USGS - science for a changing world

Northern Prairie Wildlife Research Center

  Home About NPWRC Our Science Staff Employment Contacts Common Questions About the Site

Spread, Impact, and Control of Purple Loosestrife (Lythrum salicaria) in North American Wetlands

The Case for Biological Control


DeBach (1964) identified the field of biological control as comprising the "study, importation, augmentation, and conservation of beneficial organisms for the regulation of population densities of other organisms … Particular emphasis has been laid on pest insects and mites, and weeds." The concept of classical biological control arose from the observation that some of the most damaging plant or insect pests in North America were aliens that were native to the Eurasian Continent and were spread worldwide with the rise of marine exploration and commerce. Moreover, the arrival of a potential pest species as a seed or egg enabled it to escape diseases, parasites, and predators that had coevolved with it. The assumption was made that these organisms, singly or in combination, limited the competitive vigor of the plant on its continent or area of origin. The application of this concept to the classical biological control of weeds was to search the native range of the weed for organisms (insects were the most obvious predators) that could be safely transplanted into the weed's new range. The goal of biological control was not the extirpation of the alien, but rather its reduction to a level that was acceptable. The trick was to find a natural enemy that was highly host specific, but could still survive when its target had been reduced to acceptable densities. This would produce the most cost-effective result, that is, the pest species and its control agent would oscillate within acceptable limits without further expenditure of control effort.

Although early critics of biological control of weeds (Forbes 1880; Thompson 1930) implied that it is a risky and unproven enterprise, its beginnings reach back to 1863 when an introduced cochineal insect (Dactylopius ceylonicus) was used to control naturalized cactus (Opuntia vulgaris) in southern India (Goeden 1978). In comparison, Batra (1982b) noted that the early use of toxic substances (salt, oil, plant extracts, lime, and soot) continued until the mid-1800's when simple herbicides (sulfuric acid, iron sulphate, copper sulphate, arsenic, sodium chlorate, and others) came into general use.

Taylor (1955) doubted that biological control of weeds would produce enough successful results to warrant serious consideration as an alternative to chemical control. Huffaker (1959) countered these and other detractors of the soundness of the concept by asserting that "Textbooks on plant ecology … ignore the role phytophagous insects may have on the composition of vegetation. This omission results from their emphasis on the plant environment and the climatic and edaphic factors, with a consequent neglect of biotic forces …" Wilson's (1964) apprehensions that "the main limiting factor in the application of biological control to weeds lies … in the risk that useful plants will also be attacked" have been tempered by a later assessment (Wilson 1974) that "many important successes and exciting developments" had been achieved, including "control of winter moth (Operophthera brumata) in Canada, and of Crofton weed (Eupatorium adenophorum) in Hawaii and elsewhere … " To these successes, one could add the control of alligatorweed (Alternanthera philoxeroides) in southeastern United States with insects (Maddox et al. 1971; Brown and Spencer 1973; Spencer and Coulson 1976), successful control of an alien plant (Salvinia molesta) in subtropical Australia by an imported weevil, Cyrtobagous singularis (Room et al. 1981), and promising results with the suppression of water hyacinth (Eichhornia crassipes) in Louisiana by an imported weevil, Neochetina eichhorniae (Goyer and Stark 1981).

During the past 80 years, a pattern of investigation has been constructed to guide the work of applied biological control. These procedures have become known as "Classical Biological Control" and were listed by Batra (1982a) in the following sequence.

