Evolution of different gall types in willow-feeding sawflies (Hymenoptera: Tenthredinidae).
A possible trend exists in the evolution of gall types in the nematine gallers that belong to the hymenopteran family Tenthredinidae. Within the large, cosmopolitan family Tenthredinidae, the ability to induce plant galls has apparently evolved twice, in the subfamilies Nematinae and Blennocampinae, of which the latter includes only a few species (Price and Roininen 1993). The tribe Nematini is comprised of hundreds of galling and nongalling species (Viitasaari and Vikberg 1985) that feed exclusively on the Salicaceae, mostly on Salix (Price and Roininen 1993). Like their host plants, the nematines are most common in northern latitudes, and in the Arctic they are virtually the only sawflies present (Gauld and Bolton 1988; Price and Roininen 1993).
The number of galling nematine species is probably about 200, and they belong to three genera, Phyllocolpa, Pontania, and Euura (Price et al. 1994). These genera are recognized on the basis of the gall morphology and galling site on the host plant: Phyllocolpa species form simple leaf folds or rolls, Pontania species form leaf blade galls, and Euura species are leaf midrib, petiole, bud, or stem gallers (Meyer 1987; Price and Roininen 1993). Some authors consider leaf-rolling Phyllocolpa species to be a subgenus of Pontania (Vikberg 1970; Viitasaari and Vikberg 1985; Zinovjev 1993; Zinovjev and Vikberg 1998); in this study, they are considered separate genera (e.g., Smith 1970; Price and Roininen 1993) for clarity.
The nematine sawflies can be divided into eight informal groups based on the galling site on the host plant and gall morphology: (1) species with free-living larvae; (2) leaf folders or rollers; (3) leaf blade gallers; (4) apical leaf gallers; (5) basal leaf gallers; (6) midrib and petiole gallers; (7) stem gallers; and (8) bud gallers. Each gall type can be found on several willow species, and on each willow species many gall types can coexist (Kontuniemi 1960; Vikberg 1970; Roininen 1991; Price 1992). In all species, the initial gall formation is caused by substances secreted by the female during oviposition (Price and Roininen 1993).
Price (1992) suggested that the different gall types evolved gradually in the sequence mentioned above (Fig. 1; see also Roininen 1991). Thus, the ancestor of all galling nematines would be a Nematus-like species with free-living larvae. The next step would be a Phyllocolpa-like leaf folder, and after that Pontania-like leaf blade gallers would evolve. The galling site would then "wander" on the leaf to be more closely associated with the midrib (apical and basal leaf gallers), and then to the midrib and petiole (Euura sp.). The last species to evolve would be the Euura stem and bud gallers. According to this hypothesis, radiations to new willow species occurred after the evolution of each gall type. An alternative explanation would be that similar gall types evolved independently on different willow species; however, the process might have been a mixture of both alternatives (Price and Roininen 1993).
The Price-Roininen hypothesis is appealing in its simplicity. The evolution of new gall types would occur gradually as a result of slight changes in behavior and oviposition preference of the females, which would be in concordance with the traditional gradualistic view of evolution. Since gall formation is dependent on exact positioning of the gall-forming substance (McCalla et al. 1962), gall type changes probably require multiple, coexisting changes in behavior and ovipositor shape. In the initial stage of a gall type shift, there must exist a polymorphism in the population, which seems more probable if the oviposition site of the different forms is relatively similar. These factors might act as a constraint, making large shifts improbable. However, oviposition occurs during the very early development of the willow shoots when different parts of the leaves are in very close proximity. This could enhance the possibility of "oviposition jumps" that appear more drastic than they really are (Price and Roininen 1993).
The purpose of this study was to test the validity of the Price-Roininen hypothesis. This was done by using enzyme electrophoretic data to reconstruct the phylogeny of selected nematine species representing different gall types. The evolution of the gall types could then be inferred by plotting the gall types on the tree and reconstructing the probable ancestral states.
MATERIALS AND METHODS
Sample Collection and Preservation
Four outgroup and 13 ingroup taxa were chosen for the electrophoretic survey (Table 1). In the final analysis, the number of ingroup taxa was 14 because a sample representing Phyllocolpa leucapsis included individuals with fixed differences at two loci, which were consequently split into two species. The ingroup taxa were chosen so that all gall types were represented in the sample. Two species were sampled per gall type, except for basal leaf gallers, for which Pontania pustulator was the only representative, and for leaf folders for which the final number of taxa was three as a result of the splitting of P. leucapsis.
