2007 SOUTHEASTERN NATURALIST 6(1):151–158
The Effect of Burial Depth on Removal of Seeds of
Phytolacca americana
John L. Orrock1,3,* and Ellen I. Damschen2,3,4
Abstract - Although burial is known to have important effects on seed predation in a
variety of habitats, the role of burial depth in affecting the removal of seeds in earlysuccessional
systems is poorly known. Phytolacca Americana (pokeweed) is a model
species to examine the role of burial depth in affecting seed removal because it is
common in early-successional habitats, studies suggest that seed removal is indicative
of seed predation, and seed predation is related to the recruitment of mature plants. To
determine how burial depth affects P. americana seed removal, 20 seeds of P.
americana were buried at depths of 0, 1, or 3 cm in early-successional habitats at the
Savannah River Site in South Carolina for over 6 weeks. The frequency with which
seeds were encountered (as measured by the removal of at least one seed) and the
proportion of seeds removed was significantly greater when seeds were on the soil
surface (0 cm depth) compared to seeds that were buried 1 cm or 3 cm; there was no
difference in encounter or removal between seeds at 1 cm or 3 cm. Our findings suggest
that burial may have important consequences for P. americana population dynamics,
because seed survival depends upon whether or not the seed is buried, and relatively
shallow burial can yield large increases in seed survival. Because seed limitation is
known to be an important determinant of plant community composition in earlysuccessional
systems, our work suggests that burial may play an unappreciated role in
the dynamics of these communities by reducing predator-mediated seed limitation.
Introduction
The vertical position of a seed within the soil (burial depth) influences
seed dormancy characteristics (Baskin and Baskin 1998) as well as the
likelihood that a seedling will successfully emerge from the soil (Baskin and
Baskin 1998). Burial depth may also affect the removal of seeds by seed
predators because buried seeds are less likely to be detected and removed by
seed predators in desert ecosystems (Reichman 1979), coastal dunes (Maron
and Simms 1997), grasslands (Hulme 1994, Maron and Simms 1997), temperate
forests (Crawley and Long 1995, Hulme and Borelli 1999), and
tropical forests (Andresen and Levey 2004). In early-successional species,
burial depth is known to reduce germination because many species will only
germinate near the soil surface (Baskin and Baskin 1998, Orrock et al.
2006). Moreover, burial depth influences the survival of seeds after germination
because larger-seeded species are more likely to successfully emerge
1Department of Ecology, Evolution, and Organismal Biology, Iowa State University,
Ames, IA 50011. 2Department of Zoology, North Carolina State University, Raleigh,
NC 27809. 3Current address - National Center for Ecological Analysis and Synthesis,
735 State Street, Suite 300, Santa Barbara, CA 93101. 4Current address - Marine
Science Institute, University of California, Santa Barbara, CA 93101. *Corresponding
author - orrock@nceas.ucsb.edu.
152 Southeastern Naturalist Vol. 6, No. 1
from deeper in the soil profile (Baskin and Baskin 1998, Grundy et al. 2003).
However, despite evidence that the establishment of early-successional
plant species is often limited by the number of seeds that survive, germinate,
and emerge (seed limitation; Turnbull et al. 2000), little is known about how
burial depth affects the predation of seeds in early-successional systems.
Phytolacca americana Linneaus (pokeweed) is a perennial plant whose
distribution is largely within eastern North America (Mitich 1994).
Phytolacca is typically found in early-successional habitats, forest clearings
created by disturbance, and other frequently disturbed habitats (McDonnell et
al. 1984). Birds and other vertebrates consume the fruits of P. americana and
subsequently disperse the seeds via defecation (Martin et al. 1951, McDonnell
et al. 1984, Mitich 1994). Because of reliance upon vertebrate dispersal and
the deterrent effect of P. americana fruit pulp on rodent granivores
(McDonnell et al. 1984), pre-dispersal seed predation is probably extremely
rare. Evidence suggests that post-dispersal seed predation by arthropods,
rodents, and birds removes substantial numbers of P. americana seeds
(Boman and Casper 1995, Hyatt 1998, Orrock et al. 2003, Willson and
Whelan 1990) and may affect the size of P. americana populations (Orrock et
al. 2003). The importance of seed predation in the population dynamics of
Phytolacca americana and evidence that burial depth also affects seed germination
(Orrock et al. 2006) makes P. americana a model species for the
examination of how burial depth affects seed predation of early-successional
plants. In this paper, we examine the role of burial depth in affecting the
predation of P. americana seeds. Specifically, we examine how burial depth
affects the rate at which seed predators encounter seeds, defined as the
removal of at least one seed from a particular depth treatment at a site (Hulme
1994, Willson and Whelan 1990), as well as the percentage of seeds removed.
