Pro- occurs to the exclusion of all others. The extent and teaceae with over 1000 species, and in certain kinds nature, if any, of the involvement of the mycorrhizal of habitatsŽ Powell and Bagyaraj, 1984, p. 1; Mal- symbiosis in determining these observed patterns loch et al., 1980; Pirozynski, 1981; Allen, 1991. . remains to be investigated by experimentŽ Smith and Pioneer, or colonizing weedy angiosperms of dis- Read, 1997, p. 450; Read, 1984; Allen, 1991. . turbed sites in lowland, humid tropical ecosystems, Biomes that these authors include with a low inci- both dicots and monocots, but including a number of dence of VAMs include tundra, high alpine habitats invasive forest trees that may be facultatively mycor- even including species in these environments known rhizal-forming either VA mycorrhizae or ectomycor- to be able to form them elsewhere. They reportŽ p. rhizae, for example, almost always lack mycorrhizae 28. variability associated for example, with soil con- with some few exceptionsŽ. Janos, 1987, p. 120 . At 106 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 least one angiosperm Ž.Gunnera of the Halagoraceae 108.Ž . Trappe written communication, 2000 . empha- has a symbiotic relation with the cyanobacterium sizes that AA lot has been done since 1991Ž 1000q atmospheric nitrogen-fixing Anabaena ŽPeters and papers on mycorrhizae have appeared each year Calvert, 1983. that also occurs with a broad, but since. . Most of the earlier generalizations have been restricted number of embryophytes. confirmed by the recent work. Still, only a very Some examples of variability can be provided. small proportion of the earth has been explored for Khan and BelikŽ. 1995, Tables 1 and 2 total the wide this.B variation of mycorrhizal infection not only in gen- Fitter and MerryweatherŽ. 1992 note that the ex- uinely aquatic ferns and angiosperms but in taxa tent of VAM infection varies widely within and subject to temporary inundation or living in tem- between ecosystems with a few types of ecosystems porarily water-saturated soils. Such variations in- apparently supporting Apredominantly non-mycorr- clude differences in root infection in single plants hizal plants.B Within ecosystems, drawing examples depending on whether the roots were growing to- from angiosperm taxa, they note that variation exists ward water on river banks, where they lacked mycor- Awithin habitats, within species and even within rhizal symbionts, or growing away from water where individualsB with a great range in infection density they were colonized by both ecto- and endomycor- in roots of co-existing species even on a seasonal rhizal fungi. According to AllenŽ. 1991, pp. 89–90 basis and considerable variation at any one time in facultative mycotrophic plants that can survive with closely adjacent plants of well-infected species. This or without mycorrhizae Aare probably the most com- variation includes, as they note, variation in infection mon group of plants . . . B TrappeŽ written communi- density between parts of the root systems of an cation, 2000. states that mycorrhizal plants are the individual plant. NicolsonŽ. 1967 also notes that Amost common group of taxa,B with the reservation amounts of infection can vary from total incidence to that this conclusion is sample dependent. Malloch et only a few roots within the root system and remarks al.Ž. 1980, p. 2114 state that both facultative my- on the difficulty of obtaining measurements of Ain- cotrophism and the non-mycorrhizal condition are fectionB especially in field conditions and especially correlated with herbaceous habit, a short life cycle, in these circumstances among tree species. Janos and Aevolution of root systems with abundant root Ž.1987, p. 120 also notes, as another example of hairs,B although there are notable exceptions. Cooper within ecosystem variation, that in lowland tropical Ž.Ž1975 citing relevant literature . also drew attention forests, VAM appear to be common, but that VAM to the possible correlation between copious, long in trees of humid tropical forests at high elevations persistent root hairs and well-developed, finely may be facultative or absent although too little infor- branched root systems and more limited dependency mation is available from tropical regions in compari- on mycorrhizae. son with temperate zone studies to make sweeping AllenŽ. 1991 and Smith and Read Ž. 1997 discuss generalizations. variations in distribution of mycorrhizae in different Harley and SmithŽ. 1983, p. 110 report that for biomes and differences in knowledge of mycorrhizal Central America, ectomycorrhizal forest trees are to relationships in different biomes, primarily con- be found outside of the region of 258 north and 35.58 cerned with spatial variations of VAM and ectomyc- south of the equator at all altitudinal levels, but orrhizae. It is clear that at some elevations and in within that zone they occur only at higher altitudes some vegetation types VAM dominateŽ grassland and on the equator, ectotrophs only occur above and deserts, for example.Ž . In others boreal or sub- 3000 m. TrappeŽ. written communication, 2000 notes alpine forests, for example. ectomycorrhizal trees that AAsia and Australia have many ectomycorrhizal dominate. But as Allen notes, information about the hosts at low as well as high elevations, Australia, global distribution of mycorrhizae is extremely lim- New Guinea and New Caledonia being examples.B ited and while some information is available for As Fitter and MerryweatherŽ. 1992 suggest, this temperate regions, there is insufficient work in the tendency for a range of infection to appear to all tropics and high northern and southern latitudes Ato scales poses a problem in determining whether the make even the most basic of generalizationsB Žp. intensity of infection is highly controlled and implies A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 107 tight mutualism or whether it is AstochasticB with Allen is apparently stating that mycorrhizal fungal benefits more sporadic or more diffuse. propagules are virtually eliminated or reduced fol- Peterson and BradburyŽ 1995 pp. 164–165, 174– lowing disturbanceŽ. p. 129 , and as the site recovers 175, Table 1. note that degrees of non-mycotrophy it shifts from non-mycotrophic to facultative to obli- may exist between and within species that are nor- gateŽ. p. 129 —a model since modified Ž. p. 136 ; in mally mycotrophic; they implicate genotype differ- the Galapagos notes that non-mycotrophic plants in ences, experimentally demonstrated in cultivars, to richer, lowland soils and VAMs in pyroclastic mate- explain variations in fungal colonization levels in rial; after the Krakatoa eruption the first plants to normally mycotrophic species. appear were predominantly facultative VAMs, but Some specifically regard non-mycotrophy in the later obligateŽ. p. 130 ; p. 131, the status and succes- extant angiosperm flora as a specialized, derived, sion in different biomes and relative mycotrophy of condition related to Ahabitat specialization.B Žcf. plants can vary with moisture and nutrient statusŽ his Allen, 1991, p. 33.Ž. . Trappe 1987a,b suggests that Fig. 7.2.. the facultative or non-mycorrhizal condition relates The question has been raised whether angiosperm mostly to plants considered advanced and that these floras have the potential to accelerate weathering conditions also prevail with invasive weeds. It has compared with Aconifer–evergreenB floras. Knoll and also been suggested that non-mycotrophy is charac- JamesŽ. 1987, p. 1101 argue that the advent of teristic of early successional habitatsŽ Allen, 1991, deciduous angiosperms Amay have had a significant pp. 31–32, 129. particularly in highly disturbed en- impact upon mineral weatheringB because of what vironments and among colonizing annuals of ad- they suggest are major differences in Anutrient dy- vanced angiospermous families a late evolutionary namicsB related to seasonal leaf loss and annual net condition evolving from the conservative VAM con- losses of potassium, manganese and calcium in com- ditionŽ. p. 32 . Others Ž cf. Malloch et al., 1980, p. parison with conifer–evergreen ecosystems. Knoll 2116. have raised similar arguments with regard to and JamesŽ. 1987, p 1101 conclude that post-Creta- the ectomycotroph relationship, noting the preva- ceous soil minerals should be more weathered at lence of ectotrophs in various AmarginalB habitats least in the Northern Hemisphere. They also argue and extreme environments. But Allen alsoŽ pp. 32– that the origin and diversification of grasses in the 33. emphasizes that Amycotrophy is a feature com- Cenozoic Amay have supplemented such a weather- mon to habitats that presumably most resemble early ing increase.B The caveat here is that there are a terrestrial environments,B an enigmatic statement, large number of evergreen angiosperms, even in the since these could be viewed as disturbed early suc- temperate zone, and a large number of seasonally cessional habitats, and the early land plants as colo- deciduous gymnospermsŽ. Robinson, 1990a, 1991 . nizing annuals, both characteristics that in the mod- 21.3. Extant endophytic fungal–gymnosperm rela- ern environment have led to the non-mycorrhizal and tions facultative mycorrhizal conditions. AllenŽ 1991, p. 33. states that on disturbed, fertile habitats, such as Taxa customarily included in the AgymnospermsB Mt. St. Helens they were facultative in plants and Ž.conifers, ginkgo, cycads, etc. like angiosperms, this could not be viewed as a fertile site. On p. 128, appear to have the characteristic VAM infection however, in apparent contradiction, Allen notes that Ž.Bonfante-Fasolo, 1984; Smith and Read, 1997 but Amycorrhizae and organic matter accumulation dur- some, like Pinaceae, are ectomycorrhizalŽ with many ing succession are tightly coupledB–thus in early characteristics of ectomycorrhizae but exhibiting successional coastal sand dunes, plants tended to be some intracellular penetration: Smith and Read, 1997, non-mycorrhizal, but as soil development proceeded p. 290. . But the literature presents a confusing pic- and organic matter increased, mycorrhizal activity ture of how common fungal infection is among increased. Allen also noted a study by Frydman gymnosperms, making it difficult to say how Ž.1957 where the first invading weeds in Wroclaw widespread fungal symbiosis is in extant taxa let ruins following WWII were non-mycorrhizal, that alone how widespread it may have been in the Late VAMs rapidly developed in the newly forming soil. Paleozoic and Mesozoic, when gymnosperms, in- 108 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 cluding a variety of extinct taxa, and pteridophytes, coniferales, and at least in Gnetum Ž.ŽGnetales Mal- dominated the landscape. loch et al., 1980, p. 2113. , ectomycorrhizal sym- Bonfante-FasoloŽ. 1984, pp. 26, 28 stated that bionts are well developed and sometimes exclusive documentation of fungal infection in Agymnos- Ž.Smith and Read, 1997, p. 163 including 95% of the permsB is poor and Aconcern a limited number of 100 species examined by Newman and Reddell species.B, but TrappeŽ. written communication, 2000 Ž.1987 ; some are sometimes ectendomycorrhizal Ž see comments that Athe pervasive mycorrhizal coloniza- Smith and Read, 1997.Ž ; others 2% of the 100 tion of gymnosperms is more firmly established by species examined by Newman and Reddell, 1987. far than in the 1980s.B Harley and SmithŽ 1983, p. form both ecto- and VA mycorrhizae, with the latter 19. state that most conifers, as well Aas . . . most more typical for these taxaŽ Smith and Read, 1997, other gymnosperms,B are dominantly infected by Table 6.1, pp. 167–170. , implicating a facultative VAM fungus. Wilcox’sŽ. 1996, p. 689 statement that capacity depending on local circumstances. Thus, AAll gymnosperms are reported as being mycor- Smith and ReadŽ.Ž 1997 see also Harley and Smith, rhizal . . . B incorrectly citing the compilation of New- 1983. caution, as did Newman and Reddell: AThe man and ReddellŽ. 1987 echoes Harley and Smith. record of the presence of ectomycorrhizal individuals Of the 46 unidentified species of Agymnosperms,B in a genus does not mean that all species are or may examined by Newman and ReddellŽ. 1987 , excluding be ectomycorrhizal, nor does it mean ectomycor- Pinaceae, 98% were mycorrhizalŽ 89% VAMs, 9% rhizal colonization is necessarily consistently or even ectomycorrhizal, 2% both VAM and ectomycor- normally present in any species of that genus.B rhizae. and 2% were non-mycorrhizal. Among VAM Malloch et al.Ž. 1980, p. 2115 state that ectomycor- AgymnospermsB with ancient alliances is Ginkgo rhizal phytobionts appear to take up mycobionts in a biloba, although the pattern of colonization is differ- selective manner related to such things as develop- ent from that found in angiospermsŽ Bonfante-Fasola mental phase, ecological conditions and even cli- and Fontana, 1985. . Non-mycorrhizal taxa include matic fluctuations. Cazares and TrappeŽ. 1993 recog- cycads Awhose association with the nitrogen-fixing nize vesicular endophytes in roots of the Pinaceae. blue-green alga Anabaena appears to be of great The ectomycorrhizal symbiosis in PinaceaeŽ and antiquityB Ž.Malloch et al., 1980, p. 2115 . Perraju et in some genera of Cupressacae. is anomalous and al.Ž. 1986 discuss the Cycas–Anabaena symbiosis unexplained. The earliest documentation for ectomy- and Peters and CalvertŽ. 1983 report that nine genera corrhizal symbiosis in conifers, indeed the first un- of cycadsŽ including Cycas, Zamia, Macrozamia, equivocal fossil evidence for ectomycorrhizae, oc- Encephalartos. are cyanobacterium symbionts. The curs in an Eocene Pinus Ž.LePage et al., 1994, 1997 cyanobacterium in these symbiotic associations is which may provide a minimal or maximal bench- always a Nostocaceae, a family with the ability to fix mark for infection. Molecular clock evidence sug- atmospheric N2 and provide a source of nitrogen for gests that infection in Pinaceae could have come late its symbiotic partnerŽ. Peters and Calvert, 1983 . in the MesozoicŽ. Berbee and Taylor, 1993 even Hedges and MessensŽ. 1990, p. 141 state that it is though Coniferales extend well back into the Late Nostoc and not Anabaena; MettingŽ. 1990, p. 357 Paleozoic and Pinaceae, a dominant group in the says that the phycosymbiont is thought to be more Mesozoic, into the Triassic. Berbee and TaylorŽ 1993, properly categorized as Nostoc; both are members of p. 1125. indicate an origin and radiation for the the NostaceaeŽ. Peters and Calvert, 1983, p. 109 . holobasidiomycetes, that include both aggressive However, a possible endomycorrhizal fungus has wood rotters and the major ectomycorrhizal sym- been reported in the Triassic cycad Antarcticycas bionts in PinaceaeŽ. Smith and Read, 1997, p. 163 , Ž.Stubblefield et al., 1987a,b , suggesting that the in the Cretaceous about 130 Ma ago although fossil relationship may have been facultative in this impor- holobasidiomycetes are known from the late Middle tant Late Paleozoic and Mesozoic group, or that the JurassicŽ.Ž. ca. 165 Ma . Trappe 1987b argues from relationship was not symbiotic. paleobiogeographic principles and dispersal of hypo- In Pinaceae, and apparently in some of the Cu- geous forms by animals thatŽ epigeous ancestral to pressacaeŽ. Malloch et al., 1980 , unlike other hypogeous or holobasidiomycetes. originated more A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 109 than 50 million years ago, before major separation of variety of ferns: typically in saprophytic or massive the continents, but not before the Cretaceous suggest- gametophytes; more sporadically in the leptosporan- ing, in conformity with the molecular data, that giate ferns. In the indigenous pteridophyte flora of infection in the Pinaceae is comparatively modern– New Zealand in addition to predominantly phy- see discussion in Smith and ReadŽ. 1997, p. 163 . comycetous fungi, CooperŽ. 1976 finds fungal en- Based on this combined evidence, can we assume dotrophs that are normally associated with Ericaceae that Pinaceae were non-mycorrhizal during much of or Orchidaceae and complete or partial replacement their Mesozoic history? TrappeŽ written communica- by ecotrophs where ferns are the only undergrowth tion, 2000. suggests that they may well have been in areas of pure stands of native beech Ž.Nothofagus VAM based on the work of Cazares and Trappe or exotic pine that are normally ectomycorrhizal Ž.1993 . apparently involving, however, only a single species Knoll and JamesŽ. 1987, p. 1101 have argued, as of fern. In the same forest types, but where an noted above, that mineral weathering rates should be understory of mixed broad-leaved species occurred, more limited in conifer–evergreen ecosystems than ferns were ectomycorrhizal, endomycorrhizal or a in ecosystems dominated by angiosperm–deciduous mixture. forest. Based on this type of inference, theyŽ p. Among pteridophytes—like gymnosperms a dom- 1102. conclude that Afor any given length of weath- inant Paleozoic and Mesozoic group—information is ering time, post-Cretaceous soil minerals should be equally confusing with regard to the commonness more weathered.B Due to major variables not consid- or constancy of the presence of mycorrhizae ered in their blanket characterization of modern Ž.Bonfante-Fasolo, 1984, p. 24 , but the relationship ecosystems into Aangiosperm–deciduousB and Acon- is by no means universal or consistent. Bonfante- ifer–evergreenB this conclusion may be unsupport- Fasolo writesŽ. p. 24 : A . . . fungal endophytes pass able, but it emphasizes our conclusion that it may be from an obligate state in Psilotaceae and Lycopodi- untenable to extrapolate from the mineral weathering aceae to a constant one in eurosporangiateŽ. sic rates evident in modern angiosperm floras to those ferns, a facultative one in some leptosporangiate that prevailed in the geological past. fernsŽ. Polypodiaceae, Pteridiaceae, etc. , and in oth- ersŽ those belonging to aquatic families Azollaceae, 21.4. Extant endophytic fungal–pteridophyte rela- Marsilaceae, etc.. they are completely absent.B How- tions ever, TrappeŽ written communication, 2000, comments that AAzollaceae have since been reported as There are no extant pteridophyte-dominated at least facultatively VAMB since 1984. Cooper ecosystems comparable to those of the Paleozoic. Ž.1975 states, however, from survey of leptosporan- Knoll and JamesŽ. 1987 provide no statistics compar- giate ferns in New Zealand, that Awith few excep- ing nutrient dynamics and mineral weathering for tions, they are constantly mycorrhizal in the field.B a pteridophyte-dominated ecosystem with later But experimental data demonstrated that available angiosperm–deciduous ecosystems or conifer–ever- soil phosphorusŽ and possibly other soil properties, green ecosystems that permit a comparative assess- not tested such as diffusion rate. was critical to ment with later vegetation types. BoullardŽ. 1979 whether rhizomatous ferns such as Pteridium and reviewed many aspects of pteridophyte mycorrhizal Histiopteris could grow as well without mycorrhizae relations, chiefly based on data from the present. as with them. Citing the BoullardŽ. 1979 study, The symbionts in pteridophytes are or appear to Malloch et al.Ž. 1980, p. 2114 go further to state that be characteristically VA mycorrhizae based on mor- some eusporangiate and most leptosporangiate ferns phological resemblance alone. They are found in the are facultative mycotrophs, while the non-mycorr- roots or rhizomes of pteridophytes as well as in the hizal Isoetaceae instead harbor symbiotic nitrogen- gametophytes of lycopods and Psilotales with rhi- fixing Anabaena Ž.or more correctly Nostoc ,a zoidsŽ see Pocock and Duckett, 1984, 1985a,b; Pe- cyanobacterium, as do species of Azolla ŽPerraju et terson et al., 1981; Bonfante-Fasolo, 1984. . Cooper al., 1986; Peters and Calvert, 1983; Metting, 1990. Ž.1976 reports infection at the prothallial stage in a where uniquely the endophyte is retained throughout 110 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 the life cycle implying a true and specialized mutual- 1980. notes, endotrophic symbiosis is both rare and ismŽ. Peters and Calvert, 1983 . atypical in Equisetaceae a taxon that falls into nei- According to Bonfante-FasoloŽ. 1984 , the infec- ther categoryŽ. advanced or aquatic though the last tion process in pteridophytes differs from species to scion of a formidable Carboniferous group. As re- species and from individual to individual within the viewed by Pirozynski and DalpeŽ 1989, pp. 10–12, same species. Thus, some ferns in some regions, if Table 1. much of the fossil information supporting they develop mycorrhizae at all, have been shown to Carboniferous mycorrhizal endophytes including be infected by VAMs. Cryptogams in habitats domi- those in the Carboniferous arborescent progenitors of nated by obligately ectomycorrhizal trees are more the Equisetaceae, is debatable or inconclusive. commonly ectomycorrhizal; thus the same taxon In pteridophytes, perhaps more significant than might be ectomycorrhizal when it is the undergrowth apparent morphological evidences for loss of infec- in conifer forests but mixed ectomycorrhizal and tion is the actual nature of the relationship. Thus, VAM in mixed conifer and broadleaf forestsŽ Harley Bonfante-FasoloŽ. 1984, p. 25 states ACytological and Smith, 1983, p. 107; Malloch et al., 1980, p. and physiological data are not yet sufficient to deter- 2114. Of the 100 species of Hymenophyllaceae and mine whether pteridophytes are only colonized by 30 species of Aspidiaceae examined by Newman and typical VAM fungi or whether the kind of interaction ReddellŽ. 1987 , 100% and 97%, respectively, were the ferns established with the endophytes is a mutu- mycorrhizal and predominantly VAMŽ 100% and alistic oneB like that found in angiosperms. If it is 93%, respectively. . the synergistic effects promoted by the mycorrhizal PirozynskiŽ. 1981 sees the evidence for a gra- relation in promoting weathering effects, this might dation from obligate to non-mycorrhizal in pte- be critical in a pteridophyte-dominated ecosystem, ridophytes as epitomizing the trend toward the but would this make any difference in organic acid Aevolution of phytobiont self-sufficiencyB with the production if much of it is fungal? progressive reduction in extent and duration of en- In view of the variables related to major vegeta- dotrophic symbiosis Ain what are taken to be more tional changes and to assessing the incidence of advanced groups, ending in a non-mycorrhizal condi- mycorrhizal fungi in fossil and extinct plants, plant- tion in some Pteridaceae, as well as in the aquatic induced weathering trends one might expect to see families and Isoetaceae.B TrappeŽ. 1987a,b, p. 23 through geologic time, are difficult to predict. It is finds no evidence for such a progression in the impossible, of course, to rule out specialization and angiosperms which he surveyed rather finding the habitat specialization leading to widespread non- present Aend pointB to be AB mycotrophyŽ involving mycotrophy in gymnosperms, past and present, com- ascomycetaceous and basidiomycetous mycorrhizae. . parable to what we now see in the angiosperms and TrappeŽ. 1987a,b, p. 22 concluded AThe second ma- pteridophytes, which could have led to many non- jor hypothesis, that facultative mycotrophy and au- mycorrhizal relationships. Thus, it seems clear that totrophy are the most advanced trophic types is the evolutionary assessment of endomycotrophy in rather well born out by the data presented . . . B and tracheophytes, especially those that were important addsŽ. written communication, 2000 AThe hypothesis in the geological past prior to angiosperm appear- of a progression from primitive through more ad- ance, is necessarily speculative and poorly docu- vanced fungal associations to autotrophy is not born mented, based on what can be understood for mod- out Afor most groupsB but some seem to have gone ern taxa. directly to autotrophy without the intervening stage. The Aend pointB of ABŽ. i.e. asco-basidio mycor- 21.5. Extant endophytic fungal–bryophyte relations rhizae applied only to the HamamelidaeŽ also to the Pinaceae, but they were not included in that paper. . Is there reason to assume that the relationship of The distinction is important, because the point of the fungi with tracheophytes is the first evidence for paragraph deals with whether autotrophy is an end- symbiosisrmutualism in embryophytic land plants? point. I concluded that for most groups I looked at it What about synergistic effects related to a symbiotic is.B But, as PirozynskiŽ.Ž 1981 also Malloch et al., fungal relationship in bryophytes? Could such a rela- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 111 tionship have triggered increased weathering effects of Lepidoziaceae, Calypogeiaceae, Adelanthaceae, like those prophesizedrpredicted for the tracheo- Cephaloziaceae and Cephaloziellacae. In the latter phytes? Its also worth noting that the advent of families, the ascomycetous fungal endophyte also bryophytes in terrestrial habitats would have pro- forms ericoid mycorrhiza in hair roots of a number vided a carbon source that could have precipitated a of major ericaceous genera that commonly occupy population explosion of free-living bacteria and fungi the same habitats as the hepaticsŽ Duckett and Read, with potential mineral weathering effects. It should 1995. . also be noted that bryophytes commonly form a As in other groups of plants, hepatic mycotrophy symbiosis with Anabaena Žor other nitrogen-fixing is commonly facultative depending on the environ- cyanobacteria.Ž see Schuster, 1966; Perraju et al., ment of growthŽ none are known, for example from 1986. that could have played an analogous role to ephemeral, epiphytes or taxa from wet habitats. or fungal symbiosis. Peters and CalvertŽ. 1983 say such the season. BoullardŽ. 1988, p. 112, Table 1 notes a bacterial symbiosis occurs in Anthoceras, Blasia, for example, that mycothalli are more common in the and CaÕicularia. LigroneŽ. 1988, p. 99 indicates that xerophytic Marchantiaceae than they are in the he frequently found intercellular cyanobacterial drought-intolerant Monocleales or Sphaerocarpales colonies, in addition to endophytic fungi, in the but association with hepatics of seasonally arid re- hornwort Phaeoceros. gions cannot be assumed since it is facultative or The evidence for endosymbiotic VAM relation- unknown in some such taxaŽ. Boullard, 1988, p. 121 . ships between bryophytes and zygomyceteous fungi His Table 1Ž. pp. 116–117 shows a systematic distri- and between bryophytes and ascomycetous fungi bution of hepatic mycotrophism, provides frequency presents some of the same interpretive problems information for hepatic orders and families and makes posed by the pteridophytes, based both on occur- the point that the association is Aregular or frequentB rence and function. SchusterŽ. 1966 , Pocock and in many families, but only AoccasionalB in others, DuckettŽ. 1985a,b and Boullard Ž. 1988 summarize sometimes in the same taxa. while some hepatics are the extensive literature that indicates mycorrhizal-like consistently free of fungiŽ. Boullard, 1988, p. 121 . fungi commonly associated with some hepatic game- Boullard notesŽ. p. 121 that in some species, fungal tophytes and Duckett and ReadŽ. 1991, 1995 and interactions may be either favorable or harmful de- Read et al.Ž. 1991 document rhizoid–ascomycete pending on the specific growth conditions and fungal associations for leafy liverworts in a number of parasitism is known. Duckett et al.Ž. 1991 report that jungermannialian families in some habitats. While liverworts isolated from their normal substrate are some hepatics are consistently free of infection ac- Afree from rhizoid infectionsB, but re-establish them cording to BoullardŽ. 1988 , Schuster Ž. 1966 notes when returned to their Anative substrata.B In the that Aendophytic fungi occur in large numbers in the hornwort Phaeoceros laeÕis they noteŽ p. 243 and majority of genera of Hepaticae . . . B although the citing Ligrone, 1988; Ligrone and Lopes, 1989. that relationship is facultative rather than obligatory. arbuscule-forming endophytes are not only fre- Duckett and ReadŽ.Ž 1991 also Duckett et al., quently absent from the gametophytes of wild plants 1991.Ž. report association of Zygomycete VAM , As- Žthey are totally absent from the rhizoids: Ligrone, comycete and Basidiomycete fungal hyphae with 1988, p.94. , but that fungi are noticeably more plen- both hepatics and hornworts: Zygomycetes with tiful during particular seasons. Ligrone and Lopes Metzgeriales, Marchantiales and Anthocerotales; As- Ž.1989, p. 425 report that in the hepatic Cono- comycetes with six families or 3 of the 14 orders of cephalum conicum where fungus infect both gameto- Jungermanniales; and endosymbiotic Basidiomycetes phytes and rhizoids, wholly independently of the with 3 families of 2 suborders of Jungermanniales nature of the substrateŽ. whether soil or bare rock and one family of Metzgeriales. Duckett et al.Ž. 1991 that frequency of gametophyte infection varied from confirm mycorrhizal association for 46 or about 16% 20% to over 90% of the specimens examined and of the 284 known species of British liverworts. In the was higher in spring and summer, times of most Jungermanniales Ascomycete hyphae are found in active growth, compared with the rest of the year. rhizoids of both tropical and temperate species According to LigroneŽ. 1988 , the fungal endophyte 112 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 of Phaeoceros appears to be closely related to the evolved independently potentially associated with phycomycetous EndogonaceaeŽ now Glomeraceae, colonization of varied environmental habitats, or VAM mycorrhizae of vascular plants. as does the whether it could represent an ancestral trait poten- infection in Conocephalum ŽLigrone and Lopes, tially associated with the extreme-marginal environ- 1989. . Endomycorrhizae are associated with the ments of terrestrialization in the Early Paleozoic. achlorophylous Cryptothallus mirabilis ŽSchmidt and Variations of both positions have been espoused. The Oberwinkler, 1993, as cited in Smith and Read, question is potentially important in considering 1997, p. 49; Duckett and Read, 1995. . whether mycotrophy may have been indigenous in There is no evidence for similar fungal relation- the earliest tracheophytes as Pirozynski and others ships with mosses to that recorded for hepatics. have espoused. Bonfante-FasoloŽ. 1984 states that reports of VAM Thus, noting the frequent association of hepatic infection in mosses Aare even scantierB than in hep- gametophytes with fungi, BoullardŽ. 1988, p. 107 atics, and where identified Ais not a true symbiosisB. argues for a longtime coevolutionary relationship He concludesŽ. p. 28 AThere are no characteristic between fungi and hepatics as Aa significant factor in mycorrhizae, with a mutualistic interaction between the evolutionary history of the Hepaticae.B HeŽ Boul- host and fungus.B A similar conclusion was reached lard, 1988, p. 122.Ž suggests that constant e.g. obli- by Duckett et al.Ž. 1991, p. 245 who write: AAl- gatory. mycotrophy in Hepatics may be an old, though a wide range of fungi growing on mosses AprimitiveB character what may have played a signif- have been identified . . . there is no evidence whatso- icant role in their evolution and biology, although ever that any of these form symbioses like those many modern hepatics are entirely free of fungi found in vascular plants and hepatics.B According to while others are Asomewhat plastic vis-a-vis environ- LigroneŽ. 1988, p. 92 , fungal endophytes sometimes mental conditions . . . .B Ž.Boullard, 1988, p. 122 . To reported in association with mosses are generally date, no symbiotic relationship has been detected in found in Asenescent tissuesB and are Amore likely to any fossil hepaticsŽ. Boullard, 1988, p. 113 nor is be saprophytic than mutualistic.B mycotrophism a Aprerogative of apparently primitive In hepatics, as well as in some pteridophytes, taxaB; in the Jungermanniales, for example, fungi are endophytic fungi are restricted largely to the gameto- absent from both the most primitive and most ad- phytes where they are called mycothalli at least by vanced familiesŽ Pocock and Duckett, 1985b; someŽ. Boullard, 1988 . Sometimes infection is lim- Boullard, 1988. . ited to rhizoidsŽ the mycorrhizomes of Boullard, Duckett et al.Ž. 1991 , on the other hand, argue 1988.Ž that simulate root systems Boullard, 1988, p. that the association of fungi and hepatics is more 113. . Interestingly enough, as Ligrone and Lopes likely a relatively recent one, since only a small Ž.1989, p. 430 report, sporophytes of hepatics and group of hepatics is affected plus the fact that the hornworts never harbor fungi, unlike tracheophytes, Athreewx infected orders of liverworts each contain and as they note even the gametophyte tissue representatives forming associations with three major Ajuxtaposed to the sporophyte footB remains unin- mycorrhizal-forming groups of fungiŽ namely basid- fected, thus the relationship cannot be called mycor- iomycetes, ascomycetes and zygomycetes. suggest- rhizal in bryophytes. ing that these symbioses arose after the divergence Despite the numerous caveats, the information of the different orders.B above suggests the potential for establishing a role However, Duckett et al.Ž. 1991 also raise the for fungal symbionts in early embryophytes well possibility calling on Pirozynski’s hypothesis that before the advent of the tracheophytes, with a tenta- vascular plant evolution has seen a trend toward tive benchmark in the Early to Middle Ordovician reduction in mycorrhizal symbiosisŽ see also Trappe, when fungi might first have became associated with 1987a,b, as cited in Smith and Read, 1997, p. 25. higher plants in an endosymbiotic relationship. that there could be a similar Aphasing outB of the However, concerning the potential evolutionary relationship in the hepatics since such associations significance of the relationship, there is little but are absent in potentially more advanced hepatics speculation whether bryophyte–fungal AsymbiosisB with the implication that it was an ancient trait that A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 113 first appeared among the earliest hepatics. Evidence It is also unclear what role symbiontsŽ endosym- that could be construed as consistent with Aphasing bionts. play in hepatic nutritional economy. Is it outB exists in Pteridophytes as noted aboveŽ cf. consistent with the role played in angiosperms, in- Pirozynski, 1981. . Angiosperms, on the other hand, cluding the potential for increased mineral weather- provide only incomplete evidence for an evolution- ing where the infected thallus is pressed flat on rock ary progression from endotrophy to more advanced and soil surfaces? Or where fungi are limited to symbiotic relationships to ultimate independence apices of rhizoids, is there an absorptive roleŽ Pocock from a fungal relationshipŽŽ Trappe 1987a,b, p. 23 . . and Duckett, 1984; Boullard, 1988, p. 115. ? Bon- Thus, such an evolutionary progression cannot be fante-FasoloŽ. 1984, p. 28 suggests that while fungi attributed to vascular plants as a group. In the event associated with some hepatics resemble VAM fungi that Aphasing outB is a reality, it should be more in some respects, neither the physiological or cyto- obvious in ancient embryophytertracheophyte logical details of the relationship are sufficiently groups. Most ancient groups should show both more understood in the liverworts to identify them as facultative relationships and more mycotrophic free VAM’s. Duckett and ReadŽ. 1991, 1995 also indi- relationships. Phasing out, if actually present, could cate as noted above, that functional aspects are not explain the absence in mosses and hornworts but known. Could the relationship be a controlled para- why not more phasing out in hepatics? sitism like that espoused for lichensŽ Ahmadjian, Another potential alternative is that mycotrophy is 1995. where defensive compounds stop or slow an ancient embryophytic character consistent with growth of the pathogens, and where the hepatic early terrestrial adaptation, that has persisted in some supplies the fungi with nutrients but where the role phylogenetically ancient taxa, been lost in some taxa of the mycobiont toward the plant is unclear or become facultative in time, and in some circum- Ž.Ahmadjian and Jacobs, 1983 ? Like algae, nutrients stances has been secondarily reacquired. A facul- needed by hepatics can be absorbed directly from tative relationship could be viewed as an Ain- rainwater so there would seem to be little direct need termediateB stage, consistent either with a gradual of nutrient transport from the fungal partner such as elimination or with reacquisition leading in time to occurs in tracheophytes. a more obligate relationship. Only evidence for the Attempts to justify a similar role for mycobionts presence or absence of mycotrophy in Acommon in hepatics to that in tracheophytes have been ancestorsB could help to resolve this question. broached by a number of workers using different Moreover, the highly labile, commonly facultative types of arguments. Pocock and DuckettŽ 1984, relationships between hepatics and mycorrhizal fungi, 1985a,b.Ž as cited in Ligrone, 1988 . , for example, dependent on a variety of environmental circum- note that the association of different fungi with stances, tend to confound the interpretation of evolu- different groups of hepatics may involve a range of tionary relationships, just as they do in the pte- morphological specializations, thus recalling Athe so- ridophytes and in vascular plants in general. The phisticated patterns described in vascular plants myc- facultative nature of the relationship implies an an- orrhizaeB Ž.Ligrone, 1988, pp. 92, 98 thus attempting cestral capacity for mutualism but not a co-depen- to justify a similar mutualistic and functional rela- dent, co-evolutionary association that only makes life tionship. LigroneŽ. 1988 suggests that the pioneer possible. Much evidence indicates or suggests that role of Phaeoceros on nutrient deficient substrates mycotrophism increases the ’fitness’ of the symbiont and the fact that the hornwort rhizoids are dead and but little indicates that mycotrophism is essential for unlikely to act in absorption and translocation of existence. Given the labile, facultative nature of the nutrients, may justify the assumption of such a role relationship it is difficult to apprise its evolutionary by the associated fungal hyphae. But he also notes significance and whether the relationship has evolved that cyanobacteria are as likely to be AsymbioticB as repeatedlyŽ. independently over time, or whether it is fungi in the hornwort and could account for success- a character that appeared in the early members of the ful colonization of mineral-deficient substrates. group in response to the stresses of early land life. Ligrone and LopesŽ. 1989 also argue for a nutrient 114 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 role for fungi in Conocephalum which colonizes terrestrial environments that both groups would have bare rocks or wet rocky surface, but present no faced—low available nutrients, low organic matter evidence for this assertion other than occurrence in and possibly limited habitat water availability— rhizoids that might be sites of nutrient and water would seem equally important if mycorrhizae gen- exchange. uinely made early terrestrial life possible by enhanc- The possibility of an ancient and potentially prim- ing competitive ability in presumably inhospitable itively ancestral hepatic–endosymbiotic relationship sites. could be viewed as consistent with the potential for If mycotrophy in bryophytes has evolved indepen- an early tracheophyte–endosymbiotic relationship dently, this may also be true in vascular plants and based on cladistic analysesŽ for example, Mishler et would imply no long-term coevolutionary relation- al., 1994. that support the conclusion that hepatics ship as implied in Pirozynski and Malloch’sŽ. 1975 both preceded tracheophytes in time and are ulti- hypotheses. mately ancestral to them. A pre-tracheophytic my- In modern environments, mycorrhizal relation- cotrophic association with embryophytes potentially ships are reestablished in disturbed areas where they ancestral to the tracheophytes could provide the ra- have been eliminated or reduced by the disturbance, tionale for suggesting that the earliest tracheophytes via existing inocula transported into the area by wind may have AinheritedB their endotrophic relationship or animals or surviving the disturbance in isolated Ž.see Boullard, 1988, p. 109 and even that Ater- patches from which they would be available for restrial plants are the product of this ancient and reinnoculationŽ. Allen, 1991, pp. 132–140 . If mycor- continuing partnershipB ŽPirozynski and Malloch, rhizae were indigenous to tracheophytes that inher- 1975. that began with an aquatic alga and semi- ited the algal–fungal symbiosis proposed by aquatic fungus. Consistent mycotrophy in bryophytes Pirozynski and MallochŽ. 1975 , they should have would strengthen Pirozynski and Malloch’s position occurred in all the embryophytes that preceded the for the antiquity of this relationship in tracheophy- tracheophytes and would have provided the inocu- tes. If Redecker et al.’sŽ. 2000 recognition of gloma- lum. If symbiotic fungi were not present in sister lean-type fungal remains in the Middle Ordovician groups, the relationships would have had to arise de of Wisconsin is sustained we will have further cir- novo in the tracheophytes, possibly through an origi- cumstantial evidence for a hepatic–bryophytic rela- nal parasitic relationship, as has been proposedŽ Al- tion well before the advent of the tracheophytes. len, 1991, p. 35. , and may have arisen independently This proposition would be strengthened were it over time in various groups of tracheophytes as not for the fact that paraphyletic andror potential Duckett and ReadŽ. 1995 have proposed for the sister groups to the tracheophytes, like the hornworts bryophytes. Again the lability of the relationships, and mossesŽ. Mishler et al., 1994 , appear to be tends to belie evolutionary interpretations. largely AliberatedB from such mycotrophic relation- Berner’s hypothesis of a major role for tracheo- ships with fungi, although some retain mutualistic phytes in atmospheric CO2 determination presup- relationships with cyanobacteria. The possible poses a major correlation between weathering by AVAM-likeB associations with at least one hornwort tracheophytes that necessarily assumes an ancient is interesting, but because the relationship is wholly and continuing partnership between fungi and tra- facultative, can scarcely be seen to support Athe cheophytes since he proposes it is the specific micro- hypothesis that the symbiotic association with en- habitat created by the mycorrhiza where most weath- dotrophic phycomycetous fungi has been a primary ering occurs. Berner’s emphasis on tracheophyte event in the adaptation to a terrestrial environmentB weathering negates any potential role for microbiota as advocated by Pirozynski, etc.Ž also Ligrone, 1988, and possible non-tracheophytic embryophytes that p. 99; Ligrone and Lopes, 1989. . As discussed above, preceded tracheophytes onto land and faced the same we cannot know whether mycotrophism is an ances- physical environment that millions of years later tral embryophytic character that was eliminated in were encountered by tracheophytes and tracheo- both modern groups, or whether these groups have phyte-like plants. As discussed above, the silicate never had symbiotic relations with fungi. The early weathering potential of varied microbiota and A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 115 bryophytes is not trivial, though unassessed in com- support Berner’s hypothesis about the extreme rate parison with extant angiosperms. These data tend to of CO2 drawdown and the rate of increased silicate emphasize that, among the varied organisms consid- weathering in the Late Devonian. But what is the ered, weathering potential is highest for lichens. current assessment of the Rhynie data? And what is Lichens can be viewed as analogues to the mycor- the current assessment of various of the remains that rhizal mutualismŽ. or controlled parasitism so well have appeared consistent with the early appearance documented in angiosperms. The high potential for of mycorrhizal associations in the fossil record? lichen weathering suggests that a synergistic rela- Much speculation about, and continuous reassess- tionship between fungi and their symbiotic partners ment of the role of fungal remains found in associa- may increase the potential for organic acid production with the Rhynie Chert tracheophytes, as well as tion independently of the taxa in which this relation- reassessment of the tracheophytic status of some of ship occurs and independently of the microhabitat the Rhynie plants, indicates little support for the provided by the tracheophyte root. This may occur in Pirozynski and Malloch hypothesis from the evi- hepatics infected with fungi. dence of the Rhynie ChertŽ. see Remy et al., 1994 . Although fossils resembling hyphae, vesicles and spores of the modern mycorrhizal GlomaceaeŽ originally identified as Endogone, family Endogonaceae 21.6. The Rhynie Chert and a molecular benchmark and groups now separated from the Glomales and for mycorrhizal–tracheophyte relations including saprophytes and ectomycorrhizal associates as noted above. are associated with the axes of How old is the mycorrhizal–tracheophyte rela- Rhynie plantsŽ. Kidston and Lang, 1921 , there is no tionship? Information from the fossil record has been unqualified morphological evidence that any of these used to attempt to establish a time frame for mycor- remains are endomycorrhizal or VA mycorrhizae, rhizal–tracheophyte symbiosis. This information is rather than saprophytes as Kidston and LangŽ 1921, now supplemented by molecular data. pp. 872–873.Ž see also Taylor and White, 1989; Putative evidence for endomycorrhizae in Rhynie Hass et al., 1994. originally argued indicating that tracheophytes has been commonly seen to provide a Aproof of the existence of mycorrhiza is wanting . . . B time frame for the tracheophyte–fungal symbiosis When Boullard and LemoigneŽ. 1971 restudied Ž.Pirozynski and Malloch, 1975 . In turn, antiquity of the fungi in the gametophytes and sporophytes of the putative endomycorrhizal–tracheophyte relation- Rhynia and sporophytic axes of Asteroxylon, they ship based on evidence from the Rhynie Chert pro- concluded that among the endophytic fungal re- vided the inspiration for the hypothesis that the mains, some could be endophytic, respectively my- tracheophyte–mycorrhizal relationship was the out- cothallus and mycorrhizomes, while others were more growth of prior mutualistic partnership between a likely parasitic or, in the axes of Asteroxylon, sapro- semi-aquatic chlorophyte and an aquatic fungus and phytic because of their association with dead tissues. that by implication such an obligate relationship was A saprophytic relation for some fungal remains is a necessary precursor to land plant colonization claimed to be strongly supported by the absence of ŽPirozynski and Malloch, 1975; Pirozynski, 1981; fungi from well-preserved remains and their incon- Pirozynski and Dalpe, 1989. . The implications are sistent presenceŽ see Taylor and White, 1989; Hass that endotrophy and the endomycorrhizal condition et al., 1994; Pirozynski and Dalpe, 1989. . The pres- was necessarily indigenous to the tracheophytic con- ence of Glomus-like remains in the RhynieŽ the dition and essential to tracheophyte evolution on extant Glomus and its close relatives in the gloma- landŽ. also Remy et al., 1994 . lean fungiŽ. Glomales , the group to which all en- If we lay increased silicate mineral weathering at domycorrhizal fungi, including some taxa attributed the feet of mycorrhizal tracheophytes, this hypothesis originally to the endogonaceae referred, Smith and related to the establishment of mycorrhizal root sys- Read, 1997. is consistent with a long history for the tems in tracheophytes, and the time frame then con- extant taxon that forms mycorrhizal symbiosis with jectured for terrestrial plant life, could be seen to tracheophytes in association with tracheophytes that 116 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 spans the Phanerozoic. This association, however, is with modern tracheophytes, where physiological based largely on highly recognizable wind- and wa- function of apparently mycorrhizal structures can be ter-borne sporesŽ Pirozynski and Dalpe, 1989, pp. tested, make clearŽ. Allen, 1991, p. 142 , a functional 18–19, Table 1; Smith and Read, 1997. ; few puta- not a structural approach is essential Ato understand tive mycorrhizal associations prior to the Cenozoic whether a fungus–plant association is a mycorrhiza are based on the presence of specific structures Ž.mutualist or a parasite.B within fossil roots that can be demonstrated to be The earliest unqualified fossil evidence for VA symbiotic endophytesŽ Pirozynski and Dalpe, 1989, mycorrhizae in a tracheophyte root is found in the pp. 10–12, Table 1.Ž. . Stubblefield et al. 1985 for TriassicŽ. Stubblefield et al., 1987b but may, of example, assessed in detail chlamydospores de- course, provide only a minimal age for the establish- scribed from aerial axes and roots of a number of ment of this relationship. Stubblefield et al.Ž. 1985 Paleozoic plants ranging in age from uppermost describe sporesŽ. chlamydospores from aerial axes Lower Devonian to Upper Pennsylvanian and were and roots of Paleozoic plants from localities with unable to unqualifiedly identify these as VAM or ages uppermost Lower Devonian through the Upper endomycorrhizal fungi despite the close similarity Pennsylvanian whose ultrastructural morphology is with Glomerales. consistent with interpretations that relate them to Additionally, identification of physiologically Endogonaceae, but apparently these are no longer functioning mycorrhizae in fossil remains is compli- accepted. cated by the fact that the identification is necessarily Molecular data, commonly seen as chronologi- structuralŽ. cf. Allen, 1991 while the definition is cally supportiveŽ Smith and Read, 1997; Simon et functional which means that a physiological assess- al., 1993; Remy et al., 1994. for the establishment of ment is virtually impossible with fossils. Moreover, a mycorrhizal association with Rhynie tracheophytes, if Harley’sŽ.Ž 1968 as cited in Allen, 1991, pp. 35, are inconclusive. The molecular data based on living 141–142. assertion is correct that plant–fungal inter- representatives of VAM glomerean fungiŽ Simon et actions in a plant taxon are actually gradients along a al., 1993. , provide no unique solution to this prob- time continuum from parasitism to mutualism, with a lem because of the extended range of time indicated sometimes mutualistic interaction and a sometimes as possible by their data. Their data provide no parasitic interaction, specific structures are in them- positive support for an origin of VAM consistent selves inadequate for determining when an associa- with tracheophyte origins. Contrary to confirming an tion is mutualistic and when parasitic. Or, as Allen endomycorrhizal–tracheophyte mutualism, as com- Ž.1991, p. 142 notes Adetermining a mycorrhiza sim- monly stated, their permissive data have raised imply by looking for specific structures . . . is not portant issues of timing. They can be interpreted to valid.B Additionally, even non-mycotrophic plants support the possibility that the endophytic mycor- that never form mycorrhizae, indeed taxa that ac- rhizal relationshipŽ. VAM may have originated as tively reject mycorrhizal fungi, can be colonized by early as the Middle OrdovicianŽ. 462 Ma or even in mycorrhizal fungi that appear to act as parasites the later Cambrian well before fossil evidence for Ž.Allen, 1991, p. 89 . tracheophytes or tracheophyte-like plants. It can Recent recognition of arbuscules of VA mycor- equally well be interpreted to indicate the first estab- rhizae in the Rhynie Aglaophyton major ŽRemy et lishment of such mutualisms in the Late Devonian– al., 1994. appears to confirm the VAM status of a Early Carboniferous, the maximum Ž.minimum best Rhynie Chert plant. But Aglaophyton is not a tra- time indicated by their data, which are permissive cheophyte and the arbuscules are associated with even however, as late as the Late Carboniferous. Aglaophyton’s axesŽ. protosteles where their func- These data can be taken equally to imply that the tion is speculative. If, as AllenŽ. 1991, p. 89 reports, mycorrhizal relation is not unique to the tracheo- non-mycotrophic plants can be colonized by mycor- phytes, and that a relation with a lower autotrophic rhizal fungi that appear to act as parasites, Aglao- group is reasonable in the Ordovician or earlier. phyton’s fungus may have been parasitic, despite the Alexopoulos et al.Ž. 1996 note that VAM fungus well-preserved appearance of the stems. As studies form relations with some bryophytes, pteridophytes A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 117 and even a few algae which may be either a primi- and Trappe’sŽ. 1993 finding of numerous examples tive or derived relationship. within the modern Pinaceae. According to Alexopou- Berbee and TaylorŽ. 1993, p. 1123 arrived at los et al.Ž. 1996 , VAM are most common in an- estimates based on molecular data which they inde- giosperms. Because of the latitude of the data, it is pendently have interpreted in line with both these possible to conclude that a mycorrhizal relationship conclusions: Their data are consistent with a diver- was established with the earliest land plants. On the gence of terrestrial fungi that includes endomycor- other hand, it is also possible to conclude from rhizal Glomaceae, as well as ascomycetes and basid- presently available information that widespread myc- iomycetes, from primitive aquatic chytrids Aroughly orrhizal relations arose too late in the Paleozoic to be 550 Ma agoŽ. 400–700 Ma , perhaps 50–150 Ma a factor in the Devonian drawdown or a major factor before vascular plants colonized the land . . . B al- in Paleozoic weathering phenomena. though they do not see divergence between the as- We suggest that there is varied evidence denying comycetes and basidiomycetes lineages until ca. 395 Berner’s conclusions in this regard. Earlier we pre- Ma ago at the end of the Silurian. They further sented a sample of the extensive evidence indicating conclude that Glomaceae diverged from the progeni- that lower autotrophsŽ. and heterotrophs today do tor of ascomycete and basidiomycete lineages in the generate organic acids and other compounds that are Ordovician and suggest as we have done above that active in silicate rock weathering. We suggest that the first terrestrial fungal radiation may have been the relationship between fungi and lower autotrophs correlated with the adaptive radiation of non-vascu- is an ancient one of particular advantage to the lar land plants. They have not expressly suggested, fungal pair of the partnership as suggested by Ah- as both we and BoullardŽ. 1988 have, that the rela- madjianŽ. 1995 . The earliest alleged fossil evidence tionship may already have been a mutualistic one at for a lichen is provided by Taylor et al.Ž. 1997 who that time, nor that the potential for mycorrhizal describe a lichen from the mid-Early Devonian relationships with pre-tracheophytic embryophytic Rhynie Chert, which opens up the possibility that land plants that preceded the tracheophytesŽ Gray, lichens with their associated fungi and organic acid 1985. might also have significantly increased their producing capability may well have been present far potential silicate weathering effects. earlier. However, Poinar et al.Ž. 2000 state that none In sum, Paleozoic and Mesozoic fossil evidence of the previous reports of fossil lichens, including the supporting the prevalence of the mycorrhizal rela- specimen from the Rhynie Chert, demonstrates the tionship is necessarily speculative. Pirozynski and presence of algal and fungal bionts in a single thallus DalpeŽ. 1989, p. 2 now write: ASave for a few indicating a stable physiological interaction. The al- sporadic records in the literature of fungi referred or leged fungal partner in the Rhynie Chert AlichenB is referable to VAM fungi . . . we do not know when, unknown today in any lichen. According to Poinar et where and how such fungi came to be associated al.Ž. 2000 , the first unequivocal fossil lichen is from with plants or their precursors, where they came mid-Tertiary Dominican amber. from and what and where their living relatives are.B In view of the above comments, we feel it unrea- Thus, it cannot be assumed that either the mycor- sonable to conclude that the mycorrhizal relationship rhizal niche or the geochemical implications for pro- and the enhanced weathering capability provided by moting extensive silicate rock weathering associated this relationship is a uniquely tracheophytic phe- with this niche in modern angiosperm-dominated nomenon, and that the production of organic acids by ecosystems was well established in the Paleozoic or autotrophs is a strictly tracheophytic specialty. Ad- among Paleozoic pteridophytes and gymnosperms, mittedly, there are few data that permit one to glob- because this has been assumed largely on a baseline ally generalize about the relative production capabili- for tracheophyte infection in the Early Devonian. We ties either today or in the past for organic acids may also conclude that Paleozoic and Mesozoic produced by differing groups of autotrophs. Nor are pteridophytes and gymnosperms may have had myc- there good data for generalizing about the production orrhizae-like fungi, which appears to be the circum- capabilities of the different autotrophic groups oper- stance for the Mesozoic Pinaceae based on Cazares ating under different climatic regimes. These prob- 118 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 lems make it essential that one not try to overgener- is ambiguous. We argue for the presence of terres- alize about the organic acid production capabilities trial microbiota as significant weathering agents in of the different autotrophic groups. the pre-Devonian, of sufficient importance to have had an effect on atmospheric composition. But that 21.7. Conclusions postulate necessitates evidence for the presence of a variety of pre-Devonian terrestrial organisms. Is there In addition to the evolutionary problems, there are any evidence to support this hypothesis? Where is many variables with regard to the mycorrhizal situa- the fossil record—the most definitive evidence? tions. It seems clear that the relationship between None of these organisms, with the exception of the fungi that form symbiotic relationships with plants is cyanobacteria can be definitively recognized in the in evolutionary transition, and little can be assumed pre-Devonian terrestrial environment. Both theoreti- about stability through time between specific fungi cal and molecular data support a long antiquity. and plants, or for different plants or for different biomes and soils. Most of what we know involves 22.1. Lichens the angiosperms, a group that evolved or at least whose diversification occurred in the Cretaceous and Lichens have a negligible fossil recordŽ Poinar et early Cenozoic. Read et al.Ž. 2000 review the entire al., 2000. . The recent description of a potential fossil problem of mycorrhizal ancestry, while arriving at lichen from the Early Devonian Rhynie ChertŽ Taylor the conclusion that there is still widespread uncer- et al., 1997. raises both the issue of antiquity of tainty. Some argue that some taxa have evolved out these organisms and the accuracy of the molecular of the systemŽ. pteridophytes ; other than that there is clock used to calibrate evolutionary radiations of no evidenceŽ. angiosperms . Yet the increasing vari- the fungi, including the ascomycetes and basid- ability in the types of mycorrhizaeŽ ericoid, orchids, iomycetes. Berbee and TaylorŽ. 1993, p. 1125 indi- etc.. argues for evolutionary change over time. Vari- cate that they accept that Aa molecular clock may not ability in different environments also complicates the be accurate to within 25 Ma, but hope that it is problem. Additionally, there is the serious complica- usually accurate to within 100 Ma.B The fossil record tion that a complete spectrum exists from broad to of ascomycetes in the Silurian and the lichen Winfre- narrow host specificity in ectomycorrhizal fungi to- natia in the Early Devonian, are both consistent with day, as well as large differences from biome to the presence of pre-Devonian lichens, contrary to the biomeŽ. Molina et al., 1992 , with no way to estimate molecular clock dates. The unique preservational such effects for the distant past. circumstances afforded by the Rhynie Chert may well explain why fossil lichens are virtually absent from the record, despite the reasonableness for their 22. Where is the pre-tracheophytic fossil record? presence, additional to the fact that few paleontolo- gists have worried very much about finding them. Some have argued that it is AimplausibleB that But Poinar et al.Ž. 2000 discount Taylor et al.’s land surfaces were sterile for Aseveral billion yearsB Ž.1997 claim because it does it not show the actual in the PrecambrianŽ Schwartzman and Volk, 1991, p. relationship, nor demonstrate the relationship. Poinar 360; Schwartzman and Shore, 1996. . Indeed good et al. also note that the mycobiont resembles Zy- evidence exists that they were notŽ. Gray, ms any- gomycota, a group not involved today in the lichen more than one would expect sterile ground surface in relationship. the Early Phanerozoic before evidence for early em- The similarity of a valid mid-Tertiary genus to a bryophytes. Gray found good evidence for varied, modern lichen taxonŽ. as in Poinar et al., 2000 widespread, bacterial-level, non-marine fossils in implies that the relation was not recently evolved. Precambrian strata that on associated physical evi- DNA or molecular data support the presence of both dence can be concluded to be non-marine. fungi and cyanobacteria of sufficient antiquity to We discussed above a possible evolutionary his- have formed an early relationship well in advance of tory for the mycorrhizal symbiosis. The fossil record the time they are first found in the fossil record. Cain A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 119

