Biol Invasions (2010) 12:1847–1854
DOI 10.1007/s10530-009-9596-z
ORIGINAL PAPER
Quantifying the levels of sexual reproduction and clonal
spread in the invasive plant, Rosa multiflora
Laura C. Jesse • John D. Nason • John J. Obrycki
Kirk A. Moloney
•
Received: 8 July 2008 / Accepted: 30 September 2009 / Published online: 21 October 2009
Ó Springer Science+Business Media B.V. 2009
Abstract Rosa multiflora Thunb., (Rosaceae), an
invasive plant in the eastern U.S., was introduced into
the U.S. in the early 1800s and was widely planted in
the 1940s as a living fence, for wildlife cover, and to
prevent soil erosion. This species spread rapidly from
these original plantings via seed dispersal (sexual
reproduction) and clonal spread, invading pasture
and wooded areas. In this study we used allozyme
markers to test for significant differences in the levels
of asexual and sexual spread in large ([9 m cirumference) verses small (\2 m circumference) patches
of R. multiflora and in pasture verses park settings.
Although larger patches of R. multiflora tended to be
dominated by one genotype, they exhibited significantly greater genetic diversity and inputs from
sexual reproduction than did small patches; all large
patches (N = 10) contained multiple unique genotypes. In contrast six of ten smaller patches of R.
multiflora, consisted of a single genotype, though
three patches had two genotypes and one had three
L. C. Jesse (&)
Department of Plant Pathology, Iowa State University,
Ames, IA 50011, USA
e-mail: ljesse@iastate.edu
J. D. Nason K. A. Moloney
Department of Ecology, Evolution & Organismal
Biology, Iowa State University, Ames, IA 50011, USA
J. J. Obrycki
Department of Entomology, University of Kentucky,
Lexington, KY 40546, USA
unique genotypes. Similar analyses revealed clonal
structure in R. multiflora populations both park and
pasture habitats but with significantly greater genetic
diversity and sexual inputs in the former than the
latter. These results are consistent with a model of
invasive spread involving clonal spread, sexual
reproduction, and bird-mediated seed dispersal into
established patches. Sexual inputs appear to be
highest in larger patches and park habitats where
perching sites for birds are most abundant. This
flexible reproduction system likely contributes to the
invasiveness of R. multiflora and to current management failures.
Keywords Multiflora rose Allozyme Clonal reproduction
Introduction
The negative effects of invasive taxa on native
species, biodiversity and ecosystem function has been
widely recognized and documented (Meyers and
Bazely 2003). Nonetheless, there are many non-native
species in the U.S. that are not disrupting natural
systems; it is estimated that only 0.1% of all
introduced species will become serious invaders
(Williamson and Fritter 1996). The life history
characteristics of invasive plants are of interest
because they may shed light on why certain species
are invasive and allow us to predict invasive potential
123
1848
before a plant is introduced into new habitats (Sakai
et al. 2001). In addition, understanding the life-history
characteristics, including the genetic diversity, of an
invasive plant is often key to effective control
strategies and tactics (Burdon and Marshall 1981).
Rosa multiflora was intentionally introduced
multiple times to North America from multiple
sources in eastern Asia in the 1940s and 50s to serve
as a living fence in pastures, as a means to control
soil erosion, and to provide food and cover for
wildlife (Burgess 1948; Klimstra 1956). This multistemmed shrub is now invasive in central and eastern
United States (Banasiak and Meiners 2009) and can
form impenetrable thickets that exclude native plant
species and decrease pasture quality. Rosa multiflora
is an insect-pollinated outcrossing species with seeds
dispersed primarily by birds and rodents (Klimstra
1956; Scott 1965; Jesse 2006). It also reproduces
clonally via roots and by layering (arching branches
that reach the ground) (Burgess 1948; Klimstra 1956;
USDA 1971). When R. multiflora was introduced to
the U.S., it was presumed that seeds would germinate
only under ‘‘ideal nursery conditions’’ and that
propagation by layering would occur infrequently,
when a branch was covered with soil (Burgess 1948).
