Restoration is an important tool for reducing extinction risk of endangered plants. Population viabilities of few plant restorations have been modeled over decadal time periods and linked with genetic and ecological factors that drive restoration processes. We modeled viability of restored populations of Mead's milkweed (Asclepias meadii, Asclepiadaceae), a self‐incompatible perennial herb of eastern tallgrass prairie (TGP), federally listed as threatened in the U.S. From 1994 to 2004, we planted >600 seeds and >800 juvenile plants representing >50 genotypes across seven TGP sites. Propagule type, genotype, seed source, restoration site, precipitation and fire management significantly affected establishment, growth and viability. Plants established from seed had greater mortality and greater genetic and demographic attrition than did juveniles. Seedling growth rates also projected 20–30 yrs to reach flowering stage, and their survivorship provided a metric of site suitability for life cycle completion. Seed germination and juvenile plant size were greater in burned habitat, and juvenile size was also positively correlated with spring precipitation. Seed production required presence of multiple genotypes among flowering plants. Seedlings demonstrated a heterosis effect, with greater germination among seeds derived from inter‐population crosses. However, cumulative growth of planted juveniles as well as population growth (λ) on sub‐optimal habitat conditions tended to be lower for propagules derived from inter‐population crosses, demonstrating outbreeding depression. Although flowering occurred at multiple sites, positive population growth (λ > 1) occurred at only a single site, where increasing fire frequency decreased extinction probability. These results suggest that restoration of viable Mead's milkweed populations is possible in optimal habitat. However, restoration of this species is constrained by high demographic attrition and the long period (20 or more yrs) required to complete its life cycle. Crossing among populations to increase genetic diversity and compatible mating types may result in tradeoffs, with heterosis at early life history stages, but outbreeding depression expressed in older stages. Fire and precipitation are also critical interactive processes driving A. meadii growth and reproduction. They may be most effective when precipitation, a stochastic process, results in greater than average post‐burn rainfall. These constraints may have implications for restoration of other late‐successional plant species.

Most populations ofthe federal threatened Mead's milkweed, Ascleplas meadii Torn (Asclepiadaeeae), occur primarily in prairie haymeadows in Kansas and Missouri, where annual summer mowing prevents seed production. Exceptions are large populations in fire-managed habitats at the Rockefeller Prairie, a former Kansas haymeadow, and at Weimer Hill, a glade complex in southeastern Missouri. Phis perennial rhizomatous species is self-incompatible. The few remaining small populations in Illinois, Iowa, and northern Missouri persist vegetatively but no longer produce seeds and are vulnerable to stochastic extinction processes. Allozyme electrophoresis was used to measure the amount and distribution of genetic variation in A. meadii and to provide guidance for its recovery and restoration. Samples were obtained from 19 populations encompassing the extant range of the species in Kansas, Missouri, Iowa, and Illinois. Asclepias meadii was genetically variable for most of the 12 loci examined, with a mean of 1.53 alleles per locus. 40.8% polymoqmic loci, and observed heterozygosity of 0.158. These values are comparable to published values for other milkweed species. More than half of the total 42 alleles were rare, with 15 alleles unique to single populations. About 74% of the genetic variation in A. meadii occurs within populations (F sr = 0.261), and analyses do not provide conclusive evidence for a geographic pattern in genetic variation among populations. The two fire-managed populations had more genotypes and fewer ramets per genet than the haymeadows. For the latter, inhibition of sexual reproduction may have resulted in clonal spread and attrition of genotypes, thus exacerbating the effects of sexual incompatibility and inbreeding. These factors suggest that multiple propagule sources would maximize genetic diversity for recovering depauperate populations or for restoring new populations; however, sampling from a few larger, genetically diverse populations would provide much of the species' genetic diversity. Such sources would include the large fire-managed populations. Empirical data are needed to determine the population-genetic consequences of long-distance crosses and introductions that are apparently needed to restore viable populations of this species in the eastern part of its range.

The federal threatened Asclepias meadii Torr. (Aselepiadaceae) is a perennial, self-incompatible prairie fori) imperiled by habitat destruction and population fragmentation. Many large populations persist in prairie haymeadows in Kansas and Missouri despite removal of seed pods by annual summer mowing. Only a few small populations remain in Illinois. Iowa, and northern Missouri. Recovery of these small populations and restoration of new populations are recovery objectives lor this species. This study was conducted to determine habitat differences among populations, to understand how hay mowing and fire management affect the structure of A. meadii populations, and to test the effects of different management treatments on restoration of this species. Soils analysis showed a geographic gradient, with southern populations on acid, nutrient-poor soils, and northern populations on calcareous, nutrient-rich soils. Milkweed ramet densities were lower in fire-managed prairies than in haymeadows; but burned sites had 68% (lowering ratnels while haymeadows had only 19% flowering ramets. This suggests that burning has selected for greater resource allocation toward sexual reproduction, while annual hay mowing has selected for greater resource allocation toward clonal spread. The Morton Arboretum is conducting experimental restoration of Asclepias meadii in the eastern part of its range, an objective of Federal Recovery Planning. In greenhouse and garden experiments, competition from oats significantly reduced seedling growth, with greater growth in artificially outcrossed seedlings. At seven restoration sites in Illinois and northern Indiana, significant variation in milkweed germination, survivorship, and growth was caused by weather, differences among sites, and site management. Seedlings were vulnerable to drought, with greater survivorship when rainfall was 2

No abstract