Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An examination of the evolutionary basis of range limitation Slooten, Greta 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2003-0157.pdf [ 3.35MB ]
JSON: 831-1.0090986.json
JSON-LD: 831-1.0090986-ld.json
RDF/XML (Pretty): 831-1.0090986-rdf.xml
RDF/JSON: 831-1.0090986-rdf.json
Turtle: 831-1.0090986-turtle.txt
N-Triples: 831-1.0090986-rdf-ntriples.txt
Original Record: 831-1.0090986-source.json
Full Text

Full Text

1 An Examination of the Evolutionary Basis of Range Limitation By Greta Slooten B.Sc, The University of California, Davis, 2000 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE In THE F A C U L T Y OF G R A D U A T E STUDIES DEPT OF ZOOLOGY; EVOLUTIONARY BIOLOGY P R O G R A M M E We accept this thesis as conforming to the Required standard THE UNIVERSITY OF BRITISH C O L U M B I A M A R 2003 © Greta Slooten UBC Rare Books and Special Collections - Thesis Authorisation Form Page 1 of 1 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r a n a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e a n d s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d b y t h e h e a d o f my d e p a r t m e n t o r b y h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . The U n i v e r s i t y o f B r i t i s h C o l u m b i a V a n c o u v e r , C a n a d a 13/03/2003 Abstract The thesis objective - to test the importance of gene flow in regulating species range expansion - is accomplished through two studies. The first compares the predicted and actual density distributions of two species, Cakile maritima, an outcrosser, and C. edentula, an inbreeder. Both are coastal, succulent, annual Brassicaceae whose distributions cover the Pacific coast of North America. The second study examines variation in seed oil content along this latitudinal gradient. Lipids are the primary reserve material in most seeds and are composed of triacylglycerides containing various proportions of saturated and unsaturated fatty acids. Previous work has documented variation in seed oil responding to germination temperature. These changes are presumed to represent a response to selection for optimal germination at different temperatures. It has been shown that plants occurring at higher latitudes have higher percentages of unsaturated fatty acids allowing seed oils with lower melting points to reach germination temperatures faster. Plants at lower latitudes have lower percentages of unsaturated fatty acids as they provide less energy to the plant embryo (Linder, 2000). The theory above predicts that the genes of the lower latitude plants, will overwhelm the adaptive response of border populations to the cold climate in the north, forcing these populations away from their optimum; thus, limiting adaptation and causing range limitatioa However, results provide no evidence of an effect of gene flow on density distribution or seed oil adaptation at the periphery in either species. Many other factors are known to affect adaptation at the border of these species ranges and further work investigating the effect of gene flow in natural populations is needed to clarify the contribution, if any, of gene flow on range limitation. iii Table of Contents Abstract ii Table of Contents iii List of Tables iv List of Figures v Acknowledgements viii Chapter I Review of Population Differentiation and Range Limitation by Gene Flow p i 1.1 Introduction p i 1.2 Historical review p 2 1.3 Genetic variation in outcrossing populations p 5 1.4 Ecological evidence of population differentiation p 10 1.5 Evolutionary theories of range limitation p 12 1.6 Conclusion p 14 Chapter II A Latitudinal Transect of the Pacific Coast Range of Two Closely Related Congeners, Cakile edentula and C. maritima p 15 2.1 Species suitability p 23 2.2 Materials and methods p 24 2.3 Results p 26 2.4 Discussion p 35 Chapter III Variation in Seed Oil Composition of Two Species of Cakile, C. maritima, an Outcrosser, and C edentula, an Inbreeder, Along a 1400 km Latitudinal Gradient p 44 3.1 Methods p 50 3.2 Results p 53 3.3 Generalized linear model p 54 3.4 Discussion p 58 Conclusion p 67 Literature Cited p 68 Appendix I Appendix II V List of Tables Table 1: Summary of morphological and allozymic studies on central and marginal populations of plant species p 7 Table 2: Parameter estimates and significance of the loaded model p 55 Table 3: Parameter estimates and significance of the species/ latitude model p 56 Table 4: Trend in the proportion of saturated fat in C. maritima between two locations, Sweden and Israel p 66 v i List of Figures Figure 1 A : Pacific coast average summer temperature from 1959 to 2000 p 19 Figure IB: Pacific coast average rainfall from 1959 to 2000 p 19 Figure 2: Annual degree days (base temperature 55F) at site latitudes p 20 Figure 3: Picture of C. maritima and C. edentula seedlings photographed in Oregon in July 2000 p22 Figure 4: Picture of C. maritima and C. edentula taken at Devil's Punch Bowl Beach (Beach 3) in Newport, Ore. (Site 7) p 23 Figure 5: Schematic of the sampling procedure along the pacific coast p 26 Figure 6: Masset Cemetery on Queen Charlotte Island p30 Figure 7: Average C. maritima plants per square meter vs. latitude p 32 Figure 8: Average C. edentula plants per square meter vs. latitude p 33 Figure 9: Average C. maritima seeds per plant vs. latitude p 34 Figure 10: Average C. edentula seeds per plant vs. latitude p 35 Figure 11: One-way A N O V A depicting the mean and confidence interval of the proximal and distal samples of C. edentula and C. maritima p 53 Vll Figure 12A: Changes in the proportion of saturated fat over latitude in two related species of Cakile, C. edentula an inbreeder p57 Figure 12B: Changes in the proportion of saturated fat over latitude in two related species of Cakile, C. maritima an outcrosser p57 Figure 13: One-way ANOVA comparing the mean proportion of saturated fat in C. maritima and C. edentula p 58 vm Acknowledgements I acknowledge that my co-supervisors, Dr. Jeannette Whitton and Dr. Michael Whitlock, were the major contributors to the conception and design of this study. It is not without their considerable effort and support that this study is concluded. I thank them both for their time, patience, and understanding. I thank my committee member, Dr. Sally Aitken, whose valuable comments and helpful suggestions guided me along they way. I would like to thank those persons and institutions with whose assistance procured both field and laboratory data: Dr. Ljerka Kunst and Dr. Mark Smith of the University of British Columbia for their help with seed oil assays. Dr. Michael L. Wells of the State of California Dept. of Parks and Recreation and the British Columbia Province and Parks Dept. for their assistance and recognition in the importance of ecological studies. For their assistance in data collection, I would like to thank Dave Henderson, Malcolm Wyeth, Richard Phillips, Dirk, Sally, Peter and Paul Slooten. Finally, I would like to express my deepest admiration to Andrew M. Wyeth, whose support throughout the course of this thesis has been an immeasurable source of strength. 1 Chapter 1: Review of Population Differentiation and Range Limitation by Gene Flow Marginal populations play a significant role in the study of natural selection. Observations on the behaviour of a species at the limit of its geographic range have led to many theories regarding the importance of the species border. In the thesis that follows, the historical development of these theories will be presented and discussed in relation to the critical theory of species range limitation through gene flow. Introduction In 1956, Theodosius Dobzhansky observed that the genetic composition of local populations shows remarkable versatility of adaptive evolution. Gene frequencies can undergo surprisingly rapid changes due to strong environmental influences (Dobzhansky 1956). These changes are a main component in the rise and spread of major taxonomic groups; as discussed below, evolution was viewed as a central-peripheral succession of taxa (Brown 1957). Speciation was thought to occur through an expansion and contraction of the species range originating from the evolutionary centre, the largest continuous favourable area of a species range, which serves as the chief wellspring of variation. This theory of centrifugal speciation maintains that the density difference between the large central and small peripheral populations will periodically be severe enough to cause an 'emigration effect' where the range of the species expands and the important genetic changes in the central population are carried outwards (Brown 1957). 2 It is from this infusion of novel genes and the subsequent isolation in the periphery that new species were thought to be born (Brown 1957). Thus, the mutable environment of the species border allows a closer look at the process of selection. The populations at the species border provide a unique opportunity to study the limits of natural selection and the limits of range expansion. The following review and study is designed to assess the historical and current ideas on this issue and contribute to the knowledge of marginal populations. 1.2 Historical Review "The essential stability of the species border ... would seem to contradict our belief in the powers of natural selection. One would expect a few individuals to survive in a zone immediately outside the species border and to form a new local population that would gradually become better adapted under the continuous shaping influences of local selection. One would expect the species range to grow by a process of annual accretion, like the rings of a tree. That this does not happen is particularly astonishing in the frequent cases where conditions beyond the border differ only slightly from conditions inside the border and where no drastic barriers prevent expansion." (Mayr 1963) Early genetic research focused on the heritable properties of marginal populations and found the frequency of chromosome inversion polymorphism to be variable between central and marginal populations of Drosophila species. Geographically isolated or marginal populations have been shown to possess fewer chromosomal inversions when compared to the central or continental populations of the same (da Cunha et al. 1950, da Cunha and Dobzhansky 1954, Dobzhansky 1957). These studies noted significant 3 seasonality in the incidence of specific forms of chromosomal inversions in California populations of D. pseudoobscura. However, populations exposed to harsher climates and sharply differentiated seasons show no such seasonality (Dobzhansky 1956). This discovery prompted an exploration into inversion polymorphism of temperate and tropical populations of Brazilian DrosophUa wUlistoni (da Cunha et al. 1950). It was found that populations of D. wUlistoni are geographically differentiated with respect to the relative frequencies of specific inversions. Moreover, the frequencies of heterozygous individuals vary greatly across the species range. "The highest frequencies of inversion heterozygosis occur in the central portion of the geographic distribution of the species and decline from this central region" (da Cunha et al. 1950). This chromosomal inversion cline was confirmed in other North and South American populations of D. wUlistoni (Townsend 1952, da Cunha and Dobzhansky 1954, Dobzhansky 1957, da Cunha et al. 1959). Similar clines were detected in D. subobscura from Barcelona (Prevosti 1964) and Scotland (Knight 1961), D. pseudoobscura from Columbia (Dobzhansky et al. 1963), and D. robusta in south-eastern North America (Carson 1956). This significant difference between marginal and central populations became a focal point for later studies on geographic differentiation of populations and is salient today. Several theories were advanced to explain this trend of increased inversion heterozygosity in central environments. The most widespread hypothesis maintained that "the amount of adaptive polymorphism carried in a population tends to be proportional to 4 the variety of habitats (ecological niches) which the population exploits in a territory it occupies" (da Cunha et al. 1950). Inherent in this statement is the assumption that the central portions of a species range comprise a greater diversity of habitats. This theory is challenged by a study of Trimeroptropis sparsa, a grasshopper found in western North America (White 1951). Evidence of population differentiation was found; specifically, chromosomal fusions were prevalent at the periphery and supernumerary chromosomes in the central populations. Thus, there is more chromosomal diversity in the centre of the species range yet there is no evidence that the variety of habitats available to T. sparsa decreases at the periphery. The border of the grasshoppers' range is characterised by valley habitats where there is environmental variability from the valley floor to the dividing hilltops whereas the centre of the range is characterised by a flat, uniform environment. Population differentiation does not coincide with environmental variability, refuting the theory that chromosomal diversity is directly related to ecological diversity (White 1951). Carson (1958), in view of this evidence, theorised that structural polymorphisms do not relate directly to a niche or specific ecological adaptation, but convey general vigour or "heterotic buffering" to the heterozygous individuals. Therefore, high levels of variability are to be found in the centre because they result in better performance overall. The marginal populations are too small, however, to benefit from a system of adaptation that produces a range of possibly less well adapted individuals. Homozygosity is 5 favoured based on adaptations to specific fixed characters in the high stress marginal environment (Brussard 1984). 