Consequently, invasive spread of this species was not
predicted to be a problem (Steavenson 1946). Within
10 years, however, it was apparent that R. multiflora
was spreading from the original plantings and was
considered to be a serious emerging problem as an
invasive species (Klimstra 1956).
Rosa multiflora seeds are contained within a red
fleshy hip and are dispersed primarily by birds and
rodents (Klimstra 1956; Scott 1965; Schery 1977;
Willson 1986). Rosa multiflora bushes are a common
winter habitat for a variety of birds (Steavenson 1946;
Schmid 1958; Morgan and Gates 1982), which was
one of the primary reasons it was initially planted in
the United States. In a 43 ha area in central New
Jersey, Suthers et al. (2000) observed R. multiflora
seeds in the excrement of gray catbirds, hermit
thrushes, American robin, and cedar waxwings captured in mist-nets from September to November.
Stiles (1982) observed mockingbirds, Mimus polyglottos, feeding on R. multiflora hips in New Jersey
during February and March: 31.5% of the food items
consumed by one individual mockingbird were
R. multiflora hips.
123
L. C. Jesse et al.
Reproduction by both seed and clonal growth may
contribute to the invasiveness of R. multiflora, but the
relative contribution of each mode of reproduction to
its spread has not been examined. Management of
R. multiflora requires a reduction in both the number
of plants in an area, including the seed bank, and the
potential spread of these plants. Two biological
control agents of R. multiflora that have the potential
to reduce its spread have been examined in the United
States. One is rose rosette disease (RRD), whose
causative agent has not been isolated and identified.
This disease slows the rate of seed production and
clonal spread and can kill R. multiflora plants within
several years (Epstein and Hill 1997). The other is a
wasp, Megastigmus aculeatus (Hymenoptera: Torymidae), whose larvae develop in the seeds of
R. multiflora (Weiss 1917; Jesse 2006). If sexual
reproduction does not contribute significantly to the
spread of R. multiflora, the wasp should have little
effect on its biological suppression.
Allozyme markers have been used to determine
the frequency and spatial distribution of individual
plants produced sexually (via seed) or asexually (via
clonal spread) (e.g., Hamrick et al. 1979; Ceplitis
2001; Ivey and Richards 2001). In this study we
used variation in allozymes to quantify the relative
levels of R. multiflora spread by seed dispersal and
clonal growth within two habitats: woodland park
and open pasture. In the park setting we expected a
significant contribution to spread by seed, since trees
and other plants serve as perching sites for birds, one
of the primary dispersers of R. multiflora seeds. In
pasture areas, in contrast, we expected layering to be
a more important determinant of invasive spread
because cattle often trample branches into the
ground (Klimstra 1956), potentially promoting the
production of clonal plants. As clonal patches
become larger, however, they increasingly form
suitable perching places for birds so that greater
evidence of sexual reproduction (recruitment by
seed) may be expected for larger patches as compared to smaller patches of plants in pastures.
Consequently, in addition to comparing park and
pasture effects, within each habitat we also compared patches of large ([9 m cir.) and small size
(\2 m cir.) to test the hypothesis that recruitment by
seed augments clonal reproduction in the spread of
R. multiflora.
Quantifying the levels of sexual reproduction and clonal spread
1849
Materials and methods
Allozyme analysis
Collection of material
Fresh leaves sampled from patches were refrigerated
(5°C) before 1–2 leaflets from each leaf were ground
with a mortar and pestle in the extraction buffer of
Mitton et al. (1979). Total protein extracts were
absorbed through Miracloth (Calbiochem, La Jolla,
CA) into 3 9 7 mm filter paper wicks, which were
stored at -70°C until analysis. After an initial
screening of 3 extraction buffers, 6 buffer systems,
and 18 enzyme systems, we identified 6 polymorphic
allozyme loci exhibiting expected subunit structures
and patterns of expression. Ten percent starch gels
and a 0.04 M morpholine-citrate gel-electrode buffer
system adjusted to pH 6.1 (Murphy et al. 1996) were
used to resolve aconitase (Acoh, EC 4.2.1.3), isocitrate dehydrogenase (Idh-1, Idh-2, EC 1.1.1.42),
6-phosphogluconate dehydrogenase (Pgd-1, Pgd-2,
EC 1.1.1.44), and glucose-6-phosphate isomerage
(Gpi, EC 5.3.1.9). Staining protocols for individual
loci followed Soltis et al. (1983).