Ž.1972 suggests that the lichens may well have pre- fungus and lichens at 3.8 Ga. According to this ceded the tracheophytes in time. analysis, divergence of modern lineages is thus Cam- The essential absence of lichens from the fossil brian and post-Cambrian, but this still leaves a large record is not surprising, since they only occur under window of opportunity for the activities of progeni- very unusual, very uncommon preservational situa- tors of these extant groups. tions. One approach to the question of their antiq- Doolittle et al.Ž. 1996 using a protein clock based uity, that might be raised, concerns the antiquity of on amino acid sequence data, state that plants, ani- the lichen symbiosis based on the phylogenetic con- mals and fungi, last shared a common ancestor about cepts for ascomycetes and basidiomycetes that today a billion years ago. Baldauf and PalmerŽ. 1993 see are the principal mycobionts in lichens. Berbee and animals and fungi as Asister groupsB, separate from TaylorŽ. 1993 argue from DNA sequence data, from plants, based on protein sequence data which also which they establish a relative time scale for fungal has implications for the longevity of fungi. Neither evolution, that the lineage immediately ancestral to study bears on the time of terrestrialization in the ascomycetes and basidiomycetesŽ the pre-ascomy- group or provides any information about the advent cetes–basidiomycete lineage. diverged before the of terrestrialization. Hasegawa et al.Ž. 1993 suggest Lower Devonian and probably in the Ordovician, but that animal and fungal origins are closest, based on that ascomycetes and basidiomycetes, are much mitochondrial protein sequence data. Simon et al. younger, and probably did not appear until some Ž.1993 on the other hand interpret molecular data to time in the Carboniferous. They also argue that support the advent of the Mycota at about 1 Ga, well fossils that resemble propagules from filamentous back into the mid-ProterozoicŽ. Mesoproterozoic . All ascomycetesŽ. Sherwood-Pike and Gray, 1985 could of this leaves open the question of terrestrialization. be from ancestors to higher ascomycetes and basidiomycetes rather than filamentous ascomycetes 22.3. Mycorrhizae themselves, and that the basal ascomycetes and basidiomycetes are parasitic or saprophytic on land Using a molecular approach, Simon et al.Ž. 1993 plantsŽ. p. 1121 rather than lichen symbionts. The date the advent of the Endomycorrhizae at the begin- possibility of such a relatively advanced age for ning of the Ordovician. This is well before the lichens, that are virtually unknown as fossils, would advent of paleontological evidence for tracheophytes potentially remove them as significant weatherers in in the later Silurian or earliest Devonian. Simon et the pre-Devonian. al.Ž. 1993 envisage the advent of the Mycota at about 1 BY before present, i.e. roughly mid-Proterozoic, 22.2. Fungi again employing the results of molecular data. Re- gardless of how one regards the reliability of the Fungi are equally uncommon at the appropriate molecular data and molecular AclocksB the implica- time interval, although their antiquity is supported by tion is that non-marine fungi have been available in both molecular studies and theoretical considera- the environment well before the appearance of vas- tions. How common may terrestrial fungi have been cular plants. The problem with the Simon et al. in the pre-Devonian? Again, as in the question raised Ž.1993 earlier Ordovician appearance of mycorrhizae above with regard to lichens, the antiquity of terres- is that no plants with roots or rhizomes, with which trial fungi becomes a question because of their lim- modern endomycorrhizae are coevolved, are present ited fossil record. Using a phylogeny from DNA until some time in the later Silurian and Devonian, sequence data, Berbee and TaylorŽ. 1993 estimate which suggests that terrestrial fungi have been pre- that terrestrial fungiŽ ascomycetes, basidiomycetes, sent well before the advent of the coevolved relation zygomycetes. diverged from the aquatic Chytrid- with tracheophytes was possible. There is some evi- iomycetes at least 550 Ma ago in the mid-Cambrian, dence for the presence of a AmycorrhizalB relation- but possibly as much as 700 Ma ago in the Neopro- ship between bryophytes and fungi. Duckett and terozoic, even though they discount previous Pre- ReadŽ. 1991 document the association of VA and cambrian fossil records that far back attributed to ascomycetaceous ericoid mycorrhizae with varied 120 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 liverworts, and refer to one association with an fungi and some typeŽ. s of algae, a relationship that anthoceroteŽ. Ligrone, 1988 , while Duckett and Read might have first evolved in the Precambrian well Ž.1995 further expand on the hepatic–mycorrhizal before the mid-Ordovician presence of any paleonto- association. logical evidence for hepatic-type embryophytes. There is alleged evidence for the presence of Graham and GrayŽ. 2001 have argued from mor- Endomycorrhizae, beginning with the mid-Early De- phological, biochemical and molecular evidence that vonian Rhynie Chert discussed above. The question bryophyte-like land plants were present at least 20– next arises about whether or not such fungi or their 40 million, and perhaps as long as 100 million years, ancestors were on the scene prior to the arrival of the before megafossil evidence exists for tracheophytes. tracheophytes, including potentially immediate tra- Morphological, including ultrastructural morphologi- cheophyte precursors in the later Silurian. The ques- cal information, combined with molecular and cladis- tion is conceptually important because if there were tic analyses, project phylogenetic relationships in tracheophytic mycorrhizal precursors involving the which charophyte green algae, bryophytes and tra- liverworts one can then infer that switching from a cheophytes are a monophyletic group with the land liverwort host to a primitive tracheophyte host is a plantsŽ. bryophytes and tracheophytes a well-sup- possibility. This question may be partially addressed ported monophyletic group. In some such analyses, by considering the relations of mycorrhizal fungi although neither the specific outgroup for land plants with the mosses, liverworts and anthocerotes today. nor the relationship among basal lineages are clear, Powell and BagyarajŽ. 1984, p. 28 state that mosses in others the three lineages of bryophytes appear today do not display mycorrhizal relations as do paraphyletic with respect to the tracheophytes and liverworts. SchusterŽ. 1966, pp. 249–255 has re- mosses alone as sister group to the tracheophytes viewed the question of symbiotic relations between Ž.see Mishler et al., 1994 . Regardless of the details extant fungi, and even cyanobacteria, with hepatics. of phylogeny, if we accept a middle–late Silurian After carefully reviewing the literature available at baseline for the presence of tracheophyte-like plants that time, he concludes that there is a complete and tracheophytes, the data clearly support early spectrum extending from some coevolutionary level evidences for bryophytic land plants in the Ordovi- between mycorrhizal-like fungi and hepatics to oth- cian or earlier, just as projected from the fossil spore ers where the relation is parasitic. If one takes his record for well over 20 yearsŽ. Gray, 1985 . conclusion at face value, it is then easy to potentially extend the range of this symbiotic relation back to the mid-OrdovicianŽ. Llanvirn , where there is good 23. Autotrophic productivity and sequestered or- evidence for embryophytesŽ. Gray, 1985 . On the ganic carbon, or, where is the pre-tracheophytic other hand, the presence of structurally mycorrhizal- biomass? like fungi in the lower embryophytes is not yet demonstrated functionally, and without a functional While considering the potential of lower au- demonstration the entire weathering question is still totrophs as significant organic carbon sources, one unresolved. The absence of mycorrhizae in modern must recognize that soil algaeŽ including cyanobacte- mosses may either reflect the total absence of such a ria. occur on and within soils from the tropics to the relationship from the Devonian first appearance of polar regionsŽ. Metting, 1981 . Furthermore, Lange et fossil mosses to the presence and its subsequent loss al.Ž 1992, 1994, 1998; see also Lange and West, Žkeep in mind that Calamites of the Late Paleozoic 1990. make it clear that desert crust organisms from have evidence for endomycorrhizae, whereas this arid and semi-arid environments are capable of gen- capability has been lost in Selaginella, i.e. absence erating the same order of magnitude organic carbon in modern mosses might also represent loss through as generated by higher land plants. This being the evolutionary time. . In any event, it is conceptually case today for arid and semi-arid environments is it reasonable to extend the mycorrhizal relation back to too much to ask that in the pre-tracheophytic, older the Middle Ordovician. Prior to that time, one may rocks, that the lower autotrophs were important in infer the presence of the lichenized relation between the more humid environments where they have since A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 121 been largely pushed aside by the tracheophytes? It is question that could be raised, if this is the case, is also worth noting that the organic carbon contribu- where are evidences of this organic matter? tion to modern soils outside of the desert regions has The summary of the rock-weathering capabilities not been analysed in terms of tracheophytic organic of modern heterotrophic and autotrophic microorgan- carbon as contrasted with non-tracheophytic related isms and autotrophic bryophytes indicates that they autotrophic carbon. are effective rock weathering agents. The question It is important not to confuse the potential for remains, however, whether they were capable of significant mineral weathering in soils by lower au- generating sequesterable carbon equivalent to the totrophs with the localized accumulation of large amounts produced by tracheophytes in the pre- amounts of organic carbon, such as coal. Large tracheophytic world. carbon deposits require both significant productivity Berner has emphasized the potential role of tra- plus topographic and climatic situations conducive to cheophytes in controlling sequestered organic carbon accumulation. There are well-known Meso- biomass and atmospheric composition, minimizing a zoic and Cenozoic situations where productivity was potential role for all other organisms in organic high under suitable climatic conditions, whereas car- carbon sequestration back in time. Our interest is in bon accumulation did not occur in the absence of pre-Devonian climates and in pre-tracheophytic em- suitable topographic conditions. bryophytes and other land-dwelling autotrophs and heterotrophs that we conclude were widespread in the pre-Devonian. Here we consider the major roles 23.1. Precambrian regoliths argued by Berner for tracheophytes in comparison with non-tracheophytes:Ž. 1 tracheophytes as the sole, An increasing number of regoliths and immature significant source of stores of sequestered organic AsoilsB are now identified in the Precambrian and carbon and the closely related question of whether it Early Paleozoic that show the effects of weathering is possible to adequately measure sequestered or- such as laterites, including bauxites. We argue that ganic carbon through geologic time; andŽ. 2 tracheo- the weathering effects of pre-tracheophytic organ- phytes as the only or first terrestrial organisms capa- isms that inhabited bedrock surfacesŽ. epiliths as ble of generating significant, sequesterable organic well as organisms that lived below the surface of carbon, to affect atmospheric CO2 . It is appropriate rocks, such as endolithic lichens and cyanobacteria, to consider whether marine productivity can also be could have contributed sufficient organic material to ignored in considering the effects on atmospheric these regoliths to sustain populations of autotrophic, composition. Berner’s work concludes that there is heterotrophic, and chemotrophic bacteria and hetero- little or no organic carbon sequestered prior to the trophic fungi adequate to instigate weathering of the appearance of coal basin strata in the Carboniferous type that we now find in soils where mineral solubi- although there are Late Devonian coal basins. Does lization is as much accomplished by the organic the absence of pre-Late Devonian coal basin sedi- acids produced by these free-living organisms as it is ments indicate the absence of significant sequestered by root and root mycorrhizal Aexudates.B All that is organic carbon? needed to sustain these microorganisms, independent The amount of sequestered carbon in sediments of tracheophytes, is a source of organic carbon. and sedimentary rocks is correlated with productivity Others have arrived at similar conclusionsŽ Keller of ground surface-dwelling and aquatic autotrophs and Wood, 1993. as we discuss below. and non-autotrophs with the potential for preserva- The implications of much of what we have dis- tion of their remains. Are tracheophytes always more cussed above is that there were pre-Devonian terres- productive under all environmental conditions than trial and limnic organisms capable of generating all other autotrophic organisms and are they the only sufficient organic carbon to support both bacteria organisms from which it is possible to sequester and fungi that could weather rocks and produce large amounts of organic carbon? Can reliable esti- organic-rich soils where weathering could proceed mates be made of sequestered organic carbon and, by prior to the advent of the tracheophytes. So the implication, productivity in the geological past? 122 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