  1. Weed candidate for biological control is studied to determine its probable origin, potential to cause economic damage, existing natural enemies, and beneficial uses or values.
  2. This information and basic ecological data on the pest species is reviewed by an interagency panel of scientists representing about 10 disciplines or responsibilities; Canadian and Mexican comment is also sought if control organisms are likely to disperse across an international boundary.
  3. After approval by the interagency panel, search and study are begun of the weed's natural enemies in or near the center of its native range.
  4. Ecological studies are done of the most damaging and host specific natural enemies.
  5. Exhaustive screening tests are made with the candidate control agents. Host specificity tests are made with crop, horticultural, and native plants related to the weed. If the candidate attacks only its weed target, it is subjected to starvation tests in the presence of possible alternative hosts. If it does not survive in the absence of its target, it is considered safe for trial in the United States.
  6. A report on the candidate agent is reviewed by interagency scientists; Federal and State quarantine permits for importation and release of the agent are obtained.
  7. The agent is screened in quarantine to eliminate its own parasites.
  8. The agent is released in field cages to observe behavior, survival, and reproduction in the habitat of the target species.
  9. The agent is released in undisturbed habitats that contain extensive infestations of the target weed.
  10. From these foci, the agent multiplies and spreads into adjoining infestations. Cooperators throughout the target weed's range can hasten the spread of the control agent by redistributing local materials.

An additional step should be added, namely, an evaluation of the efficacy of the introduced agents, including their impact on native biota—particularly plants that are closely related to the target weed. Klingman and Coulson (1982) prepared a detailed set of guidelines for the introductions of foreign organisms as biological control agents. Their paper codifies all procedures and identifies the missions of various Federal facilities.

The deliberate and complex nature of a biological control program requires 10 or more years before results are obtained in the field; however, cautious and time-consuming testing procedures have borne exemplary results. Batra (1982b) pointed to more than 100 hostspecific phytophagous insect species introduced worldwide without serious losses arising from introduced agents attacking beneficial plants.

The objectives and methods of biological control are ideally suited to coping with purple loosestrife's invasion of North American marshes and riparian wetlands. The high degree of host specificity demanded by testing and field trials makes biological control extremely compatible with the wildlife or natural area manager's goal of maintaining community integrity and ecological health. In addition to having a specific target, biological control is relatively gentle—it eschews extermination and aims instead at reducing the pest species density to some acceptable minimum.

Criteria of a Biological Control Target

Several criteria are useful for evaluating the need for a biological control program. First, the introduced weed must have clearly demonstrated a potential for doing serious damage over a large area. Second, alternative means of effective control must have been tried and found to be inadequate or infeasible. Third, no economically or ecologically useful plants should be near relatives of the target species. We think that our previous discussions of the spread, impact, and control of purple loosestrife have met the first two criteria. Of the third criterion, L. salicaria has no near relatives of economic value. Among more than 20 genera and 400 species of Lythraceae throughout the world, only a few species are cultivated as ornamentals. Of these, crape-myrtle (Lagerstroemia indica) is probably the most widely planted—`mostly in the southern United States where it also occurs as a naturalized shrub. In addition, several members of the genus Cuphea are cultivated in greenhouses and open plantings; however, the value of these ornamental species should be discounted by the spread of clammy cuphea (C. petiolata) into pastures, meadows, and gardens in southern New England and the Midwest (Fogg 1956). Both of these cultivated genera have tropical affinities, occur south of the probable range of spread of L. salicaria, and should not be threatened by biological control agents in northern wetlands. Nor is there a high probability of introduced biological control agents causing damage to native flora. Fernald (1950) listed only 6 genera and 14 species of Lythraceae in central and northeastern North America. Decodon, an adjacent genus to Lythrum, contains a single species: D. verticillatus (swamp loosestrife). This attractive native is probably the most frequent member of Lythraceae to occur in wetlands threatened by L. salicaria. D. verticillatus is an herbaceous perennial that shows many of the aquatic adaptations of L. salicaria. It has woody stems that develop aerenchyma below the water line. It can also spread by setting adventitious roots from stem tips that lodge on moist soil. Swamp loosestrife also has a similar growth form to L. salicaria; its stems rise from a central rootstock and reach heights (2.5 m) that equal or exceed purple loosestrife. Swamp loosestrife is not only a highly valued marsh plant, it is one of the few aquatic emergents that seems able to hold back encroaching L. salicaria in undisturbed marshes. Decodon also seems to have a higher tolerance to shade than does Lythrum and therefore is often found growing under wetland forest canopies that would exclude Lythrum salicaria. L. alatum (winged loosestrife) is another native American Lythrum that will need to be considered for the host specificity menu; it is an associate of L. salicaria in the wet prairies of the Midwest (Levin 1970). Fernald (1950) described L. alatum as extending from Ontario and northern New York to British Columbia, south to Georgia, Louisiana, and Texas; he referred to it as adventive in the Northeast and New Jersey. Bailey and Bailey (1976) described L. alatum as "more or less naturalized … " in the northeastern United States. Thus, the northward spread of L. alatum may have made the core of its native range vulnerable to biological control agents used against L. salicaria. Two other native Lythrums (L. lineare and lanceolatum) are southern in distribution—outside the range of L. salicaria infestations. All of these plants should be included in host specificity tests of potential biological control organisms; however, L. alatum and D. verticillatus should receive special attention since the spread of introduced agents will likely reach their natural habitats.