Samples of R. multiflora leaf tissue were taken from
two sites in central Iowa (Boone Co. and Story Co.)
that are heavily infested with R. multiflora. The Story
Co. site (N42°110 , W93°100 ) is a county owned park,
some wooded areas and open unmowed areas that are
infested with R. multiflora. Four of the five pairs of
R. multiflora patches sampled were growing, at least
partially, under a tree. The Boone Co. site (N42°80 ,
W93°550 ) is a privately owned and grazed cattle
pasture with grazed grassy areas with scattered trees
and an adjacent wooded area. Rosa multiflora grows
as an understory plant in the wooded areas, under
trees within the pasture and in open grassy areas. All
of the five pairs of R. multiflora patches sampled were
growing, at least partially, under a tree. At each site,
leaves were taken from five pairs of patches, each
pair consisting of a large and a small patch of plants
separated by \3 m. A small patch was defined as
what appeared to be a single multi-stemmed plant
\2 m in circumference; a large patch appeared to
consist of multiple plants and was over 9 m in
circumference (Table 1). Four samples were taken
from each small patch (\2 m cir.) at the cardinal
points and eight samples were taken from each large
patch, also at the cardinal points and at the primary
intercardinal points. Pairs of patches were separated
from other such pairs by at least 5 m.
Data analysis
For each patch, we determined the number and
frequency of multilocus allozyme genotypes. Three
measures of genotypic diversity and evenness were
used to evaluate these data (Ellstrand and Roose 1987;
Stoddart 1983). Genotypic diversity was calculated
with the Simpson Index modified for a finite sample
size (Pielou 1969). This measures the probability that
Table 1 Genetic diversity in large patches (circumference [9 m) of R. multiflora (8 samples/patch) growing in park (Story Co.) and
pasture (Boone Co.) habitats in central Iowa
Site
Patch
Circ. (m)
Distinct
genotypes
Genotypes/N
D
E
Go
Go/Ge
P
Story
1
23.8
5
0.63
0.86
0.56
4.00
0.77
0.283
Story
2
18.9
4
0.5
0.82
0.83
3.56
0.69
0.152
Story
3
18.3
4
0.5
0.64
0.00
2.29
0.44
0.024
Story
4
13.4
6
0.75
0.92
0.60
5.33
1.03
0.648
Story
5
10.4
5
0.63
0.78
0.00
3.20
0.62
0.138
Boone
1
25.6
4
0.5
0.64
0.00
2.29
0.42
0.018
Boone
2
21.9
4
0.5
0.75
0.50
2.91
0.53
0.043
Boone
3
14.6
2
0.25
0.57
0.10
2.00
0.36
0.004
Boone
4
31.7
4
0.5
0.75
0.50
2.91
0.53
0.043
Boone
5
9.1
4
0.5
0.75
0.50
2.91
0.53
0.043
D diversity calculated with a modification of Simpson index, E evenness, Go observed genotypic diversity, Ge the mean of simulated
Ge values for a patch of size N. P-values in bold are significant after sequential Bonferroni adjustment
123
1850
L. C. Jesse et al.
two ramets selected at random from a population of N
plants will have different multilocus genotypes.
Values of D range from 0 to 1:
D¼1
G
X
ni ðni 1Þ
i¼1
NðN 1Þ
where ni is the number of individuals of genotype i, N
is the number of individuals sampled, and G is the
number of distinct multilocus genotypes detected in
the population. Genotypic evenness (Frager 1972;
Ivey and Richards 2001; Novak and Mack 2000) was
measured as:
E¼
Dobs Dmin
Dmax Dmin
where Dobs ¼ D from above,
Dmin ¼
ðG 1Þð2N GÞ
NðN 1Þ
and
Dmax ¼
NðG 1Þ
GðN 1Þ
Values of E can range from 0, if the sample
population is dominated by one genotype, to 1 if each
genotype in the population is represented by the same
number of ramets.