1.3 Genetic variation in outcrossing populations Electrophoretic methods of estimating the difference in heterozygosity at allozyme loci between central and marginal populations offer another method of investigation into population differentiation. Populations of D. pseudoobscura from North America were among the first to be evaluated in this way. Prakash et al. (1969) compared a central and a marginal North American Drosophila population and concluded they show no difference in allozyme variability. The polymorphisms found in populations across North America are similar in their allelic frequencies. Comparisons of allozyme heterozygosity in both peripheral and central populations of D. obscura in central Europe (Lakovaara and Saura 1971a), D. subobscura in northern Europe (Lakovaara and Saura 1971b, Saura et al. 1973), D. willistoni in Brazil (Ayala et al. 1972b) and North America (Ayala et al. 1972a), and D. robusta in North America (Prakash 1973) all show no remarkable differences. Similarly, an examination of isozyme variation of Picea abies found a random distribution of allelic variants (Tigerstedt 1973). These studies reveal no detectable differences in genetic variation. However, conflicting evidence reports high degrees of population differentiation. Accounts of phenotypic variance with regard to geographical distribution in Lysimachia volenskii illustrate a distinct pattern with respect to environmental stress. "Populations at the centre of the 6 species range being under more nearly optimum conditions... show a higher variance than those near the edge of the range" (Agnew 1968). In other words, populations at the centre of the range exhibit more phenotypic variability. The study ascertains a pattern in phenotypic variability in inflorescence structure similar to that of the chromosomal inversions, with higher diversity in the centre. A study performed on Veronica perigrina using morphological characteristics to estimate genotypic variation also found significant population differentiation (Linhart 1974). In this case it is the marginal populations of this vernal pool annual that exhibit the increase in variability. It is possible, as the author remarks, that intraspecific competition is so intense at the centre that the central environment is more stressful than the arid geographical margins. If this were the case, then this study is consistent with the predictions of da Cunha et al. (1950) and the pattern of marginal-central differentiation is similar to the inversion frequency in Drosophila (Dobzhansky 1957). Subsequent studies reveal an interesting trend accounting for conflicting patterns in geographic differentiation and support the theory of range limitation by gene flow. A study on Hordeum jubatum, a highly self-fertilising, weedy species inhabiting the western North American continent, reveals a distinct pattern of allozymic differentiation (Shumaker 1980). In several marginal populations, allozyme phenotypes are highly conserved and lack significant variation. The central population shows much greater amounts of genetic variation. Similarly, a 1983 study on Rumex acetosella shows "significant amounts of isozyme differentiation between central and marginal 7 populations" (Farris 1983). However, similar research found no evidence of population differentiation in either Phlox drumondii (Levin 1978) or in Ranunculus adoneus (Stanton et al. 1997). Species Method of Reproduction Variation in Heterozygosity? Method of study Citation Lysimachia volkensii Self - fertilising YES. Increased variation at margins Morphological Agnew1968 Veronica perigrina Self - fertilising YES. Increased variation at centre Morphological Linhart 1974 Picea abies Outcrosser No distinct pattern of population variation Allozyme Tigerstedt 1973 Hordeum jubatum Self- fertilising YES. Increased variation at centre Allozyme Shumaker and Babble 1980 Rumex acetosella Vegetative reproduction YES. Increased variation at centre Allozyme Farris and Schaal 1983 Phlox drummondii Outcrosser No distinct pattern of population variation Allozyme Levin 1978 Ranunculus adoneus Outcrosser No distinct pattern of population variation Allozyme Stanton et al. 1997 Leptospermum scoparium Outcrosser No distinct pattern of population variation Morphological Wilson et al. 1991 Spartina patens Clonal reproduction YES. Pattern clinal variation Morphological Silander 1985 Table 1. Summary of morphological and allozymic studies comparing central and marginal populations of a plant species. Note the relationship between reproduction method and the pattern of population differentiation in all studies shown. Although these studies and those of Drosophila seem to contradict each other, one displaying a specific distributional pattern of genetic variation and the other showing more irregularity, they reveal an intriguing pattern that essentially correlates variation and gene flow (Table 1). In each study, the breeding strategy was noted and a consistent 8 relationship was discovered between this and population differentiation. Isozyme differentiation and population variation is apparent in those species where gene flow is restricted, e.g. self-fertilizers. When a species reproduces chiefly by outcrossing, no distinction between the genetic variability in the centre and the margins of the species range is found. Persuasive evidence corroborating the parallel relationship between population structure and reproduction strategy is seen in studies documenting population differentiation in a selfing species and lack thereof in a closely related outcrosser (Levin 1978). Levin (1978) performed an experiment on three species of Phlox, two of which (P. drummondii and P. roemarind) are self-incompatible and one (P.cuspidatd) which is self-compatible. Allozymic variation was measured in populations of all three species covering their entire range. It was concluded that the marginal populations did not suffer from a genetic diversity differential. Moreover, "the selfer shows significantly greater gene frequency heterogeneity than do the outbreeders" (Levin 1978). This conclusion is consistent with two other experiments comparing related inbreeders and outcrossers: Lloyd (1965) contrasted four species of Leavenworthia and found the self-compatible species to exhibit more inter-population differentiation, meaning the selfer has a higher degree of population differentiation. And, in a study performed on Trifolium, the outbreeders were found to have less among population heterogeneity than selfers (Levin 1978). In each study, genetic diversity and its distribution among populations within a species are related to reproduction strategy (Table 1). This trend is further detailed in Hamrick 9 and Godt's (1996) most recent review of plant allozyme diversity. Their analysis of the influence of such traits as breeding system, mechanisms of seed dispersal, and life form on the distribution of genetic diversity within and among species reveals the ascendancy of life history traits on population differentiation. In accordance with the trend reported in Table 1, the effect of breeding system (selfing vs. outcrossing) on the partitioning of genetic diversity within populations is highly significant. Selfing species had a two-fold higher proportion of total genetic diversity due to differences among populations (GST; see Nei (1973) for further explanation) than outcrossing species. The effect of seed dispersal relates a similar trend, indicating that the higher the potential for gene flow, the lower the potential for population differentiation. When isolating the effect of seed dispersal on genetic differentiation among populations, Hamrick and Godt (1996) reported that wind dispersed seeds have significantly lower population differentiation (GST = 0.101) than either gravity or animal dispersed (GST = 0.189 and 0.223 respectively). Moreover, when these two factors are combined, species with outcrossed seeds dispersed by the wind have four times less among population differentiation than selfing species, regardless of the dispersal mechanism. Thus, contrasting the population patterns of an inbreeder and an outcrosser will provide the opportunity to examine the effect of gene flow on the distribution of genetic variation. 1.4 Ecological evidence of population differentiation The genetic approach assumes an inherent contrast between a central and marginal population. In a review of the flora of the Great Smokey Mountains, R.H. Whittaker 10 (1956) reported that almost all species show a rounded or bell-shaped density distribution along a gradient. Whittaker (1956) theorised that at least one environmental optimum for the population may be predicted at the peak density and the further from this optimum, the more the population level declines along a normal curve. Although this work was intended to provide botanical information for its own sake to use in research into the theory of community associations and as a basis for the study of animal ecology, it spawned revolutionary ideas on species range limitation. The progression of observations and ideas concerned with species range limitation is rooted in the work of Whittaker (1956). A.R. Watkinson (1985), taking the observations of Whittaker (1956) and others, formed a theory explaining patterns of population abundance along an environmental gradient. This theory describes a curvilinear relationship between density and persistence of a population due to a non-linear relationship between population size and growth rate. This curvilinear relationship is caused by the interaction of density dependent and independent processes, as it is these forces that determine the average density of a population and ultimately its ability to persist at a particular locale (Watkinson 1985). Thus, it was thought that population distribution limits could arise simply because of small climactic fluctuations and not as a result of major physiological, genetic, and environmental incompatibilities. The normal curve became the accepted pattern of spatial variation in abundance caused by a population's reaction to an ecological gradient. J.H. Brown (1984) created a mathematical model also based on Whittaker's 1956 observations, the multidimensional 11 niche model, taken from Hutchins (1947) definition, to describe the observed density decline towards the species margins. He used several algorithms to express the relationship of a species with both the density dependent and independent processes affecting its population size. By combining these spatially variable algorithms, Brown was able to recreate and predict the expected normal distribution of species population density (Brown 1984). The above pattern of variation implies that the limit of the geographic range occurs where the population density declines to zero. The ecological view is that at the range boundary, the factors affecting survivorship and fecundity prevent the expansion of the species range. For example, Phlox drummondii in Texas only grows on sandy soils; its range is constrained to the south by dense soils with a higher percentage of minerals. The soil type affects the fecundity of the species, and thus range expansion is prevented (Levin and Clay 1984). For an evolutionist, it is impossible to consider these physiological factors as an explanation for the limitation of a species range; rather, they are a result of the boundary and not the cause. The limitation itself is caused by evolutionary constraints: either lack of genetic variation, gene flow from central populations, or negative correlations and trade offs among fitness components (Antonovics 1976b). 1.4 Evolutionary Theories of Range Limitation Lack of genetic variation is the classic explanation of range limitation. It was thought that the populations at the periphery would suffer from high homozygosity because they 12 are too small and inbred, short lived, and usually the product of a founder event. However, the empirical evidence reviewed above does not support this. Peripheral populations seem to have equal amounts of genetic variation when compared to central populations (Shumaker and Babble 1980, Farris and Schaal 1983, Blows and Hoffman 1993, Stewart and Nilsen 1995). As gene flow can mitigate the loss of genetic variation, this was a logical direction to proceed leading to a second theory on species range limitation (Richards 2000). Continuous infiltration of genes, discussed in the introduction to describe an aspect of centrifugal speciation, destroys the gene combinations necessary to meet the demands of the local border habitats (Levin and Clay 1984). Kirkpatrick and Barton (1997), though not the first to describe this theory, developed a model built upon the work reviewed above. The model describes a situation where a species is well adapted to conditions at the centre of the range. The greater the distance from the centre of the range the less well adapted that species is to the conditions in which it lives. Maladaptation to local conditions conveys a decrease in fitness; thus, the fecundity and population size are lower. The maladapted peripheral populations will produce fewer gametes than the central, well adapted population. Because of this unequal gamete production, the peripheral population can be described as a sink and the central population a source and consequently gene flow will be from the centre outwards. Genes from the centre, well adapted to conditions there, will swamp the peripheral populations and prevent the peripherally adapted genes from increasing in frequency in the local population (Garcia-Ramos and Kirkpatrick 1997). 13 The biological basis for the two main assumptions of the model was reviewed above. The model both predicts and assumes a normal distribution of fitness and population size across an environmental gradient. The work of Whittaker (1956), Watkinson (1985) and Brown (1984) provides the basis for the assumption that the central population is swamping the peripheral population, such that the peripheral population is so overwhelmed by the new genes that the mutations adapted to the peripheral environment have no chance of successfully establishing. As Table 1 illustrates, outcrossing species, where gene flow exerts influence over the population structure, show no difference in genetic variation between population in the peripheral and central environments. This supports the theory that gene flow from the centre submerges the adaptive variation in peripheral populations and differences cannot be detected. Table 1 also demonstrates that inbreeding species show clear patterns of genetic variation, which is also consistent with the theory of range limitation by gene flow. The final theory of range limitation recalls that differences between the marginal and central environment are a result of strong selection pressure on an array of characters. The result of such pressure depends on how the selected characters are genetically correlated with each other (Antonovics 1976b). Should two traits conveying increased fitness to a marginal population be negatively correlated, this could limit the spread of the population. A barrier to the spread of the species is inherent in such a negative 14 correlation because the simultaneous development of both traits in concert is physiologically and/or genetically impossible. A negative correlation between traits has been noted in several plant populations (Antonovics 1976b; p. 233 for discussion) but when a population enters a new habitat the nature of the correlations can change. For example, the trait conveying increased fitness in the new environment could be entirely different, not correlated with any other fitness characters. It is also important to point out that gene flow could influence the correlation of such traits and disrupt linkage in marginal environments (Parsons 1991) . Conclusion The evolutionary theories discussed above are inconclusive. There appear to be few consistent patterns of ecological or evolutionary difference between central and marginal populations. The regulation of density and genetic change are likely to be entirely different in central versus marginal environments (Antonovics 1976a, Levin and Clay 1984). Such variability remains at the core of understanding what limits natural selection, and it is at the species border that we can examine this fundamental question of evolutionary biology. As Mayr (1963) stated four decades ago, "The species border [remains] one of the most interesting phenomena of evolution and ecology, and as a scientific problem it has been almost totally ignored". 