Adults, prepupae, or larvae collected from galls or leaves were used for electrophoresis. The specimens were collected during the summers of 1989-1994, and stored at -80 [degrees] C, except for some samples that were initially stored at -20 [degrees] C. The sample collection dates and sites are summarized in Table 1. For each species, only larvae from one willow species were collected, but the collection was made randomly from several willow clones to get a representative sample of the population. Most species were collected at several sites.
The electrophoretic survey was performed with standard starch gel electrophoresis as outlined in Vuorinen (1984). Whole adults or larvae were homogenized in 40 [micro]L sample buffer (Ferguson 1980), and the crude extract was used for electrophoresis. The horizontal 13% starch gels were cut into five 2 mm thick slices and stained for specific enzymes.
Nine enzyme loci were scored (Table 2). Sharing of electromorphs was confirmed by side-by-side comparisons. When allele frequencies were calculated, all individuals were treated as diploid, even though hymenopteran males are haploid (Gauld and Bolton 1988). This should not affect the frequencies, however. Allele frequencies are presented in the Appendix. Exact sample sizes could not be calculated, since the larvae could not be reliably sexed. Sample sizes are presented as minimum numbers of haploid genotypes surveyed ([N.sub.min]); these were calculated by giving a weight of two to all individuals that were heterozygous at at least one locus and a weight of one to all individuals that were homozygous at all surveyed loci. [N.sub.min] is thus a conservative estimate of the true sample size in each species.
In the phylogenetic analysis, the loci were treated as characters and allelic combinations as character states, applying step matrices to define the cost of transformations between states (character = locus coding; see Mabee and Humphries 1993; Mardulyn and Pasteels 1994). All detected alleles were included in the study. This is in contrast to, for example, Futuyma and McCafferty (1990), who used a cutoff level of 0.05; alleles occurring with a frequency less than this were coded as absent. Minimum-length trees were found by using PAUP (vers. 3.1.1, Swofford 1993).
Of the nine loci studied, eight were informative in the analysis. Gains and losses were weighted equally, as Mardulyn and Pasteels (1994) recommended. To ensure the finding of the most parsimonious solutions, new (hypothetical [TABULAR DATA FOR TABLE 1 OMITTED] ancestral) character states were created when needed, as described in Mardulyn and Pasteels (1994). A branch-and-bound search was not feasible because of the computational complexity involved in analyzing step matrix data. Instead, two heuristic searches were performed using different methods (simple addition sequence with TBR branch-swapping, and random addition sequence with 10 replicates and TBR branch-swapping); both produced the same results. A bootstrap analysis (Felsenstein 1985) with 100 replicates and heuristic search was also performed, and the Bremer (1988, 1994) support index was determined for each interior node.
TABLE 2. List of enzymes studied, loci scored, and buffers used for each enzyme. E.C. Buff- Enzyme name number Locus er(1) Aspartate aminotransferase 22.214.171.124 Aat-1(2) B Aat-2 B Glucose-6-phosphate isomerase 126.96.36.199 Gpi A Isocitrate dehydrogenase 188.8.131.52 Idh B Malate dehydrogenase 184.108.40.206 Mdh-1(2) S Mdh-2 S Malic enzyme 220.127.116.11 Me S Phosphogluconate dehydrogenase 18.104.22.168 Pgdh(3) B Triose phosphate isomerase 22.214.171.124 Tpi(3) A 1 A = a discontinuous lithium hydroxide-boric acid buffer, pH 8.1 (Ridgway et al. 1970); B = an N-(3-aminopropyl)-morpholine-citrate buffer, pH 6.5 (Clayton and Tretiak 1972); S = a 0.135 M tris-citrate buffer, pH 7.1 (Shaw and Prasad 1970). 2 Cathodal locus. 3 Anodal and cathodal alleles.
Since the heuristic search does not ensure that the shortest trees are found, an alternative analysis was performed using alleles as characters (character = allele coding; Mickevich and Johnson 1976; Mickevich and Mitter 1981). A branch-and-bound search produced 23 equally parsimonious trees, but all trees were four to seven steps longer than the character = locus trees when evaluated with the character = locus data. These trees were not used for further analyses, because the character = allele approach has been widely criticized (e.g., Swofford and Olsen 1990; Murphy 1993; Meier 1994).
