Genetic constraints on macroevolution: the evolution of host affiliation in the leaf beetle genus Ophraella.
Contemporary views on the factors that determine the course of adaptive evolution diverge with respect to the importance of genetic constraints. Although adherents to each view would admit some validity in the other, internal or historical constraints (Maynard Smith et al. 1985) are considered paramountly important by many paleontologists, morphologists, and adherents to one or another "structuralist" view (e.g., Goodwin 1984; Alberch and Gale 1985; Wake and Larson 1987; Gould 1989; Wake 1991), whereas many population biologists and behavioral ecologists, drawing on revelations of genetic variation in almost every character ever examined, are inclined to affirm the sufficiency of genetic variation for responses to selection on most characters (e.g., Charlesworth et al. 1982; Barker and Thomas 1987; Weber and Diggins 1990). However, even the most optimistic optimization theorist admits and invokes constraints in the form of allocation trade-offs: negative genetic correlations are routinely invoked to explain everything from balanced polymorphism to ecological specialization (Via and Lande 1985; Futuyma and Moreno 1988). Moreover, many would agree with J. B. S. Haldane that humans will never evolve into a race of angels for lack of genetic variation in wings and the moral sense, for all recognize that selection can act only on variation in characters that exist, namely those peculiar to each lineage, bequeathed to it by its evolutionary history.
Constraints or biases in the direction of evolution are clear in trivial cases--we understand that vertebrates lack the developmental foundations for growing wings from the shoulders. In some instances, experimental developmental biology can explain lineage-specific trends (e.g., Alberch and Gale 1985). But the developmental or physiological foundations of phenotypes are usually a black box, and only by direct empirical investigation can we assess the likelihood that the course of evolution may be guided by patterns of genetic and phenotypic variation, which are, in part, functions of the history of selection and drift. In at least a few cases, limits on genetic variation seem to play a role. Threshold characters may present phenotypic uniformity despite underlying genetic variation (Kindred 1967; Rendel 1967; Scharloo 1987); some characters do not respond to artificial selection (Maynard Smith and Sondhi 1960); tolerance of heavy metals has failed to evolve precisely in species of grasses that appear to lack genetic variation in tolerance (Bradshaw 1991).
Studies of genetic variation, and of evolutionary mechanisms generally, have until recently proceeded with little reference to paleontological and systematic data on evolutionary history, which present a wealth of examples and phenomena that might repay experimental study (Futuyma 1988, 1992a). Among these are conservative characters, those that have evolved little over vast spans of time. Host associations of herbivorous insects frequently fit this pattern. In many lineages of insects, diverse species feed on closely related plants and host affiliation has changed little since the Oligocene or earlier (Ehrlich and Raven 1964; Mitter and Farrell 1991; Mitter et al. 1991). Such patterns have long been attributed (Borner 1939; Dethier 1954; Ehrlich and Raven 1964; Futuyma and Keese 1992) to the insects' responses to the similarity of related plants, especially in secondary compounds. These chemicals act at both the behavioral and deeper physiological levels: some are toxic to nonadapted insects, some have repellent effects, and some have been adopted as feeding stimuli by host-specific insects (Rosenthal and Berenbaum 1992). But for secondary compounds to restrict insect adaptation for most of the Cenozoic seems, on the face of it, surprising: we tend to think of behavior as the epitome of a phenotypically and genetically malleable trait, and hundreds of species of insects have adapted in the last few decades to toxins unprecedented in evolutionary history, namely, chemical insecticides (Georghiou and Mellon 1983). Moreover, not uncommonly, one or a few species in an otherwise highly host-specific taxon are polyphagous, such as the swallowtail butterfly genus Papilio (Feeny 1992) and the aphid genera Myzus and Aulacorthum (Eastop 1977). Such evidence that plant barriers are not insurmountable makes it all the more puzzling that most insect species have not overcome them. In the absence, then, of detailed information on detoxification mechanisms, chemoreceptors, and central-nervous-system integration of sensory input, we cannot say whether host associations are conservative because the structural foundations for genetic variation are lacking, or because stabilizing selection has maintained the associations (despite enormous spatial and temporal variation in the availability of host plants!).
The suspicion that host affiliation might be constrained by the availability of genetic variation leads us to ask if, when host shifts have occurred, they have been guided not only by ecological circumstances, but also by genetic constraints. Was the realized path of evolution more likely, in strictly genetic terms, than the paths not taken? Can studies of genetic variation serve to predict or explain the course of evolution?
OPHRAELLA, A MODEL SYSTEM
We address these questions in this, one of several related reports (Futuyma and McCafferty 1990; Futuyma et al. 1993b, 1994; Funk et al. 1995a) in which we have sought to relate genetic variation to an estimate of the evolutionary history of host affiliation in the North American leaf beetle genus Ophraella Wilcox (Chrysomelidae: Galerucinae). Each of the 14 known species in this monophyletic group (Futuyma and McCafferty 1990) feeds, as both larvae and adults, on the foliage of one or a few congeneric species of composites (Asteraceae) or, at most, on several related genera (Futuyma 1990). Futuyma and McCafferty (1990) used a cladistic analysis of morphological and electrophoretic data to estimate phylogenetic relationships within the genus and thereby to arrive at a parsimonious interpretation of the history of host affiliation. In a companion paper to this one, Funk et al. (1995a) used sequence data from two mitochondrial genes to infer a fully resolved phylogeny that conforms in the major aspects of phylogenetic structure and host affiliation to that suggested by Futuyma and McCafferty, but that provides a different and more strongly substantiated estimate of relationships among several closely related species. (For analyses of sequence evolution and congruence among data sets, also see Funk et al. [1995b].) The phylogeny of Ophraella is in part incongruent with that of the host plants, which fall into four tribes of Asteraceae; thus, changes in host affiliation are partly or perhaps entirely attributable to exclusive shifts between hosts, rather than to cospeciation (Funk et al. 1995a). Nevertheless, the inferred history of host shifts shows considerable correspondence with both the phylogeny of the hosts and their phenetic similarity of profiles of secondary compounds (Futuyma and McCafferty 1990; Funk et al. 1995a).
In the experimental program, we employed quantitative-genetic methods to screen Ophraella species for genetic variation in feeding responses to and larval survival on plants that they themselves do not naturally utilize but are host plants of their congeners. If realized host shifts, inferred from the phylogeny, have been, for reasons of the availability of genetic variation, more likely than imaginable shifts that have not been realized in evolution, then we expect to find genetic variation more frequently in responses to the host of a close relative, such as a sister species, than to the hosts of distant relatives. We might also expect asymmetries; for example, species with derived host associations might retain more variation for response to a recently ancestral host plant than species with ancestral host associations display in response to the hosts of their "apomorphic" relatives.
In this report, we describe results with O. notulata, and summarize the patterns found to date in this and three other species.
Ophraella notulata (Fabricius), referred to as O. integra (LeConte) before LeSage's (1986) revision of the genus, is almost exclusively limited to salt marshes along the Atlantic and Gulf coasts of the United States, where it feeds on the shrub Iva frutescens (tribe Heliantheae, subtribe Ambrosiinae). It has been found on I. annua in Louisiana (Futuyma 1990), and a few inland records (LeSage 1986) suggest that it may feed on this or other Iva species elsewhere, although these records might represent its sibling (and probable sister) species O. slobodkini. The latter, found in Florida and southernmost Georgia (Futuyma 1991), is only known to feed on Ambrosia artemisiifolia (also Ambrosiinae), on which we have never found O. notulata, even where the plants grow in close proximity. The phylogenetic distribution of host associations (fig. 1), and the continent-wide distribution of Ambrosiinae-associated species relative to the peripheral distribution of O. notulata, suggest that the species' association with Iva frutescens is a derived condition, probably from association with Ambrosia (Funk et al. 1995a).
