THE IMPORTANCE OF INVASIVE EARTHWORMS AS SEED PREDATORS … · earthworm, Lumbricus terrestris. Results from an earthworm-addition microcosm experiment suggest nearly 70% of seeds
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
THE IMPORTANCE OF INVASIVE EARTHWORMS AS SEED PREDATORS OF COMMON FOREST FLORA OF ONTARIO
By
Colin Cassin
A thesis submitted in conformity with the requirements for the degree of Master of Science
Graduate Department of Ecology and Evolutionary Biology University of Toronto
Effectively all (99.9%) of the 1093 seeds that were recovered from the microcosm
experiment were discovered using visual inspection aided by ultraviolet light. Although
germination trials may tend to be biased by nature, only a single seed (<0.1%) was recovered
by germination from soil incubated in the growth chamber. Consequently, the following
discussion includes only visually detected seeds.
The recovery rate for all species combined recovered <1cm from the soil surface was
27.2%±4.1% (all values are mean ± SEM) (Table 3). Nearly all of the seeds missing from the
top cm of soil were likely removed by earthworms: recovery from the top cm of control tubes
without worm addition was nearly 100% (97.8%±0.7%) for all species combined (Table 3;
Figure 1).
Results were similar for seeds recovered from the top 5cm of soil, though recovery
rates were increased, reflecting burial deeper than 1 cm (Figure 1). When all species are
combined, only 45.7%±4.3% (Table 3) of seeds remained within 5cm of the soil surface
when exposed to earthworms, while 98.0%±0.7% (Table 3) of seeds in control tubes were
recovered from the upper 5cm.
Comparatively a relatively small number of seeds were recovered at depths below
5cm than at shallower depth profiles. In the presence of earthworms only 23.8%±3.7% of
seeds were recovered buried >5cm deep while 0% were recovered in control tubes at the
same depth (Table 3).
!
!
!
33!
Finally, in the presence of earthworms, a total of 30.5%±5.0% of seeds were removed
outright and assumed digested (Table 3). Control tubes, which did not contain earthworms,
yielded a negligible loss of seeds (2%±0.7%) likely attributable to experimental error and
cracks in the soil surface.
Seed Species Preferences
All species of seed used in this experiment experienced surficial removal at some
rate, however considerable differences were apparent between species especially when
considering the depth at which the seeds were recovered (Table 3, Figure 1). For instance,
ANOVA and post-hoc results indicate significant differences among species in rates of
removal from the soil surface (<1cm: F5,48 = 9.9009, p = <0.001) (Tables 2, 3). Data poorly
met ANOVA assumptions; however Kruskal-Wallis tests produced essentially identical
results overall [X25 =24.9968, p = 0.0001]. Tukey HSD tests found P. strobus seeds were
recovered in significantly (p = < 0.0001) greater abundance (73.3%±5.7%) on the soil
surface than all other species used in this study, providing evidence for preferential feeding
habits by earthworms (Table 3). Each of the remaining species were recovered in much
lower rates; for example, <10% of A. petiolata seed was recovered within the upper 1cm;
however, there were no significant differences among other species (Table 3, Figure 1).
Recovery rates for seeds buried <5cm again differed significantly (p < 0.0001) among
species (Tables 2, 3). Kruskal-Wallis tests produced essentially identical results overall [X!!
=29.2198, p = <0.0001]. Again, white pine appears to suffer the least damage compared to
the other 5 species, as 85% of seeds were recovered within the upper 5cm of the soil profile
(Table 3, Figure 1). More generally, >50% of the two largest-seeded species were recovered
within the upper 5cm of the soil profile, compared to <25% of the while the 2 smallest-
seeded species. Tukey-Kramer results showed that again P. strobus was recovered in
significantly greater abundance than nearly all other species (p < 0.0001), again providing
evidence for preferential feeding (Table 3). Tukey-Kramer results provide further evidence
for preferential feeding with respect to A. petiolata, whose seed was removed in significantly
(p < 0.005) greater abundance than nearly all other species (Table 3, Figure 1).
Results for seeds recovered at depths >5cm again indicated significant (p < 0.0001)
differences exist among species (Tables 2, 3). Kruskal-Wallis tests produced essentially
!
!
!
34!
identical results overall [X25 =31.0968, p = <0.0001]. Seeds of the 3 smallest species were
recovered in very low abundance, compared to the 3 largest species (Figure 1).
Finally, results indicate significant (p < 0.0001) differences among species regarding
number of seeds digested and ultimately removed by earthworms (Tables 2, 3). Kruskal-
Wallis tests produced essentially identical results overall [X25 =42.6842, p = <0.0001]. Once
again Tukey- Kramer analyses indicate the two smallest seeded species were removed at
significantly (p < 0.0001) greater rates than all other species of seed (Tables 2, 3). More
specifically Alliaria and Betula were removed at 89% and 68% respectively, while the four
larger species were removed at rates lower than 15% (Table 3). This suggests the two
smallest species of seeds are largely responsible for increasing the mean rate of seed removal
across all species to 30.5% (Table 3).
! 35!
Table 2 Results of ANOVA testing the proportion of seeds recovered at notable ecological
fates after being exposed to earthworms (n=9). Significant results in bold. Depth <1cm <5cm >5cm Removed DF F Ratio Prob>F DF F Ratio Prob>F DF F Ratio Prob>F DF F Ratio Prob>F Species 5 9.9009 <0.0001 5 11.0060 <0.0001 5 12.6412 <0.0001 5 69.5863 <0.0001 Error 48 48 48 48 Total 53 53 53 53
! 36!
Table 3 Results of Lumbricus terrestris seed predation microcosm experiment for each species: mean proportion of seeds
recovered ±SEM (n=9 replicates per species for worm addition treatment, n=6) for controls) for 6 species at various ecologically
meaningful depths. Significant results are shown in bold. Species are listed in ascending order of mean seed mass. Also shown are
results of separate Tukey-Kramer tests for each worm treatment at each depth (species at that depth sharing the same letter do not
differ significantly: p > 0.05).
