Developmental trap or demographic bonanza? Opposing consequences of earlier phenology in a changing climate for a multivoltine butterfly.

A rapidly changing climate has the potential to interfere with the timing of environmental cues that ectothermic organisms rely on to initiate and regulate life history events. Short-lived ectotherms that exhibit plasticity in their life history could increase the number of generations per year under warming climate. If many individuals successfully complete an additional generation, the population experiences an additional opportunity to grow, and a warming climate could lead to a demographic bonanza. However, these plastic responses could become maladaptive in temperate regions, where a warmer climate could trigger a developmental pathway that cannot be completed within the growing season, referred to as a developmental trap. Here we incorporated detailed demography into commonly used photothermal models to evaluate these demographic consequences of phenological shifts due to a warming climate on the formerly widespread, multivoltine butterfly (Pieris oleracea). Using species-specific temperature- and photoperiod-sensitive vital rates, we estimated the number of generations per year and population growth rate over the set of climate conditions experienced during the past 38 years. We predicted that populations in the southern portion of its range have added a fourth generation in recent years, resulting in higher annual population growth rates (demographic bonanzas). We predicted that populations in the Northeast United States have experienced developmental traps, where increases in the thermal window initially caused mortality of the final generation and reduced growth rates. These populations may recover if more growing degree days are added to the year. Our framework for incorporating detailed demography into commonly used photothermal models demonstrates the importance of using both demography and phenology to predict consequences of phenological shifts.

Specific glucosinolate analysis reveals variable levels of epimeric glucobarbarins, dietary precursors of 5-phenyloxazolidine-2-thiones, in watercress types with contrasting chromosome numbers.

Watercress obtained in food stores in the United States contained significant levels of epiglucobarbarin [(R)-2-hydroxy-2-phenylethylglucosinolate] and low levels of the 2S-epimer glucobarbarin identified by an HPLC+NMR+MS/MS approach. Typical combined levels were 4-7 µmol/g dry wt. The hydrolysis product, 5-phenyloxazolidine-2-thione (barbarin), was detected at similar levels as the precursor glucosinolates after autolysis of fresh watercress in water. Fragmentation patterns in MS(2) of reference desulfoglucosinolates were side chain specific and suitable for routine identification. Watercress was of two main glucosinolate chemotypes: Material from U.S. food stores had a complex profile including glucobarbarins, gluconasturtiin, indole glucosinolates and high levels (6-28 µmol/g dry wt.) of long-chain methylsulfinylalkyl and methylthioalkyl glucosinolates. Material from European food stores had a simple profile dominated by gluconasturtiin, with low levels of epiglucobarbarin and moderate levels of indole glucosinolates. Some wild U.S. material was similar to the U.S. food store type. Both types were found to be Nasturtium officinale by floral parts morphology. Cytological analysis of one U.S. food store accession indicated that it represented a chromosome-doubled variant within N. officinale. The nutritional consequences and invasive potential of the U.S. food store chemotype are discussed.

Leaf and floral parts feeding by orange tip butterfly larvae depends on larval position but not on glucosinolate profile or nitrogen level.

In an attempt to identify chemical signals governing the general flower and silique feeding behavior of larvae of the orange tip butterfly, Anthocharis cardamines (L.), we investigated feeding behavior and chemistry of two major host plants: Cardamine pratensis L. and Alliaria petiolata (Bieb.) Cavara & Grande (garlic mustard). Larvae reportedly feed mainly on flowers and siliques rather than leaves in nature, and did so when observed on the original host plants. Behavioral experiments, using detached A. petiolata branches, however, showed that larvae readily accepted leaves and only the final instar showed a tendency for directed movement towards floral parts. To search for semiochemicals that control plant part preference and to assess possible nutritional consequences of floral parts feeding, we determined glucosinolate profiles and total nitrogen levels of floral parts and leaves. There was only moderate difference between glucosinolate profiles of leaves and floral parts within each of two host plant species. In contrast, the profiles of floral parts differed significantly between them. A. petiolata was dominated by 2-propenyl glucosinolate, while C. pratensis was dominated by aromatic glucosinolates and branched aliphatic glucosinolates, with considerable variation among populations. Nitrogen levels tended to be higher in floral parts than in leaves in A. petiolata, but not in C. pratensis, so floral feeding could not generally be attributed to higher N content. With the exception of a tendency of last instar larvae (L5) to move to the apex and ingest flowers and upper stem, we did not find either a plant chemistry basis or larval acceptance/rejection behavior that could explain the usual feeding of floral parts by orange tip larvae of all instars. However, by artificial manipulation of vertical larval position on host plants, we found that the frequency of leaf vs. flower feeding during 24 hr depended significantly on the initial larval position. Hence, we suggest that the placement of eggs on floral parts by ovipositing female butterflies is a major explanation of orange tip feeding habits previously known from field observations.

