| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reproductive Biology |
Laboratory of Ecology and Genetics, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan
Received for publication June 10, 2005. Accepted for publication September 30, 2005.
ABSTRACT
Flowers of selfing plants generally have smaller attractive structures than those of outcrossers. This phenomenon has been explained by sex-allocation theory, which considers allocation of resources to different floral functions, yet its hierarchical structures have mostly been ignored in previous studies. By assuming a three-level hierarchy, we investigate the relationships between floral resource allocation and mating system (partial selfing or obligate outcrossing) in 13 Trillium camschatcense populations from northern Japan. In general, selfing populations had smaller petals than outcrossing populations, and multiple levels in the hierarchy were essential to explain these differences in petal size. Within flowers, selfing plants allocated fewer resources to petals and more to pistils, supporting the prediction of sex-allocation theory. Moreover, selfing plants tended to increase their flower number and decrease investment per flower under the constraints of nonlinear trade-offs between these traits. Selfers may produce more flowers to enhance their female reproductive success under the nonlinear trade-offs in which more resources can be used by increasing flower number. Although our results suggest that the evolution of resource allocation associated with selfing can occur at several hierarchical levels, these allocation patterns varied widely among selfing populations and recent evolution of self-compatibility may explain this variation.
Key Words: attractive structure • flower number • geographical variation • resource allocation hierarchy • self-fertilization • sex-allocation theory • trade-off
Among animal-pollinated hermaphroditic plants, flowers of selfers generally have smaller attractive structures (such as petal and corolla) than those of outcrossers (e.g., Ornduff, 1969
; Cruden and Lyon, 1985
). This phenomenon has been explained by sex-allocation theory, in which evolution of reproductive traits is assumed to be constrained by trade-offs due to finite resources (reviewed by Brunet, 1992
). Sex-allocation models have examined evolutionarily stable allocation of reproductive resources to female, male, and attractive structures of flowers and have predicted that allocation to both male and attractive structures decreases as the selfing rate increases (e.g., Charlesworth and Charlesworth, 1987
; Lloyd, 1987
). This is because the proportion of ovules available for outcross pollen decreases with increasing selfing and so does the potential fitness gain through male and pollinator-attractive function. Many empirical works comparing resource allocation within flowers between selfers and outcrossers broadly support this prediction (e.g., Ritland and Ritland, 1989
; Parker et al., 1995
).
However, because of the modular construction of plants, resources invested in floral structures are likely to be determined through a series of hierarchical allocations (van Noordwijk and de Jong, 1986
; de Jong, 1993
; Barrett et al., 1996
; Venable, 1996
; Worley and Barrett, 2001
). Plants may allocate resources between reproductive (i.e., flowers) and vegetative (e.g., leaves and stems) functions, then subdivide reproductive resources among flowers, and further subdivide resources among floral structures within each flower (Fig. 1). In this respect, previous sex-allocation models have mostly ignored hierarchical natures of resource allocation and empirical testing of the prediction has only examined allocation at the end of the hierarchy, even though absolute amount of resources invested in floral traits is also influenced by allocation at the early levels. It is well known that resource investment per flower often decreases with increasing selfing (Ornduff, 1969
; Cruden and Lyon, 1985
; Ritland and Ritland, 1989
; Parker et al., 1995
; Sato and Yahara, 1999
). This strongly suggests that allocation early in the hierarchy also changes in response to selection by selfing, because the smaller investment per flower in selfers can only be achieved by changes in allocation fraction at the early levels.