Susceptibility of Lythrum salicaria to Biological Control

Several characteristics of L. salicaria make it an ideal candidate for biological control. Its sturdy perennial growth form guarantees that an introduced phytophagous control agent will find a sustained annual production of herbage. If Southwood's (1977) idea of arranging pest species on an r-K continuum is applied, the population dynamics of a mature monospecies stand of purple loosestrife clearly identify the weed as a K-strategist, that is, it "maintains a steady population at or near the carrying capacity of the habitat …" Furthermore, the wetlands occupied by the weed are themselves quite stable and isolated from the traumatic community disturbances that beset plant and insect populations on adjacent croplands. Moreover, the more or less continuous distribution of purple loosestrife along eastern roadsides and waterways creates natural pathways of spread for introduced control agents. The conspicuous floral displays of purple loosestrife also make the plant easy to locate and quantify in surveillance studies.

The distribution of monospecific stands of alligatorweed along waterways in the southeastern United States is similar to the distribution of purple loosestrife in the northeastern and north-central States. In 1965, the successful introduction of a South American flea beetle (Agasicles hygrophila) into a dense mat of alligatorweed in northeastern Florida was followed by a rapid expansion of the Agasicles population (Spencer and Coulson 1976). In 1966, more than 9, 000 Agasicles were captured in heavily damaged alligatorweeds near the original release site; these captives were transferred to other alligatorweed infestations in Florida, Georgia, South Carolina, and Mississippi. Spencer and Coulson (1976) noted, "Natural dispersion of the alligatorweed flea beetle occurred rapidly at release sites as the beetles destroyed the alligatorweed. In August of 1976 this ability to disperse resulted in beetles being found as far north as Waycross, Georgia, and at another site some 121 km northwest of the Ortega River release site." It was necessary to introduce two additional South American insects before substantial control was achieved; however, the rapid dispersion of these agents in aquatic and wetland habitats suggests the vulnerability of L. salicaria to introduced natural enemies in interconnected wetland complexes in the northern United States and southern Canada.

The lack of aggressiveness of purple loosestrife in western and central Europe suggested to Batra et al. (1986), that these were promising regions to explore for natural enemies of purple loosestrife. They found 120 phytophagous insects attacking L. salicaria, and considered 14 of these species to be host specific to Lythrum spp. Of these Lythrum spp. foragers, they recommended detailed ecological and host-specificity studies for six species: a cecidomyiid fly (Dasineura salicariae), whose galling can reduce purple loosestrife foliage by 75% and seed production by 80%; a stem and root boring weevil (Hylobius transversovittatus); two chrysomelids (Pyrrhalta calmariensis and P. pusilla) that can cause nearly 50% defoliation; and two weevils (Nanophyes marmoratus and N. brevis) that mine ovaries and seeds. Although additional survey work in northern Europe needs to be completed, the results of Batra et al. (1986) indicated that the chances of successful biological control of L. salicaria in North America are excellent.