Variable glucosinolate profiles of Cardamine pratensis (Brassicaceae) with equal chromosome numbers.

A novel glucosinolate, 3-(hydroxymethyl)pentylglucosinolate, was present at high levels in Cardamine pratensis L. from eastern North America and in commercially obtained seeds, but not in C. pratensis plants from southern Scandinavia. Glucosinolates in a number of accessions of C. pratensis included glucosinolates with the side chains 1-methylethyl, 1-(hydroxymethyl)ethyl, 1-methylpropyl, 1-(hydroxymethyl)propyl, 3-methylpentyl, 3-(hydroxymethyl)pentyl, benzyl, 4-hydroxybenzyl, 4-methoxybenzyl, indol-3-ylmethyl (as well as its 1-methoxy, 4-hydroxy, and 4-methoxy derivatives) and the rare side chain 1,4-dimethoxyindol-3-ylmethyl. Substantial variation was observed for four biosynthetic characters: (i) extent of chain elongation of Ile-derived glucosinolates; (ii) biosynthesis of Phe/Tyr-derived glucosinolates in general; (iii) hydroxylation of branched-chain glucosinolates; and (iv) O-methylation of 4-hydroxybenzylglucosinolate (sinalbin). Cytological analysis of pollen mother cells and root tip cells in meiosis and mitosis established the chromosome number to be 2n = 30 for all accessions, irrespective of glucosinolate profile.

Escaping an evolutionary trap: preference and performance of a native insect on an exotic invasive host.

Exotic plants may act as population sinks or evolutionary traps for native herbivores. The native butterfly Pieris oleracea lays eggs on garlic mustard, Alliaria petiolata, but larvae develop very poorly on this exotic invasive plant. We examined oviposition preference of individual females and larval performance of their offspring for individuals from one area where garlic mustard is well established and one where it is absent. These data were used to assess whether garlic mustard is being incorporated into or excluded from the diet. Females from the area without garlic mustard showed a wide range of preference, families had low larval survival on garlic mustard, and larval survivorship showed no correlation with mothers' preferences. Females from the area with garlic mustard preferred it to the native host, and larval survivorship on garlic mustard was positively correlated with the mother's preference. Individuals surviving on garlic mustard took longer to pupate and weighed >30% less compared to pupae reared on normal hosts. Our results suggest that where garlic mustard is well established P. oleracea may be adapting to this plant by both improved larval performance and increased adult female oviposition preference for it.

Modelling the impacts of two exotic invasive species on a native butterfly: top-down vs. bottom-up effects.

1. Exotic invasive species can influence population dynamics of native species through top-down or bottom-up forces. The present study examined separate and interactive effects of multiple exotic species invasions on the native mustard white butterfly, Pieris napi oleracea Harris (Lepidoptera: Pieridae), using a stochastic simulation model. 2. P. n. oleracea populations in North America have decreased regionally since the 1860s. Competition with an exotic congener (P. rapae L.), loss of native host plants and parasitism by the introduced broconid wasp (Cotesia glomerata L.), have been suggested to be independently responsible for its decline. The present study examined these hypotheses, as well as an alternative, invasion by an exotic crucifer, garlic mustard (Alliaria petiolata[Bieb.] Cavara & Grande). 3. A stochastic simulation model of P. n. oleracea population dynamics revealed that decreasing the number of host plants available for oviposition and larval development (i.e. habitat loss), sharply reduced the probability of populations persistence and decreased population size for those that persisted. 4. Simulated invasion by garlic mustard also substantially decreased both probability of persistence (= 0 at approximately 50% cover) and mean population size. Persistence probability never reached zero under any C. glomerata scenarios, even when larval mortality in the second generation due to parasitism was 100%. The impact of garlic mustard was intensified by the addition of C. glomerata parasitism. 5. Results suggest that bottom-up forces, loss of host plants through forest understorey loss and/or garlic mustard invasion are the most important forces driving P. n. oleracea population decline. Parasitism by C. glomerata may interact to reduce P. n. oleracea populations more rapidly, but appears insufficient alone to cause local extinction.

