The effects of selection of goldenrod galls’ height and diameter on Eurosta solidaginis Abstract We examined the effects of selective
agents acting on the height and diameter of Eurosta solidaginins goldenrod galls. We compared the goldenrod galls that contained live Eurosta with the galls of dead Eurosta, examining height and diameter. We hypothesized that selection favors lower galls that are less visible to avian predators and larger galls that are less likely to be successfully penetrated by the wasp Eurytoma gigantean due to their larger size. We observed significantly larger diameters of live Eurosta’s galls than dead Eurosta’s galls but did not find significant differences between the heights of galls of live and dead Eurosta. Introduction A measure of fitness in the goldenrod gall fly, Eurosta solidaginis, is larval survivorship. Gall size has been implicated in differential survivorship of Eurosta larva (Weis and Abrahamson, 1985), which makes it a likely candidate to experience selective pressures. We examined the relationship between gall diameter and larval survivorship, and gall height and larval survivorship to gain insight into the selective effects of parasitism and predation on Eurosta galls. Eurosta oviposits in June, inserting an egg into the terminal bud of the goldenrod Solidago altissima (Weis and Abrahamson, 1986). The larva tunnels into the stem immediately below the apical meristem (Weis and Abrahamson, 1985) where it releases chemicals that stimulate the host goldenrod to form a gall in which the larva will live until emergence the next May (Abrahamson and Heinrich, web source: Bucknell University). The gall forms from plant tissue, but the larva’s genotype affects gall size
(Weis and Abrahamson, 1986). The heritability of gall size in Eurosta supports the theory that Eurosta have been selected to stimulate gall formation and that the gall response is a manipulation of the goldenrod host plant by the larva (Weis and Abrahamson, 1986). Predators and parasites account for the majority of larval mortality (Weis and Abrahamson, 1985, 1986). Eurosta’s natural enemies, birds and parasitic wasps, differentially affect larva survival based on goldenrod phenotype (size measured by diameter) and exert both directional and stabilizing selection (Weis and Abrahamson, 1986). Directional selection increases the proportion of a certain phenotype in a population as a result of greater fitness or reproductive success of individuals with the phenotype. Stabilizing selection selects towards the mean of a phenotype that varies on a continuum when displaying phenotypic extremes causes decreased fitness. Stabilizing selection narrows the variance of a certain phenotype in a population. Weis and Abrahamson (1985) have shown that small gall size increases the likelihood of successful attack by the Eurytoma gigantea wasp, causing a selective pressure toward larger gall size. The selective pressures at work are not unidirectional; large size increases susceptibility to attack by the Downy Woodpecker and the Black-capped chickadee, producing an opposing selective pressure toward small gall size (Weis and Abrahamson, 1986). In our examination of selective pressures acting on Eurosta, we expected to see directional selection favoring larger gall diameter and to see stabilizing selection modifying the effects of directional selection by selecting against very large and very small gall diameter. Gall height is influenced by the height of the goldenrod at the time of oviposition and the goldenrod’s rate of growth. Eurosta females have the capacity to discriminate
when choosing goldenrods for oviposition based on signals that indicate host quality i.e. characteristics of the plant that increase her larvae’s chance of survival (Anderson et al. 1989). Horner and Abrahamson (1999) found that goldenrod genotype significantly affects gall height at the time of oviposition. Anderson et al. (1989) observed that Eurosta females select goldenrods that were growing significantly faster and were taller than the rest of the population at the time of oviposition, though 3 and 4 weeks later they were growing slower than the rest of the population (the heights remained significantly different). They identified a probable cause for the selection of faster growing goldenrods: faster growing goldenrods are more responsive to the stimulus of the Eurosta larva. Anderson et al. proposed that the subsequent reduction in growth rate was caused by allocation of limited resources to gall formation. Anderson et al. (1989) also found greater larval mortality in goldenrod genotypes that are generally more resistant to gall formation; therefore, Eurosta females do not only select goldenrods that are more susceptible to gall formation but also select goldenrods that increase the chances of larvae surviving. If heritable traits of Eurosta females affect gall height and gall height affects larval survival, natural selection can act on gall height. We predicted that natural selection favors shorter galls. We examined the effect of goldenrod gall height and gall diameter on survivorship of Eurosta in order to identify if selective pressures were acting on Eurosta, influencing gall phenotype. We measured gall height and diameter and categorized galls as containing alive or dead Eurosta larva. We analyzed the differences in gall phenotype between the alive and dead samples for height and diameter to determine if there were significant differences between the two samples that might indicate selection.
