| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Population Biology |
3Biology Department, Long Island University-Brooklyn Campus, Brooklyn, New York 11201 USA; 4Biology Department and Environmental Studies Program, Southwestern University, Georgetown, Texas 78626 USA; 5Department of Biology, University of South Florida, 4202 E. Fowler Ave., Tampa, Florida 33620-5200 USA; 6Department of Biology, Middlebury College, Middlebury, Vermont 05753 USA; 7Department of Ecology and Evolution, State University of New York, Stony Brook, New York 11794-5245 USA
Received for publication November 23, 2005. Accepted for publication May 9, 2006.
ABSTRACT
Determining the relative contributions of genetic and environmental factors to phenotypic variation is critical for understanding the evolutionary ecology of plant species, but few studies have examined the sources of phenotypic differentiation between nearby populations of woody plants. We conducted reciprocal transplant experiments to examine sources of variation in growth rate, form, survival, and maturation in a globally rare dwarf population of pitch pine (Pinus rigida) and in surrounding populations of normal-stature pitch pines on Long Island, New York. Transplants were monitored over a 6-yr period. The influence of seedling origin on height, growth rate, survival, and form (single-stemmed vs. multi-stemmed growth habit) was much smaller than the effect of transplanting location. Both planting site and seed origin were important factors in determining time to reproduction; seedlings originating from dwarf populations and seedlings planted at the normal-stature site reproduced earliest. These results suggest that many of the differences between dwarf and normal-stature pitch pines may be due more to plastic responses to environmental factors than to genetic differentiation among populations. Therefore, preservation of the dwarf pine habitat is essential for preserving dwarf pine communities; the dwarf pines cannot be preserved ex situ.
Key Words: environmental variation • genetic variation • growth rate • life history variation • nanism • pine barrens • Pinus rigida • reciprocal transplant
Phenotypic variation in plants, as in all organisms, results from a complex interplay of genetic and environmental factors (Lewontin, 1974
). Understanding the relative contribution of these factors is important due to the different consequences these sources of variation have for the ecology and evolution of plant populations. The genetic component of variation in a trait determines the potential for evolutionary response to selection (Lynch and Walsh, 1998
). Plastic responses, on the other hand, may provide functional acclimation in the face of variable or changing environments and in some instances may dampen evolutionary responses to environmental change (Berteaux et al., 2004
).
One situation in which the factors underlying phenotypic variation are of particular interest occurs when there are striking phenotypic differences between plant populations in proximity to one another. Examples include the krummholz and erect forms found in several subalpine tree species (Clausen, 1965
; Grant and Mitton, 1977
; Arno, 1984
), pygmy and normal-stature populations of Pinus contorta in coastal California (Aitken and Libby,
1994), and dwarf and normal-stature mangroves in east Africa (Dahdouh-Guebas et al., 2004
) and in the Caribbean (Lovelock et al., 2004
). In these cases, pronounced phenotypic differences between the populations exist in spite of the fact that the populations are separated by short distances in which gene flow between the populations is likely to be strong. The existence of such phenotypic differences in the face of such gene flow suggests the possibility that the phenotypic differences result from strong differential selection (Linhart and Grant, 1996
). Alternatively, the differences between populations may partly or largely result from the plastic responses of the plants to environmental factors, either in conjunction with, or with little contribution from, genetic differentiation.
There are a substantial number of documented cases in plants of selective differentiation over short distances. For example, classic ecological genetic studies in the 1960s found adaptive differentiation at very small scales in two grasses, Anthoxantum odoratum and Agrostis tenuis, in response to toxic heavy metal contamination on mine tailings (e.g., McNeilly and Antonovics, 1968
; Antonovics and Bradshaw, 1970
). Other edaphic factors, including serpentine soils, have also been shown to act as selective agents resulting in small-scale genetic differentiation among plant populations (reviewed by Brady et al., 2005
; Linhart and Grant, 1996
; Rajakaruna, 2004
). Most of these studies concern herbaceous species, with some exceptions; for example, Beckman and Mitton (1984)
, Mopper et al. (1991)
, and Mitton et al. (1977)
found genetic differentiation over small geographic scales in Pinus spp.
In contrast, a number of studies have failed to show genetic differentiation where it had been anticipated (e.g., Tonsor, 1990
; Novak et al.,
1991). Failure to show genetic differentiation is often associated with populations with high levels of phenotypic plasticity and low levels of genetic variability (Linhart and Grant, 1996
). It may also be predicted from theory to occur whenever gene flow swamps the strength of selection (e.g., Hanson, 1966
; Slatkin, 1987
). It is important for our understanding of plant evolutionary ecology to obtain a better understanding of the conditions that favor genetic differentiation within and among plant populations in response to selection. In addition, these questions can be of great interest in many applied contexts, because preserving overall genetic variation within species, as well as locally adapted ecotypes and clines in particular, may be a high conservation priority (Hughes et al., 1997
; Green, 2005
).
