Cell lineage studies in the clade Eutrochozoa, and especially the Spiralia, remains a rich and relatively untapped source for understanding broad evolutionary developmental problems; including (1) the utility of cell timing formation for phylogenetic hypotheses; (2) the evolution of cell timing changes and its relation to heterochronic patterns; (3) stereotypy or lack thereof in rates of change of cell growth during evolution and its relation to both evolutionary history and current usage; and (4) how mosaic cleavage timing variation may be expected to differ from other groups. A compilation of available cell timing information was made from previous studies where each division was explicitly followed and the total number of cells followed was greater than 24. From that compilation, we performed a series of heuristic and quantitative analyses, including a phylogenetic analysis using cell timing data as characters and analyses of timing variation across all taxa. Our results show that: (1) cell lineage data reconstructs a phylogenetic hypothesis that has similarities, especially among the Mollusca, to the patterns found in morphological and molecular analyses; (2) the mesentoblast (4d) is a unique cell compared to other cell in that it speeds up and slows down relative to other cells in taxa with both unequal and equal cell sizes; (3) some cells that form in the same quartet at the same point in the cell lineage hierarchy have much lower variations than analogous other cells, arguing for architectural constraint or stabilizing selection acting on those cells; and (4) although variation in cell timing generally increases during development, timing of formation of progeny cells in the first quartet has lower variation than the parent cells, arguing that some regulation-like behavior might be present.

Corresponding Editor: M. Zelditch

We consider the evolution of ecological specialization in a landscape with two discrete habitat types connected by migration, for example, a plant-insect system with two plant hosts. Using a quantitative genetic approach, we study the joint evolution of a quantitative character determining performance in each habitat together with the changes in the population density. We find that specialization on a single habitat evolves with intermediate migration rates, whereas a generalist species evolves with both very low and very large rates of movement between habitats. There is a threshold at which a small increase in the connectivity of the two habitats will result in dramatic decrease in the total population size and the nearly complete loss of use of one of the two habitats through a process of “migrational meltdown.” In some situations, equilibria corresponding to a specialist and a generalist species are simultaneously stable. Analysis of our model also shows cases of hysteresis in which small transient changes in the landscape structure or accidental demographic disturbances have irreversible effects on the evolution of specialization.

Corresponding Editor: M. Whitlock

This paper describes a new approach to modeling population structure for genes under strong balancing selection of the type seen in plant self-incompatibility systems and the major histocompatibility complex (MHC) system of vertebrates. Simple analytic solutions for the number of alleles maintained at equilibrium and the expected proportion of alleles shared between demes at various levels are derived and checked against simulation results. The theory accurately captures the dynamics of allele number in a subdivided population and identifies important values of m (migration rate) at which allele number and distribution change qualitatively. Starting from a panmictic population, as migration among demes decreases a qualitative change in dynamics is seen at approximately m_crit ≈ (s/4πN_T)^1/2, where N_T is the total population size and s is a measure of the strength of selection. At this point, demes can no longer maintain their panmictic allele number, due to increasing isolation from the total population. Another qualitative change occurs at a migration rate on the same order of magnitude as the mutation rate, u. At this point, the demes are highly differentiated for allele complement, and the total number of alleles in the population is increased. Because in general u ≪ m_crit, at intermediate migration rates slightly fewer alleles may be maintained in the total population than are maintained at panmixia. Within this range, total allele number may not be the best indicator of whether a population is effectively panmictic, and some caution should be used when interpreting samples from such populations. The theory presented here can help to analyze data from genes under balancing selection in subdivided populations.

