ArticlePDF Available

Experimental artifact transport by Harvester ants ( Pogonomyremx sp .): Implications for patterns in the Archaeologial record

Schoville et al.
Article JTa092. All rights reserved. *E-mail:
Journal of Taphonomy
Available online at
On the High Plains of North America the harvester ants Pogonomyrmex occidentalis and P. owyheei
build large gravel covered nest mounds in which artifacts and small fossils are frequently deposited. The
effect this mound building behavior has on the archaeological record has received little attention and has
generally been viewed as restricted to subsurface tunneling disturbances. Initial experiments highlight
the surface foraging behavior of harvester ants actively transporting artifacts during mound construction
and maintenance. Non-food related foraging behavior was investigated by placing glass beads around ant
mounds in various patterns to evaluate foraging distance, direction, density and distribution effects. Ants
were observed to forage a maximum of 48 m from the nest but the majority of foraged materials were
returned from within 20 m, regardless of density, direction, or distribution differences. Of 812 individual
mounds recorded during an extensive landscape survey, 134 contained anthropogenic debris.
Additionally, mounds tend to form near disturbed and eroding soils which enhance the opportunity for
ants to acquire actively exposed artifacts during foraging for mound construction material. These
characteristics of harvester ant foraging effectively create highly visible loci of small artifact
concentrations that are otherwise poorly represented by traditional pedestrian surveys. Understanding the
taphonomic signature of harvester ant artifact transport should aid in refining interpretations of artifact
patterning observed in archaeological contexts.
Experimental Artifact Transport by Harvester
Ants (Pogonomyrmex sp.): Implications for
Patterns in the Archaeological Record
Benjamin J. Schoville*
School of Human Evolution and Social Change, Arizona State University
Tempe, AZ 85287-2402, Arizona, USA
Lucy E. Burris
Graduate Degree Program in Ecology and Department of Anthropology
Colorado State University, Fort Collins, Colorado, USA
Lawrence C. Todd
Greybull River Sustainable Landscape Ecology, Meeteetse, Wyoming, USA
Journal of Taphonomy 7 (4) (2009), 285-303.
Manuscript received 3 September 2008, revised manuscript accepted 3 January 2009.
Artifact transport by Harvester ants
Pogonomyrmex or “bearded ant” (Table 1).
Across the western United States two species,
P. occidentalis and P. owyheei, commonly
known as western harvester ants, build
distinct gravel covered mounds covering
their subterranean nests (Figure 1a). These
mounds are frequently checked for small
bones and small artifacts by fossil collectors
and archaeological surveyors respectively
(Adams, 1984; Bass & Jefferson, 2003;
Galbreath, 1959; Hatcher, 1896; Shipman &
Walker, 1980). Reports mentioning artifacts
located on ant mounds tend to be anecdotal
in nature, rarely considering the effects of
harvester ant nest construction and maintenance
on the archaeological record. In heavily
vegetated environments, the clearings around
ant mounds as well as the materials within
the mound can often be the only indicators
for the existence of small cultural materials
on the surface.
While the social behaviors of ants
including those related to food foraging are
widely studied (Bernstein, 1975; Crist &
MacMahon, 1992; Gordon, 1995; Hölldobler
& Wilson, 1990; Usnick, 2000), comparatively
little systematic study has been concerned
with material foraging by ants and mound
content (Taber, 1998; Todd & Schoville, 2001)
or the size range of transported cultural
items (Dugas, 2001; Jorgensen & Porter,
1982; Mandel & Sorenson, 1982; Nagel, 1969;
Shipman & Walker, 1980). In his discussion
of formation processes, Schiffer (1987:208)
characterizes ants as surface foragers, primarily
influencing the archaeological record through
the creation of tunnels that form fossilized
burrows called krotovina. While Schiffer
recognizes birds and packrats may displace
surface objects, he fails to attribute this
same behavior to ants despite citing Nash
and Petraglia’s (1984:140) work that describes
the occurrence of displaced pressure flakes
Studies of formational dynamics are central
to understanding and interpreting the complex
suite of biological, physical, and cultural
processes that form the archaeological record
(Lyman, 1994; Schiffer, 1987). Small-scale
taphonomic processes such as bioturbation can
confound interpretations of site formation,
but also inform other aspects of paleoecology
and geomorphology relevant to site formation
history. Investigations into the interface of
small-scale agents such as rodents (Fowler et al.,
2004; Johnson, 1989), termites (Mcbrearty,
1990), and earthworms (Darwin, 1881; Feller
et al., 2003; Tryon, 2006), and the archaeological
record have documented the signatures left
by dynamic biological processes acting on a
static record (Balek, 2002). The interactions
between biological organisms and the material
remnants of human behavior may remove some
or all spatial patterning specifically identifiable
to anthropogenic activity. Only by understanding
the nature of these formation processes can
we begin to disentangle and quantify their
taphonomic signatures to the point where more
informed interpretations of the archaeological
record become possible. In this investigation
of the foraging behavior of harvester ants
we demonstrate that ants actively transport
and redeposit a significant portion of small
surface archaeological debris to their nest
mounds. Additionally, mounds are commonly
established within foraging distance of actively
eroding surfaces which allows for active
accumulation of previously buried deposits.
Given the ubiquity of harvester ant mounds
worldwide (Höllodobler & Wilson, 1990:Table
18-1) our results have implications for locating
artifacts through surface surveys as well as
for interpretations of intra-site artifact patterning.
Common bioturbators in North America
include members of the harvester ant genus
Schoville et al.
collecting during subsurface excavations as has
been suggested (Matthias & Carpenter, 2004).
Here we demonstrate that ants have
the ability to transport upwards of 40% of
small introduced surface artifacts within 20
m of their nests regardless of direction from
or density of the introduced artifacts through
experimental investigations of artifact transport
by Pogonomyrmex. All research was conducted
during a multi-faceted field project centered
at the Hudson-Meng Bison Bonebed Research
Facility in Northwest Nebraska, USA (Figure
1b). Experiments were designed to both
elucidate the degree to which ant transport may
be affecting archaeological and paleontological
site formation, and to understand how high-
visibility ant mounds may be better
incorporated into landscape scale surveying
techniques. Archaeological research at the
Hudson-Meng facility was performed from
a landscape taphonomy approach that considers
the formation of the archaeological record to
be a dynamic suite of interplaying processes.
This approach applies methods for investigating
site formation to landscape scale observations
(Burger et al., 2008; Todd & Rapson, 1999).
in an ant mound. Wood and Johnson (1978:
321) and Bass and Jefferson (2003:22-23)
characterized the taphonomic effect of ants
as primarily soil and consequently, artifact
mixers. In contrast, reports by Nagel (1969:
69), Reynolds (1991), Krajick (2001:202),
and Todd & Schoville (2001) stressed that
while some materials on ant mounds may be the
result of nest excavation or soil movement,
large amounts of mound gravel are purposefully
collected from exposed deposits on the
surrounding land surface by harvester ants.
Todd and Schoville (2001) placed colored
glass beads around a single ant mound to
observe ant movement of introduced mound
building material clasts. After a year, the
same ant mound was excavated in 50 cm2
units in 5 cm level increments to investigate
if the final transport location of introduced
artifacts was underground. Beads were only
located within the top 10 cm of the mound
surface and analysis of the screened sediments
indicated that the subsurface sediments did not
contain the same gravel sizes the mound
consisted of, indicating mound building
materials come from the surface and not from
Species Common name Distribution
P. badius Florida harvester ant Florida, coastal regions of Alabama, Mississippi,
Georgia, South and North Carolina.
P. barbatus Red harvester ant Mexico, Texas, southern Arizona and southern New
P. californicus California harvester ant Southern California, Nevada, Arizona, and southern
New Mexico, and northwestern Mexico.
P. maricopa Maricopa harvester ant Northern Mexico, southern Arizona, New Mexico, and
P. occidentalis Western harvester ant Western Great Plains and Great Basin including Montana,
North Dakota, South Dakota, Nebraska, Kansas, Oklahoma,
Colorado, Wyoming, New Mexico, Arizona, and Nevada.
P. salinus (also
known as P. owyheei) In older literature also referred to
as a western harvester ant Washington, Oregon, Nevada, Idaho, Montana, and
northern California.
P. rugosus Rough harvester ant Southern California, Nevada, New Mexico, Arizona,
Texas, western Oklahoma, and Mexico.
Table 1. Common Pogonomyrmex species name and distribution.
Note: Compiled from Taber (1998:104-116,132-134).
Artifact transport by Harvester ants
Figure 1. a) (upper) Typical western harvester ant nests showing mounds and cleared disks at the study site in northwestern Nebraska (Photos courtesy
of 2002 CSUAFS); b) (lower left) study site location; c) (lower right) P. occidentalis forager moving a medium size bead (Photo courtesy of Paul Burnett).
Schoville et al.
maintenance frequency all limit the amount of
time available for foraging for nest construction
material, thus impacting the likelihood
surface artifacts located nearby will be
transported. Seven species of Pogonomyrmex
are commonly recognized in North America
(Table 1). These species are all large bodied,
build graveled nests of variable size on dry
clay loam soils, exhibit large colony populations
when mature, tend to be aggressive, and to
deliver painful stings if disturbed (Taber,
1998; Schoville & Burris, personal experience).
In addition to small object collection, several
other attributes of harvester ant colonies
make them taphonomically important. First,
once colonies are established, they persist
over time and space. P. occidentalis colony
survival has been estimated between 20 and
50+ years (Keeler, 1993). Western harvester
ant nests are infrequently recolonized or
relocated, although recolonization may be far
more frequent in other ants (Gibb & Hochuli,
2003; Taber, 1998). Second, nests of mound-
building species occur in relatively high
densities. Surveys of P. occidentalis have
shown densities ranging from 6-75 nests/ha
with nest spacing from 40 m apart at the
lowest density to 12 m apart at the highest
density (Crist & Wiens, 1996; Keeler, 1993).
