Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
121
2017 NORTHEASTERN NATURALIST 24(2):121–136
Ecological Predictors and Consequences of Non-native
Earthworms in Kennebec County, Maine
Julia A. Rogers1 and Cathy D. Collins1,2,*
Abstract - Non-native earthworms are found throughout much of the United States and
southern Canada in areas glaciated during the most recent glaciation. Following invasion,
these earthworms altered nutrient cycling, soil structure, and diversity in forests throughout
the northern United States. There are no comprehensive studies of earthworm distributions
in forested areas of Maine. We surveyed earthworms in forested recreation areas in Kennebec
County, ME, and investigated ecological and landscape attributes that may predict
their presence. To examine whether the presence of worms modifies forests, we measured
environmental variables known from other studies to be affected by worms. We found
earthworms at 12 out of 23 sites. Sample sites near roads, in deciduous forests, and in small
forests were more likely to have earthworms. We also found that locations with worms have
less surface litter and more soil phosphorous, suggesting that earthworms modify soils in
Maine forests. Our study is the first to explore the distribution of earthworms in natural forests
in Maine, and our findings provide evidence that roads facilitate earthworm invasion,
with measurable consequences for soil properties.
Introduction
Earthworms currently found in previously glaciated regions of North America
are considered invasive (Bohlen et al. 2004a). Introduced earthworms can have
profound impacts on soils, plant communities, and nutrient cycling (Bohlen et al.
2004b, Davalos et al. 2013, Laossi et al. 2009). Humans facilitate the spread of
earthworms via road construction, gardening, logging, and fishing (Bohlen et al.
2004a, Hale 2007, Hendrix and Bohlen 2002, Holdsworth et al. 2007, Kalisz and
Dotson 1989). In some regions, earthworm invasion is actively monitored (Bohlen
et al. 2004a, Hale et al. 2006). However, few systematic earthworm surveys have
been conducted in Maine, and the extent to which worms have invaded—and potentially
influenced—forested landscapes is unknown.
Predicting the effects of earthworm invasion on forest properties is made challenging
by the fact that the magnitude and direction of effects depend on forest
composition, land-use history, and soil type (Bohlen et al. 2004a, Frelich et al. 2006,
Hendrix and Bohlen 2002). For instance, earthworms alter carbon, phosphorous,
and nitrogen levels through their consumption of organic matter and incorporation
of this organic matter into the mineral soil (Bohlen et al. 2004b, Frelich et al. 2006,
Scheu and Parkinson 1994). However, whether worms increase or decrease soil
phosphorous and nitrogen depends on land-use history, the species of earthworms
present, and the time since invasion (Bohlen et al. 2004b).
1Department of Biology, Colby College, Waterville, ME 04901. 2Biology Program, Bard
College, Annandale on Hudson, NY 12540. *Corresponding author - ccollins@bard.edu.
Manuscript Editor: Robert Bertin
Northeastern Naturalist
122
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
By altering nutrients, earthworms indirectly affect the abundance and diversity
of belowground microbial communities (Alban and Berry 1994, Bohlen et al.
2004b, McLean and Parkinson 1998), as well as the diversity and invasibility of
above-ground plant communities (Bohlen et al. 2004b, Hale et al. 2008). For example,
in mixed hardwood forests in Ontario, Canada, earthworms directly alter
plant community composition via selective seed consumption (Cassin and Kotanen
2016). Earthworms may also enhance seedling emergence by increasing nutrients
available to the seed (Eisenhauer and Scheu 2008, Eisenhauer et al. 2007, Milcu et
al. 2006). Given the widespread and inconsistent effects of earthworms on soils and
plants, studies from a broad range of ecosystems are needed to better understand
and predict the changes to forests following earthworm invasion.
The dramatic influence of earthworms on ecosystem properties has prompted
recent efforts to identify environmental factors that predict where earthworms occur
and will invade (Cameron et al. 2007, Costello et al. 2011, Gundale et al. 2005,
Sackett et al. 2012, Suarez et al. 2006). For instance, forest type is the strongest
predictor of earthworm presence in New York, where earthworms were more likely
to be found in mixed hardwood forests than in Fagus (beech) and Tsuga (hemlock)
forests (Suarez et al. 2006). Disturbance, too, plays a key role in predicting the
earthworm presence. Earthworms are more likely to be found near agricultural
clearings (Suarez et al. 2006), close to fishing sites (Cameron et al. 2007), and
along roads experiencing regular vehicle traffic (Cameron et al. 2007, Sackett et al.
2012). Additionally, earthworms are associated with non-wilderness sites more than
wilderness sites, a pattern likely explained by the presence of roads and logging at
the former sites (Gundale et al. 2005).
Little is known about the distribution of earthworms in natural habitats in
Maine. Approximately 90% of land in Maine is forested (Huff and McWilliams
2015), much of which is used for logging and recreation and therefore vulnerable
to human-mediated earthworm invasion (Gundale et al. 2005). Moreover most of
Maine’s forests are second-growth forests, where earthworms appear to establish
more readily relative to old-growth forests (Simmons et al. 2015). Reynolds (2008)
reported that earthworms were present in each of Maine’s counties; however his
sampling was restricted to backyards, compost piles, and towns. Owen and Galbraith
(1989) studied earthworms in relation to Scolopax minor Gmelin (American
Woodcock) populations in six townships in central and eastern Maine. They found
that land-use history and soil type were the best predictors for earthworm presence.
Areas that were farmed previously were the most likely to have earthworms
regardless of other characteristics. Additionally, earthworms were more abundant
in moderately drained fine sandy loamy soils than in other soil types (Owen and
Galbraith 1989).