The obtained character = locus trees were rooted by using four taxa with free-living larvae as an outgroup (Table 1). Nematus pavidus and N. salicis have exophytic larvae that feed on leaves, Pontopristia amentorum larvae feed inside willow catkin stems (Viitasaari 1982), and Decanematus viduatus larvae initially live inside developing shoot tops, but become free feeders in their later stages of development (Zinovjev and Vikberg 1998). The rooting required paraphyly of the outgroup.
To study the evolution of new gall types, the gall types were plotted on the trees using MacClade (vers. 3.01, Maddison and Maddison 1992). The gall types were coded as unordered multistate characters, that is, all transformations between gall types were considered equally probable.
The trees were also tested against hypothetical trees conveying different evolutionary scenarios. Templeton's (1983; see also Larson 1994) nonparametric test was used to test whether the hypothetical trees are significantly longer than the trees obtained from the phylogenetic analysis.
Randomization tests (see Liebherr and Hajek 1990) were used to test whether the observed patterns in gall type evolution are statistically significant. A gall type code was assigned for each species, from 0 (outgroup) to 7 (bud gallers) in the order presented in Table 1. A random dataset was generated by multiplying the original gall types 200 times and shuffling states in the characters (columns) between taxa, but keeping the state of the outgroup species constant (= 0). The length of the original character, when plotted on the trees obtained from the phylogenetic analysis, was then compared with the random character length distribution. The randomization analyses were performed using MacClade (vers. 3.01, Maddison and Maddison 1992).
The heuristic searches yielded two equally parsimonious solutions for the character = locus data. Both trees [ILLUSTRATION FOR FIGURE 2 OMITTED] have a length of 91 steps, and they differ only in the relative position of Euura mucronata. Both trees contain a polytomy in the Euura (all Euura species except E. atra) clade. Euura atra is included in the Pontania clade (E. atra plus all Pontania species, except for the leaf blade gallers P. bridgmanii and P. triandrae). Species representing the same gall types generally are monophyletic, and the trees could be rooted so that the ingroup remains monophyletic. Bootstrap proportions were generally low, and Bremer support indices range from one to three [ILLUSTRATION FOR FIGURE 2 OMITTED]. Bremer's (1994) total support index ti = 0.26.
Optimization of Gall Types on the Trees
The polytomy in both character = locus trees was resolved in all possible (three) ways, which led to six optional tree topologies. To find all possible solutions, the gall types were plotted on each tree using MacClade (Maddison and Maddison 1992), and three optimization options were used (Acctran, Deltran, and show all most parsimonious reconstructions).
The different options resulted in 18 alternative reconstructions. All reconstructions agreed that the first taxa to evolve from nongalling species were leaf-folding Phyllocolpa species. Leaf blade galling Pontania species evolved independently of the other (leaf galling) Pontania species, which may or may not have had a common galling ancestor with the Euura group. The alternative reconstructions show little agreement about the most probable gall type at the base of the Euura-Pontania clade.
Testing Alternative Trees
The shortest possible tree that would strictly support the Price-Roininen hypothesis has a length of 101 steps, when evaluated with the character = locus data. It is significantly less parsimonious than the character = locus trees ([T.sub.s] = 2, n = 6, one-tailed P = 0.0469; Rohlf and Sokal 1981).
Similarly, the 32 shortest possible trees that resulted from a phylogenetic analysis where the ingroup species from the same willow host species were forced to be monophyletic are significantly longer than the original trees (100 steps; [T.sub.s] = 0, n = 5, one-tailed P = 0.0312).
If the ingroup monophyly is disrupted by moving the Euura clade to the other side of the outgroup (the galling habit is polyphyletic), the tree has a length of 100 steps ([T.sub.s] = 0, n = 5, one-tailed P = 0.0312). If the bud gallers are moved to the base of the ingroup, tree length equals 102 steps ([T.sub.s] = 0, n = 5, one-tailed P = 0.0312). If the Phyllocolpa species are moved into the outgroup, so that galling arises directly from nongallers, the trees are at least 100 steps long ([T.sub.s] = 2.5, n = 7, one-tailed P [less than] 0.0391). Thus, all tested scenarios are statistically significantly less parsimonious than the trees obtained from the phylogenetic analysis (Fig. 2).
