Research Projects Pamela K. Diggle Overview Plasticity of morphological characters is an inherently developmental phenomenon. Nevertheless, the role of development in the evolution and expression of phenotypic plasticity has been virtually ignored. My research examines the role of development in the evolution and expression of phenotypic plasticity in plants. The central goal of this research program has been to examine the combined effects of genotype, environment, and ontogenetic history on the development and expression of plant phenotypes. Specific projects are described briefly below.   Phenotypic Plasticity and Architecture In previous research on the andromonoecious species Solanum hirtum (Solanaceae), I demonstrated that sex-expression, the proportions of hermaphrodite and staminate flowers borne by an individual, is phenotypically plastic and that there is genotypic variation for phenotypic plasticity. During the period of the NYI Award I reconsidered these data and developed my ideas and understanding of the importance of plant architecture and indeterminate development for understanding the expression and the evolution of phenotypic plasticity in plants. In S. hirtum, although the mechanism of plasticity of sex expression resides at the level of gynoecial development of individual flowers, plasticity of sex expression cannot be fully understood by examining flower development in isolation. The particular phenotypic expression of a flower (and by extension, any plant part) depends jointly on two factors: the developmental history (which will reflect the history of both the external and internal environment) of the individual that bears it, and the position of that plant part within the architectural ground plan of the plant. I introduced the terms "ontogenetic contingency" to describe the joint effects of these two factors on phenotypes and "architectural effects" to describe intra-individual phenotypic variation that can be ascribed solely to differences in position (Diggle 1994, 1995, 1997b). Recognition of the critical importance of architectural effects in determining the expression of plasticity in S. hirtum inspired a search for additional evidence of the effects of architecture on reproductive phenotypes in plants. My efforts to examine the broader implications of the effects of architecture and plasticity on plant phenotypes resulted in an invited article published in the Annual Review of Ecology and Systematics (Diggle 1995). In this paper, I circumscribed the concept of architectural effects in greater detail and showed that many of the patterns of resource allocation that are generally attributed (in the literature) to plastic responses to resource competition, may be due, in part or whole, to architectural effects. For example, the commonly observed proximal to distal decrease within inflorescences of fruit and/or seed maturation per flower has frequently been attributed to competition among developing fruits for resources. However, the observed variation can also be due to architecture--that is, to sources of variation inherent in plant axes. Most significantly, I demonstrated that the effects of architecture are separable experimentally from the effects of differential resource allocation, and that careful experimental analysis of these two factors is critical to understanding the physiological, developmental, and evolutionary controls of fruit and seed production in flowering plants. The hypothesis of ontogenetic contingency emphasizes that both architecture and developmental history determine phenotypes. In any analysis of phenotypic plasticity in metameric organisms, it is critical to separate these two factors; only developmental history incorporates plastic responses to the environment. I have formalized an experimental and analytical design to separate plasticity from architectural effects (Diggle 1997b). This design has been used in an analysis of floral characters in S. hirtum, two ecotypes of Arabidopsis thaliana (Diggle 1997b), and Epilobium clavatum (unpublished). I show that architectural effects are extremely common in flowering plants and that architectural effects can mimic, mask, or even cause misinterpretation of plastic variation in floral phenotypes. Finally, I have begun to consider the importance of architecture in the evolution of such life history features as the timing of reproduction. In a literature review (Diggle 1999), I show that architectural features of plants can constrain options for variation in the onset of flower production.   Origin and Evolutionary Diversification of Andromonoecy [I]t wd be great and curious blunder in dame nature.”  So wrote Charles Darwin in a letter (May 22, 1860) to Joseph Hooker on the phenomenon of andromonoecy, a plant sexual system in which individuals bear both hermaphroditic and staminate flowers.  