Methods
Study area and design
Mature P. americana fruits were collected on July 28, 2003 at the
Savannah River Site (SRS), a National Environmental Research Park
(NERP) located near Aiken, SC. Fruits contain approximately 10 seeds
(Armesto et al. 1983); each seed is 2.5–3 mm in size (Radford et al. 1968).
Seeds were removed from ripe fruits by rubbing the fruits against a sieve.
Collected seeds were then thoroughly washed and allowed to dry prior to
use. Seeds used for seed-removal trials were thus similar to the pulp-free
seed predators would encounter in the field after dispersal by frugivores
(McDonnell et al. 1984).
Seed-removal trials were conducted at three sites at the SRS, each separated
by several kilometers (Fig. 1A). At each site, we used two early-successional
patches created in 1999 by clearcutting mature pine forest (> 30 years old;
Kilgo and Blake 2005), followed with a prescribed burn (Fig. 1A). Each patch
was 1-ha in size and was connected to an identical adjacent patch by a narrow
strip of clear-cut habitat (a “corridor”). These patches represent typical early2007
J.L. Orrock and E.I. Damschen 153
successional habitat of P. americana (Radford et al. 1968), as evidenced by
natural recruitment of P. americana within this study system (Orrock et al.
2003). Based upon vegetation surveys conducted in the study sites in 2003,
vegetation was characterized by Quercus falcata Michaux, Rhus copallina
Linnaeus, Rubus cuneifolius Pursh, Sassafras albidum Nuttall, Vaccinium
stamineum Small, and Vitis spp. (for additional description of the plant
community and survey methodology, see Damschen 2005).
Within each of the six patches (three sites, two patches per site; Fig. 1),
seed removal was examined at four locations within a 50- x 50-m square area
centered on the patch. Each location was 25 m from the closest patch edge
(Fig. 1). Plastic sample cups (approximately 6-cm diameter, 9-cm height,
120-ml volume) were used to hold seeds during seed removal trials. Screwon
lids prevented rain from changing the burial depth of seeds or from
washing seeds out of the cups, and 0.5-mm diameter holes drilled in the
bottom of each cup provided drainage. Each cup had a 2.5-cm diameter hole
drilled into the side to allow seed removal by rodents and invertebrates.
Although avian granivores may also exhume seeds, cups were designed to
allow us to focus our examination on removal by rodents and invertebrates
because of past evidence of their importance in P. americana seed removal
in the study area (Orrock et al. 2003).
Within each cup, 20 P. americana seeds were placed at one of three
depths: 0, 1, and 3 cm. These depths were selected because germination of P.
americana approaches zero as burial depth increases to 3 cm (Orrock et al.
Figure 1. The experimental
landscape at the Savannah
River Site (SRS) near Aiken,
SC, where seeds were collected
and burial trials were
conducted. At each of three
sites, two patches were used.
Patches consisted of clearcuts
within a matrix of mature pine
forest that were connected
with a narrow corridor of
clearcut habitat as part of another
study (see Orrock et al.
2003). Four stations were
placed in each patch. Each
station had three cups, each
containing seeds buried at either
0, 1, or 3 cm below the
surface.
154 Southeastern Naturalist Vol. 6, No. 1
2006). Sand, chosen to match the sand-rich soils at the site (Kilgo and Blake
2005), was added to each cup until the surface of the sand was level with the
entrance hole. At each location, one cup of each depth treatment was buried
so that the lowest point of the entrance hole was flush with the ground,
leaving only 4 cm of the cup (the entrance hole and lid) visible above the soil
surface. This design resulted in 12 cups per patch, for a total of 72 observations
(3 cups per location x 4 four locations per patch x 6 patches).
Cups were placed in the field from May 20–21 until July 4–6, 2004,
which exceeds the relatively short duration of many seed-removal studies
(< 4 weeks; Hyatt 1998). Because of disturbance by feral pigs at nine of the
24 locations, only locations where all cups were undisturbed were used for
analysis (N = 45 observations from 15 locations, at least one location per
patch was not disturbed).