Corresponding Editor: A. Caballero

Plants inhabited by ants (myrmecophytes) have evolved in a diversity of tropical plant lineages. Macaranga includes approximately 300 paleotropical tree species; in western Malesia there are 26 myrmecophytic species that vary in morphological specializations for ant association. The origin and diversification of myrmecophytism in Macaranga was investigated using phylogenetic analyses of morphological and nuclear ITS DNA characters and studies of character evolution. Despite low ITS variation, the combined analysis resulted in a well-supported hypothesis of relationships. Mapping myrmecophytism on all most parsimonious trees resulting from the combined analysis indicated that the trait evolved independently between two and four times and was lost between one and three times (five changes). This hypothesis was robust when tested against trees constrained to have three or fewer evolutionary transformations, although increased taxon sampling for the ITS analysis is required to confirm this. Mapping morphological traits on the phylogeny indicated that myrmecophytism was not homologous among lineages; each independent origin involved a suite of different specializations for ant-plant association. There was no evidence that myrmecophytic traits underwent sequential change through evolution; self-hollowing domatia evolved independently from ant-excavated domatia, and different food-body production types evolved in different lineages. The multiple origins of myrmecophytism in Macaranga were restricted to one small, exclusively western Malesian lineage of an otherwise large and nonmyrmecophytic genus. Although the evolution of aggregated food-body production and the formation of domatia coincided with the evolution of myrmecophytism in all cases, several morphological, ecological, and biogeographic factors appear to have facilitated and constrained this radiation of ant-plants.

Corresponding Editor: C. Boggs

The roles of the various potential ecological and evolutionary causes of spatial population genetic structure (SPGS) cannot in general be inferred from the extant structure alone. However, a stage-specific analysis can provide clues as to the causes of SPGS. We conducted a stage-specific SPGS analysis of a mapped population of about 2000 Trillium grandiflorum (Liliaceae), a long-lived perennial herb. We compared SPGS for juvenile (J), nonreproductive (NR), and reproductive (R) stages. Fisher's exact test showed that genotypes had Hardy-Weinberg frequencies at all loci and stage classes. Allele frequencies did not differ between stages. Bootstrapped 99% confidence intervals (99%CI) indicate that F-statistic values are indistinguishable from zero, (except for a slightly negative F_IT for the R stage). Spatial autocorrelation was used to calculate f, the average kinship coefficient between individuals within distance intervals. Null hypothesis 99%CIs for f were constructed by repeatedly randomizing genotypic locations. Significant positive fine-scale genetic structure was detected in the R and NR stages, but not in the J stage. This structure was most pronounced in the R stage, and declined by about half in each remaining stage: near-neighbor f = 0.122, 0.065, 0.027, for R, NR, and J, respectively. For R and NR, the near-neighbor f lies outside the null hypothesis 99%CI, indicating kinship at approximately the level of half-sibs and first cousins, respectively. We also simulated the expected SPGS of juveniles post dispersal, based on measured R-stage SPGS, the mating system, and measured pollen and seed dispersal properties. This provides a null hypothesis expectation (as a 99%CI) for the J-stage correlogram, against which to test the likelihood that post-dispersal events have influenced J-stage SPGS. The actual J correlogram lies within the null hypothesis 99%CI for the shortest distance interval and nearly all other distance intervals indicating that the observed low recruitment, random mating and seed dispersal patterns are sufficient to account for the disappearance of SPSG between the R and the J stages. The observed increase in SPGS between J and R stages has two potential explanations: history and local selection. The observed low total allelic diversity is consistent with a past bottleneck: a possible historical explanation. Only a longitudinal stage-specific study of SPGS structure can distinguish between historical events and local selection as causes of increased structure with increasing life history stage.

Corresponding Editor: O. Savolainen

The level and distribution of genetic variation is thought to be affected primarily by the size of individual populations and by gene flow among populations. Although the effects of population size have frequently been examined, the contributions of regional gene flow to levels of genetic variation are less well known. Here I examine the effects of population size and the number of neighboring populations (metapopulation density) on the distribution and maintenance of genetic diversity in an endemic herbaceous perennial. Reductions in the proportion of polymorphic loci and the effective number of alleles per locus were apparent for many populations with a census size of less than 100 individuals, but no effects of population size on levels of inbreeding were detected. I assess the effects of regional population density on levels of diversity and inbreeding using stepwise regression analysis of metapopulation diameter (i.e., the size of a circle within which population density is estimated). This procedure provides a spatially explicit evaluation of the effects of metapopulation size on population genetic parameters and indicates the critical number of neighboring populations (fragmentation threshold) for the regional maintenance of genetic diversity. Stepwise regression analyses revealed fragmentation thresholds at two levels; at a scale of 2 km, where small metapopulations resulted in greater levels of selfing or sibling mating, and at a scale of 8 km, where metapopulation size was positively associated with higher levels of genetic diversity. I hypothesize that the smaller fragmentation threshold may reflect higher levels of selfing in isolated populations because of the absence of pollinators. The larger threshold probably indicates the maximum distance over which pollen dispersal rates are high enough to counteract genetic drift. This study demonstrates that the regional distribution of populations can be an important factor for the long-term maintenance of genetic variation.

