Controlled experiments to predict horseweed (Conyza canadensis) dispersal distances

Share Embed


Descrição do Produto

Weed Science, 54:484–489. 2006

Controlled experiments to predict horseweed (Conyza canadensis) dispersal distances Joseph T. Dauer

Corresponding author. Department of Crop and Soil Science, Pennsylvania State University, University Park, PA 16802; [email protected]

David A. Mortensen

Department of Crop and Soil Science, Pennsylvania State University, University Park, PA 16802

Robert Humston

Department of Biology, Virginia Military Institute, Lexington, VA 24450

Controlled-environment experiments were conducted to predict the dispersal distance of horseweed seed. Seed were released from a fixed height and collected at three distances from the introduction point along a 6-m wind tunnel. Dispersal potential was assessed at wind speeds of 8 and 16 km hr21 and release heights of 50.8 and 76.2 cm. In separate experiments, settlement velocity was determined to be 0.323 m sec21 (SD 5 0.0687). These data were used to parameterize a mechanistic model and compared to a quantile extrapolation (QE) of wind-tunnel results. The QE method predicted a greater mean dispersal distance than the mechanistic model, with large disparities between maximum dispersal distances. Quantile extrapolation predicted dispersal distances over 100 m, whereas the mechanistic model predicted a maximum distance of approximately 30 m. Air turbulence within the wind tunnel and complex dynamics of seed flight may have contributed to the discrepancy between models. Predicting the mean and numerical distribution of seed dispersal distance is crucial when estimating the spread of wind-dispersed seed and for the design of a field-sampling protocol. Although controlled-environment experiments lack the wind variability present in natural systems, predictions from windtunnel studies provide a better first approximation of dispersal distance than the mechanistic model. Field experiments designed on the basis of these outcomes are more likely to capture the true dispersal distribution. This should provide more accurate data to inform management decisions for wind-dispersed species. Nomenclature:

Horseweed, Conyza canadensis L. Cronq., ERICA.

Key words: Long-distance dispersal, wind dispersal, settlement velocity, glyphosate resistance, seed movement.

Quantifying long-distance dispersal (LDD) is inherently difficult because detection probabilities decline rapidly with distance from a source population (Thompson 2002). For plant species with few special adaptations for LDD (e.g., gravity-dispersed seeds), it has typically been found that most seeds disperse within a short distance of the parent plant (Willson 1993). Conversely, dispersal distances of wind-dispersed seeds may be on the order of tens of meters to kilometers. Recent work on seed dispersal of trees (Augsburger 2003; Clark 1998; Horn et al. 2001; Medjibe and Hall 2002; Nathan et al. 2001) and forbs (Bullock and Clarke 2000; Jongejans and Telenius 2001; Tackenberg et al. 2003a; 2003b; vanDorp et al. 1996) has made substantial progress in describing some of the environmental drivers that shape dispersal curves. Pollen dispersal has also received considerable attention due to concern about intra- and interspecific pollen flow (Aylor 2002; Aylor et al. 2003; Giddings et al. 1997; Rieger et al. 2002; Scheffler et al. 1993; Watrud et al. 2004). The majority of LDD research has thus far been largely phenomenological, rarely assessing the potential predictive ability of controlled environment experiments (but see Jongejans and Schippers 1999; Jongejans and Telenius 2001; Tackenberg et al. 2003a). Controlled environment (i.e., wind tunnel) experiments are commonly used in other biological sciences to explore wind-driven processes. For example, in entomology they have been used to quantify the behavioral response of insects to pheromones (Byers 1988; Kennedy 1983; Zanen and Carde` 1999) and wind speed and direction (Bohm 1995). 484



