Abstract
Asclepias (milkweed) habitat loss threatens Danaus plexippus (monarch butterfly) survival. Restoration plantings of milkweeds are needed to support monarch butterfly conservation. We investigated the effects of microsite (open gaps, edges of canopy, and under canopy), seed cover (2 cm sand or scrub oak leaves), and protection (no protection, PVC pipe, wire mesh, or PVC pipe and wire mesh) on the recruitment of Asclepias humistrata (sandhill milkweed). Seeds were sown in back dunes of the northern Gulf of Mexico coast in June and observed for 52 weeks. Mixed model analyses indicate recruitment was influenced by an interaction of microsite and seed cover and the response differed over time. Recruitment was greater in open areas than under canopy 52 weeks after seeding (WAS). Recruitment remained low in open areas and scrub edges until spring (38 WAS). Initial recruitment under canopy in the summer decreased by fall and at 52 WAS was only 2.75%. Recruitment was highest 52 WAS in open areas followed by canopy edges when seeds were covered with sand. Protection interacted with seed cover independent of microsites. PVC pipe alone was the only protection treatment with significantly higher emergence than no protection (only sand cover). There were no significant differences in emergence among the protection treatments when seeds were covered by litter. The growth of seedlings did not differ. The seedlings were small (<2 cm tall and <3 nodes per plant) 52 WAS. We recommend direct seeding Asclepias humistrata in open gaps and lightly covering seeds (2 cm) with sand.
Restoration Recap
Asclepias humistrata (sandhill milkweed) revegetation is needed to support Danaus plexippus (monarch butterfly) conservation.
We tested the effects of microsite and seed cover on seedling recruitment of Asclepias humistrata. In addition, we tested the effects of seed site protection to understand seed predation and stability of seeds within planting sites and to understand effects of these protection treatments on seedling recruitment.
Seedling recruitment was higher in open areas and at scrub oak canopy edges than under scrub oak canopies, and higher when covered with sand than leaves. Seedling recruitment under these conditions was not influenced by protection.
Protection treatments demonstrated no contribution of extant seedbank to emergence and no need to provide protection to prevent seed movement or loss.
We recommend placing seeds within backdunes in open, bare sand areas (gaps) keeping seeds outside the edge of scrub canopies and covering lightly (2 cm) with sand.
The need for restoration of coastal dune ecosystems is growing as degradation and loss associated with increased anthropogenic threats accelerates (Martinez et al. 2013, Lithogow et al. 2015). Increasing tropical storm intensity associated with human induced climate change and growing human developments and recreation in coastal areas is expected to continue. Coastal dunes are among the most vulnerable ecosystems resulting in frequent damage and need for restoration (Nordstrom 2008, Lithogow et al. 2015). Development for recreation and human dwellings has resulted in reduction and fragmentation of dune habitats for rare and endemic species. Strong hurricanes have further damaged remaining foredune and backdunes of Florida’s Gulf and Atlantic coasts, eroding and destabilizing dunes and decreasing plant diversity in some locations.
Backdunes along the northern Gulf Coast support numerous endemic and rare species that require protection. These backdunes are generally larger, older and more diverse than foredunes (Bitton and Hesp 2013). They are often eroded and continue to erode because of overwash associated with tropical storm activity and human recreation. Restoration of these dunes often includes stabilization by planting Uniola paniculata (sea oats) but these restoration efforts infrequently include re-establishment of a diversity of plant species. Additional focal species that provide resources for pollinators and other fauna are needed.
Asclepias spp. (milkweed) populations are in decline throughout North America (Hartzler 2010, Pleasants and Oberhauser 2013) from habitat loss, competition with invasive species, and increased herbicide use (Pleasants and Oberhauser 2013). An estimated 1.49 billion stems of eastern USA Asclepias, roughly 21% of stems, have disappeared between 1995 and 2013 (Flockhart et al. 2015). A decline in monarch butterfly (Danaus plexippus) [Lepidoptera: Nymphalidae] populations has been attributed to Asclepias decline. Flockhart et al. (2015) state that reducing Asclepias loss in monarch butterfly breeding grounds is a top priority for monarch butterfly conservation. In the last decade, the eastern population of monarch butterflies has diminished by over 80% and it now faces quasi-extinction (Semmens et al. 2016) and possible extinction (Flockhart et al. 2015).
