Parallel evolution of larval morphology and ... - Wiley Online Library

doi:10.1111/j.1420-9101.2006.01132.x

Parallel evolution of larval morphology and habitat in the snail-killing fly genus Tetanocera E. G. CHAPMAN,* B. A. FOOTE,* J. MALUKIEWICZ & W. R. HOEH* *Evolutionary, Population, and Systematic Biology Group, Department of Biological Sciences, Kent State University, Kent, OH, USA Program in Biological Anthropology, Department of Anthropology, Kent State University, Kent, OH, USA

Keywords:

Abstract

parallel evolution; phylogenetic comparative method; phylogenetic niche conservatism; Sciomyzidae; Tetanocera.

In this study, we sequenced one nuclear and three mitochondrial DNA loci to construct a robust estimate of phylogeny for all available species of Tetanocera. Character optimizations suggested that aquatic habitat was the ancestral condition for Tetanocera larvae, and that there were at least three parallel transitions to terrestrial habitat, with one reversal. Maximum likelihood analyses of character state transformations showed significant correlations between habitat transitions and changes in four larval morphological characteristics (cuticular pigmentation and three characters associated with the posterior spiracular disc). We provide evidence that phylogenetic niche conservatism has been responsible for the maintenance of aquatic-associated larval morphological character states, and that concerted convergence and/or gene linkage was responsible for parallel morphological changes that were derived in conjunction with habitat transitions. These habitat–morphology associations were consistent with the action of natural selection in facilitating the morphological changes that occurred during parallel aquatic to terrestrial habitat transitions in Tetanocera.

Introduction Detecting significant correlations among ecological and morphological traits is one of the primary objectives of comparative biology (Harvey & Pagel, 1991; Martins & Hansen, 1996; Armbruster, 2002; Felsenstein, 2004). Phylogenetic comparative methods (e.g. Felsenstein, 1985a; Maddison, 1990,2000; Pagel, 1994,1997,2000,2002; Martins, 2000; Maddison & Maddison, 2003; Pagel et al., 2004), which allow statistically rigorous testing of correlations among characters, help us to gain insight into which morphological characters may be adaptive in differing habitats. Few studies of dipteran taxa have used phylogenetic comparative methods to study morphological adaptations that occurred in concert with or in response to habitat transitions (e.g. Scheffer & Correspondence: E. G. Chapman, Department of Biological Sciences, Cunningham Hall, Kent State University, Kent, OH 44242-0001, USA. Tel.: (330) 672-2921; fax: (330) 672-3713; e-mail: [email protected]

Wiegmann, 2000). Fewer yet have attempted to unravel the morphological adaptations that have facilitated or accompanied transitions between aquatic and terrestrial habitats. Although Vermeij & Dudley (2000) reported that transitions between aquatic and terrestrial habitats are rare in plants and animals (with the exception of tetrapod vertebrates), there are a number of dipteran families that have sublineages that must have made such transitions (e.g. Chironomidae, Dolichopodidae, Empididae, Ephydridae, Muscidae, Sarcophagidae, Sciomyzidae, Stratiomyidae, Syrphidae, Tabanidae, and Tipulidae). Our study, which used phylogenetic comparative methods to explore morphological adaptations to both aquatic and terrestrial habitats in the sciomyzid genus Tetanocera, is one of the first to do so within a dipteran lineage. The genus Tetanocera (Diptera: Sciomyzidae: Tetanocerini) displays significant potential for studying morphological adaptations that may have occurred in concert with or in response to habitat transitions for two reasons:

ª 2006 THE AUTHORS 19 (2006) 1459–1474 JOURNAL COMPILATION ª 2006 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1459

1460

E. G. CHAPMAN ET AL.

Fig. 1 Illustrations of third instar Tetanocera larvae: (a) Lateral view of Tetanocera ferruginea (aquatic larva); (b) Posterior spiracular disc (rear view) of T. ferruginea; (c) lateral view of T. melanostigma (terrestrial larva); (d) posterior spiracular disc (rear view) of T. melanostigma. Figures from Foote (1961).

(i) larvae of its ca. 39 currently recognized species inhabit both aquatic and terrestrial habitats (Foote, 1996a,b,1999); (ii) Larvae of aquatic species share a suite of morphological characteristics that differentiates them from the terrestrial species (Fig. 1). Aquatic Tetanocera larvae live just under the surface of the water, usually against dark floating debris or plant stems (Foote, 1999), whereas terrestrial Tetanocera larvae occupy habitats ranging from damp near-shore to drier woodland habitats (Foote, 1996a,b). Aquatic Tetanocera larvae are darkly pigmented (nearly black), which is likely for crypsis. The posterior spiracular disc is typically modified in three ways: (i) the last abdominal segment is lengthened and the disc is upturned; (ii) the spiracles have a ring of long, branched, hydrophobic hairs (float hairs); and (iii) there are elongated, hirsute ventrolateral and ventral lobes extending from the disc (Fig. 1a, b). The upturned spiracular disc enables the larvae to breathe while remaining ca. horizontal under the surface of the water. The float hairs perform two functions: Aquatic Tetanocera larvae attack and eat aquatic snails just below the surface of the water (Foote, 1999). When a prey item dies or flees, it sometimes loses contact with the surface of the water, and drops through the water column, pulling the feeding Tetanocera larva down with it. When this happens, the float hairs fold over the spiracles, trapping a bubble of air (Foote, 1999). As the larva contacts the surface, the hydrophobic float hairs break the surface tension of the water and hold the spiracles (which protrude slightly from the spiracular disc) above the surface (Foote, 1999). The elongated hairy lobes around the spiracular disc likely aid in keeping the disc at the surface (B. A. Foote, unpublished data). In contrast, terrestrial Tetanocera larvae are translucent (the nonpigmented cuticle is clear; the larvae appear white to tan). The last abdominal segment is not lengthened and terminates in a rear-facing posterior spiracular disc, and the lobes and float hairs are greatly reduced (Fig. 1c, d). The distribution of these two distinct suites of larval character states, in multiple Tetanocera species, could be

the result of either a single habitat shift and subsequent speciation or multiple shifts and parallel evolution in larval morphology. There are a number of interesting evolutionary questions one can ask about Tetanocera. Do the aquatic and terrestrial species each comprise distinct lineages, or were there multiple independent habitat transitions? The answer to this question would allow us to address whether the aforementioned morphological differences between aquatic and terrestrial larvae were the result of a single habitat transition and subsequent speciation, or if multiple independent habitat transitions were each accompanied by the same morphological changes as a result of parallel evolution. If the latter were true, it would be consistent with the hypothesis that adaptation played a significant role in the transitions. Another evolutionary question deals with the polarity and order of larval character state transitions that occurred during Tetanocera phylogenesis: what was the ancestral larval habitat of Tetanocera? Did the lineage originate with aquatic larvae, subsequently transitioning to terrestrial habitats, or the reverse scenario? The nearly equal number of aquatic and terrestrial species (of the 28 species with known life cycles, 13 are aquatic, 14 are terrestrial and one is facultative) indicates that either hypothesis should be considered plausible until they can be evaluated using robust phylogenetic methods. This study has four principal objectives: (i) use DNA sequences to construct a robust estimate of phylogeny for all available species of Tetanocera, (ii) estimate the polarity and order of evolutionary transitions in larval habitat that have occurred during Tetanocera phylogenesis, (iii) test whether there is a significant correlation between larval habitat and morphology and (iv) evaluate the hypothesis that parallel habitat transitions were accompanied by parallel state shifts in four larval morphological characters. Phylogenetic comparative methods have allowed us to detect evolutionary phenomena such as parallel evolution, concerted convergence/gene linkage, and phylogenetic niche conservatism within Tetanocera.


Habitat correlated parallel evolution

Materials and methods Taxon sampling The Sciomyzidae contains two subfamilies, with over 99% of the >500 species belonging to the Sciomyzinae (Marinoni & Mathis, 2000; Knutson & Vala, 2002). The Sciomyzinae is comprised of two tribes (Sciomyzini and Tetanocerini). All of the Sciomyzini have terrestrial larvae, whereas 14 tetanocerine genera have aquatic larvae. A recent phylogenetic analysis of sciomyzid morphological data suggests that the Sciomyzinae and associated tribes are monophyletic (Marinoni & Mathis, 2000). Therefore, phylogenetic analyses were performed on DNA sequences from four loci obtained from 31 Tetanocera individuals (representing 17 species) and 23 individuals representing eight outgroup genera (six from the Tetanocerini and two from the Sciomyzini; Table 1). For 11 of the 17 Tetanocera species, multiple individuals were available and sequenced for replicate sampling purposes. In all but one case (the

1461

Sciomyza simplex terminal), the four DNA loci sequenced used to represent a particular terminal taxon were obtained from the same individual. However, one individual of S. simplex provided the three mtDNAencoded DNA sequences while another contributed the 28S sequence. Tetanocera larvae feed on (i) nonoperculate pulmonate aquatic snails in the water, (ii) nonoperculate pulmonate aquatic snails on the shoreline, (iii) slugs (genera: Deroceras, Pallifera, Philomycus), (iv) semi-terrestrial snails in the family Succineidae and (v) terrestrial snails (Knutson & Vala, 2002). Only those species that belong to group no. 1 (above) have truly aquatic larvae. The genus is Holarctic in distribution with 18 Nearctic, nine Palaearctic and 12 Holarctic species. Eight of the Tetanocera species analysed herein have aquatic larvae, and nine have terrestrial larvae (two in group no. 2, three in group no. 3, two in group no. 4, and two in group no. 5). Together, these 17 species span the range of morphological variation among known Tetanocera larvae for the characters listed above.

