| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Population Biology |
T. H. Morgan School of Biological Sciences, University of Kentucky, Lexington, Kentucky 40506 USA
Received for publication June 3, 2003. Accepted for publication December 18, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Arabidopsis thaliana • Brassicaceae • colonization • flowering time • germination • local adaptation • range expansion • seasonal dormancy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). They frequently expand their geographic range or contribute foreign genotypes to distant locations in this manner. The efficiency of long-distance dispersal to these processes, however, depends on the performance of the dispersing propagules in the new location. If local adaptation to variable microenvironments is prevalent, then dispersing propagules may not perform well in new locations (Jain and Bradshaw, 1966
; Antonovics, 1968
; Macnair, 1981
). Alternatively, if similar microenvironments are available throughout the geographic range of the plants, then dispersing propagules may more easily establish in new locations.
The ability of dispersing propagules to contribute to local genetic variation depends on how foreign genotypes are selected within the new population. Such selection will depend on additive effects of immigrant genes in the new location and also within the background of the new gene pool (Wright, 1932
; Tonsor and Goodnight, 1997
; Wade, 2002
). In highly selfing species, as in many weeds, recombination and mixing of the foreign and native genomes may be low, and adaptive gene combinations may be preserved in new locations. Thus the evolutionary ecology of weedy species depends on how divergent populations are over the geographic range of the species, how particular divergent characters are selected in new locations, whether combinations of characters and genes contribute to performance, and whether combinations of characters and genes are likely to be preserved in the new location through low recombination.
Arabidopsis thaliana (L.) Heynh. (Brassicaceae) is a highly selfing, weedy annual that has been introduced into North America, most likely from Europe, within the past few hundred years (Sharbel et al., 2000; Vander Zwan et al., 2000
; Hoffmann, 2002
). It inhabits a broad geographic range within North America, and it is considered to be highly mobile through long-distance dispersal of its tiny seeds. While genetic variation within populations is quite low (Bergelson et al., 1998
), ecotypes from different locations have been shown to differ from one another in important life history traits—especially those ecotypes from the European and Central Asian collections available from the Arabidopsis seed stock centers (e.g., The Arabidopsis Biological Resource Center and Nottingham Arabidopsis Stock Centre). Several studies have demonstrated ecotypic variation in flowering time as well as in important morphological and physiological characters (e.g., Griffing and Scholl, 1991
; Pigliucci and Schlichting, 1996
; Pigliucci, 1998
; Nordborg and Bergelson, 1999
; Mitchell-Olds, 2001
; Pigliucci and Marlow, 2001
; Kover and Schaal, 2002
). Much less is known about patterns of variation among populations in North America, the continent of its more recent introduction, although recent studies have documented differentiation among New England populations in a common greenhouse environment (Dorn et al., 2000
).
Arabidopsis thaliana, like many important weeds, typically displays a "winter annual" life history, in which seeds germinate in the autumn, plants overwinter as rosettes, and they grow, flower, set seeds, and die in the spring. In some populations, plants display a "spring annual" life history, in which seeds germinate in the early spring, and they flower, set seeds, and die all during the same season. This basic life history sequence varies among known A. thaliana ecotypes worldwide (Ratcliffe, 1965
; Effmertova, 1967
; Evans and Ratcliffe, 1972
; Nordborg and Bergelson, 1999
), as well as among natural populations in North America (K. Donohue, Harvard University, and L. Dorn, University of Wisconsin, Oshkosh, personal observation). Plants in some populations have even been observed to flower in the autumn (Thompson, 1994
), indicating that an "autumn annual" life history is also possible in A. thaliana populations. Populations in Kentucky have been characterized as winter annuals that germinate in autumn (Baskin and Baskin, 1983
), whereas New England populations have been seen to germinate either in autumn or in spring and to flower in both seasons as well (L. Dorn and K. Donohue, personal observation). Spring germination in New England populations can approach 54% in some locations. The mechanism of "summer annual," "winter annual," and "autumn annual" life-histories in A. thaliana is determined both by germination timing and vernalization requirements for flowering (Napp-Zinn, 1976
; Nordborg and Bergelson, 1999
). The expression of adult life-history characters depends on the timing of germination in natural populations of A. thaliana, and this dependency is mediated by the seasonal environment experienced by germinants (Donohue, 2002
). Therefore, characterizing basic life-history variation of natural populations of A. thaliana requires studies conducted under natural seasonal conditions, beginning with seed germination.
Arabidopsis thaliana is highly suitable for examining the dynamics of the introduction and spread of weedy species because it is fairly recently introduced, has long-distance dispersal, and harbors genetic variation in important life-history characters. In addition, it is a highly selfing species (Abbott and Gomes, 1989
), typical of many weeds, and thereby offers the chance to investigate the dynamics of adaptation to new environments in species in which linkage disequilibrium may be pronounced. It has been shown in A. thaliana that, while linkage disequilibrium declines within approximately 1 cM among globally distributed collections, it remains substantial within local populations (Nordborg et al., 2002
). Therefore, gene combinations, and therefore character combinations, may be important influences on the dynamics of adaptation in this weed. To characterize these dynamics, it is necessary to conduct artificial dispersal events, compare the performance of local and introduced populations, and determine characters and combinations of characters that influence performance of the different populations.
