Annals of Botany 83 : 11–18, 1999 Article No. anbo.1998.0782, available on line at http:\\www.idealibrary.com Theoretical Relationships Between Mean Plant Size, Size Distribution and Self Thinning under One-sided Competition KIHACHIRO KIKUZAWA Center for Ecological Research, Kyoto Uniersity, Sakyoku, Kyoto, 606-8502, Japan Received : 28 October 1997 Returned for revision : 23 February 1998 Accepted : 7 September 1998 As yet there is no comprehensive theory in plant population ecology to explain relationships between mean plant size, size distribution and self-thinning. In this paper, a new synthesis of plant monocultures is proposed. If the reciprocal relationship between plant biomass and plant population density among various stands of even-aged plant populations holds, the same reciprocal relationship must exist between cumulative mass and cumulative number of plants from the largest individual within a population, assuming strict one-sided competition (which is an extreme condition for competition for light among plants). The two parameters of the relationship between cumulative mass and cumulative number within a stand both correlate with maximum plant height in the stand. One parameter equals the reciprocal of the potential maximum plant mass per area, which is expressed by the product of maximum plant height and dry-matter density. The other parameter correlates with the potential maximum individual plant mass, which is allometrically related to maximum plant height. As a stand develops, the growth rate of the smallest individuals will become zero due to suppression from larger individuals, and they will die ; i.e. self-thinning will occur. The slope of the self-thinning line is expressed through the coefficients of allometry between height and mass and between dry matter density and height. When the former coefficient is 3 and the latter is 0, the gradient exactly corresponds to the value expected from the 3\2 power rule, but it can take various values depending on the values of the two coefficients. Competition among individuals determines size-density relationships among stands, which in turn determine the size structure of the stand. The size structure constrains the growth of individuals and results in self-thinning within the stand. # 1999 Annals of Botany Company Key words : Monoculture, plant population, self-thinning, competition, hierarchy, size-structure. INTRODUCTION There are several general quantitative relationships in plant population ecology : populations of the same species growing in similar geographical regions converge onto a constant leaf mass per unit land area (Tadaki, 1977) and a constant dry-matter density per unit volume of forest (Kira and Shidei, 1967) after canopy closure. That leaf biomass of a particular plant species per unit land area shows a more or less constant value suggesting that the amount of light energy which is shed on the land constrains the amount of leaves. Dry-matter density is the amount of above-ground organic matter per unit volume of forest ; it too is considered more or less constant (Kira and Shidei, 1967) irrespective of stand age. This trend implies the existence of an upper limit where organic matter is fully packed in the unit volume of a forest. Three types of relationships have been studied in evenaged plant monocultures (Li, Watkinson and Hara, 1996) : (1) the competition-density effect ; (2) self-thinning ; and (3) size structure. Mean plant size is related to the reciprocal of population density (Kira, Ogawa and Sakazaki, 1953 ; Shinozaki and Kira, 1956). This formulation is supported by many authors (Bleasdale and Nelder, 1960 ; Watkinson, 1984), with some modifications (Vandermeer, 1984). In a crowded but actively growing stand, natural thinning usually occurs while mean plant size increases (mean plant size is usually expressed by a power function of plant density on 0305-7364\99\010011j08 $30.00\0 unit land area). Yoda et al. (1936) reported that the exponent (the slope of the log-log plot) between population density and mean plant size takes the value of k3\2, and this thinning line was designated as the ‘ k3\2 power law ’ of natural thinning. Many empirical data sets of mean plant size and plant density have supported this rule (White and Harper, 1970 ; Westoby, 1984). The power value of 3\2 was first derived by the comparison of the dimensions of mean plant size and the reciprocal of allotted area to a mean individual (Yoda et al., 1963). This basic interpretation was extended to allometric relationships between mean plant mass and mean leaf mass (White, 1981 ; Osawa and Allen, 1993). However, Weller (1987) re-evaluated the law and found many cases in which the gradient of the line deviated from k3\2, casting doubt on the existence of a constant slope. Research on self-thinning should address the factors that determine whether the slope is k3\2 or some other value, rather than whether or not there is a ‘ law ’ (Lonsdale, 1990 ; Kikuzawa, 1993 ; Osawa and Allen, 1993 ; Hamilton, Matthew and Lemaire, 1995 ; Armstrong, 1997). Although mean population behaviour does not consider differences among individuals, size often varies considerably between individuals within a stand, even if individuals are the same age. These size differences among individuals may be the result of genetic differences in growth rate or be due to chance differences e.g. in the timing of seed-emergence (Weiner and Thomas, 1986). Initial small differences can be accelerated by competition. Competition among individuals # 1999 Annals of Botany Company 12 Kikuzawa—Hierarchy and Self-thinning for light is considered to be one-sided (Cannell, Rothery and Ford, 1984), because larger individuals shade smaller ones but the reverse seldom occurs. Hence size hierarchy among individuals becomes pronounced (Weiner, 1985) resulting in a positively skewed frequency-distribution of individual sizes. Hozumi, Shinozaki and Tadaki (1968) developed a method to describe the size frequency distribution by using cumulative number (N ) and cumulative biomass (Y ) from the largest individual in a stand and designated this as the MNY method. This method was further developed by Yamakura (1984) and applied for forest management (Kikuzawa, 1982). Several attempts have been made to determine the relationships between the above laws. Hozumi et al. (1968) and Yamakura (1984) estimated the ranges of parameter values of their size distributions under conditions of selfthinning or density effect. Westoby (1981) examined the relationship between size frequency distribution and selfthinning and argued that the mortality process depends on the suppression of the smallest individuals in the frequency distribution by larger individuals. Weiner and Thomas (1986) reviewed self-thinning data and found size-inequality among plants in a population increases with competition before mortality is extensive, but then decreases as mortality proceeds. They concluded that mortality of the smallest individuals reduces size variability. As Weiner and Thomas (1986) stated, ‘ A comprehensive theory which explains the relationships between the growth of individual plants, density-yield relationships, size distributions and selfthinning is waiting to be uncovered ’. In this paper, I will derive the size distribution from competition-density effects, or the relationship between mean plant-size (or total biomass) and plant population density, assuming one-sided competition among individuals. I will then derive the selfthinning rule under the assumptions of individual plantheight and plant weight allometry and constant dry-matter density and will show that self-thinning is the consequence of shading of small individuals by larger ones under the constraint of size hierarchy. I will test my theory using several data sets from pine and birch forests. ASSUMPTIONS In this paper, I will assume the following : (1) there is a reciprocal relation between stand biomass and plant density. Kira et al. (1953) first found a simple relationship between mean plant size and population density in an experimental population of soy bean. This relationship was later formulated by a reciprocal equation which derived from a logistic growth equation (Shinozaki and Kira, 1956), 1\w l ApjB where w is mean plant size and p is plant density per unit land area and A and B are parameters. By multiplying both sides of the equation by 1\p, we obtain the relationship between the biomass of a population per unit land area ( y) and population density, 1\y l B\pjA (1) This law governs total biomass among plant populations grown at different densities for the same period of time and appears quite general within plant populations, but it does not consider variation in individual plant sizes within a population. (2) Strictly one sided-competition. Individual plants compete for resources such as light, water and nutrients. When