We are not the first to question whether tracheo- Most non-tracheophytic autotrophs, with the excep- phytes necessarily made major changes in terrestrial tion of bryophytes in some habitats, are characterized productivity. Yapp and PothsŽ. 1994, pp. 49 and 51 by low standing crop in most regions of the world in raised the issue of concentration of CO2 generated comparison with tracheophytes. But it needs empha- by microbial respiration in soils. From their analysis sizing that non-tracheophyte productivity, at least in of the FeŽ. CO3 OH component of goethite in a Late the case of the bacteria, is far higher than that of Ordovician paleosol, they predicted that the produc- tracheophytes! Bacterial productivity in many envi- tivity of the pre-vascular autotrophic biota was simi- ronments potentially similar to some of the Early lar to that of modern soils: Aif our inference about Paleozoic, when embryophytes were evolving, may the magnitude of the CO2 flux is approximately be far higher than that of organisms higher in the correct, it indicates the presence . . . of an ancient trophic order. Bryophyte productivity can also ap- pre-vascular terrestrial biota as productive as biota of proach that of tracheophytes in some modern habi- modern tropical andror temperate soils.B Thus, they tats, as discussed below. The fact that such habitats question as we do, whether it can be automatically are environmentally and geographically restricted concluded that Apronounced productivity changes ac- under modern global climatic gradients and with companied the development of vascular plants.B They angiosperms the prevailing vegetation, does not nec- argue that the C13 isotope values for the soil carbon essarily mean that they were similarly restricted are consistent with oxidation of biologically derived throughout the Phanerozoic. It is equally important carbon. However, their conclusions regarding a 16= to note that there are significantly different levels of modern atmospheric CO2 level in the Late Ordovi- organic productivity in different climatic regions. We cian, although agreeing with Berner, strongly dis- suggest that information about productivity, as con- agrees with the geological dataŽ. our Fig. 1 . Robin- trasted with standing crop, might significantly alter sonŽ. 1990a,b makes the potentially important point conclusions about bacterialŽ. microbial and bryophy- based on a number of speculative arguments, that tic potential for providing significant, sequesterable, over Phanerozoic time the terrigenous organic car- organic carbon. bon pool has become increasingly susceptible to TappanŽ. 1968 has argued that the primary pro- biological decay due to a significant decline in ligni- ductivity of marine phytoplankton is extremely high fication in tracheophytes, and a major radiation of compared with tracheophytes. Most tracheophytes, lignin-degrading fungi. This would mean a poten- by comparison with phytoplankton, are long lived, tially significant decrease in sequestration of organic undergo long periods of dormancyŽ limited produc- carbon over Phanerozoic time, despite the potential tivity. and among some with the highest productiv- for increased productivity of organic biomass. Ac- ity, occupy limited total area. Tappan arguesŽ 1968, cording to these arguments rates of carbon sequestra- p. 190. that too great a role has been ascribed to tion for the Paleozoic may have been accomplished tracheophyte productivity and that tracheophytes by a productivity that was small by modern stan- should be seen as Apartial replacements rather than dards. This argument is based on a number of specu- additions to the total productivityB of terrestrial flo- lations and evolutionary assumptionsŽ about the time ras. Similar conclusions are reached by Longhurst et and radiation of lignin-degrading fungi, for example, al.Ž. 1995 who state that A . . . globally, the productiv- for which Robinson provides no information. but ity of the plankton is comparable to that of the land would provide an additional complication in assess- vegetation . . . B Watson and LissŽ. 1998 point out the ing the significance of the organic carbon sink and potential effects of marine biological control on cli- its impact on atmospheric pCO2 . mate via the carbon and sulphur geochemical cycles pointing out by way of example that increased ma- 23.2. Non-tracheophyte productiÕity rine productivity would lead to lower global temperatures, both through the marine biota’s effect on

The significant role for tracheophyte productivity atmospheric CO2 and dimethyl sulphide concentra- is based on information about their standing crop, tions. The potential influence of the marine biota on not their productivity, in many terrestrial ecosystems. climate is not one considered in most models of A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 123

B climate and atmospheric CO2 , but should not be ceous sediments hosted by the Huronian Mi- ignored. chigamme Formation, Michigan. They note the ex- We cannot say whether early embryophytes and tensive quantities of this coal measured in tons, varied autotrophic and non-autotrophic microorgan- uncovered by trenching operations, and refer to an- isms that preceded them in the terrestrial habitat had thracitic coal and black shale in drill cores at depths measurable effects on pre-Devonian climate or atmo- of 1000 ft. Fixed carbon content is comparable to spheric composition. There is considerable evidence that in Carboniferous coals while the carbon content that directly supports high levels of productivity and of the shales ranges from 5% to 25%. the potential for sequestering organic carbon by bac- ReimerŽ. 1978 reports carbonaceous phyllites con- teria, cyanobacteria, algae, AacritarchsB Žpresumably taining up to 46.42% carbon sometimes alternating eukaryotic algae. and presumably by fungi and with graphitic carbon, in the Neoproterozoic Precam- lichens in the Precambrian and Early Phanerozoic— brian Malmsbury Group, South Africa, an occur- in what can be interpreted as both marine and non- rence which he suggests is tempting to compare with marine environments. For example, the total organic Aboghead coals.B HeŽ. p. 37 refers to this sequence carbonŽ. TOC, kerogen in some Precambrian and as perhaps Aone of the oldest known well-preserved Cambrian non-marine and marginally non-marine examples of sapropelic coals.B Reimer et al.Ž. 1979 deposits, such as various units in the McArthur point out that even in the earlier Archaean Fig Tree Basin, AustraliaŽ. Jackson et al., 1987 ; Georgina Group of South Africa, the amount of free carbon is Basin, AustraliaŽ. Southgate, 1986, p. 345 ; Officer similar to that present in younger Precambrian and Basin, AustraliaŽ. McKirdy and Kantsler, 1980 ; Phanerozoic rocks in similar depositional basins; Mid-Continental Rift in North AmericaŽ Imbus et al., from their calculations they estimate that the Sheba 1990.Ž ; and the Torridonian in Scotland Parnell, Formation alone may contain about 52=1015 metric 1994.Ž. ; Desborough et al. 1984 discuss the presence tons of carbon, comparable to that contained in a of abundant hydrocarbons in Apache Group, Middle large coal fieldŽ. p. 97 . Various lines of evidence Proterozoic rocks from Arizona, with accompanying discussed by Reimer et al.Ž. 1979 and Tyler et al. pyrolysis data, indicating that microbial and primi- Ž.1957 are consistent with biogenic origin of this tive algal populations were capable of contributing carbon. abundantly to the TOC content of sediments well Buseck et al.Ž.Ž 1997 see also Volkova and Bog- before the appearance of the tracheophytes. The danova, 1986, 1998. discuss the characteristics of preservation of particulate and stratiform organic AshungitesB, that occur in large amounts in the Lower material is sufficiently high in some sequences that Proterozoic of Karelia. These have been mined in the some of these deposits, including most of those from past as a carbon source, and have been termed the basins indicated above, can be considered poten- AcoalsB in some papers. Although L.P. GaldovinaŽ in tial petroleum resources similar to oil shales. In the Buseck et al., 1997. has argued for a non-biotic literature, they commonly are referred to as Aoil- and origin for this material, V.V. Kovalevski and Buseck gas-prone kerogensB or as Aoil prone microbial kero- Žin Buseck et al., 1997, and Buseck personal com- gensB ŽMcKirdy and Hahn, 1982; McKirdy and Im- munication, July, 1998. base their interpretation on bus, 1992; McKirdy and Kantsler, 1980; McKirdy et the carbon isotope data. The paleobiogenic nature of al., 1984; Parnell, 1994; Imbus et al., 1990. . the organic matter in the Karelian carbonaceous se- Additional metamorphosed kerogens and kero- quences is supported by the isotopic composition of gen-bearing units, too mature for petroleum sources, the carbon, and the heterogeneous chemical composi- are widely known in the Precambrian, where they tionŽ. Volkova and Bogdanova, 1998 . Schwartzman range in age from Archaean to late Proterozoic. and VolkŽ. 1989 make a strong case for the presence Several examples illustrate. Tyler et al.Ž. 1957 dis- of significant autotrophy since at least the later Ar- cuss carbonaceous shales with graphitic bodies and chaean as do SchidlowskiŽ. 1986, 1988 and Schid- anthracitic carbonaceous sedimentary rocks that they lowski and AharonŽ. 1992 . Sozinov and Gorbachev variously refer to as Aanthracitic coal,BAanthracitic Ž.1987, p. 38 point out that while the relative abun- carbon,BAgraphitic coal,B or Acoal-like carbona- dance of highly carbonaceous deposits in Phanero- 124 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 zoic sediments is generally higher than in the Pre- tivity, organic biomass and sequestration of organic cambrian, in general there is a reduced abundance of carbon, sometimes may be unexpectedly high. carbonaceous sediments from the Precambrian to the In the endolithic community, organic carbon from Cenozoic. They state, for example, that there is more bacteria, cyanobacteria, lichens and algae is se- biogenic carbon in North American Grenville sequestered within rocks even in Antarctic dry valleys quences than in all Carboniferous deposits of the and in sandstones and granite rocks from the Mojave world! and Sonoran deserts in North America and the Negev Although there is limited direct evidence for biota in IsraelŽŽ Friedmann 1980; Friedmann et al., 1980; in the increasingly recognized number of Precam- McKay et al., 1983; Vestal, 1993; Bell, 1993. . Liv- brian and Early Paleozoic paleosols, there is a vari- ing biomass and primary production estimates for ety of kinds of direct evidence that indicates such endolithic communities living 1–3 mm inside the paleosols were far from abiotic. The evidence of rock in inhospitable cold and hot desert environ- laterites and high aluminous soilsŽ for example ments are admittedly low compared with tropical Gutzmer and Beukes, 1998. being recognized in rainforest, temperate deciduous forest or temperate increasing number in the Precambrian, clearly indi- grasslandŽ. Vestal, 1993, Table 2; Bell, 1993 but cate the necessity for what some would describe as a may compare favorably with some less productive AlushB tropical terrestrial microbial ground cover, if surface communities. Thus, Tuovila and LaRock lateritic soils were formed in the same way as they Ž.1987 report that 1 g of Antarctic rock contains at are today in tropical environmentsŽ Gutzmer and least as much living biomass as found in a liter of Beukes, 1998. . Laterites form under conditions of sea water and frequently 2 to 3 times more, with tropical, commonly monsoonal conditions. Since occasional samples approximating the biomass of there is good evidence for varied Proterozoic and some soilsŽ. Tuovila and LaRock, 1987 . In some Ordovician laterites, including ones with bauxite, extreme habitats, BellŽ. 1993 reports that the biomass there is reason to conclude that from a very early of some cryptoendolithic communities compares fa- date there was an extensive terrestrial vegetation vorably with, and in some cases equals or even capable of assisting in the laterization process, even exceeds, the Aabove ground biomassB of tracheo- though these paleosols are associated with small phytes. amounts of humusŽ. kerogen presumably related, as Endolithic communities in less stressed environ- in modern laterites, to the rapid microbial destruction ments may contribute far more significantly. Bell et of organic material. All of this is consistent with al.Ž. 1986 note that in cool–cold temperate, more pre-tracheophytic autotrophs having considerable po- mesic semi-desert environments of the Colorado tential for being involved with the subtraction of Plateau, prokaryotic cyanobacteria and eukaryotic atmospheric CO2 . All these kerogen-rich fossil ex- algae in the endolithic habitat make a significant amples sourced by microbial and algal populations contribution to their ecosystem through primary pro- clearly indicate the potential for preservation of ductivity and biomass accumulation with endolithic non-tracheophytic organic carbon. biomasses as much as 20 times higher than in the Below we provide evidence of productivity and Antarctic ecosystem. Wessels and BudelŽ. 1995 the potential for sequestration of organic carbon in studying the cryptoendolithic cyanobacterial commu- two types of modern habitats seemingly among the nity within a single sandstone cliff overhang least propitious biotic environments occupied by au- ŽMoses-Cave, Golden Gate Highlands National Park totrophic microorganisms: the endolithic and micro- in northeastern Orange Free State, South Africa. , bial crust. Both these environments likely sheltered estimated based on an average 1.5 mm thick zone pre-tracheophytic subaerial photosynthetic microbial, colonized by unicellular cyanobacteria, that the algal and fungal communities and are potentially cyanobacterial layer Arepresents 0.24 m2 of biologi- among the earliest terrestrial environments occupied cally active materialB over a surface area of 161 m2 by subaerial organisms, together with soils. Even within the cave and chlorophyll2 analysis indicated among these two seemingly least propitious biotic an appreciable quantity of cyanobacterial biomass environments, some evidence suggests that produc- within the sandstone. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 125

Bell and SommerfeldŽ. 1987 report that in com- Kidron et al.Ž. 1999 provide information about a parison with the cold temperate grassland on the modern desert crust composed of varied cyanobacte- Colorado Plateau, the cryptalgal coccoid blue-green rial taxa, and their ability to absorb moisture. The algal and coccoidrsarcinoid green algal community, other significant organic matter accumulations, de- so widespread in the sandstones of the region, Acon- rived from or composed of soil microorganisms, are tributes moderate biomassŽ. 5–10% and substantial those of translucent substrata that also often include photosynthesisŽ. 20–80% to the spare grassland cyanobacteria Ž.Azotobactor, Clostridium with the ecosystem.B Bell et al.Ž. 1988 report that cryptoen- ability to fix atmospheric nitrogenŽ Cameron and dolith communities on the Colorado Plateau shift Blank, 1966, pp. 24–25. . Diaphanous substrata also from coccoid cyanophytes to coccoid and sarcinoid encourage the most abundant and diverse popula- chlorophytes to accompany change in above ground tions of non-algal and non-bacterial microorganisms vegetation from cold-temperate desert to cold-tem- and accumulation of organic matter that is only perate forest and indicate that Athese endolithic or- Alimited by the penetration of light sufficient for ganisms may be contributing appreciable biomass algal photosynthesis and assimilationB ŽCameron and and photosynthetic activity to extreme and sparsely Blank, 1966, p. 28.Ž. . Kappen et al. 1975 point out vegetated habitats.B Bell et al.Ž. 1986, p. 435 con- that in extreme environments where there is a gener- clude that Aan ecological assessment of these habi- ally low productivity of higher plant biomass, the tats without consideration of the subsurface primary primary productivity of lichens is considerable, while producers is an incomplete one.B BellŽ. 1993 emphasizes the contribution made by With regard to desert crusts ŽAalgal and lichen soil cryptendolithic algae. crustsB.Žfound in semi-arid and arid ecosystems see Microbial crust, diaphanous substrate habitats, and Eldridge and Greene, 1994, for Australian examples. , endolithic microbial–algal communities are likely Cameron and BlankŽ. 1966 and Johansen Ž. 1993 AprimitiveB communities that have survived in simi- make a compelling case that these are good sources lar habitats from the earlier Precambrian. We may of organic carbon that accumulates in the soil below extrapolate from information about their biomass and the crusts. Favorable habitats for algalrbacterial productivity into the pre-tracheophytic past, as well growth and organic-matter accumulation are also as into the post-tracheophytic world in many areas found on the undersurfaces of translucent and trans- only poorly inhabited by tracheophytes, to conclude parent quartz and chalcedony pebbles imbedded in that such communities might have been significant A B wx the soil surface where meophilic mesophytic users of atmospheric CO2 as well as significant populations may survive because of various Amois- producers of organic carbon, some of which could ture-gathering mechanismsB associated with such have been sequestered during subsequent deposi- translucent substrates, and in any temporarily wet tional processes. habitat where water may poolŽ Cameron and Blank, Finally, using bryophytes as likely analogues to 1966. . Microbiotic crusts are water-stable surface pre-tracheophytic embryophytic land plants in the soil aggregates held together by bacteriaŽ hetero- pre-Devonian, we comment both on their commonly trophs, chemoautotrophs, aerobes, anaerobes, actino- substantial potential for high productivity and pre- mycetes.Ž , algae especially coccoid green algae . , served bryophyte phytomass as a carbon and CO2 fungiŽ. especially ascomycetous molds , lichens and sink like that hypothesized to be restricted to tra- mosses, generally dominated by filamentous cyano- cheophytes. BryophytesŽ. and lichens contribute sub- bacteria. Chemical analysis of these crusts show a stantial phytomass in a number of environments and comparatively high amount of organic matter relative vegetation typesŽ. Longton, 1992, Table 2.2 , al- to the surrounding soil, and Cameron and Blank though productivity of saxicolousŽ. rock-dwelling Ž.1966, p. 28 indicate that they are Aone of the most taxa, among other similar types, may be low. In important sources of soil organic matter in desert more oceanic cold polar regions for example, annual regionsB that often include cyanobacteria Ž.Nostoc production levels in closed moss turf and moss car- with the ability to fix atmospheric nitrogen, Athereby pet communities are comparable with those in tem- increasing the organic matter content of soil crusts.B perate grasslandŽ. Longton, 1992, p. 41 . Although 126 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 bryophyte phytomass in temperate deciduous forests Ž.1995, p. 308 report that peatlands occupy about a is often low, in Welsh oakwoodsŽ Longton, 1992, p. third of the total area of the taiga biome and can 48, citing a study by Rieley et al., 1979. bryophyte store several times more carbon per unit area than is phytomass is considerably higher than that for the implied by conventional soil carbon storage data. herb layer. Kallio and KarenlampiŽ. 1975 report on Funk et al.Ž.Ž 1994 citing others . report that boreal, the role of mosses and lichens as primary producers subarctic and arctic zone wetlands sequester about in different biomes, including regions where almost 28% of Earth’s soil carbon in undecomposed peat. the whole terrestrial ecosystem is based on their We can only speculate on the relevance of mod- primary productivity. Sveinbjornsson and Oechel ern peatlands to the tracheophyte-free pre-Devonian Ž.1992 note that in the Arctic taiga and northern bog when there is evidence for the presence of bryo- ecosystems, bryophyte productivity may easily ex- phyte-like embryophytes but limited evidence that ceed the productivity of tracheophytes, including can be extrapolated to indicate the presence of exten- black spruce forest of the taiga. Oechel et al.Ž. 1993 sive wetlands. Although there is no fossil evidence provide figures for stored carbon in soil active layers for what are today the major peatland formers and upper permafrost levels for northern ecosystems, Ž.Sphagnum until the Carboniferous—an argument including Arctic tundra, tussock tundra and wet tun- might be made to imply that the absence of bogs and dra that make clear that they are major carbon and wetlands prior to that time in geologic history—clad-

CO2 sinks of sufficient magnitude that recent change istic, molecular, and morphological evidence indi- from CO2 sink to source indicates the potential of cates that mosses alone may be the sister group to tundra ecosystems exerting a considerable positive the tracheophytes and that among the mosses, feedback on increasing atmospheric CO2 and global Sphagnum is primitive or ancestralŽ Mishler et al., warming. 1994. . The implications of such analyses are that The amount of carbon sequestered in undecom- Sphagnum or a Sphagnum-like ancestor is at least as posed peatlands is substantial in modern global old as the tracheophytes and potentially even older. ecosystems and it is possible to predict that it may The inference could also be drawn that if Sphagnum have been substantially higher during some time andror Sphagnum ancestors had the capacity for intervals of the geologic past. LongtonŽ. 1992, p. 63 forming bogs, such environments may well have Žciting the work of a number of others; see additional existed prior to the appearance of tracheophytes, and figures cited in Gorham, 1995; Woodwell et al., that such peatlands may have sequestered sufficient

1995. notes, for example, that peatlands now cover carbon to have affected atmospheric CO2 Ž although more than 1% of the land surface and can be calcu- in our statement below we say effectively that such lated to contain some 120,000=106 metric tons of peatlands cannot have been widespread because of carbon. ClymoŽ. 1987, p. 17 estimates that about 3% the broad arid climate regions so characteristic of of the land surface is peat-covered, mostly by the many past time intervals. . Although we do not re- so-called ombrotrophic peats, where bog-mosses are gard this as necessarily more speculative than argu- dominant plants, and that a single peatland in west- ing backwards in time from the significance of min- ern Siberia can cover 1800=800 km. An estimate eral weathering in today’s angiospermous soils of the for carbon presently accumulating annually in mires Temperate Zone to soils of the Paleozoic occupied from bryophyte phytomass is 80=106 metric tons by what may have been Alow weathering ecosys- and Longton notesŽ. 1992, p. 57 that a higher propor- temsB Ž.Volk, 1989 , such estimates cannot be ig- tion of moss Ž.Sphagnum production is preserved as nored for the pre-Carboniferous Phanerozoic when peat in such mires than of the associated predomi- they might even be modest for some time intervals. nantly herbaceous angiosperms. ClymoŽ. 1987 simi- Nor can it be assumed that the environmental cir- larly remarks on the slow decay of bog-mosses cumstances that would have led to bryophyte produc- compared with angiospermous remains. In the boreal tivity and sequestration of their remains are necessar- forest, 240=106 tons of carbon is fixed by annual ily the same as those related to tracheophytes, e.g. moss productionŽ. Longton, 1992 . Prentice and Sykes intervals of high bryophyte productivity and seques- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 127 tration of their remains cannot be assumed merely to Ž.1992, pp. 56–57 notes that bryophyte and lichen complement intervals of productivity and sequestra- tissues in some environments decompose more slowly tion associated with tracheophytes. than those of tracheophytes—as an example pointing The lack of a significant fossil record for bryo- out that in Fennoscandian tundra, lichen phytomass phytes until the Late Cenozoic and Quaternary, and loss after 2 years was 13%, while during the same for the taxa now involved in significant peat forma- interval, loss from tracheophyteŽ. angiosperm tissues tion, may in part be an artifact. This does not obviate was 37%. Low decomposition rates in bryophyte the possibility of peat-forming analogs even in the tissues may enable them to contribute sometimes event that it is not an artifact. The coal swamps of disproportionately to soil humus formation despite the Carboniferous and Permian have no close resem- relatively low annual productivityŽ Longton, 1992, p. blances to modern peatlands in their botanical re- 57. . It has also been noted that in some bryophytes mains or perhaps even in their chemistry, although it high levels of holocellulose and crude fiberŽ Walton, is possible to identify potential modern analogs 1985; Longton, 1992. may contribute to inhibiting Ž.Clymo, 1987 . Since many old peats are commonly decomposition much as lignin does in tracheophytes. structureless and the carbon source of the material Obviously, the point here is that whether organic can only be conjecturedŽ. Clymo, 1987, p. 21 , material contains lignin or not is irrelevant to its it is impossible to deny a potentially significant preservation as long as the time between an organ- contribution to Carboniferous peatlands by bryo- ism’s death and burial of its organic remains is rapid phytes analogous to Sphagnum in their capability for enough to preclude oxidation andror depositional differential preservation, just as other contributors to environments are anoxic like those that contribute to Carboniferous peats can be recognized in modern the slow decomposition of bryophyte tissues in acidic, analogs. waterlogged mires. LeeŽ. 1992 provides experimen- More importantly, these fossil and modern exam- tal data making it clear that the oxidation of organic ples noted above, indicate that recalcitrant tracheo- carbon sequestered in anoxic environments is a com- phyte tissues are not necessary for autotrophic and plex problem, with rates depending, among other non-autotrophic organisms to contribute significantly things, on items such as sedimentation rate and the to sedimentary marine and non-marine carbon stor- original carbon compounds present. age. These data, and perhaps especially the informa- These few modern examples, together with direct tion that deals with the high sequestration of organic information from the Precambrian, suggest caution carbon in the Precambrian, negate what Berner before eliminating a role for non-tracheophytic au- Ž.1993b, p. 374 states is among the two essential totrophs, including early embryophytes, and other contributions of tracheophytes to his atmospheric microbiota, in influencing atmospheric CO2 -levels models: A . . . the rise of vascular land plants during and as potential carbon sinks in the geological past the Paleozoic is believed to have brought about a before, and after, the appearance of the tracheo- large drop in atmospheric CO2 due to enhanced phytes. weathering by the plants and due to the increased burial of organic matter in sediments owing to a new 23.3. Summary source of refractory organic material Ž.e.g. lignin proÕided by the plantsB Ž.italics ours . We have Berner concludes that only the lignin-rich tracheo- discussed below one possible reason why there is no phytes produce large amounts of sequesterable or- indication of any significant organic carbon biomass ganic carbon ignoring the large amounts sequestered indicated in the interval of time of interest to us. in wet-land bog deposits in temperate and cool-tem- Although decomposition in bryophyte and lichen perate regions of the world and the considerable tissues may be rapid compared with tracheophytes, amounts, comparatively speaking, sequestered by decomposition rates are highly variable, and the microbial mats in arid and semi-arid regions, as assumption that they invariably exceed rates in tra- noted above. Large volumes of sequestered organic cheophyte tissue decomposition is not borne out by carbon in Precambrian sedimentary rocks, including the facts, as we have also noted above. Longton the non-marine environments, surpassing that of the 128 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Carboniferous, indicate that lignin is not a pre- mid-Paleozoic would have had far reaching biologi- requisite to organic carbon sequestration. If large cal effects in the marine ecosystem because it would volumes of organic carbon have been and can be necessarily imply a change in the nutrient source at sequestered by pre-tracheophytic organisms, it seems the base of the food chain from cellulose and hemi- equally clear that pre-tracheophytic, autotrophic or- cellulose-bearing phytoplankton to lignin-bearing de- ganisms, whether marine andror marine and non- tritus from land plants. This would mean a major marine, also can generate large amounts of organic loss of available food because the digestive capabil- carbon. The question can also be asked whether there ity for dealing with lignin is unknown in the physiol- is evidence to support the occurrence of any of these ogy of marine invertebrates and vertebrates, and such organisms in the terrestrial environment. an unprecedented change in food quality would One piece of evidence, relevant to the production doubtless have been registered in massive extinc- and sequestration of organic carbon, is that begin- tions. In this connection, RobinsonŽ. 1990a,b; 1991 ning at least by the Late Archaean, Schidlowski has argued that lignification was far higher in Paleo- Ž.1986, 1988 and Schidlowski and Aharon Ž. 1992 zoic plants than subsequently, which would have concluded that the ratio of inorganic carbon 13 to increased the impact of this terrestrial detritus on organic carbon 12 has remained relatively stable. marine invertebrates. While there is an extinction The inorganic carbon, richer in carbon 13 is gener- Ž.the Hangenberg Event followed by a significant ated from atmospheric CO2 dissolved in water that adaptive radiation near the Devonian–Carboniferous takes part in the weathering of silicate minerals in boundary, it is minor compared with what such a particular. These reactions give rise, among other major change in food resources would have entailed. things, to bicarbonate ions that subsequently precipi- If the increase of tracheophyte-generated organic tate out as limestone and dolomite, mostly in the carbon was not counterbalanced by a higher level of marine environment. The organic carbon, richer in inorganic carbonate sedimentation, the C12rC13 ra- carbon 12, is generated from varied organic pro- tio would not have remained stable over time. If cesses. The ratio indicates the relative dominance of organic carbon is increased over time, then inorganic the inorganic and organic isotope fractionating pro- carbonate has to also increase over time in order to cesses. maintain stability of the ratio. If anything there was a Since the mid-Paleozoic, the continuing stability higher level of pre-Late Devonian, platform marine of this ratio has not been affected by the rise, carbonate sedimentation than subsequently, which diversification or spread of the tracheophytes. The would mean an increased organic carbon production stability of the C12rC13 ratio throughout the Pre- on land andror in the sea prior to the diversification cambrian and Phanerozoic implies no major addition and spread of the tracheophytes. of carbon 12 into the system following advent of the In conclusion, if there has been no coeval massive tracheophytes, unless this addition was balanced by a addition of carbon 13 to the system following advent concurrent increase in inorganic carbon 13. The ab- of the tracheophytes, no massive loss in autotrophic sence of a massive increase in limestone and dolomite marine phytoplankton, and no massive addition or sedimentation in the mid-Paleozoic supports the ab- loss of inorganic carbon, it is only possible to con- sence of additive effects of organic carbon 12. Al- clude that prior to the advent of tracheophytes, the though for the Precambrian and Early Phanerozoic, it contribution of non-marine organic carbon to the C13 is commonly assumed that organic carbon-producing budget was more or less equivalent to that provided autotrophs were by and large confined to the marine by the tracheophytes that have now largely displaced environment, this is a questionable premise as we non-tracheophytic autotrophs. have discussed above and below. If we advance such a paradoxical argument to Stability of the ratio could imply a massive de- support terrestrial autotrophy in the Precambrian and crease in marine autotrophy concurrent with the ad- Early Phanerozoic, how can the limited evidence for vent of tracheophytes but without an accompanying non-marine coaly deposits in the Early Phanerozoic increase in carbonate sedimentation. The possibility be explained? One reasonable explanation can be of a massive decrease in marine autotrophy in the based on the climates of the Cambrian through the A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 129 earlier Devonian, where the majority of the pre- We shall discuss this model belowŽ. our Fig. 1 . served climate regions are warm arid regimeŽ Boucot We present a model that represents an estimate of et al., submitted. . In such regions, the organic car- Phanerozoic climatertemperature for comparison bon-generating capability of autotrophs would not with the CO2 curve generated in the Berner models. have been nearly as great as that provided by coeval Although we do not try to plot a global atmospheric humid tropical to warm-temperate regions. Pre- CO2 curve through time from these data, we assume, served, coeval, humid, tropical to warm temperate based on the greenhouse hypothesis, that relatively regions are limited although the potential areas of low global climatic gradients indicated by geological warm, humid region are largely based on paleogeo- data correspond to atmospheric CO2 levels signifi- graphic and paleoclimatic reconstructions, i.e. there cantly greater than those present during times of were probably extensive regions characterized by relatively high global climatic gradients such as the large amounts of earlier Paleozoic non-marine or- present. ganic carbon. The preserved Early Paleozoic humid There are a number of differences between esti- areas occur in regions thought to represent island arc mates of global climatertemperature implied by the type environments where basin conditions for the geological information and climatic estimates im- preservation of peats would have been very limited. plied by Berner’s CO2 curve. These discrepancies In addition, while many humid tropical to warm partly arise from the difficulty in comparing time temperate regions produce large volumes of organic units of different duration, because we use geological carbon, much of it is recycled, with sequestration time units while Berner uses an absolute time scale occurring only in regions where basin conditions are not directly tied into the relative time scale of the suitable. Evidence from the Late Cretaceous to Neo- geologist. gene lateritic bauxites bears this out since few are Below are comparison-criticisms of the other associated with peat or lignite, although they clearly models against the geological data. formed under heavy plant coverŽ Bardossy and Al- In the Berner atmospheric CO2 modelsŽ our Fig. eva, 1990. ; additionally, lateritic bauxites tend to 3. , there appear to be some key positive correlations form on more topographically high areas subject to between one or two independently verifiable climatic oxidation and recycling of organic matter. perturbations and inferences about CO2 levels. One of these verifiable climatic parameters is glaciation near the end of the Paleozoic. Another is indepen- 24. A climatic test for atmospheric models dent evidence of higher temperatures in the mid- 24.1. A geologic climatertemperature model CretaceousŽ. Berner, 1991b; Appenzeller, 1993 . Still another is higher temperatures in the Eocene. These r A B A potential climate temperature cum CO2 proxy are only three independent climatic points that can not previously considered by others is physical and be used to anchor CO2 speculations for an interval of biological evidence of climatically sensitive rocks time of about 550 Ma. There is no direct evidence

Žsuch as evaporites, coals, bauxites, kaolins, cal- for levels of atmospheric CO2 nor temperature dur- cretes, tillites, etc.. and organisms preserved in the ing these time intervals or any others prior to the geological record that permit independent estimates latest Quaternary, and estimates may range widely. of climatertemperature. Some independent Aclimatic anchors,B based on If the greenhouse hypothesis is correct, there single Aevents,B or limited time intervals, or on should be a positive correlation between changes in different methodologies, have not been so com- r atmospheric CO2 and climate temperature during pliant. For example, Raymo and RauŽ. 1992 failed to the last 600 m.y. as there appear to have been in the find a comparable match between the predicted last 160,000 years. Consequently, the best atmo- AgreenhouseB climate of the Pliocene and simulated spheric CO2 proxy would be independent estimates atmospheric curves when CO2 levels should have r = of temperature climate made from a variety of phys- been at least 1.5 to 2.0 ’s higher than modern CO2 ical and biological evidence preserved in the geolog- levels. Instead Raymo and RauŽ. 1992 and Raymo et ical record. al.Ž. 1996 found that mid-Pliocene CO2 levels were 130 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 not significantly different from today, or only slightly Middle Devonian, but in advance of that predicted higher than preindustrial values. Global warming by theoretical models. Even more telling in falsifying appears to have occurred in association with only the concept of unusually high atmospheric CO2 in slightly higher rather than significantly higher atmo- the Cambro-Devonian is the presence in the South- spheric CO2 levels, although Raymo et al.Ž. 1996 ern Hemisphere of widespread Middle Cambrian caution against possible biases associated with con- through earlier Middle DevonianŽ. Eifelian cool cli- tamination and diagenesis. Appearing to confirm the mate rocksŽ unweathered, detrital mica-bearing silici-

CO2 levels suggested by Raymo and RauŽ. 1992 and clastics; absence of evaporites, redbeds, limestone, Kurschner et al.Ž. 1996 found that maximum atmo- dolomite, lateritic products, residual kaolin, reefs. spheric CO2 levels based on stomatal parameters of and cool water faunasŽ Atlantic Realm of the Middle Late Miocene and Pliocene oak leavesŽ based, how- and Late Cambrian through Early Ordovician, ever, on only a single species. were in line with the Mediterranean Realm of the later Ordovician, Malvi- geochemical signals from the marine recordŽ Raymo nokaffric Realm of the Silurianw Afro–South Ameri- and Rau, 1992; Raymo et al., 1996. . We have al- can Subrealmx and lower half of the Devonian.Ž our ready noted above some potential difficulties with Fig. 1. . Some of these will be discussed below. interpretations based on behavior of a single species If changes in global climate and atmospheric CO2 and potentially with stomatal indices in general for are mutually involved and, causally related, a trace evaluating atmospheric CO2 . of climatic change through Phanerozoic time pro- Barron and WashingtonŽ. 1985 estimate CO2 lev- vides one possible test of atmospheric models based = els for the Cretaceous that range from 2 to 10 ’s on fluctuations in CO2 Ž. our Fig. 1 . During times of the present, while Berner estimates ranges of CO2 of continental glaciation at high latitudes, the climatic 2to3=’s the present. Arthur et al.Ž. 1985a,b, 1991 belts are compressed. These are times of high global = estimate CO2 levels in the mid-Cretaceous 8 ’s the climatic gradient. During times of warm-temperate present as opposed to Berner’s estimates of 3=’s. climate at high latitudes, the climatic belts are ex- These different levels would obviously have climatic panded. These are times of low climatic gradient. consequences at variance with those implicated by Changes in atmospheric CO2 throughout the Berner’s CO2 levels. Phanerozoic should correlate with changing global There are a number of pre-Carboniferous intervals climatesŽ the so-called greenhouse effect with CO2 - where the geologicalrclimatic data are clearly dis- linked temperature changes; MacDonald, 1990, who cordant with the predicted CO2 levels of the Berner cautions that the apparent modern correlation does model. Crowley and BaumŽ. 1991, 1995 , for exam- not necessarily reflect cause and effect; Woodwell ple, reported a Aperplexing paleoclimate paradoxB and Mackenzie, 1995. . Globally low climatic gradi- Žwidespread continental glaciation in a time of over- ent, overall warm intervals, such as the Late Pale- all very low global climatic gradient indicated by the ocene–Early Eocene, should have a different at- allegedly very high atmospheric CO2 . . Unless one mospheric CO2 concentration than globally high assumes that the Berner curve is factual, there is no climatic, overall cool–cold intervals, such as the basis for any paradox: glaciation in the Late Ordovi- Quaternary. Evidence for global climatic change, cian coincident with predicted extremely high atmo- based on diverse climatic proxiesŽ coal, evaporites, = spheric CO2 Ž 14–16 present atmospheric levels tillites, calcretes, bauxites and others. , should pro- according to Kump et al., 1995. when a greenhouse vide the best estimates of global CO2 levels on the climate should have been present according to the assumption that surface temperatures correlate with modelsŽ. we discuss this in some detail below . A CO2 . similar paradox exists between latest Late Devonian It is relevant here to mention that Owen et al. continental glaciationŽ Caputo, 1985; Caputo and Ž.1979 point out that the reduced solar luminosity Crowell, 1985; Frakes et al., 1992; Streel, 1986, present early on in Earth history, which implies 1992; Isaacson et al., 1999. and theoretical predic- lower Earth surface temperatures, could be satisfied tions of high atmospheric CO2 , although Mora et al. in terms of a more steady surface temperature by a Ž.1996 predict a rapid drawdown in CO2 by the significant increase in atmospheric CO2 , greenhouse A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 131 gas, earlier in Earth history, but this solar luminosity other and against the Berner models and the climatic effect during as short an interval as the Cambro– data, it is clear that there is actually little real accord Silurian would not be significant. If one assumes that that can be considered as anything but coincidence. the Cambro-Silurian atmospheric CO2 level was higher than at present, in order to keep surface 24.2. The OrdoÕician glaciation as an example of temperatures at more normal levels there still are the problem problems with Berner’s model. During as short an interval as the Cambro-Silurian, one would not pre- Berner, as well as Kump et al.Ž. 1995 , suggest dict significant solar changes, whereas the geological that their model’s failure to recognize the widespread data indicates a relatively warm Early Cambrian, a Late Ordovician, including the Hirnantian, Southern significantly cooler Middle Cambrian into the Late Hemisphere, glacial interval might reflect the rela- Ordovician, a glacial Late Ordovician, followed by a tively short time interval involved. However, the significantly cooler Silurian, Early Devonian and South African Pakhuis Tillite dataŽ. Boucot, 1999 earlier Middle Devonian, i.e. inconsistent with the makes it clear that the Late Ordovician glacial inter- Berner curveŽ. compare our Figs. 1 and 3 . val includes much more than the Hirnantian, the We highlight a few of the problems, principally latest Ashgillian, and may include more than the using the Berner atmospheric model. Because our underlying, mid-Ashgillian Rawtheyan and Caut- emphasis is on the earlier Paleozoic, we stress cli- leyan stages, even extending down into the basal matic discordance with the Berner model for that Pusgillian Stage or beyond into the later Caradocian time interval; we provide two other examples from ŽTheron, 1994, placed the basal Pakhuis into the later younger time intervals to make it clear that the Caradocian. . problem is not restricted to the earlier Paleozoic. If The relatively brief Hirnantian, latest Ordovician the BernerŽ. 1997 curve Ž our Fig. 3 . is superimposed glacial interval, needs to also consider the South on our Fig. 1, it is clear that there is overall poor African data. In South Africa, the pre-Hirnantian, correlation between atmospheric CO2 model esti- later Ordovician Pakhuis Tillite is overlain by the mates and the changing global climatic gradients Late Ordovician Soom Shale, which may represent a determined from geological data. For example, low pre-Hirnantian somewhat warmer interval, possibly climatic gradients occur in the Early Cambrian, the equivalent to the globally overall warmer Rawtheyan, late Middle and Late Devonian, Early Triassic, and indicated by its giant conodontsŽ conodonts are ab- Paleocene through Eocene. These are intervals when sent from truly cold water environments in the Pale- the hypothesized CO2 values are widely different, ozoic, and the presence of giant conodonts is con- yet all have similar global climatic gradients. By the sistent with a cooler water environment than is same token, the Berner CO2 values for the interval normally present with conodonts. . The Soom Shale Middle Cambrian through early Middle Devonian is in turn overlain by the cool climate, Hirnantian are basically the same, although the equivalent cli- age Disa Siltstone. All of this information suggests matic gradients show marked changes. Equivalent that in South Africa, the cold environment, including discrepancies are clear between the climatic data and the Pakhuis Tillite, probably occupied a far longer the other atmospheric composition curves. time interval than the Hirnantian alone. Supportive