Benefits versus Costs

Before a long biological control effort can be recommended, the costs, risks, and benefits of such a program need to be assessed. Harris (1979) estimated that in Canada the cost of development and implementation of a biological control program for a single weed species was $1.2 to $1.5 million. In the United States, Andres (1977) calculated that the alligatorweed program cost about $1 million as of 1976 with a satisfactory level of control achieved. Andres also estimated that the control of Klamath weed (Hypericum perforatum) in California cost $0.5 to $1.0 million with benefits of $40 million (as of 1964) projected for the first 12 years following achievement of control. Our estimate for the cost of a proposed 10-year interagency project on the biological control of purple loosestrife (Table 2) is about $0.5 million. Paralleling the work of the interagency project, Table 2 also lists cooperative efforts that will need to be undertaken by Federal, State, and local governments and various private organizations to carry on the application of an integrated control program. Included here are hidden costs, such as work reassignments to achieve a new field objective under present employment ceilings; new supply, equipment, and travel costs would also be added. We estimate that the cost of these cooperative efforts for the first 10 years would equal the cost of the U.S. Department of Agriculture contract—roughly another $0.5 million. At least 20 States would be beneficiaries of a control program; distributing the costs of the Federal and cooperative programs over a 10-year span would reduce the annual cost per State to $5,000—a fraction of the amount spent on chemical control in several of these States. The USDA's Agricultural Research Service has conducted some preliminary explorations on purple loosestrife (Batra et al. 1986) and is currently funding some additional field studies in Europe by the Commonwealth Institute for Biological Control. Table 2 provides for a year of work to conclude the first phase (Search and Testing) of the 10-year program.

Table 2. Schedule and cost estimate for 10-year interagency (USDA-USDI) project on biological control of purple loosestrife (Lythrum salicaria).
Phase USDA biological control project Year Cost Cooperative efforts by Federal, State, and local governments and private sector, in concert with USDA project
I Search and testing in Europe for biological control agents 0 + 1 $40,000 During USDA studies in Europe: develop remote sensing techniques for surveillance of L. salicaria; measure losses in wildlife and agricultural values following L. salicaria infestation.
II Quarantine and host specificity studies in U.S. 0 + 2
0 + 3
0 + 4
During USDA quarantine studies: establish baseline transects for remote sensing of impact and status of L. salicaria.
III Propagation, release, and establishment of biological control agents 0 + 5
0 + 6
0 + 7
During USDA release and establishment phase: cooperate with release program; repeat L. salicaria transects.
IV Monitoring and evaluation studies 0 + 8
0 + 9
0 + 10
During USDA evaluation studies: repeat L. salicaria transects; support USDA monitoring and followup.
Total USDI contribution to USDA $505,000  

The cost ($ 1.0 million) of a 10-year biological control program challenges us to show a reasonable prospect of benefits in excess of this proposed expenditure. Recognizing at the outset that we need better information about the quantitative and qualitative damage caused by the invasion of purple loosestrife into a wetland habitat, we nonetheless see several methods of appraisal that will give us some sense of the value of the wetland resource that is threatened.

Wetland Resource Values

All evaluations of wetlands need to be set against the background of the destabilization and continued shrinkage of this resource. Shaw and Fredine (1956) examined the fate of wetlands in California, Florida, and five central midwestern States. From an 1850 base of 15 million ha, wetlands within these States were reduced by 9.3% in 1906, 33.2% in 1922, and 45.7% around 1955. These values are distinctly lower than the shrinkage estimate we have given earlier (66%) from Auclair's studies in southern Wisconsin; however, Auclair's method was based on land survey data and was outside the area sampled by Shaw and Fredine. In addition to drainage and cultivation, Jaworski and Raphael (1979) pointed to the reduction in diversity and biological stability of remaining wetlands as a problem that has not been adequately studied.

Methods of evaluation. Methodologies for evaluating wetlands described in a symposium on wetland functions and values (Greeson et al. 1979) included a hierarchical approach (Odum), nonconsumptive uses (Reimold and Hardisky), and social values (Foster). Linder and Hubbard (1982) discussed wetland values in the prairie pothole region. We have chosen to use three appraisal methods to lend perspective to the value of wetlands threatened by purple loosestrife. We have also limited our evaluations to 12 eastern and 7 north-central States where L. salicaria has already demonstrated its potential for widespread habitat degradation.