Third, food item foraging is extensive and
thorough, spanning a large area surrounding
the nest. Individual Pogonomyrmex foragers
use foraging trails to obtain food from within
20 m of the nest (Crist & MacMahon, 1991;
Crist & Wiens, 1994; Fewell, 1988a, 1988b;
Morehead & Feener, 1998; Rogers, 1972;
Usnick, 2000). From an archaeological
perspective, this foraging area constitutes
the potential search area from which ants
may collect artifacts as mound building
material. Given that foraging trails are
relatively permanent during a single season
(Gordon, 1995), individual ants are likely to
The intended result of this approach is a
more informed model for understanding
human-environmental interaction on the
landscape. Given the prohibitively high cost
of extensive test excavations and the necessity
of large scale (landscape) datasets, archaeologists
often study surface remains through pedestrian
survey (Barton et al., 2004; Burger, 2002).
Locating artifacts is dependent on many
variables including occupation and depositional
history, erosion, and recovery biases in
survey methodology, individual experience,
and surface visibility (Burger et al., 2002;
Wandsnider & Camilli, 1992). Taphonomic
histories informed by landscape survey are
essential for the interpretation of archaeological
spatial patterns (Barton et al., 2002). The
systematic study of ant foraging behavior
and its influence on archaeological patterns can
provide archaeologists with a more complete
understanding of the relevant processes that
may modify the spatial association of cultural
materials and a baseline from which to
evaluate the cumulative impact of nest
material foraging behavior through time. By
understanding relevant features of ant nest
material foraging behavior, we seek to provide
a basis for recognizing and evaluating the
effects of artifact transport by harvester ants.
Archaeologists may recognize these patterns
and produce more compelling inferences of
surface manifestations and site formation
Genus Description
The taphonomic imprint of P. occidentalis
is determined by environmental and behavioral
attributes. Variability in mound and disk size,
foraging distance, nest construction time and
Artifact transport by Harvester ants
Foraging Experiments with Beads
Glass beads introduced into the foraging range
of harvester ants were used to evaluate ant
foraging behavior in relation to differential
availability of mound building materials
(Figure 1c). The study was conducted on a 12-ha
subset of the Hudson-Meng land parcel. All
observations were made in late-spring through
summer with initial bead seeding starting in
early June and final collections made by early
September. Opaque glass beads were used
as gravel surrogates for experimentation based
on similarities between beads and gravel
over a variety of physical characteristics.
While glass beads are cost effective and
uniform in shape and color, seeds with smooth
outlines and very rounded shapes similar to
the beads are generally avoided by harvester
ants, which may influence the overall bead
interest of the ants (Pulliam & Brand, 1975).
Additionally, beads are dissimilar to common
surface artifacts such as chipped stone,
which is often flat and thin. Therefore, we
consider these experiments as capturing the
minimum range of potential impact ants
may have on the archaeological record.
Experiments used up to 12 bead colors (e.g.,
black, white, yellow, red, orange, light
green, dark green, light blue, dark blue,
maroon, light purple, and dusty pink) and 3
sizes of beads (#10/11, ~1.5-2.5 mm diameter
weighing 11 mg; #8, 2.5 mm weighing 25
mg; and #6, 4 mm, weighing 75 mg). The
desired dispersal pattern was 0.25-0.5 m wide
with a dispense density of 400 beads/m2
(based on seed bank density from Crist &
MacMahon, 1992).
Experiments were designed to
investigate foraging distance and direction,
as well as the effect artifact density and
distribution uniformity (i.e., small artifact patch
or large scatter) has on transport response.
reuse successful foraging areas and trails
during their lifetimes (Bernstein, 1975;
Fewell, 1990), and the high turnover rate and
short lifespans of foragers (roughly 2
weeks), it seems likely that nearly all of the
surrounding nest area will eventually be
traversed (Porter & Jorgensen, 1981). Fourth,
in addition to noticeable graveled nests,
harvester ants typically remove all vegetation
from a broad disk of ground surrounding the
mound (Figure 1a). The disk cleared by P.
occidentalis is 1-3 m in diameter; however
old-world harvester ants may clear a disk up
to 20 m in diameter. These disks are frequently
utilized by other animal species for marking
(i.e., wolves and cats), dusting (i.e., bison)
and horning (bison and cattle) (Shipman &
Walker, 1980). The disk may function to
reduce forager transit time, decrease exposure
to grass fires, and increase surface exposure
to solar radiation for warming subsurface
brood chambers (MacMahon et al., 2000).
Ant Mound Distribution
In order to understand the taphonomic
implications of harvester ant artifact transport,
the spatial distribution of mound colonization
as well as the frequency of mounds containing
anthropogenic materials was evaluated. A
pedestrian survey for harvester ant colonies
documented geographic location and nest
attributes of harvester ant mounds on the
1200-ha area surrounding the Hudson-Meng
Bonebed. As part of this survey, the spatial
location of mounds was recorded with
handheld Garmin© GPS receivers. In addition
26 ecological attributes and mound metrics
including presence/absence of prehistoric
artifacts were documented (data available
upon request). In total, 812 ant mounds were
recorded within the Hudson-Meng land parcel.
Schoville et al.
within a 20 cm area 8 m from the mound. In
total, slightly more than 300,000 beads were
dispensed. In addition to monitoring the
experimental and control mounds, 15 mounds
(referred to here as "nearby" mounds) in
close enough proximity to experimental mounds
for ant collection of beads were also monitored.
Ant Mound Distribution
In total, 812 ant mounds were located during
pedestrian survey. Of these, 151 (19%)
contained evidence of chipped stone flakes.
During survey, a pattern of increased mound
frequency along arroyo rims, badland margins,
and roadway and trail edges quickly emerged.
This is consistent with mound colonization
Experiments were conducted on 14 mounds, and
8 mounds served as controls. Experiment
design and deployment has previously been
discussed in Burris (2004) and will only be
summarized here. To assess foraging distance
for mound-building materials, concentric rings
of like colored beads were placed in uniform
density (400 beads/m2) around 5 mounds at
progressively further distances (4, 8, 12, 20,
and 32 m, with one mound receiving an
interval of beads at 48 m from the mound).
The pattern for bead dispensing during a
distance experiment is shown in Figure 2.
Bead color provided a direct indication of
collection distance from the mound. As beads
are often picked up and dropped at different
points by multiple foragers, these distances
do not represent the total distance of a
single ant but rather the cumulative effect of
foraging around the mound.
A second experiment was designed
to assess if ants display any directional
foraging preferences. Different colored beads
were placed in each of 4 quadrants at a
distance of 8 m from the mound, such that
each color of bead could be matched with the
quadrant away from the center of the mound
in which it was deposited. A third experiment
designed to assess the influence of artifact
density on the likelihood of transportation
was performed at 3 mounds. Beads were
deposited 8 m from a mound at one of 3
different bead concentrations (50, 100, and
200 beads/m2). An 8 m distance was selected
to balance two objectives: 1) presenting beads
far enough from the mound to show clear
foraging behavior patterns and 2) presenting
beads close enough to the mound to provide
useful simulations for archaeological survey.
In a fourth experiment, distribution effects
were tested on 4 mounds using a single
randomly oriented deposit that consisted of
400 beads (200 large, 200 medium) dropped
Figure 2. The distance experiment consisted of
concentric bands of uniquely colored beads deposited
around the experimental mound dispensed at a uniform
rate of 400 beads/m2. Only one mound included the
48 m band of beads.
Artifact transport by Harvester ants
Table 3. Summary of Mound Building Material metrics.
disturbed areas. This appears to be true
regardless of presence or absence of artifacts
in the mound (Table 2).
Nest Material Attributes
Estimating the size range of the gravel and
artifacts being transported is an important
facet of interpreting the potential impact of
harvester ant foraging activity on the
archaeological record. Initially, a sample of
100 rocks was collected from the surface of 3
ant mounds. The individual gravel samples
were measured for length, width, and
thickness (Table 3). The average length for all
three mounds combined is 3.2 mm (range
1.5-7.0 mm). Another sample was collected
from a mound located adjacent to a recently
constructed walking trail with distinctive
patterns observed in New Mexico and North
Dakota (DeMers, 1993; Dugas, 2001). On a 1 m
resolution digital orthorectified black and
white aerial photograph, eroding surfaces of
the White River Badlands and the roadways
within the Hudson-Meng survey area appear
white against the darker vegetated grasslands.
These disturbed areas were drawn as polygons
into a GIS and the GPS derived ant mound
locations were added and buffered at 10, 20,
and 48 m to produce an estimated foraging
area of each mound. These two landscape
features were then analyzed for overlap. It
was found that at a 20 m buffer, 76% of the
ant mound foraging areas intersect areas of
erosion, trails, or roadways. At a 48 m buffer,
90% of the ant mound foraging areas intersect
disturbed soils (Table 2). As indicated by
the initial survey observations, the majority
of mounds are located near (within 48 m)
# of mounds # of mounds with
overlapping erosion artifacts overlapping
Disk radius surfaces (n=714) erosion surfaces (n=134)
10 m 458 (64%) 86 (64%)
20 m 542 (76%) 101 (75%)
30 m 589 (82%) 107 (80%)
48 m 644 (90%) 120 (90%)
Table 2. Proximity of Mounds to Disturbed Areas.
