| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2 Johnson Controls World Services Inc., National Wetlands Research Center, 700 Cajundome Blvd., Lafayette, Lousiana 70506 USA; and 3 U.S. Geological Survey, Biological Resources Division, National Wetlands Research Center, 700 Cajundome Blvd., Lafayette, Louisiana 70506 USA
Received for publication July 22, 1999. Accepted for publication December 9, 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: disturbance • exotic species • gap size • germination • invasion • prescribed burning • seedling survival
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Williamson, 1996
; Luken and Thieret, 1997
; Bright, 1998
). Although numerous authors have attempted to determine species' characteristics or community characteristics that contribute to invasion by nonnative species (Fox and Fox, 1986
; Swincer, 1986
; Rejmánek and Richardson, 1996
), there is general consensus that successful invasion is a context-specific process (Bazzaz, 1986
; Crawley, 1987
; Noble, 1989
; D'Antonio, 1993
; Hobbs and Humphries, 1995
), depending on both the invaders' life-history attributes and favorable site conditions. Thus, site-specific and species-specific studies to determine factors that contribute to ecosystem invasion are warranted for predicting the invasion potential of communities at risk from exotic species (Pierson and Mack, 1990
; Hobbs and Humphries, 1995
). This study investigates the potential for cogongrass [Imperata cylindrica (L.) Beauv.], an exotic perennial, to invade wet pine savanna communities of the southeastern United States by evaluating how gap size and disturbance type influence cogongrass establishment by seed.
Imperata cylindrica var. major was first introduced into the United States in 1911 at Mobile, Alabama, when the plant arrived accidentally as shipping material in a crate of satsuma oranges from Japan. A second, intentional introduction of cogongrass from the Philippines occurred sometime before 1920 at the McNeil Mississippi Agricultural Experiment Station, where it was introduced as a potential forage crop (Tabor, 1949
). Since its introduction, cogongrass has become established in Florida, Alabama, Mississippi, and Louisiana (MacDonald and Chandler, 1994
; Shilling, Gaffney, and Waldrop, 1995
) and was recently estimated to cover several thousand hectares in the southeastern United States (Bryson and Carter, 1993
). Shilling (1996)
reports that cogongrass in the United States can displace other vegetation in forests, rangelands, pastures, roadsides, reclaimed mining areas, and natural areas. In a study of cogongrass invasion of Florida sandhill communities, Lippincott (1997)
determined that cogongrass is not functionally equivalent to native sandhill species and is capable of altering community function by changing vegetation structure, soil processes, resource availability, fire regime, and native seedling recruitment. Other ecologically sensitive communities in the Southeast are also at risk from cogongrass invasion, including coastal wet pine savanna.
Wet pine savanna communities are some of the most diverse in North America (Peet and Allard, 1993
). They are characterized by an open canopy of pines (Pinus palustris P. Mill. and/or P. elliottii Engelm.) and a diverse understory of grasses and forbs maintained by frequent fires. Frequent fires and wet soils of low pH and relatively low nutrients provide conditions favorable to many unique endemic species including carnivorous pitcher plants (Sarracenia L. spp.) and sundews (Drosera L. spp.), and the endangered Mississippi sandhill crane (Grus canadensis pulla). Maintenance of open, wet pine savanna requires active management, particularly by prescribed burning (Brockway and Lewis, 1997
). Concern exists that the site conditions and management practices conducive to maintaining these communities may also favor the establishment and growth of cogongrass.
