ArticlePDF Available

Can sowing density facilitate a higher level of forb abundance, biomass, and richness in urban, perennial “wildflower” meadows?


Abstract and Figures

Forb species abundance and richness determine both ecological and social values in naturalistic meadows in urban landscapes. However, species loss and dominance through competition are naturally part of meadow ecological processes often leading on productive soils to large grass biomass in the absence of appropriate management. Sowing density is a design tool to manipulate the initial number of emergents of each component species however high sowing densities may not benefit community performance in terms of species richness and diversity in the longer term. This study investigated the effect of sowing density on forb species abundance, biomass and richness. Two sowing densities approximating to 500 and 1,000 emerged seedlings/m² were employed with 29 forb and one grass species. The higher sowing density did not lead to a larger grass biomass that dominated the community, as the grass species used was ultimately less competitive than the forb dominants. Increasing sowing density increased the number of forb seedlings initially but this declined, as did species richness in the longer term. In terms of subordinate forb survival, ability to access light resources to survive intense competition from dominants was key. Tall, and native species were more likely to maintain higher seedling numbers in the longer term. The research suggest that lower sowing rates are likely to be most useful on soils which are either unproductive, do not contain a significant weed seed banks, where weed free sowing mulches are employed or in rural situations where there is less immediate political pressure for rapid development of forb rich meadows.
Content may be subject to copyright.
Urban Forestry & Urban Greening 74 (2022) 127657
Available online 22 June 2022
1618-8667/© 2022 The Author(s). Published by Elsevier GmbH. This is an open access article under the CC BY license
Original article
Can sowing density facilitate a higher level of forb abundance, biomass, and
richness in urban, perennial wildowermeadows?
Mingyu Jiang
, James D. Hitchmough
Department of Landscape Architecture, The University of Shefeld, S10 2TN, UK
Horticulture and Landscape Department, Scotlands Rural College (SRUC), Kings Buildings, Edinburgh, EH9 3JG, UK
Designed meadows
Forb biomass
Sowing rates
Species competition and dominance
Forb species abundance and richness determine both ecological and social values in naturalistic meadows in
urban landscapes. However, species loss and dominance through competition are naturally part of meadow
ecological processes often leading on productive soils to large grass biomass in the absence of appropriate
management. Sowing density is a design tool to manipulate the initial number of emergents of each component
species however high sowing densities may not benet community performance in terms of species richness and
diversity in the longer term. This study investigated the effect of sowing density on forb species abundance,
biomass and richness. Two sowing densities approximating to 500 and 1,000 emerged seedlings/m
employed with 29 forb and one grass species. The higher sowing density did not lead to a larger grass biomass
that dominated the community, as the grass species used was ultimately less competitive than the forb domi-
nants. Increasing sowing density increased the number of forb seedlings initially but this declined, as did species
richness in the longer term. In terms of subordinate forb survival, ability to access light resources to survive
intense competition from dominants was key. Tall, and native species were more likely to maintain higher
seedling numbers in the longer term. The research suggest that lower sowing rates are likely to be most useful on
soils which are either unproductive, do not contain a signicant weed seed banks, where weed free sowing
mulches are employed or in rural situations where there is less immediate political pressure for rapid devel-
opment of forb rich meadows.
1. Introduction
Naturalistic meadows, inspired by the complexity of more natural
meadow communities, have become fashionable as an alternative to
conventional mown grasslands for landscape designers in both Western
and Eastern countries (Hitchmough and Dunnett, 2004; Jiang and Yuan,
2017). This approach has potential to improve ecological as well as
social value, which is of equal importance for sustainable urban land-
scapes ( ¨
Ozgüner et al., 2007; Hicks et al., 2016; Southon et al., 2018).
Within this context, forb species richness (the number of species per unit
area) becomes extremely critical not only to support a diversity of pol-
linators (Potts et al., 2009; Hicks et al., 2016) but also to deliver
aesthetic benets to the public (Hoyle et al., 2018). Highly owery,
meadow communities rich in forb species have been shown to be one of
the greenspace types most appreciated by the public (Southon et al.,
2017; Hoyle et al., 2018).
Despite the evidence that urban publics are becoming increasingly
biocentric, social acceptance and social sustainability may largely
depend on the clarity of the cues for ecological value (Linde-
mann-Matthies et al., 2010; Garbuzov et al., 2015; Hoyle et al., 2017a,
2017b). To adequately deliver these essential cues (Nassauer, 1995;
Hoyle et al., 2017a), a good level of forb species coexistence in the
longer term is required in meadow communities. It also requires in-
dividuals of forb species to have sufcient biomass to have signicant
oral visual impact, rather than be present as subordinates. The key
challenge is that species loss and dominance through intra and inter
specic competition is however naturally part of meadow ecological
processes (Grime, 2002) and in the absence of appropriate management
leads to vegetation (on productive soils) dominated by relatively few
species, often mostly grasses. On this basis Bjørn et al. (2016) have
proposed that species diverse forb dominated designed vegetation is an
illusion in the longer term. This raises the question that can the starting
point of meadow creation have any impact on longer term drift to
dominance by a few species?
* Corresponding author at: Horticulture and Landscape Department, Scotlands Rural College (SRUC), Kings Buildings, Edinburgh, EH9 3JG, UK.
E-mail addresses: (M. Jiang), j.d.hitchmough@shef (J.D. Hitchmough).
Contents lists available at ScienceDirect
Urban Forestry & Urban Greening
journal homepage:
Received 23 June 2021; Received in revised form 2 June 2022; Accepted 20 June 2022
Urban Forestry & Urban Greening 74 (2022) 127657
Sowing as a technique plays a critical role in meadow creation, as it is
the tool by which plant initial density in a designed community can be
manipulated, potentially assisting with managing competition from
invading weeds and achieving desired visual effects from an early stage
(Hitchmough and de la Fleur, 2006; Hitchmough, 2017b). It also enables
designed meadow communities to be established at large scales with
relatively low initial resource input (Dunnett and Hitchmough, 2004).
The sowing process does however require understanding of how effec-
tively the decisions made in this process, and specically the number of
seeds of each species placed into the germination environment, affect
the outcome in the short and intermediate term. Sowing density usually
refers to the quantity of seeds sown into a unit of area.
Sowing density directly affects the number of initial emergents of
each species and the dynamics of intra and inter species competition
from the outset. A high density of seedlings increases the competition
between leaves and shoots of individual plants for light especially where
water and nutrient are abundant (Grime, 2002; K¨
oppler and Hitch-
mough, 2015). Increasing sowing rate has been used in agriculture as a
non-herbicidal means to reduce weed establishment via rapid closure of
plant canopies in annual monocultures (Andrew and Storkey, 2017). A
higher sowing density can effectively compensate for the problem of low
viability seeds (James et al., 2011); and it can also create the desired
meadow effect more quickly (Stevenson et al., 1995; Hulvey and Aigner,
2014; Barr et al., 2017). In the longer term (i.e., after four years), Lubin
et al. (2019) found that doubling sowing density was still effective in
increasing the cover ratio of sown species to spontaneous species and
decreasing species diversity of the latter.
However, Nemec et al. (2013) found that the effects of sowing den-
sity were diminished after three years with no signicant effect on both
desired species and weed cover values. This was probably because that
this experimental site was highly productive. There will always be a
productivity threshold beyond which increasing sowing density ceases
to improve desired species establishment in the longer term (Stevenson
et al., 1995; Burton et al., 2006). Where conditions naturally support
species richness, i.e., where the soil is relatively unproductive or weed
seedling density low, increasing sowing density will also become less
effective (Stevenson et al., 1995; Scotton, 2019).
Del-Val and Crawley (2005) and Dickson and Busby (2009) sug-
gested grass competition from either spontaneous or sown grass seed-
lings is likely to override the original design mix, and suppresses the
performance of the forb community. Climates such as the UK favour
grass dominance especially on productive soil conditions (Pywell et al.,
2003; Walker et al., 2004). Grass species usually have advantages over
forbs in terms of abundant seedling recruited from soil seed banks
(Edwards and Crawley, 1999), high seedling survival (Hitchmough
et al., 2001; Jurado and Westoby, 2006; Ben-Hur and Kadmon, 2015),
rapid seedling growth rate (Campbell et al., 1991; Hitchmough et al.,
2001), tolerance of grazing or cutting (Pywell et al., 2003) and reduced
palatability to molluscs (Edwards and Crawley, 1999; Wilby and Brown,
2001; Del-Val and Crawley, 2005).
When grasses are left out of sowing mixes to try to slow down the
process of forb elimination, weed invasion tends to be problematic,
leading to a parallel decline in desired forb species richness and biomass
irrespective to forb sowing density (Dickson and Busby, 2009; Nemec
et al., 2013). Thus, there is a dilemma as to whether it is better to include
grasses in order to manage spontaneous weeds from the soil seed bank or
to leave grasses out and increase forb sowing density to try to better
compete with weeds. Most ndings from ecological studies, which
mostly work with existing vegetation in rural environments, reect a
condition in which scale and resources restrict the capacity to manage
site and sowing operations to maximise success from the beginning.
