Resource-based trade-offs and the adaptive significance of seasonal plasticity in butterfly wing melanism

Abstract

Phenotypic plasticity is the ability of an organism to alter its phenotype in response to environmental cues. This can be adaptive if the cues are reliable predictors of impending conditions and the alterations enhance the organism's ability to capitalize on those conditions. However, since traits do not exist in isolation but as part of larger interdependent systems of traits (phenotypic integration), trade-offs between correlated plastic traits can make phenotypic plasticity non- or maladaptive. We examine this problem in the seasonally plastic wing melanism of a pierid (Order Lepidoptera, Family Pieridae) butterfly, Pieris rapae L. Several wing pattern traits are more melanized in colder than in warmer seasons, resulting in effective thermoregulation through solar absorption. However, other wing pattern traits, the spots, are less melanized during colder seasons than in warmer seasons. Although spot plasticity may be adaptive, reduced melanism of these spots could also be explained by resource-based trade-offs. Theory predicts that traits involved in resource-based trade-offs will be positively correlated when variation among individuals in resource acquisition is greater than variation among individuals in resource allocation strategies, and negatively correlated when variation in allocation is greater than variation in acquisition. Using data from both field studies and laboratory studies that manipulate dietary tyrosine, a melanin precursor, we show that when allocation to thermoregulatory melanism (ventral hindwing, and basal dorsal fore- and hindwing "shading") varies substantially this trait is negatively correlated with spot melanism. However, when there is less variation in allocation to thermoregulatory melanism we find these traits to be positively correlated; these findings are consistent with the resource-based trade-off hypothesis, which may provide a non- or maladaptive hypothesis to explain spot plasticity. We also show that increased dietary tyrosine results in increased spot melanism under some conditions, supporting the more general idea that melanism may involve resource-based costs.

Keywords: butterflies; melanism; phenotypic integration; phenotypic plasticity; seasonal polyphenism; trade‐offs; tyrosine.

FIGURE 1

Wing pattern variation in the cabbage white butterfly (Pieris rapae L.). In cool‐season butterflies (a top) the dorsal forewing tip patch and dorsal and ventral wing spots (denoted in green in c, bottom) are small, whereas the basal portions of the dorsal fore‐ and hindwing and ventral hindwing surfaces are more heavily melanized with a diffuse “shading” of black scales (denoted in blue in c, bottom). In warm‐season butterflies (b, middle) wingtip patches and spots are large but the dorsal basal and ventral hindwing regions are less melanised. All melanised traits were quantified as the area falling above a threshold of 150, on a 0–255 (white to black) grayscale. Wingtip patches, spot (“spots” in this article) sizes, and dorsal hindwing melanism were quantified in their entirety, whereas basal dorsal forewing and ventral hindwing melanism were quantified as the area falling above the threshold in the polygon defined by wing‐vein landmarks as shown. The total area of the ventral hindwing polygon served as a proxy for overall wing size. See text for details.

FIGURE 2

Resource‐based trade‐offs may produce within‐group correlations between competing traits that differ in sign from across‐group correlations. If groups differ considerably in resource acquisition the across groups correlation may be positive despite trade‐off‐induced negative correlations within groups (inset). However, if differences among groups in allocation to traits are large the across‐groups correlation may be negative despite positive within‐group correlations due to variation in resource acquisition (main body). In this figure, darker points represent individuals that have acquired more resources.

FIGURE 3

Spots melanism (spot size) as a function of “shading” (i.e. thermoregulation) melanism and seasonal periods. Across the entire flight season, spots melanism declines as shading melanism increases for both males (a) and females (b). However, within a given shorter (~30 days) seasonal period the relationship between spots and shading melanism tends not to decline and in several cases may be positive; this is true for both males (c) and females (d). In (a) and (b), the points are partial residuals controlling for wing size, with values standardized to mean 0, standard deviation 1.0 prior to analysis (see text for details); gray bands around the regression lines indicate 95% confidence intervals. In (c) and (d), points have been eliminated for clarity but within‐period regression lines are shown, along with the 95% confidence intervals. Points (a, b) and lines with bands (c, d) are colored according to the seasonal period to which they correspond, with darker colors being the warmer periods.

FIGURE 4

The sign of the relationship between spot size and shading melanism depends upon the length of the seasonal time period under consideration. Slopes (dependent variable) are the grand slopes from linear models modeling spot size as a function of shading that include as predictors seasonal periods from 15 up to 246 days. At short time periods, spots tend to be a positive function of shading melanism (points above the horizontal dashed lines), whereas at longer periods spots tend to be a negative function of shading melanism.

FIGURE 5

Spots melanism as a function of “shading” (i.e. thermoregulation) melanism, dietary tyrosine content (HT = high tyrosine, LT = low tyrosine), and rearing temperature‐daylength. In (a) and (b), the relationship is shown within each dietary tyrosine treatment but across the rearing‐temperature‐daylength treatments. In (c) and (d), the relationship is shown within each rearing‐temperature‐daylength treatment, but across the dietary tyrosine treatments. Bands depict the 95% confidence intervals. The black dashed line is the regression line across all butterflies without consideration for experimental treatments. Points are partial residuals controlling for wing size and were also standardized to mean 0, standard deviation 1.0, prior to analysis (See text for details).