competing for light, larger individuals have an advantage since they can shade smaller ones, while smaller plants cannot shade larger ones (Cannell et al., 1984 ; Weiner, 1985). Recently, some authors have considered competition for light not to be one-sided in the strict sense, but rather asymmetrical, where smaller individuals affect larger ones to some extent (Hara, 1988), but larger ones are definitely at an advantage. Here I assume that competition is completely one-sided. (3) Constant drymatter density. Kira and Shidei (1967) divided stand biomass ( y) by the height of the stand (H ) and designated this value dry-matter density. They found that dry matter density of various stands across a range of stand heights takes a more or less constant value, around 1 kg m−$, except for extremely dwarf stands with individuals of small but bent architecture. Here I will use the height of the highest individuals in the stand (Hmax) instead of mean height (H ). Stand biomass ( y) will be expressed as, y l dHmax (2) where d is dry matter density and is independent of height, d l kHmax!. Here I will relax this formulation to obtain the more general form : (3) d l K Hmaxa " where K is a constant and a is a parameter. When a is zero, " eqn (3) coincides with the constant dry matter density. (4) Allometric relationship between mass and height. It is well known that there is a relationship, expressed by a power function, between individual plant mass (W ) and diameter or height of the plant, which is designated allometry (Niklas, 1994). When we consider plant mass and height, this is expressed as : W l K Hb (4) # where K is a constant and b takes a value around 3, since # plant mass has a dimension of [L$], while height is [L"]. The allometric relationship among individuals of the same age is shown to change when plants are competing (Weiner and Thomas, 1992 ; Weiner and Fishman, 1994 ; Nagashima and Terashima, 1995). Here I will use eqn (4) as a trajectory of development of the largest individual. THEORETICAL CONSIDERATIONS Deriation of Y–N relation Consider two populations of the same age but different planting densities and suppose one-sided competition (Fig. 1 A and B). The number of plants per unit land area in a dense stand (A) is n and in a sparse stand (B) is m. The plant biomass per unit land area and population density of the two plots must be related by eqn (1) (Fig. 1 D). Smallest n–m plants in the crowded stand do not affect the growth of largest m plants in the same stand (Fig. 1 A), since competition is one-sided and thus smaller plants have no effect on larger plants. We can say, therefore, that the larger m individuals in the crowded stand (Fig. 1 A) are identical in size to m individuals in the sparse stand (Fig. 1 B) where m 13 Kikuzawa—Hierarchy and Self-thinning A B C D Y y a b 1/Y = B/N + A 1/y = B/p + A N p Y–N Relation within a stand y–p Relation among stands F. 1. Schematic representation of comparison of two stands with different initial planting densities but the same age. A, Dense stand where smaller individuals () do not affect the size of larger individuals () due to one-sided competition. Therefore the size of larger individuals is identical to that of individuals in a sparse stand where the number of plants planted equals the number of larger plants in the dense population. B, Sparse stand where the number of plants planted equals the number of large individuals in the dense stand. They are assumed to be identical to the largest plants in the dense stand. C, Largest plant number (N ) and largest plant size (Y ) relationship in the dense stand, which is identical to plant number per plot ( p) and total plant size per plot ( y) relationship. D, Population density ( p) and population biomass ( y) relationship which was obtained by comparison of dense and sparse stands. plants were initially planted. Within the dense stand the largest m plants and the total n plants must also be related by eqn (1) since the largest m plants in the dense stand are identical to the total m plants of the sparse stand. This argument is possible upon any number (m) of largest plant in the dense stand, thus we can conclude that same relationship shown in eqn (1) must hold between cumulative plant size (Y ) and cumulative plant number (N ) from the largest individual (Fig. 1 C). 