American Journal of Botany 95(6): 672–680. 2008.
RESPONSES OF TWO BUNCHGRASSES TO NITROGEN ADDITION IN
TALLGRASS PRAIRIE: THE ROLE OF BUD BANK DEMOGRAPHY1
Harmony J. Dalgleish,2 Abigail R. Kula,3 David C. Hartnett, and Brett K.
Sandercock
Division of Biology, Kansas State University, 116 Ackert Hall, Manhattan, Kansas 66506 USA
Growth of tallgrass prairie plants, many of which maintain substantial bud banks, can be limited by nitrogen (N), water, and/or
light. We hypothesized that tallgrass prairie plants respond to increases in N through demographic effects on the bud bank. We
tested the effects of a pulse of N on (1) bud bank demography, (2) plant reproductive allocation, and (3) ramet size. We parameterized matrix models, considering each genet as a population of plant parts. Nitrogen addition significantly impacted bud bank demography in two subdominant species of bunchgrass: Sporobolus heterolepis (a C4 grass) and Koeleria macrantha (a C3 grass),
but had no effect on the size of individual ramets. Emergence from the bud bank and ramet population growth rates (λ) were significantly higher in S. heterolepis genets that received supplemental N. Nitrogen addition also affected the bud demography of
K. macrantha, but N addition decreased rather than increased λ. Prospective and retrospective demographic analyses indicated that
bud bank dynamics were the most important demographic processes driving plant responses to nutrient availability. Thus, the
variation in productivity in these tallgrass prairie species is driven principally by the demography of the bud bank rather than by
the physiology and growth of aboveground tillers. Improved understanding of bud bank dynamics may lead to improved predictive
models of grassland responses to environmental changes such as altered N deposition and precipitation.
Key words: Kansas; Koeleria macrantha; Konza Prairie Long-Term Ecological Research Site; meristem; Poaceae; Sporobolus
heterolepis.
Identifying the underlying mechanisms and the sources of
variation in plant productivity is critical to the development of
accurate predictive models of ecosystem responses to environmental change, including increased nitrogen (N) deposition and
altered precipitation patterns. A great deal of variation in aboveground net primary production (ANPP) in tallgrass prairie may
be accounted for by variation in tiller density, rather than differences in tiller size (Hartnett and Fay, 1998). In addition, many
tallgrass prairie plants maintain a substantial bud bank: a belowground population of meristems associated with rhizomes
or other perennating organs (sensu Harper, 1977). The bud bank
serves as the recruitment source for the majority of aboveground tillers in tallgrass prairie (Benson and Hartnett, 2006;
Dalgleish and Hartnett, 2006). Consequently, bud bank dynamics play a fundamental role in local plant population persistence,
structure, and dynamics. If variation in ANPP in tallgrass prairie can be explained principally through changes in tiller density, population processes such as the demography of buds and
tillers are important not only for understanding population dynamics of individual species, but for explaining temporal and
spatial variability in ANPP as well. The study of bud bank demography represents a crucial link between organismal (formation
1
and maintenance of the bud bank, rates of tiller emergence) and
ecosystem level processes (ANPP).
In tallgrass prairie, plant productivity may be limited by three
key resources: N, water, and light (Seastedt and Knapp, 1993;
Seastedt et al., 1998). The relative importance of these resources
for limiting plant growth varies considerably, both spatially and
temporally. In addition, human alteration of the N cycle through
burning fossil fuels and agricultural activities has led to at least
a doubling of the rates of terrestrial N input (Vitousek et al.,
1997). Increased N deposition has many important impacts on
plant communities (Clark et al., 2007). In this research, we
tested the role of the bud bank in plant responses to N. Grassland plants respond to available N as a key resource that can
limit plant growth, but recent theory suggests that N may also
play a role as a proximal cue for bud dormancy and emergence
(Tomlinson and O’Connor, 2004). The known importance of N
for limiting growth in tallgrass prairie plants as well as its potential to cue bud emergence, may result in ecologically significant effects of altered N availability on bud bank and tiller
demography, rather than effects on tiller growth alone. Although previous studies have examined the effects of N on tiller
demography in grasses (Noble et al., 1979; Briske and Butler,
1989; Hartnett, 1993; Derner and Briske, 1999), none have examined the effects of resource availability on the ultimate
source for tiller recruitment, the bud bank. Understanding bud
bank dynamics in conjunction with aboveground tiller responses
is important because changes in tiller recruitment or tiller size
may not necessarily lead to similar changes in the bud bank.
The maintenance of a belowground bud bank, like a seed bank,
represents a reserve population of dormant propagules that may
have important ecological consequences.
We addressed the general question of how tallgrass prairie
plants will respond to increased resource availability. We hypothesized that tallgrass prairie plant growth and production
would respond to changing resource availability principally
through changes in the demography of bud banks (changing
Manuscript received 23 August 2007; revision accepted 27 March 2008.
The authors thank J. Sinn, J. Hill, Z. Gill, M. Williamson, J. Birmingham,
and J. Jonas for field assistance and Dr. E. G. Towne for help finding
suitable populations of study plants. This work was supported by the Konza
Prairie LTER, the Kansas State University Division of Biology, NSF grant
DEB-0234159 and one REU supplement grant for DEB-0234159.
2 Author for correspondence (e-mail: h.dalgleish@usu.edu), present
address: Department of Wildland Resources, Utah State University, 5230
Old Main Hill, Logan, UT 84322 USA
3 Present address: Department of Biology, Biology-Psychology Building,
University of Maryland, College Park, MD 20742 USA
doi:10.3732/ajb.2007277
672
June 2008]
Dalgleish et al.—Bud banks and plant responses to nitrogen
numbers of buds and tillers within a genet), rather than changes
in individual tiller size. In addition to affecting bud bank
demography, nutrient supplements could alter allocation patterns between vegetative and sexual reproduction by altering
the number of flowering tillers. We experimentally manipulated
N availability for genets of two species of subdominant tallgrass prairie bunchgrasses (syn. caespitose, tussock): Sporobolus heterolepis (A. Gray, prairie dropseed) and Koeleria
macrantha (Ledeb.) Schult (Junegrass) on a watershed that is
burned annually in the autumn and is protected from grazing at
Konza Prairie Biological Station (KPBS) in northeast Kansas.
Though they are not the dominant growth form in tallgrass prairie, we selected bunchgrasses because we could apply N to a
single genet and follow the tillers within a single genet, modeling the genet as a population of plant parts (buds and tillers).
We could then apply demographic analysis methods to determine which plant parts were responding to the N addition
(if any) as well as compare tiller sizes between treatments.
Sporobolus heterolepis and K. macrantha are two of the most
common bunchgrass species at KPBS (Towne, 2002).
Our objectives were to test the effect of a pulsed increase in a
limiting nutrient (N) on (1) bud bank demography and aboveground tiller populations, (2) plant reproductive allocation and a
tradeoff between flowering and vegetative bud production, and
(3) ramet (tiller) size. If the dynamics of the belowground bud
bank are more important than the growth responses of individual
tillers, we predicted three bunchgrass responses to N fertilization: (1) Increased probability of emergence from the bud bank
and increased ramet population growth rates (λ). Additionally,
we expected that stage transitions involving the bud bank should
have the greatest effect on λ. (2) Increased vegetative bud production and an increased probability of flowering. Because of a
trade-off in sexual and vegetative reproduction, however, tillers
that flower should have decreased bud production regardless of
N treatment. (3) No change in ramet (tiller) size.
MATERIALS AND METHODS
Konza Prairie Biological Station (KPBS) is a 3487-ha tallgrass prairie research site located 10 km south of Manhattan, Kansas, USA, within the Flint
Hills region of northeastern Kansas (39°05′N, 96°35′W). Annual precipitation
averages 835 mm. Konza Prairie is representative of the Flint Hills with hard
chert- and flint-bearing limestone bedrock. Upland ridges are usually flat with
shallow, rocky soils, whereas the lowland valleys have deep permeable soils.
KPBS is divided into 52 watershed units (average size = 60 ha) subjected to
different combinations of long-term experimental fire frequency, season of fire,
and grazing treatments.
Our study was conducted in 2004 and 2005, which were above-average
years for total precipitation (2004: 988 mm, 17% above average; 2005: 891 mm,
7% above average). However, seasonal timing of the precipitation differed between the two years. March–August precipitation was 148 mm lower in 2005
(666 mm) than 2004 (814 mm). The precipitation deficit accrued from the start
of the growing season with March 2005 having 117 fewer mm of rain than
March 2004. Thus, 2005 was drier during the critical times of bud emergence,
growth, and flower development, and water availability was potentially more
limiting to plant growth.
Species and location—Sporobolus heterolepis is a subdominant warmseason, C4, perennial bunchgrass that is occasional on all KPBS sites (Towne,
2002). It flowers between August and September and is adapted to low to moderate levels of soil fertility. The average basal area of S. heterolepis genets in
our study was 953.4 ± 114.0 cm2 (N = 240, mean ±1 SE). Koeleria macrantha
is a subdominant, cool season, C3 bunchgrass. It is common in upland areas of
KPBS (Towne, 2002), flowers between late May and June, and is adapted to
low to moderate levels of soil fertility. The average basal area of K. macrantha
genets in our study was 33.9 ± 8.6 cm2 (N = 240, mean ±1 SE). This study was
673
conducted in the uplands of a 12-ha watershed burned annually in autumn but
not grazed by bison or cattle (Watershed FA). We selected this site because we
could locate at least 300 genets of each species, which made coordination of
field crews and completion of intensive field measurements feasible.
Nitrogen treatment—In late April 2004, 240 genets of each study species
were flagged and numbered. To select the plants, we started at one end of the
watershed and marked each plant we encountered until we had an adequate
sample for the study. Plants were excluded only if their boundaries were
unclear (i.e., two genets appeared to be growing together). Genets were then
randomly assigned to one of two treatments and either a nondestructive or destructive sampling method (as described in the Nondestructive aboveground
tiller sampling and Belowground bud sampling sections). Half of the marked
genets (N = 120 for each species; hereafter, “N addition” plants) received dry
ammonium nitrate powder at a rate of 10 g N/m2 during the week of 5 May
2004 (Callaham et al., 2003). Because a bunchgrass genet can take up nutrients
within a circle up to 50% greater than its diameter, we fertilized the entire genet
as well as an area 50% greater than its diameter (Derner and Briske, 1999).
Slow-release fertilizer pellets were crushed with a mortar and pestle to create
ammonium nitrate powder. Application of powder on calm days ensured an
even distribution on study plants. Rain fell within 24 h after all applications. No
evidence of tissue burn by the fertilizer was observed during the study. Control
plants were not treated.
The same nitrogen application methods and rates were used in 2005, although the timing of fertilizer application was six weeks earlier for K. macrantha to better coincide with the growth phenology of this cool-season, C3 species
(23 March 2005 for K. macrantha and 10 May 2005 for S. heterolepis). In the
second year of the study, fertilizer was applied to genets that had been fertilized
but not destructively harvested during the first year of the study. Thus, the genets sampled in the second year of the study received two applications of
nitrogen.
Nondestructive aboveground tiller sampling, 2004—For each treatment,
30 genets of each species were designated for repeated, aboveground, nondestructive sampling in 2004 (“tiller” plants, Table 1). These plants were visited
every two weeks from 5 May to 2 October 2004 (N = eight samples for each
plant with sample one used as a baseline). All vegetative and flowering ramets
(tillers) of each genet were counted. One fertilized genet of S. heterolepis was
mistakenly harvested for bud counts and was removed from the tiller plants
category. Thus, the final sample sizes for tiller plants of S. heterolepis were 29
fertilized genets and 30 control genets, with 30 plants in each treatment of
K. macrantha. The data from the S. heterolepis genets were used to parameterize the matrix models for 2004. We were not able to parameterize a model for
K. macrantha in 2004 because we did not obtain sufficient early season information. Sampling dates were shifted earlier in 2005 to remedy this problem
(Table 1).
Nondestructive aboveground tiller sampling, 2005—We changed the sampling scheme and sample sizes in 2005, informed by data from 2004 and incorporating logistical constraints (Table 1). Seven fertilized genets and eight
control genets of S. heterolepis were designated for repeated, aboveground,
nondestructive sampling in 2005. These 15 genets were separate from the 59
plants sampled in 2004, but because all genets received the same treatments,
measurements made on a sample of plants from one year should be comparable
to measurements made on a sample of plants in the next year. Genets were
visited every three weeks from 16 May to 11 October 2005 (N = 6 sampling
occasions for each genet). Due to the high number of tillers within a single
genet (often over 1000), it was time-consuming to census all tillers of S. heterolepis in 2004. In 2005, we developed a subsampling procedure to reduce sampling time and increase counting precision in the field. Five circular, 11.8-cm2
subplots were constructed from plastic-coated copper wire and fixed within
each genet using a small piece of wire. Subplots were placed before tiller emergence in March. Because ramet densities in S. heterolepis are higher at the
edges of the genet, two of the five plots were placed at the edges: one perpendicular to the shortest diameter and one perpendicular to the longest diameter,
as determined by a coin toss. One plot was placed at the sparsest location on the
genet, which was always in the interior of the genet. The remaining two plots
were randomly placed in the interior of the genet by tossing them into the genet
while looking away. At each sampling time, all vegetative and flowering tillers
were counted within the five subplots within each genet, and the density of the
tillers was then multiplied by the genet’s basal area to calculate total tillers.
Because S. heterolepis has so few flowering tillers, which are easily identified
674
[Vol. 95
American Journal of Botany
Table 1.
Summary of the sampling scheme, sample sizes, and the parameters calculated from the data. Sampling of K. macrantha began and ended earlier
in 2005 to better align with its cool-season phenology. Bud plant sample sizes were decreased in 2005 based upon data from 2004. Tiller plant sample
sizes were decreased in 2005 for logistical reasons, but note that precision of the vital rate estimates was unaffected (Table 3). Tiller plants from 2004
were used as bud plants in 2005. Tiller plants in 2005 were not sampled in 2004.
Sporobolous heterolepis
Plant type
Sampling
Tiller plants
Nondestructive sampling
N (half: N treatment, half: Control)
Use of dataa
Bud plants
Destructive sampling
N (half N, half Control)
No. harvests
No. genets/Harvest
No. tillers harvested/Genet harvested
Use of data
a See
Koeleria macrantha
5 May–2 Oct 2004
15 May –11 Oct 2005
Count all tillers
59
Sb, Sv, Sf, Gb, Gv
Count a subsample
15
Sb, Sv, Sf, Gb, Gv
5 May –2 Oct 2004
23 Mar –9 July 2005
Count all tillers
Count all tillers
60
20
Insufficient earlySb, Sv, Sf, Gb, Gv
season data
Count all buds on harvested tillers from randomly sampled genets
160
58
160
49
8
6
8
5
20
8–10
20
8–10
10
20
10
20
Tiller size
Vv, Vf, tiller size
Vv, Vf, tiller size
Vv, Vf, tiller size
Table 2 for definition of parameters.
by their large size, we counted all the flowering tillers on a genet. Before the
start of the study in 2005, both tiller counting protocols (counting all tillers vs.
subsampling) were applied to 10 S. heterolepis genets to determine that they
yielded density and total tiller estimates that were not statistically different
(Dalgleish, 2007). Thus, even though the method of determining tiller numbers
per genet was different between years, the estimates of tiller population size,
survival, and fecundity are directly comparable.
For each treatment in 2005, 10 genets per treatment of K. macrantha were
designated for repeated, aboveground, nondestructive sampling. Genets were
visited every three weeks from 23 March to 9 July 2005, to ensure our sampling
matched the phenology of this cool season species (N = 5 sampling times for
each genet).
Belowground bud sampling, 2004 and 2005—In 2004, 80 genets of each
species in each treatment were used to estimate bud production (Table 1). Every
two weeks, 10 genets of each species per treatment were randomly selected for
destructive sampling and then removed from the experiment for a total of eight
harvests. All tillers were counted on these plants, and 10 tillers per genet were
removed and placed in coolers for transport (N = 100 tillers per species per
treatment per sampling). In the laboratory, the height of each tiller was measured and the number of buds counted on each tiller. Bunchgrasses, including
the two species in this study, produce their tillers intravaginally from inside the
persistent leaf sheath of the existing tiller, making it easy to associate buds with
a specific tiller (Briske and Derner, 1998). To count the buds on bunchgrass
tillers, observers pulled back successive leaf sheaths at the base of the tiller to
reveal the buds. Little research has been done on the longevity of grass bud
banks, but buds may be viable for two or more years (Hendrickson and Briske,
1997). After dissection, tillers were dried at 60°C for 48 h and weighed to estimate biomass to the nearest 0.01 g.
Our protocol for bud sampling remained the same in 2005, but sample sizes
were reduced to 49 total genets of K. macrantha and 58 total genets of S. heterolepis. Every three weeks, approximately 10 genets per species were subsampled, with each genet being sampled only once. All tillers on K. macrantha
genets and five 11.8-cm2 subplots of tillers were counted on S. heterolepis (as
previously described) to estimate total tiller number for each genet. Twenty
tillers were removed per genet and placed in coolers for transport to the laboratory for measurement and bud counting (N ≈ 100 tillers per species per treatment per sampling occasion). The changes in timing and sample size were
informed by the results of the eight harvests in 2004.
where V, S, and G are the stage-specific vital rates for vegetative bud production, survival, and growth, respectively. The projection matrix was linear and
deterministic and did not incorporate density-dependence.
Model parameterization—We parameterized the stage-structured projection matrix to determine population growth rates for each genet. We parameterized separate models for S. heterolepis in 2004 and 2005, but were only able to
parameterize a model for K. macrantha in 2005 because the 2004 data did not
contain sufficient early season information. Sampling dates were shifted earlier
in 2005 to remedy this problem (Table 1). For each matrix, we included seven
vital rates calculated from our bud and tiller censuses (Table 2, Fig. 1).
To create matrix probabilities from our census data, we made three assumptions. First, we assumed no bud mortality within the bud bank, i.e., bud survival
to the next time step depends only on bud emergence. We did not observe any
dead or damaged buds during our sampling, and buds likely persist longer than
one growing season (Hendrickson and Briske, 1997). We have no data on bud
survival in the bud bank for bunchgrasses but made this assumption to calculate
Sb, the probability that a bud remains in the bud bank. Nevertheless, Sb may be
Table 2.
Parameterization of elements in projection matrices. VT =
vegetative tillers; FT = flowering tillers; subscript “1” indicates the
first sampling time; subscript “end” indicates the last sampling time;
subscript “peak” indicates the peak number of flowering or vegetative
tillers.
Matrix element
Definition
Equation
Sb
Survival probability of
a bud in the bud bank
1− G
b
Sv
Survival probability of a
vegetative tiller
1−
Survival probability of a
flowering tiller
1−
Sf
Gb
Probability of bud emergence
( VT
peak
VTpeak
( FT
peak
⎡ Sb
A = ⎢⎢Gb
⎢⎣ 0
Vv Vf ⎤
S v 0 ⎥⎥ ,
Gv Sf ⎥⎦
Gv
Probability of a vegetative tiller
flowering
Vv
Vegetative bud production of a
vegetative tiller
Vf
Vegetative bud production of a
flowering tiller
− FTend )
FTpeak
( VT
peak
− VT1 )
( VT ∗ ( V )
1
Matrix population model—To synthesize bud and tiller demographic rates
of the two species of bunchgrass, we developed a stage-structured matrix population model. Each genet was modeled as a population of buds and tillers over
an annual time step. Our projection matrix model had three discrete life stages:
bud (b), vegetative tiller (v), and flowering tiller (f) (Fig. 1):
− VTend − FTpeak )
v 1
+ VT1 )
FTpeak
VTpeak
∑ ( No. buds per VT )
( No. VT sampled )
∑ ( No. buds per FT )
( No. FT sampled )
June 2008]
Dalgleish et al.—Bud banks and plant responses to nitrogen
Fig. 1. Diagram of life cycle model for the tiller populations for
Sporobolus heterolepis and Koeleria macrantha. The model has three
stages (bud, vegetative tiller, and flowering tiller), two growth transitions
(Gb, bud emergence to vegetative tiller; Gv, vegetative tiller flowering),
three survival probabilities (Sb, Sv, and Sf), and two fecundity arcs that represent the number of vegetative tillers produced per vegetative (Vv) and per
flowering (Vf) tiller.
an overestimate. Second, we assumed that all changes in tiller population size
leading up to the peak number of tillers were due to emergence of new tillers
alone and that no tillers died before the peak of the growing season. Last, we
assumed that all changes in tiller population size after the peak number of tillers
were due to mortality.
Statistical analyses—To test for differences in the effects of N addition on
Vv and Vf (the average number of buds produced per tiller), we used a split-plot
design with genet as the whole plot factor and the presence or absence of a
flower on a tiller as the subplot factor using the program SAS 9.1 (Proc Mixed;
SAS, 2003). Tiller size (grams biomass) was used as a continuous covariate.
A repeated measures model was not appropriate because genets were randomly
assigned to a destructive sampling time and individual genets were not resampled. To test for differences in tiller size, we used a mixed-model ANOVA design with sampling time and treatment as fixed effects and genet as a random
effect using SAS 9.1 (Proc Mixed; SAS, 2003). In both of these models, the
Kenward–Rogers correction was applied to the degrees of freedom. Differences
between matrix elements and λ (population growth rate) between fertilized and
control genets were analyzed with a Kruskal–Wallis test using SAS 9.1 (Proc
Npar1way; SAS, 2003). All tests were two-tailed, based on type III sums of
squares and considered significant at an α level < 0.05.
Perturbation analyses—To analyze our population models, we performed
both prospective and retrospective analyses. The goal of prospective and retrospective analyses is to explore the effect of perturbation of vital rates on λ, the
finite rate of population growth (Horvitz et al., 1997). Prospective analyses consider the consequences of potential future changes in vital rates on λ (Horvitz
675
et al., 1997). Vital rates with large elasticity values would potentially have the
largest effects on λ if changed. In addition, we conducted a fixed-effect life
table response experiment (LTRE—a retrospective approach), which aims to
quantify the contribution of each of the vital rates to the treatment effect on
population growth rate (Horvitz et al., 1997; Caswell, 2001). Specifically, we
estimated how much of the difference in λ between the N addition and control
treatments was due to each of the seven vital rates (Vv, Vf, Sb, Sv, Sf, Gb, and Gv).
Vital rates with large positive or large negative contribution values contribute
the most to the difference in λ observed between treatments. Contribution values can be large either because the vital rate is different between treatments,
because the vital rate has a high sensitivity value (indicating that it would have
large effects on λ if perturbed), or both.
Perturbation analyses were conducted using algorithms of the program Matlab 6.5 (Matlab, 2002) following formulae of Caswell (2001). The one-way
fixed effect LTRE was conducted as described by Caswell (2001) on matrices
using two levels of fertilizer treatment (for S. heterolepis and K. macrantha) or
two levels of year (for S. heterolepis only). The mean control matrix or the
mean 2005 matrix was used as the reference matrix in the separate LTRE
analyses.
RESULTS
Effects of N addition on transition probabilities— Supplementary N significantly increased the probability of tiller emergence from the bud bank (Gb) by about 9% for S. heterolepis
after two N additions (Control: 0.23 ± 0.002 SE; +N: 0.25 ±
0.01, χ12 = 4.35, P = 0.04). Nitrogen did not affect the probability of emergence in 2004, the year with greater early season
precipitation (Table 3). Nitrogen had no effect on the probability of bud outgrowth and tiller emergence in K. macrantha in
2005.
The probability of a tiller flowering or the proportion of tillers within a genet that flowered is an indirect measure of resource allocation to sexual reproduction that provides a useful
estimator for treatment comparisons. Nitrogen significantly increased the probability of flowering (Gv) for both species (Table
3). In 2004 in S. heterolepis, N addition doubled the probability
2
of flowering from 0.03 ± 0.01 SE to 0.07 ± 0.01 SE in 2004 ( χ1 =
6.49, P =2 0.01), and from 0.001 ± 0.001 SE to 0.03 ± 0.01 SE in
2005 ( χ1 = 4.55, P = 0.03). Nitrogen also tended to increase the
probability of tillers flowering in K. macrantha from 0.05 ±
0.02
SE to 0.08 ± 0.02 SE, but the difference was not significant
2
(χ1 = 1.76, P = 0.18).
Effects of N addition on bud production— The absolute
number of buds produced per flowering tiller (Vf) was higher
in fertilized genets of S. heterolepis (Table 3); however, if the
additional biomass accumulated by flowering tillers is included
Summary of matrix elements and λ for two species of bunchgrass in tallgrass prairie for two years for plants treated with N (+N) and those
without N treatment (Control). Values are means ± 1 SE. Boldfaced values differ significantly between treatments within a species and within a year
at P < 0.05. Refer to Table 1 for sample sizes.
Table 3.
Sporobolus heterolepis
Koeleria macrantha
2004
Parametera
Sb
Sv
Sf
Gb
Gv
Vv
Vf
λ
a See
2005
2005
Control
+N
Control
+N
Control
+N
0.44 ± 0.01
0.69 ± 0.02
0.81 ± 0.04
0.55 ± 0.01
0.03 ± 0.01
3.38 ± 0.15
4.26 ± 0.36
1.96 ± 0.02
0.45 ± 0.01
0.72 ± 0.03
0.87 ± 0.03
0.54 ± 0.01
0.07 ± 0.01
3.92 ± 0.16
4.35 ± 0.35
1.99 ± 0.02
0.77 ± 0.002
0.87 ± 0.03
1.00 ± 0.00
0.23 ± 0.002
0.001 ± 0.001
3.37 ± 0.25
4.17 ± 0.51
1.70 ± 0.02
0.75 ± 0.01
0.86 ± 0.02
0.90 ± 0.06
0.25 ± 0.01
0.03 ± 0.01
4.00 ± 0.27
5.39 ± 0.62
1.82 ± 0.02
0.18 ± 0.03
0.84 ± 0.03
0.32 ± 0.11
0.82 ± 0.03
0.05 ± 0.02
1.70 ± 0.26
1.29 ± 0.34
1.75 ± 0.02
0.18 ± 0.04
0.79 ± 0.04
0.45 ± 0.12
0.82 ± 0.04
0.08 ± 0.02
1.20 ± 0.32
1.15 ± 0.54
1.56 ± 0.04
Table 2 for definitions of parameters.
676
[Vol. 95
American Journal of Botany
into the model as a covariate, N addition did not alter the number of buds produced per tiller. The number of buds per vegetative tiller (Vv) and the number of buds produced per flowering
tiller (Vf) did not differ between the control (Vv, 3.87 ± 0.13 SE;
Vf, 3.13 ± 0.86 SE) and +N (Vv, 3.33 ± 0.11 SE; Vf, 4.41 ± 0.50
SE) treatment of S. heterolepis in 2004 (F1,362 = 3.28, P = 0.07).
Similarly, Vv and Vf did not differ between the control (Vv, 3.81 ±
0.11 SE; Vf, 2.97 ± 1.18 SE) and +N (Vv, 3.14 ± 0.10 SE; Vf,
2.28 ± 1.02 SE) treatment of S. heterolepis in 2005, (F1,351 =
0.06, P = 0.80). Koeleria macrantha had a similar pattern in
2005: Vv and Vf did not differ between the control (Vv, 1.89 ±
0.15 SE; Vf, 1.18 ± 0.32 SE) and +N (Vv, 1.31 ± 0.18 SE; Vf,
0.73 ± 0.74 SE) treatment (F1,99 = 1.27, P = 0.26). Basal area of
the genet did not significantly affect the number of buds produced per tiller (data not shown).
Effects of N addition on tiller size— Average tiller size
(grams biomass) did not differ between the control (0.31 ± 0.02
SE) and + N treatment (0.34 ± 0.02 SE) for S. heterolepis in
2004 (F1,140 = 3.09, P = 0.08). Neither did average tiller size
differ between the control (0.29 ± 0.02 SE) and the +N treatment (0.32 ± 0.02 SE) for S. heterolepis in 2005 (F1,44 = 1.20,
P = 0.28). Similarly, average tiller size did not differ between
the control (0.09 ± 0.004 SE) and the +N treatment (0.10 ±
0.005 SE,) for K. macrantha in 2004 (F1,143 = 2.28, P = 0.13).
Last, average tiller size did not differ between the control (0.11 ±
0.01 SE) and +N treatment (0.09 ± 0.01 SE) for K. macrantha
in 2005 (F1,32.6 = 1.76, P = 0.19). In addition, the average tiller
size for each species was similar between years. Basal area of
the genet did not significantly affect the average tiller size or the
density of tillers within a genet (data not shown).
Effects of N addition on tiller population growth rate— In S.
heterolepis, N addition tended to increase the population growth
rate of tillers (λ) (from 1.96 ± 0.02 SE to 1.99 ± 0.02 SE in 2004
and from 1.70 ± 0.02 SE to 1.82 ± 0.02 SE in 2005), though the
difference was only significant in 2005 after two N additions
(2004: χ12 = 0.84, P = 0.33; 2005: χ12 = 10.5, P = 0.001). In
contrast, N addition significantly decreased λ in K. macrantha
from 1.752± 0.02 SE to 1.56 ± 0.04 SE in 2005 after two N additions ( χ1 = 9.6, P = 0.002).
Prospective analysis (elasticities)—Elasticity values for the
probability of emergence from the bud bank (Gb) and the number of buds produced per vegetative tiller (Vv) were among the
highest in both species in both years and accounted for 52–68%
of the summed elasticity values of all matrix elements (Table 4).
In 2005, elasticity values for vegetative tiller survival (Sv) were
among the highest values in both treatments for K. macrantha
and in the control plants for S. heterolepis (Table 4). All elasticities associated with flowering stages (Sf, Gv, Vf) were the
lowest, ranging from 0.001 to 0.2 (Table 4).
Retrospective analysis (LTRE)—The difference in λ between
the N addition treatment and the control for K. macrantha was
0.19 with the control plants having a higher λ on average.
The LTRE showed that vegetative bud production by vegetative tillers (Vv) contributed most to the difference in λ (Fig. 2).
In fact, the contribution value for Vv (0.17) was an order of
magnitude larger than the contribution value for any other matrix element.
The difference in λ between the N addition treatment and the
control for S. heterolepis in 2004 was only −0.02 with the N
addition plants having a slightly higher λ on average. The probability of flowering (Gv) had the highest contribution value
(−0.028), nearly double the next largest, which was for vegetative bud production by vegetative tillers (Vv) (Fig. 3A). In 2005
for S. heterolepis, the difference in λ between the N addition
and control treatments was −0.13 with the N addition treatment
having the higher λ on average. Vegetative bud production by
vegetative tillers (Vv) contributed most to the difference in λ
with a contribution value (−0.078) double that of the next highest contribution value (−0.036 for the probability of bud emergence, Gb, Fig. 3B).
The LTRE between years for S. heterolepis differed in λ by
0.17 with plants in 2004, the year with the wetter spring, having
a higher λ on average than those in 2005. Emergence from the
bud bank (Gb) had a three times greater contribution to this
difference than any other matrix element (−0.45, Fig. 3C).
Survival of buds in the bud bank (Sb) had the next highest contribution value (0.14).
DISCUSSION
Our results provided partial support for our three predictions.
First, in support of prediction 1, N addition had a positive effect
on emergence from the bud bank and genet growth rates, but
only in S. heterolepis. Second, we found support for prediction
2, as both species had an increase in bud production and flowering at the genet level, though we found no evidence of a tradeoff between bud production and flowering within an individual
Table 4.
Elasticity values for demographic parameters of S. heterolepis and K. macrantha. Parameters with the largest elasticity values would have the
greatest effect on population growth if perturbed.
Sporobolus heterolepis
Koeleria macrantha
2004
Parametera
Sb
Sv
Sf
Gb
Gv
Vv
Vf
2005
2005
Control
+N
Control
+N
Control
+N
0.10 ± 0.003
0.19 ± 0.005
0.01 ± 0.001
0.35 ± 0.003
0.01 ± 0.002
0.33 ± 0.005
0.01 ± 0.002
0.09 ± 0.003
0.19 ± 0.007
0.02 ± 0.003
0.33 ± 0.004
0.02 ± 0.003
0.31 ± 0.007
0.02 ± 0.003
0.21 ± 0.008
0.26 ± 0.010
0.001 ± 0.010
0.26 ± 0.002
0.001 ± 0.010
0.26 ± 0.002
0.001 ± 0.001
0.19 ± 0.005
0.24 ± 0.007
0.01 ± 0.003
0.27 ± 0.003
0.01 ± 0.003
0.27 ± 0.006
0.01 ± 0.003
0.03 ± 0.006
0.30 ± 0.008
0.01 ± 0.003
0.32 ± 0.005
0.01 ± 0.003
0.31 ± 0.009
0.01 ± 0.004
0.05 ± 0.012
0.31 ± 0.020
0.01 ± 0.005
0.30 ± 0.004
0.02 ± 0.004
0.28 ± 0.007
0.02 ± 0.004
Notes: Values are means ±1 SE. Values > 0.25 indicated in boldface. Refer to Table 1 for sample sizes.
a See Table 2 for definitions of parameters.
June 2008]
Dalgleish et al.—Bud banks and plant responses to nitrogen
Fig. 2. One-way life table response experiment (N addition) in Koeleria macrantha illustrating the contributions of the different matrix elements to the difference in λ between the control and N addition treatments.
C > +N indicates that for contribution values in this region, the vital
rate was greater in the control than the nitrogen addition treatment and
vice versa.
tiller, likely as a result of the small sample size. Last, prediction
3 was upheld; we found no evidence that N addition affected
tiller size in either species of bunchgrass, in contrast to previous
work in Panicum virgatum, a rhizomatous grass (Hartnett,
1993).
Nitrogen addition affects bud and tiller demography—
Nitrogen addition significantly altered bud and tiller demography in both S. heterolepis and K. macrantha, while it had no
effect on individual tiller size. In 2005, emergence from the bud
bank (Gb) and growth rate (λ) were significantly higher in
S. heterolepis genets that received N. Though the difference
in the probability of emergence was small, it was biologically
significant because it led to significant differences in λ. In addition, the elasticity analysis shows that perturbation of emergence
(Gb) can have great impacts on λ in both species. While N addition also affected the demography of K. macrantha, N decreased, rather than increased λ. The difference in bud production
in vegetative tillers (Vv) between the treatments, though not statistically significant, contributed most to the significant difference in λ between treatments. One might expect lower Vv
measures in the N addition treatment if increased N leads to increased emergence (Gb). However, Gb was not different between treatments. Nitrogen addition tended to decrease bud
production per tiller in K. macrantha, and flowering also tended
to decrease bud production. The combination of the trend of
increased flowering and decreased bud production with N addition may have led to decreased λ, perhaps from a trade-off between flowering and vegetative bud production. Further research
on the effects of reproductive allocation on population dynamics in K. macrantha is necessary to resolve the mechanisms behind the decrease in λ in this species.
Both prospective and retrospective perturbation analyses of tiller population models indicated that bud bank dynamics were key
677
demographic processes affecting genet responses to increased N
availability in both species of bunchgrasses. Bud production by
vegetative tillers (Vv) and emergence from the bud bank (Gb)
were consistently important for affecting changes in the population growth rate of plant parts for both species and for S. heterolepis in both years. In S. heterolepis, the probability of flowering
(Gv), also contributed greatly to the difference in λ between treatments in 2004, likely because Gv varied among years and was
affected by N addition. Our results support the hypothesis that
bunchgrasses respond to increased resource availability through a
demographic response principally by changing the number of
plant parts within a genet rather than by changing tiller growth
and size. Additionally, our results demonstrate that the bud bank
is crucial for plant responses to supplemental nutrients.
Neither species responded to N addition by increasing bud
natality (the number of buds per tiller). In addition, bud natality
was relatively constant within a species and between years for
S. heterolepis. Therefore, bud banks within a genet increased
solely as a function of the number of existing buds activated
and the number of tillers produced. Fertilized genets with higher
recruitment from the bud bank, such as S. heterolepis genets,
ended the season with a larger bud bank because they recruited
more tillers than control genets. While the pattern of increased
tiller recruitment with increased nutrient availability has been
demonstrated for other graminoids (Noble et al., 1979; Briske
and Butler, 1989; Derner and Briske, 1999), our study demonstrates how the demography of belowground bud banks drives
the observed aboveground tiller population responses. Understanding bud bank dynamics complements observation of
aboveground tiller responses because increased tiller recruitment may not always lead to increases in the bud bank. For
example, grazing has been shown to increase tiller recruitment
in tallgrass prairie grasses, likely as a compensatory growth
mechanism (Vinton and Hartnett, 1992). However, long-term
grazing leads to a depauperate bud bank compared to prairie
that is not grazed (H. Dalgleish and D. Hartnett, unpublished
manuscript). Tillers may be unable to replenish the bud bank
because of lack of resources in times of nutrient stress or drought
or because carbon is allocated to regrowth of photosynthetic
tissue during compensatory growth after grazing.
There was no evidence for a trade-off between sexual and
vegetative reproduction, though flowering tillers tended to produce fewer belowground buds per tiller in 2005. This same
trend was observed in 2004 for tillers that did not receive additional N. The trend was reversed, however, for tillers that received N; flowering tillers tended to increase vegetative bud
production when, presumably, neither N nor water was limited
by the timing of precipitation in 2004. The trade-off between
reproduction and growth is a foundation of plant life-history
theory, but there is little empirical evidence for such a trade-off
(Reekie and Avila-Sakar, 2005). Our study provides only
limited support for a trade-off between sexual fecundity and
vegetative reproduction or growth. Further experiments with
increased sample sizes of flowering tillers will be necessary to
tease apart the effects of N and water availability on individual
tiller allocation patterns.
Between-year differences: The potential effects of water
availability— Emergence from the bud bank and the probability
of flowering were significantly lower in 2005 than in 2004 for
S. heterolepis. Similarly, tiller population λ was much lower in
2005, and differences in emergence from the bud bank contributed most to the difference in λ between the years. The observed
678
American Journal of Botany
[Vol. 95
differences between 2004 and 2005 for S. heterolepis could be
a response to water as an important limiting resource in tallgrass prairie. The total rainfall in both years was average or
slightly above, but the timing of precipitation events in 2005
was such that the plants were likely water stressed during critical life history stages. Because the experiment was conducted
on plants in the same location and at the same time of the year,
the difference in water availability could have been the driving
force behind the annual variation. If the different response between years was due to differences in water availability, then
our data indicate that the response of S. heterolepis to water
availability was mediated through the bud bank. The LTRE
analysis indicated that probability of emergence (Gb) contributed most to the difference in λ between years. However, further research to manipulate water during the growing season is
required to test the hypothesis that S. heterolepis and other
grasses respond demographically to water availability before
we can make general conclusions about plant responses to
water.
Between-species differences: Phenology, photosynthetic
pathway, reproductive strategies— Sporobolus heterolepis and
K. macrantha are both subdominant tallgrass prairie bunchgrasses. In addition to differing phenologies and photosynthetic
pathways, this study provides evidence that S. heterolepis and
K. macrantha may have different reproductive strategies as
well. Regardless of treatment or year, S. heterolepis produced
an average of three times as many buds per tiller than K. macrantha, which produced barely enough buds per tiller to replace
the current tiller population within a genet. On the other hand,
K. macrantha tillers had a much higher probability of flowering
compared to S. heterolepis tillers, and the proportion of tillers
that flowered within a genet was much greater (~5% for S. heterolepis, >50% for K. macrantha).
Coexisting perennial grass species can vary greatly in life
history characters such as longevity, sexual and vegetative reproductive effort, dispersal, and patterns of growth (O’Connor,
1991). Understanding the relative contribution and importance
of sexual and vegetative reproduction in clonal species is important for understanding patterns of genetic diversity, as well
as spatial and temporal dynamics of populations (Benson et al.,
2004). Eriksson (1997) proposed a continuum of seedling recruitment strategies in clonal plants, from initial seed recruitment within a population with continued maintenance through
vegetative reproduction, to repeated (though perhaps low) seedling recruitment. In 40% of the 68 clonal species examined by
Eriksson (1989), seedling recruitment contributed to population
growth rates. A larger comparative demography study involving many more species of rhizomatous and bunchgrasses would
allow for generalizations to be made about the effects of growth
form, photosynthetic pathway, and sexual and vegetative reproduction on the population dynamics of grass species.
Fig. 3. Contributions to the difference in λ between treatments for the
one-way life table response experiments (LTRE) of Sporobolus heterolepis. (A) LTRE between the control and N addition treatments in 2004; (B)
LTRE between the control and N addition treatments in 2005; (C) LTRE
between years 2004 and 2005. Note the difference in scale in (B). C > +N
indicates that for contribution values in this region, the vital rate was
greater in the control than the nitrogen addition treatment and vice versa.
Linking organismal, community, and ecosystem processes— Understanding a species’ ability to mobilize the bud
bank in response to a resource pulse, such as N or an unseasonable rain event, is potentially important for predicting plant responses to community changes such as exotic species invasion.
Davis et al. (2000) developed a general theory arguing that the
fluctuation in resource availability is the key factor controlling
invasibility and that successful invasion events occur intermittently
when specific conditions of resource enrichment or release coincide with adequate propagule supply of the invasive species.
June 2008]
Dalgleish et al.—Bud banks and plant responses to nitrogen
Thus, invasions will occur during windows of opportunity when
propagules of the invader can capitalize on newly available limiting resources. If resident species have the capacity to rapidly
preempt and monopolize new resources, such as through rapid
recruitment from the bud bank, demographic mechanisms could
underlie invasion-resistant communities (Davis et al., 2000).
Our results support the hypothesis that tallgrass prairie plants
are capable of rapidly responding to resource pulses through
increasing recruitment from their bud banks. An important implication of this study is that grasslands with many plants that
maintain large bud banks may be more resistant to invasion, a
hypothesis currently under investigation (D. Hartnett and H.
Dalgleish, unpublished manuscript).
Our analyses support the hypothesis that the production potential of a genet, and hence the grass community, is driven by the
demographic dynamics of its belowground bud bank (Murphy
and Briske, 1992). Because neither S. heterolepis nor K. macrantha increased tiller size in response to N addition, an increase in
biomass production by genets was achieved solely through increases in tiller number. These results contrast with a previous
study of P. virgatum in which N addition resulted in increases in
tiller size (Hartnett, 1993). Identifying the underlying physiological and demographic mechanisms and the sources of variation in
ANPP will be critical to the development of accurate predictive
models of ecosystem responses to environmental change. Most
current models for predicting ANPP in grasslands (e.g., the models CENTURY and SOILWAT) and other terrestrial systems
(Parton et al., 1987) are based on underlying physiological responses to resources at the canopy level (e.g., photosynthesis, C3
vs. C4 physiology and aboveground vegetative growth) rather
than potentially important demographic mechanisms (e.g., bud
natality, survivorship, densities, and tillering dynamics). Our
data demonstrate that the demography of bud bank populations
can play an important role in increases in ANPP in response to
resource availability. Expanding and modifying our study for the
dominant rhizomatous grasses such as Andropogon gerardii and
Sorghastrum nutans is an important next step to developing a
more complete understanding of the importance of bud banks in
driving responses of the tallgrass prairie community to resource
availability. Investigations of bud bank dynamics in grasslands
may lead to better predictive models of productivity and potential grassland responses to environmental change such as altered
precipitation and N deposition.
Conclusions— Our study demonstrates that demographic
mechanisms of the bud bank are important for driving grass
responses to resource availability in S. heterolepis and K. macrantha. In both species of bunchgrass, perturbations in transitions involving the bud bank had the greatest potential to affect
tiller population growth rates. Understanding bud bank dynamics has important implications far beyond the individual plant.
The maintenance of a bud bank influences the dynamics of the
entire population, plays a role in plant species coexistence, contributes to the invasibility of a community, and influences ecosystem productivity. Enhanced knowledge of the bud bank will
lead to a better mechanistic and predictive understanding of
grassland dynamics.
LITERATURE CITED
Benson, E. J., and D. C. Hartnett. 2006. The role of seed and vegetative reproduction in plant recruitment and demography in tallgrass
prairie. Plant Ecology 187: 163–178.
679
Benson, E. J., D. C. Hartnett, and K. H. Mann. 2004. Belowground
bud banks and meristem limitation in tallgrass prairie plant populations. American Journal of Botany 91: 416–421.
Briske, D. D., and J. L. Butler. 1989. Density-dependent regulation of
ramet populations within the bunchgrass Schizachyrium scoparium:
Interclonal versus intraclonal interference. Journal of Ecology 77:
963–974.
Briske, D. D., and J. D. Derner. 1998. The clonal biology of caespitose
grasses. In G. P. Cheplick [ed.], Population ecology of grasses, 106–
135. Cambridge University Press, Cambridge, UK.
Callaham, M. A. Jr., J. M. Blair, T. C. Todd, D. J. Kitchen, and M.
R. Whiles. 2003. Macroinvertebrates in North American tallgrass
prairie soils: Effects of fire, mowing, and fertilization on density and
biomass. Soil Biology & Biochemistry 35: 1079–1093.
Caswell, H. 2001. Matrix population models: construction, analysis, and
interpretation. Sinauer, Sunderland, Massachusetts, USA.
Clark, C. M., E. E. Cleland, S. L. Collins, J. E. Fargione, L. Gough,
K. L. Gross, S. C. Pennings, K. N. Suding, and J. B. Grace.
2007. Environmental and plant community determinants of species
loss following nitrogen enrichment. Ecology Letters 10: 596–607.
Dalgleish, H. J. 2007. Belowground bud banks as regulators of grassland
dynamics. Ph.D. dissertation. Kansas State University, Manhattan,
Kansas, USA.
Dalgleish, H. J., and D. C. Hartnett. 2006. Below-ground bud banks
increase along a precipitation gradient of the North American Great
Plains: A test of the meristem limitation hypothesis. New Phytologist
171: 81–89.
Davis, M. A., J. P. Grime, and K. Thompson. 2000. Fluctuating resources
in plant communities: A general theory of invasibility. Journal of
Ecology 88: 528–534.
Derner, J. D., and D. D. Briske. 1999. Intraclonal regulation in a perennial caespitose grass: A field evaluation of above- and below-ground
resource availability. Journal of Ecology 87: 737–747.
Eriksson, O. 1989. Seedling dynamics and life histories in clonal plants.
Oikos 55: 231–238.
Eriksson, O. 1997. Clonal life histories and the evolution of seed recruitment. In H. de Kroon and J. van Groenendael [eds.], The ecology and
evolution of clonal plants, 211–226. Backhuys Publishers, Leiden,
Netherlands.
Harper, J. L. 1977. Population biology of plants. Academic Press, New
York, New York, USA.
Hartnett, D. C. 1993. Regulation of clonal growth and dynamics of
Panicum virgatum (Poaceae) in tallgrass prairie: Effects of neighbor removal and nutrient addition. American Journal of Botany 80:
1114–1120.
Hartnett, D. C., and P. A. Fay. 1998. Plant populations: Patterns and processes. In A. K. Knapp, J. M. Briggs, D. C. Hartnett, and S. L. Collins
[eds.], Grassland dynamics: Long-term ecological research in tallgrass
prairie, 81–100. Oxford University Press, New York, New York, USA.
Hendrickson, J. R., and D. D. Briske. 1997. Axillary bud banks of two
semiarid perennial grasses: Occurrence, longevity, and contribution to
population persistence. Oecologia 110: 584–591.
Horvitz, C., D. W. Schemske, and H. Caswell. 1997. The relative
“importance” of life-history stages to population growth: Prospective
and retrospective analysis. In S. Tuljapurkar and H. Caswell [eds.],
Structure-population models in marine, terrestrial, and freshwater systems, 247–271. Chapman and Hall, New York, New York, USA.
Matlab. 2002. MathWorks, Inc. Matlab software: Changes and enhancements through version 6.5, release 13. The MathWorks, Natick,
Massachusetts, USA.
Murphy, J. S., and D. D. Briske. 1992. Regulation of tillering by apical
dominance: Chronology, interpretive value, and current perspectives.
Journal of Range Management 45: 419–429.
Noble, J. C., A. D. Bell, and J. L. Harper. 1979. The population
biology of plants with clonal growth. I. The morphology and
structural demography of Carex arenaria. Journal of Ecology 67:
983–1008.
O’Connor, T. G. 1991. Local extinction in perennial grasslands: Lifehistory approach. American Naturalist 137: 753–773.
680
American Journal of Botany
Parton, W. J., D. S. Schimel, C. V. Cole, and D. S. Ojima.
1987. Analysis of factors controlling soil organic matter levels in
Great Plains grasslands. Soil Science Society of America Journal 51:
1173–1179.
Reekie, E. G., and G. Avila-Sakar. 2005. The shape of the trade-off
function between reproduction and growth. In E. G. Reekie and F. A.
Bazzaz [eds.], Reproductive allocation in plants, 189–214. Elsevier
Academic Press, San Diego, California, USA.
SAS. 2003. SAS/STAT software: Changes and enhancements through release 9.1, SAS Institute, Cary, North Carolina, USA.
Seastedt, T. R., B. P. Hayden, C. E. Owensby, and A. K. Knapp.
1998. Climate change, elevated CO2, and predictive modeling: Past
and future climate change scenarios for the tallgrass prairie. In A. K.
Knapp, J. M. Briggs, D. C. Hartnett, and S. L. Collins [eds.], Grassland
dynamics: Long-term ecological research in tallgrass prairie. Oxford
University Press, New York, New York, USA.
Seastedt, T. R., and A. K. Knapp. 1993. Consequences of nonequilibrium resource availability across multiple time scales: The transient
maxima hypothesis. American Naturalist 141: 621–633.
Tomlinson, K. W., and T. G. O’Connor. 2004. Control of tiller recruitment in bunchgrasses: Uniting physiology and ecology. Functional
Ecology 18: 489–496.
Towne, E. G. 2002. Vascular plants of Konza Prairie Biological Station:
An annotated checklist of species in a Kansas tallgrass prairie. Sida
20: 269–294.
Vinton, M. A., and D. C. Hartnett. 1992. Effects of bison grazing on
Androgopon gerardii and Panicum virgatum in burned and unburned
tallgrass prairie. Oecologia 90: 374–382.
Vitousek, P. M., J. D. Aber, R. Howarth, G. E. Likens, P. A. Matson,
D. Schindler, W. H. Schlesinger, and D. Tilman. 1997. Human
alteration of the global nitrogen cycle: Sources and consequences.
Ecological Applications 7: 737–750.