|
). Since the cost of producing an ovule is generally much smaller than that of developing it into a seed (Lovett Doust and Cavers, 1982
), the increased allocation to female function may require more resources at fruiting, rather than at flowering, to mature many seeds. Second, selfers may increase number of flowers that compete for resources and thus reduce their investment per flower. Sakai and Harada (2001)
theoretically showed that the size–number trade-offs of organs may be nonlinear and the total amount of resources invested in organs is expected to increase with the number of organs produced. From this perspective, it can be advantageous that selfers increase flower number to use more resources particularly for female function, whereas outcrossers may keep the large attractive structures of individual flowers as they require repeated pollinator visits for pollen export. Sakai (1993
, 2000
) developed a model with two-level allocation hierarchy that incorporated such floral size–number trade-offs. Although flower number (and investment per flower) was also set to be subject to selection in the model, he concentrated on the relationship between selfing rate and resource investment to attractive structures rather than on that between selfing rate and flower number. In Sakai's model, however, the general relationship between selfing rate and investment to attractive structures was obtained only if there are nonlinear constraints on flower number and investment per flower. Although his result stresses the importance of nonlinearity in trade-offs, no empirical study has examined the nonlinearity.
To explore how hierarchical resource allocation contributes to small attractive structures in selfers, we investigate the relationships between floral resource allocation and mating system in natural populations of an animal-pollinated hermaphroditic plant, Trillium camschatcense Ker Gawler (Trilliaceae). Breeding experiments indicate that differentiation in mating system occurs among populations from Hokkaido, Japan. Some populations are self-compatible (SC) and plants have a high selfing ability, whereas the remaining populations are self-incompatible (SI) and plants produce seeds by obligatory outcrossing (Fig. 2; after Ohara et al., 1996
). In this study, we compared the dry matter allocation of plants from 13 populations by assuming that resources were invested in floral traits according to the three-level hierarchical resource allocation as described above (Fig. 1). We ask the following specific questions. First, are there significant differences in allocation at each of the three levels of the hierarchy between SC and SI populations? Second, how do the three levels of the hierarchy contribute to the differences in size of attractive structures (i.e., petals)? Finally, are there nonlinear trade-offs between flower number and resource investment per flower? In the analysis of biomass allocation, we consider several factors that might confound the results. The resource status of plants may strongly affect not only the total quantity of resources allocated to flowers but also biomass allocation within flowers (Klinkhamer and de Jong, 1997
). Additionally, the amount of resources allocated to flowers may also be influenced by flower number due to the nonlinear size–number trade-offs of flowers (Sakai and Harada, 2001
). We account for these factors using simple statistical models.
|
Study system
Trillium camschatcense commonly occurs in the understory of broadleaved deciduous forests of Hokkaido, Japan. Plants are nonclonal, long-lived, and polycarpic (Ohara and Kawano, 1986a
). Stems and leaves emerge annually in late April to early May from the underground rhizome. In mid to late May, reproductive individuals have flowers that arise from separate stems, each of which has a whorl of three leaves. Each flower has no nectar but has three large white petals, with a sweet scent, that attract some beetles (e.g., Nitidulidae and Melandryidae) and flies (e.g., Bibionidae and Scathophagidae) as potential pollinators (Ohara et al., 1991
; Tomimatsu and Ohara, 2003
).
Four geographical groups, North, East, South, and Ishikari have been recognized in Hokkaido by chromosomal and allozyme variations (Kurabayashi, 1957
; Ohara et al., 1996
; Fig. 2). All South and North populations are SC and have lower genetic diversity. By contrast, in East populations where genetic diversity is higher, the breeding system is SI with the exception of some small SC populations in the Hidaka region. On the other hand, the small populations in the Tokachi region, also in East, are all SI probably because these populations are remants of large SI populations due to recent agricultural development (Tomimatsu and Ohara, 2002
). Ishikari is located in the Ishikari Depression, which was repeatedly submerged during interglacial periods (Fig. 2). The most recent connection between the northern and southern parts of Hokkaido was established only in the last several thousand years. Previous studies have shown that populations in the Ishikari area include substantial genetic components from all the other population groups, indicating that the other groups migrated to the depression after the connection was established (Kurabayashi, 1957
; Fukuda et al., 1960
).