Finally, we calculated the observed genotypic
diversity (Go) for each patch of plants (Stoddart
1983):
.X
p2i
Go ¼ 1
where pi is the observed frequency of genotype i
sampled from a given patch. Go can range from 1, if
all sample genotypes are identical, to N, if each
genotype in the sample is unique. Because allozyme
diversity can underestimate clonal diversity, we
tested if our observed genotypic diversity differed
significantly from a population reproducing only
sexually we used simulations to construct a distribution of expected genotypic (Ge) diversity for each
patch under the null hypothesis of free sexual
recombination. Following Ceplitis (2001), for a patch
of sample size N the null distribution of Ge (and its
mean) was determined by calculating Ge for each
of 999 simulated sets of N multilocus genotypes
constructed from study site allele frequencies assuming sexual reproduction. These simulated values of
123
Ge were ordered from lowest to highest, the rank
(r = 1–1,000) of the patch’s observed value (Go)
determined, and the value r/1,000 used to determine
the probability (P) of Go under the null hypothesis of
pure sexual reproduction. To evaluate the statistical
significance of each P-value, a sequential Bonferroni
correction (Rice 1989) was used to adjust testwise
error rates for multiple tests on patches of size N
within a site.
In addition to individual patch level tests, a global
test of Go for deviation from tie was conducted over
all patches of size N within a site using Fisher’s
combined probability test (Sokal and Rohlf 1995). To
test the effects of site (park vs. pasture), patch sixe
(large vs. small), and their interaction (site 9 patch
size) we used a non-parametric, fixed-effects twoway ANOVA for ranked data (the Schreirer Ray Hare
extension of the Kruskal–Wallis test (Sokal and Rohlf
1995). As a response variable we used the ratio
Go/Ge, a measure of the fraction of reproduction due
to sexual recombination, seed dispersal, and recruitment (as opposed to clonal spread) that takes into
account differences between sites in available variation at genetic marker loci and between patch size in
sample size. Pending a non-significant site 9 patch
size interaction we tested the effects on Go/Ge of site
(irrespective of patch size) and patch size (irrespective of site) using Mann–Whitney U-tests.
Results
The ability of genetic markers to distinguish among
clonal genotypes increases with the number of marker
loci and their allelic diversity. Taken over the six
polymorphic allozyme loci, levels of genetic diversity
provided good power to distinguish clones from
mulitlocus genotypes that were identical by chance
under sexual reproduction. Based on samples of N = 8
ramets, large patches ([9 m cir.) of R. multiflora had
2–6 distinguishable multilocus allozyme genotypes,
averaging (SE) 4.2(0.3) across both field sites in
central Iowa (Table 1). Although the diversity of
genotypes (D) in large patches ranged from 0.57 to
0.92, many patches were dominated by a single
genotype (E ranged from 0 to 0.83 with an average of
0.36). None consisted solely of a single genotype,
however, indicating some level of sexual reproduction
in each large patch.
Quantifying the levels of sexual reproduction and clonal spread
The observed genotypic diversity (Go) for large
patches in Story Co. ranged from 2.29 to 5.33
(Table 1). After sequential Bonferroni adjustment,
only one patch (Patch 3) differed significantly
(P = 0.024) from the genotypic diversity expected
from complete sexual reproduction. The observed
genotypic diversity for large patches in Boone Co.
ranged from 2.00 to 2.91 with two patches (Patches 1:
P = 0.018, and 3: P = 0.004) exhibiting significant
departures from complete sexual reproduction after
adjustment for multiple tests. We interpret these
significant results as evidence of clonal spread within
some; but not all, large patches of R. multiflora. While
only a subset of large patches exhibited significant
evidence of clonal spread, a global test of Go over
patches was significant for both sites (Story Co.:
P \ 0.05; Boone Co.: P \ 0.001) indicating an overall excess of identical genotypes consistent with a
mixture of sexual reproduction and clonal spread.