The purpose of the first randomization test was to ask if gall types change often relative to speciation. If this was the case, gall types would become effectively randomized between taxa on the trees. The six observed trees were tested separately, and the gall types were treated as unordered. The trees require seven to eight gall type changes, and with all six trees the 200 gall type replicates with randomized states required more steps (P [less than] 0.005). This shows that gall type is a relatively conservative character. Figure 3 presents the character length distribution of the tree in Figure 4.
The second test was used to see if the gall types change in a pattern that, although not identical, is closer to the Price-Roininen hypothesis than expected by random. Gall types were coded as ordered and gall types were shuffled between groups, while keeping the groups (species representing the same gall type) themselves homogeneous. The six trees required 12 to 14 steps when the gall types were ordered. In the random dataset, 0.5-5.0% of the characters were of the same length or shorter than the original character (P = 0.005-0.05). Figure 3 shows the length distribution of the randomized characters on the tree in Figure 4. The result of this test shows that gall type switches have occurred in small steps, and that there has been a trend from the leaf edge toward the more central parts of the host plant.
When considering gall type evolution in the galling sawflies, two main questions arise: (1) what is the ratio of host shifts compared to evolution of new gall types; and (2) in what sequence have the gall types evolved? If gall type shifts are more common than host plant shifts, species representing a certain gall type would not be monophyletic and, most likely, it would be impossible to infer a distinct sequence in gall type evolution.
Our phylogenetic analysis and the result from the randomization test clearly show that host shifts are more common than the evolution of new gall types. Most groups that were hypothesized to belong together based on gall morphology and location also remained together on the trees obtained from the phylogenetic analysis, although the species occur on different willow species [ILLUSTRATION FOR FIGURE 4 OMITTED]. The only notable exception is the polyphyly of the stem gallers (Euura lasiolepis and E. atra). The result is similar to Stern's (1995) finding that in cerataphidine aphids the insects, not their host plants, determine gall morphology. The main question thus becomes, what is the sequence in which the gall types evolved?
The phylogenies obtained from the character = locus analysis support the view that the galling nematines are a monophyletic group with respect to species with free-living larvae. All reconstructions agree that the first species to evolve were the leaf-folding Phyllocolpa species. This is in concordance with Smith's (1970) notion that the larvae of Phyllocolpa resemble those of the free-feeding Nematus species in form and coloration. The next species to appear would be the leaf blade galling Pontania, which, in contrast to the Price-Roininen hypothesis, are not the ancestors of apical and basal leaf-galling Pontania species, but are instead of independent origin. Thus, true closed galls apparently evolved independently at least twice. It should be noted that Zinovjev (1993) classifies the Pontania leaf blade gallers in the subgenus Pontania s. str., which includes both gallers and some leaf rollers; apical and basal leaf gallers belong to the subgenus Eupontania, which according to Zinovjev (1993) resemble Euura (e.g., by mandible morphology).
The Euura-Pontania clade is problematic because of the two equally parsimonious resolutions in the Euura clade, and the different options for resolving the polytomy. The most realistic option is probably a tree in which the bud gallers E. lanatae and E. mucronata are considered to be sister taxa, and E. lasiolepis an offshoot from petiole gallers (Fig. 4). Thus, stem gallers would have evolved twice: in the Euura group from midrib/petiole gallers and in the Pontania clade from leaf gallers. Both E. lasiolepis and E. atra are currently classified together in the Euura (Euura) lasiolepis group, based on ovipositor shape and pupation in the gall; these features may have been independently derived as a result of convergent galling habits.
An interesting fact is that E. testaceipes, which galls midribs and petioles (Price and Roininen 1993), is closer to the Pontania clade than E. venusta, which galls only petioles (Price and Roininen 1993). This is in concordance with the Price-Roininen hypothesis. The bud galling Euura (Gemmura) group has clearly evolved late, and thus Smith's (1968, 1970) suggestion that they are intermediates between Pontania and Euura gallers is incorrect. Smith based his view on similarities in ovipositor shape and pupation on the ground. However, as Price and Roininen (1993) pointed out, the larvae of the petiole galling Euura (Euura) species also pupate in the soil, and the Euura (Euura) group is closest to leaf-galling Pontania species in oviposition behavior, a trait that can be presumed to be quite complex and slow changing (Price and Roininen 1993).
