Geochimica et Cosmochimica Acta, Vol. 58, No. 21, pp. 4809-4814, 1994 Copyright 0 1994 ElsevierScience Ltd Pergamon Printed in the USA. All rightsreserved 0016-7037/94 $6.00 + .OO

0016-7037(94)00234-7

LETTER

613C of organic carbon in the Bengal Fan: Source evolution and transport of C3 and C4 plant carbon to marine sediments

CHRISTIAN FRANCE-LANORD’ and LOUIS A. DERRY’ ‘Centre de Recherches Pktrographiques et G&ochimiques, CNRS, BP20 54501, Vandoeuvre-les-Nancy, France *Cornell University, Department of Geological Sciences, Ithaca, NY 14853, USA

(Received July 8, 1994; accepted in revisedform August 29, 1994)

Abstract-Carbon isotopic measurements on organic carbon (oc) in sediment cores from the Bengal Fan (ODP Leg 116) show a dramatic 10 %Oincrease beginning ca. 7 Ma ago, and a rapid decrease after 0.9 Ma. These shifts reflect changes in the mixing ratio of terrigenous carbon derived from C3 and C4 plants. The rapid increase in 613C of Bengal Fan OC at 7 Ma shows that Late Miocene expansion of C4 plants already documented in the Siwaliks was widespread over all the Himalayan foreland. After 7 Ma, relations of 613C with sedimentological parameters show that C4 plants dominate in the foreland whereas C3 plants remain abundant in the mountain range. Variations in the source of the sediments and of the OC appear to be sensitive to climate-hydrologic conditions in the basin. Major changes in the isotopic composition of the carbon flux in one of the worlds largest river systems modify the isotopic budgets of both marine dissolved carbon and the sedimentary carbon mass.

INTRODUCHON 1988) that deep sea sediments in the Indian Ocean are an important sink for OC derived from terrestrial plants. Data THE BENGAL FAN IS one of the largest sedimentary units in from ODP Legs 116 and 117 support this hypothesis. Gas the world, containing ca. 3 X lOI kg of sediment, primarily chromatograph and pyrolysis data show that OC in the Bengal derived from the erosion of the Himalayan orogen (CURRAY, Fan is dominated by organic matter derived from tenigenous 1991; CURRAY and MOORE, 1971; REA, 1992). The rate of sources (BERTRAND et al., 1991; MEYERS and DICKENS, 1992; sediment delivery appears to have increased markedly in early COCHRAN et al., 1989; POYNTER and EGLINTON, 1990). Car- Miocene time (ca. 21 Ma; COCHRAN, 1990; LINDSAY et al., bon isotopic analyses (S13C) of paleosol carbonates, organic 199 1 ), and again at lo- 11 Ma (REA, 1992), implying an av- matter, and fossil teeth (QUADE et al., 1989, 1992; CERLING erage sediment delivery rate of cu. 1.5 X 10” kg/y during the et al., 1993; QUADE and CERLING, 1994) show evidence of late Neogene. This mean rate is similar to estimates of the a major shift in terrestrial vegetation patterns in the Hima- modem rate (MILLIMAN and SYVITSKI, 1992). Measurements layan foreland in late Miocene time. A terrestrial plant com- obtained from drill core of early Miocene to Quaternary age munity dominated by C4 plants appears to replace a C3- in the distal portion of the fan (ODP Leg 116) have an average dominated community between 7 and 6 Ma. The stable iso- total organic carbon (TOC) content of 0.85 wt% (n = 155) topic results raise several questions. Was the expansion of (COCHRAN et al., 1989, and this study). A smaller number C4 plants confined to a relatively small area in the foreland, of analyses (n = 12) from Quaternary sediments in the sub- or was it of genera1 importance in the northern Indian sub- marine portion of the proximal Bengal delta are similar (X continent? What caused the shift in the dominant photosyn- = 0.90 wt%; BOUQUILLON, 1987). If these results are repre- thetic pathway? What effects did the expansion of C4 biomass sentative of the fan as a whole, the mean burial rate of OC have on the budget of OC exported to the ocean and buried in the Bengal Fan since early the Miocene has been = 1.1 in marine sediments? Since the bulk of the OC preserved in X 10” mol/y. Recent published estimates for global OC burial the Bengal Fan is derived from the Ganga and Brahmaputra are from 4 to 10 X 10” mol/y (LASAGA et al., 1985; SMITH river basins, 613C analyses of sedimentary OC provide insight and HOLLIBAUGH, 1993); thus, the Bengal Fan may account into this question. for 10 to 25% of the total global OC burial flux during the Neogene. If the Indus Fan, Indo-Gangetic Plain, and Ganga- THE BENGAL FAN Yamuna Delta are included, the estimate of OC burial in Neogene sediments derived from the Himalaya increases by Sedimentation history of the fan can be illustrated by vari- a factor of roughly 1.5. It is evident that the global OC burial ations in rate of deposition, clay mineralogy, grain size, and budget may have been significantly perturbed by rapid erosion TOC (Fig. I). Bengal Fan sediments older than 7 Ma or of the Himalayan region. younger than 0.9 Ma are dominantly silts with an illite-rich The fraction of biologically labile particulate organic carbon clay fraction, TOC content usually I 0.5 wt%, show relatively (POC) in the Ganga river is low, which suggests (ITTEKKOT, little chemical weathering, and were deposited at high sedi-