Table 1 List of species analysed in this study with GenBank accession numbers and collecting locality information. Tribe

Species

Habitat

28S

16S

COI

COII

Locality

Tetanocerini

Tetanocera bergi Steyskal, 1954 Tetanocera ferruginea Falle´n, 1820 Tetanocera latifibula Frey, 1924 Tetanocera mesopora Steyskal, 1959 Tetanocera montana Day, 1881 Tetanocera plumosa Loew, 1847 Tetanocera robusta Loew, 1847 Tetanocera vicina Macquart, 1843 Tetanocera fuscinervis (Zetterstedt, 1838) Tetanocera silvatica Meigen, 1830 Tetanocera kerteszi Hendel 1901 Tetanocera phyllophora Melander, 1920 Tetanocera clara Loew, 1862 Tetanocera plebeja Loew, 1862 Tetanocera valida Loew, 1862 Tetanocera arrogans Meigen, 1830 Tetanocera melanostigma Steyskal, 1959 Elgiva connexa Steyskal, 1954 Elgiva solicita (Harris, 1780) Hedria mixta Steyskal, 1954 Limnia bosci Robineau-Desvoidy, 1830 Limnia ottawensis Melander, 1920 Limnia sandovalensis Fisher & Orth, 1978 Renocera amanda (Cresson 1920) Renocera johnsoni (Cresson, 1920) Sepedon armipes Loew, 1859 Sepedon fuscipennis Loew, 1859 Sepedon praemiosa Giglio-Tos, 1893 Trypetoptera canadensis (Macquart, 1843) Atrichomelina pubera (Loew, 1862) Sciomyza simplex Falle´n, 1820

A A A A A A A A T T T T T T T T T A A A T ? ? A A A A A T T T

AY875135 AY875137 AY875140 AY875142 AY875143 AY875146 AY875147 AY875150 AY875138 AY875148 AY875139 AY875144 AY875136 AY875145 AY875149 AY875134 AY875141 AY875122 AY875123 AY875124 AY875125 AY875126 AY875127 AY875128 AY875129 AY875130 AY875131 AY875132 AY875133 AY875120 AY875121




USA: AK: Kanai USA: OH: Portage County Canada: MB: Churchill USA: ID: Buffalo River Canada: AB; Banff NP USA: CO: Teller County USA: CO: Teller County USA: OH: Geauga County USA: CO: Teller County Germany USA: CO: Teller County Germany: Ludvika USA: OH: Muskingum County USA: CO: Teller County USA: WV: Tucker County Ireland USA Canada: MB: Churchill USA: OH Canada: MB: Churchill USA: MN: Cass County USA: SD: Custer Co. Custer SP Canada: AB; Banff NP USA: OH: Summit County USA: CO: Teller County USA: OH: Portage County USA: ND: Bottineau County USA: ID: Bear Lake County Canada: SK: Maple Creek Canada: SK: Cypress Hills Canada: AB: Banff NP

Sciomyzini

Habitat designations: A, aquatic; T, terrestrial. Park abbreviations: NP, National Park; SP, State Park. Parentheses around the author and year indicates that the species was described under a different genus name. No parentheses indicates that the species was described under the current genus.


1462


Laboratory protocols Field collections of adult specimens were preserved immediately in 95–100% nondenatured ethanol. In the lab, specimens were transferred to vials containing 100% hexamethyldisilazane (HMDS) for at least 24 h, after which the liquid was decanted, and the specimens were allowed to dry under a hood. In preparation for total DNA isolation, the head, legs, wings, and abdomen of each specimen were removed from the thorax. Total DNA was isolated from each thorax, and the remaining body parts (which contain the morphological characters necessary for species determination) were stored as vouchers in a vial containing 95–100% ethanol. Each specimen was given a unique number, and species identification, collecting locality information, and habitat notes were recorded in a database. Total DNA was isolated, using Qiagen DNeasy Tissue Kits, from each of the terminal taxa (¼sciomyzid species) in the analyses. Each of the DNA isolates was PCRamplified using the primer pairs listed in Table 2. The mitochondrial and nuclear amplicons were characterized by cycle sequencing analysis using the PCR amplification primers listed in Table 2. The protocols for sequencing template purification and cycle sequencing of the fragments are as presented in Folmer et al. (1994). These protocols include sequencing template purification in low-melting point agarose gels and cycle sequencing of both strands of each purified template using labelled primers. The separation of cycle sequencing reaction products was done in 3.7% and 5.5% polyacrylamide gels on LI-COR 4200L-2 and 4200S-2 automated DNA sequencers respectively. The resulting sequences were aligned initially using A L I G N I R (A L I G N I R V 2 . 0 , LI-COR Inc.) with subsequent refinement done manually using M A C C L A D E V . 4 . 0 (Maddison & Maddison, 2000). All sequences presented in this study have been deposited in the GenBank database (see Table 1 for accession numbers). The alignment of the COI and COII sequences were straightforward, as no indels have been detected in the sciomyzid sequences generated to date from these loci.

Gene primer pair

References

550

COI LCO1490/HCO2198 COI C1-J-2183/TL2-N-3014

Folmer et al. (1994) Simon et al. (1994)

700 800

COII TL2-J-3034/TK-N-3785

Simon et al. (1994) 800 J.B. Hobbs, UBC (personal communication) Park & O’Foighil (2000)

Phylogenetic analyses The mtDNA-encoded COI, COII, 16S and the nucleusencoded 28S sequences (nuDNA) were analysed using the maximum likelihood (ML) and maximum parsimony (MP) algorithms contained in P A U P * (v.4.0b10; Swofford, 2001). Bayesian inference (BI) analyses were carried out with M R B A Y E S V 3 . 1 (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003). The parsimony-based ILD test (Farris et al., 1994), as implemented in P A U P *, was used to test for incongruence between the mtDNA and nuDNA datasets. The COI, COII, 16S, and 28S sequences were analysed simultaneously, as recent literature indicates that a total evidence approach can produce the best tree topologies (Collin, 2003; Creer et al., 2003; Hassanin & Douzery, 2003; Schwarz et al., 2003). Thus, the total evidence-based trees were used as the best estimates of the phylogenetic relationships among Tetanocera species. M O D E L T E S T (V . 3 . 6 : Posada & Crandall, 1998) was used to determine which model best fit the concatenated sequence data. The GTR + G + I model was used in all BI and ML analyses. Atrichomelina pubera (Sciomyzini) sequences were used to root the trees. The ML algorithm in P A U P *, using the parameters from the output of M O D E L T E S T , was implemented to judge which of the 1001 trees sampled from the Bayesian analysis had the highest log likelihood value and, thus, represented the best topology. A total of 54 specimens, representing 31 species, were included in the initial BI analysis (four chains, two million generations, 50 000 generation burn-in, GTR + I + G), after which duplicate individuals for each

Amplicon size (bp) Notes

Mitochondrial loci 16S LR-N-13398/LR-J-12887 Simon et al. (1994)

Nuclear locus 28S D1F/D6R

While the 28S sequences contain some indels, we have found no regions that align ambiguously. However, indels that do present alignment ambiguities have been detected in the sciomyzid 16S sequences. Therefore, to avoid arbitrarily derived topologies, phylogenetic analyses were carried out with both (i) a ‘best’ alignment containing all of the 16S nucleotides, and (ii) an alignment with the ambiguous 16S characters deleted.

Table 2 Genes/primer information used in this study.

Primer sequences identical to those of ‘Locust’ Together, both COI primer pairs encompass nearly the entire gene Amplifies all of COII

1100



species were pruned (as all conspecific individuals formed monophyletic groups with 100% posterior probabilities). Four independent Bayesian analyses were performed to explore whether they all would arrive at the same topology. Subsequent MP and ML analyses were carried out with the pruned dataset, to investigate whether the use of other tree building algorithms resulted in congruent or conflicting topologies. Multiple random terminal taxa addition sequence runs, combined with global branch rearrangement options, were employed when generating topologies from the ML and MP algorithms. These options increased the probability of finding the actual best topology under each of these two optimality criteria (e.g. Hendy et al., 1988; Maddison, 1991). Standard bootstrap (Felsenstein, 1985b) analyses were carried out to evaluate the level of support for particular nodes obtained from the ML (100 bootstrap replicates) and MP (10 000 bootstrap replicates) analyses. Pairwise uncorrected p-distances were calculated, using P A U P *, for each gene. To test for significant differences in topologies between the best unconstrained tree and a topology produced by constraining the terrestrial Tetanocera species to be monophyletic, we used (i) P A U P * to do the parsimony-based Kishino-Hasegawa test (KH; Kishino & Hasegawa, 1989), Templeton test (Wilcoxon signed-ranks test; Templeton, 1983) and winning sites (sign) test (Prager & Wilson, 1988) and (ii) CONSEL (Shimodaira & Hasegawa, 2001) to do the likelihoodbased approximately unbiased (AU, Shimodaira, 2002), Kishino-Hasegawa (KH), Shimodaira-Hasegawa (SH; Shimodaira & Hasegawa, 1999), weighted KishinoHasegawa (WKH), and weighted Shimodaira–Hasegawa (WSH; Shimodaira, 2002) tests. The estimation of ancestral character states, based on our best estimate of phylogeny, was carried out using equally weighted parsimony methods (e.g. Scheffer & Wiegmann, 2000; Jousselin et al., 2003; Pauly et al., 2004) and with ML methods, both using M E S Q U I T E (V . 1 . 0 5 ; Maddison & Maddison, 2003). Although Tetanocera plumosa can be found in both aquatic and wet shoreline (terrestrial) habitats, it is principally aquatic (Foote, 1961) and was scored as such in all character optimization procedures. For the ML optimizations, both the ‘Markov k-state 1 parameter model’ (MK1 model in which ‘forward’ and ‘backward’ transition rates are equal) and the ‘Asymmetrical Markov k-state 2 parameter model’ (AsymmMK model in which ‘forward’ and ‘backward’ transition rates can be different) were used. The asymmetry likelihood ratio test was used to determine whether the AsymmMK model was significantly better than the MK1 model. Tests of correlated evolution The maximum-likelihood program D I S C R E T E (Pagel, 1994,1997,1999a,b) was used to test for correlated evolution between ecological and morphological characters (omnibus test). This test utilizes a Markov model in a

1463

Fig. 2 Flow diagram showing how habitat [aquatic (AQ) and terrestrial (T)] and a given morphological character [morphology typical of aquatic species (Maq) and morphology typical of terrestrial species (MT)] may evolve in a correlated fashion. The eight qij values are ‘forward’ and ‘backward’ transition rate parameters estimated from the data. It is assumed that both traits do not change at the same time (i.e. q14, q41, q23, and q32 all equal zero).