Nearly 150 years ago, Darwin was frustrated in his attempts to understand the selective advantages conferred by andromonoecy.  Remarkably, since Darwin’s time, efforts to understand andromonoecy within an evolutionary context have been largely inconclusive. Andromonoecy has evolved independently from hermaphroditism numerous times over the course of flowering plant history.  My research focuses on the evolutionary origin and subsequent diversification of andromonoecy within the genus Solanum.  In contrast to previous efforts over the last century, I approach the evolution of andromonoecy from a comparative developmental and morphological perspective.  In essence, I seek to understand the historical patterns of morphological and developmental transitions that underlie the origin and diversification of this sexual system.  Such an approach has, for the first time, gained meaningful insight into the evolutionary processes associated with the origin and subsequent diversification of andromonoecy. The genus Solanum was chosen for this study for four important reasons.  First, well supported phylogenetic analyses of the subgenera and sections of the genus provided the context for this comparative analysis (e.g., Olmstead and Palmer 1997; Bruneau et al. 1995).  Second, an origin of andromonoecy can be traced to the common ancestor of subgenus Leptostemonum, and the most closely related outgroups are all hermaphroditic.  Third, within the sections of subgenus Leptostemonum there is a tremendous diversity of expression of andromonoecy (the proportions of hermaphroditic and staminate flowers).  Fourth, andromonoecy has been considered to be a form of phenotypic plasticity and this had been confirmed for one species of Solanum (Diggle 1993).  Thus, comparison of andromonoecious species of Leptostemonum with closely related hermaphroditic taxa provided the means to analyze the developmental underpinnings of the origin of andromonoecy, while comparison of andromonoecious taxa within Leptostemonum identified the evolutionary developmental basis for diversification of the expression of andromonoecy.  Finally, the relationship of this reproductive strategy to one of the most important aspects of plant responses to the environment, phenotypic plasticity, could be evaluated (Diggle 2002). Analyses of the diversification of andromonoecy were focused on sections within subgenus Leptostemonum: Lasiocarpa, a monophyletic group of 12 species and a section Acanthophora (as curcuscribed by Levin et al. 2006).   Andromonoecy in Lasiocarpa and Acanthophora is expressed within a stereotypical architecture or bauplan.  Staminate flowers, when produced, are borne in a predictable pattern that encompasses both within and among inflorescence variation. Within each inflorescence, hermaphroditic flowers are borne basally and staminate flowers distally.  Among inflorescences, the transition point (number of floral nodes) from hermaphroditic to staminate flowers occurs at earlier (more basal) positions within later (more distal) inflorescences.  In all species of studied (Miller and Diggle 2003 and unpublished), this intra- and inter-inflorescence variation results in a branch-level pattern of increasing proportions of staminate flowers in later produced (distal) inflorescences.  Within the context of this highly invariant architectural pattern of staminate flower production, the strength of andromonoecy (the mean percentage of staminate flowers produced per inflorescence) varies considerably among species, from essentially zero (0.2%) to nearly 70%.  A single morphological variable explains the majority of this variation: quantitative differences in the transition from hermaphroditic to staminate flower production within inflorescences (Miller and Diggle 2003). The more strongly andromonoecious species produce staminate flowers on earlier inflorescences and, within each inflorescence, staminate flowers occur at more basal positions.  We concluded that within inflorescences of all members of the clade there is a gradient of developmental potential that determines the type of flower that will be produced.  Diversification of andromonoecy involves proximal or distal shifts of this gradient (Miller and Diggle 2003 and in prep). Plasticity of allocation to male and female function is generally thought to be an important component of the evolution of andromonoecy.  The production of staminate flowers, however is not a phenotypically plastic response in all species of section Lasiocarpa.  The weakly andromonoecious species are, indeed, phenotypically plastic; the proportion of staminate flowers increases significantly for plants bearing fruit.  In contrast, the strongly andromonoecious species are not plastic; staminate flower production is a fixed aspect of the phenotype and is not altered by fruit set (Miller and Diggle 2003).  