We explored the effects of selection on gall diameter by comparing the difference between the diameter of galls containing live larvae and the diameter of galls containing dead larvae. We predicted that galls containing live Eurosta larvae would have greater diameter than galls in which the Eurosta larvae had died because small galls are more vulnerable to parasite attack (Weis and Abrahamson, 1985). Galls have finished growth during the period of E. gigantea oviposition (Weis and Abrahamson, 1985). The smaller diameter of galls containing dead larvae does not reflect terminated growth after larval death, but would be consistent with directional selection against small gall size. We also predicted stabilizing selection as a result of the opposing selective pressures exerted by avian predators that are more likely to attack large diameter galls (Weis and Abrahamson, 1985). We predicted that natural selection acts on gall height. We hypothesized that it favors decreased gall height as a result of increased bird predation if taller galls are more attractive or apparent to birds. Materials and Methods Lab instructors collected galls on January 9th through January 13th 2006 at Hildacy Farm. We collected data on January 19, 2006: we measured gall height in centimeters with a measuring tape held straight from the base of the stem where the golden rod was cut at ground level to the midpoint of the gall. We measured gall diameter in millimeters at the widest part of the gall with digital calipers which were first zeroed in closed position. The pith of the galls were scored vertically with a knife and opened without damaging live larva inside. We examined the contents of galls and categorized the gall as containing dead Eurosta solidaginis if a living Eurosta larva was absent. We tried to ascertain fungal or
bacterial cause of death under the microscope. We categorized Eurosta as alive if we found and identified a living Eurosta larva. If the gall contained an inhabitant other than a Eurosta larva we categorized the Eurosta as dead. When possible, we identified the organism present as one of the parasitic species that invade Eurosta galls such as wasps Eurytoma gigantea and Eurytoma obtusiventris or Mordellistena. We recorded physical characteristics of the gall and its contents when pertinent. For example, after categorizing a gall as containing dead Eurosta, we noted that the gall was cracked and that a bird possibly caused the death. Data about a gall from a goldenrod gall moth was discarded as not pertinent to our experiment on goldenrod Eurosta galls. We included 173 galls in our data. We used the Shapiro-Wilk W test to test for normality in height by larval status and diameter by larval status. We used a test for non-parametric comparison, the Wilcoxon rank-sum test, to compare the median height of galls containing living Eurosta with the median height of galls containing dead Eurosta because the data for heights of galls containing live Eurosta was not normal (the data for galls containing dead Eurosta was normal). Similarly, we used the Wilcoxon test to compare the median diameter of galls containing live Eurosta with that of galls containing dead Eurosta because the data for diameters of galls containing live Eurosta was not normal (data for galls containing dead Eurosta was normal). The Wilcoxon test was an appropriate test to ascertain whether the differences between the median height and diameter of live larvae and the median height and diameter of dead larvae were significantly different. The Wilcoxon test ascertains the probability that the results of experimentation are due to random variation seen from sample size.