To compare life-history expression and performance of local and newly dispersed foreign seeds of A. thaliana, we used field-collected seeds collected from three geographic regions in North America in a common garden experiment conducted in Kentucky. We addressed the following specific questions. (1) Do populations differ in life-history expression, and is life-history variation structured at a regional scale? (2) Which life-history characters, and combinations of life-history characters, contribute most to variation in survival and fecundity? (3) Do local populations perform better than newly introduced populations? (4) What is the contribution of life-history characters to similarities or differences in performance between plants from different populations and regions?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The common garden was located at the Ecological Research Facility (ERF) of the T. H. Morgan School of Biological Sciences of the University of Kentucky in Lexington, Kentucky USA. ERF is located at latitude 38°04.834' N, longitude 084°28.397' W, at approximately 305 m above sea level. The plot was located in a grassy field and was mowed, tilled, and raked in late summer. Soil was taken from the experimental plot after tilling for use as the potting medium. The soil was autoclaved for 20 min at 121°C at 103 kPa to kill any naturally occurring seeds. The autoclaved soil was placed in Jiffy Poly-Pak (5 cm) square biodegradable peat pots. Two 0.6 m access strips were made in a 5 x 3 m2 plot, giving three 0.6 m wide by 2.8 m long strips in which the pots were planted.
In September 2000, seeds were sprinkled into pots that had been planted into the field. Approximately 20 seeds from a single plant were sprinkled on the soil surface of a pot. Except for the Kentucky populations, in which seeds were pooled over 18 individuals per population, only seeds originating from one individual were planted in each pot.
To accurately assess germination timing and to compare the performance of plants that germinated in the autumn to those that germinated in the spring, 15 pots for each population were randomly assigned to each of two treatments: "autumn" and "spring." In each "autumn" pot, one random autumn germinant was followed as the focal individual, and in each "spring" pot, one random spring germinant was chosen as the focal individual. This gave a total of 720 pots (10 populations x two treatments x 15 replicates per population per treatment). All pots were randomly distributed throughout the plot.
The number of seedlings in each pot was recorded weekly, with 25 censuses from 30 September 2000 until 17 June 2001. No censuses were conducted from 11 December 2000 until 20 February 2001. On average, approximately 80% of all seeds germinated. The focal individual in each pot was chosen as follows. In late October, after most autumn germination, the seedling that was closest to the center of the pot was left in the "autumn" pot as the focal individual while all other seedlings were plucked from the pot. If another germinant emerged later in the autumn and was closer to the center, it was chosen as the replacement focal individual. Few, however, germinated after the thinning. All new germinants were removed from autumn pots after February, as they were then considered to be spring germinants. In spring pots, all seedlings that germinated before February were removed as they germinated, and a focal spring germinant was chosen as the seedling closest to the center of the pot, as described above.
Each pot was covered by an 8-cm high, tent-shaped structure of aluminum window screen to protect the seeds from heavy rains. The effects of aluminum screening on photosynthetically active radiation (PAR) and on the red to far-red ratio (R : FR) were examined on 12 September (before planting) using an Ocean Optics S2000 spectroradiometer (Ocean Optics, Inc., Dunedin, Florida USA) attached to a Dell laptop computer with the Ocean Optics Irrad version 2.00.05 software. The energy flux of light between 400 and 700 nm for light passing through a "tent" had the same spectrum as sunlight, and PAR was reduced by 5% (with tent = 312 mmoles, without tent = 328 mmoles). The ratio of the energy flux for light between 650 and 670 nm ("red" light) to the energy flux between 720 and 740 nm ("far-red" light) was 1.39, close to the value of 1.37 obtained for sunlight at the same time.
Only spring pots were used to estimate germination timing since rosettes in autumn pots could have suppressed germination of remaining seeds. The mean day of germination of all seedlings in a pot was calculated as the number of new seedlings during each census times the census date (days after seed deposition), divided by the total number of germinants. The germination day of the focal individual was estimated as the average germination date until the autumn thinning date, revised by mortality events that established that the focal seedling germinated after those mortality events. The germination date of the focal plant was then further revised by the probability that a new germinant after that thinning date was closer to the center than the first focal plant. Very few pots actually changed the focal plant, however.
On each focal plant, we recorded the following traits: germination day (described above), rosette diameter in mid December (overwintering diameter), the day on which the plant initiated inflorescence development (bolting date), the number of days between bolting and flowering (flowering interval), the rosette diameter at the time of bolting (diameter), the number of secondary plus tertiary branches produced from the central stem (inflorescence branches), the number of basal branches produced from each rosette (basal branches), and the total number of siliques produced (total fruits) as an estimate of total lifetime fitness. If a plant were recorded as having bolted, it was defined as having survived to reproductive maturity, or having "survived to bolting."
Statistical methods
All statistical procedures used the Statistical Analysis System (1990). To test whether regions and populations differed in life-history traits, multivariate analysis of variance (MANOVA) was conducted with germination date, overwinter diameter, diameter at bolting, flowering date (combining bolting date and flowering interval), total branches (combining inflorescence and basal branches), and total fruit production as dependent variables, and region (fixed: KY, MI, or NE) and population (random) nested within region as the independent classes. The effect of region was tested over the population plus error variance. Traits were transformed to normality as necessary. Separate analyses of variance (ANOVAs; Proc GLM) using the same model were subsequently conducted to determine which traits varied significantly among regions and populations within regions. Linear discriminant function analysis was used to determine how well plants were classified into geographic regions.