Most studies of cogongrass invasion in the United States have examined communities already infested with cogongrass, and no studies have attempted to assess what site characteristics contribute to cogongrass invasion in native communities. Although it is generally known that cogongrass invades sites with disturbed soils, such as reclaimed mining lands and roadways, it is unknown whether cogongrass can invade vegetatively intact native communities by seed without some form of disturbance. The effects of small-scale disturbances on germination and establishment of many native and exotic species have been examined in numerous studies (Fenner, 1978
; Pickett and White, 1985
; Goldberg, 1987
; Peart, 1989
; Hobbs, 1991
; D'Antonio, 1993
), and in many cases, germination and seedling growth were found to be dependent on gap size or disturbance intensity (Goldberg and Werner, 1983
; McConnaughay and Bazzaz, 1987
; Rice, 1987
; Klinkhamer and DeJong, 1988
; Burke and Grime, 1996
). Hobbs and Huenneke (1992
, p. 325) point out the conundrum of disturbance and nonnative species invasions: "the continued existence of particular species or communities often requires disturbance of some type ...but disturbance may simultaneously lead to the degradation of natural communities by promoting invasions." This issue is of primary importance in wet pine savannas, which require periodic burning to maintain the unique and diverse species assemblages associated with this community type, but which may be at risk of invasion by cogongrass. The objectives of this study were to examine the potential for establishment of cogongrass in wet pine savannas by evaluating the response of cogongrass seeds and transplanted seedlings to (1) four sizes of vegetation gaps, from 0 to 100 cm in diameter; (2) four types of small-scale disturbance including no disturbance, mowing, vegetation removal with no soil disturbance, and tilling; and (3) burning of savanna vegetation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). The herbaceous vegetation of the study site is typical of wet pine savannas in the region, dominated by wiregrass (Aristida L. sp.) with a diverse mix of other herbaceous species throughout (see description of southern longleaf savannas in Peet and Allard, 1993
). The savanna contains no cogongrass, although established stands exist 
0.5 km away on the refuge property. Sphagnum moss (L.) on the soil surface is indicative of moderately wet soils of low pH, as soil analyses confirmed. The sandy loam soil has a pH of 4.3 and a low nutrient content, typical of wet pine savannas. The savanna had not been burned in over 5 yr and contains no mature pines, however pine saplings and woody shrubs including inkberry [Ilex glabra (L.) Gray] and yaupon (Ilex vomitoria Ait.) are common throughout. Relative elevation across the entire study site varies by 0.31 m.
Environmental characterization
To characterize the effects of treatments on light availability to seeds and seedlings, relative percentage of photosynthetically active radiation (PAR, 400–700 nm) reaching the soil surface at each plot was determined using LI-COR quantum sensors and a data logger (LI-COR inc., Lincoln, Nebraska, USA). Simultaneous readings were taken with a sensor placed in the plot at surface level and a sensor mounted on a pole that reached above the canopy; percentage PAR reaching the surface was determined from these two readings at each plot. We used ANOVA and Tukey's multiple comparison tests to determine whether light availability differed among treatments in the gap size experiment and the disturbance type experiment. Differences in PAR between treatments in the burning experiment were tested with a t test using Satterthwaite's approximation, because the data were distributed unequally between treatments (SAS, 1989
).
Elevation of each plot was determined using a Spectra-Physics laser level (Spectra-Physics Laserplane, Inc., Dayton, Ohio, USA) prior to planting seeds or transplants. Since surface irregularities common to this habitat type result in localized ponding, actual elevation may not give true estimates of plot flooding. Therefore, flooding depth at each plot was measured with a ruler on a single day when the site was inundated following heavy rain. Elevation and microsite flooding were analyzed with seed germination data as described below for each experiment. Statistical analyses in all experiments were performed in SAS (SAS, 1989
) using an alpha level of 0.05.
Containment barriers
All field experiments used seed plots to examine germination and seedling survival and transplant plots to examine establishment and growth. To reduce the risk of cogongrass escaping into the savanna, all seeds and transplants were planted inside of containment barriers similar to the root-exclusion tubes used by Cook and Ratcliff (1984)
and Reichenberger and Pyke (1990)
, which they used to reduce root competition between seedlings and surrounding plants. Barriers were 46-cm long cylinders of PVC (polyvinyl chloride) sewer pipe, each with two 0.79-cm drainage holes drilled opposite from the other, 1 cm from the top edge. Seed plot barriers were 10-cm in diameter and were covered with fine white nylon mesh (<0.5 mm) secured by heavy rubber bands to prevent seeds from washing out during extreme rainfall events. Transplant plot barriers were 15-cm in diameter and were left uncovered. For each seed or transplant plot in the experiments, a single barrier was pounded into the ground until drainage holes were approximately flush with the soil surface. All barriers were placed in the savanna before treatments (gap sizes, types, or burning) were administered. We conducted a pilot study in 1997 using these barriers and determined that neither the PVC tubes nor the mesh covering seed plots impeded rainfall or significantly altered surface flooding.
