Sodium shortage as a constraint on the carbon cycle in an inland tropical rainforest Michael Kasparia,b,1, Stephen P. Yanoviakc, Robert Dudleyd, May Yuane, and Natalie A. Claya Departments of aZoology and eGeography, Graduate Program in Ecology and Evolutionary Biology, University of Oklahoma, Norman, OK 73019; bSmithsonian Tropical Research Institute, Apartado 2072, Balboa Panama; cDepartment of Biology, University of Arkansas, Little Rock, AR 72204; and dDepartment of Integrative Biology, University of California, Berkeley, CA 94720 Edited by Gordon H. Orians, University of Washington, Seattle, WA, and approved September 29, 2009 (received for review June 9, 2009) Sodium (Na) is uncommon in plants but essential to themetabolism of plant consumers, both decomposers and herbivores. One con- sequence, previously unexplored, is that as Na supplies decrease (e.g., from coastal to inland forests), ecosystem carbon should accumulate as detritus. Here, we show that adding NaCl solution to the leaf litter of an inland Amazon forest enhanced mass loss by 41%, decreased lignin concentrations by 7%, and enhanced de- composition of pure cellulose by up to 50%, comparedwith stream water alone. These effects emerged after 13?18 days. Termites, a common decomposer, increased 7-fold onNaCl plots, suggesting an agent for the litter loss. Ants, a common predator, increased 2-fold, suggesting that NaCl effects cascade upward through the food web. Sodium, not chloride, was likely the driver of these patterns for two reasons: two compounds of Na (NaCl and NaPO4) resulted in equivalent cellulose loss, and ants in choice experiments underused Cl (as KCl, MgCl2, and CaCl2) relative to NaCl and three other Na compounds (NaNO3, Na3PO4, and Na2SO4). We provide experimental evidence that Na shortage slows the carbon cycle. Because 80% of global landmass lies >100 km inland, carbon stocks and consumer activity may frequently be regulated via Na limitation. biogeochemistry  biogeography  decomposition  fungi  termites Models of the carbon cycle start with the coavailability ofwater and solar energy as key constraints to photosynthesis and respiration (1?4). Lowland tropical forests, however, have ample sunshine, often ample moisture (5), and, frequently, weathered soils (6). In such forests, nutrient shortages, most notably the shortage of P, have been shown by experiment and comparative study to constrain both net primary productivity (NPP; g C/m2 per y) and decomposition (7?13). Moreover, recent theory and experiment have suggested that the rate of decomposition, a collaborative process involving thousands of species, is unlikely to be constrained, Liebig style, by a single element (14, 15). Here, we build on those studies and those of Chadwick et al. (16) to suggest that Na plays a key role in regulating decomposition in inland ecosystems. Of the 25 or so elements required for life (17), Na is unique. Most terrestrial plants have little need of Na (18). Herbivores and decomposers, in contrast, must amass Na in concentrations 100- to 1,000-fold over the plants they consume (19). In animals, costly sodium pumps maintain gradients of cell concentration and membrane voltage (1% deviations of whole body sodium are signs of pathology; ref. 17). In plants, K, not Na, performs this function (20). This biogeochemical disconnect between plants and those that eat them suggests that consumers, but not plants, should suffer when Na inputs to ecosystems decline. Decomposers metabolize 90% of terrestrial plant biomass (21). Decomposition, the breakdown of necromass into CO2 and inorganic compounds, is largely performed by microbes. Some are free living (e.g., basidiomycete and ascomycete fungi), and others live in the guts of animal decomposers (detritivores) like termites and isopods. Both fungi and animals have high Na requirements relative to plants they consume (22, 23). Both obtain energy by breaking down abundant carbon-rich macro- molecules like cellulose. Yet little is known about how the activity of an ecosystem?s decomposers is constrained by the shortage of sodium. Sodium has a geography. It accumulates locally where ground water is mined