The focus of the present study is a morphologically highly distinct population of dwarf pitch pine (Pinus rigida P. Mill.) found within a larger area of normal stature trees on Long Island, New York. Pinus rigida is the dominant tree species in the pine barrens that occur scattered throughout the Northeast USA, generally occupying nutrient poor, acidic, xeric soils of either deep glacial sands or shallow soils overlain on rocky outcrops (Seischab and Bernard, 1991
). Phenotypic variation in P. rigida is more extreme than for many other woody species. While the species typically grows to mature heights of 8–12 m, there are several distinctive dwarf populations (Olsvig et al., 1979
; Gibson et al., 1999
): two within the New Jersey pine barrens, one in eastern Long Island, New York, one in the Shawangunk Mountains near Albany, New York, and one in the Berkshire Mountains of Massachusetts. In these sites, mature pitch pines, which comprise the forest canopy, range from less than 1 m to 3 m tall, with a twisted, crooked, multistemmed growth form. In addition to differences in stature and growth form, the New Jersey populations of dwarf pines have been shown to have precocious reproduction and increased allocation to reproduction compared to normal-stature pines (Buchholz and Good, 1982
; Good and Good, 1975
). In eastern Long Island, naturally occurring P. rigida were found to begin reproduction at smaller sizes in dwarf sites than in the normal-stature sites (Landis et al., 2005
).
Studies on the causes of the differences between dwarf and normal-stature pitch pine populations have generally not distinguished between plastic responses to environmental factors and environmental factors as selective agents for evolutionary differentiation. Factors suggested to be responsible for the dwarf form include frequency and intensity of fire (Lutz, 1934
; Andresen, 1959
; Olsvig et al., 1979
; Givnish, 1981
; Jordan, 1999
; Kurczewski and Boyle, 2000
), soil fertility (Olsvig et al., 1979
; Whittaker, 1979
; Landis et al., 2005
), soil hard-pan, toxic levels of aluminum, and high wind speeds (see Andresen, 1959
, for a review of the early literature). Good and Good (1975)
reported differences in growth rate and reproductive allocation in greenhouse/nursery experiments between plants grown from seeds of dwarf and normal-stature pitch pine populations from the New Jersey pine barrens in an early study that was subject to a number of statistical limitations, including limited numbers of genotypes.
While there has been much discussion about the causes of contrasting statures of pitch pines on Long Island and in New Jersey, no studies have attempted even a crude partitioning of the sources of variation in the height or form of these trees. To evaluate the relative importance of environmental and genetic sources of variation in the growth and form of P. rigida, we conducted long-term reciprocal transplant experiments using seeds from a number of normal-stature and dwarf pine populations on Long Island. We collected seeds from a large number of maternal families from several locations within both the dwarf and normal-stature pine barrens and grew them in several sites in the field to evaluate variation at the family, population, and population-type (dwarf vs. normal-stature) levels. The experimental planting sites included a normal-stature pine barrens site; an intermediate-stature, transitional site; and two dwarf sites. In addition to studying the growth, form, and time to reproduction of the transplanted seedlings, we also compared their survival across sites and across seedlings originating from different populations.
While our approach does not allow us to evaluate genetic variation in a strict sense (because maternal effects are included in addition to additive genetic variation in the variation attributable to seed sources and maternal families, Mitchell-Olds and Rutledge, 1986
; Lynch and Walsh, 1998
), we can set an upper bound on the magnitude of the genetic variation in growth and form. By separating phenotypic plasticity in response to environmental factors from those due to genetic and maternal effects, these experimental studies also allow us to make initial inferences about the possible strength of natural selection acting on dwarf stature and related morphological and life history traits.
This study should also be useful in providing information necessary to make informed judgments about how to preserve the dwarf pines in the face of pressures from development. The dwarf pine locations comprise a globally rare community type, with high priority for conservation (Jordan, 1998
). Determining conservation strategies in the face of habitat loss and fragmentation will require an understanding of the factors that determine the nature of these distinctive pines.
MATERIALS AND METHODS
We conducted two experiments. One reason for doing so was the high mortality rate of transplanted seedlings in the first experiment. Moreover, conducting a second experiment allowed us to examine the robustness of our results by permitting us to ask whether the results depended on the particular source populations used. The same four sites were used for transplanting in both experiments.
Seed source populations and transplanting sites are listed in Table 1. Some locations served as both sources of seeds and sites for transplanting, but other locations served as only seeds sources or only sites for transplanting. One reason for this was ecological: seed collection had to be from sites with available seeds, which, particularly in the dwarf pine areas, eliminated most of the recently burned locations and especially those that had been heavily burned (cf. Landis et al., 2005
). Similarly, seeds had to be transplanted into sites that had been burned for the experiments to make ecological sense, as there is normally little germination or recruitment in P. rigida in intact forest (Little and Garrett, 1990
). Another reason for using different seed source locations and transplanting sites was that we wished to examine the correspondence of variation in growth rate, form, and time to reproduction to variation in allozyme frequencies in data that had already been collected. For this reason, seed sources for the first experiment were those used in that previous study (P. Teese and J. Gurevitch, SUNY Stony Brook, unpublished data).
|
; Olsvig et al., 1979
). We collected both serotinous and nonserotinous cones in the fall and heated the serotinous cones in the oven at 40°C for 2 h to retrieve the seeds; this produces no known effects on the seeds or embryos within (Good and Good, 1975
; Knapp and Anderson, 1980
).
Germination
Seeds from four populations (DN, DS, SH, SW) collected in 1995 were germinated in the SUNY Stony Brook greenhouse in May 1996 for the first experiment. For the second experiment, we used seeds from four populations (DE, DW, RP, and SC) collected in 1997 and two more populations (DN and SP) collected in 1995 to maintain the continuity of the study. To simulate the sandy pine barren soil, we germinated seeds in 100% sand in the greenhouse.