Models indicate mass milkweed replanting is necessary to restore eastern Asclepias spp. populations (Thogmartin et al. 2017). Asclepias humistrata (sandhill milkweed) [Apocynaceae] is a well-documented host for monarch butterfly larvae (Brower 1961, Brower 1962, Cohen and Brower 1982, Malcolm et al. 1987, Martin et al. 1992, Zalucki and Brower 1992, Knight et al. 1999, Zalucki et al. 2001, Brower et al. 2018) and is considered one of the most important Asclepias species in the Southeast United States Coastal Plain for migratory spring breeding monarch butterflies (Daniels et al. 2018). In the Florida panhandle, large populations of A. humistrata which host monarch butterflies occur on coastal dunes within the highly fragmented, degraded, and developed coastal landscape (Campbell 2020). These populations have been impacted recently by hurricanes and remain vulnerable to further erosion and loss. Asclepias humistrata is the only species of milkweed on dunes of the northern gulf. Coastal A. humistrata occur mainly on relatively intact backdunes where plants begin above-ground growth in early March, flower en masse mid-March to early-May, and fruits are mature (seed dispersed) beginning late-May (Campbell 2020). A much smaller flowering and fruiting event occurs from July to September with peak maturation from November through dormancy (Campbell 2020). Asclepias humistrata are also an important component of the sandhills plant community of longleaf pine ecosystems, another endangered plant community in need of restoration.
Direct seeding is a common restoration technique where seeds are strategically placed within a restoration site (Palma and Laurance 2015). This option is often considered cheaper and easier than the use of seedlings for restoration outplanting, but direct seeding protocols are not described for most species limiting its practical application (Ceccon et al. 2015); this is also the case for A. humistrata. Considerations in the planning of a direct seeding project include identifying appropriate seeding locations, seed burial methods, and the use of protective equipment around seeds (Ceccon et al. 2015, Palma and Laurance 2015). Defining restoration success is important too. The average seedling establishment success, defined as the percentage of total seedlings installed that survive until a defined date, within a meta-analysis of direct-seeding papers was calculated to be 11% (Ceccon et al. 2015). As such, a seedling establishment of 11% or higher can be considered a restoration success.
Asclepias humistrata occur in several back dune microsites including open sand gaps and in association with scrub oaks (Campbell 2020). Moisture is often limited in open sand of coastal dunes (Miller et al. 2003), therefore shading and/or covering by sand or leaf litter may provide increased moisture for germination and establishment. In addition, seeds may require protection from ants, beach mice, and wind for recruitment at the location where placed. Our objective was to describe the effects of microsite (open vs drip line of oaks vs oak canopy), seed cover (sand vs leaf litter vs none), and protection (cage vs no cage) on seedling establishment on coastal dunes over time. Because mature A. humistrata were most frequently found in open microsites, not associated with leaf litter and oak canopies and beach mouse tracks and ant mounds were frequent on these back dunes we hypothesized seedling recruitment would be higher when seedlings were placed in open microsites, covered with sand, and protected.
