Analysis of the floral MADS-box genes from monocotyledonous Trilliaceae species indicates the involvement of SEPALLATA3-like genes in sepal-petal differentiation.

The evolution of greenish sepals from petaloid outer tepals has occurred repeatedly in various lineages of non-grass monocots. Studies in distinct monocot species showed that the evolution of sepals could be explained by the ABC model; for example, the defect of B-class function in the outermost whorl was linked to the evolution of sepals. Here, floral MADS-box genes from three sepal-bearing monocotyledonous Trilliaceae species, Trillium camschatcense, Paris verticillata, and Kinugasa japonica were examined. Unexpectedly, expression of not only A- but also B-class genes was detected in the sepals of all three species. Although the E-class gene is generally expressed across all floral whorls, no expression was detected in sepals in the three species examined here. Overexpression of the E-class SEPALLATA3-like gene from T. camschatcense (TcamSEP) in Arabidopsis thaliana produced phenotypes identical to those reported for orthologs in other monocots. Additionally, yeast hybrid experiments indicated that TcamSEP could form a higher-order complex with an endogenous heterodimer of B-class APETALA3/DEFICIENS-like (TcamDEF) and PISTILLATA/GLOBOSA-like (TcamGLO) proteins. These results suggest a conserved role for Trilliaceae SEPALLATA3-like genes in functionalization of the B-class genes, and that a lack of SEPALLATA3-like gene expression in the outermost whorl may be related to the formation of greenish sepals.

A reevaluation of the use of rhizome scars to age plants of Trillium erectum (Melanthiaceae)1.

PREMISE OF STUDY: Herbaceous perennials are important long-lived plants in North American forests. Trillium has been used as a model organism to examine the effects of ecological processes on age structure in herbaceous forest perennials. Here, the methods of aging Trillium rhizomes are critically examined. • METHODS: Rhizomes of seedlings, single-bract plants, three-bract nonflowering plants, and flowering plants of Trillium erectum were examined. The patterns of cataphyll and scape scars on rhizomes were examined with respect to demographic category. • KEY RESULTS: Trillium rhizomes produce two cataphyll scars per year on single-bract, three-bract nonflowering, and flowering plants. Scape scars were only evident on rhizomes of three-leaved nonflowering and flowering plants. The percentage of intact rhizomes ranged from 10-67% for three-bract nonflowering plants, and 0-51% for flowering plants. Rhizomes in all demographic categories had evidence of recessing tissues from the oldest portion of the rhizome indicating that accurate age estimates are not possible on many plants. • CONCLUSIONS: Accepted methods of aging Trillium rhizomes have significant drawbacks. The primary problem is that rhizomes rot from the oldest portions in all demographic categories. A second problem is that plants producing multiple scape scars in a given year could mistakenly be counted as multiple years. Finally, confusing terminology and misrepresentations in Trillium literature suggests that many previous studies did not correctly determine age. Given the challenges of aging Trillium, we suggest that future studies use rhizome aging to study early demographic stages only and the ecological processes that influence their growth.

Seed dormancy in Trillium camschatcense (Melanthiaceae) and the possible roles of light and temperature requirements for seed germination in forests.

PREMISE OF THE STUDY: Seeds of some temperate plants require multiple treatments to break complex forms of dormancy, such as deep simple double morphophysiological dormancy (MPD), but embryo growth and radicle and cotyledon emergence have not been studied in depth for this level of MPD. Here we studied Trillium camschatcense, a species that is purported to have this MPD and that is sensitive to habitat fragmentation with decreased recruitment at forest edges. METHODS: Trillium camschatcense seeds were collected and experiments conducted in Hokkaido, Japan over 5 years. Growth of embryos and emergence of radicles and cotyledons were tracked in relation to field temperature to determine specific temperature and light requirements for these stages under laboratory conditions. KEY RESULTS: Low (winter) temperatures overcame dormancy in the underdeveloped embryos, and embryo growth and radicle emergence occurred between July and September, ca. 1 year after seed dispersal. Radicles emerged optimally over a narrow temperature range (20-25°C), in darkness, and at constant temperatures. Roots developed during the second autumn. Cotyledons needed a second low temperature (second winter) to emerge from seeds with roots, doing so in April, slightly over 1.5 years after dispersal. CONCLUSIONS: Seeds of T. camschatcense have deep simple double MPD and requirements for radicle emergence: darkness and constant temperatures. Ecologically, edges of forests may be deleterious for germination of the species due to increased light and to higher temperature fluctuations as compared to the interiors of forests. Thus, these specific requirements may play an important role in reducing seed germination of this plant at forest edges.

Demographic response of plant populations to habitat fragmentation and temporal environmental variability.

Many plant species currently exist in fragmented populations of different sizes, while they also experience unpredictable climatic fluctuation over time. However, we still understand little about how plant demography responds to such spatial and temporal environmental variability. We studied population dynamics of an understory perennial herb Trillium camschatcense in the Tokachi plain of Hokkaido, Japan, where a significant effect of forest fragmentation on seedling recruitment was previously reported. Four populations across a range of fragment sizes were studied for 6 years, and the data were analyzed using matrix population models. Per capita fecundity (the number of recruits per plant) varied greatly among populations, but the variation in population growth rates (lambda) was mainly driven by the variation in stasis and growth rates, suggesting that the general trend of reduced fecundity in fragmented populations may not be readily translated into subsequent dynamics. Temporal variation in lambda among years was more than 2 times larger than spatial variation among populations, and this result was likely attributable to the contrasting response of correlation structures among demographic rates. The among-population variation in lambda was dampened by negative covariation between matrix elements possibly due to density-dependent regulation as well as an inherent constraint that some elements are not independent, whereas positive covariation between matrix elements resulted in large temporal variation in lambda. Our results show that population dynamics responded differently to habitat fragmentation and temporal variability of the environment, emphasizing the need to discriminate these spatial and temporal variations in demographic models. Although no populations were projected to be declining in stochastic simulations, correlation between current habitat size and plant density implies historical lambda is positively related to habitat size.