Cogongrass inflorescences were collected haphazardly by hand in the first week of April 1997, from cogongrass flowering within 8 km of the study site, along the shoulder and median of Interstate Highway 10 East in Jackson County, Mississippi, USA. Cogongrass inflorescences were stored intact in paper bags at 5°C. Because production of unfilled florets is common in this species (Shilling et al., 1997
), caryopses were removed by hand from florets, counted into groups of 20, and placed in storage vials prior to planting. Each seed plot was planted by emptying the 20 cogongrass seeds onto the soil surface at the center of the containment barrier. Mesh covers were secured immediately after planting. All seed plots were planted the second week of April 1998.
Transplant plots were established at the same time as seed plots by planting two, 3-wk-old seedlings (2–3 cm tall) at the center of each containment barrier. Seedlings were started in trays of commercial potting mix (Lambert Peat Moss Inc., Quebec, Canada) in the greenhouse from the same seed collected for seed plots. After 1 wk in the field, one seedling was removed from each plot or a new seedling was transplanted if none survived, so that every transplant plot contained a single cogongrass plant.
Experiment 1
Gap size
Circular gaps of 10, 30, and 100 cm in diameter were created in the vegetation with the herbicide glyphosate, N-(phosphonomethyl)-glycine, trade name Roundup (Monsanto Co., St. Louis, Missouri, USA), with subsequent clipping to the soil surface and hand removal of clipped material. A 2% concentration of Roundup was applied once to vegetation in 30-cm and 100-cm diameter plots using a backpack sprayer and by hand application several weeks prior to planting seeds and transplants. Plots of 10-cm diameter were too small to allow the use of herbicide and were cleared by hand. All standing-dead vegetation and live vegetation not killed by the herbicide was removed from plots by clipping with hand pruners to ground level. A single seed or transplant plot was placed at the center of a each gap-size treatment. Although barriers for the transplant plots were 15-cm in diameter, the parted wiregrass canopy was repositioned after planting using hand-held cultivators (small rakes) so openings were no larger than the treatment opening sizes. An additional 0-cm gap treatment was evaluated with barriers placed in undisturbed vegetation, again with the canopy returned to its original position after parting for barrier insertion and planting. Openings were maintained canopy-free for the duration of the experiment by hand removal of regrowth as needed. The four treatments were assigned in a randomized block design with eight blocks, each containing one replicate of both seed and transplant plots.
Seed plots
The number of seedlings was recorded weekly for 2 mo. Percentage germination was determined for all plots by taking the maximum number of seedlings counted in each plot at any time during the experiment and dividing by 20 (the number of seeds planted). Data were screened for deviations from the parametric assumptions of normality and homoscedasticity (Zar, 1996
). Data fit all assumptions, and treatment means were tested by ANCOVA with microsite flooding and measured elevation as covariates. Seedling survival to 1 and 2 mo was calculated by dividing the number of living seedlings by the maximum number of germinants up to that time for each plot. To ensure that survival of seedlings was not adversely affected by seedling density, we used linear regression to test the relationship between percentage germination and survival of seedlings to 2 mo in each plot. Survival to 2 mo was square-root transformed to meet normality assumptions prior to regression analysis. Results of the regression showed that these two variables were significantly, positively related (P < 0.005, r2 = 0.73, slope = 1.24), suggesting that plot conditions favorable for seed germination were also conducive to seedling survival. Since the density of seedlings did not negatively affect seedling survival, seedling density was not used as a covariate in tests of treatment effects on survival. The survival data required no transformation to meet parametric assumptions before testing treatment effects by a repeated-measures ANOVA.
Transplant plots
Survival, number of shoots, and maximum height of each shoot were recorded once every 14 d for all transplants for 5 mo. Surviving transplants were harvested intact, rinsed with water to remove soil from roots and crowns, dried at 105°C to constant mass, and weighed to the nearest 0.001 g. Final number of surviving transplants by treatment was analyzed by chi-square analysis using Fisher's Exact Test to determine association between survival and treatment. Shoot number, shoot length, and total biomass of surviving transplants were assessed for univariate normality and homoscedasticity, and a correlation analysis was used to verify correlations between variables. All three measures of growth were linear-rank transformed and analyzed together in a MANOVA using Wilks' Lambda to detect overall differences in growth among treatments. The general linear models (GLM) procedure in SAS was used to account for the unbalanced design resulting from the unequal numbers of surviving transplants in each treatment (SAS, 1989
; Scheiner and Gurevitch, 1993
). The 100-cm treatment was excluded from the analysis because only one transplant survived to the end of the experiment resulting in no within-treatment variance.