Realty value. Table 3 presents Shaw and Fredine's (1956) estimates of the distribution of four freshwater wetland types in 12 States that compose the Atlantic Flyway North and 7 States that make up the Mississippi Flyway North. All four of these wetland types are more or less vulnerable to invasion by L. salicaria but, for our purposes, we will include only Types 3 and 4 wetlands—the shallow and deep marshes that are especially susceptible to infestation and are prime waterfowl habitats. Cowardin et al. (1979) have criticized Shaw and Fredine's estimates as being conservative, but until more accurate assessments are available, these estimates best suit our needs. If we use $750 (R. Erickson, North Central Region, USFWS, personal communication) as a modest estimate of the value of a hectare of shallow or deep freshwater marsh, the realty value of 499,000 ha threatened by L. salicaria in the Atlantic and Mississippi flyways (north) is $374 million.

Table 3. Area (hectares) of four freshwater wetland types in the northern zones of the Atlantic and Mississippi flyways.a
Wetland type Atlantic Flyway North Mississippi Flyway North
1. Seasonally flooded basins 480 453,600
2. Inland fresh meadows 12,280 953,320
3. Inland fresh shallow marshes 14,360 303,400
4. Inland deep fresh marshes 10,280 171,080
Total fresh wetland 37,400 1,881,400
a From Shaw and Fredine (1956).

Fur harvest. The dependence of the muskrat on broadleaved cattail for food and cover marks the productivity of this fur crop as threatened by the expansion of purple loosestrife. From the work of Deems and Pursley (1978) we extracted data on the harvest of muskrat fur from the 19 States that make up the northern zones of the Atlantic and Mississippi flyways (Table 4), and calculated that the mean annual value of pelts taken from 1970-71 to 1975-76 was $13,982,133. The mean annual price per pelt was $2.69.

Table 4. Muskrat fur harvest and honey sales from the Atlantic Flyway North and Mississippi Flyway North.
State Fur value in $ (thousands)a Honey sales in $ (thousands)b
Atlantic Flyway North (n = 12)
   Maine 133 36
   New Hampshire 22 1
   Vermont 385 58
   Massachusetts 144 39
   Rhode Island 7 -
   Connecticut 7 12
   New York 385 1,456
   New Jersey 1,760 93
   Pennsylvania 1,099 266
   Maryland 613 26
   Delaware 177 -
   West Virginia 134 79
   Subtotal 4,866 2,066
Mississippi Flyway North (n = 7)
   Ohio 1,698 402
   Indiana 853 144
   Illinois 1,197 242
   Iowa 1,204 1,428
   Michigan 937 1,538
   Wisconsin 1,698 2,440
   Minnesota 1,528 4,560
   Subtotal 9,115 10,754
Total 13,981 12,820
a From Deems and Pursley (1978).
b From 1974 Census of Agriculture (USBC 1977a).

Migratory bird hunting expenditures. The amounts of money people spend in pursuit of recreation have been generally accepted as a valid estimate of the resource base that supports the activity. We used regional summaries of the 1980 national survey of outdoor recreation (USFWS and USBC 1982) to calculate that 0.5 million migratory bird hunters in the 19 northeastern and northern midwestern States spent an average of $120 per year, or a total of $65.6 million on their sport. Included in this total were hunters who sought geese (about 1/3), woodcock (about 1/10), and doves (<1/10); since these birds are not often found in marsh habitat, we have halved the total migratory bird hunting estimate to $32.8 million.

Outdoor recreation expenditures. Increasing numbers of nature enthusiasts find fulfillment in outdoor settings. Their activities and expenditures have been compiled as "nonconsumptive wildlife use" (USFWS and USBC 1982). Again, we calculated from regional summaries that 12.6 million participants spent $129.86 annually, for a total of $1,639 million. These people traveled away from their residences for the primary purpose of observing or photographing native animals in their natural habitats. About 1/10 of the sites visited were marsh or wetland, hence we computed that the annual expenditure for viewing marsh or wetland wildlife was $163.9 million.