Foodplant preferences of Pieris caterpillars (Lepidoptera).

Larvae of five Nearctic Pieris butterflies accept a wide range of native and naturalized crucifers under laboratory test conditions. Preferences among crucifers are usually statistical rather than absolute. Caterpillars do not necessarily reject plants that do not support larval growth in favor of those that do. Preferences are not significantly altered by larval development or prior experience with specific foodplant species. Progeny of different females do not exhibit significantly different preferences. Retention of behavioral flexibility by Pieris may enable older larvae to exploit alternative resources not suitable for young larvae.

Population structure of pierid butterflies : I. Numbers and movements of some montane Colias species.

The numbers, dispersal behavior, aging and residence, and Wrightian neighborhood configurations of three species of Colias butterflies have been studied in central Colorado, using mark-release-recapture techniques as major tools. All populations studied have nonoverlapping generations and mature one brood each year. A brief general review of these species' autecology is given. A system for measuring degree of physical damage to the adults is introduced. This "wear rating" varies with temporal position of any given sample in the course of a brood's flight season, the insects becoming progressively more damaged with time. The sex ratio also changes with brood aging: males eclose before females, and are in the majority early in the flight season, while females may predominate at the end of flight. Local population numbers for the montane grassland species C. alexandra may reach peak levels of 700-900 insects in favorable years, but be much lower in other years as a result of, e.g., drought. Peak densities are no more than 2/ha. The montane bog species, C. scudderi, maintains comparable low density but has much smaller local populations. The subalpine/alpine grassland species C. meadii displays peak local numbers as high as 3000, with peak density as high as 120/hectare. Dispersal varies both among and within species. Those C. alexandra who disperse show an average dispersal radius of about 1.3 km, with a radius for the whole population of about 0.6 km; maximum distance moved was 8 km. Dispersal proportions among recaptures are sharply curtailed by adverse weather, but the dispersal radius of those moving is unaffected by weather. C. scudderi's dispersal is strongly influenced by the geometry of its bog and streamside habitats. Some C. meadii populations approach isolated "island" status, but others show much dispersal. Dispersal radius of those dispersing ranges from 0.3 to 0.7 km in different populations, but the proportion of dispersals varies greatly. The longest observed movement by this species is 1.3 km, although up to 2.6 km could have been detected. Colias normally display constant loss (death plus emigration) rates with average residence expectations of 4-6 days; few insects reach their maximum physiological lifespan of approximately 1 month. Bad weather can increase the loss rate drastically. Females show shorter residence than males, appearently as a result of greater mortality. Total-numbers-per-brood estimates are given for our better studied populations. The reproductive strategy of Colias is such that Wright's models for neighborhood size apply. Neighborhood size for C. alexandra varied sixfold in numbers, and from 3 to 1.3 km in physical extent, between a favorable year and a drought year. One localized C. scudderi habitat is only 200 m in diameter, but a streamside population has a neighborhood length of 4.8 km. In C. meadii, one population of 2000-2500 insects is an 8-ha "island", while another of similar numbers extends a single neighborhood across 1.9 km distance, 450 m altitude, and a major ecological boundary (timberline). Factors such as weather, individuals' visual cueing, and thermoregulatory behavior can influence population structure. For some Colias populations, selection may be very uniform within neighborhoods, while for others, single neighborhoods cross sharp discontinuities in selective forces. These patterns may differ for different selective forces, and may also vary with stages of the insects' life cycle. these populations will now prove a valuable resource for studying evolutionary population genetics.

Coevolution of pierid butterflies and their cruciferous foodplants : I. The relative quality of available resources.

Two Colorado populations of Pieris butterflies show a spectrum of larval growth responses to potential foodplant crucifer species growing in montane habitats. Analysis of larval growth responses to this array suggests potential selection for differential utilization of these species: 1) available crucifers vary considerably in the rates of larval survival and growth they support; 2) climatological factors favor larvae which develop rapidly. Food-related larval mortality and climatological factors may provide selection for utilization of crucifer species on which larvae develop rapidly.