We report evidence on genetic variation in responses of O. notulata to 1. frutescens and six other plants. Three (Solidago bicolor and S. altissima in tribe Astereae, Eupatorium perfoliatum in tribe Eupatorieae) are hosts of species that are relatively distantly related to O. notulata (fig. 1). Three (Chrysopsis villosa, tribe Astereae, Artemisia vulgaris, tribe Anthemideae, and Ambrosia artemisiifolia, subtribe Ambrosiinae) are, or are closely related to, hosts of species in the slobodkini clade that includes O. notulata. Artemisia vulgaris is a Eurasian species, naturalized in eastern North America, on which a midwestern Artemisia-feeding species, O. artemisiae, can be reared rather well (Futuyma et al. 1994). The geographic distribution of neither Chrysopsis nor the natural hosts of O. artemisiae approaches the coastal distribution of O. notutata. The other plants are geographically proximate to the coast; of these, S. bicolor, distributed in open upland forests, is perhaps least likely to be encountered by O. notulata.
A necessary although not sufficient condition for a plant to be a host of Ophraella is that it elicit oviposition by females E and feeding by both larvae and adults (females do not lay i eggs unless they have fed). The plant must also support larval growth and survival. We report whether we found genetic variation in these traits. We do not present heritabilities because they are inaccurate predictors of even short-term responses to selection (Barton and Turelli 1989) and their values are probably less relevant to macroevolutionary questions.
Methods and Results
Ophraella notulata is multivoltine, with an egg-to-egg generation time of about one month at 25[degrees]C. In all experiments, insects were reared in an environmental chamber at 25[degrees], circa 50% RH, 16:8 L:D photoperiod, either on cuttings of field-collected Iva frutescens or individually in plastic petri dishes, lined with moist filter paper, placed in closed, clear-plastic boxes, and provided with fresh leaf material every 2 d. In feeding trials, we thoroughly mixed leaf discs from at least two individual plants, distributed them to the test insects, and scored the leaf area consumed by counting squares in an ocular grid in a dissecting microscope set at 12X (100 units = 12.96 mm2). All such data are presented in these units, and were log (+1)-transformed for analysis.
We report three major experiments. Experiment 1 used progeny of wild females to screen for variation in larval feeding response to Solidago altissima, Eupatorium, and Ambrosia, and provided preliminary data on adult feeding response to the latter two plants. Experiment 2 used a half-sib breeding design to measure variation in larval feeding, adult feeding, and larval survival on S. bicolor, Eupatorium, Artemisia, Chrysopsis, and the natural host l. frutescens. Experiment 3 used a half-sib design to measure variation in oviposition on Ambrosia and 1. frutescens. We also report ancillary experiments that extend our data on responses to three of these plant species.
This experiment tested for variation in feeding responses to three plant species. Three larvae from each of 100 females captured in July 1990 at Smith Point, Long Island, N.Y., were placed, within 24 h after hatching, individually in dishes with two discs of Ambrosia artemisiifolia, and three others from each brood were individually provided with (simultaneously) one disc of Eupatorium perfoliatum and one of Solidago altissima. We scored feeding at 24 h and again (after replacing the discs) at 72 h. Survivors of the Ambrosia feeding test, together with others of the same brood, were reared in groups of seven on cuttings of Iva; progeny of each female were reared together, so variation among dams is confounded with rearing conditions. Upon eclosion, adults were deprived of food for 24 h. Beetles were then provided with Eupatorium leaf discs and feeding was scored after 48 h. These discs were then replaced by Ambrosia discs, and feeding was again scored after another 48 h. Because the wild females may have mated repeatedly, each brood consisted of an unknown mixture of full and half sibs.
In the hatchling feeding trial, no larvae ate S. altissima (n = 296). The mean (standard error, sample size) consumption of Ambrosia was 24.4 grid units (0.93, 289) after 24 h (i.e., a mean of 3.16 [mm.sub.2]) and 127.2 (3.20, 288) after 72 h (data not transformed); for Eupatorium, the corresponding values were 1.1 (0.15, 296) and 5.2 (0.66, 295). Both scores on both Ambrosia and Eupatorium yielded significant variation among families: ANOVAs of [log.sub.10]-transformed 24-h consumption yield F = 2.81; df = 97, 191; (P < 0.0001) for Ambrosia and F = 1.70; df = 99, 196; (P < 0.0001) for Eupatorium. After rearing on Iva, adults from these broods consumed 228.8 (10.47, 224) and 17.3 (2.48, 276) units of Ambrosia and Eupatorium, respectively (48-h scores). The variance among families (log-transformed data) was significant in both cases: F = 1.70; df = 59, 164; (P = 0.006) for Ambrosia; F = 2.08; df = 62, 213; (P < 0.0001) for Eupatorium. The adult data are not definitive evidence for genetic variation because sibs had been reared together. (In another experiment, with a Florida population of this species, significant variance in adult consumption of Ambrosia was found among families, each reared in two containers; Keese 1994.) The larval data, however, provide fairly strong evidence for genetic variation, because an experiment in which parents were reared as larvae on two plant species (I. frutescens and A. artemisiifolia) yielded no evidence of a direct effect of maternal host on larval progenies' consumption of these plants (Futuyma et al. 1993a). Maternal effects owing to variation in the natural host alone (Iva) might be expected to be even less pronounced.
This experiment was designed to assay for genetic variation in feeding responses to and survival on four congeners' hosts and the natural host of O. notulata. We obtained eggs from 27 females collected in Oldfield, N.Y. on June 11, 1992, reared the larvae in groups of six on Iva cuttings, and isolated the adults by sex as they emerged. Using a half-sib design (Falconer 1981), we paired two females with each male for 4 d; no sibs were included in any such trio. We obtained eggs from 29 pairs of females, removed them from the foliage onto filter paper, and, as they hatched (over the course of 1 wk), placed 40 hatchlings from each dam individually in dishes, assigning eight larvae to each of five plants: I. frutescens, S. bicolor (both from the field), Artemisia vulgaris, Eupatorium perfoliatum, and Chrysopsis villosa (from greenhouse-grown plants). We scored consumption of two leaf discs after 48 h, and then fed each larva on its test plant until death or pupation occurred. When adults eclosed (from the Iva treatment only; all others died), we scored up to four beetles from each brood (fewer in some instances, because of mortality) for 24-h consumption of Iva, and then scored each, sequentially, for 48-h consumption of Chrysopsis, Eupatorium, and Artemisia, and for 24-h consumption of S. bicolor. The beetles were provided with Iva for 1 d before rotation onto each of these plants, in the hope of reducing mortality and equalizing motivation to feed. ANOVAs were performed with the PROC GLM procedure of SAS, using Type III sums of squares. Consumption scores were [log.sub.10]-transformed in order to better approximate normality. Pooling of mean squares in F tests follows Sokal and Rohlf (1981, p. 285).
Both larvae and adults ate far more Iva, their natural host, than any other test plant, with Chrysopsis ranking a distant second (table 1). (We believe the correspondence between presentation order and mean consumption is coincidental. The variances, in any case, show no such correspondence.) The full ANOVA of larval consumption (table 2A) showed significant main effects of sire, dam, and host, and a significant dam-by-host interaction. Analyzing consumption of each host separately (table 2B-F) revealed significant sire effects that imply the existence of additive genetic variance in feeding response to Eupatorium; because our experiment adopts as its null hypothesis the existence of genetic variation in all traits, it is conservative to interpret the marginally significant sire effect on Solidago also as evidence for genetic variation. Significant dam effects on consumption of Chrysopsis, Artemisia, and possibly Eupatorium are likely to reflect genetic variance, probably nonadditive in nature (Falconer 1981).
