Printer Friendly

Evolution of parasitism among closely related species: phylogenetic relationships and the origin of inquilinism in gall wasps (Hymenoptera, Cynipidae).

partly smooth or almost smooth; (b) entirely covered with irregular sculpture.

67. Shape of metanotal trough and bar beneath it: (a) metanotal trough broad, bar narrow, ratio of maximum width of trough to minimum width of bar [greater than]3.0; (b) metanotal trough narrow, bar broad, ratio [less than]2.5.

Metapectal-propodeal Complex, Female

TABULAR DATA OMITTED

68. Position of anterior end of metapleural sulcus: (a) low, ratio of distance The family Cynipidae, as here construed, is unusual among famifies of parasitic wasps in that all its members are phytophagous. Most of the species are gall formers on various plants, but the family also includes inquilines. The inquilines cannot induce galls therapies; instead they develop inside the galls of other cynipids. The latter will hereafter simply be reffered to as the inquilines' hosts, although they are really a kind of "nest hosts" because the inquilines derive most or all of their nutrition from the plant tissue in the gall. Cynipid inquilinism can be viewed as a form of parasitism, in which the inquilines take advantage of other species' ability to form galls. Although it has been suggested that some inquilines do little harm to the original gall inducer (Mayr 1872), all inquilines studied in detail so far have been shown to affect the host gall-in-ducer negatively (Wiebes-Rijks and Shorthouse 1992). Many inquilines kill their gall-inducing host at an early stage of the development of the gall, either directly or indirectly through food deprivation (Nielsen 1903; Evans 1965, 1967; Shorthouse 1973, 1980; Wiebes-Rijks 1980; Washburn and Cornell 1981).

From an evolutionary standpoint, parasites can be divided into forms that take advantage of a phylogenetically unrelated host and forms that parasitize on species that are closely related to the parasites themselves (Wheeler 1919). Cynipid inquilinism is an example of the latter type of parasitism, here termed agastoparasitism (from Greek agastor, near kinsman). More formally, the phenomenon can be defined as follows: if there is a monophyletic group, such that it includes one or more parasites and some or all of their hosts, but few other species or lineages, then the parasites can be considered agastoparasites. In most cases, this means that the parasite will be classified in the same family or even the same genus as its host. Agastoparasitism occurs in a variety of organisms but appears to be particularly common in the aculeate Hymenoptera, showing up as social parasitism and (primary) cleptoparasitism in numerous different lineages (Bischoff 1927; Michener 1970, 1977, 1978; Askew 1971: Wilson 1971; Gauld and Bolton 1988; Alexander 1990; Holldobler and Wilson 1990). A few cases are also known in parasitic wasps, including cynipid inquilinism and the peculiar "drilling-hole parasitism" of Rhyssa by Pseudorhyssa (Couturier 1949; Spradbery 1969).

A recurrent problem in the study of agasto-parasites is the difficulty in establishing their evolutionary origin. Typically there are two competing hypotheses: either the parasites arose from one of their hosts and later radiated to exploit other related hosts (monophyletic origin), or the parasites originated repeatedly each parasite from its host (polyphyletic origin). Both hypotheses have been proposed for cynipid inquilines. Following systematists that place the cynipid inquilines in a separate subfamily (Hartig 1840; Ashmead 1896, 1903a; Burks 1979b), some have maintained that the inquilines form a monophyletic group. Among these, Roskam (1992) suggested that the inquilines evolved from gall inducers in the tribe Aylacini, and Ritchie (1984) specifically postulated an origin from the genus Diastrophus. These workers thus argue for a single transition from gall forming to inquilinism, and subsequent radiation of the inquiries to exploit different host galls. Others have questioned the monophyletic-origin scenario. Askew (1984) considered the inquilines an artificial, polyphyetic group, and Gauld and Bolton (1988) suggested that each inquiline evolved from its host. According to this view, then, there have been several transitions from gall forming to inquilinism.

The difficulty in testing these scenarios stems from a concern that phylogenetic analyses based on morphological characters will be influenced on one hand by convergent similarities among agastoparasites because of their similar mode of life, which will tend to group parasites together, and on the other hand by convergent similarities between the parasites and their hosts caused by their living in a close spatiotemporal association and therefore experiencing similar environments, which will tend to group parasites with their respective hosts. In an attempt to avoid this problem, several workers have used biochemical instead of morphological characters in phylogenetic inference (e.g., Pamilo et al. 1981; Varvio-Aho et al. 1984). However, this approach is not as straightforward as it first may seem, because most biochemical characters examined so far are exposed to natural selection and could therefore be affected by the same convergent trends as morphological characters.

Another way of attacking the problem, so far little explored, is to try to reduce or eliminate the influence of possible convergent similarities on phylogenetic inference by careful analysis and evauation of the data. One possibility is to base analyses solely on features that are unlikely to be affected by convergence, for example, features of genitalia wing venation, or life-history stages not influenced by convergent selection (Plow-fight and Stephen 1973; Michener 1977; Williams 1985; Alexander 1990). However, it is often difficult to find such features, and the likelihood of a character being affected by convergent evolution cannot easily be determined objectively (cf. Richards 1927; Wilson 1971).

In this paper, I use a different method that does not depend on assumptions about the effects of convergent evolution on different types of characters. Clearly, convergent trends will show up as apomorphic (i.e., derived) character states with a particular distribution among taxa. Thus, convergent similarities among parasites will appear as derived character states occurring in some or all of the parasites but not in the hosts. Similarly, convergence caused by shared environments produces apparent synapomorphies (shared derived similarities) uniting parasites with their respective hosts. If we are concerned that one of these two sets of convergent characters is sufficiently large to influence the result of a particular phylogenetic analysis, it is possible to remove these characters from the analysis. The problem, of course, is that true synapomorphies may have exactly the same distribution among taxa as the convergent characters. In other words, we do not know if we are removing a set of convergent characters in conflict with actual relationships, or a mixture of true synapomorphies and convergences congruent with actual relationships. However, it is possible to separate these two alternatives by studying the effect of the removal.

Consider the example presented in figure 1, in which the evolutionary origin of two agastoparasites ([P.sub.A] and [P.sub.B]) is examined. Suppose that the initial analysis results in a hypothesis consistent with, say, a separate origin for the agastoparasites, each parasite from its host. To examine whether this result is caused by parasite-host convergences, all apparent apomorphies uniting the parasites with their respective hosts (i.e., the characters supporting the branches indicated by arrows in figure 1a) are removed from the data matrix. Now, if phylogenetic analysis of the reduced data matrix results in a loss only of resolution in the best estimate of the phylogeny, such that branches supported by the characters removed are collapsed but other parts of the tree remain intact, then it is likely that the removed characters support true relationships. However, if the removal results in a completely different phylogenetic hypothesis, the suspicion that the removed characters represent convergences in conflict with true relationships is strengthened. In the special case illustrated in figure 1c, one also would have to examine whether the shift in topology was caused by parasite-parasite convergences. To do this, all apparent synapomorphies supporting the monophyly of the parasites (the branch indicated by an arrow in fig. 1c) would have to be removed in addition to the characters already excluded. Again, if the removal results in a loss only of resolution, it is likely that the removed characters support true relationships, whereas if the removal results in a topology shift, the removed characters are indicated to be convergences in conflict with true relationships. The entire test procedure is summarized in table 1.

In this paper, I test the different scenarios for the evolutionary origin of cynipid inquilines by a phylogenetic analysis of representative species of cynipid gall inducers and inquilines based on adult morphological characters. I pay particular attention to the possible influence of convergent similarities on the phylogenetic results. Among other techniques, I use the method of removing possible convergent characters outlined above.

