Prehistoric Decline in Freshwater Mussels Coincident with the Advent of Maize Agriculture

Abstract

During late prehistory, high population densities and intensive agricultural practices of Native American societies had profound effects on the pre-Columbian landscape. The degree to which Native American land use affected aquatic ecosystems is unknown. Freshwater mussels are particularly sensitive harbingers of modern-day ecosystem deterioration. We used data from prehistoric Native American shell middens to examine prehistoric trends in abundance of freshwater mussels of the genus Epioblasma in North America during the last 5000 years. The relative abundance of Epioblasma declined steadily during this period, a result that could be explained either by an increase in human impacts to streams or by long-term climatic changes unrelated to human activities. The rate of decline of Epioblasma increased significantly, however, after the advent of large-scale maize agriculture in the southeastern United States about 1000 years before the present. Our results suggest that human land-use activities in prehistory caused changes in freshwater mussel communities that were lower in magnitude but similar in direction to changes caused by recent activities.

Full text

Research Notes
Prehistoric Decline in Freshwater Mussels Coincident
with the Advent of Maize Agriculture
EVAN PEACOCK,∗WENDELL R. HAAG,†‡ AND MELVIN L. WARREN JR.†
∗Cobb Institute of Archaeology, Mississippi State University, Mississippi State, MS 39762, U.S.A.
†Center for Bottomland Hardwoods Research, U.S. Department of Agriculture Forest Service, 1000 Front Street,
Oxford, MS 38655, U.S.A.
Abstract: During late prehistory, high population densities and intensive agricultural practices of Native
American societies had profound effects on the pre-Columbian landscape. The degree to which Native American
land use affected aquatic ecosystems is unknown. Freshwater mussels are particularly sensitive harbingers of
modern-day ecosystem deterioration. We used data from prehistoric Native American shell middens to examine
prehistoric trends in abundance of freshwater mussels of the genus Epioblasma in North America during the
last 5000 years. The relative abundance of Epioblasma declined steadily during this period, a result that could
be explained either by an increase in human impacts to streams or by long-term climatic changes unrelated to
human activities. The rate of decline of Epioblasma increased significantly, however, after the advent of large-
scale maize agriculture in the southeastern United States about 1000 years before the present. Our results
suggest that human land-use activities in prehistory caused changes in freshwater mussel communities that
were lower in magnitude but similar in direction to changes caused by recent activities.
Key Words: Epioblasma,freshwater diversity, prehistoric human impacts
Declinaci´on Prehist´orica en Mejillones de Agua Dulce Coincidente con la Llegada de la Agricultura de Ma´ız
Resumen: Durante la prehistoria tard´
ıa, las altas densidades poblacionales y las pr´
acticas agr´
ıcolas in-
tensivas de las sociedades nativas de Am´
erica tuvieron efectos profundos sobre el paisaje precolombino. Se
desconoce el grado en que el uso del suelo por americanos nativos afect´
oaecosistemas acu´
aticos. Las almejas
de agua dulce son indicadores particularmente sensibles del deterioro actual de ecosistemas. Utilizamos datos
de restos de conchas prehist´
oricas para examinar patrones de abundancia de almejas de agua dulce del g´
enero
Epioblasma en Norte Am´
erica durante los ´
ultimos 5000 a˜
nos. La abundancia relativa de Epioblasma declin´
o
constantemente durante este per´
ıodo, lo que podr´
ıa explicarse por el incremento de impactos humanos en los
arroyos o por los cambios clim´
aticos de largo plazo no relacionados con actividades humanas. Sin embargo,
la tasa de declinaci´
on de Epioblasma aument´
o significativamente despu´
es de la llegada de agricultura de ma´
ız
agran escala en el sureste de Estados Unidos hace 1000 a˜
nos aproximadamente. Nuestros resultados sugieren
que las actividades humanas de uso de suelo en la prehistoria provocaron cambios en las comunidades de
almejas de agua dulce que fueron menores en magnitud, pero similares en direcci´
on, que los cambios causados
por actividades recientes.
Palabras Clave: diversidad de agua dulce, Epioblasm, impactos humanos prehist´oricos
‡Address correspondence to W.R. Haag, email whaag@fs.fed.us
Paper submitted January 27, 2004; revised manuscript accepted July 15, 2004.