Methods
We conducted an experiment to test the effects of seedling recruitment (seedling which emerged and survived) for A. humistrata direct seeded within the Bon Secour National Wildlife Refuge in Alabama on June 9, 2019. Seeds were placed on backdunes adjacent to coastal scrub in areas where A. humistrata occurs in situ (Campbell 2020). The experimental design was a split-plot randomized complete block with a full-factorial arrangement of treatments representing three levels of microsite (whole plot), two levels of seed cover, and four levels of protection (subplots). Microsite treatments included direct seeding into open gaps on shoulders of back dunes between scrub oaks Quercus geminata (sand live oak) and Q. myrtifolia (myrtle oak), on the edge (drip line) of a scrub oak, and under the canopy of a scrub oak (A, B, and C, respectively in Figure 1). Seed cover treatments included covering seeds with 2 cm of sand or litter (leaves of nearby scrub oak). Protection treatments were included to reduce the risk of predation and to reduce environmental stress via reduction of wind and windblown sand to seedlings and included a control with no protection, protection with a PVC pipe (7.6 cm length and 10.2 cm diameter), protection with a wire mesh (1.27 cm hardware cloth), and protection with both a PVC pipe and wire mesh (Figure 2). PVC pipes were buried to half their height resulting in approximately 3.5 cm remaining above-ground and the wire mesh was bent to form square cages that were positioned approximately 3.5 cm above seeds as described by Stephens and Quintana-Ascencio (2015). Each microsite × seed cover × protection treatment contained 0 or 10 seeds collected in 2017 (bi-weekly from May 2017 to July 2017) as described in Campbell-Martínez et al. (2017). A control with 0 seeds was used to detect and quantify natural seedling recruitment within each block (independent sites within the coastal landscape). Two locations (Mobile St [30.229829, –87.837584] and Ft. Morgan [30.228902, –88.007185]) separated by ~20 miles of intense coastal development, each with five blocks were used for a total of 10 replicate blocks. A total of 2,400 seeds were used (10 seeds × 10 blocks × 3 microsites × 2 seed cover treatments × 4 protection treatments).
Microsites (A) under the scrub oak canopy of Quercus geminata (sand live oak) or Q. myrtifolia (myrtle oak), (B) on the gulf-side scrub oak edge (drip line of Quercus), or (C) in open gaps of bare sand within coastal scrub located in coastal Alabama.
Protection treatments included a control with no protection, protection with a PVC pipe (7.6 cm length and 10.2 cm diameter), protection with a wire mesh (1.27 cm hardware cloth), and protection with both a PVC pipe and wire mesh.
We recorded emergence, defined as the number of seedlings with above-ground growth (live tissue) per treatment, 2, 4, 6, 7, 11, 20, 23, 38, and 52 weeks after sowing (WAS) and calculated seedling establishment [(number of live seedlings/ number of seeds sown) ×100, %]. Growth data was recorded 52 WAS and included plant height (cm), leaf length (length of largest leaf from base to apex, cm), the number of nodes per seedling and mean plant width was calculated as [(width 1 + width 2)/2, cm] where width 1 was from the widest portion of plant and width 2 perpendicular to width 1).
Additionally, at the four corners and center of each subplot the sand accretion or loss (cm) was recorded using the change in height from the initial height of PVC pipes when installed at experiment initiation and the average was calculated. Temperature and precipitation data and 30-year averages were obtained from the National Centers for Environmental Information online database (NOAA 2020) provided by the National Oceanic and Atmospheric Administration using the Dauphin Island Number 2 station (30°15′1.8″N, 88°4′39″W).
Statistical Analysis
Main effects and their interactions were analyzed using generalized linear mixed models (PROC GLIMMIX in SAS v 9.4, SAS Institute, Cary, NC) for seedling establishment and growth data (height, mean width, leaf length, and nodes per seedling) 52 WAS. GLIMMIX was used to account for fixed and random effects. Seedling recruitment (%) was coded as a fixed effect and block was coded as a random effect. Response variable was survival, and the response distribution was Gaussian. Repeat observations on experimental units were accounted for in the model and a Kenward-Rogers approximation was used for computing the denominator degrees of freedom for the fixed effects tests. Differences between means for significant main effects and interactions (p ≤ 0.05) were computed using the ilink option of the LSMEANS statement. All blocks with ≥2.5 cm sand accumulation had no seedlings (2 blocks) and were excluded from the analysis. No seedlings emerged prior to 6 WAS, hence data subject to statistical analysis were obtained between 6 and 52 WAS.