4809 C. France-Lanord and L. A. Derry

+Smectite +Illite

6 ;a 8 2 - 10 a% 12

16

18

0 200 400 600 800 0 0.3 0.6 0.9 200 400 600 0 0.5 f 1.5 2 Depth below sea floor (m) (Illite+Chlorite)/CClay Max. quartz grain size (pm) TOC fwt%)

FIG. I. Age dependence with (a) sediment thickness (GARTNER, 1990). (b) clay mineralogy (B~UQLJILLONet al., 1990), (c) maximum quartz grain size, and (d) TOC content (COCHRANet al.. 1989) at ODP Sites 717 (0) and 7 18 (0) and data in this study (+). (a) lower slopes correspond to higher sedimentation rates. (b) (Illite + chlorite)/Xclay ratios are measured on <2 pm fractions. They mostly reflect the relative abundance of illite vs. smectite. In the Bengal Fan illite and chlorite are derived from metamo~hic precursors by physical erosion whereas smectite is primarily derived from weathering of the same material (FRANCE-LANORDet al., 1993). %TOC appears to reflect grain size/ specific surface area of sediments (e.g., KEIL et al.. 1994)

mentation rates. In contrast, between 7 and 0.9 Ma, sedi- measured for TOC content, and analysed on a VG602D mass spec- mentation rates of mostly smectite-rich muds and fine silts trometer. Reproducibility is better than 0.2% and d”C of NBS-22 was measured at -29.7 i 0.1 %Orelative to PDB. Deposition ages of were lower, the degree of chemicai weathering was much sediment were estimated from biostrati~phic data (GARTNER,f 990) more intense, and TOC 2 1 wt%. However, neodymium, following the chronostratigraphy of BERGGRENet al. (1985). strontium, oxygen. and hydrogen isotopic data show that these variations were not caused by changing sediment 6°C on total organic carbon (TOC) was analyzed for sources, but rather by changing continental alteration histories twenty-eight samples from ODP Holes 7 17C and 7 1SC (Fig. (BOIJQUILLONet al., 1990; FRANCE-LANORD et al., 1993). 2). Together, these two holes provide a continuous 18 Ma Comparable variations in sedimentation rate and weathering record of turbidite sedimentation (STOW et al.. 1990). Prior intensity are also observed in the Siwaiik Himalayan foreland to ea. 7 Ma, 613C values in TQC are -27 to -25%, typical sequence (BURBANK et al., 1993; QUADE and CERLINC, of C3 plants. The Si3C data show a dramatic positive shift of 1994). These variations have been interpreted to reflect a 10% beginning near 7 Ma. 613C values in Pliocene sediments change in the rate of sediment supply and the residence time of the Bengal Fan reach -15 to -17%0. In late Quaternary of sediment in the foreland basin (BURBANK et al., 1993; sediments 6’“C values decrease rapidly to -24%. Limited FRANCE-LANORD et al., 1993). According to this hypothesis, &13C data from OC at DSDP Hole 218 show comparable between 7 and 0.9 Ma sediments were stored longer and values with the exception of one sample (ERDMAN et at., weathered more intensely in the Indo-Gangetic Plain. The 1974). The range of the 613C variations we observe is consid- beginning of this interval follows oceanographic evidence for erably greater than that which could be produced by diagen- the intensification of the Asian monsoon ca. 7-8 Ma (mOON esis (ca. 2%; e.g., MEYERS, 1994); hence, the