Benefit-Cost Ratio

From the preceding estimates of wetland resource values we constructed a benefit-cost analysis (Table 5) of the proposed biological control program on purple loosestrife. At the outset, we grant that a benefit-cost analysis is difficult to apply to a resource whose output is primarily aesthetic, that is, promotes a feeling of well being. About 86% of the resource value that we were able to identify in Table 5 was based on the amounts of money spent in pursuit of outdoor recreation. Among gains (benefits) we have listed, the annual value of wetland habitat was determined by the current price per hectare amortized at 5% over 20 years. This is a conservative estimate, in that land seldom sells for as little as $750/ha ($300/acre). Moreover, wetlands often occur as neglected swales in a surrounding matrix of farm, range, forest, or residential tracts—making them difficult to purchase or manage as productive wildlife habitat. Although the 1974 Census of Agriculture (USBC 1977b) listed 24,764 farms cutting a wild hay crop valued at $13.0 million, we were not able to identify that portion cut in marshes or riparian meadows. The annual fur statistics used to estimate this resource came from a set of years (1971-76) that included low to moderate (but not high) pelt prices.

Table 5. Benefit-cost analysis of resources and values at risk to purple loosestrife infestation versus costs of a 10-year biological control program.a
Benefit-cost Annual value in $ (millions)
   Realty value of threatened wetlands
      ($374 million amortized at 5% for 20 years)
   Wild hay and pasture no estimate
   Fur harvest (muskrat) 14.0
   Migratory bird hunting expenditures 32.8
   Wildlife observation and photography
Annual resource value
Biological control saves 20%b
of annual resource value

×   .20 

   Total annual benefit 45.9
   Combined cost of biological controlc 0.1
   10% of annual honey sales in 19 Statesd 1.3
   5% of annual sales of herbaceous ornamentalse 0.3
   Total annual cost 1.7
Benefit-cost ratio = 45.9/1.7 = 27.0
Estimated 27 units of benefit for each unit of cost
a Applied to 12 northeastern States and 7 northern Midwest States whose wetlands have been invaded by L. salicaria.
b Projected for 20 years with 75% reduction in weed densities assumed at end of 10-year biological control program.
c Includes $0.05 × 106 for USDI contract (Table 2) plus $0.05 × 106 for workload reassignments and other costs to Federal, State, and local governments and private sector.
d From USBC 1977a.
e From USBC 1973.

To give reasonable balance to our calculation of the benefits to be derived from biological control, we postulated that an acceptable result at the end of the 10-year program would be a 75% reduction in the overall densities of the weed, thereby saving 20% of the annual resource value. The projection of a 20% saving assumes that reductions in purple loosestrife densities cannot be expected to begin until the 7th year of the program. Even if dramatic reductions are achieved by the 10th year, substantial savings in resource value would not accrue until the second decade. Moreover, in the worst situation, if L. salicaria were to continue spreading at an unchecked rate, total resource value would not fall below some unknown minimum. In estimating costs or losses of biological control, we listed the annual cost of the interagency project ($0.5 × 106) and also an equal amount to cover hidden costs among many levels of cooperators who will participate in field activities. In our allowance for loss of nectar and pollen forage that might accompany biological control of L. salicaria, we note that Iowa, Michigan, Minnesota, and Wisconsin account for 78% of annual honey sales from the 19 northeastern and northern midwestern States (Table 4). Since wetlands in these States are less than half colonized by purple loosestrife (Fig. 19), we believe our estimate of potential loss to be fair. A Wisconsin honey producer (J. Mills, personal communication) advised us that although purple loosestrife was often useful forage for hives that were being "hardened" for winter, it was not important in honey production. We were not able to obtain an accurate estimate of the value of nursery stock (infertile cultivars of L. salicaria) that might be damaged by biological control organisms. Total nursery sales of herbaceous ornamentals in the United States in 1970 were $6,045,661 (USBC 1973). Purple loosestrife makes up some small part of this total; we doubt that it exceeds 5%.

GIF-Map of purple loosestrife in North American wetlands as of 1980
Fig. 19. Status of Lythrum salicaria infestations in North American wetlands as of 1980.

In summary, we conclude that a conservative projection of the benefit-cost ratio of a biological control program would yield at least 27 units of benefit for each unit of cost. As indicated earlier, we confined our analysis to the eastern half of the United States where L. salicaria has already shown that it can invade and displace native marsh vegetation.

Other Risks and Rewards

In addition to the cost of operating a biological control program, we need to consider other ecological and economic risks that would accompany these efforts against purple loosestrife. The primary risk for natural resource managers would be that an introduced organism might alter its behavior in its new environment or have some unanticipated secondary effect that would be ecologically damaging to the native biota. Serious ecological damage probably occurred in the history of biological control in Hawaii. Zimmerman (1958) noted, "Swezey was the last of the entomologists to have seen many of the endemic Hawaiian Lepidoptera in a semblance of their natural abundance. The importation of parasites to control various moths of economic importance, together with the accidental importation of other parasites has resulted in wholesale slaughter and near or complete extermination of countless species … Many are forever lost."

Another critic from Hawaii (Howarth 1983) declared his alarm that reviews of biological control have implied that the method was without environmental risk and observed, "Once the target pest appears adequately controlled researchers have turned their attention to other pests." He softened his argument by acknowledging that there have been extremely limited funds for research on the impacts of biological control introductions, and noted that the Hawaii State Department of Agriculture has begun including vulnerable native arthropods in the testing protocol for candidate biological control species. The difficulties of designing and executing field studies that would adequately monitor the impact of introduced species caused Simberloff (1985) to observe, "estimates of no effect are actually estimates of no obvious, dramatic effect." In reviewing the record of biological control, Buckingham (1984) granted that the method can never be declared absolutely safe, but cited an 80-year summary (Julien 1982) of 192 biological control agents (mostly insects) that had been used against 86 weed species without unleashing an important pest. Buckingham also noted that as ever more restrictive measures are added to assure safety (especially tests for possible damage to native plants) many potentially beneficial agents must be dropped from consideration or have their clearance unduly delayed.

To balance these ecological risks, the stress of spraying herbicides on wetland habitats would be eased by a biological control program. The high costs of product development have dictated that herbicides used on wildlife habitats be formulated from chemicals developed for application on relatively simple agricultural ecosystems. The application of a broad-leaved herbicide to a monocot crop or a pine plantation can handily achieve the manager's goal of suppressing competition from dicot weeds, but it is a mistake to apply such a simple treatment to a natural wetland that may support a complex of more than 100 species occupying several photosynthetic levels and supporting a fauna living at many trophic levels. From the viewpoint of healthy species diversity, the use of a broad-spectrum herbicide (glyphosate) is an even more drastic blow to a wetland community. Indeed, continued spraying of wetlands with herbicides will tend to shift survivorship of community members to resistant species. A second benefit from biological control would be savings on the cost of applying chemicals to wetlands infested with purple loosestrife. We estimate that the annual cost of spraying herbicides (principally 2-4,D and glyphosate) on purple loosestrife in the Northeast and northern Midwest exceeds $75,000.

On balance, the decision to apply (or not to apply) biological control to purple loosestrife hinges on weighing the uncertainty of a minor risk of environmental damage against the certainty of continued degradation of North American riparian and wetland habitats by this weed. The greatest problem facing the initiation of a biological control program is the widely scattered responsibilities and proprietary interests at private, State, Provincial, national, and international levels. Nevertheless, other weed threats have affected a wide array of potential beneficiaries and have been successfully confronted with interagency and international agreements, namely, alligatorweed (Spencer and Coulson 1976), water hyacinth (Wright and Center 1984), and musk-thistle, Carduus nutans (Linehan 1982).

Purple loosestrife was approved as a candidate for biological control by an interagency committee in 1980. As of this writing, an interagency agreement between the USDA and USDI has received favorable review; Federal funds for 1986-87 have been appropriated.