Statistical methods
We quantified seed predation using two response variables: the frequency
of seed encounter and the proportion of seeds removed. The
frequency of seed encounter, defined as the removal of at least one seed
from a particular depth treatment at a site (Hulme 1994, Willson and
Whelan 1990), was examined using chi-square tests of independence
(Quinn and Keough 2002). Due to low frequency of encounter for buried
seeds, we used randomization tests with 100,000 randomizations to generate
Monte Carlo estimates of significance when expected cell frequencies
were less than 5 (Quinn and Keough 2002); Monte Carlo results were
qualitatively identical to results using asymptotic chi square. The proportion
of seeds removed was examined using one-way analysis of variance
(ANOVA). Blocks and patches were treated as random-effects blocks, with
burial depth (0, 1, or 3 cm) treated as a fixed effect (Quinn and Keough
2002). All proportions were arcsin squareroot transformed to improve normality
prior to analysis. Examination of residuals from ANOVA suggested
that they were normally distributed and that variance was homogeneous
among groups (Quinn and Keough 2002). All analyses were performed
using SAS v. 9.1 (SAS Institute 2004).
Results
Burial significantly reduced both the encounter and removal of P.
americana seeds by seed predators (Fig. 2). Seeds on the soil surface were
more frequently encountered (100%) compared to seeds buried at 1 cm
(60%) or 3 cm (53%) depth. The depth of seed burial was not as important as
burial itself, as seeds on the surface were more frequently encountered
compared to seeds buried at 1 cm and 3 cm (2 = 9.14, d.f. = 1, P < 0.01),
while differences among encounter rates for buried seeds were not significant
(2 = 0.14, d.f. = 1, P = 0.71). The proportion of seeds removed by seed
predators was also greatest for P. americana seeds on the soil surface: 83.9,
14.4, and 14.4% for 0, 1, and 3 cm burial depth, respectively (ANOVA, F2,39
2007 J.L. Orrock and E.I. Damschen 155
= 57.83, P < 0.01; Fig. 2). As was found for encounter frequency, there was
no difference in the proportion of seeds removed between seeds buried at
depths of 1 or 3 cm (linear contrast, F1,39 = 0.01, P = 0.94; Fig. 2).
Discussion
Although burial is known to reduce seed predation in a variety of plant
communities (Andresen and Levey 2004, Crawley and Long 1995, Hulme
1994, Hulme and Borelli 1999, Maron and Simms 1997, Reichman 1979),
our study provides the first evidence that burial also affects seed predation in
early-successional communities. We show that burial itself, regardless of
whether at 1 cm or 3 cm, confers a large reduction in seed predation by
rodents and arthropods (Fig. 2). Burial is likely to decrease predation of P.
americana seeds because it reduces the likelihood that arthropod or rodent
seed predators will detect and exhume seeds (Andresen and Levey 2004,
Hulme 1994, Hulme and Borelli 1999, Maron and Simms 1997, Reichman
1979). Comparison of our findings for seeds on the soil surface (see Results)
with other studies of P. americana seed removal from the soil surface
(Boman and Casper 1995, Hyatt 1998, Willson and Whelan 1990) suggests
that our design accurately captures patterns of seed removal without biasing
granivore foraging behavior.
We assume that seeds were not removed by forces other than animal seed
predators. Field observations support this assertion, as no sign of seed washout
from undisturbed cups was detected during cup collection. Seeds were
not “removed” via germination over the course of the experiment, because
lids and drainage holes prevented moisture from collecting in cups in sufficient
amounts to cue germination; no germinants or seedlings were found in
cups when cups were collected. Seeds were also unlikely to be destroyed by
fungal pathogens because P. americana seeds are resistant to fungal attack
(Orrock and Damschen 2005). We also assume that seed removal is indicative
of seed predation. This assumption is supported by a study by Orrock et
al. (2003) in the same system that showed a strong, negative relationship
Figure 2. The effect of burial depth
(0, 1, or 3 cm) on the proportion of
P. americana seeds removed from
early-successional habitats from
May 20–21 to July 4–6, 2004. Error
bars represent ± 95% confidence intervals,
and horizontal bar indicates
means that are not significantly different.
156 Southeastern Naturalist Vol. 6, No. 1
between seed removal and the recruitment of P. americana plants: 41% of
the variation in the number of P. americana in 40 different habitat patches
was explained by seed removal. Additionally, the abundance of P.
americana plants in the 40 patches studied by Orrock et al. (2003) was
negatively related to the abundance of the most common rodent granivore at
our study sites, Peromyscus polionotus old field mouse Wagner (Pearson
correlation, r = 0.70, N = 40, P < 0.01; J.L. Orrock, unpubl. data). Moreover,
secondary dispersal by ants is unlikely because Phytolacca americana seeds
do not have eliasomes (Radford et al. 1968).