isotopic shift et al.. 1991; PRELL et al., 1992). must represent a change in the source of OC. Mixing of carbon from terrestrial C3 plants with OC derived from marine pri- METHODS AND RESULTS mary production fb’-‘C N --..20%) cannot account for values as high as - 15Z0, and in any case its relative abundance in Total Organic Carbon has been analyzed on whole-rock sediments. Bengal Fan sediments is low (BERTRAND et al., 199 I; MEYERS Data are available upon request to the authors or in NOAA data and DICKENS, 1992; COCWRAN et al., 1989; POYNTER and base. Samples were first acidified with 2 N HCI at 100°C to eliminate carbonate, rinsed, collected on a silica filter, and dried at 80°C. The EGLINTON, 1990). Thus, it is unlikely that most of the vari- sample on its filter was then oxidized under vacuum at 950°C with ation in SL3C is caused by the presence of variable amounts CuO/CuZO. The CO* released was purified at - 140°C. manomet~c~ly marine OC. Rather, we interpret these variations as primarily 6°C of OC in the Bengal Fan 4811

resulting from changing mixing ratios of terrigenous C3 and CCderived OC. Between 7 Ma and cu. 0.9 Ma, OC burial in I-- the fan appears to have been dominated by material derived from C4 biomass. In agreement with paleosol studies in the Himalayan foreland (QUADE and CERLING, 1994; QUADE, 6 et al., 1989) we find no evidence for a significant C4 com- ponent in Bengal Fan OC prior to 7 Ma. The timing and =8 magnitude of the 7 Ma 613Cshift in both carbonates and OC 5 in the Siwalik paleosols is very similar to that observed in al 10 the Bengal Fan OC. This remarkable agreement implies that the expansion of C4 plants was widespread and rapid across $12 the G-B basin. 0 14 0 DISCUSSION 16 l 0 Distribution of C3 and C4 Plants 18 l 613C of TOC recorded in the Bengal Fan reflects the C3/ 20 C4 abundance in the Himalaya-Ganga-Brahmaputra system. -27 -24 -21 -18 -15 However, in order to interpret the observed variations cor- rectly, we must first consider the processes of carbon export 6’3C of TOC(%o) and transport. Net organic carbon export from a variety of ecosystems is roughly proportional to net primary produc- FIG. 2. 613C of TOC in Bengal Fan sediments plotted against age in Holes 7 I7C and 7 18B. Data are available upon request to the tivity (KAPLAN and NEWBOLD, 1993) and productivity of authors or in NOAA data base. The increase of 6°C values at cu. 7 C4 plants is known to be significantly higher than that of C3 Ma reflects the increase of the C4/C3 plant ratio in the source of the plants (JONES, 1992). Therefore, the C4/C3 ratio recorded in organic matter. The 6°C increase near 7 Ma closely parallels that Bengal Fan OC can be amplified with respect to the ratio of observed in paleosol studies in the Himalayan foreland (e.g., CERLING, et al., 1993), but the decrease after 0.9 Ma is not observed in the C4 to C3 plants in the whole catchment area. 6’% of Bengal paleosol data. Variations in 6°C are correlated with variations in OC after 7 Ma is variable and is closely related to the other sedimentation rate, grain size and clay mineralogy (Fig. I). sedimentological variables such as TOC and water content