Mound Sample
length range
Mean maximum
Mean maximum
Mean maximum
Mound gravel
(n=300) 1.5 – 7.0 3.2 / 0.87 2.4 / 0.58 1.7 / 0.47
Mound adjacent trail
(n=100) 1.9 – 7.7 4.7 / 1.2 3.2 / 0.81 2.0 / 0.61
Trail materials
(n=100) 2.7 – 13.8 6.7 / 2.5 4.8 / 1.9 3.0 / 1.1
Lithic material
(n=389) 1.4 – 11.3 5.3 / 1.6 3.3 / 1.1 1.1 / 0.68
Schoville et al.
roughly 20 lithic flakes visible on the
surface were excavated and water screened
through 1/16” mesh screen. After screening
the lithic artifacts were removed, counted,
and measured. The average maximum
length of stone flakes was 5.3 mm with a
measured range from 1.4-11.3 mm (Figure
3, Table 3). In total, 389 pieces of chipped
stone were recovered from the two mounds,
which is noteworthy given the much lower
number of surface artifacts visible prior to
excavation. Potential subsurface abundance
should be taken into account when
interpreting the results of the mound survey
which only documented stone artifacts on
the nest surface. Therefore, the surface
survey results should be treated as minimum
estimates of the surface disturbance from
ant collection. No clear patterns in the type
of stone collected were apparent, indicating
that the flakes were the result of multiple
knapping events and were thus obtained
from multiple artifact material collection
non-local crushed stone used as trail
pavement. This mound was covered in the same
non-local stone, indicating that the harvester
ants foraged gravel from the trail for nest
construction. The mean maximum length of
gravels on the mound was 4.7 mm (range
1.9-7.7 mm), whereas the mean length of
trail gravel was 6.7 mm with a much larger
range of gravel size (2.7-13.8 mm). While
indicating a preferential selection of certain
gravel sizes for mound construction, these
data also suggest that ant foraging operates
within constraints where some stones are
too small to make transport valuable and
some are too large to effectively transport.
All of the non-cultural stones measured
are fairly cube shaped with roughly equal
average maximum length and width (3.6 and
2.6 mm respectively). This is in contrast to
the general shape of small lithic debris resulting
from tool manufacture and resharpening
which is often flat and thin. To evaluate the
size characteristics of stone artifacts
transported onto ant nests, two mounds with
Figure 3. Boxplot of maximum dimensions (MLEN= maximum length, MWID= maximum width, MTHK=
maximum thickness) of lithic debitage removed from two ant mounds (N=389).
Artifact transport by Harvester ants
greater than 19 m were collected after week
five. Additionally, 4 beads were returned from
48 m away from the single mound where this
bead distance was tested for.
Direction. Beads were recovered from
all 4 quadrants in both of the colonies tested
for directional preferences (Figure 4b, Table 5).
Although there does appear to be some
preference for collection from the Southeast
quadrant, directional differences on Mound 6
were not significant (Rayleigh's test for
randomness of direction, z=1.78, z0.01=4.6,
p>0.10, see Batschelet, 1981).
Density. Beads were recovered from
all 3 density levels (Figure 4c, Table 5). As
would be expected, beads from the lowest
density (50 beads/m2) were the least represented
while beads from the two higher density
bands were transported to the mound equally.
However, there does not appear to be a clear
pattern in variable density return rate.
Distribution. Beads were recovered
from only 2 of the distribution experiments
Foraging Experiments with Beads
The ants readily collected and returned
beads to their mounds (Figure 1c). Often
within the first 15 minutes of dispensing
beads ants began placing beads onto the
mound surface. Approximately 5700 beads
were recovered by the end of the study
period. An estimated 1600 beads remained
on the mounds after week 13. Since no
attempt was made to determine the number
of beads incorporated into the mounds, the
results presented here represent a lower
bound on ant foraging quantities. No pattern
was observed with respect to how beads
were arranged on the mounds.
Distance. Beads were collected from
20 m on all mounds and 2 of the 5 mounds
returned beads representative of every
distance band (Figure 4a, Table 4). There is
some indication that foraging distance
increased as the summer season progressed
since most beads collected from distances
Table 4. Distance Results for Experimental Mounds.
Return Rate / Bead Count
Mound 4 m 8 m 12 m 20 m 32 m 48 m
1* 10.85 % (293) 4.38 % (591) - 0.22 % (30) 0.00 % (0) -
2 14.74 % (398) 11.72 % (633) 7.42 % (601) 0.12 % (16) 0.03 % (6) 0.01 % (4)
3 22.52 % (608) 12.17 % (657) 4.19 % (339) 0.07 % (9) 0.00 % (0) -
4 11.56 % (312) 8.07 % (436) 2.93 % (237) 1.43 % (193) 0.31 % (68) -
5** 3.07 % (83) 0.78 % (42) 0.07 % (6) 0.01 % (1) 0.00 % (0) -
(95 % C.I.
of Mean)
14.9 %a
(10.4 - 19.5)
9.1 %a,b
(6.0 - 12.2)
4.84 %b
(2.2 - 7.5)
0.46 %c
(-0.1 - 1.0)
0.09 %c
(-0.04 - 0.22)
0.01 %
* During setup, about half of the dark blue beads for the 12-m band were applied at the 8-m mark
** Mound destroyed after 4 weeks, bead count and return rate are for this short period only. Excluded from
Summary values.
a,b,c Differences between means with the same superscript are not significantly different (p < 0.05).
Collected Bead Count is the sum total of all beads removed from the mound during the 13 week observation
period. Deposited Bead / Band is the number of beads dispensed during the experimental setup. Return Rate is the
ratio of collected beads to dispensed beads.
Schoville et al.
These values reflect only surface collection
and provide a lower bound on the number of
beads actually transported by ants to their
All colors and sizes of beads were
collected; although there was a clear size
preference in that 66% of the collected beads
were medium size, compared to just 22%
for small and 11% for large (Figure 5).
Large beads were retrieved from as far as 14 m
on 5 mounds (n=66). Medium size beads were
retrieved from distances out to 48 m on one
mound and from distances between 20 m and
32 m on six mounds (n=208). Five colonies
collected small beads out to the same
distance (n=87). Since bead color was used as
a distance marker and not used as a treatment
factor, there was insufficient replication to
(13 and 14) (Figure 4d, Table 5). Minimal
ant activity was noted at the other 2
mounds. Lack of collection by ants at these
mounds may have been due to decreased
colony foraging vigor or colony death.
These results suggest that discrete sources
are much less likely to be detected than are
uniform scatters of material, even at low
densities. However, collection rates can be
quite high.
Overall. In total, less than 3% of the
dispensed beads were recovered from
mounds (7,300 / 310,000) although 8.4% of the
beads dispensed within 12 m of mounds
were collected and 4.8% of beads dispensed
within 20 m were collected. In one experiment
81 of 200 available medium size beads were
collected resulting in a 40% retrieval rate.
Table 5. Direction, Density, and Disturbance Results.
Experiment Mound ID Configuration Detail Collected
Bead Count Deposited
Bead Count Return
Direction 6 East 84 800 10.50 %
6 South 16 800 2.0 %
6 West 68 800 8.5 %
6 North 20 800 2.5 %
7 Southeast 134 800 16.75 %
7 Southwest 81 800 10.1 %
7 Northeast 20 800 2.5 %
7 Northwest 87 800 10.9 %
Density 8 200 beads / m2 85 1600 5.31 %
9 100 beads / m2 83 800 10.38 %
10 50 beads / m2 31 400 7.75 %
Distribution 11 400 beads in a 20 cm pile 0 400 0
12 400 beads in a 20 cm pile 0 400 0
13 400 beads in a 20 cm pile 5 400 1.25%
14 400 beads in a 20 cm pile 85 400 21.25
a Collected Bead Count is the sum total of all beads removed from the mound during the 13 week observation
b Deposited Bead Count is the number of beads dispensed during the experimental setup.
c Return Rate is the ratio of collected beads to dispensed beads.
Artifact transport by Harvester ants
Figure 5. Return rates by size
for all mounds to a maximum
distance of 35 m.
Figure 4. a) (upper left) Distance results, Mound 5 was destroyed four weeks after experiment setup. Some 12-m beads
on Mound 1 were misapplied at the 8-m ring, graphed line reflects an estimate of the beads at 12-m (see Table 3 for
values); b) (upper right) Direction results. Bead count radiates outward from center. Outer ring is 150 beads. Lighter
wedges are Mound 6, darker Mound 7. Quadrant orientations are aligned to the door orientation of each mound. The
southeast quadrant of Mound 7 was within 7 m of similarly colored beads at Mound 3 so that the effective bead density
of that quadrant was doubled; c) (lower left) Density results. Mound numbers are given above each bar; d) (lower right)
Distribution results. Mound 13 collected an additional 20 beads from a nearby experiment. Mound 11 and Mound 12
were inactive during most of the observation periods.
Schoville et al.
either larger or smaller objects. Artifacts
removed from 2 ant mounds were on
average 5.3 mm in maximum length and 1.0 mm
to 11.6 mm in range; non-cultural building
material tends to be in the 2–3 mm range.
Ants reliably collect materials around their
mounds with densities as low as 50 items/m2
but are less reliable at locating discrete deposits
as observed during the summer time frame
of these experiments. The presence of
archaeological materials on a nest site is a
good indicator of small material on the
surrounding landscape within a 20 m radius.
Ants respond quickly to the arrival of new
objects within their foraging range although
discovery can take several days to weeks
depending upon distance of materials from
the nest. There is a strong indication that ants
forage in all quadrants around the nest such
that cardinal direction does not influence
nest material procurement behavior.
The occurrence of harvester ant mound
material foraging from prehistoric assemblages
may be evaluated in lithic assemblages for
absence of those size classes known to be
preferentially transported. It is commonly
heard in introductory archaeology courses
that archaeological assemblages are the residue
of material culture, but in this case the
archaeological assemblage may also be the
harvester ant residual assemblage of flakes
not foraged for mound building. Harvester
ant preference for the medium size beads
over small beads at distances of 4 m or more
from the nest is in agreement with previous
reports that ants will return heavier items
from further distances. This essentially may
maximize the return value for their travel
time costs. Since small and large beads were
generally as equally available as medium
beads, it appears that ants rejected small and
large beads in favor of medium beads. This
deliberate selection of medium size materials
determine any color preference, however all
colors were collected.