We surveyed forests in Kennebec County, ME, to assess the distributional extent
of earthworms and characterize the environmental factors associated with sites
where they are present. Our objectives were to (1) record the extent of earthworm
presence in Kennebec County, (2) identify landscape and soil factors that predict
earthworm presence, and (3) investigate the effects of earthworms on soils in the
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
123
invaded areas. Based on studies conducted in forests in other states, we expected
that distance to roads would be the most significant factor in predicting earthworm
presence. Moreover, we expected earthworms to reduce forest litter and alter soil
N and P quantities.
Methods
Location and sampling design
We selected 23 study sites in Kennebec County, ME (Fig. 1). Temperatures in
this area average -2.4 °C in January and 26.2 °C in June, and the average annual
precipitation is 1064 mm (US Climate Data 2016). Earthworm species respond differently
to the cold; however, worm species found in the litter layer, which may be
more susceptible to freezing than other species, can survive temperatures as low
as -14 °C (Greiner et al. 2011, Holmstrup et al. 2007). That threshold is lower than
the averarge minimum winter temperature in our study area (United States Climate
Data 2016), and though many parts of the state regularly experience winter low
air temperatures below -14 °C, the temperatures under the litter and in the ground
Figure 1. A map of Kennebec County, ME. Circles represent survey transects where worms
were present (black circles) and transects where worms were absent (white circles).
Northeastern Naturalist
124
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
where the earthworms live are generally not as extreme and the typical presence
of an additonally insulating blanket of snowcover throughout most of the winter
in these colder areas suggests the climate throughout Maine is in fact suitable for
earthworm survival.
All sample sites were located in forested conservation and recreation areas that
included contiguous forest at least 1 ha in size . On each soil type within a site, we
haphazardly selected a location for 1 transect. Consequently, the number of transects
per site was dictated by the number of soil types per conservation area, resulting in 36
total transects. Across all sites, there were 10 different soil types (Table 1). Transects
were each 50 m long and located at least 5 m from human disturbance (e.g. trails), although
all transects were within 1 km of a paved road.
Worm sampling
We sampled worms from five 25 cm x 25 cm plots located at 10-m intervals along
each of the 36 transects. To minimize heterogeneity due to above- or belowground
obstructions, we avoided placing plots within 1 m of any decomposing stumps and
trees larger than 10 cm in diameter at breast height (dbh). We first cleared the litter
and hand-collected worms on the surface. To stimulate emergence of worms from
within the ground, we poured a solution of 3.8 L of water and 40 g of mustard seed
powder on the area (Lawrence and Bowers 2002). Though all species of worms may
not respond equally to mustard extraction, this sampling technique is as effective as
digging and hand-sorting (Hale et al. 2005, Lawrence and Bowers 2002), and causes
less disturbance to the forest soils. We standardized sampling effort by collecting
emerging earthworms for 10 minutes per plot. Data from 5 plots were compiled such
that presence or absence of earthworms was expressed at the transect scale.
We sampled all locations for earthworms between 22 September 2015 and 3
November 2015, as earthworms are known to be most active during the spring
and fall months (Gates 1961). A subset of 21 transects was sampled a second time
during the latter part of that period following rain and cooler temperatures, which
we suspected might alter abundance. Not all sites were re-sampled due to freezing
weather that decreased earthworm activity. However, because we did not find earthworms
in any places where we had recorded them absent before (nor did we fail to
find earthworms in locations where they had been recorded as present), re-sampling
increased our confidence that the absence of earthworms from sites during the initial
sampling was not caused by lower worm-activity levels or ineffective methods.
Abundance varied markedly depending on the time of sampling, thus we use only
presence–absence data in our analyses.
Environmental variables
To characterize forest composition, we established belt transects by expanding
the 50-m worm-sampling transect to include 2 m on either side. Within the belt (4 m
x 50 m, 200 m2 total), we measured and identified to species all trees larger than 10
cm dbh.
To characterize soil attributes, we collected 6.2 cm3 of soil from the top 10
cm of soil (after clearing litter) at each sub-plot and combined the 5 sub-plot
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
125
Table 1. Location name, soil type, forest size, and latitude and longitude for all transects.