The ambiguity at the base of the Euura-Pontania clade is partly due to the unordered coding method, which assumes all transitions between gall types to be equally probable. Although this assumption minimizes the possibility of circular reasoning, it is not necessarily realistic. The reconstruction in Figure 4 presents the Acctran optimization of an option in which apical and basal leaf gallers are considered to be the same group, which is equivalent to giving the transition between apical and basal leaf gallers a much smaller cost (greater probability) than for the transition between other gall types. This can be justified on the basis that the oviposition behavior of apical and basal leaf gallers is relatively similar to each other. The most probable evolutionary sequence would thus be: (1) evolution of Phyllocolpa leaf folders from species with free-living larvae; (2) leaf blade galling Pontania species evolve from Phyllocolpa ancestors; (3) the common ancestor for (apical/basal) leaf galling Pontania and Euura species evolves; (4) stem gallers evolve independently from Pontania leaf gallers and Euura midrib/petiole gallers; and (5) Euura (Gemmura) bud gallers evolve directly from midrib/petiole gallers, and thus not from stem gallers as the Price-Roininen hypothesis predicts. Although it is possible that Pontania leaf gallers and Euura midrib/petiole gallers evolved independently from leaf folders, it is more realistic to assume that they share a common galling ancestor.
Although the dataset in this study is relatively small and the bootstrap analysis yielded low bootstrap proportions for many groupings, several groups are supported by the beforementioned morphological traits (Smith 1968, 1970; Zinovjev 1993; Zinovjev and Vikberg 1998). Thus, the general outline of the obtained phylogenies should be correct.
The most important implication of the phylogeny is that there is a statistically significant evolutionary trend toward oviposition in the more basal parts of the growing shoot. If the Price-Roininen hypothesis was correct, there would have been seven gall type transitions during the evolution from outgroup to bud gallers. If the gall types are coded as ordered, the obtained trees require 12 to 13 gall type changes, which is still less than expected by a random process. Since the original state is known (0, i.e., free-living larvae), this shows that the evolution of new gall types occurred in small steps and in the general direction suggested by the Price-Roininen hypothesis.
The gradual evolution of gall types is probably a result of genetical constraints in the oviposition behavior system. Since oviposition and gall formation is probably under the control of coadapted multilocus genetic systems, large shifts are unlikely. In the initial stage of a gall type shift, there must exist a polymorphic condition in the population, which is more probable if the gall types are relatively similar, as in Euura testaceipes.
Possible reasons for the observed directionality could include, for example, better control of the flow of nutrients and photosynthetic products, or greater cost of defense reactions for the host, when galls are situated more basally on the leaf. The trend could also be a consequence of intraspecific competition, if there is a "shading effect" in which galls situated closer to the main vascular system block nutrient flow to galls situated in more peripheral parts of the host plant. The effect could be important in Pontania species, where one leaf may have many galls (Price et al. 1994). The shading effect has been observed in Pemphigus gall aphids (Whitham 1978), but it has not been thoroughly studied in sawflies. Studies on interspecific competition could give an answer, but thus far they have produced ambiguous results (e.g., Fritz et al. 1986; Fritz and Price 1990).
Although evolutionary trends and biases in transition probabilities are often proposed, few cases have been documented. Janson (1992) found unequal transition probabilities between seed dispersal syndromes in Neotropical plants. In the leaf beetle genus Ophraella, host shifts are partially constrained by lack of genetical variation (Futuyma et al. 1993, 1995). Liebherr and Hajek (1990) studied habitat preference of carabid beetles, and in only one case (of eight) they found a statistically significant trend. The trend observed in the evolution of the galling nematines is thus interesting, but further studies are needed to gain a better understanding of the genealogical relationships and gall type evolution in the galling nematines.
We thank L. Kuusipalo, G. Melo, S. Lingafelter, N. E. Rank, and J. Sorjonen for various practical help during the work. P. W. Price kindly sent samples from Arizona. We also thank C. Mitter and D. L. Stern for giving very useful suggestions for the data analysis, and J. Tahvanainen for his comments on the manuscript. The text was also checked by S. Franklin, C. Anttila, and T. Carr. The work was performed at the Department of Biology at the University of Joensuu. Partial funding was provided by the Finnish Biological Society Vanamo and the Finnish Academy.
ALBRECHT, P., AND E. A. GALE. 1985. A developmental analysis of an evolutionary trend: digital reduction in amphibians. Evolution 39:8-23.
BONNER, J. T., ED. 1982. Evolution and development. Springer, Berlin, Germany.
BREMER, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42:795-803.
-----. 1994. Branch support and tree stability. Cladistics 10:295-304.
CLAYTON, J. W., AND D. N. TRETIAK. 1972. Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Bd. Can. 29:1169-1172.
FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791.
FERGUSON, A. 1980. Biochemical systematics and evolution. Blackie, Glasgow, Scotland.
FRITZ, R. S., AND P. W. PRICE. 1990. A field test of interspecific competition on oviposition of gall-forming saw flies on willow. Ecology 71:99-106.
FRITZ, R. S., C. E SACCHI, AND P. W. PRICE. 1986. Competition versus host plant phenotype in species composition: willow sawflies. Ecology 67:1608-1618.
FUTUYMA, D. J. 1988. Sturm und Drang and the evolutionary synthesis. Evolution 42:217-226.
FUTUYMA, D. J., AND S.S. MCCAFFERTY. 1990. Phylogeny and the evolution of host plant relationships in the leaf beetle genus Ophraella (Coleoptera, Chrysomelidae). Evolution 44:1885-1913.
FUTUYMA, D. J., M. C. KEESE, AND S. J. SCHEFFER. 1993. Genetic constraints and the phylogeny of insect-plant associations: responses of Ophraella communa (Coleoptera: Chrysomelidae) to host plants of its congeners. Evolution 47:888-905.
FUTUYMA, D. J., M. C. KEESE, AND D. J. FUNK. 1995. Genetic constraints on macroevolution: the evolution of host affiliation in the leaf beetle genus Ophraella. Evolution 49:797-809.
GAULD, I., AND B. BOLTON. 1988. The Hymenoptera. Oxford Univ. Press, Oxford, U.K.
JANSON, C. H. 1992. Measuring evolutionary constraints: a Markov model for phylogenetic transitions among seed dispersal syndromes. Evolution 46:136-158.
KONTUNIEMI, T. 1960. Suomen sahapistiaistoukkien ravintokasvit (Die Futterpflanzen der Sagewespenlarven [Hymenoptera, Symphyta] Finnlands). Anita. Fenn. 9:1-104.
KOPELKE, J.-P. 1994. Der Schmarotzerkomplex (Brutparasiten und Parasitoide) der Gallenbildenden Pontania-Arten (Insecta: Hymenoptera: Tenthredinidae). Senckenb. Biol. 73: 83-133.
LARSON, A. 1994. The comparison of morphological and molecular data in phylogenetic systematics. Pp. 371-390 in B. Schierwater, B. Streit, G. P. Wagner, and R. DeSalle, eds. Molecular ecology and evolution: approaches and applications. Birkhauser, Basel, Switzerland.
LIEBHERR, J. K., AND A. E. HAJEK. 1990. A cladistic test of the taxon cycle and taxon pulse hypotheses. Cladistics 6:39-59.
MABEE, P. M., AND J. HUMPHRIES. 1993. Coding polymorphic data: examples from allozymes and ontogeny. Syst. Biol. 42:166-181.
MADDISON, W. P., AND D. R. MADDISON. 1992. MacClade: analysis of phylogeny and character evolution. Vets. 3.01. Sinauer, Sunderland, MA.
MARDULYN, P., AND J. M. PASTEELS. 1994. Coding allozyme data using step matrices: defining new original states for the ancestral taxa. Syst. Biol. 43:567-572.
MCCALLA, D. R., M. K. GENTHE, AND W. HOWANITZ. 1962. Chemical nature of an insect gall growth factor. Plant Physiol. 37:48-103.
MEIER, R. 1994. On the inappropriateness of presence/absence recoding for non-additive multistate characters in computerized analyses. Zool. Anz. 232:201-212.
MEYER, J. 1987. Plant galls and gall inducers. Gebruder Borntraeger, Berlin, Germany.
MICKEVICH, M. F., AND M. S. JOHNSON. 1976. Congruence between morphological and allozyme data in evolutionary inference and character evolution. Syst. Zool. 25:260-270.
MICKEVICH, M. F., AND C. MITTER. 1981. Treating polymorphic characters in systematics: a phylogenetic treatment of electrophoretic data. Pp. 45-58 in V. A. Funk and D. R. Brooks, eds. Advances in cladistics I. New York Botanical Garden, New York.
MURPHY, R. W. 1993. The phylogenetic analysis of allozyme data: invalidity of coding alleles by presence/absence and recommended procedures. Biochem. Syst. Ecol. 21:25-38.
PRICE, P. W. 1992. Evolution and ecology of gall-inducing sawflies. Pp. 208-224 in J. D. Shorthouse and O. Rohfritsch, eds. Biology of insect-induced galls. Oxford Univ. Press, Oxford, U.K.