Results
The main effects of WAS, microsite, and seed cover all interacted to influence seedling recruitment (WAS × Microsite × Cover; F12, 668 = 3.22, p = 0.0002). Seedling emergence differed by microsite (seeds sown under scrub oak canopy, at the edge of the canopy, or in the open) and emergence within the microsite was influenced by the material covering the seed (seeds covered with leaf litter or covered with sand). Additionally, the effects of the microsite and seed cover treatment resulted in differing patterns of emergence over time. The effects of microsite and seed cover on emergence over time were independent of protection treatments (Table 1. ANOVA). However, planting site protection treatment (seeds planted in the open with no protection, within PVC pipes, covered only with a wire mesh or within PVC pipes with a wire mesh cover) did influence seedling emergence but the response to protection differed among seeds covered by litter or covered only with sand (Cover × Protection; F3, 668 = 6.10, p = 0.0004). The effects of the interaction of Cover × Protection were independent of the time of evaluation (WAS) and Microsite (Table 1).
Type III Tests of Fixed Effects and their interactions on seedling emergence (%) for Asclepias humistrata evaluated over 52 weeks in response to sowing seed within three microsite conditions, with two types of seed cover and four levels of protection. Weeks after Sowing = Seed sown June 9, 2019 and evaluated for emergence after 6, 7, 11, 20, 23, 38, and 52 weeks; Planting microsite = (seeds sown under a scrub oak canopy of Quercus geminata (sand live oak) or Q. myrtifolia (myrtle oak), on the gulf-side edge (drip line) of coastal scrub or in an open area gulf-side of coastal scrub); Seed cover = seeds covered by leaf litter or covered only with sand; Planting site protection = (seeds planted in open sand without protection, within PVC pipes, covered with a wire mesh or within PVC pipes covered with a wire mesh cover.
There was more seedling recruitment (17%) in open areas than under scrub oak canopies (2%) 52 WAS (Figure 3). Seedling recruitment remained low (<5%) in open areas and on the edges of scrub until spring (38 weeks after sowing) (Figure 3). While there was initial seedling recruitment (up to 16.75%) under scrub oak canopies in the summer (6, 7, and 11 weeks after sowing), seedling recruitment decreased to 6.75% by fall and continued to decline to 2.75% under scrub oak canopies thereafter (Figure 3). Seedling recruitment 52 WAS was highest for seeds sown in the open (17.5 %) followed by seeds sown on canopy edges (10%) when seeds were covered with sand, which exceeded recruitment for all microsites where seeds were covered with litter (9.0–0.75%) (Figure 3).
Emergence of Asclepias humistrata (N=40) over 52 weeks after sowing on June 9, 2019 in three planting microsites (seeds sown under a scrub oak canopy of Quercus geminata (sand live oak) or Q. myrtifolia (myrtle oak), on the gulf-side edge (drip line) of coastal scrub or in an open area gulf-side of coastal scrub) with seeds covered by leaf litter (dashed lines) or with sand only (solid lines).
The significant interaction of the main effects of microsite and cover over time was a result of an early peak in emergence in July (6–7 WAS) for seed sown under scrub canopy and covered only with sand (16.75%) followed by a significant decline in emerged seedlings following winter temperatures (Figure 1, Table 2). By October/November (20–23 WAS) seedling emergence was evident for all Microsite × Cover treatments but the number of emerged seedlings under canopy cover declined by 64% with sand cover (16.75 to 6%) and by 22% with litter cover (4.5 to 3.75%). Initial emergence for seeds sown at canopy edges and in open areas did not exceed 3% in July (7 WAS) but emergence continued to increase through March (38 WAS). No significant declines in seedlings were noted for canopy edge- or open area- cover combinations over the initial 38 weeks of evaluation. A decline in emerged seedlings was noted in June (52 WAS) after the peak emergence in March (38 WAS); however; emergence for seeds sown at canopy edges and in the open remained above 5% and emergence for seeds sown with a litter cover was consistently lower than emergence for seeds sown with only a sand cover.
Seedling emergence (%) of Asclepias humistrata in response to planting microsite (seeds sown under a scrub oak canopy of Quercus geminata (sand live oak) or Q. myrtifolia (myrtle oak), on the gulf-side edge (drip line) of coastal scrub or in an open area gulf-side of coastal scrub) and seed cover (seeds covered by leaf litter or covered only with sand) 7 and 38 weeks after sowing (WAS).
