Faculty Research: 4) Genetic and environmental influences on life history, floral traits, and sex allocation in Raphanus sativus (Brassicaceae; wild radish): the stability of genetic parameters across environments.

The measurement of the heritability of fitness-related traits and the estimation of the quantitative relationship between phenotype and fitness are the first steps toward determining the role of natural selection on these traits in a given population. In addition, it's necessary to discover the degree to which heritable variation in individual fitness and fitness functions are constant in time and in space. This is because evolution by natural selection is a cumulative process, and gradual evolutionary change by natural selection will only occur if the direction of selection is consistent over many generations.

Experimental and natural wild radish populations (R. sativus L.) have been used as a model system by plant evolutionary ecologists, geneticists and pollination biologists during the last decade to evaluate the strength of natural selection on fitness-related traits under natural conditions. Over the last ten years, I've been using this species to evaluate a suite of questions concerning the nature of natural selection on a variety of life history and reproductive traits. These questions all examine whether components of natural selection remain consistent from generation to generation.

1. In natural populations, is the expression of genetic variation in fitness-related traits environment-specific? Is the relationship between phenotype and fitness environment-specific? If so, do Genotype x Environment interactions contribute to the maintenance of quantitative genetic variation in these traits? To evaluate this set of questions, I have replicated seed families of known genetic background under different population densities to observe environmental effects on the magnitude of phenotypic and genetic variation in life history and reproductive traits. A primary objective of this work has been to determine the extent to which the heritability of plant traits that affect plant fitness (survival and reproduction) are constant across environments. The results of this work are summarized in several papers (Mazer and Schick, 1991a, 1991b; Mazer, 1992; Mazer and Wolfe, 1993; Mazer and Wolfe, 1998).

   One striking result is that the magnitude of the genetic component to phenotypic variation in almost all of the characters investigated depends strongly on the density at which seeds are planted. If this is a general property of natural populations, our ability to predict the rate of evolution by natural selection in the field will depend on critically on the local environmental conditions sampled and on the range and frequency of environments occupied by a given population.

   This work has also revealed that gender allocation in wild radish is strongly modified by environmental conditions. In this species, components of female reproduction (the number of ovules per flower, the proportion of ovules that develop into seeds, the number of seeds per fruit, and mean individual seed mass) are much more sensitive to population density than is pollen production per flower or pollen grain size. Female components of reproduction per flower and fruit decline drastically with increasing population density, while pollen production per flower and pollen size remain relatively buffered against environmental change. This result supports the predictions of many models of sex allocation theory, which predict that, as plant size or resource status increases, so should the proportional allocation of resources to female function.

2. Are heritable characters genetically linked, and, if so, are correlations between characters consistent across environments? Moreover, do such character correlations represent a "constraint" on the direction or rate of evolution by natural selection or are they the adaptive outcome of selection?

   For several reasons, evolutionists are interested in the degree to which characters influencing reproductive success are genetically linked. First, if traits are strongly correlated due to pleiotropy or linkage, they will not evolve independently; selection on one character will influence the evolution of correlated characters. Strong genetic correlations between characters within populations may also determine or limit the nature of phenotypic divergence among populations during the speciation process. For example, a significant negative correlation between ovule number per flower and flower diameter among genotypes of Primula stricta appears to be mirrored among Primula species means (Mazer and Hultgard, 1993), potentially reflecting such a genetic constraint acting during the speciation process.

   A second motive for measuring genetic covariation among reproductive characters is to evaluate the proposal that genes with pleiotropic effects can (under certain conditions) maintain genetic variation in fitness-related traits. Moreover, the detection of particular genetic correlations among traits within species - for example, the commonly observed negative correlation between pollen size and pollen production per flower - suggest the presence of non-random associations of trait values that result in alternate phenotypes of equal long-term fitness. In these cases, the presence and persistence within populations of alternate trait combinations suggest that natural selection has operated to maintain these genetic correlations.

   Finally, while theoretical work in resource allocation and quantitative genetics predicts that negative genetic correlations should be common in nature, the results of available empirical studies are sparse and often equivocal. Our ongoing work with wild radish (Raphanus sativus), sand-spurrey (Spergularia marina), and Clarkia are helping to fill this gap.

3. Are environmentally-induced changes in phenotype adaptive? Is phenotypic plasticity the adaptive result of natural selection acting in different microenvironments, or is it simply the inevitable result of being raised in stressful vs. favorable conditions?

   Plant ecologists have long observed that plants raised in different environmental conditions exhibit different phenotypes for a variety of life history and reproductive traits. These observations naturally lead one to ask whether environment-specific phenotypes are themselves adaptive, or do they reflect direct environmental effects on phenotype that overwhelm the direction of selection? I have recently completed the analysis of a common garden experiment on Raphanus sativus designed to ask whether the phenotypic changes observed as population density increases reflect the outcome of density-specific selection or, rather, a direct environmental effect (this research was presented at the International Botanical Congress meeting in St. Louis, Missouri, in August, 1999). I used a multiple regression approach (phenotypic selection gradient analysis) to detect in each density the strength and direction of selection on each of 15 life history and reproductive traits independently of all other measured traits. I found that, for most traits, the phenotype observed in each population density was not what one would predict based on the direction and strength of selection on the trait in each density. Therefore, for most traits, it appears that phenotypic change in response to increasing density is a direct environmental effect that overwhelms the ability of selection to cause the evolution of the "optimum" phenotype for each density.

   Ongoing and Future Work: In an effort to address the question, "Is plasticity adaptive?", I am also determining whether the magnitude of phenotypic plasticity expressed in these traits influences expected fitness over a range of population densities. For example, preliminary results suggest that maternal families that respond to high population densities by producing relatively small seeds actually have higher fitness - across densities - than maternal families that do not exhibit an environmental effect on seed size (Mazer and Wolfe, 1999). This is because the former have lower seed abortion rates and ultimately produce higher numbers of viable seeds and higher total seed mass per fruit than the latter. I am continuing to analyze these data to ask whether phenotypic plasticity can be interpreted as a trait that is itself under selection in wild radish.

Current Research Projects

1. Testing the assumptions of sex allocation theory: quantitative genetic variation and covariation among floral traits in the Sand-Spurrey, Spergularia marina (Caryophyllaceae)
2. The evolution of gender-related traits in species with different mating systems: a quantitative genetic comparison of selfing and outcrossing species of Clarkia (farewell-to-spring: Onagraceae)
3. The Evolutionary Significance of Variation in Traits Subject to Ontogenetic Change and Maternal Environmental Effects
4. Genetic and environmental influences on life history, floral traits, and sex allocation in Raphanus sativus (Brassicaceae; wild radish): the stability of genetic parameters across environments.
5. The Detection of the Long-term Outcome of Natural Selection and the Ecological Sorting of Species Among Habitats: Comparative Studies of Plant Reproductive Characters
6. Ecological adaptation, gene flow, and the potential for hybrid breakdown in restoration projects.

Susan Mazer | Research | Publications | Curriculum Vitae

<< Return to Faculty Page