547
Conservation Biology, Pages 547–551
Volume 19, No. 2, April 2005

548 Prehistoric Mussel Decline Peacock et al.
Introduction
Freshwater ecosystems worldwide are beleaguered by
the increasing demands of a growing human population.
In the southeastern United States, a variety of pressures
threaten highly diverse aquatic communities, including
the richest freshwater mussel fauna (order Unionoida)
on Earth (297 North American species, 269 in the south-
eastern United States; Neves et al. 1997). Historically,
freshwater mussels have been highly sensitive to human-
induced changes in aquatic habitats and today they rep-
resent the most endangered group of organisms in North
America. At least 21 species have become extinct in the
last 100 years (7% of the total fauna), and 194 (65%) are
considered imperiled (Williams et al. 1993). Members of
the genus Epioblasma have suffered a particularly severe
decline. Of 20 recognized species, mostly endemic to
the southeastern United States, 13 became extinct in the
twentieth century, indicating that these species were es-
pecially intolerant of human-induced changes to rivers.
Although localized mussel declines are often attributable
to reservoir construction, stream channelization, or point-
source pollution (Neves et al. 1997), the widespread na-
ture of declines worldwide implies that other, large-scale
disturbances associated with human land use, such as sed-
imentation and nonpoint-source pollution, have equally
large impacts on these animals.
Beginning about 5000 years before present (BP), steady
increases in the human population of the southeastern
United States resulted in an increased rate of land clear-
ance and disturbance that was associated with agricul-
ture, acquisition of wood for fuel, game management,
and other activities (Delcourt 1987; Johannessen 1993;
Peacock 1998). These disturbances intensified coincident
with the advent of large-scale maize agriculture beginning
about 1000 BP (Lopinot 1992). The impact of these ac-
tivities on water quality and aquatic organisms, including
freshwater mussels, has not been examined.
Native Americans exploited freshwater mussels as food
throughout the Holocene, and extensive shell deposits
(middens) associated with human habitation sites occur
along major rivers in the midwestern and south cen-
tral United States. Because prehistoric humans likely har-
vested mussels indiscriminate of species (Matteson 1960;
Peacock 2000), these data provide important records
of prehistoric mussel assemblages and are used for re-
constructing paleoenvironmental conditions (Morey &
Crothers 1998) and establishing ranges of species that pre-
dated modern human impacts (Peacock & James 2002).
Although few species extinctions or extirpations are doc-
umented in the archaeological record (Bogan 1990; but
see Williams & Fradkin 1999), little is known about long-
term prehistoric trends in mussel abundance and species
composition. We used data from archaeological shell mid-
dens to examine prehistoric trends in the abundance of
freshwater mussels of the genus Epioblasma in rivers in
the southeastern United States during the last 5000 years.
To assess potential impacts of changes in Native American
land-use patterns on freshwater mussel communities, we
examined differences in the rate of change of Epioblasma
abundance before and after the advent of maize agricul-
ture.
Methods
We compiled data from published and unpublished ar-
chaeological reports on mussel community composition
from 41 temporally distinct prehistoric shell assemblages
representing 27 sites and 12 rivers, mostly in the south-
eastern United States (Table 1). A complete list of data
sources is available from the senior author. A variety of
contexts, from general cultural layers to the fill of fea-
tures (e.g., pits, postholes) are represented in these data
sets. Assemblages were excluded if they appeared to be
of mixed cultural origin (e.g., if pits intruded into earlier
strata, mixing the shell), if an insufficient number of depo-
sitional contexts were represented (e.g., shell collected
from the surface of a site), or if shell was recovered using
nonstandardized methods (Peacock 2000). For samples
without a specific date assigned only to a general cultural
period, we used the midpoint of the reported cultural
period as the date for that sample. We used the follow-
ing cultural period definitions (BP): Late Archaic (5000–
2500), Early Woodland (2500–2000), Middle Woodland
(2000–1500), Late Woodland (1500–1000), and Mississip-
pian (1000–500) (Steponaitis 1986). We did not include
mussel assemblages assigned to “transitional periods.”