Conflict Resolution

DeLoach (1978) recognized that the generally beneficial effects of biological control programs can also have negative economic or ecological impacts on nontarget organisms. Andres (1980) suggested some guidelines for the resolution of these conflicts. He also pointed to the need for improved methods of predicting the effects of biological agents on nontarget plants. We are aware of three potential conflicts that may result from the decimation of L. salicaria by an introduced biological agent.

Honey Production

Although a successful biological control program would further reduce L. salicaria's minor role as a nectar source, an analysis of the 10 leading honey-producing States in 1974 (Table 6) shows that none of the top producers have high acreages of loosestrife infestation (Fig. 19). Minnesota and Wisconsin have very large areas of glacial marsh, but less than half of these wetlands are colonized by purple loosestrife. A further analysis of honey production in these top 10 States shows that they contain 34% of the Nation's farms with bee colonies and 65% of the national total of colonies; they also account for 68% of national honey sales. These data reflect the migratory pattern of large honey producers. They winter their colonies in southern States and move them north each spring in search of foraging areas. Nevertheless, to the small honey producer in the Northeast and northcentral United States, the prospect of additional losses of bee foraging plants is difficult to face. Since the advent of carbaryl insecticides, northeastern and northcentral beekeepers have sought wetland foraging areas for relief from inadvertent mortality caused by insecticides applied to upland forage crops.

Table 6. Number of colonies of bees and value of honey on farms with sales of $2,500 or more in top 10 honey-producing States.a
Rank State Inventory of colonies Honey sales ($1,000)
No. of farms No. of colonies
1 California 447 305,753 8,137
2 South Dakota 91 115,904 4,505
3 Idaho 104 106,376 3,178
4 Minnesota 237 103,222 4,560
5 Washington 157 97,655 1,816
6 Nebraska 150 98,877 3,667
7 Texas 368 77,717 2,651
8 Florida 174 75,338 2,339
9 Wisconsin 435 70,330 2,440
10 North Dakota 53 61,552 2,926
Total 2,216 1,112,724 36,219
U.S. Total (48 States) 6,459 1,708,383 53,614
a From 1974 Census of Agriculture (USBC 1977a).

The resolution of this conflict might turn on the realization that a successful biological control program would not seek the extermination of purple loosestrife. A 75% reduction in the density of purple loosestrife would probably allow wetland managers to cope with the species; surviving L. salicaria would provide a substantial base of pollen and nectar forage amidst a healthier mix of native plants. The American Bee Journal has already opened its pages to proponents of both sides of this conflict (Bunch 1977a, b; Hayes 1979; Hughes 1977; USFWS 1979). This is a wholesome beginning.

Horticultural Sales

Several aspects of the sales of horticultural stocks of Lythrum spp. need to be considered. First, the annual national sales of purple loosestrife are probably a minor item compared with the environmental costs imposed by the inadvertent escape of wild-type stock into adjoining wetlands. Although they make very attractive perennial displays, the various cultivars of Lythrum are not a significant part of the annual sales of nursery products. Total nursery sales of all herbaceous ornamentals in the United States in 1969 were estimated at $6,045,661 (USBC 1973). Lythrum stock was not important enough to be identified. Second, a ban on the shipment and sale of Lythrum salicaria should not block the growth and sale of infertile hybrid stock. Examples of infertile hybrid stock that we have seen have small, delicately set leaves and flowers in greatly elongated inflorescences that do not produce seed capsules. Third, most landscape settings that include infertile hybrid stock would be far enough removed from wetland habitats to screen the plantings from the full impact of control agents. Moreover, a protective insecticide spray program could probably be developed to reduce damage to an acceptable minimum.

Recent actions of the Minnesota Nurseryman's Association (B. Harper, personal communication) have demonstrated that they are anxious to participate in a resolution of this potential conflict. Other states would do well to follow their responsible example.

Previous Section -- Recent Control Efforts
Return to Contents
Next Section -- Containment Strategy and Tactics

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: Webmaster
Page Last Modified: Saturday, 02-Feb-2013 06:29:54 EST
Sioux Falls, SD [sdww54]