Northern Prairie Wildlife Research Center
Although we do not know enough about L. salicaria's growth requirements or genetic plasticity to make an accurate forecast of the limits of its spread in North America, some reasonable projections can be made from examining its spread throughout the world. The worldwide distribution of L. salicaria clearly identifies it as a plant of temperate climates with a strong affinity for moist or saturated soils. The northern limits of L. salicaria's distribution in Europe (Fig. 2) and the experimental work of Shamsi and Whitehead (1974a, 1974b, 1977a, 1977b) indicate that purple loosestrife will be a vigorous competitor with native American flora throughout the marshes and alluvial wetlands of the northern United States and southern Canada.
By 1900 (Fig. 6), purple loosestrife was established along the eastern seaboard as far north as Nova Scotia and as far south as North Carolina. The infestations in Nova Scotia have continued to spread slowly but have not shown the vigor of more southerly establishments. The plant may not have reached its ultimate northerly distribution, but its rate of spread has slowed markedly in comparison with its continued rapid expansion into western North America. Similarly, the early establishment on the Cape Fear estuary (near Wilmington, North Carolina) has been static, if indeed it has survived. Radford et al. (1964) did not mention L. salicaria in the Cape Fear area, but described the plant as very rare in marshes in northeastern North Carolina (Watauga County). Again, neither T. latifolia nor P. arundinacea has been a successful plant as far south as North Carolina. Furthermore, despite the early establishment of L. salicaria on the Delaware River ballast grounds, purple loosestrife has not been successful in invading disturbed coastal marshes in New Jersey. From the Hudson River estuary southward, reed (Phragmites australis) dominates fresh to brackish marshes. The presence of this vigorous competitor may, in part, explain the failure of L. salicaria to invade southern coastal wetlands.
Purple loosestrife's obvious success in displacing cattail inside the limits of Wisconsin glaciation (Fig. 7) marks it as a serious threat to wetland habitats throughout the glacial basins and riparian wetlands that are the primary producers of waterfowl and other marsh wildlife in the Atlantic and Mississippi flyways. Shaw and Fredine (1956) defined eight types of inland fresh wetlands in their assessment of wetlands of the United States. Purple loosestrife has been highly successful in colonizing at least three of these types: Type 2—Inland fresh meadow; Type 3—Inland shallow freshwater marsh; and Type 4—Inland deep freshwater marsh. The area estimates by Shaw and Fredine of these three wetland types in the seven north-central States exceed 1.4 million ha (3.6 million acres). More than 90% of these wetlands would fall within the boundaries of the Wisconsin glacial drift line and are highly susceptible to degradation from L. salicaria invasion.
The potential threat of purple loosestrife invasion in prairie pothole wetlands of North Dakota, South Dakota, Manitoba, and Saskatchewan is difficult to evaluate at this time. The plant has been unable to invade saline wetlands in New York State's lower Hudson and Long Island areas; this strongly suggests that L. salicaria's growth is somehow inhibited by high soil conductance. Similarly, L. salicaria faces a gradient of salinity as it invades new habitats west of the 100th meridian. Nevertheless, the recent report (Thompson and Yates 1985) of an infestation in an off-water swale near Powell, Wyoming, indicates that purple loosestrife will be troublesome in prairie wetlands of pH 7.5 or less and able to invade but not dominate open habitats with soil pH values up to 8.0. The Garrison Project in North Dakota lies in a critical position across this zone of neutral to alkaline soils. Purple loosestrife is already established east and north of the project, but has not been reported upstream from Garrison Dam.
Wetland and riparian habitats vary greatly in their resistance to invasion by L. salicaria. Since natural spread of purple loosestrife is primarily by floating seeds or propagules, the continuity and configuration of a watershed strongly influence the rate and ease of expansion of a local infestation. Isolated wetland basins (as in northern prairie potholes) are relatively secure from the spread of seeds or propagules of purple loosestrife, whereas wetland complexes connected by a common waterway (e.g., the Mohawk waterway in upstate New York, Upper Mississippi navigation pools, and the Klamath River wetland complex in Oregon) are highly susceptible sites. The gradient and cross-section of a stream also strongly influence its resistance to L. salicaria infestation. Mountain or high plateau streams with steep gradients and narrow canyons (e.g., Snake River in Oregon, the Colorado River in Colorado, and the Wind River in Wyoming) are frequently scoured, with few eddies or slack-water areas where emergent aquatic plants can take hold. In contrast, low-gradient streams with broad alluvial deposits offer many sites for wetland plants to colonize and are susceptible to purple loosestrife infestation and spread. Many streams have shade-covered banks (riverbottom hardwoods in the East; spruce, willow, and alder borders in many western streams) where the high light requirements of L. salicaria seedlings preclude development and make these sites unsusceptible to invasion.
The vulnerability of a habitat to invasion by L . salicaria can be judged from the occurrence of its most frequent plant associates (Table 1). The presence of cattails, reed canarygrass, sedges, or rushes in marsh basins or along riparian wetlands identifies a habitat that can be invaded by L. salicaria. Last, one of the most important influences on the security of habitats against invasion by alien species is the degree of disturbance or environmental stress suffered by its soil, water, and biota. In the Midwest, Auclair (1976) estimated that 92% of the original prairie (upland) surface was converted to a cropland-temporary pasture system between 1833 and 1934; during the same period, 66% of the original marshlands were "transformed." In the subsequent 50 years, the widespread and pervasive effects of soil, water, and air pollution suggest that no wetlands have escaped some form of perturbation.