Weed Science 54, May–June 2006

A few recent studies have used wind tunnels to examine seed dispersal. VanDorp et al. (1996) quantified dispersion of six perennial grassland species that lacked special adaptation to wind transport using wind gusts up to 50 km hr21. Jongejans and Schippers (1999) used wind-tunnel experiments to predict dispersal distances of seeds at variable wind speeds. They demonstrated that these experiments could be used to predict dispersal dynamics of 10 Apiaceae species and guide development of a field-sampling protocol. Horseweed is a winter annual, wind-dispersed composite that has become increasingly problematic in agricultural fields as well as field edges and roadsides throughout the United States. Mature plants reach heights of 1.8 m, and can produce hundreds of thousands of seeds (Buhler 1992; Bhowmik and Bekech 1993; Weaver 2001). Mulligan and Findlay (1970) showed that horseweed is self-compatible and Smisek (1995) found an average 96% of florets were self-pollinated, which the author attributed to pollen being released before capitula are fully open, encouraging exposure of their own pollen before pollen from other plants can reach the stigma. Like many wind-dispersed plant species, seeds of horseweed have a small (2–3 mm) achene and pappus to slow settlement velocity. Andersen (1993) reported a settlement velocity of 0.278 m sec21 (SE 5 0.0564) for horseweed, the lowest of 19 composite species analyzed in the study. Slow settlement velocity increases the period during which wind can affect seed movement. Although horseweed dispersal is not well quantified, Regehr and Bazzaz (1979) documented that seeds may travel more than 100 m in corn (Zea mays L.).

TABLE 1. Total seed collected (N ), mean, and standard deviation of seed height (s) in the wind tunnel at each wind speed, release height, and collection distance.

FIGURE 1. Schematic image of wind-tunnel apparatus and dimensions. Positions of screens, dispersal tube, wind-speed sensor, fan, and baffles are approximate, but proportionate relative to dimensions of the wind tunnel.

Quantifying horseweed dispersal has become important because glyphosate resistance was recently confirmed in this species in Delaware (VanGessel 2001). Primarily found in soybean [Glycine max (L.) Merr.] fields, 12 states have reported glyphosate-resistant horseweed (Heap 2005). The existence of a wind-dispersed, glyphosate-resistant weed threatens the efficacy of a widely adopted weed management practice and will likely increase the cost of management. Understanding factors that influence the rate of spread of this species could help direct in-field and area-wide resistance management strategies. Establishing an effective field-sampling protocol is critical to quantifying dispersal distance (Skarpaas et al. 2005). Controlled-environment experiments can provide first approximations of dispersal distances and increase seed sampling efficiency in field tests. The objective of this study was to construct and compare two models predicting the dispersal distance of horseweed seed: a mechanistic model derived from measured settlement velocity, and a QE method derived from data collected in windtunnel experiments. The models predict mean and greatest dispersal distances that will provide insight into the dispersal capabilities of this species and help to better design field dispersal studies.

Materials and Methods Horseweed plants from the same seed source were greenhouse grown in 15-cm pots under ambient light from January to July, 2002 (wind-tunnel experiments) and October to March, 2003–2004 (settlement-velocity experiments). Plants were rotated among greenhouse benches twice weekly to reduce effects of environmental variation on plant growth and development. Plants were between 1 and 1.5 m tall and in full seed set when experiments were conducted.

Settlement-Velocity Experiments Settlement-velocity experiments were conducted in a laboratory with low humidity (14 to 19%) and followed the protocol used by Andersen (1993). Individual seed from 12 plants (n 5 111) were removed from the capitula and seeds with an intact pappus were dropped through a 76-cm clear vertical plexiglass tube with a sealed lid containing a 2.5-cm entry hole. Air movement was minimal within the tube and settlement velocity was determined by timing the duration of the descent.