We calculated the relative abundance of Epioblasma
spp. as a percentage of total shells represented in each
sample. We tested for relationships between time and
relative abundance of Epioblasma in two ways. First,
we conducted a randomized linear regression of relative
abundance (as arc-sine transformed proportions) on time
(10,000 randomizations; Manly 1997). We also fit curvi-
linear models to the data and, although significant, none
were an improvement over the linear model. Second,
we conducted a randomized Mantel test between two
pair-wise distance matrices of differences in time and dif-
ferences in relative abundance between samples (10,000
randomizations; Manly 1997).
To examine effects of changes in human land use on
mussel communities, we calculated the rate of change
(percent/100 years) in Epioblasma spp. relative abun-
dance between pairs of temporally successive mussel as-
semblages found at the same site (n=13 pairs). We
grouped pairs of observations into two categories: before
and after the advent of maize agriculture (1000 BP). We in-
cluded pairs of observations that straddled this boundary
in the “after-maize” category. We calculated the mean rate
of change and percentile bootstrapped 95% confidence
Conservation Biology
Volume 19, No. 2, April 2005

Peacock et al. Prehistoric Mussel Decline 549
Table 1. Prehistoric mussel assemblages from the eastern United States examined in this study.
Cultural period (sample size,%
River, State, CountyaEpioblasma)bEpioblasma spp. presentc
Clinch, TN, Roane Woodland (20238, 13.7), arc, bre, cap, flo, hay, obl, pro, ste, tor, tri
Mississippian (2713, 17.8)
Cumberland, KY, Livingston Mississippian (4913, 21.47; 1551, 7.22) arc, fle, tor
Cumberland, TN, Davidson Mississippian (4548, 18.6; 1642, 12.3) arc, bre, cap, fle, flo, hay, pro, ste, tor
Cumberland, TN, Jackson Archaic (12060, 12.6; 4548, 15.4) arc, bre, fle, hay, obl, pro, tor, tri
Cumberland, TN, Smith Woodland ( 715, 11.3) arc, bre, cap, flo, hay, obl, ste, tor
Cumberland, TN, Smith Woodland (827, 8.3) arc, bre, cap, fle, flo, hay, lew, obl, ste, tor
Cumberland, TN, Davidson Mississippian (762, 2.6) arc, bre, hay, pro
Green, KY, Butler Archaic (2215, 21.0; 2248, 18.0) obl, per, pro, tor,dtri
Green, KY, Butler Archaic (16279, 37.7) tor, dtri
Hiwassee, TN, Bradley, McMinn Mississippian (167, 1.2) tor
Kentucky, KY, Woodford Mississippian (>400,e2.5) sam, tord
Little Pigeon, TN, Sevier Mississippian (3855, 2.0) bre, cap, flo, hay, ste, tor
Little Tennessee, TN, Monroe Mississippian (707, 1.1) bre, hay, lew, ste
Pond, KY, Hopkins Mississippian (2162, 2.4) tor,dtri
Scioto, OH, Ross Woodland (1977, 21.4) obl, tor, tri
Tennessee, AL, Lauderdale Archaic (1185, 50.5), Woodland (830, 34.8) arc, bie, bre, fle, flo, hay, lew, obl, per, pro, ste, tor
Tennessee, AL, Lauderdale Archaic (10473, 28.1), arc, bie, bre, cap, fle, flo, hay, lew, obl, per, pro,
Woodland ( 7451, 21.2) ste, tor
Tennessee, AL, Jackson Woodland (3148, 1.5) arc, bre, fle, lew, pro, ste, tor, tri
Tennessee, TN, Meigs, Rhea Woodland (945, 6.7; 11363, 6.2), arc, bre, cap, fle, flo, hay, lew, obl, pro, ste, tor, tur
Mississippian (2794, 1.0)
Tennessee, AL, Jackson Archaic (1938, 9.4), Woodland (48196, 4.9) arc, bie, bre, cap, fle, hay, lew, pro, ste, tor, tri
Tennessee, AL, Lauderdale Archaic (2537, 22.0) arc, bie, bre, cap, fle, flo, hay, lew, pro, ste, tor
Woodland (630, 12.7)
Tennessee, AL, Lauderdale Archaic (2980, 30.6) arc, bie, bre, cap, flo, hay, lew, per, pro, ste, tor
Tennessee, TN, Loudon Woodland (593, 19.1), arc, bre, cap, hay, lew, pro, ste, tor, tri
Mississippian (931, 14.3; 9206, 14.3)
Tombigbee, MS, Clay, Woodland (46801, 6.4) pen
Lowndes and AL, Pickens Mississippian (6911, 1.2)
Wabash, IL, Crawford Archaic (8135, 15.5) obl, tor,dtri
Wabash, IL, Crawford Archaic (6557, 30.6) obl, tor,dtri
Wabash, IL, Lawrence Archaic (18516, 15.2) obl, tor,dtri
aState abbreviations: TN, Tennessee; KY, Kentucky, OH, Ohio; AL, Alabama; MS, Mississippi; IL, Illinois.