Transplantation
Experimental plantings were carried out in areas of pitch pine forest that had experienced severe crown fires in August 1995. All sites were in areas of forest dominated by pitch pine, with dense understories comprised largely of scrub oak (Quercus ilicifolia) and ericad shrubs. The transplanting sites were chosen to span the range of heights of mature P. rigida. The site with the tallest trees was the RP site (40°54.5' N, 72°54.5' W), dominated by 10–15 m tall pitch pine. Two of the transplanting sites (D1 and D2; 40°51.1' N, 72°39.3' W and 40°51.3' N, 72°39.4' W, respectively) were within the dwarf pine plains, with mature pitch pine heights of 1–3 m. The fourth transplanting site SP (40°51.5' N, 72°42.1' W) has pitch pine of an intermediate-stature (6–10 m tall).
In early May 1997, seedlings for the first experiment were transplanted to each of the transplanting sites; hereafter we refer to this as Exp 97. For each seed collection source, we transplanted seedlings from 12–14 maternal sibships, with the number of sibships planted and the number of seedlings per sibship depending on germination rate and survival up to that point. Within each site, 35 circular plots (radius of 0.55 m) were randomly chosen in open space between scrub oaks. Eight seedlings (two from each population source) chosen at random were planted equidistantly along the circle.
In late April 1999, seedlings from the second experiment were transplanted to the same four sites (D1, D2, RP, and SP); hereafter we refer to this as Exp 99. For each seed collection source, we transplanted seedlings from 12 maternal sibships. Our experience with Exp 97 suggested that mortality would be substantially greater in plantings in the dwarf sites than in the NS sires. We therefore set 50 circular plots each (with a radius of 0.55 m) in D1 and D2, and 20 circular plots each in RP and SP. Along the circles eight seedlings were planted, using at least one seedling from each of six population sources, with one additional dwarf and one NS seedling chosen randomly.
Data recording
Census and measurement of the plants was performed at least twice annually from 1997 through 2000 and once annually in 2001 and in 2003. At each census, seedling survival, height (from the ground to the apical meristem), and the number of cones borne were recorded. Growth form (single vs. multiple stems) was recorded beginning with the fall 1999 census.
Data analysis
We analyzed the natural log of final heights using PROC MIXED in SAS-PC v.9.1 (SAS Institute, Cary, North Carolina, USA). The fixed effect factors (source type [i.e., normal or dwarf], source population within source type, site, source type x site, and source within source type x site) were tested over the appropriate random effect terms (family within source x source type, plot within site, site x family within source x source type) calculated using restricted maximum likelihood estimates (REML); in those cases where the REML estimates were zero, the residual error was used in the F tests. Degrees of freedom in these unbalanced designs were calculated using the Satterthwaite approximation. We evaluated growth using profile analysis (multivariate repeated measures; Von Ende, 2001
). For the 1997 experiment, heights (untransformed) taken in the spring of 1997, 1998, 1999, 2000, and summers of 2001 and 2003 were used with the simplifying assumption that the measurements were equally spaced in time. For the 1999 experiment, heights from the latter four dates were used in the profile analysis. Residuals were tested for normality for all analyses, and models were compared using Akaike's information criterion (AIC) when appropriate.
We analyzed both survival data and data on time to reproduction with accelerated failure time (AFT) models (Fox, 2001
; Kalbfleisch and Prentice, 2002
). These models can accommodate censored data—that is, cases for which the exact timing of the event is unknown. There are three kinds of censorship in these data. Plants whose tags were lost were considered right-censored as of the last census at which they were located—that is, the true death date or time to reproduction is sometime after that date. Plants still living (or still not having produced cones) at the end of the census were considered right-censored as of the last census. All other plants were considered to be interval-censored—that is, the true date of the event was between the census at which it was recorded and the prior census.
AFT models can use a variety of error distributions and are therefore a standard approach for event-time data, which characteristically do not have symmetric errors. We fit models using loglogistic, lognormal, and Weibull distributions, as well as the exponential (constant mortality) distribution, and used the AIC to choose among the models.
To examine whether variation at the level of plots within sites affected the fit of the models, we included a "frailty" term for plots in the survival models and compared the resulting AICs with models that did not include this term. Frailty models are analogous to mixed-model ANOVA in many respects: the key assumption is that there is unmeasured heterogeneity at the level of the frailty term. Unlike mixed-model ANOVA, current algorithms permit only a single frailty term per model.
We explored the set of predictors of survival with a stepwise AIC procedure—that is, we sequentially removed and added terms to the models to find the set of predictors that yielded the lowest AIC. The full set of predictors we used was source type (normal vs. dwarf), source population within source type, maternal sibship within source within source type, and growing site, as well as all interactions between these terms. Given the model with the lowest AIC, we then used standard significance tests for these terms.
Decomposition of the sources of variation in the selected model was performed by analysis of deviance. Analysis of deviance is a generalization of analysis of variance (McCullagh and Nelder 1989
), used for cases (such as with survival data and in logistic regression) in which the residuals are not normally distributed. When this is the case, sums of squares (as used in ANOVA) are no longer useful to measure the discrepancy between model and data. Instead, the appropriate measurement is the model's deviance, defined as –2 times its loglikelihood (Edwards 1992
).
We used a similar approach for analyzing time to reproduction. In referring to this variable as time to reproduction, we do not mean to imply that these plants have a threshold age for reproduction rather than a threshold size; the proximate controls over the transition to reproduction are not known for this species. Because very few plants from the 1999 experiment had produced cones by the end of our experiment, we restricted our analysis to the 1997 experiment. This also means that our time-to-reproduction analyses capture only the earliest-maturing individuals rather than the time to reproduction of the entire planted cohorts, which cannot yet be determined. The relatively few events in this data set made it impossible to estimate models using interaction terms or sibships. Thus we evaluated models using source type, source population within source type, site, and frailty terms for plot within site.
