RESPONSES TO SELECTION ON FLIGHT BEHAVIOR IN A MIGRATORY POPULATION OF MILKWEED BUG ( ONCOPELTUS FASCIATUS )

Abstract

The nature of genetic variation and covariation in-fluencing life-history traits has been the subject of an increasing n umber of theoretical and empirical studies (Steams, 1983; Arnold, 1987). Knowledge of the so-called "genetic structure" among traits is essential to understanding both the potential and constraints for phenotypic evolution in response to a particular selec-tive regimen, be it natural or artificial (Loeschcke, 1987). Indeed, artificial selection is widely used to detect ge-netic associations between traits, as revealed by cor-related changes in phenotype when selection is applied to a single trait (Falconer, 1981 pp. 286-287). We have used artificial-selectio n experiments to compare the genetic structure of migratory and non-migratory populations of the large milkweed bug, On-copeltus fasciatus (Palmer and Dingle, 1986; Dingle et al., 1988). In these studies, we chose wing length as the target of directional selection because previous work indicated that it should respond rapidly and that other traits might show correlated responses (Hegmann and Dingle, 1982). Migratory (Iowa) and nonmigratory (Puerto Rico) populations showed similar rapid direct responses to selection on wing length but differed in their pattern of correlated responses. Of particular in-terest was the result that the migratory population showed positive correlated responses in the propensity for long-duration (presumed migratory) flight behavior (Palmer, 1985) and early reproductive output (Palmer and Dingle, 1986) while the nonmigratory population showed no significant correlated responses in these same traits (Dingle and Evans, 1987; Dingle et al., 1988). In both populations, adult age at first reproduction failed to show any correlated response to selection on wing length. Thus, for the migratory Iowa population the positive genetic correlations among body size, flight potential, and fecundity provide the potential for these traits to evolve together in response to selection pressures that favor a colonizing life history. However, in nature, wing length may not be the direct, or only, target of selection. For example, the flight potential ofmigratory milkweed bugs, which allows them to reach and col-onize more northern habitats such as those occurring in Iowa, may be a more direct target of natural selec-tion. The purpose of this study was to see whether selection on flight propensity in this migratory popu- lation would yield a genetic correlation structure among body size, flight, and fecundity similar to that revealed when wing length was the target of artificial selection. MATERIALS AND METHODS Milkweed bugs, Oncopeltus fasciatus, used in these experiments were descended from adults that were col-lected from field sites in eastern Iowa (Johnson County) in September 1981. Offspring were reared in full-sib families at 27" ± OSC under a 14L: 10D photoperiod, according to the experimental protocol described by Palmer and Dingle (1986). Bidirectional selection on wing length was performed for 13 generations (Palmer and Dingle, 1986; Fig. I). In generation 8, eggs from unselected control lines were collected to become the base population for selection on flight propensity. These progeny were reared as full-sib families under a 14L: IOD photoperiod at 23° ± 0.5OC. At eclosion, adults were separated by sex and marked so that the flight behavior of individuals could be monitored. Each generation, approximately 50 bugs of each sex were tested in each line. Adults were flight-tested according to the protocol described by Palmer (1985): individuals were flight-tested on days 8, 10, and 12 posteclosion. Each test consisted of five consecutive flights, and the durations of these flights were summed to give the total duration of flight per bug per day. Individuals whose flight times totaled 30 minutes were scored as fliers and were not tested any more on that day. The longest of the three daily flight totals was the selection criterion for flight propensity among individuals. A mass-selection pro-tocol was employed in this experiment: the 10 female and 10 male bugs that displayed the longest and short-est duration flights based on the selection criterion were selected to be parents of the next generation of "flier" and "nonflier" lines, respectively. Parents in the flier line invariably had at least one flight score of 30 min-utes or more, and parents in the non flier line always had flight scores of less than two minutes (frequently o minutes). Their progeny became generation I of the