-15 i I

Y A 1

AA IrA

A A AA

AAA AAA 0 0 0 A 0 . A -24 .t 0

0 1 2 0 5 10 15 20 Total Organic Carbon (wt%) H20+ (wt%)

FIG. 3. Variations of 613Cof TOC with (a) TOC content and (b) HzO+ content for sediments of Holes 717C and 7 l8B. HrO+ is calculated from loss on ignition of whole rock analysis corrected for carbonate and TOC contents, after drying at 110°C for 24 h. Symbols: 0 samples > 7 Ma; 0 = 7 to 0.9 Ma; A < 0.9 Ma. In samples < 7 Ma, the relationships between b”C, TOC and H20+ suggest that Bengal Fan OC contains a mixture of C3derived material associated with coarser, unweathered sediments and C4derived carbon associated with weathered, fine-pained sediments. 4812 C. France-Lanord and L. A. Deny of the sediments (Fig. 3). The increase of TOC in sediments that the close coupling of the shifts in both vegetation and younger than 7 Ma (Fig. 1) likely results from the decrease weathering in the Himalayan foreland can only be reasonably of the mean grain size (and increase of specific surface area) explained by climate change. The marked change in sediment of the particles delivered to the fan, as has recently been shown supply beginning near 0.9 Ma correlates with a significant for other marine elastic sediments (KIX et al., 1994). Whole- cooling of global climate and ice volume expansion, and the rock water contents (I&O’) reflect the relative proportion of beginning of the strong 100 kyr periodicity in the marine hydrous supergene minerals (mainly smectite and vermicu- &I80 record (RUDDIMAN et al., 1986). The corresponding lite) derived via weathering in the Indo-Gangetic Plain IX decrease in 6°C in Bengal Fan OC could indicate lower less-hydrated minerais produced by physical erosion in the abundance or pr~uctivity for C4 plants in the G-B basin Himalaya (quartz, feldspars, micas, illite; FRANCE-LANORD precipitated by a cooler and/or wetter climate. However, the et al., 1993). The relation between 613C and H,O+ implies relationship between sediment source and 613C for post-O.9 that the secondary mineral assemblage is associated with OC Ma samples (Fig. 3) suggests that the light carbon was pri- derived from a C4dominated ecosystem, while the “primary” marily derived from higher in the G-B drainage system. This mineral assemblage carries carbon from a C3-dominant eco- relationship further suggests that the hydrology and transport system. Thus, the relationship for post-7 Ma samples shown dynamics of the G-B river system exert first-order control in Fig. 3 can be interpreted as a mixing line between coarse. over the type of OC exported to the oceans after 7 Ma. relatively unaltered sediments associated with C3 carbon, and fine-grained, weathered sediments carrying mostly C4 carbon. Implications for the Marine and Sedimentary Carbon The 613C variability of Bengal Fan OC contrasts with the Isotope Budget relatively constant 613C of Siwalik paleosol carbonate and OC through this interval (QUADE et al., 1989; CERLING et The present-day Ganga, Brahmaputra, and Indus rivers al., 1993; QUADE and CERLING, 1994). This suggests that, carry 3 to 5% of the river dissolved inorganic carbon (DIC) after 7 Ma, OC 613C values were primarily controlled by flux (MEYBECK, 1979), and >lO% of the river TOC flux changes in the source of particulate organic matter rather (SUBRAMANIANand ITTEKKOT, 1991). Although the Bengal than by C3/C4 change over the whole source region. The Fan records only refractory PQC from the G-B river system, coarser-grained, less weathered sediments, which we interpret it is likely that DOC and labile POC were isotopically similar to have been derived more or less directly from the Himalayan (In the Amazon, e.g., G13Cpoc- S’3Cr.oeusually _12%0,QUAY hinterland during intervals of rapid