Effects of protection treatments were independent of planting microsites (Table 1). Protection treatments interacted with seed cover to influence seedling emergence (Cover × Protection; F3, 668 = 6.10, p = 0.0004); within plots where seed were covered by sand, both PVC pipe alone and PVC pipe in conjunction with wire mesh resulted in emergence greater than when seeds were planted with litter cover, regardless of the level of protection. PVC pipe alone was the only protection treatment where emergence was significantly higher than emergence for seeds sown with only sand cover and no protection. Among seeds covered with litter the presence of a PVC pipe resulted in the greatest emergence; however, there were no significant differences among the protection treatments for seeds sown with litter. The consistently lower emergence values for treatments with mesh present suggest the mesh was a deterrent to emergence.
Seedlings were small when observed 52 WAS. Growth data means and standard errors are reported for each level of main effects in Table 3. Analysis of growth data indicated no significant difference among main effects or their interactions. No seedlings were observed within control plots that received no seeds suggesting seedlings within plots were not from an existing seed bank within these coastal sites. Monitoring of sand loss and accumulation within planting sites demonstrated a minimal shift in sand regardless of the level of protection (data not presented). Two blocks were eliminated from the study following sand accumulation exceeding 2.5 cm.
Effects of microsite, seed cover, and protection on growth characteristics of Asclepias humistrata (sandhill milkweed) seedlings 52 weeks after direct seeding into backdunes of coastal Alabama on 6/9/2019. Seeds were placed in a microsite under a scrub oak canopy of Quercus geminata (sand live oak) or Q. myrtifolia (myrtle oak), on the gulf-side edge (drip line) of coastal scrub or in an open area gulf-side of coastal scrub. Seeds were covered with 2 cm of sand or nearby litter (Quercus leaves). Seeds were either unprotected or protected with PVC pipes (7.6 cm length and 10.2 cm diameter) buried to ½ their height above ground, protected with 1.27 cm wire-mesh that was bent to form square cages positioned approximately 3.5 cm above seeds, or protected with both pvc and wire mesh cages.
The recorded high, low, and average monthly temperatures were generally higher than 30-year normal temperatures (Supplemental Figure 1). There were notable exceptions for November 2019 and May and June 2020 when temperatures were cooler than 30-year normals (Supplemental Figure 1). In the first three months of the experiment temperatures were ~2.5 °C or less above 30-year normals. Precipitation was generally near or above 30-year normal values in the first five months of planting and below or well below 30-year normals until the last month of observation when precipitation was 6 cm above 30-year normals (Supplemental Figure 1).
Discussion
Asclepias humistrata was successfully seeded directly into coastal backdunes during the summer using simple restoration outplanting techniques without specialized equipment. Seedling recruitment (17.5% 52 WAS) was greater than the mean establishment rate (11%) of direct seeding projects within ecosystem restoration literature (Ceccon et al. 2015). Covering seed with leaf litter provided no benefit for A. humistrata emergence and reduced emergence for seeds sown in the open. Seedling recruitment was highest when seeds were placed in open gaps (17.5%) or at canopy edges (10%) and covered with sand. This pattern of recruitment is consistent with plant survey data, where A. humistrata plants were documented in open areas and were not documented under shrub canopies or in areas with high litter accumulation (Campbell 2020). A similar trend was noted for a commonly co-occurring species, Balduina angustifolia (coastalplain honeycombhead), which had a higher proportion of seedlings established when seeds were placed in bare sandy spots of ancient dunes (inland scrub) than when placed under shrubs or when placed in open areas with litter (Stephens and Quintana-Ascencio 2015). This contrasts with A. curtissii (Curtiss’ milkweed) which had enhanced seed germination in shade but not by leaf litter (Mondo et al. 2010).