Wind-Tunnel Experiments Wind-tunnel experiments were conducted in a wind tunnel measuring 90 cm high and 60 cm wide with a 6-m-long

Wind speed

Release height

Collection distance

kmph 8 8 8 8 8 8 16 16 16 16 16 16

50.8 50.8 50.8 76.2 76.2 76.2 50.8 50.8 50.8 76.2 76.2 76.2

133 272 455 133 272 455 133 272 455 133 272 455

Mean seed height

SD seed height

N

38.1 27.8 23.7 64.0 46.1 39.8 49.8 47.0 41.7 72.0 70.8 62.4

8.48 12.06 11.35 6.61 8.44 11.50 7.06 9.40 10.77 6.28 7.59 9.29

471 602 801 534 461 546 738 806 673 492 657 623

cm

test section (see Figure 1). Air was moved through a closed circuit by a fan and baffles were inserted upwind of the test section to reduce wind turbulence (Leon et al. 1998). Wind speed was monitored with the use of a Series 475 Mark III Digital Manometer1 located approximately 30 cm downwind of the beginning of the test section and suspended from the ceiling. Seeds were introduced into the tunnel via an 80-cm-long, 7.6-cm-diameter PVC tube inserted through the tunnel ceiling approximately 60 cm downwind of the test section entrance. The release height of the seeds was measured from the bottom of the wind tunnel to the point where seed exited the release tube. Fiberglass mesh screens (1.6 mm) were coated with a sticky resin (TangleTrap2 aerosol) and positioned at 133, 272, or 455 cm downwind of the release point to trap seeds and preserve their vertical distribution in the tunnel. Seven to ten randomly chosen capitula from 15 plants were collected, and their seed (;200) were released into the introduction tube once a constant wind speed was reached. Horseweed seeds reach terminal velocity nearly instantaneously (Andersen 1993) and descended very little upon exiting the tube before being moved horizontally with the wind (J. Dauer, personal observation). Wind speed, release height, and distance of collection screen from release point were manipulated in a completely randomized design replicated three times. Seeds were counted and assigned a position in a 2.5 by 2.5-cm grid (24 by 36). Vertical profiles of seed distribution at each distance were combined to provide a sequential representation of the downwind seed travel for each treatment (Table 1, Figure 2). The linear model representing seed movement through the tunnel was defined as a function of wind speed (W, kmph) and distance from the release point (D, cm): H(W,D) 5 R 1 a · W 1 b · D 1 t · W · D 1 « [1] where H is height of seed at the collection distance (cm), R is the seed release height (cm), and « is the error associated with the linear model. The constants: a, b, and t, were estimated by fitting the general linear model to the data collected with the intercept fixed at the release height (R Development Core Team 2005). Average height of seed on the collection screen at each distance was compared over Dauer et al.: Conyza canadensis seed dispersal



485

TABLE 2. Mean, standard deviation, and greatest dispersal distance of the lognormal distribution of dispersal distances from the mechanistic model (Equation 2) and the mean, standard deviation, and greatest dispersal distance from the quantile extrapolation model. Quantile extrapolation

Mechanistic model Wind speed

FIGURE 2. Mean (6SD) for 50.8-cm release height (V) and 76.2-cm release height (n) at 8 kmph (unfilled symbols) and 16 kmph (filled symbols). Although seed were collected at fixed distances, distance values have been offset slightly to improve the visualization of means and errors.

wind speed and release height with the use of Tukey’s honest significant difference test.

Simulation Studies We used results of settlement-velocity observations to parameterize a simple, mechanistic model of individual seed dispersal. A mechanistic model for an object falling at a constant velocity can be defined as: D5

R ·W , F

[2]

with D the predicted dispersal distance (m), R the release height (m), W the constant wind speed (m sec21), and F a normal variate representing settlement velocity (m sec21). Values of R and W reflected wind speeds and release heights tested in wind-tunnel experiments. This model was assumed to approximate the data collected in wind-tunnel experiments since R and W were constant and F would vary equally in wind-tunnel and settlement-velocity experiments. To simulate a single seed dispersal event, a value of F was randomly selected from the normal probability distribution of settlement velocities quantified in fall rate experiments. Ten thousand individual dispersal events were simulated for each treatment. The resulting distributions of D quantified seed deposition for the mechanistic model. A third release height of 150 cm was included to simulate dispersal of seed from full-sized plants. Simulated dispersal events resulted in a frequency histogram of seeds as a function of distance from the release point. A lognormal probability distribution:

y5

1 Ï2p·s

exp

5

21·[ln(x) 2 x¯ ] 2 2·s 2

x

6

[3]

where x is distance from the source (m), and y is the probability of a seed reaching that distance, was fitted to the frequency histogram with the use of nonlinear least squares (R Development Core Team 2005). The mean distance (m) 486