TABULAR DATA OMITTED

Biology and Systematics of Cynipid Wasps

Before turning to the analysis I will summarize current knowledge of cynipid gall inducers and inquilines. The gall inducers are conveniently divided into five groups, based partly on morphological similarities and partly on host plant preferences (Ashmead 1903a,b; Kinsey 1920): Cynipini, Rhoditini, Aylacini, Eschatocerini, and Pediaspidini. Cynipini is a species-rich and phylogenetically diverse group of gall inducers confined to Fagaceae, with the vast majority of species occurring on Quercus. Rosa is the only plant genus attacked by gall inducers belonging to Rhoditini. Aylacini is a somewhat heterogeneous group of species that gall a variety of plants, such as Papaver, Rubus, and Potentilla, but the majority of species occur on plants in the family Asteraceae. The remaining groups, Eschatocerini and Pediaspidini, are recognized for two morphologically and biologically aberrant genera of gall formers. Eschatocerus is a genus of South American gall inducers on Acacia and Prosopis (Fabaceae), and Pediaspis includes one Palaearctic species on Acer.

The galls of Rhoditini and Cynipini belong to the most complex and well-organized insect galls known, whereas the galls of Aylacini are generally less complex (Kinsey 1920; Askew 1984). Rhoditini and Cynipini galls are frequently attacked by cynipid inquilines, but in the Aylacini only some species of Diastrophus are known to host such inquilines. No cynipid inquilines are known from galls of Eschatocerus or Pediaspis.
TABLE 2. Overview of cynipid gall inducers based on data in Ashmead (1903a,b),
Weld (1952) and Diaz (1980).

                    No. of
Tribe               genera                    Galls on

Cynipini              40              Fagaceae (mostly Quercus)
Rhoditini              2(*)           Rosa (Rosaceae)
Aylacini              14              Asteraceae, Rosaceae, and others
Eschatocerini          1              Acacia, Prosopis (Fabaceae)
Pediaspidini           1              Acer (Aceraceae)

* Including the genus Liebelia.


Cynipid inquilines have phytophagous larvae that develop inside the galls of other cynipids. Although the inquilines cannot induce galls from normal plant tissue, they can modify the host gall tissue such that larval chambers similar to those of normal cynipid galls are formed around the developing larvae (Shorthouse 1973, 1980; Wiebes-Rijks 1980). The host gall is often conspicuously modified by the inquiline attack, either enlarged (Shorthouse 1973, 1980; Wiebes-Rijks 1980) or dwarfed (Nielsen 1903; Washburn and Cornell 1981; Wiebes-Rijks 1982). Many inquilines kill the host larva early in the development of the gall, but the host larva is apparently not eaten by the inquiline (Evans 1965, 1967; Shorthouse 1973, 1980; Wiebes-Rijks 1980). Some inquilines develop in the peripheral tissues of the host gall and appear to do little harm to the original gall inducer (Mayr 1872). However, the original gall inducer may still suffer from the inquiline attack in these cases, for example, through increased risk of succumbing to fungus infections (Wiebes-Rijks and Shorthouse 1992).

Systematically the inquilines are either placed in a separate subfamily of Cynipidae (Hartig 1840; Ashmead 1896, 1903a; Burks 1979b), or included in the Aylacini (Roskam 1992), or lumped together with the gall inducers in a large group without subdivisions (Weld 1952; Eady and Quinlan 1963).

The cynipid inquilines are restricted to galls TABULAR DATA OMITTED on trees and bushes; cynipid galls on herbs are not attacked. Among the inquilines, Synergus, Saphonecrus, and Ceroptres occur in oak galls produced by gall wasps in the tribe Cynipini; Periclistus is assorted with Diplolepis (Rhoditini) galls on roses, and Synophromorpha is found in Diastrophus (Aylacini) galls on Rubus. The genera Rhoophilus and Synophrus probably also belong to the cynipid inquilines. These genera were originally described as gall inducers (Hartig 1843; Mayr 1881), but they are morphologically similar to Synergus and Saphonecrus, and the biology of Rhoophilus and Synophrus has never been subject to detailed investigations. It is therefore quire possible that they represent inquilines incorrectly identified as gall inducers; inquilines are prone to be mistaken as gall inducers because of their life histories (Weld 1952; Wiebes-Rijks 1980).

In addition to the true cynipid inquilines, there is another group of cynipoids that have been roared from the galls of other insects. This group might be called the "figitoid inquilines" and includes Euceroptres, Thrasorus, Myrtopsen, Pegacynips, Plectocynips, and an undescribed genus. The species in this group occur in both cynipid and chalcidoid galls. Some of the genera of figitoid inquilines are currently placed in Cynipidae, others in the parasitoid family Figitidae, and Thrasorus has alternated between the two families (Weld 1952; Barbotin 1963; Rick 1970, 1971; Diaz 1976; Nordlander 1984; Ritchie 1984; Naumann 1991). However, the characters that place figitoid inquilines in Cynipidae are either clearly symplesiomorphies, like the primitive segmentation of the adult metasoma (Weld 1952), or of dubious value, as exemplified by the "hypopygial spine" (Riek 1971). The latter is a feature that occurs in almost all cynipoids but is well developed only in some gall wasps, some cynipid, and some figitoid inquilines, and in a few other unrelated groups of cynipoids. The only strong character bearing on the question known at present, namely the position of the Rs + M vein in the fore wing, indicates that the figitoid inquilines do not belong to Cynipidae but instead form a monophyletic group together with a large subsection of the cynipoid parasitoids, that is, the group formed by the current families Figitidae, Anacharitidae, Charipidae, and Eucoilidae. The biology of the figitoid inquilines has not been studied in detail, and it is quite possible that the figitoid "inquilines" are really parasitoids on some insect in the galls from which they have been reared and not true inquilines. The figitoid inquilines will not be considered further in this paper.
TABLE 4. Cynipid gall inducers and inquilines selected for analysis, and the
relation between the genera to which they belong. Abbreviations of genera used
in figures 2, 4, and 6 are given in parentheses.

Tribe of
gall inducer          Gall inducer                        Inquiline

Cynipini       Andricus quercusradicis (And)
               Neuroterus numismalis (Neu)        Synergus crassicornis (Sye)
               Biorhiza pallida (Bio)             Ceroptres arator (Cer)

Rhoditini      Diplolepis rosae (Dip)             Periclistus brandtii (Per)

Aylacini       Diastrophus turgidus (Dia)         Synophromorpha terricola
(Syo)                Isocolus rogenhoferi (Iso)         none
               Aylax papaveris (Ayl)              none
               Phanacis phoenixopodos (Pha)       none


MATERIALS AND METHODS

For the initial survey, dried specimens representing all genera of known or presumed cynipid and figitoid inquilines were studied in addition to a comprehensive sample of cynipid gall inducers and cynipoid parasitoids. For the detailed analysis of cynipid inquilines and gall inducers, 12 species were selected to represent morphologically different types of inquilines, gall inducers hosting these inquilines and gall inducers without associated inquilines. Gall-inducing hosts for the studied inquiline species are as follows (Eady and Quinlan 1963; Ritchie and Shorthouse 1987): Synergus crassicornis, Andricus fecundatrix; Ceroptres arator, Andricus quercusradicis and Andricus kollari; Periclistus brandtii, Diplolepis rosae; Synophromorpha terricola, Diastrophus bassetti and Diastrophus radicum. Ibalia rufipes (Ibaliidae), a parasitoid on wood-boring siricid larvae, was used as the out-group. Although ibaliids are not very closely related to cynipids, the unresolved status of higher-level relationships in the Cynipoidea makes it difficult to find a better outgroup at present (cf. Ronquist and Nordlander 1989).