Aesthetic outcomes are also a low priority in these studies.
In urban landscape architecture works, which mostly aim to create
communities in a relatively smaller scale, there are more resources
available than ecological restorations in rural environments for site
preparation, design and management, increasing the likelihood of forb
dominated communities persisting in the longer term (Hitchmough,
2017a). The starting point can be more controlled through, for example,
stripping off topsoil or sowing into a low productivity mineral mulch
layer to greatly reduce the establishment of spontaneous weeds and
grass competition (Hitchmough et al., 2004; Hitchmough, 2017a). Plant
communities can be designed to utilise different canopy layers and
species composition to maximise cover and competitiveness with
invading grasses (Hitchmough, 2009; Hitchmough et al., 2017). Weed
management can be utilised at critical phases to reduce the development
of dominance by undesired species. Within the category of urban land-
scape design, sowing density, as a gradient, from very high to very low,
can be used as a tool to try to achieve the most desirable outcomes for
meadow communities given other environmental and management
factors. Greater levels of control may allow a relatively low density
sowing mix to be sustainable and yet also reduce competitive dominance
within the sown vegetation. However, this may delay the delivery of
visual benets which may potentially undermine social support for the
Previous urban meadow studies based in Shefeld, north England,
show that higher sowing densities can deliver aesthetic benet sooner
given sufcient longer-term management (Hitchmough and de la Fleur,
2006). Without this management, high sowing density often leads to a
dominance by a few competitive sown or spontaneous species (Hitch-
mough and Wagner, 2013; Hitchmough et al., 2017). Many studies in
urban landscape contexts have mostly involved tall productive forb-only
communities or those with minimal grass within the sowing mix. Much
less is known about the dynamics of forbs and grasses within urban
This study explored how two sowing densities of both grasses and
forbs under urban conditions affected longer term retention (i.e., over a
three-year experiment period) of forbs within sown communities of
Western Europe and Inner Mongolian species, many species of which are
naturally co-distributed in both of these geographical regions. The
theoretic position underpinning the research is the widely observed
ecological paradox that as seedling density increases post a disturbance
event, nite resource availability inevitably leads (particularly on pro-
ductive soils) to self-thinning, i.e. increased mortality (Morris and
Myerscough, 1991; Burton et al., 2006; Frances et al., 2010; Kli-
mek-Kopyra et al., 2020), and that this mortality is not evenly spread
across the community but is asymmetric, i.e. impacting least on the
species that have the best capacity to compete for the key resources and
most on the species that do not (Stevenson et al., 1995; Lawson et al.,
2004; Dickson and Busby, 2009; Jaksetic et al., 2017). As a result we
explored three core research questions in this study i) does increasing
sowing density inevitably accelerate sown grasses eliminating sown
forbs? ii) to what extent can initial sowing density inuence forb per-
formance in terms of forb abundance, richness and biomass? iii) which
subordinate forb species tend to gain advantages in terms of number of
seedlings and biomass in the higher sowing density and vice versa?
2. Methods
The eld experiment was conducted from May 2017 to August 2019,
three growing years, at Manor Top (533782′′N 14351′′W), Shefeld,
UK, on a west facing slope previously used for cultivation with a highly
productive clay loam topsoil. Soil nutrient analysis was not available for
the site, however as biomass harvesting was a key part of the research
methodology and provided a more meaningful measure of potential
productivity. Typical peak standing shoot biomass was in the region of
1200 g/m
, which corresponds to the upper levels possible in non-
wetland sites in both Inner Mongolia, China and Britain (Ni, 2004; Qi
et al., 2018). Highly productive soils are common in urban areas and the
site represented a worst-case scenario in terms of likely competitive
Ground preparation work took place in March 2017 to control
perennial rhizomatous grasses (herbicide applied twice) and cultivating
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
the ground. Experimental plots and the sand substate were installed in
May 2017, and the seed mixes were sown in the 23rd May 2017. Sharp
sand was used as a sowing mulch(see Hitchmough, 2017a) substrate
to restrict weed emergence from the soil seed bank and to achieve a high
percentage emergence and lower potential productivity in the immedi-
ate rooting environment. Sand is often used in urban meadow sowings in
practice for these reasons (Hitchmough, 2017a).
The eld experiment employed a fully randomised factorial design
involving a total of 96 experimental plots. The factorial design and
experimental procedures used were similar to that employed by other
workers exploring competitive interactions between forbs and grasses in
mixed communities, for example; Del Val and Crawley (2005) and
Dickson and Busby (2009). In this paper only the effect of sowing density
and relevant interactions, rather than the full set of experimental factors,
are discussed. Two sowing densities, calculated to achieve approxi-
mately 500 or 1000 emerged seedlings/m
were used within the 1×1 m
plots. These rates are relatively high but were used to ensure that at the
various ratios of grasses to forbs present in the study (1:9, 1:1 and 9:1),
there would be sufcient minimum numbers of seedlings of all species to
analyse. All plots were separated by 1 m wide (downslope) and 0.5 m
wide (across slope) weed mat covered cross paths.
Within the communities sown into each plot, there were three cat-
egories of forb canopy height (low; medium; tall) to test the signicance
of light competition and to create meadow communities that were likely
to be appealing to people in urban environments. Using the methodol-
ogy devised by Hitchmough (2017a) seed weight and seed emergence
data were used to calculate the amount of seed necessary per plot to
arrive at a target number of seedlings per m
(see Table 1). The ratio of
emerged seedlings designed into each layer (low: medium; tall) was
4:2:1. This ratio was to reduce the impact of taller species on shorter
species in terms of light competition and dominance. These approaches
are most effective in practice at ratios >20:10:1 (Hitchmough, 2017a).
These latter ratios could not be used in the experiment because on small
plots they require either excessively high overall sowing densities or
acceptance of the absence of some species.
The meadow species used in the study involved 15 species which are
currently distributed in meadows in both the UK and Inner Mongolia and
14 species only found in Inner Mongolia. The study was designed as two-
site research with the experiments replicated in the UK and Inner
Mongolia, however due to problems of site management in Inner
Mongolia data was not ultimately available from the latter. All forbs
were selected based on three criteria of ecological feasibility to the UK
environment (a history of cultivation), attractiveness for urban land-
scape (the appearance and robustness), and availability from commer-
cial suppliers (species selection procedure is shown in Fig. 1). Species
are listed in Table 1. Deschampsia cespitosa was used as the competitor
grass because of its wide distribution in both the UK and Inner Mongolia
and its tussock form and capacity to remain structurally intact post
owering. The viability to emerge and establish in cultivations under UK
climate conditions were tested in Hitchmough (2010, 2017a).
Species that could not be established from a late spring sowing
without winter chilling (Aconitum carmichaelii, Angelica sylvestris and
Stachys ofcinalis) were sown in seed trays and transplanted into the
plots according to designed seedling numbers in late November 2017.
Due to insufcient seedlings of Aconitum carmichaelii and Angelica syl-
vestris, seedling density of these two species were below the designed
seedling number.
2.1. Experiment management and data recording
As late spring sowing can lead to low species emergence and estab-
lishment due to high temperatures and moisture stress, hessian was
stretched over each plot to create approximately 50% shade. Plots were
initially irrigated every two days in the absence of rain.
The rst year of the experiment was used to establish the commu-
nities with data collection (not reported in this paper) commencing two
months after the rst seedling emergences in June 2017. To retain sown
community richness and achieve a relatively uniform starting point for
the longer term study, plots were mown (at 50 mm) approximately
every 10 days within summer to disadvantage the largest and fastest
seedling (mainly Achillea millefolium, Echinops ritro, Echinops sphaer-
ocephalus, Geranium pratense and Deschampsia cespitosa) and reduce early
dominance. This is a standard technique used in landscape practice (Du
Gard Pasley, 1990; Schmithals, and Kühn, 2014). Seedling numbers of
A. millefolium, E. ritro and E. sphaerocephalus were then thinned down by
hand to the designed seedling numbers. Biomass harvesting of both all
species and individual species was used as mean to assess competitive
behaviour of species in experimental communities, in line with other
research in this eld (e.g., Davies et al., 1999; Del Val and Crawley,
2005; Bjørn et al., 2019).
Plots were manually irrigated 3 times in summer 2018 (Met Ofce,
Table 1
Target numbers of forbs and grasses used in the experiment.