Weed Science 54, May–June 2006

Release height

Mean

SD

Longest distance

kmph cm 50.8 8 76.2 8 150.0 8 50.8 16 76.2 16 150.0 16

3.8 5.4 10.6 7.2 10.8 21.1

1.01 1.22 1.22 1.22 1.22 1.21

15 54 65 65 29 104

Mean

SD

Longest distance

58.87 6.14 12.09 38.37 120.40 237.01

740 65 101 141 370 728

m 14.1 10.1 19.9 36.1 78.3 154.0

and standard deviation (s) associated with a treatment were recorded for comparison with wind-tunnel results (Table 2). Data collected in the wind tunnel were extrapolated to estimate dispersal distance. Seed heights on collection screens were assumed to be normally distributed and defined by the mean and standard deviations (Table 1, Figure 2). Quantiles were then determined and used to define the range in seed heights at the first seed trap (133 cm). Heights were then selected from a normal distribution at the second (272 cm) and third (455 cm) seed traps to match the quantile from the first location. A line was then fitted through the release height and heights at the three locations with the ues of general linear model methods (R Development Core Team 2005). We thus extrapolated the quantiles of the horizontal (seed rain) distribution by projecting seed dispersal paths along corresponding quantiles of the vertical height data on collection screens. The increasing variation in vertical distribution of seed with distance from the release point resulted in some slope values greater than zero (infinite dispersal distances). Because it is likely that positive slopes are an artifact of the statistical analysis, they were not used in determining the mean or longest dispersal distance. So as not to underestimate dispersal distance, the quantity of trajectories with the most negative slope (shortest dispersal distance) equal to those with a positive slope were also not included in calculation of the mean. Mean dispersal distance and standard deviation were calculated from the truncated data set (Figure 3, Table 2). The physical limitations of the wind tunnel restricted the potential release heights that could be used in the study. To simulate a 150-cm seed release height, seed distribution data from the 76.2-cm dispersal height were offset upward by 73.8 cm. Seed heights were selected and dispersal distances predicted using the method described above. Frequency histograms of dispersal distance were calculated for both settlement-velocity–based mechanistic model and quantile extrapolation. For the QE, a variable-width histogram was used to represent frequency of seed deposition (vertical height of bars) over discrete distance intervals (i.e., quantiles, horizontal width of bars).

Results and Discussion Both wind-tunnel and settlement-velocity experiments can be useful in quantifying dispersal distance, but this research shows large differences in estimated distance based

FIGURE 3. Quantile extrapolation of wind-tunnel collected seed for 8 kmph (left) and 16 kmph (right) and release heights of 50.8 cm (top), 76.2 cm (middle), and 150 cm (bottom). Area between lines represents quantiles (n 5 20). The x axis has been rescaled for the 150-cm release height to better represent the predicted dispersal pattern. Mean values are denoted by arrows (↑).

FIGURE 4. Comparison of predicted dispersal distances with the use of the quantile-extrapolation (grey shaded bars) and settlement-velocity (black bars) methods at 8 kmph (left) and 16 kmph (right) and release heights of 50.8 cm (top), 76.2 cm (middle), and 150 cm (bottom). The x axis has been rescaled for the 16 kmph, 150-cm release height to better represent the predicted dispersal pattern.