What does this disparity between CO2 concentra- of a Caradoc age for the Pakhuis Tillite is widespread tion and Phanerozoic climates indicate? If we elimi- evidence from North Africa in MoroccoŽ Hamoumi, nate the question of a sampling problem, we must 2001.Ž and Libya Massa et al., 1977 . for glacioma- conclude either that the geological climatic data have rine sediments that include varied dropstone occur- been wrongly interpreted or that model assumptions rences. about changes in fluctuations in the Earth’s atmo- Yapp and Poths’Ž. 1996 data, referred to previ- spheric CO2 concentration through time are in error. ously, led them to conclude that the Late Ordovician Although there are also some intervals of accord was a time characterized by 16= present levels of between Berner’s curve and the climatic data, when atmospheric CO2 , whereas the geological dataŽ our one considers the other five models against each Fig. 1. falsifies their conclusion. The Late Ordo- 132 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 vician, latest Ashgillian Hirnantian and earlier Ashgill, that might be equivalent in age to the South Ashgillian Pakhuis, glaciation occurred in the middle African Pakhuis Tillite. The Pakhuis, although com- of a supposedly high atmospheric CO2 interval monly assumed to be of Hirnantian age, is probably Ž.Crowley and Baum, 1991, 1995 . This glaciation significantly older as discussed earlier. The Pakhuis occurred when atmospheric CO2 was projected is older than the Hirnantian, but there still is evi- Ž.Berner, 1997 to have been approximately 16 Ž our dence for a cold South African, Hirnantian climate Fig. 3. times what it is in the present atmosphere. represented by the Mediterranean Realm fauna of the Isotopic proxies from goethite in a paleosol also Disa Siltstone Member of the Cedarberg Formation, predict high atmospheric pCO2 Ž Yapp and Poths, which bears on the position of the Pole well to the 1992; Kump et al., 1995. . All of this high atmo- south of where Crowley and BaumŽ. 1991, 1995 spheric CO2 for the Late Ordovician is falsified by placed it. the geological dataŽ. our Fig. 1 . One could, of SmithŽ. 1997 likewise concluded that the Hirnan- course, have both a moderately high global climatic tian South Pole in Africa should have been far to the gradient and very high atmospheric CO2 if the gas southeast of the position assigned it by Crowley and had no greenhouse effect, a possibility that most BaumŽ. 1991, 1995 . But Smith too failed to place the consider very unlikely. pole far enough to the south because he overlooked A glacial interval would normally call for a rela- the demonstrated presence of later Ordovician glacia- tively low concentration of atmospheric CO2 . The tion in South Africa. Sutcliffe et al.’sŽ 2000, their CO2 -enriched atmosphere would normally be ex- Fig. 1A. outlined occurrences of Late Ordovician pected to have favored a AgreenhouseB climate. Sev- tillites in a manner indicating a pole position in eral workersŽ Brenchley et al., 1994; Kump et al., central Africa consistent with the data. 1995; Wang et al., 1997. have remarked on the Wang et al.Ž. 1997 made another attempt to rec- limited time span of the Hirnantian glaciation. Al- oncile the allegedly discordant, predicted CO2 levels though two of these papers did not suggest that the suggested by the Berner models. They suggested that discordance was related to the short time spanŽ Kump an increase in the marine organic carbon isotope et al., 1995 do suggest that possibility. , some might ratio in Hirnantian sedimentary rocks, with examples suppose that this is the case. We see no basis for from the Yangtze Sea, may have resulted from a such a conclusion as detailed here. decrease in the dissolved surface water CO2 , which Crowley and BaumŽ. 1991, 1995 provided several in turn could have reflected a significant decrease in r potential explanations for this alleged climatic ocean atmospheric pCO2 levels. As they note, anomaly. They suggested that the Ordovician South however, none of this information can be quantified Pole was located in northwestern Africa, in a near Abecause there still exist so many uncertainties with shoreline position. Such a polar position might per- respect to some important variables . . . B ŽWang et mit a General Circulation ModelŽ. GCM explanation al., 1997, p. 157. . consistent with both continental glaciation and a high Kump et al.Ž. 1995 and Gibbs et al. Ž. 1997 tried atmospheric CO2 concentration. However, Crowley to explain the Hirnantian glacial interval as the pos- and BaumŽ. 1991, 1995 did not consider the pres- sible result of the sequestering of a large amount of ence of coeval glacial deposits in the mid-Andean organic carbon that led to significant drawdown of regions of Bolivia, southern Peru and northern Ar- atmospheric CO2 , information that they state they do gentina, or the Parana´ Basin of southern Brazil, nor not have. They also prepared varied GCMs consis- the pre-Hirnantian, probably earlier Ashgillian tent with non-glacial to heavily glacial conditions by

Pakhuis Tillite of South Africa, all of which invali- varying the amounts of atmospheric CO2 . Their con- date the position of their South Pole, and make a clusions are reasonable in principle, although the South Pole situated somewhere in Central Africa far absence of any geological evidence for unusually more reasonable. Wang et al.Ž. 1997 illustrate pre- large amounts of sequestered organic carbon in the Hirnantian, Ashgillian isotopic anomalies in several Hirnantian beds globally is unsupportive. Still, one of their sectionsŽ. Wangjiawan, Fengxian that are could appeal to abyssal organic carbon storage, al- consistent with glaciation in the pre-Hirnantian though the non-recognition of any Hirnantian abyssal A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 133 strata means that the idea is based on negative In conclusion, the Late Ordovician glaciation is evidence. Gibbs et al.Ž. 1997 use Crowley and discordant with alleged high atmospheric CO2 levels Baum’sŽ. 1991, 1995 conclusions based on the projected for some models in the Early Phanerozoic. latter’s erroneous placement of the Late Ordovician There is no objective evidence for recognizing high

South Pole. atmospheric CO2 levels in the Late Ordovician shown Kump et al.Ž. 1999 attempt still another possible in most models. Attempts to support a hypothetical explanation for Late Ordovician continental glacia- projection and not something for which there is tion. They suggest that Taconic age orogenyŽ more objective reality—the glaciation—are not needed. or less Middle and Late Ordovician. made possible There is no positive evidence to support a failure in more extensive continental rock weathering that drew any of the models to show a CO2 drawdown in the down atmospheric CO2 significantly, plus an inter- Late Ordovician. Use of the word anomaly in refer- pretation of oxygen and carbonŽ organic and inor- ence to Late Ordovician, high southern latitude ganic. isotope ratios consistent with both Hirnantian glaciation, is a misleading choice of words. continental glaciation and significantly higher levels of atmospheric CO2 than during the Quaternary. 24.3. A Permo–Triassic boundary example They do not provide evidence that there was enough orogenic activity of Taconic age to provide the Visscher et al.Ž. 1996 provide palynological evi- needed amount of rock weathering, nor that the dence for a large increase in fungal activity at the significantly earlier Taconic Orogeny and alleged Permo–Triassic boundary which they interpret as weathering is consistent with latest Ordovician, Hir- indicating a massive higher land plant extinction. If nantian continental glaciation. Good evidence for the contemporary vulcanismŽ. Kozur, 1998 was respon-

Taconic Orogeny is largely restricted to the North sible for a marked increase in atmospheric CO2 in Atlantic region from northerly parts of the Central the Permo–Triassic boundary region, as Visscher et Appalachians through the Northern Appalachians, al. conclude, it would have been still more increased and parts of the Caledonides in Scandinavia and by the woody plant decay, fungally assisted, or if Britain. If one were to select a highly orogenic they are mistaken about the importance of a volcanic Ordovician interval, it would probably be the component, by woody plant decay alone. If a mas- widespread orogenies of the Late Arenig–Llanvirn, sive increase in atmospheric CO2 is combined with well known from many parts of North America, the information about the virtual absence of coal de- northern European Caledonides, Central and South- posits in the Early Triassic, their rarity in the Middle ern Europe, plus the Himalayas, the Andean region Triassic, and their abundance in the Late Triassic, it of South America, possibly the Urals and southeast- is possible to predict that the Early and Middle ern Kazakhstan at a minimum; there is no positive Triassic should be relatively high atmospheric CO2 evidence for widespread continental glaciation dur- intervals, whereas the reverse might be the case for ing the late Arenig–Llanvirn interval. If glaciation the Late Triassic. Berner’sŽ.Ž 1997 our Fig. 3 . Trias- were so tightly coupled with widespread orogeny sic, and possibly part of his Late Permian curve, is one wonders why the widespread, Southern Hemi- inconsistent with this geological information. His sphere, later Cambrian phase of the Pan-African Early and Middle Triassic atmospheric CO2 level Orogeny does not coincide with widespread conti- should have been much higher, especially if one nental glaciation. As they indicate, there are alterna- prefers the additive effects of volcanic and woody tive explanations for their isotopic data. plant decomposition, and his Late Triassic much Sutcliffe et al.’sŽ. 2000 correlation of the Pakhuis lower. Tillite with the North African tillites does not explain the possibly warmer water later Ordovician 24.4. A Cretaceous example Soom Shale large conodonts. The precise age of the Pakhuis is still a problem, with pre-Hirnantian ages Krassilov’sŽ. 1973, 1975, 1994, his Fig. 2 paleob- as low as later Caradoc reasonable possibilities, rather otanical-based conclusions about changing Creta- than Hirnantian. ceous temperature trends in the Northern Hemi- 134 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 sphere do not support the Berner atmospheric CO2 water to high latitudes by oceanic surface currents is curve. In comparing geologically precise data points, a means of raising the global temperature. such as those provided by Krassilov, it is difficult to However, theyŽ. Pearson and Palmer, 2000 later make direct comparisons with Berner’s atmospheric suggest, employing the same methodology, that at-

CO2 curves, because Berner used a purely numerical mospheric CO2 was much higher overall during the rather than a geological time scale without making Paleogene than in the Neogene. Their methodology clear where the different stage boundaries were may well provide superior estimates of atmospheric placed. Krassilov notes a relative temperature decline CO2 than have been previously available. from a Late Jurassic high to a Barremian low, fol- Beerling and JolleyŽ. 1998 discuss the possibility lowed by an Aptian increase, then an Albian in- that a large negative carbon 13 excursion together crease, ultimately leading to a Campanian high fol- with the incoming of AparatropicalB pollen taxa, lowed by a Maastrichtian decrease. Our comparison coinciding with an increase in detrital kaolin, are all of Krassilov’s conclusions with Berner’sŽ 1997, p. indicative of a Paleocene–Eocene warming event, 545. curve shows disagreement in terms of Berner’s based on data from northwestern Europe. They hy- which suggests a more or less steady decrease from pothesize a global Ashort-term extreme warm eventB the Late JurassicŽ. 155 Ma into the Paleogene, which with marked perturbation of the ocean–atmosphere also disagrees with the available geological–climatic system related to the transfer of 12 C in the form of dataŽ. our Fig. 1 . CO24 or CH from a large ocean carbon reservoir into a smaller atmospheric reservoir that resulted in a

24.5. Eocene examples short-lived high CO2 episode, i.e. a correlation between elevated atmospheric CO2 and elevated tem- Pearson and PalmerŽ. 1999, p. 1824 point out that perature. levels of Middle to Late Eocene atmospheric CO2 Pearson and PalmerŽ. 1999 suggest an alternative A B A are . . . uncertain . . . , but that . . . atmospheric to using estimated atmospheric CO2 as a proxy for pCO2 was probably similar to modern concentrations temperature is to employ the transport by ocean or slightly higher.B Their data suggests a decoupling surface currents of warm water to higher latitudes between atmospheric CO2 and surface temperature. under a significantly differing paleogeographic situa- CowlingŽ. 1999; see also Royer et al., 2001 also tion than that of the present. Pearson and Palmer raises the distinct possibility of surface temperature Ž.1999 also indicate that their boron isotope method and atmospheric CO2 being decoupled. The paleon- would benefit from further testing, while pointing tological and lithofacies data, the geological informa- out that they are assuming constant seawater compo- tion, all indicate that the Eocene was one of the sition through time. Data presented by Pagani et al. overall warmest intervals in Earth history. Using this Ž.1999 tend to support Pearson and Palmer’s conclu- geological information, many, including Berner sions. Pagani et al.’sŽ. 1999 find that C12rC 13 ratios

Ž.1998 , have inferred high levels of atmospheric CO2 in the relatively warm Lower and Middle Miocene while employing varied modeling approaches. were significantly lower than pre-industrial values, Pearson and PalmerŽ. 1999 employ boron isotopic correlated with a glacial maximum near the Oli- data to estimate dissolved CO2 and pH in seawater, gocene–Miocene boundary. Pagani et al.Ž 1999, p. which in turn is employed to estimate atmospheric 289. conclude that the major controls over climate in A CO2 . They note Seasonal temperature cycles and this time interval were tectonic, paleogeographic and biological processes mean that the pCO2 of seawater ocean current circulation patterns rather than levels in the surface mixed layer is not in perfect equilib- of atmospheric CO2 . rium with the atmosphere, but this deviation is gen- - erally 10% in the absence of upwelling of CO2 -rich 24.6. A Pliocene example deep waters.B This means that oceanic seawater is a reasonable proxy for atmospheric CO2 . If, as they We have discussed above an example from the suggest, atmospheric CO2 in the Middle Eocene was Pliocene that indicates that a similar type of discor- about the same as at present, then transport of warm dance can be found very late in time when Raymo A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 135 and RauŽ. 1992 failed to find a comparable match Ž.Fig. 1 . This similarity casts further doubt on between the predicted AgreenhouseB climate of the Berner’s conclusions and the proxies from which

Pliocene and simulated atmospheric CO2 curves. they have been drawn. This example indicated that global warming appears If Shackleton’sŽ. 2000 conclusion that the overall to have occurred in association with only slightly climatic control involves the Milankovitch mecha- higher rather than significantly higher atmospheric nisms, particularly the 100,000 year cycles, strongly

CO2 levelsŽ Raymo and Rau, 1992; Raymo et al., correlating atmospheric CO2 and temperature rather 1996. . than continental ice volume, the latter seen as an effect of the first two, then one is left to conclude that major Phanerozoic climatic gradient changes are 25. Summary and conclusions chiefly a result of the interactions of changing geography. These geographic changes involve changing In summary, considerable disparity exists between land mass and continental positions, changing orog- the curves generated by the varied modelsŽ compare raphy and changing ocean current strengths and posi- our Figs. 2–8. . The wide disparities between the tions. Autotrophs from this perspective are viewed various published curves suggests that the presently more as a result than a cause, which also concurs published models are inadequate. Considerable dis- with Schidlowski’s carbon isotope investigations. parity also exists between all the models and the BoucotŽ. 1990, pp. 543–545 has briefly made the geological climatic evidence indicating changing cli- point that during times of globally low climatic matic gradients through the PhanerozoicŽ. our Fig. 1 . gradient the Milankovitch effects, which will have This indicates, based on present knowledge of cli- been present since the very beginning, will be hard mates of the geological past, that there is no simple to make out in the rock record, as contrasted with straightforward relationŽ see MacDonald, 1990, for times of very high global climatic gradient such as caution concerning this correlation. between levels the Quaternary. of atmospheric CO2 , as estimated by the various In the Berner models, perturbations in atmo- modelers and changes in the global climatic gradient. spheric CO2 are causally correlated with the appear- If climate is the nearest proxy to atmospheric CO2 ance of tracheophytes and related to assumptions and level, then the geologically generated climatic infor- speculations about tracheophyte productivity, tra- mation should agree closely with the models; this is cheophyte-enhanced weathering rates, sequestration not the case. We submit that the modelers need to of tracheophyte-generated carbon, and amounts of rethink many of their basic assumptions. We suspect limestone and dolomite. We suggest that these corre- that by combining the information on the locationŽ. s lations may be largely fortuitous. The atmospheric through time of major upland regions, the Raymo models that Budyko et al., Tappan, Berner, Raymo and Ruddiman approach, with information about ma- and Francois et al. have proposed are all different. jor landmass positions through time, the Worsley et All are unsubstantiated. al. approach, one might achieve a far better approxi- The overall weathering problem may be encapsu- mation than has been achieved by any prior model. lated by looking at the extremes. At freezing tem- RaymoŽ. 1997 places the modeling problems in the peratures, very little chemical rock weathering will proper context: AData collection has not kept up with occur regardless of other factors, such as uplift, the modeling efforts in this field and more data, not precipitation, plant cover, etc. At tropical tempera- more models, is what’s needed.B We concur whole- tures, very little chemical rock weathering will occur heartedly with her statement. beneath a thick cover of lateritic materials in the We are favorably impressed with the high level of absence of considerable relief, regardless of plant concurrence between Veizer et al.’sŽ. 2000 account cover, etc. In the absence of adequate precipitation, of Phanerozoic oxygen isotopic fluctuations and our regardless of temperature and uplift, there will be Fig. 1, indicating that temperatures deduced from very little rock weathering. In other words, it is their isotopic work changed in a manner reasonably clearly the interaction between a reasonable amount close to our own changes in global climatic gradient of uplift that exposes unweathered silicate minerals, 136 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 plus adequate precipitationrrunoff, all taking place control will turn out to have been multifactorial. at reasonably high temperatures that permits rock Currently, there is much interest being shown in weathering to go forward. The role of both higher intermittent release of methane hydrates from the and lower autotrophs under varied climatic condi- continental shelves and margins, just as in the past tions extending from the tropical to the cold polar, to interest has centered more on organic carbon burial, the arid, is not well understood in a quantitative strontium isotopes, levels of silicate weathering, both manner. However, the relative stability of the continental and oceanic, changing terrestrial relief, C12rC13 ratio over time indicates that major sources changing continental positions, and so forth. Hayden of organic carbon have been present since the very Ž.1998 points out the importance of biogenically beginning, although clearly the relative contributions generated greenhouse gases and of biogenic water made by higher and lower autotrophs is uncertain. vapor in modulating climate, with no prior attempt Berner’s thesis assumes that during the reign of by modelers to take these factors into consideration the tracheophytes, beginning in the Devonian, there for past time intervals. Some of the uncertainties in will have been a high level of weathering affecting this complex system might be partially removed by silicate minerals through time that has remained solid summaries of varied classes of geological data essentially fixed regardless of changes affecting any such as carbonate rocks through time, kerogen other variables. Is this reasonable? The basic ques- through time, coal through time, etc. Still, it will be a tion is whether or not one can extrapolate from the long time before we are in command of adequate weathering characteristic of the modern, North Tem- data to uniquely solve the dilemma. In the meantime, perate zone rhizosphere to the entire Phanerozoic? we do have the summary of the geological–climatic One must consider whether or not the ever-changing gradient indicatorsŽ. our Fig. 1 . positions of the continents, platforms and montane We argue that, even if it is theoretically possible regions have remained constant during the Phanero- to provide reliable estimates for some of the factors zoic, with the answer being that they have not. involved in such models, the overall number of Global Phanerozoic geology suggests that most of complex variables and the potential for integrating the continental areas were relatively peneplaned for them in any meaningful way, particularly in the lengthy time intervals, and relatively close to sea pre-Cenozoic, makes such models uncertain. There level, i.e. montane regions of high relief and high are uncertainties even in the Late Cenozoic, but more levels of rock weathering that characterize the Qua- or less realistic estimates can be made for some of ternary are abnormal geologically. When one com- the factors involved. BolinŽ.Ž 1986, p. 423 see also bines this information with the relatively stable Sigman and Boyle, 2000. emphasizes that we have C12rC13 ratio through time, it may well be that the difficulty understanding the present carbon situation: photic zone of the seas and oceans plays a significant APresent models, however, are inadequate for more role in controlling the levels of atmospheric CO2 precise projections into the future. Even though the during tracheophytic and pre-tracheophytic times. role of the oceans is better understood than that of During intervals of globally maximum orogeny, there the terrestrial biota, considerable improvements in will obviously be more weathering, particularly if the the modelling of all parts of the carbon cycle are areas affected are largely within the contemporary essential. It is not until the response characteristics warm, humid regions. are well described on all time scales for a few years BroeckerŽ. 1998 has summed up the problem of to one or two centuries that we will be able to overall atmospheric CO2 control admirably, when analyze more subtle processes, which may be quite pointing out that there has to be a ApolicemanB important when projecting until the time when atmo- through time that maintains a relatively stable level, spheric CO2 concentration will have doubled. De- just as Schidlowski has done over the years with his tailed models are needed for comparison of what is massive studies of the organic:inorganic carbon iso- implied by a synthesis of our present knowledge in tope ratio relative stability questions. It is clear that terms of a model and the real behavior of the carbon there is no simple answer concerning the identity of cycle as revealed by observations.B SchimelŽ. 1998 , the ApolicemanB. It is probable that the ultimate Sarmiento et al.Ž. 1998 and Cao and Woodward A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 137

Ž.1998 provide still more insight into the serious ing with organisms without taxonomic and ecologi- problems involved with understanding the carbon cal models in the modern world, arrive at indepen- cycle at the present time, so critical to any under- dent conclusions rather than conclusions that are standing of atmospheric CO2 through time. This predicated on correlation with a single atmospheric being the case for the present, there is little wonder model. that major uncertainties attend any consideration of We suggest that the assumptions that underlie all the distant past! Finally, McCauley and DePaolo such models, based on sweeping, commonly simplis- Ž.1997 in an elegant paper make it very clear how tic generalizations through vast amounts of geologi- many of the variables involved in the different mod- cal time, need testing in terms of some of the els remain poorly constrained. geological methodologies discussed above to provide Ramstein et al.Ž. 1997 emphasize the real impor- more realistic information about the amounts of or- tance for understanding past climates of having as ganic carbon and carbonate rocks preserved in Early detailed, reliable paleogeography as possible. They Devonian and pre-Devonian strata. We urge testing point out that Oligocene to present climatic models as well by means of experiments with living or that ignore such things as the major climatic effect of AcomparableB representatives of early embryophytes the Oligocene–earlier Miocene Paratethys, as well as and lower autotrophs to provide information about Himalayan uplift, give rise to climatic conclusions in their potential for contributing to sedimentary or- disagreement with paleobotanical data from central ganic carbon storage during the Cambrian, Ordovi- Eurasia, as well as with geological data. The climatic cian and Silurian. We urge extensive greenhouse models we have summarized pay far too little atten- experimentation with modern autotrophs from varied tion to the details of past paleogeography, even when cyanobacteria to varied angiosperms, using the right they are readily available, i.e. there is no easy, quick mineral substrates under carefully controlled temper- road to climatic understanding, including potentially ature and moisture conditions, as well as using Amix- B changing levels of atmospheric CO2 of past time tures of some of the lower autotrophs to see if there intervals. One must be willing to carefully synthesize are any synergistic effects. We have provided an the massive, available geological data. The com- estimateŽ. our Fig. 1 of changing Phanerozoic cli- monly employed alternative of trying to substitute matic gradients. We have not tried to use this infor- inadequate, although readily available, proxies re- mation to estimate changing levels of atmospheric sults in lack of both reliability and understanding. CO2 because of the many, many unevaluated vari- The complexities of mycorrhizal distribution to- ables which this would involve. day in relation to taxonomic position and ecology dictate that caution be the rule when trying to evaluate the role of mycorrhizae during the past. It is Acknowledgements unreasonable to assume that all past and present tracheophytes, living in all environments and cli- This is a totally cooperative paper in which the mates, were intimately coevolved with mycorrhizae. authors shared joint responsibility for most of the We regard it as unfortunate that paleobotanists, ideas expressed. Gray’s untimely demise in January, among others, have attempted to justify conclusions 2000, prevented her from seeing the final draft. about tracheophyte evolution, diversification and Boucot has done his best to complete the paper, morphology in the Paleozoic from Berner‘s atmo- including the non-geological portions for which Gray spheric models, using apparent correlations in sup- had the chief responsibility. port of cause and effect relationships. A more rea- We are grateful to Christopher Scotese for criti- listic approach to test paleobotanical expectations cally reviewing the information shown on our Fig. 1, involves use of direct climatic data from the geologi- and for making useful suggestions about the for- cal record. This would encourage paleobotanists to mat and the manuscript. Heather Viles, University consider Amultiple working hypothesesB, a concept of Oxford, generously guided us to pertinent litera- that seems to have fallen out of favor. It is essential, ture, and provided helpful advice concerning the however, that paleobotanists, especially those work- manuscript. Patrick V. Brady, Sandia National Labo- 138 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 ratories, Albuquerque, NM, provided very extensive Amy, P.S., Haldeman, D.L.Ž. Eds. , 1997. The Microbiology of the helpful advice regarding a number of questions and Terrestrial Deep Subsurface. CRC Press LLC, Boca Raton, provided two reviews of the entire manuscript. James 356 pp. Anagnostidis, K., Gehrmann, C.K., Gross, M., Krumbein, W.E., M. Trappe, Oregon State University, contributed Lisi, S., Pantazidou, A., Urzi, C., Zagari, M., 1991. Biodeteri- considerably to the consideration of mycorrhizae and oration of marbles of the Parthenon and Propylaea, Acropolis, the mycorrhizal condition, and contributed gener- Athens Aassociated organisms, decay and treatment sugges- ously from his wide fund of knowledge. Kermit tionsB. In: Decrouez, D., Chamay, J., Zezza, Z.Ž. Eds. , The Cromack, Oregon State University, was very helpful Conservation of Monuments in the Mediterranean Basin. Mu- seum d’Histoire naturelle and Musee´ d’art et d’histoire, Ville with information regarding mycorrhizae and their de Geneve,` pp. 305–325. weathering potentialities. We are indebted to Robert Andrews, J.E., Tandon, S.K., Dennis, P.F., 1995. Concentration of Berner, Yale University, for his careful review of an carbon dioxide in the Late Cretaceous atmosphere. Journal of earlier draft, despite the fact that we disagree basi- the Geological SocietyŽ. London 152, 1–3. cally on many points. Finally, we are very indebted Aplet, G.H., Hughes, R.F., Vitousek, P.M., 1998. Ecosystem development on Hawaiian lava flows: biomass and species to Anthony Hallam for his careful editorial review of composition. Journal of Vegetation Science 9, 17–26. the entire manuscript, and for making a number of Appenzeller, T., 1992. Deep-living microbes mount a relentless penetrating suggestions, as well as providing numer- attack on rock. Science 258, 222. ous suggestions for improving the overall presenta- Appenzeller, T., 1993. Searching for clues to ancient carbon tion. dioxide. Science 259, 908–909. April, R., Newton, R., 1992. Mineralogy and mineral weathering. In: Johnson, D.W., Lindberg, S.E.Ž. Eds. , Atmospheric Depo- sition and Forest Nutrient Cycling: A Synthesis of the Inte- References grated Forest Study. Springer-Verlag, New York, pp. 378–425. Arino, X., Ortega-Calvo, J.J., Gomez-Bolea, A., Saiz-Jimenez, C., Adamo, P., Marchetiello, A., Violante, P., 1993. The weathering 1995. Lichen colonization of the Roman pavement at Baelo of mafic rocks by lichens. Lichenologist 25, 285–297. ClaudiaŽ. Cadiz, Spain : biodeterioration vs. bioprotection. The Ahmadjian, V., 1995. Lichens—specialized groups of parasitic Science of the Total Environment 167, 353–363. fungi. In: Kohmoto, K., Singh, U.S., Singh, R.P.Ž. Eds. , Arthur, M.A., Fahey, T.J., 1993. Controls on soil solution chem- Pathogenesis and Host Specificity in Plant Diseases, Vol. II. istry in a subalpine forest in North-Central Colorado. Journal Eucaryotes. Pergamon, pp. 277–288. of the Soil Science Society of America 57, 1122–1130. Ahmadjian, V., Jacobs, J.B., 1983. Algal–fungal relationships in Arthur, M.A., Dean, W.A., Pratt, L.M., 1985a. Geochemical and lichens: recognition, synthesis, and development. In: Goff, L.J. climatic effects of increased marine organic carbon burial at Ž.Ed. , Algal Symbiosis. Cambridge Univ. Press, Cambridge, the CenomanianrTuronian boundary. Nature 335, 714–717. pp. 147–172. Arthur, M.A., Dean, W.E., Schlanger, S.O., 1985b. Variations in Ahn, J.H., Cho, M., Buseck, P.R., 1999. Interstratification of the global carbon cycle during the Cretaceous related to

carbonaceous material within illite. American Mineralogist 84, climate, volcanism, and changes in atmospheric CO2 . In: 1967–1970. Sundquist, E.T., Broecker, W.S.Ž. Eds. , The Carbon Cycle and

Alexopoulos, C.J., Mims, C.W., Blackwell, M., 1996. Introduc- Atmospheric C02 : Natural Variations Archean to Present. tory Mycology. 4th edn. Wiley, New York, 868 pp. American Geophysical Union, Geophysical Monograph, vol. Algeo, T.J., Berner, R.A., Maynard, J.B., Scheckler, S.E., 1995. 32, pp. 504–529. Late Devonian oceanic anoxic events and biotic crises: Arthur, M.A., Allard, D., Hinga, K.R., 1991. Cretaceous and ArootedB in the evolution of vascular land plants? GSA Today Cenozoic atmospheric carbon dioxide variations and past global 5Ž. 45 , 64–66. climatic change. Geological Society of America Program with Algeo, T.J., Scheckler, S.E., Maynard, J.B., 2001. Effects of the Abstracts 23, A178. Middle to Late Devonian spread of vascular land plants on Ascaso, C., Galvan, J., Ortega, C., 1976. The pedogenic action of weathering regimes, marine biotas, and global climate. In: Parmelia conspersa, Rhizocarpon geographicum and Umbili- Gensel, P.G., Edwards, D.Ž. Eds. , Plants Invade the Land. caria pustulata. Lichenologist 8, 151–171. Columbia Univ. Press, NYC, pp. 213–236. Avakyan, Z.A., Karavalko, G.I., Mel’Nikova, E.O., Drutkso, V.S., Allen, M.F., 1991. The Ecology of Mycorrhizae. Cambridge Ostroushko, Y., 1981. Role of microscopic fungi in weather- Studies in Ecology. Cambridge Univ. Press, Cambridge, 184 ing of rocks and minerals from a pegmatite deposit. Mikrobi- pp. ologiya 50, 156–162. Allen, E.B., Allen, M.F., 1990. The mediation of competition by Bagyaraj, D.J., 1984. Biological interactions with VA mycorrhizal mycorrhiza in successional and patchy environments. In: fungi. In: Powell, C.L., Bagyaraj, D.J.Ž. Eds. , VA Mycorrhiza. Grace, J.B., Tolman, G.D.Ž. Eds. , Perspectives on Plant Com- CRC Press, Boca Raton, pp. 131–186. petition. Academic Press, New York, pp. 367–389. Bakker, J.P., 1970. Zur okologie einer blaualgenkruste in den A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 139

feuchten tropen auf dem biotitgranit einer inselberglandschaft Eocene transition in NW Europe. Journal of the Geological Ž.Voltzberggebeit-Surinam . Colloquium Geographicum 12, Society of London 155, 591–594. 152–176. Beerling, D.J., Woodward, F.I., 1996. Stomatal density responses Bakker, J.C., 1991. Effects of humidity on stomatal density and its to global environmental change. In: Stanhill, G.Ž. Ed. , Ad- relation to leaf conductance. Scientia Horticulturae 48, 205– vances in Bioclimatology, vol. 4. Springer-Verlag, Berlin, pp. 212. 171–221. Baldauf, S.L., Palmer, J.D., 1993. Animals and fungi are each Beerling, D.J., Osborn, C.P., Chaloner, W.G., 2001. Evolution of

others closest relatives—congruent evidence from multiple leaf-form in land plants linked to atmospheric CO2 decline in proteins. Proceedings of the National Academy of Sciences of the Late Palaeozoic era. Nature 410, 352–354. the United States of America 90, 11558–11562. Bell, R.A., 1993. Cryptoendolithic algae of hot semiarid lands and Balkwill, D.L., 1989. Numbers, diversity, and morphological deserts. Journal of Phycology 29, 133–139. characteristics of aerobic, chemoheterotrophic bacteria in deep Bell, R.A., Sommerfeld, M.R., 1987. Algal biomass and primary subsurface sediments from a site in South Carolina. Geomicro- production within a temperate zone sandstone. American Jour- biology Journal 7, 33–52. nal of Botany 74, 294–297. Banfield, J.F., Nealson, K.H.Ž. Eds. , 1997. Geomicrobiology: Bell, R.A., Athey, P.V., Sommerfeld, M.R., 1986. Cryptoen- Interactions between Microbes and Minerals. Reviews in Min- dolithic algal communities of the Colorado Plateau. Journal of eralogy, vol. 35. Mineralogical Society of America, Washing- Phycology 22, 429–435. ton, DC, 448 pp. Bell, R.A., Athey, P.V., Sommerfeld, M.R., 1988. Distribution of Bardossy, G., Aleva, G.J.J., 1990. Lateritic Bauxite. Elsevier, endolithic algae on the Colorado Plateau of northern Arizona. Amsterdam, 624 pp. The Southwestern Naturalist 33, 315–322. Bar-Yosef, B., 1996. Root excretions and their environmental Bennett, P.C., Casey, W., 1994. Chemistry and mechanisms of effects: influence on availability of phosphorus. In: Waisel, Y., low-temperature dissolution of silicates by organic acids. In: Eshel, A., Kafkafi, U.Ž. Eds. , Plant Roots the Hidden Half. Pittman, E.D., Lewan, M.D.Ž. Eds. , Organic Acids in Geologi- Marcel Dekker, New York, pp. 581–605. cal Processes. Springer-Verlag, Berlin, pp. 162–200. Barker, W.W., Banfield, J.F., 1996. Biologically versus inorgani- Bennett, P.C., Hiebert, F.K., Choi, W.J., 1996. Microbial colo- cally mediated weathering reactions: relationships between nization and weathering of silicates in a petroleum-con- minerals and extracellular microbial polymers in lithobiontic taminated groundwater. Chemical Geology 132, 45–53. communities. Chemical Geology 132, 55–69. Berbee, M.L., Taylor, J.W., 1993. Dating the evolutionary radia- Barker, W.W., Banfield, J.F., 1998. Zones of chemical and physi- tions of the true fungi. Canadian Journal of Botany 71, cal interaction at interfaces between microbial communities 1114–1127. and minerals: a model. Geomicrobiology Journal 15, 223–244. Berner, R.A., 1982. Burial of organic carbon and pyrite sulfur in Barker, W.W., Welch, S.A., Banfield, J.F., 1997. Biogeochemical the modern ocean: Its geochemical and environmental signifi- weathering of silicate minerals. In: Banfield, J.F., Nealson, cance. American Journal of Science 282, 451–473. K.H.Ž. Eds. , Geomicrobiology: Interactions Between Microbes Berner, R.A., 1990. Atmospheric carbon dioxide levels over and Minerals. Reviews in Mineralogy, vol. 35. Mineralogical Phanerozoic time. Science 249, 1382–1386. Society of America, Washington, DC, pp. 391–428. Berner, R.A., 1991a. Land plants and weathering: discussion and Barker, W.W., Welch, S.A., Chu, S., Banfield, J.F., 1998. Experi- response. Science 252, 860.

mental observation of the effect of bacteria on aluminosilicate Berner, R.A., 1991b. A model for atmospheric CO2 over Phanero- weathering. American Mineralogist 83, 1551–1563. zoic time. American Journal of Science 291, 339–376. Barnola, J.M., Raynaud, D., Korotkevich, Y.S., Lorius, C., 1987. Berner, R.A., 1992. Weathering, plants, and the long-term carbon Vostok ice core provides 160,000-year record of atmospheric cycle. Geochimica et Cosmochimica Acta 56, 3225–3231.

