2006 SOUTHEASTERN NATURALIST 5(3):515–522
Range and Ecology of Helenium virginicum
in the Missouri Ozarks
Rhonda L. Rimer1,* and James W. (Bill) Summers1
Abstract - Helenium virginicum (Virginia sneezeweed) is a federally listed threatened
herb found at unshaded sites in seasonally inundated ponds in the Shenandoah
Valley of Virginia and in the Ozark Highlands of Missouri. In the Ozarks, one
population of H. virginicum was discovered in the 1950s, though it was thought to be
a hybrid of Helenium flexuosum and Helenium autumnale until genetics work resolved
its identity in 2000. Since that time, it was the only known population outside
of Virginia. In this study, we systematically identified potential H. virginicum habitat
in the Ozarks by using topographic maps and aerial photographs and surveyed these
areas for H. virginicum. Our objectives were to locate new populations of the plant
and to better define its ecology in Missouri. By the end of the growing season of
2004, we had located 42 new populations of H. virginicum in the lower Missouri
Ozarks—more populations than are currently known in the state of Virginia. The
results of this study greatly expand the known range of H. virginicum in Missouri and
provide important information regarding its ecology.
Introduction
Helenium virginicum Blake (Virginia sneezeweed; Asteraceae) is a federally
listed threatened herb considered to be an obligate of moist borders of
seasonally wet sinkhole ponds that occur in the Shenandoah Valley of
Virginia (US Fish and Wildlife Service 2000) and as one disjunct population
(Knox et al. 1995, Simurda and Knox 2000) and two reintroduced populations
in Howell County, MO (Rimer and McCue 2005). The species occurs
in wetland habitat on acid clay soil in a matrix of sand, gravel, and cobble
that overlies limestone bedrock. Erosion of the bedrock leads to formation of
shallow basins in which water levels vary both seasonally and annually from
no standing water to continuous inundation for months (US Fish and Wildlife
Service 2000). Though morphology and habitat were shown to be similar
for Missouri and Virginia populations (Knox 1997, Knox et al. 1995),
botanists were uncertain whether the Missouri population merited
treatement as H. virginicum. Internal Transcribed Spacer Region (ITS)
sequence comparisons completed in 2000 established that there were no
significant genetic variations between the population in Missouri and populations
in Virginia, leading Simurda and Knox (2000) to recognize the
Missouri population as H. virginicum.
Helenium virginicum was first described in Virginia by Blake (1936).
The plant is 70.00–110.00 cm tall with a branching stem within the inflorescence,
a simple stem below, and winged (0.3–2.5 mm wide) throughout
1Missouri Department of Conservation, Wildlife Division, 551 Joe Jones Boulevard,
West Plains, MO 65775. *Corresponding author - Rhonda.Rimer@mdc.mo.gov.
516 Southeastern Naturalist Vol. 5, No. 3
by decurrent stem tissue. The basal leaves can be both toothed and
untoothed, are widest in the upper half and tapered at both ends, form a
large basal rosette, and are dotted with glands (Blake 1936). Several field
characters distinguish H. virginicum from its closest relative H. autumnale
L. and other Helenium species. Helenium virginicum has lance-linear,
mostly entire, cauline leaves that taper only slightly toward the leaf base.
The stem and rosette leaves are coarsely hairy, and the corollas abscise at
maturity, leaving each head with conspicuous silvery pappus scales (Knox
1987, Knox et al. 1995).
Threats to habitat, limited range, and a small number of occurrences led
to the designation of H. virginicum as “threatened” by the US Fish and
Wildlife Service on December 3, 1998 (US Fish and Wildlife Service 1998).
Since that time, both Virginia and Missouri have formulated recovery strategies
for the species in their states, and a federal recovery plan is in review. In
Missouri, the recovery tactics for the species included a reintroduction
project that resulted in two new populations on public lands (Rimer and
McCue 2005). The continuing vigor of the one natural H. virginicum population
near Pomona and the success of the Missouri reintroduction project
fueled lingering questions as to how H. virginicum came to have just one
natural location in the Ozarks of Missouri and whether additional natural
populations might exist in Missouri. The purpose of this study was to locate
new populations of H. virginicum in the Ozark Highlands and to better
define its ecology in Missouri.
Methods
This study initially focused on Howell County, MO, since H. virginicum
was known from no other locations in the state. We used topographic maps
and aerial photos of the county to delineate natural sinkhole features.
