Climatic and topographic controls on long-term erosion rates
This project links cosmogenic nuclide techniques and element depletion profiles to systematically explore how climate and topography affect rates of weathering and erosion.  Graduate student Cliff Riebe is using cosmogenic nuclides in stream sediment to measure long-term erosion rates at eight unglaciated Sierra Nevada sites, spanning an eight-fold range in annual precipitation and a range of 10°C in annual average temperature.  Each site includes a series of small subcatchments, with a five-fold range of average hillslope gradients.  With this experimental design, one can explore how weathering and erosion rates vary with topography under diverse climatic conditions.  Conversely, one can also explore how climate affects weathering and erosion in a range of topographic situations.

Results to date include the following:

 Long-term (10,000-20,000 year) erosion rates at typical hillslope gradients (30-40 percent) differ by only a factor of 2.5, across the diverse climates of our study sites, and do not vary systematically with either temperature or precipitation.

 Site-to-site climatic differences have much smaller effects on long-term rates of mineral weathering than conventional geochemical models would predict.
 At catchments with proximate tectonic forcing (from fault scarps and deeply incised river canyons), erosion rates increase exponentially with hillslope gradient.  At catchments that are isolated from local tectonic forcing, long-term erosion rates do not increase systematically with hillslope gradient, implying that the landscape is in dynamic equilibrium.  We hypothesize that rapidly eroding slopes become mantled with a lag deposit of granitic boulders, which inhibits further erosion beyond the boulder decomposition rate.

Collaborators:  Cliff Riebe (Ph.D. candidate, University of California), Darryl Granger (Purdue University) and Robert Finkel (Lawrence Livermore National Laboratory).

Related publications:
Granger, D.E., J.W. Kirchner, and R.C. Finkel, Quaternary downcutting rate of the New River, Virginia, measured from 26Al and 10Be in cave-deposited alluvium, Geology, 25, 107-110, 1997.
Granger, D.E., J.W. Kirchner, and R.C. Finkel, Spatially averaged long-term erosion rates measured from in situ cosmogenic nuclides in alluvial sediment, Journal of Geology, 104, 249-257, 1996.

Abstracts:
Riebe, C. S., J. W. Kirchner, D. E. Granger and R. C. Finkel, Tectonic control of erosion rates in the Sierra Nevada, California inferred from cosmogenic nuclide concentrations in alluvial sediment, Eos, Transactions, American Geophysical Union, 80, 1999.
Granger, D.E., C.S. Riebe, J.W. Kirchner, and R.C. Finkel, Erosional dynamic equilibrium in the Diamond Mountains, California, maintained by boulder armoring of hillslopes, Eos, Transactions, American Geophysical Union, 79, 1998.
Riebe, C.S. and J.W. Kirchner, Quantifying how topography, soil depth and bedrock erodibility affect long-term erosion rates, using cosmogenic nuclides in alluvial sediment, Eos, Transactions, American Geophysical Union, 78, F288, 1997.
Granger, D. E., J. W. Kirchner and C. S. Riebe, Inferring exhumation rates and processes from cosmogenic nuclides in sediments, GSA Abstracts with Programs, 1997.
Kirchner, J.W., D.E. Granger and C.S. Riebe, Cosmogenic isotope methods for measuring catchment erosion and weathering rates, Journal of Conference Abstracts, 2, 217, 1997.
Riebe, C.S., D.E. Granger and J.W. Kirchner, Quantifying effects of climate and topography on long-term erosion rates using cosmogenic nuclide concentrations in alluvial sediment, Eos, Transactions, American Geophysical Union, 77, F251, 1996.
Granger, D.E. and J.W. Kirchner, Downcutting rate of the New River, Virginia, from 26Al/10Be in buried river gravels, Eos, Transactions, American Geophysical Union, 76, F689, 1995.
Granger, D.E. and J.W. Kirchner, Erosional response to tectonic forcing inferred from cosmogenic isotopes in alluvial sediment, Eos Trans. Amer. Geophys. Union 75, 289, 1994.
Granger D.E., and Kirchner J.W.  Estimating catchment-wide denudation rates from cosmogenic isotope concentrations in alluvial sediment: Fort Sage Mountains, California, in Abstracts of the Eighth International Conference on Geochronology, Cosmochronology, and Isotope Geology (M. A. Lanphere, G. B. Dalrymple and B. D. Turrin, eds.), pp. 116. U.S. Geological Survey Circular 1107, 1994.
 