Self-pollinated flowers of SI plants produce no seeds at all, indicating that the SI system is not leaky. SC plants have a high ability of selfing and self-pollinated flowers produce abundant seeds (Ohara et al., 1996
). Selfing may occur largely through autonomous selfing because flowers from which pollinators are excluded set as many seeds as those hand-outcrossed (K. Takenaka, H. Tomimatsu, and M. Ohara, unpublished data). In SC populations from East, a preliminary analysis revealed high selfing rates (s = 0.74–0.86). Although virtually no polymorphism at 13 allozyme loci precluded measurement of selfing rate in the other groups, the high ability of autonomous selfing and homozygosity at the isozyme loci suggest high levels of selfing. Thus, the self-incompatibility in this species can be treated as a dichotomous trait (i.e., either high or no ability of selfing), rather than a continuous, quantitative trait as suggested in some species (e.g., Good-Avila and Stephenson, 2002
).
Biomass estimation
We estimated the biomass allocation of plants from 13 T. camschatcense populations (Table 1; Fig. 2). Reproductive plants with one or two flowers were sampled for this study because they predominated. In each population, 10–29 plants (total N = 280) were harvested to cover the full range of sizes of reproductive plants (Table 1). In advance, we selected two predictor variables, flower number and leaf width, for multiple regression models that estimate the biomass of plants according to the AIC criterion (Akaike, 1973
). These preliminary analyses were done in four populations (total N = 85) and the coefficients of determination (R2) of the models ranged from 0.86 to 0.96. Using these variables, we were able to exhaustively search plants with a full range of sizes in the field. Plants were sampled in May 2002, at the beginning of anthesis, to prevent any pollen loss. In two populations (pops. 9 and 11), however, plants were sampled shortly after anthers dehisced, so that the data were not included in the analysis of stamen biomass. After the plants were harvested, their tissues were dried in an oven at 80°C for at least 48 h and weighed to the nearest 0.1 mg. To obtain comparable biomass measures among populations, the dry weight of the whole plant, including underground rhizomes, was used for estimating plant size, rather than flower number and leaf width.
|
). If b = 1, the increases in biomass of flowers and plants are isometric (LaBarbera, 1989
) and the biomass allocation ratio to flowers is constant and independent of the plant biomass. Otherwise, the biomass of the flowers is a decelerating (b < 1) or accelerating (b > 1) function of the biomass of the plants. Analysis was performed by using the GLM procedure of SAS version 8.2 (SAS Institute, 1999
) and we used Type III sums of squares.
Allocation to flowers
In the analysis described, the biomass of flowers was clearly related to plant biomass and flower number (see Results). Then, we compared total investment in flowers among the 13 populations by statistically controlling for plant biomass and flower number. We obtained adjusted means of biomass of flowers as "least square means" which can be derived from the ANCOVA model by using the LSMEANS statement of the GLM procedure of SAS. These are values from the specified model where plant biomass and flower number are controlled by being set to sample means (SAS Institute, 1999
). The absence of significant interaction terms indicates that the differences in the log of biomass of flowers between any pair of populations do not change over a range of plant biomass, justifying our comparison. We compared adjusted means between populations by Fisher's LSD method, with 
-levels calculated according to the Bonferroni technique (Rice, 1989
) to limit overall experiment-wise error rate. To test the effect of mating system, adjusted means of nine SC populations were compared with those of four SI populations by Mann–Whitney U test. Although the biomass estimates were logarithmically transformed, we report back-transformed values with asymmetrical 95% confidence intervals.
Flower number
The strategy of flower number and investment per flower was examined by comparing the ranges of biomass of single- and double-flowered plants among populations. The smallest value of biomass of single-flowered plants implies the threshold size at which sexual reproduction occurs. Similarly, the smallest value of biomass of double-flowered plants implies the threshold size at which flower number increases. The existence of such threshold sizes was shown by a previous life-history study of this species (Ohara and Kawano, 1986b
). The differences in average number of flowers among populations could be inferred by comparing these thresholds. We tested their differences between SC and SI populations by Mann–Whitney U tests.