Corresponding Editor: O. Savolainen

In the weedy plant species Allium vineale (wild garlic), individuals may simultaneously produce sexually and asexually derived offspring, by seed and bulbils, respectively. In this study, genetic and genotypic diversity was determined in samples from 14 European A. vineale populations using nuclear (RAPD) and cytoplasmic (PCR-RFLP of cpDNA) markers to investigate the importance of the different reproductive modes. In the whole sample, 77 nuclear multilocus genotypes and four chloroplast haplotypes (chlorotypes) were found. Populations exhibited a high degree of subdivision for nuclear and cytoplasmic markers as estimated from hierarchical F-statistics; at the same time, identical chlorotypes could be found in populations separated by large distances. Genotypic diversity was significantly lower than expected under free recombination in almost all populations, indicating that recruitment into populations is mostly by asexually produced offspring. Nevertheless, within each chlorotype, the distribution of markers from pairs of nuclear loci was incompatible with a purely clonal structure, suggesting that many multilocus genotypes have originated by sexual recombination rather than by mutation within asexual lineages. It is argued that the weedy habit of A. vineale is likely to have favored bulbil reproduction, whereas sexually generated genotypes may have facilitated local adaptation during the species' expansion across Europe.

Corresponding Editor: M. Dudash

Because of their extensive functional interaction, mitochondrial DNA (mtDNA) and nuclear genes may evolve to form coadapted complexes within reproductively isolated populations. As a consequence of coadaptation, the fitness of particular nuclear alleles may depend on mtDNA genotype. Among populations of the copepod Tigriopus californicus, there are high levels of amino acid substitutions in both the mtDNA genes encoding subunits of cytochrome c oxidase (COX) and the nuclear gene encoding cytochrome c (CYC), the substrate for COX. Because of the functional interaction between enzyme and substrate proteins, we hypothesized that the fitness of CYC genotypes would depend on mtDNA genotype. To test this hypothesis, segregation ratios for CYC and a second nuclear marker (histone H1) unrelated to mitochondrial function were scored in F₂ progeny of several reciprocal interpopulation crosses. Genotypic ratios at the CYC locus (but not the H1 locus) differed between reciprocal crosses and differed from expected Mendelian ratios, suggesting that CYC genotypic fitnesses were strongly influenced by cytoplasmic (including mtDNA) background. However, in most cases the nature of the deviations from Mendelian ratios and differences between reciprocal crosses are not consistent with simple coevolution between CYC and mtDNA background. In a cross in which both newly hatched larvae and adults were sampled, only the adult sample showed deviations from Mendelian ratios, indicating that genotypic viabilities differed. In two of six crosses, large genotypic ratio differences for CYC were observed between the sexes. These results suggest that significant variation in nuclear-mtDNA coadaptation may exist between T. californicus populations and that the relative viability of specific cytonuclear allelic combinations is somehow affected by sex.

Corresponding Editor: S. Edwards

Little is known about physiological mechanisms that underlie the cost of reproduction. We tested the hypothesis that stress susceptibility is a cost of reproduction. In one test of our hypothesis, Drosophila melanogaster females were exposed to a juvenile hormone analog (methoprene) to stimulate egg production followed by stress assays. A sterile stock of D. melanogaster was employed as a control for reproduction. Exposure of fertile females to methoprene resulted in an increase in female reproduction and increased susceptibility to oxidative stress and starvation (compared to solvent controls). Sterile females did not exhibit a decrease in stress resistance. Mating also stimulated egg production. As a second test of our hypothesis, mated females were compared to virgin females. Mated fertile females were relatively susceptible to oxidative stress, but this relationship was not evident when mated and virgin sterile females were compared. The results of the present study support the hypothesis that stress susceptibility is a cost of reproduction.