for agriculture (19) and where roads are salted to prevent icing (24). Regionally, extensive rainfall promotes leach- ing (25), and the Na content of that rainfall decreases exponen- tially as one travels inland from sources of oceanic aerosols (26, 27). This decline in aerosol deposition has consequences for ecosystem levels of sodium: coastal forests in Panama have higher concentrations of Na than the Peruvian Amazon both in freshwater streams and rivers [149 vs. 20?60 mol/L (26, 28)] and soil [0.008 vs. 0.005 ppm (29)]. Consistent with the hypoth- esis of Na limitation, here we show that litter decomposition rates, and the abundance of decomposers and their predators, increase with NaCl fertilization in an inland Amazon forest. Results We tested the hypothesis that Na shortage limits decomposers in a forest near Iquitos Peru, 2,000 km from its ocean source of aerosols (26). The first experiment compared decomposition and decomposer densities across 35 paired 0.5  0.5-m plots through old-growth forest. One of each pair was fertilized with a 0.5% solution of NaCl and stream water every other day, whereas the other received only stream water. After 18 days litter concentrations of Na were 3-fold higher on NaCl plots (paired t329.1, P 0.0001; Fig. 1).* Litter volume onNaCl plots, in contrast, was 71% lower than controls (paired t34 3.99, P  0.0003; Fig. 1); the concentration of recalcitrant lignin was 7% lower (paired t32  9.1, P  0.0001; Fig. 1). There was marked increase of termites on NaCl plots (Fig. 1). Termites that nest in high densities within the forest f loor (30) emerged from the soil and were observed consuming litter. Systematic scans over the course of the experiment revealed termites on the surface of six NaCl plots but no control plots (Kruskal Wallis X2  6.5, P  0.011). At harvest, 7-fold more termites were extracted from NaCl plots than controls (paired t344.13, P 0.0002; Fig. 1). Monitoring revealed 8-foldmore ants on NaCl plots (paired t34  3.8, P  0.0006) and 2-fold more ants when the plots were harvested (paired t342.2, P 0.03; Fig. 1). Thus, both a key decomposer and a key predator of tropical brown food webs accumulated on NaCl plots. A second experiment confirmed higher mass loss of cellulose with increased access to Na (Fig. 2). After 13 days, mass loss varied across the landscape (F19,60  12.6, P  0.0001) and was 53% higher on NaCl-treated cellulose than water-treated con- trols; mass loss of Na3PO4-treated cellulose was 29% higher (F2,60 10.3, P 0.0003, Tukey minimum significant difference Author contributions: M.K., S.P.Y., R.D., and N.A.C. designed research; M.K., S.P.Y., and N.A.C. performed research; M.K. and M.Y. analyzed data; and M.K. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1To whom correspondence should be addressed. E-mail: mkaspari@ou.edu. *The litter chemistryof twoNaClplotswasunknownas termiteshadeatenall of the litter. www.pnas.orgcgidoi10.1073pnas.0906448106 PNAS Early Edition  1 of 5 EC O LO G Y EN V IR O N M EN TA L SC IE N CE S test: NaCl  Na3PO4  H2O). Cellulose disappeared from the forest f loor more quickly in the presence of Na, whether coupled with chloride or phosphate. A third experiment verified that the Na in NaCl was the primary attractant to ants. Vials with solutions of seven com- pounds (plus urine) were presented along five transects across the forest f loor (Fig. 3). Both solution type (F7,39  9.22, P  0.0001) and transect (F4,39  7.47, P  0.0003) accounted for variation in ant use of vials. Ants used all but the lowest-ranked Na solution (NaNO3), more than three solutions with Cl but no Na (KCl, MgCl2, and CaCl2; Fig. 3). Urine is likely the most common Na supplement to the brown food web, attracting a variety of arthropods. NaCl and urine vials received similar numbers of visits. Discussion In tropical rainforests, where solar energy and precipitation are often in ample supply, biogeochemical gradients of elements like phosphorus can limit both NPP and decomposition (11) and may account for the deeper litter in the Amazon interior compared with more coastal forests (13). Here, we provide experimental evidence that shortage of another