Allocation within flowers
We analyzed patterns of biomass allocation within flowers by using ANCOVA models, in which the biomass of five floral structures (petals, sepals, stamens, pistils, and peduncles) were analyzed separately. In these models, the biomass of flowers was set as a covariate and the effect of population was a categorical, fixed effect. Because flowers on the same plant are equal-sized and equally sex-allocated (H. Tomimatsu and M. Ohara, unpublished data), we did not consider the effect of flower number here. The regression coefficients of the covariate were tested for isometry by t test against the null hypothesis H0: b = 1, by which size-dependent sex allocation could be examined. Interaction terms could be removed from the final models because they were not significant at P > 0.20. Then, we compared adjusted means among the 13 populations using Fisher's LSD method and between SC and SI populations by Mann–Whitney U test. To evaluate phenotypic correlations, we calculated two-tailed tests of Pearson's correlations between adjusted means of the five floral structures. The table-wide significance was assessed by using -levels calculated according to the Bonferroni technique (Rice, 1989
).
Natural variations in petal size and flower number
We also assessed natural variations in petal size and flower number across the populations. Although we systematically harvested T. camschatcense individuals in the analysis of biomass allocation to cover the full range of sizes, we used random (not systematic) samples to investigate variations in petal size and flower number. In each population we randomly established four 2 x 5 m quadrats on the forest floor and analyzed all reproductive plants within the quadrats (total N = 743). We sampled the petals of all flowers, optically scanned them into a computer, and analyzed them with the NIH Image Program (version 1.62, the U.S. National Institutes of Health, http://rsb.info.nih.gov/nih-image/) to determine the mean area of a petal for each individual.
RESULTS
Biomass allocation within populations
The ANCOVA model explained much of the variance (93%) in biomass of flowers across the 13 populations (Table 2). Biomass of flowers was clearly related to plant biomass and flower number. Because no significant interactions were observed, the relationship among biomass of flowers and plant biomass and flower number was qualitatively the same for all populations. Thus, we show this relationship for only pop. 3 in Fig. 3A. Biomass of flowers was positively related to plant biomass, and only large plants could have two flowers. The resulting fitted ANCOVA model showed that double-flowered plants invested significantly more resources (1.4 times) in flowers than did single-flowered plants when we compared them at the same plant biomass (Fig. 3B). As a result, investment per flower of double-flowered plants was only 0.7 times that of single-flowered plants. Because the regression coefficient of the covariate was significantly smaller than 1 (Table 2), the biomass of flowers appeared to be a decelerating function of plant biomass (Fig. 3B). This indicates that the ratio of total biomass allocated to flowers gradually decreased with increasing plant biomass. The biomass of the five floral structures was positively related to the biomass of flowers (Table 3). The ANCOVA models showed that the allocation patterns of reproductive resources to the five floral structures changed with the biomass of flowers, as the biomass estimate of each floral structure was either a decelerating (b < 1) or accelerating (b > 1) function of the biomass of flowers. This reflects size-dependent sex allocation because the biomass of flowers was strongly related to plant biomass (Table 2). Pistil and stamen biomass were strongly accelerating and decelerating functions, respectively (Table 3), indicating that larger plants were more female-biased in their proportional sex allocation. Biomass of attractive structures was also a decelerating function. Again, these relationships were qualitatively the same for all populations because no significant interactions were observed.
|
|
|
|
|
|
|
This study indicates how mating system affects floral resource allocation that is probably made in a hierarchical manner. In T. camschatcense, SC and SI populations differed in their resource allocation at two hierarchical levels. Plants from SI populations tended to decrease their flower number and increase investment per flower (Fig. 5) and allocated more resources to petals and fewer to pistils within flowers (Fig. 6). Both of these patterns indicate that the absolute sizes of petals are larger in SI populations and this expectation well agrees with patterns of natural variations in petal size (Fig. 7). However, these allocation patterns were more closely related to the geographical location than to mating system. In the analysis of biomass allocation, we accounted for several possible confounding factors including size–number trade-offs and size-dependent sex-allocation. As generally observed in animal-pollinated plants (Klinkhamer and de Jong, 1997
), larger plants were more female-biased in their proportional sex allocation (Table 3; for causal mechanisms see Wright and Barrett, 1999
). The important point here is that such patterns were highly conserved across populations, thereby making it possible to accurately compare resource allocation among populations. In addition, to our knowledge, this study provides the first empirical evidence for nonlinear size–number trade-offs of flowers. We discuss these results and their implications in more detail.