We analyzed the relationship between stem form (whether plants had a single or multiple stems) and population source and growing site using logistic regression, again using a stepwise AIC approach to find the best model and then evaluating significance tests for each model term.
Data figures
The underlying variance structure in our experimental design (and in the data obtained) is more complicated than can be captured by error bars, as error bars cannot take into account the levels of nesting and multiple factors present in the experimental design and data. We provide figures to enable a quick visual understanding of the major patterns in the data (even though the point estimates provided in these figures also do not take into account the underlying variance structure), but providing error bars would imply that meaningful inferential analysis could be obtained by eye. We do not believe this is the case given the complexity of the experimental design and instead refer the reader to the results from appropriate inferential analyses detailed in the text and tables, rather than providing potentially misleading error bars.
RESULTS
Final plant height
Most of the variation in final heights in both Exp 97 and Exp 99 was due to site (Figs. 1, 2); the effect of site on ln final height was highly significant (F = 149.24, P < 0.0001, df = 3, 129 for Exp 97; F = 22.20, P < 0.0001, df = 3, 131 for Exp 99). There was no overall difference in either experiment in final heights for seeds from normal and dwarf sources (source type was NS; F = 0.08, P = 0.78, df = 1, 425 for Exp 97; F = 1.20, P = 0.27, df = 1, 204 for Exp 99. In Exp 97 normal source seedlings were taller than dwarf source seedlings at some sites and shorter at others (Fig. 1; there was a statistically significant interaction between source type [normal or dwarf] and site: F = 2.77, P = 0.041, df = 3, 414). Unlike the 1997 experiment, in Exp 99 there was no statistical interaction between source type (normal or dwarf) and site (Fig. 2; F = 1.08, P = 0.38, df = 12,149). For both experiments, other terms were small and not statistically significant.
|
|
For both the 1997 and 1999 experiments, the residuals for the natural log of height and for the profile analyses were non-normal, although they did not differ greatly from normality. Untransformed and other transformations examined were also non-normal, but qualitative results were the same for untransformed data and other measures examined, indicating that the F tests were robust to the limited degree of violation of this assumption found in the data.
Single vs. multiple stems
The only significant influence on stem habit (single vs. multiple stems) in Exp 97 was the growth environment, with a higher proportion of seedlings exhibiting multiple stems at dwarf vs. non-dwarf sites (Fig. 3). The stepwise AIC procedure using logistic regression led to a model including only site as a predictor. The regression coefficients are shown in Table 2. A classical approach—using a full model and its associated F tests—leads to the same conclusion: only site was a significant predictor of multiple vs. single stems. We conducted the same analyses for the 1999 experiment, with qualitatively the same conclusion (data not shown).
|
|
|
|
Survival in the 1997 experiment was strikingly lower at dwarf sites than in nondwarf sites and was also influenced by seedling origin (Fig. 5). The best-fitting model included terms for source type, source population within source type, maternal family within source population within source type, site, and site x source type. The analysis of deviance is shown in Table 4. Overwhelmingly, the model deviance is reduced by site and by the frailty term for plot within site. Source type makes almost no contribution to reducing model deviance. Much smaller, but still statistically significant, contributions are made by factors that may involve genetic differentiation: source population within source type, and family within source population within type. There is also a small but statistically significant contribution from the interaction between source type and site.
|
|
|
Growth environment far exceeded site of seed origin in determining the height, growth form, and survival of the pine seedlings and saplings in these experiments. The only trait in which seed origin had a similar magnitude of effect as growth environment was time to reproduction. The differences observed between plantings in dwarf and nondwarf sites largely paralleled those seen in natural stands and in our previous observations of naturally occurring seedlings (Landis et al., 2005
); plants in dwarf sites grew more slowly, survived at a lower rate, and were more likely to have multiple branches than plants at nondwarf sites. Many of the striking phenotypic differences between dwarf and normal-stature stands of pitch pines seen on Long Island therefore appear to be less a matter of genetic differentiation among populations than of plastic phenotypic responses to environmental factors.
We did find substantial differences due to seed origin in the life history trait of age at first reproduction. It is believed that short fire-return intervals have historically been characteristic for the dwarf populations on Long Island (Jordan et al., 2003
); if so, early reproduction in these populations could represent adaptation in a demographic parameter at very local scales. If additive genetic variance is an important component of these differences—which cannot be distinguished with our design—this may indicate strong selection for early maturation in the dwarf areas. Given that these pines are wind-pollinated and the distances between populations of the different types are very small (Olsvig et al., 1979
), gene flow is likely to be great between dwarf and normal-stature pines. When gene flow is high, maintenance of genetic differentiation between populations requires strong selection (Endler, 1977
). While we have observed only relatively slight differences in height based on seed origin, it is possible that the earlier reproduction of trees from dwarf populations will ultimately affect their height, as these plants may invest relatively more in reproduction and less in height growth than trees originating from normal-stature populations (Ledig et al., 1976
).
While some of the variation in time to reproduction was associated with seedling origin, we also found substantial environmental variation in life-history characteristics, including significant variation in both size and reproductive status among transplanting sites and between plots within these sites. This indicates a heterogeneity in demographic parameters that can have important consequences for increasing population growth and reducing risk of population extinction (Fox and Kendall, 2002
; Kendall and Fox, 2002
, 2003; Fox, 2005
).