bSample size is the total number of shells (all species) present in each assemblage; assemblages with two different samples size entries indicate
two temporally distinct samples within that cultural period.
cSpecies abbreviations: arc, arcaeformis; bie, biemarginata;bre,brevidens; cap, capsaeformis;fle, flexuosa;flo,florentina;hay,haysiana;lew,lewisi;
obl, obliquata; pen, penita; per, personata;pro, propinqua; sam, sampsoni; ste, stewardsoni; tor, torulosa;tri,triquetra; tur, turgidula.
dReported as Epioblasma cincinnatiensis, E. phillipsi, or E. rangiana.
eExact sample size not given.
intervals of Epioblasma relative abundance before and af-
ter maize. We tested for differences in the rates of change
during these two time periods with a randomized one-
way analysis of variance (10,000 randomizations; Manly
1997).
Results
Relative abundance of Epioblasma in mussel commu-
nities in the southeastern United States declined signif-
icantly from 5000 to 500 BP. Relative abundance was lin-
early and negatively related to time (R2=0.369, p<
0.0001, Fig. 1), and distance matrices of time and rela-
tive abundance were significantly associated (Mantel r=
0.142, p<0.0004). Although the slope of the linear re-
lationship is low (0.0043; 95% CI, 0.0027–0.0061), over
time this gradual decrease in abundance resulted in ma-
jor changes in mussel community composition. Mean rel-
ative abundance of Epioblasma decreased from 23.6%
(95% CI, 17.9–30.0) in the Archaic to 8.0% (4.4–11.9) in
the Mississippian.
The rate of decline in the relative abundance of Epi-
oblasma differed over time. Although we could fit both
linear and curvilinear functions to the relationship be-
tween time and Epioblasma abundance, a conspicuous
cluster of low values of Epioblasma abundance was ev-
ident after the advent of maize agriculture (Fig. 1). The
mean rate of decline in Epioblasma abundance between
paired, temporally successive samples was significantly
higher after the advent of maize agriculture (F1,11 =5.07,
Conservation Biology
Volume 19, No. 2, April 2005

550 Prehistoric Mussel Decline Peacock et al.
Figure 1. Relationship between time and relative
abundance of Epioblasma spp. in the eastern United
States (arc sine [relative abundance] =12.119 +
0.0043 time, R2=0.369,p<0.0001). The regression
line was created using back-transformed predicted y
values generated from the regression equation.
p<0.017; mean rate of decline/100 years [95% CI]: be-
fore maize =1.38 [–0.79–3.01], n=7; after maize =14.81
[4.81–27.51], n=6). The confidence interval around the
mean decline before maize included zero, suggesting lit-
tle or no decline during this time period. In contrast, the
confidence interval after maize did not include zero but
was wide, indicating high regional variation in the rate of
decline.
Discussion
Although the climate of eastern North America changed
continually throughout the Holocene, including small
but abrupt temperature fluctuations during the last 1000
years (Gates 1993), the major climate changes occurred
prior to 9000 BP (Webb et al. 1993). After the hypsither-
mal climatic optimum (6000 BP), distribution of vege-
tative communities has been similar to the present day
(Delcourt & Delcourt 1987; Webb et al. 1993), suggest-
ing a relatively stable climate during our study period.