Comparing short-term and long-term sediment yields from forested terrain
By comparing cosmogenic nuclide measurements of sediment yield (which are averaged over 10,000-20,000 years) with direct measurements of sediment yield by sediment trapping or sediment gauging methods (which typically cover 1-30 years), we can clarify and quantify the mechanisms that control sediment yield over different timescales.

We sampled 30 catchments in the mountains of central Idaho, ranging from small experimental watersheds (0.2 km2) to the entire Salmon River (35,000 km2).  Our analyses show that across all 30 sites, long-term sediment yields are 4 to 40 times greater than sediment yields measured over the last 4-84 years.  Our results indicate that 70-97 percent of the sediment yield in this mountainous landscape comes from very large, episodic erosional events.  These events are too infrequent (recurrence intervals of roughly 50-200 years) to be captured by typical monitoring programs.  Our results imply that catastrophic events are a natural part of mountain erosion, subjecting aquatic ecosystems to episodic disturbance even in the absence of human interference.

Collaborators:  Cliff Riebe (Ph.D. candidate, University of California), Darryl Granger (Purdue University), Robert Finkel (Lawrence Livermore National Laboratory), Jim Clayton (U.S. Forest Service), Walt Megahan (National Council for Air and Stream Improvement).

Publications:
Kirchner, J. W., R. C. Finkel, C. S. Riebe, D. E. Granger, J. L. Clayton and W. F. Megahan, Episodic mountain erosion inferred from sediment yields over decades and millennia, Nature (in review).

Abstracts:
Kirchner, J.W., R.C. Finkel, C.S. Riebe, D.E. Granger, J.L. Clayton, and W.F. Megahan, Episodic erosion of the Idaho Batholith inferred from measurements over 10-year and 10,000-year timescales, Eos, Transactions, American Geophysical Union, 79, 1998.
 

Comparing short-term and long-term rates of chemical weathering
At intensively monitored watersheds in central Idaho, we are comparing rates of chemical weathering over timescales of decades and millennia.  Our decade-scale estimates of chemical weathering are derived from input-output watershed mass balances.  Our millennial-scale chemical weathering estimates are derived from cosmogenic nuclides (which measure the long-term rate of erosion and thus, assuming erosional equilibrium, the long-term rate that rock is converted to soil) and element depletion profiles (which measure how much of each element is weathered as rock is converted to soil).  Preliminary results suggest our methods yield well-constrained estimates of long-term chemical weathering rates, and that these are broadly consistent with present-day weathering rates estimated from watershed mass balances.

Collaborators:  Cliff Riebe (Ph.D. candidate, University of California), Laura Glaser (Undergraduate, University of California), Jim Clayton (U.S. Forest Service).
 

Effects of engineered channel modifications on flood stage and bank erosion of the Ganges River
Varanasi, the holiest Hindu city, receives pilgrims from around the world who come to ritually bathe in the Ganges.  Unfortunately, Varanasi--a city of over 1.2 million inhabitants--lacks adequate sewage treatment facilities, and consequently fecal coliform counts in the religious bathing area are 10,000-1,000,000 times higher than acceptable levels.  The pollution of the river presents an obvious health hazard to bathers, particularly to those who perform the complete bathing ritual, which includes drinking the river water.  Varanasi's sewage treatment facilities may soon be improved, by building sewage oxidation ponds in a small high-flow channel of the Ganges several kilometers downstream of the city.  These sewage treatment facilities would require closing off a 7m-deep, 400m-wide channel of the Ganges entirely.

Our study evaluates whether closing this channel will significantly alter flood heights and patterns of bank erosion on the Ganges.  Using simple, robust 'back-of-the-envelope' calculations, we have found that these facilities will indeed alter flood stages and bank erosion rates, but not by much: flood stages will rise by a maximum of 9cm and bank erosion rates will increase by a maximum of 6 percent (Kirchner et al., 1997).  Effects this small are likely to undetectable, given the wide range of natural variability.

Collaborators:  Mallickarjun Joshi, D. K. Sundd, Veer Bhadra Mishra and S. K. Mishra (Sankat Mochan Foundation and Banaras Hindu University).

Technical report:
Kirchner, J.W., Expected effects of closing the Sota channel on flood stage and bank erosion of the Ganges River, near Varanasi, Uttar Pradesh, U.S.-Asia Environmental Partnership Environmental Exchange Program, 23 pp., 1997.
 