CO2 . Nature 329, 408–414. Berner, R.A., 1993a. Paleozoic atmospheric CO2 : importance of Barron, E.J., Washington, W.M., 1985. Warm Cretaceous cli- solar radiation and plant evolution. Science 261, 68–70.

mates: high atmospheric CO2 as a plausible mechanism. In: Berner, R.A., 1993b. Weathering and its effect on atmospheric Sundquist, E.T.W., Broecker, W.S.Ž. Eds. , The Carbon Cycle CO2 over Phanerozoic time. Chemical Geology 107, 373–374. and Atmospheric CO2 : Natural Variations Archean to Present. Berner, R.A., 1994. GEOCARB II: a revised model of atmo- American Geophysical Union Geophysical Monograph, vol. spheric CO2 over Phanerozoic time. American Journal of 32, pp. 546–553. Science 294, 56–91. Bazin, M.J., Markham, P., Scott, E.M., Lynch, J.M., 1990. Popu- Berner, R.A., 1995. Chemical weathering and its effect on atmo-

lation dynamics and rhizosphere interactions. In: Lynch, J.M. spheric CO2 and climate. In: White, A.F., Brantley, S.L. Ž.Ed. , The Rhizosphere. Wiley, Chichester, pp. 99–127. Ž.Eds. , Chemical Weathering Rates of Silicate Minerals. Re- Beerbower, J.R., 1985. Early development of continental ecosys- views in Mineralogy, vol. 31. Mineralogical Society of Amer- tems. In: Tiffney, B.H.Ž. Ed. , Geological Factors and the ica, Washington, DC, pp. 565–583. Evolution of Plants. Yale Univ. Press, New Haven, pp. 47–91. Berner, R.A., 1997. The rise of plants and their effect on weather-

Beerling, D.J., Jolley, D.W., 1998. Fossil plants record an atmo- ing and atmospheric CO2 . Science 276, 544–546. 12 spheric CO2 and temperature spike across the Palaeocene– Berner, R.A., 1998. The carbon cycle and CO2 over Phanerozoic 140 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

time: the role of land plants. Philosophical Transactions of the mycorrhizae. In: Powell, C.L., Bagyaraj, D.J.Ž. Eds. , VA Royal Society of London, Series B 353, 75–82. Mycorrhiza. CRC Press, Boca Raton, FL, pp. 5–33. Berner, R.A., 2001. The effect of the rise of land plants on Bonfante-Fasola, P., Fontana, A., 1985. VAM fungi in Ginkgo

atmospheric CO2 during the Paleozoic. In: Gensel, P.G., biloba roots and their interactions at the cellular level. Sym- Edwards, D.Ž. Eds. , Plants Invade the Land. Columbia Univ. biosis 1, 53–67. Press, NYC, pp. 173–178. Bormann, B.T.D., Wang, F.H., Bormann, F.H., Benoit, G., April, Berner, R.A., Berner, E.K., 1997. Silicate weathering and climate. R., Snyder, M.C., 1998. Rapid, plant-induced weathering in an In: Ruddiman, W.F.Ž. Ed. , Tectonic Uplift and Climate Change. aggrading experimental ecosystem. Biogeochemistry 43, 129– Plenum, New York, pp. 353–365. 155. Berner, R.A., Canfield, D.E., 1989. A new model for atmospheric Boucot, A.J., 1990. Evolutionary Paleobiology of Behavior and oxygen over Phanerozoic time. American Journal of Science Coevolution. Elsevier, Amsterdam, 725 pp. 289, 333–361. Boucot, A.J., 1999. Southern African Phanerozoic marine inverte- Berner, R.A., Lasaga, A.C., Garrels, R.M., 1983. The carbonate- brates: biogeography, palaeoecology, climatology and com- silica geochemical cycle and it’s effect on atmospheric carbon ments on adjacent regions. Journal of African Earth Sciences dioxide over the past 100 million years. American Journal of 28, 129–143. Science 283, 641–683. Boucot, A.J., Gray, J., Fang, R., Yang, X., Li, Z.-P., Zhang, N., Berry, W.B.N., Wilde, P., 1978. Progressive ventilation of the 1982. Devonian calcrete from China: its significance as the oceans—an explanation for the distribution of the Lower first Devonian calcrete in Asia. Canadian Journal of Earth Paleozoic black shales. American Journal of Science 278, Sciences 19, 1532–1534. 257–275. Boullard, B., 1979. Considerations sur la symbiose fongique chez Berthelin, J., 1983. Microbial weathering processes. In: Krumbein, les Pteridophytes. Syllogeus 19, 1–59. W.E.Ž. Ed. , Microbial Geochemistry. Blackwell, Oxford, pp. Boullard, B., 1988. Observations on the coevolution of fungi with 223–262. hepatics. In: Pirozynski, K.A., Hawksworth, D.L.Ž. Eds. , Co- Berthelin, J., 1988. Microbial weathering processes in natural evolution of Fungi with Plants and Animals. Academic Press, environments. In: Lerman, A., Meybeck, M.Ž. Eds. , Physical London, pp. 107–124. and Chemical Weathering in Geochemical Cycles. Kluwer Boullard, B., Lemoigne, Y., 1971. Les champignons endophytes Academic Publishing, Dordrecht, pp. 33–59. du ARhynia gwynne-vaughaniiB K et L. Le Botaniste 54, Bianciotto, V., Bandi, C., Minerdi, D., Sironi, M., Tichy, H.V., 49–89. Bonfonte, P., 1996. An obligately endosymbiotic mycorrhizal Bouwman, A.F.Ž. Ed. , 1990. Soils and the Greenhouse Effect. fungus itself harbours obligately intracellular bacteria. Applied Wiley, Chichester, 575 pp. and Environmental Microbiology 62, 3005–3010. Bowen, G.D., 1980. Misconceptions, concepts and approaches in Bickel, M.J., 1998. The need for mass balance and feedback in the rhizosphere biology. In: Ellwood, D.C., Latham, M.J., Hedger, geochemical carbon cycle: comment and reply. Geology 26, J.N., Lynch, J.M., Slater, J.H.Ž. Eds. , Contemporary Microbial 411–414. Ecology. Academic Press, London, pp. 283–304. Blazquez, F., Calvet, F., Vendrell, M., 1995. Lichenic alteration Boyle, J.R., Voigt, G.K., 1973. Biological weathering of silicate and mineralization in calcareous monuments of northeastern minerals. Plant and Soil 38, 191–201. Spain. Geomicrobiology Journal 13, 223–247. Brady, P.V., 1991. The effect of silicate weathering on global

Blum, J.D., 1997. The effect of Late Cenozoic glaciation and temperature and atmospheric CO2 . Journal of Geophysical tectonic uplift on silicate weathering rates and the marine Research 96Ž. B11 , 18101–18117. 87r 86 Sr Sr record. In: Ruddiman, W.F.Ž. Ed. , Tectonic Uplift Brady, P.V., Carroll, S.A., 1994. Direct effects of CO2 and and Climate Change. Plenum, New York, pp. 259–288. temperature on silicate weathering: possible implications for Blum, J.D., Gazis, C.A., Jacobson, A.D., Chamberlin, C.P., 1998. climate control. Geochimica et Cosmochimica Acta 58Ž. 8 , Carbonate versus silicate weathering in the Raikhot watershed 1853–1856. within the High Himalayan Crystalline Series. Geology 26, Brady, P.V., Gislason, S.R., 1997. Seafloor weathering controls

411–414. on atmospheric CO2 and global climate. Geochimica et Cos- Bluth, G.J.S., Kump, L.R., 1991. Phanerozoic paleogeology. mochimica Acta 61Ž. 5 , 965–973. American Journal of Science 291, 284–308. Brady, P.V., House, W.A., 1996. Surface-controlled dissolution Bois, C., Bouche, P., Pelet, R., 1982. Global geologic history and and growth of minerals. In: Brady, P.V.Ž. Ed. , Physics and distribution of hydrocarbon reserves. American Association of Chemistry of Mineral Surfaces. CRC Press, Boca Raton, pp. Petroleum Geologists Bulletin 66, 1248–1270. 225–305. Bolin, B., 1986. Requirements for a satisfactory model of the Brady, P.V., Zachara, J.M., 1996. Geochemical applications of global carbon cycle and current status of modeling efforts. In: mineral surface science. In: Brady, P.V.Ž. Ed. , Physics and Trabalka, J.R., Reichle, D.E.Ž. Eds. , The Changing Carbon Chemistry of Mineral Surfaces. CRC Press, Boca Raton, pp. Cycle: A Global Analysis. Springer-Verlag, New York, pp. 307–356. 403–424. Brady, P.V., Dorn, R.I., Brazel, A.J., Clark, J., Moore, R.B., Bonfante-Fasolo, P., 1984. Anatomy and morphology of VA Glidewell, T., 1999. Direct measurement of the combined A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 141

effects of lichen, rainfall, and temperature on silicate weather- In: Whelan, J.K., Farrington, J.W.Ž. Eds. , Organic Matter: ing. Geochimica et Cosmochimica Acta 63, 3293–3300. Productivity, Accumulation, and Preservation in Recent and Brasseur, G.P., Orlando, J.J., Tyndall, G.S.Ž. Eds. , 1999. Atmo- Ancient Sediments. Columbia Univ. Press, New York, pp. spheric Chemistry and Global Change. Oxford Univ. Press, 231–263. Oxford, 654 pp. Cameron, R.E., Blank, G.B., 1966. Desert algae: soil crusts and Breger, I.A., Brown, A., 1962. Kerogen in the Chattanooga Shale. diaphanous substrata as algal habitats. Technical Report No. Science 137, 221–224. 32-971. Jet Propulsion Laboratory, California Institute of Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, Technology, Pasadena, 41 pp. D.B.R., Long, D.G.F., Meidla, T., Hints, L., Anderson, T.F., Campbell, R., Greaves, M.P., 1990. Anatomy and community 1994. Bathymetric and isotopic evidence for a short-lived Late structure of the rhizosphere. In: Lynch, J.M.Ž. Ed. , The Rhizo- Ordovician glaciation in a greenhouse period. Geology 22, sphere. Wiley, Chichester, pp. 11–34. 295–298. Cao, M., Woodward, F.I., 1998. Dynamic responses of terrestrial Brock, T.D., Madigan, M.T., Martinko, J.M., Parker, J., 1994. ecosystem carbon cycling to global climate change. Nature Biology of Microorganisms. 7th edn. Prentice-Hall, Engle- 393, 249–252. wood Cliffs, NJ, 909 pp. Capo, R.C., Chadwick, O.A., 1999. Sources of strontium and Broecker, W., 1997. Will our ride into the greenhouse future be a calcium in desert soil and calcrete. Earth and Planetary Sci- smooth one? GSA Today 7Ž. 5 , 1–7 May. ence Letters 170, 61–77.

Broecker, W.S., 1998. Does atmospheric CO2 police the rate of Caputo, M.V., 1985. Late Devonian glaciation in South America. chemical weathering? Global Biogeochemical Cycles 12, 403– Palaeogeography, Palaeoclimatology and Palaeoecology 51, 408. 291–317. Broecker, W., 1999. Climate change prediction. Science 283 Caputo, M.V., Crowell, J.C., 1985. Migration of glacial centers Ž.5399 , 179. across Gondwana during Paleozoic era. Geological Society of

Broecker, W.S., Sanyal, A., 1998. Does atmospheric CO2 police America Bulletin 96, 1020–1036. the rate of chemical weathering. Global Biogeochemical Cy- Cawley, J.L., Burruss, R.C., Holland, H.D., 1969. Chemical cles 12, 403–408. weathering in central Iceland: an analog of pre-Silurian weath- Brundrett, M.C., 1991. Mycorrhizas in natural ecosystems. Ad- ering. Science 165, 391–392. vances in Ecological Research 21, 171–313. Cazares, E., 1997. Mycorrhizal fungi and their relationship to Budyko, M.I., 1986. The Evolution of the Biosphere. Reidel plant succession in subalpine habitats. PhD thesis, Oregon Publishing, Dordrecht, 423 pp.Ž translated from the 1984 State University. edition by M.I. Budyko, S.F. Lemeshko and V.G. Yanuta. . Cazares, E., Trappe, J.M., 1993. Vesicular endophytes in roots of Budyko, M.I., Ronov, A.B., Yanshin, A.L., 1987. History of the the Pinaceae. Mycorrhiza 2, 153–156. Earth’s Atmosphere. Springer-Verlag, Berlin, 139 pp.Ž Trans- Cerling, T.E., 1991. Carbon dioxide in the atmosphere: evidence lated from Russian by S.F. Lemeshko and V.G. Yanuta. . from Cenozoic and Mesozoic paleosols. American Journal of Bull, P.A., Laverty, M., 1982. Observations on phytokarst. Science 291, 377–400. Zeitschrift fur¨ Geomorphologie N.F. 26, 437–457. Cerling, T.E., 1992. Use of carbon isotopes in paleosols as an Burns, R.G., 1986. Interaction of enzymes with soil mineral and indicator of the PŽ. CO2 of the paleoatmosphere. Global Bio- organic colloids. In: Huang, P.M., Schnitzer, M.Ž. Eds. , Inter- geochemical Cycles 6, 307–314. actions of Soil Minerals with Natural Organics and Microbes. Cerling, T.E., Quade, J., 1993. Stable carbon and oxygen isotopes Soil Science Society of America Special, vol. 17, pp. 430–451. in soil carbonates. In: Swart, P.K., Lohmann, K.C., McKenzie, Buseck, P.R., Galdobina, L.P., Kovalevski, V.V., Rozhkova, N.N., J., Savin, S.Ž. Eds. , Climate Change in Continental Isotopic Valley, J.W., Zaidenberg, A.Z., 1997. Shungites: the C-rich Records. Geophysical Monograph, vol. 78, pp. 217–231. rocks of Karelia, Russia. The Canadian Mineralogist 35, Chamberlin, T.C., 1899. An attempt to frame a working hypothe- 1363–1378. sis on the cause of glacial periods on an atmospheric basis. Cain, R.F., 1972. Evolution of the fungi. Mycologia 64, 1–14. Journal of Geology 7, 545, 584, 667–685, 751–787. Caldeira, K., 1995. Long-term control of atmospheric carbon Chapelle, F.H., Bradley, P.M., 1997. Alteration of aquifer geo- dioxide: low-temperature seafloor alteration or terrestrial sili- chemistry by microorganisms. In: Hurst, C.J., Knudsen, G.R., cate-rock weathering. American Journal of Science 295, McInerney, M.J., Stetzenbach, L.D., Walter, M.V.Ž. Eds. , 1077–1114. Manual of Environmental Microbiology. American Society for Caldeira, K., Kasting, J.F., 1992. The life span of the biosphere Microbiology Press, Washington, DC, pp. 558–564. revisited. Nature 360, 721–723. Chapelle, F.H., Zelibor Jr., J.L., Grimes, D.J., Knobel, L.L., 1987. Caldeira, K., Arthur, M.A., Berner, R.A., Lasaga, A.C., 1993. Bacteria in deep Coastal Plain sediments of Maryland: a

Cooling in the late Cenozoic. Nature 361, 123–124. possible source of CO2 in groundwater. Water Resources Caldwell, B.A., Jumpponen, A., Trappe, J.M., 2000. Utilization of Research 23, 1625–1632. major detrital substrates by dark-septate, root endophytes. Chapelle, F.H., Morris, J.T., McMahon, P.B., Zelibor Jr., J.L., Mycologia 92, 230–232. 1988. Bacterial metabolism and the delta 13 C composition of Calvert, S.E., Pedersen, T.F., 1992. Organic carbon accumulation ground water, Floridan aquifer system, South Carolina. Geol- and preservation in marine sediments: how important is anoxia? ogy 16, 117–121. 142 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Clymo, R.S., 1987. Rainwater-fed peat as a precursor of coal. In: Curl, E.A., Truelove, B., 1986. The Rhizosphere. Springer-Verlag, Scott, A.C.Ž. Ed. , Coal and Coal-Bearing Strata: Recent Ad- Berlin, 288 pp. vances. Geological Society Special Publication, vol. 32, Danin, A., Garty, J., 1983. Distribution of cyanobacteria and their Blackwell, Oxford, pp. 17–23. impact on biogenic weathering. Zeitschrift fur¨ Geomorpholo- Cochran, M.F., 1995. Weathering, plants and the geochemical gie N.F. 27, 423–444. carbon-cycle. PhD Dissertation, Yale University, New Haven, Danin, A., Gerson, R., Garty, J., 1983. Weathering patterns of Connecticut. hard limestone and dolomite by endolithic lichens and Cochran, M.F., Berner, R.A., 1993a. Enhancement of silicate cyanobacteria: supporting evidence for aeolian action of crus- weathering rates by vascular land plants; quantifying the ef- taceous lichens to Terra Rossa soil. Soil Science 1361, 213– fect. Chemical Geology 107, 213–215. 217. Cochran, M.F., Berner, R.A., 1993b. Reply to the comments on Davis, C., Pratt, L., Sliter, W., Mompart, L., Murat, B., 1999. AWeathering, plants, and the long-term carbon cycleB. Factors influencing organic carbon and trace metal accumula- Geochimica et Cosmochimica Acta 57, 2147–2148. tion in the Upper Cretaceous La Luna Formation of the Cochran, M.F., Berner, R.A., 1996. Promotion of chemical weath- Western Maracaibo Basin, Venezuela, Special Paper 332. ering by higher plants: field observations on Hawaiian basalts. Geological Society of America, 203–230. Chemical Geology 132, 71–77. Debaud, J.C., Marmeisse, R., Gay, G., 1999. Intraspecific genetic Cochran, M.F., Berner, R.A., Snyder, M.C., Wang, D., 1994. variation and populations of ectomycorrhizal fungi. In: Varma, Chemical weathering enhancement by pines and grasses: re- A., Hock, B.Ž. Eds. , Mycorrhiza. Springer, Berlin, pp. 75–110. sults from the Hubbard Brook Sandbox experiment. Transac- De la Torre, M.A., Gomez-Alarcon, G., 1994. Manganese and tions of the American Geophysical Union, EOS Supplement iron oxidation by fungi isolated from building stone. Microbial 75, 697. Ecology 27, 177–188. Cooks, J., Fourie, Y., 1990. The influence of two epilithic lichens De la Torre, M.A., Gomez-Alarcon, G., Palacios, J.M., 1993. In on the weathering of quartzite and gabbro. South African vitro biofilm formation by Penicillium frequentans strains on Geographer 17Ž. 1r2 , 24–33. sandstone, granite, and limestone. Appl. Microbiol. Biotech- Cooks, J., Otto, E., 1990. The weathering effects of the lichen nol. 40, 408–415. Lecidea aff. sarcogynoidesŽ. Koerb. on Magaliesberg quartzite. Desborough, G.A., Poole, F.G., Daws, T.A., Scarborough, R., Earth Surface Processes and Landforms 15, 491–500.2. 1984. Hydrocarbon source rock evaluation of the Middle Cooper, K.M., 1975. Growth responses to the formation of en- Proterozoic Apache Group, Gila County, Arizona. In: Wood- dotrophic mycorrhizas in Solanum, Leptospermum, and New ward, J., Meissner, F.F., Clayton, J.L.Ž. Eds. , Rocky Mountain Zealand ferns. In: Sanders, F.E., Mosse, B., Tinker, P.B. Association of Geologists, Hydrocarbon Source Rocks of the Ž.Eds. , Endomycorrhizas. Academic Press, London, pp. 391– Greater Rocky Mountain Region, pp. 57–81. 407. Diakumaku, E., Ausset, P., Sterflinger, K., Wollenzien, U., Krum- Cooper, K.M., 1976. A field survey of mycorrhizas in New bein, W.E., Lefevre, R.A., 1994. On the problem of rock Zealand ferns. New Zealand Journal of Botany 14, 169–181. blackening by fly-ash, fungal and other biogenic particles, and Cooper, K.M., 1984. In: Bagyaraj, D.J., Powell, C.L.Ž. Eds. , their detection in Mediterranean marbles and monuments. In: Physiology of VA Mycorrhizal Associations. CRC Press, Boca Fassina, V., Ott, H., Zezza, F.Ž. Eds. , The Conservation of Raton, FL, pp. 155–186. Monuments in the Mediterranean Basin. Museum d’Histoire

Cowling, S.A., 1999. Plants and temperature—CO2 uncoupling. naturelle and Musee´` d’art et d’histoire, Ville de Geneve, pp. Science 285, 1500–1501. 305–310. Cromack, K., Sollins, P., Graustein, W.C., Speidel, K., Todd, Doolittle, J.H., Feng, D.F., Tsang, S., Cho, G., Little, E., 1996. A.V., Spycher, G., Li, C.Y., Todd, R.L., 1979. Calcium Determining divergence times of the major kingdoms of living oxalate accumulation and soil weathering in mats of the organisms with a protein clock. Science 271, 470–477. hypogeous fungus Hysterangium crassum. Soil Biology and Dormaar, J.F., 1968. Infrared absorption spectra of mineral matter Biochemistry 11, 463–468. in saxicolous lichens and associated mosses. Canadian Journal Cronan, C.S., 1985. Chemical weathering and solution chemistry of Earth Sciences 5Ž. 2 , 223–230. in acid forest soils: differential influence of soil type, bio- Dorn, R.I., 1998. Rock coatings. Developments in Earth Surface tic processes, and Hq deposition. In: Drever, J.I.Ž. Ed. , Processes, vol. 6 Elsevier, Amsterdam, 429 pp. The Chemistry of Weathering. Reidel, Dordrecht, pp. 175– Dorn, R.I., Brady, P.V., 1995. Rock-based measurement of tem- 195. perature-dependent plagioclase weathering. Geochimica et Crowell, J.C., 1983. Ice ages recorded on Gondwanan continents. Cosmochimica Acta 59Ž. 13 , 2847–2852. Transactions of the Geological Society of South Africa 86, Drever, J.I., 1994. The effect of land plants on weathering rates of 238–262. silicate minerals. Geochimica et Cosmochimica Acta 58Ž. 10 , Crowley, T.J., Baum, S.K., 1991. Toward reconciliation of Late 2325–2332. y Ordovician Ž.440 Ma glaciation with very high CO2 levels. Drever, J.I., Clow, D.W., 1995. Weathering rates in catchments. Journal of Geophysical Research 96, 22597–22610. In: White, A.F., Brantley, S.L.Ž. Eds. , Chemical Weathering Crowley, T.J., Baum, S.K., 1995. Reconciling Late Ordovician Rates of Silicate Minerals. Reviews in Mineralogy, vol. 31, = Ž.440 Ma glaciation with very high Ž. 14 CO2 levels. Journal Mineralogical Society of America, Washington, DC, pp. 464– of Geophysical Research 100, 1093–1101. 483. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 143

Drever, J.I., Vance, G.F., 1994. Role of soil organic acids in the Guayana Shield in Venezuela, Colombia, and Brazil. mineral weathering processes. In: Pittman, E.D.Ž. Ed. , Organic Geochimica et Cosmochimica Acta 59, 3301–3325. Acids in Geological Process. Springer-Verlag, Berlin, pp. Edwards, H.G.M., 1997. FT-Raman microscopy of lichen encrus- 138–161. tations. USA Microscopy and Analysis, May, pp. 13–15. Drever, J.I., Zobrist, J., 1992. Chemical weathering of silicate Edwards, D., Axe, L., 1992. Stomata and mechanics of stomatal rocks as a function of elevation in the southern Swiss Alps. functioning in some early land plants. Courier Forschungsinsti- Geochimica et Cosmochimica Acta 56, 3209–3216. tut Senckenberg 147, 59–73. Drew, M.C., 1990. Root function, development, growth and min- Edwards, D., Abbott, G.D., Raven, J.A., 1996. Cuticles of early eral nutrition. In: Lynch, J.M.Ž. Ed. , The Rhizosphere. Wiley, land plants: a palaeoecophysiological evaluation. In: Kerstiens, Chichester, pp. 35–57. G.Ž. Ed. , Plant Cuticles. BIOS Scientific Publishers, Oxford, Driese, S.G., Mora, C.I., 1993. Introduction and primary objec- pp. 1–31. tives of field excursion. In: Driese, S.G.Ž. Ed. , Paleosols, Ehrlich, H.L., 1995. Geomicrobiology. 3rd edn. Marcel Dekker,

Paleoclimate and Paleoatmospheric CO2 ; Paleozoic Paleosols New York, 719 pp. of Central Pennsylvania. University of Tennessee, Department Ehrlich, H.L., 1996. How microbes influence mineral growth and of Geological Sciences. Studies in Geology, vol. 22, pp. dissolution. Chemical Geology 132, 5–9. 22–33. Ehrlich, H.L., 1998. Geomicrobiology: its significance for geol- Driese, S.G., Mora, C.I., Cotter, E., 1993. Buffalo Mountain ogy. Earth-Science Reviews 45, 45–60. paleosol, Duncannon Member of Catskill FormationŽ Upper Ekart, D.D., Cerling, T.E., Montanez, I.P., Tabor, N.J., 1999. A Devonian.Ž. near Newport, PA. In: Driese, S.G. Ed. , Pale- 400 million year carbon isotope record of pedogenic carbon-

osols, Paleoclimate and Paleoatmospheric CO2 ; Paleozoic Pa- ate: implications for paleoatmospheric carbon dioxide. Ameri- leosols of Central Pennsylvania. University of Tennessee, can Journal of Science 299, 805–827. Department of Geological Sciences. Studies in Geology, vol. Eldridge, D.J., Greene, R.S.B., 1994. Microbiotic soil crusts: a 22, pp. 50–66. review of their roles in soil and ecological processes in the Duckett, J.G., Read, D.J., 1991. The use of the fluorescent dye, rangelands of Australia. Australian Journal of Soil Research X 3,3 dihexyloxacarbocyanine iodide, for selective staining of 32, 389–415. ascomycete fungi associated with liverwort rhizoids and eri- Elick, J.M., Driese, S.G., Mora, C.I., 1998. Very large plant and coid mycorrhizal roots. New Phytologist 118, 259–272. root traces from the Early to Middle Devonian: implications

Duckett, J.G., Read, D.J., 1995. Ericoid mycorrhizas and for early terrestrial ecosystems and atmospheric pŽ.CO2 . Ge- rhizoid–ascomycete associations in liverworts share the same ology 26Ž. 2 , 143–146. mycobiont: isolation of the partners and resynthesis of the Ellwood, D.C., Hedger, J.N., Latham, M.J., Lynch, J.M., Slater, associations in vitro. New Phytologist 129, 439–447. J.H., 1980. Contemporary Microbial Ecology. Academic Press, Duckett, J.G., Renzaglia, K.S., Pell, K., 1991. A light and electron London, 438 pp. microscope study of rhizoid–ascomycete associations and Erbacher, J., Thurow, J., 1997. Influence of oceanic anoxic events flagelliform axes in British hepatics with observations on the on the evolution of mid-Cretaceous radiolaria in the North effects of the fungi on host morphology. New Phytologist 118, Atlantic and western Tethys. Marine Micropaleontology 30, 233–257. 139–158. Duff, R.B., Webley, D.M., Scott, R.O., 1963. Solubilization of Erbacher, J., Thurow, J., Littke, R., 1996. Evolution patterns of minerals and related materials by 2-ketogluconic acid-produc- radiolaria and organic matter variations: a new approach to ing bacteria. Soil Science 95, 105–114. identify sea-level changes in mid-Cretaceous pelagic environ- Easton, R.M., 1997. Lichen–rock–mineral interactions: an ments. Geology 24, 499–502. overview, 1997. In: McIntosh, J.M., Groat, L.A.Ž. Eds. , Bio- Erbacher, J., Huber, B.T., Norris, R.D., Markey, M., 2001. In- logical–Mineralogical Interactions. Mineralogical Association creased thermohaline stratification as a possible cause for an of Canada, Short Course, vol. 25, pp. 209–239. ocean anoxic event in the Cretaceous period. Nature 409, Eckhardt, F.E.W., 1985. Solubilization, transport, and deposition 325–327. of mineral cations by microorganisms—efficient rock weath- Erdman, J.A., Gough, L.P., White, R.W., 1977. Calcium oxalate ering agents. In: Drever, J.I.Ž. Ed. , The Chemistry of Weather- as source of high ash yields in the terricolous lichen Parmelia ing. Reidel, Dordrecht, pp. 161–173. chlorochroa. The Bryologist 80, 334–339. Edmond, J.M., 1992. Himalayan tectonics, weathering processes, Ferris, R., Taylor, G., 1994. Stomatal characteristics of four native

and the strontium isotope record in marine limestones. Science herbs following exposure to elevated CO2 . Annals of Botany 258, 1594–1597. 73, 447–453. Edmond, J.M., Huh, Y., 1997. Chemical weathering yields from Ferris, F.G., Wiese, R.G., Fyfe, W.S., 1994. Precipitation of basement and orogenic terrains in hot and cold climates. In: carbonate minerals by microorganisms: implications for sili- Ruddiman, W.F.Ž. Ed. , Tectonic Uplift and Climate Change. cate weathering and the global carbon dioxide budget. Geomi- Plenum, New York, pp. 329–351. crobiology Journal 12, 1–13. Edmond, J.M., Palmer, M.R., Measures, C.I., Grant, B., Stallard, Fiol, L., Fomos, J.J., Gines, A., 1996. Effects of biokarstic R.F., 1995. The fluvial geochemistry and denudation rate of processes on the development of solutional rillenkarren in 144 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

limestone rocks. Earth Surface Processes and Landforms 21, topes by phytoplankton and estimates of ancient CO2 levels. 447–452. Global Biogeochemical Cycles 6, 185–198. Fitter, A.H., Merryweather, J.W., 1992. Why are some plants Friedmann, E.I., 1971. Light and scanning electron microscopy of more mycorrhizal than others?, An ecological inquiry. In: the endolithic desert algal habitat. Phycologia 10Ž. 4 , 411–428. Read, D.J., Lewis, D.H., Fitter, A.H., Alexander, I.J.Ž. Eds. , Friedmann, E.I., 1980. Endolithic microbial life in hot and cold Mycorrhizas in Ecosystems. Cambridge Univ. Press, Cam- deserts. Origins of Life 10, 223–235. bridge, pp. 26–36. Friedmann, E.I., 1982. Endolithic microorganisms in the Antarctic

Flower, B.P., 1999. Warming without high CO2 . Nature 399, Cold Desert. Science 215, 1045–1053. 313–314. Friedmann, E.I., 1986. The Antarctic cold desert and the search Fluckiger, J., Dallenbach, A., Blunier, T., Stauffer, B., Stockton, for traces of life on Mars. Advances in Space Research 6, H., Raynaud, D., Barnola, J.-M., 1999. Variations in atmo- 265–268.