Nonetheless, we cannot discount the possibility that pre-
historic declines in relative abundance of Epioblasma
represent mussel community responses to long-term cli-
matic changes and are unrelated to human effects on the
landscape. Temporal variation in the rate of Epioblasma
decline, however, is coincident with changes in prehis-
toric Native American land use. Together with the sen-
sitivity of Epioblasma to recent anthropogenic stream
alterations, prehistoric land-use patterns provide a com-
pelling explanation for changes in abundance of these
species seen over the last 5000 years.
Throughout the Late Archaic and Woodland periods
(5000–1000 BP), prehistoric Americans gradually but
steadily increased their dependence on a suite of na-
tive cultigens (the Eastern Agricultural Complex; Yarnell
1993). Concomitant increases in land clearing for cultiva-
tion of these crops resulted in conspicuous increases in
the pollen record of disturbance-favored plants (e.g., rag-
weed, Ambrosia spp. L.; American cane, Arundinaria
gigantea [Walt.] Muhl.) (Chapman et al. 1982; Delcourt
1987; Delcourt et al. 1998). During this period of grad-
ual but steady increases in anthropogenic environmental
disturbance, abundance of Epioblasma showed at most
agradual, low rate of decline.
About 1000 BP, most Native American groups in the
southeastern United States adopted maize-based agricul-
ture, supplemented by beans, squash, and continued
cultivation of Eastern Agricultural Complex crops (Fritz
1990; Lopinot 1992). Adoption of maize occurred simulta-
neously with a rapid increase in the intensity and scale of
agriculture throughout the Mississippian. Fields reaching
tens to hundreds of hectares in size became a common
feature across the landscape (Peacock 1998), and settle-
ment hierarchies based on an agricultural surplus were
established (Peebles 1978). During this time, indicators
of anthropogenic disturbance such as charcoal influx and
sedimentation rates increased markedly (Chapman et al.
1982), showing widespread intensification of land use
and soil erosion (Steponaitis 1986; Delcourt 1987; Del-
court 1997). Similarly, the mean rate of decline in Epi-
oblasma abundance increased sharply during this period
of rapid intensification of anthropogenic disturbance and
wasanorder of magnitude higher than before the advent
of large-scale maize agriculture in the region.
Epioblasma declined even more dramatically in the
twentieth century, relative to prehistoric declines, and
these species are now extirpated from all our study sites
or extinct (Neves et al. 1997). Although the causes of
widespread, modern-day declines in mussel populations
remain unclear, stream sedimentation associated with
large-scale land disturbance is implicated as a primary
cause (Bogan 1993). Our results from prehistory support
the notion that increases in land clearance and distur-
bance have measurable effects on freshwater mussel com-
munities. Further, our results suggest that prehistoric hu-
man populations exerted substantial pressures on aquatic
ecosystems that were similar to, but less acute than, pres-
sures exerted by modern-day society.
Acknowledgments
We thank J. G. McWhirter, E. Futato, and R. Hurst for their
contributions to this study. This study was supported in
part by the U.S. Department of Agriculture Forest Service,
Southern Research Station.
Conservation Biology
Volume 19, No. 2, April 2005

Peacock et al. Prehistoric Mussel Decline 551
Literature Cited
Bogan, A. E. 1990. Stability of recent unionid (Mollusca: Bivalvia) com-
munities over the past 6000 years. Pages 112–136 in W. Miller III,
editor. Paleocommunity temporal dynamics: the long-term develop-
ment of multispecies assemblies. Special Publication 5. The Paleon-
tological Society, Museum of Natural History, Washington, D.C.
Bogan, A. E. 1993. Freshwater bivalve extinctions (Mollusca: Unionida):
asearch for causes. American Zoologist 33:599–609.
Chapman, J., P. A. Delcourt, P. A. Cridlebaugh, A. B. Shea, and H. R.
Delcourt. 1982. Man-land interactions: 10,000 years of American
Indian impact on native ecosystems in the lower Little Tennessee
River valley. Southeastern Archaeology 1:115–121.
Delcourt, H. R. 1987. The impact of prehistoric agriculture and land
occupation on natural vegetation. Trends in Ecology & Evolution
2:39–44.
Delcourt, H. R. 1997. Pre-Columbian Native American use of fire on
southern Appalachian landscapes. Conservation Biology 11:1010–
1014.