1\Y l B\NjA (5) To describe skewed size structure within a forest, Hozumi et al. (1968) independently obtained eqn (5) as an empirical function in which N and Y are expressed as follows : Nl & Wmax W f(W ) dW (6) Yl & Wmax Wf(W )dW (7) W where f(W ) is a distribution density function of individual weight W (Hozumi et al. 1968). Deriation of the relationship between parameters A and B I will examine the meanings of parameters A and B in eqn (5). Parameter A is the reciprocal of stand biomass (Ymax ) when N reaches infinity A l 1\Ymax , (8) while B is the reciprocal of mean plant size when N reaches zero, or the reciprocal of the maximum individual mass (Wmax) of the stand B l 1\Wmax (9) Both parameters A and B, together with Wmax and Ymax will change with the development of the stand and are thus 14 Kikuzawa—Hierarchy and Self-thinning considered to be functions of time t. They can be expressed as functions of maximum plant height [H(t)max] at each time period (t), since plant height increases with time. By applying eqn (2) to the Ymax–Hmax relationship, we obtain Y(t)max l dH(t)max. By substituting eqns (3) and (8), the following is obtained : ") A(t) l K−" H(t)−(a+ (10) max " By applying eqn (4) to HmaxkWmax allometry, we obtain : W(t)max l K H(t)bmax # Substitution of eqns (9) and (10) into eqn (11) gives (11) B(t) l K A(t)b/(a+") $ where K l Kb/(a+") K−" $ " # (12) Deriation of self-thinning line Here I will consider changes in stand biomass and population density under conditions where stand structure is expressed by eqn (5) and its parameters change with time following eqn (12). We can assume that the order of individual sizes within a stand will not change with time since we presume from assumption (2) (one-sided competition) that larger individuals have greater growth than smaller individuals. By differentiating eqn (5) with W, we obtain : dY\dW l (dY\dN ) (dN\dW ) l [B\(ANjB)#](dN\dW ). From definitions of Y and N in eqns (6) and (7), dY\dW l Wf and dN\dW l f, we obtain the following (Hozumi et al. 1968) : W(t) l B(t)\[A(t)NjB(t)]#. Substituting eqn (12), we get : W(t) l oK A(t)b/(a+")q\oA(t)NjK A(t)b/(a+")q# $ $ Differentiation of this equation with respect to t gives : dW\dt l (dA\dt)[K N(b\(aj1)k2)Ab/(a+") $ k(bK #\(aj1))A(#b/(a+")−")]\(ANjK Ab/(a+"))$. (13) $ $ Now let consider the condition where self-thinning occurs in eqn (13). Assume individuals will die when the growth rate or dW\dt equals zero. In an actively growing stand, dA\dt is not zero, nor is (ANjK Ab(a+"))$. The condition which $ satisfies dW\dt l 0 in eqn (13), therefore, is obtained as : N(t) l K A(t)(b/(a+")−") ! % where N(t) is number of individuals which satisfy dW\dt l ! 0 and K l bK (aj1)\[b\(aj1)k2]. Substituting this % $ equation and eqn (12) into eqn (5), and letting Y which satisfies dW\dt l 0 be Y(t) , we obtain : ! (14) Y(t) l K N(t)[−(a+")/(b−(a+"))] ! ! & where K l oK−(a+")/(b−(a+")jK K−b/(b−(a+"))q−". When a l 0 & % $ % [assumption (2)] and b l 3 [assumption (4)], the power in eqn (14) becomes k1\2, which corresponds to the k3\2 power law. MATERIALS AND METHODS Birch stands in Hokkaido, northern Japan (Kikuzawa, 1988) and a pine stand in Tochigi, central Japan (Tadaki et al. 1979) were used for analysis. The questions addressed here are as follows : does the cumulative plant size cumulative plant number relationship follow equation (5) ? ; is parameter a in eqn (3) nearly zero in stands where selfthinning is occurring ? ; is there an allometric relationship expressed by (4) between observed maximum plant mass and observed maximum plant height ? if so, does the parameter b in eqn (4) take a value near 3 ? ; does the relationship expressed by eqn (12) exist between parameters A(t) and B(t) in eqn (5) ? ; is the gradient of the A(t)–B(t) relationship in naturally thinned stands on a double logarithmic scale approximated by b\(aj1) ? ; and can the self-thinning line be approximated by eqn (14) ? Data from four plots were analysed. The surveyed area was 100 m# in plots 1, 2 and 4, and 25 m# in plot 3. Plots 1–3 are birch stands in which Japanese mountain birch (Betula ermanii Cham.) and a few other tree species (B. maximowicziana Regel and Phellodendron amurense Rupr.) regenerated naturally after a scarification. Plots 1 and 2 are in Tobetsu (western Hokkaido) and plot 3 is in Ashibetsu (west-central Hokkaido). Ages of stands at establishment were 12 years in plots 1 and 2 and 6 years in plot 3. In plot 2, trees were artificially thinned from 200 trees per plot to 19 trees per plot at the time of establishment. Plots 1 and 2 were