element, sodium, slows the degradation of cellulose and lignin and promotes carbon storage, all with likely no direct impact on the construction of new plant tissue. Sodium has long been suspected of regulating inland popu- lations of herbivores, from North American meadow voles (31) to elephants of the African savannah (32). We found that termites, which occupy 68% of Earth?s land area representing 77% of terrestrial NPP (33, 34), increased 7-fold on plots that Fig. 1. Effects of Na treatments on soil salinity, litter decomposition (volume and litter content), and the abundance of two common litter taxa in an Amazon forest. Lines from left to right denote differences between control andNaCl treatment in 35 paired plots. Circles are treatmentmeans: black control (stream water), white  0.5% NaCl solution made from stream water. Fig. 2. Average mass loss of cellulose treated with stream water versus two forms of sodium (least square means 2 SEMs). Circles are treatment means: black control (stream water), white 0.1 M Na solution made from stream water. Fig. 3. Ant use (least square means percentage of 15 baits SEMs) of eight solutions offered along five transects across the forest floor. Gray bars repre- sent higher (A) and lower (B) use of baits using Tukey criteria of minimum differences (P  0.05). 2 of 5  www.pnas.orgcgidoi10.1073pnas.0906448106 Kaspari et al. averaged 3-fold more Na. Termites transform carbon from soil, leaf litter, and coarse woody debris (34) into significant quan- tities of methane and carbon dioxide (33). They also apparently crave sodium: termites of an African savannah 250 km inland increase Na concentrations 6 m down into the soil (35); pest control agents in North America are taught to add Na-rich sports drinks to termite baits (Edward Vargo, personal communica- tion). Our experimental evidence shows behavior consistent with Na limitation (i.e., increased densities and activity with added Na) in an animal decomposer (i.e., detritivore). Furthermore, the concentration of lignin, a recalcitrant mac- romolecule that inhibits cellulolysis, also decreased on NaCl plots, implicating the action of powerful peroxidase systems used by a subset of free-living microbes and termite symbionts (34, 36?38). At present, we do not have strong evidence that fungal activity increased on NaCl plots. However, in this inland Amazon forest we observed none of the dense mycelial mats characteristic of coastal tropical forests (39), suggesting some form of fungal inhibition. Decomposition is the conversion of detritus to microbial and animal food (29) that in turn supports a large fraction of the forest fauna (40). The increase of ants on NaCl plots suggests that ant populations in the inland Amazon are Na-limited directly (through metabolic requirements) and/or indirectly via the increase in prey on Na-rich patches. Our observations suggest both are true. The gradient of increasing Na preference from coastal to inland ant communities (41), combined with intense recruitment in this study to Na (but not Cl) by herbiv- orous ants likeCephalotes atratus, supports the role of Na deficits in directly influencing ant activity. Moreover, predators like the trap-jaw ants of the genera Pyramica and Odontomachus, which are less constrained by Na deficits because of their high-sodium diets (41), also increased on the NaCl plots. The increased decomposer activity brought about by added Na cascaded up the brown food web. In forest ecosystems, leaf litter and coarse woody debris can represent a significant carbon sink (42, 43) although much uncertainty remains regarding the dynamics and standing stocks of both (44). If the local experiments reported here scale up, then as one moves inland away from ocean inputs of sodium (16) stores of leaf litter and coarse woody debris should increase, and the release of CO2 from forests should decrease. The resulting imprint on the global carbon cycle would not be trivial as 80% of Earth?s terrestrial surface is 100 km inland (the distance where ant communities begin to show signs of Na deficit; Table 1 and ref. 41). The Earth?s vast inland boreal forests (45) and fire-maintained