To test whether pre-reproductive traits influenced survival to bolting, logistic regression was used (Liao, 1994
; SAS Proc Catmod). First, the effect of season of germination was determined. Next, the effects of overwintering size and germination date in the autumn were determined using autumn germinants only (because only they expressed overwintering size). Population was used as a factor in these analyses.
To measure the strength of natural selection on adult life-history characters, Lande and Arnold's (1983
; see also Fairbairn and Reeve, 2001
) phenotypic selection analysis was conducted on autumn germinants using individual phenotypic values. Only four spring germinants expressed adult life-history characters, so they could not be analyzed. Selection differentials, or the strength of total directional selection, were estimated as the covariance between standardized traits (mean = 0, 1 SD) and relative fitness [total fruits/(mean number of fruits of sample)], and standard errors were calculated from a simple regression of each trait on relative fitness. The selection gradient, or strength of direct selection while controlling for correlated selection acting through other characters, was estimated as the partial regression coefficient of standardized characters on relative fitness. Nonlinear and correlational selection was estimated in a multiple regression that included all two-way interactions among variables.
To test whether the strength of directional selection on characters depended on which region or population the plant was from, analysis of covariance was conducted. An interaction between the trait and region (or population) would indicate that selection on that trait varied among plants from different regions (or populations).
To determine the contribution of life-history characters to similarity or differences in performance of the plants from different regions or populations, logistic regression and ANOVA were conducted, with either survival to bolting or fruit production by those that survived as the dependent variables, respectively. One analysis was conducted with no traits as covariates, and another analysis was conducted with the traits as covariates. The effect of region or population and the least-square means (residual means) of regions and populations were compared across models. If the difference in performance among regions or populations depends on the model, then the measured traits can be interpreted to alter the degree of differences in performance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
F = 9.67, df = 12, P = 0.02) and among populations within geographic regions (Wilk's F = 1.74, df = 42, P = 0.003). Individual ANOVAs indicated that germination date in the autumn did not vary across regions, but it did vary among populations within regions (Table 1, Fig. 1a). One population in KY germinated significantly later than the others, and RI germinated earlier than the other NE populations. Surprisingly, the proportion of spring germinants did not differ across regions nor among populations within regions (Fig. 1b). The populations that had later germination in the autumn also had higher proportions of spring germination (Fig. 2), suggesting that spring germination in this sample could be simply the tail end of the distribution of late-autumn germinants. NE populations also tended to bolt in the autumn (Fig. 1c), and this tendency has the potential to alter the seasonal conditions experienced by maturing seeds in a manner that could influence germination (see Discussion).
|
|
|
|
|
|
|
2 = 11.64, P = 0.23, N = 272), no interaction between germination timing and overwinter diameter (2 = 0.56, P = 0.46, N = 272), but a significant effect of germination timing (2 = 11.06, P = 0.001) and overwintering diameter (2 = 28.41, P = 0.0001; marginally significant when the non-significant interaction was included in the model) on survivorship to reproduction. Overwinter mortality was substantial. Of those seedlings that survived until the onset of winter, 35% of them died before they could reproduce in the spring. Twenty-nine percent of the plants that bolted in the autumn died over the winter, and 45% of the plants that did not bolt in the autumn died.
|
|
Significant disruptive selection was detected for inflorescence branch production, such that those plants that made an intermediate number of branches had the lowest fitness. In addition, significant correlational selection was detected. Plants with larger diameters at bolting had higher fitness if they also made more basal branches (Fig. 7). Basal branch production had a larger positive effect on fruit production in larger plants than in smaller plants (coefficient = 0.28, SE = 0.14, P < 0.05). Also, those plants that had early bolting had higher fitness if they also had shorter flowering intervals (coefficient = 0.65, SE = 0.31, P < 0.05; Fig. 8).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Life-history variation
Life-history characters varied among populations, but the geographic structure of the variation, while detectable, was not especially pronounced. Much variation was found among populations within regions, indicating that A. thaliana can express local variation even within a larger geographic pattern of life-history variation. Traits that varied regionally included size and timing of reproduction. The pre-reproductive characters of germination date and overwintering size varied substantially at the local scale.
In contrast to field observations of local populations in situ, we found no significant variation among populations in the proportion of spring germinants. In natural populations in their native location, spring germination has been seen to be between 2% to less than 10% in most KY locations, but it has been documented to be up to 54% in some locations in New England (K. Donohue, unpublished data). Therefore a lower proportion of spring germination was seen in this study than is typically observed in the field in New England.
The results of this study indicate that two different mechanisms may contribute to spring germination and the "spring annual" life history. First, we found a strong positive association between germination timing in the autumn and the proportion of spring germination. The spring germination observed in this study therefore may simply be a continuation of severely postponed autumn germination. Seeds may require a certain number of days of cold to germinate (Baskin and Baskin, 1998
; Nordborg and Bergelson, 1999
), and this cumulative cold requirement may cause some seeds to postpone germination until they experience prohibitive temperatures for germination in the winter. Germination would then occur only after temperatures become permissive again in the spring.