At the end of the observation period,
15 of the 17 experimental mounds contained
beads and 12 of the 19 nearby mounds
included beads, while only 1 of the 8 control
mounds had beads. It is apparent that the
experimental treatment given to each mound
influences the bead return rate. Based on a
G-test (log-likelihood ratio test) for
independence with Williams correction
(Sokal & Rohlf, 1995), a null hypothesis
that beads appear on mounds independent of
treatment or distance from beads can be
rejected (G adjusted=13.2 tested against a
critical value of χ2.005,2=0.6). The number of
beads found on experimental mounds ranged
from 0 (in the 2 mounds tested for distribution
effects) to 1659, average number of beads
per mound was 285 (95% CI: 184 to 386,
sd=193, using a square root transformation).
The number of beads on control mounds
ranged from 0 (n=8) to 2 (n=1).
Western harvester ants are active taphonomic
agents with respect to small anthropogenic
materials located near their nests, moving
minimally from 8-40% of materials introduced
in their foraging area onto the surface of
their nests. They routinely forage for gravel-
like material within 20 m of their nests and
may go considerably farther if conditions are
favorable. However, the majority of foraging
occurs within 15 m of the nest. Ants were
observed transporting heavy and awkward
loads (such as the large 75 mg beads used
for testing) from as far away as 12 m. In
general, when 3 different item sizes are
offered, loads in a weight range and/or size
range of 25 mg or 2.5 mm are preferred over
Artifact transport by Harvester ants
documented fully, is that the artifacts found
in mounds often have a very high lithic raw
material diversity, which probably results
from the ants having sampled multiple tool
reduction events. One would expect much
lower raw material diversity at in situ tool
maintenance localities.
The spatial distribution of mound
colonization also has interpretive implications
for recognizing the potential impacts ant foraging
may have on deposited cultural materials.
Both anthropogenic and non-anthropogenic
occupation and behavioral residues left on
the edges of disturbed areas (e.g., trails, paths,
or ant disks and rodent backdirt piles), along
transition margins (e.g., from sod table to
badland exposure) or arroyo edges tend to be
the regions where ant mounds are more
frequently established and in greater densities
than in wetter, more heavily vegetated, or
unbroken landscapes. As erosion or vegetation
clearing encroaches onto previously buried
deposits, it is also the ant colonies in closer
proximity to these disturbance edges that
will more readily transport artifacts. This
could result either from exposure of buried
deposits or from “poaching” construction
materials from abandoned mounds that may
contain artifacts. Through geomorphological
assessments of the landscape the original
deposition environment may be inferred
which could cue researchers to the possibility
of increased artifact winnowing due to
harvester ant collection. With further research
this interpretive process could potentially
work in reverse, whereby the observed gaps
in flake sizes (on average 25 mg in weight
and 5.3 mm in maximum length, or the size
of the “preferred” medium-size beads and the
average lithic flake size found in mounds)
could indicate areas that may have been
more frequently traversed by harvester ants
(i.e., trails or arroyo edges).
has the potential to help discern whether a
small debris scatter encountered during
archaeological excavation is intact or has
been subjected to size winnowing by ant
foraging. Conceivably, any debris scatter
harvested by ants will be depleted in
"medium" size materials relative to "large"
and "small" materials. The nearby harvester
ant mounds in contrast will be enriched in
“medium” size materials that maximize the
ant return rate of gravel size given the costs
associated with transportation. A preferred
"medium" size is likely to be on the order of
25 mg and on average 5.3 mm in length. The
association of natural material of a similar
size with "medium" anthropogenic material is
an indication of a collapsed ant mound rather
than an intact cultural deposit. Since the number
of large beads dropped off rapidly with distance,
the ratio of small : medium : large anthropogenic
materials on a mound can potentially provide
some indication of how far the cultural
deposit is from the mound. For example, if
"large" materials (e.g., 75 mg beads) are on a
mound cultural deposits may more likely be
less than 12 m away. These results emphasize
the need for fine grained screening and
collection methods during archaeological
excavation and survey.
Conversely, areas in which mounds
are or have been formerly active may create
aggregates of like-sized artifacts. It is fairly
easy to envision a moribund mound being
misinterpreted as a stone tool rejuvenation
area. The “medium” sized debitage often
found in mounds is often composed of retouch
flakes, sometimes with several hundred
clustered in less than several square meters.
As with the “lag” deposits from which these
materials have been removed, a distinctive
pattern of item size may be one of the keys
to identifying such clusters. A second potential
feature that we have noted, but not yet
Schoville et al.
insight into material sourcing, occupancy
duration, and site extent. Similarly, other small
items such as ceramic fragments and glass
trade beads can indicate local and distant trade
patterns as well as site occupation period.
Although this study was specifically designed
to address the archaeological implications
of harvester ant artifact transport, the same
suite of foraging behaviors would seem to
apply to the transport of paleontological
fossils, local vegetation seeds, small geologic
gravels, and perhaps a suite of other
regionally discrete materials that may be
foraged for nest building materials and be of
a more general interest. In this light,
paleontologists have long targeted mounds
for the same size range of fossils that we
report for artifacts and would include many
skeletal elements of micromammals -especially
high density elements such as teeth and
tarsals (Lyman, 1994). Examination of ant
mounds has long provided vertebrate
paleontologists with glimpses of many of
the smaller members of fossil communities.
Fossil remains would be likely to appear in
mounds along disturbed areas and foraged
from within a 20 m radius of the nest. While
the absence of fossil evidence on mounds
certainly doesn’t necessitate an absence of
larger size fossils, small fossil remains should
cue researchers in to the surrounding
surface for more remains. A systematic
assessment of harvester ant foraging for
paleontological fossils is beyond the scope
of this paper, however such a study would
be valuable and it is hoped that the research
presented here would benefit such an
endeavor. Whereas traditional use of ant mounds
in paleontological work has largely been as
Incorporating systematic harvester
ant mound inspection as a component of
surface pedestrian surveys may also enable
archaeologists to quickly and efficiently not
only locate surface artifacts where they are
exposed, but also document very small
artifacts which may otherwise be
underrepresented by even the most stringent
of survey methodologies (Burger et al., 2002).
In a situation where stone tools are highly
curated, it is not difficult to imagine that the
main indications of landscape use will come
from small, non-diagnostic flakes and isolated
tools that are difficult to locate during
pedestrian surveys (Wandsnider & Camilli,
1992). Including small artifacts found
within mounds into broader landscape scale
interpretations of human land use patterns
may provide support for interpretations of
behavioral and taphonomic processes
(Barton et al., 2002). While not being a clear
indicator for the presence of artifacts,
harvester ant mounds possess many features
beneficial to archaeological surveys including
high visibility even at coarse transect spacing,
broad geographic distribution, and reliable
foraging behavior within 20 m of their nest.
Finally, inspection of ant mounds
allows archaeologists to make inference about
parts of the record which are no longer
present. In many parts of the western Great
Plains, collectors and hobbyists have been
actively removing both diagnostic projectile
points and worked lithic material since at
least the 1920s (LaBelle, 1997, 2003; Westfall,
2007). While some collectors may actively
record and research their collections many
collect for trade or sale and information
about archaeological context is often not
collected and lost. Lithic microdebitage may
be the only remaining indicator of originally
more diverse sites. Although small, the variety
and distribution of this debitage can provide
Artifact transport by Harvester ants
In addition given the tendency for
mounds to be located near erosional contacts,
the materials aggregated in mounds may
represent archaeological sources derived
from multiple stratigraphic levels. This potential
time averaging in mound formation means
that a multiplicity of contextually distinct
episodes of stone tool use and maintenance
may be represented within any given mound
artifact collection. In developing interpretations
of regional records that have been modified
by harvester ants, differing size ranges of
artifactual material may represent different
ranges of temporal specificity.
The unobserved action of continual
artifact transport by harvester ants across
landscape surfaces over millennia has
affected the surface record in ways directly
related to the ongoing cultural, biological,
and physical processes that affect mound
colonization, human occupation, and landscape
taphonomy. Rather than being deleterious to
our understanding of cultural landscapes,
understanding the impacts of harvest ant
foraging has the ability to enhance our
interpretations of complex landscape processes
and patterns of human land use behavior as
documented by large-scale pedestrian
surveying of surface artifacts and small-scale
examination of harvester ant mounds.
We would like to thank the valuable input
from Briana Pobiner and one anonymous
reviewer which greatly improved the quality
of this manuscript. Additional valuable
comments on previous versions of this draft
were provided by Oskar Burger, Paul Burnett,
Erik Otárola-Castillo, Curtis Marean, and
Sarah Lansing. All remaining errors are the
sole responsibility of the authors. We thank
a discovery tool, we argue that having a better
understanding of the foraging behaviors of the
ants will enable better integration of specimens
collected from mounds into more comprehensive
paleoecological interpretations.
In The Formation of Vegetable Mould
Through the Action of Worms, Darwin (1881)
describes the cumulative action of earthworms
on soil. Darwin observed earthworms
completely bioturbating the soil, highlighting
just how dramatically small reoccurring
processes can have very large, seemingly
disproportionate relative to their size, effects
on their environment (Brown et al., 2003;
Johnson, 2002). Given that harvester ants
also operate at small spatial scale, but may
also have long-term cumulative impacts on
both the biophysical and archaeological
components of landscapes, there are several
fundamental reasons why they must be
considered a significant taphonomic factor
on the North American Plains.