15 Chapter 2: A latitudinal transect across the Pacific coast range to two closely related congeners, Cakile edentula and C. maritima Species range boundaries are critical features of our natural world, yet the question of why this fundamental limitation occurs is still unanswered. Why doesn't a species become adapted to the peripheral environment? Moreover, if it can become locally adapted, what is preventing it from spreading from that local population to areas beyond its current range? There are a number of hypotheses addressing this problem (da Cunha et al. 1959, Soule 1973, Antonovics 1976, Kirkpatrick and Barton 1997). The first two are endogenous and involve the marginal population irrespective of outside influences. For example, the populations at the margins are unable to adapt due to a lack of variation resulting from a small population size. This theory, although popular, has little experimental support (Shumaker and Babble 1980, Farris and Schaal 1983, Brussard 1984, Blows and Hoffman 1993, Stewart and Nilsen 1995). Alternatively, perhaps negative correlations or trade-offs in fitness components prevent adaptation to harsh environments. However, this case requires equivalent selection pressure on the correlated traits (Antonovics 1976). A third theory concerns the effects of gene flow on marginal populations. It proposes that at the centre of the range, a species is well adapted to local conditions. 16 The environment changes further from the centre and the species becomes unfit, as it is not adapted to marginal conditions. The fecundity, population size, and reproductive rate accordingly decrease. Thus, the central well-adapted population will produce high numbers of gametes, and the maladapted peripheral populations will produce fewer gametes. This disproportionate production causes the marginal population to be a sink and the central population a source; the movement of genes is from the centre outwards. The genes from the centre, well adapted to conditions there, will inundate the peripheral population and prevent genes adapted to the margins from increasing in frequency. A stable range limit is found when adaptation is balanced by gene flow at the periphery (Kirkpatrick and Barton 1997, Case and Taper 2000). This qualitative theory offers a number of testable predictions. The rate of evolution is affected by population density; asymmetrical gene flow impedes the accumulation of fitting adaptations in peripheral populations. Density depends on how well adapted the population is to local conditions; thus, there is a feedback between demography and evolution. Should the species have a normal density distribution, the higher density populations would be well adapted to local conditions resulting in the growth of the population. The smaller populations will be maladapted to local conditions and, as a result, not increase in size. Thus, the theory both predicts and assumes a 'bell-shaped' density distribution over the range of the species. 17 An example of half this 'bell-shaped' distribution is in the density decline in the grass Anthoaxanthum odoratum across a "woodland-field ecotone," observed by Grant (1974). There was a marked difference in population turnover and size between the peripheral and central sites however; the marginal population density remained constant over the three years of the study. This stability suggests steady population density regulation rather than simply distended density dependence or severe density independent factors. This chapter, based on the predictions listed above, has two main objectives: first, to compare the predicted and actual density distribution of a species' range and test the assumption of normal density distribution and second, to determine to what extent gene flow affects density distributions by contrasting inbreeding and outcrossing species. LimUations and Requirements To fulfil this project's intention, a system must be chosen which complies with the many limitations and requirements of the theory. The first two requirements are that the conditions at the margin are different from those at the centre of the range and that the cline is linear. In other words, the environment changes one-dimensionally. The species must be continuously distributed in space and dispersing randomly. The traits that cause limitation of the species must respond to selection and must change with the linear environmental cline. The limitation of this theory is in its applicability; the trait cline and selection must not be too strong 18 or the population will go extinct. Neither can the trait slope be too gradual or the population will spread to cover the entire gradient (Kirkpatrick and Barton 1997). Habitat No natural arrangement can meet these exact requirements; however, the proposed system below attempts to satisfy the major requirements and provide a reasonable system with which to test these predictions. The threatened Pacific North American coastal dune community begins at the top of the fore dune, closest to the ocean, and continues towards the ocean to the mean high tide line. Characterised by high exposure to sand and salt spray, shifting substrate, and low water and organic matter, this distinct habitat is arranged along a linear coastal environmental gradient from south to north (Keddy 1981) and is dominated by a temperate maritime climate. This Pacific maritime climate, though not as sharply differentiated as the Atlantic coast, varies from south to north. "A typical west-coast sub-tropical climate is found in California", whereas the climate of the Pacific Northwest is rainy and cool most of the year (Ward 1920; p.565). The modern difference in average summer temperature shows a similar, subtle slope decreasing from south to north (Fig IA). The temperature changes at a rate of about 0.73C per degree in latitude. The slope is subtler in winter, at approximately 0.3 8C per degree of latitude (Ward 1920). Fig 19 IB shows the current difference in rainfall between the north and south Pacific coasts. A slope of 14.73 cm per degree in latitude again shows a climactic difference between the north and south Pacific coasts. Pacific coast average summer temperature from 1959 to 2000 100 80 ^ cr eo § | 40 I S 1 20 ~ o y = -1.3174x+ 120.38 R2 = 0.6918 32.3 22.3 12.3 2.3 -7.7 -17.7 a — = <-> 2 8 a) a> E QJ 40 42 44 46 48 50 52 54 56 58 60 62 64 Latitude Pacific coast average rainfall from 1959 to 2000 R 2 = A 5 9 9 9 40 42 44 46 48 50 52 54 56 58 60 62 64 Latitude Figure 1 Average temperature (A) and rainfall (B) along the shores of the Pacific Coast from 1959 to 2000. Regression lines show the trend of increasing moisture and decreasing temperature going up the coast. Data provided by the Western Regional Climate Center, B A more effective method to discriminate between these two climates is through degree-days, a botanical estimate of physiological time. Degree-days measure temperature differences between the average daily temperature and a base temperature. Plant development requires a specific total heat, which can be estimated through an accumulated number of degree-days. For example, if the base temp is 55F and the average daily temp is 56F, one degree-day will accrue. As seen 20 in Fig. 3, the average number of degree-days above 55F in San Diego, CA is 3,494 degrees above the average per year. However, in the Pacific Northwest, Longbeach, WA, the average number of degree-days above 55F is 481 (N.O.A.A., 2002). This drastic decrease as one moves north identifies one of the many differences between the north and south Pacific coastal environments. Annual Degree Days (Base Temperature 55 F) at Site Latitudes 4000 to >> co O <i> ~a> 0) co »-fe™ E 2000 » li- a O n & • * i 3000 1000 30 35 40 45 50 55 Latitude F i g 2. Average annual degree-days by latitude along the shores of the Pacific Coast from 1971 to 2000. A trend is apparent of decreasing Degree-days going up the coast. Data provided by the National Oceanic and Atmospheric Administration of the U.S. Department of Commerce, National Climactic Data Centre, June 20, 2002. The suitability of the zone for this study is reflected in the obvious linear temperature and moisture gradient that insures the peripheral environment is distinct from the central (Fig. 1). The decreasing trend in degree-days also supports the assumption that the environments of the north and south Pacific coast are distinct (Fig 2). The restriction of a species by gene flow requires a difference in 21 selective environments great enough to prevent successful recruitment at the margins. Study system Cakile edentula (Bigel.) Hook. var. edentula was introduced to the west coast in the San Francisco Bay Area in 1880-1882 and migrated at an amazing rate of 64 km per year for 50 years (Barbour 1970a). Cakile edentula is a succulent, annual member of the Brassicaceae. The seeds germinate synchronously in the spring and mature in late summer to produce large, easily counted segmented fruits. Each fruit comprises a proximal and deciduous distal segment each containing one seed. Distal seeds are on average larger than proximal seeds (distal seeds = 0.009g, SE= 0.002, proximal seeds = 0.007g, SE= 0.001; Rodman 1974). The deciduous distal segment detaches readily from the parent plant when mature and disperses over large distances. Its ability to float for upwards of a month in salt water aids in this coastal plant's dispersal ability (Rodman 1974, Payne and Maun 1981). Self-compatibility is high in this species as indicated by high rates, 94%, of autogamy in the greenhouse (Barbour 1970a, Donohue 1998) and reduced or non-existent perianth (Rodman 1974; field observations). Cakile maritima Scop. ssp. maritima, the presumed evolutionary progenitor of Cakile edentula, was first collected on the west coast 15 miles north of San Francisco in 1935, about 50 years after the introduction of C. edentula. At this 22 time, the two species were growing intermingled on the California coast (Barbour 1970b). C. maritima spread at a rate of 33 miles per year and within 50 years replaced C. edentula in the southern latitudes of its range. Reports on the historical biogeography of the two species indicate a replacement of C. edentula by C. maritima not only on the Pacific coast of North America but also on the south Australian coast (Rodman 1974, 1986). Fig 3. Picture of C. maritima (top right) and C. edentula (bottom left) seedlings take in Oregon in July 2001. Note the difference in the lobed leaves of C. edentula compared to the pinnatifid leaves of C. maritima. C. maritima differs from C. edentula in two major aspects. The first and most noticeable is the shape of the succulent leaves as seen in Fig 3. In C. maritima they are deeply pinnatifid and greyish green, whereas in C. edentula they are ovulate or lobed and a bright green. Moreover, as seen in Fig 4, the flowers of C. maritima are large with pale purple to rose petals and a large nectary. The pollinator is unknown however wind is suspected to play a large part in pollen dispersal (Rodman, 1974). The flowers of C. edentula are small. Seventy-five percent of the flowers observed 23 by Rodman (1974) were incomplete in both perianth and ovary, and largely unproductive compared to C. maritima. Fig 4. Picture of C. maritima (top) and C. edentula (bottom) taken at Devil's Punch Bowl Beach (Beach 3) in Newport, Ore. (Site 7). Contrasting leaves are shown as well as the absence of inflorescence in C. edentula. Species Suitability C. maritima and C. edentula coincide for a large portion of their range and are wide ranging, opportunistic plants. Due to this wide-ranging habit, the deleterious effects of gene flow are presumed severe, as foreign input constitutes a large portion of the fruit density, especially in peripheral populations (Donohue 1998). They are similar in vegetative and reproductive structures allowing exact physiological comparisons. They occur along a continuous linear gradient of temperature and rainfall, and the width of the range is narrow, therefore not confounding the source/sink dynamics so 24 important to this study. Although the species was recently introduced, the annual nature of the species provides for the maximum number of generations post-introduction allowing populations to move towards equilibrium and adapt to local environmental conditions. Finally, the two species are distinct in terms of their reproductive strategy, one an obligate outcrosser and the other highly selfing, making these species suitable for the proposed study. 2.2 Materials and Methods As reviewed above and in reference to the objectives listed, the theory of range limitation by gene flow rests on a number of testable predictions. The first prediction of a decreasing trend in abundance from the centre to the margins can be tested by a simple measure of density distribution over the range of the species. Comparing the predicted and actual density distribution During the summer of 2001, the North American Pacific coastal dune habitat was surveyed for both C edentula and C. maritima along the latitudinal temperature and rainfall gradient ranging from San Diego, CA to Masset on the Queen Charlotte Islands, BC (Ward 1920, Hutchins 1947); Fig. 1). Each of thirteen approximately evenly spaced sites (160km) consisted of three distinct beaches spanning at most 26 km wherein three separate mixed populations, at least 0.5 km apart, were measured (Fig 5). The results of Donohue (1997) indicate that this spacing allows these to be considered distinct populations. For each of the 117 populations arranged south to north along the coast (summarised in Appendix II), the Global Positioning System 25 co-ordinates, area, fitness of the plants as measured by seeds/plant, and population size were counted and recorded. Where possible, 100 mature fruits were collected and stored in paper bags at room temperature until used (Boyd and Barbour 1993). Fig 5. Schematic of the sampling procedure along the Pacific coast. Thirteen sites along the Pacific coast were sampled; three beaches per site within 26 km and three populations per beach with at least 0.5 km between them. Pictures courtesy of Good Nature Publishing Company c. 1998. 