Forest
Location Soil type size (ha) Lat., Long. (°)
Runnals Hill, Colby College Paxton Charlton very 115 44.560, 69.668
stony fine sandy loam
Runnals Hill, Colby College Hollis fine sandy loam 115 44.558, 69.665
Perkins Arboretum, Colby College Buxton silt loam 76 44.558, 69.657
Perkins Arboretum, Colby College Buxton silt loam 76 44.560, 69.654
Quarry Road Ski Area, Waterville Scantic silt loam 81 44.578, 69.654
Quarry Road Ski Area, Waterville Hollis fine sandy loam 81 44.576, 69.652
Quarry Road Ski Area, Waterville Buxton silt loam 81 44.573, 69.653
Mount Phillip, Rome Lyman loam 275 44.586, 69.884
Mount Phillip, Rome Berkshire very stony 275 44.579, 69.883
fine sandy loam
Round Top, Rome Lyman loam 1019 44.533, 69.928
Round Top, Rome Berkshire very stony 1019 44.530, 69.924
fine sandy loam
Round Top, Rome Peru fine sandy loam 1019 44.530, 69.923
Sanders Hill, Rome Lyman loam 294 44.561, 69.930
Sanders Hill, Rome Berkshire very stony 294 44.567, 69.923
fine sandy loam
French Mountain, Rome Lyman loam 205 44.574, 69.919
Seaward Mills Vassalboro Buxton silt loam 43 44.400, 69.634
Davidson Nature Preserve, Vassalboro Hollis fine sandy loam 397 44.453, 69.645
Vassalboro Wildlife Habitat Scantic silt loam 15 44.409, 69.668
Vassalboro Wildlife Habitat Buxton silt loam 15 44.409, 69.672
Woodsmen Field, Colby College Woodbridge very stony 11 44.565, 69.668
fine sandy loam
Jamie’s Pond WMA, Hallowell Paxton Charlton very 290 44.286, 69.852
stony fine sandy loam
Reynolds Forest, Sidney Suffield silt loam 223 44.420, 69.715
Oxbow, Waterville Scantic silt loam 9 44.546, 69.642
Vaughan Woods, Hallowell Suffield silt loam 98 44.276, 69.797
Woodbury Pond State Park, Litchfield Paxton Charlton very 38 44.202, 69.957
stony fine sandy loam
Mt. Pisgah Conservation Area, Winthrop Paxton Charlton very 1144 44.301, 70.035
stony fine sandy loam
Small-Burnham Conservation Area, Litchfield Woodbridge very stony 706 44.157, 69.927
fine sandy loam
Small-Burnham Conservation Area, Litchfield Hinckley gravelly 706 44.156, 69.933
sandy loam
Parker Pond Headlands, Fayette Paxton very stony fine 136 44.487, 70.036
sandy loam
Torsey Pond, Mt. Vernon Woodbridge very stony 92 44.418, 70.000
fine sandy loam
MacDonald Conservation Area, Readfield Woodbridge very stony 424 44.373, 70.013
fine sandy loam
Hutchinson Pond, Manchester Woodbridge very stony 166 44.267, 69.883
fine sandy loam
Wyman Memorial Forest, Readfield Hollis fine sandy loam 418 44.360, 69.868
Augusta Arboretum Hollis fine sandy loam 39 44.299, 69.762
Augusta Nature Center Suffield silt loam 24 44.315, 69.753
Augusta Nature Center Hollis fine sandy loam 24 44.315, 69.754
Northeastern Naturalist
126
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
samples into a single sample per transect. Soils were air-dried for 2 weeks prior
to analysis. Soil moisture-holding capacity (SMHC; our proxy for soil moisture)
was measured by taking the difference in mass between soil wetted to field capacity
and soil oven dried at 105 °C for 72 hours (Brudvig and Damschen 2011).
Nutrient and texture analyses were conducted by Brookside Laboratories, Inc.
(New Knoxville, OH; www.blinc.com). We focused our analyses on soil texture
Table 2. Landscape and local habitat variables that influence and thus serve as predictors in our statistical
models) or are influenced by (response variable in our statistical models) earthworm presence.
Variable Location System Study
Predictors of earthworm presence
Soil pH Puerto Rico Forest González et al. 2007
Maine Forest Owen and Gailbraith 1989
India Agricultural area Singh et al. 2015
Georgia Forest Lobe et al. 2013
New York Forest Homan et al. 2015
Europe Forest Wandeler et al. 2016
Soil texture Germany Agriculture areas Palm et al. 2013
Ontario, Canada Forests Sackett et al. 2012
Maine Forests Owen and Galbraith 1989
Soil moisture Himalayas Agricultural field Kaushal et al. 1999
Distance to roads Alberta, Canada Forests Cameron et al. 2007
Minnesota and Wisonsin Forests Holdsworth et al. 2007
Ontario, Canada Forests Sackett et al. 2012
Distance to water New York Forests Suarez et al. 2006
Minnesota and Wisonsin Forests Holdsworth et al. 2007
Forest composition New York Forests Suarez et al. 2006
Maine Forests Owen and Galbraith 1989
Influenced by earthworm presence
Soil N Minnesota Forests Alban and Berry 1994
Alberta, Canada Forests Scheu and Parkinson 1994
New York Forests Burtelow et al. 1998
New York Forests Bohlen et al. 2004b
Minnesota Forests Frelich et al. 2006
Greenhouse Greenhouse–forests Hale et al. 2008
Quebec, Canada Forests Wironen and Moore 2006
Soil C Minnesota Forests Alban and Berry 1994
New York Forests Burtelow et al. 1998
New York Forests Bohlen et al. 2004b
Michigan Forests Gundale et al. 2005
Soil P New York Forests Groffman et al. 2004, 2015
Minnesota Forests Frelich et al. 2006
Michigan Forests Hale et al. 2007
Litter depth Illinois Forests Heneghan et al. 2007
Puerto Rico Forests Gonzalez et al. 2003
Puerto Rico Forests and fields Liu and Zou 2002
Minnesota Forests Frelich et al. 2006
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
127
(percent silt, sand, and clay), pH, nitrogen levels, organic matter, and phosphorus
levels because other studies have shown that they affect or are affected by earthworm
presence (Table 2).
Statistical analysis
All statistical analyses were conducted in R version 3.2.2 (R Core Team 2015)
and ArcGIS (ESRI, Redlands, CA). We used ArcGIS to determine Euclidean
distance to roads and water, and to confirm soil types for each transect based on
soil-map data from the USGS Web Soil survey (USDA 2013). Based on previously
published studies, we divided the soil variables according to whether they were
more likely to influence or be influenced by earthworm presence (Table 2). Using
logistic regression, we determined whether earthworm presence was predicted by
landscape-level variables (distance to roads, distance to water, and forest size),
environmental factors unlikely to be changed by worms (soil texture, SMHC, and
pH), and tree composition (specifically the proportion of deciduous trees). Because
transect locations signify unique soil types and associated vegetation, we treated
each transect independently.