First, they have the potential for
adding a systematic size-sorting to cultural
materials on or near the surface. This sorting
is not just of the vertical type observed
among earthworms, pocket gophers, and other
burrowing creatures, but also, as demonstrated
here, can have a substantial pattern of
horizontal displacement. Not only do
harvester ants sometimes transport materials
for nest construction and maintenance over
considerable distances, by the very nature of
their task, these materials can be
simultaneously dispersed from their
locations of primary deposition and aggregated
into secondary storage at the mounds. Over
century to millennial time scales, the multiple
life cycles of spatially disjunct, but often
proximate, Pogonomyrmex colonies and their
mounds, can play a major role in the
formation of a regional archaeological
Schoville et al.
Burris, L.E. (2004). Harvester ant mounds: utility for
small object detection in archaeology. Unpublished
master's thesis, Colorado State University, Fort
Crist, T.O. & MacMahon, J.A. (1991). Individual foraging
components of harvester ants: movement patterns
and seed patch fidelity. Insectes Sociaux, 38: 379-
Crist, T.O. & MacMahon, J.A. (1992). Harvester ant
foraging and shrub-steppe seeds: interactions of
seed resources and seed use. Ecology, 73: 1768-
Crist, T.O. & Wiens, J.A. (1994). Scale effects of
vegetation on forager movement and seed harvesting
by ants. Oikos, 69: 37-46.
Crist, T.O. & Wiens, J.A. (1996). The distribution of
ant colonies in a semiarid landscape: Implications
for community and ecosystem processes. Oikos,
76: 301-311.
Darwin, C. (1881). The formation of vegetable mould
through the action of worms with some
observations on their habits. John Murray, London.
DeMers, M.N. (1993). Roadside ditches as corridors
for range expansion of the western harvester ant
(Pogonomyrmex occidentalis Cresson). Landscape
Ecology, 8: 93-102.
Dugas, D.P. (2001). The influence of arroyo edges on
Pogonomyrmex rugosus nest distribution in the
Chihuahuan Desert, New Mexico. Journal of Arid
Environments, 47: 153-159.
Feller, C., Brown, G. G., Blanchart, E., Deleporte, P.,
& Chernyanskii, S.S. (2003). Charles Darwin,
earthworms and the natural sciences: various
lessons from past to future. Agriculture Ecosystems
and Environment, 99: 29-49.
Fewell, J.H. (1988a). Energetic and time costs of
foraging in harvester ants, Pogonomyrmex occidentalis.
Behavioral Ecology and Sociobiology, 22: 401-408.
Fewell, J.H. (1988b). Variation in foraging patterns of
the western harvester ant, Pogonomyrmex occidentalis,
in relation to variation in habitat structure. In
(Jeanne, R.L., ed.) Interindividual behavioral variability
in social insects, Boulder: Westview Press, pp.
Fewell, J.H. (1990). Directional fidelity as a foraging
constraint in the western harvester ant, Pogonomyrmex
occidentalis. Oecologia, 82: 45-51.
Fowler, K. D., Greenfield, H.J. & van Schalkwyk, L.O.
(2004). The effects of burrowing activity on
archaeological sites: Ndondondwane, South Africa.
Geoarchaeology, 19: 441-470.
Galbreath, E.C. (1959). Collecting fossils from harvester
ant mounds. Transactions of the Kansas Academy
of Science, 62: 173-174.
the 2001-2003 Colorado State University
Field School students at Hudson-Meng for
their hard work during hot Nebraska
summers watching ants when they thought
they would be doing archaeology. Funding
for this project was provided by Colorado
State University, the Nebraska National
Forest and Hudson-Meng research facility.
Adams, D.B. (1984). Fossil hunters best friend is an
ant called pogo: paleontologists use insects to find
small bones. Smithsonian, 15: 99-104.
Balek, C.L. (2002). Buried artifacts in stable upland
sites and the role of bioturbation: a review.
Geoarchaeology, 17: 41-51.
Barton, C.M., Bernabeu, J., Aura, J.E., Garcia, O. &
Roca, N.L. (2002). Dynamic landscapes, artifact
taphonomy, and landuse modeling in the western
Mediterranean. Geoarchaeology, 17: 155-190.
Barton, C.M., Bernabeu, J., Aura, J.E., Garcia, O., Schmich,
S. & Molina, L. (2004). Long-term socioecology and
contingent landscapes. Journal of Archaeological
Method and Theory, 11: 253-295.
Bass, B. & Jefferson, J. (2003). Death's acre: inside the
legendary forensic lab - the Body Farm - where the
dead do tell tales. G.P. Putnam's Sons, New York.
Batschelet, E. (1981). Circular statistics in biology.
Academic Press, New York.
Bernstein, R.A. (1975). Foraging strategies of ants in
response to variable food density. Ecology, 56:
Brown, G.G., Feller, C., Blanchart, E., Deleporte, P. &
Chernyanskii, S.S. (2003). With Darwin, earthworms
turn intelligent and become human friends.
Pedobiologia, 47: 924-933.
Burger, O. (2002). A multi-scale perspective for
archaeological survey. Unpublished master's thesis,
Colorado State University, Fort Collins.
Burger, O., Todd, L.C. & Burnett, P. (2008). The
behavior of surface artifacts: building a landscape
taphonomy on the high plains. In (Clark, B. & Sheiber,
L.L., eds.) Archaeological Landscapes on the High
Plains, Boulder: University Press of Colorado.
Burger, O., Todd, L.C., Burnett, P., Stohlgren, T.J. &
Stephens, D. (2002). Multi-scale and nested-
intensity sampling techniques for archaeological
survey. Journal of Field Archaeology, 29: 409-
Artifact transport by Harvester ants
Morehead, S.A. & Feener, Jr., D.H. (1998). Foraging
behavior and morphology: seed selection in the
harvester ant genus, Pogonomyrmex. Oecologia,
114: 548-555.
Nagel, H.G. (1969). Western harvester ants in Kansas:
colony founding, nest structure and function,
factors affecting density, and effect on soil
formation. Ph.D. dissertation, Kansas State University,
Nash, D.T. & Petraglia, M.D. (1984). Natural disturbance
processes: a preliminary report on experiments in
Jemez Canyon, New Mexico. Haliksa'i: University
of New Mexico Contributions to Anthropology, 3:
Porter, S.D. & Jorgensen, C.D. (1981). Foragers of the
harvester ant, Pogonomyrmex owyheei: a disposable
caste? Behavioral Ecology and Sociobiology, 9:
Pulliam, H.R. & Brand, M.R. (1975). The production
utilization of seeds in plains grassland of southeastern
Arizona. Ecology, 56: 1158-1166.
Reynolds, T.D. (1991). Movement of gravel by the
'Owyhee' harvester ant, Pogonomyrmex salinus
(Hymenoptera: Formicidae). Entomological News,
102: 118-124.
Rogers, L.E. (1972). The ecological effects of the western
harvester ant (Pogonomyrmex occidentalis) in the
short grass plains ecosystem. Ph.D. dissertation,
University of Wyoming, Laramie.
Schiffer, M. B. (1987). Formation processes of the
archaeological record. University of Utah Press,
Salt Lake City.
Shipman, P. & Walker, A. (1980). Bone-collecting by
harvester ants. Paleobiology, 6: 496-502.
Sokal, R.R. & Rohlf, F.J. (1995). Biometry. 3rd ed. W. H.
Freeman and Co., San Francisco.
Taber, S.W. (1998). The world of the harvester ants. 1
ed. Natural History Series 23. Texas A&M University
Press, College Station.
Todd, L.C. & Rapson, D.J. (1999). Formational analysis of
bison bonebeds and interpretation of Paleoindian
subsistence. In (Brugal, J.P., David, F., Enloe, J.G. &
Jaubert, J., eds.) Le Bison: Gibier et Moyen de
Subsistance des Hommes du Paléolithique aux
Paléoindiens des Grandes Plaines. France: APDCA,
Antibes, pp. 479-499.
Todd, L.C. & Schoville, B.J. (2001). Small-scale processes,
long-term archaeological consequences: harvester
ant nest maintenance. Poster Presented at the 59th
Annual Plains Archaeological Conference, Lincoln,
Tryon, C.A. (2006). The destructive potential of earthworms
on the archaeological record. Journal of Field
Archaeology, 31: 199-202.
Gibb, H. & Hochuli, D.F. (2003). Nest relocation in the
golden spiny ant, Polyrhachis ammon:
environmental cues and temporal castes. Insectes
Sociaux, 50: 323-329.
Gordon, D.M. (1995). The development of an ant colony's
foraging range. Animal Behaviour, 49: 649-659.
Hatcher, J.B. (1896). Some locations for Laramie
mammals and horned dinosaurs. The American
Naturalist, 30: 112-120.
Hölldobler, B. & Wilson, E.O. (1990). The ants. The
Belknap Press of Harvard University Press, Cambridge.
Johnson, D.L. (1989). Subsurface stone lines, stone
zones, artifact-manuport layers, and biomantles
produced by bioturbation via pocket gophers
(Thomomys-bottae). American Antiquity, 54: 370-389.
Johnson, D.L. (2002). Darwin would be proud:
bioturbation, dynamic denudation, and the power
of theory in science. Geoarchaeology, 17: 7-40.
Jorgensen, C.D. & Porter, S.D. (1982). Foraging behavior
of Pogonomyrmex owyheei in southeast Idaho.
Environmental Entomology, 11: 381-384.
Keeler, K.H. (1993). Fifteen years of colony dynamics
in Pogonomyrmex occidentalis, the western harvester
ant, in western Nebraska. Southwestern Naturalist,
38: 286-289.
Krajick, K. (2001). Barren lands: an epic search for
diamonds in the North American Arctic. Henry
Holt and Company, New York.
LaBelle, J.M. (2003). Coffee cans and Folsom points: Why
we cannot continue to ignore the artifact collectors.