[TABULAR DATA 1 and 2 OMITTED]
Because the adult progeny were scored after unequal times of exposure to different plants, we present only the separate analysis for each plant (table 3). Significant sire effects are
Table 3. ANOVA of area ([log.sub.10]-transformed) consumed by adult Ophraella notulata. ConsUmption measured over 24 h for Iva and Solidago, over 48 h for others. Notation as for table 2. evident on Chrysopsis and Artemisia; no significant dam effects were obtained. The presence of significant sire but not dam effects is surprising and could be due to rare alleles of major effect; however, examination of sire means revealed no notable outliers (except on Iva, on which three sires had unusually low means.) Source df MS d F P A. Iva frutescens S 28 0.1735 D + E 1.41 >0.05 D(S) 27 0.1032 E 0.8 10.73 E 137 0.1270 B. Chrysopsis villosa S 28 2.2469 D + E 2.91 <0.001 D(S) 27 0.7710 E 1.00 0.47 E 130 0.7703 C. Eupatorium perfoliatum S 28 0.3226 D + E 0.83 >0.50 D(S) 27 0.4524 E 1.20 0.25 E 117 0.3766 D. Artemisia vulgaris S 28 0.7716 D 2.53 <0.025 D(S) 25 0.3054 E 1.14 0.32 E 84 0.2685 E. Solidago bicolor S 26 0.3235 D 0.88 >0.50 D(S) 24 0.3684 E 1.19 0.28 E 68 0.3100
When different characters (such as feeding responses to different plants) are scored on different individuals in each family, correlations of family means are the only approximations to genetic correlations to which significance tests can currently be applied (Via 1984). For larval consumption, three of ten pairwise product-moment correlations of sire means are statistically significant; two of these provide significant Spearman rank correlations (table 4). Three of ten correlations of adult consumption, by sire, dam, or both, are significant. All significant correlations are positive except for the correlation between adult consumption of Iva and Artemisia, which is negative (and for which the rank correlation, - 0.319, was not significant).
[TABULAR DATA 4 OMITTED]
Of the larvae initially placed on Iva, 72.9% survived to pupation; none survived to pupation on any other plant. A few individuals survived more than 4 d on Eupatorium (5 of 436), Artemisia (6 of 452), and Chrysopsis (32 of 444); some individuals on the latter plant survived as long as 12 d, feeding throughout and passing into the second instar, but growing very slowly. This observation suggests but does not prove that properties of this plant not only deter feeding, but also affect metabolism, since death occurred despite feeding. Using the dam mean square to test for variance among sires in the (arcsine-transformed) proportion of larvae surviving to pupation on Iva, we found no significant effect (F = 4.20; df = 10, 2; P = 0.21).
Virgin adult progeny from 48 females collected in Mt. Sinai, N.Y., were paired in a half-sib design (one male and two females). Progeny were reared individually in dishes on Iva. Upon eclosion, two females from each brood were provided with Iva and two with Ambrosia artemisiifolia, and each was given a male (almost no matings were between sibs). We replaced leaf material and counted eggs every 2 d through 16 d or death of the female. Few eggs were found before day 6, and none before day 4. We analyzed the (square root-transformed) total number of eggs/census + 0.5, where "census" is the number of egg counts made for each female (including the count made when a female was found dead).
Of 95 females placed on Iva, 100% laid eggs, and of 91 on Ambrosia, 69.2% laid eggs [Mathematical Expression Omitted]. The temporal distribution of first clutches on Ambrosia lagged (by 0.72 d) behind that on Iva [Mathematical Expression Omitted], but the distribution of female survival did not differ [Mathematical Expression Omitted]. The mean number of eggs per census was lower on Ambrosia [Mathematical Expression Omitted] than on Iva (x= 23.61, SE = 1.166, n = 95). Analyses of variance (table 5) show a significant effect of host plant, but not of dam.
TABLE 5. ANOVAs of oviposition on Iva frutescens and Ambrosia artemisiifolia by Ophraella notulata. The variate is the square root of (total eggs/census + 0.5), counted every 2 d through 16 d or death of the female. A. Full analysis. B. Data for /va. C. Data for Ambrosia. Notation as for table 2. A. Full analysis S 29 2.5270 D 1.52 >0.10 D(S) 23 1.6595 E 0.52 >0.50 H 1 197.2062 S x H 76.61 <0.001 S x H 29 2.5741 (D x H) + E 1.37 >0.10 D x H(S) 22 2.4538 E 1.43 >0.10 E 8 11.7178 B. Iva frutescens S 29 1.3851 D 0.64 >0.75 D(S) 23 2.1670 E 1.33 0.21 E 42 1.6343 C. Ambrosia artemisiifolia(*) S 29 3.2560 D 1.61 >0.10 D 22 2.0223 E 1.12 0.37 E 39 1.8078
The sire term is not significant for Iva, and is significant for Ambrosia only if tested over the pooled mean square, which is questionably valid in this case (Sokal and Rohlf 1981, p. 285). However, the magnitude of the sire mean square on Ambrosia suggests that a larger sample might well yield evidence of genetic variation. Product-moment (rp) and Spearman rank ([r.sub.s]) correlations of sire means (rp = 0.237, n = 30, P = 0.208; [r.sub.s] = 0.170, P = 0.368) and of dam means ([r.sub.p] = - 0.090, n = 52, P = 0.525; [r.sub.s] = - 0.078, P = 0.585) were not significant, as expected from the ANOVAs.
Some progeny of the wild females that initiated this experiment were used to assess the likelihood of oviposition on Chrysopsis. We placed 49 newly eclosed female progeny, from 31 broods, individually in dishes with Chrysopsis foliage, and 62 others on Ambrosia for comparison. A newly eclosed male was placed in each dish. Foliage was changed and eggs were counted at 3-d intervals for 9 d. On Ambrosia, 55% of females surviving to day 6 had oviposited, and 83% of the 40 survivors to day 9 had done so. Despite evidence of feeding in almost all cases, no females laid eggs on Chrysopsis (survivorship was 71% to day 6 and 43% to day 9).
(A) Because the behavior of free-ranging larvae might differ from those confined in small dishes, we placed 120 hatchling larvae (5 from each of 24 wild females) in a cage (constructed of plexiglass and larva-proof mesh) with 10 large Eupatorium perfoliatum plants. After 7 d, a few leaves displayed grazing damage, but no live larvae were recovered. (B) In preliminary tests that add to our data on responses to Solidago altissima, we placed 190 larvae, representing 33 broods from wild-caught females, on cuttings of this plant, and examined them 11 d later. No larvae had survived, and there was no evidence of feeding. We provided 150 wild caught adults individually with discs of S. altissima. None fed within 24 h. (C) Larval survival was compared on Ambrosia artemisiifolia and Iva frutescens. From each of 19 wild-caught females, four hatchlings were placed in each of two dishes with Ambrosia and one dish with Iva, foliage was replaced every 3 d, and survival and development were monitored for 25 d. The proportion surviving on Ambrosia varied significantly among families (F = 2.52; df = 18, 19; P = 0.026) at 25 d, although not earlier. Mortality was higher ([chi square] = 41.43, P < 0.001) on Ambrosia (91/148) than on Iva (12/ 75), a lower proportion reached pupation on Ambrosia (0.52) than on Iva (0.87) within the period of the experiment ([chi square] = 25.82, P < 0.001), and those entering pupation on Ambrosia did so with a mean lag of 3.56 d relative to those on Iva ([t.sub.140] = 7.593, P < 0.001).