The environmental factors responsible for the phenotypic differences between dwarf and normal-stature populations of pitch pine appear likely to be edaphic, because climate differences between these sites seem unlikely, considering that short (<1 km) distances separate such sites and the region has only slight topographic relief. This raises the question of which edaphic factors might be involved and by what mechanisms these factors exert such profound effects on the plants. Several edaphic factors have been suggested to differ between sites with dwarf and nondwarf pitch pine populations on Long Island. We previously found that organic matter and concentrations of aluminum and of several macronutrient elements (P, K, Ca, Mg) were higher at the RP and SP sites than in several dwarf pine sites (Landis et al., 2005
). Olsvig et al. (1979)
found that dwarf site soils had a shallower A horizon than the soils in normal-stature sites and suggested that this was a result of frequent fires consuming soil organic matter in the dwarf sites. While soil organic matter and macronutrient concentrations are plausible explanations for differences in the growth rate of plants between sites, it is not obvious how these or other edaphic factors might influence the tendency of the trees to have multiple vs. single stems.
Short fire-return intervals have sometimes been invoked as a causative agent in the formation of the dwarf pitch pine communities. As our results indicate a relatively small role for genetic differentiation between dwarf and normal-stature pines, fire as a selective agent does not appear likely to be an important factor underlying the dwarf phenotype, as has been suggested by several authors (e.g., Olsivg et al., 1979; Givnish, 1981
; Jordan et al., 2003
). We cannot, however, rule out the hypothesis that exposure to frequent fires may contribute to dwarfing in individual plants (Andresen, 1959
). If this is the case, environmental factors may play an even larger role in determining the phenotypic differences between the dwarf and normal-stature pines than is indicated by our experiment.
While much of the variation that we observed in P. rigida was environmental rather than genetic in origin, it is unknown whether phenotypic plasticity is as important in other cases in which striking phenotypic differences are observed between populations of woody plants over short distances. Several authors have addressed this question by using molecular markers to examine the genetic distinctiveness of phenotypically distinct populations of woody plants, with varying results. A number of studies have failed to find molecular differences between populations that are phenotypically quite distinct. In pitch pines, Guries and Ledig (1982)
found no significant differences in allozyme allele frequencies between two dwarf populations from New Jersey and nine normal-stature populations sampled from across a wide portion of the species range. Rogers et al (1999)
found much stronger genetic differentiation in allozyme loci between individuals stands of krummholz and normal-stature whitebark pine (Pinus albicaulis) than between the two types of stands. Using RAPD markers, Heaton et al. (1999)
found no evidence for genetic differentiation between dwarf and normal populations of Manilkara zapota (Sapotaceae) in the Yucatan peninsula, despite their considerable phenotypic differentiation. Plants in swamps have a short, shrubby growth form (as well as smaller fruits, leaves, and seeds), while plants in nearby forests are tall, straight trees. In contrast, in several studies, differentiation in molecular markers coincides with population differences in morphological traits. Dahdouh-Guebas et al. (2004)
found differentiation in RAPD markers between nearby landward and seaward populations of the mangrove Avicennia marina that differ in plant height (mean of 3.2 m in landward population and 10.2 m in the seaward population), as well as leaf and pneumatophore morphology. Similarly, Grant and Mitton (1977)
showed differences in allozyme allele frequencies between krummholz and regular growth forms of Abies lasiocarpa and Picea engelmanii over relatively short distances.
Studies on population differentiation in molecular markers in natural populations are valuable in elucidating questions of gene flow and population structure, but they cannot resolve whether genetic factors underlie the observed differences between populations in particular quantitative genetic characters. A lack of variation between the populations in molecular markers might indicate that the phenotypic differences between populations are due to plastic responses to the environment. Alternatively, it might indicate that there is differentiation in characters subject to natural selection, but sufficient gene flow between populations to prevent a buildup of differentiation in selectively neutral genetic markers (Aitken and Libby,
1994). Merila and Cnorkrak (2001)
performed a meta-analysis of studies that compared differentiation between populations for both neutral molecular marker loci and quantitative traits in a wide variety of organisms. They found that differentiation among populations was typically greater for quantitative characters than for neutral loci, particularly for morphological characters. Conversely, differentiation may exist between populations for neutral molecular loci without indicating that this variation is correlated with genetic differences in morphological or life-history characters (Linhart et al., 1989
; Premoli, 2003
).
In addition to this type of molecular marker comparison between populations, results from common-garden, greenhouse, and nursery experiments have also sometimes been used to suggest that genetic differences underlie observed phenotypic differences between populations. In a common-garden trial with pitch pine, Ledig et al. (1976)
found that by the second year, seedlings originating from two dwarf populations in New Jersey were significantly shorter than seedlings originating from 11 normal-stature provenances from the New Jersey coastal plain. Good and Good (1975)
in a greenhouse/nursery experiment with New Jersey P. rigida found that seedlings originating from dwarf populations were shorter, had less biomass, and reproduced earlier than seedlings originating from normal-stature populations. Using a common-garden approach, Steiner and Berrang (1990)
found greater cold tolerance in P. rigida seedlings originating from a localized "cold pocket" depression in Pennsylvania than in seedlings originating from nearby, warmer habitats.
These experiments have demonstrated that there is some genetic (or maternal effects) component to various phenotypic trait differences between populations. However, they were unable to estimate how the magnitude of these genetic differences compared to the environmental component of variation because they did not grow the plants across the range of natural environmental variation.