Interestingly, we found significant variation among populations in their timing of germination in the autumn even though we did not detect variation in the proportion of spring germination. This indicates that populations may differ genetically in their cold requirement for germination. This interpretation is supported by subsequent studies of germination of seeds matured in a common environment in which the populations differed significantly in the proportion of seeds that germinated without cold stratification (up to 72% difference; F = 11.80, P < 0.0001; C. Griffith, C. Polisetty, K. Donohue, unpublished results). These results suggest that if populations were dispersed into more northerly latitudes in which prohibitive winter temperatures arrived earlier, populations may exhibit higher proportions of spring germination, and variation in the proportion of spring germination may then vary genetically among populations. Counterintuitively, it may actually be those populations that have responded to selection for delayed autumn germination (due perhaps to protracted summer drought conditions associated with warmer climates) that would exhibit a greater proportion of spring germination if they happened to be dispersed to more northerly latitudes. Rather than being an adaptive mechanism to escape overwinter mortality during conditions of harsh winters, spring germination may be simply a correlated response to selection for delayed autumn germination.
The second possible mechanism for spring germination includes maternal effects on germination. The season of seed maturation and dispersal determines the seasonal cues that seeds are exposed to and can influence the morphology and behavior of the seeds (Junttila, 1971
; Gutterman, 1978
, 1994
; Pourrat and Jacques, 1978
; Lacey and Pace, 1983
, Lacey, 1991
, 1996
). As mentioned previously, spring germination in New England is more prevalent in the field than was observed in this study conducted in KY. In this study, only seeds collected in the spring were used. However, New England populations have also been observed to flower and set seeds in autumn (K. Donohue and L. Dorn, personal observation). Even when grown in KY, the New England populations exhibited a pronounced tendency to initiate reproduction in the autumn. This habit is genetically based to some degree, because subsequent studies, conducted under growth chamber conditions and controlling for maternal effects, detected significant differences among these populations in the proportion of plants flowering with no rosette vernalization (up to 75% difference), or the proportion of plants flowering with only one month of rosette vernalization (up to 64% difference; 2 = 91.68, P < 0.001). Therefore, New England populations are more likely to mature seeds in the autumn. A previous study of germination in A. thaliana conducted on recombinant inbred lines under controlled conditions (Munir et al., 2001
) demonstrated that seeds matured under autumn-like seasonal conditions—namely seeds matured under short days which subsequently receive cold stratification as they would during winter—germinate to high percentages. Thus, autumn seasonal conditions of seed maturation can promote spring germination, even without the typical summer warm afterripening period required by many winter annuals (Baskin and Baskin, 1972
, 1974
, 1983
, 1990
). The high proportion of spring germination in New England may therefore not be caused by genetic differences in germination requirements per se, but rather by genetic differences in the probability of experiencing the seasonal maternal effects that induce spring germination, through genetic variation in flowering time.
Spring germination and the "spring annual" life history can therefore be caused by (a) a cumulative cold requirement for germination, followed by the early onset of prohibitive winter temperatures that results in the postponement of germination until permissive temperatures are experienced in the spring, or (b) seasonal maternal effects accompanying autumn seed maturation that promote germination in the spring. The degree to which these two mechanisms account for variation in spring germination observed in the field requires further study of populations grown under New England seasonal conditions. Because of the importance of germination timing to subsequent life-history expression (Donohue, 2002
), such studies of mechanisms of seasonal dormancy would contribute to our understanding of life-history variation expressed at all life stages.
In summary, the life-history variation observed among A. thaliana populations in situ cannot be explained entirely by genetic differences among regions or populations in life history characters. Rather, genetically based differences interact with local seasonal conditions to determine the life history that is expressed in the field.
Natural selection on life-history characters
Several life-history traits that we measured influenced survival or fruit production. Plants that germinated earlier in the autumn and overwintered at a larger size had a higher probability of surviving to initiate reproduction. Postponing germination until spring caused an extreme reduction in the probability of survival, indicating that the spring annual life history that is frequently observed in New England was not adaptive in KY, at least during that year, despite overwinter mortality (see also Donohue, 2002
). We found evidence that, of those that did survive the winter, later germination in the autumn was associated with somewhat higher fruit production. Therefore, selection on germination timing might act differently at different life stages.
Earlier bolting and bolting at a larger size was associated with the highest fruit production, as has frequently been observed in A. thaliana (Clauss and Aarssen, 1994
; Mitchell-Olds, 1996
; Dorn et al., 2000
). Early bolting permits fruit maturation before the summer drought, while bolting at a larger size could insure that more resources are available for reproduction. Total selection was stronger than direct selection on both characters, indicating that selection on correlated characters reinforced direct selection on these characters. In particular, selection favoring increased branch production would reinforce selection favoring larger diameters, because the two were positively correlated (r = 0.68, P < 0.05 for inflorescence branches; r = 0.87, P < 0.01 for basal branches).