We first interpret briefly the results on Ophraella notulata, and then turn to a synthesis and interpretation of data from the four species of Ophraella that we have studied. The principal aim of this research is to determine whether responses (feeding, survival, oviposition) to host plants of each species' congeners are uniformly genetically variable. Having found evidence of genetic variation in some instances but not in others, we wish to determine whether the pattern of genetic variation bears any relation to the phylogenetic history of host affiliation that we have inferred for the genus (Funk et al. 1995a). That is, do the data of population biology help to explain a macroevolutionary history?
We must first address two major limitations of these studies. First, if any of these characters are threshold traits (Falconer 1981), there may exist genetic variation that is not expressed, but which nonetheless provides potential capacity for evolution. Throughout this paper, "genetic variation" should be read as "expressed genetic variation." Second, we clearly cannot, in any instance, definitively demonstrate that genetic variation is absent, for larger samples or samples from other populations might reveal variation, especially if variants are rare. Moreover, even if rare alleles are captured in samples used for breeding, they may not be detected by the statistics of quantitative genetics unless the environmental variance is negligible. Selection experiments may provide a more powerful approach but are impractical for screening numerous traits for genetic variation in each of several species. We note, however, that even a very small sample should, on average, include most ([n-1]/n for a sample of n genes) of the additive genetic variance in the source population (Lewontin 1965; Nei et al. 1975). In all the populations used in our experiments, we have detected genetic variation in responses to some plants, and most of the populations have been examined by electrophoresis and found to be highly heterozygous (Futuyma and McCafferty 1990), providing evidence that they are not highly inbred. Nonetheless, we recognize that however adequately our experiments may estimate present levels of additive genetic variance, they cannot characterize the evolutionary potential provided by rare alleles. On the other hand, if we take our estimates of genetic variation to imply a capacity to adapt to new hosts, our methodology almost surely overestimates this capacity, at least for behavioral traits. An insect confined with a novel plant in a small dish is more likely to feed than is a free-ranging animal, the common response of which is to disperse in search of its natural food. We observed this response in three species when we placed large numbers of larvae in small "gardens" of a test plant (Futuyma et al. 1993b, 1994; this paper, Ancillary Experiment A). In two of these cases, we surrounded the "garden" with a sticky barrier and recovered many larvae from it. Most such individuals would probably perish in nature.
Interpretation of Results from Ophraella notulata
Among the six plants used in these experiments (other than the natural host, Iva), Ambrosia artemisiifolia is closely related to Iva, and has the same major classes of secondary compounds (Seaman 1982; Futuyma and McCafferty 1990). Compared to these Ambrosiinae, Solidago altissima, S. bicolor, and Chrysopsis villosa, in the Astereae, chemically differ most, notably in lacking sesquiterpene lactones, whereas Eupatorium and especially Artemisia are chemically more similar. Ambrosia is the host of Ophraella slobodkini, to which O. notulata is most closely related (Futuyma and McCafferty 1990; Futuyma 1991; Funk et al. 1995a). Although these species can form viable (but nearly infertile) hybrids (Keese 1994), we tentatively estimate their divergence to have occurred 5-9 mya (Funk et al. 1995a).
The mean responses of O. notulata to its congeners' hosts bear some relation to their classification and perhaps to their chemistry. Only Ambrosia supported larval survival to pupation; Ambrosia elicited oviposition but Chrysopsis did not; consumption of Ambrosia far exceeded that of the other plants, with the Solidago species ranking lowest (table 1 and Ancillary Experiments B, C). The relatively high consumption of Chrysopsis was not expected; although we have found no reports on the chemistry of this plant, it belongs to the Astereae, which generally differ greatly from the Ambrosiinae, as noted above. It may or may not be significant that another Ambrosiinae-associated species, O. communa, has apparently given rise to the Chrysopsis-associated species O. bilineata (Funk et al. 1995a).
Hatchling consumption of congeners' hosts provided evidence for genetic variation in all cases except Solidago altissima, the plant that elicited the least feeding. Adult consumption showed genetic variation only on Ambrosia, Artemisia, and Chrysopsis. Survival to the late larval or to the pupal stage provided evidence of genetic variation only on Ambrosia (mortality on all the other plants was complete). Thus, genetically variable behavioral responses in both life stages were found for only three plants, and genetically variable survival on only one of these. The data therefore provide prima facie evidence of limitations on genetic variation that could constrain the evolution of diet, or at least bias its direction. Moreover, the only plant on which all traits (perhaps including oviposition) were genetically variable was Ambrosia, which is most closely related and phenetically similar to the natural host, and is probably the plant from which this insect's present association with Iva was derived. We cannot say whether the pronounced, genetically variable responses of O. notulata to this plant are better explained by the history of the insect lineage or by the similarity of Ambrosia to Iva. A similar pattern was displayed by a relatively distantly related member of the same clade, O. communa (figure 1), which feeds on Ambrosia and on Iva axillaris: it showed more genetically variable responses, with higher means, to Iva frutescens than to any other test plant (Futuyma et al. 1993b).
In insects that oviposit on the host plant and have relatively immobile larvae, the variety of plants acceptable to the larva is often broader than that on which females will oviposit (Wiklund 1975). Our data conform to this pattern (viz., oviposition on Ambrosia but not Chrysopsis). Moreover, genetic variation in larval consumption extended to more plants than did adult consumption (the first "hurdle" to be passed in colonization of a novel plant, if this is accomplished by dispersing adults). Fewer plants still (viz., Ambrosia) supported genetically variable larval survival. In view of models in which avoidance of, rather than adaptation to, a novel plant evolves if behavior is more genetically variable than performance (Futuyma 1983; Castillo-Chavez et al. 1988; Rausher 1993), expansion of diet to include most of these plants would appear unlikely.
Most of the estimates of genetic correlation in feeding responses (table 4), as well as in oviposition response to Ambrosia vs. Iva, did not differ significantly from zero. When significant pair wise correlations were observed, the plants involved (e.g., Iva, Chrysopsis) failed to maintain significant correlations in other combinations (e.g., Iva, Eupatorium), even when the responses to both plants were genetically variable. This suggests independence of feeding responses to the several plants, that is, that the genetic variation observed is not merely in a generalized trait such as "motivation to feed" or "consumption rate" that might be attributable to body size or "vigor." Only one instance of a negative genetic correlation in feeding response was observed. The incidence of significant genetic correlations in table 5 shows no clear correspondence with either the relationships among the plants or their overall similarity in secondary compounds.
Patterns of Genetic Variation in the Genus Ophraella
We now summarize and interpret our data on Ophraella communa (Futuyma et al. 1993b), O. conferta, O. artemisiae (both in Futuyma et al. 1994), and O. notulata (this paper). Each was screened for genetic variation, using methods similar to those described for O. notulata. For none of the traits examined (neonate larval consumption, adult consumption, larval survival) is the beetle-by-plant matrix complete, because of the necessary trade-offs in allocation of effort. Our purpose here is to determine whether the pattern of genetic variation is related to the insects' phylogenetic history, to plant relationships, or both, in order to infer whether or not the evolution of host affiliation may have been affected by genetic constraints.