(a) Forb species
Target Seedlings/m
Species Low
Low canopy Shared Anemone sylvestris 14 29
Galium verum 14 29
Potentilla rupestris 14 29
Pulsatilla vulgaris 14 29
Veronica teucrium 14 29
Mongolian Dracocephalum
14 29
14 29
14 29
14 29
Veronica incana 14 29
Shared Achillea
7 14
7 14
Origanum vulgare 7 14
7 14
Stachys ofcinalis 7 14
Mongolian Campanula
7 14
7 14
Kalimeris incisa 7 14
7 14
7 14
Tall canopy Shared Echinops ritro 4 7
Geranium pratense 4 7
4 7
4 7
Veronica longifolia 4 7
Mongolian Aconitum
5 9
Angelica sylvestris 5 9
5 9
5 9
(b) Grass species
250 500
Data forCampanula glomerata not included in results due to poor emergence
(close to zero) in experimental trials
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
2018), which was extraordinarily hot and dry. Tall ruderal weed seed-
lings were removed prior to the commencement of the study in spring
2018. Meadow cutback to about 20 mm above the ground was under-
taken prior to the commencement of growth in February 2018. The same
practice was applied in August and the cut material was sorted, dried,
and weighed to generate biomass data to reect standard meadow
management in practice.
Data was collected from an 800×800 mm permanent quadrat placed
in the centre of each plot approximately one year post the rst seedling
emergents. In April 2018 and 2019, forb seedling numbers present in
each plot were counted for each species. Forb and grass cover values in
each plot were measured in May in both years. Biomass for each species
was collected in August 2018 and 2019. The biomass samples were
placed in a drying cabinet at 75 for 24 h and left for another 24 h
before weighing to achieve some consistency between weighing as
weight increased quickly in the rst few hours when moisture from the
air was absorbed.
2.2. Statistical analysis
The statistical tests were undertaken with SPSS version 26. Gener-
alized Estimating Equations (the GEEs) were applied to build the 2-level
factorial models (i.e. the model tests with all possible 2-factor interac-
tive combination), which included all four designed factors and ‘Year(i.
e. to represent the data difference between 2018 and 2019 as Within-
subject Variable). Sequential Sidak correction was applied for the
comparison of estimated means to obtain the signicance levels.
The accumulated biomass data and cover value data in each plot
were analysed with the Linear model type within the GEEs, where the
tests were valid regarding the standardised residuals and the data size.
Fig. 1. Forb species selection procedure (E-, 2016; Hauck and Solongo, 2010; Liu, 2010; Liu et al., 2015).
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
Seedling number data were treated as ‘countsand were analysed with
Poisson Loglinear type models.
To meet the assumption of data distribution and validate the tests,
other model types within the GEEs were applied and data were trans-
formed to optimise normality. Due to the different intrinsic size of
species there was a need to standardize the scores. To test the difference
of forb biomass between the treatments or years, the mean and standard
deviation for each species were calculated as a best guess at the
normative behaviour of the species. To standardize the raw scores, the
following equation was used to obtain the z-score;.
=(Sample biomass – Mean
biomass/ species
)/ Standard Deviation
This reduced the direct effect of the factor species but still allowed
assessment of interaction with respect to species.
3. Results
3.1. Effect of sowing density on forb seedling numbers, forb and grass
biomass and cover values in 2018 and 2019
As shown in Fig. 2, high sowing density led to signicantly more forb
seedlings in both April 2018 (p =0.000) and 2019 (p =0.000) but did
not double forb seedling number. The difference of forb seedling number
between low and high sowing density had diminished by 2019.
In terms of forb biomass, the high sowing density did not double this
but did lead to signicantly more forb biomass (p =0.015) in August
2018 (Fig. 3). There was no signicant difference by 2019 (p =0.710).
This pattern was also reected in coverage measurements in April for
these two years (Fig. 4). There was a signicant decrease of forb seedling
numbers across this time period in both sowing densities (p =0.000).
Mortality was higher in the high sowing density (47.4%) than the low
sowing density (40.6%) between 2018 and 2019. In 2018, High density
sowings of Deschampsia cespitosa supported signicantly more forb
biomass (1003.40 g), than the low density grass treatment (666.29 g,
In terms of grass biomass, the low sowing density treatment had
higher but not signicantly higher grass biomass in 2018 (p =0.401). In
2019 grass biomass became signicantly (p =0.002) greater in the low
sowing density treatment, suggesting the forb dominants were out-
competing the grasses. Forb biomass was signicantly higher than grass
biomass in both sowing densities and both years (p =0.000).
The higher sowing density led to a signicantly higher forb
(p =0.000) cover value in 2018 (Fig. 4), but this did not increase grass
coverage (p =0.960). In 2019, despite forb coverage appearing higher
in the high sowing density while grass coverage was higher in the low
sowing density treatment, these treatments had no signicant effect on
the cover values.
3.2. Effect on dominant forbs (Achillea millefolium and Echinops
sphaerocephalus) and subordinate forbs on biomass in 2018 and 2019
Biomass distribution within the community was highly asymmetric
for individual species. This pattern is normal in many eld experiments
and real-life projects. The dominant forbs were mainly A. millefolium in
2018 and then both A. millefolium and E. sphaerocephalus in 2019. The
rest of 26 forb species (Campanula glomerata is excluded due to no valid
data collected) refer to the subordinates in both 2018 and 2019.
High sowing density led to signicantly more biomass of
A. millefolium in August 2018 (p =0.004) but did not signicantly in-
crease A. millefolium biomass in 2019 (p =0.056), as shown in Fig. 5.
The biomass of E. sphaerocephalus was not signicantly different be-
tween the two density treatments in both 2018 (p =0.997) and in 2019
(p =0.171) despite low sowing density leading to greater
E. sphaerocephalus biomass (467.44 g) than the high sowing density
treatment (358.76 g) in 2019. Subordinate biomass showed a similar
Fig. 2. Effect of sowing density on forb seedling number/ plot in April 2018
and April 2019 (*p 0.05; **p 0.01; ***p 0.000 and ns=not signicant.
Error bar =2 Standard Errors).
Fig. 3. Effect of sowing density on forb and grass biomass/ plot in August 2018
and August 2019 (*p 0.05; **p 0.01; ***p 0.000 and ns=not signicant.
Error bar =2 Standard Errors).
Fig. 4. Effect of sowing density (on forb and grass cover values in May 2018
and May 2019 (*p 0.05; **p 0.01; ***p 0.000 and ns=not signicant.
Error bar =2 Standard Errors).
Fig. 5. Effect of sowing density (low and high density) on dominant and sub-
ordinate forb biomass/ plot in August 2018 and August 2019 (*p 0.05;
**p 0.01; ***p 0.000 and ns=not signicant. Error bar =2 Stan-
dard Errors).
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
pattern to E. sphaerocephalus (p =0.622 in 2018 and p =0.139 in 2019)
and showed more biomass in the low sowing density treatment
(104.88 g compared with 85.01 g in the high sowing density) despite
not being signicant (Fig. 5).
High sowing density led to signicantly more forb biomass in 2018,
mostly because of A. millefolium, which comprised most of this forb
biomass. Biomass change between the two years, was signicantly
different for E. sphaerocephalus (p =0.000 for both densities). The sub-
ordinate biomass was also signicantly different (p =0.000) for both
low density and high density (p =0.019). Table 2.
3.3. Effect of sowing density on forb seedling richness in 2018 and 2019
In 2018, 18 out of 23 subordinate forb species (data for Platycodon
gradiorum, Scutellaria baicalensis in both years and Aconitum carmi-
chaelii in 2018 was not validated for use in the statistical model) had
signicantly higher seedling numbers in high sowing density treatments
(Table 3). However, in 2019, this went down to 9 out of 24. The effect of
increasing sowing density to increase numbers of seedlings was reduced
with the passage of time in most of the species especially the lower
canopy forb species. Eight low canopy subordinates had more seedlings
in high density in 2018 and only 2 species retained this advantage in
2019 (Potentilla rupestris, p=0.001 and Veronica teucrium, p=0.021).
Overall, six subordinate forb species had signicantly higher
numbers of seedling in the high sowing density treatment in both 2018
and 2019; Potentilla rupestris (p =0.003 in 2018, p =0.001 in 2019),
Origanum vulgare (p =0.000, p =0.007) Polemonium caeruleum
(p =0.000, p =0.003), Kalimeris incisa (p =0.000, p =0.000), Stachys
ofcinalis (p =0.000, p =0.000) and Thalictrum aquilegifolium
(p =0.001, p =0.000).
3.4. Effect on subordinate forb biomass per species in 2018 and 2019
Increasing sowing density generally had limited capacity to increase
subordinate forb biomass for most of the species in both years. As shown
in Table 4, Stachys ofcinalis was the only subordinate forb that had
signicantly more biomass in the high sowing density treatment in 2018
(p =0.002). Low sowing density increased the biomass of Patrinia sca-
biosifolia (p =0.032) in 2018, Origanum vulgare (p =0.006) in 2019 and
Dracocephalum rupestre in both years (p =0.018 in 2018 and p =0.005
in 2019). Despite no statistical difference (p =0.106), P. rupestris pro-
duced far more biomass in low sowing density (6.17 g) comparing with
high sowing density (2.23 g).