Material preserved in 70% ethanol was used except for Diastrophus and Synophromorpha in which only dried material secondarily transferred to 70% ethanol was available. External structure of female body pans, antennae, wings, legs, mouthparts, and ovipositor, and male antennae, metasoma, and genitalia was studied using stereo microscopy, light microscopy, and low-power scanning electron microscopy (except for Synophromorpha in which scanning electron microscopy was not used because of the lack of sufficient material). In addition, some internal skeletal features of the female head and the female and male metasoma were studied by dissection under a stereo microscope. Qualitative differences between species were coded as characters for phylogenetic analysis. Except as noted in table 5, two or three specimens of each species were studied for every character. In Cynipini, there is usually both an agamic and a sexual generation. In Neuroterus numismalis only the agamic generation was studied, in Biorhiza pallida only the sexual generation and in Andricus quercusradicis both generations. In the latter case, the state differed between generations in six characters; in these instances, the species was coded as being polymorphic. Multistate characters were ordered if the states appeared to form a natural sequence (morphocline), otherwise they were left unordered. The state in the outgroup was coded as unknown when it did not correspond to any of the states in the ingroup, In the final matrix phylogenetically uninformative characters were removed (cf. Farris 1991). The data matrix is available on diskette from the author.

PAUP (Phylogenetic Analysis Using Parsimony) version 3.0s (Swofford 1991) was used for computer analyses. Extra steps caused by terminal polymorphisms were included in calculations of tree lengths and fit measures.

RESULTS

The morphological study resulted in a set of 108 informative characters, with a sum of minimum possible lengths of 130 and a sum of maximum possible lengths of 474. Fitch-Wagner parsimony analysis of the data using the branch-and-bound algorithm of PAUP produced five minimum-length trees. The trees had a length of 228 steps, a CI (consistency index) of 0.57, and a RI (retention index) of 0.72. The strict consensus tree of the five trees is shown in figure 3.

The result of the analysis shows considerable stability to alternative methods of analyzing the data. Running all characters as unordered (Fitch parsimony) produced the same set of five trees as in the original analysis. All but one of the branches in the consensus tree were supported in at least 99% of 1000 bootstrap replications of the analysis performed with the branch-and-bound algorithm of PAUP. All branches in the consensus tree were stable to successive weighting, whether based on best values of retention indices, consistency indices, or rescaled consistency indices.

The analysis indicates that the oak gall wasps in the tribe Cynipini, here represented by Andricus, Neuroterus, and Biorhiza, form a monophyletic group and that they are closely related to Diplolepis, the rose gall wasps. Aylacini is shown to be a paraphyletic group, comprising some of the basal lineages of gall wasps. The monophyly of the inquilines is particularly well supported by the data. At least 17 unambiguous character changes support this grouping in the most parsimonious trees, and it takes 15 extra steps to break up the monophyly of the inquilines. In 1000 bootstrap replications of the analysis, the inquilines appeared as a monophyletic group in nearly all, or 99.98%. The close relationship between Diastrophus and the inquilines is also well supported by the data. At least 10 character changes support the monophyly of Diastrophus and the inquilines; this grouping occurred in 98.8% of the 1000 bootstrap replications of the analysis; and it takes eight extra steps to break up the group.

To evaluate more directly the support or lack of support for the polyphyletic hypothesis, the length of the shortest tree consistent with the inquilines originating independently from their respective cynipid hosts was calculated. This was done by using the branch-and-bound algorithm of PAUP and keeping only trees satisfying the constraint that Periclistus + Diplolepis, Synophromorpha + Diastrophus and Synergus + Ceroptres + Andricus + Neuroterus + Biorhiza appeared as monophyletic groups. Six trees were obtained, the consensus tree of which is shown in figure 4. The trees have a length of 321, which is shorter than a random tree (100,000 trees drawn randomly from a universe of all possible, fully resolved trees had a mean length of 413 steps), but still 93 steps longer than the minimum-length trees.

To examine whether the result of the analysis was influenced by convergent similarities among inquilines caused by their similar mode of life, I removed possible convergent characters from the analysis. First, I excluded all characters with one state (putative synapomorphy but possible convergence) occurring in all the inquilines but not in the gall inducers or the outgroup. These characters, if homoplastic, represent strong convergent trends in that they show up in all inquilines. In multistate ordered characters, the state for the inquilines was replaced with the closest state in the transformation series to preserve information that would otherwise have been deleted (see Appendix, first reanalysis). The data matrix now had an informative variation (cf. Farris 1991) of 425 - 117 steps, that is, 308 steps. Fitch-Wagner parsimony analysis using the branch-and-bound algorithm of PAUP resulted in five trees with a length of 215, CI = 0.54 and RI = 0.68. The five trees were identical to the five trees obtained with the original data.

In a subsequent analysis I also excluded characters with a state occurring in some inquilines but not in the gall inducers or the outgroup (see Appendix, second reanalysis). These characters, if not true indicators of relationship, represent weak convergent trends in that they are expressed only in some inquilines. Multistate ordered characters were treated as before. The informative variation of the data matrix was now reduced to 411 - 111 steps, that is, 300 steps. For this matrix, there were six most-parsimonious trees of length 207, CI = 0.54, RI = 0.68. Three of the trees were identical with trees obtained with the original data; the other three differed in placing Diastrophus within the inquilines. None of the trees broke up the monophyly of the group Diastrophus + inquilines.

Returning to the original data. I wanted to determine exactly how much of the nonrandomness in the shortest trees consistent with the polyphyletic hypothesis was actually caused by the assumption of a polyphyletic origin of the inquilines from their gall-inducing hosts. To do this, I started with an unresolved bush with the maximum possible length for the data, 474 steps. Grouping Synergus with Ceroptres and Biorhiza with Neuroterus and then with Andricus reduced the length of the tree to 359 steps. These are groupings within inquilines and within gall inducers and are not directly relevant to the testing of the polyphyletic hypothesis. Now, grouping inquilines with their hosts according to the polyphyletic hypothesis reduced the tree length to 337, that is, by another 22 steps. Almost all of that, 18 steps, is caused by the grouping together of Diastrophus and Synophro-morpha. Thus. there are almost no data to support the origin of Periclistus from Diplolepis or the origin of Synergus and Ceroptres from the Cynipini. By contrast, the grouping together of all inquilines reduced the length of the bush by 95 steps to a length of 379.

DISCUSSION

Cynipid Phylogeny and the Origin of the Inquilines

An obvious conclusion from the analysis presented here is that the polyphyletic hypothesis for the origin of cynipid inquilines is incorrect. The polyphyletic scenario conflicts with the morphological data, and the conflict is not caused by convergent similarities among inquilines because of their similar mode of life but by a lack of support for the groupings predicted by the polyphyletic hypothesis.

Two other conclusions emerge from the present analysis. First, the cynipid inquilines form a monophyletic group, which means that the step from the gall-inducing to the inquiline strategy must have been taken by the stem species of the inquilines. Second, the inquilines evolved from gall inducers in or related to the genus Diastrophus, one of the current host groups.

The family Cynipidae is speciose, and phylogenetic relationships hardly have been studied at all (but see Kinsey 1920). Therefore, it was necessary in this study to use exemplar taxa. An important concern in drawing evolutionary conclusions from an exemplar study is the placement of taxa not included in the analysis. Will the conclusions hold when the analysis is expanded to include more and eventually all taxa? If we assume a certain phylogenetic homogeneity, if not strict monophyly, of cynipid genera, the prime concern is for genera not included in the present analysis.