Corresponding Editor: T. Mousseau

The extent of genetic variation in fitness and its components and genetic variation's dependence on environmental conditions remain key issues in evolutionary biology. We present measurements of genetic variation in preadult viability in a laboratory-adapted population of Drosophila melanogaster, made at four different densities. By crossing flies heterozygous for a wild-type chromosome and one of two different balancers (TM1, TM2), we measure both heterozygous (TM1/ , TM2/ ) and homozygous ( / ) viability relative to a standard genotype (TM1/TM2). Forty wild-type chromosomes were tested, of which 10 were chosen to be homozygous viable. The mean numbers produced varied significantly between chromosome lines, with an estimated between-line variance in log_e numbers of 0.013. Relative viabilities also varied significantly across chromosome lines, with a variance in log_e homozygous viability of 1.76 and of log_e heterozygous viability of 0.165. The between-line variance for numbers emerging increased with density, from 0.009 at lowest density to 0.079 at highest. The genetic variance in relative viability increases with density, but not significantly. Overall, the effects of different chromosomes on relative viability were remarkably consistent across densities and across the two heterozygous genotypes (TM1, TM2). The 10 lines that carried homozygous viable wild-type chromosomes produced significantly more adults than the 30 lethal lines at low density and significantly fewer adults at the highest density. Similarly, there was a positive correlation between heterozygous viability and mean numbers at low density, but a negative correlation at high density.

Corresponding Editor: T. Kawecki

Stress resistance traits in Drosophila often show clinal variation. Although these patterns suggest selection, there is generally no attempt to test how large differences at the geographical level are relative to levels of variation within and between local populations. Here we compare these levels in D. melanogaster from temperate Tasmania versus tropical northern Queensland by focusing on adult resistance to desiccation, cold and starvation stress, as well as associated traits (size, lipid content). For starvation and desiccation resistance, levels of variation were highest among strains from the same population, whereas there was little differentiation among local populations and a low level of differentiation at the geographic level. For adult cold resistance, there was local differentiation and strain variation but no geographic variation. For size (thorax length), geographic differentiation was higher despite some overlap among strains from the tropical and temperate locations. Finally, for lipid levels there was only evidence for variation among strains. The low level of differentiation among geographic locations for stress resistance was further verified with the characterization of isofemale strains from 18 locations along a coastal transect extending from Tasmania to northern Queensland. Crosses among some of the isofemale strains showed that results were not confounded by inbreeding effects. Strains derived from a cross between a tropical and temperate strain differed for all traits, and variation among strains for body size was higher than strain variation within the geographic regions. Unlike in previous studies, lipid content and starvation resistance were not correlated in any set of strains, but there was a correlation between cold resistance and lipid content. There was also a correlation between desiccation resistance and size but only in the geographic cross strains. These findings suggest a large amount of variation in stress resistance at the population level and inconsistent correlation patterns across experimental approaches.

Corresponding Editor: W. T. Starmer

Recent studies, primarily in Drosophila, have greatly advanced our understanding of Haldane's rule, the tendency for hybrid sterility or inviability to affect primarily the heterogametic sex (Haldane 1922). Although dominance theory (Turelli and Orr 1995) has been proposed as a general explanation of Haldane's rule, this remains to be tested in female-heterogametic taxa, such as the Lepidoptera. Here we describe a novel example of Haldane's rule in Heliconius melpomene (Lepidoptera; Nymphalidae). Female F₁ offspring are sterile when a male from French Guiana is crossed to a female from Panama, but fertile in the reciprocal cross. Male F₁s are fertile in both directions. Similar female F₁ sterility occurs in crosses between French Guiana and eastern Colombian populations. Backcrosses and linkage analysis show that sterility results from an interaction between gene(s) on the Z chromosome of the Guiana race with autosomal factors in the Panama genome. Large X (or Z) effects are commonly observed in Drosophila, but to our knowledge have not been previously demonstrated for hybrid sterility in Lepidoptera. Differences in the abundance of male versus female or Z-linked versus autosomal sterility factors cannot be ruled out in our crosses as causes of Haldane's rule. Nonetheless, the demonstration that recessive Z-linked loci cause hybrid sterility in a female heterogametic species supports the contention that dominance theory provides a general explanation of Haldane's rule (Turelli and Orr 2000).