For model selection, we started with a full model including soil pH, soil
texture (as percent sand and percent silt), SMHC, distance to roads, distance to
water, and forest composition (Table 3). Because some soil variables may be influenced
by roads, we included interactions between environmental variables and
distance to roads in the initial model (Table 3). We then removed each factor individually
in order of least significance and tested for model significance to create
the simplest model. We used the Akaike information criterion (AIC) to select the
best-fitting model.
To determine the effect of earthworms on soil properties we first used nonmetric
multidimensional scaling (NMDS) to visualize the data. We included percent soil
Table 3. Results from the full model, prior to variable selection. All predictors in this model are ecological
and landscape level factors previously found to influence the presence of earthworms (Table
2). SMHC is soil moisture holding capacity, DisRoad is the distance to roads, SA.per is the percent
sand, Prop.Decid is the proportion of deciduous trees. We included only interactions between variables
that may be influenced by distance to roads (see text).
Estimate Std. Error z value P
Intercept -41.98 45.72 -0.92 0.36
pH 6.23 8.18 0.76 0.05
DisRoad 0.04 0.14 0.24 0.81
SA.per 0.07 0.19 0.35 0.73
SMHC 0.16 0.21 0.75 0.46
Prop.Decid 4.98 10.67 0.47 0.64
Distance to Water 0.002 0.01 0.33 0.74
Forest Size -0.01 0.01 -1.46 0.15
DisRoad * pH -0.001 0.03 -0.04 0.97
DisRoad * SA.per -0.0002 0.001 -0.29 0.77
DisRoad * SMHC -0.001 0.001 -0.77 0.44
DisRoad * Prop.Decid 0.004 0.05 0.07 0.94
Northeastern Naturalist
128
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
organic matter, average litter depth (cm), estimated nitrogen release via organic
matter decomposition (N/ha), and soil phosphorous levels (mg/kg), all of which
have previously been shown to be affected by earthworms (Table 2). To test whether
the combined effect of soil variables generated statistically significant differences
between sites with and without worms, we used a permutation multivariate analysis
of variance (PERMANOVA) from the “vegan” package in R. We followed this
multivariate approach with 2-sample t-tests to identify individual soil properties
that differed between sites with and without worms.
Results
Earthworms were found at 12 out of 23 sites (16 out of 36 transects; Fig. 1).
Three sites included transects with and transects without worms. In the 22 transects
that were re-sampled, all presences and absences were confirmed.
Following model selection, our final model included distance to roads, forest
composition, forest size, and pH (AIC final model = 33.61, AIC full model = 45.37;
Table 4). Distance to roads, forest composition, and forest size were significant
predictors (Table 4). Although not statistically significant, retaining pH improved
model fit.
Specifically, invasive earthworms were more likely to be found near roads. Transects
varied from 26 m to 870 m from roads; the mean distance from transects to
roads was 156 m (± 26 SE) where earthworms were present and 373 m (± 62 SE)
where they were absent. The pH ranged from 4.6 to 5.5 across all sites. The mean
pH was higher where earthworms were present (5.08 ± 0.06 SE) than where they
were absent (4.92 ± 0.05). Forests where earthworms were present were on average
smaller (mean ± SE = 145 ± 44 ha), than those without earthworms (413 ± 91 ha).
Lastly, the proportion of deciduous trees was slightly higher under the presence of
earthworms (mean ± SE = 0.76 ± 0.05), than without earthworms (0.63 ± 0.06).
Table 5. Results from two sample t-tests for soil factors including nitrogen, phosphorus, litter depth,
and percent organic matter. * indicates significant results.
Test t df P
N~Presence 1.28 34.00 0.21
P~Presence -2.25 33.78 0.03*
LD~Presence 4.96 20.59 less than 0.001*
OM~Presence 1.57 32.32 0.13
Table 4. Results from final model with soil pH, distance to roads, the proportion of deciduous trees,
and forest size. * indicates significant results.
Estimate SE z value P
Intercept -23.00 12.96 -1.87 0.08
pH 4.24 2.32 1.83 0.07
DisRoad -0.01 0.01 -2.07 0.03*
Prop.Decid 7.94 3.32 2.39 0.02*
Forest Size -0.01 0.002 -2.22 0.02*
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
129
Transects with and without worms differed with respect to soil attributes previously
shown to be influenced by earthworm presence (PERMANOVA: F35 = 5.69,
P = 0.002; Fig. 2, Table 5). When soil attributes were analyzed separately, transects
with earthworms had significantly higher levels of phosphorus and lower litter
depths than those without worms (Fig. 3, Table 5). We did not detect a difference in
levels of organic matter and nitrogen between sites with and without worms (Fig. 3,
Table 5).
Discussion
Non-native earthworms are widespread in Kennebec County forests, though not
present at all sites we sampled. Soil texture appeared to have little bearing on the
likelihood of worm invasion or persistence relative to landscape attributes, as we
found earthworms in distinct soil types even within a single forests. Worms were
detected most frequently near roads, in smaller forests with fewer conifers, and in
soils with higher pH. Soils where worms were present differed with respect to several
variables related to nutrient cycling (litter depth and phosphorous).
Figure 2. Non-metric dimensional scaling (NMDS) ordination of soil variables potentially
impacted by worms (estimated nitrogen release, soil phosphorous, litter depth, and percent
organic matter; based on literature in Table 2). Black circles depict sites where worms were
detected. Gray circles are sites where worms were surveyed, but not found.