PRICE, P. W., AND H. ROININEN. 1993. Adaptive radiation in gall induction. Pp. 229-257 in M. Wagner and K. F. Raffa, eds. Sawfly life history adaptations to woody plants. Academic Press, New York.
PRICE, P. W., K. M. CLANCY, AND H. ROININEN. 1994. Comparative population dynamics of the galling sawflies. Pp. 1-11 in P. W. Price, W. J. Mattson, and Y. N. Baranchikov, eds. The ecology and evolution of gall-forming insects. United States Department of Agriculture, St. Paul, MN.
RIDGWAY, G. J., S. W. SHERBURNE, AND R. D. LEWIS. 1970. Polymorphism in the esterases of Atlantic herring. Trans. Am. Fish. Soc. 99:147-151.
ROHLF, F. J., AND R. R. SOKAL. 1981. Statistical tables. 2d ed. Freeman, San Francisco, CA.
ROININEN, H. 1991. The ecology and evolution of the host plant relationships among willow-feeding sawflies. Univ. Joensuu Publ. Sci. 23:1-21.
SCHLUTER, D. 1996. Adaptive radiation along genetic lines of least resistance. Evolution 50:1766-1774.
SHAW, C. R., AND R. PRASAD. 1970. Starch gel electrophoresis of enzymes - a compilation of recipes. Biochem. Genet. 4:297-320.
SMITH, E. L. 1968. Biosystematics and morphology of Symphyta. I. Stem-galling Euura of the California region and new female genitalia nomenclature. Ann. Ent. Soc. Am. 6:1389-1407.
-----. 1970. Biosystematics and morphology of Symphyta. II. Biology of gall-making nematine sawflies in the California region. Ann. Ent. Soc. Am. 63:35-51.
STERN, D. L. 1995. Phylogenetic evidence that aphids, rather than plants, determine gall morphology. Proc. R. Soc. Lond. B Biol. Sci. 260:85-89.
SWOFFORD, D. L. 1993. PAUP: phylogenetic analysis using parsimony. Vers. 3.1.1. Illinois Natural History Survey, Champaign.
SWOFFORD, D. L., AND G. J. OLSEN. 1990. Phylogeny reconstruction. Pp. 411-501 in D. M. Hillis and C. Moritz, eds. Molecular systematics. Sinauer, Sunderland, MA.
TEMPLETON, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37:221-244.
VISTASAARI, M. 1982. Sahapistiaiset 1: Yleinen osa. University of Helsinki, Department of Agricultural and Forest Zoology, report no. 3.
VIITASAARI, M., AND V. VIKBERG. 1985. A checklist of the sawflies (Hymenoptera, Symphyta) of Finland. Not. Entomol. 65:1-17.
VIKBERG, V. 1970. The genus Pontania O. Costa (Hym., Tenthredinidae) in the Kilpisjarvi district, Finnish Lapland. Ann. Entomol. Fenn. 36:10-24.
VUORINEN, J. 1984. Electrophoretic expression of genetic variation and duplicate gene activity in vendace, Coregonus albula (Salmonidae). Hereditas 101:85-96.
WAKE, D. B. 1991. Homoplasy: the result of natural selection, or evidence of design limitations. Am. Nat. 138:543-567.
WHITHAM, T. G. 1978. Habitat selection by Pemphigus gall aphids in response to resource limitation and competition. Ecology 59: 1164-1176.
ZINOVJEV, A. G. 1993. Subgenera and Palaearctic species groups of the genus Pontania, with notes on the taxonomy of some European species of the viminalis-group (Hymenoptera: Tenthredinidae). Zoosyst. Ross. 2:145-154.
ZINOVJEV, A. G., AND V. VIKBERG. 1998. On the biology of Nematinae with hiding habits (Hymenoptera, Symphyta, Tenthredinidae). Beitr. Entomol. 48:145-155.
[TABULAR DATA FOR APPENDIX OMITTED]
|Printer friendly Cite/link Email Feedback|
|Author:||Nyman, Tommi; Roininen, Heikki; Vuorinen, Jukka A.|
|Date:||Apr 1, 1998|
|Previous Article:||Mortality plateaus and the evolution of senescence: why are old-age mortality rates so low?|
|Next Article:||Stabilizing selection and variance in fig wasp sex ratios.|