4. Discussion
4.1. Does increasing sowing density inevitably accelerate sown grasses
eliminating forbs?
Unlike many previous studies (Pywell et al., 2003; Del-Val and
Crawley, 2005; Silvertown et al., 2006), the higher sowing rate did not
lead to a larger grass biomass that dominated the community. Grass
biomass diminished as a percentage of forb biomass from 2018 on, with
the highest grass biomass associated with the low sowing density in
2019 (p =0.002). The grass species used in this study generally showed
lower competitiveness than the most competitive forbs that dominated
the forb biomass. Deschampsia cespitosa is less competitive in terms of
relative growth rate than the ubiquitous weedy grasses that invade
designed meadow, such as Arrhenatherum elatius, Holcus lanatus and
Lolium perenne (Del-Val and Crawley, 2005; Hitchmough et al., 2008).
These common highly competitive grasses were eliminated from the site
as part of the initial site preparation protocols and the nutrient and
moisture stress created by the 150 mm sand mulch treatment, plus the
high density sown community inhibited subsequent re-establishment
over the three years of this study.
Biomass production of D. cespitosa (a species of moist to wet envi-
ronments) appeared to have been reduced by the sowing mulch relative
to the most dominant forb, A. millefolium, a relatively aggressive, colo-
nising species (Burton et al., 2006; Bjørn et al., 2019).
This study provides an interesting meadow design model; forbs were
not eliminated by grass competition whilst the shade tolerant
D. cespitosa was able to persist under the summer canopy of the taller
forbs and deliver functional benets, for example early emergence, and
quick recovery after late summer cutback to reduce soil exposure. Both
sowing density treatments had an extremely low biomass of weed
species (1.1% and 0.5% of the sown biomass in the low and high sowing
treatments in 2018; 0.8% and 0.7% in 2019) after three growing seasons
with no weed removal. In the longer term, weedy, more competitive
grasses may, in the absence of management, gradually become more
abundant in the created meadows. Thus, increasing sowing density does
not always accelerate sown grass eliminating forbs, it depends on the
relative competitiveness of the grass species in relation to the most
dominant forb species.
4.2. To what extent, can initial sowing density inuence forb performance
in terms of forb abundance, richness and biomass?
In conventional ecological studies, diversity is seen as the key mea-
sure of success in meadows creation, however in urban areas, the
number of forb seedlings present and their biomass is also important and
potentially underpins oral performance as much or more than richness
does (Hoyle et al., 2018).
As in many previous studies (Dickson and Busby, 2009; Hitchmough
et al., 2017; Lubin et al., 2019), increasing forb sowing density increased
the number of forb seedlings present for at least for three years. Having
more forb seedlings often leads to more attractive initial appearance and
offers greater competition to invading species from an early stage, and
provides at least an opportunity for a more sustainable end point
(Hitchmough and de la Fleur, 2006; Lauenroth and Adler, 2008;
Hitchmough, 2017a). Even if only temporary this is worth having
Table 2
Effect of sowing density on overall results in 2018 and 2019 (SE =Standard Error of Mean).
2018 2019
Low density High density P value Low density High density P value
Mean SE Mean SE Mean SE Mean SE
Forb seedling present number 180.04 18.64 269.71 25.71 0.000 * ** 107.35 8.98 141.88 11.54 0.000 * **
Forb biomass (g) 604.34 57.46 803.63 80.64 0.015 * 1113.59 71.84 1081.95 64.27 0.710 ns
Grass biomass (g) 320.70 39.91 286.43 41.47 0.401 ns 150.21 19.39 90.04 13.96 0.002 * *
Forb cover value (%) 33.56 2.94 46.08 4.32 0.000 * ** 67.02 3.75 73.38 3.54 0.098 ns
Grass cover value (%) 34.79 4.67 35.02 4.93 0.960 ns 29.58 3.88 23.67 3.61 0.139 ns
Bare ground cover value (%) 31.65 3.35 18.90 2.77 0.000 * ** 3.63 0.56 2.96 0.51 0.331 ns
Achillea millefolium biomass (g) 426.06 47.77 620.49 68.48 0.004 * * 541.27 44.50 638.18 44.00 0.056 ns
Echinops sphaerocephalus biomass (g) 112.56 14.39 112.48 15.00 0.997 ns 467.44 67.95 358.76 61.11 0.171 ns
Subordinate forb biomass (g) 65.73 9.38 70.66 9.67 0.622 ns 104.88 12.75 85.01 10.14 0.139 ns
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
especially in very politically contested urban environments, as a sign of
initial success. Whether this opportunity leads to longer term success or
not, depends on site productivity and management.
Forb seedling number increased, but not linearly with sowing den-
sity; and the effect of higher sowing density diminished with the time
from the initial sowing because of greater competition leading to greater
self-thinning (Yoda et al., 1963). Whilst absolute forb seedling
numbers in the high sowing density were still higher in the third year
(2019, p =0.000) than in the low density, the numbers were likely to
become similar in subsequent years, as this is ultimately determined by
competition between individuals and species, and herbivory.
In this study, competition for light is believed to be the major factor
behind decline in the number of forb seedlings, with many of these being
low growing shade intolerant species. ‘Sunscan PAR measurements
indicated lower solar energy on the soil surface in high sowing density
(2.04%, ground solar radiation level/ ambient, watts/m
) than low
sowing density plots (2.71%). Earlier in the year cutting and removal of
biomass from the meadow or selectively thinning of competitive species
would probably have slowed down the loss of forb seedlings.
Mortality of forb seedlings was more marked in this study than in
Hitchmough et al. (2008). This was probably because in this study the
sowing densities and subsequent emergence were higher, and a greater
proportion of the forb species were more sensitive to shade. Although
species diversity gradually declined in both sowing density treatments in
both 2018 and 2019 (Shannon Wiener index: in 2018; 3.04 in low
density, 3.08 in high density; in 2019, 2.94 in low density and 2.97 in
high density) differences between densities remained relatively small.
Research in rural herbaceous communities in British Colombia,
shows that doubling sowing density shortens the time taken to reach a
forb biomass ceiling (Burton et al., 2006). Higher cover values in spring
reduce potential weed invasion and deliver visual evidence of ‘ecolog-
ical value that is of both social and ecological importance (Bergelson
et al., 1993; Hoyle et al., 2017a). Increasing sowing density is unlikely to
affect grassland community biomass in the long term as it is not possible
to override ecological processes such as self-thinning, by adding more
seeds. The increase in forb biomass with higher sowing density in this
study was primarily due to the increase of A. millefolium, the most
dominant species. This suggests that increasing sowing density exacer-
bates asymmetric competition with dominance effects occurring sooner
and with subordinate forbs being suppressed or even eliminated in a
short time period. The greater biomass of A. millefolium in the high
sowing density was likely to be the main cause of low grass and subor-
dinate forb biomass (Dwyer, 1958). In this study, A. millefolium funda-
mentally adopted the ‘weedy grass role and suppressed the
subordinates through similar mechanisms to grass competition.
Subordinate forb biomass was greater in the low sowing density in
2019 despite the difference being non-signicant (p =0.139). This was
presumably because there were fewer A. millefolium to intercept light
resources in this treatment. Stevenson et al. (1995) recommended a
lower sowing density to be used in unproductive conditions and where
competitive perennial grasses are both less productive and present at
lower density. If too low a seed density is sown, the impact maybe too
limited in closely viewed urban landscapes.
The substantial basal foliage and leafy stems of E. sphaerocephalus
allowed it to take the role of second dominant. Since light competition is
the major process by which dominants suppress subordinates in the
Table 3
Effect of sowing density on seedling number of each forb species in 2018 and 2019 (SE =Standard Error of Mean; na =not applicable).