Selection for hierarchical resource allocation
The results that plants from SC populations generally invest fewer resources to petals and more to pistils support the prediction of sex-allocation theory. Allocation to stamens showed less clear differentiation probably because of the lack of data from two populations and ambiguity of biomass estimates in pop. 6 with a large confidence interval. The prediction of sex-allocation theory has been generally confirmed in many studies that broadly compared multiple hermaphroditic species (e.g., Lovett Doust and Cavers, 1982
; Cruden and Lyon, 1985
) or compared two or more closely related taxa (e.g., Ritland and Ritland, 1989
; Parker et al., 1995
), but very few studies have conducted intraspecific comparison of a number of populations with different mating systems (Schoen, 1982
; Cumaraswamy and Bawa, 1989
; Morgan and Barrett, 1989
). In addition, no study has explicitly considered hierarchies in the context of mating system, even though variations in within-flower allocation may not fully explain natural variations in size of floral traits. In T. camschatcense, allocation to petals within flowers differed only 
1.4 times among populations although petal size varied more than threefold. This clearly indicates that allocation early in the hieararchy is also responsible for these differences in petal size.
Given that all changes in resource allocation reflect natural selection associated with selfing, the results demonstrate that evolutionary change in flower number occurred as frequently as that in within-flower allocation. Flower number was generally expected to be higher in SC populations than in SI populations at the same plant mass (Fig. 5), suggesting that the increase in flower number also substantially contributes to small attractive structures in selfers. Allocation of total resources to flowers was only reduced in several SC populations (Fig. 4), implying that allocation at this hierarchical level had little impact on the relationship between petal size and mating system. One explanation is that there may be only little or no selection on this allocation fraction. Alternatively, heritable genetic variation at this level may be small, so that it would require more generations for this allocation fraction to respond to selection on floral traits. A simulation study of hierarchical allocation showed that initial evolutionary change mainly proceeds in directions with ample genetic variation (Worley et al., 2003
).
Contrary to the patterns expected, the number of flowers produced by plants randomly sampled in the field did not differ between SC and SI populations. However, flower number may not necessarily be higher in selfing taxa unless variation in resource levels is accounted. Plants from SC populations generally began to produce at smaller sizes than those from SI populations (Fig. 5). Therefore, SC populations may include many small reproductive individuals with only a single small flower, and this may have obscured the differences in flower number between the two mating systems. Hence, the evolution of floral resource allocation associated with selfing is more complex than previously appreciated, and studies of allocation hierarchies require an accounting of confounding factors including resource levels. The strategy of early reproduction in selfers is likely when considering the fact that selfing is more common in shorter-lived plants (Stebbins, 1950
; Barrett et al., 1996
). Early reproduction probably occurs at the cost of future reproduction and survival (Galen, 1993
; Silvertown and Dodd, 1999
) and may therefore select for shorter life span.
Mating system and geography
Allocation patterns were more closely associated with geographical groups than with mating system. This is because SC and SI populations within the East had similar patterns of resource allocation at all the hierarchical levels. This result has two evolutionary implications. First, one plausible explanation for this dissociation is that the SC populations in the East were more recently derived from their outcrossing ancestors and may have experienced a shorter history of selection favoring small attractive structures. All SC populations from this region are small enough to suffer from pollen limitation caused by low pollinator attractiveness (Tomimatsu and Ohara, 2002
). Thus, selfing modifiers may have rapidly spread through the small populations under high selection gradients without changes in resource allocation. Second, the close association between genetic and allocation patterns suggests that our results have a genetic basis. Although the results of this study were based on phenotypic variations, previous studies have shown that the geographical groups are genetically differentiated (Kurabayashi, 1957
; Ohara et al., 1996
). Moreover, allocation patterns differed between two populations from Ishikari (Fig. 4) where genetic composition also differs among populations (Kurabayashi, 1957
; Fukuda et al., 1960
). Although mating system appeared confounded with geography, but not always, in our system, a priori predictions of sex-allocation theory justify testing for the effect of mating system on resource allocation.