Treatment Species Surface: Buried <1cm
Tukey-Kramer
Germinable: Buried <5cm
Tukey-Kramer
Buried at Depth: Buried >5cm
Tukey-Kramer
Removed: Seed Not
Recovered
Tukey-Kramer
Worm Betula alleghaniensis 0.185±0.052 B 0.252±0.038 BC 0.067±0.031 B 0.681±0.051 B Addition Alliaria petiolata 0.089±0.038 B 0.111±0.046 C 0 ±0 B 0.889±0.046 A
Pinus strobus 0.733±0.057 A 0.852±0.047 A 0.074±0.034 B 0.074±0.026 CD Berberis thunbergii 0.296±0.117 B 0.467±0.102 B 0.4±0.086 A 0.133±0.037 C Prunus serotina 0.230±0.057 B 0.504±0.083 B 0.496±0.083 A 0±0 D Acer saccharum 0.096±0.073 B 0.556±0.103 AB 0.393±0.101 A 0.052±0.024 CD
Combined Sp. Mean 0.272±0.041 0.457±0.043 0.238±0.037 0.305±0.050 Worm-free Betula alleghaniensis 0.956±0.022 A 0.956±0.022 A 0±0 A 0.044±0.022 A
Controls Alliaria petiolata 0.944±0.027 A 0.956±0.022 A 0±0 A 0.044±0.022 A Pinus strobus 1.000±0.000 A 1.000±0.000 A 0±0 A 0±0 A Berberis thunbergii 0.978±0.022 A 0.978±0.022 A 0±0 A 0.022±0.022 A Prunus serotina 1.000±0.000 A 1.000±0.000 A 0±0 A 0±0 A Acer saccharum 0.989±0.011 A 0.989±0.011 A 0±0 A 0.011±0.011 A
Combined Sp. Mean 0.978±0.008 0.98±0.007 0±0 0.02±0.007
! 37!
Figure 1 Mean proportion of seeds recovered ± SEM, at various depth profiles within
microcosms exposed to (worm addition) (n=9) and protected from (controls) (n=6) the
granivorous earthworm Lumbricus terrestris. In part A, species at each depth sharing the
same letter do not significantly differ; in part B, no significant differences among species
were detected Seed species are ordered by increasing seed size.
0.0
0.1
0.20.3
0.40.5
0.6
0.70.8
0.91.0
Prop
ortio
n R
ecov
ered
B
BB
Betu
la a
llegh
anie
nsis
Allia
ria p
etio
lata
Pinu
s st
robu
s
Berb
eris
thun
berg
ii
Prun
us s
erot
ina
Acer
sac
char
um
0.00.1
0.2
0.30.4
0.50.6
0.7
0.80.9
1.0
Prop
ortio
n R
ecov
ered
A) Worm addition
B) Controls
<1cm
<5cm
>5cm
Not recovered
B
B
BC
B C
B
A
A
A
BCD
B
B
A
C
B
B A
D
B
AB
A
CD
!
!
!
38!
Granivore Guild Experiment Results
Summer Trial
For the summer trial of the granivory guild experiment, exclusion of belowground
seed predators was compared to control plots in which seed was available to both
belowground and aboveground seed predators. At the 2-week sampling all fixed factors were
significant (p<0.05) in the full ANOVA model (Table 4). More seeds were recovered from
plots that excluded belowground granivores than from control plots. The model displays a
significant interaction between treatment and species, indicating that the effect of treatment
differs between species (Table 4). When analyzed separately differences between treatments
at the species level became apparent (Table 5). Individual F-tests indicate that after 2 weeks
of exposure to predation, 9 of the 21 species had significantly (p<0.005) greater recovery
when belowground seed predators were excluded (Table 5). Wilcoxon Rank-Sum tests also
identified each of these species as significantly different between treatments; in addition,
results for Pinus banksiana were marginally significant (p=0.049). In all cases, recovery was
higher in below-ground exclosure treatments than controls (Table 5).
After 4 weeks of exposure to granivores both exclusion treatment (p=0.0007) and
species (p<0.0001) were found to have significant effects, while the interaction between was
not significant (Table 4). As expected, many species showed recovery rates that are lower
than those found after just 2-weeks exposure to granivores; however, only one species
(Betula allegheniensis) differed among treatments according to parametric statistical tests,
likely because higher background rates of removal by vertebrate seed predators obscured any
patterns (Figure 2). In addition to this example, Wilcoxon Rank-Sum tests identified
differences among treatments for Larix laricina (P = 0.007), Maianthemum racemosum (P =
0.020), Berberis thungberii (P = 0.044), Betula pendula (P = 0.048). In all cases, recovery
was higher in below-ground exclosure treatments than controls (Table 5, Figure 2).
! 39!
Table 4 Results of comprehensive randomized block factorial ANOVAs from summer
granivore exclosure experiment (n=12). Site used as a blocking factor; "Treatment" =
Control, Above-ground exclosure, or Below-ground exclosure; "Species" = species tested
(Table 1). Significant results (p < 0.05) are indicated in bold.
Sampling Week 2 Week 4 DF F P DF F P Treatment 1,451 38.37 <0.0001 1,451 11.54 0.0007 Species 20,451 47.89 <0.0001 20,451 45.55 <0.0001 Treatment x Species 20,451 2.12 0.0035 20,451 0.87 0.6244
!
!
!
40!
Table 5 Results of randomized block ANOVAs for recovery of seeds of each species.
Significant results (p<0.05) are indicated in bold. Also shown are results of F-tests
comparing controls (C) and below-ground exclosures (treatment "B"); "-" means no
significant result was detected (p>0.05).
2 Weeks 4 Weeks
Origin Species F1,11 P Treatments F1,11 P Treatments Native
Table 6 Results of comprehensive randomized block factorial ANOVAs with site as a
blocking factor; "Treatment" = Control, Above-ground exclosure, or Below-ground
exclosure; "Species" = species tested (Table 1) (n=12). Significant results (p<0.05) are
indicated in bold.
Sampling Week 2 Week 4 DF F P DF F P Treatment 2,715 422.88 <0.0001 2,695.6 289.22 <0.0001 Species 21,715 15.88 <0.0001 21,693.1 16.75 <0.0001 Treatment x Species 42,715 19.40 <0.0001 42,693.1 14.22 <0.0001
! 45!
Figure 3 Mean proportion of seeds recovered ± SEM, after 2 and 4 weeks exposure to
various granivore treatments in a fall exclosure experiment (n=12). Treatments: "Above" =
above-ground exclosure, "Below" = below-ground exclosure, "Control" = no protection. For
both natives and exotics, species are ordered by increasing seed size.
Natives Exotics
Betu
la p
apyr
ifera
Betu
la a
llegh
anie
nsis
Larix
laric
ina
Tsug
a ca
nade
nsis
Loni
cera
can
aden
sis
Pinu
s ba
nksi
ana
Pinu
s st
robu
s
Acer
rubr
um
Mai
anth
emum
race
mos
um
Frax
inus
am
eric
ana
Prun
us s
erot
ina
Acer
sac
char
um
Tilia
am
eric
ana
Car
ya c
ordi
form
is
Car
ya o
vata
Que
rcus
rubr
a
Allia
ria p
etio
lata
Pinu
s sy
lves
tris
Rob
inia
pse
udoa
caci
a
Berb
eris
thun
berg
ii
Rha
mnu
s ca
thar
tica
Acer
pla
tano
ides
0.0
0.2
0.4
0.6
0.8
1.0
A) Week 2
B) Week 4
0.0
0.2
0.4
0.6
0.8
1.0Proportion
Proportion
Above
Below
Control
! 46!
Table 7 Results of randomized block ANOVAs for recovery of seeds of each species.