transport, may have de- et al., 1992). Because the G-B river system is such an im- rived more of their OC from C3-dominated forests upstream. portant source of OC, it is thus very likely that the mean Sediments which became incorporated into soils in the C4- global d13Cof the riverine TOC input to the ocean increased dominated foreland and were weathered during intervals of significantly cu. 7 Ma. The mean 613Cof the inorganic carbon slower transport carry C4 carbon. We propose that, once C4 input to the ocean was probably similarly affected. A large plants became established in the G-B basin at 7 Ma, S’“C fraction of DIC in rivers is derived from respiration of organic variations in Bengal OC result from changes in the dynamics matter, which results in river DIC with low Si3C values relative of the river transport system leading to changes in the source to the atmosphere (HITTHON and KROUSE, 1972: LONGI- region of rivet-me POC. Essentially, under different hydrologic NELLIand EDMOND, 1983). Since isotopically heavy carbon conditions the rivers appear to “sample” different parts of derived from C4 plants began to contribute significantly to the drainage basin. For example, two mid-Pliocene samples the carbon load of the G-B rivers at ca. 7 Ma, the mean 613C show a brief return to low 6°C values indicative of increased of marine DIG should also have increased. A 5%0increase in C3 plant carbon. These samples are illite-rich silts. associated the d”C of 10% of the global river TOC flux would be suf- with an interval of increased sedimentation rate and grain ficient to change the mean 613C of marine bicarbonate by size. The post-O.9 Ma shift to lower 613C values is also as- +0.5%0. The 613C record of marine carbonates is complex sociated with a return to rapidly deposited, coarse, illite-rich during this interval, and not all records agree (BROECKER sedimentation. and WOC)DRUFF,1992). We note that 613C of bulk marine The increase in the ratio of chemical weathering to physical carbonate, which has been proposed to be the best estimate erosion appears to be synchronous with the expansion of C4 of the global ocean carbon budget, increases by ca. OS%0 biomass recorded both in the Bengal Fan and the Siwalik near 7 Ma (SHACKLETONand HALL, 1984), as do several foreland. Oceanographic evidence indicates a more pro- foraminiferal records from both the Atlantic and Pacific nounced monsoonal climate in the region beginning 7-8 Ma (WRIGHT et al., 1991). (KROONet al., 1991; PRELLet al., 1992). Increased seasonality The expansion of C4 photosynthetic plants in the terrestrial should favour C4 plants, and the expansion of C4 vegetation environment has implications for interpreting the Late Cen- in the Siwaliks could be a consequence of the development ozoic carbon isotope record as well. As noted above, the of the monsoon (QUADE et al., 1989). More recent work has Neogene OC burial flux in the Bengal and Indus Fan, asso- raised the possibility that the late Miocene expansion of C4 ciated deltas and foreland basins probably represents 2 15% plants was a global event, perhaps driven by falling atmo- of the global total. From cu. 7 to 1 Ma. the OC buried in the spheric CO* levels (CERLINGand QUADE, 1993; EHLERINGER Bengal Fan was 7-lo%0 less depleted in 13Cthan it had been et al., 199 I), but African evidence (MORGAN et al., 1994; prior to 7 Ma. Consequently, the global r3C depletion of sed- KINGSTON et al., 1994) suggests a diachronous history. Sim- imentary organic carbon relative to carbonate, b”CcZrt, ilarly, changes in the weathering regime could reflect either - b’3Corg= AB (HAYESet al., 1989), must have decreased by climatic or tectonic control (BURBANKet al., 1993). We argue l-2%0. If C4 plant-delved carbon was an important