Considered in light of evidence that seed predation is related to the
density of mature P. americana (Orrock et al. 2003) and additional evidence
that high levels of P. americana seed predation occur on or near the soil
surface (Boman and Casper 1995, Hyatt 1998, Willson and Whelan 1990),
our results suggest that burial may affect seed limitation of P. americana by
decreasing the likelihood that seeds will be consumed by seed predators.
Despite removal of nearly 90% of the seeds presented on the soil surface, our
results are likely to be conservative because P. americana dispersed in
summer would be susceptible to predators for longer periods of time than
used in this study (4–6 months compared to the 6 weeks used in this study).
The importance of seed predators in affecting P. americana over longer time
periods is reflected by additional work in this study system: recruitment of
P. americana did not increase despite the addition of over 10,000 seeds to a
patch, and instead, the abundance of P. americana in each patch was lower
in patches where seed removal was greatest (Orrock et al. 2006).
The immediate benefit of reduced predation conferred by burial may be
offset by the effect of burial depth on successful germination. A study by
Orrock et al. (2006) has shown that % germination of P. americana is 39, 3,
and 3% at 0, 1, and 3 cm, respectively (Orrock et al. 2006). Combined with
our estimates of seed predation, these data suggest that the proportion of seeds
on the soil surface that are likely to recruit to the seedling stage is 0.161 x 0.39
= 0.063, or 6.3%. The proportion of seeds buried at either at 1 or 3 cm that are
likely to recruit to the seedling stage is 0.856 x 0.03 = 0.026, or 2.6%.
Although a smaller proportion of recruits will come from buried seeds, the
reduced mortality of buried seeds yields greater contribution to future P.
americana recruitment when the total number of seeds is finite, assuming that
losses of buried seeds to soil pathogens and birds is negligible (an assumption
supported by the resistance of P. americana to fungal attack; Orrock and
Damschen 2005). These data suggest that, for early successional plants like P.
americana, the best overall strategy would combine rapid germination on the
soil surface (i.e., when predation is greatest), but high levels of dormancy
when seeds are buried, because deep seed burial is likely to affect emergence
if seedlings die while pushing through the soil (Baskin and Baskin 1998,
Grundy et al. 2003). We suggest that a general model of P. americana
population dynamics is that seeds are stored deep in the soil at depths greater
than those from which they will germinate, which are also depths where they
2007 J.L. Orrock and E.I. Damschen 157
are effectively safe from seed predators. Phytolacca americana may be
particularly well-suited to this strategy because its resistance to fungal pathogens
(Orrock and Damschen 2005) would promote persistence in the
seedbank. Upon disturbance (e.g., treefall, soil erosion), seeds are moved
closer to the soil surface, where germination becomes probable. At this
critical stage, when recruitment is either successful or thwarted, subtle differences
in burial depth yield large differences in predation.
Within the broader context of early-successional plant communities,
seed limitation has been shown to be an important determinant of community
composition (Turnbull et al. 2000). As such, burial depth may also play
an important role in determining the predation, and subsequent limitation, of
seeds of other early-successional plant species. However, because the role of
burial in protecting seeds is likely to be a function of the value of seeds to
granivores (Hulme 1994) and seed size (Hulme and Borelli 1999), different
seeds are likely to reap different levels of protection via burial. Similarly,
although burial may provide protection from seed predators, the likelihood
of successful germination and emergence is influenced by depth and varies
among species (Baskin and Baskin 1998, Grundy et al. 2003). We have
shown that burial depth is important for affecting the removal of one earlysuccessional
species, Phytolacca americana. Future studies manipulating
burial depth of a suite of early-successional plant species are needed to
provide insight into the role of burial depth in affecting the seed limitation
common to early-successional communities (Turnbull et al. 2000).
Acknowledgments
Candice Schneberger and Christy Prenger provided excellent assistance. We
thank John Blake, US Forest Service Savannah River, for logistic support. Funding
and support was provided by the Department of Energy-Savannah River Operations
office through the US Forest Service Savannah River under Interagency Agreement
DE-IA09-00SR22188. Funding was also provided by NSF grant DEB-9907365, a
Professional Advancement Grant from Iowa State University, and a STAR Fellowship
to J.L. Orrock from the US Environmental Protection Agency. Portions of this
work were conducted while J.L. Orrock was a Postdoctoral Associate at the National
Center for Ecological Analysis and Synthesis, a Center funded by the National
Science Foundation (Grant #DEB-0072909), the University of California, and the
Santa Barbara campus.