Among the inquilines, only Rhoophilus, Saphonecrus, and Synophrus were not represented in the analysis. The strongest evidence for monophyly of the inquilines selected for the present study is provided by the 11 character states (in characters 1, 5, 8, 14, 31, 33, 68, 74, 87, 95, 104; see Appendix and table Al) that were uniquely derived on all most-parsimonious trees. Of these 11 synapomorphies. Saphonecrus (at least the type-species Saphonecrus connatus) shares 10 (all but character 104), Rhoophilus shares 8 (all but characters 8, 95, and 104), and Synophrus shares 8 (all but characters 68, 74, and 104). Furthermore, Synophrus and Saphonecrus appear to share several apomorphies with Synergus, notably in the structure of the petiole and in the fusion of the third and fourth abdominal tergum in the male, and Rhoophilus seems to be the sister group of Synergus + Synophrus + Saphonecrus (Ronquist, unpubl. data). Thus, inquiline monphyly almost certainly will hold even if all genera of inquilines are considered. Furthermore, the suggested relationships imply that even if Synophrus politus and Rhoophilus loewi turned out to be true gall inducers instead of inquilines, one would have to conclude that they secondarily reverted from being inquilines back to being gall inducers.

Although relatively few representative species of gall inducers were included in the present analysis, the result is clearly indicative of an origin of the inquilines from gall inducers related to Diastrophus. It is particularly the strong support for the close relationship between Diastrophus and the inquilines, and the basal or subbasal placement within the inquilines of Synophromorpha, the inquiline associated with Diastrophus, that support this conclusion.

Additional evidence indicates that it is likely that Diastrophus is the genus of gall inducers most closely related to the inquilines. Diastrophus was unique among gall inducers hosting inquilines included in this analysis in showing a close relationship to the inquilines; the other genera hosting inquilines (in the groups Rhoditini and Cynipini) seem to belong to a different lineage of cynipids. Furthermore, Rhoditini and Cynipini are both morphologically and biologically rather homogeneous groups (Weld 1952), making it unlikely that anomalous gall inducers that are closely related to the inquilines will be discovered in these groups.

Weld (1952) recognized 15 genera of Aylacini in the most recent generic revision of this heterogeneous group. Of these genera, only four were included in the present analysis. Few characters have been studied across a comprehensive sample of species belonging to Aylacini. One of them, however, is the presence or absence of a basal tooth or lobe on the claw. In the analysis reported here, the presence of a basal tooth is inferred to be a synapomorphy uniting Diastrophus and the inquilines (character 82 in Appendix, fig. 39). Among the Aylacini a toothed claw is found only in a few genera, namely Diastrophus, Xestophanes, and Gonaspis (Eady and Quinlan 1963; Quinlan 1968; own observation). These genera are also unique among Aylacini gall inducers in being associated with host plants in the Rosaceae. Furthermore, cynipid inquilines are restricted to galls on trees and bushes, and Diastrophus is the only genus of Aylacini that includes some species associated with bushes; all other genera occur exclusively on herbs. Again, these facts indicate that Diastrophus might actually be the genus of gall inducers most closely related to the inquilines.

The relationships of Eschatocerini and Pediaspidini are unknown, but they show some morphological and biological affinities to Rhoditini and Cynipini (Weld 1952; Folliot 1964). None of these genera have toothed claws, and none of them host inquilines.

To summarize then, although much work remains to be done on the phylogeny of cynipids, there is at least some support now for a monophyletic origin of the cynipid inquilines from gall inducers in the genus Diastrophus, as hypothesized by Ritchie (1984). Askew's (1984) and Gauld and Bolton's (1988) contention that the cynipid inquilines form a polyphyletic group is correct only in the sense that some anomalous elements, that is, some figitoid inquilines, were previously included among the cynipid inquilines. When these elements are removed, the cynipid inquilines form a natural, monophyletic group.

Even if the figitoid inquilines were considered, it is clear that inquilinism has not evolved many times among cynipids. A maximum of six times seems possible: one for the true cynipid inquilines, and one each for Euceroptres, Thrasorus, Myrtopsen, Pegacynips + Pegacynips and the undescribed genus of figitoid inquilines. However, considering the appreciable structural similarity among genera of figitoid inquilines, two or three origins of inquilinism would seem more plausible. Furthermore, as has been mentioned, there is uncertainty about the biology of the figitoid inquilines; if the figitoid "inquilines" are really parasitoids, then inquilinism has evolved only once in Cynipoidea; that is, in the true cynipid inquilines.

Evolution of Agastoparasitism

Convergent evolution will undoubtedly remain a major theme in the study of agastoparasites in general (Wilson 1971; Michener 1977, 1978; Alexander 1990; Fisher and Sampson 1992). However, the difficulties that these convergences pose for phylogenetic inference based on morphological characters may have been overestimated. It remains to be shown that convergent trends are strong enough in any group of agastoparasites to mislead ordinary parsimony analysis of a representative sample of characters, be they morphological or molecular. Techniques for removing possible convergent characters based on the distribution of their states among taxa should be particularly useful for studying the influence of convergent evolution on phylogenetic inference in these organisms.

Are there any general patterns in the evolution of agastoparasitism? Social parasitism is by far the best studied form of agastoparasitism. According to Wilson (1971; see also Holldobler and Wilson 1990), the single most important generalization about the evolution of social parasitism in aculeate wasps is that it follows "Emery's rule," that is, the parasites are more closely related to their host species than to any other free-living form. This is interpreted to imply an extreme version of the polyphyletic scenario, in which almost every single species of parasite had an independent origin from its host (see also Le Masne 1956; Pearson 1981). Such a pattern would be produced if agastoparasitism were associated with high extinction and low speciation rates, that is, if agastoparasitism was an evolutionary dead end. Undoubtedly, there are several groups to which Emery's rule applies (e.g., Holldobler and Wilson 1990), but it is now clear that it does not hold for agastoparasites in general. For instance, in addition to the cynipid inquilines, major radiations of agastoparasites have occurred in cleptoparasitic bees (Alexander 1990; Rozen 1991) and cleptoparasitic sphecid wasps (Bohart and Menke 1976). Even among social parasites there are notable exceptions to Emery's rule, for instance, in bumble bees (Plowright and Stephen 1973; Williams 1985; Pamilo et al. 1987). In many cases, the phylogenetic relationships between the parasites and their hosts have not been determined, and whether Emery's rule applies in these cases is still an open question.

An apparent general feature in the evolution of agastoparasitism is that the parasites originated from one of their hosts (Askew 1971; Packer 1986), and cynipid inquilines seem to fall into this pattern. In other words, the stem species of the agastoparasites was originally associated with its sister species or another very closely related species. Two models have been proposed for the origin of such parasitism between very closely related species. Wheeler (1919) suggested that intraspecific parasitism, in which some individuals parasitize other individuals of the same species, could lead to sister-species parasitism through sympatric speciation. Wilson (1971), however, considered this to be an unlikely scenario and instead proposed that the species first separated through allopatric speciation, and then upon secondary sympatry, one of the species evolved to become a parasite of the other. Sympatric speciation is controversial among some biologists, but intraspecific parasitism is known to occur in several groups with agastoparasites (e.g., Eickwort 1975; Packer 1986; Field 1992), and several workers favor the sympatric model for the origin of agastoparasitism (Buschinger 1986, 1990; Bourke and Franks 1991). Interestingly, the two alternative models predict different macroevolutionary patterns. If the sympatric model is correct, the original parasite and its host must be sister species, whereas this is not the case with the allopatric model. Actually, if allopatric speciation is the dominant mode of speciation, then most cases of sympatry will not involve sister species. Therefore, the original parasite and its host would not be expected to be sister species under the allopatric model. Future studies will have to indicate which of these patterns is more common in the evolution of agastoparasitism. For the cynipid inquilines, more detailed studies are clearly needed before it is possible to distinguish between the two models.