Selection on bolting time depended on the interval between bolting and flowering, as indicated by significant correlational selection between the two traits. The positive correlational selection between bolting time and flowering interval is most likely because most of the variance in the relationship between bolting and flowering interval within this sample was due to those plants, primarily from New England, that bolted in the autumn but did not flower until spring. Earlier bolting was favored most strongly in New England populations, a result which could be interpreted as selection favoring autumn bolting. Autumn bolters did in fact produce more fruits than spring bolters on average (213 ± 228 for autumn bolters vs. 187 ± 220 for spring bolters). The pattern of correlational selection suggests that those autumn bolters that flowered the earliest in spring had higher fitness. Indeed, delayed flowering would not give any reproductive advantage to those plants that bolted early since allocation to vegetative meristems stops at the time of bolting.
Plants that produced many branches also produced more fruits, although those that produced the fewest inflorescence branches did not have the lowest fitness, as indicated by significant disruptive selection. The smallest plants therefore appeared to set more fruits if they did not allocate resources to branches. No tradeoff was observed between inflorescence branch and basal branch production (r = 0.49, P > 0.05), so larger plants overall produced more fruits. Increased basal branch production had particularly strong positive effects on plants that were larger at the time of bolting, and these plants were able to make the largest number of basal branches. Smaller bolters did not gain as much by making more basal branches, in part because they were able to make fewer of them.
The strength of natural selection sometimes depended on the population or region the plant was from. In particular, selection on inflorescence branches was strongest in MI populations, and selection on basal branches was strongest in KY populations. The differences in the strength of selection among populations can be attributed in part to differences in the opportunity for selection caused by higher phenotypic variance in some regional samples than in others. The strong selection for basal branch production in KY populations appears to be attributable to the fact that they inhabit a region on the adaptive landscape in which increased branch production has a strong effect on fitness (that is, they have larger diameters), and they have a large variance in branch number. The marginally significant difference among regions in the direction of selection on flowering interval is due to the fact that NE populations bolted in the autumn, and autumn bolting tended to induce selection for shorter flowering intervals (as indicated by the pattern of correlational selection). The other populations flowered in the spring, and selection on flowering interval was negligible (in KY) or was in the opposite direction (in MI). Thus when bolting differs among individuals from different regions, selection on flowering interval can differ among individuals from different regions as a consequence.
Local adaptation
If the KY populations had been exposed to and responded to past natural selection that was specific to conditions in KY, then one would expect that plants from the KY populations would perform better in an experiment conducted in KY. However, we found no significant variation among populations in either survival or fecundity. All populations performed variably but equally well when grown in a common environment in KY. This suggests either that all populations have been exposed to similar selective influences across a broad geographic range, or that no population, including those from KY, is locally adapted to KY conditions due to lack of adequate response to natural selection after introduction.
The similarity in performance of all populations may reflect a similarity in the past natural selection experienced by the ancestors of all the populations. Selection may act in similar directions in different populations because the habitat of A. thaliana may be fairly similar over a wide geographic area. Arabidopsis thaliana frequently occurs in open areas such as fields, yards, roadsides, or agricultural field margins (Fernald, 1950
; Radford et al., 1968
; Strausbaugh and Core, 1977
; Correll and Johnson, 1979
; Gleason and Cronquist, 1991
). Such open areas could experience similar microclimates, regardless of their geographic locations. On a broader geographical scale, in contrast, the locations of the seed sources are under different climatic regimes, as defined by the Koppen–Geiger climatic classification. Kentucky and Rhode Island are described as mild, moist and warm climates, the Michigan locations as severe, moist and warm, and the Massachusetts locations as severe, moist, and cold (maps in Howarth et al., 1998
). The difference in climate may not be as important to the performance of A. thaliana as the similarity in microenvironment experienced by all populations.
The lack of evidence for local adaptation may also be because even the KY populations are not especially adapted to KY conditions. That is, instead of all populations displaying a comparable degree of adaptation to KY, all populations could be displaying a comparable degree of maladaptation to KY. The lack of stabilizing selection indicates that no population is optimally adapted. Populations that had favorable phenotypes for some characters had unfavorable ones for others. For example, MI populations were the largest at the time of bolting, which was favorable, but they bolted later, which was unfavorable. The opposite was true of NE populations. Stabilizing selection is much less frequently documented than directional selection in most systems (Kingsolver et al., 2001
), and this result, when statistically substantiated, is often interpreted as evidence for genetic constraints on achieving optimal phenotypes (Lande and Arnold, 1983
; Arnold and Wade, 1984
). The well-known tradeoff between size and timing of reproduction in A. thaliana (Mitchell-Olds, 1996
) does appear to have constrained performance in these plants. In historically mobile species such as A. thaliana, however, the lack of adaptation is also likely to be due to the non-equilibrium status of many introduced populations that simply haven't had time to adapt to new conditions.
We still do not know whether the observed regional and population differentiation of life history traits represents historical processes of relatedness and colonization patterns or adaptation to local or regional conditions. For example, the lack of autumn flowering plants in KY and MI populations could be because autumn-flowering genotypes never dispersed into those regions or because they have been selected against in those regions. While our results did not find evidence of stronger selection against life-history strategies exhibited by foreign populations, this experiment measured selection in only one year and under one set of experimental conditions. Longer-term selection studies under different experimental conditions would provide more information on the quality of natural selection on life history traits that vary among populations. In addition, genetic information on the structure of relatedness among populations would provide valuable information on the importance of historical processes in creating geographic patterns of life histories.