2018 2019
Low density High density P value Low density High density P value
Mean SE Mean SE Mean SE Mean SE
Low canopy
Anemone sylvestris 6.63 1.17 9.63 1.36 0.003 * * 4.92 0.88 5.00 0.67 0.325 ns
Galium verum 7.44 1.15 11.17 1.64 0.012 * 6.02 0.82 7.75 1.01 0.061 ns
Potentilla rupestris 8.19 1.02 11.19 1.22 0.003 * * 5.81 0.60 8.15 0.82 0.001 * *
Pulsatilla vulgaris 5.27 0.86 9.85 1.53 0.001 * * 2.46 0.47 2.90 0.68 0.363 ns
Veronica teucrium 7.73 1.17 9.65 1.40 0.134 ns 5.33 0.59 6.83 1.02 0.492 ns
Dracocephalum rupestre 13.21 1.58 19.67 2.24 0.000 * ** 4.96 0.80 4.75 0.77 0.393 ns
Dracocephalum ruychiana 9.83 1.54 13.44 2.18 0.017 * 2.42 0.53 2.42 0.47 0.362 ns
Thalictrum petaloideum 8.75 1.08 14.04 1.38 0.000 * ** 4.98 0.59 6.94 0.86 0.021 *
Thermopsis lanceolata 5.98 0.89 10.79 1.60 0.000 * ** 1.60 0.32 2.19 0.41 0.059 ns
Veronica incana 0.33 0.14 0.54 0.29 0.984 ns 0.00 0.00 0.00 0.00 0.844 ns
Medium canopy
Achillea millefolium 4.27 0.41 8.96 0.81 0.000 * ** 4.27 0.41 8.96 0.81 0.000 * **
Origanum vulgare 21.79 2.15 32.42 3.38 0.000 * ** 13.60 1.15 17.73 1.38 0.007 * *
Polemonium caeruleum 5.54 0.72 8.98 1.05 0.000 * ** 2.71 0.45 4.23 0.84 0.003 * *
Stachys ofcinalis 4.33 0.42 9.00 0.83 0.000 * ** 2.90 0.34 5.38 0.60 0.000 * **
Campanula punctata 21.85 2.78 24.65 2.98 0.128 ns 14.27 1.85 14.60 1.90 0.802 ns
Delphinium grandiorum 11.15 1.58 14.67 1.94 0.002 * * 7.67 1.12 9.35 1.46 0.082 ns
Kalimeris incisa 6.52 0.81 14.19 1.58 0.000 * ** 3.90 0.53 6.73 0.90 0.000 * **
Platycodon grandiorum 3.33 0.70 5.33 0.93 na 0.00 0.00 0.06 0.06 na
Scutellaria baicalensis 2.15 0.48 2.29 0.56 na 0.23 0.09 0.15 0.06 na
Tall canopy
Echinops ritro 1.04 0.20 1.92 0.40 0.060 ns 0.48 0.13 1.08 0.28 0.051 ns
Geranium pratense 5.17 0.51 7.67 0.73 0.004 * * 5.56 0.61 7.83 0.82 0.039 *
Sanguisorba ofcinalis 1.52 0.23 3.15 0.49 0.000 * ** 1.60 0.23 2.40 0.33 0.040 *
Thalictrum aquilegifolium 5.92 0.68 9.21 1.12 0.001 * * 4.67 0.47 7.38 0.75 0.000 * **
Veronica longifolia 2.44 0.46 3.27 0.60 0.210 ns 2.23 0.43 2.46 0.50 0.830 ns
Aconitum carmichaelii 1.00 0.00 1.00 0.00 na 0.83 0.05 0.83 0.05 0.980 ns
Angelica sylvestris 2.00 0.12 3.00 0.24 0.000 * ** 0.75 0.12 0.50 0.09 0.119 ns
Echinops sphaerocephalus 3.75 0.27 5.85 0.41 0.000 * ** 2.85 0.31 4.77 0.49 0.000 * **
Patrinia scabiosifolia 2.92 0.36 4.21 0.51 0.001 * * 0.33 0.10 0.52 0.32 0.271 ns
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
seedling stage in urban meadows (K¨
oppler and Hitchmough, 2015)
increasing seedling density of the subordinate species is probably less
effective than reducing the density of sown dominants such as
A. millefolium and E. sphaerocephalus where weed biomass is low. Dick-
son and Busby (2009) argued that spatial separation is an effective way
to encourage the growth of less competitive species in a community but
this is not feasible in meadow-like vegetation.
4.3. Which subordinate forb species tend to gain advantages in terms of
number of seedlings and biomass in the higher sowing density?
For most of subordinate forbs, increasing sowing density increases
the number of seedlings but was unlikely to enhance their biomass in
2018 and 2019 in this study. Hitchmough et al. (2017) found the same
patterns. Morphological abilities to access light resources through early
emergence or taller leafy foliage are essential to survive the more intense
competition from dominants. Tall, and native species with presumed
superior tness were more likely to maintain higher numbers of sur-
viving seedling in the higher sowing density in the longer term.
Increasing sowing density was least likely to enhance the number of
survivals of low canopy forb species.
Although the competition of grasses and the dominant forb
A. millefolium was greater at the higher sowing density, this treatment
retained most of the subordinate forb species (18 out of 23) in 2018.
Although a few species such as Veronica teucrium and Campanula punc-
tata did not show a statistically signicant difference, seedling numbers
were initially higher at the high sowing density. Benecial effects of
sowing density on subordinate seedling number were greatly reduced in
2019. Only 9 species still had signicantly more seedling numbers in the
high-density treatments. Seven species, out of these 9, were native to the
UK suggesting that species with native distributions were potentially
more adapted and more persistent in the UK climate. This may be also
because the shared species have a wider geographically distribution
implying adaptiveness to a more generalist habitat conditions, which
tend to make them more able to survive in semi-natural grasslands
(Pywell et al., 2003). More importantly, most native species had the
advantages of greater light competitiveness from the seedling stage due
to earlier emergence (P. rupestris and P. caeruleum), rapid seedling
growth (O. vulgare) or advantages in architecture, for example, long
petioles (G. pratense), tall leafy stems (Sanguisorba ofcinalis), or clam-
bering stems as in Galium verum.
Sowing density was largely ineffective as a means of increasing
subordinate forb biomass. In 2018, 11 species increased their biomass in
the high sowing density while 13 species decreased. In 2019, 10 species
increased their biomass and 14 species declined in terms of biomass,
with Thalictrum petaloideum and V. incana poorly represented in both
years. Stachys ofcinalis was the only species that had a signicant in-
crease of biomass among the subordinates. Due to seed dormancy
problems, Stachys ofcinalis was one of the species established in the
experiment by planting, so they were larger than some other subordinate
forbs in the plots, and this may have helped them to cope with shading
stress generated by dominant species. The biomass of this species almost
increased linearly within the high density treatment in both years.
Transplanting did not improve the growth of all species established in
this way. For example, Angelica sylvestris remained very small (0.02 and
0.03 g in low and high sowing density treatments). This species is largely
associated with wet sites (8 on the Ellenberg scale for moisture) and the
experiment was probably too dry for this species to establish. Lack of
Table 4
Effect of sowing density on biomass of each subordinate forb species in 2018 and 2019 (SE =Standard Error of Mean).
2018 2019
Low density High density P value Low density High density P value
Mean SE Mean SE Mean SE Mean SE
Low canopy
Anemone sylvestris 0.08 0.02 0.06 0.02 0.536 ns 0.06 0.04 0.04 0.02 0.481 ns
Galium verum 2.34 0.73 3.69 1.20 0.243 ns 6.76 1.34 8.52 2.09 0.391 ns
Potentilla rupestris 3.70 1.09 2.35 0.46 0.190 ns 6.17 2.61 2.23 0.58 0.106 ns
Pulsatilla vulgaris 0.18 0.06 0.13 0.04 0.429 ns 0.03 0.02 0.05 0.03 0.672 ns
Veronica teucrium 1.21 0.27 1.42 0.49 0.664 ns 2.32 0.66 1.64 0.48 0.305 ns
Dracocephalum rupestre 2.94 0.49 1.84 0.31 0.018 * 0.16 0.05 0.03 0.01 0.005 * *
Dracocephalum ruychiana 1.07 0.23 0.66 0.13 0.066 ns 0.43 0.22 0.09 0.05 0.091 ns
Thalictrum petaloideum 0.00 0.00 0.00 0.00 na 0.00 0.00 0.00 0.00 0.140 ns
Thermopsis lanceolata 1.13 0.40 1.22 0.27 0.839 ns 0.30 0.14 0.60 0.22 0.214 ns
Veronica incana 0.00 0.00 0.00 0.00 0.400 ns 0.00 0.00 0.00 0.00 1.000 ns
Medium canopy
Origanum vulgare 20.15 3.59 15.84 3.10 0.257 ns 52.53 7.19 31.63 4.89 0.006 * *
Polemonium caeruleum 0.56 0.26 2.01 0.86 0.056 ns 0.77 0.24 1.66 0.65 0.142 ns
Stachys ofcinalis 0.26 0.05 0.54 0.10 0.002 * * 0.76 0.16 1.57 0.51 0.102 ns
Campanula punctata 2.21 0.58 1.64 0.56 0.415 ns 1.24 0.39 0.84 0.61 0.545 ns
Delphinium grandiorum 2.26 0.62 2.59 0.85 0.713 ns 2.01 0.73 2.18 0.97 0.875 ns
Kalimeris incisa 15.66 3.31 22.33 4.31 0.111 ns 10.85 3.34 10.21 2.98 0.866 ns
Platycodon grandiorum 0.10 0.03 0.07 0.01 0.315 ns 0.02 0.01 0.00 0.00 0.068 ns
Scutellaria baicalensis 0.73 0.26 0.50 0.17 0.392 ns 0.10 0.04 0.08 0.05 0.609 ns
Tall canopy
Echinops ritro 2.71 0.85 4.84 1.54 0.199 ns 2.73 1.05 3.71 1.47 0.564 ns
Geranium pratense 3.66 0.63 3.48 0.82 0.847 ns 11.87 2.14 13.65 3.77 0.636 ns
Sanguisorba ofcinalis 2.48 0.62 2.90 0.92 0.678 ns 3.55 0.96 4.36 1.25 0.575 ns
Thalictrum aquilegifolium 0.08 0.02 0.07 0.02 0.674 ns 0.07 0.03 0.02 0.01 0.077 ns
Veronica longifolia 1.14 0.46 1.87 0.77 0.369 ns 1.53 0.48 1.04 0.41 0.402 ns
Aconitum carmichaelii 0.29 0.05 0.22 0.03 0.231 ns 0.42 0.10 0.24 0.05 0.087 ns
Angelica sylvestris 0.02 0.01 0.03 0.02 0.741 ns 0.08 0.03 0.55 0.52 0.329 ns
Patrinia scabiosifolia 0.78 0.20 0.37 0.10 0.032 * 0.10 0.04 0.06 0.02 0.363 ns
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
ecological adaptiveness to the community appears more important than
means of establishment.