Northeastern Naturalist
130
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
Predictors of earthworm presence
Proximity to roads was a strong predictor of worm presence in Kennebec County,
similar to studies from other regions (Cameron and Bayne 2009, 2015; Gundale
et al. 2005; Holdsworth et al. 2007; Sackett 2012; Shartell et al. 2015). Roads
provide an avenue for earthworm invasion during the construction phase when
bulk gravel or other fill is transported from other locations (Cameron et al. 2007,
Hendrix and Bohlen 2002). Both during and after road construction, earthworm
cocoons can travel along roadways in substrate attached to vehicle tires (Marinissen
and van den Bosch 1992). In a sparsely populated state like Maine, lack of roads,
or infrequently traveled roads may have slowed the spread of earthworms relative
to more-populated states. Nonetheless, introductions through fishing, greenhouses,
composting, gardens, road construction, and other factors have led to worms being
found in every county in Maine (Reynolds 2008). If road construction, travel,
and recreation continue to spread to remote areas in the state (MaineDot 2015), the
capacity for earthworms to spread to natural areas will also increase.
Figure 3. Boxplots of soil factors potentially affected by earthworms. Asterisks denote statistically
significant differences (*P < 0.05, **P < 0.001; see Table 5). For each, whiskers
represent max and min values, top of the box is the third quartile, bottom of the box is the
first quartile, and the darker line within the box is the median .
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
131
Our data may reflect co-occurring disturbances, as many sites with worms
were in small forests near the towns of Waterville and Augusta (Fig. 1). Roads and
forest fragmentation are typically coupled with other disturbances that facilitate
earthworm invasion such as logging (Gundale et al. 2005, Sackett 2012), agricultural
fields (Shartell et al. 2015), urban development (Beausejour et al. 2015),
and recreational facilities (Bohlen et al. 2004a, Holdsworth et al. 2007). Smaller
patches of forests have a greater edge-to-area ratio; closer proximity to anthropogenic
activity and disproportionate access for worm introduction on forest edges
may be why some species of earthworms occur more frequently on edges of forests
relative to the interior (Gibson et al. 2013).
While roads facilitate the spread of worms, and small forests likely increase access
for invasion, our data show that not all forest types are equally likely to have
worms. Both the proportion of deciduous trees and soil pH played a role in predicting
earthworm presence, perhaps in part because conifer-dominated forests have
more acidic soils than deciduous forests (Frelich et al. 2006, Suarez et al. 2006).
However, it is worth noting that we found earthworms both in sites dominated by
conifers and sites dominated by deciduous trees, suggesting that access to sites via
roads is a stronger determinant of earthworm presence.
It is certainly plausible that earthworm presence in Kennebec county may reflect
factors that we did not measure such as land-use history (Simmons et al. 2015),
proximity to wet areas (Suarez et al. 2006), proximity to agricultural fields (Shartell
et al. 2015, Suarez et al. 2006), or proximity to logging operations (Costello et al.
2011 Gundale et al. 2005, Sackett et al. 2012). Beginning in the 1800s, logging has
occurred in the majority of forests in Kennebec County and elsewhere in the state
of Maine (Moore and Whitham 1996). Following the logging in the 1800s, much
of the land was used as farmland, both for crops and for sheep farming. Some of
these farms were abandoned and returned to forests, while others remain as functional
farms (Moore and Whitham 1996). Mapping worm distribution in relation to
proximity to agriculture, especially crop farming, and land-use history may help us
better interpret current patterns, as well as predict the location of future earthworm
invasions (Owen and Galbraith 1989, Suarez et al. 2006).
Effects of earthworms on forest soils
Consistent with other studies (Bohlen et al. 2004b, Burtelow et al. 1998), soil
nitrogen levels did not differ between sites with and without earthworms. However,
in contrast to studies in forests in New York (Suarez et al. 2004) and Minnesota
(Resner et al. 2015), we found that soil phosphorus was higher where earthworms
were present. Earthworms may modify phosphorous to varying degrees depending
on the species of earthworms present and how long they have been at a site (Bohlen
et al. 2004a, Frelich et al. 2006, Resner et al. 2015). For instance, the presence of
Lumbricus terrestris L. (Nightcrawler) is thought to bring soils from deeper horizons
to the surface, increasing the available phosphorous (Frelich et al. 2006).
We did not identify worms to the species level, so we cannot say for sure whether
species composition explains higher soil phosphorus in our study. Following
Northeastern Naturalist
132
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
earthworm invasion, however, soil phosphorous initially increases, then decreases
(Bohlen et al. 2004a); it is possible we measured forests during this early phase and
that over time, differences will fade.
Our finding that sites with earthworms had less surface litter was expected given
that earthworms consume litter (Bohlen et al. 2004a). Litter reduction by earthworms
modifies soil nutrient availability and understory plant composition (Frelich et al.
2006, Gonzalezir et al. 2003, Heneghan et al. 2007, Liu and Zhou 2002). It follows
that worms invading Maine forests may ultimately have consequences for plant
community diversity and ecosystem function. Furthermore, litter consumption
and nutrient cycling are not the only mechanisms by which worms may influence
aboveground plant communities—including plants of economic importance to the
state of Maine. For instance, Lawrence et al. (2003) found that earthworms reduce
the colonization and presence of hyphae in mycorrhizal fungi associated with Acer
saccharum Marsh (Sugar Maple). Moreover, earthworms may create conditions
conducive for invasive plant species, including Rhamnus carthartica L. (Common
Buckthorn), Alliaria petiolata M. Bieb. (Garlic Mustard), and Rosa multiflora
Thunb. (Multiflora Rose) (Clause et al. 2015, Hopfensperger and Hamilton 2015,
Nuzzo et al. 2015, Quakenbush et al. 2012, Roth et al. 2015, Whitfeld et al. 2014).