In (Zimmerman, L.J., Vitelli, K.D. & Holloweel-
Zimmer, J., eds.) Ethical issues in archaeology. Walnut
Creek: AltaMira Press, pp. 115-127.
LaBelle, J.M. (1997). "Uncle Bill" and Panhandle
Paleoindians. Electronic document acceessed
4/15/2008 at
Lyman, R.L. (1994). Vertebrate taphonomy. Cambridge
University Press, Cambridge.
MacMahon, J.A., Mull, J.F. & Crist, T.O. (2000).
Harvester ants (Pogonomyrmex spp.): their community
and ecosystem influences. Annual Review of
Ecology and Systematics, 31: 265-291.
Mandel, R.D. & Sorenson, C.J. (1982). The role of the
western harvester ant (Pogonomyrmex occidentalis)
in soil formation. Soil Science Society of America
Journal, 46: 785-788.
Matthias, A. & Carpenter, K. (2004). Experimental fossil
and glass bead collecting by the harvester ant.
Neues Jahrbuch Fur Geologie Und Palaontologie-
Monatshefte, 2: 80-86.
Mcbrearty, S. (1990). Consider the humble termite - termites
as agents of postdepositional disturbance at African
archaeological sites. Journal of Archaeological
Science, 17: 111-143.
Schoville et al.
Westfall, T. (2007). Sacred Ground: Treasures in
Family and Stone. Airleaf Publishing.
Wood, W.R. & Johnson, D.L. (1978). A survey of
disturbance processes in archaeological site formation.
In (Schiffer, M.) Advances in Archaeological
Method and Theory, Vol. 1. Academic Press, New
York, pp. 315-318.
Usnick, S.J. (2000). Foraging distance of Pogonomyrmex
occidentalis (Hymenoptera: Formicidae) on grazed
and ungrazed shortgrass prairies in Colorado.
Entomological News, 111: 201-205.
Wandsnider, L. & Camilli, E.L. (1992). The character
of surface archaeological deposits and its influence
on survey accuracy. Journal of Field Archaeology,
19: 169-188.
... The second relationship needed is a measure of how the local effects of ants decline with distance from a mound. As central place foragers, western harvester ants concentrate most activity close to their mounds, with most foraging activity occurring within 10 meters of their mound (Crist and MacMahon 1991a, b, MacMahon et al. 2000, Schoville et al. 2009). While this general pattern is well-known, previous studies of harvester ant effects have generally not focused on spatial issues, and thus have not explicitly quantified this declining rate of activity or effects of ants with distance. ...
... Western harvester ants and other harvester ant species can have a significant effect on vegetation composition and soil nutrient content (MacMahon et al. 2000, Whitford et al. 2008, DeFalco et al. 2009), with individual colonies surviving for up to 30 years and maintaining these effects over long time spans (Gordon 2010). The majority of P. occidentalis foraging, and hence their likely influence, occurs at limited distances ranging out to about 10 m from each mound center (Crist and MacMahon 1991a, b, MacMahon et al. 2000, Schoville et al. 2009). There is, however, also variability in the distance at which P. occidentalis ants have been shown to concentrate their foraging activity. ...
... There is, however, also variability in the distance at which P. occidentalis ants have been shown to concentrate their foraging activity. For example, an intensive study of western harvester ants moving beads (to simulate grains) observed ants foraging up to 48 m away, but found the majority of foraging activity within 20 m of each the mound (Schoville et al. 2009). Previous studies put foraging distances for this same species at between three meters (Carlson andWhitford 1991) and25 m (MacMahon et al. 2000), with multiple studies finding a range of ;10 m (Crist and MacMahon 1991a, b). ...
Full-text available
In many ecosystems, foundational species create spatial patterns that structure a broader community. It is unclear, however, how robust these patterns are across large areas and strong environmental gradients, and how the landscape-level consequences of these patterns may vary. We investigated the robustness of non-random patterning in the dispersion of the western harvester ant (Pogonomyrmex occidentalis), a widely recognized ecosystem engineer of western North America. We used remote imagery to characterize the spatial structure and densities of western harvester ant mounds at sites spanning their range within the sagebrush steppe and short-grass prairie areas of Wyoming (581 3 450 km area). We found that ant mound densities varied substantially across the study region, but that mounds were strongly and consistently overdispersed (regularly patterned) across both climatic gradients and mound densities. Precipitation was the only abiotic factor that significantly affected either density or pattern, with stronger patterning among mounds at drier sites. This robustness in ecological patterning is likely to have strong effects on community function; mound dispersion increased the fraction of the landscape within typical ant foraging distances up to 30% over what density alone would predict. We estimated how patterning can modify one key ant effect at a landscape level by combining mound dispersion data with information from a seed removal experiment. Randomization tests based on these results showed that in a representative area, overdispersion could increase the mean landscape-wide seed removal rate by 16%, and decrease its spatial variance by 50%. Western harvester ants are known to affect multiple aspects of community function and structure at a relatively fine scale, and our results show that their spatial dispersion may therefore influence many features of interspecific interactions and community dynamics.
... In areas with exposures of fossiliferous rocks, microvertebrate fossils can be concentrated on these mounds through the foraging behavior of the ants, allowing paleontologists to recover a large number of otherwise difficult to locate fossils in a short period of time. Originally sampled to speed the recovery of typically rare mammalian fossils from Late Cretaceous rocks in Western North America (e.g., Hatcher, 1896;Lull, 1915), the practice of targeting P. occidentalis ant mounds (and other species with similar habits) for microvertebrate fossils was eventually adopted by paleontologists studying a wide range of taxa from a broad span of geologic time, as well as archaeologists searching for small bones and artifacts (e.g., Shipman and Walker, 1980;Bass and Jefferson, 2003;Schoville et al., 2009). ...
... After one year he returned to the ant mound and was able to estimate the ants had foraged beads from at least 50 feet (15.2 meters) away from the mound. An extensive study of the foraging habits of P. occidentalis was conducted by Schoville et al. (2009) in an area just a few miles to the west of the current study site (Figure 2: AS). One of the experiments conducted in that study involved the placement of rings of different colored beads at set distances from four ant mounds (a fifth mound was originally included but was destroyed after four weeks), allowing the foraging distance of each ant mound to be easily visually assessed for distances up to 48 meters away. ...
... One of the experiments conducted in that study involved the placement of rings of different colored beads at set distances from four ant mounds (a fifth mound was originally included but was destroyed after four weeks), allowing the foraging distance of each ant mound to be easily visually assessed for distances up to 48 meters away. At the conclusion of the study (13 weeks) they noted that the vast majority of the beads foraged (~95 percent) came from within 20 meters of the mound, with only two of the mounds foraging beads from between 20 and 48 meters away (Schoville et al., 2009). For the purposes of this study, we calculated two stratigraphic ranges of surface sampling for each ant mound: the stratigraphic range sampled within 20 meters of the mound and stratigraphic range sampled out to 48 meters from the mound (Figure 44). ...
Full-text available
For over a century the mounds of harvester ants (Pogonomyrmex spp.) have been targeted by paleontological field crews as sources of concentrated microvertebrate fossils that facilitate the collection of large numbers of specimens with a small investment of time. This study describes a collection of over 6,000 identifiable micromammal teeth and jaws recovered from 19 ant mounds in Sioux County, Nebraska south of Toadstool Geologic Park. All of these ant mounds are situated on rocks of the Big Cottonwood Creek Member and their stratigraphic posit ions span from seven meters below to five meters above the upper purplish-white layer (UPW). The large sample size and stratigraphic position of this collection allows for the documentation of the micromammal faunae of this region, investigation of patterns of faunal change across the Chadronian-Orellan boundary, and identification of the local stratigraphic position of the Chadronian-Orellan boundary. Over 80 mammalian taxa are recognized in this collection, including ten new species and four new genera. Among the rodents nine new species, including three new genera, are described. The new genera include the cylindrodont Siouxlindrodon (S. sullivani n. sp.) and the aplodontids Cosepeiromys (C. attasorus n. sp.) and Protansomys (P. gulottai n. sp.). The additional new rodent species consist of the ischyromyid Ischyromys brevidens n. sp., the eomyids Paradjidaumo patriciae n. sp., Yoderimys massarae n. sp., and Litoyoderimys grossus n. sp., the florentiamyid Kirkomys miriamae n. sp., and the sciurid Cedromus modicus n. sp. Among the Lipotyphla, the soricid Domnina compressa Galbreath, 1953, is referred to a new genus Noritrimylus (as N. compressus), along with N. dakotensis (Repenning, 1967) and N. metaxy (Korth, 2020a), and a new species of oligoryctid, Oligoryctes tenutalonidus n. sp., is recognized. The large sample size available in this study also improved the understanding of the morphology and distribution of several previously described taxa. Tooth positions that were previously unknown for several taxa were identified and described for the first time (e.g., Adjidaumo intermedius; Eutypomys hibernodus, Ecclesimus tenuiceps, Micropternodus montrosensis). In some cases, the number of specimens here referred to specific taxa exceeds the total number of specimens previously known for that taxon (e.g., A. intermedius). A single chiropteran specimen (partial dentary with m3) and three isolated omomyid primate teeth are present in the collection but cannot be confidently referred to any known species. A partial M2 is referred to the borophagine canid Otarocyon macdonaldi, potentially representing the earliest occurrence of borophagines in North America. The majority of the non-lagomorph specimens identified in this study (~69 percent) are referred to taxa that elsewhere first appear in Orellan or younger faunae and Orellan taxa dominate the faunal lists of all 19 ant mounds regardless of their stratigraphic position. This prevalence of Orellan taxa results from downslope transport of Orellan specimens into the foraging range of the ant mounds, an important factor that must be considered when working with ant mound collections. To compensate for that effect the local last appearance datums of 14 Chadronian taxa (including three new species) were used to estimate the local stratigraphic position of the Chadronian-Orellan boundary instead of the local first appearance datums of Orellan taxa. It is estimated that the boundary is situated at the distinctive layer derived from volcanic ash (=UPW; ±2 meters), in close agreement with previous studies.
... Harvester ants (e.g. Pogonomyrmex sp.) take coarse material (including human artefacts) to their mounds from many metres away (Spangler and Rettenmeyer 1966;Schoville et al. 2009). ...
Ants are abundant in most of the world's terrestrial environments. They are energetic, strong for their size, numerous, and socially cooperative. They play many geomorphologically important roles. In particular, they construct mounds and subterranean galleries, create patterned ground, play a role in bioturbation, affect vegetation cover and soil properties (such as infiltration rate) and influence runoff and erosion. They also play roles in biogeochemical cycling and rock and mineral weathering. Here, we review and reanalyse data collected from over 80 studies on ant contributions to geomorphology from around the world. The clearest manifestation of the geomorphological role of ants is found in their various constructions, such as mounds. There can be hundreds or thousands of mounds per hectare, with a median density of 125 ha⁻¹ recorded in the studies reviewed. The longevity of these features varies and some are stable while others are highly erodible. The construction of mounds and galleries causes bioturbation (pedoturbation), a role which ants share with termites, worms and many mammals. A median rate of 1.5 t ha⁻¹ a⁻¹ is derived from the studies reviewed. Ants also produce patterned ground through their effects on vegetation. The relationships between ant activity and runoff and erosion are complex and not consistent. Bioturbation of soil, tunnelling activity, the construction of underground chambers, galleries and macro-pores, the removal and/or accumulation of organic material, and changes in vegetation cover, are all mechanisms by which ants might modify soil infiltration characteristics. Because of their effect on soil infiltration rates, sediment provision and on vegetation cover, ants can have a profound influence on runoff and soil movement on slopes. Only a modest amount of work has been done to investigate the role that ants play in rock weathering. Ants are greatly affected by human activities (especially land cover changes), and some geomorphologically-active species have proved to be highly invasive. The response of ants to future climate changes needs further investigation.
... Bioturbation has been reported in similar contexts at other sites, limiting the use of optical dating of the sediment: the presence of different age populations among the dated grains affects the accuracy of the ages and limits their precision (e.g., Tribolo et al., 2010), in particular in unconsolidated sands (e.g., Schoville et al., 2009;Chazan et al., 2013;Kristensen et al., 2015;Williams, 2019). Indeed, the lack of any sedimentary structures in the sands at Bestwood 1 indicates that bioturbation occurred. ...
The transition from the Earlier Stone Age (ESA) to the Middle Stone Age (MSA) in the interior of southern Africa is associated with the Fauresmith Industry. Major cultural developments found in the Fauresmith include regular use of ochre and other coloured minerals, prepared core technology including blade and point production, and the use of hafted spears. Chronological control for the Fauresmith is weak so that critical questions regarding the relationship of this industry to the evolution of modern humans remain unresolved. Here we present ages for the Bestwood 1 site, an open-air locality in the Northern Cape Province (South Africa) where an extensive Fauresmith occupation is found underlying sand deposits.Optically stimulated luminescence (OSL) was first applied to samples from the sands overlying the Bestwood 1 occupation horizon, and from the occupation horizon itself, in order to establish the chronology of the site. However, sediment mixing resulting from bioturbation processes has been observed, causing post-depositional bleaching of the majority of the grains, thus limiting the use of OSL. In addition, given the identification of the lithic assemblage to the Fauresmith, it seems likely that the sands were beyond the dating range of conventional OSL. Due to its hard-to-bleach properties, the thermally transferred-optically stimulated luminescence (TT-OSL) signal was deemed suitable for detecting the least-bleached grains.Single grain TT-OSL analyses combined with the finite mixture model (FMM) were conducted in order to isolate the oldest grains that could be contemporaneous with the time of deposition of the sediment associated with the ESA assemblage. High scattering of the equivalent doses is consistent with bioturbation processes that mixed sediment; the distribution of the equivalent dose values suggests that younger grains were incorporated into the ESA layers, thus supporting the use of the oldest component determined using the FMM to calculate the TT-OSL ages. This approach allowed us to establish the time for the Fauresmith occupation at 366 ± 32 ka, and the age of the overlying sand deposits, spanning from 350 ± 22 ka to 226 ± 13 ka.
... for the role of ants in bioturbation and transport of artefacts see Rink et al. 15 and Schoville et al. 16 ). However, the fact that some artefacts are found embedded in a clay matrix at the top of the gravels argues against such a scenario. ...
In order to investigate the buried landscape at the Fauresmith locality of Bestwood 1, outside the town of Kathu in the Northern Cape Province, we performed ground-penetrating radar and magnetometry surveys across the sand-filled central portion of the valley. The radar images a strong continuous reflector which we can assign to the boundary between the Kalahari sands and underlying Banded Ironstone Formation gravels. Moreover, the thickness of the sand delineates a buried depression in the centre of the valley with flat plateaus at the sides. Subtracting the sand thickness from the current topography produces a map of a small stream channel in the northern part of the valley. Analysis of the magnetic gradient data allows us to extend this buried channel further to the south. Our geophysical survey provides a valuable contribution towards understanding the context of hominin occupation along the banks of a small stream in the Kathu Complex. Significance: • We provide an example of combining two geophysical methods to map overburden thickness, useful for archaeological landscape interpretation. .
... Behavioral interpretations of prehistoric patterning are complicated by the effects of post-depositional processes. Natural processes influence the burial, modification, and patterning observed on all archeological materials at multiple scales (Flenniken and Haggarty 1979;Villa and Courtin 1983;Gifford-Gonzalez et al. 1985;Behrensmeyer et al. 1986;Olsen and Shipman 1988;Nielsen 1991;Shea and Klenck 1993;McBrearty et al. 1998; Barton et al. 2002;Schoville et al. 2009;Eren et al. 2010;Pargeter 2011). Although stone tools are the most common surviving artifact from most Pleistocene archeological contexts, they are subject to the same trampling, bioturbation, and displacement processes that impact the archeological visibility of other artifact classes (Lyman 1994;Dibble et al. 2006). ...
Full-text available
Controlled experiments in lithic technology tend to focus on controlling the human component of lithic tool manufacturing and use; however, animal disturbance can move and alter artifacts in non-random ways, thus altering the behavioral meaning assigned to artifacts and their contexts. The patterning visible in archeological debris on a horizontal plane can provide evidence for activity zones, pathways, and site formation processes. While the effects of trampling actors on the vertical displacement of artifacts have shown that artifacts can be dramatically displaced, the horizontal movement due to trampling is relatively less studied, particularly the effect over extended time periods. Here, an experimental investigation of experimentally produced lithic tools in three contexts with varying degrees of animal trampling intensity is described, and the resulting patterns of artifact displacement are presented. Animal trampling can produce directed, non-random patterning in how artifacts are moved from their original location. The role that bedding slope plays in transport direction given different degrees of activity is also explored. These results show that trampling can produce patterned artifact scatters similar to activity centers and should be taken into consideration for spatial analyses of archeological formation processes.
... Many isolated teeth as well as micro-vertebrate remains were collected through screen-washing of in-situ sediment, as well as from anthills developed on Wa-0 strata, using mesh down to 0.5 mm. While it is impossible to know the exact source of anthill fossils, recent observations suggest that harvester ants usually forage within 20 m of the nest (Schoville et al., 2009). It is generally assumed that anthill samples, if taken from relatively flat areas, are quite limited in stratigraphic range and do not mix samples from significantly different levels. ...
... They have a demonstrated capacity to move or rework large amounts of sediment in a short period of time and to cause selective movement of elements of a deposit, including the creation of stone lines (Williams 1978; Young 1976). Studies outside of Australia have established the potential for ants, termites and earthworms to act as taphonomic agents in archaeological sites (Shipman and Walker 1980; McBrearty 1990; Tryon 2006; Schoville et al. 2009). There is no reason to suppose that archaeological sites in Australia would not be subject to similar taphonomic agents. ...
Full-text available
During salvage excavations of an Aboriginal shell midden at Hollywell, on the Gold Coast of Queensland, ant activity was noted as a contributor to both bioturbation, and to the introduction of modern material, including metal fragments, plastic, nylon fishing line and cotton thread into the deposit. This material was found at depths of up to 400 mm and adjacent to excavation units with shells with a calibrated age of 1050–900 BP. These observations prompted the development of a small experiment to illustrate the impact that one species of common Australian ant observed on site, the green-head ant (Rhytidoponera metallica), can have on cultural material in sandy deposits.
Full-text available
1. FUNDAMENTACIÓN Y OBJETIVOS Los estudios referidos a los materiales líticos (v.g. artefactos, ecofactos y fuentes de materias primas) representan, en arqueología, una de las líneas de investigación con mayor tradición y desarrollo. Entre las razones que explican el importante rol que han cumplido y cumplen los estudios líticos en nuestro campo, se encuentran la ubicuidad de tales materiales, la diversidad de su uso en el pasado y su perdurabilidad, lo cual contrasta con las propiedades típicas de otros constituyentes del registro arqueológico. A lo largo de las décadas, los estudios líticos han crecido a través de la incorporación de nuevas perspectivas que han enriquecido y ampliado, de diferente modo, la aproximación básica —de carácter fundamentalmente descriptivo y comparativo— al conocimiento de los conjuntos artefactuales. De este modo, a la perspectiva descriptivo/clasificatoria tradicional (usualmente tipológica o técno-tipológica) se han sumado, entre otros, estudios funcionales, experimentales, geoquímicos, petrológicos, de residuos o adherencias, espaciales, tafonómicos y morfométricos. En muchos casos, las diversas líneas de evidencia se han integrado en marcos interpretativos particulares, tales como los de cadena operativa, secuencias de reducción, organización tecnológica y, más recientemente, de transmisión cultural. En este contexto, el objetivo central del presente curso es introducir a los doctorandos y profesionales interesados (procedentes de las distintas ramas de la Antropología) en un conjunto seleccionado de aproximaciones actuales —definidas en función de preguntas, perspectivas teóricas, aplicaciones metodológicas y técnicas— en el campo de los estudios líticos. Tales problemas (tafonomía lítica, análisis espaciales basados en el uso de SIG, perspectiva fenotípica en el estudio de los patrones de variación artefactual, análisis de morfometría geométrica y cladísticos) se caracterizan por su relevancia en términos de la interpretación y explicación arqueológica, así como por su relativa novedad y escasa, aun, popularización dentro de la comunidad académica. A este respecto, se considera importante que los alumnos adquieran conocimientos específicos en relación con nuevas tendencias teórico-metodológicas y, sobre todo, desarrollen criterios que les permitan integrar en forma coherente tales conocimientos a las prácticas que resultan habituales dentro de la especialidad (i.e. definición de problemas, especificación de líneas de evidencia relevantes, elección de vías teórico-metodológicas de abordaje, evaluación de resultados), así como explorar la posible
The interpretation of taphonomic and behavioral lithic edge wear formation is complicated by equifinality of edge damage morphologies. Rejecting hypotheses that edge damage originates from taphonomic processes is standard practice for many archaeological analyses and should be incorporated into lithic use-wear more explicitly. Quantitative hypothesis testing is advocated here, and facilitated by recording edge wear observations in an image referenced GIS spatial environment. A taphonomic predictive model was generated using trampling and flint-knapping experiments. Trampling experiments were conducted to determine how edge damage is distributed along tool edges due to non-use related, taphonomic processes. Experiments designed to test the assumption that undisturbed flakes do not preferentially orient either surface side-up (dorsal or ventral) were performed. Furthermore, it is argued that artifact orientation data, if available, can also be used to assess whether the frequency of edge damage is correlated with the degree of disturbance. This taphonomic predictive model is then statistically compared with frequency and distribution edge damage data from two South African Middle Stone Age sites. The research presented here illustrates the usefulness of edge damage distribution analysis for accounting for taphonomic processes as causal agents of edge damage formation, and strengthening behavioral interpretations regarding tool function. Bringing tool wear observations into a uniform spatial structure is one avenue for standardization of lithic use-wear analysis.
Full-text available
RESUME Plusieurs etudes recentes sur les gisements de hisons d'Ameriquc du Nord se sont anachees a alTiner les interpretations concernant l'histoire de la formation des sites paleoindiens. Durant cette derniere decennie, plusieurs gise mems paleoindiens o nt ete examines dans une perspective taphonomique et heaucoup de presupposes archeologiques ont connu des revireme nts spectaculaires. Un exemple en est fourni par les fouilles taphonomiques et archeologiques menees sur le site de Hudson-Meng (Nebraska, USA) depuis 1991. Ce site, q ui contient les restes de pres de 600 hisons concentres surplus de l 000 m 2 de superficie, a d 'abord ete fouille et interprete dans !cs annees 1970. Une quantile succincte d 'outils lithiques et de dehitage indique la prese nce de Paleoindiens sur le site, et une serie de 11 dates AMS ohtenues ii la fois sur des morceaux de charhon et sm des os de hisons donne un age d'environ 9563 ± 37 BP pom le gisement. Lors des fouilles precedenres 0971-1977), ce gisement avail ete interprete comme une zone de traiteme nt seco ndaire pour des q uartiers d 'animaux abattus sur un abrupt proche, hypothetique. Des travaux de terrain supplementaires et une analyse taphonomique mettem en doute ces interpretatiom. La reconstitution de la topographie au Pleistocene final demontre qu 'aucune denivellation adjacente n'aur.iit pu correspondre ~ un abrupt. L'analyse de la desarticulation. de la dispersion et de la fracturation des squelettes ainsi que des do mmages et de l"usure des os montre que Jes animaux etaient morts sm le gisemem et non que celle des segmenL~ de carcasses partiellemem depeces y avaient ete ras-sembles par !es hommes. ABSTRACT Several recent studies C>f North American bison boneheds have focused on refining inletpretations ahout the fonnational bistories of Paleoindian sites. Within !he last ten years, a number C>f Paleoindian bonebeds have been examined from a taphonomic perspective and many ideas about the archaeolof!,ical record have chanf!,ed dramatically. An example of this d![fwence is provided by the taphonomiclarchaeolo-/!,ical excavations that have been conducted at the Hudson-Meng site (Nebraska, USA) since 1991. Hudson-Meng, which wntains the remains of perhaps as many as 600 bison within a concentrated honebed
Full-text available
Bone-collecting behavior in the harvester ant, Messor barbarus , has been observed for the first time. Excavation to the depth of the packed substrate of a single ant hill yielded 1167 modern bones derived from nine or more species of small vertebrates. Since harvester ants of various species are well-known from the Americas, Africa, and Eurasia, ants may have served as important agents of concentration for micromammal assemblages in the fossil record. Analysis of the faunal and skeletal representation of the ant hill assemblages reveals characteristic patterns that might be used to distinguish such assemblages from those collected by avian or mammalian predators.
Full-text available
To determine whether the fossils found in mounds of the Harvester ants (Pogonomyrmex occidentalis) originated from subsurface deposits and brought to the surface or collected from the surface during the construction of their tunnels, 300 ants were placed in a controlled environment. Numerous tunnels were excavated in a layered artificial mound. Beads and fossils were placed on the surface and in the layers. The conclusion was that fossils and beads were incorporated into the mounds from the surface and subsurface layers resulting in a significant guide to a hidden fossil resource. © 2004 E. Schweizerbart'sche Verlagsbuchhandlung, D-70176 Stuttgart.
This chapter presents a survey of disturbance processes in archaeological site formation. A study of archaeological context must go hand in hand with an understanding of the matrix in which remains are embedded. Although soil scientists have long been aware of the dynamic nature of soil, only recently have prehistorians begun to apply this concept systematically to field situations, at least in the New World. Two contrasting sets of processes operate in soil development: horizonation, where soil materials are differentiated into profiles having horizons, and homogenization, or haploidization, where horizon formation is impeded, or where horizons and their contents may be mixed or otherwise disturbed. Faunalturbation refers to the mixing of the soil by animals. Surely every archaeologist has seen the effects of burrowing animals time and again.
Ecological communities respond to environ- mental changes as the individuals of the component species die and are replaced. Thus, pat- terns of population turnover form an important aspect of community processes. Much less is known about species of long-lived individuals than of short-lived ones (Likens, 1989). Instantaneous observations of age structure can be used to infer long-term dynamics but not all species can be aged retrospectively. Inferring life- history dynamics from current populations requires assumptions that are hard to verify. Following marked colonies of long-lived species is slow but provides direct, non-inferential data on population dynamics, although those are specific to the period observed. This note reports 15 years of observations aimed at determining survivorship of individual harvester ant colonies. Harvester ants are important arid grassland herbivores whose dynamics are crucial parameters for patterns of change in the rest of the community (Brown et al., 1979; Coffin and Lauenroth, 1990). Fifty-six mounds of Pogonomyrmex occidentalis Cresson (Hymenoptera: Formicidae), the western harvester ant, were permanently marked with aluminum tags in August 1977. The ant colonies were checked each August from 1977 to 1991; colony deaths were noted and new colonies were marked. Death was determined based on the following information: absence of foragers at times in which neighboring colonies were foraging and poor condition of the mound. It was verified by deterioration of the colony site in subsequent years. Unoccupied sites remain marked for a study of succession. The site, about one hectare in extent, just south of the University of Nebraska's Cedar Point Biological Station, Keith County, Nebraska, was within a pasture which received moderate, half-summer grazing during the period studied. The vegetation is typical of shortgrass prairie, dominated by Buchloe dactyloides, Bouteloua hirsuta and B. gracilis, interspersed with Stipa comata, Aristida purpurea and A. oligantha, and forbs such as Artemesia and Psoralea and woody perennials, including Yucca glauca and Juniperus virginiana (Kaul et al., 1983). The site lies between eroded canyons, which give the study area an irregular shape. Rock outcrops produce some areas with insufficient soil depth to support a harvester colony.
Granivore-seed interactions involve a feedback between granivore seed selectivity and seed availability. We examined this feedback to determine how seed preferences by the western harvester ant, Pogonomyrmex occidentalis, related to seed availability and, in turn, affected the soil seed pool. Preferences were estimated from natural diets as well as from experiments that controlled seed size, relative availability, and distance from ant nests. Seed availability to ants varied with season and over 2 yr. Colony activity and seed intake rates were correlated with seed availability. Seed preference by ants was correlated with the seasonal availability of preferred species, but not with unpreferred seeds. From the soil seed pool, ants preferentially harvested small, sound seeds. They removed 9-26% of the potentially viable seed pool each year, and as much as 100% of available preferred species. Seed densities were lower 2-7 m from nests, where foraging activity was concentrated, than 7-12 m from nests. In controlled preference experiments, P. occidentalis was unselective near nests, but preferred large seeds with higher assimilable energy content in trials 10 m from nests. A relatively low foraging activity > 7 m, however, suggests that this distance-dependent preference is rarely manifested in natural conditions and does not measurably affect soil seed dynamics. Our results point to the importance of studying diet choice in a natural context; preferences measured under experimental conditions may not correspond to natural diets. Such discrepancies in food preference measurements will affect predictions about how consumers influence the population dynamics of resource organisms.