Interspecific differences in above-ground growth patterns result in spatial and temporal partitioning of light among species in a tall-grass meadow

Authors

  • Niels P. R. Anten,

    Corresponding author
       Present address and correspondence: Niels Anten, Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA (fax 650 7236132; e-mail nanten1@leland.stanford.edu).
    Search for more papers by this author
  • Tadaki Hirose

    1. Biological Institute, Graduate School of Science, Tohoku University, Aoba, Sendai 980–77, Japan
    Search for more papers by this author

 Present address and correspondence: Niels Anten, Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA (fax 650 7236132; e-mail nanten1@leland.stanford.edu).

Abstract

1 We compared allometric growth patterns, canopy structure and light interception for individual shoots of different species in a tall-grass meadow.

2 The vertical distributions of above-ground biomass, leaf area, height and leaf angles were measured, both early and late in the season, for individual shoots of Miscanthus sinensis (the dominant species), Lespedeza bicolor, Lysimachia clethroides, Astilbe thunbergii and Potentilla freyniana. A canopy model was developed to calculate light absorption by individual shoots. Light absorption per unit mass (Φmass) was used to quantify the efficiency with which plants utilized biomass to capture light.

3 The leaf mass ratio (LMR), average specific leaf area (SLA) and therefore the leaf area ratio (LAR) decreased with shoot height and light availability. Light absorption per unit leaf area (Φarea) increased with shoot height, and this increase was observed to be much stronger at greater than at smaller shoot heights.

4 In the taller species (Miscanthus and Lespedeza) Φmass (the product of LAR and Φarea) increased, while in the shortest species (Potentilla) it decreased with shoot height. Clones of Miscanthus and Lespedeza may thus increase total light capture by allocating shoot biomass among fewer taller shoots, and a clone of Potentilla by producing a larger number of shorter shoots.

5 Shoots of the shorter species were equally efficient in capturing light (i.e. they had similar Φmass) early in the season, but less efficient later in the season than shoots of the taller species. Shorter species appear to be able to use the earlier part of the season for efficient light capture, while shoots of taller species gain an advantage from their height later in the season.

6 This study shows how different above-ground growth patterns of the species in a tall-grass meadow allow them to use different positions in vertical space and different periods of the season to absorb light efficiently. This is a clear example of niche separation and helps to explain the coexistence of these species.

Introduction

Tall-grass meadows are herbaceous plant communities that are generally a mixture of the dominant tall-grass species, which make up the largest fraction of the total biomass, and shorter species, which are mostly forbs. In such communities, high plant species diversity results from a large number of forb species rather than the presence of a large number of grass species ( Grime 1974, 1979; Turner & Knapp 1996). It is therefore interesting to analyse the mechanisms that allow these forbs to coexist with the grasses. Mechanisms of species coexistence in plant communities are often explained by the theory of niche separation, which states that different species in a community exhibit contrasting structural, phenological or physiological characteristics that allow them to partition resources among themselves on a spatial and/or temporal basis ( Grime 1974, 1979; Kemp & Williams 1980). Here, we focus on interspecific differences in the utilization of above-ground biomass in relation to the acquisition of light.

In stands of vegetation, photon flux density (PFD) decreases exponentially with increasing depth within the canopy ( Monsi & Saeki 1953). Taller dominant species will consequently have a dual advantage in that they can both project their leaves in the highest positions of the canopy, where they receive the highest PFD, and simultaneously shade shorter subordinate species. However, as height increases, plants have to invest a disproportionate amount of biomass in support tissue (i.e. stems, branches and petioles) ( McMahon 1973) and, as a consequence, leaf mass per unit of total above-ground mass (leaf mass ratio; LMR) generally decreases with increasing plant height ( Givnish 1982; Stutzel et al. 1988 ). In addition, leaf area per unit leaf mass (specific leaf area; SLA) is generally found to decrease with increasing PFD ( Dijkstra 1989; Anten & Werger 1996). Thus, leaf area per unit of above-ground mass (leaf area ratio; LAR), which is the product of LMR and SLA, can be expected to decrease as plant height and light availability per unit leaf area increase.

Hirose & Werger (1995) developed a method to compare above-ground biomass allocation and light capture between dominant and subordinate species in a grass stand. They calculated light absorption per unit above-ground biomass (Φmass), the product of the average light absorption per unit leaf area (Φarea) and LAR, and considered Φmass to be a measure of the efficiency of biomass investment to capture light. Surprisingly, they found the short subordinate species to have a slightly higher Φmass than the tall dominant species: the considerably higher LAR values for subordinate species more than compensated for their lower Φarea. Apparently, their contrasting morphologies enable dominant and subordinate species to utilize different positions in the canopy to absorb light with almost equal efficiencies ( Hirose & Werger 1995).

Hirose & Werger (1995) focused on species and not on individual plants within a species, i.e. they determined average LAR and Φmass per species. However, the allometric relationships between the above-ground biomass, height and projected leaf area of plants which, as noted above, are crucial in determining their ability to capture light, differ considerably both within and between species ( Corré 1983; Weiner & Thomas 1992). Within-species variation in plant size will therefore have different consequences for light interception in different species.