Physical protection of seedlings during the establishment phase is commonly used in restoration projects to reduce herbivory pressure from insects, birds, and small mammals (Willoughby et al. 2004, Madsen and Löf 2005, Jinks et al. 2006, St-Denis et al. 2013, Ceccon et al. 2015, Reque and Martin 2015). Several others have demonstrated the effectiveness of physical protection at increasing seedling recruitment rates during direct seeding restoration projects (Willoughby et al. 2004, Ceccon et al. 2015, Reque and Martin 2015). Others have also reported that protection did not increase or decrease seedling recruitment (Jinks et al. 2006, St-Denis et al. 2013). For example, St-Denis et al. (2013) reported no increase in seedling recruitment when protection was used for six tree species direct seeded into abandoned agricultural fields of Canada. Likewise, the use of protection was reported to reduce seedling recruitment for Fraxinus excelsior (European ash) compared to seeds sown without protection in old pastures of England (Jinks et al. 2006). We included protection treatments to determine if seed predation or seed loss from shifting sand was a factor in seedling recruitment. Seed protection did result in a higher level of seedling recruitment compared to only covering seed with sand (Table 1). However, the lack of an interaction of protection treatments with the main effects of microsite and seed cover suggests herbivory, sand loss or seed burial were not significant factors in the recruitment of A. humistrata within these relative positions of the coastal scrub landscape gaps. However, the higher recruitment in conjunction with the PVC pipe protection suggests microclimate effects such as reduced wind, windblown sand, and perhaps an increase in humidity or soil moisture may have been present. Interestingly, protection involving wire mesh appeared to reduce recruitment even when used in conjunction with the PVC pipe. These results are informative but use of physical protection for seedlings is not recommended as it can be costly, time and labor intensive to install, and requires removal.
A lack of herbivory pressure on the seeds of A. humistrata could be explained by toxins (cardenolides) present in seeds of Asclepias which can deter foraging by many insects (Moore and Scudder 1985). There is a caveat in that Oncopeltus fasciatus (milkweed bug) predates seeds and sequesters toxins but has only been observed eating seeds still attached to the mother plant, not seeds that have dispersed and are in the soil (Moore and Scudder 1985). As such, the use of protection when direct seeding A. humistrata was useful in understanding herbivory pressures at this site and provides some insight into microclimate effects but would not be operationally recommended.
Despite a documented lack of primary seed dormancy and quick germination under ideal conditions (Campbell-Martínez et al. 2017), and above average rainfall for a few months after sowing, most seeds did not emerge until sometime between winter and spring (23 and 38 weeks after sowing, respectively). A notable exception was for seeds placed under a Quercus canopy, which emerged in the summer soon after access to moisture. Initial air temperatures during summer and fall were within optimal germination temperatures of 24–28°C (Campbell-Martínez et al. 2017) and simulated summer (33/24°C) and fall (29/19°C) air temperatures (Campbell 2016). It is likely that microsites subjected to full days of direct sunlight (open and edge) had soil temperatures 2 cm below the soil surface (where seeds were sown) well above air temperatures and surface soils dried quickly. Coastal dune soils dry quickly in response to high soil temperatures during periods of sunlight and water is a limiting factor for survival of restoration plantings in coastal dunes (Martinez et al. 2013). For example, Miller et al. (2003) reported <2% soil moisture in the top 15 cm layer of soil after a 5 cm rainfall in a coastal foredune in northwest Florida. Thus, summer conditions may not have been conducive to germination at open and edge microsites.
In contrast, soil temperatures under the Quercus (scrub oak) canopy were likely more moderate and soils retained more moisture due to the shading effects of the canopy and the presence of persistent leaves (litter) which are known to retain moisture. Therefore, conditions were conducive to germination at this microsite during the summer. This is supported by a study conducted by Lakshmi et al. (2003) that documented an increase in surface (top 1–10 cm) soil temperatures occurred with a decrease in soil moisture. It is also supported by a direct seeding study of the California coastal dune plant, Lupinus nipomensis (Nipomo lupine), which had increased seedling recruitment for seeds placed on north facing slopes with less sunlight and more soil moisture compared to other slopes with high sunlight and less soil moisture (Luong et al. 2019). However, although germination occurs in the dark, A. humistrata requires 6–8 hours of sunlight for seedlings to establish and through time seedlings did not survive (Campbell 2020).
