Skip booklet index and go to page content

Recovery Strategy for the Western Spiderwort (Tradescantia occidentalis) in Canada – 2013

3. Species Information

3.1 Species Description

Western Spiderwort is a perennial herb from the Spiderwort Family (Commelinaceae) whose common name originates from the sticky substance secreted by injured leaves or stems which hardens into a cobweb-like material (Kershaw et al. 2001). It has semi-succulent stems up to 60 cm high, and leaves that are grass-like (Looman and Best 1979). Young plants can resemble grass seedlings. Flowers have three petals ranging in colour from pink to violet with dark-blue being the most common type (Fig. 1; Scoggan 1957, Looman and Best 1979). Flowers are in clusters at the top of stems with each flower lasting about one day, opening in early morning and usually closing by noon (Faden 2000, C. Neufeld pers. obs.). It flowers in late June to mid July, with most mature seed capsules produced by late July (Kershaw et al. 2001). Reproduction is only by seeds, and plants lack rhizomes for vegetative reproduction (Scoggan 1978, Great Plains Flora Association 1991, Remarchuk 2006).

Figure 1. Flowering Western Spiderwort © Candace Neufeld

Figure 1 is a photograph of a flowering Western Spiderwort.

3.2 Population and Distribution

The range of Western Spiderwort in North America extends east to west from New York to Arizona, and north to south from Saskatchewan to Texas (Fig. 2). In Canada, Western Spiderwort is limited to four populations[2] in Alberta, Saskatchewan, and Manitoba (Fig. 3). These are located at the Pakowki Lake Sand Hills in southeastern Alberta, the Elbow Sand Hills in south-central Saskatchewan, and the Routledge and Lauder Sand Hills in Manitoba. Although COSEWIC (Smith 2002) recognized an east and west population within the Lauder Sand Hills, they will be considered as one population for the purposes of this recovery strategy for the following reasons: they are in the same dune complex; the east and west occurrences are separated by under 1.5 km with approximately 1 km of cultivation and a gravel road separating the suitable dune habitat (NatureServe 2010b); there is a high degree of genetic similarity, indicating there likely is, or has been, genetic exchange between them (Remarchuk 2006); the habitat and threats between the two are similar.   

Figure 2. Range of Western Spiderwort in North America

Figure 2 shows the current distribution of Western Spiderwort in North America. There are several locations spread out over central United States and up to southern Manitoba, Saskatchewan and Alberta.

In the three provinces where Western Spiderwort occurs, it has a restricted area of occupancy[3], and restricted extent of occurrence[4] (COSEWIC 2002). Numerous targeted surveys in recent years since the COSEWIC status update (Smith 2002) have greatly increased the known area of occupancy and estimates of population sizes, but the extent of occurrence, number of total populations and threats to habitat remain very similar to what they were prior to the COSEWIC status update. The 1990 COSEWIC assessment estimated 3,500 plants on 5 quarter sections[5] in two provinces (Smith and Bradley 1990) which was updated in the 2002 COSEWIC assessment to 22,000 plants on 19 quarter sections in three provinces with an area of occupancy of < 10 km2 and an extent of occurrence of approximately 500 km2 (Smith 2002). As of 2010, the Canadian population was estimated to be over 100,000 plants on 39 quarter sections in 3 provinces with an index area of occupancy [6] of 76 km2 (Table 1; Environment Canada, Saskatchewan Conservation Data Centre 2010 Element Occurrence data, Alberta Conservation Information Management System 2010 Element Occurrence data, Manitoba Conservation Data Centre 2010 Element Occurrence data).

Figure 3. Known range of Western Spiderwort in Canada.

Figure 3 shows the current known range of the four Western Spiderwort Canadian populations. Two populations occur near Brandon (Manitoba), one near Elbow (Saskatchewan), and one near Medicine Hat (Alberta).

To date, there has been inconsistency in methods used for surveying and monitoring Western Spiderwort. In addition, the few available estimates of population abundance are subject to wide variability or questionable accuracy and precision. This makes it difficult to infer any trends from past information that has been collected. Differences and potential sampling error arise in survey or monitoring techniques, search effort, enumeration methods, detectability at time of searching, observer ability, and perhaps inter-annual environmental factors (e.g., Goulet and Kenkel 1997, Peters 2003a and 2003b, Godwin and Thorpe 2004-2007, Remarchuk 2006, Peters et al. 2009). Therefore, due to a lack of standardized and consistent data, current population trends are unknown but there is no evidence for a continuing decline in numbers or area of occupancy. Large increases of population size and area of occupancy between decades to date should not be interpreted as an increase in population size due to biological phenomena, but rather a function of the cumulative search effort undertaken.