Light capture is not only determined by the height and total leaf area of a plant but also by the geometric arrangement of its leaf area. For example, because vertical components of radiation predominate over horizontal ones, incident PFD will be higher on horizontally than on vertically inclined leaves ( Monsi & Saeki 1953). Forbs generally have more horizontally orientated leaves than grasses and this contributes to their ability to intercept light efficiently ( Monsi & Saeki 1953; Anten et al. 1995 ). Plants that arrange their leaf area in a planar array with minimal leaf overlap, rather than having it more evenly distributed along the length of the shoot, will have an advantage in light capture in that they avoid self-shading ( Horn 1971). However, such a leaf arrangement requires a considerable investment of biomass in horizontal branches and this biomass cannot then be invested in height growth or leaf production ( King 1981; Givnish 1995; Aiba & Kohyama 1997). It is therefore generally assumed that shade-tolerant species (those growing in the lowest and darkest layers of a vegetation) should produce their leaf area in a single layer, while sun-adapted species should produce their leaves more along the length of the stem ( King 1981; Givnish 1995). Some degree of self-shading may in fact confer an advantage in high light environments in that it reduces water loss ( Givnish 1995).

Light availability in the lower layers of a canopy decreases during the course of growth as the leaf area index (LAI) of the dominant species increases. Consequently, if subordinate species produce their leaves early in the season, they may increase light acquisition by escaping the main impact of the dominant species. However, high susceptibility to photo-inhibition leads to a strong reduction of both photosynthesis and growth in many shade-adapted plants in high light ( Watling et al. 1997 ). Such species are probably unable to make efficient use of the high light conditions early in the season.

In this study we analysed (i) whether above-ground allometric growth patterns and canopy structure (i.e. leaf angles and vertical leaf area distribution) differ among species in a tall-grass meadow, and (ii) whether these differences allow the species to utilize different positions in the canopy and/or different parts of the growing season to invest biomass efficiently in order to capture light (expressed as Φmass). For this purpose, we developed a model for light distribution with which we could estimate light absorption of individual shoots as a function of their position in the canopy and their leaf area and leaf angle distributions. With this model we analysed light absorption of individual shoots of the component species in a tall-grass meadow dominated by Miscanthus sinensis, both early and late in the season. We focused on individual shoots rather than plants (in clonal species a plant may consist of more than one shoot) because shoots are the most apparently distinguishable above-ground unit. The results are discussed in relation to the theory of niche separation.

Model

The model developed here is an extension of those developed for light distribution in canopies of uniform stands ( Goudriaan 1977, 1988; Gates 1980; Spitters 1986; Anten 1997). Modifications were made to accommodate shoots of different species and of different height. The model divides the canopy into horizontal layers. Within each layer, different classes of foliage are defined, in order to represent the foliage at that height of shoots of different species. The foliage classes are allowed to have different leaf angle distributions and leaf absorbance.

Total light absorption by individual shoots

For a particular height class j, the total light absorbed by a shoot of a given species (Φj) is found as:

image(eqn 1)

where Φij is the light absorbed by the foliage of shoots of class j that are found in the ith canopy layer (counted from the top to the bottom), and nj is the number of shoots in class j. Each value of Φij has two components:

image(eqn 2)

where Φdr,ij and Φdf,ij are the amounts of direct and diffuse light absorbed by foliage class ij, respectively. Direct light is incident on the foliage with a single angle equal to the angle of the sun, but the fact that diffuse light is incident under various angles considerably complicates the calculation of its distribution in canopies. For direct light, Φdr,ij can be found as:

image(eqn 3)

where Φdr,i is direct light absorbed by layer i, fij the LAI (one-sided leaf area per unit ground area) of foliage class ij, and Kdr,j and αj the extinction coefficient of ‘black’ non-scattering leaves for direct light and the leaf absorbance of class j, respectively. The 0.5 power in the term αj0.5 is used to model the effects of leaf reflectance and transmittance on the light climate in the canopy according to Goudriaan (1977).

Φdr,i can be quantified according to Spitters (1986):

image(eqn 4)

where Iodr,i is the direct PFD at the top of layer i and is found by subtracting the light absorbed by layer i − 1 from the PFD at the top of that layer [i.e. Iodr,i = Idr,(i−1) − Φdr,(i−1)].

Goudriaan (1977) showed that the calculation for diffuse light can be simplified by assuming that it can be represented by a summation of radiation components each of which originates from a different ring zone of the sky, and has an angle of incidence equal to the zone elevation angle. Absorption of diffuse light by plants has to be calculated for each component separately ( Goudriaan 1988). The total diffuse light absorbed by foliage class ijdf,ij) is thus calculated as:

image(eqn 5)

where Φdf,ijk is the absorbed diffuse light originating from sky zone k. Φdf,ijk can be found as:

image(eqn 6)

where Φdf,ik is the amount of diffuse light originating from sky zone k which is absorbed by layer i, and Kdf,jk is the extinction coefficient of ‘black’ non-scattering leaves of plant class j for this radiation component. Φdf,ik is quantified according to Spitters (1986):

image(eqn 7)

where Iodf,ik is the diffuse PFD from the kth sky zone at the top of layer i and is found by subtracting the photon flux absorbed by layer i– 1 from the PFD at the top of that layer [i.e. Iodf,ik = Iodf,(i−1)k − Φdf,(i−1)k].