grasslands (46) might also be particular candi- dates for Na limitation of decomposers. Furthermore, road salts (which collectively exceed oceanic inputs of Na by 4-fold in the United States; ref. 24) and hurricanes (that can transport the equivalent of 0.3 mm of seawater inland; ref. 47)? may leave their own Na-based imprint on consumer communities and carbon stocks. ?Murphy SF, Stallard RF, The Third Interagency Conference on Research in the Watersheds, September 8?11, 2008, Estes Park, CO. Table 1. Area of terrestrial landmass (km2) found at four zones representing distances to the ocean and potential oceanic aerosols Location Distance % 100 km inland10 km 10?100 km 100?1,000 km  1,000 km Continent Africa 1.66E05 2.85E06 2.11E07 5.40E06 89.8 Asia 6.12E05 6.83E06 3.00E07 5.58E06 82.7 Australia 6.99E04 1.13E06 6.30E06 0.00E00 84.0 North America 9.67E05 6.40E06 1.47E07 6.09E05 67.5 Oceania 2.92E04 2.25E05 7.05E03 0.00E00 2.7 South America 1.42E05 1.87E06 1.23E07 3.14E06 88.5 Antarctica 6.79E05 3.74E06 7.81E06 0.00E00 63.9 Europe 3.19E05 2.71E06 6.31E06 0.00E00 67.6 Ecoregion domain Dry 3.87E05 3.92E06 3.33E07 9.12E06 90.8 Humid temperate 8.92E05 6.03E06 1.48E07 7.23E05 69.1 Humid tropical 9.26E05 7.08E06 2.60E07 4.75E06 79.4 Polar 2.40E06 1.14E07 2.44E07 1.30E05 64.1 Ecoregion divisions with largest inland regions Temperate desert regime mountains 0.00E00 0.00E00 6.15E04 5.53E05 100.0 Temperate steppe regime mountains 0.00E00 1.10E03 6.65E05 4.02E05 99.9 Temperate steppe 2.21E04 2.09E05 4.03E06 5.33E05 95.2 Tropical/subtropical desert regime mountains 2.55E04 2.06E05 1.59E06 1.37E06 92.7 Tropical/subtropical desert 1.32E05 1.35E06 1.32E07 2.58E06 91.4 Temperate desert 4.43E04 4.33E05 3.32E06 1.69E06 91.3 Prairie regime mountains 1.12E04 1.04E05 8.50E05 2.93E05 90.9 Prairie 3.61E04 4.03E05 3.82E06 1.67E05 90.1 Tropical/subtropical steppe regime mountains 4.87E04 4.68E05 3.33E06 6.96E05 88.6 Savanna 2.83E05 2.49E06 1.49E07 2.92E06 86.5 Tropical/subtropical steppe 1.15E05 1.25E06 7.16E06 1.29E06 86.1 Subtropical regime mountains 3.25E04 2.19E05 1.05E06 2.42E05 83.7 Sub-Arctic division 1.90E05 2.09E06 1.00E07 7.78E03 81.4 Data summarized for continents and ecoregions follow Bailey and Ropes (48). Coastlines were extracted from the Digital Chart of the World (54) and used to estimate area within five rings of distances from the sea at 5  5-km resolution. Kaspari et al. PNAS Early Edition  3 of 5 EC O LO G Y EN V IR O N M EN TA L SC IE N CE S Materials and Methods The experiments took place December 16, 2008 to January 6, 2009 at the Amazon Center for Tropical Studies (ACTS) field station 67 km NE of Iquitos, in Loreto Province, Peru (3.25?S, 72.91? W). ACTS is embedded in a humid tropical lowland forest of clay oxisols andultisols (49) that receives3,000mm of rainfall spread relatively evenly throughout the year (50). ACTSwas chosen as a likely candidate for decomposer limitation as it received low concentra- tions of Na in its rainfall (26) and among the highest use of NaCl baits in a 17-site survey of New World ant assemblages (41). In all three experiments we made solutions from the water of a small (2 m wide, 50 cm deep) low-flow, low-gradient, sandy-bottom tributary of the Rio Sucusari that abutted our study site. Analysis of this water with an inductively coupled plasma-optical emission spectrophotometer revealed a Na concentration of 81 mol/L. First Experiment. In the first experiment, NaCl solution was added directly to the litter on0.25-m2 (0.50.5m)plots. Thirty-five22-mblockswereflagged on either side of anACTS trail running along and 4mabove the stream. Blocks werearrayedonboth sidesof the trail. Blockswere separated3mfromthose on the same side of the trail and5 m from those on the opposite side of the trail. All blockswere located onflat expanses of open litter. Two 0.25-m2 plots were positioned in each block so as to avoid coarse woody debris (branches 10 cm in diameter). These plots were flagged and assigned to control or NaCl treatments. Beginning December 18, 2008 control plots received 250 mL of stream water on alternate days or 250 mL of 0.5% (by weight) NaCl solution dribbled slowly over the