ML framework, taking branch length information into account, but does not rely on ancestral character state reconstruction. Given a pair of binary characters and a tree topology, the program calculates the log-likelihoods for two models: (i) a model in which the two characters are allowed to evolve independently (independent model) and (ii) a model in which the two characters evolve in a correlated manner (dependent model: Fig. 2). A Monte Carlo simulation study, in which character states are repeatedly assigned independently and randomly to the terminals of the tree, approximates the null hypothesis distribution for the characters at hand. The outcome of the simulation is used to determine whether the independent or dependent model of character evolution best fits the data, via a likelihood ratio test. If the dependent model (correlated evolution model) fits the data significantly better than the independent model, then the null hypothesis that the two traits evolved independently is rejected. The omnibus test was also used to test for correlated evolution among morphological characters. When evidence for correlated evolution was found, D I S C R E T E was used to test whether a given trait changes from state 0 to state 1 before the other (temporal order test). This is also tested via a likelihood ratio test that compares the log-likelihood of the full eightparameter (dependent) model to that of a seven-parameter model in which q12 and q13 are set to be equal to one another (see Fig. 2). In other words, if two times the difference in log-likelihoods between the two models is larger than 3.84 (v2 value with one degree of freedom; a ¼ 0.05), then the null hypothesis that neither character tends to change before the other can be rejected. Rejecting the null hypothesis of the omnibus test, combined with failure to reject the null hypothesis of


1464


the temporal order test indicates that while the two traits are evolving in a correlated fashion, it is not possible to tell which trait changes first. When two morphological traits are being tested, this result indicates that the two traits are linked by (i) pleiotropy, (ii) indirect selection through another trait or (iii) concerted evolution (Armbruster, 2002), but other methods are necessary to distinguish between these. We tested whether there were significant correlations among all possible combinations of larval habitat and the four morphological characters outlined above. These tests were conducted on both the entire phylogeny and the Tetanocera portion only. As Tetanocera plumosa can be found in both habitats, and the spiracular disc orientation of T. silvatica could not be determined, all possible scenarios involving these taxa were tested. All variables were scored as discrete binary characters as there is a clear distinction between each scored character’s character states; [coding: larvae with pigmented cuticle ¼ 0, unpigmented larvae ¼ 1; float hairs long (extending well beyond the base of the spiracular tube), branched, conspicuous ¼ 0, short (not or barely extending beyond the base of the spiracular tube), unbranched ¼ 1; ventrolateral lobes longer than width at base ¼ 0, shorter/ equal to basal width ¼ 1; posterior spiracular disc upturned ¼ 0, posterior spiracular disc not upturned ¼ 1]. Larval character states and habitat designations were obtained from Foote (1959,1961,1971,1976, unpublished data), Foote et al. (1960), Knutson (1963), Knutson & Berg (1964), Neff & Berg (1966), and Knutson & Vala (2002). The sequential Bonferroni technique (Rice, 1989) was used to minimize type I statistical error, the likelihood of which is increased when performing multiple statistical tests (also see Holm, 1979).

Results Phylogenetic analyses Pairwise uncorrected p-distances for 16S and 28S rDNA sequences are given in Table 3, and those for COI and COII are given in Table 4. The tree topologies obtained from the three phylogenetic analyses of the concatenated dataset were largely congruent, with the BI tree fully resolved (Fig. 3) and the MP and ML bootstrap trees (not shown) recovering the same major clades as indicated in the BI analysis (albeit with lower intra-clade resolution). All four independent BI analyses converged on the same topology. The best BI (as judged by ML) and ML trees were identical (Ln L ¼ )37275.65). The MP analysis of 1217 parsimony-informative characters in the concatenated dataset, produced two equally parsimonious trees (Ln L ¼ )37322.21 and )37323.07 as judged by ML) which had identical topologies within Tetanocera, only differing from the BI and ML trees in the placement of T. latifibula and T. mesopora: In the MP trees, T. latifibula is sister to T. montana + T. kerteszi, and T. mesopora is sister to

a clade comprised of the three aforementioned species, whereas T. latifibula and T. mesopora are sister taxa on the BI and ML trees. As T. kerteszi is the only terrestrial species among these four (L.V. Knutson, personal communication), either topology would give nearly identical results with respect to both character optimizations and correlated evolution tests. Concatenating the nuDNA and mtDNA sequences was legitimized by the lack of significant incongruence between the datasets (as indicated by the ILD test, P ¼ 0.974). In the BI analysis, of the 29 internal nodes, 23 were supported by posterior probabilities ‡0.90, three were between 0.80 and 0.89, and there was one each in the 0.70, 0.60 and 0.50s (lowest pp ¼ 0.59). Tetanocera is clearly supported as a monophyletic group based on very high nodal support values (BI posterior probability ¼ 1.00; ML bootstrap percentage ¼ 100; MP bootstrap percentage ¼ 90; Fig. 3). Within Tetanocera, three well-supported subclades were inferred, each containing at least one aquatic and one terrestrial species (Fig. 3 and Table 1). The robusta-silvatica clade (BI pp ¼ 1.00, ML bootstrap percentage ¼ 100, MP bootstrap percentage ¼ 100) is sister to the remaining Tetanocera species. It contains two species: one aquatic and one terrestrial. The plumosa-mesopora clade (BI pp ¼ 1.00, ML bootstrap percentage ¼ 82, MP bootstrap percentage ¼ 65) contains four aquatic, one terrestrial, and one species (T. plumosa) that can be found both in water and on shorelines (Foote, 1961). The phyllophora-valida clade (BI pp ¼ 1.00, ML bootstrap percentage ¼ 97, MP bootstrap percentage ¼ 74) contains two aquatic and seven terrestrial species. Larval habitats (aquatic vs. terrestrial) of the sciomyzid species were mapped onto the BI topology using ML (Fig. 4) and parsimony (Fig. 5) methods. The asymmetry likelihood ratio test showed that the AsymmMK model was not significantly better than the MK1 model. Of the 25 internal nodes on the tree, 19 had ancestral states that were statistically significant as judged by the AsymmMk optimization model (nodes with an asterisk on Fig. 4a), and 13 had ancestral states that were statistically significant as judged by the MK1 optimization model (nodes with an asterisk on Fig. 4b). From this analysis, it can be inferred that the ancestral larvae of Tetanocera was very likely aquatic (AsymmMk ML estimate ¼ 0.97; MK1 ML estimate ¼ 0.82). From this ancestral condition, there were at least three independent transitions to terrestrial existence, with one reversal (also see parsimony optimization in Fig. 5). Since the evidence supporting a terrestrial ancestral state for the phyllophora-valida clade is relatively weak (AsymmMk ML estimate ¼ 0.54; MK1 ML estimate ¼ 0.64) as it is for the subsequent node after T. phyllophora arose (AsymmMk ML estimate ¼ 0.62; MK1 ML estimate ¼ 0.72), it is possible that the ancestor of the phyllophora-valida clade was, in fact, aquatic. If this were true, then terrestrial existence arose five times independently within Tetanocera, with no reversals.