26 Contrasting the density distributions of an inbreeding and outcrossing species In contrasting the density distributions of inbreeding and outcrossing species, the information collected in the 2001 field season was used to compare the northern portion of both species ranges. Graphs and 95% confidence intervals were used to determine if a change in population density occurred across the range of each species. A direct comparison of the northern distribution of the density of the two species was not performed. 2.3 Results The field observations showed a clear pattern of density decline across the geographic range of both species. Populations of C maritima in southern California (Sites One and Two, San Diego and Santa Barbara) were very dense with high numbers of seeds per plant. A typical population consisted of three to four very large individuals (1 meter in diameter) producing upwards of 2,000 seeds per plant and many, perhaps hundreds, of smaller individuals (less than 10 cm in diameter) producing not more than ten seeds, many with no seeds at all. It is possible this "second crop" was produced from the early mature seeds of the large individuals. Populations were situated on the landward side of the first dune or within sheltered areas, and only a winter storm or high wind would likely coax the seeds from the base of the parent plant. Thus seeds would likely remain where they fell and germinate later in the season producing a smaller "bumper crop". 27 Moving north, Site Three (Monterey) had smaller populations with high numbers of aborted seeds and signs of herbivory. Sandy cliffs were visible here and C. maritima was seen growing quite high on their face. These cliff faces provided a sharp interruption of the beach habitat and were a conduit for rainwater run-off; C. maritima was observed growing in or near these areas of fresh water. Sites Four and Five (Pt. Reyes and Eureka) showed significantly reduced densities, and plants were seen developing runners rather than growing upright. Plants typically occupied less than ten cm in diameter; many were dry and brown with immature, dry seeds. Populations were found in sheltered areas behind driftwood, at the base of sandstone cliffs, or further up the slope. An interesting feature of Site Five (Eureka) is that many of the C. maritima plants seemed to show intermediate characteristics with lobed leaves, rounded seeds, and shorter reproductive branches reminiscent of C. edentula. These plants were counted as C. maritima in future data as they exhibited many typical C. maritima features such as obvious flowers; moreover, no true C. edentula was detected in this area. In southern Oregon, Site Six, C. edentula was found in small numbers with many drying and immature plants less than 10 cm in diameter. The reproductive stalks contained some smaller seeds. However, many populations of C. edentula were larger than C. maritima, and in the more exposed areas C. edentula was the predominant species. The C. maritima populations were found only in sheltered 28 locations, and they were still dominant here; however, an estimated 95% of the observed seeds were aborted. In northern Oregon and southern Washington (Sites Seven and Eight) healthier C. edentula with more seeds per branch remain. Microclimatic differentiation was still observed with larger C. maritima plants seen in sheltered locals, but there was no longer a density difference between the two species. C. edentula plants were larger here than found farther south (over 10 cm in diameter), and the reproductive stalks were longer and supported over 100 seeds per plant. Both species were found growing near organic matter at the tide line. Southern Vancouver Island (Sites Nine and Ten), having a rocky shoreline with pebbly beaches, did not show a reduction in either size or seed production of either species when compared to southerly populations. However, the health of C. maritima was consistently low and aborted seeds, desiccated, and dead plants were more prevalent; C. edentula was also dried although not as severely. Microclimatic differentiation observed farther south, where C. maritima was dominant in more sheltered sites, was no longer apparent. Site Nine was unusually successful as seen in Figures 8, 9, 10, and 11. Seed count and densities were likely affected by the climate in the highly sheltered Juan de Fuca Strait. The conditions of the strait are among the mildest in Canada with a summer mean of 14C or 57.2 F (Environment Canada, 2002). The environment at 29 Site Nine is similar to the climate conditions of the south Oregon coast, at 54.4F or 12.4C (Fig. 1); thus, it would not be unusual to find density and fecundity aberrations here. Northern Vancouver Island (Site 11) populations of C. edentula were large, as was typical of the southern California C. maritima. Population samples consisted of four or five plants more than a meter across producing high numbers of seeds. In some populations, no C. maritima was found. Site 12 contained no Cakile; a search of suitable sites was conducted by not only the author and field assistant but also others familiar with the area and produced no Cakile. It is unknown why Cakile was absent at this site as the beaches here seemed no different from southern Vancouver Island and sightings have been reported in previous years (J. Pojar, personal communication). Perhaps a search in a milder year would have proven fruitful. Fig 6. Masset Cemetery on Queen Charlotte Island, Site 13. Note the delimitation of plant growth along the beach and the abundance of C. maritima in this northernmost location. 30 The final site on the Queen Charlotte Islands had populations of both species and a slight increase in sightings of C. maritima compared to other northern sites. Beaches were distinctly stratified with Cakile only occurring within a restricted zone along the beach (Fig 6). C. edentula was found here in comparable numbers, although there were more mature, reproductive individuals. The last population studied (Rose Spit) contained one dried C. edentula, with no leaves and one reproductive stalk in an area where no other plants were observed for miles. Comparing the predicted and actual density distributions Figure 7 shows the density distribution of C. maritima from the southernmost site (San Diego, California) to the northernmost site (Masset, Queen Charlotte Islands, British Columbia). There is a gradual sloping decline from what is presumably near the centre of the range, San Diego, to the northern limits. The southernmost population, with a density of 4.11 plants per square meter, has a 95% confidence interval of ±1.344 and the northern most populations, with a density of 0.207 plants per square meter, has a 95% confidence interval of ± 0.152. Thus, there is a distinct and significant density difference between the northern and southern populations of C. maritima. Moreover, a linear regression performed on C. 31 maritima's density decline is significant (r2=0.683, p=0.0009) and gently sloping (slope=-0.1405), extending over the whole of the measured range. « 6 F i g 7. Average C. maritima plants per meter square vs. latitude. A gradual density decline is seen throughout the examined range. Site numbers are labeled on each point. Above 40 degrees latitude, the density distribution follows a clear decline. Below this point, the density continues to decrease with one outlier at 48.37 degrees latitude, Site 9. Fig 8 shows the population density curve of the selfing species C. edentula. There is a plateau-like relationship between the centre and northern densities; the distribution reveals a realm of little variation in density from northern Oregon to British Columbia, following an initial increase in southern Oregon. The 32 northernmost sites emphasise this plateau shape distribution as the density falls from its highest point of 2 plants/m2 to zero in a short 300 miles. Although there is a curvilinear density distribution, by comparing the confidence intervals of the central (sites 8 and 9) and peripheral sites (southern = Site 5 and northern = Site 13), it is apparent that there is no clinal difference in population size. 3.5 2 c SS a. 'I £ 2.5 n CO ~-" 5 fc. •B CD I I CD fc. • CD O Q. CD CD CO fc. CD > < 1.5 0.5 32 -2-*T3-*-4 37 -•-5 Fig 8. Average C. edentula plants per meter square vs. latitude. The graph shows a plateau shaped density distribution with a relatively steep northern decline. Site numbers are labeled on each point Fecundity changes in C. maritima are obvious. The southern populations have approximately 2.5 plants per square meter and about 300 seeds per plant whereas the Queen Charlotte populations of C. maritima have approximately 0.1 plants per square meter and an average of 30 seeds per plant (Figs 7 and 9). Northwards, the number of seeds per branch decreases, and the number of aborted seed increases. 33 Fig 9 shows this significant decrease in seed production, reaching very low numbers in the northern latitudes. Site 9 again shows unusually high production. The mean and confidence interval of the southern populations plant productivity is 240.59 ± 192.05 and for the northern populations are 7.99 ± 8.34. There is a difference in the number of seeds per plant between the central and peripheral populations providing support for the assumption of asymmetrical gene flow. q> 800 -i Latitude Fig 9. Average number of C maritima seeds per plant vs. latitude. Data were collected by counting the number of seeds on each of ten plants per population, on each of three beaches for 12 sites and averaged. Thus each point is an average of ninety observational counts. Each site is labeled. The graph shows a decline in the number of seeds per plant. Cakile edentula suffered less dramatic reductions in density moving north but there is still a visible decrease in abundance from 0.8 plants per square meter on southern Vancouver Island to 0.4 plants per square meter (Fig. 8). However, the number of 34 seeds per plant increases moving north, with roughly 50 seeds per plant in the south to approximately 250 seeds per plant in the Queen Charlottes (Fig. 7 and 9). Figure 9 shows this dramatic increase in seed production by C. edentula, with its highest point at the edge of the geographic range. The mean number of seeds per plant in the northern population is 5.1 ± 2.8 and the southern populations is 294.4 ± 244.3; thus, the average number of seeds per plant significantly increases moving northwards. Fig 10 Average number of C. edentula seeds per plant vs. latitude. Data was collected by counting the number of seeds on each of ten plants per population on each of three beaches for 12 sites and averaged. Thus each point is an average of ninety observational counts. Each site is labeled. The graph shows an increase in number of seeds per plant moving north. Contrasting the density distributions of an inbreeding and outcrossing species As seen in Fig 7 and Fig 8, a northward population decline is apparent in both species. Cakile maritima shows a difference in population density between southern 35 and northern populations whereas C. edentula shows no such pattern. The graph of C. maritima is gently sloping (r = 0.683, p=0.0009) with a decline of 1 plant per square meter over roughly 800 km. The C. edentula graph shows no significant evidence of a pattern of density decline. 2.4 Discussion The results of the density distribution analysis indicate a clear pattern of decline from south to north in C. maritima. No significant pattern of decline was found in C. edentula; however, a steep drop in population density at the northern range limit was noted. There are a number of possible explanations for this pattern or lack thereof. Both C. edentula and C maritima were introduced to the west coast no earlier than 1880, and it is possible that they have not yet reached equilibrium. The results might show the progress of expansion of these species and the trend observed might be transient; high seed set in C. edentula in the north might confirm this explanation. However, Dudley (1996) reported evidence of local adaptation in C. edentula in response to water availability at wet and dry sites along the coast of the Great Lakes. This experiment showed a selective response in physiological traits concerned with water loss in C. edentula found in wet and dry sites along a coast line, and it is likely that these adaptive responses developed within the time C. edentula has been observed on this coast. Comparing the predicted and actual density distributions 36 Based on the proffered model of range limitation, predictions of density distribution in an inbreeder and outcrosser match these two clinally varying species. To review, this model of species limitation states that for a species occurring along a line of environmental or geographic transition (where the species is most fit at the centre of its range), the fitness of the species will decrease as it moves further from the conditions of the central environment. Healthy individuals at the centre of the range produce a large number of propagules, and smaller numbers of offspring are produced from the less fit peripheral populations, resulting in asymmetrical dispersal. This unbalanced gene flow will cause the populations at the margins to receive greater amounts of alleles originating from the centre. The central propagules disseminating towards the margins of the range carry alleles adapted to the central environment. As these alleles spread into the peripheral populations they prevent the increase of alleles conferring local adaptation at the periphery. This keeps fitness low as it prevents adaptations to cope with marginal conditions. Thus, the rate of evolution in the peripheral populations is inversely proportional to the rate of gene flow from the centre of the population. The density of a population depends on how well adapted a population is to its local environment, which is determined by the rate of evolution. The rate of gene flow will therefore indirectly affect population density. 