Effects of riparian vegetation on river bank stability and erosion rates
It is widely believed that vegetation affects river bank erosion rates, but this effect remains poorly quanitifed.  Long-term planning for riparian zones requires a better understanding of how, and how much, vegetation contributes to bank stability.  Graduate student Lisa Micheli is pursuing this question through three lines of research (Micheli and Kirchner, 1997):
1. Direct measurements of mechanical strength of banks with similar soils but differing vegetation densities.  Preliminary results indicate that roots of herbaceous vegetation can account for up to 80 percent of the mechanical strength of stream banks.
2. Indirect measurements of effective bank strength, derived by measuring the geometry of stable and unstable banks and modeling the stress distributions within the bank profile.
3. Comparisons of bank vegetation patterns and long-term rates of bank erosion, as measured by comparing aerial photographs from different periods.  From surveys of channel geometry, the flow patterns (and thus the bank stresses) during floods are reconstructed.  By comparing the observed rate of bank erosion and the stresses imposed during floods, we assess how patterns of bank erosion are related to (a) shear stresses imposed by flood flows, (b) bank material characteristics, and (c) bank vegetation cover and rooting density.  This provides a large-scale field test of both whether vegetation significantly alters bank erodibility under real-world conditions, and whether our mechanistic models of bank erosion adequately predict this effect.

Collaborators:  Lisa Micheli (Ph.D. candidate, University of California)

Abstracts:
Micheli, E. and J.W. Kirchner, Quantifying how herbaceous vegetation stabilizes stream banks: Monache Meadow, South Fork of the Kern River, Southern Sierra Nevada, Eos, Transactions, American Geophysical Union, 78, F306, 1997.
 

Non-linear mechanisms of hillslope erosion
Understanding the mechanisms controlling hillslope erosion is crucial for assessing the impact of timber harvesting, grazing, and other land uses.  Hillslope erosion is usually modeled as a 'linear diffusion' process, in which soil creeps downhill at a rate that is proportional to the topographic gradient.  Over long periods of time, linear diffusion creates parabolic hillslopes, in which convexity is roughly constant along the hillslope profile.  Most hillslopes, however, become straighter as they become steeper, suggesting that hillslope erosion is a non-linear diffusion process.  Graduate student Josh Roering is exploring non-linear diffusion as a model for hillslope erosion.  Our theoretical analysis of erosional energy dissipation suggests that downslope transport of sediment should obey a particular non-linear diffusion law, in which downslope transport rates rise toward infinity as slope angles approach a critical gradient.  Tests of this erosion model, using high-precision digital topographic data derived from airborne laser altimetry, strongly support the non-linear diffusion law (Roering, Kirchner and Dietrich, 1999).  This non-linear transport law will be important for estimating basin-scale sediment yields and modeling the linkages between climate, land use, and landscape evolution.

Collaborators:  Josh Roering (Ph.D. candidate) and Bill Dietrich (University of California).

Publications:
Roering, J. J., J. W. Kirchner, L. E. Sklar and W. E. Dietrich, Nonlinear creep and landsliding on an experimental hillslope, Nature (in review).
Roering, J.J., J.W. Kirchner and W.E. Dietrich, Evidence for non-linear, diffusive sediment transport on hillslopes and implications for landscape morphology, Water Resources Research, 35, 853-870, 1999.

Abstracts:
Roering, J.J., L. Sklar, J.W. Kirchner, and W.E. Dietrich, A laboratory simulation of diffusive sediment transport on hillslopes: non-linear transport and the evolution of convex hilltops, Eos, Transactions, American Geophysical Union, 79, 1998.
Roering, J.J., J.W. Kirchner and W.E. Dietrich, Evidence for a non-linear diffusive mass wasting transport law and implications for hillslope evolution in the Oregon Coast Range, Eos, Transactions, American Geophysical Union, 78, F286, 1997.
 

Transport and fate of mercury in mine tailings, Marin County, California
The Gambonini mine is typical of many small mines throughout the west.  In the early 1970's it was mined for cinnabar and then abandoned, leaving behind a 200,000 m3 tailings pile containing over five tons of mercury.  This tailings pile is eroding, carrying mercury-rich tailings into a stream that drains into Tomales Bay, where oysters are grown for human consumption.  Graduate student Dyan Whyte measured fluxes of mercury in the stream draining the mine site, and is now assessing the ultimate fate of mine tailings in the stream and bay.  Our measurements showed that in only two months, this small mine site released over 80 kilograms of mercury to downstream waters.  Our measurements triggered a multi-million-dollar EPA emergency Superfund intervention to prevent further mercury losses from the site.  We are currently designing a monitoring program to evaluate the effectiveness of the Superfund remediation efforts.