Reference to a phylogeny of the section (Bohs in press) shows that weak andromonoecy and plasticity are plesiomorphic and that strong andromonoecy, in association with the loss of plasticity, is derived (ms. in prep).  Ironically, although plasticity is associated with the origin of andromonoecy (see below) diversification and enhancement of andromonoecy involve loss of plasticity.  These detailed analyses of andromonoecy demonstrate that two critical morphological and developmental features underlie the diverse expression of andromonoecy within Leptostemonum: (1) developmental plasticity and (2) gradients of developmental potential within and among inflorescences that I have termed “architectural effects” (Diggle 1999, 2003). These same key features were examined in three hermaphroditic species of Solanum in order to reconstruct the features that were present in the common ancestor of the andromonoecious clade (greenhouse experiments and data analysis completed June 2003).  Each of the three species show inter- and intra-inflorescence patterns of architectural variation in flower function that are qualitatively similar to those described for the andromonoecious species of subgenus Leptostemonum.  Although all flowers are morphologically hermaphroditic, plants respond to the presence of developing fruits by producing functionally staminate flowers in distal positions.  That is, within the inflorescences of fruit-bearing plants, basal flowers set fruit whereas distal flowers do not.  Moreover, the proportion of distal flowers that do not set fruit (that are functionally staminate) increases within successive inflorescences. Thus, these hermaphroditic species of Solanum display the same functional and architectural level responses to the presence of developing fruit as the plastic, weakly andromonoecious (and for these characters, plesiomorphic) species within Lasiocarpa. Based on these developmental and morphological comparisons of andromonoecious and hermaphroditic Solanum species, we can, for the first time, reconstruct the evolutionary events associated with an origin and diversification of andromonoecy. Andromonoecy arose as a phenotypically plastic, functional response to the resource demands of developing fruit.  As a result of existing developmental gradients within inflorescences of the ancestral hermaphrodite, this plastic response was specific to ovary function of distal flowers.  In the common ancestor of subgenus Leptostemonum, the developmental innovation of a mechanism for pre-anthesis termination of gynoecial maturation in distal flowers resulted in the evolution of morphologically staminate flowers.  This mechanism drew upon a gradient of developmental potential already present in the hermaphroditic ancestor.  The production of morphologically staminate flowers in subgenus Leptsostemonum was initially a plastic response to the presence of developing fruit. Subsequently, this same pattern of staminate flower production in distal positions became a fixed aspect of the phenotype inmore derived taxa within section Lasiocarpa (ms. in prep).   Architectural Effects and the Analysis of Sexual Dimorphism Among taxa with unisexual flowers, sexual dimorphism of flowers is common (Darwin, 1877; Lloyd and Webb, 1977; Delph, 1996; Delph et al., 1996; Eckhart, 1999).  In particular, sexual dimorphism of corolla size is well documented and has played an important role both in the development of theory underlying sex allocation models and in stimulating empirical research in plant evolutionary ecology.  The widespread distribution of floral size dimorphism (occurring in 85% of species with unisexual flowers; Delph et al., 1996) may suggest common underlying explanations for the evolution of sexual size dimorphism, and indeed, a variety of hypotheses have been proposed (summarized in Delph et al., 1996). These range from “nonfunctional” hypotheses that postulate developmental correlations between stamens and corolla (Darwin, 1877; Plack, 1957, 1958) to “functional” hypotheses that emphasize presumptive biological roles for the perianth in protection (Bawa and Opler, 1975) and pollinator attraction (Bell, 1985) or deal with optimizing resource allocation for male or female function (Eckhart, 1992; Ashman, 1994; Costich and Meagher, 2001; Miller and Venable, 2003). I have explored an additional hypothesis, that architectural variation may underlie apparent sexual dimorphism.  In andromonoecious Solanum, hermaphroditic flowers are significantly larger than staminate flowers for all features measured.  Thus, flowers could be characterized as sexually size dimorphic.  However, when size variation due to flower position (architecture) is controlled experimentally, differences between the floral genders for the nongynoecial characters disappear; there is no difference in corolla or androecium size.  Staminate flowers appear to be generally smaller than hermaphroditic flowers, not because of any difference related to primary sexual function, but because they tend to occur in the distal regions of each inflorescence.  