on the type of controlled-environment experiment that is conducted. Settlement-velocity experiments are occasionally conducted as a surrogate measure of dispersal ability, and only rarely as a basis for field-level experiments (Matlack 1987; Andersen 1993; Jongejans and Telenius 2001). Settlement velocities ranged from 0.134 to 0.512 m sec21 and were normally distributed (W 5 0.9931, P 5 0.85) with a mean of 0.323 m sec21 and standard deviation of 0.0687 m sec21. These values represent a considerably faster settlement velocity than the rate of 0.2778 m sec21 6 0.0564, mean 6 SD reported by Andersen (1993), though the variance is nearly identical. Were the settlement velocities too fast? It is possible that seed weight played a role, because Andersen used field-collected seeds, whereas this study used greenhouse-grown seeds. Greenhouse plants were grown in an optimal environment; it is likely nearly all seeds were filled, increasing the settlement velocity. Parameterizing the mechanistic model with the lower settlement velocity reported by Andersen increased mean dispersal distances only 1 to 2 m in each simulation. However, settlement velocity would need to be reduced by approximately 50% for distances predicted by the mechanistic model to approach those estimated by the QE (Table 2). The linear model was significant for all wind-speed and release-height treatments and the fitted model (constants 6 SE) was

Simulation studies using the mechanistic model found that increases in release height or wind speed resulted in a proportional increase in mean dispersal distance to a maximum of 21 m (Table 2). Mean dispersal distances estimated with the QE method were 2–7 times greater than those estimated with the mechanistic model, ranging from 10 to 154 m, also increasing with wind speed and height (Table 2, Figure 4). Mean seed height along the wind tunnel decreased (Table 1, P , 0.001), whereas variance increased (adj. R2 5 0.544, P , 0.01) at each successive distance for both release heights and wind speeds. The QE was affected by the increasing variance in seed height with distance from the seed source resulting in greater estimated dispersal distances. The longest distances were often associated with a few seeds collected high on the screens at the first location. For example, at 8 kmph and 50.8-cm release height, three seeds were recorded at the first screen location at a height nearly identical to the release height. In this case, the projected dispersal distance of these seeds was 740 m, skewing the mean and standard deviation (Table 2). The longest dispersal distances estimated by the mechanistic model were generally below 65 m, whereas the QE predicted greater distances, sometimes by an order of magnitude, and ranging from 65 to 740 m. The large discrepancy in dispersal distances between models is the likely result of one of two things—turbulence in the wind tunnel and dynamics in the flight characteristic of the seed—which are poorly represented by settlement-velocity measurements. Capone and Lauchle (2000) found that

H 5 R 2 0.049(60.022)·W 2 0.128(60.001)·D 1 0.007(60.0001)·W·D.