Collaborator:  Dyan Whyte (M.S. candidate, University of California).

Publications:
Whyte, D. C. and J. W. Kirchner , Assessing water quality impacts and cleanup effectiveness in streams dominated by episodic mercury discharges, Science of the Total Environment (in review).

Abstracts:
Whyte, D.C. and J.W. Kirchner, Assessing water quality impacts and cleanup effectiveness in streams dominated by episodic mercury discharges (Fifth International Conference on Mercury as a Global Pollutant, Rio de Janeiro, May 1999)
Whyte, D. and J.W. Kirchner, Fluvial transport of mercury in a watershed impacted by mining (Third Annual Mercury Conference, Coastal Advocates, Asilomar, California, February 1998)
 

Inferring watershed flowpaths and residence times from chemical tracers in rainfall and streamflow
The residence time of water in watersheds controls the retention of soluble contaminants, and thus the downstream consequences of pollution episodes.  We have recently used detailed time series of chloride (a nonreactive tracer) to study the travel time of water through watersheds.  Across diverse sites in Wales, Norway, and the northeastern United States, rainfall chloride concentrations have a white noise spectrum, but streamflow chloride concentrations follow fractal 1/f scaling on timescales from days to years (Kirchner, Feng, and Neal, in review and manuscript in preparation).  This indicates that watersheds do not have a characteristic residence time in the conventional sense; instead, their travel times follow an approximate power-law distribution with a long upper tail.

We plan to extend this work in two ways.  First, we want to make detailed, long-term measurements of 18O or 2H in rainfall and streamflow, to unambiguously confirm the fractal scaling observed in our chloride time series, and to rule out the possibility that as-yet-unknown chemical reactions are inhibiting the mobility of chloride in our study watersheds.  Second, we want to uncover the mechanistic reason why travel times in watersheds tend to follow a power-law distribution.  Soil pores and bedrock fractures exhibit fractal scaling, which may be linked to the fractals we observed in stream tracer concentrations.  But more prosaic factors must also be considered: the spatial distribution of rainfall, the length and tortuosity of the flowpaths connecting the stream to each point on the surface, dispersion along and between flowpaths, and diffusive exchange between mobile water in macropores and immobile water in the matrix between them.  With mathematical models of these various mechanisms, we can test whether they could potentially give rise to the scaling behavior we observe in the field.  We also hope to build a series of laboratory-scale physical models, both to confirm our theoretical analyses, and to directly test how subsurface flowpath configurations are reflected in streamflow tracer time series.

Collaborators:  Xiahong Feng (Dartmouth College), Colin Neal (Institute of Hydrology, UK)

Publications:
Kirchner, J. W., X. Feng and C. Neal, Fractal stream chemistry and its implications for contaminant transport in catchments, Nature (in press).

Abstracts:
Kirchner, J. W., X. Feng and C. Neal, Fractal stream chemistry and its implications for contaminant transport in catchments (abstract), Eos,Transactions, American Geophysical Union, 80, 1999.
Kirchner, J.W., Simplicity and complexity, in model watersheds and real ones (Gordon Conference on Hydrological, Biological, and Geochemical Processes in Forested Catchments, July 1999)
 

Long-term base cation dynamics of forest soils under declining acid deposition
For studying catchment base cation dynamics, the Solling forest data set is unique worldwide.  Long-term studies ongoing since 1966 at Solling have included not only monitoring of deposition, throughfall, and seepage water chemistry and fluxes, but also seven inventories of exchangeable cations in forest soils.  Measured changes in soil chemistry over decade timescales can therefore be compared directly with observed changes in seepage water chemistry.  These comparisons are important for testing models linking soil chemistry and runoff chemistry, and only at Solling can such comparisons be made directly from measured data.  We are using the Solling data set to directly test Kirchner's (1992) theory linking base cation depletion from catchment soils, and long-term acidification of catchment runoff.  There are no adjustable parameters, making this a particularly strict test.  Preliminary results from the Solling data set agree with the theoretical predictions.