0.139 0.134 0.112 0.149 0.147 0.118 0.135

0.139 0.140 0.142 0.163 0.149 0.118 0.137

0.155 0.135 0.158 0.159 0.149 0.129 0.142

0.140 0.136 0.140 0.158 0.148 0.131 0.145

0.149 0.144 0.172 0.126 0.142 0.159 0.151 0.154 0.142 0.162 0.145 0.144 0.159

0.147 0.150 0.162 0.146 0.152 0.165 0.168 0.157 0.159 0.187 0.172 0.161 0.168

0.161 0.169 0.177 0.149 0.162 0.172

0.122 0.143 0.161 0.145 0.083 0.174 0.169

0.151 0.154 0.142 0.162 0.145 0.144 0.159

0.147 0.150 0.162 0.146 0.152 0.165

0.112 0.158 0.139 0.168 0.118 0.154 0.150

0.135 0.155 0.139 0.144 0.141 0.133 0.139

0.152 0.148 0.159 0.143 0.145 0.162

0.124 0.160 0.163 0.172 0.128 0.164 0.150

0.051 0.039 0.039 0.049 0.039 –

0.130 0.137 0.123 0.138 0.156 0.129 0.158

0.145 0.143 0.153 0.129 0.134 0.159

– 0.135 0.141 0.143 0.133 0.146 0.143

0.048 0.033 0.030 0.042 0.030 0.043

0.160 0.153 0.173 0.180 0.171 0.159 0.158

0.188 0.161 0.171 0.149 0.163 0.198

0.043 – 0.104 0.091 0.146 0.180 0.163

0.037 0.035 0.033 0.041 0.033 0.049

0.153 0.158 0.164 0.185 0.172 0.164 0.176

0.166 0.152 0.156 0.154 0.164 0.195

0.046 0.024 – 0.094 0.145 0.176 0.152

0.038 0.032 0.031 0.037 0.031 0.041

0.146 0.146 0.163 0.187 0.168 0.161 0.185

0.179 0.154 0.164 0.142 0.164 0.191

0.044 0.004 0.021 – 0.133 0.185 0.156

0.036 0.032 0.032 0.038 0.032 0.047

0.148 0.154 0.145 0.174 0.164 0.153 0.164

0.147 0.151 0.175 0.146 0.154 0.176

0.028 0.030 0.029 0.030 – 0.161 0.151

0.033 0.007 0.005 0.023 0.005 0.037

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

0.130 0.119 0.131 0.136 0.153 0.113 0.157

0.112 0.129 0.151 0.118 0.133 0.161

0.032 0.029 0.027 0.026 0.020 – 0.127

0.032 0.020 0.023 0.018 0.023 0.037

0.115 0.120 0.093 0.117 0.144 0.120 0.138

0.116 0.002 0.108 0.114 0.122 0.138

0.031 0.032 0.030 0.030 0.019 0.010 –

0.030 0.018 0.022 0.016 0.022 0.039

0.115 0.124 0.130 0.134 0.131 0.118 0.156

– 0.116 0.120 0.118 0.131 0.145

0.034 0.034 0.032 0.031 0.022 0.011 0.014

0.036 0.021 0.025 0.019 0.025 0.038

0.114 0.118 0.091 0.118 0.142 0.120 0.136

0.010 – 0.106 0.112 0.120 0.139

0.029 0.030 0.027 0.027 0.015 0.007 0.004

0.027 0.014 0.018 0.015 0.018 0.035

0.117 0.126 0.112 0.125 0.150 0.114 0.142

0.014 0.007 – 0.120 0.122 0.133

0.031 0.031 0.029 0.030 0.016 0.011 0.010

0.029 0.015 0.019 0.015 0.019 0.037

0.081 0.018 0.114 0.123 0.107 0.104 0.118

0.021 0.012 0.015 – 0.085 0.137

0.032 0.034 0.029 0.032 0.020 0.015 0.016

0.032 0.021 0.023 0.025 0.023 0.037

0.089 0.089 0.114 0.134 0.125 0.098 0.123

0.016 0.008 0.010 0.009 – 0.141

0.030 0.032 0.029 0.029 0.017 0.012 0.011

0.031 0.018 0.020 0.019 0.020 0.037

0.143 0.143 0.124 0.150 0.162 0.126 0.156

0.015 0.014 0.018 0.024 0.016 –

0.038 0.034 0.031 0.032 0.027 0.013 0.018

0.039 0.026 0.029 0.025 0.029 0.045

– 0.093 0.104 0.110 0.105 0.095 0.109

0.014 0.006 0.009 0.007 0.003 0.017

0.028 0.030 0.027 0.027 0.015 0.010 0.010

0.030 0.015 0.018 0.017 0.018 0.035

0.009 – 0.120 0.129 0.113 0.102 0.132

0.024 0.015 0.018 0.003 0.012 0.027

0.035 0.037 0.032 0.034 0.023 0.017 0.019

0.035 0.024 0.026 0.027 0.026 0.040

0.014 0.023 – 0.119 0.140 0.108 0.130

0.017 0.010 0.014 0.020 0.016 0.019

0.036 0.036 0.032 0.034 0.023 0.013 0.012

0.036 0.023 0.026 0.019 0.026 0.042

0.010 0.016 0.011 – 0.142 0.131 0.144

0.011 0.007 0.011 0.013 0.012 0.014

0.034 0.028 0.024 0.025 0.020 0.008 0.010

0.031 0.019 0.023 0.016 0.023 0.036

0.002 0.011 0.013 0.009 – 0.123 0.139

0.012 0.004 0.009 0.009 0.005 0.016

0.027 0.028 0.024 0.024 0.014 0.008 0.008

0.028 0.015 0.017 0.017 0.017 0.032

0.012 0.021 0.015 0.014 0.011 – 0.114

0.018 0.010 0.015 0.018 0.014 0.022

0.033 0.035 0.031 0.033 0.022 0.015 0.012

0.034 0.023 0.025 0.019 0.025 0.040

0.011 0.021 0.019 0.017 0.012 0.017 –

0.020 0.011 0.016 0.018 0.013 0.024

0.027 0.034 0.030 0.033 0.017 0.018 0.013

0.035 0.022 0.019 0.023 0.019 0.040

0.018 0.028 0.020 0.017 0.017 0.021 0.024

0.013 0.014 0.019 0.025 0.020 0.020

0.040 0.038 0.035 0.037 0.025 0.017 0.018

0.040 0.026 0.027 0.026 0.027 0.045

0.004 0.013 0.016 0.011 0.002 0.014 0.015

0.015 0.007 0.011 0.010 0.007 0.017

0.029 0.031 0.025 0.027 0.017 0.010 0.010

0.031 0.018 0.020 0.019 0.020 0.035

0.135 0.131 0.157 0.160 0.153 0.161 0.174 0.161 0.154 0.131 0.171 0.171 0.170 0.165 0.132 0.111 0.104 0.111 0.124 0.108 0.092 0.143 0.102 0.116 0.103 0.119 0.136 0.117 0.138 – 0.018 0.137 0.135 0.146 0.143 0.156 0.146 0.172 0.146 0.153 0.148 0.162 0.170 0.172 0.162 0.142 0.139 0.123 0.137 0.138 0.103 0.123 0.153 0.125 0.111 0.141 0.150 0.079 0.102 0.133 0.122 –

0.139 0.126 0.112 0.139 0.151 0.118 0.147

0.160 0.142 0.159 0.125 0.145 0.161

0.112 0.158 0.139 0.168 0.118 0.154 0.150

Lower left, 16S rDNA p-distances; upper right, 28S p-distances.

30. Tetanocera valida 31. Tetanocera vicina

mesopora montana phyllophora plebeja plumosa robusta silvatica

0.156 0.136 0.134 0.136 0.140 0.143

0.118 0.135 0.136 0.137 0.107 0.147 0.144

0.036 0.009 0.000 0.025 – 0.114

13

Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera

0.138 0.126 0.138 0.120 0.129 0.142

0.114 0.136 0.128 0.148 0.122 0.156 0.142

0.038 0.021 0.025 – 0.106 0.136

12

23. 24. 25. 26. 27. 28. 29.

0.138 0.112 0.130 0.116 0.133 0.139

0.131 0.132 0.135 0.134 0.145 0.169 0.136

0.036 0.009 – 0.106 0.000 0.114

11

Trypetoptera clara Trypetoptera ferruginea Tetanocera fuscinervis Tetanocera kerteszi Tetanocera latifibula Tetanocera melanostigma

0.096 0.126 0.133 0.132 0.122 0.136 0.126

0.032 – 0.086 0.093 0.086 0.102

10

17. 18. 19. 20. 21. 22.

0.092 0.128 0.125 0.134 0.132 0.138 0.112

– 0.096 0.113 0.127 0.113 0.118

9

Renocera johnsoni Sepedon armipes Sepedon fuscipennis Sepedon praemiosa Trypetoptera canadensis Trypetoptera arrogans Trypetoptera bergi

0.110 0.107 0.133 0.137 0.133 0.114

8

10. 11. 12. 13. 14. 15. 16.

0.100 0.078 0.110 0.118 0.110 0.105

7

0.098 0.090 0.108 0.132 0.108 0.117

6

Elgiva solicita Hedria mixta Limnia bosci Limnia ottawensis Limnia sandovalensis Renocera amanda

5

4. 5. 6. 7. 8. 9.

4

– 0.020 0.052 0.059 0.054 0.054 0.058 0.054 0.065 0.063 0.057 0.054 0.056 0.051 0.054 0.053 0.055 0.050 0.053 0.052 0.048 0.057 0.047 0.054 0.059 0.051 0.048 0.052 0.056 0.060 0.049 0.033 – 0.046 0.054 0.050 0.048 0.057 0.048 0.062 0.055 0.053 0.053 0.050 0.048 0.051 0.049 0.053 0.045 0.050 0.050 0.045 0.054 0.044 0.052 0.055 0.050 0.044 0.049 0.051 0.054 0.045 0.096 0.104 – 0.020 0.020 0.022 0.025 0.022 0.040 0.036 0.032 0.030 0.029 0.019 0.020 0.020 0.022 0.016 0.019 0.021 0.020 0.028 0.018 0.024 0.025 0.019 0.015 0.024 0.024 0.027 0.018

3

1. Atrichomelina pubera 2. Sciomyza simplex 3. Elgiva connexa

2

1

Species

Table 3 Pair-wise uncorrected p-distances between sciomyzid species appearing in Table 1.

Habitat correlated parallel evolution 1465

0.183 0.212 0.204 0.177 0.184 0.188

0.198 0.220 0.199 0.234 0.201 0.204 0.218

8. Limnia sandovalensis 9. Renocera amanda 10. Renocera johnsoni 11. Sepedon armipes 12. Sepedon fuscipennis 13. Sepedon praemiosa

Trypetoptera canadensis Trypetoptera arrogans Trypetoptera bergi Trypetoptera clara Trypetoptera ferruginea Tetanocera fuscinervis Tetanocera kerteszi

Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera

Tetanocera Tetanocera Tetanocera Tetanocera Tetanocera

14. 15. 16. 17. 18. 19. 20.

21. 22. 23. 24. 25. 26.

27. 28. 29. 30. 31.

0.197 0.174 0.185 0.219 0.190

0.186 0.166 0.172 0.195 0.177

0.204 0.243 0.197 0.211 0.204 0.202

0.186 0.208 0.195 0.213 0.192 0.193 0.213

0.176 0.201 0.191 0.177 0.161 0.184

0.114 – 0.178 0.163 0.180 0.173 0.159

2

0.182 0.165 0.170 0.200 0.176

0.186 0.234 0.188 0.185 0.192 0.199

0.191 0.212 0.188 0.213 0.189 0.193 0.193

0.187 0.210 0.197 0.186 0.187 0.190

0.188 0.208 – 0.168 0.178 0.183 0.180

3

0.182 0.162 0.170 0.196 0.181

0.185 0.230 0.181 0.193 0.188 0.216

0.179 0.214 0.178 0.203 0.179 0.188 0.193

0.184 0.210 0.184 0.179 0.173 0.179

0.169 0.188 0.182 – 0.176 0.184 0.156

4

0.191 0.177 0.179 0.200 0.184

0.183 0.231 0.187 0.194 0.205 0.198

0.187 0.217 0.191 0.207 0.194 0.196 0.194

0.185 0.209 0.205 0.199 0.198 0.200

0.235 0.235 0.202 0.200 – 0.184 0.173

5

Lower left, COI p-distances; upper right, COII p-distances.

plumosa robusta silvatica valida vicina

0.204 0.258 0.190 0.214 0.212 0.219

– 0.122 0.172 0.166 0.196 0.181 0.168

Atrichomelina pubera Sciomyza simplex Elgiva connexa Elgiva solicita Hedria mixta Limnia bosci Limnia ottawensis

1. 2. 3. 4. 5. 6. 7.

latifibula melanostigma mesopora montana phyllophora plebeja

1

Species

0.188 0.177 0.188 0.200 0.191

0.192 0.234 0.198 0.211 0.195 0.188

0.180 0.224 0.199 0.221 0.197 0.185 0.210

0.005 0.224 0.190 0.189 0.195 0.193

0.197 0.215 0.196 0.202 0.211 – 0.165

6

0.181 0.157 0.161 0.191 0.174

0.175 0.240 0.180 0.191 0.182 0.199

0.151 0.205 0.192 0.211 0.192 0.183 0.200

0.168 0.200 0.184 0.156 0.161 0.161

0.197 0.208 0.202 0.182 0.190 0.193 –

7

0.190 0.179 0.190 0.200 0.194

0.191 0.238 0.197 0.212 0.199 0.192

0.180 0.221 0.198 0.221 0.196 0.185 0.213

– 0.223 0.190 0.194 0.194 0.195

0.199 0.214 0.194 0.203 0.209 0.001 0.191

8

0.213 0.207 0.213 0.227 0.205

0.206 0.262 0.207 0.219 0.223 0.238

0.225 0.236 0.217 0.231 0.220 0.228 0.217

0.239 – 0.227 0.212 0.212 0.215

0.234 0.230 0.236 0.238 0.263 0.241 0.222

9

0.192 0.188 0.186 0.192 0.190

0.194 0.224 0.189 0.201 0.190 0.217

0.202 0.227 0.188 0.189 0.189 0.210 0.203

0.212 0.238 – 0.195 0.194 0.200

0.235 0.232 0.223 0.221 0.233 0.214 0.212

10

0.191 0.181 0.186 0.219 0.189

0.193 0.258 0.194 0.208 0.190 0.217

0.204 0.227 0.199 0.223 0.200 0.200 0.211

0.217 0.229 0.220 – 0.151 0.147

0.197 0.200 0.208 0.181 0.212 0.215 0.179

11

0.188 0.175 0.190 0.195 0.184

0.193 0.243 0.186 0.196 0.186 0.214

0.198 0.211 0.170 0.204 0.173 0.188 0.198

0.221 0.220 0.232 0.182 – 0.160

0.191 0.215 0.214 0.172 0.227 0.223 0.199

12

0.193 0.188 0.184 0.216 0.186

0.203 0.253 0.197 0.215 0.204 0.221

0.198 0.225 0.194 0.228 0.196 0.201 0.218

0.202 0.210 0.226 0.167 0.184 –

0.187 0.218 0.181 0.179 0.221 0.203 0.196

13

Table 4 Pairwise uncorrected p-distances between sciomyzid species appearing in Table 1.