In the extreme (and implausible) case in which the history of host shifts has been determined entirely by the availability of genetic variation, we might expect to find genetically variable responses only to plants that represent phylogenetically local, recent host shifts (i.e., in the response of a species to its most recent ancestral host, or to the apomorphic host of a closely related species). This expectation assumes that host shifts have not been mediated by rare mutations of large effect, which we should not expect to find, but instead by polygenic variation. It assumes, moreover, that the genetic variances of characters in contemporary populations reflect those in their recent ancestors, that is, that the variances remain relatively constant for appreciable periods of evolutionary time. Lande (1975) has argued that mutation-selection balance will maintain reasonably constant genetic variances. Alternative models for the maintenance of genetic variation have been proposed, motivated largely by observed levels of genetic variation that exceed the levels predicted by Lande's model, even in presumably small populations (Turelli 1984; Barton and Turelli 1989; Burger et al. 1989; Houle 1989). Both data and even relatively conservative theory, then, suggest that if a character has a polygenic foundation, it may be expected to manifest substantial genetic variation unless the population is inbred or selection has been unusually strong. This conclusion suggests, in turn, that if some measurement on an organism (a "character") is not found to be genetically variable, either the expression of variation is suppressed ("canalization"), or the "character" lacks a developmental or structural foundation, and in a sense does not exist.
A few authors have examined populations of phytophagous arthropods for genetic variation in responses to plants that the population or species does not naturally feed on, but which are used as hosts by conspecific populations or related species (Futuyma et al. [1993b] cited seven such studies). Evidence for genetic variation was reported in these cases, as well as in most of the more numerous studies of responses to the several hosts actually used by a population (Futuyma and Peterson 1985; Jaenike 1990; Via 1990). This study appears to be the first to report both positive and negative results, and also the first to compare patterns of genetic variation in a phylogenetic context.
In our estimate of the phylogeny of Ophraella, with its inferred history of host associations (fig. 1), we distinguish the pilosa, conferta, and slobodkini clades, the latter including a commune subclade of very closely related species. (Our "conferta clade" equals the "notate" clade of Funk et al. 1995a.) An ancestral association with tribe Astereae (Solidago) has been retained in the pilosa and conferta clades, with an autapomorphic switch to Eupatorieae (Eupatorium) in one member of the conferta clade and a transition to Heliantheae (especially Ambrosia, Iva) in the origin of the slobodkini clade, within which are observed a transition to Anthemideae (Artemisia) and a reversal to Astereae (Chrysopsis, Solidago).
The tribes Astereae (within which Solidago and Chrysopsis are quite closely related) and Anthemideae are more closely related to each other (fig. 2) than to the Heliantheae (including Ambrosiinae), which is paraphyletic with respect to the Eupatorieae (Bremer et al. 1992; Kim et al. 1992; Xiaoping and Bremer 1993). Other tribes, not known to include hosts of Ophraella, are intercalated among these on the phylogeny of the Asteraceae as a whole. Thus, the phylogeny of the host plants does not entirely correspond to the phylogeny of hostplant transitions in Ophraella, providing evidence for host shifts rather than cospeciation (Futuyma and McCafferty 1990; Funk et al. 1995a). However, as noted below, a majority of host shifts have been between between plants in the same tribe, that is, between closely related and chemically similar plants.
Table 6 summarizes the results of our tests for genetic variation in O. conferta (in the conferta clade), O. notulata (in the slobodkini clade), and O. artemisiae and O. communa (both in the commune subclade). Evidence for genetic variation is based on significant sire (S) or dam (D) terms in ANOVAs. In some instances (all tests of O. conferta; tests of O. notulata on S. altissima, Ambrosia), the data are based on offspring of wild-caught females, rather than a half-sib breeding design. The sample sizes of the species of Ophraella other than O. notulata are at least equal to and in most cases substantially greater than those reported for O. notulata in this paper. We treat Artemisia vulgaris, a Eurasian plant closely related to the hosts of O. artemisiae, as if it were a host of O. artemisiae, in order to avoid bias. Feeding responses of species to their natural hosts (including responses of O. artemisiae to Artemisia vulgaris) were genetically variable in all cases, and larval survival was variable in one of the three cases tested.
[TABULAR DATA 6 OMITTED]
Possibly the most important result is that, in 18 of 39 tests of larval or adult feeding responses to congeners' host plants, and in 14 of 16 tests of larval survival, no evidence of genetic variation was discerned. In many instances (marked NN in table 6), only trace (probably exploratory) feeding was exhibited by at most a few individuals, and/or no larvae survived more than a few (ca. 4) days. As described above for O. notulata, the incidence of genetic variation is higher for behavioral responses than for larval survival. Bearing in mind that we cannot definitively conclude that genetic variation is absent, the data nonetheless suggest that genetic constraints could substantially affect the likelihood of adapting to many of these plants. The opportunity for such adaptation has surely existed, for with few exceptions, the geographic ranges of these beetle species broadly overlap those of all the test plants (table 1 of Funk et al. 1995a).
Is there any pattern in the incidence of genetic variation? Is it related to the evolutionary history of host affiliation? We have tested for association between the presence or absence of genetic variation and each of several criteria, calculating exact probabilities by maximum likelihood with the STATEXACT statistical package (Cytel Software Corp. Cambridge, Mass.). In order to obtain an adequate sample size. we treat the 39 larval and adult feeding responses a. independent data, although there is a low but statistically, significant association between them (P = 0.0373, one-tailed test). Because the phylogenetic position of a species is irrelevant under the null hypothesis that the supply of genetic variation does not limit adaptation, it is not clear that a correction for phylogeny (Felsenstein 1985) should be applied to the data. In any case, we cannot do so, for we do not have adequate data for phylogenetic reconstruction of the history of genetic variation in the response to any of the test plants. Table 7 reports univariate tests of association with three criteria, which are correlated with each other to a degree that their relative effects cannot be dissociated. (For this reason, significant effects of the variables were not found in a multivariate analysis by the linear model of logistic regression [procedure CATMOD in SAS], whether the independent variables were treated as binary or multivalued scores. For the latter, we used the DNA "distance" between Ophraella species [Funk et al. 1995a], and a "taxonomic distance" [as in Cheverud et al. 1985] between plant species, based on their phylogeny, for lack of sufficiently comprehensive phylogenetic or distance data.)
Whether a feeding response to a test plant is genetically variable is positively associated (P = 0.0373, one-tailed test) with whether the plant is in the same tribe as the natural host of the beetle species tested (table 7A). This suggests that related plants are more likely to share features, probably chemical, that elicit genetically variable feeding and, moreover, that adaptation to closely related plants may be more likely than adaptation to plants distantly related to the insect's natural host. The genetic data accord with the phylogeny of 1 host shifts in Ophraella, in which four inferred shifts are between plant tribes and six within tribes (nine, if three inferred shifts in geographic populations of O. communa are counted [Funk et al. 1995a]). They are also consonant with the abundant taxonomic and phylogenetic evidence that related insects very frequently feed on related plants (Ehrlich and Raven 1964; Mitter and Farrell 1991; Mitter et al. 1991), instantiating the gradualist Darwinian doctrine that natura non facit saltum.