Corresponding Editor: C. López-Fanjul

Multiple mating by females (polyandry) is taxonomically widespread but the evolution of such behaviors is not clearly understood given potential costs of polyandry such as time, energy, or predation risk. The genetic variability versus parasites hypothesis predicts a reduction of parasitism due to increased genetic variability among offspring and an associated fitness gain. We tested this hypothesis with a field experiment in the bumblebee, Bombus terrestris L. Worker heterogeneity within the colony was experimentally altered by artificially inseminating queens with sperm from one male, four brothers, two males, or four unrelated males. We found genetic variability to be effective, because intensity and prevalence of the most common parasite, Crithidia bombi, a trypanosome, decreased with increasing levels of colony heterogeneity. Fitness differed between treatments but did not increase in a simple way, with increasing genetic heterogeneity among colony workers. Instead, fitness followed a U-shaped function with a minimum for small amounts of genetic heterogeneity. We therefore suggest that polyandry also induces a cost, perhaps as a result of the social structure within the colony. In evolutionary terms, singly mated females appear to be constrained by an adaptive valley that needs to be crossed before high degrees of mating frequency can be reached. This may help to explain why B. terrestris and most other social insects are often monandrous.

Corresponding Editor: D. Wheeler

Variation among females in mate choice may influence evolution by sexual selection. The genetic basis of this variation is of interest because the elaboration of mating preferences requires additive genetic variation in these traits. Here we measure the repeatability and heritability of two components of female choosiness (responsiveness and discrimination) and of female preference functions for the multiple ornaments borne by male guppies (Poecilia reticulata). We show that there is significant repeatable variation in both components of choosiness and in some preference functions but not in others. There appear to be several male ornaments that females find uniformly attractive and others for which females differ in preference. One consequence is that there is no universally attractive male phenotype. Only responsiveness shows significant additive genetic variation. Variation in responsiveness appears to mask variation in discrimination and some preference functions and may be the most biologically relevant source of phenotypic and genetic variation in mate-choice behavior. To test the potential evolutionary importance of the phenotypic variation in mate choice that we report, we estimated the opportunity for and the intensity of sexual selection under models of mate choice that excluded and that incorporated individual female variation. We then compared these estimates with estimates based on measured mating success. Incorporating individual variation in mate choice generally did not predict the outcome of sexual selection any better than models that ignored such variation.

Corresponding Editor: L. Bernatchez

The trade-off between reproductive investment and migration should be an important factor shaping the evolution of life-history traits among populations following their radiation into habitats with different migratory costs and benefits. An experimentally induced difference in migratory rigor for families of chinook salmon (Oncorhynchus tshawytscha), of approximately 86 km and 413 m elevation, exacted a cost to somatic energy reserves (∼ 17% reduction in metabolizable mass) and ovarian investment (13.7% reduction in ovarian mass). This cost was associated with a reduction in egg size and paralleled the phenotypic pattern of divergence between two introduced New Zealand populations of common origin, presently breeding at sites with different migration distances. The genetic pattern of divergence of these same populations, detected under common rearing, was consistent with compensation for migratory costs (the population that migrates farther invested more in ovarian mass), but egg number more than egg size was associated with this evolution. These evolutionary patterns are consistent with what is known of the inheritance of these traits and with trade-offs and constraints favoring initial evolution in offspring number over offspring size. Analysis of egg number-size patterns of other Pacific salmon populations in their native range supported the hypothesis that migration strongly influences patterns of reproductive allocation, favoring a higher ratio of egg number to egg size with greater migration distance.

Corresponding Editor: M. Zelditch

Metamorphosis is assumed to be beneficial because it can break developmental links between traits in the different phases of a complex life-cycle and thereby allow larval and adult phases to adapt independently. I tested the prediction that correlations between the larval and adult phases are smaller than within stages. I estimated phenotypic and additive genetic variances and correlations for tadpole swimming speed, frog jump distance, body size, and larval period in a single population of the Pacific tree frog, Hyla regilla. These traits are known or reasonably assumed to be important for survival in this and other anuran species from temporary ponds. Only the three size variables were affected by sire identity. Heritabilities for locomotor performance, larval period, and size-independent performance were low (0.00–0.23) and not significant. Body size measurements showed somewhat higher and statistically significant heritabilities (0.24–0.34). Most traits were phenotypically correlated. On average, phenotypic correlations were larger between phases than within phases (0.41 vs. 0.28). Genetic correlations involving body-size traits were positive and large, and average within- and between-phase genetic correlation coefficients were identical (0.81). These results do not support the adaptive decoupling hypothesis, and they indicate that a paucity of additive genetic variation is a likely constraint on the evolution of traits measured for this population.

Corresponding Editor: T. Garland Jr.