37 If gene flow is high, the deviation between the optimum adaptation and the mean phenotype of the marginal populations will be substantial. This deviation is expected to increase as the local optimum phenotype moves further and further away from the optimum central phenotype, i.e. moving towards the geographic periphery of the species range. The population is expected to show a decline in the density distribution, as it will be more range limited by the homogenising effects of gene flow. However, if gene flow is low, as in an inbreeding species, the peripheral populations are free to evolve to the local optima. The distribution is expected to have more of a plateau shape with the population at the carrying capacity throughout the flat summit area and escarpments where density declines rapidly to zero at the distribution's edge. The current analysis illustrates these patterns and lends support to the above theory. Figure 7 demonstrates a significant pattern of density decline throughout the sampled portion of the range of C. maritima. The 95% confidence interval of the northernmost population (4.11 ± 1.344) does not overlap with the southernmost population's confidence interval of (0.207 ± 0.152); thus, there is a significant difference in density between these two populations. Moreover, a linear regression performed on C. maritima's density decline lends support to the theory (r2=0.683, p=0.0009), a significant, gently sloping (slope=-0.1405), and consistent decline 38 conforms to the predicted pattern of decline described in the gene flow theory of range limitation. There is a significant difference between the seed number per plant in the southern sites (240.59 ± 106.90 seeds per plant, Fig 9) and the northern sites of C. maritima (7.92 ± 8.34 seeds per plant). This further supports the conclusion that C. maritima displays a gradual declining trend with qualities similar to those predicted and assumed in the gene flow model of range limitation. These features are that C. maritima has a large seed source (with over 1400 seeds on one plant in the southern population) and population density at the centre of the range and, as observed in the field, the periphery of the range has lower population density and seed production and shows more cases of maladaptation (immature seeds, aborted seeds, immature plants dead and/ or drying). Moreover, the decline in fecundity and density occurs in a gradual fashion over the approximately 1187.5 km measured range. This pattern of gradual decline in organism abundance at the edges of a range has been shown in many species. In 1956, Whittaker reported that most organisms studied in the Great Smoky Mountain ecosystem show this type of density curve (Whittaker 1956). Similarly, and in support of the theory of range limitation by gene flow, Emlen (1978) in his survey of density patterns of avifauna in North America reported undifferentiated species showing greater density reductions towards the margins of their range as compared to polytypic species. 39 Finally, this pattern of density decline has been found in populations of Anthoaxanthum odoratum (Grant and Antonovics 1978). This experiment not only surveyed the range of A. odoratum across a North Carolina woodland ecotone but also compared the genetic structure of central and marginal populations. Realising the importance of gene flow to the range limitation of the population, this study also attempted to quantify gene flow by determining the range of influence that seed and pollen dispersal from the centre might have. It is important to note that the species used is a strict outcrosser and individuals are distinct and identifiable. In all years measured, the central population had the highest density, survival rate, and lowest mortality rate. This is in contrast to the marginal site where significantly fewer individuals were found and mortality rate was significantly higher. This shows a potential maladaptation to the local environment in the peripheral population. As an indication of gene flow in the A. odoratum study, the marginal recruitment rate was more than twice that of the centre. Indirect, phenological measurements of gene flow, such as the onset of pollen dispersal and flowering time overlap in the marginal and central populations, revealed the proportion of pollen flow into the margin was 35% whereas no pollen flow was detected in the opposite direction. However, Grant and Antonovics (1978) concluded that the proportion of central gametes dispersed was not enough to swamp the marginal population and keep it from adapting. Most relevant to the current study is the density of both the marginal and the central populations. It was found that the density of the margins 40 was significantly lower than that of the centre. This is similar to the findings of the current study. Although there are many reports of similar density patterns, there is little evidence of local populations maintained away from the locally optimum phenotype. Camin and Ehrlich (1958) reported anomalies in banding patterns of the water snake Natrix spedion on the islands of Lake Erie. On islands where snakes faced restricted habitats, adult un-banded snakes were the most common whereas mainland population adults exhibited dark banding patterns. However, a number of island snakes were patterned to some degree, and it was hypothesised that populations of banded snakes were maintained on the islands although maladapted to the local island environment. As it was found that brood composition on the islands was not uniformly banded or un-banded and the adult island population is strongly biased towards un-banded snakes, this suggests gene flow from the mainland to the islands is responsible for maintaining those few populations of patterned snakes on the islands. Stearns and Sage (1980) performed another experiment investigating local maladaptation caused by gene flow in their study of the mosquitofish, Gambusia affinis, in the Armand Bayou in Texas. This experiment compared populations of fish in brackish and fresh water and found populations of fish maintained in freshwater were maladapted to that situation and better suited to the saline 41 condition. Electrophoretic data show a slight genotypic differentiation between brackish and freshwater populations; however, further experimentation proved this difference to be non-adaptive. Thus, Stearns and Sage (1980) concluded that life history differences between the two populations are caused by maladaption to the freshwater environment. Contrasting the density distributions of an inbreeding and outcrossing species The above-described experiments suggest that gene flow may indeed restrict local adaptation. If one allele is favoured locally and another is introduced in high frequencies from a large gene pool, the introduced allele will reach high frequencies when selection pressure is smaller than the effects of migration (Haldane 1930). This describes a situation where gene flow is overwhelming local adaptation and is the essence of the range limitation hypothesis. The theory of asymmetric gene flow leads to the following specific predictions: the outcrossing C. maritima, constrained by gene flow, is expected to exhibit a gradual decline in density towards the range periphery. In contrast, the inbreeder C. edentula, unconstrained by gene flow, is predicted to have adapted to the local conditions; population fitness and density is expected to remain high throughout the range. Thus, the expected density distribution of C. edentula is a plateau with steep slopes at the boundary. 42 C. maritima shows a significant decrease in population size moving from south to north (Fig. 7). Moreover, average seed number per plant also decreases (Fig. 9). This reduced fitness seen in the marginal northern populations of C. maritima is similar to the situation in Natrix spedion (Camin and Ehrlich 1958) and Gambusia affinis (Stearns and Sage 1980) and could be a result of gene flow causing maladaptation to the marginal environment. It is therefore reasonable to hypothesise that C maritima, suffering from decreased fitness at the margins, might be restricted by gene flow from more central populations. Cakile edentula also shows a pattern similar to that predicted by the theory. The decline at the northern edge of the population is steep, and the apex of the density curve is flat, spanning the majority of the C. edentula range. However, as there is no difference between the density of the central and peripheral populations, no clear resolution is possible. It is certainly conceivable that these species may be constrained by gene flow; they are capable of long distance dispersal and a large number of the seeds produced travel great distances. Donohue (1998) reported 51% of C. edentula fruit disperse "beyond recovery", and Payne (1981) reports C. edentula to be capable of travel upwards of 600 km on ocean currents. As reported here, the central environment of C edentula can support 0.45 plants / m and each plant can produce from 100-300 seeds. Thus, for every square metre of sandy beach, 135 seeds are produced and, with a low estimate of 51% dispersal rate as reported above, approximately 68.8 43 seeds per m of sandy beach are dispersed. This is astonishing when one considers the 1400 km of coastline this species covers. The dispersal capabilities of C. edentula are only equalled by C. maritima (Rodman 1980). However, the density decline observed does not confirm the theory of range limitation by gene flow. The effects of gene flow do not exclusively cause such patterns (Endler 1973). Several models characterising the influence of gene flow found no detectable difference between populations experiencing various levels of gene flow. "Gene flow may be unimportant in the differentiation of populations along an environmental gradient" (Endler 1973; p. 249). To demonstrate specifically that C. maritima is restricted in its range by gene flow from more central populations, a reciprocal transplant between central and marginal populations would be required. Thus, more work is needed to clarify the above evidence of population differentiation. 44 Chapter 3 -Variation in Seed Oil Composition in Two Species of Cakile, C. maritima, an Outcrosser, and C. edentula, an Inbreeder, along a 1400 km Latitudinal Gradient Gene flow is the "change in gene frequency of a population due to the movement of gametes, individuals, or groups of individuals from one place to another" (Slatkin 1987; p.787). Gene flow may restrict the development of high levels of genetic differentiation over the range of a species and may limit local adaptation. Through this migration, the species as a whole may be prevented from expanding its range. Mayr (1963) theorized that the process of local adaptation by selection in the peripheral populations would be interrupted by migration of novel genes and gene combinations from the center of the species range. This prevents gene arrangements adapted to the border conditions from increasing in frequency. The ideas of Mayr (1963) and others have been formalised by Kirkpatrick and Barton (1997) into a model that predicts range limitation as a result of gene flow along a linear environmental gradient. This theory predicts a 'bell-shaped' density distribution in species restricted by gene flow and posits, symptomatic of this migration, that marginal populations will be maladapted to marginal conditions. Thus, an adaptive trait subject to selection that varies along a linear gradient is needed to test this theory. Linder (2000) identified such a trait and performed a series of analyses testing the prediction that "germination temperature is an important selective agent causing seed oils of higher-latitude plants to contain proportionally more unsaturated fatty acids than lower-latitude plants" (Linder 2000; p. 442). 45 Linder performed three assays investigating the response of seed oil composition to changes in germination temperature. Using published data on seed oil composition of over 700 species of angiosperms, Linder's first analysis showed that temperate plants have consistently higher proportions of unsaturated fatty acids (FAs) in seed triacylglycerides (TAGs) than tropical plants. Furthermore, species of Helianthus existing on a latitudinal gradient show a remarkable declining trend in the proportion of saturated fat in seed oils as latitude increases. In support of this theory that the proportion of saturated fat in seed TAGs responds to changes in latitude, Linder's second experiment determined that differences existing between the northernmost and southernmost populations of H. annum and H. maximiliani. Third, germination performance experiments confirmed this trend, and the results show that seeds with a higher percentage of saturated fats germinate faster at higher temperatures. This evidence supports the hypothesis that a latitudinal gradient in unsaturated FAs in TAGs exists in Helianthus caused by selection for an adaptation to local germination temperatures. Building on the conclusions of Linder (2000), the following chapter examines patterns of plant seed oil composition in both Cakile maritima and Cakile edentula throughout their Pacific coastal range. In comparing the seed oil composition of two closely related species, one outcrossing and one selfing, the effects of gene flow on peripheral populations will be examined. The theory of range limitation by gene flow predicts that the homogenizing effects of gene flow will temper the adaptive response in peripheral populations of an outcrossing species when compared to a selfing species. As there exists evidence of a relationship between the FA content of seeds and a latitudinal/ temperature gradient, the following predictions can be made as to the proportion of unsaturated FAs in the central and peripheral populations of both C. maritima and C. edentula. C. maritima experiences large amounts of gene flow both from seed migration and pollen dispersal; thus, peripheral and central populations are expected to have similar proportions of saturated FAs. Gene flow from the large, central population is predicted to keep the marginal populations maladapted to the peripheral environment. Specifically, high levels of saturated FAs found in the central environment of southern California will also be found in the marginal environment of northern British Columbia. Gene flow in C. edentula is assumed to be relatively minor as it is only accomplished through seed migration. Therefore, C. edentula is predicted to show a cline of decreasing proportions of saturated FAs moving northwards. Central populations of C. edentula should have little influence over the northern populations; thus, gene flow should not keep the marginal populations from adapting to their local environment. Low levels of saturated FAs are expected in 47 northern latitudes such as northern British Columbia, and southern latitudes such as the Oregon coast are predicted to have higher levels of saturated FAs. The existence of this