established in 1983 and plot 3 in 1985. The number of trees in each plot, stem diameter at breast height (DBH) or at the ground surface (D ) (when trees were smaller than 1n3 ! m in plot 3) and height of each tree were monitored every year for 10, 9 and 8 years for plots 1, 2 and 3, respectively (Kikuzawa, 1988, 1993). Plot 4 is a pine (Pinus dinsiflora Sieb. et Zucc.) stand described by Tadaki et al. (1979) in which pines were naturally regenerated after a clear logging. The plot was established in 1953 and data sets from 1957 to 1976 were used in this study. The number of trees in the plot, tree height and DBH were monitored every 2 years for 20 years beginning when the stand was 14 years old. Individual tree sizes (stem volume or above-ground phytomass) were estimated using allometric relationships between sizes and DBH or D and height obtained from ten ! to 160 sample trees which were felled near the plots (allometries for birch stands were given in Kikuzawa, 1988). For plots 1–3, stem volume was used and for plot 4 above ground phytomass was used. Trees were arranged from largest to smallest in each plot, and cumulative number (N ) and cumulative size (Y ; volume or mass) were calculated from the largest. Dry matter density (or stand volume\mass density, d ) was calculated by dividing total tree mass or volume in a plot by the maximum tree height in the plot for each plot in each age. Simple least square linear regressions were applied on log transformed x, y data sets. Non linear regressions on "! eqn (5) were carried out using quasi Newton’s method. Significance levels adopted in this paper were P 0n001 and P 0n0001. 15 Kikuzawa—Hierarchy and Self-thinning RESULTS Regressions of eqn (5) to actual data sets were fairly good, as exemplified in Fig. 2. There was no significant correlation between stand volume density (d ) and maximum tree height (Hmax) in plot 1 in any year (Fig. 3 A) ; i.e. parameter a in eqn (3) is nearly zero, which supports assumption (3). W(t)maxkH(t)max allometry in plot 1 (Fig. 3 B) also fits the expectation from assumption (4) i.e. the allometric coefficient is nearly 3. As a result, parameter b in the B(t)–A(t) relationship shown in eqn (12) also takes a value of approx. 3 (Fig. 3 C), which fits fairly well with expectations. As expected from eqns (9) and (8), the stem volume of the largest tree in the plot, and the total stem volume within the plot were almost proportional to reciprocals of parameters B and A in eqn (5), respectively (data not shown). Total stand volume (Y ) was regressed ! against total stand number (N ) in each year, resulting in the ! self-thinning line. The slope of the self-thinning line was k0n424 (Fig. 3 D). However, the self-thinning line predicted by eqn (14) deviated somewhat from the data. Similar results were obtained in plots 3 and 4, where parameter a in the d–Hmax regression is very small and natural thinning actively occurs, although the observed thinning lines deviated slightly from the expected ones. The natural thinning line in plot 4 (Fig. 4 I) is shallower than 1\2. In plot 2, parameter a in the d–Hmax relationship is greater than zero (Fig. 4 A). In this plot, self-thinning hardly occurs (Fig. 4 C), so no thinning line could be obtained. DISCUSSION Nearly half a century has passed since the first monumental work on the effects of density on plant size by Kira et al. (1953). Since then, many studies (Watkinson, 1984 ; Begon, Harper and Townsend, 1996) have been carried out to support the density effect on plant size expressed by eqn (1). Yoda et al. (1963) studied the mean plant size (or total plant size per unit land area) and plant density trajectory of actively growing plant populations. The trajectory on a double logarithmic scale usually takes a value around k3\2. They considered the basis for this trajectory to be the ratio of dimensions of mean plant mass to mean area occupied by a plant. The relationship between the two laws has not yet been clarified. The solution must be pursued not in the mean size of a population but in the size structure of a plant population, since interactions among individuals bring about density effects and self-thinning. Two kind of 0.2 0.6 B 0.15 Y (m3 100 m–2) Y (m3 100 m–2) A 0.1 0.4 0.2 0.05 0 10 5 15 20 0 N (100m ) 20 N (100m ) 1200 C D 1000 Y (kg 100 m–2) 1000 Y (kg 100 m–2) 15 2 1200 800 600 400 200 0 10 5 2 800 600 400 200 20 40 60 2 N (100 m ) 80 100 0 20 40 60 80 100 2 N (100 m ) F. 2. Cumulative tree size (Y, stem volume or above ground mass) and cumulative tree number (N ) relationships in representative stands and ages. A, Y–N relationship in plot 2 at age 12. 