There was evidence that low sowing density could benet sub-
ordinates that have poor capacity to compete for light. Dracocephlum
rupestre and Patrinia scabiosifolia showed a signicantly negative
response to the doubling of sowing density. Where competitive pressure
was lower, they produced more biomass in 2018. Dracocephlum rupestre
has short basal foliage and relatively slow growth placing it at a disad-
vantage when competing for light in taller vegetation. Although differ-
ences were not statistically signicant, species such as P. rupestris or
C. punctata which also have short foliage also showed a negative
response to increased sowing density. Patrinia scabiosifolia is also a late
emerging species, forcing it to compete for light with already actively
growing species, and hence was sensitive to high sowing density.
Enhanced ability to access light resources improves the likelihood of
subordinate species survival but does not increase their biomass under
the competition with more dominant species. This is an important
nding for how meadows might be perceived by the public in practice,
enhance survival is unlikely to be perceived positively if those seedlings
are too small to be ower or even to be perceived as being present.
5. Conclusion
Sowing density is one of the relatively few levers available to
practitioners to try to design meadow communities with a specic
preferred initial composition. To practitioners with a horticultural world
view, it seems intuitive that having more plants of a desired species post
sowing is a good thing, and likely to be a positive in the future devel-
opment of the community. The actual benets of higher sowing density
are mostly associated with how the developing meadow community is
perceived in the rst and second year. More seedlings resulting from
higher sowing densities confer a sense of success, and are likely to result
in a orally enhanced display in the second year.
Higher sowing densities inevitably result in increased intra and
interspecic competition leading to more rapid onset of dominance from
within the sown cohort, and a gradual loss of species diversity, and the
loss of the slowest growing and most shade intolerant subordinate spe-
cies in particular. Sown species with tall leafy stems, or other means of
competing for light were best able to persist at high sowing densities.
These effects became evident particularly rapidly in this study
because both sowing rates were relatively high (in order to reduce
missing values for subordinate species) with mean established densities
of forb seedlings of 180 (low sowing rate) and 270 per m
(high sowing
rate) in the second year (2018), a decline from approximately 500 and
1000 respectively in the rst growing season. The direction of travel for
forb density was clearly to decline more rapidly at higher sowing den-
sities. The increase in observed biomass of some subordinate species at
low sowing density took place under conditions in which sand mulching
essentially eliminated weed emergence and competition from the soil
seed bank. Where these mulches are absent low density sowings will
generally experience greater competition from weeds emerging from the
soil seed bank. This may lead to low sowing density meadows having
low sown forb densities than those achieved with higher sowing den-
sities. The effect of sowing density is hence highly contingent on local
conditions and practices. The most appropriate designed sowing den-
sities need to be based on assessment of these local factors, in order to
balance positives and negative outcomes as best as is possible for specic
site conditions.
CRediT authorship contribution statement
Mingyu Jiang: Conceptualization, Methodology, Software, Data
curation, Visualization, Investigation, Writing original draft. James D
Hitchmough: Supervision, Writing review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Andrew, I.K.S., Storkey, J., 2017. Using simulation models to investigate the cumulative
effects of sowing rate, sowing date and cultivar choice on weed competition. Crop
Barr, S., Jonas, J.L., Paschke, M.W., 2017. Optimizing seed mixture diversity and seeding
rates for grassland restoration. Restor. Ecol.
Ben-Hur, E., Kadmon, R., 2015. An experimental test of the relationship between seed
size and competitive ability in annual plants. Oikos 124 (10), 13461353. https://
Bergelson, J., Newman, J.A., Floresroux, E.M., 1993. Rates of weed spread in spatially
heterogeneous environments. Ecology.
Bjørn, M.C., et al., 2019. Increasing local biodiversity in urban environments: community
development in semi-natural species-rich forb vegetation. Landsc. Urban Plan.
Bjørn, M.C., Weiner, J., Ørgaard, M., 2016. Is colourful self-sustaining forb vegetation
mere fantasy? Urban For. Urban Green.
Burton, C.M., et al., 2006. Determining the optimal sowing density for a mixture of
native plants used to revegetate degraded ecosystems. Restor. Ecol.
Campbell, B.D., Grime, J.P., Mackey, J.M.L., 1991. A trade-off between scale and
precision in resource foraging. Oecologia 87 (4), 532538.
Davies, A., Dunnett, N., Kendle, T., 1999. The importance of transplant size and gap
width in the botanical enrichment of species-poor grasslands in britain. Restor. Ecol.
7 (3), 271280.
Del-Val, E., Crawley, M.J., 2005. What limits herb biomass in grasslands: Competition or
herbivory? Oecologia 142 (2), 202211.
Dickson, T.L., Busby, W.H., 2009. Forb species establishment increases with decreased
grass seeding density and with increased forb seeding density in a Northeast Kansas,
U.S.A., experimental prairie restoration. Restor. Ecol. 17 (5), 597605. https://doi.
Du Gard Pasley, A., 1990. ‘Herbaceous Plants and Bulbs. In: Clouston, B. (Ed.),
Landscape Design with Plants, second ed. Butterworth Architecture, Oxford,
pp. 8385.
Dunnett, N., Hitchmough, J., 2004. The dynamic landscape: Design, ecology and
management of naturalistic urban planting, The Dynamic Landscape: Design. Ecol.
Manag. Nat. Urban Plant.
Dwyer, D.D., 1958. Competition between Forbs and Grasses. J. Range Manag. https://
Edwards, G.R., Crawley, M.J., 1999. Herbivores, seed banks and seedling recruitment in
mesic grassland. J. Ecol. 87 (3), 423435.
E- (2016) Home. [online] Available at: <http://www.e
>[Accessed 2 November 2016].
Garbuzov, M., Fensome, K.A., Ratnieks, F.L.W., 2015. Public approval plus more wildlife:
twin benets of reduced mowing of amenity grass in a suburban public park in
Saltdean, UK. Insect Conserv. Divers.
Grime, P., 2002. Plant strategies, vegetation processes and ecosystem properties, second
ed. John Wiley & Sons Ltd, Chichester.
Hauck, M., Solongo, Z., 2010. Flowers of Mongolia. R. Biermann, Telgte.
Hicks, D.M., et al., 2016. Food for pollinators: quantifying the nectar and pollen
resources of urban ower meadows. PLoS ONE.
Hitchmough, J., 2009. Diversication of grassland in urban greenspace with planted,
nursery-grown forbs. J. Landsc. Archit.
Hitchmough, J. (2010) Characterising eld emergence in cultivated herbaceous vegetation.
Hitchmough, J., Dunnett, N., 2004. Introduction to naturalistic planting in urban
landscapes. Dyn. Landsc. Des. Ecol. Manag. Nat. Urban Plant.
Hitchmough, J., de la Fleur, M., 2006. Establishing North American prairie vegetation in
urban parks in northern England: Effect of management and soil type on long-term
community development. Landsc. Urban Plan.
Hitchmough, J., Wagner, M., 2013. The dynamics of designed plant communities of
rosette forming forbs for use in supra-urban drainage swales. Landsc. Urban Plan.
Hitchmough, J., Kendle, T., Paraskevopoulou, A.T., 2001. Seedling emergence, survival
and initial growth of forbs and grasses native to Britain and central/southern Europe
in low productivity urban wastesubstrates. Urban Ecosyst.
Hitchmough, J., de la Fleur, M., Findlay, C., 2004. Establishing North American prairie
vegetation in urban parks in northern England: part 1. Effect of sowing season,
sowing rate and soil type. Landsc. Urban Plan.