Finally, because earthworms consume small-seeded species (Cassin and Kotenen
2016), their presence may influence aboveground plant composition. While we did
not address plant communities in this study, the fact that we observed dramatic differences
in soil properties in invaded versus uninvaded forests warrants future studies
on the aboveground consequences of worms in Maine forests.
Overall, we found that earthworms are present—particularly in small, deciduous
forests near roads—and induce measurable changes to soils in Kennebec County
forests. Larger-scale systematic surveys are needed to document the extent of the
invasion in Maine and better predict the ecosystem consequences for forests that
developed in the absence of worms for most of the last 10,000 years.
Acknowledgments
We are grateful to the land owners and conservation organizations that allowed access
to their property for this research. We would also like to thank Judy Stone and Herb Wilson
for their input on early drafts of this manuscript. This project was supported by funds from
the Colby Biology Department and the Dean of Students Special Funds at Colby College.
Literature Cited
Alban, D.H., and E.C. Berry. 1994. Effects of earthworm invasion on morphology, carbon,
and nitrogen of a forest soil. Applied Soil Ecology 1:243–49.
Beauséjour, R., I.T., Handa, M.J. Lechowicz, B. Gilbert, and M. Vellend. 2015. Historical
anthropogenic disturbances influence patterns of non-native earthworm and plant invasions
in a temperate primary forest. Biological Invasions 17:1267–1281.
Bohlen, P.J., S. Scheu, C.M. Hale, M.A. McLean, S. Migge, P.M. Groffman, and D. Parkinson.
2004a. Non-native invasive earthworms as agents of change in northern temperate
forests. Frontiers in Ecology and the Environment 2:427–435.
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
133
Bohlen, P. J., D.M. Pelletier, P.M. Groffman, T.J. Fahey, and M.C. Fisk. 2004b. Influence
of earthworm invasion on redistribution and retention of soil carbon and nitrogen in
northern temperate forests. Ecosystems 7:13–27.
Brudvig, L.A., and E.I. Damschen. 2011. Land-use history, historical connectivity, and land
management interact to determine Longleaf Pine woodland understory richness and
composition. Ecography 34:257–266.
Burtelow, A.E., P.J. Bohlen, and P.M. Groffman. 1998. Influence of exotic earthworm invasion
on soil organic matter, microbial biomass, and denitrification potential in forest
soils of the northeastern United States. Applied Soil Ecology 9:197–202.
Cameron, E.K., and E.M. Bayne. 2009. Road age and its importance in earthworm invasion
of northern boreal forests. Journal of Applied Ecology 46:28–36.
Cameron, E.K., and E.M. Bayne. 2015. Spatial patterns and spread of exotic earthworms at
local scales. Canadian Journal of Zoology 93:721–726.
Cameron, E.K., E.M. Bayne, and M.J. Clapperton. 2007. Human-facilitated invasion of
exotic earthworms into northern boreal forests. Ecoscience 14:482–490.
Cassin, C.M., and P.M. Kotanen. 2016. Invasive earthworms as seed predators of temperate
forest plants. Biological Invasions 18:1–14.
Clause, J., E. Forey, C.J. Lortie, A.M. Lambert, and S. Barot. 2015. Non-native earthworms
promote plant invasion by ingesting seeds and modifying soil properties. Acta Oecologica,
64:10–20.
Costello, D.M., S.D. Tiegs, and G.A. Lamberti. 2011. Do non-native earthworms in southeast
Alaska use streams as invasional corridors in watersheds harvested for timber?
Biological Invasions 13:177–187.
Dávalos, A., V. Nuzzo, J. Stark, and B. Blossey. 2013. Unexpected earthworm effects on
forest understory plants. BMC Ecology 13:48.
Eisenhauer, N., and S. Scheu. 2008. Invasibility of experimental grassland communities: The
role of earthworms, plant functional group identity, and seed size. Oikos 117:1026–1036.
Eisenhauer, N., S. Partsch, D. Parkinson, S. Scheu. 2007. Invasion of a deciduous forest
by earthworms: Changes in soil chemistry, microflora, microarthropods, and vegetation.
Soil Biology and Biochemistry 39:1099–1110.
Frelich, L.E., C.M. Hale, S. Scheu, A.R. Holdsworth, L. Heneghan, P.J. Bohlen, and P.B.
Reich. 2006. Earthworm invasion into previously earthworm-free temperate and boreal
forests. Biological Invasions 8:1235–1245.
Gates, G.E. 1961. Ecology of some earthworms with special reference to seasonal activity.
The American Midland Naturalist 66:61–86.
Gibson, K.D., P.M. Quackenbush, N.C. Emery, M.A. Jenkins, and E.J. Kladivko. 2013. Invasive
earthworms and plants in Indiana old- and second-growth forests. Invasive Plant
Science and Management 6:161–174.
González, G., E. García, V. Cruz, S. Borges, M. Zalamea, and M.M. Rivera. 2007. Earthworm
communities along an elevation gradient in Northeastern Puerto Rico. European
Journal of Soil Biology 43:24–32.
Gonzalezir, V.G., T.R. Seastedt, and Z. Donate. 2003. Earthworms, arthropods, and plant
litter decomposition in Aspen (Populus tremuloides) and Lodgepole Pine. Pedobiologia
47:863–869.