Based on samples of N = 4 ramets, small patches
of R. multiflora had 1–3 distinguishable genotypes,
with an average (SE) of 1.5(0.25) genotypes across
both field sites (Table 2). The diversity of genotypes
(D) in small patches ranged from 0.00 to 0.83: with 6
of 10 patches consisting of a single multilocus
genotype (E ranged from 0 to 1.00 with an average
of 0.1). Of note, in 3 of 5 cases at both study sites,
small patches of R. multiflora had unique genotypes
compared to their paired large patch located nearby.
The observed genotypic diversity (Go) for small
patches of R. multiflora in Story Co. ranged from 1.0
to 2.67 (Table 2). After sequential Bonferroni
Table 2 Genetic diversity in small patches (\2 m cir.) of
R. multiflora (4 samples/patch) growing in two sites (Story Co.
and Boone Co.) in central Iowa
E
Go
Table 3 Nonparametric two-way ANOVA examining the
effects of site, patch size, and their interaction on Go/Ge
Source
df
SS
2
X1;0:05
Site
1
115.2
3.32
Patch size
1
168.2
4.85
Size 9 patch size
1
0.8
0.00
16
374.3
Error
Bold indicates significance at the P \ 0.05 level
adjustment, three patches (Patch 1, 4, and 5) differed
significantly from the genotypic diversity expected
under pure sexual reproduction. Small patches of
R. multiflora growing at the Boone Co. site had
observed genotypic diversities ranging from 1.0 to
1.6 (Table 2), with three patches (Patches 1, 2 and 4)
differing significantly from expected values with pure
sexual reproduction. These six patches with significant Go values are the same six patches consisting of
a single genotype, a pattern indicative of complete
asexual reproduction via clonal spread. In contrast the
four other small patches consisted of 2 or 3 different
sample genotypes, requiring at least some sexual
reproduction. A global test of Go over patches was
highly significant for both sites (P \ 0.001), not
surprising given the number of patches at each site
consisting of a single genotype.
The two-way ANOVA indicated a significant
effect of patch size (0.05 [ P [ 0.025) but not site
(0.10 [ P [ 0.05) or their interaction (P [ 0.05) on
the ratio Go/Ge (Table 3). Subsequent Mann–Whitney
U-tests indicate significant effects on Go/Ge of both site
and patch size with mean Go/Ge lower and hence clonal
spread greater in pasture (Boone Co.) than park sites
(Story Co.) and in small patches compared to large.
Go/Ge P
Site
Patch Genotypes D
Story
1
1
0.00 0.00 1.00 0.30
0.007
Story
2
3
0.83 0.00 2.67 0.80
0.46
Story
3
2
0.67 1.00 2.00 0.60
0.07
Story
4
1
0.00 0.00 1.00 0.30
0.007
Story 5
Boone 1
1
1
0.00 0.00 1.00 0.30
0.00 0.00 1.00 0.29
0.007
0.003
Boone 2
1
0.00 0.00 1.00 0.29
0.003
Boone 3
2
0.50 0.00 1.60 0.46
0.04
Boone 4
1
0.00 0.00 1.00 0.29
0.003
Boone 5
2
0.50 0.00 1.60 0.46
0.04
Parameters are as described in Table 1
1851
Discussion
Our analysis of two sites in central Iowa indicate that at
the scale of contiguous plants (patches up to 32 m in
circumference) R. multiflora is spreading via a combination of sexual and clonal reproduction. Although
overall genetic diversity of large patches was less than
that expected under pure sexual reproduction at both
field sites in Iowa, multiple genotypes were found in
all large patches. Small patches of R. multiflora, in
contrast, often consisted of a single genotype, with six
of ten small patches having genotypes not detected in
123
1852
the nearby large patch. Taken together these results are
consistent with the hypothesis that patches are established by seed, spread clonally, and at larger sizes are
augmented by additional recruitment via seed. Indeed,
ANOVA and Mann–Whitney U-tests indicate Go/Ge,
and hence the combined effect of sexual reproduction,
seed dispersal, and subsequent recruitment, is greater
in larger patches than small.