Resource allocation trade-offs
The presence of allocation trade-offs is a major assumption of sex-allocation theory, but it has been difficult to demonstrate because genetic and/or environmental variation in resource levels among individuals may obscure trade-offs (Campbell, 2000
;Worley and Barrett, 2000
). We clearly detected some evidences for trade-offs by statistically controlling for resource levels. Within flowers, allocation to petals and pistils showed a strong negative phenotypic correlation. Moreover, the trade-offs between flower number and investment per flower appeared to be nonlinear as predicted by Sakai and Harada (2001)
. Our data support their prediction that the total mass of flowers increases with the number of flowers produced. Double-flowered plants of T. camschatcense invested much more resources in flowers than did single-flowered plants with the same mass (Fig. 3). Although this study does not provide any explanation for the nonlinearity, Sakai and Harada (2001)
provides a metabolic reason that an organism can reduce the time needed to complete organ growth by increasing the number of organs, and faster completion of organ growth reduces the loss of resources due to maintenance respiration. A caveat of these results is that allocation measured in biomass may not accurately reflect the allocation of critical nutrients. Measuring trade-offs in other currencies such as nitrogen may also be needed to determine the extent of nonlinearity.
Evolutionary changes in flower number should reflect selection on the size of floral traits or on the allocation at the focal hierarchical level, or both. As mentioned in the introduction, the nonlinearity in floral size–number trade-offs would select for larger number of flowers in selfing plants. Sakai's (1993
, 2000
) model including the trade-offs suggested that the size of attractive structures depends on the selfing rate only if the trade-offs are nonlinear. Although he did not examine how flower number varies with the selfing rate, selfers would produce more flowers under the nonlinear trade-offs because they can use more resources. In T. camschatcense, the number of ovules is positively correlated with pistil mass (r = +0.91; H. Tomimatsu and M. Ohara, unpublished data), so that selfers can increase the total number of ovules by producing more flowers and this may enhance individual fitness by an increase in selfed seeds. By contrast, outcrossers may favor larger attractive structures of individual flowers to attract more pollinators. However, producing more flowers could also attract pollinators and/or increase the cost of geitonogamy (e.g., Harder and Barrett, 1995
; Ohashi and Yahara, 1998
). Sakai (2000)
considered the former effect and showed that this effect does not alter the main conclusion of his model. Unfortunately, only little is known about the relationship between mating system and flower number from not only theoretical but also empirical perspectives. Consistent evidence comes from a comparative study of two varieties of Impatiens hypophylla (Sato and Yahara, 1999
), in which the outcrossing variety had fewer flowers that invest more resources to attractive structures, whereas the selfing variety had more flowers with less attractiveness. More studies of this sort are clearly needed to draw any conclusions about the generality of producing more flowers in selfers and its contribution to the small attractive structures. In T. camschatcense, we were able to show that the size–number trade-offs of flowers strongly constrain the evolution of floral display and producing more flowers in selfers is likely to be responsible for its small petals. Therefore, the hierarchy of floral resource allocation gives new insight into how the size of attractive structures has been repeatedly reduced with the evolution of selfing.
FOOTNOTES
We thank K. Takenaka, S. Kobayashi, S. Sakai, and M. T. Kimura for valuable comments and suggestions; D. Taylor and anonymous reviewers for constructive criticism on previous versions of the manuscript; and H. Yamagishi, Y. Kameyama, and J. Ishikawa for help with our field work. This research was funded by grants from the Suhara Memorial Foundation and Ministry of Education, Culture, Sports, Science, and Technology for 21st Century COE programs to both authors, and JSPS for Scientific Research (15370006, 16370007) to M.O. and for Young Scientists to H.T. 