Table 1. Summary of known Western Spiderwort populations in Canadaa. Accessible version of Table 1
SiteYear First ObservedRecent Pop. Estimate [Year]bCOSEWIC pop. estimategHighest Pop. Estimate [Year]b
Routledge Sand Hills192313 402 [2005]c9422 [2001]26 550 [1996]d
Lauder Sand Hills
• Lauder Sand Hills West1950775 [2005]c619 [2001]h854 [2001]
Lauder Sand Hills East19954 024 [2005]c4321 [2002]19 540 [1996]d
Pakowki Lake Sand Hills198637 195 [2007/08]f7450 [2002]37 195 [2007/08]
Elbow Sand Hills
AAFC-AESB Elbow Community Pasture199144 000 [2006]c42 [1991]69 000 [2005]c
Douglas Provincial Park20014 686 [2009]e100 [2002]4 686 [2009]

a Values and occurrences in the table are those known to Environment Canada as of January 2011. Sources: Smith and Bradley 1990, Hohn 1994, Goulet and Kenkel 1997, Hughes 2001, Peters 2003a and 2003b, Godwin and Thorpe 2004-2007, T. Sample (pers. comm.), Remarchuk 2005, Remarchuk 2006, Peters et al. 2009, S. Vinge (pers. comm.), C. Neufeld (unpubl. data), Neufeld 2008, Neufeld 2010, Manitoba Conservation Data Centre element occurrence records (unpubl. data), Alberta Natural Heritage Information Centre element occurrence records (unpubl. data).
b Note that population sizes are most often estimates because of distribution over a large area or high numbers of plants in an area. Not all occurrences within a population are enumerated in a year with most surveys. In addition, some surveyors counted individual stems as a unit while others counted clumps of plants as a unit. This makes it impossible to compare population sizes between years. Most populations have been revisited since the last population estimates, but the visits were for the purposes of delimiting area of occupancy, not population counts.
c These values are based on counts made along transects, and then extrapolating average density to the area of available suitable habitat on each sand dune. To view methods and standard error associated with the estimates, refer to the documents (Godwin and Thorpe 2004 and 2006, Remarchuk 2006).
d Only flowering plants were counted. The count was divided by 0.3 to account for only 1/3 of plants flowering at any time; this may have inflated the estimate (Goulet and Kenkel 1997, Hughes 2001).
e Not all occurrences were revisited in this estimate so this is only an estimate of part of the population.
f Some occurrences were counted. Population counts for the remainder were extrapolated from patch area based on GPS boundaries, and average population density (Remarchuk 2005, Peters et al. 2009).
g From Smith (2002).
h There are two different numbers reported for population size for 2001 in the source report (Hughes 2001); COSEWIC reports the population size from the table in the report while the Manitoba Conservation Data Centre reports the population size from the text in the report.

3.3 Needs of the Western Spiderwort

Western Spiderwort inhabits eolian[7] landscape complexes derived from lacustrine or glaciolacustrine deposits comprised of sandy and loamy sandy soils (David 1977). The habitat is in a transitional state between recently disturbed and fully stabilized, but probably tending towards earlier stages of vegetative succession and partial stabilization. Active sand dunes become stabilized with vegetation through natural succession and only remain open or free of woody plants through repeated fire, grazing disturbances and fluctuations in climate (Geological Survey of Canada 2001). Thus, Western Spiderwort needs eolian landscapes with native vegetation subject to fire and grazing disturbance patterns that mimic historical regimes.

Regionally, in the Pakowki Sand Hills in Alberta, Western Spiderwort is most common in level slacks between dune features that are stabilized by mixed grass prairie vegetation, or along the south-facing edges of partially stabilized or active dunes (Peters 2003a; Remarchuk 2006). In the Elbow Sand Hills in Saskatchewan and the Lauder and Routledge Sand Hills in Manitoba, Western Spiderwort is most common on south and southwest facing slopes of partially stabilized dunes, and less commonly on sandy flats with moderate vegetation (Hohn 1994, Goulet and Kenkel 1997, Hughes 2001, Godwin and Thorpe 2006, Remarchuk 2006). In these landscapes, a scrubby forest cover of Trembling Aspen (Populus tremuloides), Chokecherry (Prunus virginiana) and, in Manitoba, Bur Oak (Quercus macrocarpa), forms a matrix within which the sandy grassland and barren sand patches appear isolated from each other. Historically, it is possible the sand dunes and grassland may have been the dominant land class with the aspen and shrubs being isolated patches within.

A detailed survey by Godwin and Thorpe (2006) in the Elbow Sand Hills in Saskatchewan found preferred Western Spiderwort habitat to be moderate to steep south to southwest facing dune slopes with partial exposures of bare sand and no topsoil or organic material (soil A-horizon). These preferred slopes contained grassland vegetation associated with earlier successional or stable dune environments and had little cover by taller vegetation, including shrubs and trees. This description appears to agree with descriptions provided from the occupied dunes in Alberta and Manitoba as well (Hohn 1994, Goulet and Kenkel 1997, Peters 2003a and 2003b, Remarchuk 2006).