Helenium virginicum is documented as an acid-soil, shade-intolerant, sinkhole-
pond obligate in Virginia located in habitats ranging from pristine to
disturbed. These factors led us to sample a variety of karst habitats including
sinkholes with seasonal standing water, sinkhole ponds, and dry collapses in
woods, open fields, and meadows.
Topographical quadrangle maps, and associated aerial photographs, from
6 counties were examined. When natural sinkhole-pond features were identified
along county edges, maps and photographs for adjoining counties were
included in the survey.
Topographical maps and aerial photographs revealed 113 natural sinkhole
features meriting field survey. Each survey site was assigned a unique
letter-number combination. Sites were defined as landscapes at least 0.80
km apart. If several karst features existed within 0.80 km of each other, they
were defined as one site. By the same token, a population was defined as a
group of H. virginicum that was more than 0.80 km from another such group.
Sites were surveyed for H. virginicum from July–October 2003 and 2004.
2006 R.L. Rimer and W. Summers 517
Geospatial values for elevation, slope, aspect, geosurface (e.g., the geology
at the site), ecological subsection (categories based on characteristics
such as lithology, relief, slope, elevation, geomorphic process, potential
natural vegetation, and soils), and land cover (using the National Land
Cover Database) were generated for specific survey sites using ArcView 3.0.
Values for individual parameters were compared for sites where H.
virginicum was present and a random subset of 35 sites where H. virginicum
was absent using a one-way ANOVA for the continuous variables elevation,
slope, and aspect and a 2 x 2 chi-square test of independence for the
categorical variables geosurface, ecological subsection, and land cover. The
35-site “H. virginicum absent” subset was originally randomly selected to
balance our data with the 35 H. virginicum sites that we discovered during
the first field season of 2003. Because categorical variables consisted of
many types, some types were combined into broader descriptive groups to
avoid cell frequencies of zero and expected cell frequencies of less than five
in the chi-square table. Geosurface data encompassed four geosurface types
that were simplified into “dolomite” and “Roubidoux overlaying dolomite”
groups; land cover encompassed six types that were simplified into “open”
and “forested” groups. Fisher’s exact test was used with chi-square 2 x 2
classification tables to determine significance. Significance for all tests was
set at P ≤ 0.05.
Results
We discovered forty-two H. virginicum populations in the six counties
(Howell, Shannon, Oregon, Texas, Wright, and Webster) we investigated in
the Ozark Highlands of Missouri. Howell County, the location of the original
Missouri H. virginicum population, supports the majority of sites (Fig. 1).
Survey sites at higher elevations supported significantly more H.
virginicum populations than did sites at lower elevations (P < 0.01; Table 1).
Mean elevation for H. virginicum sites was 338.48 m ± 4.15 SE, while mean
elevation for H. virginicum-absent sites was 317.28 m ± 4.38 (Table 1).
Slope and aspect were not significantly different between sites (Table 1).
All survey sites were located within the Ozark Highlands section of
Missouri. This area is further divided into ecosubsections based on characteristics
such as lithology, relief, slope, elevation, geomorphic process,
potential natural vegetation, and soils. (Nigh and Schroeder 2002). Our
survey sites fell in four ecosubsections: White River Hills, Central Plateau,
Gasconade River Hills, and Current River Hills. We found that 79% of all
sites surveyed were in the Central Plateau subsection, with 69% of H.
virginicum sites and 91% H. virginicum-absent sites located in this subsection
(P < 0.01; Fig. 2).
Survey sites fell into two broad geosurface categories of either dolomite
or Roubidoux sandstone overlaying dolomite. We found that 70% of survey
sites were located on dolomite bedrock. However, only 38% of H.
518 Southeastern Naturalist Vol. 5, No. 3
virginicum sites were located on dolomite bedrock, while 61% were located
on Roubidoux sandstone overlaying dolomite bedrock (P < 0.01; Fig. 2).
Data from the National Land Cover Database indicated survey sites fell
into six different land-cover classes including open water, commercial/
Figure 1. Current locations of the 45 known H. virginicum populations in Missouri.
Forty-two sites were located during this study, 2 are reintroduction sites (Rimer and
McCue 2005), and 1 is the original site discovered by Steyermark (1960).