In contrast, significant differences between hermaphroditic and staminate flowers for primary female traits (ovary, style, and stigma) remain after controlling for position: the two floral types are truly dimorphic for these characters. We show that consideration of architectural effects can direct and refine hypotheses concerning the evolution of andromonoecy. More generally, if architectural effects on flower size are common among taxa with unisexual flowers, then these effects may contribute to the common perception of size dimorphism in taxa with unisexual flowers (Diggle and Miller 2004). This approach has been extended to analyses of various members of the family Apiaceae.  Within this family, some genera include both hermaphroditic and andromonoecious taxa.  Comparison of architecture and size dimorphism in such congeneric species is underway.   Preformation and Plasticity in Alpine Tundra Perennials The development of many perennial plants is characterized by preformation, the initiation of structures one or more years before they fully mature and function, a phenomenon absent in annuals and most short lived perennials. I investigated the hypothesis that preformation can inhibit or even preclude the expression of morphological plasticity. Developmental and architectural analyses of Polygonum viviparum (Polygonaceae), an herbaceous perennial of the alpine tundra, resulted in a complete temporal model of the developmental dynamics of this organism (Diggle 1997a). Each leaf and inflorescence borne by an individual is preformed and requires a four-year period of development from initiation to maturation and function. This developmental model of P. viviparum was then used to make explicit predictions about potential short and long term phenotypic responses to environmental variation. Recent analyses in the alpine tundra (at the Niwot Ridge NSF LTER Research Site) demonstrate a delay of one or more years in community level plant responses to environmental variation. A central unresolved issue for alpine ecologists had been to link observed community level responses in the alpine to processes manifest at the level of individuals. Developmental models provide the critical link: as a consequence of the prevalence of extreme preformation in the alpine, developmental responses to environmental change are not "expressed" for an additional one to four years, when those preforming organs mature and function (Diggle 1997a). Studies of the connections between organismic patterns of development and community level response to the environment in the alpine are continuing. In collaboration with two of my graduate students, investigations of preformation, and the constraints of preformation on plasticity, have been extended to other species (Aydelotte and Diggle 1997, Meloche and Diggle, 2001). Manipulative experiments have also tested the mechanistic link between preformation and the delayed response to environmental variation observed in the alpine. Experimental defoliation was used to dramatically alter the resource environment of individuals of both P. viviparum and Caltha leptospeala (Ranunculaceae) (Aydelotte and Diggle 1997, Diggle in preparation). Despite the presence of numerous preformed leaf primordia within the apical bud, neither species is able to replace the photosynthetic tissue lost to defoliation within the growing season. Plants did respond developmentally to the defoliation treatment, however all responses occurred in leaf and inflorescence primordia developing below-ground in the apical bud. The responses were not manifest on an ecological (or mature phenotypic) scale, until one and two years following the treatment. These experiments confirm the importance of developmental models for understanding patterns of response observed at ecological scales. Investigations of Acomastylis rossii (Rosaceae) have shown that the complex morphology, including the presence of a well developed preformed bud bank can ameliorate the constraints imposed by preformation (Meloche and Diggle, 2003). Finally, we have extended our work in the alpine to examine levels of genetic variation within populations. Genetic variation among individuals in the alpine has been predicted to be low due to the prevalence of asexual reproduction. However, during the course of our developmental analyses of plants in the alpine, we discovered that there is considerable phenotypic variation among individuals. This in turn suggested that genetic variation might be significantly greater than expected. An isozyme analysis of P. viviparum confirmed that genotypic diversity in this species is much greater than had been predicted; and may serve as a reservoir for continued population-level responses to environmental fluctuation (Diggle et al. 1998).

Return to Lab Home Page