The values above the canopy (Iodf,1k as the top layer in the canopy has i = 1) are found as:

image(eqn 8)

with Iodf,1 the total diffuse PFD above the canopy and airk the fraction of the total diffuse PFD coming from sky zone k. Following Goudriaan (1988), we distinguish three 30° sky zones (k = 3) and the values of airk are calculated according to his standard overcast sky ( Goudriaan 1977).

Extinction coefficients and leaf absorbance

The extinction coefficient for direct light (Kdr,j) for a given species is calculated as a function of the solar elevation angle and the leaf angle distribution, which is allowed to vary between species. This is done following Goudriaan (1988) (but see equations 3–5 in Anten 1997) by assigning leaves to three inclination classes (0–30°, 30–60° and 60–90°) and assuming all leaves in a class to have an inclination angle equal to the centre angle of that class (i.e. 15°, 45° and 75°, respectively). The extinction coefficients for diffuse light (Kdf,jk) are calculated in the same way, with the solar elevation angle being replaced by sky zone elevation angles.

Leaf absorbance (αj) is calculated following Evans (1993):

image(eqn 9)

where chlj is the average chlorophyll content per unit leaf area (μmol m−2) for shoots of class j.

Daily light absorption per unit leaf area and shoot mass

The light absorbed by a shoot of each species × height class j (Φ, note that the subscript j has been omitted) is integrated over the day, from sunrise to sunset, to obtain daily values of light absorption (ΦD). In this calculation, the daily courses of direct and total diffuse PFD above the canopy and the solar inclination angle are calculated as a function of the latitude and the date according to equations 6.27 and 6.31 in Gates (1980) which assume a totally clear day. Part of the PFD incident on the canopy is reflected and is not available for absorption. PFD levels above the canopy are therefore pre-multiplied by one minus the canopy reflectance (τ) and τ is assumed to be 0.05, which comes close to the values calculated by Goudriaan (1977).

The daily light absorption per unit leaf area (Φarea) can be found following Hirose & Werger (1995):

image(eqn 10)

where F denotes the total leaf area of a shoot. Similarly, daily light absorption per unit shoot mass (Φmass) can be found as:

image(eqn 11)

with M being the total mass of a shoot. Φmass can be considered to represent the efficiency of using above-ground biomass to capture light ( Hirose & Werger 1995). The following relationship holds between Φmass and Φarea:

image(eqn 12)

where LAR is the amount of leaf area per unit shoot mass ( Hirose & Werger 1995). The LAR can be divided into its two components, LMR, which is the amount of leaf dry mass per unit shoot mass, and SLA:

image(eqn 13)

Methods

Study area

This study was carried out in a Miscanthus sinensis grassland at the Kawatabi Research Station of the Tohoku University, 60 km north of Sendai (38°44′N, 140°15′E), where the grassland has been maintained by mowing in autumn every other year since 1967. This is one of the most common types of tall-grass meadows in Japan ( Lieth et al. 1973 ). The soils are derived from volcanic ash and have adequate supplies of nitrogen and phosphorus. The mean monthly temperatures and precipitation for June, July and August at the station are 16.8 °C and 187 mm, 21.5 °C and 247 mm, and 22.5 °C and 284 mm, respectively ( Lieth et al. 1973 ).

The stand that was analysed contained a total of 10 species, all of which were perennials. We primarily focused on the five most common species, which also represented a range of structural characteristics: (i) Miscanthus sinensis (Gramineae), a tall (1–2 m) grass that was the dominant species in the stand; (ii) Lespedeza bicolor (Fabaceae), a leguminous forb (although it is often woody at the base) with strongly branching stems, which occupied the middle and upper layers of the stand; (iii) Lysimachia clethroides (Primulaceae), a forb with an erect non-branching stem that was found in the lower to middle layers of the stand; (iv) Astilbe thunbergii (Saxifragaceae), a profusely branching forb that formed relatively broad canopies in the lower to middle layers of the stand; (v) Potentilla freyniana (Rosaceae), a very short forb that occupied the lowest layers of the canopy. The above-ground parts of vegetative Potentilla plants consist of petioles, each of which bears a trifoliate leaf. The shoots of all five focal species are completely erect. The other five species were Aster agretoides, Iris ensata, Rubus parvi, Spodiopogon sibiricus and Viola sp. Nomenclature and description of species follow Ohwi (1965). Hereafter, plant species are referred to by their genus name only.

Canopy structure and light distribution

Canopy structure and the distribution of diffuse PFD were determined on 26 June and 7 August 1996. No plants had started flowering except for a few Lysimachia plants observed on 7 August. A 1 × 1 m quadrat was established and PFD (400–700 nm) was measured at height increments of 10 cm at five replicates points, using an SF80 line sensor (Decagon Devices Ltd, Pullman, WA) under an overcast sky. Reference PFD at the top of the canopy was measured simultaneously using a point sensor (LI190SA; LiCor, Lincoln, NE) connected to a datalogger (LI1000; LiCor). Leaf inclination angles for representative individuals of each species were measured with a protractor. In Miscanthus inclination angles were measured at various points along the length of each leaf because leaves in this species are noticeably curved.