In testing the hypothesis that genetic variation might most frequently be manifested in responses to plants that represent phylogenetic host shifts, several criteria might be used. By a stringent criterion, a test plant presented to a beetle species represents a realized host shift (according to the inferred history of host affiliation in fig. 1, cf. also fig. 2A in Funk et al. 1995a) if the species tested and the species that normally feeds on the test plant are derived from an immediate common ancestor, such that the test plant represents either an immediately ancestral host association (e.g., Ambrosia, with respect to the Iva association of O. notulata), or an immediately derived host association of the closest relative (e.g., Chrysopsis, with respect to use of Ambrosia by O. communa). By this criterion, the incidence of genetic variation in feeding responses is not correlated with the phylogeny of host shifts (P = 0.3116, table 7B). A less conservative test would include all realized host shifts in which the most recent common ancestor of the species tested and the species normally feeding on the test plant plant gave rise to both host associations, without intervening host shifts. However, because some host shifts defined by this criterion are phylogenetically "deep" (e.g., O. conferta with respect to Ambrosia), we might expect loss of the capacity to respond to the host of a lineage that diverged long ago (e.g., 6.0-11.9 my for divergence of the conferta clade from the Ambrosia-associated slobodkini clade; Funk et al. 1995a). Perhaps unsurprisingly, the incidence of genetic variation in feeding response is not correlated with host shifts defined by this criterion (P = 0.5306, table 7C).
In a still less stringent test, we specify whether or not the Ophraella species that normally feeds on the test plant is a close or a distant relative of the species tested, where a "close" relative is defined as a member of the same major clade (pilosa, conferta, or slobodkini clades). Under this criterion, a significant positive association exists (P = 0.0288, table 7D). Thus, there is evidence that a capacity for adaptation to a plant, at least at the behavioral level, is reflected in the phylogenetic distribution of host utilization, although not in the inferred history of host shifts as such.
From the inferred history of associations with plant genera (fig. 1; also see Funk et al. 1995a), we determined for each experiment whether the genus of the test plant represents an ancestral host association (one or two steps removed), a derived host association (one or two steps removed), or neither, with reference to the insect's current host. No relationship to the polarity of host shift is evident in the pattern of genetic variation (table 8).
Table 8. Distribution of the incidence of genetic variation in larval and adult feeding responses to congeners' hosts, classified by the polarity of host shifts among plant genera as inferred from the phylogeny of Ophraella. "One step" refers to instances in which the host association of the Ophraella species tested is immediately ancestral to or derived from a beetle lineage associated with the test plant; "two steps" refers to instances in which another plant is historically interposed between the species' host association and the lineage that feeds on the test plant. Ancestral Derived One Two One Two Ambigu step steps step steps Neither ous Not detected 3 6 4 3 2 0 Detected 4 1 4 1 9 2(*) (*) Shift within one plant genus (marks responses of O. corferta to Solidago bicolor).
We are unable to test for correspondence between genetic variation and chemical similarities among plants, for lack of adequate information on the secondary compounds of all the species. Phenetic analysis of the available data suggested that host shifts in Ophraella may correspond more strongly to chemical similarities among plants than to their phylogeny, although plant phylogeny and chemical profiles are themselves related (Futuyma and McCafferty 1990). Thus, although the incidence of genetic variation in responses to congeners' hosts is correlated with both the phylogenetic propinquity of Ophraella species and the phylogenetic propinquity of plant species, the latter, perhaps reflecting the distribution of chemical and other features to which the insects respond, may largely account for the pattern of genetic variation. The correlations of genetic variation with plant relationships and, perhaps through these, with the insect phylogeny suggest that the evolution of host associations in Ophraella has been guided not only by ecological agents of selection but also by internal constraints on the capacity for adaptation.
This study of genetic variation was devised explicitly in a phylogenetic perspective, in the conviction that estimates of the history of character evolution and of species' relationships can inform studies of evolutionary mechanisms at the population level. The phylogenetic analysis has provided evidence on speciation and on the relationship of host shifts to plant phylogeny (Funk et al. 1995a), and provides an important basis for several general implications for evolution.
For all variables examined, the mean performance of an Ophraella species was invariably lower on a congener's host than on the species' natural host. In almost all cases the disparity is large, and even in instances in which the responses are genetically variable, there is little evidence of genotypes with levels of performance comparable to the mean on the natural host. This was usually true even of responses to the hosts of very closely related species, for example, the responses of O. communa to Chrysopsis, the occupant of which (O. bilineata) apparently arose recently from O. communa by peripatric speciation (Funk et al. 1995a). If a population's natural host is rare or absent (as encountered by dispersing individuals, for example), selection for adoption of a novel plant is strong, and adaptation may occur if all relevant features are either genetically variable or sufficiently "preadapted" that the population growth rate is positive. However, in populations that have access to the normal host, the low mean performance on novel hosts implies strong selection against feeding or ovipositing on them. Thus, sympatric expansion of host range is unlikely unless there is strong counterselection (e.g., by enemies associated with the normal host; Bernays and Graham 1988; Rausher 1992) and the difference in performance is relatively slight. Our observations, thus, provide no evidence in favor of models of sympatric speciation by host shift (Maynard Smith 1966; Diehl and Bush 1989), which assume a strong fitness advantage of some genotypes on a novel host. These models, moreover, assume a negative genetic correlation in performance on different plants (i.e., a trade-off), for which the literature to date provides little evidence (Jaenike 1990; Futuyma and Keese 1992). Although most of our genetic correlations are calculated for behavioral (feeding) responses, larval consumption rates do affect fitness (witness the high mortality in most cases), and we have found almost no instances of negative genetic correlations.
Few authors have cited lack of genetic variation, except in inbred populations, as a constraint on evolution. Many would agree that "the simplest possible genetic constraint, viz. lack of genetic variation, would appear not to be important" (Barker and Thomas 1987, p. 6), although Rausher (1992, p. 38) noted that selection may deplete genetic variation and that "little is known about how commonly evolution in natural populations is retarded by lack of genetic variation." However, the existence of genetic variation in a character requires that the character exist, that is, that the structural and developmental foundations exist for a trait that we may otherwise only reify by our imagination. Certain imaginable alterations of the carpals and digits of amphibians have not evolved, for example, because the developmental pathways are unlikely to yield the requisite variants (Alberch and Gale 1985; Wake 1991). Growth and survival of insects on plants require metabolic defenses against toxic plant allelochemicals (Brattsten 1992), and the feeding responses of insects to plants are based largely on activation of chemoreceptors by compounds acting both as inhibitors and stimulants, and on central nervous integration of the complex signal pattern (Chapman and Bernays 1989; Dethier 1982; Stadler 1992; Feeny 1992). In these complex, little understood systems, absence of a structural element such as an acceptor protein for a feeding stimulant, or an enzyme capable of metabolizing a toxin, could mean that the foundation for a feeding response or growth simply does not exist.
It is perhaps surprising that among these closely related species of insects dating only to the Miocene, genetic variation in responses to their congeners' hosts, all in the same plant family, should not have been detected in 18 of 39 tests of behavioral reactions and 14 of 16 tests of larval survival (many of the latter cases immediately ascribable to lack of feeding). In 11 of the 18 negative instances for behavior, and 9 of the 15 for survival, effectively all individuals refused to feed, or mortality was complete (table 6). It is at least possible that these represent instances in which the structural requisites for the existence of genetic variation are lacking. Whether or not this is the case, our data imply the existence of genetic constraints on ecologically important characters and provide a genetic rationale for the macroevolutionary conservatism of diet in herbivorous insects. Other phylogenetically conservative traits, if examined, might well also prove to be highly canalized (Render 1967) or genetically invariant.
Behavioral acceptance of a novel plant (or, generally, a resource) and the ability to grow and survive on it together generate strong epistasis for fitness (Castillo-Chavez et al. 1988; Rausher 1993). If a population has access to both the normal and a novel host, acceptance versus avoidance of the novel plant can constitute distinct adaptive peaks, with alleles for avoidance increasing in frequency unless the frequencies of alleles that provide physiological adaptation are sufficiently high. In our data, genetic variation in survival seemingly occurs considerably less often than in behavior. This may imply stabilizing selection on host preference, helping to explain the prevalence of host specificity in this genus. However, we caution that our measure of performance (survival) is not independent of feeding, our criteria for genetic variation in survival (to late larval life or pupation) may be more stringent than for feeding (1- or 2-d tests), and free-ranging animals are likely not to feed as readily as those confined in dishes.