Dauer et al.: Conyza canadensis seed dispersal



487

unsteady pressure levels were greatest at the ends of a cylinder and on the downwind side when a constant flow is applied perpendicular to the length. Although baffles in the tunnel were intended to minimize turbulence, the low-pressure zone downwind of the release tube may have resulted in stochastic air movement and an ‘‘incoherent recirculation zone’’ (Capone and Lauchle 2000) that could affect seed movement. At the lowest release height (50.8 cm) there were 5 and 20% (8 and 16 kmph, respectively) of the seeds moving upwards (slope greater than zero), suggesting that the increased length of pipe present in the tunnel may have contributed to turbulent air movement (Figure 3). Simulating the effects of turbulence in a mechanistic model (Equation 2) would likely result in longer predicted distances, but without observing individual seed movements, any additional parameters to the model would have lacked a realistic basis. Settlement-velocity studies can be conducted repeatedly and with accuracy to provide a first approximation of dispersal ability. Unfortunately, settlement velocity may not capture the complex dynamics of wind dispersal that windtunnel experiments may better approximate. In addition to turbulence, little is known about the flight characteristics of individual seed. In the mechanistic model, wind speed and release height are fixed and the normal variate of the measured velocity is used to calculate the dispersal distribution of seed. This assumes the aerodynamics of seed movement for seed dropping vertically are similar to those moving horizontally. Calculations of settlement velocity based on windtunnel results find that increasing the wind speed lowers the settlement velocity to 0.212 m sec21 at 8 kmph and 0.033 m sec21 at 16 kmph for seed released at 50.8 cm. Similarly, the calculated settlement velocity for seeds released at 76.2 cm, decreases to 0.203 m sec21 at 8 kmph and 0.140 m sec21 at 16 kmph. This may be the result of turbulence. It is also possible that seed flight efficiency is enhanced under conditions of heavy winds, where the pappus orientation to the wind will result in significantly greater dispersal distances than those estimated with the settlement velocity. Continued research of individual seed movement in air may provide insight into the relative effects of seed orientation and turbulence. Both modeling approaches provide a foundation for fieldbased experiments. Cain et al. (2000) noted that few procedures are able to ‘‘maximize the chance of observing longdistance seed dispersal events.’’ Often, research does not report how collection distances are selected (Baker et al. 1986; Bullock and Clarke 2000; McEvoy and Cox 1987; Regehr and Bazzaz 1979) and it is possible that sampled distances do not represent potential dispersal distances. For example, McCallum (1989) collected musk thistle (Carduus nutans L.) seed at 1.7 m from the source population, the furthest extent of the sampling design. In contrast, a larger sampling effort has found that musk thistle seed can travel 96 m and possibly farther (O. Skarpaas, unpublished data). Although the predicted distances for the QE method may be greater than realized seed movement for horseweed, sampling to the longest distances increases the likelihood of capturing rare long-distance events. For an average plant that produces 130,000 seeds, 6,500 seeds (95th percentile) may be considered long-distance dispersers that can readily begin a new infestation in an adjoining field. Sampling to the longest 488



Weed Science 54, May–June 2006

distances predicted using settlement-velocity experiments might miss the majority of seed that are being dispersed. Therefore model selection is a critical step in designing field studies. Controlled-environment studies allow researchers to direct sampling efforts focused on capturing long-distance dispersal events, which are critical for determining rate of spread (Portnoy and Willson 1993; Kot et al. 1996). Increasing the chances of capturing rare events, that is, sampling to 150 m or beyond, will provide better information about the dynamics of the dispersal curve for horseweed. Wind-tunnel experiments alone cannot substitute for field studies, but insights from such experiments can help determine effective field-sampling designs to quantify dispersal distances. The complex nature of air movement in the field, including gusts, updrafts, and boundary-layer interactions, affect the dispersal distance of wind-borne objects. In this work we estimate that horseweed seed may move hundreds of meters, estimates corroborated by recent field studies (Dauer et al. 2004). Continued work must examine the role of complex air movement on seed dispersal distances. As a discipline, the repercussions of poor weed management are thought to impact the field in which the infestation resides. Long-distance dispersal effectively decouples weed management and the resulting weed infestation. In this way, wind-dispersed weeds connect a heterogeneous matrix of agricultural fields. Although not practiced commonly in weed management, area-wide management (coordinated at a scale larger than the individual field) may be the only way to slow the spread of glyphosate-resistant horseweed.

Sources of Materials 1

Series 475 Mark III Digital Manometer, Dwyer Instruments, P.O. Box 373, 102 Indiana Hwy. 212, Michigan City, IN 46361. 2 TangleTrap2 aerosol, The TangleFoot Company, 314 Straight Avenue, S.W., Grand Rapids, MI 49504-6485.

Acknowledgments We thank S. Shirtliffe and two anonymous reviewers for their insightful comments on earlier versions of the manuscript. We would like to thank the Pennsylvania State University College of Engineering and R. Auhl, who assisted us in operation of the wind tunnel, B. Jones and N. Peskin for logistical assistance, and O. Bjornstad, K. Shea, W. Curran, and O. Skarpaas for their many helpful suggestions.