spheric N2 O concentration during abrupt climatic changes. Friedmann, E.I., Galun, M., 1974. Desert algae, lichens, and Science 285, 227–230. fungi. In: Brown Jr., G.W.Ž. Ed. , Desert Biology. Academic Folk, R.L., 1994a. Nannobacteria and chemistry on the face of the Press, New York, pp. 165–212. Earth. Texas Bureau of Economic Geology Seminar Series, Friedmann, E.I., Ocampo, R., 1976. Endolithic blue-green algae in Wednesday, March 16, Austin. the dry valleys: primary producers in the Antarctic desert Folk, R.L., 1994b. Dwarf bacteriaŽ. Nannobacteria and the pre- ecosystem. Science 193, 1247–1249. cipitation of silica and clay minerals. Abstract, 14th Interna- Friedmann, E.I., Weed, R., 1987. Microbial trace-fossil formation, tional Sedimentological Congress, Recife, Brasil, VIII. biogenous, and abiotic weathering in the Antarctic cold desert. Folk, R.L., 1999. Nannobacteria and the precipitation of carbonate Science 236, 703–705. in unusual environments. Sedimentary Geology 126, 47–55. Friedmann, E.I., Lipkin, Y., Ocampo-Paus, R., 1967. Desert algae Folk, R.L., Roberts, H.H., Moore, C.H., 1973. Black phytokarst of the NegevŽ. Israel . Phycologia 6, 185–195. from Hell, Cayman Islands, B.W.I. Bulletin of the Geological Friedmann, E.I., LaRock, P.A., Brunson, J.O., 1980. Adenosine Society of America 84, 2351–2360. triphosphateŽ. ATP , chlorophyll, and organic nitrogen in en- Folk, R.L., Lynch, F.L., Rasbury, E.T., 1994. Evidence for bacte- dolithic microbial communities and adjacent soils in the dry rial precipitation of clay minerals upon sand grains, in soils, valleys of southern Victoria Land. Antarctic Journal of the and in the subsurface. Geological Society of America Annual United States 15Ž. 5 , 164–167. MeetingŽ. Seattle , Program with Abstract, p. A-508. Friend, P.F., 1966. Clay fractions and colours of some Devonian Follmi,¨ K.B., Weissert, H., Lini, A., 1993. Nonlinearities in red beds in the Catskill Mountains, USA. Quarterly Journal of phosphogenesis and phosphorus–carbon coupling and their the Geological Society of London 122, 273–292. implications for global change. NATO ASI Series, vol. 14, Fritz, E., Knoch, D., Meyer, D., 1994. A new approach for Springer-Verlag, Berlin, pp. 447–474. rhizosphere research by X-ray microanalysis of micro-liter soil Fox, T.R., Comerford, N.B., 1992. Influence of oxalate loading on solutions. Plant Soil 161, 219–223. phosphorus and aluminum solubility in spodosols. Soil Sci- Fry, E.J., 1927. The mechanical action of crustaceous lichens on ence Society of America Journal 56, 290–294. substrates of shale, schist, gneiss, limestone, and obsidian. Frakes, L.A., Francis, J.E., Syktus, J.I., 1992. Climate modes of Annals of Botany 41, 437–460. the Phanerozoic. Cambridge Univ. Press, Cambridge, 274 pp. Frydman, I., 1957. Myktrofizm roslinnosci pokrywajacej gruzy; France-Lanord, C., Derry, L.A., 1997. Organic carbon burial ruiny domow Wroclawia.w Mycotrophy of plants growing on forcing of the carbon cycle from Himalayan erosion. Nature ruins in Wroclawx Soc. Bot. Pol. Acta 26, 45–60. 390, 65–67. Funk, D.W., Pullman, E.R., Peterson, K.M., Crill, P.M., Billings, Francis, A.J., Slater, J.M., Dodge, C.J., 1989. Denitrification in W.D., 1994. Influence of water table on carbon dioxide, deep subsurface sediments. Geomicrobiology Journal 7, 103– carbon monoxide, and methane fluxes from taiga bog micro- 116. cosms. Global Biogeochemical Cycles 8Ž. 3 , 271–278. Francois, L.M., Walker, J.C.G., 1992. Modelling the Phanerozoic Gaffin, S., 1987. Ridge volume dependence on seafloor generation carbon cycle and climate: constraints from the 87Srr86 Sr rate and inversion using long term sealevel change. American isotope ratio of seawater. American Journal of Science 292, Journal of Science 287, 596–611. 81–135. Gaillardet, J., Allegre, C., Dupre, B., 2000. What the geochem- Francois, L.M., Walker, J.C.G., Opdyke, B.N., 1993. The history istry of large rivers tells us about chemical weathering rates

of global weathering and the chemical evolution of the and long-term CO2 consumption rates. Abstract, 31st Interna- ocean–atmosphere system. In: Takahashi, S.E., Jeanloz, R., tional Geological Congress, Rio de Janeiro, unpaginated on Rubie, D.Ž. Eds. , Evolution of the Earth and Planets. Geophys- disc. ical Monograph, vol. 74, pp. 143–159, IUGG Vol. 14. Gao, G., 1993. The temperature and oxygen-isotope composition Fredrickson, J.K., Garland, T.R., Hicks, R.J., Thomas, J.M., Li, of early Devonian oceans. Nature 361, 712–714. S.W., McFadden, K.M., 1989. Lithotrophic and heterotrophic Garbaye, J., 1994a. Tansley Review No. 76 Helper bacteria—a bacteria in deep subsurface sediments and their relation to new dimension to the mycorrhizal symbiosis. New Phytologist sediment properties. Geomicrobiology Journal 7, 53–66. 128, 197–210. Freeman, K.H., Hayes, J.M., 1992. Fractionation of carbon iso- Garbaye, J., 1994b. Mycorrhization helper bacteria—a new di- A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 145

mension to the mycorrhizal symbiosis. Acta Botanica Gallica in color change and biodeterioration of antique marbles. Ge- 141, 517–521. omicrobiology Journal 11, 205–221. Garg, K.L., Garg, N., Mukerji, K.G.Ž. Eds. , 1993. Recent Ad- Gorham, E., 1995. The biogeochemistry of northern peatlands and vances in Biodeterioration and Biodegradation, vol. 1, Naya its possible responses to global warming. In: Woodwell, G.M., Prokash, Calcutta, 496 pp. MacKenzie, F.T.Ž. Eds. , Biotic Feedbacks in the Global Cli- Gates, D.M., Strain, B.R., Weber, J.A., 1983. 15 Ecophysiological matic System: Will the Warming Feed the Warming? Oxford

effects of changing atmospheric CO2 concentration. In: Lange, Univ. Press, Oxford, pp. 169–187. O.L., Nobel, P.S., Ziegler, O.H.Ž. Eds. , Physiological Plant Gorham, E., Vitousek, P.M., Reiners, W.A., 1979. The regulation Ecology IV. Springer-Verlag, pp. 503–526. of chemical budgets over the course of terrestrial ecosystem Gehrmann, D.K., Krumbein, E., 1994. Interactions between succession. Annual Review of Ecology and Systematics 10, epilithic and endolithic lichens and carbonate rocks. In: 53–84. Fassina, V., Ott, H., Zezza, F.Ž. Eds. , The Conservation of Graham, L.E., Gray, J., 2001. The origin, morphology and eco- Monuments in the Mediterranean Basin. Museum d’Histoire physiology of early embryophytes: neontological and paleon- naturelle and Musee´` d’art et d’histoire, Ville de Geneve, pp. tological perspectives. In: Gensel, P., Edwards, D.Ž. Eds. , 311–316. Plants invade the land: Evolutionary and Environmental Per- Gehrmann, C., Krumbein, W.E., Petersen, K., 1988. Lichen spectives. Columbia Univ. Press, New York, pp. 140–158. weathering activities on mineral and rock surfaces. Studia Graustein, W.C., Cromack Jr., K., Sollins, P., 1977. Calcium Geobotanica 8, 33–45. oxalate: occurrence in soils and effect on nutrient and geo- Ghosh, P., Bhattacharya, S.K., Jani, R.A., 1995. Palaeoclimate chemical cycles. Science 198, 1252–1254. and palaeovegetation in central India during the Upper Creta- Gray, J., 1985. The microfossil record of early land plants: ceous based on stable isotope composition of the palaeosol advances in understanding of early terrestrialization. Philo- carbonates. Palaeogeography, Palaeoclimatology, Palaeoecol- sophical Transactions of the Royal Society of London, Series ogy 114, 285–296. B 309, 167–195. Gibbs, R.J., 1967. Geochemistry of the Amazon River system. Gray, J., 1993. Major Paleozoic land plant evolutionary bio-events. Geological Society of America Bulletin 78, 1203–1232. Palaeogeography, Palaeoclimatology, Palaeoecology 104, Gibbs, M.T., Barron, E.J., Kump, L.R., 1997. An atmospheric 153–169.

pCO2 threshold for glaciation in the Late Ordovician. Geology Griffiths, R.P., Baham, J.E., Caldwell, B.C., 1994. Soil solution 25, 447–450. chemistry of ectomycorrhizal mats in forest soil. Soil Biology Giles, M.R., de Boer, R.B., Marshall, J.D., 1994. How important and Biochemistry 26, 331–337. are organic acids in generating secondary porosity in the Gutzmer, J., Beukes, N.J., 1998. Earliest laterites and possible subsurface? In: Pittman, E.D., Lewan, M.D.Ž. Eds. , Organic evidence for terrestrial vegetation in the Early Paleozoic. Acids in Geological Process. Springer-Verlag, Berlin, pp. Geology 26, 263–266. 449–470. Gwiazda, R.H., Broecker, W.S., 1994. The separate and combined

Gold, T., 1992. The deep, hot biosphere. Proceedings of the effects of temperature, soil pCO2 , and organic acidity on National Academy of Science of the United States of America silicate weathering in the soil environment: formulation of a 89, 6045–6049. model and results. Global Biogeochemical Cycles 8Ž. 2 , 141– Golubic, S., 1973. The relationship between blue-green algae and 155. carbonate deposits. In: Carr, N., Whitton, B.A.Ž. Eds. , The Hall, K., Otte, W., 1990. A note on biological weathering on Biology of Blue-green Algae. Blackwell Science Publ., Lon- nunataks of the Juneau Icefield, Alaska. Permafrost Periglacial don, pp. 434–472. Process 1, 180–196. Golubic, S., Friedmann, I., Schneider, J., 1981. The lithobiontic Hallbauer, D.K., Jahns, H.M., 1977. Attack of lichens on quartzitic ecological niche, with special reference to microorganisms. rock surfaces. Lichenologist 9, 119–122. Journal of Sedimentary Petrology 51, 475–478. Hambrey, M.J., Harland, W.B.Ž. Eds. , 1981. Earth’s Pre-Pleisto- Gomez-Alarcon, G., Munoz, M., Arino, X., Ortega-Calvo, J.J., cene Glacial Record. Cambridge Univ. Press, Cambridge, 1995a. Microbial communities in weathered sandstones: the 1004 pp. case of Carrascosa del Campo church, Spain. The Science of Hamoumi, N., 2001. The Gondwanan platform during Ordovician the Total Environment 167, 249–254. times: Climatic eustatic and geodynamic evolution. Field trip Gomez-Alarcon, G., Cilleros, B., Flores, M., Lorenzo, J., 1995b. guide book, Ministere`´ de l’Enseignement Superieur de la Microbial communities and alteration processes in monuments Formation des Cadres et de la Recherche Scientifique sur les at Alcala de Henares, Spain. The Science of the Total Envi- fonds geres para le CNCPRST, 120 pp. ronment 167, 231–239. Harley, J.L., 1968. Presidental address, Fungal symbiosis. Trans- Gorbushina, A.A., Krumbein, W.E., 2000. Subaerial microbial actions of the British Mycological Society 51, 1–11. mats and their effects on soil and rock. In: Riding, R.E., Harley, J.L., Smith, S.E., 1983. Mycorrhizal Symbiosis. Academic Awramik, S.M.Ž. Eds. , Microbial Sediments. Springer-Verlag, Press, London, 483 pp. pp. 161–170. Hart, M.B., 1999. The evolution and biodiversity of Cretaceous Gorbushina, A.A., Krumbein, W.E., Hamman, C.H., Panina, L., planktonic Foraminiferida. Geobios 32, 247–255. Soukharjevski, S., Wollenzien, U., 1993. Role of black fungi Hasegawa, M., Hashimoto, T., Adachi, J., Iwabe, N., Miyata, T., 146 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

1993. Early branchings in the evolution of eucaryotes—an- Hochuli, P.A., Menegatti, A.P., Weissert, H., Riva, A., Erba, E., cient divergence of Entamoeba that lacks mitochondria re- Premoli Silva, I., 1999. Episodes of high productivity and vealed by protein sequence data. Journal of Molecular Evolu- cooling in the early Aptian of the Alpine Tethys. Geology 27, tion 36, 380–388. 657–660. Hass, H., Taylor, T.N., Remy, W., 1994. Fungi from the Lower Hofmann, P., Ricken, W., Schwark, L., Leythaeuser, D., 1999. Devonian Rhynie chert: mycoparasitism. American Journal of Coupled oceanic affects of climatic cycles from late Albian Botany 81, 29–37. deep-sea sections of the North Atlantic. Geological Society of Hattori, T., 1973. Microbial Life in the Soil: An Introduction. America, Special Paper 332, 143–159. Marcel Dekker, New York, 427 pp. Hofmann, P., Ricken, W., Schwark, L., Leythaeuser, D., 2000. Hayden, B.P., 1998. Ecosystem feedbacks on climate at the Carbon–sulfur–iron relationships and delta 13 C of organic landscape scale. Philosophical Transactions of the Royal Soci- matter for late Albian sedimentary rocks from the North ety of London, Series B 353, 5–18. Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology Hazen, T.C., Jimenez, L., de Victoria, G.L., 1991. Comparison of 163, 97–113. bacteria from deep subsurface sediment and adjacent ground- Holland, H.D., 1978. The Chemistry of the Atmosphere and the water. Microbiology and Ecology 22, 293–304. Oceans. Wiley, New York, 351 pp. Hedges, J.I., Keil, R.G., 1995. Sedimentary organic matter preser- Holland, H.D., 1984. The Chemical Evolution of the Atmosphere vation: an assessment and speculative synthesis. Marine and Oceans. Princeton Univ. Press, Princeton, NJ, 582 pp. Chemistry 49, 81–115. Holland, H.D., Lazar, B., McCaffrey, M., 1986. Evolution of the Hedges, R.W., Messens, E., 1990. Genetic aspects of rhizosphere atmosphere and oceans. Nature 320, 27–33. interactions. In: Lynch, J.M.Ž. Ed. , The Rhizosphere. Wiley- Holm, P.E., Nielsen, P.H., Albrechtsen, H.-J., Christensen, T.A., Interscience, Chichester, pp. 129–176. 1992. Importance of unattached bacteria and bacteria attached Heinen, W., 1967. Iron accumulation in bacterial systems: I. to sediment in determining potentials for the degradation of Isolation of two particulate fractions participating in silicon xenobiotic organic contaminants in an aerobic aquifer. Applied metabolism, from Proteus mirabilis cell-free extracts. Environmental Microbiology 58, 3020–3026. Archives of Biochemistry and Biophysics 120, 86–92. Huang, W.H., Kiang, W.C., 1972. Laboratory dissolution of pla- Heinen, W., 1968. The distribution and some properties of accu- gioclase feldspars in water and organic acids at room tempera- mulated silicate in cell-free bacterial extracts. Acta Bot. Neerl. ture. American Mineralogist 57, 1849–1859. 17, 105–113. Huang, P.M., Schnitzer, M.Ž. Eds. , 1986. Interactions of soil Henderson, M.E.K., Duff, R.B., 1963. The release of metallic and minerals with natural organics and microbes. Soil Science silicate ions from minerals, rocks, and soils by fungal activity. Society of America Special Publication, vol. 17, 606 pp. Journal of Soil Science 14, 236–246. Huang, P.M., Violante, A., 1986. Influence of organic acids on Herrera, R., Jordan, C.F., Klinge, H., Medina, E., 1978. Amazon crystallization and surface properties of precipitation products ecosystems. Their structure and functioning with particular of aluminum. In: Huang, P.M., Schnitzer, L.Ž. Eds. , Interac- emphasis on nutrients. Interciencia 3, 223–231. tions of Soil Minerals with Natural Organics and Microbes. Hesselbo, S.P., Grocke, D.R., Jenkyns, H.C., Bjerrum, C.J., Farri- Soil Science Society of Americam, Special Publication, vol. mond, P., Bell, H.S.M., Green, O.R., 2000. Massive dissocia- 17, pp. 159–221. tion of gas hydrate during a Jurassic oceanic anoxic event. Hunt, J.M., 1972. Distribution of carbon in crust of Earth. Bulletin Nature 406, 392–395. of the American Association of Petroleum Geologists 56, Hicks, R.J., Fredrickson, J.K., 1989. Aerobic metabolic potential 2273–2277. of microbial populations indigenous to deep subsurface envi- Hunt, J.M., 1979. Petroleum Geochemistry and Geology. Free- ronments. Geomicrobiology Journal 7, 67–77. man, San Francisco, 617 pp. Hiebert, F.K., Bennett, P.C., 1992. Microbial control of silicate Hunt, C.B., Durrell, L.W., 1966. Distribution of fungi and algae. weathering in organic-rich ground water. Science 258, 278– United States Geological Survey Professional Paper 509, pp. 281. 55–66. Hirsch, P., Eckhardt, F.E.W., Palmer Jr., R.J., 1995a. Fungi active Imbus, S.W., Engel, M.H., Elmore, R.D., 1990. Organic geochem- in weathering of rock and stone monuments. Canadian Journal istry and sedimentology of Middle Proterozoic Nonesuch For- of Botany 73Ž. Suppl. 1 , S1384–S1390. mation—hydrocarbon source rock assessment of lacustrine rift Hirsch, P., Eckhardt, F.E.W., Palmer Jr., R.J., 1995b. Methods for deposits. In: Katz, B.J.Ž. Ed. , Lacustrine Basin Exploration: the study of rock-inhabiting microorganisms—a mini review. Case Studies and Modern Analogs. American Association of Journal of Microbiologial Methods 23, 143–167. Petroleum Geologists Memoir, vol. 50, pp. 197–208. Hochella Jr., M.F., Banfield, J.F., 1995. Chemical weathering of Indermuhle, A., Stocker, T.F., Joos, F., Fischer, H., Smith, H.J., silicates in nature: A microscopic perspective with theoretical Wahlen, M., Deck, B., Mastroiani, D., Tschumi, J., Blunier, considerations. In: White, A.F., Brantley, S.L.Ž. Eds. , Chemi- T., Meyer, R., Stauffer, B., 1999. Holocene carbon-cycle

cal Weathering Rates of Silicate Minerals. Reviews in Miner- dynamics based on CO2 trapped in ice at Taylor Dome, alogy, vol. 31, Mineralogical Society of America, Washington, Antarctica. Nature 398, 121–126. DC, pp. 353–406. Isaacson, P.E., Hladil, J., Shen, J.-W., Kalvoda, J., Grader, G., A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 147

1999. Late DevonianŽ. Famennian glaciation in South America R., Melkerud, P.-A., Olsson, M., 2000. Rock-eating fungi. and marine offlap on other continents. Abhandlungen der Nature 389, 682–683. Geologischen Bundesanstalt 54, 239–257. Jordan, C.F., 1982. The nutrient balance of an Amazonian rain Jacks, G.V., 1965. The role of organisms in the early stages of forest. Ecology 63, 647–654. soil formation. Proceedings of the Easter School in Agricul- Jordan, C.F., Herrera, R., 1981. Tropical rain forest: are nutrients tural Science. Butterworth, London, pp. 219–226. really critical? American Naturalist 117, 167–180. Jackson, T.A., 1971. A study of the ecology of pioneer lichens, Jouzel, J., Barkov, N.I., Barnola, J.M., Bender, M., Chappellaz, J., mosses, and algae on Recent Hawaiian lava flows. Pacific Genthon, C., Kotlyakov, V.M., Lyenkov, V., Lorius, C., Petit, Science XXV, 22–32. J.R., Raynaud, D., Raisbeck, G., Ritz, C., Sowers, T., Stieve- Jackson, T.A., 1993. Comment on AWeathering, plants, and the nard, M., 1993. Extending the Vostok ice-core record of long-term carbon cycleB by Robert A. Berner. Geochimica et palaeoclimate to the penultimate glacial period. Nature 364, Cosmochimica Acta 57, 2141–2144. 407–412. Jackson, T.A., 1996. Comment on AThe effect of land plants on Kallio, P., Karenlampi, L., 1975. Photosynthesis in mosses and weathering rates of silicate mineralsB by J.I. Drever. Geochim- lichens. In: Cooper, J.P.Ž. Ed. , Photosynthesis and Productivity ica et Cosmochimica Acta 60, 723–724. in Different EnvironmentsŽ International Biologicial Pro- Jackson, T.A., Keller, W.D., 1970. A comparative study of the gramme 3. . Cambridge Univ. Press, Cambridge, pp. 393–423. role of lichens and ‘inorganic’ processes in the chemical Kappen, L., Lange, O.L., Schulze, E.-D., Evenari, M., Buschbom, weathering of Recent Hawaiian lava flows. American Journal U., 1975. Primary production of lower plantsŽ. lichens in the of Science 269, 446–466. desert and its physiological basis. In: Cooper, J.P.Ž. Ed. , Jackson, M.J., Muir, M.D., Plumb, K.A., 1987. Geology of the Photosynthesis and Productivity in Different Environments southern McArthur Basin, Northern Territory. Bulletin 220, Ž.International Biological Programme 3 . Cambridge Univ. Bureau of Mineral Resources, Geology and Geophysics, De- Press, Cambridge, pp. 133–143. partment of Resources and Energy. Australian Government Kawano, M., Tomita, K., 2001. Microbial biomineralization in Publishing Service, Canberra 173 pp. weathered volcanic ash deposit and formation of biogenic Jain, K.K., Mishra, A.K., Singh, T., 1993. Biodeterioration of minerals by experimental incubation. American Mineralogist stone: a review of mechanisms involved. In: Garg, K.L., Garg, 86, 400–410. N., Mukerji, K.G.Ž. Eds. , 1993. Recent Advances in Biodeteri- Keller, C.K., Wood, B.D., 1993. Possibility of chemical weather- oration and Biodegradation, vol. 1, Naya Prokash, Calcutta, ing before the advent of vascular land plants. Nature 364, pp. 323–354. 223–225. Janos, D.P., 1987. VA mycorrhizas in humid tropical ecosystems. Kennett, J.P., Cannariato, K.G., Hendy, I.L., Behl, R.J., 2000. In: Safir, G.R.Ž. Ed. , Ecophysiology of VA Mycorrhizal Plants. Carbon isotopic evidence for methane hydrate instability dur- CRC Press, Boca Raton, FL, pp. 107–134. ing Quaternary interstadials. Science 288, 128–133. Jenkyns, H.C., 1999. Mesozoic anoxic events and palaeoclimate. Kenrick, P., 2001. Turning over a new leaf. Nature 410, 309–310. Zentralblatt fur¨¨ Geologie und Palaontologie, Teil I, H 7–9, Kerr, A.C., 1998. Oceanic plateau formation: a cause of mass 943–949. extinction and black shale deposition around the Jennings, D.H., 1995. The Physiology of Fungal Nutrition. Cam- Cenomanian–Turonian boundary? Journal of the Geological bridge Univ. Press, Cambridge, 622 pp. Society of London 158, 619–626. Johansen, J.R., 1993. Cryptogamic crusts of semiarid and arid Kerr, R.A., 1999. Slide into ice ages not carbon dioxide’s fault? lands of North America. Journal of Phycology 29, 140–147. Science 284, 1743–1746. Jones, D., 1988. Lichens and pedogenesis. In: Galun, M.Ž. Ed. , Kerr, S., Zavada, M.S., 1989. The effect of the lichen Acarospora 1988. CRC Handbook of Lichenology, vol. 3. CRC Press, sinopica on the elemental composition of three sedimentary Boca Raton, FL, pp. 109–124. rock substrates in South Africa. Bryologist 92, 407–410. Jones, D., Wilson, M.J., 1985. Chemical activity of lichens on Khain, V.E., Ronov, A.B., Balukhovski, A.N., 1975. Melovye mineral surfaces—a review. International Biodeterioration litologicheskie formatsii mira. Sovetskaya Geologiya 11, 10– Bulletin 21Ž. 2 , 99–104. 39. Jones, D., Wilson, M.J., Tait, J.M., 1980. Weathering of a basalt Khain, V.E., Ronov, A.B., Seslavinski, K.B., 1977. Siluriskie by Pertusaria corallina. Lichenologist 12Ž. 3 , 277–289. litologicheskie formatsii mira. Sovetskaya Geologiya 5, 21–43. Jones, D., Wilson, W.J., McHardy, W.J., 1981. Lichen weathering Khan, A.G., Belik, M., 1995. Occurrence and ecological signifi- of rock-forming minerals: application of scanning electron cance of mycorrhizal symbiosis in aquatic plants. In: Varma, microscopy and microprobe analysis. Journal of Microscopy A., Hock, B.Ž. Eds. , Mycorrhiza. Springer-Verlag, Berlin, pp. 124, 95–104. 627–666. Jones, R.E., Beeman, R.E., Suflita, J.M., 1989. Anaerobic Kidron, G.J., Yaalon, D.H., Vonshak, A., 1999. Two different metabolic processes in the deep terrestrial subsurface. Geomi- causes for runoff initiation on microbiotic crusts: hydrophobic- crobiology Journal 7, 117–130. ity and pore clogging. Soil Science 164, 18–27. Jongmans, A.G., van Breemen, N., Lundstrom, U., van Hees, Kidston, R., Lang, W.H., 1921. On Old Red Sandstone plants P.A.W., Finlay, R.D., Srinivasan, M., Unestam, T., Giesler, showing structure, from the Rhynie Chert bed, Aberdeenshire. 148 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Part V. Transactions of the Royal Society of Edinburgh 52, Krumbein, W.E.Ž. Ed. , Microbial Geochemistry. Blackwell, 855–902. Oxford, pp. 132–143. Kitayama, K., Mueller-Dombois, D., 1995. Vegetation changes Krumbein, W.E., Grote, G., Petersen, K., 1987. Metal biotransfer during long-term soil development in the Hawaiian montaine and crust formation in building stones. In: Morton, L.H.G. rainforest zone. Vegetatio 120, 1–20. Ž.Ed. , Biodeterioration of Constructional Materials. Biodeteri- Kitayama, K., Mueller-Dombois, D., Vitousek, P.M., 1995. Pri- oration Society, Lancashire Polytechnic, UK, p. 15. mary succession of Hawaiian montane rain forest on a Kump, L.R., Gibbs, M.T., Arthur, M.A., Patzkowsky, M.E., Shee- chronosequence of eight lava flows. Journal of Vegetation han, P.M., 1995. Hirnantian glaciation and the carbon cycle. Science 6, 211–222. In: Cooper, J.D., Droser, M.L., Finney, S.C.Ž. Eds. , Ordovian Klappa, C.F., 1979. Lichen stromatolites: criterion for subaerial Odyssey. Pacific Section, SEPM, Fullerton, CA, pp. 299–302. exposure and a mechanism for the formation of laminar Kump, L.R., Arthur, M.A., Patzkowsky, M.E., Gibbs, M.T., calcretesŽ. caliche . Journal of Sedimentary Petrology 49, 387– Pinkus, D.S., Sheehan, P.M., 1999. A weathering hypothesis

400. for glaciation at high atmospheric pCO2 during the Late Knoll, A.H., James, W.C., 1987. Effect of the advent and diversi- Ordovician. Palaeogeography, Palaeoclimatology, Palaeoecol- fication of vascular land plants on mineral weathering through ogy 152, 173–187. geologic time. Geology 15, 1099–1102. Kurschner, W.M., van der Burgh, J., Visscher, H., Dilcher, D.L., Koppes, S., 1990. Learning to Like Lichens. ASU Research 1996. Oak leaves as biosensors of late Neogene and early

Ž.Arizona State University , Fall issue, pp. 13–15. Pleistocene paleoatmospheric CO2 concentrations. Marine Mi- Kozur, H.W., 1998. Some aspects of the Permian–Triassic bound- cropaleontology 27, 299–312. aryŽ. PTB and the possible causes for the biotic crisis around Kurschner, W.M., Wagner, F., Dilcher, D.L., Visscher, H., 1997. this boundary. Palaeogeography, Palaeoclimatology, Palaeoe- Photomorphogenesis in deciduous and evergreen angiosperm

cology 143, 227–272. leaves and its consequences for palaeoatmospheric CO2 recon- Krassilov, V.A., 1973. Climatic changes in eastern Asia as indi- structions. Climates, Past, Present and Future, 2nd European cated by fossil floras: I. Early Cretaceous. Palaeogeography, Palaeontological Congress, July 10–12, Abstracts, pp. 25–26. Palaeoclimatology, Palaeoecology 13, 261–273. Kutuzova, R.S., 1969. Release of silica from minerals as a result Krassilov, V.A., 1975. Climatic changes in eastern Asia as indi- of microbial activity. Mikrobiologiya 38, 714–721. cated by fossil floras. II. Late Cretaceous and Danian. Palaeo- Kuypers, M.M.M., Blokker, P., Erbacher, J., Kinkel, H., Pancost, geography, Palaeoclimatology, Palaeoecology 17, 157–172. R.D., Schouten, S., Damste, J.S.S., 2001. Massive expansion

Krassilov, V.A., 1994. Rol kolebanii CO2 v evolutsii klimata I of marine Archaea during a Mid-Cretaceous oceanic anoxic bioti. Bool. Mosk. Obshestva Ispitatelei Prirodi otdel. Ge- event. Science 293, 92–94. ologiya T 69Ž. Vip 6 , 75–84. Kuypers, M.M.M., Pancost, R.D., Damste, J.S.S., 1999. A large

Krassilov, V.A., 1999. Global climatic changes and redistribution and abrupt fall in atmospheric CO2 concentration during Cre- of biomass. Doklady Russian Academy of Science 369Ž. 2 , taceous times. Nature 399, 342–345. 234–237. Lange, O.L., West, N.E., 1990. Structure and function of micro- Krinsley, D., 1998. Models of rock varnish formation constrained phytic soil crusts in wildland ecosystems of arid to semi-arid by high resolution transmission electron microscopy. Sedimen- regions. Advances in Ecological Research 20, 179–223. tology 45Ž. 4 , 711–725. Lange, O.L., Kidron, G.J., Budel, B., Meyer, A., Kilian, E., Krumbein, W.E., 1968. Geomicrobiology and geochemistry of the Abeliovich, A., 1992. Taxonomic composition and photosyn- ANari-limecrustsB Ž.Israel . In: Friedman, G.M., Muller, G. thetic characteristics of the ‘biological soil crusts’ covering Ž.Eds. , Recent developments in carbonate sedimentation in sand dunes in the western Negev Desert. Functional Ecology central Europe. Springer-Verlag, pp. 138–147. 6, 519–527. Krumbein, W.E., 1979. Photolithotrophic and chemooorgan- Lange, O.L., Meyer, A., Zellner, H., Heber, U., 1994. Photosyn- otrophic activity of bacteria and algae as related to beachrock thesis and water relations of lichen soil crusts: field measure- formation and degradationŽ. Gulf of Aquaba, Sinai . Geomicro- ments in the coastal fog zone of the Namib Desert. Functional biology Journal 1, 139–203. Ecology 8, 253–264. Krumbein, W.E.Ž. Ed. , 1983. Microbial Geochemistry. Blackwell, Lange, O.L., Belnap, J., Reichenberger, H., 1998. Photosynthesis Oxford, 330 pp. of the cyanobacterial soil crust lichen Collema tenax from arid Krumbein, W.E., Jens, K., 1981. Biogenic rock varnishes of the lands in southern Utah, USA: role of water content on light

Negev DesertŽ. Israel : an ecological study of iron and man- and temperature responses of CO2 exchange. Functional Ecol- ganese transformation by cyanobacteria and fungi. Oecologia ogy 12, 195–202. 50, 25–38. Lee, C., 1992. Controls on organic carbon preservation: the use of Krumbein, W.E., Urzi, C., 1991. Biologically induced phenomena stratified water bodies to compare intrinsic rates of decomposi- of antique marbles—some general considerations. In: De- tion in oxic and anoxic systems. Geochimica et Cosmochimica crouez, D., Chamay, J., Zezza, F.Ž. Eds. , The Conservation of Acta 56, 3323–3335. monuments in the Mediterranean basin. Museum d’Histoire Lee, M.R., 1999. Organic–mineral interactions studied by con- naturelle and Musee´` d’art et d’histoire, Ville de Geneve, pp. trolled pressure SEM. The Americas Microscopy and Analy- 219–235. sis, March, pp. 19–21. Krumbein, W.E., Werner, D., 1983. The microbial silica cycle. In: LePage, B.A., Currah, R.S., Stockey, R.A., Rothwell, G.W., 1994. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 149

Fossil ectomycorrhiza in Eocene Pinus roots. The fossil fungi Stomatal density and aperture length in four plant species

of the Princeton chert. International Journal of Plant Sciences grown across a subambient CO2 gradient. American Journal 155, 828–836. of Botany 80, 1413–1418. LePage, B.A., Currah, R.S., Stockey, R.A., Rothwell, G.W., 1997. Marathe, K.V., Chaudhari, P.R., 1975. An example of algae as Fossil Ectomycorrhizae from the Middle Eocene. American pioneers in the lithosere and their role in rock corrosion. Journal of Botany 84, 410–412. Journal of Ecology 63, 65–69. Lewan, M.D., Pittman, E.D., 1994. Introduction to the role of Marschner, H., Romheld, V., 1996. Root-induced changes in the organic acids in geological processes. In: Pittman, E.D., Lewan, availability of micronutrients in the rhizosphere. In: Waisel, M.D.Ž. Eds. , Organic Acids in Geological Processes. Y., Eshel, A., Kafkafi, U.Ž. Eds. , Plant Roots, the Hidden Springer-Verlag, Berlin, pp. 1–21. Half, 2nd edn. Marcel Dekker, New York, pp. 557–579. Lewin, S.Z., Charola, A.E., 1981. Plant life on stone surfaces and Martinez, P., Castro, J., Sanchez, A., 1993. Biodeterioration of its relation to stone conservation. Scanning Electron Mi- cultural property in Cuba. In: Garg, K.L., Garg, N., Mukerji, croscopy 1981, 563–568. K.G.Ž. Eds. , 1993. Recent Advances in Biodeterioration and Leyval, C., Berthelin, J., 1991. Weathering of a mica by roots and Biodegradation, vol. 1, Naya Prokash, Calcutta, pp. 379–397. rhizospheric microorganisms of pine. Soil Science Society of Massa, D., Havlicek, V., Bonnefous, J., 1977. Stratigraphic and America Journal 55, 1009–1016. faunal data on the Ordovician of the Rhadames BasinŽ Libya Ligrone, R., 1988. Ultrastructure of a fungal endophyte in Phaeo- and Tunisia. . Bull. Centres Rech. Explor-Prod. Elf-Aquitaine ceras laeÕis Ž.L. Prosk, Ž Anthocerophyte . . Botanical Gazette 1, 3–27. 149, 92–100. McCarroll, D., Viles, H., 1995. Rock-weathering by the lichen Ligrone, R., Lopes, C., 1989. Cytology and development of Lecidea auriculata in an arctic alpine environment. Earth mycorrhiza-like infection in the gametophyte of Cono- Surface Processes and Landforms 20, 199–206. cephalum conicum Ž.L. Dom. Ž Marchantiales, Hepatophyta . . McCauley, S.E., DePaolo, D.J., 1997. The marine 87Srr86 Sr and New Phytologist 111, 423–433. delta 18 O records, Himalayan alkalinity fluxes and Cenozoic Livingston, D.A., 1963. Sodium cycle and the age of the ocean. climate models. In: Ruddiman, W.F.Ž. Ed. , Tectonic Uplift and Geochimica et Cosmochimica Acta 27, 1055–1069. Climate Change. Plenum, New York, pp. 427–467. Longhurst, A., Sathyendranath, S., Platt, T., Caverhill, C., 1995. McConnaughey, T.A., Whelan, J.F., 1997. Calcification generates An estimate of global primary production in the ocean from protons for nutrient and bicarbonate uptake. Earth-Science satellite radiometer data. Journal of Plankton Research 17, Reviews 42, 95–117. 1245–1271. McElwain, J.C., 1998. Do fossil plants signal palaeoatmospheric