Greiner, H.G., A.M.T. Andrew, and S.D. Tiegs. 2011. Cold tolerance among composting
earthworm species to evaluate invasion potential. American Midland Naturalist 166:
349–357.
Groffman, P.M., P.J. Bohlen, M.C. Fisk, and T.J. Fahey. 2004. Exotic earthworm invasion
and microbial biomass in temperate forest soils. Ecosystems 7:45–54.
Northeastern Naturalist
134
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
Groffman, P.M., T.J. Fahey, M.C. Fisk, J.B. Yavitt, R.E. Sherman, P.J. Bohlen, and J.C.
Maerz. 2015. Earthworms increase soil microbial biomass carrying capacity and nitrogen
retention in northern hardwood forests. Soil Biology and Biochemistry 87:51–58.
Gundale, M. J., W.M. Jolly, and T.H. Deluca. 2005. Susceptibility of a northern hardwood
forest to exotic earthworm invasion: Earthworm invasion susceptibility. Conservation
Biology 19:1075–1083.
Hale, C.M. 2007. Evidence for human-mediated dispersal of exotic earthworms: Support
for exploring strategies to limit further spread. Molecular Ecology 17:1165–1169.
Hale, C.M., L.E. Frelich, P.B. Reich, and J. Pastor. 2005. Effects of European earthworm
invasion on soil characteristics in northern hardwood forests of Minnesota, USA. Ecosystems
8:911–927.
Hale, C. M. L.E. Frelich, and P.B. Reich. 2006. Changes in hardwood forest understory plant
communities in response to European earthworm invasions. Ecology 87:1637–1649.
Hale, C. M., L.E. Frelich, P.B. Reich, and J. Pastor. 2008. Exotic earthworm effects on hardwood
forest floor, nutrient availability, and native plants: A mesocosm study. Oecologia
155:509–518.
Hendrix, P.F., and P.J. Bohlen. 2002. Exotic earthworm invasions in North America: Ecological
and policy implications. BioScience 52:801–811.
Heneghan, L., J. Steffen, and K. Fagen. 2007. Interactions of an introduced shrub and introduced
earthworms in an Illinois urban woodland: Impact on leaf-litter decomposition.
Pedobiologia 50:543–551.
Holdsworth, A.R., L.E. Frelich, and P.B. Reich. 2007. Regional extent of an ecosystem
engineer: Earthworm invasion in northern hardwood forests. Ecological Applications
17:1666–1677.
Holmstrup, M., J. Overgaard, A.M. Bindesbo, C. Pertoldi, and M. Bayley. 2007. Adaptations
to overwintering in the earthworm Dendrobaena octaedra: Genetic differences
in glucose mobilization and freeze tolerance. Soil Biology and Biochemistry
39:2640–2650.
Homan, C., C. Beier, T. McCay, and G. Lawrence. 2016. Application of lime (CaCO3) to
promote forest recovery from severe acidification increases potential for earthworm
invasion. Forest Ecology and Management 368:39–44.
Hopfensperger, K.N., and S. Hamilton. 2015. Earthworm communities in previously glaciated
and unglaciated eastern deciduous forests. Southeastern Naturalist 14:66–84.
Huff, E.S., and W.H. McWilliams. 2016. Forests of Maine, 2015. Resource Update FS-86.
US Department of Agriculture, Forest Service, Northern Research Station, Newtown
Square, PA.
Kalisz, P.J., and D.B. Dotson. 1989. Land-use history and the occurrence of exotic
earthworms in the mountains of Eastern Kentucky. American Midland Naturalist
122:288–297.
Kaushal, B.R., B. Kandpal, S.P. Bisht, S. Bora, and R. Dhapola. 1999. Abundance and
seasonal activity of earthworms in croplands of Kumaon Himalayas. European Journal
of Soil Biology 35:171–176.
Laossi, K.R., D.C. Noguera, A. Bartolomé-Lasa, J. Mathieu, M. Blouin, and S. Barot. 2009.
Effects of an endogeic and an anecic earthworm on the competition between four annual
plants and their relative fecundity. Soil Biology and Biochemistry 4:1668–1673.
Lawrence, A.P., and M.A. Bowers. 2002. A test of the “hot” mustard extraction method of
sampling earthworms. Soil Biology and Biochemistry 34:549–552.
Lawrence, B., M.C. Fisk, T.J. Fahey, and E.R. Suarez. 2003. Influence of nonative earthworms
on mycorrhizal colonization of Sugar Maple (Acer saccharum). New Phytologist
157:145–153.
Northeastern Naturalist Vol. 24, No. 2
J.A. Rogers and C.D. Collins
2017
135
Liu, Z.G., and X.M. Zou. 2002. Exotic earthworms accelerate plant-litter decomposition in
a Puerto Rican pasture and a wet forest. Ecological Applications 12:1406–1417.
Lobe, J.W., M.A. Callaham Jr., P.F. Hendrix, and J.L. Hanula. 2013. Removal of an invasive
shrub (Chinese Privet: Ligustrum sinense Lour) reduces exotic earthworm abundance
and promotes recovery of native North American earthworms. Applied Soil Ecology
83:133–139.
MaineDOT. 2015. Work plan calendar years 2015–2016–2017. Available online at: http://
www.maine.gov/mdot/projects/workplan/docs/2015/WorkPlan2015-2016-2017.pdf.
Accessed 24 April 2016.
Marinissen, J.C.Y, and F. van den Bosch. 1992. Colonizaiton of new habitats by earthworms.
Oecologia 91:371–376.