litter. Before watering, each plot was monitored for ants and termites on the surface of the litter. Abundance was classified on a log10 scale (i.e., 1  1?9 individuals, 2  10?99, 3  100?999). Plots were harvested to quantify litter volume, percentage of lignin, per- centage of NaCl, and the abundance of termites and ants. Five blocks were harvested on December 25; the other 30 blocks were harvested, in batches of five blocks per day, January 1?6. At harvest, litter depth was measured within each cornerofaplotby insertingawireflagpresseddownward tomineral soil. All litterwasharvesteddown tomineral soil and shaken for 30 s through1-cm2 mesh. The siftatewas stored in a cloth bag andplaced for 24 h in aminiwinkler (51) to extract invertebrates. This siftate was further hand-sorted to remove any remaining invertebrates2 mm. The samples were sorted in M.K.?s lab to quantify the abundance of ants and termites. To determine Na and fiber composition, sampleswere thendried at 50 ?C and a 10-g subsample analyzed by the Oklahoma State Soil, Water, and Forage Analytical Laboratory. Here, we report results for percentage of Na (via a Spectro CirOs ICP spectrometer) and percentage of lignin (estimated by digesting cellulose  lignin in 72% sulfuric acid). See ref. 52 for further details. Pairwise t tests were used to compare changes in litter volume, chemistry, and termite and ant abundance with NaCl. We pooled the five plots har- vested on December 25 with those harvested from January 1?6; thus we refer to harvest results after 18 days. Second Experiment. We also examined the effect of increased sodium on the decomposition of cellulose, the world?s most abundant biomolecule (21). Cellulose filter paper (9-cm disks, 100% cellulose; Fisherbrand) was saturated in one of two Na compounds (0.1 M Na3PO4 or 0.1 M NaCl made with stream water) or a control treatment (stream water). On December 23, 2008, disks were arranged in 20 replicate rows along a 40-m east?west transect in the forest understory 100 m NE of the ACTS station. Rows were established every 1?2 m along the transect; each included a control disk, a sodium phosphate disk, and a sodium chloride disk, 0.5 m apart. For each disk, a small patch of litter was cleared down to mineral soil, and the disk (folded twice to form a four-layered quarter circle) was secured to the soil with a surveyor flag and recoveredwith litter. Cellulose diskswere recovered after 13 days anddried at 50 ?C to stable mass. The percentage mass loss was used as an estimate of decomposition rate and analyzed by two-way ANOVA, with treatment and block effects. Third Experiment. We explored which chemical elements were attractive to consumers by presenting the local ant community with a choice of eight compounds in solution, patterned after a similar choice experiment in aNorth American butterfly population (53). In addition to NaCl, three solutions contained Cl (CaCl2, MgCl2, and KCl) and three contained Na (NaNO3, Na2SO4, andNa3PO4). All such solutionsweremade by dissolving compounds in stream water. The eighth solution was human urine contributed the morning of the experiment by S.P.Y. and N.A.C. (equal portions subsequently homogenized). Ants were offered these solutions in transects of labeled 2 mL of Eppendorf vials. Each vial was half-stuffed with cotton saturated with one of eight solutions. For each of five replicate transects, 15 vials of each solution were snapped shut and thoroughly mixed in plastic bag. Each transect was run on one of five different trails. Every 1 m along each transect, a vial was randomly selected from the bag, uncapped, and placed on the litter surface 1moff trail. We collected the vials after 1 h, snapping the cap shut and capturing the ants by using the baits within. For each transect, we calculated the percentage of vials for each solution visited by ants. These percentages were compared with ANOVA, using transect as a block. How Much of Continents Are Inland? To estimate the potential importance of sodium shortage in oceanic aerosols at the global scale, we quantified the fraction of the terrestrial surface at different distances from the ocean. We used the Digital Chart