Though not as strong as patch size effects, the
influence of site was also significant. The ratio of
Go/Ge not only differed between park (Story Co.) and
pasture (Boone Co.) sites but was greater in the
former. This is consistent with the hypothesis that by
providing more perching sites for frugivorous birds,
more wooded settings, like parks, increase the
likelihood that seed dispersal and recruitment will
augment clonal spread in forming invasive patches of
R. multiflora.
More generally, Eriksson (1989) examined demographic studies of 68 plant species that reproduce
clonally to determine rates of seedling recruitment
into their populations. Forty percent of the clonal
plants repeatedly included seedling recruitment into
their populations. In contrast to our findings for
R. multiflora, Eriksson (1989) found greater seedling
recruitment was observed in species growing in
grasslands (verses woodlands), with above ground
clonal growth, phalanx style genet structure (verses
guerilla style), and short distance seed dispersal
(Eriksson 1989). The increases of R. multiflora may
be enhanced by bird dispersal of seeds which
provides long distance seed dispersal.
The benefits of clonal growth include rapid increase
in size, capture of resources, and reduced mortality of
young ramets, relative to sexual reproduction (Silvertown and Doust 1993). Plant species seldom completely loose the ability to reproduce sexually,
however, presumably because the advantages of
maintaining genetic diversity outweigh the costs of
sex (Silvertown and Doust 1993). In simulation
models with the clonal plant Ranunculus repens, Soan
and Watkinson (1979) found that even occasional
establishment of sexually produced seedlings successfully maintained genetic variability within populations. Such levels of sexual reproduction can be quite
high. In Mahonia aquifolium, for example, a clonal
invasive shrub in Europe, approximately 50% of new
ramets are estimated to arise via seedling recruitment
(Auge and Brandl 1997). In a review of genotypic
123
L. C. Jesse et al.
diversity in clonal plants Ellstrand and Roose (1987)
found that most populations contained multiple unique
clones consistent with recurrent sexual reproduction
and recruitment. Such multiclonal populations usually
showed intermediate levels of diversity and evenness.
Species that regularly produced sexual progeny had
relatively higher levels of genetic diversity (0.29–1.0
clones/sample) (Ellstrand and Roose 1987). We
observed a similar range (0.25–0.75) in the number
of clones/sample in the large patches of R. multiflora,
where evidence of sexual inputs were greatest.
Plants growing from bird-disseminated seeds often
grow under fencerows and isolated trees in fields
(McAtee 1947). Generally, relatively few bird dispersed seeds are found in areas without a perch site
(Ferguson and Drake 1999). In a study conducted in
New Jersey to determine the effect of vegetation
height on bird-dispersed seeds, McDonnel (1986)
observed that 16.4% of seeds found in bird excrement
collected in traps placed under artificial saplings were
from R. multiflora. During the winter of 2003–2004,
we collected an average of 125 seeds and 151 hips in
four 1 m2 boxes placed underneath four R. multiflora
bushes in eastern Iowa (Jesse, unpublished data). The
whole hips may have been knocked off the bush by
wind or foraging mammals. The separate seeds may
have been in a hip consumed by a bird or chewed
apart by a rodent or other mammal. In either case,
these findings together with our genetic data demonstrate that R. multiflora seeds collect and successfully
recruit new individuals under established plants.
Clonally reproducing invasive plants that have
been introduced multiple times into a new environment will presumably have higher genotypic diversity
than species with a single introduction (Pappert et al.
2000; Khudamrongsawat et al. 2004). For example,
kudzu, Pueraria lobata, an invasive ornamental that
was introduced multiple times into North America
was found to have an average Simpsons D of 0.69
and evenness (E Fragers) of 0.74 (Pappert et al. 2000)
indicating a diversity of different genotypes. Compared to kudzu, R. multiflora had a slightly higher
level of diversity, but lower evenness (average
E = 0.36). Rosa multiflora was introduced multiple
times into the U.S., with multiple shipments of seeds
entering the U.S. from Japan being reported in just
the spring of 1917 (Weiss 1917).