Experiment 2
Type of disturbance
This experiment examined types of site disturbance that occur naturally and with active management of this community: no disturbance, mowing, vegetation removal without soil disturbance (bare), and tilling. All treatments were applied as circles of 100 cm in diameter with either a seed or transplant plot placed at the center. For the no-disturbance treatment, seed and transplant plots were inserted without removing or altering vegetation or ground litter. These plots were the same as the 0-cm gap plots described previously. The mowing treatment was created using a gas-powered weed eater to reduce standing vegetation to a height of 10 cm. Cut vegetation was left on the plot where it fell. The bare treatment was the same as the 100-cm diameter treatment in the gap-size experiment. Tilling was accomplished by weed-eating and removing standing vegetation, then turning the soil over with shovels to a depth of 30 cm. Treatments were assigned in a randomized block design with eight blocks, each containing one replicate for both seed and transplant plots. Data from seed and transplant plots were collected and analyzed as in the gap-size experiment. Again, linear regression of percentage germination and seedling survival to 2 mo showed a significant, positive relationship (P < 0.005, r2 = 0.78, slope = 1.18) suggesting no density-dependent constraints on seedling survival. Data from seed and transplant plots met parametric assumptions for all tests, except for MANOVA of growth responses for surviving transplants. These data were inverse transformed prior to analysis with the GLM procedure. A priori contrasts identified which treatments differed. Analysis of surviving transplant growth did not include the bare treatment because only one plant survived to the end of the experiment.
Experiment 3
Burning
Six 10 x 10 m areas of savanna were burned in a prescribed burn on 6 April 1998, and six comparable areas of the same size were left unburned. Ten transplant plots were planted in each of the treatment areas 2 d after the burn, and ten seed plots were planted in each treatment area 3 d after the burn. The basic statistical design was a randomized block design with two blocks, each containing three treatment replicates per block with ten subsamples (seed and transplant plots) per replicate.
Seed plots
The number of seedlings was recorded weekly for 2 mo. Seed germination data fit parametric assumptions and required no transformation before analysis. To determine whether germination success differed between burned and unburned savanna, percentage germination was analyzed in a nested ANCOVA with burning treatment as the main effect and water depth and measured elevation as covariates. Survival of germinants to 1 and 2 mo was calculated by dividing the number of surviving seedlings by the maximum number of germinants up to that time for each plot. As in the other two experiments, to determine whether seedling density adversely affected seedling survival, we tested the relationship between percentage germination and survival of seedlings to 2 mo (square-root transformed) with linear regression. Survival was not constrained by the number of germinants in a plot (P < 0.005, r2 = 0.57, slope = 1.09), and data on survival to 1 and 2 mo were arcsine square-root transformed and analyzed in a repeated-measures nested ANOVA (Zar, 1996
).
Transplant plots
As in the previous experiments, transplant survival, number of shoots, and maximum height of each shoot were recorded at 14-d intervals for 5 mo. Aboveground and belowground biomass were harvested, rinsed with water to remove soil from roots and crowns, dried at 105°C, and weighed to the nearest 0.001 g. Categorical data analysis of a 2 x 2 x 2 contingency table (block x treatment x survival) with a chi-square test was used to detect differences in the total number of surviving transplants by block and treatment. Transplant growth data were screened for outliers and adherence to normality assumptions. A single outlier was detected in the burned treatment and was removed prior to analysis. This excessively robust plant was over twice as tall and had over three times greater total biomass than the next largest plant in the experiment. The data were linear rank transformed to meet parametric assumptions, and a correlation analysis on number of shoots, shoot height, and biomass confirmed that these growth variables were correlated. All three responses were analyzed with the GLM procedure using a nested MANOVA and Wilks' Lambda to detect differences in growth between treatments.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Seed plots
Percentage germination was 40% across all treatments (Fig. 1) and did not differ by gap size, microsite flooding depth, or measured elevation (P > 0.05). Survival of germinants to 1 and 2 mo differed by time, but not by gap size. A near-significant interaction effect of Time x Gap Size (P = 0.059) suggested that survival between months may differ among treatments (Table 1). The greatest drop in survival between months was observed in the 100-cm gaps (Fig. 1).
|
|
|
89% of above-canopy light reaching the surface in tilled and bare plots and 56 and 20% light penetration in mowed and undisturbed plots, respectively. Overall, elevation varied by 13 cm, and microsite flooding differed by 7.1 cm between the most flooded and least flooded plots. In seed plots alone, elevation varied by 11 cm, and flooding varied by 5.6 cm.
Seed plots
Germination did not differ significantly by disturbance type, microsite flooding depth, or measured elevation (P > 0.05) and averaged 44% across all treatments (Fig. 3). Survival of germinants to 1 and 2 mo differed by time (P < 0.001) but not by disturbance type (P = 0.45; Fig. 3).
|
38% across all treatments, ranging from one (bare treatment) to four survivors (no-disturbance and mowed treatments). The tilled treatment had three surviving transplants at the end of the experiment. Differences in survival among treatments were not statistically significant (P = 0.40). However, growth of transplants differed among no-disturbance, mowed, and tilled treatments (Table 2). Total shoot length and final biomass were five and ten times greater, respectively, in the tilled treatment than the other disturbance types (Fig. 4).
|
|
Seed plots
Percentage germination did not differ between burned (23%) and unburned (28%) savanna and was not affected by microsite flooding or elevation (P > 0.05); however, survival of germinants to both months was greater in burned than unburned savanna (Fig. 5, Table 3).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
; Goldberg, 1987
; D'Antonio, 1993
). Sajise (1972)
reported that cogongrass germination requires light and is favored by light quality characterized by a red to far-red ratio of 1.0 or greater, qualities that are characteristic of open or partly shaded environments. Although cogongrass has tiny seeds (1–1.3 mm long) and requires light to germinate, in this study seeds germinated just as well in undisturbed savanna as in created gaps of all sizes and types, despite receiving only 20–30% of available PAR. Because we scattered bare caryopses under the vegetation canopy, observed germination rates may be higher in this study than would be expected from natural seed dispersal events where spikelets or caryopses would have to fall through the vegetation to reach the soil surface. Pierson and Mack (1990)
reported that recruitment of seedlings of the invasive grass Bromus tectorum in forest habitats may be determined largely by whether a caryopsis falls through the forest litter layer and germinates in mineral soil. Thus, although disturbance of savanna had no direct facilitating effect on cogongrass seed germination, disturbance of the canopy would likely promote invasion simply by improving the chances for seed to reach the soil surface unimpeded.
Survival of cogongrass germinants was not significantly affected by gap size or any disturbance type other than burning. Significantly more seedlings survived in the burned treatments to months 1 and 2. Although burning can result in an overall loss of nutrients due to oxidation, volatilization, and increased leaching, fire is reported to cause a short-term increase in mobility and availability of various nutrients. Christensen (1993)
reported that burned wiregrass savanna soils were enriched in PO4-P, K, Ca, and Mg 2 mo following fire, particularly in the top 10 cm of the soil profile where new seedlings would be rooted. This agrees with the immediate post-fire soil fertility increases and greater plant growth in burned vs. unburned soils reported in an earlier study by the same author (Christensen, 1977
). Other studies report the positive effects of burning on soil microbial activity and microclimate (Old, 1969
; Peet, Anderson, and Adams, 1975
), which may be important for seedbed preparation. Christensen (1993)
suggested that microclimatic conditions for successful establishment of native savanna species are most favorable in the second post-fire year when native herb seed production is highest due to increased flowering in the first post-fire growing season. The results of this study suggest that cogongrass seedlings could already be established by the second post-fire year since cogongrass seed ripens quickly, requires no dormancy for germination, and germination and early seedling survival were high in burned savanna. Thus, spring burning of pine savannas in proximity to flowering stands of cogongrass could promote invasion by giving cogongrass seedlings an early advantage over other species. Long-term studies are needed to further evaluate cogongrass seedling survival and seedling interactions with native savanna vegetation.
We used cogongrass transplants in this study to examine how older seedlings, which had successfully grown past the two-leaf stage (as per Shilling et al., 1997
), would respond to site conditions. Overall, cogongrass transplant survival was high (40%), and no differences among treatments were detected in any experiment, despite only one transplant surviving in the bare, 100-cm gaps. Most seedling deaths began to occur at the end of July, 9 wk after transplants were put in the field. Thus, we believe that seedling deaths were not related to transplant shock, rather they were associated with stress from the hot summer weather conditions. Had our seedlings been older at the time of transplanting, they may have had even higher rates of survival. Interestingly, transplants in the largest gaps, as well as in the tilled and burned plots, appeared stressed during most of the experimental period, having red leaves and high leaf turnover (King, personal observation). Although not included in the growth analyses, the single surviving plant from the 100-cm gap was smaller (in shoot length and biomass) than the survivors from the other gap sizes and disturbance types (Figs. 2, 4). The poor performance of cogongrass in the more open treatments may have been caused by exposure to more stressful microsite temperature, light, and/or moisture conditions. Franco and Nobel (1989)
reported improved performance for plants in microsites moderated by neighboring vegetation, and Suding and Goldberg (1999)
found that direct effects of either vegetation or litter on seedling establishment were consistently facilitative. Fowler (1986)
reported that safe-site requirements for germination, survival, and growth of two native grass species were microhabitat characteristics that prevented desiccation: mainly the presence of litter and/or rocks. Interestingly, although transplants in the tilled plots also appeared stressed for much of the experiment, this did not result in decreased performance. In fact, growth as measured by shoot number, shoot length, and total biomass was greatest with tilling. Results also suggest that cogongrass growth was significantly greater in unburned savanna than in burned savanna, however, this was likely just a morphological effect of shoot elongation from shading, because total biomass showed no corresponding increase in the unburned treatment (Fig. 6). A similar trend of shoot elongation was discernible in the undisturbed treatments of the other experiments. Thus, the only significant effect on cogongrass growth and subsequent spread may be from tilling, which undoubtedly enhances invasion.
The positive effects of soil disturbance on cogongrass growth and spread may be due in part to loosening of the upper soil layer, which in this ecosystem is very dense and may physically impede root and rhizome penetration. Tilling also may improve drainage and may reduce root competition of seedlings with neighboring vegetation, which was effectively eliminated from the plot in the process of turning over the soil. Reichenberger and Pyke (1990)
demonstrated that root competition can impact seedling survival and subsequent reproductive performance for semiarid species in the Great Basin sagebrush ecosystem. Their study used PVC tubes, similar to the containment barriers used in this study, to segregate roots of seedlings and neighboring vegetation. Root competition in our study also may have been reduced by the PVC tubes used to prevent the escape of cogongrass into the savanna, although native vegetation that grew inside of the containment barriers with seedlings was allowed to grow unaltered. The greatest impact of the barriers would have been the exclusion of wiregrass roots competing with seedlings since most plots were located between wiregrass tussocks. However, in this ecosystem soil moisture is readily available in spring and early summer, thus we would likely not see the same effects of root competition observed by Reichenberger and Pyke (1990)
in their study of arid rangeland in Utah. The effects of the barriers on modifying the interactions of cogongrass with the native community might have been problematic if we had wanted to measure the development of our seedlings and transplants longer, through all life stages. Although the barriers may have lessened the consequences of underground competition, they effectively prevented the escape of cogongrass into this yet-uninfested savanna, providing valuable insight on invasion of this species into a community before it has occurred, information that is sorely lacking for most exotic species and community types.
Although more replicates in our transplant experiments might have allowed us to include all treatments in our analyses, providing stronger statistical support for our findings, we believe that our results address the issue of Imperata cylidrica's potential to establish in wet pine savanna. This study did not evaluate cogongrass survival and performance beyond the first growing season; however, the number of survivors and the vigorous appearance of the surviving seedlings and transplants when harvested suggests that a large number of these plants would have successfully established if left in place. Clearly, the fact that cogongrass was able to germinate and grow in all treatments, including the undisturbed treatments, suggests that coastal wet pine savanna is indeed at risk of invasion by this species. Further study on the long-term persistence of cogongrass seedlings under closed-canopy conditions and the competitive ability of cogongrass seedlings with native seedlings is warranted for predicting how invasion will proceed in these communities. Results from this study suggest that efforts to prevent cogongrass invasion should focus on preventing cogongrass disseminules (seeds and rhizomes) from reaching sensitive communities.
|
FOOTNOTES |
|---|
4 Author for correspondence (e-mail: Jim_grace{at}usgs.gov
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Bright, C. 1998 Life out of bounds. W.W. Norton, New York, New York, USA
Brockway, D. G., and C. E. Lewis. 1997 Long-term effects of dormant-season prescribed fire on plant community diversity, structure and productivity in a longleaf pine wiregrass ecosystem. Forest Ecology and Management 96: 167–183[CrossRef][ISI]
Bryson, C. T., and R. Carter. 1993 Cogongrass, Imperata cylindrica, in the United States. Weed Technology 7: 1005–1009[ISI]
Burke, M. J. W., and J. P. Grime. 1996 An experimental study of plant community invasibility. Ecology 77: 776–790[CrossRef][ISI]
Christensen, N. L. 1977 Fire and soil-plant nutrient relations in a pine-wiregrass savanna on the Coastal Plain of North Carolina. Oecologia 27: 27–44
———. 1993 The effects of fire on nutrient cycles in longleaf pine ecosystems. In S. M. Hermann [ed.], Proceedings of the 18th Tall Timbers Fire Ecology Conference, The longleaf pine ecosystem: ecology, restoration, and management, 205–214. Tall Timbers Research, Inc., Tallahassee, Florida, USA
Cook, S. J., and D. Ratcliff. 1984 A study of the effects of root and shoot competition on the growth of green panic (Panicum maximum var. trichoglume) seedlings in a an existing grassland using root exclusion tubes. Journal of Applied Ecology 21: 971–982[CrossRef][ISI]
Crawley, M. J. 1987 What makes a community invasible? In A. J. Gray, M. J. Crawley, and P. J. Edwards [eds.], Colonization, succession, and stability, 429–453. Blackwell Scientific Publications, Boston, Massachusetts, USA
D'Antonio, C. M. 1993 Mechanisms controlling invasion of coastal plant communities by the alien succulent Carpobrotus edulis. Ecology 74: 83–95[CrossRef][ISI]
Fenner, M. 1978 A comparison of the abilities of colonizers and closed-turf species to establish from seed in artificial swards. Journal of Ecology 66: 953–963[CrossRef]
Fowler, N. L. 1986 Microsite requirements for germination and establishment of three grass species. American Midland Naturalist 115: 131–145[CrossRef][ISI]
Fox, M. D., and B. J. FOX. 1986 The susceptibility of natural communities to invasion. In R. H. Groves, and J. J. Burdon [eds.], Ecology of biological invasions, 57–66. Cambridge University Press, London, UK
Franco, A. C., and P. S. Nobel. 1989 Effect of nurse plants on the microhabitat and growth of cacti. Journal of Ecology 77: 870–886[CrossRef][ISI]
Goldberg, D. E. 1987 Seedling colonization of experimental gaps in two old-field communities. Bulletin of the Torrey Botanical Club 114: 139–148[CrossRef][ISI]
———, and P. A. Werner. 1983 The effects of size of opening in vegetation and litter cover on seedling establishment of goldenrods (Solidago spp.). Oecologia 60: 149–155[CrossRef][ISI]
Hobbs, R. J. 1991 Disturbance a precursor to weed invasion in native vegetation. Plant Protection Quarterly 6: 99–104
———, and L. F. Huenneke. 1992 Disturbance, diversity and invasion: implications for conservation. Conservation Biology 6: 324–337[CrossRef][ISI]
———, and S. E. Humphries. 1995 An integrated approach to the ecology and management of plant invasions. Conservation Biology 9: 761–770[CrossRef][ISI]
Hurlbert, S. H. 1984 Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54: 187–211[CrossRef]
Kareiva, P. [ed.]. 1996 Special feature: developing a predictive ecology for non-indigenous species and ecological invasions. Ecology 77: 1651–1652[CrossRef][ISI]
Klinkhamer, P. G. L., and T. J. De Jong. 1988 The importance of small-scale disturbance for seedling establishment in Cirsium vulgare and Cynoglossum officinale. Journal of Ecology 76: 383–392
Lippincott, C. L. 1997 Ecological consequences of Imperata cylindrica (cogongrass) invasion in Florida sandhill. Ph.D. dissertation, University of Florida, Gainesville, Florida, USA
Luken, J. O., and J. W. Thieret [eds.]. 1997 Assessment and management of plant invasions. Springer-Verlag, New York, New York, USA
MacDonald, S. K., and I. E. Chandler. 1994 Element stewardship abstract for Imperata cylindrica cf. var. major. Nature Conservancy, Arlington, Virginia, USA
McConnaughay, K. D. M., and F. A. Bazzaz. 1987 The relationship between gap size and performance of several colonizing annuals. Ecology 68: 411–416[CrossRef][ISI]
Noble, I. R. 1989 Attributes of invaders and the invading process: terrestrial and vascular plants. In J. A. Drake, and H. A. Mooney [eds.], Biological invasions, a global perspective, 301–310. John Wiley and Sons, New York, New York, USA
Old, S. 1969 Microclimate, fire and plant production in an Illinois prairie. Ecological Monographs 39: 335–384[CrossRef][ISI]
Peart, D. R. 1989 Species interactions in a successional grassland. III. Effects of canopy gaps, gopher mounds and grazing on colonization. Journal of Ecology 77: 267–289[CrossRef]
Peet, R. K., and D. J. Allard. 1993 Longleaf pine vegetation of the Southern Atlantic and Eastern Gulf Coast regions: a preliminary classification. In S. M. Hermann [ed.], Proceedings of the 18th Tall Timbers Fire Ecology Conference, The longleaf pine ecosystem: ecology, restoration, and management, 45–81. Tall Timbers Research, Inc., Tallahassee, Florida, USA
———, M. R. anderson, and M. S. Adams. 1975 Effect of fire on big bluestem production. American Midland Naturalist 94: 15–26[CrossRef][ISI]
Pickett, S. T. A., and P. S. White. 1985 The ecology of natural disturbance and patch dynamics. Academic Press, San Diego, California, USA
Pierson, E. A., and R. N. Mack. 1990 The population biology of Bromus tectorum in forests: effect of disturbance, grazing, and litter on seedling establishment and reproduction. Oecologia 84: 526–533[ISI]
Reichenberger, G., and D. A. Pyke. 1990 Impact of early root competition on fitness components of four semiarid species. Oecologia 85: 159–166[CrossRef][ISI]
Rejmánek, M., and D. M. Richardson. 1996 What attributes make some plant species more invasive? Ecology 77: 1655–1661[CrossRef][ISI]
Rice, K. J. 1987 Interaction of disturbance patch size and herbivory in Erodium colonization. Ecology 68: 1113–1115[CrossRef][ISI]
Sajise, P. E. 1972 Evaluation of cogon [Imperata cylindrica (L.) Beauv.] as a seral stage in Philippine vegetational succession. I. The cogonal seral stage and plant succession. II. Autecological studies on cogon. Ph.D. dissertation, Cornell University, Ithaca, New York, USA
SAS. 1989 SAS/STAT user's guide, version 6, 4th ed., SAS Institute, Cary, North Carolina, USA
Scheiner, S. M., and J. Gurevitch [eds]. 1993 Design and analysis of ecological experiments. Chapman and Hall, New York, New York, USA
Shilling, D. G. 1996 Integrated management of cogongrass. Integrated Pest Management Florida, Winter 1996, 2(4): 1–2. University of Florida Cooperative Extension Service, Gainsville, Florida, USA
———, T. A. Bewick, J. F. Gaffney, S. K. McDonald, C. A. Chase, and E. R. R. L. Johnson. 1997 Ecology, physiology, and management of cogongrass (Imperata cylindrica). Final report prepared for the Florida Institute of Phosphate Research, Bartow, Florida, USA
———, J. F. Gaffney, and P. Waldrop. 1995 Cogongrass: problem and solutions. Alabama's treasured forests, Fall 1995: 8–9
Suding, K. N., and D. E. Goldberg. 1999 Variation in the effects of vegetation and litter on recruitment across productivity gradients. Journal of Ecology 87: 436–449
Swincer, D. E. 1986 Physical characteristics of sites in relation to invasions. In R. H. Groves and J. J. Burdon [eds.], Ecology of biological invasions, 67–76. Cambridge University Press, London, UK
Tabor, P. 1949 Cogon grass, Imperata cylindrica (L.) Beauv. in the southeastern United States. Agronomy Journal 41: 270
United States Department Of Agriculture. 1997 PLANTS database. http://plants.usda.gov. National Plant Data Center, Baton Rouge, Louisiana, USA (7 April 1998)
Williamson, M. 1996 Biological invasions. Chapman and Hall, New York, New York, USA
Winn, A. A. 1985 Effects of seed size and microsite on seedling emergence of Prunella vulgaris in four habitats. Journal of Ecology 73: 831–840[CrossRef]
Zar, J. H. 1996 Biostatistical analysis, 3rd ed. Prentice Hall, Inc., Upper Saddle River, New Jersey, USA
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