0.204 0.189 0.187 0.216 0.202

0.204 0.242 0.204 0.212 0.203 0.216

– 0.216 0.208 0.221 0.211 0.204 0.219

0.203 0.245 0.227 0.203 0.205 0.233

0.223 0.236 0.206 0.184 0.184 0.202 0.142

14

0.216 0.221 0.203 0.207 0.209

0.206 0.213 0.216 0.210 0.216 0.200

0.244 – 0.207 0.205 0.210 0.203 0.218

0.261 0.270 0.261 0.259 0.253 0.256

0.270 0.265 0.235 0.252 0.259 0.259 0.249

15

0.170 0.181 0.180 0.194 0.177

0.168 0.229 0.181 0.181 0.181 0.188

0.221 0.238 – 0.210 0.011 0.177 0.197

0.236 0.236 0.232 0.223 0.230 0.232

0.226 0.229 0.220 0.215 0.236 0.235 0.226

16

0.201 0.191 0.208 0.195 0.202

0.189 0.217 0.199 0.188 0.196 0.210

0.269 0.262 0.245 – 0.210 0.205 0.191

0.253 0.248 0.242 0.272 0.260 0.254

0.257 0.271 0.218 0.251 0.262 0.251 0.254

17

0.171 0.180 0.179 0.195 0.177

0.172 0.229 0.179 0.179 0.181 0.188

0.221 0.240 0.010 0.241 – 0.177 0.197

0.241 0.235 0.238 0.227 0.235 0.235

0.232 0.235 0.218 0.220 0.230 0.239 0.227

18

0.192 0.174 0.179 0.190 0.196

0.187 0.226 0.186 0.190 0.190 0.192

0.250 0.280 0.208 0.245 0.203 – 0.187

0.247 0.254 0.244 0.221 0.236 0.230

0.248 0.268 0.226 0.227 0.268 0.248 0.221

19

0.186 0.181 0.179 0.196 0.190

0.168 0.224 0.176 0.102 0.191 0.206

0.244 0.261 0.210 0.240 0.213 0.225 –

0.253 0.234 0.242 0.237 0.226 0.229

0.237 0.246 0.207 0.229 0.243 0.255 0.225

20

0.163 0.168 0.185 0.201 0.162

– 0.226 0.162 0.168 0.174 0.199

0.193 0.252 0.194 0.245 0.197 0.212 0.198

0.223 0.208 0.242 0.230 0.216 0.224

0.220 0.235 0.197 0.203 0.236 0.224 0.185

21

0.227 0.226 0.221 0.223 0.220

0.235 – 0.234 0.227 0.229 0.228

0.272 0.271 0.260 0.257 0.260 0.226 0.245

0.262 0.256 0.272 0.236 0.247 0.265

0.259 0.268 0.260 0.244 0.259 0.263 0.256

22

0.162 0.167 0.175 0.186 0.174

0.161 0.239 – 0.167 0.170 0.204

0.211 0.249 0.214 0.223 0.220 0.229 0.199

0.230 0.215 0.224 0.200 0.193 0.188

0.209 0.224 0.185 0.200 0.230 0.232 0.190

23

0.174 0.179 0.184 0.187 0.177

0.196 0.250 0.201 – 0.192 0.206

0.237 0.263 0.235 0.249 0.236 0.226 0.136

0.268 0.228 0.254 0.228 0.211 0.232

0.237 0.251 0.234 0.222 0.259 0.269 0.211

24

0.182 0.172 0.184 0.188 0.181

0.211 0.253 0.202 0.217 – 0.195

0.238 0.264 0.220 0.245 0.225 0.204 0.225

0.240 0.244 0.245 0.208 0.241 0.240

0.238 0.259 0.231 0.229 0.249 0.241 0.220

25

0.199 0.199 0.201 0.195 0.209

0.238 0.275 0.246 0.269 0.250 –

0.246 0.248 0.211 0.253 0.205 0.251 0.246

0.252 0.266 0.265 0.250 0.234 0.264

0.238 0.268 0.240 0.249 0.261 0.253 0.267

26

– 0.172 0.179 0.194 0.151

0.166 0.259 0.166 0.190 0.198 0.241

0.208 0.245 0.182 0.244 0.190 0.193 0.199

0.224 0.242 0.224 0.202 0.203 0.209

0.205 0.209 0.196 0.196 0.230 0.226 0.179

27

0.170 – 0.159 0.195 0.176

0.187 0.232 0.203 0.211 0.214 0.249

0.197 0.253 0.215 0.250 0.218 0.206 0.208

0.227 0.226 0.211 0.200 0.185 0.203

0.184 0.194 0.199 0.182 0.212 0.229 0.172

28

0.178 0.208 – 0.195 0.175

0.196 0.262 0.188 0.210 0.195 0.234

0.214 0.250 0.193 0.251 0.202 0.232 0.223

0.256 0.251 0.232 0.208 0.235 0.221

0.233 0.242 0.211 0.226 0.256 0.254 0.211

29

0.245 0.221 0.227 – 0.186

0.238 0.236 0.232 0.256 0.239 0.269

0.251 0.282 0.236 0.221 0.242 0.250 0.244

0.263 0.247 0.254 0.245 0.262 0.256

0.274 0.265 0.244 0.244 0.253 0.265 0.233

30

0.170 0.166 0.199 0.239 –

0.191 0.253 0.178 0.222 0.213 0.256

0.206 0.274 0.208 0.247 0.205 0.196 0.223

0.238 0.247 0.230 0.221 0.215 0.217

0.235 0.233 0.224 0.218 0.230 0.239 0.193

31

1466 E. G. CHAPMAN ET AL.



1467

Fig. 3 Best tree topology (as judged by ML) from a sample of 1001 trees from a Bayesian analysis of the concatenated dataset, showing posterior probabilities above the lines and ML bootstrap and MP bootstrap values below the lines, the latter in parentheses. Numbers in parentheses after the taxon names indicate how many specimens of each species were sequenced and analysed. Genus abbreviations from top to bottom: Sciomyzini: Atrichomelina (A); Sciomyza (S); Tetanocerini: Sepedon (S); Elgiva (E); Hedria (H); Limnia (L); Trypetoptera (T); Renocera (R); Tetanocera (T).

To further test the robustness of our phylogenetic inferences, we performed a variety of topology tests comparing the best unconstrained tree topology with a tree in which the terrestrial Tetanocera were constrained to be monophyletic. The best ML constrained tree was identical to the unconstrained ML topology, except for the arrangement of Tetanocera species (tree not shown). All tests strongly rejected the null hypothesis that the difference between the two trees was no greater than expected from sampling error (P < 0.0001 for all; Table 5). Habitat-morphology correlations Figure 5 shows four mirror trees visualizing the correlations between larval habitat (left side of each tree) and

each of four morphological characters: (i) larval colour (Fig. 5a), (ii) float hair length (Fig. 5b), (iii) ventrolateral lobe length ratios (Fig. 5c), (iv) and orientation of the posterior spiracular disc (Fig. 5d), each on the right side of their respective trees. The parsimony-inferred ancestral character states for Tetanocera are (i) aquatic habitat, (ii) dark colour, (iii) long float hairs, (iv) relatively long ventrolateral lobes and (v) lengthened last abdominal segment with upturned spiracular disc (Fig. 5; the orientation of the spiracular disc in Tetanocera silvatica could not be determined, therefore was coded as unknown in Fig. 5d). Thus, the initial direction of change for the polarized characters within Tetanocera was from (i) aquatic to terrestrial habitat, (ii) dark to light colour, (iii) long to short float hairs, (iv) from relatively long to


1468


Fig. 4 Maximum likelihood optimizations of habitat analysed with M E S Q U I T E : (a) asymmetrical two-parameter model, (b) MK1 model. An asterisk (*) to the left of a node indicates significance (ancestral states with raw likelihood scores >2.0).

relatively short ventrolateral lobes and (v) from upturned to rear-facing spiracular disc. The trees in Fig. 5 illustrate the relatively tight correlations we observed between habitat and morphology. Dark colour is perfectly correlated with aquatic habitat within Tetanocera; the outgroup genus Sepedon is the exception (Fig. 5a). The only exception to a perfect correlation between aquatic habitat and long float hairs occurs if Tetanocera plumosa (Fig. 5b) is considered aquatic. There is a perfect correlation between aquatic habitat and relatively long ventrolateral lobes throughout the entire phylogeny (Fig. 5c). Finally, the orientation of the spiracular disc has two mismatches among the outgroup genera (Hedria and Renocera) and a single mismatch within Tetanocera (T. fuscinervis; Fig. 5d). The correlated evolution tests between habitat and the above four morphological characters all yielded significant correlations (P £ 0.01), as did tests of correlations among morphological characters (P £ 0.02; Table 6). All correlations were significant, and all remained significant after applying the sequential Bonferroni test. Furthermore, of the 18 possible temporal order tests among the four morphological characters (coding T. silvatica with or without an upturned spiracular disc, and examining both the entire phylogeny and the Tetanocera portion), only one was significant: cuticular pigmentation is reduced before float hair length is shortened on the entire phylogeny (v21 ¼ 4.76, P ¼ 0.029). Of the 20 possible temporal order tests between habitat and the four morphological characters (coding T. silvatica with or without an upturned spiracular disc, coding T. plumosa aquatic or terrestrial, and examining both the entire phylogeny and the Tetanocera portion), only one was significant: habitat changes from aquatic to terrestrial before cuticular pigmentation is reduced on the entire

phylogeny if T. plumosa is coded as terrestrial (v21 ¼ 5.17, P ¼ 0.023). Neither of these results remained significant after applying the sequential Bonferroni test.