M. Jiang and J.D. Hitchmough
Urban Forestry & Urban Greening 74 (2022) 127657
Hitchmough, J., Paraskevopoulou, A., Dunnett, N., 2008. Inuence of grass suppression
and sowing rate on the establishment and persistence of forb dominated urban
meadows. Urban Ecosyst.
Hitchmough, J., Wagner, M., Ahmad, H., 2017. Extended owering and high weed
resistance within two layer designed perennial prairie-meadow vegetation. Urban
For. Urban Green.
Hitchmough, J.D., 2017a. Sowing Beauty: Designing Flowering Meadows From Seed,
rst ed. Timber Press, Portland.
Hitchmough, J.D., 2017b. The plant community: a model for horticultural thought and
practice in the 21st century? Acta Hortic.
Hoyle, H., et al., 2018. Plant species or ower colour diversity? Identifying the drivers of
public and invertebrate response to designed annual meadows. Landsc. Urban Plan.
Hoyle, H., Hitchmough, J., Jorgensen, A., 2017a. All about the wow factor? The
relationships between aesthetics, restorative effect and perceived biodiversity in
designed urban planting. Landsc. Urban Plan.
Hoyle, H., Hitchmough, J., Jorgensen, A., 2017b. Attractive, climate-adapted and
sustainable? Public perception of non-native planting in the designed urban
landscape. Landsc. Urban Plan.
Hulvey, K.B., Aigner, P.A., 2014. Using lter-based community assembly models to
improve restoration outcomes. J. Appl. Ecol.
Jaksetic, N., Foster, B., Bever, J., Schwarting, J., Alexander, H., 2017. Sowing density
effects and patterns of colonization in a prairie restoration. Restor. Ecol. 26 (2),
James, J.J., Svejcar, T.J., Rinella, M.J., 2011. Demographic processes limiting seedling
recruitment in arid grassland restoration. J. Appl. Ecol.
Jiang, Y., Yuan, T., 2017. Public perceptions and preferences for wildower meadows in
Beijing, China. Urban For. Urban Green.
Jurado, E., Westoby, M., 2006. Seedling growth in relation to seed size among species of
arid Australia. J. Ecol. 80 (3), 407.
Klimek-Kopyra, A., Bacior, M., Lorenc-Kozik, A., Neugschwandtner, R., Zając, T., 2020.
The Intraspecic competition as a driver for true production potential of soybean.
Ital. J. Agron. 16 (1).
oppler, M.R., Hitchmough, J.D., 2015. Ecology good, aut-ecology better; improving the
sustainability of designed plantings. J. Landsc. Archit.
Lauenroth, W.K., Adler, P.B., 2008. Demography of perennial grassland plants: survival,
life expectancy and life span. J. Ecol. 96 (5), 10231032.
Lawson, C., Ford, M., Mitchley, J., 2004. The inuence of seed addition and cutting
regime on the success of grassland restoration on former arable land. Appl. Veg. Sci.
7 (2), 259266.
Lindemann-Matthies, P., Junge, X., Matthies, D., 2010. The inuence of plant diversity
on peoples perception and aesthetic appreciation of grassland vegetation. Biol.
Conserv. 143 (1), 195202.
Liu, C., 2010. Plants in Saihanba. China Forestry Publishing House,, Beijing.
Liu, S., Zhang, Y., Jiang, J., Duan, B., 2015. Wild Flowers of Inner Mongolia, rst ed.
China Forestry Publishing House, Beijing.
Lubin, T.K., et al., 2019. Are two strategies better than one? Manipulation of seed density
and soil community in an experimental prairie restoration. Restor. Ecol. https://doi.
Met Ofce (2018) Summer 2018 UK.
Morris, E., Myerscough, P., 1991. Self-thinning and competition intensity over a gradient
of nutrient availability. J. Ecol. 79 (4), 903923.
Nassauer, J.I., 1995. Messy ecosystems, orderly frames. Landsc. J.
Nemec, K.T., et al., 2013. Inuence of richness and seeding density on invasion resistance
in experimental tallgrass prairie restorations. Ecol. Restor.
Ni, J., 2004. Estimating net primary productivity of grasslands from eld biomass
measurements in temperate northern China. Plant Ecol. 174 (2), 217234. https://
Ozgüner, H., Kendle, A.D., Bisgrove, R.J., 2007. Attitudes of landscape professionals
towards naturalistic versus formal urban landscapes in the UK. Landsc. Urban Plan.
Potts, S.G., et al., 2009. Enhancing pollinator biodiversity in intensive grasslands.
J. Appl. Ecol.
Pywell, R.F., et al., 2003. Plant traits as predictors of ecological performance. J. Appl.
Ecol. 40 (1), 6577.
Qi, A., et al., 2018. Grassland futures in Great Britain Productivity assessment and
scenarios for land use change opportunities. Sci. Total Environ. 634, 11081118.
Schmithals, A., Kühn, N., 2014. To Burn or Not to Burn? Effect of Management Strategy
on North American Prairie Vegetation for Public Urban Areas in Germany. PLoS
ONE 9 (10), e108588.
Scotton, M., 2019. Mountain grassland restoration: Effects of sowing rate, climate and
soil on plant density and cover. Sci. Total Environ.
Silvertown, J., et al., 2006. The Park Grass Experiment 1856-2006: Its contribution to
ecology. J. Ecol.
Southon, G.E., et al., 2017. Biodiverse perennial meadows have aesthetic value and
increase residents perceptions of site quality in urban green-space. Landsc. Urban
Southon, G.E., et al., 2018. Perceived species-richness in urban green spaces: Cues,
accuracy and well-being impacts. Landsc. Urban Plan.
Stevenson, M.J., Bullock, J.M., Ward, L.K., 1995. Re-creating Semi-natural Communities:
Effect of Sowing Rate on Establishment of Calcareous Grassland. Restor. Ecol.
Walker, K.J., et al., 2004. The restoration and re-creation of species-rich lowland
grassland on land formerly managed for intensive agriculture in the UK. Biol.
Conserv. 118.
Wilby, A., Brown, V.K., 2001. Herbivory, litter and soil disturbance as determinants of
vegetation dynamics during early old-eld succession under set-aside. Oecologia.
Yoda, K., et al., 1963. Self-thinning in overcrowded pure stands under cultivated and
natural conditions (Intraspecic competition among higher plants XI). J. Biol.
M. Jiang and J.D. Hitchmough
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Phenotypic plasticity of agricultural plants is considered to be one of the main means by which plants cope with the variability of environmental factors. A major contributor to plant plasticity is sowing density, which has a relevant impact on competitive intensity concerning plant density in different environments (CI) and absolute severity of competition (ASC) concerning plant-plant responses to each other in canopy. A field experiment with soybean was set up at the Experimental Station in Prusy, Krakow, to determine the impact of intraspecific competition on growth, plant architecture, nitrogen accumulation, and yield of soybean as an effect of seven different plant densities and weather conditions. The study showed that intraspecific competition in soybean was conditioned by sowing density and access to water, thus revealing the true plant productive potential. Low intraspecific competition increased with plant density causing an increase in the yield of plants. In the wet year of 2014, strong intraspecific competition resulted in high yield and nitrogen accumulation only up to a density of 42 plants m-2, compared to dry years when nitrogen uptake of soybean increased with plant density in full range. The CI and ASC competition indices were sensitive to the varying amount of rainfall. Greater rainfall during crop vegetation increased the intensity of competition as well as the absolute severity of competition and decreased the relative yield with increasing density. In contrast, drought reduced intraspecific competition, eliminating it entirely at over 52 plants m-2.
Full-text available
There is increasing evidence of the benefits of introducing urban meadows as an alternative to amenity mown grass in public greenspaces, both for biodiversity, and human wellbeing. Developing a better understanding of the meadow characteristics driving human and wildlife response is therefore critical. We addressed this by assessing public and invertebrate response to eight different annual meadow mixes defined by two levels of plant species diversity and two levels of colour diversity, sown in an urban park in Luton, UK, in April 2015. On-site questionnaires with the visiting public were conducted in July, August and September 2015. Invertebrate responses were assessed via contemporaneous visual surveys and one sweep net survey (August 2015). Flower colour diversity had effects on human aesthetic response and the response of pollinators such as bumblebees and hoverflies. Plant species diversity, however, was not a driver of human response with evidence that people used colour diversity as a cue to assessing species diversity. Plant species diversity did affect some invertebrates, with higher abundances of certain taxa in low species diversity meadows. Our findings indicate that if the priority for sown meadows is to maximise human aesthetic enjoyment and the abundance and diversity of observable invertebrates, particularly pollinators, managers of urban green infrastructure should prioritise high flower colour diversity mixes over those of high plant species diversity. Incorporating late-flowering non-native species such as Coreopsis tinctoria (plains coreopsis) can prolong the attractiveness of the meadows for people and availability of resources for pollinators and would therefore be beneficial.