In varying degrees across the Canadian range, Western Spiderwort is found amongst shrubs or beneath an aspen or more open oak canopy but almost always immediately adjacent to other occupied habitat on open slopes (Goulet and Kenkel 1997, Godwin and Thorpe 2006). Potential reasons for establishment and survival in these “secondary” semi-shaded habitats include protection amongst the taller vegetation from grazing pressure, spill-over from abundant seed production of plants on adjacent slopes, or remnant plants from a previously larger occurrence on more sandy, open habitat that is now being stabilized through succession. Occasionally spiderwort plants are observed along sandy game trails within wooded areas which join isolated dune patches. These plants are most likely the result of seed dispersal by grazers. Although some of these occurrences in shrubby or wooded areas are likely more transitory in nature, they may act as dispersers of seed and pollen between occurrences in a population or metapopulation by deer or pollinators. Deer in this region can disperse long distances in short periods, and may be responsible for historical short and long distance colonization of plants in sand dune complexes (Skelton 2010). While it has been suggested that some low shrub cover within the sand dune complex may be important for spiderwort in exposed habitat, complete encroachment of woody vegetation leading to dune stabilization should be avoided (Goulet and Kenkel 1997, Smith 2002); the many physiological adaptations of spiderwort to drought conditions suggest it is better suited to warmer, drier habitats than moist, cool habitats (Remarchuk 2006).

For detailed lists of plant species that have been found growing near Western Spiderwort, refer to Smith (2002) and Remarchuk (2006)[8]. Remarchuk (2006) did not find Western Spiderwort to be associated with any particular plant taxa; slight differences in species composition among populations were attributed to differences in the amount of dune stabilization or slope height among provinces, as well as differences in regional species communities. Godwin and Thorpe (2006) found the most frequently occurring species in Western Spiderwort habitat to be affiliated with early to mid-successional dune habitats or stabilized low-dune grasslands.

Limiting Factors

Western Spiderwort lacks rhizomes for vegetative reproduction (Goulet and Kenkel 1997, Remarchuk 2006); reproduction occurs only by seed. Therefore, Western Spiderwort likely relies on a range of pollinating insects, particularly sweat bees (Smith 2002, Alberta Western Spiderwort Recovery Team 2005). Furthermore, for successful pollination and reproduction, suitable habitat should be sufficiently connected for the dispersal of those insects and any vectors for dispersal of seed. This dispersal of both pollinators and seed is needed to maintain gene flow among occurrences within a population (Remarchuk 2006). Mechanisms for long distance dispersal of Western Spiderwort seed are not known, although plants are frequently grazed. Thus, herbivory may have a role in seed dispersal although no studies have been done. Seed capsules hang in a cluster so most seeds are naturally released close to the parent plant, resulting in a clumped distribution of plants around the parent plant with some seeds dispersed downslope by gravity or runoff (Smith 2002, Remarchuk 2006).

A genetic diversity study was conducted within and among all Canadian populations; the study found populations to be isolated from each other with no genetic flow among them (Remarchuk 2006). There was high genetic similarity within each provincial population potentially indicating a future risk of inbreeding depression if there is a large decline in population numbers; the populations in Alberta and Saskatchewan are large enough that as long as habitat conservation and compatible management occur, inbreeding depression is unlikely to occur (Remarchuk 2006).

2 Using the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) definition, a population is defined as “geographically or otherwise distinct group within a wildlife species that has little demographic or genetic exchange with other such groups. Theoretically, populations maintain genetic distinction if there is typically less than one successful breeding immigrant individual or gamete per generation.” (COSEWIC 2010). NatureServe considers occurrences within 1 km of each other, or within 3 km if there is less than 1 km of unsuitable habitat between them, to be from the same element occurrence (NatureServe 2010b). For the purposes of the recovery strategy, we will consider the term “element occurrence” equivalent to population. The Canadian population is the total number of mature individuals in Canada.

3 Area of occupancy is the portion within the 'extent of occurrence', or range of a species, that is actually occupied by the species (COSEWIC 2010). This can also be viewed as the area occupied by each occurrence.

4 Extent of occurrence, as defined by COSEWIC, is “the area included in a polygon without concave angles that encompasses the geographic distribution of all known populations of a wildlife species” (COSEWIC 2010).

5 The Dominion Land Survey system (McKercher and Wolfe 1986) is the grid system used in the Prairie Provinces to describe land locations. The provinces are divided into townships and each township is divided into 36 sections. Each section is further divided into four 0.8 km x 0.8 km quarter sections.

6 The index area of occupancy is calculated by counting the total number of 2 x 2 km grid squares that contain the species (COSEWIC 2010). This count was made by using either the area of occupancy for each occurrence where this had been mapped, or centroid coordinates for those populations lacking more detailed survey work, and is current up to December 2010. Occurrences included in the calculation of index area of occupancy needed to meet the following criteria: occurrences were reported using precise and accurate geographic referencing systems; habitat still exists at the location to support the species; occurrences have been confirmed at the location within the past 25 years. A breakdown of index area of occupancy per population is: Lauder Sand Hills = 20 km2, Routledge Sand Hills = 12 km2, Elbow Sand Hills = 32 km2, Pakowki Sand Hills = 12 km2.

7 Eolian means borne, deposited, produced, or eroded by the wind.

8 For associated plant lists for each province refer to Goulet and Kenkel (1997) and Hohn (1994) for Manitoba, Godwin and Thorpe (2006) for Saskatchewan, and Peters (2003a and 2003b) and Remarchuk 2005 for Alberta.