The area of the quadrat was then divided into four 0.5 × 0.5 m subquadrats and all shoots were cut at ground level. The term shoot is used to describe a unit that could be distinguished at ground level, i.e. individual petioles with a trifoliate leaf in the case of Potentilla, and individual stems with leaves in the case of the other species. Several shoots may come from a single individual (i.e. they are part of the same clone). Such pseudo-replication is not, however, a serious problem unless there are significant differences between the clones in the characteristics investigated. This was not the case (see the Results). Harvested shoots were sealed in polythene bags and brought to the laboratory. There they were sorted to species and their height was measured from the base to the highest leaf. Shoots with heights of 0–20 cm were then divided into 2-cm height classes, those of 20–80 cm into 5-cm classes and those taller than 80 cm into 10-cm classes. To fit the model assumptions, the height of a shoot was assumed to be equal to the median height of its class.

Every individual shoot was divided into segments by clipping every 2 cm from the base to a height of 20 cm, every 5 cm from 20 to 80 cm, and every 10 cm above 80 cm, keeping stem petiole and leaf angles as natural as possible. There was a very large number of Miscanthus shoots and height segments from all shoots in a given height class from a given subquadrat were combined. For the other species height segments of individual shoots were kept separate. Leaves were separated from stems, branches, petioles and leaf sheaths, which were all considered to be support tissue. Leaf area was measured with a leaf area meter (LI-3100; LiCor). Dry mass was determined after oven-drying at 70 °C for at least 72 h.

Chlorophyll contents per unit area (chl mmol m−2) were determined with a chlorophyll meter (SPAD-502, Minolta, Osaka, Japan). SPAD values were calibrated with independent measurements of chlorophyll according to Hikosaka (1996). For all species the correspondence between SPAD values and measured chlorophyll contents was very good. with r2 ranging from 0.94 to 0.99.

There were insufficient shoots of Lespedeza, Lysimachia, Astilbe and Potentilla in the quadrats to determine allometric relationships at either date, so any additional shoots of these species that could be found within 1.5 m of the quadrats were also sampled so that n was at least 17. Relative incident PFD on these shoots was measured, after which they were harvested and leaf area and dry mass was measured. These additional shoots were used only as extra data points to determine allometric relationships, and were not included in analyses of light interception with respect to architecture.

Results

Stand characteristics: biomass, leaf area and light distribution

Total above-ground biomass and LAI were about 2.5- and 1.9-fold higher, respectively, for the late (7 August) than for the early (26 June) season stand ( Table 1). Miscanthus constituted the largest part of both the biomass and LAI of the stands, followed by Lespedeza, Astilbe, Lysimachia and Potentilla. Miscanthus also absorbed the largest fraction of total incident PFD. Species other than the five focal ones constituted less than 1% of the total biomass and LAI.

Table 1.  Total above-ground biomass, leaf area index and calculated daily light absorption and their partitioning among species in a tall-grass meadow both early (26 June) and late (7 August) in the season. Values in parentheses denote standard errors (n = 4)
 Dry mass (g m−2) Leaf area (m m−2) Light absorption (mol day−1)
Mean%Mean%Mean%
26 June
Miscanthus151.79 (7.66)84.231.928 (0.108)78.2939.81 (2.23)83.43
Lespedeza7.86 (3.16)4.360.155 (0.059)6.271.89 (0.73)3.97
Lysimachia4.63 (1.11)2.570.099 (0.016)3.881.07 (0.18)2.25
Astilbe15.36 (4.46)8.520.271 (0.145)11.004.66 (2.49)9.78
Potentilla0.44 (0.31)0.240.015 (0.011)0.620.13 (0.09)0.28
Others0.14 (0.12)0.080.004 (0.004)0.180.03 (0.02)0.06
Total180.21 (10.10) 2.468 (0.206) 47.58 (3.973) 
7 August
Miscanthus389.51 (51.66)84.453.099 (0.370)69.2547.51 (5.67)84.51
Lespedeza33.23 (2.10)7.210.678 (0.051)15.136.16 (0.46)10.97
Lysimachia15.05 (3.40)3.260.271 (0.050)6.040.97 (0.18)1.73
Astilbe20.75 (8.66)4.490.348 (0.226)7.771.39 (0.90)2.46
Potentilla1.24 (0.30)0.270.039 (0.007)0.940.09 (0.01)0.16
Others1.55 (0.67)0.330.039 (0.01)0.870.09 (0.02)0.16
Total461.18 (43.39) 4.475 (0.197) 56.18 (2.47) 

The leaf area in both stands was mostly concentrated in the middle layers of the canopy ( Fig. 1a). Measured values of diffuse relative PFD changed very little with increasing height in the lowest layers of the canopy, but increased sharply in the middle to higher layers ( Fig. 1b).

Figure 1.

Vertical distribution of (a) leaf area index and (b) relative diffuse photon flux density (PFD) measured early (26 June) and late (7 August) in the season in the canopy of a tall-grass meadow. Error bars indicate standard errors (n is 4 and 5 for (a) and (b), respectively).