2 Author for correspondence (e-mail: tomimatsu-hiroshi{at}c.metro-u.ac.jp
) phone: +81-426-77-2585, fax: +81-426-77-2559; current address: Department of Biological Sciences, Graduate School of Science, Tokyo Metropolitan University, Hachioji 192-0397, Japan
LITERATURE CITED
Akaike H. 1973 Information theory and an extension of the maximum likelihood principle. In B. N. Petrov and F. Csaki [eds.], Second international symposium on information theory, 267–281. Akademiai Kiado, Budapest, Hungary
Barrett S. C. H. L. D. Harder A. C. Worley 1996 The comparative biology of pollination and mating in flowering plants. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 351: 1271-1280[CrossRef]
Brunet J. 1992 Sex allocation in hermaphroditic plants. Trends in Ecology and Evoution 7: 79-84[CrossRef]
Campbell D. R. 2000 Experimental tests of sex-allocation theory in plants. Trends in Ecology and Evolution 15: 227-232
Charlesworth D. B. Charlesworth 1987 The effect of investment in attractive structures on allocation to male and female functions in plants. Evolution 41: 948-968[CrossRef][ISI]
Cruden R. W. D. L. Lyon 1985 Patterns of biomass allocation to male and female functions in plants with different mating systems. Oecologia 66: 299-306[ISI]
Cumaraswamy A. K. S. Bawa 1989 Sex allocation and mating systems in pigeonpea, Cajanus cajan (Fabaceae). Plant Systematics and Evolution 168: 59-69[CrossRef][ISI]
Fukuda I. Y. Hiraizumi T. Narise M. Kurabayashi 1960 Evolution and variation in Trillium. VI. Migrations among natural populations of T. kamtschaticum across the Ishikari depression. Evolution 14: 224-231[CrossRef][ISI]
Galen C. 1993 Costs of reproduction in Polemonium viscosum: phenotypic and genetic approaches. Evolution 47: 1073-1079[CrossRef][ISI]
Good-Avila S. V. A. G. Stephenson 2002 The inheritance of modifiers conferring self-fertility in the partially self-incompatible perennial, Campanula rapunculoides L. (Campanulaceae). Evolution 56: 263-272[CrossRef][ISI][Medline]
Harder L. D. S. C. H. Barrett 1995 Mating cost of large floral displays in hermaphrodite plants. Nature 373: 512-515[CrossRef]
de Jong G. 1993 Covariances between traits deriving from successive allocations of a resource. Functional Ecology 7: 75-83[CrossRef][ISI]
Klinkhamer P. G. L. T. J. de Jong 1997 Size-dependent allocation to male and female reproduction. In F. A. Bazzaz and J. Grace [eds.], Plant resource allocation, 211–229. Academic Press, San Diego, California, USA
Kurabayashi M. 1957 Evolution and variation in Trillium. IV. Chromosome variation in natural populations of Trillium kamtschaticum Pall. Japanese Journal of Botany 16: 1-45
LaBarbera N. 1989 Analyzing body size as a factor in ecology and evolution. Annual Review of Ecology and Systematics 20: 97-117[CrossRef][ISI]
Lloyd D. G. 1987 Allocations to pollen, seeds and pollination mechanisms in self-fertilization plants. Functional Ecology 1: 83-89
Lovett Doust J. P. J. Cavers 1982 Biomass allocation in hermaphrodite flowers. Canadian Journal of Botany 60: 2530-2534[CrossRef][ISI]
Morgan M. T. S. C. H. Barrett 1989 Reproductive correlates of mating system variation in Eichhornia paniculata (Spreng.) Solms (Pontederiaceae). Journal of Evolutionary Biology 2: 183-203
van Noordwijk A. J. G. de Jong 1986 Acquisition and allocation of resources: their influence on variation in life-history tactics. American Naturalist 128: 137-142[CrossRef][ISI]
Ohara M. S. Kawano 1986a Life history studies on the genus Trillium (Liliaceae). I. Reproductive biology of four Japanese species. Plant Species Biology 1: 35-45
Ohara M. S. Kawano 1986b Life history studies on the genus Trillium (Liliaceae). IV. Stage class structures and spatial distribution of four Japanese species. Plant Species Biology 1: 147-161