The major purpose of this research program has been to determine whether or not patterns of genetic variation shed any light on the macroevolution of ecological diversification (Futuyma and McCafferty 1990; Futuyma 1992a,b; Futuyma et al. 1993b, 1994). The history of host associations of an insect clade might be entirely ascribable to ecological agents of selection (e.g., plant abundance, predation, competition) if the genetic variation required to adapt to any plant whatever were plentiful. But both the phylogenetic distribution of host associations and our study of genetic variation imply that this is not the case. In our studies, genetic variation in responses to congeners' hosts was associated with both the propinquity of relationships among the beetle species and the propinquity of relationships among the plants, the latter being imperfectly correlated with similarity of overall profiles of secondary compounds. Although we cannot determine the relative impact of these intercorrelated variables, the pattern of genetic variation accords with the hypotheses that similarity among plants facilitates host shifts in phytophagous insects (Borner 1939; Ehrlich and Raven 1964) and that the phylogenetic history of host associations has been guided not only by ecological factors but also by genetic constraints, that undoubtedly are the results of the prior history of host associations. This study, finally, shows that the data of macroevolution and of population biology can illuminate each other (Futuyma 1988): that synthesis of historical and synchronic evolutionary biology is not entirely beyond reach.
[Figures 1 to 2 ILLUSTRATION OMITTED]
We are grateful to C. Herrman, S. Milstein, T. Morton, and P. Quintana-Ascensio for abundant, indispensable assistance with the experiments on O. notulata, to C. Herrera, R. R. Sokal, and B. Thomson for advice and assistance on statistics, to K. Bremer and R. K. Jansen for information on plant phylogeny, to M. Axelrod for greenhouse management, and to the National Science Foundation (BSR-8817912) for support. This is contribution no. 916 in Ecology and Evolution from the State University of New York at Stony Brook
Alberch, P., and E. A. Gale. 1985. A developmental analysis of an evolutionary trend: digital reduction in amphibians. Evolution 39:8-23.
Barker, J. S. F., and R. H. Thomas. 1987. A quantitative genetic perspective on adaptive evolution. Pp. 3-23 in V. Loeschcke, ed. Genetic constraints on adaptive evolution. Springer, Berlin.
Barton, N. H., and M. Turelli. 1989. Evolutionary quantitative genetics: How little do we know? Annual Review of Genetics 23:337-370.
Bernays, E., and M. Graham. 1988. On the evolution of host specificity in phytophagous arthropods. Ecology 69:886-892.
Borner, C. 1939. Anfalligkeit, Resistenz und Immunitat der Reben gegen die Reblaus. Allgemeine Gesichtspunkte zur Frage der Spezialisierung von Parasiten: die harmonische Beschrankung des Lebensraumes. Zeitschrift fur hygienische und Schadlungsbekampfung 31:274-285, 301-308, 325-334.
Bradshaw, A. D. 1991. The Croonian Lecture, 1991: Genostasis and the limits to evolution. Philosophical Transactions of the Royal Society of London B 333:289-305.
Brattsten, L. B. 1992. Metabolic defenses against plant allelo-chemicals. Pp. 175-242 in Rosenthal and Berenbaum 1992.
Bremer, K. 1994. Asteraceae: cladistics and classification. Timberland Press, Portland, Oreg.
Bremer, K., R. K. Jansen, P. O. Karis, M. Kallersjo, S. C. Keeley, K.-J. Kim, H. J. Michaels, J. D. Palmer, and R. S. Wallace. 1992. A review of the phylogeny and classification of the Asteraceae. Nordic Journal of Botany 12:141-148.
Burger, R., G. P. Wagner, and F. Stettinger. 1989. How much heritable variation can be maintained in finite populations by mutation-selection balance? Evolution 43:1748-1766.
Castillo-Chavez, C., S. A. Levin, and F. Gould. 1988. Physiological and behavioral adaptation of insects to varying environments: a mathematical model. Evolution 31:568-579.
Chapman, R. F., and E. Bernays. 1989. Insect behavior at the leaf surface and learning as aspects of host selection. Experientia 45:215-222.
Charlesworth, B., R. Lande, and M. Slatkin. 1982. A neo-Darwinian commentary on macroevolution. Evolution 36:474-498.
Cheverud, J. M., M. M. Dow, and W. Leutenegger. 1985. The quantitative assessment of phylogenetic constraints in comparative analyses: sexual dimorphism in body weight among primates. Evolution 39:1335-1351.
Dethier, V. G. 1954. Evolution of feeding preferences in phytophagous insects. Evolution 8:33-54.
--. 1982. Mechanisms of host-plant recognition. Entomologia Experimentalis et Applicata 31:49-56.
Diehl, S. R., and G. L. Bush. 1989. The role of habitat preference in adaptation and speciation. Pp. 345-365 in D. Otte and J. A. Endler, eds. Speciation and its consequences. Sinauer, Sunderland, Mass.
Eastop, V. F. 1977. Worldwide importance of aphids as virus vectors. Pp. 3-61 in K. F. Harris and K. Maramorosch, eds. Aphids as virus vectors. Academic, New York.
Ehrlich, P. R., and P. H. Raven. 1964. Butterflies and plants: a study in coevolution. Evolution 18:586-608.
Falconer, D. S. 1981. Introduction to quantitative genetics, 2d ed. Longman, New York.
Feeny, P. 1992. The evolution of chemical ecology: contributions from the study of herbivorous insects. Pp. 1-44 in Rosenthal and Berenbaum 1992.
Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist 125:1-15.
Funk, D. J., D. J. Futoyma, G. Orti, and A. Meyer. 1995a. A history of host associations and evolutionary diversification for Ophraella (Coleoptera: Chrysomelidae): new evidence from mitochondrial DNA. Evolution 49:1017-1022.
--. 1995b. Mitochondrial DNA sequences and multiple data sets: a phylogenetic study of phytophagous beetles (Chrysomelidae: Ophraella). Molecular Biology and Evolution 12:627-640.
Futuyma, D. J. 1983. Selective factors in the evolution of host choice by phytophagous insects. Pp. 227-244 in S. Ahmad, eds. Herbivorous insects: host-seeking behavior and mechanisms. Academic Press, New York.
--. 1988. Sturm und Drang and the evolutionary synthesis. Evolution 42:217-226.
--. 1990. Observations on the taxonomy and natural history of Ophraella Wilcox (Coleoptera: Chrysomelidae), with a description of a new species. Journal of the New York Entomological Society 98:163-186.
--. 1991. A new species of Ophraella Wilcox (Coleoptera: Chrysomelidae) from the southeastern United States. Journal of the New York Entomological Society 99:643-653.
--. 1992a. History and evolutionary processes. Pp. 103-130 in M. Nitecki and D. V. Nitecki, eds. History and evolution. University of Chicago Press, Chicago.
--. 1992b. Genetics and the phylogeny of insect-plant interactions. Pp. 191-200 in S. B. J. Menken, J. H. Visser, and P. Harrewijn, eds. Proceedings of the Eighth International Congress on Insect-Plant Relationships. Kluwer, Dordrecht.
Futuyma, D. J., and M. C. Keese. 1992. Evolution and coevolution of plants and phytophagous arthropods. Pp. 439-475 in Rosenthal and Berenbaum 1992.