Even if the lack of evidence for local adaptation is due to purely historical factors in populations not in adaptive equilibrium, we nevertheless found evidence that life-history differences among populations do in fact influence the relative performance of populations. All populations performed comparably despite natural selection on life histories and despite population differentiation in life-history traits. This result and the significant correlational selection imply the fitness of a plant depends on the particular combinations of characters that the plant expresses. While variable life histories can be under natural selection, different combinations of life-history and other characters can lead to similar fitness of different populations.
In fact, we found that differences in life-history expression actually contributed to the similarity of performance of plants from different regions. That is, life-history characters compensated for deficiencies in other characters that were not included in the study. Variation among regions in the timing of bolting, size at bolting, and basal branch production all contributed to similarity of performance. For example, the ability of plants from NE to bolt in the autumn (which was associated with higher fitness) was able to compensate for deficiencies in other traits that would have decreased their fitness. One such trait may be slower growth rates in the spring. NE plants that did not bolt in the autumn were smaller at bolting than the other populations (24 cm compared to 42 cm for MI plants and 35 cm for KY plants). Because larger plants had higher fitness, NE spring bolters had lower fitness than the other plants, but NE autumn bolters had higher fitness. Thus bolting in the autumn could have compensated for slower growth rates in the spring, although spring growth rate was not measured in this study. Similarly, MI plants had the largest diameters and produced the most basal branches (7.0 compared to 4.7 for KY and 3.6 for NE), yet they did not have higher fitness than the others. MI plants therefore appear to have some unmeasured traits that were disadvantageous in KY, and their size advantage compensated for those deficiencies. These observations indicate that, while all A. thaliana populations performed similarly, they did so because they expressed different combinations of characters. These combinations of characters can act in an additive manner to preserve fitness, as is the case of autumn bolting apparently compensating for slower growth rates, or in a nonadditive manner, as in the case of the observed correlational selection. In particular, basic life-history traits appear to compensate for deficiencies in other traits.
Character combinations therefore appear to be important determinants of performance in novel locations within the range of this highly mobile species. Arabidopsis thaliana is highly selfing (Abbott and Gomes, 1989
). Rare outcrossing events followed by many generations of selfing would effectively create suites of recombinant inbred lines that contain many different combinations of traits from the hybridizing parental lines. The results from this study suggest that such patterns in mating system result in populations with different viable combinations of characters that are comparably adapted to environments that are distributed over a broad geographic range. Such a mating system pattern may be a particularly effective manner of adaptation to broad geographic ranges while maintaining genetic variation among populations.
Conclusion
We found evidence for genetic differences among populations in fundamental life-history traits, but life-history expression in situ is likely to be mediated by local environmental conditions as well. Despite differences in life history, and despite selection on life-history traits, we found no significant differences in the performance of different populations from many locations in a common environment in KY. The lack of local adaptation and stabilizing selection suggests that these populations may not be in adaptive equilibrium, a result that accords with the suspected historical mobility of the populations. Different populations expressed different combinations of life-history characters, and life-history traits apparently compensated for deficiencies in other traits. Thus different combinations of characters caused plants from different populations to perform equally well. In highly inbreeding and mobile species such as A. thaliana, variation in life-history expression may contribute to its ability to expand into novel locations while maintaining genetic variation at local and regional scales.
|
FOOTNOTES |
|---|
2 Current address: National Forests and Grasslands of Texas, 701 North First Street, Lufkin, TX 75901 USA
3 Current address: Department of Organismic and Evolutionary Biology, Harvard University, 22 Divinity Ave. Cambridge, MA 02138 USA
4 Reprint requests: kdonohue{at}oeb.harvard.edu
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Antonovics J. 1968 Evolution in closely adjacent populations VI. Manifold effects of gene flow. Heredity 23: 507-524[Web of Science]
Arnold S. J. M. J. Wade 1984 On the measurement of natural and sexual selection: theory. Evolution 38: 709-719[CrossRef][Web of Science]