Literature Cited Andersen, M. C. 1993. Diaspore morphology and seed dispersal in several wind-dispersed Asteraceae. Am. J. Bot. 80:487–492. Augsburger, C. K. 2003. Morphology and dispersal potential of wind-dispersed diaspores of neotropical trees. Am. J. Bot. 73:353–363. Aylor, D. E. 2002. Settling speed of corn (Zea mays) pollen. J. Aerosol Sci. 33:1601–1607. Aylor, D. E., N. P. Schultes, and E. J. Shields. 2003. An aerobiological framework for assessing cross-pollination in maize. Agric. For. Meteorol. 119:111–129. Baker, P. S., A.S.T. Chan, and M.A.J. Zavala. 1986. Dispersal and orientation of sterile Ceratitis capitata and Anastrepha ludens (Tephritidae) in Chiapas, Mexico. J. Appl. Ecol. 23:27–38. Bhowmik, P. C. and M. M. Bekech. 1993. Horseweed (Conyza canadensis) seed production, emergence, and distribution in no-tillage and conventional-tillage corn (Zea mays). Agronomy (Trends Agric. Sci.) 1: 67–71.

Bohm, H. 1995. Dynamic properties of orientation to turbulent air current by walking carrion beetles. J. Exp. Biol. 198:1995–2005. Buhler, D. D. 1992. Population dynamics and control of annual weeds in corn (Zea mays) as influenced by tillage systems. Weed Sci. 40:241– 248. Bullock, J. M. and R. T. Clarke. 2000. Long distance seed dispersal by wind: measuring and modeling the tail of the curve. Oecologia 124: 506–521. Byers, J. A. 1988. Upwind flight orientation to pheromone in Western pine-beetle Dendroctonus brevicomis Lec (Coleoptera, Scolytidae) tested with rotating windvane Traps. J. Chem. Ecol. 14:189–198. Cain, M. L., B. G. Milligan, and A. E. Strand. 2000. Long-distance seed dispersal in plant populations. Am. J. Bot. 87:1217–1227. Capone, D. E. and G. C. Lauchle. 2000. Modeling the unsteady lift and drag on a finite-length circular cylinder in cross-flow. J. Fluids Struct. 14:799–817. Clark, J. S. 1998. Why trees migrate so fast: confronting theory with dispersal biology and the paleorecord. Am. Nat. 152:204–224. Dauer, J. T., D. A. Mortensen, B. P. Jones, and M. Van Gessel. 2004. Long-distance aerial transport of horseweed (Conyza canadensis) seed. Weed Sci. Soc. Am. Abstr. 44:267. Giddings, G. D., N.R.S. Hamilton, and M. D. Hayward. 1997. The release of genetically modified grasses 1. Pollen dispersal to traps in Lolium perenne. Theor. Appl. Genet. 94:1000–1006. Heap, I. 2005. International weed resistance monitoring site. http:// www.weedscience.com. Horn, H. S., R. Nathan, and S. R. Kaplan. 2001. Long-distance dispersal of tree seeds by wind. Ecol. Res. 16:877–885. Jongejans, E. and P. Schippers. 1999. Modeling seed dispersal by wind in herbaceous species. Oikos 87:362–372. Jongejans, E. and A. Telenius. 2001. Field experiments on seed dispersal by wind in ten umbelliferous species (Apiaceae). Plant Ecol. 152:67– 78. Kennedy, J. S. 1983. Zigzagging and casting as a programmed response to wind-borne odour: a review. Phys. Entomol. 8:109–120. Kot, M., M. A. Lewis, and P. Van den Driessche. 1996. Dispersal data and the spread of invading organisms. Ecology 77:2027–2042. Leon, J. C., B. Babin, and C. Y. Choi. 1998. Design, construction, and evaluation of a compact recirculating wind tunnel for agricultural experiments. Trans. ASAE 41:213–218. Matlack, G. R. 1987. Diaspore size, shape, and fall behavior in winddispersed plant species. Am. J. Bot. 74:1150–1160. McCallum, K. 1989. Seed biology of nodding thistle and its interaction with biological control. University of Canterbury, Christchurch, New Zealand. McEvoy, P. B. and C. S. Cox. 1987. Wind dispersal distances in dimorphic achenes of Ragwort, Senecio jacobaea. Ecology 68:2006–2015. Medjibe, V. and J. S. Hall. 2002. Seed dispersal and its implications for silviculture of African mahogany (Entandrophragma spp.) in undis-