Reciprocal transplant studies provide a methodology that can be used to estimate the relative contributions of plastic responses to the environment vs. plant origin (genetic + maternal effects) as components of phenotypic variation. A number of reciprocal transplant experiments have been performed with populations of woody plant species, examining questions such as whether there is evidence of a home-site advantage (e.g., Sork et al., 1993
; Kittelson and Maron, 2002
) or the suitability of various source populations for forestry plantings across a range of habitats (e.g., Wu and Ying, 2004
; Mylecraine et al., 2005
).
A relatively small number of studies have been performed specifically to compare seed source (genetic + maternal effects) vs. environmental influences on known phenotypic differences between populations of woody plants. For example, Boege and Dirzo (2004)
studied variation in growth and phenolic production in Dialium guianense (Caesalpiniaceae), a canopy tree in SE Mexico. Floodplain plants grew faster and had higher phenolic concentrations than plants in hills nearby. Their reciprocal transplant experiment pointed to phenotypic plasticity as the major source of this variation. Davy and Gill (1984)
reciprocally transplanted seedlings originating from morphologically distinct bog and heath populations of the Betula pendula–B. pubescens species complex. They found that plastic variation and seed source variation predominated for different characters. Plant size and leaf size were largely environmentally determined, while heritabilities were high for leaf morphometric measurements.
The striking phenotypic plasticity responsible for local differentiation that we found in growth and life history characteristics in P. rigida may therefore be more general for woody plants than is commonly assumed. While selection may act in opposition to gene flow to create genetically based phenotypic differentiation between populations (Linhart and Grant, 1996
), it may also act to create adaptive phenotypic plasticity (Pigliucci, 2001
). The result of selection for phenotypic plasticity may be to create phenotypic differentiation at a local scale that is not genetically based. Further study in P. rigida and in other cases is needed to better understand the interplay of these types of selection and of other factors (such as epistasis and maternal effects) in determining the nature of phenotypic variation at a range of spatial scales. In addition, as we have pointed out here, these aspects of the ecology and evolutionary biology of many species have important implications for conservation biology. Although the mechanisms for the strong environmental influence on dwarfing in P. rigida are not well understood, the results of the current study suggest that conservation efforts seeking to preserve the globally rare dwarf pine communities (e.g., Jordan et al., 2003
) should focus on preserving the sites where dwarf plants are found rather than on preserving their genes.
FOOTNOTES
1 The authors thank P. Teese, T. Howard, R. Misra, K. Brown, E. Woo, and others who have aided in the fieldwork and the anonymous reviewers and the editors for their thoughtful suggestions. This work was supported by National Science Foundation grant DEB-9806923 to J. G. and G. A. F. and a grant from The Nature Conservancy to J. G. This is contribution 1150 in Ecology and Evolution, State University of New York at Stony Brook. 
2 The order of the names of the first two authors is arbitrary; W.F. and D.R.T. equally contributed as first authors.
8 Author for correspondence (taubd{at}southwestern.edu
)
LITERATURE CITED
Aitken S. N. Libby W. J. Evolution of the pygmy-forest edaphic subspecies of Pinus contorta across an ecological staircase. Evolution 48: 1009-1019.
Andresen J. W.. 1959. A study of pseudo-nanism in Pinus rigida Mill. Ecological Monographs 29: 309-332.[CrossRef]
Antonovics J. Bradshaw A. D.. 1970. Evolution in closely adjacent plant populations. VIII. Clinal patterns at a mine boundary. Heredity 25: 349-362.[ISI]
Arno S. F.. 1984. Timberline: mountain and arctic forest frontiers Mountaineers Books, Seattle, Washington, USA.
Beckman J. S. Mitton J. B.. 1984. Peroxidase allozyme differentiation among successional stands of ponderosa pine. American Midland Naturalist 112: 43-49.[CrossRef][ISI]
Berteaux D. Réale D. McAdam A. G. Boutin S.. 2004. Keeping pace with fast climate change: can arctic life count on evolution?. Integrative and Comparative Biology 44: 140-151.
Boege K. Dirzo R.. 2004. Intraspecific variation in growth, defense and herbivory in Dialium guianense (Caesalpiniaceae) mediated by edaphic heterogeneity. Plant Ecology 175: 59-69.[CrossRef][ISI]
Brady K. U. Kruckeberg A. R. Bradshaw H. D.. 2005. Evolutionary ecology of plant adaptation to serpentine soils. Annual Review of Ecology, Evolution and Systematics 36: 243-266.
Buchholz K. Good R. E.. 1982. Density, age structure, biomass and net annual aboveground productivity of dwarfed Pinus rigida Mill. from the New Jersey Pine Barren Plains. Bulletin of the Torrey Botanical Club 109: 24-34.[CrossRef][ISI]
Buchholz K. Motto H.. 1981. Abundances and vertical distributions of mycorrhizae in Plains and Barrens forest soils from the New Jersey Pine Barrens. Bulletin of the Torrey Botanical Club 108: 268-271.[CrossRef][ISI]
Clausen J.. 1965. Population studies of alpine and subalpine races of conifers and willows in the California high Sierra Nevada. Evolution 19: 56-68.[Medline]
Dahdouh-Guebas F. R. De Bondt R. Abeysinghe P. D. Kairo J. G. Cannicci S. Triest L. Koedam N.. 2004. Comparative study of the disjunct zonation pattern of the grey mangrove Avicennia marina (Forsk.) Vierh. in Gazi Bay (Kenya). Bulletin of Marine Science 74: 237-252.[ISI]
Davy A. J. Gill J. A.. 1984. Variation due to environment and heredity in birch transplanted between heath and bog. New Phytologist 97: 489-505.[CrossRef][ISI]
Edwards A. W. F.. 1992. Likelihood Johns Hopkins University Press, Baltimore, Maryland, USA.