Our studies have documented that R. multiflora is
spreading clonally, but like many clonal plants, has
Quantifying the levels of sexual reproduction and clonal spread
retained genetic diversity. However, these results
may not be representative of how R. multiflora is
spreading elsewhere as we analyzed populations only
in Iowa. The genotypic diversity of invasive weeds
may affect levels of control, e.g., high levels of
genetic diversity may facilitate resistance to management practices, including biological control organisms (Khudamrongsawat et al. 2004; Ahmad et al.
2008). Burdon and Marshall (1981) concluded that it
was probably due to genetic homogeneity that
biological control was more effective against asexually reproducing weeds. Intentionally introduced
weeds that are also capable of sexual reproduction,
like R. multiflora, will be more difficult to control as
they will have the genetic diversity to adapt to
biological control agents (Burdon and Marshall
1981). Genetic diversity may be a contributing factor
to current difficulties in managing R. multiflora. For
example, R. multiflora in Iowa is susceptible to rose
rosette disease and to seed predation by the larvae of
the introduced wasp Megastigmus aculeatus (Hymenoptera: Torymidae) (Jesse 2006). Resistant individuals produced by sexual recombination may be
selected for, however, making these sources of
control less effective in the future. Seedling recruitment occurs in R. multiflora populations, so biological controls that reduce seed output may also reduce
spread. We observed a reduction in viable seeds per
rosehip at R. multiflora sites when M. aculeatus
larvae consume viable seeds (Jesse 2006). It will be
necessary to combine mechanical controls with any
biological control for effective R. multiflora management. Dead R. multiflora plants are woody and will
remain upright for several years, serving as a focal
point for bird perches and will aid in protecting new
seedlings from grazing animals. Therefore, removing
dead plants by mowing or burning needs to be
included as a component of R. multiflora management.
References
Ahmad R, Liow PZ, Spencer DF, Jasieniuk M (2008) Molecular evidence for a single genetic clone of invasive
Arundo donax in the United States. Aquat Bot 88:113–120
Auge H, Brandl R (1997) Seedling recruitment in the invasive
clonal shrub. Mahonia aquifolium Pursh (Nutt.). Oecologia 110:205–211
Banasiak SE, Meiners SJ (2009) Long term dynamics of
Rosa multiflora in a successional system. Biol Invasions
11:215–224
1853
Burdon JJ, Marshall DR (1981) Biological control and the
reproductive mode of weeds. J Appl Ecol 18:649–659
Burgess CH (1948) Living fence. Farm Q 3:50–54
Ceplitis A (2001) The importance of sexual and asexual
reproduction in the recent evolution of Allium vineale.
Evolution 55:1581–1591
Ellstrand NC, Roose ML (1987) Patterns of genotypic diversity
in clonal plant species. Am J Bot 74:123–131
Epstein AH, Hill JH (1997) Augmentation of rose rosette
disease for biocontrol of multiflora rose (Rosa multiflora).
Weed Sci 45:172–178
Eriksson O (1989) Seedling dynamics and life histories in
clonal plants. Oikos 55:231–238
Ferguson RN, Drake DR (1999) Influence of vegetation
structure on spatial patterns of seed deposition by birds.
NZ J Bot 37:671–677
Frager EW (1972) Diversity: a sampling study. Am Nat
106:293–310
Hamrick JL, Linhart YB, Mitton JB (1979) Relationships
between life history characteristics and electrophoretically
detectable genetic variation in plants. Ann Rev Ecol Syst
10:173–200
Ivey CT, Richards JH (2001) Genotypic diversity and clonal
structure of Everglades sawgrass, Cladium jamaicense
(Cyperaceae). Int J Plant Sci 162:1327–1335
Jesse LC (2006) The biology of Rosa multiflora (Rosaceae) and
two of its biotic mortality factors. Thesis, Iowa State
University, Ames, IA
Khudamrongsawat J, Tayyar R, Holt JS (2004) Genetic
diversity of giant reed (Arundo donax) in the Santa Ana
River, California. Weed Sci 52:395–405
Klimstra WD (1956) Problems in the use of multiflora rose. Ill.