Longman, K.A., Jenik, J., 1987. Tropical Forest and its Environ- CO2 concentration in the geological past? In: Zeki, S.Ž. Ed. , ment. 2nd edn. Longman, Harlow, Essex, 347 pp. Vegetation–Climate–Atmosphere Interactions: Past, Present Longton, R.E., 1988. Biology of Polar Bryophytes and Lichens. and Future. Philosophical Transactions of the Royal Society of Cambridge Univ. Press, Cambridge, 391 pp. London, Series B, vol. 353, pp. 83–96. Longton, R.E., 1992. The role of bryophytes and lichens in McElwain, J.C., Chaloner, W.G., 1995. Stomatal density and terrestrial ecosystems. In: Bates, J.W., Farmer, A.M.Ž. Eds. , index of fossil plants track atmospheric carbon dioxide in the Bryophytes and Lichens in a Changing Environment. Claren- Palaeozoic. Annals of Botany 76, 389–395. don Press, Oxford, pp. 32–76. McElwain, J.C., Chaloner, W.G., 1996. The fossil cuticle as a Lucas, Y., Luizao, F.J., Chauvel, A., Rouiller, J., Nahon, D., skeletal record of environmental change. Palaios 11, 376–388. 1993. The relation between biological activity of the rain McElwain, J.C., Beerling, D.J., Woodward, F.I., 1999. Fossil forest and mineral composition of soils. Science 260, 521–523. plants and global warming at the Triassic–Jurassic boundary. Lynch, J.M.Ž. Ed. , 1990. The Rhizosphere. Wiley, Chichester, 458 Science 285, 1386–1390. pp. McIntosh, J.M., Silver, M., Groat, L.A., 1997. Bacteria and the MacDonald, G.J., 1990. Global climate change. In: MacDonald, breakdown of sulfide minerals. In: McIntosh, J.M., Groat, G., Sertorio, L.Ž. Eds. , Global Climate and Ecosystems Change L.A.Ž. Eds. , Biological–Mineralogical Interactions. Mineralog- Ž.NATO ASI Series B: Physics Vol. 240 . Plenum, New York, ical Association of Canada, Short Course, vol. 25, pp. 63–92. pp. 1–95. McKay, C.P., Weed, R., Tyler, D.A., Vestal, J.R., Friedmann, Machel, H.G., Foght, J., 2000. Products and depth limits of E.I., 1983. Studies of cryptoendolithic microbial communities microbial activity in petroliferous subsurface settings. In: Rid- in the Antarctic cold desert. Antarctic Journal of the United ing, R.E., Awramik, S.M.Ž. Eds. , Microbial Sediments. States 18, 227–228. Springer-Verlag, Berlin, pp. 105–120. McKeague, J.A., Cheshire, M.V., Andreux, F., Berthelin, J., 1986. Madsen, E.L., Bollag, J.-M., 1989. Aerobic and anaerobic micro- Organo-mineral complexes in relation to pedogenesis. In: bial activity in deep subsurface sediments from the Savannah Huang, P.M., Schnitzer, M.Ž. Eds. , Interactions of Soil Miner- River Plant. Geomicrobiology Journal 7, 93–101. als with Natural Organics and Microbes. Soil Science Society Malloch, D.W., Pirozynski, K.A., Raven, P.H., 1980. Ecological of America, Special Publication, vol. 17, pp. 549–592. and evolutionary significance of mycorrhizal symbioses in McKinley, J.P., Stevens, T.O., Westall, F., 2000. Microfossils and vascular plantsŽ. a review . Proceedings of the National paleoenvironments in deep subsurface basalt samples. Geomi- Academy of Sciences 77Ž. 4 , 2113–2118. crobiology Journal 17, 43–54. Malone, S.R., Mayeux, H.S., Johnson, H.B., Polley, H.W., 1993. McKirdy, D.M., Hahn, J.H., 1982. The composition of kerogen 150 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

and hydrocarbons in Precambrian rocks. In: Holland, H.D., Moulton, K.L., West, J., Berner, R.A., 2000. Solute flux and Schidlowski, M.Ž. Eds. , Mineral Deposits and the Evolution of mineral mass balance approaches to the quantification of plant the BiosphereŽ. Dahlem Konferenzen 1982 . Springer-Verlag, effects on silicate weathering. American Journal of Science Berlin, pp. 123–154. 300, 539–570. McKirdy, D.M., Imbus, S.W., 1992. Precambrian petroleum: a Mueller, P.J., Suess, E., 1979. Productivity, sedimentation rate, decade of changing perception. In: Schidlowski, M., Golubic, and sedimentary organic matter in the oceans. I. Organic S., Kimberley, M.M., McKirdy, D.M., Trudinger, P.A.Ž. Eds. , carbon preservation. Deep-Sea Research, Part A, Oceano- Early Organic Evolution: Implications for Mineral and Energy graphic Research Paper, vol. 26, no. 12, Pergamon, pp. 1347– Resources. Springer-Verlag, Berlin, pp. 176–192. 1362. McKirdy, D.M., Kantsler, A.J., 1980. Oil geochemistry and poten- Nadelhoffer, K.J., Emmett, B.A., Gundersen, P., Kjonass, O.J., tial source rocks of the Officer Basin, South Australia. Aus- Koopmans, C.J., Schleppi, P., Tietema, A., Wright, R.F., tralian Petroleum Exploration Association Journal 20, 68–86. 1999. Nitrogen deposition makes a minor contribution to McKirdy, D.M., Kantsler, A.J., Emmett, J.K., Aldridge, A.K., carbon sequestration in temperate forests. Nature 398, 145– 1984. Hydrocarbon genesis and organic facies in Cambrian 148. carbonates of the eastern Officer Basin, South Australia. In: Nash III, T.H., White, S.L., Marsh, J.E., 1977. Lichen and moss Palacas, J.G.Ž. Ed. , Petroleum Geochemistry and Source Rocks distribution and biomass in hot desert ecosystems. The Bryolo- Potential of Carbonate Rocks. American Association of gist 80, 470–479. Petroleum Geologists Studies in Geology, vol. 18, pp. 13–32. Nesbitt, H.W., 1997. Bacterial and inorganic weathering processes McMahon, P.B., Chapelle, F.H., 1991. Microbial production of and weathering of crystalline rocks. In: McIntosh, J.M., Groat, organic acids in aquitard sediments and its role in aquifer L.A.Ž. Eds. , Biological–Mineralogical Interactions. Mineralog- geochemistry. Nature 349, 233–235. ical Association of Canada, Short Course, vol. 25, pp. 113– McMahon, P.B., Chapelle, F.H., Falls, W.F., Bradley, P.M., 1992. 142. Role of microbial processes in linking sandstone diagenesis Newman, E.I., Reddell, P., 1987. The distribution of mycorrhizas with organic-rich clays. Journal of Sedimentary Petrology 62, among families of vascular plants. New Phytologist 106, 1–10, Nature 349, 233–235. 745–751. Metting, B., 1981. The systematics and ecology of soil algae. Nicolson, T.H., 1967. Vesicular–arbuscular mycorrhiza—a uni- Botanical Review 47, 56–60. versal plant symbiosis. Science Progress, Oxford 55, 561–581. Metting, B., 1990. Soil algae. In: Lynch, J.M.Ž. Ed. , The Rhizo- Nicolson, T.H., 1975. Evolution of vesicular–arbuscular mycor- sphere. Wiley, Chichester, pp. 355–368. rhizas. In: Sanders, F.E., Mosse, B., Tinker, P.B.Ž. Eds. , Meybeck, M., 1979. Concentrations des eaux fluviales en ele- Endomycorrhizas. Academic Press, London, pp. 25–34. ments majeurs et apports en solution aux oceans. Revue de Nieboer, E., Richardson, D.H.S., Tomassini, F.D., 1978. Mineral Geologie Dynamique et de Geographie Physique 21, 215–246. uptake and release by lichens: an overview. The Bryologist 81, Meybeck, M., 1980. Pathways of major elements from land to 226–246. oceans through rivers. In: Martin, J.M., Burton, J.B., Eisma, Nobel, P.S., 1999. Physicochemical and Environmental Plant D.Ž. Eds. , River Inputs to Ocean Systems. UNEP, IOC, SCOR Physiology. 2nd edn. Academic Press, New York, 474 pp. Workshop Proc., United Nations Press, New York, pp. 18–30. Norris, R.D., Rohl, U., 1999. Carbon cycling and chronology of Mishler, B.D., Mishler, L.A., Buchheim, M.A., Renzaglia, K.S., climate warming during the PalaeocenerEocene transition. Garbary, D.J., Delwiche, C.F., Zechman, F.W., Kantz, T.S., Nature 401, 775–778. Chapman, R.L., 1994. Phylogenetic relationships of the Agreen Nugent, M.A., Brantley, S.L., Panano, C.G., Maurice, P.A., 1998. algaeB and AbryophytesB. Annals of the Missouri Botanical The influence of natural mineral coatings on feldspar weather- Garden 81, 451–483. ing. Nature 395, 588–591. Molina, R., Massicotte, Trappe, J.M., 1992. Ecological role of O’Brien, B.J., Stout, J.D., 1978. Movement and turnover of soil specificity phenomena in ectomycorrhizal plant communities: organic matter as indicated by carbon isotope measurements. potentials for interplant linkages and guild development. In: Soil Biology and Biochemistry 10Ž. 4 , 309–317. Read, D.J., Lewis, D.H., Fitter, A.H., Alexander, I.J.Ž. Eds. , Ochs, M., 1996. Influence of humified and non-humified natural Mycorrhizas in Ecosystems. CAB International, Oxford, pp. organic compounds on mineral dissolution. Chemical Geology 106–112. 132, 119–124. Moore, P.D., 1998. Life in the upper Crust. Nature 393, 419–420. Oechel, W.C., Hastings, S.J., Vourlitis, G., Jenkins, M., Riechers, Mora, C.I., Driese, S.G., Seager, P.G., 1991. Carbon dioxide in G., Grulke, N., 1993. Recent change of Arctic tundra ecosys- the Paleozoic atmosphere: evidence from carbon-isotope com- tems from a net carbon dioxide sink to a source. Nature 361, position of pedogenic carbonate. Geology 19, 1017–1020. 520–523. Mora, C.I., Driese, S.G., Colarusso, L.A., 1996. Middle to late Oeschger, H., Stauffer, B., 1986. Review of the history of atmo-

Paleozoic atmospheric CO2 levels from soil carbonate and spheric CO2 recorded in ice cores. In: Trabalka, J.R., Reichle, organic matter. Science 271, 1105–1107. D.E.Ž. Eds. , The Changing Carbon Cycle: A Global Analysis. Moulton, K.L., Berner, R.A., 1998. Quantification of the effects Springer-Verlag, New York, pp. 89–108. of plants on weathering: studies in Iceland. Geology 26, Orial, G., Brunet, A., 1991. Elne: the old cathedral cloister—the 895–898. capitals up against a bacterial attack. In: Decrouez, D., A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 151

Chamay, J., Zezza, F.Ž. Eds. , The Conservation of Monuments Pedersen, T.F., Calvert, S.E., 1990. Anoxia vs. productivity: what in the Mediterranean Basin. Museum d’Histoire naturelle and controls the formation of organic-carbon-rich sediments and Musee´` d’art et d’histoire, Ville de Geneve, pp. 237–245. sedimentary rocks? American Association of Petroleum Geol- Ortega-Calvo, J.J., Hernandez-Marine, M., Saiz-Jimenez, C., 1991. ogists Bulletin 74, 454–466. Biodeterioration of building materials by Cyanobacteria and Perraju, B.T.V.V., Rai, A.N., Kumar, A.P., Singh, H.N., 1986. Algae. International Biodeterioration 28, 165–185. Cycas circinalis – Anabaena cycadeae symbiosis: photo- Ortega-Calvo, J.J., Hernandez-Marine, M., Saiz-Jimenez, C., 1993. synthesis and the enzymes of nitrogen and hydrogen Cyanobacteria and algae on historic buiildings and monu- metabolism in symbiotic cultured Anabaena cycadeae. Sym- ments. In: Garg, K.L., Garg, N., Mukerji, K.G.Ž. Eds. , 1993. biosis 1, 239–250. Recent Advances in Biodeterioration and Biodegradation, vol. Peters, G.A., Calvert, H.E., 1983. The Azolla–Anabaena azollae 1, Naya Prokash, Calcutta, pp. 173–203. symbiosis. In: Goff, L.J.Ž. Ed. , Algal Symbiosis: A Continuum Ortega-Calvo, J.J., Arino, X., Stal, L.J., Saiz-Jimenez, C., 1994. of Interaction Strategies. Cambridge Univ. Press, Cambridge, Cyanobacterial sulfate accumulation from black crust of a pp. 109–145. historic building. Geomicrobiology Journal 12, 15–22. Peterson, R.L., Bradbury, S.M., 1995. Use of plant mutations, Ortega-Calvo, J.J., Arino, X., Hernandez-Marine, M., Saiz- intraspecific variants, and non-hosts. In: Varma, A., Hock, B. Jimenez, C., 1995. Factors affecting the weathering and colo- Ž.Eds. , Mycorrhiza: Structure, Function, Molecular Biology nization of monuments by phototrophic microorganisms. The and Biotechnology. Springer-Verlag, Berlin, pp. 153–176. Science of the Total Environment 167, 329–341. Peterson, R.L., Howarth, M.J., Whittier, D.P., 1981. Interactions Otto-Bliesner, B., 1995. Continental drift, runoff, and weathering between a fungal endophyte and gametophyte cells in Psilo- feedbacks: implications from climate model experiments. Jour- tum nudum. Canadian Journal of Botany 59, 711–720. nal of Geophysical Research 100Ž. D6 , 11537–11548. Petit, J.R., Basile, I., Leruyuet, A., Raynaud, D., Lorius, C.,

Owen, T., Cess, R.D., Ramanathan, V., 1979. Enhanced CO2 Jouzel, J., Stievenard, M., Lipenkov, V.Y., Barkov, N.I., greenhouse to compensate for reduced solar luminosity on Kudryashov, B.B., Davis, M., Saltzman, E., Kotlyakov, V., early Earth. Nature 277, 640–642. 1997. Four climate cycles in Vostok ice core. Nature 387, Pagani, M., Arthur, M.A., Freeman, K.H., 1999. Miocene evolu- 359–360. tion of atmospheric carbon dioxide. Paleoceanography 14, Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.-M., 273–292. Basile, I., Bender, M., Chappellaz, J., Davis, M., Delaygue, Palmer Jr., R.J., Hirsch, P., 1991. Photosynthesis-based microbial G., Delmonte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, communities on two churches in Northern Germany: weather- V.Y., Lorius, C., Pepin, L., Ritz, C., Saltzman, E., Stievenard, ing of granite and glazed brick. Geomicrobiology Journal 9, M., 1999. Climate and atmospheric history of the past 420,000 103–118. years from the Vostok ice core, Antarctica. Nature 399, 429– Paradise, T.R., 1997. Sandstone weathering from lichen over- 436. growth on Red Mountain, Arizona. Geografiska Annaler, A Petsch, S.T., Eglinton, T.I., Edwards, K.J., 2001. 14C-dead living 79, 177–184. biomass: Evidence for microbial assimilation of ancient or- Parnell, J., 1994. Late Proterozoic, northwest Scotland, U.K. In: ganic carbon during shale weathering. Science 292, 1127– Gierlowski-Kordesch, E., Kelts, K.Ž. Eds. , 1994. Global Geo- 1131. logical Record of Lake Basins, vol. 1. Cambridge Univ. Press, Petushkova, J., Lyalikova, N., 1993. The microbial deterioration Cambridge, pp. 63–64. of historical buildings and mural paints. In: Garg, K.L., Garg, Paul, E.A., Clark, F.E., 1996. Soil Microbiology and Biochem- N., Mukerji, K.G.Ž. Eds. , 1993. Recent Advances in Biodeteri- istry. Academic Press, 340 pp. oration and Biodegradation, vol. 1, Naya Prokash, Calcutta, Paul, E.A., Voroney, R.P., 1980. Nutrient and energy flows pp. 145–171. through soil microbial biomass. In: Ellwood, D.C., Hedger, Phelps, T.J., Fliermans, C.B., Garland, T.R., 1989a. Methods for J.N., Latham, M.J., Lynch, J.M., Slater, J.H.Ž. Eds. , Contem- recovery of deep terrestrial subsurface sediments for microbio- porary Microbial Ecology. Academic Press, London, pp. 215– logical study. Journal of Microbiological Methods 9, 267–279. 237. Phelps, T.J., Raione, E.G., White, D.C., Fliermans, C.B., 1989b. Payton, C.E.Ž. Ed. , Seismic stratigraphy—applications to hydro- Microbial activities in deep subsurface environments. Geomi- carbon exploration. American Association of Petroleum Geol- crobiology Journal 7, 79–91. ogists Memoir, vol. 26, 516 pp. Phelps, T.J., Murphy, E.M., Pfiffner, S.M., White, D.C., 1994a. Pearson, P.N., Palmer, M.R., 1999. Middle Eocene seawater pH Comparison between geochemical and biological estimates of and atmospheric carbon dioxide concentrations. Science 284, subsurface microbial activities. Microbial Ecology 28, 335– 1824–1826. 349. Pearson, P.N., Palmer, M.R., 2000. Atmospheric carbon dioxide Phelps, T.J., Pfiffner, E.M., Sargent, K.A., White, D.C., 1994b. concentrations over the past 60 million years. Nature 406, Factors influencing the abundance and metabolic capacities of 695–699. microorganisms in eastern Coastal Plain sediments. Microbial Pedersen, K., 1993. The deep subterranean biosphere. Earth-Sci- Ecology 28, 351–364. ence Reviews 34, 243–260. Piervittori, R., Salvadori, O., Laccisaglia, A., 1994. Literature on Pedersen, K., 1999. De sma under jorden. Geologiskt Forum 23, lichens and biodeterioration of stone work: I. Lichenologist 26 3–8. Ž.2 , 171–192. 152 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Piervittori, R., Salvadori, O., Laccisaglia, A., 1996. Literature on Ramstein, G., Fluteau, F., Besse, J., Joussaume, S., 1997. Effect lichens and biodeterioration of stonework: II. Lichenologist 28 of orogeny, plate motion and land–sea distribution on Eurasian Ž.5 , 471–483. climate change over the past 30 million years. Nature 386, Pietrini, A.M., Ricci, S., 1993. Occurrence of a calcareous blue- 788–795. green alga. Scytonema julianum Ž.Kutz. Meneghina, on the Raymo, M.E., 1991. Geochemical evidence supporting T.C. frescoes of a church carved from the rock in Matera, Italy. Chamberlin’s theory of glaciation. Geology 19, 344–347. Cryptogamic Botany 3, 290–295. Raymo, M.E., 1994a. The Himalayas, organic carbon burial, and Pirozynski, K.A., 1981. Interactions between fungi and plants climate in the Miocene. Paleoceanography 9, 399–404. through the ages. Canadian Journal of Botany 59, 1824–1827. Raymo, M.E., 1994b. The initiation of Northern Hemisphere Pirozynski, K.A., Dalpe, Y., 1989. Geological history of the glaciation. Annual Reviews of Earth and Planetary Sciences Glomaceae with particular reference to mycorrhizal symbiosis. 22, 353–383. Symbiosis 7, 1–36. Raymo, M.E., 1997. Carbon cycle models: how strong are the Pirozynski, K.A., Hawksworth, D.L., 1988a. Coevolution of fungi constraints? In: Ruddiman, W.F.Ž. Ed. , Global Tectonics and with plants and animals: introduction and overview. In: Climate Change. Plenum, New York, pp. 367–381.

Pirozynski, K.A., Hawksworth, D.L.Ž. Eds. , Coevolution of Raymo, M.E., Horowitz, M., 1996. Organic carbon Paleo-pCO2 Fungi with Plants and Animals. Academic Press, London, pp. and marine ice-core correlations. Geophysical Research Let- 1–29. ters 23, 367–370.

Pirozynski, K.A., Hawksworth, D.L.Ž. Eds. , Coevolution of Fungi Raymo, M.E., Rau, G., 1992. Plio-Pleistocene atmospheric CO2 with Plants and Animals. Academic Press, Harcourt Brace levels inferred from POM13 C at DSDP site 607. EOS 73Ž. 43 , Jovanovich, London, 285 pp. 95–96Ž abstract, Transactions of the American Geophysical Pirozynski, K.A., Malloch, D.W., 1975. The origins of land Union, Fall Meeting, 1992. . plants: a matter of mycotrophism. Biosystems 6, 153–164. Raymo, M.E., Ruddiman, W.F., 1992. Tectonic forcing of late Pittman, E.D., Lewan, M.D.Ž. Eds. , Organic Acids in Geological Cenozoic climate. Nature 359, 117–122. Processes. Springer-Verlag, Berlin, 482 pp. Raymo, M.E., Ruddiman, W.F., 1993a. Tectonic forcing of late Pocock, K., Duckett, J.G., 1984. A comparative ultrastructural Cenozoic climate. Nature 359, 117–122. analysis of the fungal endophyte in Cryptothallus mirabilis Raymo, M.E., Ruddiman, W.F., 1993. See Volk et al., 1993. Malm. and other British thalloid hepatics. J. Bryology 13, Raymo, M.E., Ruddiman, W.F., Froelich, P.N., 1988. Influence of 227–233. late Cenozoic mountain building on ocean geochemical cycles. Pocock, K., Duckett, J.G., 1985a. Fungi in hepatics. Bryology Geology 16, 649–653. Times 31, 2–3. Raymo, M.E., Grant, B., Horowitz, M., Rau, G.H., 1996. Mid- Pocock, K., Duckett, J.G., 1985b. On the occurrence of branched Pliocene warmth: stronger greenhouse and stronger conveyor. and swollen rhizoids in British hepatics: their relationships Marine Micropaleontology 27, 313–326. with the substratum and associations with fungi. The New Raymo, M.E., Ganley, K., Carter, S., Oppo, D.W., McManus, J., Phytologist 99, 281–304. 1998. Millenial-scale climate instability during the early Pleis- Poinar Jr., G., Peterson, E.B., Platt, J.L., 2000. Fossil Parmelia in tocene epoch. Nature 392, 699–702. New World amber. Lichenologist 32, 263–269. Read, D.J., 1984. Interactions between ericaceous plants and their Poole, I., Weyer, J.D.B., Lawson, T., Raven, J.A., 1996. Varia- competitors with special reference to soil toxicity. Aspects of tions in stomatal density and index: implications for palaeocli- Applied Biology 5, 195–209. matic reconstructions. Plant, Cell and Environment 19, 705– Read, D.J., 1993. Plant-microbe mutualisms and community struc- 712. ture. In: Schultz, E.-D., Mooney, H.A.Ž. Eds. , Biodiversity and Powell, C.L., Bagyaraj, D.J.Ž. Eds. , VA Mycorrhiza. CRC Press, Ecosystem Function. Ecological Studies, vol. 99. Springer, pp. Boca Raton, FL, 234 pp. 181–210. Prentice, I.C., Sykes, M.T., 1995. Vegetation geography and local Read, D.J., 1999. Mycorrhiza—the state of the art. In: Varma, A., carbon storage changes. In: Woodwell, G.M., MacKenzie, Hock, B.Ž. Eds. , Mycorrhiza. Springer, Berlin, pp. 3–34. F.T.Ž. Eds. , Biotic Feedbacks in the Global Climatic System: Read, D.J., Kerley, S., 1999. The status and function of mycor- Will the Warming Feed the Warming? Oxford Univ. Press, rhizal systems. In: Varma, A., Hock, B.Ž. Eds. , Mycorrhiza. Oxford, pp. 304–312. Springer, Berlin, pp. 499–520. Preto, B., Edwards, H.G.M., Seaward, M.R.D., 2000. A fourier Read, D.J., Duckett, J.G., Francis, R., Ligrone, R., Russell, A., transform-raman spectrographic study of lichen strategies on 2000. Symbiotic fungal associations in ‘lower’ land plants. granite monuments. Geomicrobiology Journal 17, 55–60. Philosophical Transactions of the Royal Society of London, Proctor, J.Ž. Ed. , Mineral Nutrients in Tropical Forest and Sa- Series B 355, 815–831. vanna Ecosystems. Blackwell, Oxford, 473 pp. Redecker, D., Kodner, R., Graham, L.E., 2000. Glomalean fungi Quade, J., Roe, L., DeCelles, P.G., Ojha, T.P., 1997. The Late from the Ordovician. Science 289, 1920–1921. Neogene 87Srr86 Sr record of lowland Himalayan rivers. Sci- Reid, C.P.P., 1990. Mycorrhizas. In: Lynch, J.M.Ž. Ed. , The ence 276, 1828–1831. Rhizosphere. Wiley, Chichester, pp. 281–315. Raich, J.W., Schlesinger, W.H., 1992. The global carbon dioxide Reimer, T.O., 1978. Stromatolitic dolomites in the Late Precam- flux in soil respiration and its relationship to vegetation and brian Malmesbury Group, Cape Province, South Africa. Sedi- climate. Tellus 44B, 81–99. mentary Geology 20, 29–40. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 153

Reimer, T.O., Barghoorn, E.S., Margulis, L., 1979. Primary pro- Ronov, A.B., Khain, V.E., 1961. Triasovye litologicheskie format- ductivity in an Early Archean microbial ecosystem. Precam- sii mira. Sovetskaya Geologiya 59, 27–48. brian Research 9, 93–104. Ronov, A.B., Khain, V.E., 1962. Yurskie litologicheskie formatsii Remy, W., Taylor, T.N., Hass, H., Kerp, H., 1994. Four hundred- mira. Sovetskaya Geologiya 60Ž. 1 , 9–34. million-year-old vesicular arbuscular mycorrhizae. Proceed- Ronov, A.B., Seslavinski, K.B., Khain, V.E., 1974. Kembriskie ings of the National Academy of Sciences 91, 11841–11843. litologicheskie formatsii mira. Sovetskaya Geologiya 12, 10– Retallack, G.J., 1986. The fossil record of soils. In: Wright, V.P. 33. Ž.Ed. , Paleosols: Their recognition and interpretation. Prince- Ronov, A.B., Khain, V.E., Seslavinski, K.B., 1976. Ordovikskie ton Univ. Press, Princeton, NJ, pp. 1–57. litologicheskie formatsii mira, Sovetskaya Geologiya 18Ž. 1 , Retallack, G.J., 2001. A 300-million-year record of atmospheric pp. 7–28, 7–27. carbon dioxide from fossil plant cuticles. Nature 411, 287–290. Ronov, A.B., Khain, V.E., Balukhovsky, A.N., 1978. Paleogen- Reynolds, R.C., Johnson, N.M., 1972. Chemical weathering in the ovye litologicheskie formatsii kontinentov. Sovetskaya Ge- temperate glacial environment of the northern Cascade Moun- ologiya 3, 10–42. tains. Geochimica et Cosmochimica Acta 36, 537–554. Ronov, A.B., Khain, V.E., Balukhovsky, A.N., Seslavinsky, K.B., Rice, C.M., Ashcroft, W.A., Batten, D.J., Boyce, A.J., Caulfield, 1980. Quantitative aspects of Phanerozoic sedimentation. Sed- J.B.D., Fallick, A.E., Hole, M.J., Jones, E., Pearson, M.J., imentary Geology 25, 311–325. Rogers, G., Saxton, J.M., Stuart, F.M., Trewin, N.H., Turner, Rosenzweig, C., Hillel, D., 1998. Climate Change and the Global G., 1995. A Devonian auriferous hot spring system, Rhynie, Harvest. Oxford Univ. Press, Oxford, 324 pp. Scotland. Journal of the Geological Society of London 152, Rossinsky, V., Swart, P.K., 1993. Influence of climate on the 229–250. formation and isotopic composition of calcretes. In: Swart, Riding, R.E., Awramik, S.M., 2000. Microbial Sediments. P.K., Lohmann, K.C., McKenzie, J., Savin, S.Ž. Eds. , 1993. Springer, 331 pp. Geophysical Monograph, vol. 78, pp. 67–75. Rieley, J.O., Richards, P.W., Bebbington, A.D.L., 1979. The Round, F.E., 1981. The Ecology of Algae. Cambridge Univ. ecological role of bryophytes in a north Wales woodland. Press, Cambridge, 653 pp. Journal of Ecology 67, 497–527. Royer, D.L., 2001. Stomatal density and stomatal index as indica-

Robert, M., Berthelin, J., 1986. Role of biological and biochemi- tors of paleoatmospheric CO2 concentration. Review of cal factors in soil mineral weathering. In: Huang, P.M., Palaeobotany and Palynology 114, 1–28. Schnitzer, M.Ž. Eds. , Interactions of Soil Minerals with Natu- Royer, D.L., Wing, S.L., Beerling, D.J., Jolley, D.W., Koch, P.L., ral Organics and Microbes. Soil Science Society of America Hickey, L.J., Berner, R.A., 2001. Paleobotanical evidence for