of the World (54), developed from the Operational Navigational Chart at 1:1,000,000 scale from the U.S. Defense Mapping Agency (reorganized into the National Geospatial-Intelligence Agency) and nationalmapping agencies fromAustralia, Canada, and the United Kingdom. ACKNOWLEDGMENTS. We thank P. Bucur, P. Jensen, and S. Madigosky for logistical support; the Peruvian Instituto Nacional de Recursos Naturales for permits; the Amazon Conservatory for Tropical Studies and Amazon Explorama Lodges for access to the field site; B. Nairn for analysis of streamwater; E. Vargo, B. Thorne, and R. Constantino for consultations on termite ecology; and J. Gillooly, J. Powers, andT.Valone for commentson themanuscript. Thisworkwas supported by a grant from the National Geographic Society. 1. Meetemeyer V (1978) Macroclimate and lignin control of litter decomposition rates. Ecology 59:465?472. 2. Gholz HL, Wedin DA, Smitherman SM, Harmon ME, Parton WJ (2000) Long-term dynamics of pine and hardwood litter in contrasting environments: Toward a global model of decomposition. Global Change Biol 6:751?765. 3. Allen AP, Gillooly JF, Brown JH (2005) Linking the global carbon cycle to individual metabolism. Funct Ecol 19:202?213. 4. Cox PM, Betts RA, Jones CD, Spall SA, Totterdell IJ (2000) Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature 408:184?187. 5. Phillips OL, et al. (2009) Drought sensitivity of the Amazon rainforest. Science 323:1344?1347. 6. Walker TW, Syers JK (1976) The fate of phosphorus during pedogenesis. Geoderma 15:1?19. 7. Elser JJ, et al. (2007) Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecol Lett 10:1135?1142. 8. Reich PB, Oleksyn J (2004) Global patterns of plant leaf N and P in relation to temperature and latitude. Proc Natl Acad Sci USA 101:11001?11006. 9. Wardle DA, Walker LR, Bardgett RD (2004) Ecosystem properties and forest decline in contrasting long-term chronosequences. Science 305:509?513. 10. Thingstad TF, et al. (2005) Nature of phosphorus limitation in the ultraoligotrophic eastern Mediterranean. Science 309:1068?1071. 11. Vitousek PM (2004) Nutrient Cycling and Limitation: Hawai?i as a Model System (Princeton Univ Press, Princeton). 12. Cleveland CC, Reed SC, Townsend AR (2006) Nutrient regulation of organic matter decomposition in a tropical rain forest. Ecology 87:492?503. 13. Kaspari M, Yanoviak S (2008) The biogeography of litter depth in tropical forests: Evaluating the phosphorus growth rate hypothesis. Funct Ecol 22:919?923. 14. Kaspari M, et al. (2008) Multiple nutrients regulate litterfall and decomposition in a tropical forest. Ecol Lett 11:35?43. 15. Saito MA, Goepfert TJ, Ritt JT (2008) Some thoughts on the concept of colimitation: Three definitions and the importance of bioavailability. Limnol Oceanogr 53:276?290. 16. Chadwick OA, Derry LA, Vitousek PM, Huebert BJ, Hedin LO (1999) Changing sources of nutrients during four millions years of ecosystem development. Nature 397:491?497. 17. Frausto da Silva JJR, Williams RJP (2001) The Biological Chemistry of the Elements: The Inorganic Chemistry of Life (Oxford Univ Press, Oxford). 18. Marschner H (1995) Mineral Nutrition in Higher Plants (Academic, San Diego). 19. National Research Council (2005) Mineral Tolerance of Animals (Natl Acad Press, Washington, DC). 20. Taiz L, Zeiger E (1998) Plant Physiology (Sinauer, Sunderland, MA), 2nd Ed. 21. Chapin FS, Matson PA, Mooney HA (2002) Principles of Ecosystem Ecology (Springer, New York). 22. Cromack FJ, et al. (1977) in The Role of Arthropods in Forest Ecosystems, ed Mattson WJ (Springer, New York), pp 78?84. 23. Stark N (1972) Nutrient cycling pathways and litter fungi. Bioscience 22:355?360. 24. Jackson RB, Jobbgy EG (2005) From icy roads to salty streams. Proc Natl Acad Sci USA 102:14487?14488. 25. Vitousek PM, Sanford RL (1986) Nutrient cycling inmoist tropical forest.Annu Rev Ecol Syst 17:137?167. 26. Stallard RF, Edmond JM (1981) Geochemistry of the Amazon 1. Precipitation chemistry and the marine contribution to the dissolved load at the time of peak discharge. J Geophys Res 86:9844?9858. 27. National Atmospheric Deposition Program (2006) Sodium Ion Wet Deposition (Cham- paign, IL). 