Results We examined the diameter and height of 173 goldenrod galls induced by Eurosta Solidaginis. The galls were divided into two samples which were comprised of 120 galls containing live Eurosta and 53 galls of dead Eurosta. Galls containing live Eurosta had greater diameter than galls of dead Eurosta. Gall height was similar across larval status. We observed galls containing live Eurosta (median=21.7 mm, upper IQR=1.8 mm, lower IQR=2.47 mm; n=120) had significantly greater median diameter than galls of dead Eurosta (median=19.1 mm, upper IQR=1.6 mm, lower IQR=2.3 mm; n=53, P <.0001; Figure 1). The slightly greater median height of galls of dead Eurosta (median=82.6 cm, upper IQR=17.0 cm, lower IQR=14.85 cm; n=53) was not significantly different from the median height of galls containing live Eurosta (median=80 cm, upper IQR=11.75 cm, lower IQR=8.15 cm; n=120, P=.474; Figure 2). Discussion Weis and Abrahamson (1986) showed that the mean size of selected individuals’ galls, those that survived and presumably reproduced, was significantly greater than the mean of the whole gene pool. E. gigantea prey on larvae after the galls have reached their full size, therefore small gall size is not a result of larva death terminating growth (Weis and Abrahamson, 1986). Their data indicated that the effect of upward pressure on gall size by E. gigantea outweighed that of the opposing pressure toward smaller gall size that birds exert. Their data also supported the presence of stabilizing selection acting to reduce variance. Similarly our data showed significantly greater median gall size for live Eurosta galls than for galls of dead Eurosta, confirming selection toward larger gall diameter. The distribution of diameters of live Eurosta galls appears to reflect stabilizing
selection but we did not perform addition statistical tests to identify the characteristics of the non-normal distributions. A dearth of literature on the effects of selection on Eurosta gall height, especially when compared to the preponderance of research on the effects of selection on gall size, suggests that selection does not affect gall height as strongly as it affects gall diameter. The absence of selection on gall height might result from an unfulfilled requirement for selection—that in order for natural selection to occur there must be genetically based variation in the trait (Purrington et al. Swarthmore College Bio 2 lecture). Gall height must be a heritable character of Eurosta for selection to act on it. Selection could act on height if females’ selection of goldenrods for oviposition, which is likely a heritable trait, affected gall height and if height influenced the survivorship of larvae. Horner and Abrahamson (1999) investigated goldenrod response to simulated drought stress and found that goldenrod genotype modulates and causes varying degrees of response. In some genotypes, goldenrods flowered significantly earlier under control conditions than under drought stress but others displayed a reverse trend, while some experienced no significant effects. Ovipositing females discerned between droughtsimulation and control plants, selecting more control plants. Females also responded significantly to genotype in their selection, but there was no significant effect of the interaction between treatment and genotype, despite the demonstrated effect of genotype on reaction to treatment. We found no significant difference between the median height of galls of dead Eurosta and the median height of live Eurosta galls. We propose that because the signals used by Eurosta in goldenrod selection do not incorporate the effects of the interaction of
genetics and the environment that selective pressures do not have a strong effect on gall height. Height must be heritable for selection to act and, though the ability to select more suitable goldenrods is likely a heritable trait, goldenrod genotypes interact with the environment in diverse ways, which most likely reduces the effect of selective pressures (if any) on goldenrod height. Literature Cited Abrahamson, W. G., Heinrich, P. The Solidago Eurosta Gall Homepage: A Resource for Teaching and Research. http://www.facstaff.bucknell.edu/abrahmsn/solidago/main.html>. [accessed 2005 January 18, Jan 31.] Horner, J. D., Abrahamson W. G. 1999. Influence of Plant Genotype and Early-Season Water Deficits on Oviposition Preference and Offspring Performance in Eurosta solidaginis (Diptera: Tephritidae). Am. Midl. Nat. 142 No. 1: 162-172. Weis, A. E., Abrahamson, W. G. 1985. Potential Selective Pressures by Parasitoids on a Plant-Herbivore Interaction. Ecology. 66 No. 4: 1261-1269. Weis, A. E., Abrahamson, W. G. 1986. Evolution of Host-Plant Manipulation by Gall Makers: Ecological and Genetic Factors in the Solidago-eurosta System. Am. Midl. Nat. 127 No. 5: 681-695. Anderson, S. S., McCrea, K. D., Abrahamson W. G., Hartzel, L. M. 1989. Host Genotype Choice by the Ball Gallmaker Eurosta Solidaginis (Diptera: Tephritidae). Ecology. 70 No. 4.: 1048-1054.