com- d13Cof OC in the Bengal Fan 4813 ponent of OC burial in other regions (CERLINGet al., 1993), evidence from the Bengal Fan. In Himalayan Tectonics (ed. P. J. this effect would have been correspondingly greater. Carbon TRELOARand M. SEARLE),Vol. 74, pp. 603-62 I. Geol. Sot. Lond. isotope depletion produced by marine algae appears to have GARTNERS. (1990) Neogene calcareous nannofossil biostratigraphy, Leg I 16 (Central Indian Ocean). In Proc. ODP Sci. Res. (ed. J. R. decreased by 5-7’3~0,mostly by early Neogene time (POPP et COCHRANet al.), Vol. 116, pp. 165-187. Ocean Drilling Program. al., 1989). The value of AB at any given time depends on the HAYESJ. M., POPP B. N., TAKIGIKUR., and JOHNSONM. W. (1989) relative proportions of marine and terrestrial organic carbon An isotopic study of biogeochemical relationships between car- burial as well as the depletion associated with each. Because bonates and organic carbon in the Greenhorn Formation. Geochim. of the large changes in carbon isotope depletion associated Cosmochim. Acta 53,296 I-2972. HITCHONB. and KROUSEH. R. (1972) Hydrogeochemistry of the with photosynthesis in both the marine and terrestrial habitats surface waters of the Mackenzie River drainage basin, Canada- models of the Cenozoic sedimentary carbon isotope mass III. Stable isotopes of oxygen, carbon and sulfur. Geochim. Cos- balance must take these effects into account (KUMP, 1989; mochim. Acta 36, 1337-1357. DERRY, 1993). ITTEKKOTV. (1988) Global trends in the nature of organic matter in river suspensions. Nature 332,436-438. JONESH. G. ( 1992) Plants and Microclimate: A Quantitative Approach Acknowledgments-We thank Todd Dawson and Jay Quade for to Environmental Plant Physiology. Cambridge Univ. Press. helpful discussions. This study was carried out under the auspices of KAPLANL. A. and NEWBOLDJ. D. (1993) Biogeochemistry of dis- the CNRS program “Dynamique et Bilan de la Terre-erosion et solved organic carbon entering streams. In Aquatic Microbiology: HydrogCologie”. CNRS-INSU-DBT contribution #7 15 and CRPG- An Ecological Approach (ed. T. FORD), pp. 139-165. Blackwell. CNRS contribution # 1069. KEIL R. G., TSAMAKISE., BOR FUH C. B., GIDDINGS J. C., and HEDGESJ. I. (1994) Mineralogical and textural controls on the Editorial handling: J. D. Macdougall organic composition of coastal marine sediments: Hydrodynamic separation using SPLITT-fractionation. Geochim. Cosmochim. Acta 58, 879-894. REFERENCES KINGSTONJ. D., MARINOB. D., and HILLA. (1994) Isotopic evidence BERCGRENW. A., KENT D. V., FLYNNJ. J., and VAN COUVERING for Neogene hominid paleoenvironments in the Kenya Rift Valley. J. A. ( 1985) Cenozoic geochronology. GSA Bull. 96, 1407- I4 18. Science 264,955-959. BERTRANDP., LALLIER-VERGESE., and GRALL H. (199 I) Organic KROOND., STEENST., and TROELSTRAS. R. (1991) Onset of mon- petrology of Neogene sediments from North Indian Ocean (Leg soonal related upwelling in the western Arabian Sea as revealed 117): Amount, type, and preservation of organic matter. In Proc. by planktonic foraminifers. In Proc. ODP Sci. Res. (ed. W. L. ODP Sci. Rex (ed. W. L. PRELL et al.), Vol. I 17, pp. 587-594. PRELL et al.), Vol. I 17, pp. 257-263. Ocean Drilling Program. KUMP L. R. (1989) Alternative modeling approaches to the geo- BOUQU~LLONA. (1987) Influence continentales et marines dans les chemical cycles of carbon, sulfur, and strontium isotopes. Amer. skdiments C&ozoiques de l’ocian Indien Nord Oriental. Thesis, J. Sri. 289, 390-4 IO. UniversitC de Lille Flandres-Artois. LASAGAA. C., BERNERR. A., and GARRELSR. M. (1985) An im- BOUQUILLONA., FRANCE-LANORDC., MICHARDA., and TIERCELIN proved model of atmospheric CO2 fluctuations over the past 100 J.