Discussion The results of our phylogenetic and correlated evolution analyses indicate that larval habitat changes are robustly linked to larval pigmentation, float hair length, spiracular disc orientation, and ventrolateral lobe length. These analyses are consistent with the hypothesis that terrestrial habitat preference and the morphological character states typical of terrestrial Tetanocera species all arose at least three times independently (Figs 3 and 4). Knutson & Vala (2002) conclude ‘terrestrial behaviour and morphology in the Tetanocerini are apomorphic features of that tribe’. Our species-level phylogeny, ancestral character state optimizations, and topology tests clearly support this conclusion with respect to Tetanocera. Habitat–morphology correlations Because our optimization analyses (Figs 3 and 4) indicate at least three independent transitions from aquatic to terrestrial larval habitat within Tetanocera, and all correlated evolution tests were significant, we are able to infer that changes in habitat preference were typically accompanied by morphological transitions in four larval characters. All four of these morphological character state changes can be viewed as transitions to more vestigial conditions, as they are either a reduction of cuticular pigmentation or a reduction in the size of the structure. In his comprehensive work on the breeding habits and immature stages of cyclorrhaphan Diptera (the division to which the Sciomyzidae belongs), Ferrar (1987) stated that



1469

Fig. 5 Mirror trees showing the correlations between habitat and (a) larval colour, (b) float hair length, (c) ventrolateral lobe ratios and (d) orientation of the posterior spiracular disc. Character optimizations done using equally weighted parsimony methods in M E S Q U I T E .

Table 5 Results of the parsimony-based Kishino-Hasegawa (KH), Templeton (Wilcoxon signed-ranks) and winning sites (sign) tests calculated using P A U P *, and the likelihood-based approximately unbiased (AU), Kishino–Hasegawa (KH), Shimodiara-Hasegawa (SH), weighted KishinoHasegawa (WKH), and weighted Shimodiara-Hasegawa (WSH) tests calculated using CONSEL. Trees compared were the best topology from the unconstrained Bayesian analysis vs. an analysis where the terrestrial Tetanocera were constrained to be monophyletic. Test Tree

Length

Difference

KH

Templeton

Winning sites

Parsimony-based tests Unconstrained Terrestrials constrained

8506 8719

213

P < 0.0001

P < 0.0001

P < 0.0001

)Ln L

Difference

AU

KH

SH

WKH

WSH

)37275.65 )37587.04

311.39

P ¼ 3e)22

P¼0

P¼0

P¼0

P¼0

Likelihood-based tests Unconstrained Terrestrials constrained


1470


Table 6 Tests for correlated evolution using

DISCRETE

(omnibus test) showing likelihood ratios (LR) and associated P-values for each test.

Habitat–morphology correlations

Habitat vs.

Scenario

Pigmentation

Float hair length

Ventrolateral lobe length

Spiracular disc orientation

LR ¼ 10.40, P < 10)6 LR ¼ 10.59, P < 10)6

LR ¼ 13.77, P < 10)6 LR ¼ 8.09, P < 10)6

LR ¼ 13.44, P < 10)6 LR ¼ 10.59, P < 10)6

LR ¼ 10.11, P < 10)6 LR ¼ 8.06, P < 10)6

LR ¼ 6.86, P < 10)6 LR ¼ 8.02, P < 10)6

LR ¼ 16.21, P < 10)6 LR ¼ 9.86, P < 10)6

LR ¼ 10.42, P < 10)6 LR ¼ 8.11, P < 10)6

LR ¼ 8.30, P < 10)6 LR ¼ 6.57, P < 10)6

N/A N/A

N/A N/A

N/A N/A

LR ¼ 6.70, P £ 0.01 LR ¼ 5.34, P £ 0.01

N/A N/A

N/A N/A

N/A N/A

LR ¼ 8.30, P < 10)6 LR ¼ 6.37, P < 10)6

Pigmentation

Float hair length


LR ¼ 10.46, P < 10)6 LR ¼ 8.09, P < 10)6

– –

– –

LR ¼ 11.62, P < 10)6 LR ¼ 10.43, P < 10)6

LR ¼ 10.42, P < 10)6 LR ¼ 8.11, P < 10)6

– –

LR ¼ 6.03, P < 10)6 LR ¼ 6.37, P < 10)6

LR ¼ 6.70, P < 10)6 LR ¼ 5.33, P £ 0.02

LR ¼ 8.38, P < 10)6 LR ¼ 6.38, P £ 0.01

LR ¼ 8.33, P < 10)6 LR ¼ 8.08, P < 10)6

LR ¼ 8.30, P < 10)6 LR ¼ 6.57, P < 10)6

LR ¼ 8.38, P < 10)6 LR ¼ 8.07, P < 10)6

Tetanocera plumosa coded aquatic Entire phylogeny Tetanocera portion only T. plumosa coded terrestrial Entire phylogeny Tetanocera portion only T. silvatica with upturned disc Entire phylogeny Tetanocera portion only T. silvatica with rear-facing disc Entire phylogeny Tetanocera portion only Morphology–morphology correlations Scenario Float hair length Entire phylogeny Tetanocera portion only Ventrolateral lobe length Entire phylogeny Tetanocera portion only Spiracular disc orientation T. silvatica with upturned disc Entire phylogeny Tetanocera portion only T. silvatica with rear-facing disc Entire phylogeny Tetanocera portion only

Top half, correlation tests between habitat (aquatic or terrestrial) and four morphological characters under a number of different possible scenarios. Correlated evolution tests were conducted on both the entire phylogeny and the Tetanocera portion only. Because T. plumosa can be found in both aquatic and shoreline habitats, both scenarios were tested. Because the spiracular disc orientation of T. silvatica could not be determined, both character states were tested. Each test was repeated to insure accuracy; Bottom half, correlation tests between morphological characters.

‘I still believe that to some extent one can draw conclusions on [phylogenetic] relationship from immature stages, but I now consider that larval morphology is predominantly functional and that larvae show a number of interesting examples of parallel evolution’. Our analyses of Tetanocera phylogenesis and character evolution, which quite strikingly corroborate Ferrar’s conclusion, indicate at least three instances where larval morphology changed in the same way during habitat shifts. Table 7 lists the character states of the Tetanocera species with known habitat preferences that we were unable to include in the analyses presented herein. Upon examination of these data, it is clear that no matter where these species eventually insert into the phylogeny of Tetanocera, they will only render our highly significant habitat–morphology correlations stronger. Furthermore, given the high nodal support values for the three Tetanocera subclades and the highly significant results of the constrained/unconstrained topology tests, it seems very unlikely that the addition of these taxa would rearrange the phylogeny such that the aquatic and/or

terrestrial species would form monophyletic groups. Thus, the parallel evolution of terrestrial habit and associated morphological changes appears to be the best explanation for the observed character distributions within Tetanocera. Because of the tight correlations among the morphological characters examined (i.e. all temporal order tests were nonsignificant after applying the sequential Bonferroni test), we were not able to distinguish among three possible scenarios regarding these correlations: (i) pleiotropy, (ii) indirect selection via another trait, or (iii) concerted evolution. However, the two habitat–morphology mismatches within Tetanocera may be indicative of the general ordering of larval morphological character state transitions or of current adaptation to novel habitats. Tetanocera plumosa has larvae that are principally aquatic but can also be found in shoreline (¼terrestrial) situations (Foote, 1961). This species exhibits one morphological character state (short float hairs) that is typically found in terrestrial larvae. Tetanocera fuscinervis, a shoreline predator, has an upturned spiracular disc,



Table 7 Character states of the Tetanocera species with known habitat preference not included in the phylogeny presented herein.


Spiracular disc orientation

Length

Pigmentation

Float hair length

Aquatic species T. annae T. loewi T. obtusifibula T. punctifrons T. soror T. spreta T. stricklandi

0 0 0 1 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0

Terrestrial species T. elata T. hyalipennis T. oxia T. rotundicornis T. spirifera

1 1 1 1 1

1 1 1 1 1

1 1 1 1 1

1 ? 1 1 1

State 0, darkly pigmented, long float hairs, long ventrolateral lobes, and lengthened terminal abdominal segment with upturned disc; State 1, unpigmented (appearing white), very short float hairs, short ventrolateral lobes, and terminal abdominal segment not lengthened, disc not upturned. Data from Foote (1961); unpublished data) and Knutson (1963).

which is the typical character state in aquatic species. These two exceptions to the rule could represent the two ends of an aquatic-to-terrestrial transition series. For example, the reduction of float hair length and the evolution of a posterior-facing spiracular disc could be the first and last, respectively, of the four characters to change during the evolutionary transition to a terrestrial larval habit in Tetanocera. This interpretation suggests that the larval stage of T. plumosa may be in the initial phase of transitioning to land and thus the short float hairs represent an adaptation to the much drier shoreline habitat. Furthermore, T. fuscinervis larvae could be viewed as being in the penultimate phase of transitioning to land with the upturned spiracular disc being viewed as historical baggage. The observation that the two, possibly transitional, Tetanocera species possessing mismatched character states occur in shoreline situations is consistent with a stepping stone role for this habitat in aquatic to terrestrial habitat transitions. Alternatively, these habitat–morphology mismatches may represent currently advantageous characteristics for specialized ecological circumstances. This hypothesis seems reasonable due to the ecotonal nature of the shoreline habit in these two species. For example, the upturned spiracular disc in T. fuscinervis could allow this species to breathe during periodic flooding of its shoreline habitat, enabling it to swim back to the shoreline. Ongoing, taxonomically more inclusive, phylogenetic/PCM studies within the Sciomyzidae, combined with evaluations of current

