We censused plants bearing ripe fruit along eight permanently marked transects to describe patterns of ripe fruit availability in a spatially interspersed patchwork of seasonally dry forests in the northern Florida Keys. Plants were sampled every four to six weeks between May 1998 and April 1999. We noted every tree, shrub, or vine with ripe fleshy fruits present and rated each stem for the proportion of fruit produced. This sampling scheme allowed us to describe periods of ripe fruit availability in the forest community, as well as for the fourteen most frequently encountered species. Ripe fruit availability peaked during the wet season (July–September), deciduous trees and evergreen shrubs produced most of the fruit in most months, and most of the species producing fruit were evergreen trees in all months.

Monitoring programs have been established by agencies to assist in the conservation of biological diversity and to assess the impacts of a variety of threats to species and natural systems. However, poorly designed programs may fail to detect changes in the system they were intended to monitor. We evaluated and revised the long-term monitoring program for vascular plants at Shenandoah National Park (SHEN) and used our evaluation and revision to illustrate common problems associated with monitoring programs. Our evaluation of data collected through SHEN's 14-year monitoring program indicated that the design and sampling effort lacked the statistical power to detect changes in stem density of shrubs or saplings, although there was sufficient power to detect changes in tree basal area for some species. The inability of the program to detect changes was due, in part, to the lack of specific management and sampling objectives, stratifying the park according to forest cover types that changed over time, errors associated with misidentification of species, and an inappropriate temporal schedule for visiting sampling plots. Our revised sampling design attempted to correct these deficiencies by using strata that are unlikely to change over time (elevation, bedrock geology, and aspect), streamlining data collection to allow a greater number of sampling plots, and visiting all plots over a 2-year period (interpenetrating sampling) to increase sample size. Long-term monitoring programs should: (1) establish specific, quantifiable, and attainable management and statistical sampling objectives; (2) use permanent characteristics to define strata (e.g., elevation or bedrock geology), if stratification is warranted; and (3) perform a statistical evaluation of the monitoring program once sufficient data have been collected. Prompt evaluation of an implemented monitoring program will allow timely modifications of either sampling objectives or the sample design with minimal delay in collecting meaningful long-term trend information.

Considerable variation in seed dormancy has been reported for Colubrina species. Most freshly-matured seeds of Colubrina oppositifolia have water-impermeable seed coats (physical dormancy). At 28(12 h)/20(12 h), 25(12 h)/15(12 h) °C and ca. 24 °C (room temperatures), 99–100% of the seeds imbibed, and 96–100% of the seeds that germinated did so in 27–34 days. Daily weighing of individual seeds showed that start of imbibition (= 5% increase in mass) at room temperature ranged from < 1 to 20 days. Thus, at high temperature, physical dormancy is broken so rapidly that technically most seeds would be classified as nondormant (i.e. germinate in ≤30 days). At 15(12 h)/6(12 h) and 20(12 h)/10(12 h) °C, 93 and 96%, respectively, of the seeds had imbibed after 30 days, but only 12 and 50%, respectively, of the seeds that germinated did so in 30 days. Eleven months of dry storage at room temperatures increased rates of imbibition and germination of seeds incubated at 28/20 and 20/10 °C. Based on the relatively rapid rates at which seeds become water-permeable and germinate when incubated on a moist substrate, we predict that the soil seed bank of this endangered species may be only short-lived. Thus, seeds would imbibe and germinate (or die) during the first wet season following their dispersal.

Intensive agricultural practices across the midwestern United States cleared nearly all of the original broadleaf deciduous forest, and today restrict the selection of protected areas to regrowth stands across human-dominated landscapes. This study focused on the: (1) physical-site conditions and past land use influence woody plant diversity and the canopy structure of protected regrowth forests, and (2) relationships between these measures of structural diversity and community resistance to an invasive non-native shrub. We measured woody vegetation at 30 points in the 268 ha Bachelor Reserve, Southwest Ohio, and examined statistical relationships between physical-site conditions, historical land use since 1938, and the structural diversity of woody plants. Stand age, with topographic and soil attributes, provide unique and significant contributions to multiple regression models, explaining canopy species richness and height variability. Certain physical settings were managed in certain ways and released at similar times, confounding the interpretation of unique management effects. A diverse canopy structure shows a negative relationship with abundances of non-native Lonicera maackii (Rupr.) Herder, but the shrub is positively related with understory species richness and physical-site conditions that explain diversity in the understory. The study documents positive trends over time toward the development of regrowth broadleaf deciduous forests in the reserve, following a history of cultivation and grazing, and highlights the importance of environmental heterogeneity and size in protecting diversity across intensively managed landscapes.