-J. (1990) Sedimentology and isotopic chemistry of the Bengal million years. In The Carbon Cycle and Atmospheric C02: Natural fan sediments: the denudation of the Himalaya. In Proc. ODPSci. Variations Archean to Present (ed. E. T. SUNDQUISTand W. S. Res. (ed. J. R. COCHRANet al.), Vol. 116, pp. 43-58. BROECKER),pp. 397-4 IO. Amer. Geophys. Union. BROECKERW. S. and WOODRUFFF. (1992) Discrepancies in the LINDSAYJ. F., HOLLIDAYD. W., and HULBERTA. G. (1991) Se- oceanic carbon isotope record for the last fifteen million years? quence Stratigraphy and the Evolution of the Ganges-Brahmaputra Geochim. Cosmochim. Acta 56, 3259-3264. Delta Complex. AAPG Bull. 75, 1233- 1254. BURBANKD. W., DERRY L. A., and FRANCE-LANORDC. (1993) LONGINELLIA. and EDMONDJ. M. (1983) Isotope geochemistry of Reduced Himalayan sediment production 8 Myr ago despite an the Amazon basin: a reconnaissance. J. Geophys. Res. 88, 3703- intensified monsoon. Nature 364,48-50. 3717. CERLINGT. E. and QUADEJ. (1993) Stable carbon and oxygen iso- MEYBECKM. (1979) Concentrations des eaux fluviales en Cltments topes in soil carbonates. In Climate Change in Continental Isotopic majeurs et apports en solution aux octans. Rev. G.&o/.Dyn. Gkogr. Records (ed. P. K. SWARTet al.), Vol. 78, pp. 217-231. AGU. Phys. 21, 220-227. CERLINGT. E., WANG Y., and QUADE J. (1993) Expansion of C4 MEYERSP. A. (1994) Preservation of elemental and isotopic source ecosystems as an indicator of global ecological change in the late identification of sedimentary organic matter. Chem. Geol. 114, Miocene. Nature 361, 344-345. 289-302. COCHRANJ. R. (I 990) Himalayan uplift, sea level, and the record MEYERSP. A. and DICKENSG. R. (I 992) Accumulation of organic of Bengal fan sedimentation at the ODP Leg I I6 Sites. In Proc. Matter in sediments of the Indian Ocean: A synthesis of results ODPSci. Res. (ed. J. R. COCHRANet al.), Vol. 116, pp. 397-414. from scientific deep sea drilling. In Synthesis of Results from Sci- Ocean Drilling Program. en&@ Drilling in the Indian Ocean, Geophysical Monograph (ed. C~CHRANJ. R. et al. (1989) Proc. ODP Init. Rep&, Vol. 116. College R. A. DUNCAN et al.), Vol. 70, pp. 295-309. Amer. Geophys. Ocean Drilling Program. Union. CURRAYJ. R. (1991) Postcollision sediments in the Bay of Bengal. MILLIMANJ. D. and SYVITSKIP. M. (1992) Geomorphic/Tectonic Eos 72, 250. control of sediment discharge to the ocean: the importance of small CURRAY J. R. and MOORED. G. (197 1) Growth of the Bengal Deep mountainous rivers. J. Geol. 100,525-544. Sea Fan and denudation in the Himalayas. GSA Bull. 82, 563- MORGANM. E., KINGSTONJ. D., and MARINOB. (1994) Carbon 572. isotopic evidence for emergence of C4 plants in the Neogene from DERRYL. A. (1993) Interpreting the Neogene d13Crecord: the effect Pakistan and Kenya. Nature 367, 162- 165. of variable isotopic fractionation. Eos 74, 328 (abstr.). POPP B. N., TAKIGIKUR., HAYESJ. M., LOUDAJ. W., and BAKER EHLER~NGERJ. R., SAGE R. F., FLANAGANL. B., and PEARCY E. W. (1989) The post-Paleozoic chronology and mechanism of R. W. (1991) Climate change and the evolution of C4 photosyn- 13C depletion in primary marine organic matter. Amer. J. Sci. thesis. Trends Ecol. Evol. 6, 95-99. 289,436-454. ERDMANJ. G., SCHRONOK. S., and SCALANR. S. (1974) Geochem- POYNTERJ. and EGLINTONG. (1990) Molecular composition of istry of carbon: Legs 22,24, 26,27, and 28. In Init. Repts. DSDP, three sediments from Hole 717~: the Bengal Fan. In ?roc. ODP Vol. 24, pp. I 169-I 176. US Govt. Printing Office. Sci. Res. (ed. J. R. COCHRANet al.). Vol. 116. DD.. . 155-161. Ocean FRANCE-LANORDC., DERRYL., and MICHARDA. (1993) Evolution Drilling Program. of the Himalaya since Miocene time: isotopic and sedimentologic PRELLW. L., MURRAYD. W., and CLEMENSS. C. (1992) Evolution 4814 C. France-Lanord and L. A. Derry