Table 1. Mean ± SE, minimum, and maximum values for elevation (m), slope (%), and aspect (º)
at sites where H. virginicum was present and where H. virginicum was absent. Significant
differences (P < 0.05) are starred (*).
Variable H. virginicum Mean ± SE Maximum Minimum
Elevation (m) Present 338.48 ± 4.15* 377.34 297.18
Absent 317.28 ± 4.38* 384.05 276.76
Slope (%) Present 8.12 ± 1.63 48.00 1.00
Absent 10.60 ± 1.72 38.00 1.00
Aspect (º) Present 184.87 ± 17.57 360.00 5.00
Absent 182.69 ± 18.54 343.00 1.00
2006 R.L. Rimer and W. Summers 519
industrial/transportation, deciduous forest, mixed forest, pasture/hay, and
woody wetlands. We combined these classes into general categories of open
land and forest land. Survey sites were located almost equally in open land
and forest land (54% and 49%, respectively). However, open-land sites
supported significantly more populations of H. virginicum than did forestland
sites (P < 0.01; Fig. 2).
Discussion
Helenium virginicum is one of 1260 species on the US Federal Threatened
and Endangered species list (US Fish and Wildlife Service 2003). This
study demonstrates, with the discovery of 42 new populations in Missouri,
that H. virginicum is far more prevalent than previously thought. Our analysis
of geospatial site data facilitates a better understanding of H. virginicum
ecology throughout its known range in Missouri.
Helenium virginicum occurs more often at higher elevations among
the sites sampled in this study. It is not unusual for plants to show elevation
preferences (Luteyn 2002). In the Ozarks, this preference may be
linked to geology and associated soil pH. A geologic shift occurs between
Gasconade dolomite and Roubidoux sandstone overlaying dolomite at
elevations roughly between 183–366 m throughout the survey area.
Limestone and dolomite geologies give rise to alkaline soils while chert,
Figure 2. Percent of sites where H. virginicum was present and where H. virginicum
was absent: in the Central Plateau subsection, with Roubidoux sandstone overlaying
dolomite geology, and located in open land.
520 Southeastern Naturalist Vol. 5, No. 3
sandstone, such as Roubidoux, and granite produce acid soils
(Steyermark 1963). Preference for higher elevations may be related to a
preference for areas where Roubidoux sandstone overlaying dolomite
dominates over that of dolomite geology because the resulting soils may
be more acidic. Other studies have demonstrated that distinct pH preferences
may limit plant range and distribution (Bradshaw et al. 1960, Cole
and Weltzin 2004, Pearsall and Wray 1927, Stone 1944), and H.
virginicum is known only from acidic soils throughout its range in
Virginia (Knox 1997). The geosurface data we present here appears to
support soil-pH data collected in Virginia, though field measurement of
soil pH at Missouri sites is needed to validate this assertion.
Our data showed no significant variation in slope and aspect between
sites that supported H. virginicum and sites where H. virginicum was absent.
The natural habitats where H. virginicum is most commonly found (sinkhole
ponds) and, hence, the types of sites targeted by our search, rarely occur on
portions of the landscape with dramatic relief. Our survey sites reflect the
natural occurrence of these habitats on the landscape, regardless of whether
they supported H. virginicum.
We observed that, within the Ozark Highlands section of Missouri, the
Central subsection was most likely to support H. virginicum. This may be
due to the relatively higher proportions of open lands supporting karst
features in this subsection (Nigh and Schroeder 2002). In addition, the
Central subsection once had expanses of Ozark tallgrass prairie, and many of
the sites which supported H. virginicum occurred on fescue pastures that
were considered to have been previously covered in prairie (Nigh and
Schroeder 2002). This information, together with the habitat preferences of
H. virginicum suggests that, in Missouri, this plant may have been a part
of the pre-settlement Ozark prairie community. Our data are incomplete and
may be potentially biased due to the limited area we surveyed. Should
additional H. virginicum sites be discovered in Missouri or other states, data
from these sites will be crucial in further defining the relationship between
this species’ distribution and ecological sections and subsections.
Helenium virginicum is known to prefer open growing conditions in
Virginia (Blake 1936, Knox 1997), and our results show that H. virginicum
also prefers open-land sites to those that are forested in Missouri. Further,
the variety of sites where we found this plant flourishing—including undisturbed
sinkhole ponds, rural airports, roadside ditches, and cattle ranches—
is similar to habitats where the plant grows in Virginia (US Fish and Wildlife
Service 2000). Knowing that H. virginicum grows in the open, in undisturbed
as well as heavily disturbed wetlands, and on soil that is known or
suspected to be acidic lying upon limestone or dolomite bedrock, should
guide future efforts to locate additional H. virginicum sites in Missouri and
to assist other states with their search efforts.