A heterogeneity test for fine-scale genetic structure.

For organisms with limited vagility and/or occupying patchy habitats, we often encounter nonrandom patterns of genetic affinity over relatively small spatial scales, labelled fine-scale genetic structure. Both the extent and decay rate of that pattern can be expected to depend on numerous interesting demographic, ecological, historical, and mating system factors, and it would be useful to be able to compare different situations. There is, however, no heterogeneity test currently available for fine-scale genetic structure that would provide us with any guidance on whether the differences we encounter are statistically credible. Here, we develop a general nonparametric heterogeneity test, elaborating on standard autocorrelation methods for pairs of individuals. We first develop a 'pooled within-population' correlogram, where the distance classes (lags) can be defined as functions of distance. Using that pooled correlogram as our null-hypothesis reference frame, we then develop a heterogeneity test of the autocorrelations among different populations, lag-by-lag. From these single-lag tests, we construct an analogous test of heterogeneity for multilag correlograms. We illustrate with a pair of biological examples, one involving the Australian bush rat, the other involving toadshade trillium. The Australian bush rat has limited vagility, and sometimes occupies patchy habitat. We show that the autocorrelation pattern diverges somewhat between continuous and patchy habitat types. For toadshade trillium, clonal replication in Piedmont populations substantially increases autocorrelation for short lags, but clonal replication is less pronounced in mountain populations. Removal of clonal replicates reduces the autocorrelation for short lags and reverses the sign of the difference between mountain and Piedmont correlograms.

Evolutionary dynamics as a component of stage-structured matrix models: an example using Trillium grandiflorum.

Evolution by natural selection improves fitness and may therefore influence population trajectories. Demographic matrix models are often employed in conservation studies to project population dynamics, but such analyses have not incorporated evolutionary dynamics. We project evolutionarily informed population trajectories for a population of the perennial plant Trillium grandiflorum, which is declining due to high levels of herbivory by white-tailed deer. Individuals with later flowering times are less often consumed, so there is selection on this trait. We first incorporated selection analyses into a deterministic matrix model in three ways (corresponding to different methods that have been used for analyzing evolution in structured populations). Because it is not clear which of these methods works best for stage-structured models, we compared each with a more realistic, individual-based model. Deterministic models using fitness averaged over the phenotypic distribution gave trajectories that were similar to those of the individual-based model, whereas the deterministic model using fitness at the mean phenotype gave a much faster rate of evolution than that which was observed. This illustrates that subtle differences in the way in which one splices evolution into demographic models can have a large effect on expected outcomes. This study demonstrates that, by combining demographic and selection analyses, one can gauge the potential relevance of evolution to population dynamics and persistence.

Timing of canopy closure influences carbon translocation and seed production of an understorey herb, Trillium apetalon (Trilliaceae).

BACKGROUND AND AIMS: The light availability on a temperate, deciduous-forest floor varies greatly, reflecting the seasonal leaf dynamics of the canopy trees. The growth and/or reproductive activity of understorey plants should be influenced by the length of the high-irradiance period from snowmelt to canopy closure. The aim of the present study was to clarify how spring-blooming species regulate the translocation of photosynthetic products to current reproduction and storage organs during a growing season in accordance with the changing light conditions. METHODS: Growth pattern, net photosynthetic rate, seed production, and shoot and flower production in the next year of Trillium apetalon were compared between natural and experimentally shaded conditions. Furthermore, translocation of current photosynthetic products within plants was assessed by a labelled carbon-chase experiment. KEY RESULTS: During the high-irradiance period, plants showed high photosynthetic ability, in which current products were initially used for shoot growth, then reserved in the rhizome. Carbon translocation to developing fruit occurred after canopy closure, but this was very small due to low photosynthetic rates under the darker conditions. The shading treatment in the early season advanced the time of carbon translocation to fruit, but reduced seed production in the current year and flower production of the next year. CONCLUSIONS: Carbon translocation to the storage organ had priority over seed production under high-irradiance conditions. A shortened bright period due to early canopy closure effectively restricts carbon assimilation, which greatly reduces subsequent reproductive output owing to low photosynthetic products for fruit development and small carbon storage for future reproduction. As populations of this species are maintained by seedling recruitment, acceleration of canopy closure timing may influence the maintenance and dynamics of populations.

Floral density, pollen limitation, and reproductive success in Trillium grandiflorum.

Decreases in floral density can disrupt mutualistic interactions between plants and their pollinators, and decrease reproductive success. I addressed the relationship between floral density and plant reproductive success using two experimental approaches: a pollen supplementation experiment in 12 populations of Trillium grandiflorum that naturally varied in floral density, and a transplant experiment in which floral density was manipulated in plots at four experimental sites. In the pollen supplementation experiments, the degree of pollen limitation, in terms of fruit set and seed set, decreased with floral density. Further, in the experimental sites, plant reproductive success increased asymptotically with floral density. These results demonstrate the value of simultaneously conducting experiments in both experimental sites and natural populations to understand how population density influences plant reproductive success. Factors that reduce the density of this perennial herb, such as habitat fragmentation and herbivory by white-tailed deer ( Odocoileus virginianus), should be expected to limit its reproduction.