Literature Cited
Andresen, E., and D.J. Levey. 2004. Effects of dung and seed size on secondary
dispersal, seed predation, and seedling establishment of rain forest trees.
Oecologia 139:45–54.
Armesto, J.J., G.P. Cheplick, and M.J. McDonnell. 1983. Observations on the
reproductive biology of Phytolacca americana (Phytolaccaceae). Bulletin of the
Torrey Botanical Club 110:380–383.
Baskin, C.C., and J.M. Baskin. 1998. Seeds: Ecology, Biogeography, and Evolution
of Dormancy and Germination. Academic Press, New York, NY. 666 pp.
158 Southeastern Naturalist Vol. 6, No. 1
Boman, J.S., and B.B. Casper. 1995. Differential postdispersal seed predation in
disturbed and intact temperate forest. American Midland Naturalist 134:107–116.
Crawley, M.J., and C.R. Long. 1995. Alternate bearing, predator satiation, and
seedling recruitment in Quercus robur. Journal of Ecology 83:683–696.
Damschen, E.I. 2005. Plant Community Response to Landscape Connectivity and
Patch Shape. Ph.D. Dissertation. North Carolina State University, Raleigh, NC.
123 pp.
Grundy, A.C., A. Mead, and S. Burston. 2003. Modelling the emergence response of
weed seeds to burial depth: Interactions with seed density, weight, and shape.
Journal of Applied Ecology 40:757–770.
Hulme, P.E. 1994. Post-dispersal seed predation in grassland: Its magnitude and
sources of variation. Journal of Ecology 84:645–652.
Hulme, P.E., and T.Borelli. 1999. Variability in post-dispersal seed predation in
deciduous woodland: Relative importance of location, seed species, burial, and
density. Plant Ecology 145:149–156.
Hyatt, L.A. 1998. Spatial patterns and causes of overwinter seed mortality in
Phytolacca americana. Canadian Journal of Botany 76:197–203.
Kilgo, J.C., and J.I. Blake (Eds.). 2005. Ecology and Management of a Forested
Landscape: Fifty Years on the Savannah River Site. Island Press, Washington,
DC. 479 pp.
Maron, J.L., and E.L. Simms. 1997. Effect of seed predation on seed bank size and
seedling recruitment of bush lupine (Lupinus arboreus). Oecologia 111:76–83.
Martin, A.C., H.S. Zim, and A.L. Nelson. 1951. American Wildlife and Plants: A
Guide to Food Habits. Dover Publications, New York, NY. 500 pp.
McDonnell, M.J., E.W. Stiles, G.P. Cheplick, and J.J. Armesto. 1984. Bird-dispersal
of Phytolacca americana L. and the influence of fruit removal on subsequent
development. American Journal of Botany 71:895–901.
Mitich, L.W. 1994. Common pokeweed. Weed Technology 8:887–890.
Orrock, J.L., and E.I. Damschen. 2005. Fungi-mediated mortality of seeds of two
old-field plant species. Journal of the Torrey Botanical Society 132:613–617.
Orrock, J.L., B.J. Danielson, M.J. Burns, and D.J. Levey. 2003. Spatial ecology of
predator-prey interactions: Corridors and patch shape influence seed predation.
Ecology 84:2589–2599.
Orrock, J.L., D.J. Levey, B.J. Danielson, and E.I. Damschen. 2006. Seed predation,
not seed limitation, explains the landscape-level abundance of an early-successional
plant. Journal of Ecology 94:838–845.
Quinn, G.P., and M.J. Keough. 2002. Experimental Design and Data Analysis for
Biologists. Cambridge University Press, Cambridge, UK. 520 pp.
Radford, A.E., H.E. Ahles, and C.R. Bell. 1968. Manual of the Vascular Flora of the
Carolinas. University of North Carolina Press, Chapel Hill, NC. 1245 pp.
Reichman, O.J. 1979. Desert granivore foraging and its impact on seed densities and
distributions. Ecology 60:1085–1092.
SAS Institute. 2004. SAS User’s Guide. Version 8.1. SAS Institute, Cary, NC. 3809 pp.
Turnbull, L.A., M.J. Crawley, and M. Rees. 2000. Are plant populations seedlimited?
A review of seed sowing experiments. Oikos 88:225–238.
Willson, M.F., and C.J. Whalen. 1990. Variation in postdispersal survival of vertebrate-
dispersed seeds: Effects of density, habitat, location, season, and species.
Oikos 57:191–198.