Our work has led to the discovery of 42 new H. virginicum sites in the
state of Missouri. Knowledge of these new populations and the 25 known
2006 R.L. Rimer and W. Summers 521
populations in Virginia (US Fish and Wildlife Service 2000), as well as the
information we gained regarding the species’ ecology, will further steer
conservation efforts for this species. Protection of key habitat and the cooperation
of many private landowners whose property supports H. virginicum
gives us hope that this species will persist as a valued component of our
biodiversity.
Acknowledgments
We thank Dr. Jeff Briggler, Peggy Horner, Lee Hughes, Dr. Paul McKenzie, Larry
Rieken, C.D. Scott, Donald Smith, Timothy Smith, David Urich, and Dr. George
Yatskievych for technical and logistical support with the field and analysis portions of
this project. Special thanks to manuscript editor Dr. John Knox and the two anonymous
reviewers whose thoughtful suggestions much improved this manuscript. Funding for
this work was provided by the Missouri Department of Conservation Wildlife Division.
Vouchers for all populations discovered in this study were deposited at the
Missouri Botanical Garden, St. Louis, MO.
Literature Cited
Blake, S. 1936. A new Helenium from Virginia. Claytonia 3:5–6.
Bradshaw, A., R. Lodge, D. Jowett, and J. Chadwick. 1960. Experimental investigations
into the mineral nutrition of several grass species: Part II. pH and calcium
level. Ecology 48:143–150.
Cole, P., and J. Weltzin. 2004. Environmental correlates of the distribution and
abundance of Microstegium vimineum, in East Tennessee. Southeastern Naturalist
3:545–562.
Knox, J. 1987. An experimental garden test of characters used to distinguish
Helenium virginicum Blake from Helenium autumnale L. Castanea 52:52–58.
Knox, J. 1997. A nine-year demographic study of Helenium virginicum (Asteraceae),
a narrow endemic seasonal wetland plant. Journal of the Torrey Botanical Society
124:236–245.
Knox, J., M. Gutowski, D. Marshall, and O. Rand. 1995. Tests of the genetic bases of
character differences between Helenium virginicum and H. autumnale
(Asteraceae) using common gardens and transplant studies. Systematic Botany
20:120–131.
Luteyn, J. 2002. Diversity, adaptation, and endemism in neotropical Ericaceae:
Biogeographical patterns in the Vaccinieae. The Botanical Review 68:55–87.
Nigh, T., and W. Schroeder. 2002. Atlas of Missouri Ecoregions. Missouri Department
of Conservation, Jefferson City, MO. 212 pp.
Pearsall, W., and E. Wray. 1927. The physiology and ecology of the calcifuge habit
in Eriophorum Angustifolium. Ecology 15:1–32.
Rimer, R., and K. McCue. 2005. Restoration of Helenium virginicum Blake, a
threatened plant of the Ozark Highlands. Natural Areas Journal 25:86–90.
Simurda, M., and J. Knox. 2000. ITS sequence evidence for the disjunct distribution
between Virginia and Missouri of the narrow endemic Helenium virginicum
(Asteraceae). Journal of the Torrey Botanical Society 127:316–323.
Steyermark, J. 1960. An unusual hybrid Helenium. Rhodora 62:343–346.
Steyermark, J. 1963. Flora of Missouri. Iowa State University Press, Ames, IA.
1725 pp.
522 Southeastern Naturalist Vol. 5, No. 3
Stone, M. 1944. Soil reaction in relation to the distribution of native plant species.
Ecology 25:379–386.
US Fish and Wildlife Service. 1998. Endangered and threatened wildlife and plants:
Determination of threatened status for Virginia sneezeweed (Helenium
virginicum), a plant from the Shenandoah Valley of Virginia. Federal Register
63:59239–59244.
US Fish and Wildlife Service. 2000. Virginia sneezeweed (Helenium virginicum)
recovery plan. Technical/Agency Draft. Hadley, MA. 54 pp.
US Fish and Wildlife Service. 2003. Endangered Species Bulletin 28:44.