1\Y l 3n609\Nj41n22 (m$ 100 m−#, 100 m−#) ; r# l 0n994 ; P 0n0001. B, Y–N relationship in plot 2 at age 20. 1\Y l 1n002\Nj11n98 (m$ 100 m−#, 100 m−#) ; r# l 0n9906 ; P 0n0001. C, Y–N relationship in plot 4 at age 14. 1\Y l 0n00157\Nj0n277 (kg 100 m−#, 100 m−#) ; r# l 0n9984 ; P 0n0001. D, Y–N relationship in plot 4 at age 33. 1\Y l 0n000715\Nj0n0140 (kg 100 m−#, 100 m−#) ; r# l 0n9978 ; P 0n0001. 16 Kikuzawa—Hierarchy and Self-thinning 0.1 0.4 Wmax (m3) Y (m3 100 m–2 m) A 0.01 7 10 0.1 0.01 17 B 3 10 Hmax (m) 10 300 30 Hmax (m) D C B (m–3) Yo (m3 100 m–2) 100 1 10 3 0.3 1 2 3 –3 A (100 m m ) 0.1 10 100 1000 –2 Ho (100 m ) F. 3. Relationships between parameters and variables of a young birch stand in plot 1. A, A relationship between maximum tree height (Hmax) n!*)( in a stand and stand volume density (d ) along stand development. d l 0n0535H!max (m$ 100 m−$, m) ; r# l 0n102 ; P l 0n368. B, A relationship n!)& between Hmax and largest tree volume (Wmax) along stand development. Wmax l 0n0000215H$max (m$, m) ; r# l 0n921 ; P 0n0001. C, A relationship between parameters A and B along stand development B l 26n35A#n)$# (m−$, 100 m# m−$) ; r# l 0n9914 ; P 0n0001. D, A relationship between total tree number (N ) and total tree volume (Y ) in the stand along stand development. Y l 6n524N−!n%#* (m$ 100 m−#, 100 m−#) ; r# l 0n9715 ; ! ! ! ! P 0n0001. Broken line is an expected thinning line from eqn (14). interactions, one- or two-sided, have been shown (Weiner and Thomas, 1986). If plants compete for light, competition is more one-sided, since larger plants suppress small ones but smaller plants cannot suppress larger ones. However, if larger plants are shaded by smaller neighbours in the lower part of their crown, competition cannot be absolutely onesided, but asymmetrically two-sided (Yokozawa and Hara, 1992 ; Kikuzawa and Umeki, 1996). Moreover, if plants compete for underground resources, competition is twosided (Hara, 1986 ; Weiner and Thomas, 1986). Therefore, one sided-competition is an extreme case. However, in plant populations, competition for light is most important and thus one-sided competition is not rare. Westoby (1981) suggested that crowding will increase the size hierarchy among individuals in a population and suppression of smaller individuals by larger ones will bring about natural thinning. Weiner and Thomas (1986) revealed that smaller individuals actually die in the process of self-thinning. Kikuzawa (1993) found that smaller trees died by suppression from larger trees in a birch stand. In this study, I derived the Y–N relationship in eqn (5), first proposed by Hozumi et al. (1968), from the density effect expressed in eqn (1), assuming of one-sided competition Under assumptions of independence of dry matter density on maximum plant height (Hmax) in the stand and an allometric relationship between maximum plant mass (Wmax) and Hmax, parameters A and B will change with time following the relationship expressed in eqn (12). When the Y–N relation shifts on a trajectory constrained by eqn (12), smaller plants will die because their growth rate becomes zero as a consequence of suppression by larger plants. These are the causal relationships between density effect, size structure and self-thinning. During stand development when competition has not yet occurred, the trajectory of plant biomass density rises vertically, since no mortality usually occurs. When interactions among individuals start to occur, the trajectory deviates left of the vertical line. As competition becomes more severe, the gradient of the trajectory line becomes shallower, and at last reaches a steady state where the selfthinning line is expressed as an oblique line with a fixed negative gradient. Yoda et al. (1963) first considered the gradient of self-thinning to be k3\2 in mean plant size s. plant population relations and k1\2 in total mass s. 17 Kikuzawa—Hierarchy and Self-thinning 1 A B 0.1 0.01 5 10 0.1 0.01 20 Yo (m3 100m–2) 5 Wmax (m3) 1 10 Hmax (m) D 0.1 10 15 3 –2 Yo (m 25m ) 0.01 0.001 1 10 100 F 0.1 0.01 10 100 100 G 100 2 10 Hmax (m) 20 No (25 m ) H 10 1 1000 –2 10000 50 100 10 Hmax (m) Wmax (kg) d (kg 100 m–2 m) 1 1 E Hmax (m) 200 0.1 No (100 m–2) 0.01 0.005 1 Hmax (m) Wmax (m3) d (m3 25 m–2 m) 0.1 100 C Yo (kg 100m–2) d (m3 100 m–2 m) 0.5 100 10 Hmax (m) I 1000 100 10 100 1000 10000 No (100 m–2) F. 4. Relationships between parameters and variables in young birch (plots 2 and 3) and pine (plot 4) stands. A, Relationship between maximum n&% tree height in a stand (Hmax) and stand volume density (d ) in plot 2. d l 0n0000739 H#max (m$ 100 m−$, m) ; r# l 0n690 ; P l 0n006. B, Relationship n!