Moreover, phenological data corroborate a spring germination event for seeds of extant plants as there was a relatively large number of small plants (plant index < 3 cm) during the spring, hence a higher proportion of new seedlings in spring than in the fall (Campbell 2020). Anecdotal evidence also supports a larger number of seedlings in the spring than in the fall (personal observation). Evidence of a spring germination event may indicate temperature and moisture requirements for germination are not met until spring but could also indicate the possibility of secondary dormancy which could be initiated upon exposure to high temperatures (Baskin and Baskin 2014, Gao et al. 2018) found in the upper portion of coastal dune soils. This secondary dormancy could prevent seeds from germinating during the harsh summer months when survival may be less likely, instead delaying germination through the fall until the cooler conditions between the winter and spring.
Temporal variability inherent within these three microsites is likely a significant driver of the patterns of A. humistrata seedling emergence and recruitment success. The initial positive effect of shrub cover on seedling emergence was likely a result of reduced temperatures and higher soil moisture associated with the scrub oak canopy microclimate. Hence, shrub canopy was initially facilitative to seedling emergence. However, the shrub canopy microclimate became competitive with seasonal reductions in rainfall and increased temperature resulting in severe loss of early summer seedling recruits. The distinct wet and dry seasons of Florida, and the facilitative effects of the canopy cover on emergence in the summer were detrimental to long term recruitment because of subsequent competition for resources between the oak scrub and A. humistrata during the dry season. Mondo et al. (2010) describe a different response to shade and litter for A. curtissii in which shade increased germination, litter had no effect on germination, and seedlings transplanted under shrub cover had greater survival (40%) when planted under shrub canopy compared to planting in the open sand (2%). Despite these differing responses to shade and litter, Mondo et al. (2010) also concluded that severe summer conditions lead to reduced soil water availability beneath the shrub canopy such that initial facilitative effects of shrubs on seedlings are lost or become competitive effects. This temporal variability in the microclimate associated with shrubs and canopy cover may only be evident in field studies that census populations beyond an initial recruitment evaluation and should be considered in subsequent research.
Temporal variation may also include sand movements in the Florida dune scrub. Petru and Menges (2004) investigated the effects of seed burial on seedling emergence with three dune species common to the present field sites (Balduina angustifolia, Paronychia and Polygonella) and showed that significantly more seedlings emerge with burial from 0 to 5 mm sand depths compared to 20 mm. Petru and Menges (2010) further showed that seedling survival did not differ among burial levels, nor among species. Each of these three species have very small seed and seedlings are much smaller and shorter than those of Asclepias. It is likely that the threshold depth of burial for Asclepias is much greater when compared to these three test species and sand accumulation at our test sites did not exceed the threshold for Asclepias.
Asclepias humistrata can be successfully direct seeded into backdunes using standard restoration techniques. We recommend placing seeds in areas of bare sand within backdunes in open areas or on the edge of scrub and lightly covering seed with sand. Seeds planted 2 cm deep in summer had successful seedling establishment after 12 months of observation. Future germination studies of A. humistrata should investigate whether the delay in germination from summer to the next spring reported here is in response to environmental conditions (temperature or moisture) during summer, fall and winter or if germination was delayed by the imposition of secondary dormancy in seeds in response to the high temperature and low moisture conditions of the Florida scrub.
Asclepias spp. (milkweed) flower showing pollination. The pollen coheres into twin waxy masses (pollinia, image b), and the handle or caudicle lies in a chink on the side of the stigma. L.H. Bailey, 1917. Standard Cyclopedia of Horticulture. New York, NY: The MacMillan Company. The Florida Center for Instructional Technology, fcit.usf.edu.
Acknowledgements
This work was supported, in part, by US Fish and Wildlife Service project F16AC00519, the USDA National Institute of Food and Agriculture McIntire Stennis project FLA-JAY-005222 and McIntire Stennis project FLA-WFC-005653. Additional support was provided by the Gary Henry Florida Wildflower Research Endowment (Florida Wildflower Foundation, Inc.). A special thanks to Anna Dicks, Steve Wagner, Barbara Corey, Ashley Moore, Teri Vandenberg, and Cheyenne Brodie-McAleister for assistance with experiment management and data collection, and Gina Mangold for assistance with manuscript review.
This open access article is distributed under the terms of the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0) and is freely available online at: http://er.uwpress.org
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.