The selection pressures involved in the origin of agastoparasitism are poorly understood in many cases (e.g., Eickwort 1975). Several authors perceive severe competition for a limited resource as a key factor (Wheeler 1919; Pearson 1981). Competition would be aggravated by the competitors being closely related, thus sharing similar ecological requirements and similar constraints in their abilities to adopt alternative strategies. However, there is a possibility that mutualism or mutualistic forces are involved, at least in some cases. For instance, the prevalent model for the evolution of social parasitism in aculeate wasps includes a tendency for young mated females to become adopted into old colonies as an important first step (Wilson 1971; Holldobler and Wilson 1990). A necessary pre-condition for the evolution of parasitism is that it be difficult for young mated females to start new nests on their own such that they do better (or at least as well) when they try to enter old nests. However, the evolution of parasitism would be greatly facilitated, especially in the early stages, if it were also advantageous for established colonies to adopt young queens (cf. Brockmann 1993). Such an advantage is easily perceived, for example, if the young queen's workers contributed to nest building and defense.

Evolution of Inquilinism

Even though the origin of cynipid inquilinism appears to be a unique historical event, it is tempting to speculate about possible adaptive mechanisms involved. Assuming a common origin with some species of Diastrophus, cynipid inquilines would have evolved from gall inducers on Rubus bushes. Unfortunately, little is known about the biology of Diastrophus or the Diastrophus-Synophromorpha association. Some general patterns in the biology of cynipid gall formers suggest that competition for oviposition sites might be involved. Cynipid gall inducers are highly organ specific (Rohfritsch 1992), especially those on trees and bushes (Askew 1984). Tree and bush gallers also tend to have short activity periods synchronized with host plant development (Shorthouse 1973; Wangberg 1975; Washburn and Cornell 1981). Thus, resources for tree and bush gallers are restricted both in time and space. However, this does not necessarily lead to increased competition, and although competition has been suggested as an important force in structuring oak gall wasp communities (Askew 1984), field studies have not revealed severe intra- or interspecific competition among oak gall inducers (Washburn and Cornell 1981; Hails and Crawley 1991). Furthermore, under the competition model, it is difficult to explain why inquilinism has not originated repeatedly among cynipid tree and bush gall inducers. The strong concentration in space and time of cynipid galls on trees and bushes may be more relevant as a factor in explaining why the inquilines have not colonized cynipid galls on herbs; these galls may simply be too dispersed in space and time to support inquiline populations (cf. Wcislo 1987).

An alternative model for the origin of cynipid inquilinism is that it evolved from mutualistic communal oviposition behavior. The species of Diastrophus hosting inquilines all induce multichambered galls on Rubus bushes. In galls of Diplolepis rosae and Diastrophus kincaidii, both multichambered bush galls, survivorship of the gall inducer is positively correlated with gall size, which in turn is determined by the number of larval chambers in the gall (Jones 1983; Stille 1984). Furthermore, larger galls produce larger individuals and larger individuals are more fecund, at least in D. rosae (Schroder 1967; Stille 1984). Thus, in these species there seems to be a selection pressure for large galls. If individual females cannot themselves produce galls of optimal size, selection may favor aggregation of females at oviposition sites. Inquilinism could then originate if some individuals, through random genetic drift, lost the capability to induce galls on their own. Alternatively, an inquiline strategy could evolve if females arriving late gained more in fitness by displacing existing eggs or larvae in favorable sites in the cluster than by laying their eggs in unoccupied but less favorable sites. Communal oviposition has been observed in Diastrophus kincaidii (Wangberg 1975; Jones 1983), but this species does not have any inquilines associated with it, and it is unknown if other species of Diastrophus display the same behavior. Whatever the significance of this observation, it is clear that further study of the biology of Diastrophus and the Diastrophus-Synophromorpha association could produce important new insights into the evolutionary origin of cynipid inquilinism.

ACKNOWLEDGMENTS

I am grateful for technical assistance by G. Wife and input from several undergraduate students of entomology at Uppsala University. N. Fergusson, J. L. Nieves, G. Prinsloo, and J. Read generously provided specimens for the study. I owe special thanks to J. L. Nieves for bringing the case of Aulacidea nigripes to my attention. Early versions of the manuscript were much improved thanks to comments from J. Heraty, J. Huber, A. Kuris, G. Nordlander, E. Sandnes. M. Sharkey, M. Siddall, C. Solbreck, B. G. Svensson, and an anonymous reviewer. Part of the work was performed at the Canadian National Collection of Insects in Ottawa during a one-year visit sponsored by the Sweden-America Foundation.

LITERATURE CITED

Alexander, B. 1990. A cladistic analysis of the nomadine bees (Hymenoptera: Apoidea). Systematic Entomology 15:121-152.

Ashmead, W. H. 1896. Descriptions of new parasitic Hymenoptera. Transactions of the American Entomological Society (Philadelphia) 23:179-234.

-----. 1903a. Classification of the gall-wasps and the parasitic cynipoids, or the superfamily Cynipoidea. III. Psyche (Cambridge) 10:140-155.

-----. 1903b. Classification of the gall-wasps and the parasitic cynipoids, or the superfamily Cynipoidea. IV. Psyche (Cambridge) 10:210-215.

Askew, R. R. 1971. Parasitic insects. Heinemann Educational Books, London.

-----. 1984. The biology of gall wasps. Pp. 223-271 in T. N. Ananthakrishnan, ed. Biology of gall insects. Edward Arnold, London.

Barbotin, F. 1963. Sur une nouvelle galle et deux nouveaux cynipides en provenance d'Algerie. Marcellia 31:151-157.

Bischoff, H. 1927. Biologie der hymenopteren. Springer, Berlin.

Blair, K. G. 1949. Ceroptres arator Hartig (Hym., Cynipidae): inquiline or parasite? Entomologist's Monthly Magazine 85:254-255.

Bohart, R. M., and A. S. Menke. 1976. Sphecid wasps of the world. University of California Press, Berkeley.

Bourke, A. F. G., and N. R. Franks. 1991. Alternative adaptations, sympatric speciation, and the evolution of parasitic, inquiline ants. Biological Journal of the Linnean Society 43:157-178.

Brockmann, H. J. 1993. Parasitizing conspecifics: comparisons between Hymenoptera and birds. Trends in Ecology and Evolution 8:2-4.

Burks, B. D. 1979a. Eupelmidae. Pp. 878-889 in K. V. Krombein, P. D. Hurd, Jr., D. R. Smith, and B. D. Burks, eds. Catalog of Hymenoptera in North America north of Mexico, vol. 1. Smithsonian Institution Press, Washington, D.C.

-----. 1979b. Cynipoidea. Pp. 1045-1107 in K. V. Krombein, P. D. Hurd, Jr., D. R. Smith, and B. D. Burks, eds. Catalog of Hymenoptera in North America north of Mexico, vol. 1. Smithsonian Institution Press. Washington, D.C.

Buschinger, A. 1986. Evolution of social parasitism in ants. Trends in Ecology and Evolution 1:155-160.

-----. 1990. Sympatric speciation and radiative evolution of socially parasitic ants -- heretic hypotheses and their factual background. Zeitschrift fur Zoologische Systematik und Evolutionsforschung 28:241-260.