Endler J. A.. 1977. Geographic variation, speciation, and clines Princeton University Press, Princeton, New Jersey, USA.
Fox G. A.. 2001. Failure-time analysis: studying times to events and rates at which events occur. In S. M. Scheiner and J. Gurevitch [eds.] Design and analysis of ecological experiments 235-266 Oxford University Press, New York, New York, USA.
Fox G. A.. 2005. Population viability and extinction risk of heterogeneous populations. Ecology 86: 1191-1198.[CrossRef][ISI]
Fox G. A. Kendall B. E.. 2002. Demographic stochasticity and the variance reduction effect. Ecology 83: 1928-1934.[ISI]
Gibson D. J. Zampella R. A. Windisch A. G.. 1999. New Jersey pine plains: the "true barrens" of the New Jersey pine barrens. In R. C. Anderson, J. S. Fralish and J. M. Baskin [eds.] Savannas, barrens, and rock outcrop plant communities of North America 52-66 Cambridge University Press, New York, New York, USA.
Givnish T. J.. 1981. Serotiny, geography, and fire in the Pine Barrens of New Jersey. Evolution 35: 101-123.[CrossRef][ISI]
Good R. E. Good N. F.. 1975. Growth characteristics of two populations of Pinus rigida Mill. from the Pine Barrens of New Jersey. Ecology 56: 1215-1220.[CrossRef][ISI]
Good R. E. Good N. F. Andresen J. W.. 1979. The pine barren plains. In R. T. T. Forman [ed.] Pine Barrens: ecosystem and landscape 283-295 Academic Press, New York, New York, USA.
Grant M. C. Mitton J. B.. 1977. Genetic differentiation among growth forms of Engelmann spruce and subalpine fir at tree line. Arctic and Alpine Research 9: 259-263.[CrossRef][ISI]
Green D. M.. 2005. Designatable units for status assessment of endangered species. Conservation Biology 19: 1813-1820.[CrossRef][ISI]
Guries R. P. Ledig F. T.. 1982. Genetic diversity and population structure in pitch pine (Pinus rigida Mill). Evolution 36: 387-402.[CrossRef][ISI]
Hanson W. D.. 1966. Effect of partial isolation (distance) migration, and different fitness requirements among environmental pockets upon steady state gene frequencies. Biometrics 22: 453-468.[CrossRef][ISI][Medline]
Heaton H. J. Whitkus R. Go'mez-Pompa A.. 1999. Extreme ecological and phenotypic differences in the tropical tree chicozapote (Manilkara zapota (L.) P. Royen) are not matched by genetic divergence: a random amplified polymorphic DNA (RAPD) analysis. Molecular Ecology 8: 627-632.
Hughes J. B. Daily G. C. Ehrlich P. R.. 1997. Population diversity: its extent and extinction. Science 278: 689-692.
Jordan M.. 1998. Ecological effects of a large and severe summer wildfire in the Long Island Dwarf Pine Barrens Report on two years of post-wildfire research and monitoring. Funded in part by a 1996 grant from the Rodney Johnson and Katharine Ordway Stewardship Endowment. The Nature Conservancy, Long Island Chapter, Cold Spring Harbor, New York, USA.
Jordan M.. 1999. Conceptual ecological models for the Long Island pine barrens The Nature Conservancy, Long Island Chapter, Cold Spring Harbor, New York, USA.
Jordan M. Patterson A. III Windisch A. G.. 2003. Conceptual ecological models for the Long Island pine barrens: implication for managing rare plant communities. Forest Ecology and Management 185: 151-168.[CrossRef][ISI]
Kalbfleisch J. Prentice R.. 2002. The statistical analysis of failure time data Wiley, Hoboken, New Jersey, USA.
Kendall B. Fox G. A.. 2002. Variation among individuals and reduced demographic stochasticity. Conservation Biology 16: 109-116.[CrossRef][ISI]
Kittelson P. M. Maron J. L.. 2001. Fine-scale genetically based differentiation of life-history traits in the perennial shrub Lupinus arboreus. Evolution 55: 2429-2438.[CrossRef][ISI][Medline]
Knapp A. K. Anderson J. E.. 1980. Effect of heat on germination of seeds from serotinous lodgepole pine cones. American Midland Naturalist 104: 370-372.[CrossRef][ISI]
Kurczewski F. E. Boyle H. F.. 2000. Historical changes in the pine barrens of central Suffolk County, New York. Northeastern Naturalist 7: 95-112.
Landis R. M. Gurevitch J. Fox G. A. Fang W. Taub D. R.. 2005. Variation in recruitment and early demography in Pinus rigida following crown fire in the pine barrens of Long Island, New York. Journal of Ecology 93: 607-617.[CrossRef]
Ledig F. T. Fryer J. H.. 1972. A pocket of variability in Pinus rigida. Evolution 26: 25-266.
Ledig F. T. Lambeth C. C. Linzer D. I. H.. 1976. Nursery evaluation of a pitch pine provenance trial. Proceedings of the Twenty-third Northeast Forest Tree Improvement Conference August 4–7 1975,New Brunswick, New Jersey, USA,93-108.
Ledig F. T. Little S.. 1979. Pitch pine (Pinus rigida Mill.): ecology, physiology, and genetics. In R. T. T. Forman [ed.] Pine barrens: ecosystem and landscape 347-371 Academic Press, New York, New York, USA.
Lewontin R. C.. 1974. The genetic basis of evolutionary change Columbia University Press, New York, New York, USA.