Special Publication, vol. 17, pp. 453–495, Madison, WI. near-present-day levels of atmospheric CO2 during part of the Robinson, J.M., 1990a. The burial of organic carbon as affected Tertiary. Science 292, 2310–2313. by the evolution of land plants. Historical Biology 3, 189–201. Ruddiman, W.F.Ž. Ed. , 1997. Tectonic Uplift and Climate Change. Robinson, J.M., 1990b. Lignin, land plants, and fungi: biological Plenum, New York, 535 pp. evolution affecting Phanerozoic oxygen balance. Geology 18, Ryan, W.B.F., Cita, M.B., 1977. Ignorance concerning episodes 607–610. of ocean-wide stagnation. Marine Geology 23, 197–215. Robinson, J.M., 1991. Land plants and weathering: discussion and Safir, G.R.Ž. Ed. , 1987. Ecophysiology of VA Mycorrhizal Plants. response. Science 252, 860, Technical Comment. CRC Press, Florida, 224 pp. Robinson, D.A., Williams, R.B.G.Ž. Eds. , 1994. Rock Weathering Saiz-Jimenez, C., Garcia-Rowe, J., 1991. Biodeterioration of mar- and Landform Evolution. Wiley, Chichester, 519 pp. bles and limestones in Roman pavements. In: Decrouez, D., Robinson, A.B., Baliunas, S.L., Soon, W., Robinson, Z.W., 1998. Chamay, J., Zezza, F.Ž. Eds. , The Conservation of Monuments Environmental effects of increased atmospheric carbon diox- in the Mediterranean Basin. Musee´ d’Histoire naturelle and ide. Med. Sent 3Ž. 5 , 171–178. Musee´` d’art et d’histoire, Ville de Geneve, pp. 263–271. Rogers, J.R., Bennett, P.C., Choi, W.J., 1998. Feldspars as a Sanchez, P.A., 1976. Properties and Management of Soils in the source of nutrients for microorganisms. American Mineralo- Tropics. Wiley-Interscience, New York, 618 pp. gist 83, 1532–1540. Sanders, W.B., Ascaso, C., Wierzchos, J., 1994. Physical interac- Ronov, A.B., 1958. Organic carbon in sedimentary rocksŽ in tions of two rhizomorph-forming lichens with their rock sub- relation to the presence of petroleum.Ž . Geochemistry now strate. Botanica Acta 107, 432–439. Geochemistry International.Ž. , volume for 1958 5 , 510–536. Sargent, K.A., Fliermans, C.B., 1989. Geology and hydrology of Ronov, A.B., 1976. Global carbon geochemistry, volcanism, car- the deep subsurface microbiology sampling sites at the Savan- bonate accumulation, and life. Geochemistry International 13 nah River plant, South Carolina. Geomicrobiology Journal 7, Ž.4 , 172–195. 3–13. Ronov, A.B., Khain, V.E., 1954. Devonskie litologicheskie for- Sarmiento, J.L., Hughes, T.M.C., Stouffer, R.J., Manabe, S., matsii mira. Sovetskaya Geologiya 41, 46–76. 1998. Simulated response of the ocean carbon cycle to anthro- Ronov, A.B., Khain, V.E., 1955. Kamennougolynye litologich- pogenic climate warming. Nature 393, 245–249. eskie formatsii mira. Sovetskaya Geologiya 48, 92–117. Schidlowski, M., 1983. Evolution of photoautotrophy and early Ronov, A.B., Khain, V.E., 1956. Permskie litologicheskie format- atmospheric oxygen levels. Precambrian Research 20, 319– sii mira. Sovetskaya Geologiya 54, 20–36. 335. 154 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Schidlowski, M., 1986. 13 Cr12 C ratios as indicators of biogenic- Schutz, H., Seiler, W., Rennenberg, H., 1990. Soil and land use

ity. In: Johns, R.B.Ž. Ed. , Biological Markers in the Sedimen- related sources and sinks of methaneŽ. CH4 in the context of tary Record. Elsevier, Amsterdam, pp. 347–361. the global methane budget. In: Bouwman, A.F.Ž. Ed. , Soils Schidlowski, M., 1988. A 3,800-million-year isotopic record of and the Greenhouse Effect. Wiley, Chichester, pp. 269–285. life from carbon in sedimentary rocks. Nature 333, 313–318. Schwartzman, D., 1993. Comment on AWeathering, plants, and Schidlowski, M., Aharon, P., 1992. Carbon cycle and carbon the long-term carbon cycleB by Robert A. Berner. Geochimica isotope records: geochemical impact of life over 3.8 Ga of et Cosmochimica Acta 57, 2145–2146. Earth history. In: Schidlowski, M., Golubic, S., Kimberley, Schwartzman, D.W., Shore, S.N., 1996. Biotically mediated sur- M.M., McKirdy, D.M., Trudinger, P.A.Ž. Eds. , Early Organic face cooling and habitability for complex life. In: Doyle, L.R. Evolution: Implications for Mineral and Energy Resources. Ž.Ed. , Circumstellar Habitable Zones. Proceedings of the First Springer-Verlag, Heidelberg, pp. 147–175. International Conference. Travis House Publications, Menlo Schimel, D.S., 1998. The carbon equation. Nature 393, 208–209. Park, pp. 421–443. Schimel, D.S., Braswell, B.H., Holland, E.A., McKeown, R., Schwartzman, D.W., Volk, T., 1989. Biotic enhancement of Ojima, D.S., Painter, T.H., Parton, W.J., Townsend, A.R., weathering and the habitability of Earth. Nature 340, 457–460. 1994. Climatic, edaphic, and biotic controls over storage and Schwartzman, D.W., Volk, T., 1991. Biotic enhancement of turnover of carbon in soils. Global Biogeochemical Cycles 8 weathering and surface temperatures on earth since the origin Ž.3 , 279–293. of life. Palaeogeography, Palaeoclimatology, Palaeoecology Schlesinger, W.H., 1977. Carbon balance in terrestrial detritus. 90, 385–394Ž. Global and Planetary Change 4, 357–371 . Annual Review of Ecology and Systematics 8, 51–81. Schwartzman, D., McMenamin, M., Volk, T., 1993. Did surface Schlesinger, W.H., 1985. The formation of caliche in soils of the temperatures constrain microbial evolution? Bioscience 43, Mojave Desert, California. Geochimica et Cosmochimica Acta 390–393. 49, 53–66. Seaward, M.R.D., 1988. Lichen damage to ancient monument: a Schlesinger, W.H., 1986. Changes in soil carbon storage and case study. Lichenologist 20, 291–295. associated properties with disturbance and recovery. In: Tra- Seaward, M.R.D., 1996. Lichens and the environment. In: Sutton, balka, J.R., Reichle, D.E.Ž. Eds. , The Changing Carbon Cycle, B.C.Ž. Ed. , A Century of Mycology. Cambridge Univ. Press, A Global Analysis. Springer-Verlag, New York, pp. 194–220. Cambridge, pp. 293–320. Schlesinger, M.E.Ž. Ed. , 1991. Greenhouse-Gas-Induced Climatic Sellwood, B.W., Valdes, P.J., 1997. Geological evaluation of Change: A Critical Appraisal of Simulations and Observations. climate General Circulation Models and model implications Elsevier, Amsterdam, 615 pp. for Mesozoic cloud cover. Terra Nova 9, 75–78. Schlesinger, W.H., 1995. Soil respiration and changes in soil Shachak, M., Jones, C.G., Granot, Y., 1987. Herbivory in rocks carbon stocks. In: Woodwell, G.M., MacKenzie, F.T.Ž. Eds. , and the weathering of a desert. Science 236, 1098–1099. Biotic Feedbacks in the Global Climatic System: Will the Shackleton, N.J., 2000. The 100,000-year ice-age cycle identified Warming Feed the Warming? Oxford Univ. Press, Oxford, pp. and found to lag temperature, carbon dioxide, and orbital 159–168. eccentricity. Science 289, 1897–1902. Schlunz, B., Schneider, R.R., 2000. Transport of terrestrial or- Sherwood-Pike, M.A., Gray, J., 1985. Silurian fungal remains: ganic carbon to the oceans by rivers: re-estimating flux and probable records of the Class Ascomycetes. Lethaia 18, 1–20. burial rates. International Journal of Earth Sciences 88, 599– Shine, K.P., Forster, P.M. de F., 1999. The effect of human 606. activity on radiative forcing of climate change: a review of Schmidt, E., Oberwinkler, F., 1993. Mycorrhiza-like interactions recent developments. Global and Planetary Change 20, 205– between the achlorophyllous gametophyte of Lycopodium 225. claÕatum L. and its fungal endophyte studied by light and Sigman, D.M., Boyle, E.A., 2000. Glacialrinterglacial variations electron microscopy. New Phytologist 124, 69–82. in atmospheric carbon dioxide. Nature 407, 859–869. Schneider, J., 1976. Biological and inorganic factors in the de- Sillitoe, R.H., Folk, R.L., Saric, N., 1996. Bacteria as mediators of struction of limestone coasts. Contributions to Sedimentology, copper sulfide enrichment during weathering. Science 272, 6, 112 pp.Ž. Translated by Peter Sadler E. Schweizerbart’sche 1153–1155. Verlagsbuchhandlung, Stuttgart. Silver, M., Ehrlich, H.L., Ivarson, K.C., 1986. Soil mineral trans- Schouten, S., Van Kaam-Peters, H.M.E., Rijpstra, W.I.C., Schoell, formation mediated by soil microbes. In: Huang, P.M., M., Damste, U.S.S., 2000. Effects of an oceanic anoxic event Schnitzer, L.Ž. Eds. , Interactions of Soil Minerals with Natural on the stable carbon isotope composition of Early Toarcian Organics and Microbes. Soil Science Society of America, carbon. American Journal of Science 300, 1–22. Special Publication, vol. 17, pp. 497–519. Schussler, A., 2000. Glomus claroideum forms an arbuscular Silverman, M.P., Munoz, E.F., 1970. Fungal attack on rock: mycorrhiza-like symbiosis with the hornwort Anthoceros solubilization and altered infrared spectra. Science 169, 985– punctatus. Mycorrhiza 10, 15–21. 987. Schuster, R.M., 1966. The Hepaticae and Anthocerotae of North Simon, L., Bousquet, J., Levesque, R.C., Lalonde, M., 1993. America East of the Hundredth Meridian, vol. I Columbia Origin and diversification of endomycorrhizal fungi and coin- Univ. Press, New York, 802 pp. cidence with vascular land plants. Nature 363, 67–69. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 155

Sinclair, J.L., Ghiorse, W.C., 1989. Distribution of aerobic bacte- of Silicate Minerals. Reviews in Mineralogy, vol. 31. Miner- ria, protozoa, algae, and fungi in deep subsurface sediments. alogical Society of America, Washington, DC, pp. 544–564. Geomicrobiology Journal 7, 15–31. Stallard, R.F., Edmund, J.M., 1983. Geochemistry of the Amazon: Singh, A., Sinha, A., 1993. Corrosion of natural and monument 2. The influence of the geology and weathering environment stone with special reference to lichen activity. In: Garg, K.L., on the load. Journal of Geophysical Research 88, 9671–9688. Garg, N., Mukerji, K.G.Ž. Eds. , Recent Advances in Biodeteri- Stauffer, B., Blunier, T., Dallenbach, A., Indermuhle, A., Schwan- oration and Biodegradation, vol. 1. Naya Prokash, Calcutta, der, J., Stocker, T.F., Tschumi, J., Chappellaz, J., Raynaud,

pp. 355–377. D., Hammer, C.U., Clausen, H.B., 1998. Atmospheric CO2 Sinha, A., Stott, L.D., 1994. New atmospheric pCO2 estimates concentration and millenial-scale climate change during the from paleosols during the late Paleocenerearly Eocene global last glacial period. Nature 392, 59–62. warming interval. Global and Planetary Change 9, 297–307. Stephens, B.B., Keeling, R.F., 2000. The influence of Antarctic

Sliter, W.V., 1989. Aptian anoxia in the Pacific Basin. Geology sea ice on glacial–interglacial CO2 variations. Nature 404, 17, 909–912. 171–174. Smith, A.G., 1997. Estimates of the Earth’s spinŽ. geographic axis Sterflinger, K., Krumbein, W.E., 1995. Multiple stress factors relative to Gondwana from glacial sediments and paleomag- affecting growth of rock-inhabiting black fungi. Botanica Acta netism. Earth-Science Reviews 42, 161–179. 108, 490–496. Smith, S.E., Read, D.J., 1997. Mycorrhizal Symbiosis. 2nd edn. Sterflinger, K., Krumbein, W.E., 1997. Dematiaceous fungi as a Academic Press, San Diego, Harcourt Brace and Company, major agent for biopitting on Mediterranean marbles and 605 pp. limestones. Geomicrobiology Journal 14, 219–230. Smith, T.M., Shugart, H.H., 1993. The transient response of Stevens, T.O., 1997. Subsurface microbiology and the evolution terrestrial carbon storage to a perturbed climate. Nature 361, of the biosphere. In: Amy, P.S., Haldeman, D.L.Ž. Eds. , The 523–526. Microbiology of the Terrestrial Deep Subsurface. CRC Press Soukharjevski, S., Gorbushina, A.A., Krumbein, W.E., Panina, L., LLC, Boca Raton, pp. 205–223. 1994. Recognition and identification of marbles and marble Stevens, T.O., McKinley, J.P., 1995. Lithoautotrophic microbial infecting black fungi using EPR-techniques. In: Fassina, V., ecosystems in deep basalt aquifers. Science 270, 450–454. Ott, H., Zezza, F.Ž. Eds. , III International Symposium on the Stevenson, F.J., 1967. Organic acids in soils. In: McLaren, A.D., Conservation of Monuments in the Mediterranean Basin. Peterson, G.H.Ž. Eds. , Soil Biochemistry, vol. 1. Dekker, pp. Musee´´ d’Histoire naturelle and Musee d’Art et d’Histoire, 119–146. Ville de Geneve,` pp. 335–341. Stevenson, F.J., 1994. Humus Chemistry: Genesis, Composition, Southgate, P.N., 1986. Proterozoic and Cambrian phosphorites— Reactions. 2nd edn. John Wiley, NY. specialist studies: middle Cambrian phosphatic hardgrounds, Stevenson, F.J., Fitch, A., 1986. Chemistry of complexation of phoscrete profiles and stromatolites and their implications for metal ions with soil solution organics. In: Huang, P.M., phosphogenesis. In: Cook, P.J., Shergold, J.H.Ž. Eds. , Phos- Schnitzer, M.Ž. Eds. , Interactions of Soil Minerals with Natu- phate Deposits of the World, vol. 1. Cambridge Univ. Press, ral Organics and Microbes. Soil Science Society of America, Cambridge, pp. 327–351. Special Publication, vol. 17, pp. 29–58. Sozinov, N.A., Gorbachev, O.V., 1987. Precambrian carbona- Stotzky, G., 1986. Influence of soil mineral colloids on metabolic ceous formations: their evolution and metal content. In: Kro- processes, growth, adhesion and ecology of microbes and ner, A.Ž. Ed. , Proterozoic Lithospheric Evolution. American viruses. In: Huang, P.M., Schnitzer, M.Ž. Eds. , Interactions of Geophysical Union, Geodynamic Series, vol. 17, pp. 35–42 soil minerals with natural organics and microbes. Soil Science ŽPublication No. 0130 of the International Lithosphere Pro- Society of America, Special Publication, vol. 17. gram. . Strakhov, N.M., 1967. Principles of Lithogenesis, vol. 1. Consul- Spark, I., Patey, I., Duncan, B., Hamilton, A., Devine, C., Mc- tants Bureau, New York, Oliver and Boyd, Edinburgh, 245 pp. Govern-Traa, C., 2000. The effects of indigenous and intro- English Translation. duced microbes on deeper buried hydrocarbon reservoirs, North Streel, M., 1986. Miospore contribution to the Upper Famen- Sea. Clay Minerals 15, 5–12. nian–Strunian event stratigraphy. Annales de la Societe´ Stallard, R.F., 1985. River chemistry, geology, geomorphology, Geologique´ de Belgique 109, 75–92. and soils in the Amazon and Orinoco Basins. In: Drever, J.I. Streel, M., 1992. Climatic impact on Famennian miospore distri- Ž.Ed. , The Chemistry of Weathering. Reidel, Dordrecht, pp. bution. Fifth International Conference on Global Bioevents 293–316. Ž.IGCP 216 , Gottingen,¨ Germany, Abstracts, pp. 108–109. Stallard, R.F., 1992. Tectonic processes, continental freeboard, Striegl, R.G., McConnaughey, T.A., Thorstenson, D.C., Weeks, and the rate controlling step for continental denudation. In: E.P., Woodward, J.C., 1992. Consumption of atmospheric Butcher, S.S., Charlson, R.J., Orians, G.H., Wolfe, G.V.Ž. Eds. , methane by desert soils. Nature 357, 145–147. Global Biogeochemical Cycles. Academic Press, London, pp. Stubblefield, S.P., Taylor, T.N., Beck, C.B., 1985. Studies of 93–121. Paleozoic fungi: IV. Wood-decaying fungi in Callixylon new- Stallard, R.F., 1995. Relating chemical and physical erosion. In: berryi from the Upper Devonian. American Journal of Botany White, A.F., Brantley, S.L.Ž. Eds. , Chemical Weathering Rates 72, 1765–1774. 156 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Stubblefield, S.P., Miller, C.E., Trappe, J.M., 1987a. Vesicular– Tissot, B.P., Welte, D.H., 1984. Petroleum Formation and Occur- arbuscular mycorrhizae from the Triassic of Antarctica. Amer- rence. 2nd edn. Springer-Verlag, Berlin, 699 pp. ican Journal of Botany 74, 1904–1911. Trabalka, J.R., Reichle, D.E.Ž. Eds. , 1986. The Changing Carbon Stubblefield, S.P., Taylor, T.N., Trappe, J.M., 1987b. Fossil myc- Cycle, A Global Analysis. Springer-Verlag, New York, 592 orrhizae: a case for symbiosis. Science 237, 59–60. pp. Sundquist, E.T., 1986. Geologic analogs: their value and limita- Trappe, J.M., 1962. Fungus associates of ectotrophic mycorrhizae. tions in carbon dioxide research. In: Trabalka, J.R., Reichle, Botanical Review 28, 538–603. D.E.Ž. Eds. , The Changing Carbon Cycle: A Global Analysis. Trappe, J.M., 1977. Selection of fungi for ectomycorrhizal inocu- Springer-Verlag, New York, pp. 371–402. lation in nurseries. Annual Review of Phytopathology 15, Sutcliffe, O.E., Dowdeswell, J.A., Whittington, R.J., Theron, J.N., 203–222. 2000. Calibrating the Late Ordovician glaciation and mass Trappe, J.M., 1987a. Phylogenetic and ecological aspects of my- extinction by the eccentricity cycles of Earth’s orbit. Geology cotrophy in the angiosperms from an evolutionary standpoint. 28, 967–970. In: Safir, G.R.Ž. Ed. , Ecophysiology of VA Mycorrhizal Plants. Sveinbjornsson, B., Oechel, W.C., 1992. Controls on growth and CRC Press, Boca Raton, FL, pp. 5–25. productivity of bryophytes: environmental limitations under Trappe, J.M., 1987b. Lessons from Alpine fungi. Mycologia 80, current and anticipated conditions. In: Bates, J.W., Farmer, 1–10. A.M.Ž. Eds. , Bryophytes and Lichens in a Changing Environ- Tretiach, M., Pecchiari, M., 1995. Gas exchange rates and chloro- ment. Clarendon Press, Oxford, pp. 77–102. phyll content of epi- and endolithic lichens from the Trieste Syers, J.K., Iskandar, I.K., 1973. Pedogenetic significance of KarstŽ. NE Italy . New Phytologist 130, 585–592. lichens. In: Ahmadjian, V., Hale, M.E.Ž. Eds. , The Lichens. Trewin, N.H., 1994. Depositional environment and preservation of Academic Press, New York, pp. 225–248. biota in the Lower Devonian hot-springs of Rhynie, Aberdeen- Tan, K.H., 1986. Degradation of soil minerals by organic acids. shire, Scotland. Transactions of the Royal Society of Edin- In: Huang, P.M., Schnitzer, M.Ž. Eds. , Interactions of Soil burgh: Earth Sciences 84, 433–442. Minerals with Natural Organics and Microbes. Soil Science Tuovila, B.J., LaRock, P.A., 1987. Occurrence and preservation of Society of America, Special Publication, vol. 17, pp. 1–27. ATP in Antarctic rocks and its implications in biomass deter- Tanner, L.H., Hubert, J.F., Coffey, B.P., McInnerney, D.P., 2001. minations. Geomicrobiology Journal 5Ž. 2 , 105–118. r Stability of atmospheric CO2 levels across the Triassic Tyler, S.A., Barghoorn, E.S., Barrett, L.P., 1957. Anthracitic coal Jurassic boundary. Nature 411, 675–677. from Precambrian upper Huronian black shale of the Iron Tappan, H., 1968. Primary production, isotopes, extinctions and River District, Northern Michigan. Geological Society of the atmosphere. Palaeogeography, Palaeoclimatology, Palaeoe- America Bulletin 68, 1293–1304. cology 4, 187–210. Ullman, W.J., Kirchman, D.L., Welch, S.A., Vandevivere, P., Tappan, H., Loeblich Jr., A.R., 1970. Geobiologic implications of 1996. Laboratory evidence for microbially mediated silicate fossil phytoplankton evolution and time–space distribution. mineral dissolution in nature. Chemical Geology 132, 11–17. Geological Society of America, Special Paper 127, 247–340. Ulmishek, G.F., Klemme, H.D., 1990. Depositional controls, dis- Tardy, Y., N’Kounkou, R., Probst, J.-L., 1989. The global water tribution, and effectiveness of world’s petroleum source rocks. cycle and continental erosion during Phanerozoic timeŽ 570 U.S. Geological Survey Bulletin, 1931, 59 pp. my. . American Journal of Science 289, 455–483. Unwins, P.J.R., Webb, R.I., Taylor, A.P., 1998. Novel nano- Taylor, T.N., White Jr., J.F., 1989. Fossil fungiŽ. endogonaceae organisms from Australian sandstones. American Mineralogist from the Triassic of Antarctica. American Journal of Botany 83, 1541–1550. 76Ž. 3 , 389–396. Upchurch Jr., G.R., Otto-Bliesner, B.L., Scotese, C., 1998. Vege- Taylor, T.N., Hass, H., Kerp, H., 1997. A cyanolichen from the tation–atmosphere interactions and their role in global warm- Lower Devonian Rhynie Chert. American Journal of Botany ing during the latest Cretaceous. Philosophical Transactions of 84, 992–1004. the Royal Society of London B 353Ž. 1365 , 97–112. Theron, J.N., 1994. The ordovician system in South Africa. In: Urey, H.C., 1952. On the early chemical history of the Earth and Williams, H.S.Ž. Ed. , International Union of Geological Sci- the origin of life. Proceedings of the National Academy of ences Publication, vol. 29, pp. 1–5. Sciences, United States of America 38, 351–363. Thorseth, H., Furnes, H., Feldal, M., 1992. The importance of Urzi, C., Krumbein, W.E., Warscheid, T., 1991. On the question microbiological activity in the alteration of natural basaltic of biogenic colour changes of Mediterranean monuments glass. Geochimica et Cosmochimica Acta 56, 845–850. Žcoatings–crust–microstromatolite –patina –scialbatura –skin – Thorseth, I.H., Furnes, H., Tumyr, O., 1995. Textural and chemi- rock varish.Ž. . In: Decrouez, D., Chamay, J., Zezza, F. Eds. , cal effects of bacterial activity on basaltic glass: an experimen- The Conservation of monuments in the Mediterranean basin. tal approach. Chemical Geology 119, 139–160. Museum d’Histoire naturelle and Musee´ d’art et d’histoire, Tiano, P., 1993. Biodeterioration of stone monuments: a critical Ville de Geneve,` pp. 397–420. review. In: Garg, K.L., Garg, N., Mukerji, K.G.Ž. Eds. , Recent Urzi, C., Criseo, G., Krumbein, W.E., Wollenzien, U., Gor- Advances in Biodeterioration and Biodegradation, vol. 1, Naya bushina, A.A., 1993. Are colour changes of rocks caused by Prokash, Calcutta, pp. 301–321. climate, pollution, biological growth, or by interactions of the A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 157

three? In: Thiel, M.-J.Ž. Ed. , 21 Proceedings of the Interna- tufa deposits of the Napier Range, N.W. Australia. Earth tional RILEMrUNESCO Congress on the Conservation of Surface Processes and Landforms 15, 425–443. Stone and Other Materials, vol. 1. E & FN Spon, London, pp. Viles, H., Pentecost, A., 1994. Problems in assessing the weather- 279–286. ing action of lichens with an example of epiliths on sandstone. Urzi, C., Krumbein, W.E., Lyalikova, N., Petushkova, J., Wollen- In: Robinson, D.A., Williams, R.B.G.Ž. Eds. , Rock Weather- zien, U., Zagari, M., 1994. Microbiological investigations of ing and Landform Evolution. Wiley, Chichester, pp. 99–116. marbles exposed to natural and anthropogenic influences in Villamil, T., Arango, C., Hay, W.W., 1999. Plate tectonic paleo- northern and southern climates. In: Fassina, V., Ott, H., Zezza, ceanographic hypothesis for Cretaceous source rocks and cherts F.Ž. Eds. , III International Symposium on the conservation of of northern South America, Special Paper 332. Geological monuments in the Mediterranean Basin. Museum d’Histoire Society of America, 191–202. naturelle and Musee´` d’art et d’histoire, Ville de Geneve, pp. Visscher, H., Brinkhuis, H., Dilcher, D.L., Elsik, W.C., Eshet, Y., 297–304. Looy, C.V., Rampino, M.R., Traverse, A., 1996. The terminal Van der Burgh, J., Visscher, H., Dilcher, D.L., Kurschner, W.M., Paleozoic fungal event: evidence of terrestrial ecosystem 1993. Paleoatmospheric signatures in Neogene fossil leaves. destabilization and collapse. Proceedings of the National Science 260, 1788–1790. Academy of Sciences of the United States of America 93, Vandevivere, S.A., Welch, S.A., Ullman, W.J., Kirchman, D.L., 2155–2158. 1994. Enhanced dissolution of silicate minerals by bacteria at Volk, T., 1989. Rise of angiosperms as a factor in long-term near-neutral pH. Microbial Ecology 27, 241–251. climatic cooling. Geology 17, 107–110. Varma, A., 1995. Ecophysiology and application of arbuscular Volk, T., Caldeira, K., Arthur, M.A., Berner, R.A., Lasaga, A.C., mycorrhizal fungi in arid soils. In: Varma, A., Hock, B. Raymo, M.E., Ruddiman, W., 1993. Cooling in the Late Ž.Eds. , Mycorrhiza—Structure, Function, Molecular Biology Cenozoic. Nature 361, 123–124, Scientific Correspondance. and Biotechnology. Springer-Verlag, Berlin, pp. 561–591. Volkert, R.A., Johnson, C.A., Tamashausky, A.V., 2000. Meso- Varma, A., 1999. Functions and applications of arbuscular mycor- proterozoic graphite deposits, New Jersey Highlands: geologic rhizal fungi in arid and semi-arid soils. In: Varma, A., Hock, and stable isotopic evidence of possible algal origins. Cana- B.Ž. Eds. , Mycorrhiza. Springer, pp. 521–556. dian Journal of Earth Sciences 37, 1665–1675. Varma, A., Hock, B.Ž. Eds. , 1995. Mycorrhiza—Structure, Func- Volkova, I.B., Bogdanova, M.V., 1986. Petrology and genesis of tion, Molecular Biology and Biotechnology. Springer Verlag, Karelian shungite—high rank coal. International Journal of Berlin, 747 pp. Coal Geology 6, 369–379. Veizer, J., Godderis, Y., Francois, L.M., 2000. Evidence for Volkova, I.B., Bogdanova, M.V., 1998. Genesis of organic matter

decoupling atmospheric CO2 and global climate during the in old Karelian carbonaceous sequences. Carbonaceous Forma- Phanerozoic eon. Nature 408, 698–701. tions in Geologic History, Section of Geology, Geophysics, Velbel, M.A., 1993. Temperature dependence of silicate weather- Geochemistry, and Mining Sciences, RAS, pp. 37–38. ing in nature: how strong a negative feedback on long-term Wagner, M., Schwartz, W., 1967. Geomicrobiological studies:

accumulation of atmospheric CO2 and global greenhouse VIII. Effect of bacteria on the surface of rocks and minerals warming? Geology 21, 1059–1062. and their role in erosion. Zeitschrift fur Allgemeine Mikrobi- Vestal, J.R., 1993. Cryptoendolithic communities from hot and ologie 7, 33–52. cold deserts: speculation on microbial colonization and succes- Wagner, F., Below, R., de Klerk, P., Dilcher, D.L., Joosten, H., sion. In: Miles, J., Walton, D.W.H.Ž. Eds. , Primary Succession Kurschner, W.M., Visscher, H., 1996. A natural experiment on LandŽ Special Publication No. 12 of the British Ecological on plant acclimation: lifetime stomatal frequency response of

Society. . Blackwell, Oxford, pp. 5–16. an individual tree to annual atmospheric CO2 increase. Pro- Viles, H.A., 1984. Biokarst: review and prospect. Progress in ceedings of the National Academy of Sciences of the United Physical Geography 8, 523–542. States of America 93, 11705–11708. Viles, H.A., 1987a. Blue-green algae and terrestrial weathering on Waisel, Y., Eshel, A., Kafkafi, U., 1996. Plant Roots, The Hidden Aldabra Atoll: an SEM and light microscope study. Earth Half. Marcel Dekker, New York, 1002 pp. Surface Processes and Landforms 12, 319–330. Walker, J.C.G., Hays, P.B., Kasting, J.F., 1981. A negative Viles, H., 1987b. A quantitative scanning electron micrscope feedback mechanism for the long-term stabilization of Earth’s study of evidence for lichen weathering of limestone, Mendip surface temperature. Journal of Geophysical Research 86, Hills, Somerset. Earth Surface Processes and Landforms 12, 9776–9782. 467–473. Walton, D.W.H., 1985. A preliminary study of the action of Viles, H.A., 1990. ‘The agency of organic beings’: a selective crustose lichens on rock surfaces in Antarctica. In: Siegfried, review of recent work in biogeomorphology. In: Thornes, J.B. W.R., Condy, P.R., Laws, R.M.Ž. Eds. , Antarctic Nutrient Ž.Ed. , Vegetation and Erosion: Processes and Environments. Cycles and Food Webs. Springer-Verlag, Berlin, 700 pp. Wiley, Chichester, pp. 5–24. Walton, D.W.H., 1993. The effects of cryptogams on mineral Viles, H., 1995. Ecological perspectives on rock surface weather- substrates. In: Miles, J., Walton, D.W.H.Ž. Eds. , Primary ing: towards a conceptual model. Geomorphology 13, 21–35. Succession on LandŽ Special Publication No. 12 of the British Viles, H.A., Goudie, A.S., 1990. Reconnaissance studies of the Ecological Society. . Blackwell, Oxford, pp. 33–53. 158 A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159

Wang, K., Chatterton, B.D.E., Wang, Y., 1997. An organic carbon basidiocarp dehiscence in Geastrum saccatum ŽGastro- isotope record of Late Ordovician to Early Silurian marine mycetes. . Mycologia 78, 649–656. sedimentary rocks, Yangtze Sea, South China: implications for Wignall, P.B., 1994. Black Shales. Clarendon Press, Oxford, 127

CO2 changes during the Hirnantian glaciation. Palaeogeogra- pp. phy, Palaeoclimatology, Palaeoecology 132, 147–158. Wilcox, H.E., 1996. Mycorrhizae in plant roots, the hidden half. Watanabe, Y., Martini, J.E.J., Ohmoto, H., 2000. Geochemical In: Waisel, Y., Eshel, A., Kafkafi, U.Ž. Eds. , Plant Roots—The evidence for terrestrial ecosystems 2.6 billion years ago. Na- Hidden Half. Marcel Dekker, New York, pp. 689–721. ture 408, 574–578. Williams, M.E., Rudolph, E.D., 1974. The role of lichens and Watson, A.J., Liss, P.S., 1998. Marine biological controls on associated fungi in the chemical weathering of rock. Mycolo- climate via the carbon and sulphur geochemical cycles. Philo- gia 66, 648–660. sophical Transactions of the Royal Society of London, Series Wilson, M.J., Jones, D., 1983. Lichen weathering of minerals: B 353, 41–51. implications for pedogenesis. In: Wilson, R.C.L.Ž. Ed. , Resid- Webley, D.M., Henderson, M.E.K., Taylor, I.F., 1963. The micro- ual Deposits: Surface Related Weathering Processes and Mate- biology of rocks and weathered stones. Journal of Soil Science rials. Geological Society of London Special Publication, vol. 14, 102–112. 11, Blackwell, Oxford, pp. 5–12. Weed, S.B., Davey, C.B., Cook, M.G., 1969. Weathering of mica Wollenzien, U., de Hoog, G.S., Krumbein, W.E., Urzi, C., 1995. by fungi. Soil Science Society of America Proceedings 33, On the isolation of microcolonial fungi occurring on and in 702–706. marble and other calcareous rocks. The Science of the Total Wein, R.W., Speer, J.E., 1975. Lichen biomass in Acadian and Environment 167, 287–294. Boreal Forest of Cape Breton Island, Nova Scotia. The Bryol- Woodward, F.I., 1987. Stomatal numbers are sensitive to in-

ogist 78, 328–333. creases in CO2 from pre-industrial levels. Nature 327, 617– Weissert, H., 2000. Deciphering methane’s fingerprint. Nature 618. 406, 356–357. Woodward, F.I., Thompson, G.B., McKee, I.F., 1991. The effects Welch, S.A., Ullman, W., 1993. The effects of organic acids on of elevated concentrations of carbon dioxide on individual plagioclase dissolution rates and stoichiometry. Geochimica et plants, populations, communities and ecosystems. Annals of Cosmochimica Acta 57, 2725–2736. Botany 67Ž. Suppl. 1 , 23–38. Welte, D.H., 1970. Organische Kohlenstoff und die Entwicklung Woodwell, G.M., 1995. Biotic feedbacks from the warming of the der Photosynthese auf der Erde. Naturwissenschaften Jahrgang Earth. In: Woodwell, G.M., MacKenzie, F.T.Ž. Eds. , Biotic 57, 17–23. Feedbacks in the Global Climatic System: Will the Warming Wessels, D.C.J., Budel, B., 1995. Epilithic and cryptoendolithic Feed the Warming? Oxford Univ. Press, Oxford, pp. 3–21. cyanobacteria of Clarens Sandstone Cliffs in the Golden Gate Woodwell, G.M., Mackenzie, F.T.Ž. Eds. , 1995. Biotic Feedbacks Highlands National Park, South Africa. Botanica Acta 108, in the Global Climatic System: Will the Warming Feed the 220–226. Warming? Oxford Univ. Press, Oxford, 416 pp. Wessels, D., Kappen, L., 1993. Photosynthetic performance of Woodwell, G.M., Mackenzie, F.T., Houghton, R.A., Apps, M.J., rock-colonizing lichens in the Mount Zebra National Park, Gorham, E., Davidson, E.A., 1995. Will the warming speed South Africa. Koedoe 36, 27–48. the warming? In: Woodwell, G.M., MacKenzie, F.T.Ž. Eds. , West, N.E., 1990. Structure and function of microphytic soil Biotic Feedbacks in the Global Climatic System: Will the crusts in wildland ecosystems of arid to semi-arid regions. Warming Feed the Warming? Oxford Univ. Press, Oxford, pp. Advances in Ecological Research 20, 179–223. 393–411. Weyers, J.D.B., Lawson, T., 1997. Heterogeneity in stomatal Worsley, T.R., Moore, T.L., Fraticelli, C.M., Scotese, C.R., 1994.

characteristics. Advances in Botanical Research 26, 317–352. Phanerozoic CO2 levels and global temperatures inferred from Whelan, J.K., Farrington, J.W.Ž. Eds. , 1992. Organic Matter: changing paleogeography. Geological Society of America, Productivity, Accumulation, and Preservation in Recent and Special Paper 288, 57–73.

Ancient Sediments. Columbia Univ. Press, New York, 533 pp. Yapp, C.J., Poths, H., 1992. Ancient atmospheric CO2 pressures Whipple, K.X., Kirby, E., Brocklehurst, S.H., 1999. Geomorphic inferred from natural goethites. Nature 355, 342–344. limits to climate-induced increases in topographic relief. Na- Yapp, C.J., Poths, H., 1994. Productivity of pre-vascular continen-

ture 401, 39–43. tal biota inferred from the FeŽ. CO3 OH content of goethite. Whipps, J.M., 1990. Carbon economy. In: Lynch, J.M.Ž. Ed. , The Nature 368, 49–51. Rhizosphere. Wiley, pp. 59–97. Yapp, C.J., Poths, H., 1996. Carbon isotopes in continental weath-

White, A.F., Blum, A.E., 1995. Effects of climate on chemical ering environments and variation in ancient atmospheric CO2 weathering in watersheds. Geochimica et Cosmochimica Acta pressure. Earth and Planetary Science Letters 137, 71–82. 59Ž. 9 , 1729–1747. Yatsu, E., 1988. The Nature of Weathering: An Introduction. White, A.F., Brantley, S.L.Ž. Eds. , 1995. Chemical weathering Sozosha, Tokyo, 624 pp. rates of silicate minerals. Reviews in Mineralogy, vol. 31, 583 Zwieniecki, M.A., Newton, M., 1995. Roots growing in rock pp. fissures: their morphological adaptation. Plant and Soil 172, Whitney, K.D., Arnott, H.J., 1986. Calcium oxalate crystals and 181–187. A.J. Boucot, J. GrayrEarth-Science ReÕiews 56() 2001 1–159 159

Jane Gray, 1929-2000. ShearŽ NA- A.J. Boucot. Born in 1924, received his TURE, 405:34. provided an obituary that university education at HarvardŽ PhD in documents Gray’s career. She was pre- 1953. , with a thesis involving Silurian dominantly a paleoecologist of the non- and Devonian rocks and fossils in the marine, both terrestral and aquatic. Her Northern Appalachians. Initial employ- course in nonmarine paleoecology was ment was with the former Stratigraphy unique and could have been the basis and Paleontology Branch of the U.S. of a first rate text. She was capable Geological Survey in Washington, fol- of taking on a new area, exploring it lowed by teaching stints at M.I.T., Cal in depth, and then making major con- Tech and Oregon State University. His tributions; for example, recognizing specialty is the morphology-taxonomy that Middle Ordovician-earlier Silurian of later Ordovician-Devonian articulate spores, commonly occurring in obligate tetrads until near the end brachiopods. He works with global biogeography of the mid- of the Early Silurian belonged to higher land plants, embryophytes Paleozoic and its community paleoecology. Also concerned with at an hepatic level. Another noteworthy example is her Palaeo- what the fossil record can provide about rates of behavioral geography, Palaeoclimatology, PalaeoclimatologyŽ. 1988 treat- evolution and coevolution. More recent work on Phanerozoic ment of the evolution and contents of the freshwater ecosystem climatic belt positions globally, in cooperation with Chen Xu, during the Phanerozoic. She was also a very capable palynologist, Nanjing Institute of Geology & Palaeontology and Christopher thoroughly familiar with the Neogene pollen floras of the Pacific Scotese, University of Texas at Arlington, is aimed at providing Northwest, her doctoral thesis area. more reliable global paleogeographies on which to plot biogeographic information as it becomes available.