Baker H. G. 1974 The evolution of weeds. Annual Review of Ecology and Systematics 5: 1-24
Baskin C. C. J. M. Baskin 1998 Seeds: ecology, biogeography and evolution of dormancy and germination. Academic Press, San Diego, California, USA
Baskin J. M. C. C. Baskin 1972 Ecological life cycle and physiological ecology of seed germination of Arabidopsis thaliana. Canadian Journal of Botany 50: 353-360
Baskin J. M. C. C. Baskin 1974 Germination and survival in a population of the winter annual Alyssum alyssoides. Canadian Journal of Botany 52: 2439-2445[Web of Science]
Baskin J. M. C. C. Baskin 1983 Seasonal changes in the germination responses of buried seeds of Arabidopsis thaliana and ecological interpretation. Botanical Gazette 144: 540-543[CrossRef]
Baskin J. M. C. C. Baskin 1990 Seed germination biology of the narrowly endemic species Lesquerella stonensis (Brassicaceae). Plant Species Biology 5: 205-214
Bergelson J. E. Stahl S. Dudek M. Kreitman 1998 Genetic delete variation within and among populations of Arabidopsis thaliana. Genetics 148: 1311-1323
Clauss M. J. L. W. Aarssen 1994 Phenotypic plasticity of size-fecundity relationships in Arabidopsis thaliana. Journal of Ecology 82: 63-68
Correll D. S. M. C. Johnston 1979 Manual of the vascular plants of Texas, 2nd ed. University of Texas at Dallas, Richardson, Texas, USA
Crone E. E. 2001 Is survivorship a better fitness surrogate than fecundity?. Evolution 55: 2611-2614[CrossRef][Web of Science][Medline]
Donohue K. 2002 Germination timing influences natural selection on life history characters in Arabidopsis thaliana. Ecology 83: 1006-1016[CrossRef][Web of Science]
Dorn L. E. Hammond-Pyle J. Schmitt 2000 Plasticity to light cues and resources in Arabidopsis thaliana: testing for adaptive value and costs. Evolution 54: 1982-1994[CrossRef][Web of Science][Medline]
Effmertova E. 1967 The behaviour of "summer annual", "mixed", and "winter annual" natural populations as compared with early and late races in field conditions. Arabidopsis Information Service 4
Evans J. D. Ratcliffe 1972 Variation in "after-ripening" of seeds of Arabidopsis thaliana and its ecological significance.. Arabidopsis Information Service 9: 3-5
Fairbairn D. J. J. P. Reeve 2001 Natural selection. In Charles W. Fox, Derek A. Roff, and Daphne J. Fairbairn [eds.], Evolutionary ecology, 29–43. Oxford University Press, New York, New York, USA
Fernald M. L. 1950 Gray's manual of botany, 8th ed. American Book Company, New York, New York, USA
Gleason H. A. A. Cronquist 1991 Manual of vascular plants of the northeastern United States and adjacent Canada, 2nd ed. The New York Botanical Garden, Bronx, New York, USA
Griffing B. R. L. Scholl 1991 Qualitative and quantitative genetic studies of Arabidopsis thaliana. Genetics 129: 605-609[Web of Science][Medline]
Gutterman Y. 1978 Seed coat permeability as a function of photoperiodical treatments of the mother plants during seed maturation in the desert annual plant: Trigonella arabica, del. Journal of Arid Environments 1: 141-144
Gutterman Y. 1994 Effect of day length during plant development and caryopsis maturation on flowering and germination, in addition to temperature during dry storage and light during wetting, of Schismus arabicus (Poaceae) in the Negev desert, Israel. Journal of Arid Environments 33: 439-448[CrossRef]
Hardy M. 1993 Regression with dummy variables. Sage University Paper series on Quantitative Applications in the Social Sciences, 07–093. Thousand Oaks, Newbury Park, California, USA
Hoffman M. H. 2002 Biogeography of Arabidopsis thaliana (L.) Heynh. (Brassicaceae). Journal of Biogeography 21: 125-134
Howarth D. A. D. G. Conner C. A. Leuthart L. M. L. Trapasso R. Ulack J. F. Watkins 1998 The natural environment. In R. Ulack, K. Raitz, and G. Pauer [eds.], Atlas of Kentucky, 15–46. The University Press of Kentucky, Lexington, Kentucky, USA
Jain S. K. A. D. Bradshaw 1966 Evolutionary divergence among adjacent plant populations. I. Evidence and its theoretical analysis. Heredity 23: 407-441
Junttila O. 1971 Effect of mother plant temperature on seed development and germination in Syringa reflexa Schneid. Meld. Norges Landbruk. Shogsk 50: 1-16
Karssen C. M. 1970 The light promoted germination of the seeds of Chenopodium album L. III. Effect of the photoperiod during growth and development of the plants on the dormancy of the produced seeds. Acta Botanica Neerlandica 19: 81-94
Kingsolver J. G. H. E. Hoekstra J. M. Hoekstra D. Berrigan S. N. Vignieri C. E. Hill A. Hoang P. Gibert P. Beerli 2001 The strength of phenotypic selection in natural populations. The American Naturalist 157: 245-261[CrossRef][Web of Science]
Kover P. B. Schaal 2002 Genetic variation for disease resistance among Arabidopsis thaliana ecotypes. Proceedings of the National Academy of Sciences, USA 99: 11270-11274
Lacey E. P. R. Pace 1983 Effect of parental flowering and dispersal times on offspring fate in Daucus carota (Apiaceae). Oecologia 60: 274-278[CrossRef][Web of Science]
Lacey E. P. 1991 Parental effects on life history traits in plants,. In E. C. Dudley [ed.], The unity of evolutionary biology, 735–744. Dioscorides Press, Portland, Oregon, USA