McLean M.A., and D. Parkinson. 1998. Impacts of the epigeic earthworm Dendrobaena
octaedra on orbatid mite community diversity and microarthropod abundances in a pine
forest floor: A mesocosm study. Applied Soil Ecology 7:125–136.
Milcu, A., J. Schumacher, and S. Scheu. 2006. Earthworms (Lumbricus terrestris) affect
plant seedling recruitment and microhabitat heterogeneity. Functional Ecology
20:261–268.
Moore, E.H., and J.W. Witham. 1996. From forest to farm and back again: Land-use history
as a dimension of ecological research in coastal Maine. Forest History 1:50–69
Nuzzo, V., A. Dávalos, and B. Blossey. 2015. Invasive earthworms shape forest seed bank
composition. Diversity and Distributions 21:560–570.
Owen, R.B., Jr., and W.J. Galbraith. 1989. Earthworm biomass in relation to forest types,
soil, and land use: Implications for woodcock management. Wildlife Society Bulletin
17:130–136.
Palm, J., N.L.M.B. van Schaik, and B. Schröder. 2013. Modeling distribution patterns of
anecic, epigeic, and endogeic earthworms at catchment-scale in agro-ecosystems. Pedobiologia
56:23–31
Quackenbush, P.M., R.A. Butler, N.C. Emery, M.A. Jenkins, E.J. Kladivko, and K.D. Gibson.
2012. Lumbricus terrestris prefers to consume Garlic Mustard (Alliaria petiolata)
seeds. Invasive Plant Science and Management 5:148–154.
R Core Team. 2015. R: A language and environment for statistical computing. R Foundation
for Statistical Computing, Vienna, Austria. Available online at http://www.Rproject.
org/.
Resner, K., K. Yoo, S.D. Sebestyen, A. Aufdenkampe, C. Hale, A. Lyttle, and A. Blum.
2015. Invasive earthworms deplete key soil inorganic nutrients (Ca, Mg, K, and P) in a
northern hardwood forest. Ecosystems 18:89–102.
Reynolds, J.W. 2008. The earthworms (Oligochaeta: Acanthodrilidae, Lumbricidae and
Megascolecidae) of Maine, USA. Megadrilogica 12:61–72.
Roth, A.M., T.J.S. Whitfeld, A.G. Lodge, N. Eisenhauer, L.E. Frelich, and P.B. Reich. 2015.
Invasive earthworms interact with abiotic conditions to influence the invasion of Common
Buckthorn (Rhamnus cathartica). Oecologia 178:219–230.
Sackett, T.E., S.M. Smith, and N. Basiliko. 2012. Exotic earthworm distribution in a mixeduse
northern temperate forest region: Influence of disturbance type, development age,
and soils. Canadian Journal of Forest Research 42:375–381.
Scheu, S., and D. Parkinson. 1994. Effects of earthworms on nutrient dynamics, carbon
turnover, and microorganisms in soils from cool temperate forests of the Canadian
Rocky Mountains: Laboratory studies. Applied Soil Ecology 1:113–125.
Northeastern Naturalist
136
J.A. Rogers and C.D. Collins
2017 Vol. 24, No. 2
Shartell, L.M., R.G. Corace, A.J. Storer, and D.M. Kashian. 2015. Broad and local-scale
patterns of exotic earthworm functional groups in forests of national wildlife refuges of
the Upper Midwest, USA. Biological Invasions 17:3591–3607.
Simmons, W., A. Dávalos, and B. Blossey. 2015. Forest successional history and earthworm
legacy affect earthworm survival and performance. Pedobiologia 58:153–164.
Singh, S., J. Singh, and A. Pal Vig. 2015. Earthworm as ecological engineers to change the
physic-chemical properties of soil: Soil vs vermicast. Ecological Engineering 90:1–5.
Suarez, E.R., T.J Fahey, P.M. Groffman, P.J. Bohlen, and M.C. Fisk. 2004. Effects of exotic
earthworms on soil phosphorous cycling in two broadleaf temperate forests. Ecosystems
7:28–44.
Suárez, E.R., G.L. Tierney, T.J. Fahey, and R. Fahey. 2006. Exploring patterns of exotic
earthworm distribution in a temperate hardwood forest in south-central New York, USA.
Landscape Ecology 21:297–306.
United States Climate Data. 2016. Climate data for Augusta, ME. Available online at:
http://www.usclimatedata.com/climate/augusta/maine/united-states/usme0013/2016/1.
Accessed 31 December 2016.
United States Department of Agriculture (USDA). 2013. Web soil survey. Available online at
http://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm. Accessed 2 September 2015.
Wandeler, H.D., R. Sousa-Silva, E. Ampoorter, H. Bruelheide, M. Carnol, S.M. Dawud, G.
Danila, L. Finer, S. Hattenschwiler, M. Hermy, B. Jaroszewicz, F.X. Joly, S. Muller, M.
Pollastrini, S. Ratcliffe, K. Raulund-Rasmussen, F. Selvi, F. Valladares, K. Van Meerbeek,
K. Verheyen, L. Vesterdal, and B. Muys. 2016. Drivers of earthworm incidence
and abundance across European forests. Soil Biology and Biochemistry 99:167–178.
Whitfeld, T.J.S., A.M. Roth, A.G. Lodge, N. Eisenhauer, L.E. Frelich, and P.B. Reich. 2014.
Resident plant diversity and introduced earthworms have contrasting effects on the success
of invasive plants. Biological Invasions 16:2181–2193.
Wironen, M., and T.R. Moore. 2006. Exotic earthworm invasion increases soil carbon and
nitrogen in an old-growth forest in southern Quebec. Canadian Journal of Forest Research
36:845–854.