4 of 5  www.pnas.orgcgidoi10.1073pnas.0906448106 Kaspari et al. 28. Stallard RF (2002) Comparative biogeochemistry of catchments with steep and gentle slopes, Barro Colorado Island, Panama. Tropical Forests, Past, Present, Future: Abstract Volume of TheAssociation for Tropical Biology AnnualMeeting (Smithsonian Tropical Research Institute, Balboa), p 111. 29. Kaspari M, Yanoviak SP (2009) Biogeochemistry and the structure of tropical brown food webs. Ecology, in press. 30. Vargo EL, Husseneder C (2009) Biology of subterranean termites: Insights from mo- lecular studies of Reticulitermes and Coptotermes. Annu Rev Entomol 54:379?403. 31. AumannGD, Emlen JT (1965) Relation of population density to sodiumavailability and sodium selection by microtine rodents. Nature 208:198?199. 32. Weir JS (1972) Spatial distribution of elephants in an African national park in relation to environmental sodium. Oikos 23:1?13. 33. Zimmerman PR, Greenberg JP, Wandiga SO, Crutzen PJ (1982) Termites: A potentially large source of atmospheric methane, carbon dioxide, and molecular hydrogen. Science 218:563?565. 34. Wood TG (1976) in The Role of Terrestrial and Aquatic Organisms in Decomposition Processes, eds Anderson JM, Macfayden A (Blackwell Scientific, Oxford), pp 145?168. 35. Watson JP (1962) The soil below a termite mound. Eur J Soil Sci 13:46?59. 36. Kirk TK, Farrell RL (1987) Enzymatic ??combustion??: The microbial degradation of lignin. Annu Rev Microbiol 41:465?501. 37. Kato K, Kozaki S, Sakuranaga M (1998) Degradation of lignin compounds by bacteria from termite guts. Biotechnol Lett 20:459?462. 38. Pasti MB, Pometto AL, Nuti MP, Crawford DL (1990) Lignin-solubilizing ability of actino- mycetes isolated from termite (Termitidae) gut. Appl Environ Microbiol 56:2213?2218. 39. Lodge DJ (1996) in The Food Web of a Tropical Rain Forest, eds Reagan DP, Waide RB (Univ Chicago Press, Chicago), pp 53?108. 40. Fittkau EJ, Klinge H (1973) On biomass and trophic structure of the central Amazonian rain forest ecosystem. Biotropica 5:2?14. 41. KaspariM, Yanoviak S, Dudley R (2008)On the biogeography of salt limitation: A study of ant communities. Proc Natl Acad Sci USA 105:17848?17851. 42. KellerM, PalaceM,AsnerGP, PereiraR, Silva JNM(2004) Coarsewoodydebris inundisturbed and logged forests in the eastern Brazilian Amazon. Global Change Biol 10:784?795. 43. Chambers JQ, Higuchi N, Schimel JP, Ferreira LV, Melack JM (2000) Decomposition and carbon cycling of dead trees in tropical forests of the central Amazon. Oecologia 122:380?388. 44. Clark DB, Clark DA, Brown S, Oberbauer SF, Veldkamp E (2002) Stocks and flows of coarse woody debris across a tropical rain forest nutrient and topography gradient. Forest Ecol Manage 164:237?248. 45. Dixon RK, et al. (1994) Carbon pools and flux of global forest ecosystems. Science 263:185?190. 46. Stephenson NL (1990) Climatic control of vegetation distribution: The role of water balance. Am Nat 135:649?670. 47. Wentz FJ, Ricciardulli L, Hilburn K, Mears C (2007) How much more rain will global warming bring? Science 317:233?235. 48. Bailey RG, Ropes L (1998) Ecoregions: The Ecosystem Geography of the Oceans and Continents (Springer, Berlin). 49. Holdridge LR, GrenkeWC, HathewayWH, Liang T, Tosi JAJ (1971) Forest Environments in Tropical Life Zones: A Pilot Study (Pergamon, New York). 50. Madigosky SR, Vatnick I (2000) Microclimatic characteristics of a primary tropical Amazonian rain forest, ACEER, Iquitos, Peru. Selbyana 21:165?172. 51. Agosti D, Majer JD, Alonso LE, Schultz TR (2000) Measuring and Monitoring Biological Diversity: Standard Methods for Ground-Living Ants (Smithsonian Press, Washington DC). 52. Oklahoma State University Soil, Water, and Forage Analytical Laboratory (2006) Avail- able at www.soiltesting.okstate.edu/. Accessed February 1, 2009. 53. ArmsK, FeenyP, LederhouseRC (1974) Sodium: Stimulus for puddlingbehavior by tiger swallowtail butterflies, Papilio glaucus. Science 185:372?374. 54. Department of Defense (1992) Digital Chart of the Worlds (Department of Defense, Washington, DC). Kaspari et al. PNAS Early Edition  5 of 5 EC O LO G Y EN V IR O N M EN TA L SC IE N CE S