& between Hmax in a stand and stem volume of the largest tree (Wmax) in plot 2. Wmax l 0n00003097 H$max (m$, m) ; r# l 0n873 ; P 0n001. C, Relationship between number of trees in a stand (N ) and total tree volume (Y ) in the stand in plot 2. D, d against Hmax in plot 3. ! ! n#"& n!" d l 0n0155 H!max (m$ 25 m−$, m) ; r# l 0n583 ; P l 0n027. E, Wmax against Hmax in plot 3. Wmax l 0n00002379 H$max (m$, m) ; r# l 0n966 ; P 0n0001. F, Y against N in plot 3. Y l 1n225 N−!n%!% (m$ 25 m−#, 25 m−#) ; r# l 0n947 ; P 0n0001. Broken line is an expected thinning line from eqn (14). ! ! ! ! !n!$'* (kg m−$, m) ; r# l 0n047 ; P l 0n54. H, W against H n$& G, d against Hmax in plot 4. d l 118n5 H−max in plot 4. Wmax l 0n0294 H$max (kg, m) ; max max r# l 0n943 ; P 0n0001. I, Y against N in plot 4. Y l 3625 N−!n#'% (kg 100 m−#, 100 m−# ) ; r# l 0n953 ; P 0n0001. Broken line is an expected ! ! ! ! thinning line from eqn (14). population relations. In this study, the self-thinning line is expressed by eqn (14) and the slope by k(aj1)\ [bk(aj1)]. If b l 3 and a l 0, as expected from assumptions (3) and (4), the gradient becomes k1\2. However, the gradient can take various values depending on a and b. If a equals zero but b is greater than 3, the self-thinning line becomes shallower than k1\2 (Fig. 4 I). On the other hand, when a has a positive value, the slope of the self-thinning line becomes steeper. From eqn (14), no plant death is expected to occur when b aj1. When b l 3, if a is greater than 2, no self-thinning will be found as is seen in plot 2 (Fig. 4 C). Hence eqn (14) could simulate the deeper to shallower change of the thinning gradient by the systematic changes in parameters. In the fully stocked forest stands, parameter a in eqn (3) may be zero and dry matter density of stands reaches an upper limit of constant value. However, dry matter density could be smaller than the upper limit in stands which do not reach full stock and parameter a in eqn (3) could take a positive value. Parameter b implies the dependency of plant mass (W ) on plant height (H ). This dependency has been shown to change when plants are competing (Weiner and Thomas, 1992 ; Weiner and Fishman, 1994 ; Nagashima and Terashima, 1995). However, the allometry between mass 18 Kikuzawa—Hierarchy and Self-thinning and height of the largest individual is considered to be more or less constant under strict one-sided competition, since neighbouring individuals do not affect growth of the largest individual. In summary, parameter a is considered to decrease and b to be constant with the development of stand. Thus the slope of eqn (14) is expected to become shallower with the development of a stand. This trend is consistent with the trajectory of actual forest stands (Tadaki, 1963). A C K N O W L E D G E M E N TS Field work was supported by the late Norio Mizui, as well as T. Asai, K. Seiwa, M. Ishida, M. Shibuya, N. Nitta, K. Umeki, and H. Koyama. T. Takada, T. Hara and Y. Harada provided helpful mathematical suggestions. I also thank N. Yamamura and A. Osawa for reading the draft of the manuscript and anonymous reviewers for helpful suggestions. This work was supported by the fund T09304073 of Japanese Ministry of Education, Science, Sports and Culture. LITERATURE CITED Armstrong JD. 1997. Self-thinning in juvenile sea trout and other salmonid fishes revisited. Journal of Animal Ecology 66 : 519–526. Begon M, Harper JL, Townsend CR. 1996. Ecology, indiiduals, populations and communities. 