latitudinal gradient in the proportion of saturated fat is determined by the composition of lipids within a seed. Lipids, specifically seed oils, are the primary reserve material in most seeds and are found in oleosomes or oil bodies within the cotyledons or endosperm (Esau 1977). They are present as intracellular triacylglycerides (TAGs) and are made of chains of fatty acids (FAs) consisting of three FAs esterified to a molecule of glycerol (McMurry 1994). Each TAG molecule is composed of a few major and many different minor FAs (Huang et al. 1988). During the early stages of germination, the TAGs are mobilised by lipases which hydrolyze the ester bond that connects the FA to the glycerol releasing the FAs, which are then catabolised to a ready form of energy. The composition of the TAGs is important in this final oxidation as the energy gain associated with catabolism differs as a function of the constituent FAs (McMurry 1994). Specifically, the unsaturated FAs (those with double bonds in the carbon chain) require extra steps and enzymes in their production and breakdown; thus, less energy is gained from this form of storage. However, unsaturated FAs possess significantly lower melting points than saturated FAs. Lipase activity is required to release the energy stored in TAGs for the developing plant embryo. Lipases are diverse and distinct in their substrate specificity, optimal pH, hydrophobicity, and subcellular location (Huang et al. 1988). Hydrolysis is performed efficiently on TAG molecules in the liquid state; thus, liquid TAGs within the seed optimize rapid energy release for germination (Linder 2000). When seeds germinate at warm temperatures, the FAs are in the liquid state, and it is hypothesized that selection should favor the seeds that contain the most energy for growth. Those seeds with a higher proportion of saturated FAs will have a selective advantage. However, at cooler germination temperatures, below the melting point of saturated FAs, it is expected that seeds containing a higher proportion of unsaturated FAs (which lowers the melting point) will germinate earlier and more rapidly than seeds without unsaturated FAs. Therefore, it is reasonable to assume that the ratio of saturated to unsaturated FA will be under selection for local adaptation to germination temperature (Linder 2000). This assumption allows a direct examination of local adaptation in peripheral populations of Cakile. This hypothesis, that seed oils respond to latitudinal changes, builds on the presupposition that selection can affect the quantity and composition of seed oils. In order for a trait to respond to selection it must be heritable. Plant seed oil content and composition is under strict genetic control and historical breeding and selection reports a heritability of over 50% in many oil producing crops (Levin 1974). Moreover, Kondra and Stefansson (1970) discovered through a complex system of experiments that the composition of seed glucosinolates (related seed oils also possessing a glycerol backbone) are under polygenic control in Brassicaceae. 49 Experiments performed by Appelqvist (1971) have shown plant seed oil composition in Cakile to be highly conserved within species. Four samples of C. maritima from several locations in Sweden were studied and shown to be very similar in their composition. Moreover, the proportion of saturated fats as well as the proportion of specific oils of C. maritima reported in Appelqvist (1971) is similar to those who performed comparable analyses (Mikolajczak et al. 1961, Miller et al. 1965). Although the composition of C maritima seeds from locations within Sweden not more than 30 km apart are indistinguishable, the linoleic acid content of the samples from Sweden is slightly higher than samples collected from Israel (Miller et al. 1965). Appelqvist (1971) attributed this rise in unsaturated fatty acid to the extreme difference in climate between Israel and Sweden. Rodman (1980) performed a study of Cakile glucosinolates, comparing over 94 North American samples of roughly 50 plants each, and also found seed oil content to be species-specific. He states "the two species [C. maritima and C. edentula] are qualitatively similar but differ quantitatively in the percentage composition of oleic, linoleic, behenic, and erucic acids" (Rodman 1974; p.49). The two species also differ significantly in their seed oil to dry weight ratio with C. maritima at 44% and C. edentula at 49% (Rodman 1980). 50 .1 Methods Sample Procedure To determine the fatty acid composition of the seeds collected along the Pacific coast in the 2001 field season, 20 seeds (ten proximal and ten distal) were assayed from populations throughout the latitudinal range sampled. Not all sites surveyed contained mature seeds; thus, only sites where 20 seeds were available were examined. From these 20 seeds, the first ten seeds analysed were from the southernmost population on the southernmost beach of every site. The second group of ten seeds for each site was collected from the northernmost beach and population available at each site. The specific locations of the sites are summarised in Appendix I. Each sample comprises five distal and five proximal seeds. Seeds were kept in Petri plates and stored in paper bags until needed. Proximal and distal seeds were first compared to determine if differences in seed oil content disallowed combining the samples in further analyses. To perform the proximal and distal seed comparisons, seeds were again taken from the 2001 field season collection. For the first assay, an unpaired analysis, nine C. maritima proximal, nine C. maritima distal, ten C. edentula proximal, and ten C. edentula distal seeds were taken from the same population, beach, and site. In the second paired analysis seeds were again all from site 9 beach 3 population 2. Ten seeds of each 51 type were collected; the distal and proximal seed from the same fruit were considered a pair and evaluated as such. Gas Chromatography To extract the seed oils for the GC analysis seeds were shelled and crushed slightly to crack the seed coat. Whole seeds were placed in test tubes and treated with 1.0 ml of IM HCI in ethanol and 3.0 ml of hexane. This mixture was then incubated at 80C for 2 hrs and cooled again to room temperature for 0.5 hr to extract the TAGs from the seed body and sever them from the carbon backbone. One ml of 9% NaCl was then added to separate the seed oils from the solution and ease collection. Seed oils were then carefully extracted and pipetted into small 1.0 ml aluminum sealed vials, which were then inserted into the gas chromatograph. One ul of the resulting seed oil from the hexane fraction was injected into a GC 5890 Hewlett Packard gas chromatographer to begin the analysis. The injector port was kept at 250C to volatize the sample immediately. Part of the sample entered the HP-23 cis/ trans FAME column of 50% cyanopropyl-methylpolysiloxane that is 30 m long with a 0.25 mm diameter and a 0.25 um film thickness. The sample was carried gently down the capillary column by the flow of a carrier gas, in this case helium at 2 ml per minute. While passing through the column, the oven temperature was kept at 180C for one minute and then increased from 180 to 240C at a rate of four degrees per minute with a final five minutes isotherm at 240C. Fatty acid methylesters 52 (FAMEs) were detected using a flame ionisation detector kept at 350C. The flame ionization detector burned the FAMEs in a mix of air and hydrogen and detected the resulting ions. A description of the standards and quantification procedure, as well as a more detailed description of the process can be found in Kunst et al. (1992). Analysis The proportion of saturated fat per seed was calculated by adding the relative abundance of all saturated FAs; these significant chemical species within Cakile seeds are 16:0, 18:0, 20:0, and 22:0 with 24:0 a minor contributor. In order to determine whether to pool the two seed types, the difference in the proportions of saturated FA between the proximal and distal seeds was analysed. Both unpaired and paired experiments were performed and used to determine whether the mean of the proximal and distal populations was different. By combining the p-values of both the t-tests and the GLM analysis of the factor seed type and analysing this data with a Fisher's combined probability test, a corrected measure of the similarity in seed oil composition between the two seed types was obtained. The latitudinal comparison of the percentage of saturated fat in the seeds was done using both separate regression analyses and a generalised linear model (GLM). This characterises and tests the significance of the relationship between several independent factors such as latitude and species and analyses their contribution to variation in the proportion of saturated fat in a seed. The effect of species on seed oil 53 composition was also analysed using an ANOVA to further describe their relationship. : Results Seed Type Analysis The seed type analysis determines if a difference exists between the percentage seed oil in the two different types of seeds, proximal and distal, within a fruit. If no such difference exists seed types can be pooled in later analyses. Results of the unpaired t-test (Fig. 11, C. maritima p=0.864, df = 16, C. edentula p= 0.282, df = 18) and the more powerful paired t-test (C. maritima p=0.9505 df = 9 C. edentula, p= 0.4009, df = 9) indicate no difference between the seed oil composition of the two seed types. F i g . 11. A One-way ANOVA analysis depicting the mean and confidence interval of the proximal and distal samples of C. edentula (top) and C. maritima (bottom). distal proximal 54 To combine the proximal and distal samples in further tests, two Fisher's combined probability analyses were performed with both discrete and combined proximal and distal data from the GLM analysis. Using only the significance values from the two analyses discussed previously (the paired and unpaired analyses), both the calculated X (C. edentula=3.518 and C. maritima= 1.154) are less than the critical % (a = 0.05, df=4, x2=9.488). Furthermore, using the GLM data, the calculated %2 value for both C. edentula (8.967) and C. maritima (5.00) is again less than the critical %2, 12.53 (a = 0.05, df=6, see Appendix II for calculations). As all calculated values are less than critical values, there is no evidence of a difference between the two seed types in either species. Generalized linear model To analyze the proportion of saturated fat along the latitudinal gradient of the Pacific coast, a linear model was constructed using the Separate Slopes analysis in JMPIN software (2000). The purpose of this study is to identify relationships between several independent factors or predictor values, such as latitude and species, and the dependant value, % saturated fat. Loaded Model A loaded model where all measured factors are included, both categorical (species, seed type (either proximal or distal)) and continuous (latitude), was constructed first. This was done so as to include seed type in the combined probability analysis. The parameter values shown below (Table 2) express the estimated coefficients of the independent factors and crosses in this model. Parameter Estimates Factor Parameter Estimate t P Latitude -0.0108 -0.18 0.8622 Species -0.9758 -3.08 0.0049 Seed Type 0.5408 1.71 0.0999 Latitude *Species 0.0182 0.29 0.7705 Species *Seed Type 0.3075 0.97 0.341 Latitude *Seed Type -0.0754 -1.22 0.2337 Latitude * Seed Type * Species -0.0795 -1.29 0.21 Table 2. Parameter estimates and their significance in the loaded model. Parameter estimates are the factor's coefficient in the model equation. The t column tests the difference between the coefficient and 0; significance is reported in the p column. If the coefficient is significantly different from 0 (p < 0.05), the factor affects % saturated fat. Only species has a significant relationship with seed oil content (p= 0.0049; Fig. 12). Species/ Latitude A second model was constructed without the seed type factor. It has been shown that there is no difference between the seed oil content of proximal and distal seeds of either species; thus, the data for each seed type within species were combined. The 56 p-values for the two independent factors and their interactions are seen in Table 3. Species is again the only significant factor with p less than 0.05, confirming the difference in seed oil composition between the species (p=0.009, Table 3). The latitude and species interaction term shows no relationship between the change in proportion of saturated fat and latitude between the outcrossing and inbreeding species over the latitudinal gradient of the Pacific Coast (p= 0.6446, Table 3). Thus, while C. maritima and C. edentula differ in seed oil content, neither shows a significant latitudinal gradient. Parameter Estimates Factor Parameter Estimate t P S p e c i e s - 1 . 0 9 7 4 - 3 . 0 3 0 . 0 0 9 L a t i t u d e 0 . 0 0 1 9 0 . 0 3 0 . 9 7 8 8 S p e c i e s * L a t i t u d e 0 . 0 3 4 7 0 . 4 7 0 . 6 4 4 6 Table 3. Parameter estimates and their significance in the Species/ Latitude Model. Parameter estimates shown above correspond to the factor's coefficient in the model equation describing the proportion of saturated fat in a seed. Figure 12 below illustrates the Species * Latitude result seen in Table 3 allowing a better perspective for the difference between the two species in their change in seed oil composition by latitude. Neither species shows a significant relationship between seed oil content and latitude (C. maritima p=0.2413; C. edentula p = 0.7435). 57 A. Fig. 12 - Changes in the proportion of saturated fat over latitude in two related species of Cakile, C. edentula, an inbreeder (A) and C. maritima an outcrosser (B). The regression line (-) is compared to the mean line (-) in each species separately, and no significant difference is found between the changes in seed oil content and the mean line in either species. 