and variability of the Indian Ocean Summer Monsoon: Evidence Drilling in the Indian Ocean, tieophysical Monograph (ed. R. A. from the Western Arabian Sea Drilling Program. In Synthesis qf DUNCANet al.), Vol. 70, pp. 387-402. Amer. Geophys. Union. Resultsfrom ScientiJc Drilling in the Indian Ocean, Geophysical RUDDIMANW. F., SHACKLETONN. J., and MCINTYREA. (1986) Monograph (ed. R. A. DUNCANet al.), Vol. 70, pp. 447-469. Amer. North Atlantic sea-surface temperature for the last I. I million years. Geophys. Union. Geological Society Special Publication 21, 15% 173. QUADEJ. and CERLINGT. E. ( 1994) Expansion of C4 grasses in the SHACKLETONS. J. and HALLM. A. (1984) Carbon isotope data from Late Miocene of Northern Pakistan: Evidence from stable isotopes Leg 74 sediments. In Initial Reports qfthe Deep Sea Drilling Pro- in paleosols. Paleo. Pale0 Paleo. gram (ed. T. C. MOOREet al.), Vol. 74, pp. 613-619. US Govt. QUADEJ., CERLINGT. E., and BOWMANJ. R. (1989) Development Printing Office. of Asian monsoon revealed by marked ecological shift during the SMITHS. V. and HOLLIBAUGHJ. T. (1993) Coastal metabolism and latest Miocene in northern Pakistan. Nature 342, 163-166. the oceanic organic carbon balance. Rev. Geophys. 31, 75-89. QUADE J. et al. (1992) A l6-Ma record of paleodiet using carbon STOWD. A. V., et al. (I 990) Sediment facies and processes on the and oxygen isotopes in fossil teeth from Pakistan. Chem. Geol. 94, distal Bengal Fan, Leg 116. In Proc. ODP Sci. Rex (ed. J. R. 183-192. COCHRANet al.). Vol. 116, pp. 377-396. Ocean Drilling Program. QUAY P. D., WILBUR D. O., and RICHEYJ. E. (I 992) Carbon cycling SUBRAMANIANV. and ITTEKKOT V. ( 1991) Carbon transport by the in the Amazon River: implications from the 13Ccompositions of Himalayan rivers. In Biogeochemistry of Major World Rivers. particles and solutes. Limnol. Oceanog. 37, 857-87 I. SCOPE (ed. E. T. DEGENSet al.), Vol. 42, pp. 157-168. Wiley. REA D. K. (1992) Delivery of Himalayan sediment to the Northern WRIGHTJ. D.. MILLERK. G., and FAIRBANKSR. G. (1991) Evolution Indian Ocean and its relation to global climate, sea level, uplift. of modern deepwater circulation: evidence from the Late Miocene and seawater strontium. In Synthesis of Results /km Scient$c southern ocean. Paleoceanography 6,275-290.