1471

utility for the mismatched character states within Tetanocera, offer possible means to evaluate these hypotheses. Both ancestral character state optimization methods suggested one reversal from terrestrial to aquatic habit in Tetanocera (Figs 4 and 5). Three parallel shifts to terrestrial habit plus one reversal is a more parsimonious explanation than the five parallel shifts to terrestrial habit that would be required if the ancestor to the phyllophoravalida clade were aquatic. Considering that the four morphological character states of the two species in question (T. ferruginea and T. bergi) closely resemble those of other aquatic species, this habitat reversal suggests that this lineage, minimally, reacquired cuticular pigmentation, long float hairs and long ventrolateral lobes. As a lengthened last abdominal segment with upturned spiracular disc occurs in the extant shoreline-inhabiting T. fuscinervis (the sister taxon to T. bergi + T. ferruginea), it may be that the T. ferruginea + T. bergi lineage moved to the shoreline and then back into the water without ever changing this character. Whiting et al. (2003) demonstrated that fully developed wings re-evolved in as many as four independent lineages of stick insects. His study demonstrated that a complex structure such as wings, requiring interactions between nerves, muscles, sclerites, and wing blades, could be lost and subsequently reacquired. In this light, the reacquisition of aquatic habitatassociated morphologies in Tetanocera is not surprising, especially if the genes responsible for the formation of these traits are not functionally ‘degraded’ during and after previous habitat transitions. Phylogenetic niche conservatism Pigmented cuticle, long float hairs, long ventrolateral lobes, and a lengthened last abdominal segment with an upturned spiracular disc are the plesiomorphic character states for Tetanocera (Fig. 5a–d). There are two possible explanations for the retention of plesiomorphic traits within a lineage: (i) phylogenetic constraint (McKitrick, 1993; Brooks & McLennan, 1994) or (ii) phylogenetic niche conservatism (Harvey & Pagel, 1991). McKitrick (1993) defined phylogenetic constraint as ‘any result or component of the phylogenetic history of a lineage that prevents an anticipated course of evolution in that lineage’. Phylogenetic constraint may result from the lack of genetic variation for a given trait, from coadaptation among traits that impose a genetic burden on the trait of interest, or from developmental correlations among traits (Wagner, 1995; Givnish, 1997). These constraints often result in a lag between environmental shifts and phenotypic change (Johnson et al., 1999). On the other hand, phylogenetic niche conservatism involves the action of stabilizing selection on phenotypic traits when ancestral ecological conditions are maintained within a lineage (Lord et al., 1995). Given that (i) there was a minimum of three independent shifts to terrestrial habit within Tetanocera, each accompanied by


1472


the same changes in these four morphological characters (clearly demonstrating the ability to evolve), (ii) all temporal order tests of habitat and morphology were nonsignificant and (iii) each plesiomorphic character state is apparently fundamental to aquatic existence, it seems apparent that phylogenetic niche conservatism, rather than phylogenetic constraint, continues to maintain these ancestral morphologies in the extant aquatic Tetanocera species. Adaptation to terrestrial existence Multiple parallel morphological character state changes accompanying a given habitat shift can be considered as evidence that the characters are adaptive (Baum & Larson, 1991; Brooks & McLennan, 1991,1994; Givnish, 1997; Patterson & Givnish, 2002). We are interpreting these multiple parallel changes in Tetanocera larval morphology, which are significantly correlated with habitat, as evidence consistent with an adaptation hypothesis. Extrapolation from a correlative to a cause-andeffect relationship is difficult in historical evolutionary studies as random associations of character states can potentially account for any observed pattern. However, we hypothesize that the inferred changes in the four morphological characters that accompanied the three transitions from the aquatic to the terrestrial habitat represent adaptive peak shifts, as each of the aquatic character states may typically be selected against in the terrestrial setting (e.g. Harvey & Pagel, 1991, Martins, 2000). For instance, the long ventrolateral lobes and upturned posterior spiracular discs of aquatic Tetanocera larvae may generally be impediments to prey-seeking movements in the vegetative tangles of the terrestrial habitat. An alternative cause-and-effect relationship, in which changes in larval morphology caused habit transitions, seems highly unlikely due to the physiological requirements of aquatic sciomyzid larvae. Studies of Anolis lizards in the Greater Antilles demonstrated that distinct lineages evolved similar morphologies as they made similar habitat shifts on different islands (Losos, 1992; Losos et al., 1994,1998). At least 17 parallel habitat transitions resulted in similar communities on each island. This landmark work demonstrated that similar solutions to similar evolutionary problems can be arrived at multiple times independently. Tetanocera exhibits a comparable phenomenon in that multiple lineages have transitioned to terrestrial larval foraging during phylogenesis (Figs 3 and 4) with each transition typically accompanied by the same four changes in larval morphology. These derived character states likely represent adaptations to the terrestrial environment. One of the interesting evolutionary questions generated by this study is ‘why did the larvae of multiple Tetanocera lineages leave the water for a terrestrial existence?’ This is not a general evolutionary tendency for the tribe as several tetanocerine genera have

exclusively aquatic larvae (e.g. Dictya, Elgiva, and Sepedonea). There are biotic and abiotic factors that may have been involved in these habitat transitions. Potential biotic factors for the transition from the aquatic to the terrestrial larval habit in Tetanocera include the following: (i) eliminating competition with other aquatic pulmonate snail predators, (ii) escaping aquatic predators/parasitoids, (iii) compensating for prolonged declines in aquatic pulmonate snail populations and (iv) low dispersal ability. Very little is known about the current levels of competition, predation, and dispersal ability, let alone about the historical levels, that act or have acted upon the larval stages of Tetanocera. An abiotic factor that could influence transitions from aquatic to terrestrial habitats is a generally drying climate. This could markedly affect the fitness of aquatic Tetanocera larvae directly by reducing the amount of suitable aquatic habitat. Wiegmann et al. (2003) estimated that the Schizophora [the dipteran subclade (series) to which the Sciomyzidae belongs] arose between 142 and 70 mya with the oldest known fossil sciomyzid dated to the Eocene/Oligocene epochs (55-24 mya; Hennig, 1965) and the oldest known fossil Tetanocera from the Oligocene epoch (34-24 mya; Theobald, 1937; Fo¨rster, 1891). The time period between the beginning of the Cenozoic era through the end of the Pliocene epoch (65-5 mya) is generally characterized by an increasingly cooler and drier climate. As wetlands began to dry (or at least change from permanent to temporary wetlands), sciomyzid lineages were likely pressured to cope with such conditions, and some populations may have responded by gradually shifting to more and more terrestrial habitats. The multiple aquatic-to-terrestrial larval habitat transitions identified herein for Tetanocera may have occurred during this time period in response to this general climatic trend. Additionally, because not all tetanocerine genera display aquatic to terrestrial habitat transitions (e.g. Dictya), factors intrinsic to Tetanocera may also be potentiating the changes (e.g. dispersal ability). However, lineages such as Dictya may have diversified after the drying time period, but the lack of fossils for such lineages prevents us from speculating further. Subsequent phylogenetic/comparative investigations, including those concentrating on dating the behavioural, ecological, and morphological changes we have described, will likely be of significant utility in further elucidating the causal processes responsible for these evolutionary transitions.

Acknowledgments We thank Lita Greve-Jensen, Lloyd Knutson, Wayne Mathis, Rory McDonnell, Joe Keiper, Ladislav Roller, and Rudolph Rozkosny for donating specimens for this project. We thank Diana Senyo, Judy Santmire, and Jennifer Walker for helping with lab work. We thank Bob Androw, Steve Chordas III, Laura Lawler-Chapman,



Joe Keiper, Austin Richards and Rob Roughley for assistance in the field. This work was partially funded by NSF grant DEB-0237175 (to W.R.H.), and by a Kent State University Graduate Student Senate research grant. We thank Ferenc DeSzalay, Joe Keiper, Patrick Lorch, Austin Richards, Andrea Schwarzbach, and Jennifer Walker for reviewing drafts of this manuscript. Finally, we thank Allen Moore and the anonymous reviewer for their comments on the manuscript.

References Armbruster, W.S. 2002. Can indirect selection and genetic context contribute to trait diversification? A transition-probability study of blossom-colour evolution in two genera. J. Evol. Biol. 15: 468–486. Baum, D.A. & Larson, A. 1991. Adaptation reviewed: a phylogenetic methodology for studying character macroevolution. Syst. Zool. 40: 1–18. Brooks, D.L. & McLennan, D.A. 1991. Phylogeny, Ecology, and Behavior: A Research Program in Comparative Biology. The University of Chicago Press, Chicago, IL. Brooks, D.L. & McLennan, D.A. 1994. Historical ecology as a research programme: scope, limitations, and the future. In: Phylogenetics and Ecology (P. Eggleton & R. Vane-Wright, eds), pp. 1–27. Academic Press, London. Collin, R. 2003. Phylogenetic relationships among calyptraeid gastropods and their implications for the biogeography of marine speciation. Syst. Biol. 52: 618–640. Creer, S., Malhotra, A. & Thorpe, R.S. 2003. Assessing the phylogenetic utility of four mitochondrial genes and a nuclear intron in the Asian pit viper genus, Trimeresurus: separate, simultaneous and conditional data combination analysis. Mol. Biol. Evol. 20: 1240–1251. Farris, J.S., Kallersjo, M., Kluge, A.G. & Bult, C. 1994. Testing significance of incongruence. Cladistics. 10: 315–319. Felsenstein, J. 1985a. Phylogenies and the comparative method. Am. Nat. 125: 1–15. Felsenstein, J. 1985b. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. Felsenstein, J. 2004. Inferring Phylogenies. Sinauer Assoc., Inc., Sunderland, MA. Ferrar, P. 1987. A guide to the breeding habits and immature stages of Diptera Cyclorrhapha (part 1: text), Sciomyzidae, p. 329–340 (part 2: figures), Sciomyzidae. In: Entomonograph. (L. Lyneborg, ed.), pp. 815–827. Vol 8, E. J. Brill/Scandanavian Science Press, Leiden & Copenhagen. Folmer, O., Black, M., Hoeh, W.R., Lutz, R. & Vrijenhoek, R. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Biotechnol. 3: 294–299. Foote, B.A. 1959. Biology and life history of the snail-killing flies belonging to the genus Sciomyza Fallen (Diptera). Ann. Entomol. Soc. Am. 52: 31–43. Foote, B.A. 1961. Biology and Immature Stages of the Snailkilling Flies Belonging to the Genus Tetanocera (Diptera: Sciomyzidae), PhD thesis. Cornell University, Ithaca, NY. Foote, B.A. 1971. Biology of Hedria mixta (Diptera: Sciomyzidae). Ann. Entomol. Soc. Am. 64: 931–941. Foote, B.A. 1976. Biology and larval feeding habits of three species of Renocera (Diptera: Sciomyzidae) that prey on