Collaborators:  Karl-Joseph Meiwes and Henning Meesenburg (Lower Saxony Forest Experiment Station, Göttingen), and Michael Bredemeier (Forest Ecosystem Research Center, Göttingen University)

Related publications:
Kirchner, J. W., Acid rain revisited (Letter), Science, 273, 293-294, 1996.
Kirchner, J.W., R.P. Hooper, C. Kendall, C. Neal, and G. Leavesley, Testing and validating environmental models, Science of the Total Environment , 183, 33-47, 1996.
Kirchner, J. W. and E. Lydersen, Base cation depletion and potential long-term acidification of Norwegian catchments, Environmental Science and Technology, 29, 1953-1960, 1995.
Kirchner, J. W., Heterogeneous geochemistry of catchment acidification, Geochimica et Cosmochimica Acta, 56, 2311-2327, 1992.
Kirchner, J. W., P. J. Dillon and B. D. LaZerte, Predicted response of stream chemistry to acid loading tested in Canadian catchments, Nature, 358, 478-482, 1992.

Abstracts:
Kirchner, J.W., Long-term acidification of streams and watersheds revealed by catchment monitoring data, Eos, Transactions, American Geophysical Union, 78, F326, 1997.
Kirchner, J.W., Long-term acidification resulting from catchment base cation depletion: geochemical theory and field observations, Journal of Conference Abstracts, 1, 312, 1996
Kirchner, J.W., Geochemical methods for interpreting runoff chemistry from manipulated catchments, in Experimental Manipulations of Biota and Biogeochemical Cycling in Ecosystems
(L. Rasmussen, T. Brydges and P. Mathy, eds.), pp. 128-130, Commission of the European Communities, 1993.
 

Sulfur isotope dynamics in highly polluted Czech catchments
We are using 34S to constrain the mass balance for sulfur at two highly polluted catchments (Jezeri and Nacetin) in the 'Black Triangle' region of the Czech Republic.  Atmospheric deposition of sulfur is now declining due to 'acid rain' emission controls, but in many regions, outputs of sulfur in streamwater are higher than--and declining more slowly than--inputs in precipitation and dry deposition.  At these two sites, our isotopic measurements show that over half of the sulfur in runoff is coming from organic sulfur pools in the catchments' soils (Novak et al, in press).  The inventory of sulfur in the soils is roughly 30 times the annual flux, so release of stored sulfur will significantly delay catchment recovery as pollution controls are introduced.

Collaborators:  Martin Novak and Hana Groscheova (Czech Geological Survey).

Publications:
Novak, M., J. W. Kirchner, H. Groscheova, M. Havel, J. Cerny and R. Krejci, Sulphur isotope dynamics in two Central European watersheds affected by high atmospheric deposition of SOx, Geochimica et Cosmochimica Acta (in press).

Abstracts:
Novak, M., J.W. Kirchner and H. Groscheova, Sulphur isotope dynamics in two mountaintop forest catchments in the Black Triangle, Central Europe, Journal of Conference Abstracts, 2, 261, 1997.
 

Long-term base cation dynamics of forested catchments at Hubbard Brook, New Hampshire
The biogeochemistry of the Hubbard Brook catchment has been intensively monitored for over 30 years.  At Hubbard Brook, acid deposition is now declining, but because base cation concentrations in streamwater have declined at roughly equal rates, overall water quality has not improved.  We are applying methods developed by Kirchner (1992) and Kirchner and Lydersen (1995) to assess how water quality would have evolved over time, if 'acid rain' emission controls had not been implemented.  We are also testing whether geochemical methods (Kirchner, 1992) can accurately predict the mobilization of base cations, aluminum, and hydrogen ions in response to nitrate pulses triggered by forest clearcutting.

Collaborator:  Gene Likens (Institute of Ecosystem Studies)
 

Trace element mobilization by clearcutting and acid deposition in Welsh upland catchments
Hydrological and chemical monitoring by the Institute of Hydrology has compiled an unprecedented data set for studying catchment biogeochemistry: 14 years of weekly measurements of 35 major, minor, and trace elements in rainfall and streamflow from six Welsh catchments.  We are using this data set to:
-evaluate different data analysis methods, with the aim of assessing the hydrological and geochemical effects of clearcutting (in some catchments) and forest aggradation (in other catchments),
-test whether these multiple chemical signals can help to reveal the hydrological flowpaths within these catchments, and,
-test whether acid anions (supplied either by acid deposition or forest clearcutting) measurably mobilize trace elements in these catchments.  Initial results show that Kirchner's (1992) heterogeneous equilibrium geochemical theory accurately predicts the mobilization of a range of major, minor, and trace elements, whose concentrations span over four orders of magnitude.