Significant results (p<0.05) are indicated in bold. Also shown are results of Tukey HSD tests
comparing controls (C), above-ground exclosures (treatment "A"), and below-ground
exclosures (treatment "B"); "-" means no significant result was detected (p>0.05). Results
failed to converge for Betula allegheniensis; however, means were very similar across all
three treatments.
2 Weeks 4 Weeks
Origin Species F2,22 P Treatments F2,21 P Treatments Native
Analysis of variance confirmed that neither seed losses nor the number of fragmented
seeds was affected by the marking treatment (p > 0.15) (Table 8). The fraction of seeds
removed by seed predators was very high (Figure 4), resulting in the inability to homogenize
variances. However, it was clear that UV marking did not present a deterrent to seed
predators.
Again significant differences in removal rates between all species were apparent (p <
0.05). For instance, viable seeds of the relatively small-seeded Lonicera canadensis were
recovered more than the other, larger-seeded species (Quercus rubra, Prunus serotina, Acer
saccharum, Carya cordiformis) (Figure 4).
! 48!
Table 8 Results from randomized block factorial ANOVA for UV marking on seed removal
(n=6). For the analysis site was used as a blocking factor; "Treatment" = Marked or
Unmarked; "Species" = species tested (Table 1). Significant results (p<0.05) are indicated in
bold.
Viable Seeds Remaining DF F P Treatment 1,45 1.16 0.287 Species 4,45 3.03 0.027 Treatment x Species 4,45 0.56 0.693
! 49!
Figure 4 Proportion of viable seeds remaining ± SEM, for marked and unmarked seeds, after
7 days exposure to granivores (n=6). Treatments: "marked" = UV-marked, "Unmarked" =
unmarked. Species are ordered by increasing seed size.
Loni
cera
can
aden
sis
Pinu
s st
robu
s
Prun
us s
erot
ina
Acer
sac
char
um
Car
ya c
ordi
form
is
0.0
0.2
0.4
0.6
0.8
1.0MarkedUnmarked
Proportion
!
!
!
50!
Leaf Litter & Granivory Experiment Results
Analysis of variance indicated that neither seed losses nor the number of fragmented
seeds was affected by the litter treatment; as well, there were no significant interactions
(Table 9). In contrast, there were differences in removal rate between species (Figure 5). As
found in the summer and fall granivore exclosure experiments (Figures 2, 3), removal rates
were lowest in smaller-seeded species (Betula alleghaniensis, Alliaria petiolata, Lonicera
canadensis) compared to larger-seeded species (Quercus rubra, Prunus serotina, Acer
saccharum) (Figure 5).
Germination From Burial Depth Experiment Results
Of the 6 species used in the germination from depth experiment, only Pinus strobus,
Berberis thunbergii and Betula alleghaniensis were successfully germinated at any depth
(Table 10). The inability to successfully germinate Acer saccharum, Prunus serotina, and
Alliaria petiolata can likely be attributed to a failure to properly address specific
pregermination requirements (e.g. scarification, cold dormancies, etc.) for each species.
In each of the 3 species that were successfully germinated (Pinus, Berberis, and
Betula), germination rates were greatest for seeds buried <1cm from the soil surface. As
burial depth increased, rates of successful germination decreased for all species (Table 10).
Both Berberis and Betula were unable to successfully germinate from depths greater than
3cm. Pinus was able to successfully germinate as deep as 7cm, however germination rates
from depths >5cm were minor. Only 2.8% of seeds buried at 7cm germinated compared with
80.5% at the soil surface (Table 10).
! 51!
Table 9 Effects of litter on seed removal (n=10). Results of randomized block factorial
ANOVAs with site as a blocking factor; “Treatment”=Litter present or Litter Removed,
“Species = species tested (Table 1). Significant results (p<0.05) are indicated in bold.
Viable Seeds Remaining Fragments DF F P DF F P Treatment 1,117 0.00 0.985 1,117 1.60 0.208 Species 6,117 152.60 <0.0001 6,117 156.68 <0.0001 Treatment x Species 6,117 0.81 0.561 6,117 1.70 0.127
! 52!
Figure 5 Mean proportions of viable seeds remaining ± SEM, in plots with and without leaf
litter layers (n=10). Treatments: "Litter" = litter present, "No Litter" = litter removed. Species
are ordered by increasing seed size.
Betu
la a
llegh
anie
nsis
Allia
ria p
etio
lata
Loni
cera
can
aden
sis
Pinu
s st
robu
s
Prun
us s
erot
ina
Acer
sac
char
um
Que
rcus
rubr
a
0.00.10.20.30.40.50.60.7
0.80.91.0
LitterNo Litter
Prop
ortio
n re
mai
ning
! 53!
Table 10 Proportion of seeds successful in germinating from various burial depths in a
greenhouse control study (n=3). Species are listed in order of decreasing mean seed mass
Potential Of Earthworms As Granivores Earthworms are potentially significant agents of mortality for seeds of forest species.
When all plant species used in the microcosm experiments are combined, 72.8% of seeds
placed on the soil surface were removed when exposed to earthworms. Separately, each of
the 6 species used in this experiment suffered considerable increases in removal when
exposed to earthworms relative to their respective worm-free controls. These losses occurred
over the relatively short period of only two weeks, suggesting that earthworms could be
removing a remarkably large proportion of seeds over longer periods of time. These rates of
seed loss are comparable with losses observed in seed predation studies (Eisenhauer and
Scheu 2008; Quackenbush et al. 2012).
Seed removal from the soil surface is not necessarily fatal. Seeds remaining at or near
the soil surface almost certainly have the potential to successfully germinate, although they
may be exposed to higher rates of predation by rodents and birds compared to buried seeds
(Chambers and MacMahon 1994). Results for seeds transported to greater depths are more
complicated. First, some seeds were fragmented or vanished entirely (>30%). As
experimental controls indicate seed detection methods were highly effective (<2% of seeds
escaped detection), it is assumed that vanished seeds suffered a fatal outcome. It is probable
vanished seeds were ingested, pulverized within the digestive tract, and ultimately destroyed
by earthworms, as seen in other studies (Quackenbush et al. 2012). In other cases it is
possible seeds were ingested, transported, and excreted by earthworms. Interpretation on how
ingestion and excretion impacts subsequent germination rates is unclear, being shown to both
positively and negatively impact germinability on different seeds (Eisenhauer et al. 2009).
Transportation and burial at depth perhaps has a clearer impact on the likelihood of
germination, as many species are unable to germinate from depths at which seeds are often
deposited by earthworms.
Of the remaining buried seeds, the chance of survival likely depends strongly on
burial depth. Few species likely emerge from depths >5cm, as indicated by the growth
chamber experiment; most of these likely are ecologically dead, at least unless they are
further transported to the soil surface once again. Many smaller seeds, such as those of
yellow birch, are only capable of germinating at the soil surface due to specific requirements
!
!
!