3rd edn. Oxford : Blackwell Science Ltd. Bleasdale JKA, Nelder JA. 1960. Plant population and crop yield. Nature 188 : 342. Cannell MGR, Rothery P, Ford ED. 1984. Competition within stands of Picea sitchensis and Pinus contorta. Annals of Botany 53 : 349–362. Hamilton NRS, Matthew C, Lemaire G. 1995. In defense of the k3\2 boundary rule : a re-evaluation of self-thinning concepts and status. Annals of Botany 76 : 569–577. Hara T. 1986. Effects of density and extinction coefficient on size variability in plant population. Annals of Botany 57 : 885–892. Hara T. 1988. Dynamics of size structure in plant populations. Trends in Ecology and Eolution 3 : 129–133. Hozumi K, Shinozaki K, Tadaki Y. 1968. Studies on the frequency distribution of the weight of individual trees in a forest stand. I. A new approach toward the analysis of the distribution function and the k3\2th power distribution. Japanese Journal of Ecology 18 : 10–20. Kikuzawa K. 1982. Yield-density diagram for natural deciduous broadleaved forest stands. Forest Ecology and Management 4 : 341–358. Kikuzawa K. 1988. Intraspecific competition in a natural stand of Betula ermanii. Annals of Botany 61 : 727–734. Kikuzawa K. 1993. Self-thinning line and B-point line of the yielddensity diagram in a young birch stand. Forest Ecology and Management 58 : 287–298. Kikuzawa K, Umeki K. 1996. Effect of canopy structure on degree of asymmetry of competition in two forest stands in northern Japan. Annals of Botany 77 : 565–571. Kira T, Shidei T. 1967. Primary production and turnover of organic matter in different forest ecosystems of the western Pacific. Japanese Journal of Ecology 17 : 70–87. Kira T, Ogawa H, Sakazaki N. 1953. Intraspecific competition among higher plants. I. Competition-yield-density interrelationship in regularly dispersed populations. Journal of the Institute of Polytechnics, Osaka City Uniersity 4 : 1–16. Li B, Watkinson AR, Hara T. 1996. Dynamics of competition in populations of carrot (Daucus carota). Annals of Botany 78 : 203–214. Lonsdale WM. 1990. The self-thinning rule : dead or alive ? Ecology 71 : 1373–1388. Nagashima H, Terashima I. 1995. Relationships between height, diameter and weight distributions of Chenopodium album plants in stands : effect of dimension and allometry. Annals of Botany 75 : 181–188. Niklas K. 1994. Plant allometry. Chicago and London : The University of Chicago Press. Osawa A, Allen RB. 1993. Allometric theory explains self-thinning relationships of mountain beech and red pine. Ecology 74 : 1020–1032. Shinozaki K, Kira T. 1956. Intraspecific competition among higher plants. VII Logistic theory of the C-D effect. Journal of the Institute of Polytechnics, Osaka City Uniersity 12 : 69–82. Tadaki Y. 1963. The pre-estimating of stem yield based on the competition-density effect. Bulletin of the Goernment Forest Experiment Station 154 : 1–19. Tadaki Y. 1977. Leaf biomass. JIBP Synthesis 16 : 39–44. Tokyo : University of Tokyo. Tadaki Y, Takeuchi I, Kawahara T, Sato A, Hatiya K. 1979. Growth analysis on the natural stands of Japanese red pine (Pinus densiflora Sieb. et Zucc.) III Results of experiment. Bulletin of Forest and Forest Products Institute 305 : 125–144. Vandermeer J. 1984. Plant competition and the yield-density relationship. Journal of Theoretical Biology 109 : 393–399. Watkinson AR. 1984. Yield-density relationships : the influence of resource availability on growth and self-thinning in populations of Vulpia fasciculata. Annals of Botany 53 : 469–482. Weiner J. 1985. Size hierarchies in experimental populations of annual plants. Ecology 66 : 743–752. Weiner J, Fishman L. 1994. Competition and allometry in Kochia scoparia. Annals of Botany 73 : 263–271. Weiner J, Thomas SC. 1986. Size variability and competition in plant monocultures. Oikos 47 : 211–222. Weiner J, Thomas SC. 1992. Competition and allometry in three species of annual plants. Ecology 73 : 648–656. Weller DE. 1987. A reevaluation of the k3\2 power rule of plant selfthinning. Ecological Monographs 57 : 23–43. Westoby M. 1981. The place of the self-thinning rule in population dynamics. American Naturalist 118 : 581–587. Westoby M. 1984. The self-thinning rule. Adances in Ecological Research 14 : 167–225. White J. 1981. The allometric interpretation of the self-thinning rule. Journal of Theoretical Biology 89 : 475–500. White J, Harper JL. 1970. Correlated changes in plant size and number in plant populations. Journal of Ecology 58 : 467–485. Yamakura T. 1984. Experimental model of the distribution function of individual plant weight in even-aged populations restricted by the reciprocal equation of the C-D effect. Japanese Journal of Ecology 34 : 35–45. Yoda K, Kira T, Ogawa H, Hozumi K. 1963. Self-thinning in overcrowded pure stands under cultivated and natural conditions. Journal of Biology, Osaka City Uniersity 14 : 107–129. Yokozawa M, Hara T. 1992. A canopy photosynthesis model for the dynamics of size structure and self-thinning in plant populations. Annals of Botany 70 : 305–316.
© Copyright 2026 Paperzz