Couturier, A. 1949. Observations sur Rhyssa approximator F. cleptoparasite de Rh. persuasoria L. Bulletin de la Societe Entomologique de France 54:62-63.

Dettmer, H. 1930. Beschreibung von funf neuen Cynipidenarten, worunter eine neue Gattung. Broteria, Serie Zoologica 26:54-68.

Diaz. N. B. 1975. Anotaciones sobre cinipoideos argentinos III (Hymenoptera). Neotropica (La Plata) 21:32-36

-----. 1976. Estudio ecologico y sistematico de cinipoideos neotropicales I. (Hymenoptera). Plectocynips longicornis gen. y sp. n. Neotropica (La Plata) 22:99-102

-----. 1980. Cinipoideos galigenos e inquilinos de la Republica Argentina. Revista de la Sociedad Entomologica Argentina 39:221-226.

Eady, R. D., and J. Quinlan. 1963. Hymenoptera: Cynipoidea Key to families and subfamilies and Cynipinae (including galls). Handbooks for the Identification of British Insects 8(1a):1-81.

Eickwort. G. C. 1975. Gregarious nesting of the mason bee Hoplitis anthocopoides and the evolution of parasitism and sociality among megachilid bees. Evolution 29:142-150.

Evans, D. 1965. The life history and immature stages of Synergus pacificus McCracken and Egbert (Hymenoptera: Cynipidae). Canadian Entomologist 97: 185-188.

-----. 1967. The bisexual and agamic generations of Besbicus mirabilis(Hymenoptera:Cynipidae), and their associate insects. Canadian Entomologist 99: 187-196.

Farris, J. S. 1991. Excess homoplasy ratios. Cladistics 7:81-91.

Field, J. 1992. Intraspecific parasitism as an alternative reproductive tactic in nest-building wasps and bees. Biological Reviews of the Cambridge Philosophical Society 67:79-126.

Fisher, R. M., and B. J. Sampson. 1992. Morphological specializations of the bumble bee social parasite Psithyrus ashtoni (Cresson) (Hymenoptera: Apidae). Canadian Entomologist 124:69-77.

Folliot, R. 1964. Contribution a l'etude de la biologie des cynipides gallicoles (Hymenopteres, Cynipoidea). Annales des Sciences Naturelles, Zoologie (Serie 12) 6:407-564.

Gauld, I., and B. Bolton, eds. 1988. The Hymenoptera. Oxford University Press, Oxford.

Girault, A.A. 1937. New naturals unorthodoxies and non-pollutions viz new hexapods. Privately printed, Brisbane, 3 pp. Cited from pp. 326-328 in G. Gordh, A. S. Menke, E.C. Dahms and J. C. Hall, eds. The privately printed papers of A. A. Girault. Memoirs of the American Entomological Institute (Gainesville) 28 (1979).

Hails, R. S., and M. J. Crawley. 1991. The population dynamics of an alien insect: Andricus quercuscalicis (Hymenoptera: Cynipidae) Journal of Animal Ecology 60:545-562.

Harris, R.A. 1979. A glossary of surface sculpturing. California Department of Food and Agriculture Laboratory Services-Entomology Occasional Papers 28:1-31.

Hartig, T. 1840. Uber die Familie der Gallwespen. Zeitschrift fur die Entomologie, herausgegeben von Ernst Friedrich Germar 2:176-209.

-----. 1843. Zweiter Nachtrag zur Naturgeschichle der'Gallwespen. Zeitschrift fur die Entomologie, herausgegeben von Ernst Friedrich Germar 4:395-422.

Hedicke, H., and G. J. Kerrich. 1940. A revision of the family Liopteridae (Hymenopt., Cynipoidea). Transactions of the Royal Entomological Society of London 90:177-225.

Holldobler, B., and E. O. Wilson. 1990. The ants. BelknapPress, Cambridge, Mass.

Jones, D. 1983. The influence of host density and gall shape on the survivorship of Diastrophus kincaidii Gill. (Hymenoptera: Cynipidae). Canadian Journal of Zoology 61:2138-2142.

Kinsey, A. C. 1920. Phylogeny of cynipid genera and biological characteristics. Bulletin of the American Museum of Natural History 42:357-402.

LaSalle, J. 1987. New World Tanaostigmatidae (Hymenoptera, Chalcidoidea). Contributions of the American Entomological Institute 23:1-181.

LeMasne, G. 1956. Recherches sur les fourmis parasites. Plagiolepis grassei et l'evolution des Plagiolepis parasites. Comptes Rendus Hebdomadaires des Seances de l'Academie des Sciences, Paris 243: 673-675.

Mayr, G. 1872. Die Einmiethler der mitteleuropaischen Eichengallen. Verhandlungen der Zoologisch-Botanischen Gesellschaft in Wien 22:669-726.

-----. 1881. Die Genera der gallenbewohnenden Cynipiden. Jahresberrichte der Communal-Oberrealschule im I. Bezirke 20:1-38.

Michener. C. D. 1970. Social parasites among African allodapine bees (Hymenoptera, Anthophoridae, Ceratinini). Zoological Journal of the Linnean Society 49: 199-215.

-----. 1977. Discordant evolution and the classification of allodapine bees. Systematic Zoology 26: 32-56.

-----. 1978. The parasitic groups of Halictidae (Hymenoptera, Apoidea). University of Kansas Science Bulletin 51:291-339.

Naumann, I. D. 1991. Hymenoptera. Pp. 916-1000 in The insects of Australia, 2d ed. Melbourne University Press, Carlton, Victoria.

Nielsen, J. C. 1903. Untersuchungen uber Lebensweise einiger Arten der Cattung Synergus. Allgemeine Zeitschrift fur Entomologie 8:35-37.

Nordlander, G. 1982. Identities and relationships of the previously confused genera Odonteucoila, Coneucoela, and Trichoplasta (Hymenoptera, Cynipoidea: Eucoilidae). Entomologica Scandinavica 13: 269-292.

-----. 1984. What do we know about parasitic cynipoids (Hymenoptera)? Entomologisk Tidskrift 105:36-40 [In Swedish].

Osten Sacken, C. R. 1865. Contributions to the natural history of the Cynipidae of the United States and of their galls. Proceedings of the Entomological Society of Philadelphia 4:331-380.

Packer, L. 1986. The biology of a subtropical population of Halictus ligatus. IV. A cuckoo-like caste. Journal of the New York Entomological Society 94: 458-466.

Pamilo, P., A Pekkarinen, and S.-L. Varvio-Aho. 1981. Phylogenetic relationships and the origin of social parasitism in Vespidae and in Bombus and Psithyrus as revealed by enzyme genes. Pp. 37-48 in P. E. Howse and J.-L. Clement, eds. Biosyste-matics of social insects. Systematics Association, Special Vol. 19. Academic Press, London.

-----. 1987. Clustering of bumblebee subgenera based on interspecific genetic relationships (Hymenoptera, Apidae: Bombus and Psithyrus). Annales Zoologici Fennici 24:19-27.

Pearson, B. 1981. The electrophoretic determination of Myrmica rubra microgynes as a social parasite: possible significance in the evolution of ant social parasites. Pp. 75-84 in P. E. Howse and J.-L. Clement. eds. Biosystematics of social insects. Systemaries Association, Special vol. 19. Academic Press, London.

Plowright, R. C., and W. P. Stephen. 1973. A numerical taxonomic analysis of the evolutionary relationships of Bombus and Psithyrus (Apidae: Hymenoptera). Canadian Entomologist 105:733-743.

Quinlan, J. 1968. Cynipinae (Hymenoptera) occuring on Phlomis L. Transactions of the Royal Entomological Society of London 120:275-286.