Ohara M. K. Okazaki S. Kawano 1991 Predominant inbreeding and resource limitation in four Japanese Trillium species. Miyabea 1: 21-30
Ohara M. H. Takeda Y. Ohno Y. Shimamoto 1996 Variations in the breeding system and the population genetic structure of Trillium kamtschaticum (Liliaceae). Heredity 76: 476-484[ISI]
Ohashi K. T. Yahara 1998 Effects of variation in flower number on pollinator visits in Cirsium purpuratum (Asteraceae). American Journal of Botany 85: 219-224[Abstract]
Ornduff R. 1969 Reproductive biology in relation to systematics. Taxon 18: 121-133[CrossRef]
Parker I. M. R. R. Nakamura D. W. Schemske 1995 Reproductive allocation and the fitness consequences of selfing in two sympatric species of Epilobium (Onagraceae) with contrasting mating systems. American Journal of Botany 82: 1007-1016[CrossRef][ISI]
Rice W. R. 1989 Analyzing tables of statistical tests. Evolution 43: 223-225[CrossRef][ISI]
Ritland C. K. Ritland 1989 Variation of sex allocation among eight taxa of the Mimulus guttatus species complex (Scrophulariaceae). American Journal of Botany 76: 1731-1739[CrossRef][ISI]
Sakai S. 1993 Allocation to attractive structures in animal-pollinated flowers. Evolution 47: 1711-1720[CrossRef][ISI]
Sakai S. 2000 Size–number trade-off and allocation to flower production in animal-pollinated flowers. Plant Species Biology 15: 261-268
Sakai S. Y. Harada 2001 Sink-limitation and the size-number trade-off of organs: production of organs using a fixed amount of reserves. Evolution 55: 467-476[CrossRef][ISI][Medline]
SAS Institute. 1999 SAS version 8.2. SAS Institute, Cary, North Carolina, USA
Sato H. T. Yahara 1999 Trade-offs between flower number and investment to a flower in selfing and outcrossing varieties of Impatiens hypophylla (Balsaminaceae). American Journal of Botany 86: 1699-1707
Schoen D. J. 1982 Male reproductive effort and breeding system in an hermaphroditic plant. Oecologia 53: 255-257[CrossRef][ISI]
Silvertown J. M. Dodd 1999 The demographic cost of reproduction and its consequences in Balsam fir (Abies balsamea). American Naturalist 154: 321-332[CrossRef][ISI][Medline]
Sokal R. R. F. J. Rohlf 1995 Biometry, 3rd ed. Freeman, New York, New York
Stebbins G. L. 1950 Variation and evolution in plants. Columbia University Press, New York, New York
Tomimatsu H. M. Ohara 2002 Effects of forest fragmentation on seed production of the understory herb Trillium camschatcense. Conservation Biology 16: 1277-1285[CrossRef][ISI]
Tomimatsu H. M. Ohara 2003 Floral visitors of Trillium camschatcense (Trilliaceae) in fragmented forests. Plant Species Biology 18: 123-127[CrossRef]
Venable D. L. 1996 Packaging and provisioning in plant reproduction. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 351: 1319-1329[CrossRef]
Worley A. C. S. C. H. Barrett 2000 Evolution of floral display in Eichhornia paniculata (Pontederiaceae): direct and correlated responses to selection on flower size and number. Evolution 54: 1533-1545[ISI][Medline]
Worley A. C. S. C. H. Barrett 2001 Evolution of floral display in Eichhornia paniculata (Pontederiaceae): genetic correlations between flower size and number. Journal of Evolutionary Biology 14: 469-481[CrossRef][ISI]
Worley A. C. D. Houle S. C. H. Barrett 2003 Consequences of hierarchical allocation for the evolution of life-history traits. American Naturalist 161: 153-167[CrossRef][Medline]
Wright S. I. S. C. H. Barrett 1999 Size-dependent gender modification in a hermaphroditic perennial herb. Proceedings of the Royal Society of London, B, Biological Sciences 266: 225-232[CrossRef]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