55!
such as light (Burns and Honkala 1990). In other cases, burying a seed more than 1cm deep
in soil can lead to exponential declines in ability to successfully germinate since they likely
contain insufficient resources to enable seedlings to reach the soil surface (Burns and
Honkala 1990). This again may be particularly problematic for small-seeded species such as
yellow birch, whose ability to germinate successfully even when only buried by leaf litter is
negligible; more than 80% of yellow birch seeds in this experiment appear to be unavailable
for germination when in the presence of earthworms. Garlic mustard is another small seeded
species that appears to suffer a similar fate, as 91.1%±3.8% of seeds were removed from the
soil surface, and thus largely unavailable for germination.
In contrast, some larger-seeded species, notably white pine, may be able to rarely
emerge from greater depths: in the greenhouse experiment, < 3% of seed emerged from
>5cm burial depth. Burial greater than 5cm is likely fatal for the remaining species used in
this study as zero individuals were able to successfully germinate from such depths. Thus for
the purpose of this study, it is reasonable to assume that most seed buried more than 5cm
below the soil surface is effectively removed and unavailable for germination. Though the
possibility some seeds may ultimately emerge clouds the interpretation of my results, similar
issues apply to seed-caching rodents such as squirrels (Hadj-Chikh et al. 1996).
In the microcosm experiment it was clear that earthworms preferentially removed
some species over others. On one hand, white pine appears to be considerably less prone to
removal than other species: as 73.3% of seeds remained on the soil surface. This lack of
removal suggests this species of seed may be relatively immune to earthworm seed predation
compared to other species of seed. Previous studies have linked decreases in seed palatability
to earthworms to factors such as seed size, morphology, and plant functional group
(Eisenhauer and Scheu 2008, Quackenbush et al. 2012). Coupled with the ability of seeds
that do become buried to germinate from deeper depths, these results suggest white pine may
be relatively resistant to earthworm granivory.
Rates of removal were much higher for the remaining species, even the relatively
large seeded P. serotina, B. thunbergii, and A. saccharum. Nearly 50% of seed from each of
these species was either buried below a depth that allows successful germination, or assumed
digested. The relatively large size of these seeds, particularly P. serotina, and A. saccharum,
may suggest that earthworms are capable of directly or indirectly removing at least some
!
!
!
56!
quantity of the majority of species commonly found in temperate hardwood forests. Results
are even more striking for the small-seeded A. petiolata and B. alleghaniensis, which suffer
tremendously high rates (88.9% and 74.8% respectively) of effective seed removal from
germinable depths. This reduction in the amount of available seed for smaller seeded species
could perhaps influence the number of mature individuals in future plant communities.
It is notable that the problematic invader A. petiolata was removed in significantly
greater abundance than all other species. This agrees with the results of Quackenbuch et al.
(2012), who also found that Alliaria was highly palatable to introduced earthworms. This
may suggest that, rather than producing invasional meltdown (Simberloff and Von Holle
1999), earthworms may actually decease the success of this invasive plant. However, this is
not true for all invaders: B. thunbergii experienced much lower rates of removal, while the
native B. alleghaniensis displays similarly large rates of predation.
Relative Role Of Earthworms & Other Granivores
Twenty-one species of seed were used to assess the relative role of belowground
predators in the summer granivore exclusion experiment. Samples collected after 2 weeks of
exposure to granivores show that excluding belowground seed predators yielded a significant
increase in the fraction of seeds recovered. Therefore, in the absence of belowground
granivores, such as L. terrestris, seeds stand a better chance at survival, which indicates that
this suite of granivores may play a more influential role in community seed dynamics than
previously thought. This relationship appears to hold true over a longer period of time as
samples collected after 4 weeks of exposure to granivory continued to show reduced but
significantly greater overall seed recovery when L. terrestris was excluded, though there
were few species-level effects. Although a handful of studies have shown that L. terrestris is
indeed capable of ingesting, egesting, and even digesting some species of seed (Quackenbush
et al. 2012, Eisenhauer and Scheu 2008), evidence of seed removal by earthworms has not
previously been demonstrated in a realistic field experiment.
Belowground granivores such as L. terrestris clearly exerted significant (p < 0.0001)
preferential feeding on some species over others in this experiment. After 2 weeks, nearly
half (9 of 21 species) were removed at significantly different (p < 0.05) rates between
treatments. The ability of L. terrestris to preferentially feed on some species over others has
!
!
!
57!
also been demonstrated before using a series of simple choice experiments comparing garlic
mustard to 6 other herbaceous species (Quackenbush et al. 2012). In this study L. terrestris
ingested A. petiolata at significantly (p < 0.05) greater rate than 3 of 6 other species tested.
How these preferences can influence and manipulate soil seed banks has yet to be directly
evaluated.
In the fall trial of this experiment the addition of an aboveground exclusion allowed
effects of worms to be compared with other seed predators. Using a suite of 21 different
species of seed it was evident that generally aboveground granivores remove a considerably
greater quantity of seed than belowground seed predators, although after a greater length of
exposure to seed predators the discrepancies between granivore guilds became less
pronounced. When exposed exclusively to belowground seed predators 35.0%±2.02% of
seeds were removed after an exposure period of 2 weeks. However plots exposed to
aboveground seed predators and control units (exposed to above & belowground predators)
exhibited nearly identical levels of seed removal (75.0%±1.79% and 76.6%±1.76%
respectively). These results point to the importance of aboveground predators such as
numerous species of rodents as the prominent granivores in the system, likely attributable to
the speed and efficiency at which they remove recently dispersed seed.
After 4-weeks of exposure to seed predators, it appears that belowground granivores
played an increasingly important role, removing 47.0%±2.29% of seeds when they are the
exclusive seed predators. As expected, aboveground predators and control units also saw
increases in the amount of seeds removed, with removal rates of 79.5%±2.29% and
84.2%±1.65% respectively, again pointing to their importance at removing large quantities of
seed.
To some extent, the lack of strong earthworm effect at the fall sampling may be a
result of very strong rodent effect statistically masking weaker worm predation. This
interpretation is supported by the fact that deleting the above-ground exclosure treatment
from the analyses added several significant results. Despite this, it is clear that seed predation
by rodents was far more frequent, stronger, and easier to detect than these relatively subtle
earthworm effects.
Collectively these findings suggest aboveground granivores are especially effective at
removing seeds that have been recently dispersed, especially those seeds that are medium to
!
!
!
58!
large in size. It appears that belowground granivores, likely mainly consisting of earthworms
in this system, may play a more important role in the removal of seeds of smaller mass. This
is supported by the low recovery rates of smaller species such as Betula alleghensis and
Lonicera canadensis in treatments exposed to earthworms compared with other treatments.
Earthworms likely play a measurable but relatively small role in seed predation on a
community scale in this forest system. Generally speaking, temperate forests tend to host
species with reasonably large seeds compared to other habitats, many of which are quickly
targeted by aboveground granivores such as rodents and birds. This combined with the fact
that earthworm populations tend to exhibit decreased activity towards the end of the summer
months (Edwards and Bohlen 1996), when seed dispersal is beginning but when small
mammal populations tend to peak (Falls et al. 2007), suggests that earthworms are likely
responsible for relatively little seed removal from a community perspective. These life
history traits may be at least partially responsible for an apparent decrease in earthworm seed
removal in the fall trial compared with the summer trial. The relatively small role that
earthworms appear to play in forest seed predation may not hold true in other systems,
especially those with a different range in seed size and different patterns of dispersal.
Although both aboveground and belowground seed predator guilds clearly remove
seeds at different rates, they are similar in that they clearly attack different species of seed at
different rates. Stark differences between different species of seed are evident after both
shorter (2 week) and longer (4 week) exposure to seed predators. A number of trends begin
to emerge when similar species are grouped together into different functional groups. For
instance, when looking at congeneric sets of species within the same granivore treatment,
highly comparable seed removal rates are evident. The 3 Pinus species (P. strobus, sylvestris
and banksiana) used in this study were subject to comparatively similar rates of recovery in
aboveground excluded (37.5%±10.93%, 30.6%±8.72%, and 13.9%±5.16%), belowground
excluded (9.1%±4.69%, 12.1%±4.26% and 6.1%±2.54%), and control treatments
(3.5%±1.91%, 6.9%±2.68% and 1.4%±0.94%). The relatively similar rates of granivory
between members of a genus was also observed in the Acer, Betula and Carya genera.
Of course other likely important factors confound with genera, such as the size and
nutritional value of a seed. A few interesting trends arise when looking at the role of seed
size on granivory. For instance, when exposed to aboveground granivores, larger sized seeds
!
!
!
59!
such as those of the genera Quercus, Carya, and Acer tend to suffer remarkably high rates of
seed removal. In fact each of the largest 8 species of seed used in this experiment were
subject to >90% removal when exposed to aboveground granivores. Conversely, smaller and
less resource-rich seeds tended to escape high rates of removal from aboveground predators.
In this experiment only 2 of the smallest 7 species of seeds experienced predation rates
greater than 30%.
Interestingly, earthworms appear to be driving similar seed selection trends to those
observed in the earthworm microcosm experiment. In the aboveground exclusion treatments
that were subject only to belowground seed predators such as earthworms, many of the larger
seeds were recovered at remarkably high proportions. This may suggest that belowground
predators have a relatively minor impact on larger seeded genera such as Acer, Quercus, and
Carya. Conversely, these belowground seed predators again appear to be driving the removal
of smaller seeded genera such as Alliaria, Larix and Pinus, where earthworms were
attributed with the removal of >60% of seeds.
Potential Biases
All seeds that were used in each of these experiments detailed were marked using an
ultraviolet fluorescent dye to aide in seed recovery. ANOVA results indicate the number of
recovered seeds and seed fragments recovered are not significantly different between marked
and unmarked treatments. Also, as expected, differences in removal rates between species
were significant (p < 0.05). These findings suggest that the methods utilized in seed marking
in this experiment have a non-significant impact on the interpretation and application of
these findings to naturally occurring, unmarked seed.
The leaf-litter removal experiment provided remarkably similar results between
control and litter-removed treatments. These similarities were unexpected and quite
surprising to discover. Intuitively it would appear logical for the removal of the leaf litter
layer to expose seeds to increased rates of granivory, however these results were unfounded
in this study. An increased sample size, range of seeds (including seeds subject to lower
removal rates than those used) and duration may have increased the success of this study.
The germination from depth experiment provided useful insight into the interpretation
of the earthworm microcosm results. As expected, burial at increasing depth generally leads
!
!
!
60!
to reductions in germinability. Although differences between species of seeds were quite
strong, such as B. papyrifera's inability to germinate when buried below 1cm depth and P.
strobus’ ability to in rare cases germinate buried below 5cm, the species of seeds used are
largely unable to germinate successfully when buried at depths of 5cm and greater. This
suggests that it is reasonable to assume that seeds buried below 5cm of soil are effectively
unavailable for germination, and for the purpose of interpretation for this study, effectively
removed from the soil seed bank.
2.5 CONCLUSIONS
In the presence of earthworms seeds appear to be restricted to 1 of 4 possible fates.
They may (1) remain at the soil surface, (2) be buried at a shallow depth that does not
eliminate the possibility for successful germination, (3) be buried at a depth that restricts
successful germination, or (4) removed outright and likely digested.
The results of this experiment clearly indicate that when in the presence of
earthworms seeds potentially can be removed from the soil surface in substantial quantities.
All but one species used in my microcosm experiment saw more than 70% of their seeds
removed from the soil surface when in the presence of earthworms. Despite these
tremendous rates, many seeds (>50% from 4 of 6 species) were found to be buried within the
upper 5cm of the soil profile, suggesting that the overall reduction in fitness may not be as
pronounced as originally believed. Still 4 of the 6 species of seed used in this experiment
maintained less than 60% of their original quantity of seed within a germinable depth of soil.
I also found evidence that earthworms are capable of removing different species of
seed at significantly different rates. This supports the conclusion that earthworms are capable
of exhibiting preferential selection as granivores (Quackenbush et al. 2012; Eisenhauer et al.
2010). In contrast, I found that in a more realistic setting that earthworms are eclipsed as
granivores by more efficient seed predators, namely small rodents (Garett and Graber 1994;
Falls et al. 2007; Schnurr et al. 2002). Earthworms were still capable of removing seeds with
certain traits however, and may be viewed as size-restricted specialists.
Some of the interesting trends in earthworm seed preference would benefit from
further study. For instance, how earthworm seed selection manifests itself over time to
influence the structure and composition of a plant community would be a fascinating angle
!
!
!
61!
for research. Specifically a duration of 2-3 growing seasons may help us better understand
longer-term implications of earthworm activity on soil seed bank processes.
Broader Implications The results of this study provide evidence that invasive earthworms may have direct
impacts on invaded plant communities through seed consumption, as well as indirect impacts
via soil modification (Larsson et al. 2010; Bohlen et al. 2004a). These impacts are additional
to depredations by rodents, and similarly may involve both negative (mortality) and positive
(protection) components. These pressures likely differ among species of seed, depending on
their palatability and capacity to emerge from depth, and perhaps suggest that earthworms
may act as ecological filters in their potential ability to restrict success of certain species.
Earthworms are known to have negative effects on many understory plants, though
the mechanisms are not entirely clear. Many explanations have been proposed (Eisenhauer
and Scheu 2008; Griffith et al. 20013), notably impacts on mycorrhizae and complex
consequences of removal of the litter layer (Lawrence et al. 2003; Hopfensperger et al.
2011). Although Darwin was aware of the ability of earthworms to act as seed predators as
long ago as 1881, this aspect of earthworm ecology has been largely overlooked in recent
ecological studies. The results of this study suggest that although earthworms are certainly
capable of removing remarkable quantities of seed under controlled conditions, they often
appear to be eclipsed as seed predators at the community level in temperate forests where
rodents are common and seeds are larger than those in other habitat types. In less rodent-
dominated systems, they may be capable of playing a greater role in seed predation,
especially for small-seeded species.
!
!
!
62!
References Addison JA. 2009. Distribution and impacts of invasive earthworms in Canadian forest
ecosystems. Biological Invasions 11:59-79
Bewley JD, M Black, and P Halmer. 2007. The encyclopedia of seeds: science, technology
and uses. CABI Publishing
Bohlen P, PM Groffman, TJ Fahey, MC Fisk, E Suarez, DM Pelletier, and RT Fahey.
2004a. Ecosystem consequences of exotic earthworm invasion of north temperate
forests. Ecosystems 7:1-12
Bohlen PJ, S Scheu, CM Hale, MA McLean, S Migge, PM Groffman, and D Parkinson.
2004b. Non-native invasive earthworms as agents of change in northern temperate
forests. Frontiers in Ecology and the Environment 2:427-435
Bouché M. 1977. Strategies lombriciennes. Ecological Bulletins 25:122-132
Burns RM, and BH Honkala. 1990. Silvics of North America Volume 2: Hardwoods. United
States Department of Agriculture, Forest Service
Chambers JC, and JA MacMahon. 1994. A day in the life of a seed: movements and fates of
seeds and their implication for natural and managed systems. Annual Review of
Ecological Systems 25:263-292
Clark CJ, JR Poulsen, DJ Levey, and CW Osenberg. 2007. Are plant populations seed
limited? A critique and meta-analysis of seed addition experiments. American
Naturalist 170:128-142
Connell JH. 1971. On the role of natural enemies in preventing competitive exclusion in
some marine animals and in rain forest trees. Dynamics of Populations (eds PJ den
Boer and GR Gradwell) 289-312. Centre for Agricultural Publishing and
Documentation, Wageningen, Netherlands
Darwin C. 1881. The formation of vegetable mould through the action of worms.
Cambridge University Press
Edwards CA and PJ Bohlen. 1996. Biology and ecology of earthworms 3rd ed. Chapman
and Hall, London
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
!
!
!
63!
Eisenhauer N, M Schuy, O Butenschoen, and S Scheu. 2009. Direct and indirect effects of
endogeic earthworms on plant seeds. Pedobiologia 52:151-162
Evers A, A Gordon, P Gray, and W Dunlop. 2012. Implications of a potential range
expansion of invasive earthworms in Ontario’s forested ecosystems: A preliminary
vulnerability analysis. Ontario Ministry of Natural Resources Climate Change
Research Report CCRR-23
Falls JB, EA Falls, and JM Fryxell. 2007. Fluctuations of deer mice in Ontario in relation to
seed crops. Ecological monographs 77:19-32
Fenner M. 1985. Seed ecology. Chapman and Hall, London
Figueroa JA, AA Munoz, JE Mella, and MTK Arroyo. 2002. Pre- and post-dispersal seed
predation in a Mediterranean-type climate montane sclerophyllous forest in central
Chile. Australian Journal of Botany 50:183-195
Garett PW, and RE Graber. 1994. Sugar maple seed production in Northern New
Hampshire. United States Department of Agriculture. Research Paper NE-697
Griffith B, M Turke, WW Weisser, and N Eisenhauer. 2013. Herbivore behaviour in the
anecic earthworm species Lumbricus terrestris L.? European journal of Soil Biology
55:62-65
Hadj-Chikh LZ, MA Steele, and PD Smallwood. 1996. Caching decisions by grey squirrels:
A test of handling time and perishability hypotheses. Animal Behaviour 52:941-948
Holdsworth AR, LE Frelich, and PB Reich. 2007. Regional extent of an ecosystem engineer:
earthworm invasion in northern hardwood forests. Ecological Applications 17:1666-
1677
Honek A, Z Martinkova, and V Jarosik. 2003. Ground beetles (Carabidae) as seed predators.
European Journal of Entomology 100:531-544
Hopfensperger KN, GM Leighton, and TJ Fahey. 2011. Influence of invasive earthworms on
above and belowground vegetation in a northern hardwood forest. The American
Midland Naturalist 166:53-62
Hsia JF, and KE Francl. 2009. Postdispersal sugar maple (Acer saccharum) seed predation
by small mammals in a Northern hardwood forest. American Midland Naturalist
162:213-233
!
!
!
64!
Janzen DH. 1970. Herbivores and the number of tree species in tropical forests. American
Naturalist 104:501-528
Jansen PA, F Bongers, and L Hemerik. 2004. Seed mass and mast seeding enhance dispersal
by a neotropical scatter-hording rodent. Ecological Monographs 74:569-589
Jones CG, JH Lawton and M Shachak. 1994. Organisms as ecosystem engineers. Oikos
69:373-386
Larson ER, KF Kipfmueller, CM Hale, LE Frelich, and PB Reich. 2010. Tree rings detect
earthworm invasions and their effects in northern Hardwood forests. Biological
Invasions 12:1053-1066
Lawrence B, MC Fisk, TJ Fahey, and ER Suarez. 2003. Influence of nonnative earthworms
on mycorrhizal colonization of sugar maple (Acer saccharum). New Phytologist
157:145-153
Lee KE. 1985. Earthworms, their ecology and relationships with soils and land use.
Academic Press, Sydney
Maron JK, DE Pearson, T Potter, and YK Ortega. 2012. Seed size and provenance mediate
the joint effects of disturbance and seed predation on community assembly. Journal
of Ecology 100:1492-1500
McCay TS, DH McCay, and JL Czajka. 2009. Deposition of exotic bird-dispersed seeds into
three habitats of a fragmented landscape in the northeastern United States. Plant
Ecology 203:59-67
Meiners SJ. 2005. Seed and seedling ecology of Acer saccharum and Acer platanoides: a
contrast between native and exotic congeners. Northeastern Naturalist 12:23-32
Mull JF. 2003. Dispersal of sagebrush-steppe seeds by the western harvester ant
(Pogonomyrmex occidentalis). Western North American Naturalist 63:358-362
Myster RW, and STA Pickett. 1993. Effects of litter, distance, density and vegetation patch
type on postdispersal tree seed predation in old fields. Oikos 66:381-388
Nienstaedt H, and JC Zasada. 1990. Picea glauca (Moench) Voss. White spruce. In Silvics
of North America. Volume 1: Conifers. Edited by RM Burns and BH Honkala. U.S.
Department of Agriculture. Agriculture Handbook 654:204–226
!
!
!
65!
Peters S, S Boutin, and E Macdonald. 2003. Pre-dispersal seed predation of white spruce
cones in logged boreal mixedwood forest. Canadian Journal of Forest Research
33:33-40
Pizo MA, and EM Vieira. 2004. Granivorous birds as potentially important post-dispersal
seed predators in a Brazilian forest fragment. Biotropica 36:417-423
Reynolds JW. 1977. The Earthworms (Lumbricidae and Sparganophilidae) of Ontario.
Royal Ontario Museum Toronto, Ontario
Schnurr JL, RS Ostfeld, and CD Canham. 2002. Direct and indirect effects of masting on
rodent populations and tree seed survival. Oikos 96:402-410
Simberloff DJ, and B Von Holle. 1999. Positive interactions of nonindigenous species:
Invasional meltdown? Biological Invasions 1:21-32
Quackenbush P, RA Butler, NC Emery, MA Jenkins, EJ Kladivko and KD Gibson. 2012.
Lumbricus terrestris prefers to consume garlic mustard (Alliaria petiolata) seeds.
Invasive Plant Science and Management 5:148-154
Vander Wall SB. 1995. Dynamics of yellow pine chipmunk (Tamias amoenus) seed caches
underground traffic in bitterbrush seeds. Ecoscience 3:261-266
Vander Wall SB. 2001. The evolutionary ecology of nut dispersal. The botanical review
67:74-117
Vander Wall SB, KM Kuhn, and JR Gworek. 2005. Two-phase seed dispersal: linking the
effects of frugiverous birds and seed-caching rodents. Oecologia 145: 282-287
!
!
!
66!
Chapter 3 General Conclusions
!
3.1 REVISITING & INTERPRETING THESIS OBJECTIVES
The negative impacts inflicted by exotic earthworms on the forests of Ontario have
been well documented in the literature. The extent to which they negatively impact the
structure and function of Ontario ecosystems has been so severe that they are deemed
ecological engineers (Jones et al. 1994, Holdsworth et al. 2007) for their ability to alter the
structure and function of the habitats they invade. Their presence has been observed to alter
numerous foundational aspects of North American habitats including geochemical cycling,
and plant and animal diversity (Sackett et al. 2012; Addison 2009; Bohlen et al. 2004).
Increasing focus has been given to the biotic consequences of these invasive species.
A few studies (Eisenhauer et al. 2010; Quackenbush et al. 2012) have noted the ability of
earthworms to act as seed predators across a range of habitats. Other seed predators such as
mice and squirrels have been observed to act as biotic filters, preferentially consuming
certain species of seed over others. With this in mind those concerned with the impacts that
these invasive species may be inflicting on forests of North America could use a deeper
understanding on the potentially similar role that these invasive earthworms play in the
availability of seeds in forest systems.
This thesis has two core stated goals: to better understand the potential that invasive
earthworms have to act as seed predators, and to measure the actual role that they play in
seed predation, compared to other more well documented granivores. To achieve the first
goal, I used a microcosm experiment to address 3 questions that collectively outline the
potential that invasive earthworms hold to act as important seed predators in the forests of
Ontario.
Question 1. When isolated from other granivores are earthworms capable of removing a
substantial quantity of seed from the soil surface?
!
!
!
67!
Outcome. Six species of seeds were added to the soil surface in microcosms that contained a
single earthworm. The earthworms removed a significant proportion of seeds in five of the
six species. Irrespective of seed species, only 27.8% of seeds placed on the soil surface
remained after 7 days of exposure to an earthworm. When seeds were applied to control
mesocosms lacking earthworms, 97.8% of seeds were recovered from the soil surface over
the same time period.
Question 2. What happens to those seeds removed from the soil surface? What proportion is
removed outright (digested)? Buried to a depth enabling germinable? Buried to a depth
inhibiting germination?
Outcome. Most (45.7%) of the seeds that are removed from the soil surface can be recovered
within 1-5cm from the soil surface. Nearly all species of seed used in this study are capable
of successfully germinating at this depth. Roughly one quarter (23.8%) of seeds were
recovered at depths between 5cm and 30cm. Burial at this depth would effectively render
most of these species of seed ecologically “dead” with very little chance at successfully
germinating. Perhaps most importantly, in the presence of earthworms, nearly one-third
(30.5%) of seeds are removed entirely. These seeds are likely fully digested by earthworms.
This number appears to be at least partially driven by particularly high losses in smaller-
seeded species such as Betula and Alliaria.
Question 3. Do earthworms remove certain species of seed over others? Do traits such as
seed mass or species provenance seem to guide preferences?
Outcome. Although all species of seed are removed from the soil surface in significant
numbers compared to microcosms containing no earthworms, considerable differences
between species are clear. For instance, earthworms remove >90% (91.1%) of Alliaria
petiolata from the soil surface, compared to only 26.7% of Pinus strobus. Although only 6
species of seeds were used in this experiment, it is clear that species with a smaller seed mass
suffer from higher rates of earthworm granivory than those of larger mass. Since only 6
species were used, disentangling the effect of species provenance is challenging.
The second of the two core stated goals of this thesis is to quantify the relative
importance of invasive earthworms compared with more established seed predators. To
answer this issue I designed a seed-addition granivore-exclusion field experiment.
!
!
!
68!
Question 4. What proportion of seed predation can be attributed to earthworms, relative to
other guilds of seed predators?
Outcome. Seeds from more than 20 species of forest plants were added to a series of
exclosures which inhibited predation from either aboveground granivores (squirrels, mice,
birds, etc.) or belowground granivores (mainly earthworms). Results for different species
were highly variable between treatments, suggesting that different groups of granivores
target different sized seeds. Despite the differences between species, it is clear that plots
targeted by aboveground predators yielded significantly fewer recovered seeds than plots
subject to only earthworm predation. These findings were largely driven by medium to large
seeded species which were efficiently removed by aboveground granivores. Trends in
smaller seeded species were less clear, however for a number of species plots excluding
earthworms had a greater proportion of seeds recovered than those exposed to earthworms.
This may suggest that in at least some cases, earthworms are contributing a measurable effect
to seed removal. However, the trends in medium and large sized seeds strongly suggest that
rodents are primarily responsible for driving predation in many species.
Question 5. Are different species of seed removed at different rates? If so, what traits may
help explain these preferences? Seed mass? Species provenance?
Outcome. As mentioned previously seed removal was highly variable between both species
of seeds and granivore treatments; all species of seed tended to respond to granivores in
different ways. With that said, seed size appears to play a large role in predicting which
granivore guilds will remove a seed. Larger genera, such as Carya and Quercus, were often
removed at rates approaching 100% when exposed to aboveground granivores. Earthworms
often seem to prefer seeds that are less-desirable to efficient above-ground predators, such as
seeds of smaller mass.
3.2 SUMMARY OF FINDINGS
A small handful of studies (Eisenhauer et al. 2010; Quackenbush et al. 2012) have
pointed to earthworms as seed predators, and even made reference to potential traits that may
increase or decrease a seed's desirability to a granivorous earthworm. These results show that
indeed in an isolated environment, such as a microcosm or petri dish, earthworms are capable
of removing a remarkable quantity of seeds off the soil surface. They are also capable of
!
!
!
69!
exhibiting strong preferences, removing certain species of seeds in much greater quantity
than others (Quackenbush et al. 2012). At least some of these preferences can be explained
by seed size, although a more complete list of seeds representing realistic ranges in size
would be useful. Despite these interesting patterns in earthworm granivory, results from the
granivore exclusion experiment clearly indicate that earthworms likely provide a rather small
influence on granivory at the community level; aboveground granivores clearly remove seed
in considerably greater rates within Ontario forests.
A different system that still contains similarly abundant earthworm populations yet is
home to a considerably less prominent aboveground seed predator population could
potentially provide a situation where earthworms are a dominant granivore. Perhaps a
disturbance rich, possibly novel ecosystem, such as an active agricultural habitat may
provide such conditions. The findings of this thesis however demonstrate that although
exotic earthworms are a serious threat as an invasive species to Ontario’s temperate forests,
they likely have a modest impact on seed predation in a temperate forest and for most species
are unlikely to influence forest seedbank dynamics in a meaningful way.
3.3 DIRECTIONS FOR FUTURE RESEARCH During the development, execution, and analysis of this research project, a number of
interesting questions have arisen that I was unable to answer within the scope of this project.
Addressing these issues could perhaps lead to useful developments in our understanding of
this study system.
Earthworms have clearly demonstrated the ability to significantly reduce seed
availability in isolation, but appear to be easily eclipsed by more effective seed predators in
temperate forest systems. Could their effectiveness as a major seed predator be depressed in
this habitat compared to others? How do differences between earthworm densities in
different habitats factor into these conclusions? Could their impacts on seed survival likely
to be greater in a habitat with smaller seeds such as a meadow or old-field? Are their impacts
greater in habitats where granivorous rodents are scarce? Similar methods and approaches
could be used to answer these questions and more, to better understand the depth that these
invasive species impact the temperate region of North America.
!
!
!
70!
Very little is also known about the long-term implications for those seeds that are
transported vertically through the soil seed bank. Although it can be a very challenging
system to experimentally manipulate, a developed understanding of the long-term effects of
seed burial would certainly help us understand the scope of earthworm/seed dynamics.
Finally, recent efforts have also been made to better understand the role that ants of
temperate North America play in community level seed predation. Ants tend to prefer
smaller seeds, especially those with elaisomes, perhaps enabling both earthworms and ants to
function as competitors for certain species of seed.
!
!
!
71!
References Addison JA. 2009. Distribution and impacts of invasive earthworms in Canadian forest
ecosystems. Biological Invasions 11:59-79
Bohlen PJ, S Scheu, CM Hale, MA McLean, S Migge, PM Groffman, and D Parkinson.
2004. Non-native invasive earthworms as agents of change in northern temperate
forests. Frontiers in Ecology and the Environment 2:427-435
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
Holdsworth AR, LE Frelich, and PB Reich. 2007. Regional extent of an ecosystem engineer:
earthworm invasion in northern hardwood forests. Ecological Applications 17:1666-
1677
Jones CG, JH Lawton and M Shachak. 1994. Organisms as ecosystem engineers. Oikos
69:373-386
Quackenbush P, RA Butler, NC Emery, MA Jenkins, EJ Kladivko and KD Gibson. 2012.
Lumbricus terrestris prefers to consume garlic mustard (Alliaria petiolata) seeds.
Invasive Plant Science and Management 5:148-154
Sackett TE, SM Smith, and N Basiliko. 2014. Exotic earthworm distribution in a mixed- use
northern temperate forest region: influence of disturbance type, development age, and
soils. Canadian Journal of Forest Research 42:375-381
!
!
!
72!
Appendix APPENDIX 1 Rationalizations for the selection of each species of seed.
Acer platanoides Exotic Exotic congener of native maples Acer rubrum Native Widespread deciduous tree Acer saccharum Native Common (often dominant) deciduous tree species Alliaria petiolata Exotic Problematic herbaceous invader, possibly preferred by worms Berberis thunbergii Exotic Problematic invasive shrub Betula alleghaniensis Native Small seeded deciduous tree Betula pendula Exotic Exotic congener of native birch Betula papyrifera Native Common, small seeded deciduous tree Carya cordiformis Native Large-seeded nut tree Carya ovata Native Large-seeded nut tree Fraxinus americana Native Common deciduous tree Larix laricina Native Small seeded deciduous tree Lonicera canadensis Native Widespread deciduous shrub Maianthemum racemosum Native Native herbaceous groundcover in mixed forests Pinus banksiana Native Native congener of exotic pine Pinus strobus Native Common coniferous tree species in mixed forests Pinus sylvestris Exotic Exotic congener of other native pines Prunus serotina Native Common deciduous tree species, seeds an important food source Quercus rubra Native Large-seeded nut tree Rhamnus cathartica Exotic Problematic invasive shrub/tree Robinia pseudoacacia Exotic Larger-seeded exotic tree Tilia americana Native Widespread deciduous tree Tsuga canadensis Native Small-seeded native conifer tree
!
!
!
73!
APPENDIX 2 Extended abstract.
THE IMPORTANCE OF INVASIVE EARTHWORMS AS SEED PREDATORS OF COMMON FOREST FLORA OF ONTARIO
Colin M Cassin
Master of Science
Ecology and Evolutionary Biology Department University of Toronto
2015 ABSTRACT Soil seedbanks play a vital role in forest plant communities, having long been viewed as a refuge for
seeds that are vulnerable to aboveground seed predators. This study provides evidence that seeds
entering the soil seedbank may be subject to previously underestimated rates of granivory by a
common species of invasive earthworm, Lumbricus terrestris. We report that nearly 55% of seeds
from 6 ecologically important forest species were either digested or buried below a germinable depth
when in the presence of earthworms in a microcosm experiment. L. terrestris was also found to
preferentially remove certain species of seed over others, for instance smaller sized seeds were often
removed in higher abundances than larger sized seeds. The common forest invader Garlic Mustard,
Alliaria petiolata, was subject to the highest removal rates in this study, as nearly 90% of seeds were
effectively destroyed by earthworms.
In contrast, results from a field exclusion experiment indicate that seed predation by rodents may
eclipse that of earthworms under natural conditions, for most temperate forest seeds. While seed
predation by rodents was high in mid to large seeded species, earthworms produced a greater relative
effect in seeds of smaller mass.
These findings suggest that although rodents are the main driver of seed predation, earthworms may
have the potential to act as an additional layer of ecological filter, and potentially further influence
the species composition of future forest plant communities by selectively targeting certain seeds, or
seed traits, over others. Although the effects may be small, the application of these findings may
prove useful to management of forests where threat of earthworm invasion exists.