Delcourt, P. A., and H. R. Delcourt. 1987. Long-term forest dynamics in
the temperate zone. Springer-Verlag, New York.
Delcourt, P. A., H. R. Delcourt, C. R. Ison, W. E. Sharp, and K. J.
Gremillion. 1998. Prehistoric human use of fire, the eastern agri-
cultural complex, and Appalachian oak-chestnut forests: paleoecol-
ogy of Cliff Palace Pond, Kentucky. American Antiquity 63:263–
278.
Fritz, G. J. 1990. Multiple pathways to farming in pre-contact eastern
North America. Journal of World Prehistory 4:387–435.
Gates, D. M. 1993. Climate change and its biological consequences.
Sinauer Associates, Sunderland, Massachusetts.
Johannessen, S. 1993. Farmers of the late woodland. Pages 57–77 in C.
M. Scarry, editor. Foraging and farming in the eastern woodlands.
University of Florida Press, Gainesville.
Lopinot, N. H. 1992. Spatial and temporal variability in Mississippian sub-
sistence: the archaeobotanical record. Pages 44–94 in W. I. Woods,
editor. Late prehistoric agriculture. Illinois Historic Preservation
Agency, Springfield, Illinois.
Manly, B. F. J. 1997. Randomization, bootstrap,and monte carlo methods
in biology. Chapman and Hall, London.
Matteson, M.A. 1960. Reconstruction of prehistoric environments
through the analysis of molluscan collections from shell middens.
American Antiquity 26:117–120.
Morey, D. F., and G.M. Crother s. 1998. Clearing up clouded waters: pale-
oenvironmental analysis of freshwater mussel assemblages from the
Green River shell middens, western Kentucky. Journal of Archaeo-
logical Science 25:907–926.
Neves, R. J., A. E. Bogan, J. D. Williams, S. A. Ahlstedt, and P. D. Hartfield.
1997. Status of aquatic mollusks in the southeastern United States:
adownward spiral of diversity. Pages 43–86 in G. W. Benz and D. E.
Collins, editors. Aquatic fauna in peril: the southeastern perspective.
Lenz Design & Communications, Decatur, Georgia.
Peacock, E. 1998. Historical and applied perspectives on prehistoric
land use in eastern North America. Environment & History 4:1–29.
Peacock, E. 2000. Assessing bias in archaeological shell assemblages.
Journal of Field Archaeology 27:183–196.
Peacock, E., and T. R. James. 2002. A prehistoric unionid assemblage
from the Big Black drainage in Hinds County, Mississippi. Journal of
the Mississippi Academy of Science 47:119–123.
Peebles, C. S. 1978. Determinants of settlement size and location in the
Moundville phase. Pages 369–416 in B. D. Smith, editor. Mississip-
pian settlement patterns. Academic Press, New York.
Steponaitis, V. P. 1986. Prehistoric archaeology in the southeastern
United States, 1970–1985. Annual Review of Anthropology 15:363–
404.
Webb,T., III, P. J. Bartlein, S. P. Harrison, and K. H. Anderson. 1993.
Veg etation, lake levels, and climate in eastern North America for the
past 18,000 years. Pages 415–467 in T. Webb III, W. F. Ruddiman, F.
A. Street-Perrott, and P. J. Bartlien, editors. Global climates since the
last glacial maximum. University of Minnesota Press, Minneapolis.
Williams, J. D., and A. Fradkin. 1999. Fusconaia apalachicola,a
new species of freshwater mussel (Bivalvia:Unionidae) from Pre-
columbian archaeological sites in the Apalachicola basin of Alabama,
Florida, and Georgia. Tulane Studies in Zoology & Botany 31:51–62.
Williams, J. D., M. L. Warren, K. S. Cummings, J. L. Harris, and R. J.Neves.
1993. Conservation status of freshwater mussels of the United States
and Canada. Fisheries 18:6–22.
Yar nell, R. A. 1993. The importance of native crops during the Late
Archaic and Woodland periods. Pages 13–26 in C. M. Scarry, editor.
Foraging and farming in the eastern woodlands. University of Florida
Press, Gainesville.
Conservation Biology
Volume 19, No. 2, April 2005