The light model used in this paper was validated by comparing the measured and calculated distributions of relative diffuse PFD in the canopy. Correspondence between the measured (x) and calculated (y) values was very good (for both dates r > 0.99). The regression lines were: y = −0.037 + 1.11x for the early season stand and y = −0.0098 + 1.025x for the late season stand, which indicates that the model slightly overestimated relative diffuse PFD in the upper layers and underestimated it in lower layers of the canopy.

Shoot characteristics: biomass allocation, canopy structure and leaf absorbance

Figure 2 shows the allometric relationship between the height and above-ground biomass of individual shoots. There was a significant species effect on the slope of this relationship ( Table 2) and Potentilla appeared to have the smallest slope ( Table 3). For a given biomass, Astilbe appeared to be shortest while Miscanthus appeared to be tallest ( Fig. 2). This was also indicated by the fact that Astilbe had the second highest and Miscanthus the lowest intercept of the allometric mass vs. height relationship ( Table 3). Miscanthus and Lespedeza shoots in the late season stand tended to be taller, relative to their biomass, than those in the early season stand. No seasonal differences in shoot allometry were observed for the other three species.

Figure 2.

Allometric relationships between above-ground biomass and shoot height for shoots of the five focal species in a tall-grass meadow both early (26 June) and late (7 August) in the season. (○) Miscanthus sinensis; (&U25CF;) Lespedeza bicolor; (□) Lysimachia clethroides; (▪) Astilbe thunbergii; and (◆) Potentilla freyniana. Lines indicate linear regressions (coefficients are given in Table 3). For the four forb species (i.e. all species except Miscanthus) additional shoots were harvested from outside the experimental quadrats and are included in this figure.

Table 2.  Results of two-way analysis of covariance ( ancova) with shoot height and relative diffuse PFD incident on shoots as covariates, and species and harvest date as factors. All values are P-values, * indicates significant effects (P < 0.05) and log that data have been log-transformed
Dependent variableCovariateFactorAmong slopesAmong intercepts
HeightlogMasslogSpecies<0.001*
Harvest 0.952 0.531
Species × harvest 0.020*
LMRlogHeightSpecies 0.019*
Harvest 0.925 0.457
Species × harvest 0.611 0.116
SLAHeightSpecies<0.001*
Harvest 0.534 0.183
Species × harvest 0.124 0.012*
Rel. PFDlogSpecies<0.001*
Harvest 0.715 0.100
Species × harvest 0.321 0.004*
LARlogHeightSpecies<0.001*
Harvest 0.488 0.064
Species × harvest 0.288 0.163
Rel. PFDlogSpecies<0.001*
Harvest 0.828 0.070
Species × harvest 0.370<0.001*
Table 3.  Coefficients of the allometric relationships between shoot height and mass (i.e. log mass = a + b log height)
  Constant (a) Slope (b) r2
Miscanthus26 June−4.172.420.941
7 August−4.812.710.992
Lespedeza26 June−3.152.160.881
7 August−4.192.590.951
Lysimachia26 June−3.812.410.846
7 August−2.971.830.960
Astilbe26 June−2.892.090.892
7 August−2.662.080.876
Potentilla26 June−2.541.220.923
7 August−2.561.410.911

Fractional allocation of above-ground biomass to leaves (LMR), the average specific leaf area (SLA) and the above-ground leaf area ratio (LAR, i.e. the product of SLA and LMR) all decreased with increasing shoot height ( Fig. 3), shoot mass and relative PFD incident on shoots (data not shown). The shortest shoots in the stands had three- to sixfold higher LAR values than the tallest shoots ( Fig. 3e,f). Within each stand, a given shoot height always corresponded to one relative PFD value ( Fig. 1). Consequently, in order to analyse separately the effects of these factors on shoot characteristics, the data from the early and late season stands had to be combined. Combining these data revealed that in all species, except Lespedeza, LMR, SLA and LAR were more strongly correlated with shoot height than with relative PFD or shoot biomass; in Lespedeza both SLA and LAR were most strongly correlated with incident PFD ( Table 4).

Figure 3.

Above-ground leaf mass ratio (LMR; a,b), specific leaf area (SLA; c,d) and above-ground leaf area ratio (LAR; e,f) as a function of shoot height for shoots of the five focal species in a tall-grass meadow both early (26 June; a,c,e) and late (7 August; b,d,f) in the season. Symbols are as given in Fig. 2. Lines indicate exponential curve fits in (a), (b), (e) and (f) and linear fits in (c) and (d). For the four forb species (i.e. all species except Miscanthus) additional shoots were harvested from outside the experimental quadrats and are included in this figure.

Table 4.  Correlation coefficients between leaf mass ratio (LMR), specific leaf area (SLA) and leaf area ratio (LAR) and shoot height, above-ground dry mass and the relative PFD incident on shoots calculated for each species separately and for all species combined (i.e. total). NS indicates that there is no significant correlation (P > 0.05). Note that the data for two harvests have been combined
Dependent
variable
Independent
variable
MiscanthusLespedezaLysimachiaAstilbePotentillaTotal
LMRHeight−0.842−0.727−0.843−0.903−0.759−0.903
Above-ground mass−0.730−0.486−0.673−0.677−0.574−0.569
Rel. PFD−0.416−0.701−0.403−0.518−0.152 NS−0.692
SLAHeight−0.873−0.604−0.867−0.649−0.644−0.755
Above-ground mass−0.271−0.589−0.776−0.540−0.530−0.317
Rel. PFD−0.489−0.717−0.130 NS−0.361−0.209 NS−0.693
LARHeight−0.896−0.729−0.876−0.772−0.800−0.865
Above-ground mass−0.768−0.574−0.739−0.669−0.705−0.517
Rel. PFD−0.536−0.812−0.254 NS−0.413−0.159 NS−0.702

When the data for all the species within each stand were combined, the LMR–height and LAR–height relationships appeared to be negatively exponential, i.e. LMR and LAR decreased more rapidly with shoot height at small shoot heights than at large shoot heights ( Fig. 3). This pattern was also indicated by significantly positive second-order polynomial terms ( Table 5). The same pattern appeared to hold for the LMR–height and LAR–height relationships for individual species, although in these cases the second-order polynomial terms were not always significantly different from zero ( Table 5).

Table 5.  Significance levels of the second-order polynomial terms of the relationship between leaf mass ratio (LMR), leaf area ratio (LAR) and light absorption per unit leaf area (Φarea) and above-ground mass (Φmass), and shoot height. NS denotes non-significant and asterisks significant (*P < 0.05, **P < 0.01, ***P < 0.001) terms. Note that the second-order polynomial terms were always positive, indicating that the relationships are convex
  LMRLARΦareaΦmass
Miscanthus26 June***NSNSNS
7 August************
Lespedeza26 June*NS******
7 August**********
Lysimachia26 JuneNSNS**NS
7 August**********
Astilbe26 June********
7 AugustNSNS***NS
Potentilla26 JuneNSNSNSNS
7 AugustNSNSNSNS
Total stands26 June************
7 August************

ancova revealed a significant species effect on the slopes of the LMR–, SLA– and LAR–shoot height relationships, and also on the slopes of the SLA– and LAR–incident relative PFD relationships ( Table 2). Logarithmic transformation was applied to the values of LMR, LAR and relative PFD in order to obtain linear relationships. Despite the significant species effect, the LMR–height relationship appeared to differ very little between species. Lespedeza shoots had higher SLA and LAR values than shoots of other species of the same height ( Fig. 3).

Figure 4(a) shows the fraction of leaves in each of the three 30° leaf inclination classes (0–30°, 30–60° and 60–90°) as well as the average leaf inclination angle. Potentilla had the most horizontally orientated leaves, with 80% of the leaves in the 0–30° class and an average leaf angle of 20°, followed by Astilbe, Lysimachia and Lespedeza in that order. Miscanthus had the most vertically inclined leaves, with its leaves fairly evenly distributed among the inclination classes and an average leaf angle of 47°.

Figure 4.

(a) Average leaf angle distribution, i.e. fraction of leaves (or leaf segments in the case of Miscanthus) in each of the three 30° leaf inclination classes, as well as the average leaf angle (the number above each set of bars), and the average vertical leaf area distribution, i.e. fraction of leaf area in the lower, middle and upper third of shoots for the early (b) and late season stand (c) for shoots of the five focal species in a tall-grass meadow. Abbreviations on the x-axis are Mi for Miscanthus sinensis, Le for Lespedeza bicolor, Ly for Lysimachia clethroides, As for Astilbe thunbergii and Po for Potentilla freyniana. Error bars indicate standard errors (n = 10 in a and equal to the number of shoots for each species in each stand in b and c).

The average vertical leaf area distribution along the heights of the shoots for each species is shown in Fig. 4(b,c). In Potentilla shoots all leaf area was concentrated in the upper third of the height of the shoots, which was expected because a Potentilla‘shoot’ is actually a petiole with three leaflets at its end. In Astilbe shoots, more than 90% of the leaf area was concentrated in the upper third, while in Lysimachia, Lespedeza and Miscanthus shoots leaf area was distributed much more evenly along the length of the stem. Within species, vertical leaf area distribution did not differ very much between shoots of different height.

Average leaf chlorophyll content varied from 120 mmol m−2 in some of the Potentilla and Astilbe shoots, to 450 mmol m−2 in some of the Miscanthus shoots (data not shown). As a result average leaf absorbance (α, equation 9) ranged from about 0.6 to 0.85 ( Fig. 5).

Figure 5.

Average leaf absorbance as a function of shoot height for shoots of the five focal species in a tall-grass meadow both early (26 June) and late (7 August) in the season. Symbols are as in Fig. 2.

To check that pseudo-replication was not a problem, we re-analysed our data to see if there was a significant clone effect on any of the characteristics investigated. In the four forb species (Lespedeza, Lysimachia, Potentilla and Astilbe) we did not find any significant differences between clones (based on a one-way anova, P > 0.05). Unfortunately, in Miscanthus such an analysis was not possible because we had combined height segments from different shoots. However, three of the four subquadrats in the early season stand and two of the four subquadrats in the late season stand did not share shoots from the same clone. There was no significant difference in any characteristic between these subquadrats (based on a one-way anova and a studentized t-test, respectively, P > 0.05), indicating that there was no difference between clones. The differences reported are therefore statistically valid.

Light absorption per unit leaf area and per unit mass

Average light absorption per unit leaf area (Φarea) increased with increasing shoot height ( Fig. 6a,b). The difference in Φarea between the tallest and shortest shoots in a stand was greater in the late season than in the early season stand. In all species, except Miscanthus and Potentilla in the early season stand, the relationship between Φarea and shoot height was significantly convex ( Table 5). This was also the case when the data of all the shoots in a stand were combined. Astilbe and Potentilla shoots had higher Φarea values than shoots of other species of the same height.

Figure 6.

Daily light absorption per unit leaf area (Φarea; a,b) and per unit above-ground mass (Φmass; c,d) as a function of shoot height for shoots of the five focal species in a tall-grass meadow both early (26 June) and late (7 August) in the season. Symbols are as in Fig. 2. Lines indicate second-order polynomial curve fits. Note the difference in vertical scales for the two dates.

The relationship between light absorption per unit shoot mass (Φmass) and shoot height differed considerably between species. In Miscanthus and Lespedeza there was a strong positive correlation, in Lysimachia and Astilbe no correlation, and in Potentilla a negative correlation between Φmass and shoot height ( Fig. 6c,d and Table 6). When the data were combined per stand, there was no overall correlation between Φmass and shoot height in the early season stand but there was an overall positive correlation in the late season stand ( Table 6). However, plots such as those in Fig. 6 show that in both stands Φmass appeared first to decrease and then to increase with shoot height, with shoots of intermediate height having the lowest Φmass. Second-order polynomial regression revealed that the Φmass–height relationships were significantly convex ( Table 5). In June, in the intermediate height positions of the stand, where Astilbe is found, Φmass values found for this species were considerably higher than for other species and these values were similar to those of the shortest (Potentilla) and tallest (Miscanthus) shoots in the stand. In August, Lespedeza shoots had the highest Φmass values.

Table 6.  Results of linear regressions between light absorption per unit above-ground mass (Φmass) and shoot height; + and − denote positive and negative correlations, respectively, and * indicates significant correlations (P < 0.05)
  SignP-value
Miscanthus26 June+<0.001*
7 August+<0.001*
Lespedeza26 June+ 0.026*
7 August+<0.001*
Lysimachia26 June 0.945
7 August 0.715
Astilbe26 June+ 0.992
7 August 0.520
Potentilla26 June 0.002*
7 August<0.001*
Total stands26 June 0.400
7 August+<0.001*

Discussion

Above-ground biomass allocation

The allometric relationships between above-ground mass and shoot height differed considerably between species. For a given mass, Miscanthus shoots were taller than the shoots of the other species, indicating that they allocated relatively more mass to height growth. Such an allocation pattern can be considered a shade-avoidance strategy typical of shade-intolerant species, whereas a more conservative allocation of biomass to height growth is generally associated with a greater shade tolerance ( King 1981, 1991; Corré 1983). Miscanthus is a C4 species and thus has a higher light requirement for photosynthesis than the forb species, which are C3 ( Black 1971; Kemp & Williams 1980; Anten et al. 1995 ). Astilbe shoots exhibited the greatest degree of branching, forming relatively broad canopies in middle to lower parts of the stands, but for a given above-ground biomass they were shorter than shoots of the other species. These results indicate that there is a trade-off between lateral growth and height growth, which is in accordance with the findings of King (1981) for woody species.

Allocation of mass to leaves (LMR) decreased with increasing shoot height. This result is consistent with various other studies ( Givnish 1982; Menges 1987; Stutzel et al. 1988 ) and indicates that with increasing height plants have to invest a disproportionate amount of biomass in support tissue (stems and branches) in order to maintain mechanical stability ( McMahon 1973; Givnish 1982). Givnish (1982, 1995) suggested that the LMR–height relationship is strongly influenced by the branching patterns of plants. For example, plants that exhibit a strong degree of horizontal branching forming relatively broad canopies, need more support tissue to carry a certain mass of leaves than plants which produce their leaves along the length of the stem, because they have longer more expensive lever arms. However, in the present study, there was surprisingly little difference in the LMR–height relationships between the species. For example, among the forbs Astilbe shoots, which had the greatest degree of horizontal branching, had similar LMR values as Lespedeza or Lysimachia shoots of the same height, which produced their leaves more along the length of the stem. Care should be taken, however, when comparing LMR between Miscanthus and the four forbs. In contrast to dicots, many monocots like Miscanthus produce long leaves that cover a large fraction of the length of the shoot. The lower leaf parts have support and transport functions similar to those of stems ( Hirose et al. 1989 ). Nevertheless, it is interesting that Miscanthus shoots had lower average SLA and also lower leaf area per unit mass values (LAR) than shoots of the forb Lespedeza of the same height but with leaves that did not have a similar support function.

SLA decreased with both shoot height and light availability but, with the exception of Lespedeza, it was more strongly correlated with the former than with the latter. This result is in contrast with most previous studies ( Dijkstra 1989; Anten & Werger 1996), which suggested light availability to be the predominant factor in control of SLA. However, Lysimachia, Astilbe and Potentilla shoots produced relatively few leaves between the early and late harvests (K. Hikosaka, personal communication). The SLA of these species were probably predominantly determined by the light availability at the time and level at which leaves were formed, and are therefore correlated with shoot height, which continues to reflect their position in the canopy at that time. In the case of Miscanthus (as noted above), lower leaf parts increasingly have a support function as leaf length increases with shoot height, and a negative correlation between shoot height and SLA is therefore to be expected.

Light absorption per unit leaf area and per unit mass

The different species in the tall-grass meadow had different patterns of biomass allocation and used different layers of the canopy to acquire light efficiently. In the tallest species Miscanthus and Lespedeza, light absorption per unit mass (Φmass, a measure of the efficiency of light acquisition in terms of biomass) increased with shoot height, whereas in the shortest species, Potentilla, it decreased. When the data of all the shoots within the stands were combined, Φmass first decreased and then increased with shoot height so that shoots of intermediate height had the lowest Φmass values. Φmass is the product of the LAR and light absorption per unit area (Φarea) and this result reflects the convex relationships of LAR and Φarea with shoot height. LAR decreased with shoot height but the reduction was stronger at smaller than at greater height. Φarea, on the other hand, increased with shoot height, and this increase became stronger with increasing height. Thus, in Potentilla, LAR decreased very strongly while Φarea increased little over its height range, resulting in the observed reduction in Φmass. By contrast, in Miscanthus and Lespedeza, Φarea increased strongly over their height range overriding the relatively small reduction in LAR.

Astilbe in the early season stand was somewhat exceptional in that it produced shoots of intermediate height that had considerably higher Φmass values than shoots of other species of the same height. This result can be attributed to the high Φarea of these shoots, which resulted from the fact that they produced broad canopies with horizontal leaves that were concentrated in a single layer. This was in contrast to Lespedeza, Lysimachia and Miscanthus, which distributed their leaves more evenly along the length of the stem, while Miscanthus also had more vertical leaves than any of the four forb species. The latter is a commonly observed difference between grasses and forbs ( Monsi & Saeki 1953; Anten et al. 1995 ). As noted in the introduction, the projection of leaves in a single layer towards the top of the shoot minimizes self-shading and may reduce shading by neighbouring shoots. In addition, incident PFD is higher on horizontal leaves than on vertical leaves. In this context, it is interesting to note that model calculations in which Astilbe shoots were assumed to have leaf area distributions and leaf angles similar to those of Miscanthus resulted in 30–50% lower Φarea and Φmass values than the observed values for Astilbe reported here.

The importance of these interspecific differences in the relationship between Φmass and shoot height for species coexistence can be understood if we assume that an individual clone of a given species has a limited amount of biomass to allocate to shoots, and that total light capture is the product of Φmass and total mass allocated to shoots. Miscanthus and Lespedeza clones will increase their total light capture by producing a small number of taller shoots rather than a large number of shorter shoots. By contrast, Potentilla clones will absorb more light by producing a large number of shorter shoots. A third possibility is to produce relatively few shoots of intermediate height with broad canopies in which leaf area can be displayed in a single layer, thus enhancing light capture, as seen in the case of Astilbe. So, different species in a tall-grass meadow can maximize their light capture by occupying different layers in the canopy.

In the early season stand, shoots of the dominant species Miscanthus absorbed similar amounts of light per unit mass (Φmass) to the shorter subordinate forbs, while in the late season stand both Miscanthus and the tallest forb Lespedeza had somewhat higher Φmass values than the shortest species. Total light absorption (Φmass × shoot mass; data not shown) of the three subordinate forbs was higher early rather than late in the season, whereas Miscanthus and Lespedeza shoots absorbed more light late rather than early in the season. The difference in the result between the early and late season stands can be attributed to the higher total LAI (of which Miscanthus constituted the largest part) of the late season stand. Light availability and light absorption per unit area (Φarea) of shoots in the lower layers of the stand were consequently much lower and the higher LAR of these shoots did not compensate for their lower Φarea. This result could be typical for tall-grass meadows. Because these are communities of mainly herbaceous plants, all above-ground parts are produced during the course of a growing season. Consequently, it takes time for dominant species to produce a large and tall canopy, and light availability in the lower layers is considerably higher early rather than late in the season. In the present study, shoot growth of the subordinate species mainly occurred in the earlier part of the season. Thus, a direct consequence of the differences in shoot size and biomass allocation between dominant and subordinate species is a temporal partitioning of light acquisition between them.

Concluding statement

The results in this study indicate how different species in a tall-grass meadow have different patterns of above-ground biomass allocation, canopy structures and phenologies, allowing them to use different positions in vertical space and different periods of the season to absorb light efficiently. This is a clear example of niche separation ( Grime 1974, 1979; Kemp & Williams 1980) and helps to explain how these species are able to coexist.

Acknowledgements

We thank David Ackerly, Kouki Hikosaka, Takahide Kurosawa, Katherine Preston and Marinus Werger for valuable comments on the manuscript, and Cindy Kranendonk and Nanchin for technical assistance. This work was supported by grant P-95091 from the Japan Society for the Promotion of Science (JSPS) and by a grant-in-aid from the Japan Ministry of Education, Science and Culture. N.P.R.A. also wishes to acknowledge National Science Foundation grant IBN 96-04030.

Received 17 November 1997revision accepted 24 November 1998

Ancillary

Advertisement