Collaborator:  Colin Neal (Institute of Hydrology, U.K.).
 

Macroevolutionary dynamics inferred from the fossil record
Paleontologist Anne Weil and I are using time-series and spectral methods to look for signatures of macroevolutionary processes in the Phanerozoic marine fossil record.  Using the cross-correlation between rates of extinction and subsequent rates of origination, we have shown recoveries from mass extinctions and "background" extinctions alike require an average of roughly 10 million years.  We have also shown that origination rates are significantly more autocorrelated than extinction rates, although neither exhibits significant autocorrelation beyond roughly 20 million years.  The Fourier power spectra of fossil extinction and origination do not unambiguously conform to any simple model, such as white noise, fractals, or autoregressive time series.

Collaborator:  Anne Weil (Duke University).

Publications:
Kirchner, J. W. and A. Weil, Delayed biological recovery from extinctions throughout the fossil record, Nature (in review).
Kirchner, J. W. and A. Weil, Autocorrelations through time in rates of fossil extinction and origination, Proceedings of the Royal Society of London B: Biological Sciences, (in review).
Kirchner, J.W. and A. Weil, No fractals in fossil extinction statistics, Nature, 395, 337-338, 1998.
Abstracts:
Kirchner, J.W. and A. Weil, Time scales of recovery from extinction, inferred from lags between extinctions and originations through Phanerozoic time, GSA Abstracts with Programs, 31, 1999.
Kirchner, J.W. and A. Weil, The fossil record does not support fractal extinctions or self-organized criticality of the biosphere, GSA Abstracts with Programs, 30, 1998
 

Evolutionary ecology and genetics of host-pathogen interactions
We are building simple models to explore how host-pathogen interactions affect the co-evolutionary genetics of host and pathogen traits.  Pathogens need host organisms in order to survive and reproduce, but pathogen infection usually decreases a host's ability to reproduce itself.  These interconnected mechanisms of natural selection can create complicated--and often counterintuitive--evolutionary dynamics.  Work to date has yielded the following results:

-Host-pathogen interactions can affect the evolution of pathogen infectiousness and lethality, as well as host longevity and reproduction rate.  The degree of host-pathogen genetic specificity--that is, the degree to which particular pathogen genotypes infect some host genotypes more readily than others--is a crucial factor controlling the rate, and even the direction, that host and pathogen traits evolve.

-Pathogens can affect the evolution of host traits in counterintutive ways.  For example, selection by pathogens can give shorter-lived host organisms an evolutionary advantage over longer-lived hosts.

-Host organisms respond to the threat of disease either by inibiting infection (resistance) or by limiting the damage it causes (tolerance).  Although tolerance and resistance may have equivalent short-term benefits to the host, their evolutionary dynamics are fundamentally different.  No gene conferring complete resistance can become universal in a host population, but any tolerance gene conferring a net benefit will do so.  The co-evolutionary dynamics are different as well; pathogens have an evolutionary incentive to evade a hosts' resistance defenses, but they have an incentive not to outwit a hosts' tolerance mechanisms.  These observations suggest a new mechanism for the evolution of mutualism from parasitism, and strengthen the arguments for a tolerance-based approach to managing agricultural pathogens.

Collaborator:  Barbara Roy (Geobotanical Institute, Swiss Federal Institute of Technology, Zürich).

Publications:
Kirchner, J.W. and B.A. Roy, Evolutionary implications of host-pathogen specificity 1.  Fitness consequences of host life history traits, Journal of Evolutionary Biology (in review).
Kirchner, J.W. and B.A. Roy, Evolutionary implications of host-pathogen specificity 2.  Fitness consequences of pathogen virulence traits, Journal of Evolutionary Biology (in review).
Roy, B.A., J.W. Kirchner, C. Christian and L. Rose, Disease incidence and disease tolerance in a Great Basin plant community, Oecologia, (in review).
Roy, B.A. and J.W. Kirchner, Evolutionary dynamics of pathogen resistance and tolerance, Evolution  (in press).
Kirchner, J.W. and B.A. Roy, The evolutionary advantages of dying young: epidemiological implications of longevity in metapopulations, The American Naturalist , 154, 140-159, 1999.