1473

fingernail clams (Mollusca: Sphaeriidae). Ann. Entomol. Soc. Am. 69: 121–133. Foote, B.A. 1996a. Biology and immature stages of snail-killing flies belonging to the genus Tetanocera (Insecta: Diptera: Sciomyzidae). 1. Introduction and life histories of predators of shoreline snails. Ann. Carnegie Mus. 65: 1–12. Foote, B.A. 1996b. Biology and immature stages of snail-killing flies belonging to the genus Tetanocera (Insecta: Diptera: Sciomyzidae). 2. Life histories of predators of snails of the family Succineidae. Ann. Carnegie Mus 65: 153–166. Foote, B.A. 1999. Biology and immature stages of snail-killing flies belonging to the genus Tetanocera (Insecta: Diptera: Sciomyzidae) 3. Life histories of the predators of aquatic snails. Ann. Carnegie Mus. 68: 151–174. Foote, B.A., Neff, S.E. & Berg, C.O. 1960. Biology and immature stages of Atrichomelina pubera (Diptera: Sciomyzidae). Ann. Entomol. Soc. Am. 53: 192–199. Fo¨rster, B. 1891. Die Insekten des ‘Plattigen Steinmergels’ von Brunstatt. Abh. Geol. Spez.-Karte Elsass-Lotheringen 3: 333–594. Givnish, T.J. 1997. Adaptive radiation and molecular systematics: issues and approaches. In: Molecular Evolution and Adaptive Radiation (T. J. Givnish & K. J. Systma, eds), pp. 1–54. Cambridge University Press, Cambridge. Harvey, P.H. & Pagel, M. 1991. The Comparative Method in Evolutionary Biology. Oxford University Press, Oxford. Hassanin, A. & Douzery, E. 2003. Molecular and morphological phylogenies of Ruminantia and the alternative position of the Moschidae. Syst. Biol. 52: 206–228. Hendy, M.D., Steel, M.A., Penny, D. & Henderson, I.M. 1988. Families of trees and consensus. In: Classification and Related Methods of Data Analysis (H. H. Bock, ed.), pp. 355–362. Elsevier, Amsterdam. Hennig, W. 1965. Die Acalyptratae des baltischen Bernsteins und ihre Bedeutung fu¨r die Erforschung der phylogenetischen Entwicklung dieser Dipteren-Gruppe. Stuttg. Beitr. Naturkd. 145: 1–215. Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6: 65–70. Huelsenbeck, J.P. & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogeny. Bioinformatics 17: 754–755. Johnson, K.P., McKinney, F. & Sorenson, M.D. 1999. Phylogenetic constraint on male parental care in the dabbling ducks. Proc. R. Soc. Lond. B 266: 759–763. Jousselin, E., Rasplus, J.-Y. & Kjellberg, F. 2003. Convergence and coevolution in a mutualism: evidence from a molecular phylogeny of Ficus. Evolution 57: 1255–1269. Kishino, H. & Hasegawa, M. 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order of the Hominoidea. J. Mol. Evol. 29: 170–179. Knutson, L.V. 1963. Biology and Immature Stages of Snailkilling Flies of Europe (Diptera: Sciomyzidae). PhD thesis, pp. 390.Cornell University, Ithaca, NY. Knutson, L.V. & Berg, C.O. 1964. Biology and immature stages of snail-killing flies: the genus Elgiva (Diptera: Sciomyzidae). Ann. Entomol. Soc. Am. 57: 173–192. Knutson, L.V. & Vala, J.C. 2002. An evolutionary scenario of Sciomyzidae and Phaeomyiidae (Diptera). Ann. Soc. Entomol. France 38: 145–162. Lord, J., Westoby, M. & Leishman, M. 1995. Seed size and phylogeny in six temperate floras: constraints, niche conservatism, and adaptation. Am. Nat. 146: 349–364.


1474


Losos, J.B. 1992. The evolution of convergent structure in Caribbean Anolis communities. Syst. Biol. 41: 403–420. Losos, J.B., Duncan, J.I. & Schoener, T.W. 1994. Adaptation and constraint in the evolution of specialization of Bahamian Anolis lizards. Evolution 48: 1786–1798. Losos, J.B., Jackman, T.R. Larson, A., de Qqeiroz, K. & Rodriguez-Schettino, L. 1998. Contingincy and determinism in replicated adaptive radiations of island lizards. Science 279: 2115–2118. Maddison, W.P. 1990. A method for testing the correlated evolution of two binary characters: are gains or losses concentrated on certain branches of a phylogenetic tree? Evolution 44: 539–557. Maddison, D.R. 1991. The discovery and importance of multiple islands of most-parsimonious trees. Syst. Zool. 40: 315–328. Maddison, W.P. 2000. Testing character correlation using pairwise comparisons on a phylogeny. J. Theor. Biol. 202: 195–204. Maddison, W.P. & Maddison, D.R. 2000. MacClade: Analysis of Phylogeny and Character Evolution. Sinauer Association Inc., Sunderland, MA. Maddison, W.P. & Maddison, D.R. 2003. Mesquite: a modular system for evolutionary analysis. Version 1.05 http://mesquiteproject.org.. Marinoni, L. & Mathis, W.N. 2000. A cladistic analysis of Sciomyzidae Fallen (Diptera). Proc. Entomol. Soc. Wash. 113: 162–209. Martins, E.P. 2000. Adaptation and the comparative method. Trends Ecol. Evol. 15: 296–299. Martins, E.P. & Hansen, T.F. 1996. A microevolutionary link between phylogenies and comparative data. In: New Uses for New Phylogenies (P. H. Harvey, A. J. L. Brown, J. M. Smith & S. Nee, eds), pp. 273–288. Oxford University Press, Oxford. McKitrick, M.C. 1993. Phylogenetic constraint in evolutionary theory: has it any explanatory power? Ann. Rev. Ecol. Syst. 24: 307–330. Neff, S.E. & Berg, C.O. 1966. Biology and immature stages of malacophagous Diptera of the genus Sepedon (Sciomyzidae). Va. Agr. Exp. Stat. B. 566: 1–113. Pagel, M. 1994. Detecting correlated evolution on phylogenies: a general method for the comparative analysis of discrete characters. Proc. Roy. Soc. Lond. B 255: 37–45. Pagel, M. 1997. Inferring evolutionary processes from phylogenies. Zool. Scr. 26: 331–348. Pagel, M. 1999a. Inferring the historical patterns of biological evolution. Nature 401: 877–884. Pagel, M. 1999b. The maximum likelihood approach to reconstructing ancestral character states on phylogenies. Syst. Biol. 48: 612–622. Pagel, M. 2000. Discrete, Version 4.0. A computer program distributed by the author. Pagel, M. 2002. Multistate, Version 0.6. A computer program, distributed by the author. Pagel, M., Meade, A. & Barker, D. 2004. Bayesian estimation of ancestral character states on phylogenies. Syst. Biol. 53: 673– 684. Park, J-K. & O’Foighil, D. 2000. Sphaeriid and corbiculid clams represent separate heterodont bivalve radiations into freshwater environments. Mol. Phylogenet. Evol. 14: 75–88.

Patterson, T.B. & Givnish, T.J. 2002. Phylogeny, concerted convergence, and phylogenetic niche conservatism in the core Liliales: insights from rbcL and ndhF sequence data. Evolution 56: 233–252. Pauly, G.B., Hillis, D.M. & Cannatella, D.C. 2004. The history of a Nearctic colonization: molecular phylogenetics and biogeography of the Nearctic toads (Bufo). Evolution 58: 2517–2535. Posada, D. & Crandall, K.A. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818. Prager, E.M. & Wilson, A.C. 1988. Ancient origin of lactabumin from lysozyme: analysis of DNA and amino acid sequences. J. Mol. Evol. 27: 326–335. Rice, W.R. 1989. Analyzing tables of statistical tests. Evolution 43: 223–225. Scheffer, S.J. & Wiegmann, B.M. 2000. Molecular phylogenetics of the holly leafminers (Diptera: Agromyzidae: Phytomyza): species limits, speciation, and dietary specialization. Mol. Phylogenet. Evol. 17: 244–255. Schwarz, M.P., Bull, N.J. & Cooper, S.J.B. 2003. Molecular phylogenetics of allodapine Bees, with implications for the evolution of sociality and progressive rearing. Syst. Biol. 52: 1–14. Shimodaira, H. 2002. An approximately unbiased test of phylogenetic tree selection. Syst. Biol. 51: 492–508. Shimodaira, H. & Hasegawa, M. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol. Biol. Evol. 16: 1114–1116. Shimodaira, H. & Hasegawa, M. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17: 1246–1247. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & Flook, P. 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87: 651–701. Swofford, D.L. 2001. PAUP*. Phylogenetic Analysis Using Parsimony (*and other Methods). Sinauer Associates, Sunderland, MA. Templeton, A.R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the humans and apes. Evolution 37: 221–244. Theobald, N. 1937. Les insectes fossiles des terrains oligoce`nes de France. G. Thomas, Nancy. 473 + [1] p. Vermeij, G.J. & Dudley, R. 2000. Why are there so few evolutionary transitions between aquatic and terrestrial ecosystems. Biol. J. Linn. Soc. 70: 541–554. Wagner, P.J. 1995. Testing evolutionary constraints hypotheses with early Paleozoic gastropods. Paleobiology 21: 248–272. Whiting, M.F., Bradler, S. & Maxwell, T. 2003. Loss and recovery of wings in stick insects. Nature 421: 264–267. Wiegmann, B.M., Yeates, D.K., Thorne, J.L. & Kishino, H. 2003. Time flies, a new molecular time scale for brachyceran fly evolution without a clock. Syst. Biol. 52: 745–756. Received 21 November 2005; revised 3 March 2006; accepted 9 March 2006