Futuyma, D. J., and S. S. McCafferty. 1990. Phylogeny and the evolution of host plant associations in the leaf beetle genus Ophraella (Coleoptera: Chrysomelidae). Evolution 44:1885-1913.
Futuyma, D. J., and G. Moreno. 1988. The evolution of ecological specialization. Annual Review of Ecology and Systematics 19: 207-233.
Futuyma, D. J., and S. C. Peterson. 1985. Genetic variation in the use of resources by insects. Annual Review of Entomology 30: 217-238.
Futuyma, D. J., C. Herrmann, S. Milstein, and M. C. Keese. 1993a. Apparent transgenerational effects of host plant in the leaf beetle Ophraella notulata (Coleoptera: Chrysomelidae). Oecologia 96: 365-372.
Futuyma, D. J., M. C. Keese, and S. J. Scheffer. 1993b. Genetic constraints and the phylogeny of insect-plant associations: responses of Ophraella communa (Coleoptera: Chrysomelidae) to host plants of its congeners. Evolution 47:888-905.
Futuyma, D. J., J. Walsh, T. Morton, D. J. Funk, and M. C. Keese. 1994. Genetic variation in a phylogenetic context: responses of two specialized leaf beetles (Coleoptera: Chrysomelidae) to host plants of their congeners. Journal of Evolutionary Biology 7: 127-146.
Georghiou, G. P., and R. Mellon. 1983. Pesticide resistance in time and space. Pp. 1-46 in G. P. Georghiou and T. Saito, eds. Pest resistance to pesticides. Plenum, New York.
Goodwin, B. C. 1984. Changing from an evolutionary to a generative paradigm in biology. Pp. 99-120 in J. W. Pollard, ed. Evolutionary theory: paths into the future. Wiley, New York.
Gould, S. J. 1989. A developmental constraint in Cerion, with comments on the definition and interpretation of constraint in evolution. Evolution 43:516-539.
Jaenike, J. 1990. Host specialization in phytophagous insects. Annual Review of Ecology and Systematics 21:243-273.
Jansen, R. K., K. E. Holsinger, H. J. Michaels, and J. D. Palmer. 1990. Phylogenetic analyses of chloroplast DNA restriction site data at higher taxonomic levels: an example from the Asteraceae. Evolution 44:2089-2105.
Houle, D. 1989. The maintenance of polygenic variation in finite populations. Evolution 43:1767-1766.
Keese, M. C. 1994. Genetic and ecological determinants of host range in leaf feeding beetles (Coleoptera: Chrysomelidae). Ph.D. diss. State University of New York, Stony Brook.
Kim, K.-J., R. K. Jansen, R. S. Wallace, H. J. Michaels, and J. D. Palmer. 1992. Phylogenetic implications of rbcL sequence variation in the Asteraceae. Annals of the Missouri Botanical Garden 79:428-445.
Kindred, B. 1967. Selection for an invariant character, vibrissa number in the house mouse. V. Selection on non-Tabby segregants from Tabby selection lines. Genetics 55:365-373.
Lande, R. 1975. The maintenance of genetic variability by mutation in a polygenic character with linked loci. Genetical Research 26:221-235.
LeSage, L. 1986. A taxonomic monograph of the Nearctic galerucine genus Ophraella Wilcox (Coleoptera: Chrysomelidae). Memoirs of the Entomological Society of Canada 133:1-75.
Lewontin, R. C. 1965. [Comment]. P. 481 in H. G. Baker and G. L. Stebbins, eds. The genetics of colonizing species. Academic Press, New York.
Maynard Smith, J. 1966. Sympatric speciation. American Naturalist 100:637-650.
Maynard Smith, J., and K. C. Sondhi. 1960. The genetics of a pattern. Genetics 45:1039-1050.
Maynard Smith, J., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin, R. Lande, D. Raup, and L. Wolpert. 1985. Developmental constraints and evolution. Quarterly Review of Biology 60:265-287.
Miao, B. 1993. A molecular study of the Ambrosiinae (Asteraceae: Heliantheae), emphasizing chloroplast DNA variations. M. S. thesis. University of Texas, Austin.
Mitter, C., and B. Farrell. 1991. Macroevolutionary aspects of plant/insect interactions. Pp. 35-78 in E. Bernays, ed. Insectplant interactions, Vol. 3. CRC Press, Boca Raton, Fla.
Mitter, C., B. Farrell, and D. J. Futuyma. 1991. Phylogenetic studies of insect-plant interactions: insights into the genesis of diversity. Trends in Ecology and Evolution 6:290-293.
Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic variability in populations. Evolution 29:1-10.
Rausher, M. D. 1992. Natural selection and the evolution of plantinsect interactions. Pp. 20-88 in B. D. Roitberg and M. B. Isman, eds. Insect chemical ecology: an evolutionary approach. Chapman and Hall, New York.
--. 1993. The evolution of habitat preference: avoidance and adaptation. Pp. 259-283 in K. C. Kim and B. A. McPheron, eds. Evolution of insect pests: patterns of variation. Wiley, New York.
Rendel, J. M. 1967. Canalization and gene control. Logos, London.
Rosenthal, G. A., and M. R. Berenbaum, eds. 1992. Herbivores: their interactions with secondary plant metabolites, 2d ed. Academic Press, San Diego, Calif.
Scharloo, W. 1987. Constraints in selection response. Pp. 125-149 in V. Loeschcke, ed. Genetic constraints in adaptive evolution. Springer, Berlin.
Seaman, F. C. 1982. Sesquiterpene lactones as taxonomic characters in the Asteraceae. Botanical Reviews 48:121-595.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2d ed. Freeman, San Francisco.
Stadler, E. 1992. Behavioral responses of insects to plant secondary compounds. Pp. 45-88 in Rosenthal and Berenbaum 1992.
Turelli, M. 1984. Heritable genetic variation via mutation-selection balance: Lerch's zeta meets the abdominal bristle. Theoretical Population Biology 25:138-193.
Via, S. 1984. The quantitative genetics of polyphagy in an insect herbivore. 11. Genetic correlations in larval performance within and among host plants. Evolution 38:582-595.
--. 1990. Ecological genetics and host adaptation in herbivorous insects: the experimental study of evolution in natural and agricultural systems. Annual Review of Entomology 35:421446.
Via, S., and R. Lande. 1985. Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39:505522.
Wake, D. B. 1991. Homoplasy: the result of natural selection, or evidence of design limitations? American Naturalist 138:543567.
Wake, D. B. and A. Larson. 1987. Multidimensional analysis of an evolving lineage. Science 238:42-48.
Weber, K. E., and L. T. Diggins. 1990. Increased selection response in larger populations. II. Selection for ethanol vapor resistance in Drosophila melanogaster at two population sizes. Genetics 125:585-597.
Wiklund, C. 1975. The evolutionary relationship between adult oviposition preference and larval host plant range in Papilio machaon L. Oecologia 18:185-197.
Xiaoping, Z. (Zhang, X.), and K. Bremer. 1993. A cladistic analysis of the tribe Astereae (Asteraceae) with notes on their evolution and subtribal classification. Plant Systematics and Evolution 184:259-283.
|Printer friendly Cite/link Email Feedback|
|Author:||Futuyma, Douglas J.; Keese, Mark C.; Funk, Daniel J.|
|Date:||Oct 1, 1995|
|Previous Article:||A units-of-evolution perspective on the endosymbiont theory of the origin of the mitochondrion.|
|Next Article:||Local adaptation and agents of selection in a mobile insect.|