40 Latitude B. The means of the two species were compared using an ANOVA. This provides a clear picture of the differences between the species' seed oil composition (Fig. 13). This analysis of variance (p = 0.0002) shows a clear difference in seed oil composition between the species, with C. maritima having a greater proportion of saturated fat than C. edentula. C. edentula C. maritima Species Fig. 13 Graph of One-way A N O V A comparing the mean proportion of saturated fat in C. maritima and C. edentula. The means and confidence intervals are shown by the green diamonds and indicate a significant difference between the proportion of saturated fat in C. maritima and C. edentula. Discussion No significant difference in the percentage of saturated fat within the seeds was found between the two types of seed, proximal and distal, within a fruit. A notable difference was found in the proportion of saturated fat in C. maritima vs. C. edentula. However, the G L M analysis revealed no difference in the proportion of saturated fat across latitude in the seeds of either species. Proximal and Distal Comparison 59 A comparison of proximal and distal seeds revealed no difference in the proportion of saturated fat between the two seed types in any of the tests performed (Fig. 11, Table 3). Thus, the data from both seed types was combined in further analyses. Other biological evidence supports this conclusion. Rodman (1974) reported in his extensive observational investigation on Cakile that the segmented fruit develop from a unilocular or single ovary. It is only after fertilization that the ovary wall grows inward and separates the two seeds to form the fruit segments. Thus it follows that the seeds are similar in their fatty acid content as they are separated only during embryonic development. Moreover, seed oil composition in Cakile is determined by the maternal genotype (Kondra and Stefansson 1970). As the distal and proximal seeds are within a single fruit their seed oil profiles are expected to be similar. Loaded Linear Model The only significant term in the loaded linear model containing all independent factors (PD, S, and L) is species (S) with p = 0.0049 (Table 2); hence, the mean percentage of saturated fat in the seeds is significantly different between the two species. This result is supported by several other analyses (Barbour and Rodman 1970, Appelqvist 1971) and indicates that seed oil composition is conserved within species and differs significantly even between closely related taxa. Species/ Latitude Model 60 In the reduced model, species remains the only significant parameter affecting seed oil composition (p = 0.0090, Table 3, Fig. 12). The C. maritima population, with a mean and confidence interval of 13.52 ± 0.30, contains significantly more saturated fat than the C. edentula, with a mean and confidence interval of 11.52 ± 0.38 (p = 0.0002, df = 32). Again, this supports the conclusion made by Rodman (1974) and others that seed oil content is highly conserved within species. A phylogeny proposed by Rodman (1974) and based on glucosinolate profiles groups species of Cakile into three biochemical "patterns": the Mediterranean, Amphi-Atlantic, and Caribbean. Cakile edentula is classified within the Amphi-Atlantic clade and shows a typical "northern" glucosinolate composition whereas C. maritima displays a Mediterranean pattern distinctive of the ancestral line within the genus. The interaction term between latitude and species determines if the proportion of saturated fat in the seed changes differentially over latitude between the species. It is this term in which we would see significance were gene flow to affect local TAG composition in the outcrossing and not the inbreeding species. This term is not significant (p = 0.6446); thus, we conclude there is no difference in the change in proportion of saturated fat over latitude between the outcrossing and inbreeding species (Table 3.). Therefore, there is no evidence of local adaptation in seed oil composition in either species. This argues against the conclusions made by Linder (2000). In Cakile, seed oils of higher latitude populations do not contain proportionally less saturated FAs than seed 61 oils from the lower latitude populations. The data presented shows no clines in percent saturated or unsaturated fat for the outcrosser or the inbreeder. There are several possible reasons why these seed oil data show no response to changes in latitude. The model system chosen for this study is a relatively recent introduction (Barbour 1970a, b), and it is possible seed oil composition has not responded significantly to the rapid northward expansion. Huey et al. (2000) have proved that it is possible for the re-establishment of a cline in adaptive characteristics in Drosophila, such as wing size, to occur in a relatively small number of generations. When compared to the ancestral "Old World" populations, an introduced population of D. subobscura in North America displays a parallel response in wing size evolution along a latitudinal gradient. It was concluded that the direction of evolution could be predictable. However, it is clear that the results of Huey et al. (2000) cannot be generalized as Pacific coast Cakile, although benefiting from a similar number of generations (-100), does not display clinal variation in seed oil content. It is possible that the strength of selection on seed oil composition in Cakile is lessened by a number of factors preventing the establishment of a detectable cline. Linder (2000) concluded through experimentation that "species with broad latitudinal or altitudinal distributions should exhibit patterns of saturated fat/ unsaturated fat ratios such that higher latitude populations have a lower proportion of saturated fat" (Linder 2000; p.442). The results of the present study do not concur. Specifically, 62 Linder's experiment documents a decrease in saturated FAs across a broad latitudinal range (Texas to Canada) within the genus Helianthus. However, populations of sunflowers in mid-continental North America experience greater temperature fluctuations than Cakile, a coastal species. Documentation of seed oil differences in C. maritima between equally abstracted temperatures (Sweden and Israel; Appelqvist 1971) suggest the need to examine a larger temperature span to detect differences of this kind on the Pacific coast. There are many different factors contributing to the melting point of seed oils in natural populations of Cakile that would alter the response in seed oil composition to changes in latitude. The anatomy of the seed provides one ready explanation as to why no response to the latitudinal gradient was detected. The Cakile fruit is capable of water migration, as a buoyant nucamentaceous (corky) layer surrounds it, providing the seed with a waterproof vessel and insulation (Barbour and Rodman 1970). This insulation could allow the lipids to maintain their fluid state or reach their fluid state rapidly at the onset of germination in cold climates, thus requiring no change in the ratio of saturated to unsaturated FAs. If this were true, it is predicted that C. edentula could show a clinal change in the thickness or density of the corky layer. Although there is no information available on the latitudinal change in the pericarp, there is comparative data showing large variations in the weight of this corky layer among species of Cakile world-wide. Observations made by Ridley (1930, as related by Rodman 1974) report that variation in the "spongy" tissue of the 63 pericarp is responsible for the variation in floating ability among species suggest diversity in this fruiting body feature. Another possible reason that the data show no temperature cline is due to lipase character. Lipase structure is variable and highly specific to the type of glycerol backbone; lipases also display stereo-specificity to their position on the glycerol molecule (Huang et al. 1988). Moreover, the lipase is under concurrent selection with seed oil content, thus resulting in a 'chicken or egg' scenario. For hydrolysis to occur, either the fatty acid content of a seed must conform to the requirements of the lipase (i.e. fluidity) or the lipase must conform to the state of the seed oil. It is possible that Cakile contains lipases capable of hydrolysis in the solid state, as is common in other species, e.g. maize; consequently, the seed oil content need not change significantly in northern populations (Huang et al. 1988). The hypothesis of lipase adaptation is supported by observations noting an extreme specificity in the elm lipase as caused by inherited structural properties of the enzyme (Huang et al. 1988). The melting point of seed oil lipids is not only dependant on chemical structure but also on size, organization and distribution of oleosomes (which carry the lipids) and glyoxysomes (which aid in the glyoxylate cycle) to provide the seed with the energy from hydrolysis (Huang et al. 1988; M.Rosenberg, personal communication). These characters are also subject to selective pressure. 64 Finally, Cakile is a species that lives in the coastal dune community and is highly influenced by the unique environmental conditions found there. Salinity can affect the germination requirements and fatty acid profile of the seed (Younis et al. 1987). Experiments performed on the effect of salinization on the hydrolysis of TAGs show that high concentrations of NaCl accelerate lipase activity, thus germination. Younis et al. (1987) provided a detailed investigation into the affects of NaCl on glycerol contents and lypolitic activity in several seeds high in fat: flax, cotton, and castor beans. They compared factors relating to enzyme activity and germination in the three species for three different concentrations of NaCl. Results show that NaCl affects germination, water accumulation, respiration, oil content, and lipase activity. Interestingly, oil content decreased faster in the seeds treated with salt, and respiration rates significantly increased. This provides evidence that salt causes an increase in the oxidation of metabolites (Younis et al. 1987). Moreover, salt caused a rapid decrease in oil content through the course of germination and an increase in lipase activity at low NaCl. Thus, it is possible that the seaside environment actually accelerates the degradation of TAGs and aids in germination. Apart from factors affecting the melting point of TAGs, there are other possible reasons why the results do not support the hypothesis that species at higher latitudes will contain proportionally more unsaturated FAs than species at lower latitudes. 65 Insect herbivory increases at lower latitudes (Wilf and Labandeira 1999); consequently, there is a strong selection for acrid oils at lower latitudes to prevent such predation (Rodman 1980). Glucosinolates, glucose derivatives containing nitrogen and/or sulfur, and FAs acting as deterrents, are predominantly unsaturated (Rodman 1980). The lower latitude plants would contain higher proportions of unsaturated FA to deter pests. In higher latitude plants the importance of this ratio of saturated to unsaturated FA in combating the effect of colder temperatures on germination, could also result in higher unsaturated FAs. Regardless of the agent of selection, comparable levels of unsaturated FAs could conceivably be found in both northern and southern populations. Conclusion The data show no trend in the proportion of saturated fat across the Pacific coast latitudinal gradient. As there are many factors which influence seed oil content, it is proposed that the extent to which the hypothesis that higher latitude plants will exhibit lower proportions of saturated FAs depends largely on the range and character of the species. Moreover, Cakile, a relatively recent introduction to the Pacific coast ecosystem, does not support the predictions of the theory of range limitation by gene flow. There is no evidence of a difference in adaptation between the central and marginal populations of C. maritima, nor is there evidence of local adaptation in C. edentula. 66 Proportion of saturated fat in C. maritima Location % Saturated Fat Author Lund, Sweden 6.90% Appelqvist, L.A., 1971 Lund, Sweden 7.00% Appelqvist, L.A., 1971 Lund, Sweden 8.10% Appelqvist, L.A., 1971 Svalov, Sweden 7.50% Appelqvist, L.A., 1971 Israel 12.00% Miller, R., 1965 Table 4. Trend in proportion of saturated fat between two locations, Sweden and Israel. Sweden and Israel have distinctly different temperature regimes and, as such, show a detectable difference in the proportion of saturated fat in C. maritima seeds Species spanning extreme differences in temperature, like those found between Sweden and Israel in Cakile, might display fitness patterns predicted by the theory (Table 4); however, the current analysis does not display this difference. 67 Conclusion Gene flow occurs both randomly and asymmetrically due to density differences between central and peripheral populations. It can moderate an evolutionary response to selection in small peripheral populations by introducing a deluge of foreign genes poorly adapted to the border region. Thus, a species whose range is constrained by gene flow should show a gradual decline in density towards the margins where the integrative effects of gene flow counter selection. The population distribution of C. maritima displays this decreasing pattern of abundance. However, the density distribution of C. edentula does not coincide with the predicted pattern. And, there is evidence that the gradual decline in abundance seen in C. maritima is not exclusively a result of range limitation by gene flow. The fitness of border populations should reflect the impact of gene flow from the central regions, maladapted to marginal conditions. Therefore, a species whose range is constrained by gene flow should show a diluted response to local selection at the borders and little differentiation in adaptive characteristics through out the range. In contrast, a species unfettered by gene flow would display variation in response to local environmental selection. Seed oil composition has been shown to be adaptive and respond to changes in temperature (Linder 2000). This thesis predicted Cakile maritima, an outcrosser, to show little difference in seed oil composition between central and marginal populations. However, C. edentula, an inbreeder, should display marked differences in seed oil composition along the Pacific coast environmental gradient. No such pattern was detected in C. edentula, disallowing a comparison of the adaptive response in seed oil content to marginal conditions between an outcrossing and inbreeding species. As many possible factors affected this adaptive response, further work is needed to clarify the relationship between gene flow and species range limitation in natural Cakile populations. 68 Literature Cited Agnew, A. D. Q. 1968. Variation and selection in an isolated series of populations of Lysimachia volkensii Engl. Evolution 22: 228-236. Appelqvist, L. A. 1971. Lipids in Cruciferea: VIII. The fatty acid composition of some wild or partially domesticated species. Journal of the American Oil Chemists' Society 48:740-744. Antonovics, J. 1976a. The input from population genetics: "The new ecological genetics". Systematic Botany 1:233-245. Antonovics, J. 1976b. The nature of limits to natural selection. Annals of the Missouri Botanical Garden 63:224-247. Ayala, F. J., J. R. Powell, and M. L. Tracey. 