Anthropogenic fire regimes obscure natural fire regimes, reducing the ability to manage fire-frequented habitats ecologically. To address this problem, we attempted to decouple natural and anthropogenic fire regimes by comparing them to seasonal climatic patterns and landscape characteristics in Everglades National Park (1948–1999). Of the total area burned by lightning fires, 57% resulted from ignitions seven days within onset of the wet season, 11% from ignitions starting 7–21 days before onset, and 36% from ignitions > 7 days after onset. In contrast, of the total area burned by incendiary fires, 89% resulted from ignitions > 7 days before onset, and 40% resulted from ignitions > 35 days before onset. Moreover, ~100% of the total area burned by prescribed fires resulted from ignitions > 7 days after onset. Lightning fires occurred most frequently in wet seasonal savanna that had limited accessibility to humans; incendiary fires were most frequent in wet seasonal savanna that had ready accessibility to humans. In addition, 35% of the total area burned by incendiary fires in areas of limited accessibility occurred when incendiary fires spread from readily accessible areas. We propose that, because incendiary fires occurred at the end of the dry season, they burned drier fuels and burned more intensely than lightning fires, which generally occurred following the first rains of the wet season. Incendiary fires thus should be more likely to burn lower elevation areas that normally hinder fire spread. Finally, by occurring later in the wet season, prescribed fires may have burned patchily and insufficiently intensely to achieve restoration goals. Decoupling anthropogenic and natural fire regimes using seasonal climate patterns and landscape characteristics leads us to propose strategies to guide fire management in the park.

One of the threats to native biodiversity in native grasslands is the abundance of exotic plants. The objectives of this study were to evaluate the efficacy of imazapic, clethodim, and sulfosulfuron for removing tall fescue (Festuca arundinacea Schreb.) and releasing the native warm-season grasses (NWSG) and to document which species of broadleaf plants survive a post-emergence treatment of 0.2 kg ai/ha of imazapic or 0.03 kg ai/ha sulfosulfuron. All three herbicides reduced the amount of total vegetative cover after the first growing season, but there were no differences by the end of the second growing season. All herbicide-treated plots had less tall fescue than untreated control plots after the first and second growing seasons. After the first growing season, imazapic- and sulfosulfuron-treated plots had less tall fescue than clethodim-treated plots, but there was no difference by the end of the second growing season. NWSG increased the most in the imazapic-treated plots and second most in the sulfosulfuron-treated plots. NWSG did not increase in the clethodim-treated plots when compared to the controls after the first growing season. All herbicide treatments increased the amount of bare ground, and the clethodim was the least harmful to broadleaf plants. These herbicides show promise as a useful tool for removing tall fescue from remnant grasslands and enhancing NWSG and selected broadleaf plants in native grassland systems.

The Barrens of the Eastern Highland Rim in south-central Tennessee is a unique area and home to several rare species and plant communities. An arthropod diversity study conducted on Arnold Air Force Base sampled 11 sites that exemplify habitats found within the Barrens. This diversity study documented adults of 37 species of Tenebrionidae (darkling beetles) representing 24 genera; 14 of the 37 species collected are new Tennessee state records. The greatest number of species was collected from Sinking Pond (n = 16), Goose Pond (n = 14), and the Pine site (n = 11), which were all forested sites. No adult darkling beetles were collected in four (Grassland sites 1, 2, and 3 and Maidencane site) of the 11 sites sampled; these four sites are all open and lack a forest component. Of five collection methods that captured tenebrionids, canopy fogging collected the greatest number of species (n = 14), followed by light trapping (n = 13), and under-bark sampling (n = 13). About 50% (n = 84) of all tenebrionid specimens were collected utilizing under-bark sampling, due, in part, to the collection of 24 specimens of Uloma imberbis LeConte and 37 specimens of six species collected solely under bark. This study documented about 28% of the approximate 80 species of Tenebrionidae previously recorded in the literature from Tennessee, while adding 14 species records to the literature. This study helps to characterize the species known from habitats within the Barrens.

Bottomland hardwood forest has been greatly diminished by conversion to agriculture. Less than 25% of the pre-Columbian bottomland hardwood forests remain in the southeastern United States. Because of the valuable ecological and hydrological functions performed by these forests, their conservation and restoration has been a high priority. Part of these restoration efforts has focused on developing tools that can be used for both assessments at the landscape level and policy implementation at the local level. The distribution of bottomland hardwood forests in the Cache and White River watersheds in eastern Arkansas were examined using existing GIS databases. Criteria were developed to select areas that should be conserved or augmented for wildlife habitat. Over 67% of the study area was classified as agriculture, with bottomland hardwood forest the next largest habitat class. The thickness of a forest fragment was defined as the radius of the largest circle that can be inscribed in a fragment. Thickness was used in three ways. First, individual forest fragments were identified and selected based on ecological function using criteria we established. Second, individual fragments that were too small to support interior species, but large enough that if moderately augmented they could recover that function, were identified and selected. These augmentable fragments were further prioritized by adjacency to habitat that might be suitable for reforestation, namely agriculture. Third, watersheds were prioritized for conservation and augmentation based on the size and distributions of forest fragment thickness and area within each watershed.

To identify natural resources in need of conservation, and assess the effectiveness of ongoing management practices, a ‘reliable’ monitoring program is necessary. It is critical to assess the reliability of our data, and our data analyses, so that we draw the appropriate conclusions regarding the natural resource of interest. One way to evaluate this reliability is through the use of statistical power analysis. Although power analysis may provide valuable insights into the design and results of a study or monitoring program, its misuse may lead to inappropriate conclusions and management actions. This review describes the appropriate use of statistical power analysis in the context of natural areas management, and points out numerous misuses, some of which are not widely recognized. Alternative approaches to traditional power analyses are presented, along with a discussion of their advantages and disadvantages.