Trans Ill State Acad Sci 58:66–72
McAtee WL (1947) Distribution of seeds by birds. Am Midl
Nat 38:214–223
McDonnel MJ (1986) Old field vegetation height and the dispersal pattern of bird-disseminated woody plants. Bull
Torrey Bot Club 133:6–11
Meyers JH, Bazely DR (2003) Ecology and control of introduced plants. Cambridge University Press, Cambridge, UK
Mitton JB, Linhart YB, Sturgeon KB, Hamrick JL (1979) Allozyme polymorphisms detected in mature needle tissue
of ponderosa pine. J Hered 70:86–89
Morgan KA, Gates JE (1982) Bird population patterns in forest
edge and strip vegetation at Remington Farms, Maryland.
J Wildl Manage 46:933–944
Murphy RW, Sites JW, Buth DG, Haufler CH (1996) Proteins:
isozyme electrophoresis. In: Hillis DM, Moritz C, Mable
BK (eds) Molecular systematics. Sinauer Associates,
Sunderland, MA, pp 51–120
Novak SJ, Mack RN (2000) Clonal diversity within and among
introduced populations of the apomictic vine Bryonia alba
(Cucurbitaceae). Can J Bot 78:1469–1481
Pappert RA, Hamrick JL, Donovan LA (2000) Genetic variation in Pueraria lobata (Fabaceae), an introduced, clonal,
invasive plant of the southeastern United States. Am J Bot
87:1240–1245
Pielou EC (1969) An introduction to mathematical ecology.
John Wiley & Sons, Inc, New York
Rice W (1989) Analyzing tables of statistical tests. Evolution
43:223–225
123
1854
Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J,
With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand
NC, McCauley DE, O’Neil P, Parker IM, Thompson JN,
Weller SG (2001) The population biology of invasive
species. Annu Rev Ecol Syst 32:305–332
Schery R (1977) The curious double life of Rosa multiflora.
Horticulture 55:56–61
Schmid FC (1958) Cedar waxwings and Fox sparrows feed
upon Multiflora rose. Wilson Bull 70:194–195
Scott RF (1965) Problems of multiflora rose spread and control. Thirtieth North American Wildlife Conference 13:
360–378
Silvertown JW, Doust JL (1993) Introduction to plant population biology. Blackwell, Oxford, UK
Soan ID, Watkinson AR (1979) Clonal variation of Ranunculus
repens. New Phytol 82:557–573
Sokal RR, Rohlf FJ (1995) Biometry: the principles and
practice of statistics in biological research. W.H. Freeman
and Company, New York
Soltis DE, Haufler CH, Darrow DC and Gastony GJ (1983)
Starch gel electrophoresis of ferns: a compilation of
grinding buffers, gel and electrode buffers, and staining
schedules. Am Fern J 73:9–27
123
L. C. Jesse et al.
Steavenson HA (1946) Multiflora rose for farm hedges. J Wildl
Manag 10:227–234
Stiles EW (1982) Expansions of Mockingbird and multiflora
rose in the northeastern United States and Canada. Am
Birds 36:358–364
Stoddart JA (1983) A genotypic diversity measure. J Hered
74:489–490
Suthers HB, Bickal JM, Rodewald PG (2000) Use of successional habitat and fruit resources by songbirds during
autumn migration in central New Jersey. Wilson Bull
112:249–260
USDA (1971) Common weeds of the United States. United
States Department of Agriculture—Agricultural Research
Service. Dover Publications, New York, NY
Weiss HB (1917) Megastigmus aculeatus Swed., introduced
into New Jersey from Japan. J Econ Entomol 10:448
Williamson M, Fritter A (1996) The varying success of
invaders. Ecology 77(6):1661–1666
Willson MF (1986) Avian frugivory and seed dispersal in
eastern North America. Curr Ornithol 3:223–279