1972a. Enzyme variability in the Drosophila willistonii group. V. Genie variation in natural populations of Drosophila equinoxialis. Genetical Research 20:19-42. Ayala, F. J., J. R. Powell, M. L. Tracey, C. A. Mourao, and S. Perez-Salas. 1972b. Enzyme variability in the Drosophila wUlistoni group. IV. Genie variation in natural populations of Drosophila wUlistoni. Genetics 70:113-139. Barbour, M. G. 1970a. Germination and early growth of the strand plant Cakile maritima. Bulletin of the Torrey Botanical Club 97:13-22. Barbour, M. G. 1970b. Seedling ecology of Cakile maritima along the California coast. Bulletin of the Torrey Botanical Club 97:280-289. Barbour, M. G., and J. E. Rodman. 1970. Saga of the west coast sea-rockets: Cakile edentula ssp. californica and C. maritima. Rhodora 72:370-386. Blows, M. W., and A. A. Hoffman. 1993. The genetics of central and marginal populations of Drosophila serrata I. Genetic variation of stress resistance at species borders. Evolution 47:1255-1270. Boyd, R. S., and M. G. Barbour. 1993. Replacement of Cakile edentula by C. maritima in the strand habitat of California. American Midland Naturalist 130:209-228. 69 Brown, J. 1984. On the relationship between abundance and distribution of species. American Naturalist 124:255-279. Brown, J. H. et. al. 1995. Spatial variation in abundance. Ecology 76:2028-2043. Brown, W., Jr. 1957. Centrifugal speciation. Quarterly Review of Biology 32:247-277. Brussard, P. 1984. Geographic patterns and environmental gradients: The central-marginal model in Drosophila revisited. Annual Review of Ecological Systematics 15:25-64. Camin, J. H., and P. R. Ehrlich. 1958. Natural selection in water snakes (Natrix sipedon L.) on islands in Lake Erie. Evolution 12:504-511. Case, T. J., and M. L. Taper. 2000. Interspecific competition, environmental gradients, gene flow, and the coevolution of species' borders. American Naturalist 155:583-605. Carson, H. L. 1956. Marginal homozygosity for gene arrangement in Drosophila robusta Science 123:630-631. Carson, H. L. 1958. Increase in fitness in experimental populations resulting from heterosis. Proceedings of the National Academy of Sciences of the United States of America 44:1136-1141. da Cunha, A. B., H. Burla, and T. Dobzhansky. 1950. Adaptive chromosomal polymorphism in Drosophila wUlistoni. Evolution 4:212-235. da Cunha, A. B., and T. Dobzhansky. 1954. A further study of chromosomal polymorphism in Drosophila wUlistoni in relation to the environment. Evolution 8:119-134. da Cunha, A. B., T. Dobzhansky, O. Pavlovsky, and B. Spassky. 1959. Genetics of natural populations XXVIII. Supplementary data on the chromosomal polymorphism in Drosophila wUlistoni in relation to the environment Evolution 13:389-404. 70 Dobzhansky, T. 1956. Genetics of natural populations XXV. Genetic changes in populations of Drosophila pseudoobscura and Drosophila persimilis in some localities in California. Evolution 10:82-92. Dobzhansky, T. 1957. Genetics of natural populations XXVI. Chromosomal variation in island and continental populations of Drosophila willistoni from Central America and the West Indies. Evolution 11:208-293. Dobzhansky, T., A. S. Hunter, O. Pavlovsky, B. Spassky, and B. Wallace. 1963. Genetics of natural populations XXXI. Genetics of an isolated marginal population of Drosophila pseudoobscura. Genetics 48:91-103. Donohue, K. 1997. Seed dispersal in Cakile edentula var. lacustris: decoupling the fitness effects of density and distance from the home site. Oecologia 110:520-527. Donohue, K. 1998. Maternal determinants of seed dispersal in Cakile edentula: Fruit, plant, and site traits. Ecology 78:2271-2788. Dudley, S. A. 1996. The response to differing selection on plant physiological traits: Evidence for local adaptation. Evolution 50:103-110. Emlen, J. T. 1978. Density anomalies and regulatory mechanisms in land bird populations on the Florida peninsula. American Naturalist 112:265-286. Endler, J. A. 1973. Gene flow and population differentiation. Science 179:243-250 Esau, K. 1977. Plant Anatomy, 2 edition. John Wiley & Sons, Inc. Farris, M. A., and B. A. Schaal. 1983. Morphological and genetic variation in ecologically central and marginal populations of Rumex acetosella L.(Polygonaceae). American Journal of Botany. 70:246-255. Garcia-Ramos, G. a. M. K. 1997. Genetic models of adaptation and gene flow in peripheral populations. Evolution 51:21-28. Grant, M. C. 1974. Genetic properties of biologically marginal populations of Anthoxanthum odoratum. PhD. Duke University. 71 Grant, M. C., and J. Antonovics. 1978. Biology of ecologically marginal populations of Anthoxanthum odoratum. I. Phenetics and dynamics. Evolution 32:822-838. Haldane, J. B. S. 1930. A note on Fisher's theory of the origin of dominance and on a correlation between dominance and linkage. American Naturalist 64:87-90. Hamrick, J. L., and M. J. W. Godt. 1996. Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society in London: Biological Sciences 351:1291-1298. Huang, A. H. C , Y. Lin, and S. Wang. 1988. Characteristics and biosynthesis of seed lipases in maize and other plant species. Journal of the American Oil Chemists' Society 65:897-899. Huey, R. B., G. W. Gilchrist, M.L.Carlson, D. Berrigan, and L.Serra. 2000. Rapid evolution of a geographic cline in size in an introduced fly. Science 287:308-309. Hutchins, L. W. 1947. The bases for temperature zonation in geographical distribution. Ecological Monographs 17:325-335. JMPIN 4.0.3 (Academic) software. Distributed by SAS Institute Inc. Copyright © 1989 -2000. Keddy, P. A. 1981. Experimental demography of the sand dune annual, Cakde edentula, growing along an environmental gradient in Nova Scotia. Journal of Ecology 69:615-630. Kirkpatrick, M., and N. H. Barton. 1997. Evolution of a species' range. The American Naturalist 150:1-23. Kondra, Z. P., and B. R. Stefansson. 1970. Inheritance of the major glucosinolates of rapeseed (Brassica napus) meal. Canadian Journal of Plant Science 50:643-647. Kunst, L., D. C. Taylor, and E. W. Underhill. 1992. Fatty acid elongation in developing seeds of Arabidopsis thaliana. Plant Physiological Biochemistry 30:425-434. Knight, G. R. 1961. Structural polymorphism in Drosophila subobscura Coll. from various localities in Scotland. Genetical Research 2:1-9. ^ 72 Lakovaara, S., and A. Saura. 1971a. Genetic variation in natural populations of Drosophila obscura. Genetics 69:377-384. Lakovaara, S., and A. Saura. 1971b. Genie variation in marginal populations of Drosophila subobscura. Hereditas 69: 77-82 Levin, D. A. 1974. The oil content of seeds: An ecological perspective. American Naturalist 108:193-206. Levin, D. A. 1978. Genetic variation in annual Phlox: Self-compatible versus Self-incompatible species. Evolution 32:245-263. Levin, D. A., and K. Clay. 1984. Dynamics of synthetic Phlox drummondii populations at the species margin. American Journal of Botany. 71:1040-1050. Linder, C. R. 2000. Adaptive evolution of seed oils in plants: Accounting for the biogeographic distribution of saturated and unsaturated fatty acids in seed oils. American Naturalist 156:442-458. Linhart, Y. R. 1974. Intra-population differentiation in annual plants I. Veronica perigrina L. raised under non-competitive conditions. Evolution 28:232-243. Lloyd, D. G. 1965. Evolution of self-compatibility and racial differentiation in Leavenworthia (Cruciferae). Contributions of the Gray Herbarium 195:3-134. Mayr, E. 1963. Populations, Species, and Evolution. Harvard University Press, Cambridge, Mass. McMurry, J. 1994. Fundamentals of Organic Chemistry, 3 edition. Brooks/ Cole Publishing Co., Belmont, Ca. Mikolajczak, K. L., T. K. Miwa, F. R. Earle, and I. A. Wolff. 1961. Search for new industrial oils V. Oils of Cruciferae. Journal of the American Oil Chemists' Society 38:678-681. Miller, R. W., F. R. Earle, and I. A. Wolff. 1965. Search for new industrial oils XIII. Oils from 102 species of Cruciferae. Journal of the American Oil Chemists' Society 42:817-821. 73 Nei, M. 1973. Analysis of gene diversity in subdivided populations. Proceedings of the National Science Academy 70:3321-3323. Parsons, P. A. 1991. Evolutionary rates: Stress and species boundaries. Annual Review of Ecology and Systematics 22:1-18. Payne, A. M., and M. A. Maun. 1981. Dispersal and floating ability of dimorphic fruit segments of Cakile edentula var. lacustris. Canadian Journal of Botany 59:2595-2602. Prakash, S. 1973. Patterns of gene variation in central and marginal populations of Drosophila robusta. Genetics 75:347-369. Prakash, S., R. C. Lewontin, and J. L. Hubby. 1969. A molecular approach to the study of genie heterozygosity in natural populations IV. Patterns of genie variation in central, marginal and isolated populations of Drosophila pseudoobscura. Genetics 61:841-858. Prevosti, A. 1964. Chromosomal polymorphism in Drosophila subobscura populations from Barcelona (Spain). Genetical Research 5:27-38. Richards, C. 2000. Genetic and demographic influences on population persistence: Gene flow and genetic rescue in Silene alba. Pages 271-291 in A. a. C. Young, GM, editor. Genetics, Demography and Viability of Fragmented Populations. Cambridge University Press, Cambridge. Rodman, J. E. 1974. Systematics and evolution of the genus Cakile (Cruciferae). Contributions of the Gray Herbarium at Harvard University 205:3-146. Rodman, J. E. 1980. Population variation and hybridization in Sea-rockets (Cakile, Cruciferae): Seed glucosinolate characters. American Journal of Botany. 67:1145-1159. Rodman, J. E. 1986. Introduction, establishment and replacement of sea-rockets (Cakile, Cruciferae) in Australia. Journal of Biogeography 13:159-171. Saura, A., S. Lakovara, J. Lokki, and P. Lankinen. 1973. Genie variation in central and marginal populations of Drosophila subobscura. Hereditas 75:33-46. 74 Shumaker, K. M., and G. B. Babble. 1980. Patterns of allozymic similarity in ecologically central and marginal populations of Hordeum jubatum in Utah. Evolution 34:110-116. Silander, J. A. J. 1985. The genetic basis of ecological amplitude of Spartina patens. II. Variance and correlation analysis. Evolution 39:1034-1052. Slatkin, M. 1987. Gene flow and the geographic structure of natural populations. Science 236:787-792. Stanton, M. L., C. Galen, and J. Shore. 1997. Population structure along a steep environmental gradient: Consequences of flowering time and habitat variation in the snow buttercup, Ranunculus adoneus. Evolution 51:79-94. Stearns, S. C , and R. D. Sage. 1980. Maladaptation in a marginal population of the Mosquito Fish, Gambusia affinis. Evolution 34:65-75. Stewart, C. N. J., and E. T. Nilsen. 1995. Phenotypic plasticity and genetic variation of Vaccinium macrocarpon (American cranberry) II. Reaction norms and spatial clonal patterns in two marginal populations. International Journal of Plant Sciences 156:698-708. Soule, M. 1973. The epistasis cycle: A theory of marginal populations. Annual Review of Ecology and Systematics. 4:165-187. Tigerstedt, P. M. A. 1973. Studies on isozyme variation in marginal and central populations of Picea abies. Hereditas 75:45-60. Townsend, J. I. 1952. Genetics of marginal populations of Drosophila willistonii. Evolution 6:428 - 442. Ward, R. D. 1920. The essential characteristics of United States climates. Scientific Monthly 11:555 -568. Watkinson, A. R. 1985. On the abundance of plants along an environmental gradient. Journal of Ecology 73:569-578. 75 White, M.J.D. 1951. Structural heterozygosity in natural populations of the grasshopper Trimerotropis sparsa. Evolution 5: 376-394. Whittaker, R. H. 1956. Vegetation of the Great Smokey Mountains. Ecological Monographs 26:1-80. Wilf, P., and C. C. Labandeira. 1999. Response of plant-insect associations to Paleocene Eocene warming. Ecology 284:2153-2156. Wilson, J. B., Y. Ronghua, A. F. Mark, and A. D. Q. Agnew. 1991. A test of the low marginal variance (LMV) theory, in Leptospermum scoparium (Myrtaceae). Evolution 45:780-784. Younis, M. E., M. N. A. Hasaneen, and M. M. Nemet-Alla. 1987. Plant growth, metabolism and adaptation in relation to stress conditions IV. Effects of salinity on certain factors associated with the germination of three different seeds high in fats. Annals of Botany 60:337-344. 76 Appendix I Summary of Site and Beach locations for 2001 field season Site Number Site Beach Latitude 1 San Diego, CA Border Field State Park 32.57 Silver Strand State Beach 32.62 Coronado Beach 32.68 2 Santa Barbara, CA Jalama State Park 34.51 Lompoc State Beach 34.67 Nippomo Dunes State Park 34.95 3 Monterrey, CA Pfeiffer State Beach 36.24 Carmel City Beach 36.55 Sand City Beach 36.62 4 Point Reyes, CA Limantour State Beach 38.02 South Beach 38.07 McClures Beach 38.19 5 Eureka, CA Samoa Dunes past Fairhaven 40.77 Sonora State Beach 40.82 Manila State Beach 40.85 6 Bandon, OR Bandon State Beach 43.07 Face Rock Beach 43.09 Bandon City Beach 43.12 7 Newport, OR Newport Beach 44.63 Agate Beach 44.66 Devil's Punch Bowl 44.74 8 Long Beach, WA Chatauqua Lodge Beach 46.36 Seaside City Beach 46.50 Pacific Pins Beach 46.56 9 Sooke, BC Wier Beach 48.35 Wiffen Spitt 48.37 Esquimalt Lagoon 48.43 10 Toffino, BC Floriencia Provincial Beach 49.00 C Beach 49.03 Long Beach 49.07 11 Cape Scott, BC San Josef Beach 50.67 Guise Bay 50.77 Experiment Blight 50.78 13 Queen Charlotte Island, BC Skidigate Beach 53.38 Tellel Beach 53.58 Masset Cemetery 54.03 Note: At each beach three populations were measured for density and fecundity. Populations are distinct and identifiable; thus, density measurements were simply counts of a single population's size. Populations were chosen randomly after 0.5 km from the previous population was reached. 77 Appendix II Fisher's Combined Probability Analysis n1 n2 df ts P C. edentula unpaired 10 10 18 1.109 0.282 C. edentula paired 10 10 9 0.8816 0.4009 GLM 0.0999 In(p1) -1.265848 In(p2) -0.914043 In(p3) -2.303585 SUM -4.483477 *-2 8.966954 The critical %2 for a = 0.05 and 6 df (including the GLM analysis) is 12.53. This test fails to reject the null because the calculated %2 (8.96695) is less than the critical %2 (12.53). Interpretation: Null hypothesis: The p values are from the same %2 distribution, thus the seed types are not different. n1 n2 df ts P C. maritima 9 9 16 0.174 0.8638 unpaired C. maritima 10 10 9 -0.0638 0.9505 paired GLM 0.0999 In(p1) -0.146414 In(p2) -0.050767 In(p3) -2.303585 Sum -2.500766 *-2 5.001533 The critical %2 for a = 0.05 and 6 df is 12.53 Fail to reject the null because the calculated %2 (5.001) is less than the critical x2 (12.53) Interpretation: Null hypothesis: The p values are from the same %2 distribution, thus are not different. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items