Full-text available
Editor: R Ludwig To optimise trade-offs provided by future changes in grassland use intensity, spatially and temporally explicit estimates of respective grassland productivities are required at the systems level. Here, we benchmark the potential national availability of grassland biomass, identify optimal strategies for its management, and investigate the relative importance of intensification over reversion (prioritising productivity versus environmental ecosystem services). Process-conservative meta-models for different grasslands were used to calculate the baseline dry matter yields (DMY; 1961-1990) at 1 km 2 resolution for the whole UK. The effects of climate change, rising atmospheric [CO 2 ] and technological progress on baseline DMYs were used to estimate future grassland productivities (up to 2050) for low and medium CO 2 emission scenarios of UKCP09. UK benchmark productivities of 12.5, 8.7 and 2.8 t/ha on temporary, permanent and rough-grazing grassland, respectively, accounted for productivity gains by 2010. By 2050, productivities under medium emission scenario are predicted to increase to 15.5 and 9.8 t/ha on temporary and permanent grassland, respectively, but not on rough grassland. Based on surveyed grassland distributions for Great Britain in 2010 the annual availability of grassland biomass is likely to rise from 64 to 72 million tonnes by 2050. Assuming optimal N application could close existing productivity gaps of ca. 40% a range of management options could deliver additional 21 * 10 6 tonnes of biomass available for bioenergy. Scenarios of changes in grassland use intensity demonstrated considerable scope for maintaining or further increasing grassland production and sparing some grassland for the provision of environmental ecosystem services.
Full-text available
Evidence that urban green-space promotes health and well-being of urban residents is increasing. The role of biodiversity is unclear: perceived biodiversity may be important, but how accurately it is perceived and the factors influencing this accuracy are poorly understood. We use experimental perennial urban meadows in southern England to investigate the impact of creating biodiverse habitats on green-space users’ i) physical and mental health, psychological well-being, ii) factors moderating health and well-being outcomes (site satisfaction and nature connectedness), and iii) perceived biodiversity. We explore whether ‘nature dose’ (time spent at a site) influences these relationships. We then assess whether green-space users can estimate botanical diversity accurately across meadow treatments differing in plant species richness and vegetation structure, and determine the environmental cues and personal characteristics associated with these estimates. Sites with experimental meadows did not increase respondents’ perceptions of site level biodiversity, their self-rated physical and mental health or psychological well-being relative to control sites lacking meadows. However, there were significant associations between perceived site level biodiversity per se, and site satisfaction and feeling connected to nature. Moreover, we observed a positive association between nature dose and self-estimated mental health. We found that actual and perceived botanical richness in individual meadow plots were strongly positively correlated. Perceived richness was positively associated with vegetation height, evenness, and colourfulness suggesting that these are cues for estimating species richness. The accuracy of estimates varied, but respondents with higher levels of eco-centricity were more accurate than people who were less connected to nature.
Restoration practitioners have a variety of practices to choose from when designing a restoration, and different strategies may address different goals. Knowledge of how to best use multiple strategies could improve restoration outcomes. Here we examine two commonly suggested strategies in a single tallgrass prairie restoration experiment: increased forb sowing density and prairie soil inoculation. We designed a study with two different forb seeding densities. Within these densities, we transplanted seedlings into 1 m² plots that had been grown in either a whole prairie soil inoculum or sterilized prairie soil. After 4 years, we found positive effects of both high forb sowing density and inoculation treatments on the ratio of seeded to non‐seeded plant cover in these plots, and negative effects of both treatments on non‐seeded plant diversity. No effects of either treatment were seen on seeded plant diversity. Each strategy also affected the plant community in different ways: high forb sowing density increased seeded forb richness and decreased native non‐seeded plant cover, while inoculation decreased non‐native cover, and tended to increase average successional stage of the community. These effects on restoration outcome were typically independent of each other, with the result that plots with both manipulations had the most positive effects on restoration outcomes. We thus advocate the combined use of these restoration strategies, and further studies which focus on both seeding and soil community manipulation in restoration. This article is protected by copyright. All rights reserved.
There is much interest in supporting biodiversity through eye-catching low maintenance requiring herbaceous vegetation types in urban environments. Community development of forb vegetation is not well understood, and experimental studies investigating the potentials and limitations of establishing such vegetation are needed. We performed a field experiment including three mixtures of 19 forb species (three annuals, three biennials, 13 perennials) and analysed the development of the community over 3 years. In spring 2010 a total of 3630 seedling plug plants were transplanted in a randomized block experiment with 30 plots; non-planted control plots were monitored as reference. Management was limited to removal of dead plant material in February. We documented transplant survivorship, gap colonisation and soil seed bank composition. Standing biomass was harvested in a subset of plots in August 2012, and in all plots in July 2013. After four growing seasons, 1592 transplants were still alive. Fifteen experimental species had spread to gaps, but grasses, Achillea millefolium and Trifolium pratense var. sativum were the most frequent gap-colonizers. On average, biomass of non-experimental species contributed only 11% to the total aboveground biomass in treatment mixtures after four growing seasons. The initial proportions of species (annuals vs perennials) had no significant effect on the community's resistance to colonisation by non-experimental species. Biomass of non-experimental species was negatively correlated to transplant survival. Thus, high plant density increases resistance to invasion, but long-term persistence of forb vegetation is species-specific. In practice some disturbance is necessary to facilitate regeneration of desirable forbs.
Semi-natural grasslands are of great importance to European biodiversity. Their restoration in mountain regions often presents specific problems of raw soil stabilisation and difficult climatic conditions. Methods used to tackle these problems include increasing sowing density and fertilisation. However, past studies were performed at only one site or for a short time and do not clearly show how site traits and time affect restoration results. In six field experiments undertaken on raw soils in the Italian Alps, the effects of sowing density, fertilisation, soil, and climate on the development over time of plant density and cover were analysed over three to six years. The seed was collected from nearby grasslands and sown at rates of 335–21,410 seeds m⁻². During the first one-two years, the percentage of seeds producing established plants at low seed rates was higher in stonier surfaces; however, at high rates a larger absolute plant number was recorded in finer soils. Two to five years after sowing, climatic conditions became more important. Full vegetation cover was achieved in all experiments. However, at high altitudes, individual plant cover became greater (and plant density decreased), probably because the grasses needed to accumulate more non-structural carbohydrates necessary to survive during the longer winter. Fertilisation accelerated the establishment of full vegetation cover by 1 to 3 years as it increased the mean cover per plant without affecting plant density. Grassland restoration is possible in mountain raw soils by using the seeds collected from nearby donor sites. High sowing densities are not required even in coarse or higher altitude soils because, in the short term, stones on the ground surface increase the percentage of seed producing established plants and, over the long term, plant density is not dependent on the sowing density. In low-fertile soils, fertilisation accelerates the establishment of full vegetation cover.
Interactions between two canopy layers in a designed perennial herbaceous plant community were investigated over a period of four and a half years to see whether it was possible to create an urban landscape vegetation that was both flower rich for an extensive time period and resistant to weed colonization at very low levels of maintenance by sowing seed in situ. The ecologically novel plant community involved a tall over-canopy layer of 18 species of North American prairie and woodland edge forbs, and a shade tolerant under-canopy of eight European and North American, mainly woodland forbs. After 5 years the community was dominated by four over-canopy and two winter green under-canopy species. Interspecific competition generated by the sown biomass restricted weed colonization to very low levels, despite the experiment being surrounded by a weedy brownfield. The winter green canopies of the two dominant under-storey forbs closed down gaps within a winter deciduous, prairie-like vegetation, improving winter appearance and providing a major flowering display in spring. This vegetation is an attractive design model for more sustainable herbaceous planting in urban landscape space.
A cost-effective approach in plant restorations could be to increase sowing density for species known to be challenging to establish, while reducing sowing density for species that easily colonize on their own. Sowing need not occur evenly across the site for rapidly dispersing species. We explored these issues using a prairie restoration experiment on a high-school campus with three treatments: plots sown only to grasses (G plots), to grasses and forbs (GF1), and to grasses and forbs with forbs sown at twice the density (GF2). In year 2, GF1 and GF2 plots had higher diversity than G plots, as expected, but GF2 treatments did not have twice the sown forb cover. However, high forb sowing density increased forb richness, probably by reducing stochastic factors in establishment. Cover of nonsown species was highest in G plots and lowest in GF2 plots, suggesting suppressive effects of native forbs on weedy species. Colonization of G plots by two sown forbs (Coreopsis tinctoria and Rudbeckia hirta) was apparent after 2.5 years, providing evidence that these species are self-sustaining. Colonization was greater in edges than in the central areas of G plots. Through construction of establishment kernels, we infer that the mean establishment distance was shorter for R. hirta (6.7 m) compared to C. tinctoria (21.1 m). Our results lead us to advocate for restoration practices that consider not only seed sowing but also subsequent dispersal of sown species. Furthermore, we conclude that restoration research is particularly amenable for outdoor education and university-high school collaborations.