Linhart Y. B. Grant M. C. Montazer P.. 1989. Experimental studies in ponderosa pine. I. Relationship between variation in proteins and morphology. American Journal of Botany 76: 1024-1032.[CrossRef][ISI]
Linhart Y. B. Grant M. C.. 1996. Evolutionary significance of local genetic differentiation in plants. Annual Review of Ecology and Systematics 27: 237-277.[CrossRef][ISI]
Little S. Garrett P. W.. 1990. Pinus rigida Mill. Pitch pine. In R. M. Burns and B. H. Honkala [eds.] Silvics of North America 1. Conifers. Agriculture Handbook No. 654, 877 U.S. Department of Agriculture, Forest Service, Washington, D.C., USA.
Lovelock C. E. Feller I. C. Mckee M. L. Engelbrect B. M. J. Ball M. C.. 2004. The effect of nutrient enrichment on growth, photosynthesis and hydraulic conductance of dwarf mangroves in Panamá. Functional Ecology 18: 25-33.[CrossRef][ISI]
Lutz H. L.. 1934. Ecological relations in the pitch pine plains of southern New Jersey. Yale University School of Forestry Bulletin 38: 1-80.
Lynch M. Walsh B.. 1998. Genetics and analysis of quantitative traits Sinauer, Sunderland, Massachusetts, USA.
McCullagh P. Nelder J. A.. 1989. Generalized linear models, 2nd ed Chapman and Hall, London, U.K.
McNeilly T. Antonovics J.. 1968. Evolution in closely adjacent plant populations. IV. Barriers to gene flow. Heredity 23: 205-218.[ISI]
Merila J. Cnorkrak P.. 2001. Comparison of genetic differentiation at marker loci and quantitative traits. Journal of Evolutionary Biology 14: 892-903.[CrossRef][ISI]
Mitchell-Olds T. Rutledge J. J.. 1986. Quantitative genetics in natural plant populations: a review of the theory. American Naturalist 127: 379-402.[CrossRef][ISI]
Mitton J. B. Linhart Y. B. Hamrick J. L. Beckman J. S.. 1977. Observations on the genetic structure and mating system of ponderosa pine in the Colorado Front Range. Theoretical and Applied Genetics 51: 5-13.
Mopper S. Mitton J. B. Whitham T. G. Cobb N. S. Christensen K. M.. 1991. Genetic differentiation and heterozygosity in pinyon pine associated with resistance to herbivory and environmental stress. Evolution 45: 989-995.[CrossRef][ISI]
Mylecraine K. A. Kuser J. E. Zimmermann G. L. Smouse P. E.. 2005. Rangewide provenance variation in Atlantic white-cedar (Chamaecyparis thyoides): early survival and growth in New Jersey and North Carolina plantations. Forest Ecology and Management 216: 91-104.[CrossRef][ISI]
Novak S. J. Mack R. N. Soltis D. E.. 1991. Genetic variation in Bromus tectorum (Poaceae): population differentiation in its North American range. American Journal of Botany 78: 1150-1161.[CrossRef][ISI]
Olsvig L. S. Cryan J. F. Whittaker R. H.. 1979. Vegetational gradients of the pine plains and barrens of Long Island, New York. In R. T. T. Forman [ed.] Pine barrens: ecosystem and landscape 265-282 Academic Press, New York, New York, USA.
Pigliucci M.. 2001. Phenotypic plasticity: beyond nature and nurture Johns Hopkins University Press, Baltimore, Maryland, USA.
Premoli A. C.. 2003. Isozyme polymorphisms provide evidence of clinal variation in Nothofagus pumilio. Journal of Heredity 94: 218-226.
Rajakaruna N.. 2004. The edaphic factor in the origin of plant species. International Geology Review 46: 471-478.[ISI]
Rogers D. L. Millar C. I. Westfall R. D.. 1999. Fine-scale genetic structure of whitebark pine (Pinus albicaulis): associations with watershed and growth form. Evolution 53: 74-90.[CrossRef][ISI]
Seischab F. K. Bernard J. M.. 1991. Pitch pine (Pinus rigida Mill) communities in central and western New York. Bulletin of the Torrey Botanical Club 118: 412-423.[CrossRef][ISI]
Slatkin M.. 1987. Gene flow and the geographic structure of natural populations. Science 236: 787-792.
Sork V. L. Stowe K. A. Hockwender C.. 1993. Evidence for local adaptation in closely adjacent subpopulations of northern red oak (Quercus rubra L.) expressed as resistance to leaf herbivores. American Naturalist 142: 928-936.[CrossRef][ISI]
Steiner K. C. Berrang P. C.. 1990. Microgeographic adaptation to temperature in pitch pine progenies. American Midland Naturalist 123: 292-300.[CrossRef][ISI]
Tonsor S. J. Kalisz S. Fisher J. Holtsford T. P.. 1993. A life history based study of population structure: seed bank to adults in Plantago lanceolata. Evolution 47: 833-843.[CrossRef][ISI]
von Ende C. N.. 2001. Repeated-measures analysis: growth and other time dependent measures. In S. Scheiner and J. Gurevitch [eds.] The design and analysis of ecological experiments 134-157 Oxford University Press, Oxford, UK.
Whittaker R. H.. 1979. Vegetational relationships of the pine barrens. In R. T. T. Forman [ed.] Pine barrens: ecosystem and landscape 315-331 Academic Press, New York, New York, USA.
Wu H. X. Ying C. C.. 2004. Geographic pattern of local optimality in natural populations of lodgepole pine. Forest Ecology and Management 194: 177-198.[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