Nucleotide sequence data from the mitochondrial control region were used from a phylogenetic context to investigate the long-term history of a population of bowhead whales (Balaena mysticetus). In addition, the coalescence time of these sequences was used to estimate the age of the inferred patterns of population size change. The results indicate that mitochondrial genetic polymorphism was not affected by a recent bottleneck that occurred near the turn of the 20th century, thereby preserving the signature of historical population size change in the mitochondrial genome. Further analysis showed that this population underwent an expansion initiated in the Middle to Late Pleistocene. As such, early Holocene changes in Arctic sea ice distribution appear to have had little influence on patterns of genetic variability in this population.

Corresponding Editor: S. Karl

This study investigates hybridization and population genetics of two species of macaque monkey in Sulawesi, Indonesia, using molecular markers from mitochondrial, autosomal, and Y-chromosome DNA. Hybridization is the interbreeding of individuals from different parental taxa that are distinguishable by one or more heritable characteristics. Because hybridization can affect population structure of the parental taxa, it is an important consideration for conservation management. On the Indonesian island of Sulawesi an explosive diversification of macaques has occurred; seven of 19 species in the genus Macaca live on this island. The contact zone of the subjects of this study, M. maura and M. tonkeana, is located at the base of the southwestern peninsula of Sulawesi. Land conversion in Sulawesi is occurring at an alarming pace; currently two species of Sulawesi macaque, one of which is M. maura, are classified as endangered species. Results of this study indicate that hybridization among M. maura and M. tonkeana has led to different distributions of molecular variation in mitochondrial DNA and nuclear DNA in the contact zone; mitochondrial DNA shows a sharp transition from M. maura to M. tonkeana haplotypes, but nuclear DNA from the parental taxa is homogenized in a narrow hybrid zone. Similarly, within M. maura divergent mitochondrial DNA haplotypes are geographically structured but population subdivision in the nuclear genome is low or absent. In M. tonkeana, mitochondrial DNA haplotypes are geographically structured and a high level of nuclear DNA population subdivision is present in this species. These results are largely consistent with a macaque behavioral paradigm of female philopatry and obligate male dispersal, suggest that introgression between M. maura and M. tonkeana is restricted to the hybrid zone, and delineate one conservation management unit in M. maura and at least two in M. tonkeana.

Corresponding Editor: M. Hammer

Partial Mantel tests were designed to test for correlation among three matrices of pairwise distances. We show through an example that these tests may be inadequate, because the associated P-value is not indicative of the type I error.

Corresponding Editor: D. Waller

Ecological studies suggest that hummingbird-pollinated plants in North America mimic each other to increase visitation by birds. Published quantitative trait locus (QTL) data for two Mimulus species indicate that floral traits associated with hummingbird versus bee pollination results from a few loci with major effects on morphology, as predicted by classical models for the evolution of mimicry. Thus, the architecture of genetic divergence associated with speciation may depend on the ecological context.

Corresponding Editor: O. Savolainen

Wolbachia-induced cytoplasmic incompatibility (CI) can have two consequences in haplodiploid insects: fertilized eggs either die (female mortality, FM) or they develop into haploid males (male development, MD). Origin of this diversity remains poorly understood, but current hypotheses invoke variation in damage suffered by paternal chromosomes in incompatible eggs, thus intermediate CI types should be expected. Here, we show the existence of such a particular CI type. In the parasitoid wasp Leptopilina heterotoma, we compared CI effects in crosses involving lines derived from a single inbred line with various Wolbachia infection statuses (natural tri-infection, mono-infection, or no infection). Tri-infected males induce a FM CI type when crossed with either uninfected or mono-infected females. Crossing mono-infected males with uninfected females results in almost complete CI with both reduced offspring production, indicating partial mortality of fertilized eggs, and increased number of sons, showing haploid male development of others. Mono-infected males thus induce an intermediate CI type when mated with uninfected females. The first evidence of this expected particular CI type demonstrates that no discontinuity separates MD and FM CI types, which appear to be end points of a phenotypic continuum. Second, different CI types can occur within a given species and even within offspring of a single pair. Third, phenotypic expression of the particular CI type induced by a given Wolbachia variant depends on other bacterial variants that co-infect the same tissues. These results support the idea that haplodiploids should be helpful in clarifying evolutionary pathways of insect-Wolbachia associations.

Corresponding Editor: K. Ross

Corresponding Editor: D. Futuyma