Field studies were conducted to characterize the predominant nonstructural carbohydrates in roots and crowns of purple loosestrife and to determine the seasonal fluctuation of root and crown carbohydrates in three Minnesota wetland habitats. Starch was the primary nonstructural carbohydrate present, and concentrations were consistently higher in roots than in crowns. Starch content decreased following shoot emergence and declined until flower bud formation, at which time levels reached seasonal lows. Following bud initiation, levels of starch increased during flowering, and this trend continued through to plant senescence in late September or October. Sucrose was the predominant soluble sugar in roots and crowns of purple loosestrife, but fructose and glucose were detected. Levels of sucrose in roots and crowns followed the same seasonal trends as starch.

Starch levels, used as a measure of plant stress, were not consistently reduced in root or crown tissue of purple loosestrife plants after 2 yr of severe Galerucella calmariensis or Galerucella pusilla (Coleoptera: Chrysomelidae) defoliation. Early in the season, defoliation from Galerucella spp. approached 100%, but the majority of Lythrum salicaria plants regrew by the end of August, resulting in an average reduction of 81% of the aboveground biomass compared to the control. The stress imposed by Galerucella spp. defoliation was less than that achieved from more severe stress imposed by mechanical shoot clipping at 2- or 4-wk intervals from June to October. Both shoot-clipping treatments killed the majority of plants after one growing season. Galerucella spp. feeding reduced plant stature, which may reduce competitiveness. However, considering the extensive carbohydrate reserves present in the large woody crowns of Lythrum salicaria, it will require in excess of 2 yr of consistent, severe leaf defoliation to cause plant mortality. A combination of stresses, such as winter crown injury, or other biological control agents in addition to Galerucella leaf defoliation may be required for plant mortality.

Field experiments were used to assess how distance mediates the nontarget effect on crepe myrtle by two chrysomelid beetles that were introduced to the United States in 1992 for biological control of purple loosestrife. Previous laboratory tests in Germany and concurrent tests in Oregon showed that although the control organisms can feed on crepe myrtle, they cannot complete development. Therefore, we predicted that negative effects on crepe myrtle would decrease with distance from the purple loosestrife stand. To test this prediction, cohorts of both plant species were transplanted at increasing distances (0, 5, 15, 30, and 50 m) from the colonization source. We found that leaf damage inflicted by the beetles was negatively correlated with increasing distance. Damage was significantly lower at each distance for crepe myrtle plants than for purple loosestrife plants, with a mean difference of 22% and a 95% confidence interval ranging from 12 to 31%. Extensive defoliation of crepe myrtle was limited to within 30 m of the edge of the loosestrife stand. Plant yield was negatively correlated with damage: the closer plants were to the purple loosestrife stand, the greater the suppression of biomass in both plant species. However, loosestrife biomass decreased significantly more quickly than crepe myrtle biomass, with a mean difference in slopes of 0.035 and a 95% confidence interval ranging from 0.022 to 0.048. Our results suggest that release of the Galerucella beetles in North America poses little risk to crepe myrtle. Beetles can feed but cannot complete their life cycle on crepe myrtle, and damage to crepe myrtle approaches zero approximately 50 m from the beetle colonization source.

In wetlands, drought or managed late-summer drawdowns create exposed mudflats that provide an excellent substrate for germination of purple loosestrife seeds. If late-emerging purple loosestrife seedlings survive the winter, new or expanding populations of purple loosestrife will result. Spring survival was determined for overwintered purple loosestrife seedlings from seeds planted at weekly intervals in late summer or fall of the previous year. Seedlings of purple loosestrife that emerged from late July to early August had the greatest survival rates and the greatest shoot dry weight, and they were the tallest the following spring. However, 37% of purple loosestrife seedlings that emerged in late August, although stunted, generated a crown that was able to overwinter successfully and regrow the following spring. The number of growing degree days accumulated from planting date to October 6 (the average date of first frost for Minneapolis and St. Paul, MN) was 1,424 for seedlings from seeds planted on July 21 but only 219 for seedlings from seeds planted on September 15. Purple loosestrife seedlings that emerge during late summer through early September in Minnesota may survive the winter to create additional purple loosestrife weed problems in wetland mudflats caused by artificial drawdowns or droughts.