Lacey E. P. 1996 Parental effects in Plantago lanceolata L. I. A growth chamber experiment to examine pre- and postzygotic temperature effects. Evolution 50: 865-878[CrossRef][Web of Science]
Lande R. L. S. J. Arnold 1983 The measurement of selection on correlated characters. Evolution 37: 1210-1226[CrossRef][Web of Science]
Liao T. F. 1994 Interpreting probability models: Logit, Probit, and other generalized linear models. Sage University Paper series on Quantitative Applications in the Social Sciences, 07-101. Thousand Oaks, Newbury Park, California, USA
Macnair M. R. 1981 Tolerance of higher plants to toxic materials. In J. A. Bishop and L. M. Cook [eds.], Genetic consequences of man-made change, 177–207. Academic Press, New York, New York, USA
Mitchell-Olds T. 1996 Genetic constraints on life-history evolution: quantitative trait loci influecing growth and flowering in Arabidopsis thaliana. Evolution 50: 140-145[CrossRef][Web of Science]
Mitchell-Olds T. 2001 Arabidopsis thaliana and its wild relatives: a model system for ecology and evolution. Trends in Ecology and Evolution 16: 693-699[CrossRef]
Munir J. L. Dorn K. Donohue J. Schmitt 2001 The influence of maternal photoperiod on germination requirements in Arabidopsis thaliana. American Journal of Botany 88: 1240-1249
Napp-Zinn K. 1976 Population genetical and geographical aspects germination and flowering in Arabidopsis thaliana. Arabidopsis Information Service 13
Nordborg M. J. O. Borevitz J. Bergelson C. C. Berry J. Chory J. Hagenblad M. Kreitman J. N. Maloof T. Noyes P. J. Oefner E. A. Stahl D. Weigel 2002 The extent of linkage disequilibrium in Arabidopsis thaliana. Nature Genetics 30: 190-193[CrossRef][Web of Science][Medline]
Nordborg M. J. Bergelson 1999 The effect of seed and rosette cold treatment on germination and flowering time in some Arabidopsis thaliana (Brassicaceae) ecotypes. American Journal of Botany 86: 470-475
Pigliucci M. 1998 Ecological and evolutionary genetics of Arabidopsis. Trends in Plant Science 3: 485-489[CrossRef][Web of Science]
Pigliucci M. E. T. Marlow 2001 Differentiation of flowering time and phenotypic integration in Arabidopsis thaliana in response to season length and vernalization. Oecologia 127: 501-508[CrossRef][Web of Science]
Pigliucci M. C. D. Schlichting 1996 Reaction norms of Arabidopsis thaliana (Brassicaceae). III. Response to nutrients in 26 populations from a worldwide collection. American Journal of Botany 82: 1117-1125[CrossRef][Web of Science]
Pourrat Y. R. Jacques 1978 The influence of photoperiodic conditions received by the mother plant on morphological and physiological characteristics of Chenopodium polyspermum L. seeds. Plant Science Letters 4: 273-279
Radford A. E. H. E. Ahles C. R. Bell 1968 Manual of the vascular flora of the Carolinas. University of North Carolina Press, Chapel Hill, North Carolina, USA
Ratcliffe D. 1965 Germination characteristics and their inter- and intra-population variability in Arabidopsis. Arabidopsis Information Service 13
Ratcliffe D. 1965 Germination characteristics and their inter- and intra-population variability in Arabidopsis. Arabidopsis Information Service 13
SAS. 1990 SAS/STAT User's Guide. SAS Institute, Cary, North Carolina, USA
Sharbel T. F. B. Haubold T. Mitchell-Olds 2000 Genetic isolation by distance in Arabidopsis thaliana: biogeography and postglacial colonization of Europe. Molecular Ecology 9: 2109-2118[CrossRef][Medline]
Strausbaugh P. D. E. L. Core 1977 Flora of West Viriginia, 2nd ed. Seneca Book, Morgantown, West Virginia, USA
Thompson L. 1994 The spatiotemporal effects of nitrgen and litter on the population dynamics of Arabidopsis thaliana. Journal of Ecology 82: 63-68
Tonsor S. J. C. J. Goodnight 1997 Evolutionary predictability in natural populations: do mating system and nonadditive genetic variance interact to affect heritabilities in Plantago lanceolata?. Evolution 51: 1773-1784[CrossRef][Web of Science]
Vander Zwan C. S. A. Brodie J. J. Campanella 2000 The intraspecific phylogenetics of Arabidopsis thaliana in worldwide populations. Systematic Botany 25: 47-59
Wade M. J. 2002 A gene's eye view of epistasis, selection, and speciation. Journal of Evolutionary Biology 15: 337-346[CrossRef][Web of Science]
Wright S. 1932 The roles of mutation, inbreeding, crossbreeding and selection in evolution. Proceedings of the Sixth International Congress of Genetics 1: 356-66
This article has been cited by other articles:
|
|
 |
|
  P. H. Leinonen, S. Sandring, B. Quilot, M. J. Clauss, T. Mitchell-Olds, J. Agren, and O. Savolainen Local adaptation in European populations of Arabidopsis lyrata (Brassicaceae) Am. J. Botany, June 1, 2009; 96(6): 1129 - 1137. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Donohue Completing the cycle: maternal effects as the missing link in plant life histories Phil Trans R Soc B, April 27, 2009; 364(1520): 1059 - 1074. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. J. Wender, C. R. Polisetty, and K. Donohue Density-dependent processes influencing the evolutionary dynamics of dispersal: a functional analysis of seed dispersal in Arabidopsis thaliana (Brassicaceae) Am. J. Botany, June 1, 2005; 92(6): 960 - 971. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
