Download Effects of Garden Attributes on Ant (Formicidae) Species Richness

Published July 28, 2016
urban agriculture & regional food systems
dl.sciencesocieties.org/publications/ua
Effects of Garden Attributes on Ant (Formicidae)
Species Richness and Potential for Pest Control
Naim Edwards*
abstract
Ecology and Evolutionary Biology, Univ. of Michigan,
Ann Arbor, MI 48109-1048 USA. Received 13 Nov. 2014.
*Corresponding author ([email protected]).
Abbreviations: CFA, County Farms Park site A; CFB,
County Farms Park site B; CSS, Catholic Social Services;
GV, Greenview; ST, self-tilled; TT, tractor tilled.
Urban vegetable gardens serve as sources of social, economic, and biodiversity
conservation opportunities. Determining how gardens can be developed to
enhance biodiversity and ecosystem services is important for organic food production. This study investigated the diversity of ant (Formicidae) species in relation to various garden attributes in three distinct gardening sites in Ann Arbor,
MI. Further, predation experiments were conducted to explore the potential biocontrol services of commonly encountered ant species on Cabbage Looper moths
(Trichoplusia ni) and Squash Bugs (Anasa tristis). Results indicated that community gardens have the capacity to support a multitude of ant species; 19 species
from 12 genera were sampled. However, soil texture, intense tilling, and distance
from less disturbed areas significantly reduces species richness. Less intensively
tilled garden plots averaged nearly double the number of ant species collected
in tractor-tilled plots. Plots bordered by grass lawns and closer to wooded areas
had significantly more ant species than plots surrounded by other garden plots.
Predation experiments indicated that ants consume Cabbage Looper moth eggs
and larvae, although predation on Squash Bugs was less common. Some of the
species that demonstrated levels of predation are rare in gardens due to sensitivity to disturbance. Thus, reducing disturbance within and around gardens can
promote ant biodiversity and beneficial predators leading to increased predation
on garden pests.
A
s cities and agricultural areas continue to expand, converting
natural areas into managed landscapes, public interest in restoration gardening and urban agriculture increases (McMahan, 2006;
Hodgson et al., 2011). Urban gardens can strengthen local food systems and environmental conservation as well as stimulate economic,
social activity, and food security (Yadav et al., 2012). In addition to
producing locally grown food, vegetable gardens reconnect people
with nature and create habitat for a wealth of organisms (Goddard et
al., 2010). Moreover, some of these organisms contribute valuable ecosystem services such as pollination, decomposition, pest control, and
Published in Urban Agric. Reg. Food Syst. Vol. 1 (2016).
doi: 10.2134/urbanag2015.01.1405
© American Society of Agronomy and Crop Science Society of America.
This is an open access article distributed under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
urban agriculture & regional food systems
1
nutrient cycling, which contribute to the viability of the gardens. Thus, understanding how gardens can be designed and
managed to support biodiversity is critical for utilizing urban
systems for conservation and ecosystem services.
Species diversity in any landscape is influenced by endogenous and exogenous factors acting at different scales (Yitbarek
et al., 2011). Identifying these factors will enable urban gardeners to enhance the diversity of desirable organisms via
different management techniques that improve soil quality
and reduce costs associated with inputs like pesticides and fertilizers (Peck et al., 1998). For example, surrounding habitat,
soil texture, tillage, and chemical inputs impact the types of
organisms that can persist within gardens. The potential for
biodiversity depends on the availability of resources and the
level of disturbance within gardens, as well as the surrounding
habitats’ capacity to serve as a reservoir or additional resource.
The intensity of tillage in particular has an immediate
impact on garden biodiversity. Heavy tilling has been shown
to decrease soil species richness and community complexity
of microorganisms and arthropods (DuPont et al., 2009).
The decline is attributed to reduced soil moisture, removal
of biomass, and compaction, and tilling is associated with
exaggerated rates of erosion and labor costs (Peck et al., 1998;
Altieri, 1999; Pimentel et al., 2012). Thus, examining the
extent and severity of different tilling methods on biodiversity
is an integral component of garden management.
Pest control of insects that damage crops is a primary concern of vegetable growers (Ashebir et al., 2007). In many parts
of the world, applying synthetic insecticides is the standard
for addressing this issue, though studies have shown chemical
controls present an array of public and environmental health
consequences (Isman, 2006). An alternative method is biological control via predators and parasitoids of garden pests. These
natural controls can persist in gardens that are managed organically and designed to sustain them. In this study, we investigated
correlations between ant species diversity in gardens and their
potential ecosystem service as pest control agents.
Ants are ideal organisms for assessing both biodiversity
and habitat quality because they comprise numerous species
that participate in a variety of ecosystem functions (Uno et
al., 2010; Pacheco et al., 2013). Along with other soil-dwelling
fauna, ground-nesting ants play a crucial role in soil structure
by increasing moisture, water infiltration, aeration, and organic
matter (Lal, 1988). Studies investigating nest preferences, species richness, and predation have indicated that ants commonly
occur in gardens and can be valuable predators of other arthropod pests at multiple ontogenetic stages (Way and Khoo, 1992;
López and Potter, 2000; Yadav et al., 2012). Nevertheless, the
bio-control services ants provide have been scarcely studied in
community gardens, and ants are rarely recognized as natural
predators in gardening literature in temperate habitats.
Here, we examined how different endogenous and exogenous variables affect ant diversity in community gardens. We
assessed the effects of soil texture, tillage, proximity to natural
and disturbed habitats, and the number of garden sides adjacent to open lawns on species richness within garden plots.
Additionally, we explored how commonly encountered ant
species contribute biological control via predation on different
life stages of Squash Bug (Anasa tristis) and Cabbage Looper
moth (Trichoplusia ni) eggs and nymphs, and eggs and larvae,
respectively. Both pests are economically important due to
crop damage and their wide distribution throughout North
America (Beard, 1940; Shorey et al., 1962). We expected
greater levels of disturbance within and around garden plots to
negatively affect ant species diversity, and we anticipated that
multiple species would inflict high levels of mortality on insect
eggs and juveniles.
Methods
Site Selection
A “community garden” is defined as an aggregation of organic
vegetable garden plots where individual gardeners manage
each plot independently. Three community garden sites in
Ann Arbor, MI (42°16¢53² N, 83°44¢54² W) were selected
based on the number of plots within each site that were tilled
either by a tractor (tractor tilled, TT) or less intensively tilled
with garden tools (self-tilled, ST). The site names are: County
Farms Park (CFA and CFB), Greenview (GV), and Catholic
Social Services (CSS) and are shown in Fig. 1. County Farms
Park is divided into two sections because the gardening space
is divided by a concrete trail and mixed vegetation approximately 20 m wide. The two sections also differ in soil texture.
Randomization of plot selection was limited by access and
permission from all of the gardeners to sample in their plot.
Still, the plots sampled accounted for a minimum of 25% of
any given site and represented 31% of all available plots.
Project Grow, a non-profit organization that promotes
organic community gardens, administered the sites. The organization leases plots to local gardeners and has mandatory,
organic guidelines that all participants must follow, which
controls for non-organic practices. Plots are rectangular and
vary by area: half (~30 m2), full (~60 m2), and double (~160
m2). Each garden site had at least three full ST plots and nine
TT plots. TT plots were strip tilled by a tractor once in the
spring before planting and once in the fall after harvesting.
Gardeners tilled ST plots with hand-held farming tools at their
discretion or not at all. All sites were over 10 yr old, except for
one section of County Farms (CFA), estimated to be approximately 6 yr old (personal communication with director). The
sites varied by soil texture. The number of plots with different
characteristics is summarized in Table 1.
We recorded garden characteristics including type of till,
distance from nearest wooded and concrete area, ground cover
type, and the number of sides of the plot bordering a grass lawn.
A wooded area is considered to be five mature trees within 20
m2. Number of sides bordering a grass lawn is relevant because
many ant species forage over 1 m from the nest, and thus are
2
urban agriculture & regional food systems
Fig. 1. Google images of the three gardening sites. 1: County Farms Park sections (CFA and CFB), 2: Greenview (GV), 3: Catholic
Social Services (CSS).
Table 1. Number of plots per site for different categories. Each column designates a plot characteristic and the number of
plots that correspond within a given site.
ST
TT
0 Sides
1 Side
2 Sides
Half
(~30 m2)
Full
(~60 m2)
Double
(~150 m2)
Total
plots
CFA
4
5
4
5
0
4
4
1
9
CFB
7
2
4
3
2
1
4
4
9
CSS
1
5
0
5
1
4
2
0
6
GV
3
8
2
4
5
4
7
1
11
15
20
10
17
8
13
16
6
35
Site†
Total
† CFA, County Farms Park site A; CFB, County Farms Park site B; CSS, Catholic Social Services; GV, Greenview; ST, self-tilled, TT, tractor tilled.
likely to disperse into gardens from surrounding habitat. Concrete areas consisted of playgrounds, paved walkways, or areas
where the soil was compacted and covered with gravel.
We also noted vegetation within plots, which varied considerably between plots and also changed within plots during
the study. Variability ranged from plots growing one or two
of the same crops all season to plots where gardeners regularly weeded, seeded, and transplanted new plants ranging to
up to nine different crops and ornamental plants. The average number of crops per plot was four and the most common
urban agriculture & regional food systems
crops were tomatoes (g. Solanum), beans (F. Fabaceae), garlic
and onions (g. Allium), and lettuce (g. Lactuca). Many gardens
also had basil (Ocimum L.), sunflower (Helianthus L.), cucurbits, and ornamental plants like marigolds (Calendula L.) and
peonies (Paeonia L.). Due to the variability in both plants and
their proportions and locations within plots no analysis examined their correlation with ant species.
3
Sampling for Ants
Ant diversity was based on the number of species collected
nesting or foraging within each garden plot throughout
the study period. Direct sampling and tuna baiting for ants
occurred once every 2 wk per plot between 7:30 AM and
12:00 PM from May to August. Pitfall traps were installed
and left in plots for 48 h once every 2 wk as well in the months
of June and July. These three sampling techniques were applied
to optimize species inventory (Gotelli and Colwell, 2011).
We used standard tuna baiting methods, placing pieces of
tuna directly on the soil surface along transects at 2-m intervals within garden plots (Brown, 2000). Half plots had three
transects, while full and double plots had four. All baits were
checked 30 min and 1 h after placement. Ant specimens were
generally collected after 1 h for identification, unless rare species were observed at the 30-min check.
Direct sampling was done in two separate square meter
quadrats every 2 wk during the study. The procedure consisted
of scanning plants, soil, and under stones for ants. Specimens
were collected with an aspirator and later identified to species.
Pitfall traps consisted of a Hefty plastic cup (9 cm in diameter
and 12 cm in height) (Reynolds Consumer Products, Lake
Forest, IL) placed into the ground so that the lip of the cup
was flush with the soil surface. Traps were filled with 150 mL
propylene glycol solution with a few drops of Seventh Generation dish detergent (Seventh Generation, Burlington, VT)
to reduce surface tension (Brown, 2000). A plastic roof was
secured over the cup to reduce entry from rain and the likelihood of flying and larger fauna disturbing the trap. Traps were
placed at least 1 m from the plot edge and in different locations
within the plot for each sampling period; full and double plots
had two pitfall traps placed within them each sample. Pitfall
traps created the opportunity to sample for ant species active
outside the direct sampling and tuna-baiting time frames.
Specimens were keyed out and verified by an ant taxonomist
(B. DeMarco, Michigan State University, August 2013); other
insects in the pitfall traps were identified to family.
Predation Experiments
Predation experiments consisted of 10-min trials to quantify
how many eggs, larvae, or nymphs of Cabbage Looper moths
or Squash Bugs ants could remove or damage. Seven ant species were tested, and all trials were conducted on a cleared
soil surface within 20 cm of ant nests or in areas of high ant
traffic—a space where one or multiple species were frequently
observed foraging. A. tristis eggs and nymphs were collected
from garden plants of the genus cucurbita (spp. maxima, moschata, and pepo). The eggs were laid in clusters, which were
cut out of the leaves (approximately 15 eggs/4 cm2). Eggs and
larvae of T. ni were ordered from BioServe, an online supplier
of reared insects.
Cabbage Looper eggs were oviposited on a paper towel
and 3 cm2 pieces of paper towel were cut out (approx. 100
eggs per square). Squash Bug egg clusters, still attached to the
leaves, were also cut out and placed on the soil surface. The
nymphs of A. tristis and larvae of T. ni were placed on squares
of orange marker tape (2.54 cm2) for visibility. Groups of three
nymphs or larvae were placed on each orange square per trial.
The number of eggs or larvae removed, attempts (unsuccessful
removals), and passes (walking by prey item with no change of
behavior) by ants was recorded over 10 min.
Results
Analysis
All statistical analyses were conducted using SPSS and EstimateS. Each garden plot was considered the basic sampling unit
for ant species diversity, and the data from May to August for
each plot was combined for analysis. Independent t tests compared the mean number of ant species encountered between
tillage groups (TT and ST). One-way ANOVAs compared
the mean number of ant species per plot between sites (which
varied by soil type) and between plots varying in the number
of sides bordered by a grass lawn. Multiple regression analyses
assessed the relationship between ant species per plot and plot
characteristics including distance from a concrete or wooded
area and the number of sides adjacent to a lawn.
Post hoc analysis indicated that GV, which was classified
by heavy clay soil texture, had significantly fewer ant species than the other sites (p < 0.05; Table 2). Only five species
total were encountered sporadically and at low frequencies
throughout the study at GV. Most ant species are sensitive to
soil texture, and heavily textured or wet soils do not support
ground-nesting ants (Ellison et al., 2012). Thus, due to the
likelihood that the heavy clay texture was a limiting factor to
ant species abundance, GV data was excluded from statistical
analyses for the effects of other characteristics.
Species richness estimates for the garden sites were calculated using Chao 2, ICE, and first and second order Jackknife
Table 2. Species richness between sites. The sites did not differ significantly between the average number of species per
plot except for Greenview (GV) which was characterized by heavy clay soil texture.
Site†
(soil texture)
CFA (clay loam)
CFB (sandy loam)
CSS (clay loam)
GV (heavy clay)
Species accumulation
Avg. no. of species/plot
SD‡
Sig.
P < 0.05
8
9
9
5
4.00
4.89
4.50
1.18
2.160
2.369
3.209
0.751
A
A
A
B
† CFA, County Farms Park site A; CFB, County Farms Park site B; CSS, Catholic Social Services.
‡ SD, standard deviation.
4
urban agriculture & regional food systems
formulas with incidence data from all of the plots sampled in
the study (Gotelli and Colwell, 2011). Here we report only on
the total species accumulation estimates for all sites combined
because estimates based on less than 20 samples generally are
not dependable for accurate estimates, and we had less than
twenty samples for plots with regard to shared characteristics
(Gotelli and Colwell, 2011).
Due to the limited number of pest control trials, we report
only the occurrence of attempts and removals during the 10
min trials with Cabbage Looper and Squash Bugs at different
pre-adult ontogenic stages by a given ant species. The number
of passes and the average number of removals are not included.
Predation rates are reported as “low” or “high:” 10–25% or <
25% prey damaged or removed, respectively.
more extreme temperatures or highly disturbed habitats where
humans are present, such as gardens (Andersen, 1997).
Twelve species were collected at tuna baits, 13 in pitfall
traps, and 14 from direct sampling; each sampling method
yielded two unique species not sampled via other methods.
Regression analysis for distance to concrete (R2 =0 .004, p =
0.383) and plot area (R2 = 0.141, p =0 .071) did not correlate
significantly with species accumulation per plot. Additionally,
ant species abundance in pitfall traps was significantly correlated with abundance of other insect families (Fig. 2).
Ant species diversity was significantly higher in ST plots
than in TT plots, averaging nearly double the number of species (Fig. 3). In total, 17 ant species were collected from ST
plots and 14 from TT plots. Of the 19 ant species collected,
five rarely sampled species (found in two or fewer plots) were
encountered solely in ST plots and an additional six species
were encountered ³60% more frequently in ST plots than in
TT plots (Table 3). The total number of ant species per plot
ranged from 0 to 9. Excluding site GV, which had low ant
activity throughout the entire study, the mean number of ants
per plot for all sampling methods was 5.8 ± 2.4 (n = 12) in ST
plots and 3.2 ± 1.8 (n = 12) in TT plots (Fig. 3).
Exogenous factors such as the number of plot sides
adjacent to grass lawns and the distance from wooded areas
significantly affected the species diversity per plot (Fig. 4 and
5). The average number of species per plot increased with
Species Richness
We collected 19 ant species from 12 genera and 4 subfamilies
representing 7 functional groups in 33 of the 35 plots (Table
3). Functional groups reveal community structure based on the
environment and behavioral dominance (Andersen, 1997). Of
the seven functional groups, three were represented by one species and one group was represented by the genus Camponotus
(Table 3). All other species were categorized as opportunists and
hot or cold climate specialists. Climate specialists and opportunists are generally weak competitors, but have an advantage in
Table 3. Ant species and proportion of incidence in self-tilled (ST) plots and plots varying in sides adjacent to grass lawns. Species in bold text were found nesting in plots. Blank spaces represent no occurrence. Functional groups (Andersen, 1997): Cold
climate specialist (cc), hot climate specialist (hc), Opportunists (opp), Cryptic species (cry), Generalized Myrmicinae (gm), Subordinate Camponitini (sc), Specialized predator (sp). Less frequently encountered species were generally larger ants and cryptic
species, and they were also found more often or exclusively in ST plots and plots that bordered grass lawns.
Species (functional group)
Total plots (n = 35)
ST
0
Sides adj. lawn
1
2
Lasius neoniger (cc)
28
0.46
0.80
0.82
0.75
Tetramorium caespitum (opp)
24
0.50
0.80
0.76
0.38
Tapinoma sessile (opp)
15
0.73
0.30
0.41
0.63
Prenolepis impairs (cc)
10
0.60
0.10
0.41
0.25
Solenopsis molesta (hc)
10
0.60
0.20
0.41
0.13
Aphaenogaster rudis (opp)
7
0.86
0.10
0.24
0.13
Crematogaster cerasi (gm)
5
0.80
0.18
0.25
Formica nitidiventris (cc)
4
0.50
0.18
0.13
Formica obscuripes (cc)
4
0.75
Camponotus pennsylvanicus (sc)
2
1.00
0.10
0.12
0.13
0.25
Formica argentea (opp)
2
0.50
0.06
0.13
Formica subsericea (opp)
2
0.50
0.06
0.13
Lasius flavus (cc)
2
0.50
0.12
Myrmica lobifrons (opp)
2
0.00
Ponera pennsylvanica (cry)
2
1.00
Camponotus noveboracensis (sc)
1
1.00
0.13
Lasius umbratus (cc)
1
1.00
0.13
Myrmica americana (opp)
1
1.00
0.13
Polyergus lucidis (sp)
1
0.00
urban agriculture & regional food systems
0.25
0.06
0.13
0.10
5
Fig. 2. Ant species richness correlates with other insect family per
plot. There is a positive correlation between ant species richness
per plot and the number of other insect families encountered in
pit fall traps for the same plot r(19) = 0.54, p < 0.01.
Fig. 3. Average species per plot for TT (tractor-tilled) and ST
(self-tilled) plots. Less intensively tilled plots averaged nearly
double the number of species per plot than plots tilled by a
tractor (t = 2.72, df = 22, p < 0.01).
increasing number of adjacent sides to grass lawns (Fig. 4), and
we observed an overall decrease in species accumulation per
plot with increasing distance from wooded areas (Fig. 5). Separate regression analyses for TT and ST plots regarding average
species per plot with distance to a wooded area were not significant (p = 0.27 and p = 0.22 respectively). This may be due
to a small sampling size for both groups (n = 12), imprecision
in the multi-regression analysis, or it may indicate that within
groups the effect of tilling was stronger than the effect of proximity to a forest.
Lasius neoniger was the most abundant and widely distributed ant (Table 3). However, in plots where present,
Tetramorium caespitum was the most widely distributed and
dominant ant, occupying and taking over the largest proportion
of tuna baits. These two species were sampled the most from all
Fig. 4. Ant species per plot with increasing number of adjacent
grass sides. The average number of ant species per plot
correlated significantly with the number of sides of a plot that
bordered a grass lawn. F(2, 21) = 9.72, p = 0.001.
Fig. 5. Ant species number per plot decreased with distance
from wooded area. The distance from a forest fragment
significantly predicted the average ant species per plot, b =
–0.461, t(23) = –3.35, R2 = 0.27, p <0.01. Note no self-tilled (ST)
plots beyond 25 m were sampled.
three sampling methods. Species encountered in less than five
plots were sampled most at tuna baits located near plot edges,
although pitfall trap collections showed that many species do
forage throughout garden plots (personal observation).
Based on incidence data, EstimateS calculated species
richness for the gardens to range from 19.83 to 22.89. The
ICE estimate of 22.75 may be the most appropriate because its
function considers the incidence of frequent and infrequent
species observed across samples, whereas the other estimator
formulas assume homogeneity between samples and focus
on the numbers of singletons and doubletons which varied
between sites.
Predation
Formica obscuripes and Tetramorium caespitum were the only
two species tested to predate at least one life stage of both
6
urban agriculture & regional food systems
Table 4. Predation rates. The predator potential of seven species was studied and the occurrence of an event is marked with
an “x.” The proportion of prey damaged or removed is given for the 10-min predation trials. A blank box signifies no action
was observed. “n/a” signifies a species was either not active during the trial or not tested: P. impairs was not active when T.
ni trials were conducted and C. cerasi was not tested for T. ni predation.
Removal
T. ni
larvae
Damage
T. ni
eggs
Attempt
T. ni
larvae
C. cerasi
n/a
n/a
C. pennsylvanicus
Low
Low
Species
Attempt
A. tristis
eggs
x
F. obscuripes
F. subsericea
Attempt
A. tristis
nymphs
x
L. neoniger
P. impairs
x
T. caespitum
x
Removal
A. tristis
eggs
Removal
A. tristis
nymph
Low
n/a
Low
High
x
Low
High
x
Low
x
n/a
n/a
Squash Bugs and Cabbage Looper (Table 4). F. obscuripes successfully removed A. tristis eggs, T. ni larva, and damaged T. ni
eggs. T. caespitum successfully removed A. tristis nymphs and
also removed and damaged T. ni larva and eggs, respectively.
Crematogaster cerasi demonstrated no predatory activity in
three trials, showing no change in behavior in the presence of
pests. Formica obscuripes, Formica subcericea, L. neoniger, and
T. caespitum recruited and swarmed on T. ni eggs, with 3 to
10+ individuals. Ant species that attacked T. ni eggs ruptured
them with their mandibles and apparently ingested the fluid.
Smaller ants (T. caespitum and L. neoniger) removed T. ni
larvae with two or more workers, while Formica and Camponotus species could carry whole larvae individually. T. caespitum
and F. subsericea carried nymphs of A. tristis individually.
Discussion
Species Richness in Garden Plots
We found 19 species of ants in community gardens in Ann
Arbor, and species estimates were as high as 22.89 overall.
Low
High
n/a
High
Given that there are 113 described species of ants in the state
of Michigan (Wheeler et al., 1994) results suggest that three
garden sites in Ann Arbor serve as habitat for 20% or more of
the ant fauna found in the state. Ant species number per plot
also correlated with the diversity of other arthropods caught
in pitfall traps (33 families), which suggests that gardens
with higher ant diversity also favor higher diversity of other
arthropods (Fig. 2). This does not mean ants directly influence
biodiversity, but we argue that garden attributes that favor
higher ant species richness also favor diversity in other groups
of organisms (Altieri, 1999).
Of the nine most abundant ant species, the two most
common were encountered equally or more often in TT plots,
while the other seven species were found 50% or more often in
ST plots than in TT plots (Fig. 6). Both Lasius neoniger and
Tetramorium caespitum were found in all of the plots sampled
in each site except for GV, where L. neoniger was encountered
in four plots and T. caespitum was not observed at all. The
abundance of these two species corresponds with other studies
Fig. 6. Percent encounters of species within self-tilled (ST) and tractor-tilled (TT) plots. Aside from the two most commonly encountered
species, all of the other species appeared more often in ST plots except for F. nitidiventris, which was found equally in both.
urban agriculture & regional food systems
7
that surveyed ant species richness in urban areas (Pećarević et
al., 2010; Uno et al., 2010).
Five species were found exclusively in ST plots. These five
species were rare, being encountered in only one or two plots
throughout the study. Two species were found exclusively in
TT plots—Polyergus lucidis in which one individual specimen
was encountered in the entire study, and Myrmica lobifrons
was collected in two TT plots at GV. These observations support that less disturbed plots have a greater capacity for ant
species diversity.
It is important to note that site GV, which was characterized by heavy clay soils, averaged significantly fewer ants per
plot than the other garden sites (Table 2). In fact, no ants were
encountered in two of the GV plots throughout the entire
study. GV soils were muddy in May and June and hard and dry
in July and August. One indicator of the conditions at the site
was the collection of Myrmica lobifrons, which is a bog specialist and also nests in swamps (Wheeler et al., 1994). Although
we did not test for the effects of soil texture or other factors
that may have caused this reduction in ant species diversity,
Ellison et al. (2012) report that heavily textured, clay or wet
soils do not support ground-nesting ants.
Along with soil texture, the way the soil is tilled also
influences ant diversity. TT may eliminate or damage nests,
immediately decreasing ant presence and activity in gardens.
If nests are not eliminated, ants may take weeks to reestablish
tunnels and foraging trails. Further, tilling may bury resources,
which would further reduce the likelihood of survival. Tilled
soils are also less habitable for ants due to reduced moisture,
as ants are highly sensitive to desiccation (Hölldobler and
Wilson, 1994).
Figure 2 also shows that less intensively tilled (ST) gardens had higher diversities of other insect families despite
having fewer ant species, which suggests that the consequences
of disturbance from tractor-tilling reduces the diversity of
other organisms within plots as well. The arthropods in pitfall
traps consisted of both beneficial groups like spiders, myriapods, pollinators, and predatory beetles as well as pests like
grasshoppers and stinkbugs.
Our results reveal that more intense tilling has a negative
effect on ant species diversity. Less intensively tilled (ST) plots
averaged nearly double the number of species encountered in
TT plots (Fig. 3). No ants were encountered in TT plots after
tractor tilling in early May, except for one plot which may have
reintroduced a species via a nest in compost that was added
shortly after tilling. Gardens that were tractor-tilled showed
no signs of ant activity until June, nearly 4 wk after tilling.
Nevertheless, the strongly negative effect of tilling on ant
presence can be balanced by optimizing undisturbed space
outside of plots as well as proximity to wooded areas, which
both correlated with greater ant species number in plots (Fig.
4 and 5). Once again, excluding the two most common species,
the other species have a higher frequency in plots that border
a grass lawn than plots that do not (Fig. 7). For example, we
sampled an intensively tilled plot near a wooded area and bordered by grass lawns with up to 8 ant species– nearly triple the
average for TT plots. These results reveal that the surrounding
habitat can compensate for the loss of ant species within plots
due to dispersal or migration into gardens from the exterior.
Ants as Bio-Control
Predation experiments demonstrated that multiple species of
ants do consume garden pests. Cabbage Looper moth eggs
were damaged by every species tested that encountered them
and the larvae were removed by all but one species tested
(Table 4). Squash Bugs experienced less predation during the
10-min trials with only one species successfully removing eggs
oviposited on leaves, and two species removing nymphs (Table
4). The experiments were conducted on the soil surface, which
does not represent predation on vegetation. However, all species tested are known to forage on vegetation. In fact, Formica
oscuripes and L. neoniger workers were observed capturing
prey while foraging on plants during the study. It is worth
mentioning that we anticipated much more activity from Prenolepis impairs, but this species was absent from most of our
predation trials probably due to the time of day, year, and high
temperatures during trials (Talbot, 1943).
Another factor that affects different ant species’ pest
control potential is the size and morphology of the prey. Cabbage Looper eggs average 0.4 mm3 (Jackson et al., 1969) and
are much fleshier than Squash Bug eggs. The size and texture
increases the ability of small and large ants to damage eggs.
Large ants removed larvae with little difficulty, while smaller
ants were observed to carry the larvae with as few as two workers. Squash Bug eggs are on average 1.5 × 1.1 mm (Beard,
1940), which is up to half of the size of the medium and small
ants collected (Ellison et al., 2012). In addition to not being
able to carry the eggs in their mandibles, the adhesion of the
eggs to host plant leaves was often too great a force for small
ants to overcome. Still, with recruitment and more time to
remove prey, smaller species can be valuable predators.
Collignon and Detrain (2010) found that T. caespitum
foragers recruit more workers at higher rates to food items
that require lower transport costs and higher energy pay off.
This may indicate that smaller ants are less likely to predate or
recruit to large or difficult to remove prey items like Squash
Bug eggs or Cabbage Looper larvae. Squash bug nymphs are
comparable in size to small and medium ants, but nymphs may
be more vulnerable to predation than eggs because ant mandibles can easily cling to their legs.
From egg to adult, Cabbage Looper and Squash Bugs
require at least 3 wk of development (Shorey et al., 1962;
Jackson et al., 1969). Since multiple ant species demonstrated
measurable levels of predation during 10-min trials, prolonged
exposure of prey may have led to greater predation levels. Since
different ant species are active at different times of day, areas
with high species diversity may confer a greater likelihood for
pest suppression of herbivorous insects.
8
urban agriculture & regional food systems
Larger ants such as Formica subcericea, Formica obscuripes,
and Camponotus pennsylvanicus showed levels of predation
and easily removed prey when available. However, an important trend to note is that these larger ants were not encountered
in plots that had no adjacent grass sides, with one exception
(Table 3). Thus, large ants occur less often in plots further
from lawns and wooded areas. One reason that explains this
low frequency is that species of the genera Camponotus and
Formica generally nest in wood or shallow nests respectively
(Mikheyev and Tshinkel, 2004; Coovert, 2005; Ellison et al.,
2012). Such nesting preferences make larger ants more sensitive to soil disturbances. Still, gardens can be designed in a
way to allow these species to persist through management that
does not eliminate or remove nests.
This further supports the hypothesis that tillage and other
disturbances that reduce ant species diversity also reduce a plot’s
potential for pest control by ants. Six of the species sampled in
this study nest in leaf litter, rotting logs, or dead branches of
trees (Coovert, 2005; Oswalt, 2007). Given the observation
that most gardeners intentionally remove leaf litter, woody
debris, and woody plants from their plots, it is not surprising
that species that depend on those mediums for nests are rarely
encountered in gardens. Simply adding or leaving these nesting sites in plots would likely contribute to more ant species,
as observed in an experiment conducted by Armbrecht et al.
(2004), where a greater diversity of nesting resources led to a
greater diversity of ant species.
How a garden and its surrounding area are managed greatly
influences the community of ant species within it. The most
common species in our sites are probably a consequence of disturbance caused by tilling and removal of biomass. Most species
collected were ruderal and stress tolerant as reflected by their functional groups (Table 3). In fact, the five most encountered species
are characterized by their ability to tolerate disturbance and temperature stress, which is often a consequence of gardening.
The community structure presents another factor to
consider regarding pest control. T. caespitum was the most
dominant competitor and second most abundant species based
on the tuna baiting samples, as it dominated most baits at the
hour mark, taking them over and guarding them from other
species. T. caespitum is an introduced species, highly competitive in disturbed areas, and negatively correlated with native
ant species where it occurs (Uno et al., 2010). It was rarely
observed foraging on plants, and its dominance may force
other ant species to forage on plants due to competition on the
soil surface. Interestingly, T. caespitum had a low frequency in
plots bordering two grass lawns, which averaged more species
than plots with one or zero sides bordering a lawn (Table 3;
Fig. 4). This may suggest that gardens surrounded by habitats
that support greater diversity allow other ant species to coexist
with T. caespitum due to increased nesting sites and resources.
Other studies support the potential for ant species as
predators on agricultural pests. Yusa (2001) found that mortality rates of Apple Snail egg masses were higher in areas of
urban agriculture & regional food systems
Fig. 7. Percent of encounters of the nine most abundant species
(ranked numbers 1–9) in plots that bordered grass lawns or one
or two grass lawns. Bars represent the percentage of encounters
of certain species within plots of the two groups. Species varied
in frequency within these groupings, but most species were
encountered more often in plots that bordered grass lawns than
plots that were neighbored on all sides by other plots.
high ant activity due to direct predation by Solenopsis geminata. These results also suggest that ants may be more effective
predators on soft-bodied eggs and insects than nymphs or
hard-encased eggs. Additionally, ants may influence below
ground herbivory, which may have an even greater effect on
plant performance and yield (Tao and Hunter, 2013).
Ant species are active at different temporal, seasonal, temperature, and spatial scales (Stuble et al., 2012). P. impairs,
which we expected to be an active predator based on recruitment and defense of tuna baits, showed no level of predation
(Table 4). This may be due to the time of day and year when
predation experiments were conducted. P. impairs nests are
less active if not dormant during the day, in summer months,
and at higher temperatures (Talbot, 1943; Stuble et al., 2012).
Indeed, our predation experiments were conducted under
conditions that were not favorable to P. impairs and possibly
other ants as well. Species of Camponotus are known to be
most active from dusk until dawn, while Aphaenogaster and
some species of Formica are most active in the late morning
and early afternoon (Andersen, 1997; Stuble et al., 2012).
Thus, the species tested and not tested may show higher rates
of predation at different times of day and in different seasons.
Not only does species activity vary with respect to temporal
changes, but also spatially. Functional groups shift in behavioral dominance ranking based on habitat structure (Andersen,
1997). Dominance hierarchy changes with temperature, time
of day, and levels of disturbance. For example, T. caespitum has
less of an advantage at cooler temperatures, is less active at night,
and possibly less competitive in plots with more ant species.
9
Within the context of pest control, the management and
surrounding habitat of gardens will affect which species are
present and how effective they are at competing for resources.
If one species is abundant and dominates the soil surface, it
may lead to niche partitioning where weaker species forage
on plants, thereby increasing the probability of predation on
pests. Thus, tillage intensity, soil texture, and surrounding
habitat should be taken into account to promote ant species
biodiversity. By supporting a variety of species active at different seasonal, temporal and spatial scales, pest suppression and
other ant services can persist throughout the growing season.
Conclusions
Community gardens have the potential to provide habitat
for a diversity of ants and other organisms. Soil texture and
level of disturbance are probably the most limiting factors
affecting ant diversity, with heavy clay soil and higher level
of disturbance to soil decreasing species diversity. However,
minimizing soil disturbance outside of garden plots and maximizing proximity to wooded areas may increase ant activity
within gardens because ants can disperse into gardens from
the surrounding habitat.
Multiple ant species are opportunistic predators of garden
pests; although experiments need to be conducted directly on
plants of interest to further support this. Moreover, a greater
understanding of species interactions within gardens, how
community structure affects foraging behavior, and ant species
ability to control pest populations is necessary for determining
the value of ants as pest control agents.
It is important to note that there is a continuum of tilling
techniques– not simply with or without a tractor. Understanding how different levels of till affect different species will be
useful in promoting species diversity. Finally, minimizing tilling and habitat disturbance both within and around garden
plots, along with avoiding or amending clay soils is likely to
lead to increased ant species diversity, and other insect diversity within gardens.
References
Altieri, M.A. 1999. The ecological role of biodiversity in agroecosystems. Agric. Ecosyst. Environ. 74(1):19–31. doi:10.1016/S01678809(99)00028-6
Andersen, A. 1997. Functional groups and patterns of organization in
North American ant communities: A comparison with Australia. J.
Biogeogr. 24(4):433–460. doi:10.1111/j.1365-2699.1997.00137.x
Armbrecht, I., I. Perfecto, and J. Vandermeer. 2004. Enigmatic biodiversity correlations: Ant diversity responds to diverse resources.
Science 304(5668):284–286. doi:10.1126/science.1094981
Ashebir, D., M. Pasquini, and W. Bihon. 2007. Urban agriculture in
Mekelle, Tigray state, Ethiopia: Principal characteristics, opportunities and constraints for further research and development.
Cities 24(3):218–228. doi:10.1016/j.cities.2007.01.008
Beard, R. L. (1940). The biology of Anasa tristis DeGeer with particular
reference to the Tachinid parasite, Trichopoda pennipes Fabr. Bull.
Conn. Agric. Exp. Stn. 440:595–680.
Brown, W.L. 2000. Diversity of ants. In: B.T. Bestelmeyer, D. Agosti,
L.E. Alonso, C.R.F. Brandão, W.L. Brown, Jr., J.H.C. Delabie,
and T.R. Schultz, editors, Ants: Standard methods for measuring and monitoring biodiversity. Smithsonian Institution
Press, Washington, DC. p. 49–53.
Collignon, B., and C. Detrain. 2010. Distributed leadership and
adaptive decision-making in the ant Tetramorium caespitum.
Proc. R. Soc. B: Biol. Sci. 277(1685):1267–1273. doi:10.1098/
rspb.2009.1976
Coovert, G.A. 2005. The ants of Ohio (Hymenoptera: Formicidae).
Bull. Ohio Biol. Surv. 15(2):1–202.
DuPont, S. T., H. Ferris, and M. Van Horn. 2009. Effects of cover crop
quality and quantity on nematode-based soil food webs and nutrient cycling. Appl. Soil Ecol. 41(2):157–167.
Ellison, A.M., N.J. Gotelli, E.J. Farnsworth, and G.D. Alpert. 2012. A
field guide to the ants of New England. Yale Univ. Press, New
Haven, CT.
Goddard, M.A., A.J. Dougill, and T.G. Benton. 2010. Scaling up from
gardens: Biodiversity conservation in urban environments. Trends
Ecol. Evol. 25(2):90–98. doi:10.1016/j.tree.2009.07.016
Gotelli, N. J., and R.K. Colwell. 2011. Estimating species richness.
In: A.E. Magurran and B.J. McGill, editors, Biological diversity:
Frontiers in measurement and assessment. Oxford Univ. Press,
New York. p. 39–54.
Hodgson, K., M.C. Campbell, and M. Bailkey. 2011. Urban agricul-
ture: Growing healthy, sustainable places. American Planning
Association, Chicago, IL .
Hölldobler, B., and E.O. Wilson. 1994. Journey to the ants: A story of
scientific exploration. Harvard Univ. Press, Cambridge, MA .
Isman, M.B. 2006. Botanical insecticides, deterrents, and repellents in modern agriculture and an increasingly regulated
world. Annu. Rev. Entomol. 51:45–66. doi:10.1146/annurev.
ento.51.110104.151146
Jackson, C.G., G.D. Butler, Jr., and D.E. Bryan. 1969. Time required
for development of Voria ruralis and its host, the cabbage looper, at
different temperatures. J. Econ. Entomol. doi:10.1093/jee/62.1.69
Lal, R. 1988. Effects of macrofauna on soil properties in tropical ecosystems. Agric. Ecosyst. Environ. 24(1):101–116.
doi:10.1016/0167-8809(88)90059-X
López, R., and D.A. Potter. 2000. Ant predation on eggs and larvae
of the black cutworm (Lepidoptera: Noctuidae) and Japanese
beetle (Coleoptera: Scarabaeidae) in turfgrass. Environ. Entomol.
29(1):116–125. doi:10.1603/0046-225X-29.1.116
McMahan, L.R. 2006. Understanding cultural reasons for the increase
in both restoration efforts and gardening with native plants.
Native Plants J. 7(1):31–34. doi:10.1353/npj.2006.0013
Mikheyev, A.S., and W.R. Tschinkel. 2004. Nest architecture of the
ant Formica pallidefulva: Structure, costs and rules of excavation.
Insectes Soc. 51(1):30–36. doi:10.1007/s00040-003-0703-3
Oswalt, D.A. (2007). Nesting and foraging characteristics of the black
carpenter ant Camponotus pennsylvanicus DeGeer (Hymenoptera: Formicidae). Ph.D. diss., Clemson University.
Pacheco, R., H.L. Vasconcelos, S. Groc, G.P. Camacho, and T.L. Frizzo.
2013. The importance of remnants of natural vegetation for maintaining ant diversity in Brazilian agricultural landscapes. Biodivers. Conserv. 22(4):983–997. doi:10.1007/s10531-013-0463-y
Pećarević, M., J. Danoff-Burg, and R.R. Dunn. 2010. Biodiversity on
Broadway—Enigmatic diversity of the societies of ants (Formicidae) on the streets of New York City. PLoS One 5(10):e13222.
doi:10.1371/journal.pone.0013222
Peck, S.L., B. McQuaid, and C.L. Campbell. 1998. Using ant species (Hymenoptera: Formicidae) as a biological indicator of
10
urban agriculture & regional food systems
agroecosystem condition. Environ. Entomol. 27(5):1102–1110.
doi:10.1093/ee/27.5.1102
Pimentel, D., D. Cerasale, R.C. Stanley, R. Perlman, E.M. Newman,
L.C. Brent, and D.T.I. Chang. 2012. Annual vs. perennial grain
production. Agric. Ecosyst. Environ. 161:1–9. doi:10.1016/j.
agee.2012.05.025
Shorey, H.H., L.A. Andres, and R.L. Hale. 1962. The biology of
Trichoplusia ni (Lepidoptera: Noctuidae): I. Life history and
behavior. Ann. Entomol. Soc. Am. 55(5):591–597. doi:10.1093/
aesa/55.5.591
Stuble, K.L., M.A. Rodriguez-Cabal, G.L. McCormick, I. Jurić, R.R.
Dunn, and N.J. Sanders. 2012. Tradeoffs, competition, and coexistence in eastern deciduous forest ant communities. Oecologia
171(4):981–992. doi:10.1007/s00442-012-2459-9
Talbot, M. 1943. Response of the ant Prenolepis imparis Say to temperature and humidity changes. Ecology 24(3):345–352.
doi:10.2307/1930536
Tao, L., and M.D. Hunter. 2013. Allocation of resources away from
sites of herbivory under simultaneous attack by aboveground
and belowground herbivores in the common milkweed, Asclepias
syriaca. Arthropod-Plant Interact. 7(2):217–224. doi:10.1007/
s11829-012-9235-y
urban agriculture & regional food systems
Uno, S., J. Cotton, and S.M. Philpott. 2010. Diversity, abundance, and
species composition of ants in urban green spaces. Urban Ecosyst.
13(4):425–441. doi:10.1007/s11252-010-0136-5
Way, M.J., and K.C. Khoo. 1992. Role of ants in pest management.
Annu. Rev. Entomol. 37(1):479–503. doi:10.1146/annurev.
en.37.010192.002403
Wheeler, G.C., J.N. Wheeler, and P.B. Kannowski. 1994. Checklist of
the ants of Michigan (Hymenoptera: Formicidae). Great Lakes
Entomol. 26(4):297–310.
Yadav, P., K. Duckworth, and P.S. Grewal. 2012. Habitat structure
influences below ground biocontrol services: A comparison
between urban gardens and vacant lots. Landscape Urban Planning 104(2):238–244.
Yitbarek, S., J.H. Vandermeer, and D. Allen. 2011. The combined effects
of exogenous and endogenous variability on the spatial distribution of ant communities in a forested ecosystem (Hymenoptera:
Formicidae). Environ. Entomol. 40(5):1067–1073. doi:10.1603/
EN11058
Yusa, Y. 2001. Predation on eggs of the apple snail Pomacea canaliculata
(Gastropoda: Ampullariidae) by the fire ant Solenopsis geminata.
J. Molluscan Stud. 67(3):275–279. doi:10.1093/mollus/67.3.275
11