Richards, O. W. 1927. The specific characters of the British bumblebees. Transactions of the Royal Entomological Society of London 75:233-268.

-----. 1977. Hymenoptera. Introduction and key to families, 2d ed. Handbooks for the Identification of British Insects 6(1):1-100.

Riek, E. F. 1970. Hymenoptera. Pp. 867-959 in I. M. Mackerras, ed. The insects of Australia. Melbourne University Press, Carlton, Victoria, Australia.

-----. 1971. A new subfamily of cynipoid wasps (Hymenoptera: Cynipoidea) from Australia. Pp. 107-112 in Entomological essays to commemorate the retirement of Professor K. Yasumatsu. Hokuryukan, Tokyo.

Ritchie, A. J. 1984. A review of the higher classification of the inquiline gall wasps (Hymenoptera: Cynipidae) and a revision of the Nearctic species of Periclistus Forster. Ph.D. diss. Carleton University, Ottawa.

Ritchie, A. J., and J. D. Shorthouse. 1987. Revision of the genus Synophromorpha Ashmead (Hymenoptera: Cynipidae). Canadian Entomologist 119: 215-230.

Rohfritsch, O. 1992. Patterns in gall development. Pp. 60-86 in J. D. Shorthouse and O. Rohfritsch, eds. Biology of insect-induced galls. Oxford University Press, New York.

Ronquist. F., and G. Nordlander. 1989. Skeletal morphology of an archaic cynipoid, Ibalia rufipes (Hymenoptera: Ibaliidae). Entomologica Scandinavica Supplement 33:1-60.

Roskam, J. C. 1992. Evolution of the gall-inducing guild. Pp. 34-49 in J. D. Shorthouse and O. Rohfritsch, eds. Biology of insect-induced galls. Oxford University Press. New York.

Rozen, J. G., Jr. 1991. Evolution of cleptoparasitism in anthophorid bees as revealed by their mode of parasitism and first instars (Hymenoptera: Apoidea). American Museum Novitates 3029:1-36

Schroder, D. 1967. Diplolepis (= Rhodites) rosae (L.) (Hym: Cynipidae) and a review of its parasite complex in Europe. Commonwealth Institute of Biological Control Technical Bulletin 9:93-131.

Shorthouse, J. D. 1973. The insect community associated with rose galls of Diplolepis polita (Cynipidae, Hymenoptera). Quaestiones Entomologicae 9:55-98.

-----. 1980. Modification of galls of Diplodepis polita by the inquiline Periclistus pirata. Bulletin de la Societe Botanique de France Actualites Botaniques 127:79-84.

Spradbery, J. P. 1969. The biology of Pseudorhyssa sternata Merrill (Hym., Ichneumonidae), a cleptoparasite of siricid woodwasps. Bulletin of Entomological Research 59:291-297.

Stille, B. 1984. The effect of hostplant and parasitoids on the reproductive success of the parthenogenetic gall wasp Diplolepis rosae (Hymenoptera, Cynipidae). Oecologia 63:364-369.

Swofford, D. L. 1991. PAUP: phylogenetic analysis using parsimony, Version 3.0s. Computer program distributed by the Illinois Natural History Survey, Champaign.

Varvio-Aho, S.-L., P. Pamilo, and A. Pekkarinen. 1984. Evolutionary genetics of social wasps (Hymenoptera, Vespidae, Vespula). Insectes Sociaux 31:375-386.

Wangberg, J. K. 1975. Biology of the thimbleberry gallmaker Diastrophus kincaidii. Pan-Pacific Entomologist 51:39-48.

Washburn, J. O., and H. V. Cornell. 1981. Parasitoids, patches, and phenology: their possible role in the local extinction of a cynipid gall wasp population. Ecology 62:1597-1607.

Wcislo, W. T. 1987. The roles of seasonality, host synchrony, and behaviour in the evolutions and distributions of nest parasites in Hymenoptera (Insecta), with special reference to bees (Apoidea). Biological Reviews of the Cambridge Philosophical Society 62:515-543.

Weld, L. H. 1926. Field notes on gall-inhabiting cynipid wasps with descriptions of new species. Proceedings of the United States National Museum 68:1-131.

-----. 1952. Cynipoidea (Hym.) 1905-1950. Privately printed, Ann Arbor, Mich.

Wheeler, W. M. 1919. The parasitic Aculeata, a study in evolution. Proceedings of the American Philosophical Society 58: 1-40.

Wiebes-Rijks, A. A. 1980. The identity of the gallwasp causing stunted acorns (Hymenoptera, Cynipidae). Netherlands Journal of Zoology 30:243-253.

-----. 1982. Early parasitation of oak-apple galls (Cynips quercusfolii L., Hymenoptera). Netherlands Journal of Zoology 32:112-116.

Wiebes-Rijks, A. A., and J. D. Shorthouse. 1992. Ecological relationships of insects inhabiting cynipid galls. Pp. 238-257 in J. D. Shorthouse and O. Rohfritsch. eds. Biology of insect-induced galls. Oxford University Press, New York.

Williams, P. H. 1985. A preliminary cladistic investigation of relationships among the bumble bees (Hymenoptera; Apidae). Systematic Entomology 10: 239-255.

Wilson, E. O. 1971. The insect societies. Belknap Press, Cambridge, Mass.

APPENDIX

Characters used for phylogenetic analyses. Morphological terminology follows Ronquist and Nordlander (1989) with additional terms from Nordlander (1982), Richards (1977), and Hedicke and Kerrich (1940). Sculptural terms are from Harris (1979). Most character states are illustrated in figures 7-50. Abbreviations (x refers to a number): Fx, flagellomere x; xtg, abdominal tergum x; xst, abdominal sternum x. Transformation series hypotheses are given for multistate characters.

Head, Anterior View, Female

1. Shape of ventral clypeal margin: (a) rounded, projecting over mandibles; (b) straight, not projecting over mandibles.

2. Clypeo-pleurostomal lines: (a) present, at least visible by different sculpture; (b) absent.

3. Epistomal sulcus: (a) present, at least marked by a distinct change in curvature of lower face; (b) absent.

4. Facial strigae radiating from clypeus: (a) present, at least close to clypeus; (b) absent.

5. Extent of facial strigae: (a) laterally terminating well before compound eye (not illustrated); (b) laterally reaching or almost reaching compound eye.

6. Subocular impression: (a) absent; (b) present.

7. Shape of antennal rim: (a) distinctly widened lateroventrally; (b) not distinctly widened lateroventrally.

8. Distance between antennal socket and eye: (a) longer than maximum width of antennal sockets including rim; (b) shorter than maximum width of antennal sockets.

9. Size of antennal sockets: (a) small ratio of maximum width including antennal rim to maximum width between outer margins of eves [less than]0.13; (b) large, ratio [greater than]0.13.

10. Orbits: (a) flat, not impressed; (b) slightly but distinctly impressed.

11. Lateral frontal carinae: (a) present; (b) absent.

12. Shape of surface of upper face: (a) flat; (b) slightly raised medially; (c) distinctly raised medially. Ordered abc.

13. Sculpture on vertex dorsad compound eye: (a) punctate or irregular; (b) acinose-colliculate.

Head, Posterior View, Female

14. Distance between occipital and oral foramina (a) shorter than height of occipital foramen including postoccipital rim; (b) longer than height of occipital foramen.

15. Gular sulci and gular ridges: (a) united well before reaching hypostomata; (b) free, but meeting at hypostomata; (c) free, well separated at hypostomata. Ordered abc.

16. Median hairy strip of gula: (a) broad; (b) narrow; (c) absent. Unordered.

17. Shape of ventral part of hypostoma: (a) not or only slightly projecting from cranial margin, italy slightly raised; (b) distinctly projecting from cranial margin, distinctly raised.