turbed forest in the Central African Republic. For. Ecol. Manage. 170: 249–257. Mulligan, G. A. and J. N. Findlay. 1970. Reproductive systems and colonization in Canadian weeds. Can. J. Bot. 48:859–860. Nathan, R., U. N. Safriel, and I. Noy-Meir. 2001. Field validation and sensitivity analysis of a mechanistic model for tree seed dispersal by wind. Ecology 82:374–388. Portnoy, S. and M. F. Willson. 1993. Seed dispersal curves—behavior of the tail of the distribution. Evol. Ecol. 7:25–44. R Development Core Team. (2005). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-00-3. http://www.R-project.org Regehr, D. L. and F. A. Bazzaz. 1979. The population dynamics of Erigeron canadensis, a successional winter annual. J. Ecol. 67:923–933. Rieger, M. A., M. Lamond, C. Preston, S. B. Powles, and R. T. Roush. 2002. Pollen-mediated movement of herbicide resistance between commercial canola fields. Science 296:2386–2388. [SAS] Statistical Analysis Systems. 1999. SAS/STAT User’s Guide. Version 8. Cary, NC: Statistical Analysis Systems Institute. Scheffler, J. A., R. Parkinson, and P. J. Dale. 1993. Frequency and distance of pollen dispersal from transgenic oilseed rape (Brassica napus). Transgenic Res. 2:356–364. Skarpaas, O, K. Shea, and J. M. Bullock. 2005. Optimizing dispersal study design by Monte Carlo simulation. J. Appl. Ecol. 42(4):731–739. Smisek, A. 1995. The evolution of resistance to paraquat in populations of Erigeron canadensis L. M.S. thesis. University of Western Ontario, London, Canada. Tackenberg, O., P. Poschlod, and S. Bonn. 2003a. Assessment of wind dispersal potential in plant species. Ecol. Monogr. 73:191–205. Tackenberg, O., P. Poschlod, and S. Kahmen. 2003b. Dandelion seed dispersal: the horizontal wind speed does not matter for long-distance dispersal—it is updraft! Plant Biol. 5:451–454. Thompson, S. K. 2002. Sampling. 2nd ed. New York: Wiley. 367 p. vanDorp, D., W.P.M. vandenHoek, and C. Daleboudt. 1996. Seed dispersal capacity of six perennial grassland species measured in a wind tunnel at varying wind speed and height. Can. J. Bot. 74:1956–1963. VanGessel, M. J. 2001. Glyphosate-resistant horseweed from Delaware. Weed Sci. 49:703–705. Watrud, L. S., E. H. Lee, A. Fairbrother, C. Burdick, J. R. Reichman, M. Bollman, M. Storm, G. King, and P. K. Van de Water. 2004. Evidence for landscape-level, pollen-mediated gene flow from genetically modified creeping bentgrass with CP4 EPSPS as a marker. Proc. Nat. Acad. Sci. USA 101:14533–14538. Weaver, S. E. 2001. The biology of Canadian weeds. 115. Conyza canadensis. Can. J. Plant. Sci. 81:867–875. Willson, M. F. 1993. Dispersal mode, seed shadows, and colonization patterns. Vegetation 108:261–280. Zanen, P. O. and R. T. Carde`. 1999. Directional control by male gypsy moths of upwind flight along a pheromone plume in three wind speeds. J. Comp. Physiol. A 184:21–35.

Received February 10, 2005, and approved March 13, 2006.

Dauer et al.: Conyza canadensis seed dispersal



489

Lihat lebih banyak...

Comentários

Copyright © 2017 DADOSPDF Inc.