18. Depression laterad hypostoma, close to ventral margin of cranium: (a) absent; (b) present.

Mandibles, Female

19. Number of teeth with corresponding internal rods on right mandible: (a) three; (b) two.

20. Basal swelling on anterior side of mandible: (a) small or indistinct; (b) large, conspicuous.

21. Size of oval window: (a) large; (b) small or almost absent.

22. Shape of dorsal margin of upper tooth on left mandible: (a) evenly rounded, no extra tooth indicated; (b) produced into an extra tooth.

23. Posterior region of ventral surface of mandible: (a) horizontal, set off from posterior surface of mandible by a distinct carina; (b) oblique, gradually continued in posterior surface of mandible.

Labiomaxillary Complex, Female

24. Longitudinal, mesal carina on posterior surface of stipes: (a) present; (b) absent. 25. Shape of cardo: (a) bent distally such that part is visible in posterior view of head; (b) straight, not bent distally, not visible in posterior view of head.

26. Shape of apical peg on last segment of maxillary and labial palpus: (a) long and narrow, situated subapically; (b) short and broad, situated apically.

27. Shape of apical segment of maxillary palpus: (a) distinctly asymmetrical; (b) almost symmetrical.

28. Articulation between fourth and fifth segments of maxillary palpus: (a) normal, free articulatio; (b) fifth segment rigidly inserted into fourth; (c) fifth and fourth segment fused . Ordered abc.

29. Length of third segment of maxillary palpus: (a) short, ratio of length of third segment to length of second segment [less than]0.90; (b) long, ratio [greater than]0.95.

30. Length of second segment of maxillary palpus: (a) short, ratio of length of second segment to length of third to fifth segment combined [less than]0.5; (b) long, ratio [greater than]0.5.

31. Shape of first segment of maxillary palpus: (a) subrectangular. broader than long; (b) triangular, longer than broad.

32. Number of segments of labial palpus: (a) three normal segments; (b) three segments, second strongly reduced in size; (c) two segments. Ordered abc.

33. Shape of first segment of labial palpus: (a) long, gradually tapering towards base; (b) short, abruptly tapering towards base.

Female Antenna

34. Number of completely separated flagellomeres; (a) 10; (b) 11; (c) 12; (d) 13. Ordered abcd.

35. Length of F1: (a) short, ratio of length of F1 to length of F2 [less than or equal to]1.00; (b) long, ratio [greater than or equal to]1.10.

Male Antenna

36. Number of flagellomeres: (a) 13; (b) 12.

37. Length of F1: (a) short, ratio of length of F1 to length of F2 [less than]1.25; (b) long, ratio [greater than]1.40.

38. Size of excavated part of F1: (a) involving major part of F1; (b) involving only small basal part of F1.

39. Longitudinal ridge on F1: (a) absent; (b) present, extending part of length of F1; (c) present, extending entire length of F1. Ordered abc.

Pronotum, Female

40. Admedian depressions of pronotum: (a) separated medially; (b) united medially, forming a transverse impression anteriorly on the pronotum.

41. Shape of dorsal pronotal margin medially: (a) sharp, with a distinct dorsal edge; (b) blunt, rounded dorsally.

42. Shape of pronotum: (a) long medially ratio of median to posterior distance between dorsal and ventral margins of pronotum [greater than or equal to]0.22: (b) short medially, ratio [less than]0.20.

43. Lateral margin of posterior part of pronotal plate: (a) marked entirely; (b) marked only ventrally; (c) not marked Ordered abc.

44. Lateral pronotal carina: (a) present: (b) absent.

45. Sculpture on lateral surface of pronotum: (a) irregular, without linear component; (b) with irregular, horizontal costulae; (c) with many regular, horizontal costulae . Unordered.

46. Shape of lateroventral margin of pronotum: (a) distinctly concave: (b) straight or very slightly concave.

47. Angle of lateroventral margin of pronotum: (a) oblique, ratio of vertical distance between highest and lowest points of ventral pronotal margin to dorsal length of pronotum [greater than]0.85; (b) less oblique, ratio [less than]0.85.

Prosternum, Female

48. Position of profurcal pit: (a) anterior to middle of furcasternum; (b) at or behind middle of furcasternum.

49. Shape of profurcal pit: (a) rounded, small; (b) transverse, large.

Mesonotum, Female

50. Impression mesad parascutal carina: (a) anteriorly ending just in front of tegula; (b) anteriorly continuing to anterior end of notaulus.

51. Shape of anterolateral margin of mesoscutum and dorsolateral margin of pronotum: (a) mesoscuturn not projecting over pronotum, dorsolateral part of pronotum not impressed; (b) mesoscutum projecting over pronotum, pronotum impressed along its dorsolateral margin; (c) mesoscutum not projecting but partly overhanging pronotum, pronotum impressed at a point just anterior to the tegula . Unordered.

52. Surface sculpture of mesoscutum : (a) glabrous, at least medially; (b) dull.

53. Transverse ridges on mesoscutum: (a) present; (b) absent.

54. Pubescence on mesoscutum : (a) about as dense as pubescence laterally on pronotum; (b) distinctly less dense; (c) hairs almost absent. Ordered abc.

55. Lateral bar: (a) present; (b) absent .

56. Posterodorsal margin of axillula: (a) not marked: (b) distinctly marked.

57. Shape of subaxillular bar: (a) broad, vertical, evenly continuing posteriorly in shining strip of scutellum; (b) narrow, horizontal, rapidly expanding posteriorly in shining strip of scutellum.

Mesopectus (Mesopleuron and Mesosternum), Female

58. Sculpture on speculum: (a) glabrous with or without punctures; (b) rugulose ; (c) longitudinally, horizontally costate-costulate. Unordered.

59. Shape of posterior part of dorsal margin of speculum: (a) rounded, peak at or immediately anterior to posterior subalar pit, margin posterior to peak straight; (b) pointed, peak some distance anterior to posterior subalar pit, margin posterior to peak concave.

60. Shape of acetabular carina medially: (a) absent, not marked; (b) distinct, only slightly raised ; (c) distinct, strongly raised. Ordered abc.

61. Shape of surface of acetabulum posteromedially, adjacent to acetabular carina: (a) slightly longitudinally raised; (b) flat.

62. Longitudinal carina from mesocoxal foramen towards acetabular carina, laterally delimiting mesosubpleuron: (a) absent; (b) present.

63. Shape of rim surrounding mesocoxal foramen: (a) narrow throughout or slightly expanded posteriorly, ratio of posterior width to anterior width [less than]2.0; (b) distinctly expanded posteriorly, ratio [greater than]2.0.

64. Position of mesocoxal foramen: (a) removed from posterior margin of mesosubpleuron, ratio of distance between posterior margin of mesocoxal foramen and posterior margin of mesosubpleuron to longitudinal width of mesocoxal foramen [greater than]0.32; (b) close to posterior margin of mesosubpleuron, ratio [less than]0.32.

Metanotum, Female

65. Shape of metascutellum: (a) subrectangular: (b) distinctly constricted medially. 66. Sculpture on bar ventral to metanotal trough: (a) at least
COPYRIGHT 1994 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1994 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ronquist, Fredrik
Publication:Evolution
Date:Apr 1, 1994
Words:9584
Previous Article:DNA evolution and colonization sequence of island lizards in relation to geological history: mtDNA RFLP, cytochrome B, cytochrome oxidase, 12s rRNA...
Next Article:Genetics of physiological differentiation within the marine mussel genus Mytilus.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters