Position of Cassidinae within Chrysomelidae

Latreille (1802) first proposed the family name Chrysomelines for leaf beetles. Chapuis (1874) subdivided Chrysomelines into four sections and 15 “tribes”, with his section IV, Cryptostomes, consisting of two tribes, Hispides and Cassidides. Subsequent attempts to refine this arrangement include splitting or combining groups and altering ranks at all levels (Jacoby, 1908; Pierce, 1916; Chen, 1940, 1964; Gressitt, 1942b; Chûjo, 1953; Monrós, 1959; Gressitt and Kimoto, 1963), but Chapuis' (1874) system and group concepts are still used today (Seeno and Wilcox, 1982).

Cassidinae s.str. and Hispinae s.str. have been traditionally regarded as forming a single clade, Cryptostomes, either most closely related to Chrysomelidae or as one or more groups within Chrysomelidae. Viewpoints on the position, rank, and formal names of Cassidinae s.str. and Hispinae s.str. are summarized in table 1.

A hypothesized relationship between Cassidinae s.str. and Coccinellidae (Gage, 1920; Westwood, 1920) never gained support. However, other authors have commented on the resemblance between cassidine and coccinellid larvae and pupae (Paterson, 1931, 1941); some cassidine species were even described in Coccinella (Fabricius, 1775).

When Cassidinae s.str. and Hispinae s.str. have been considered as a distinct clade most closely related to Chrysomelidae, they have been ranked as a family or superfamily, and termed Cassidiadae, Cassidiformes, Cassididae, or superfamily Cassidoidea. When treated as members of Chrysomelidae, Cassidinae s.str. and Hispinae s.str. have been ranked as the section Cryptostoma or division Fronticornis (Chapuis, 1875). They have also been ranked as a single subfamily (Hispinae or Cassidinae), as two distinct subfamilies, or as two distinct tribes called Cassididae s.str., Hispidae s.str., Cassidinae s.str., or Hispinae s.str. (these names were used before ICZN regulated the “ini” suffix for tribes). Based on larval data, Lee (1993) proposed a dissident view that Cassidinae s.str. and Hispinae s.str. are not each other's closest relatives within Chrysomelidae.

Except for Reid (1995) and Hsiao and Windsor (1999), these views on phylogenetic concepts, ranks, and nomenclature have been universally based on limited taxon samples, few characters (table 2), personal views on hypothetical ancestral groundplans, and a priori decisions on polarity (primitive versus derived). Many interpretations are essentially single character system phylogenies. For example, Verma's view (1985, 1988, 1992, 1996) is based only on male genitalia, and Suzuki's view (1969a, 1969b, 1970a, 1970b, 1994, 1996) is based on hindwing venation.

Table 2

Support for Relationships of Cassidinae s.str. and Hispinae s.str.

Current phylogenetic analyses emphasize a more comprehensive approach that uses extensive taxon samples, outgroup comparisons, multiple character sources (morphology, molecules, behavior and ecology), and global parsimony analysis (Kitching et al., 1998; Schuh, 2000). Modern character argumentation requires more rigorous homology assessment with consideration of outgroups (Nixon and Carpenter, 1993).

Recent higher level studies of chrysomelid phylogenetics (Lee, 1993; Hsiao, 1994a, 1994b, 1994c; Reid, 1995, 2000; Farrell, 1997; Duckett et al., 2004) provide a framework for addressing lower level problems. These studies are relevant to phylogenetic relationships in Cassidinae and are discussed briefly below.

Lee (1993) provided the first cladistic analysis of chrysomelid subfamily relationships. Thirty-four larval morphological characters were coded for 77 Japanese species (17 chrysomelid subfamilies; 0.21% species diversity) and a hypothetical ancestor. Monophyly of the hispine–cassidine clade was not supported, and Alticinae, Galerucinae, and Chrysomelinae were identified as possible relatives. Some problems of Lee's (1993) analysis include a priori character polarization, ambiguous character state demarcations (e.g., states of characters 2, 15 and 22 overlap) and analysis with an obscure program by Sawada. Reanalysis of these data by Reid (1995) produced different topologies from those published by Lee (1993).

Hsiao (1994a) presented an unrooted chrysomelid phylogeny with the relationship ((Cassidinae s.str. + Hispinae s.str.) + Orsodacninae) + other subfamilies. His immediate goal was to draw attention to sequence data, namely 12S mtDNA, as a tool in chrysomelid systematics. Only 11 chrysomelid species (representing 11 chrysomelid subfamilies) were sampled and no outgroups were used. Hsiao (1994b) again promoted the use of 12S mtDNA and 16S mtDNA in chrysomelid systematics. His sample of 13 Chrysomelid genera included 2 cassidine genera but no outgroups; under both parsimony and neighbor-joining analyses, these cassidines were placed as sister taxa, basal to a large clade that included Galerucinae s.l., Chrysomelinae, Criocerinae, Synetinae, Megalopodinae and Orsodacninae.

Reid's analysis (1995) marked a significant advance in the quality of chrysomelid phylogenetic analyses. He used outgroups (Curculionoidea and Cerambycoidea) for the first time, sampled all chrysomelid subfamilies, reviewed traditional morphological characters, introduced many new character systems (71 characters total), sampled an unprecedented number of taxa (ca. 0.2% species diversity [C. Reid, personal commun.]), and used current phylogenetic methods consistently (i.e., repeating the analysis recovered published topologies [Schmitt, 1996; Chaboo, unpubl. data]). Reid did not design his 1995 study to address the historical controversy surrounding (1) Cassidinae s.str. + Hispinae s.str. and (2) Alticinae s.str. + Galerucinae s.str., and he assumed them to be single monophyletic terminals (Hispinae s.l. and Galerucinae s.l., respectively) prior to analysis. Weighted analysis of the adult dataset resulted in the relationship Eumolpini + Hispinae s.l. Weighted analysis of the larval dataset resulted in a tritomy (Criocerini + Hispinae s.l. + Chrysomelinae). The unweighted analysis of combined adult and larval datasets resulted in Hispinae s.l. appearing as sister to the clade Galerucinae s.l. + Chrysomelinae. Various permutations of weighted characters were used on the combined data, with three results for Hispinae: sister to Galerucinae s.l. + Chrysomelinae, sister to Chrysomelinae, or sister to Criocerinae. Reid (1995) proposed that Hispinae s.l. and Criocerinae were probable sister taxa, but he did not present strong evidence. Reid (2000) re-analyzed his dataset by omitting non-chrysomeloid outgroups, correcting mis-scored characters, adding missing data, and exploring the effects of partitioning and weighting. The position of Hispinae s.l. remained inconclusive in all analyses of larval, adult and combined data. Criocerinae was suggested as the sister group based on plesiomorphic support.

Farrell (1998) presented a combined analysis of morphological data of Chrysomelidae (Reid, 1995) and Curculionidae (weevils) (Kuschel, 1995) with his own 18S nuclear ribosomal DNA data for 115 species (ca. 0.09 % of Phytophaga species diversity). Species from six genera of Cassidinae were sampled (ca. 0.02 % generic diversity). The preferred topology suggested the relationship Cassidinae + (Donaciinae + Criocerinae + Pachymerinae + Amblycerinae + Bruchinae). Reid (2000) critically discussed problems of taxon and character sampling, theory, analysis, and interpretation in Farrell's (1998) work. Additionally, Farrell (1998) stated that sequences were aligned using Sequencher 3.0 (Genecodes, 1995); however, this program only shows raw chromatographs. One could produce coarse alignments using Sequencher 3.0, but more rigorous programs like Clustal (Thompson et al., 1997), MALIGN (Wheeler and Gladstein, 1994, 1994–2000), or POY 3.0 (Wheeler et al., 2002) are commonly used for alignment. Farrell and Sequeira (2004) subsequently expanded Farrell's (1998) study by adding 11 sequences for eight new in-group genera and four more outgroup genera, further refining the morphological characters of Reid (1995, 2000) and incorporating the morphological dataset of Svacha et al. (1997). Various analyses were performed with mixed results for the sister taxon and placement of Cassidinae (table 2). A most surprising result was the separation of Imatidium from other cassidines under Bayesian analyses of the molecular data.

In another higher level chrysomelid analysis, Duckett et al. (2004) combined sequence data from small subunit ribosomal DNA, 18S nuclear ribosomal subunit DNA (Farrell, 1999) with the morphological data of Reid (1995, 2000). Seven cassidine species from four tribes were included in their 100 species sample. Results suggested that Cassidinae was paraphyletic with respect to Cryptocephalinae. The placement of the cassidine Imatidiini as sister to Cryptocephalinae appears to corroborate Farrell and Sequeira (2004).

Genomic data are a powerful tool in evolutionary biology, and recent chrysomelid studies have used 12S mtDNA sequences (Hsiao and Windsor, 1999), 18S nuclear DNA (Farrell, 1995), and small subunit ribosomal DNA (Duckett et al., 2004). These are the earliest explorations of a wide pool of potential molecular markers, and it is premature to determine what each sequence together or in combination reveals about chrysomelid phylogenies or at what level they are most effective. In some other insect groups, where molecular systematics is more advanced, it is evident that certain sequences work better at lower or higher taxonomic levels because of differential rates of evolution. For example, Danforth et al. (2004) have found that mitochondrial sequences are not the best choice for recovering deep divergences in bees.

All of the recent cladistic studies help narrow the range of phylogenetic hypotheses for higher-level chrysomelid relationships. They collectively provide a context for lower level research, as evidenced by studies of Galerucinae s.l. (Lingafelter and Konstantinov, 2000; Kim et al., 2003; Gillespie et al., 2003; Duckett et al., 2004; Gillespie et al., in press), new world Criocerinae (Vencl and Morton, 1999; Vencl et al., 2004), Cryptocephalinae (Gómez-Zurita et al., 1999; Termonia et al., 2006), Cassidinae (Borowiec, 1995; Hsiao and Windsor, 1999; McKenna and Farrell, 2005), and Eumolpinae (Gómez-Zurita et al., 2005).

Cassidinae s.str. and Hispinae s.str.: Two Distinct Clades?

The boundary is ambiguous between Cassidinae s.str. and Hispinae s.str. because a number of taxa exhibit intermediate features (Chapuis, 1875; Maulik, 1919; Crowson, 1953, 1955, 1967), particularly in the immature stages (Sanderson, 1902; Böving and Craighead, 1931; Gressitt and Kimoto, 1961a, 1961b; Lawrence, 1982, 1991; Jolivet and Hawkeswood, 1995). In these publications, cassidine larvae were considered as external leaf-feeders, whereas hispine larvae were considered as miners or cryptic feeders. The larval eight abdominal spiracles were considered vestigial in Cassidinae s.str. but developed in Hispinae s.str. Sanderson (1902) stated that cassidine larvae have the last abdominal tergum modified with a caudal process whereas hispine larvae lack this modification, and cassidines retain fecal shields on these processes whereas hispine larvae carry no shields. These distinctions were conceptualized early in the history of Cassidinae and have not been reevaluated or have become invalid as more taxa have become known.

Placement of certain taxa is controversial because their adults and larvae display a combination of these classical cassidine and hispine features. Such enigmatic taxa include the hispoid Oediopalpa negligens (Weise) whose larvae are free-living and retain a shield (Bruch, 1906). Notosacanthini (broadly flattened animals) was long classified in Cassidinae s.str. based on adult morphology; discovery of their mining larvae led to the tribe's transfer to Hispinae s.str. (Medvedev and Eroshkina, 1988). Borowiec (1999) nonetheless maintained Notosacanthini under Cassidinae s.str. Plautyauchenia Sturm was placed in Cassidinae s.str., but larval morphology suggests its position in the hispine Alurnini Weise (Maulik, 1933683). Placement of Delocrania Guérin-Ménéville in Cassidinae s.str. has also been questioned (Maulik, 1919, 1933683), although all subsequent classifications of Cassidinae s.str. have included Delocrania.

Borowiec (1995) presented the first modern phylogenetic analysis of Cassidinae. This study used 19 binary characters—14 adult morphological characters, 3 larval morphological characters, 1 larval behavioral character, and 1 host plant character. The morphological characters, derived mostly from Hincks (1952), were scored for nine cassidine tribes, two hispine tribes and an all-zero hypothetical ancestor. The use of higher taxa as terminals does not test monophyly of those groups and limits the range of characters that could be sampled. The single published tree (fig. 3) indicated that Cassidinae s.str. and Hispinae s.str. were paraphyletic with respect to each other. Five of the 19 characters were uninformative autapomorphies, which do not resolve relationships (Hennig, 1966). Borowiec (1995) synonymized Cassidinae s.str. and Hispinae s.str. as Hispinae s.l. without discussing the name priority. In his subsequent world catalog of Cassidinae s.str., Borowiec (1999) reerected the traditional name Cassidinae sensu Seeno and Wilcox (1982), omitted the tribe Imatidiini, and rearranged the 15 remaining tribes into 12 tribes. No argument was made for his novel arrangement nor any explanation offered for the obvious contradictions with his earlier phylogenetic concept and nomenclature.

Figure 3

Phylogenetic hypothesis for Cassidinae proposed by Borowiec (1995) based on analysis of 12 tribal terminals, a hypothetical ancestor as outgroup, and 19 characters (5 autapomorphies, 11 binary and 3 multi-state).

Hsiao and Windsor (1999) used 12S mtDNA in their phylogenetic analysis of 48 species representing 18 tribes of Cassidinae s.str. and Hispinae s.str., and they presented two alternative hypotheses of Cassidine relationships (fig. 4A, B). They concluded that Cassidinae s.str. and Hispinae s.str. were each paraphyletic and synonymized them as Hispinae s.l. This study initiated the use of molecules for cassidine systematics; however, problems in analysis and taxon sampling undermine the proposed hypothesis. Of the ca. 450–535 aligned bases used it is unclear how many were actually informative. A single outgroup, Donaciinae, was included, despite the equal likelihood of multiple possibilities from other cladistic studies (e.g., Lee, 1993; Reid, 1995; Farrell, 1998) and the apparent contradiction with Hsiao's (1994a) earlier study that resolved Orsodacninae as sister to Cassidinae s.str. + Hispinae s.str. Outgroup selection influences polarity of ingroup states, so a different selection or, even better, an expanded set of outgroups could change their topology. By selecting only the aquatic Donaciinae, Hsiao and Windsor (1995) constrained their analysis to support a conclusion of an aquatic ancestor of Hispinae s.l. + Donaciinae. The ingroup taxon sample comprised one Old World species and 47 Panamanian species, including four Stolas Billberg species, three Charidotella Weise species, and three Cassida species. These latter three genera were found to be polyphyletic, a result suggesting that the sequence choice may not be conservative enough to be phylogenetically meaningful at tribal levels. The sampling of 18 of the 43 recognized tribes, and the bias towards Panamanian taxa, missed important taxonomic and geographic diversity. Both parsimony and neighbor-joining were used to develop tree topologies. The consensus derived by parsimony was poorly resolved with a large basal polytomy (fig. 4A). Instead of reevaluating their data, Hsiao and Windsor (1999) did a neighbor-joining analysis and selected that result as the basis of their phylogenetic and evolutionary discussions (fig. 4B). Farris et al. (1996) have already demonstrated inadequacies in the neighbor-joining method (e.g., finding only one exact solution) and the existence of better methods (jackknifing).

Figure 4

A, B. Phylogenetic hypotheses for Cassidinae proposed by Hsiao and Windsor (1999) based on 48 taxa, Donaciinae as outgroup, and 12s mtDNA data. C. Phylogeny presented by Wilf et al. (2000).

Wilf et al. (2000) altered the Hsiao and Windsor (1999) neighbor-joining topology by adding terminals to represent six chrysomelid outgroup subfamilies and four Cassidine tribes, Delocraniini, Chalepini, Sceloenoplini, and Uroplatini (fig. 4C). These taxa were inserted into the topology of Hsiao and Windsor (1999) based on opinion, and no sequence data were submitted to Genbank (as of 2007). Although no reanalysis was done, phylogenetic structure was found among outgroups. The host plant evolutionary scenario constructed by Wilf et al. (2000) is intriguing, and it is unfortunate that this Science cover story was built on a contrived dataset.

Borowiec (1995) and Hsiao and Windsor (1999) provided two alternative phylogenetic hypotheses:

Currently, there are several competing classification schemes for Cassidinae s.str. and Hispinae s.str.: the classical one (Seeno and Wilcox, 1982), two recent schemes based on weakly supported phylogenetic hypotheses (Borowiec, 1995; Hsiao and Windsor, 1999), and the scheme proposed in the catalog of Cassidinae s.str. (Borowiec, 1999).

Fundamental Nomenclatural Issues

The lack of a well-supported phylogenetic hypothesis has created many nomenclatural problems. Borowiec (1995) and Hsiao and Windsor (1999) arrived at different topologies, but agreed that Cassidinae s.str. and Hispinae s.str. were paraphyletic, and that the unified clade should be called Hispinae s.l. Borowiec (1995) provided no argumentation for selecting the name Hispinae s.l. over Cassidinae. Hsiao and Windsor (1999) justified their choice as a convention supposedly established by Crowson (1955), Lawrence and Britton (1994) and Reid (1995). (Crowson [1955] is a collection of papers that appeared in the Entomologist's Monthly Magazine during the 1950s; chrysomelid relationships were discussed in Crowson [1953] and that text was reproduced exactly in Crowson [1955, 1967]). Crowson (1953) used the name Hispinae for Cassidinae + Hispinae, but in subsequent publications he recognized both Cassidinae s.str. and Hispinae s.str. as distinct entities (Crowson, 1981; Mann and Crowson, 1981, 1983, 1986). In any case, the first revisor of Chrysomelidae was Chen (1940) and we must follow his convention (article 24.2, ICZN [2000]).

Lawrence and Britton (1994) recognized Cassidinae, but Lawrence (1982, 1991) recognized Cassidinae s.str. and Hispinae s.str. Reid's (1995) use of Hispinae s.l. for Cassidinae s.str. + Hispinae s.str. was for convenience of analysis (C. Reid, personal commun.). “Cryptostomes” appears to be the closest to a conventional name for the clade Cassidinae s.str. + Hispinae s.str.

Several group terms are currently in use: Cassidinae s.str.; Hispinae s.str.; “hispoid Hispinae” (Borowiec, 1995; Hsiao and Windsor, 1999); “cassidoid Hispinae” (Heron and Borowiec, 1997; Hsiao and Windsor, 1999; Wilf et al., 2000). Use of the terms “hispoid Hispinae” and “cassidoid Hispinae” is not recommended as these lack any apomorphic support and are not natural groups. The “-oid” suffix is also poorly since as it implies superfamily rank. As stated earlier, Cassidinae is used here after Chen (1940) for the clade Cassidinae s.str. + Hispinae s.str.

Tribal Relationships of Cassidinae

Ambiguous tribal boundaries have been discussed: Cassidini Hincks, Charidotini Hincks, Aspidimorphini Hincks, and Basiptini Hincks (Riley, 1986; Borowiec, 1999; Riley et al., 2002); Goniocheniini Hincks and Omocerini Hincks (Viana, 1968; Seeno and Wilcox, 1982); Physonotini Hincks, Asterizini Hincks, and Ischyrosonychini Hincks (Seeno and Wilcox, 1982; Boroweic, 1999; Riley et al., 2002); Stolaini Hincks and Eugenysini Hincks (Seeno and Wilcox, 1982); Aproidini Weise and Eurispini Weise (Würmli, 1975a; Samuelson, 1968, 1989); Alurnini Weise (Maulik, 1933a, 1933b); Cephaloleiini Weise and Arescini Weise (Strong, 1977a); Notosacanthini Hincks and Aspidimorphini Hincks (Hawkeswood, 1989); Notosacanthini Hincks, Epistictini Hincks, and Basiprionotini Hincks (Zaitzev and Medvedev, 1982; Borowiec and Świętojańska, 2004); Delocraniini Hincks, Notosacanthini Hincks, Hemisphaerotini Hincks, and Spilophorini Hincks (Monrós and Viana, 1951); and Imatidiini Hincks and Cephaloleiini Weise (Spaeth, 1914, 1938; Bondar, 1940c; Blackwelder, 1946; Monrós and Viana, 1947, 1951; Papp, 1953; Aslam, 1965; Uhmann, 1957; Seeno and Wilcox, 1982; Windsor et al., 1992; Borowiec, 1995, 1999; Hsiao and Windsor, 1999; Borowiec, 2000; Wilf et al., 2000; McKenna and Farrell, 2005). Only the synonymy of Aspidimorphini, Charidotini and Cassidini appears to be generally accepted. Duckett et al. 's (2004) finding of Imatidiini as sister to Cryptocephalinae and separated from other cassidines needs reconsideration.

Confusion about the monophyly, positions, relationships and taxonomic names of groups stems from several factors. Historically there has been no more than one specialist of Hispinae s.str. per generation; the retirement of Charles Staines, the principal current authority, has resulted in his greater productivity. Expertise on Cassidinae s.str. has been marginally better, and currently there are several active taxonomic researchers. In general, a new generation of specialists must be trained, as in any other taxonomic group. Morphological systematics in Cassidinae has been limited and few phylogenetic characters have been proposed. Molecular systematics is a novel and rapidly evolving field but promises an explosion of genomic data. Ecological and behavioral data await exploration. The research tradition has been separated along subfamily lines and/or along geographic lines (Old World versus New World) and thus characters have not been optimized across the entire group, much less across all relevant outgroups. Consequently, traditional characters tend to become ambiguous as more taxa are compared. Expanding and refining the character dataset is critical to resolving systematics issues.

Morphology of Eggs (Fig. 15)

Egg biology has been discussed for ecological guilds of Cassidinae. In general, egg descriptions are available for fewer taxa than for larvae. Oviposition sites may be on the surface or in maternal excavations of stems and leaves, and eggs may be deposited singly, in small groups (20–40 eggs), or in large groups of more than 100 eggs. Eggs may be naked or they may have coverings of colleterial secretions, fecal deposits, or chewed leaf fragments prepared by the mother. Suspended masses of eggs are known so far only in Acromis and Omaspides (Chaboo, 2001). These features can be developed as phylogenetic hypotheses; however, I only include two egg characters here because information is lacking for many taxa.

In species where maternal guarding has been recorded (table 6, fig. 21), females guard all immature stages. Maternal guarding may be a stereotypical behavior, but it is quite complex. As we know more about maternal care and care-providing species, variation in the behavior may be partitioned into multiple characters. Instead of treating maternal care as several characters to represent each stage that is guarded, I used maternal care as a single character for adults. We presently lack detailed comparative information on how guarding might vary from one immature stage to another, or how species and genera vary.

Morphology of larvae (Figs. 16, 18, 20, 24–Figure 2526)

Larval sizes range from 1 to 2 mm up to 6 cm and their color ranges vary from creamy white to black, with some aposematically colored yellow and black forms. First-instar larvae can differ from later instars in color (usually darker), asperities (usually more), spines (more), setation (more), and in possessing egg bursters. Egg bursters occur in many first-instar larvae throughout Insecta and help the larva escape from its eggshell (van Emden, 1946). In taxa lacking egg bursters, larvae presumably exit the egg by chewing their way out (Cox, 1988). In Chrysomeloidea, egg bursters may be found dorsad of spiracles on the thoracic and/or abdominal segments IX (Cox, 1988). Askevold's (1990a) character 25 treated the presence/absence of egg bursters. Reid's (1995) characters 61 and 62 refined the states as presence/absence on specific segments, on the meso- and metathoracic segments only, or on abdominal segment I only, or on additional abdominal segments, up to segment VIII.

These previous state definitions do not take into account that egg bursters vary in segment position (on the head, thorax, or abdomen, or in some other combination) and can comprise spines, setae, or tubercles. Egg burster per se refers to structures that are not homologous among insects (indeed, snakes also emerge with the aid of egg bursters), and the structures covered by the term need to be reviewed and brought into line with phylogenetic concepts. I have used egg bursters on the abdomen according to Reid (1995), but including or excluding this character from the present analysis does not affect the topology.

Generally, cassidine larvae tend to be dorsoventrally flattened, with the extreme condition being found in the platyform “water-penny”-like immatures of rolled-leaf cassidines (Maulik, 1932, 1933683). Variations in head exposure and mouth position relative to the antennal bases appear to parallel the condition on the adult head. The antennomeres vary in number from two to three (Maulik, 1919; Paterson, 1931b). The lateral margins of thoracic and abdominal segments can be expanded and bordered with spinules and hairs, and the prothoracic segment can be explanate frontally, covering the head dorsally. Mining insects appear to show reductions in head, mouthparts, lateral projections, and legs (Frost, 1924; Maulik, 1933683; Ford and Cavey, 1985). Variation and function of larval stemmata have not been widely studied (Gilbert, 1994) but could be diverse across Cassidinae.

Mouthparts. Sanderson (1900) and Paterson (1931) briefly discussed differences of larval mouthparts at the subfamily level and pointed out some unusual features of Donaciinae. Paterson (1931) found resemblances between Lema (Criocerinae) and Cassidinae in setal arrangements. From my preliminary examinations, mouthparts do vary greatly. Triangular unidentate mandibles in hemisphaerotines (Chaboo and Nguyen, 2004) and palmate four-dentate mandibles in Acromis (Chaboo, 2002189) suggest a wide variation in mandible morphology. Preliminary examinations of mouth morphology and observations of nearly stereotypical feeding patterns suggest great variation in mouth morphology and feeding mechanics to deal with the diversity of host plants (from grasses to morning glories) and diversity of feeding modes (boring, mining, skeletonizing, scraping, and chewing).

Thorax. Thoracic segments frequently have scoli (pleural lateral projections) that vary in their number, arrangment, branching, and setation (Gressitt, 1963; Maulik, 1931679680; Zaitsev and Medvedev, 1982). Additionally, the margins of all thoracic segments may be explanate laterally (e.g., Coelaenomenodera; Maulik, 1931679680), paralleling the pattern in adults. Anterior extensions of the prothorax may also cover the head, as in many adults. Thoracic spiracles, especially the prothoracic spiracle, vary in proportions.

Legs. Exophagous Cassidinae always have three-segmented legs but mining forms appear to be more labile in leg development (Roberts, 1930; Maulik, 1926, 1931679680; Jolivet and Hawkeswood, 1995). Miners can have legs soft with reduced sclerotization or vestigial and clawless, for example, Craspedonispa saccharina Maulik (Callan, 1954) and Sceloenopla (Jolivet and Hawkeswood, 1995). Entire segments may be lost, for example, Wallacea dactylifera has two segments (Maulik, 1919). Some mining larval cassidines lack legs altogether, for example, Octotoma and Odontota (Needham et al., 1928; Ford and Cavey, 1985).

Abdomen. The number of larval abdominal segments varies with 8, 9, and 10 segments known. In nonshield-retaining cassidines, there is much confusion and disagreement about the terminal segment and the urogomphi. Mining larvae have been reported as having 8, 9, 10, and 11 segments (Chen et al., 1986); Coleoptera larvae can have up to 10 abdominal segments but never 11 segments (Lawrence, 1982, 1991; Lawrence and Britton, 1994). This information is ambiguous even for other chrysomelid subfamilies; for example, data for Donacia larvae (Donaciinae) are similarly conflicting with eight segments being commonly reported, but nine segments (Schmidt-Schwedt, 1887) and a rudimentary segment X (Sanderson, 1900) are also reported.

The last two abdominal segments of many mining forms are commonly fused, with the suture distinct or reduced. The terminal segment can therefore appear to be one segment or to comprise two segments (fig. 24A, C, F and H). It is frequently heavily sclerotized, sometimes with bladelike (razor sharp in some species) or blunt edges, and/or it has a large dorsal concavity. Maulik (1931)679680 termed this peculiar posterior a “shovel” and suggested that the dorsal concavity may provide a pocket of air for respiration, may be used for slicing within the leaf mine, or it may help keep the mine clear of excreta and debris. It is also possible that these “shovels”, termed here as urogomphal plates, may act as blocking device within a mine.

Figure 24

Larvae of Cassidinae, hind end, dorsal view. A. Anisodera sp. B. Basiprionota sp. C. Botryonopa sp. D. Hemisphaerota palmarum. E. Physonota sp. 1. F. Prosopodonta corallina. G. Coptocycla sp. H. Undetermined mining species. I. Eugenysa columbiana, lateral view. J. Cleared abdomen, lateral view of sagittal section.

The fusion and loss of abdominal segments in Cassidinae poses special problems in discussing homologies of posterior structures. Urogomphi are projections of the tergum of segment IX and they appear in unrelated immatures throughout Coleoptera (Crowson, 1981). Urogomphus may be an umbrella term for a variety of structures found on the terminal abdominal segment. Within Chrysomelidae, Cassidinae appears to be the only group with hind-end projections. More remarkably, Cassidinae exhibits two types, the urogomphal plate (“tail shovels”) and caudal processes. Chaboo and Nguyen (2004) provided an illustrated discussion of the various terminology, morphology, and homology of caudal processes. Tail shovels are not homologous with caudal processes as the former involve one or more segments and form a single functional unit. A case could be made for homologizing spines, small marginal lobes, elongate lobes, or other secondary projections on tail shovels (figs. 24A, C, F, 26B) with caudal processes, but this idea needs more detailed morphological examination. Illustrations of Leptispa by Chen et al. (Chen, 1973; Medvedev, 1982; Chen et al., 1986) show larvae with a single elongate bifid process, which, however, appears on an abdominal segment IX in their illustration. Caudal processes range from one to four, are not articulated, and may be bifid. In addition to numbers, processes vary in length, position, orientation, morphology and setation (Chaboo and Nguyen, 2004).

Figure 26

Anus of mature larval Cassidinae. A. Acrocassis gibbipennis, lateral view. B. Brontispa sp. 1, ventral view. C. Chalepus ater, ventral view.

Scoli. Larval abdominal segments can be similar to paired thoracic pleural projections in having scoli. These may be present or absent across Cassidinae. When present, their lengths are highly variable from tubercle-like to long. Scoli may be simple or branched and smooth or with sharp barb-like setation. Świętojańska et al. (2005) discussed scoli number as a phylogenetic character, with 13 representing the most plesiomorphic condition in “true cassids”, then 14 as the next evolutionary step, and 16 as the most derived. However, scoli number varies within genera (e.g., Chlamydocassis Spaeth; Świętojańska et al., 2005), so any hypothesized transformation series should consider variation at all hierarchic levels.

Spiracles. These vary in number and position. The thoracic spiracles can be positioned laterally or ventrally. Laterally directed prothoracic spiracles were first described in Oediopalpa larvae (Maulik, 1933683). Hemisphaerotines also have similar prothoracic spiracles that are elongate tubes directed laterad and extend slightly beyond the lateral margins of the exuvio-fecal shield (Chaboo and Nguyen, 2004). The fifth abdominal spiracle is long and conical in Dactylispa (Maulik 1929, 1931679680), unlike the short ones in most other cassidine larvae. The terminal pair of abdominal spiracles may be well developed or vestigial, functional or not, lateral or dorsal, and may be biforous or annular (Böving and Craighead, 1931; Jolivet and Hawkeswood, 1995; Chu, 1949; Peterson 1960). Abdominal spiracles, especially the terminal pair, can be laterally or dorsally positioned. Spiracles associated with urogomphal plates may be present and well developed (fig. 24A, E), reduced or absent. When present, they may be positioned on the suture between segments VIII and IX, or laterally like other abdominal spiracles.

Anus. This varies greatly in shape and position across Cassidinae (fig. 26). In shield-retaining cassidines, the anus complex comprises the anal pore, which is at the end of an elongate rectum that is telescoped with segment X. These two sections are narrowed relative to preceding segments (fig. 26A). Eurypepla is unique among shield-retaining cassidine larvae in having segment X and the rectal extension resembling the preceding abdominal segments (Chaboo, 2004; fig. 25). In many nonshield-retaining forms the anus is a narrow transverse ventral slit (fig. 26B) or an apical pore (fig. 26C). The functional significance of the telescoped anus appears to be for shield construction and repairs, but the significance of this wide variation in anal position is unclear. Shrivastava and Verma (1983) briefly examined rectal morphology, and this is worth pursuing to understand how rectal morphology and physiology influence fecal production and shield construction.

Figure 25

The hind end of larval Cassidinae as a character system. A, B. Eurypepla calochroma, instar V. A. Lateral view. B. Dorsal view.

Chaetotaxy could be an important tool in the taxonomy of immatures (Paterson, 1931). Borowiec and Świętojańska (2003) developed a chaetotaxonomy for cassidine immatures but their terminology is different from that utilized for other chrysomelids (e.g., Cox, 1996253) and in Coleoptera. Their study provides a starting point for anatomical study and homology determination.

Modifications for Semiaquatic Living. The larvae of rolled-leaf and bract cassidines are not only extremely dorsoventrally flattened, but they also tend to have the venter and margins densely pubescent (Maulik, 1937). Cephaloleia species vary in having glabrous or pubescent larvae. Larvae of the Old World genus Leptispa also have fine ventral hairs and live in rice paddy fields where they can withstand periodic submersion (Maulik, 1919; Chaboo and Prathapan, unpubl. data). The significance of these hairs (whether locomotion or plastron) is unclear.

Behavior. Gregariousness, cycloalexy, and exuvio-fecal shield retention occur in cassidine larvae and pupae and are complex character systems. Shields (fig. 20) are a particularly interesting aspect of immature stages and are actively being studied by many researchers. Shield retention can vary among larval instars and in pupae and thus very precise observational data are needed for exact scoring as character states. Steinhausen (1969) was the first to define shield-related characters in larvae and pupae of Cassida species; his characters 14 and 28 concerned shield presence or absence and composition of exuviae or feces.

Fecal recycling occurs in five clades of Chrysomelidae: Cassidinae, Cryptocephalinae, Criocerinae, Galerucinae, and Lamprosomatinae (Chaboo et al., 2007). The diverse fecal cases and shields, the life-history stages involved, and the elaborate morphology associated with constructions (anal and rectal modifications) and retention (setation and caudal processes) promise rich ground for developing novel character hypotheses.

Morphology of Pupa (Figs. 17, 19)

Cassidine pupae parallel larvae in their morphology, with a tendency to dorsoventral flattening and with lateral scoli of various sizes. Many morphological characters of larvae can be compared in the pupal stage. Pupation patterns (singles or groups, sites) are also of interest. As far as is known, all exophagous cassidines pupate externally. Cassidines with mining larvae have pupation sites within the mine or externally. Pupae may be naked or they may retain the exuvio-fecal shield of the fifth larvae or only the fifth larval exuviae on its abdomen. Stenispa appears to be unique in having pupation occur within the last larval exuviae (Ford and Cavey, 1985); this genus is also unusual because it has an external feeding larvae, but it is currently classified in the tribe Cephaloleiini, whose members are otherwise cryptic closed leaf feeders.

From my examination of specimens, it appears that the development of caudal processes and shield retention in the pupa cannot be predicted from the larval condition. Pupae may be naked (fig. 17E), may retain the last larval exuviae (fig. 17F), or may retain the shield of the last larva (fig. 19). Shield retention requires caudal process development, but retention of the last larval exuviae does not. When caudal processes are developed, their morphology can be very different from the corresponding larval processes. For example, hemisphaerotine larvae have four caudal processes, but pupae have two (Chaboo and Nguyen, 2004).

Figure 17

Pupae of Cassidinae, dorsal aspect. A. Basiprionota cerata, shield status unknown. B. Eugenysa columbiana, shield removed. C. Hemisphaerota palmarum, shield removed. D. Megapyga sp., fifth larval exuviae removed. E. Microrhopala rubra, no shield retained. F. Physonota helianthi, fifth larval exuviae in situ.

Only two characters are proposed for pupae. Available descriptions of pupae are fewer than those for larvae, and these suffer from the same problems in adequacy and accuracy as in larval descriptions.

Head (Figs. 29–Figure 30Figure 3132, 41–Figure 42Figure 43Figure 44Figure 45Figure 46Figure 47Figure 48Figure 49Figure 50Figure 51Figure 52Figure 5354; Chars. 26–83)

The cassidine head differs from that of other chrysomelids in some fundamental ways, and their structural diversity is exceptional. Because the cassidine head tends to be deeply inserted into the prothorax, complete disarticulation is required for a detailed examination. For this study, head disarticulation was possible for only a subset of the taxa studied. This body region accounts for most of the missing data in the character matrix. Cassidine head and mouthparts are compared briefly below, and a more comprehensive account will be given elsewhere (Chaboo, unpubl. data).

The cassidine head appears to have become shortened and/or retracted into the prothorax. It may be completely exposed or completely hidden in dorsal aspect. Exposure depends on several independent features: head length, head insertion into the prothorax, emargination or extension of the pronotal anterior margin (hiding the head dorsally), and extension of the prosternal anterior margin (partially or completely covering mouthparts ventrally). Researchers have tended to emphasize different aspects of head exposure; for example, Monrós and Viana (1951) focused on the pronotal emargination, whereas Würmli (1975) focused on head length. Promecothecini was diagnosed as having the head short due to retraction up to the eyes (Würmli, 1975).

Heads (figs. 29–30, 41–Figure 42Figure 43Figure 4445) are commonly quadriform in dorsal aspect but may also be rounded (e.g., Callispa, Delocrania, Hemisphaerota, and Spaethiella). In lateral profile, the ventral surface is flattened or diagonal in shape, so that the posterior margin of the mouth is obscured or visible in anterior view (figs. 43, 44). The vertex may be short or long, and smooth, coarse, or punctate. Posteriorly it may be finely transversely striate with the “pars stridens” or stridulatory file (fig. 44A, C). The vertex is commonly dissected longitudinally by the mid-cranial suture, which terminates between the antennal calli or is continuous frontally with the mid-frontal sulcus. The antennal calli ranges from indistinct to well defined and protuberant.

Figure 41

Head, dorsal view. A, D. Chrysomelidae. A. Ophraella sp. 1. B. Pachybrachis circumcincta (in situ). C. Pachybrachis circumcincta (cleared). D. Plataumaris rufa. E, L. Cassidinae. E. Alurnus bipunctatus. F. Aproida balyi (in situ). G. Aproida balyi (exposed). H. Arescus sp. 1. I. Aslamidium sp. 3.

Figure 42

Head, dorsal view. A. Callistola speciosa fasciata. B. Ceratispa palmicola. C. Delocrania panamensis. D. Demotispa sp. 1. E. Exothispa reimeri. F. Hispodonta chapuisi. G. Metriona bilimeki. H. Notosacantha badia.

Figure 43

Head, anterior view. A. Lamprosoma sp. 1. B. Lema trivittata. C. Ophraella sp. 1. D. Pachybrachis circumcincta. E. Plataumaris rufa. F. Sagra sp. 1. G. Calliaspis rubra. H. Jonthonota maculata.

Figure 44

Head of Cassidinae, anterior view. A. Botryonopa foveicollis. B. Callistola speciosa fasciata. C. Physonota sp. D. Plautyauchenia deyrollei. E. Prosopodonta dorsata. F. Stolas lebasi. Scale bars  =  0.5 mm.

Figure 45

Head of Cassidinae, ventral aspect. A. Botryonopa foveicollis. B. Callistola speciosa fasciata. C. Physonota sp. D. Plautyauchenia deyrollei. E. Prosopodonta dorsata. F. Stolas lebasi. Scale bars  =  0.5 mm.

Head Sutures. Frontal and ocular sutures or grooves are important landmarks in determining chrysomelid subfamily relationships (Schmitt, 1985a, 1985b, 1988; Askevold, 1990a, 1990b; Reid, 1995; Santisteban, 1997; Lingafelter and Konstantinov, 2000). In his review of head sutures and terms in Donaciinae, Askevold (1990b) determined that the frontal sutures ( =  frontoclypeal suture) form the lower /\ shape in the X-shape grooves on some chrysomelid heads, namely Sagra. Gular sutures converge anteriad but never meet, terminating at the posterior tentorial pits. In some taxa (e.g., Poecilaspis) they extend to the margin of the mouth. Gular sutures vary in their length, curvation, and orientation so the gular size is also variable. The transfrontal suture (Konstantinov and Vandenberg, 1996) may be present or absent in Cassidinae. The coronal suture (Snodgrass, 1935) has been variously termed in Chrysomelidae: epicranial suture (Rivnay, 1928), median longitudinal groove (Schmitt, 1989), orbital sulcus (Staines, 1989), frontal furrow (Santisteban, 1997), frontal groove (Riley, 1985), median groove (Reid, 1995), and midcranial suture (Lingafelter and Konstantinov, 2000). Alexander Konstantinov's (personal commun.) distinction of the coronal suture into two sections, the midcranial suture posteriorly and the midfrontal sulcus anteriorly, is followed here. Schmitt (1989) indicated the midcranial suture as present in Criocerinae, Donaciinae, and Cassidinae. This sulcus is interrupted in Sagra tristis, which Schmitt (1989) interpreted as a remnant of the condition in other subfamilies. Lingafelter and Konstantinov (2000) however coded the suture as absent in Sagrinae. Reid's (1995) character 8 involving the coronal suture is refined here as three characters: presence/absence of the mid-cranial suture (char. 29), presence/absence of frontoclypeal sutures (char. 37), and presence/absence of antennal tubercles (char. 48). The relationship between the positions of the tubercles and suture is unclear so these are treated as separate characters. It is also ambiguous whether the X-grooves and the coronal groove are homologous so these are also treated separately.

The midfrontal sulcus is commonly present in cassidines, but it is absent in Prosopodonta and Pharangispa. Some cassidines have the sulcus everted as the coronal carina (char. 32). In some (e.g., Asmangulia and Cephalodonta) the coronal carina and midcranial suture are present. This carinate projection has been referred to as a rostrum (Würmli, 1975). The midcranial suture may be slightly concave or deeply sulcate, and short or long.

Eyes may be small or large relative to the head. They are situated anterolaterally, with the facets directed dorsally, ventrally, laterally, and anteriorly. Eye shapes include round and irregular quadrate forms with the margins straight or sinuate. Individual facets may be flat or convex, giving the eye a granular appearance. Usually the eyes are widely separated by the frontoclypeus, but in Plesispa their medial margins are very close due to a narrowed frontoclypeus. The eye margin can be elevated medially (e.g., Notosacantha), posteriorly (e.g., Delocrania), or frontally (e.g., Calliaspis). The elevation is slightly different among these taxa; for example, Notosacantha has a sclerotized ridge around the orbit, whereas in Imatidium, the eye orbit is elevated so much so that the entire eye is projected and the anterior eye margin forms a discontinuous line with the anterior margin of the head. In Callistola, Chalepus and Xenochalepus the vertex is protuberant, projecting the margin toward the antennal insertions.

Frons and Clypeus (figs. 43–Figure 4445; chars. 67–68). This region is variable and complex. Some cassidines appear to lack both frons and clypeus so that the mouth and antennal insertions are proximate or are separated by a thin sclerite. The clypeus is commonly fused to the frons, forming the frontoclypeus. Sometimes the epistomal suture between the frons and clypeus is barely discernible, but most often it is not apparent. The frontoclypeal length varies, influencing the position of the mouth and the gular length. Borowiec's (1993) character 3 has two clypeal states, long and short, however these are not so discrete, but variation in clypeal length is conveyed here as relative to hypostomal length (char. 65). Borowiec (1995: 545) stated “horizontalization of the clypeus is a linear trend in Cassidinae.” Further study and detailed illustrations are needed to understand this trend. The clypeus may be discernible by a faint groove or suture, or by a slight change of angulation, however in many cases it cannot be defined since the anterior tentorial pits are not obvious. Sagittal sectioning may be necessary to determine locations of these structures. The frontoclypeus may be distinct, with developed margins. Anteriorly it may be dissected by the midfrontal sulcus. Its surface may be flat, depressed, or protuberant, and its texture may be smooth, striate, wrinkled, or punctate, and sparsely to densely pubescent.

Antenna (figs. 46–Figure 47Figure 48Figure 4950; chars. 48–64). Cassidine antennal morphology has been used in a limited way in the past. Traditional characters include close insertions of antennae, relative proportions of antennomeres, presence of grooves, weakly or strongly developed club, and fusion and loss of distal antennomeres.

Figure 46

Antennae of Cassidinae. A. Anisotena inaequalis. B. Aproida balyi.C. Arescus sp. 1. D. Arescus sp. 1, antennomere I. E. Aslamidium sp. 1. F. Aspidimorpha miliaris, scanning electron micrograph. G. Asteriza flavicornis, scanning electron micrograph. H. Baliosus californicus. I. Basipta stolida, scale bar  =  1mm. J. Botryonopa foveiocollis. K. Callispa bowringi. L. Callistola speciosa.M. Calyptocephala brevicornis. N. Cephaloleia sp. 1. O. Ceratispa sp. 1. P. Chalepus acuticornis. Q. Cyperispa thorachostachyi kolombangana. R. Dactylispa angusta.

Figure 47

Antennae of Chrysomelidae. A. Delocrania panamensis, antennomere I. B. Delocrania panamensis, antennomere VIII. C. Dorcathispa sp. 1. D. Echoma dichroa. E. Epistictina viridimaculata, scanning electron micrograph. F. Epistictina viridimaculata, apex. G. Lema trivittata. H. Notosacantha badia, scanning electron micrograph. I. Palmispa parallela. J. Philaspis bivulnerata. K. Physonota exclamationis, scanning electron micrograph. L. Prosopodonta dorsata. M. Sceloenopla mantecada. N. Spaethiella sp. 1.

Figure 48

Antenna of Cassidinae, surface features. A. Uroplata girardi. B. Uroplata girardi, apex. C. Xenochalepus sp. 1. D, E. Asteriza flavicornis, surface of scape. F. Baliosus californicus, pore on antennomere VII. G. Calyptocephala brevicornis, groove on antennomere VI. H. Calyptocephala brevicornis, groove on antennomere X. I. Calyptocephala brevicornis, sensilla within groove on antennomere X.

Figure 49

Antenna of Cassidinae, scanning electron micrographs. A. Baliosus sp., antennomere III. B. Physonota sp., base of antennomere I. C. Physonota sp., base of antennomere I. D. Physonota sp., simple sensilla. E. Physonota sp., bifid sensilla. F. Physonota sp., trifid sensilla. G. Microrhopala vittata, antennomere IV. H. Chalepus sp. 1, sensilla trichodea I on base of antennomere IV. I. Calyptocephala brevicornis, sensilla trichodea I on antennomere III.

Figure 50

Antennal sensilla, scanning electron micrographs. A. Baliosus sp. 1, antennomere 6. B. Physonota sp. 1, antennomere VI. C. Physonota sp. 1, antennomereVIII. D. Physonota sp. 1, grooved peg sensilla (short), antennomere VIII. E. Physonota sp. 1, grooved peg sensilla VIII (long), antennomere VIII. F. Notosacantha badia, sensilla trichodea II on antennomere. G. Notosacantha badia, sensilla type VI on antennomere 8. H. Asteriza flavicornis, sensilla chaetica on antennomere VII.

Position. The close insertion of the antennal tubercles (char. 51) was treated as a synapomorphy of Cassidinae (Chen, 1985; Schmitt, 1989; Reid, 1995) but it also applies in Galerucinae (e.g., char. 10 of Lingafelter and Konstantinov, 2000). Monrós (1959) used the cephalic width as a measure of interantennal distance.

Cassidine antennal insertions vary relative to the eye margin and in interantennal distance. Length is also variable and is measured relative to pronotal length. Uhmann (1954) described the “twisting” of the antenna to distinguish among species of Botryonopa, but this is a difficult feature to observe (Würmli, 1975). Antennomeres vary in relative proportions, relationship to each other (how they fit), shapes (cylindrical or compressed), numbers (fusion and loss), pubescence, striations, and grooves (presence, number, and extent). Relative proportions of antennomeres were first used by Baly (1969a). Monrós (1959) also used antennal length (relative to body length) for diagnoses. Hincks (1952) included five antennal features in his key to cassidine tribes: (1) presence or absence of parallel longitudinal grooves on distal antennomeres; (2) antennomere shape cylindrical (round cross-section) or broader at apex; (3) division of the antenna into proximal and distal portions, based on shape and length; (4) presence or absence of pubescence; and (5) relative lengths of the second and third antennal segments. Of these features, I use the first, second and fifth as characters in this study. Hincks’ (1952) third and fourth characters are difficult to apply consistently. Striated antennomeres are more widespread than as originally defined; for example, they also occur in Xenochalepus (C. Staines, personal commun.) and Emdenia (L. Borowiec, personal commun.).

Number of Antennomeres. Antennomere number is surprisingly variable at generic and species levels. For example, coelaenomederine genera have 6–11 antennomeres (Gressitt and Samuelson, 1990); cephaloleine genera have 10 antennomeres (Calliaspis Dejean; Borowiec, 2003) or 11 antennomeres in others; and uroplatine genera vary with 3–8 segments (Staines, 1986a, 1989). These variations may be due to loss or fusion of antennomeres. The degree of fusion also varies; for example, sutures are apparent but reduced in Acanthodes Baly (Uroplatini). Acanthodes displays intraspecific variation, with some having up to nine antennomeres fused together (Baly, 1864a). Fusion of distal antennomeres appears to have also occurred in Octotoma, Callistola, and Microrhopala. Distal antennomeres may also be thickened and have the appearance of a gradual club, for example, Chalepini (Staines, 1993), Brachycoryna (Staines, 1986a) and Notosacantha (Borowiec, 1995).

Antennomere I (scape) is the most divergent morphologically from other antennomeres. It is variable in shape, length, sensilla (presence, density, arrangement, and types), and surface texture. In Delocrania its shape is highly irregular, with the mesal surface having horizontal ridges (appearing like a stridulatory file). In some cassidines, the apical margin is projected into a spine of variable length (e.g., Arescus, Asmangulia, and Hispellinus). This spine may be forked (e.g., Hispa donckieri (Pic)) or unforked (e.g., Hispa atra Linnaeus) (Gressitt, 1950). This projection was suggested to be sex-related (Würmli, 1975). In a few genera, the apical and basal margins are sinuate. Other antennomeres generally resemble one another in shape; that is, elongate and cylindrical with the apex slightly wider than the base. Distal antennomeres tend to be shorter, giving the antenna the appearance of gradual thickening distally.

Surface Texture. Cassidinae antennal microsculpture includes pores, grooves, pits, striations, and diverse sensilla. Pits vary in their sizes, shapes, and presence on different antennomeres. Callistola antennae appear to be uniquely densely pitted. Gressitt and Samuelson (1990) characterized Pharangispa by the numerous sensory pits on antennomere XI. Shaw (1961) placed Capillocassis close to Trichaspis Spaeth based primarily on the presence of large sensory pits on the underside of distal antennomeres.

Sensilla. A few galerucine sensilla were studied (Ritchey and McIver, 1990; Baker, 1987) and these provide a guide to cassidine sensilla types. Sensilla trichodea I (fig. 50A–C) is the most common type on the cassidine antenna. It occurs on most antennomeres but varies markedly in numbers from base to apex, increasing sharply in density between antennomeres IV and V. This sensillum is long, slender, and tapered, and it is aligned to the long axis of the antenna. Its surface may be smooth or ridged, with ridges ranging from barely perceptible to distinct. The basal attachment is variable, being simple or slightly depressed. This sensillum probably serves a mechanoreception function. Sensilla trichodea II is also very common and is long and stout with well-developed ridges. Its base commonly has a crescentic pore on the anterior side, and its apex is slightly narrowed and compressed, with the apical margin straight or sinuate. This may also serve a mechanoreception function. Sensilla basiconica occur on all taxa examined. These are smooth, medium-sized, and stout, and they are found basally on antennomeres I–II. The most common form has a simple apex, but within Physonota sp. 1 (of the taxa sampled here) (figs. 49B–F), two variations were found with bi- or tri- furcated apices. The split apex is suggestive of an aperture, and a closer examination of this tip is needed to determine if one exists. Sensilla type 3 (fig. 48H) is found on all taxa and on all antennomeres. They are medium-sized, tapered to a simple apex, curved towards the antenna, and with a smooth surface. The basal attachment may be flat or prominent. They are found as singles or in fields, especially within grooves of some taxa. Grooved peg sensilla (fig. 48I, 50D–E) are double-walled, single branched, olfactory sensilla. They are slightly mushroom-shaped, with a slender basal stem and an inflated grooved apical portion. Some are short, lacking the stem. The shape of the basal disc varies from a rounded protuberance to a prominent angular base. Sensilla Type 6 (fig. 50G) is unique to Notosacantha. It is medium-sized and double-branched, with the bifurcate section being longer than the basal stem. It is loosely set in a prominent socket and the surface is entirely porous. These wall pores suggest a chemosensory function. This sensilla type does not appear to be documented for any insects. Sensilla chaetica (fig. 50H) are the largest sensilla on the cassidine antennae and occur in all species examined. Within the antenna, they are found in a sparse row along the apex of distal antennomeres, oriented perpendicularly to the antenna. They are long, tapered, and unibranched; the surface has faint longitudinal ridges, and the base is in a circular, depressed, porous pit. These pits may represent an olfactory pore plate.

Antennal Grooves. These occur on some antennomeres in a few tribes. On basal antennomeres, grooves are short and incomplete, not spanning the length of the antennomere. Grooves tend to be longer distally, and terminal ones span the length of the antennomere. Fields of mixed sensilla types occur within these grooves (fig. 48F–H). The number of antennomeres involved varies among taxa. For example, grooves occur on antennomeres VII–XI in Stenispa, Anisostena, Ceratispa, Cyperispa, and Palmispa; on antennomeres VIII–XI in Callistola; on antennomeres V–XI in Ischyrosonyx and Calliapsis; and on antennomeres III–XI in Calyptocephala. Grobbelaar (1993) illustrated similar antennal grooves in some Galerucinae s.l., but Lingafelter and Konstantinov (2000) did not discuss this feature. Antennal grooves occur in Emdenia Spaeth (L. Borowiec, personal commun.), currently classified in the tribe Cassidini. Its larvae resemble those of Basiprionota (C. Reid, personal commun.) more than those of other Cassidini, so its tribal placement may need reevaluation. Sensory pits, sensory fields and other antennal structures of cassidine antennae are interesting and could be phylogenetically significant.

Mouth (Figs. 30–Figure 3132, 51–Figure 52Figure 5354; Chars. 69–83)

Position. Cassidines have the mouth positioned ventrally, in contrast to the prognathous condition in other chrysomelids. Their mouth position varies from extreme anteroventral to a posteroventral position; for example, the Exothispa mouth is in an entreme anterior position whereas others have the mouth in an extreme posterior position. As a consequence of this shift in position, tracing the areas of the gena, vertex, frons, and corona on the cassidine head is problematic. I refer to the dorsal surface of the head as the vertex, the laterofrontal areas as the gena, and the area between the antenna and clypeus as the frontoclypeus. As a consequence of hypognathy, the landscape of the head has completely altered, making comparisons and homologies with prognathous heads difficult. Several mouth and head characters defined for hypognathous heads are inapplicable on prognathous heads. This positional variation also influences the development and relationships of the antennal insertions, frons, clypeus and gula.

Gressitt (1950) used some interesting mouth features (e.g., relationships to antennal insertions and to clypeus) in his key to Chinese hispine tribes. A thorough comparative analysis of the cassidine head is needed to separate characters and states. The present study is only skimming the surface of this diversity.

Mouth Cavity (figs. 44A, D–E, 45). The shape of the cavity may be rounded (fig. 45A) or irregularly quadrate. The margins may be simple (fig. 45C) or heavily sclerotized and protuberant (fig. 45A, B, E). The margin may be extended into the head, providing a shelf for mouthparts (e.g., figs. 44A, D, E, 45A). This shelf is variously modified with cavities that receive the mandibular, maxillary and labial articulations, and a detailed study is needed to determine the homology of articulations and sockets.

Mouthparts are complex and varied in Cassidinae and provide rich ground for characters. Material for disarticulation of the head and mouthparts was not available for many species examined here, and this region is a source of missing information in my dataset. Mouthpart structures are described only briefly here.

The cassidine labrum (fig. 51; chars. 76–78) has shifted ventrad, resulting in a reorientation of its surfaces: the dorsal surface has become the anterior or external surface, the distal margin has become the ventral margin, and the posterior margin has become the dorsal or internal surface. Cassidine labra are deeply inserted into the head, leaving a short exposed section that may be quadrate, hemisphaerical or trapezoidal. They may be narrow and overlap only the mandibular apices, or they may be wide and extend to the mandibular bases. They are also symmetrical, with the lateral margins parallel, sub-parallel, or rounded. The apical margin may be simple (fig. 51D) or emarginated (fig. 51A, E), and narrowed (fig. 51E) or thickened (fig. 51B). When the distal margin is emarginate, the emargination may be broad or narrow and shallow or deep. The external (anterior) surface of the labrum may be asetose (fig. 51C) or setose (fig. 51B, D, E); setae may be soft or bristle-like and randomly scattered at the apex or in paired lateral groups. Botryonopa (fig. 51B) has the apical margin thickened and with two types of setae; long flexible setae occur in a narrow band across the upper portion while short, closely spaced, bristle-like setae occur in the lower portion. A few genera have the labrum carinate dorsally, with one or two carina extending from apex to the exposed basal margin, or diagonally from apex to lateral margins (e.g., fig. 51C). This carina variation was proposed as character 4 in Borowiec (1995) and is used here as character 77. Some cassidine labra are modified with a thickened epipharyngeal surface, for example, Hemisphaerota (fig. 31C, D) and Botryonopa (fig. 51B). In Hemisphaerota, this thickened epipharyngeal surface spans the labral width and its surface is smooth. In Botryonopa, the epipharyngeal surface is exposed as a narrow, rounded thickening located medially along the labral margin, and it is densely microsetose.

Figure 51

Labrum of Cassidinae, anterior view. A. Arescus sp. B. Botryonopa foveiocollis. C. Callistola speciosa fasciata. D. Prosopodonata dorsata. E. Stolas lebasi. Scale bars  =  0.5 mm.

Mandible (figs. 31E, F, 52; chars. 79–82). Migration of the mouth ventrad has altered the articulation, shape and position of the cutting edge of the cassidine mandible. The dorsolateral socket has moved to an anterior position and the ventrolateral condyle has moved to a posterior position. The cutting edge is mesal, not apical. Generally, cassidine mandibles overlap, with no obvious asymmetry except for slight size differences. Many genera have the mandibles trigonal in shape with a broad base and narrowed apex (e.g., fig. 52G–I), often with a single incisor alone, or with the incisor combined with a secondary dens or a sub-apical notch. Some trigonal-shaped mandibles may be slightly constricted medially or may have the dorsal (inner) surface shallowly convex. In a few genera, the mandibles are extremely thick throughout with the mesal margin wide, flattened, and transversely grooved; these mandibles abut directly, suggesting a grinding surface. Some taxa have the internal (dorsal) mandibular surface with paired, heavily sclerotized ridges that form a groove or channel (e.g., fig. 52A–C). Many genera have the mandibles flattened, and palmate-shaped (e.g., fig. 52M–O), with the internal surface depressed and the mesal margin developed in three to nine dens of varying sizes. Palmate mandibles often have the external surface with a thickened medial or basal ridge or one to three basal tubercles.

Figure 52

Mandible of Cassidinae. A. Arescus sp., ventral view. B. Dorsal view. C. Posterolateral view. D. Botryonopa foveicollis, ventral view. E. Dorsal view. F. Anterolateral view. G. Callistola speciosa fasciata, ventral view. H. Dorsal view. I. Anterolateral view. J. Prosopodonta dorsata, ventral view. K. Dorsal view. L. Anterolateral view. M. Stolas lebasi, ventral view. N. Dorsal view. O. Anterolateral view. Scale bars  =  1 mm.

The dramatic variation observed in adult mouthpart morphology across Cassidinae is paralleled by similar variation in larval mouthparts. Given the wide taxonomic and morphological diversity of host plants and plant parts selected as food within Cassidinae, it is intriguing if and how host choice has affected mouthpart morphology.

Maxilla (fig. 53). This has a cardo, one- or two-segmented galea, a one-segmented lacinia and a four-segmented palpus. The cardo is usually constricted medially; the basal portion may be deflected into the head, leaving a quadrate apical section exposed. The stipes varies in shape and sclerotization, with the mesal margin longer than the external margin. Sclerotization may be uneven (e.g., Callistola) with a heavily sclerotized external portion and a lightly sclerotized mesal portion. The surface is commonly flat but may also have a prominent longitudinal ridge. The apical margin may project laterally beyond the lacinia. The palpomeres are commonly cylindrical and they vary in proportions with the distal most one being the longest. They are commonly setose, with setae varying in density and arrangement. In a few taxa, the palpomeres are flattened. The medial and lateral margins of palpomeres are usually similar in length, but in some the lateral margin is longer, giving a curved appearance to the palpus. The galea is cylindrical, and in two-segmented forms, the basigalea and distigalea vary in relative proportions and density of vestiture. Setation on the basigalea may be absent, scant, or dense. Setation on the distigalea may be scant (e.g., Callistola), confined to a narrow apical row (e.g., Arescus) or a mesal band (e.g., Botryonopa), or cover the entire surface and form a dense brush. The lacinia may be small and rounded, larger and rounded, or quadrate. The apex is commonly rounded. In Botryonopa (fig. 53B), it is flattened, and the enlarged surface is densely covered with short setation.

Figure 53

Maxilla of Cassidinae, external view. A. Arescus sp. B. Botryonopa foveiocollis. C. Callistola speciosa fasciata. D. Physonota sp. Scale bars  =  0.5 mm

Labium (fig. 54). Cassidine labia comprise distinct menta, prementa, ligula, and a three-segmented palpus. The mentum may be quadrate, triangular, or rounded, with lateral margins parallel, subparallel, or diverging apically. The surface is generally smooth, sometimes with a few sparse setae. The apical margin of the mentum may be continuous with, slightly narrower than or slightly wider than the basal margin of the prementum. In Exothispa (fig. 54F) the apical margin is slightly projected laterally. The prementum is quadriform and may be longer than wide or wider than long. The surface is usually flattened and with sparse setation. The labial palpus is generally present and three-segmented with palpomeres varying in their relative proportions. In Choeridiona Baly, this palpus is absent (Gressitt, 1950). Palpomere I is usually the shortest. Palpomere II or III may be the longest. Palpomeres may be cylindrical or, more rarely, flattened. In Cephaloleia (fig. 54C) the palpus is flattened, and the external margin of palpomere II is almost twice as long as the mesal margin. The palpus insertion can vary from mediolateral to basal in position. The cassidine ligula is relatively large compared to other chrysomelids (e.g., fig. 54A) and ranges in general shape, sclerotization and surface pubescence. It can be rounded or triangular and lightly to heavily sclerotized. The apical margin may be acuminate or rounded, but never bilobed. The ligular surface is commonly flat and can be glabrous or sparsely to densely setose. The ligula of Exothispa is slightly depressed medially (fig. 54F).

Figure 54

Labium of Chrysomelidae. A. Plateumaris rufa (Donaciinae). B, H. Cassidinae. B. Basiprionota chinensis. C. Cephaloleia sp. D. Botryonopa foveicollis. E. Calliaspis rubra. F. Exothispa reimeri. G. Eurispa vittata. H. Hispodonta chapuisi.

Endoskeleton of Head

Stickney (1923) found that the tentorium is generally membranous in the heads of Anoplites gracilis and Chelymorpha argus. The absence of the clypeal sclerite was not discussed but the term “postclypeus” was introduced for the frontoclypeus. Although Stickney's (1923) illustrations indicate some variation in head endoskeleton, I have not explored this feature here because of lack of materials for the sagittal sectioning needed to examine it.

Prothorax (Figs. 55–Figure 56Figure 57Figure 58Figure 59Figure 60Figure 61Figure 62Figure 63Figure 64Figure 6566; Chars. 84–113)

This segment varies across Cassidinae in general shape, margin form, development and number of angles, orientation of various sections, prosternal process form, coloration (dorsally and ventrally), and texture patterns (dorsally and ventrally). Some pronota are transparent, but with cuticular patterns within the sclerite (e.g., Coptocycla). The presence of the anterior and posterior trichobothria has been useful for diagnosing cassidine tribes (Hincks, 1952; Würmli, 1975; Staines, 2002b). The prothorax can be longer or shorter than the pterothorax.

Pronotum (figs. 55–Figure 56Figure 57Figure 5859, chars 84–102). Pronotal shape in dorsal aspect ranges from transverse rectangular (almost square) to hemisphaerical. It has anterior ( =  apical) and posterior ( =  basal) margins and anterolateral, posterolateral, and posterior angles. Lateral margins may be absent, slightly developed or greatly explanate. In pronota with a deeply convex anterior margin, anteromedial angles are also present. In pronota with bisinuate basal margins, medioposterior angles are formed (e.g., fig. 55C, D). The length of pronotal margins varies in relative proportions; frequently the apex and base are more or less equal in width or the base is wider than the apex. In Basipta (fig. 55D) the pronotum has the apex appearing wider than the base, but the basal margin is actually much longer than the apical margin. In lateral aspects, cassidine pronota appear flattened (e.g., Callistola), shallowly arched (e.g., Spilophora), or strongly arched with the basal margin highly elevated from the apical margin (e.g., Oxynodera).

Figure 55

Pronotum of Cassidinae, dorsal aspect. A. Anisodera guerini. B. Arescus sp. C. Basiprionota quadriimpressa. D. Basipta glauca. E. Botryonopa foveiocollis. F. Calliaspis rubra.

Figure 56

Pronotum of Cassidinae, dorsal aspect. A. Callistola speciosa. B. Canistra plagosa. C. Cephaloleia sp. D. Ceratispa palmicola. E. Conchylotenia hybrida. F. Dactylispa sp.

Figure 57

Pronotum of Cassidinae, dorsal aspect. A. Epistictina viridimaculata. B. Eurispa sp. C. Hispodonta chapuisi. D. Imatidium fasciatum. E. Ischyrosonyx oblonga. F. Jonthonota novemmaculata.

Figure 58

Pronotum of Cassidinae, dorsal aspect. A. Klitispa opacicolllis. B. Lasiochila gestroi. C. Notosacantha badia. D. Physonota alutacea. E. Promecotheca papuana. F. Scoloenopla mantecada.

Figure 59

Pronotum of Cassidinae, dorsal aspect. A. Trichispa sp. B. Hybosa sp., anterolateral trichobothrium. C. Calyptocephala nigricans, anterolateral and posterolateral trichobothrium. D. Calyptocephala nigricans, scanning electron micrograph showing anterolateral trichobothrium.

The anterior pronotal margin may be simple (fig. 55A, B) or expanded (fig. 55D, 56B, E). In many species, this margin is either straight or broadly concave but in some it is broadly and deeply convex, partially or completely covering the head (fig. 56E). The relationship between the marginal extension and head retraction into the prothorax must be studied further since these two distinct processes may have occurred together. In convex anterior margins, the prothorax has both anterolateral and anteromedial angles, and the head is exposed dorsally. Because of variations in the marginal extension, the anterior prothoracic foramen can appear wide, bounded completely by the anterior margin, or narrow and bounded by the discal margin. In some genera (e.g., Elytrogona and Hilarocassis) the anterior margin is thickened and slightly upturned.

Ventrally, the basisternum may be very narrow longitudinally so that the anterior margin is very close to the procoxal margin (e.g., fig. 64A–C). More commonly, the basisternum is well developed and long, with the anterior and procoxal margins well separated (e.g., fig. 64C). In Basiprionota, the basisternum is very long and partially covers the venter of the head, hiding the mouth completely. In Paraselenis, the prepectus is approximately three times longer than the length of the hypomeral lobe. The mouth may be partially hidden in other taxa because of a medial extension of the anterior margin. This flange over the mouth has been called a “chin plate” (Würmli, 1975), or prosternal collar. The prosternal collar often has a transverse groove posteriorly that defines it from the rest of the prosternum. It may be short or long, and the surface texture may be smooth or transversely wrinkled. In Basipta, the prosternal collar covers the mouth up to the labrum, and it is unclear how the animal feeds, if at all.

Figure 64

Prosternal process in Cassidinae, ventral view. A. Asteriza flavicornis. B. Basiprionota chinensis. C. Hispodonta chapuisi. D. Estigmena chinensis. E. Lasiochila gestroi. F. Notosacantha badia. G. Promecotheca papuana. Scale bars  =  1.0 mm

The anterior pronotal margin may be smooth, spinose, setose, or serrate. Asamangulia has the pronotal disc aspinose but the lateral and anterior margins spinose; these spines branch basally. In Cyperispa the anteroventral margin has a single median tooth and a row of evenly spaced, erect hairs.

The lateral margins of cassidine pronota are commonly margined but some (e.g., Promecotheca) have rounded sides that are also protuberant medially. In marginned forms, margins are usually complete, extending from the anterior margin to the basal margin, and simple or slightly thickened (e.g., Chelymorpha). More rarely, some marginned forms have the margin incomplete (fig. 60A, D). The margin is evanescent posteriorly in Chalepus, evanescent medially in Cyperispa, and evanescent anteriorly and posteriorly in Exothispa. Margin development varies widely, as in the elytra. Some have a very narrow bead or gutter along the edge of the pronotal disc. In others, the margin is explanate, and its width can be less than, equal to, or greater than the width of the disc. Development of the explanate margin is also accompanied by enlargement of the hypomeron ventrally. Laterally, some pronota have a humeral protuberance (e.g., Sagra, Plateumaris), defined by a shallow projection of the disc and margin and/or with a slight depression behind the humerus (e.g., Plateumaris). The surface texture of the lateral margin includes setation, serration, and spines. In Anisodera (fig. 55A, 60A), the marginal surface texture is irregular. Lateral margins of Botryonopa (fig. 55E) have large, irregular serration. Dactylispa has spinose lateral margins (fig. 56F).

Figure 60

Pronotum of Cassidinae, lateral view. A. Anisodera guerini. B. Aspidimorpha quadrinodosa. C. Callistola speciosa fasciata. D. Exothispa reimeri.

Posterior ( =  Basal) Margin. This may be tranverse (fig. 55A, 58A, B), appearing truncate at the posterior angle. The posterolateral angle may or may not be projected, and with a single trichobothrium (fig. 55B) (char. 9 of Borowiec [1995]  =  char. 101 herein). Where projected, this angle may be directed laterad, or posteriad (e.g., Chelymorpha). In cassidines with bisinuate posterior margins (fig. 55C, D, 56B), the latter may be distinguished into two sections separated medially by the medioposterior angle. Both sections may be concave, corresponding with convexities of the elytral basal margin (e.g., fig. 55C, 56B) or the outer section may be straight and the inner section may be convex (e.g., fig. 55D). The posterior margin may be inflected, forming a secondary dorsal margin (fig. 58A). This secondary margin corresponds to the elytral basal margin while the true basal margin is hidden under the elytra. In these cases, the primary pronotal posterior margin is exposed medially at the mesoscutellum.

The pronotal posterior angle may be absent (i.e., the margin is straight) or present; when present it is rounded or acute and may overlap the mesoscutellum. In Dorynota, the anterior margin of the mesoscutellum is expanded and overlaps the pronotum at this point. The posterior margin is commonly smooth (e.g., Basipta) but can also be crenulate (e.g., Basiprionota).

Pronotal Disc. Where the pronotum has explanate margins, the disc and margin may be clearly defined by changes in angulation, a discal margin, ridges, wrinkles or punctation. When they are not easily demarcated dorsally, one must examine the venter of the prothorax to determine the hypomeron, tergosternal suture and prosternum.

Pronotal surface texture varies across Cassidinae from smooth to having irregular patterns of depressions or protuberances, to having various patterns of punctation and vestiture. Botryonopa has a slightly carinate medial longitudinal line and a shallow, uneven, transverse medial depression. In Asamangulia and Dactylispa, a slight transverse medial projection and the inflection of the basal margin create two parallel transverse grooves. There is a transverse basal depression in Cephalodonta (weakly developed) and Dactylispa (well developed; fig. 56F). Exothispa has a slight transverse, sub-basal ridge. Klitispa has a deep, narrow transverse slit in the sub-basal medial area, and paired arcuate ridges running longitudinally in the lateral areas. Lasiochila has short depressions running longitudinally in the posterolateral area.

Punctation can vary greatly, being shallow or deeply impressed and even fenestrate (e.g., fig. 58C), small or large, with regular round forms or irregular shapes, widely separated or adjoining, and in many different distribution patterns over the pronotal surface. Ceratispa has a highly regular arrangment of evenly spaced, round, similar-sized punctures. Punctation patterns on the disc and explanate margin can be independent; for example, Calyptocephala has a smooth disc and a shallowly impunctate margin. In Arescus, the disc has fine, shallow dispersed punctations, whereas in Anisostena it has deeply impressed, closely spaced punctation.

Dorcathispa has one of the most complicated pronotal surfaces of any cassidine. The lateral margins are evanescent anteriorly and have multibranched spines medially. The disc also has a pair of subapical multibranched spines and a pair of sharply defined protuberant discs medioposteriorly. The anterolateral and posterolateral angles both have protuberant trichobothria with a single long bristle. Erect, somewhat flattened setae are also dispersed on the body, including the pronotum especially along the anterior and posterior margins and in a medial field.

Pubescence. Generally pronotal pubescence is restricted to trichobothria located on the angles with single or multiple bristles. Basipta has fine, dense, recumbent hairs oriented anteromedially. Asamangulia has erect, somewhat flattened setae with swollen apices similar to that of Dorcathispa; these are distributed on the disc and along the anterior margin of the pronotum. Eurispa has a deeply punctate pronotum, and the lateral and anterior punctures have a single, recumbent, anteriorly directed hair. In Trichispa, the body is generally clothed in dense recumbent pubescence.

Trichobothria (figs. 33A, 55A, B, 58F, 59B, E). Hincks (1952) distinguished the tribes Spilophorini and Imatidiini by, among other features, trichobothria on the anterior pronotal angles alone (e.g., Imatidium) and on both anterior and posterior pronotal angles (e.g., Calyptocephala). No other tribes within Cassidinae s.str. possess this structure. Within the Hispinae s.str., the occurrence of these trichobothria on the pronotal angles is widespread and has been useful in tribal diagnoses (e.g., Würmli, 1975), with some taxa having “sensory tufts” with multiple pores and/or multiple bristles. These sensory tufts, or clusters of trichobothria, occur on both the anterior and posterior corners of Hispini, on only the anterior corners of Aproidini, Anisoderini, and Eurispini (Würmli, 1975), and on only the posterior angle of Alurnus. Alurnus has three pores, each with a single bristle. Anterior and posterior trichobothria can occur together (e.g., Asamangulia) or independently. They are only on the anterolateral angles in Baliosus, on the anteromedial angles in Calliaspis, and on the posterolateral angles in Callispa. The pore base may be flat, protuberant or even slightly elongated. Cases where trichobothria occur on both anterolateral and anteromedial angles are not known. I follow Borowiec (1995; chars. 8, 9) and treat these as independent characters (chars. 99–101) based on their topographic differences.

The hypomeron may be oriented vertically (e.g., Ceratispa), or more commonly, ventrolaterally (e.g., Calyptocephala). In the former condition, the procoxae may be in a lateral position so that the profemur can be retracted vertically against the propleuron (e.g., Hemisphaerota). The propleuron surface can be indented, often in conjunction with changes in the orientation of the mesopleuron, and together they form a cavity where the legs are retracted (e.g., Pachybrachis). The hypomeral surface texture varies from smooth (e.g., Calyptocephala) to wrinkled (e.g., Desmonota) to punctate (e.g., Callistola and Ceratispa).

Prosternum (figs. 61–Figure 62Figure 63Figure 6465; chars. 103–113). The length of the prepectus in cassidines is usually longer than that of the hypomeral lobe; however in a few taxa (e.g., Ceratispa) the latter is longer than the former. The anterior prosternal margin may be smooth, serrate, dentate, or with a row of fine hairs; the edge may be simple or thickened, and continuous or with an anterior convexity called the prosternal collar. This collar forms a broadly rectangular flange, which overlaps the mouth. In Basiprionota, the collar is greatly expanded anteriad, covering the mouth as far as the labrum. In expanded margins, the edge may be directed anteriad or anteroventrad, with a slight groove behind the collar. The prepectus (section of prosternum anteriad of procoxae) is usually as long as or longer than the length of the hypomeral lobe behind the procoxae, but in Delocrania it is shorter than the length of the latter. The anterolateral section of prepectus is oriented vertically (Sagra, Plateumaris) or ventrally (in the same plane as process). In Deloyala, a pair of weakly-developed carina extend from the hypomeron, longitudinally alongside the eye and gena. In Dorynota, this point on the hypomeron is slightly protuberant and forms a groove with the prosternal collar; these lateral grooves may hold the antennae but this has not been observed.

Figure 61

Basipta glauca. A. Prosternal process, scanning electron micrograph. B. Meso- and metasternum.

Figure 62

Thorax of Cassidinae, ventral view. A. Brontispa mariana. B. Botryonopa foveicollis. C. Callistola speciosa fasciata. D. Epistictina viridimaculata. E. Estigmena chinesnsis. F. Hispodonta chapuisi.

Figure 63

Thorax of Chrysomelidae, ventral view. A, E. Cassidinae. A. Lasiochila gestroi. B. Pharangispa palmicola. C. Physonota alutacea. D. Promecotheca papuana. E. Spilophora aequatoriensis. F. Sagra sp. (Sagrinae).

Figure 65

Mesoscutellum of Cassidinae. Dashed lines represent underlying sclerite. A. Aproida balyi. B. Coptocycla dorsoplagiata. C. Dorynota sp. D. Spaethiella sp.

The prosternal process (figs. 33B, C, 61–Figure 62Figure 6364; chars. 108–113) varies in length, protuberance, and general shape especially in the apical region. In most cassidines the base, stem, and apex of the process are in the same plane; however, some may have the medial section protuberant (e.g., Arescus). This differs from the condition in some outgroups (e.g., Sagra) where the apex is protuberant from the base and stem. The general shape of the process apex varies across Cassidinae. Commonly, the apex is expanded laterally, extending below the procoxae and meeting the hypomeral lobe directly or overlapping the latter. The apex may also be simple, with the process width being even from base to apex.

The process surface can be smooth or have grooves, wrinkles, punctures and pubescence. The base, stem and apex can also have different patterns of textures. In Botryonopa, the base is smooth, the stem has a slightly convex line medially and longitudinally, and the apex has five short, deeply impressed, parallel, longitudinal grooves. In Charidotella and Chalepus, the process has sparse fine long hairs.

Tergosternal Suture ( =  notosternal suture; Chamorro-Lacayo and Konstantinov, 2004) (fig. 33B). This is usually apparent but is absent in Plautyauchenia and Acromis. It may be long, extending to the anterior margin (e.g., Basiprionota). In many other cassidines, it is short, terminating before the anterior margin. Its orientation may be longitudinal or anterolateral.

Procoxal Cavity. The position, shape, and orientation vary in subtle ways that are difficult to define. In Hemisphaerota the cavities are in a laterovertical plane, but more commonly the cavities are in the ventral plane. This must certainly affect the orientation of legs. The margins of the procoxal cavities are simple or marginate and protuberant. Procoxal cowlings are not prominent in cassidines.

Cassidine procoxae are closed posteriorly by the hypomeral lobe ( =  post-coxal bridge) and lateral expansions of the prosternal process apex. The point of contact between hypomeron and process varies depending on their widths. The medial margin of the lobe may extend under the prosternal process, and its apex is therefore hidden. Its length and width vary, as does the surface texture, which may be smooth or wrinkled.

Pterothorax (Figs. 65–Figure 67Figure 68Figure 69Figure 70Figure 71Figure 7273; Chars. 114–159)

Dorsally, the pterothorax appears to be fairly uniform but further study with additional taxa is needed. Ventrally, relative proportions, shape, form of margins, mesocoxal process form, ventral coloration and ventral texture patterns show much diversity.

Mesonotum. This region will not be described in great detail here, only to note that the mesoscutellum is usually elevated from the mesoscutum. The mesoscutum is commonly relatively weakly sclerotized and smooth. In Plautyauchenia, the mesoscutum is heavily sclerotized, deeply impunctate, and elevated from the mesoscutellum.

Mesoscutellum (fig. 65; chars. 114–117). This sclerite is always exposed in Cassidinae and shows variation in shape, margin exposure, and surface texture. It may be triangular (e.g., Canistra), rounded (e.g., Aproida), or quadrate (e.g., Calyptocephala). In Cephalodonta and Anisostena, it is quadrate, with the apex depressed and the base protuberant. The pronotal basal margin may be inflected onto itself, creating a small cavity around the depressed mesoscutellum. The anterior margin of the mesoscutellum is usually straight, but it is convex in Dorynota (fig. 65C) and Spaethiella (fig. 65D) and overlaps their posterior pronotal angle. The lateral margins of the mesoscutellum may be expanded and overlap the sutural edge of the elytra (e.g., fig. 65A, C). The mesoscutellum surface can be smooth, punctate, pubescent, or finely microreticulate. In most cassidines, the surface is glabrous, but in Trichispa it is densely pubescent, like the rest of the body.

Mesosternum (figs. 62–Figure 6364; chars. 118–122). The mesosternal surface may be flat (e.g., Estigmena) or deeply notched and receiving the prosternal process (e.g., Canistra). Anterolateral lobes of the mesobasisternum close the mesocoxae laterally. These lobes can be broad (e.g., Basipta) or tapered to a rounded point. The lateral portion of the mesosternum may be oriented vertically such that the mesosternum has a curved appearance from ventral to lateral. Mesosternal surface texture can be smooth, wrinkled, grooved, striate or punctate. In Ceratispa this surface is punctate anteriorly but smooth medially and on the mesosternal process. Botryonopa has the mesosternum with a deep medial trough, the apex slightly protuberant, and the surface with shallow lateral punctures and shallow medial longitudinal grooves. This medial trough is not regarded as homologous with the notched surface since it does not fit around the prosternal process, as in the latter condition.

The mesosternal process (figs. 62–Figure 6364; chars. 120–121) may be short or long (relative to its width), with the posterior margin rounded (e.g., Botryonopa), broadly quadriform (e.g., Anisostena), tapered and acuminate (e.g., Plautyauchenia), or quadrate (e.g., Callispa). Overall, the process is commonly flat, but in Estigmena and Dactylispa the apex is protuberant from the base. It is frequently longer than wide; in Callispa it is wider than long. The surface may be smooth, wrinkled, punctate, or with shallow parallel longitudinal grooves (e.g., Botryonopa). The posterior margin of the process may be concave, convex, or straight. The mesosternal process is free in most cassidines, but in Hispodonta it is fused to the metasternal intercoxal process with the suture being poorly defined (fig. 62F). Mesocoxal cavities are usually rounded, with the intercoxal distance varying according to the width of the mesosternal process. The sclerites closing the cavity laterally also vary according to the development of the lateral lobe of the mesoscutellum.

Cassidine mesopleura (fig. 66) have the mesepisternum small and roughly triangular in shape and they are commonly exposed ventrally or, infrequently, hidden in lateral view by explanate pronotal and elytral margins. This sclerite may be curved or flattened, and positioned in a ventral or ventrovertical plane. Its surface texture includes smooth or punctate and glabrous or pubescent. In Alurnus, a thin transverse carina divides the sclerite into a narrow rectangular anterior section and broad triangular posterior section. In Anepsiomorpha a shallow gully runs alongside the lateral margin. The mesepimeron may be triangular or quadrate in shape, with the anterior margin hidden by the elytra (e.g., Asmangulia), meeting the epipleural ridge directly (e.g., Alurnus) or slightly expanded and overlapping the latter (e.g., Acromis). Surface texture varies as in the mesepisternum. The orientation may be ventral or mostly lateral.

Figure 66

Mesopleuron of Cassidinae. A. Alurnus sp. B. Anoplites sp. C. Asmangulia cuspidata. D. Aspidimorpha miliaris.

The mesepisternum and mesepimeron are commonly separate sclerites, but in a few cassidines they are very tightly fitted together and appear almost fused. Hinck's (1952)475 proposed the fused state as a character but Borowiec (1995) considered this assessment incorrect. I agree with Borowiec (1995). Based on examination of both intact and cleared, disarticulated specimens, I have found no cases of fusion between the mesepisternum and mesepimeron. Both mesopleural sclerites may be smooth, wrinkled, punctate, and glabrous or pubescent. In Asamangulia the surface has fine microreticulations in addition to larger punctation.

Metapleuron (fig. 66, char. 123). The metepimeron is hidden under the elytra as in other beetles. The metepisternum is usually exposed, but in Anisostena, Delocrania, and Dorcathispa it is hidden because the elytra extends to the lateral margin of the metasternum. In Anisostena the elytra fits into a groove under the length of the lateral margin that completely covers the metepisternum. In Delocrania and Dorcathispa, the elytra abut directly with the metasternal lateral margin, hiding the metepisternum.

The surface texture of the metepisternum is usually smooth (e.g., Acromis), but it can be finely or coarsely punctate posteriorly (e.g., Chalepus) or throughout (e.g., Charidotella). Some cassidines have the anterior surface protuberant or the entire surface broadly depressed (e.g., Chelymorpha). In some taxa a protuberance from the anterior surface metepisternum overlaps the epipleural ridge, thus locking the elytra into position (e.g., fig. 66D). In Basiprionota, the lock is slightly different with a slight lobe from the lateral margin of the metepisternum extending over the elytral ridge. Epistictina has marginal lobes off both the mesepimeron and metepisternum that overlap the epipleural ridge. In Paratrikona and Dorynota, the lock is reversed with a lobe off the epipleural ridge nested in a cavity in the anterior mesepisternal surface.

The relationship of the meso- and meta-pleural sclerites provides a significant character in Cassidinae s.str. Spaeth, as translated by Hincks (1952), separated tribes (couplet 1) according to the fusion or separation of the metepisternum and mesepimeron, and diagnosed the fused state for Delocraniini, Hemisphaerotini, Imatidiini, Notosacanthini and Spilophorini. Borowiec (1995) indicated that Hincks' assessment was incorrect, and his character 0 treated the metepisternum as fused with the metepimeron in this group of tribes, and the metepisternal suture as lacking. In my examination of both intact and disarticulated specimens, I found that the mesepimeron and mesepisternum may be fused, with the mesopleural suture discernible or evanescent. However, I found these sclerites to be separated in Calyptocephala and Spilophora (Spilophorini), and in Hemisphaerota and Spaethiella (Hemisphaerotini), taxa indicated as fused by Hincks (1952). Borowiec (1995) also appears to be incorrect since I found the mesepimeron and metepisternum to be distinct sclerites throughout Cassidinae (e.g., Hemisphaerota, fig. 36B). Finally the term “metepisternal suture” should be referred to by the standard terminology, metapleural suture (Snodgrass, 1935: fig. 102). The metepimeron is usually covered by the elytra, with the elytral margin meeting a well-defined metapleural suture. This region of the cassidine body warrants further study.

Metasternum (chars. 124–127). This is usually rounded or flattened, with the anterior and posterior margins sub-equal, and the medial and lateral lengths equal or sub-equal. More precisely, the base is slightly wider than the apex, with the lateral margins curved and slightly convergent anteriorly. In most cassidines, the lateral margin is slightly longer than the length along the medial line. In Eurispa, the medial length appears longer than the lateral marginal length. In Elytrogona, the metasternum is highly distorted and protuberant, with the lateral margin almost twice as long than the medial length. The median longitudinal groove is impressed (e.g., Cephalodonta) or faint (e.g., Charidotella), and it may extend from the anterior to the posterior margin or become evanescent before the anterior margin. The relative proportions of length and width vary, with some having the metasternum longer at the midline than half its width (e.g., Callistola), and others having the medial length shorter than half the width (e.g., Calyptocephala). In Exothispa, the anterolateral corner of the metasternum is projected into a rounded lobe that overlaps the mesepimeron. The metasternal surface is shallowly convex in some cassidines (e.g., Callistola and Calyptocephala). In some (e.g., Cephalodonta) the surface is rounded, with the lateral margins in the vertical plane. In Basipta the metasternal surface is protuberant posterolaterally. In addition to such protuberances, the surface may be punctate and/or striate (e.g., Ceratispa). In Charidotella, the lateral area has fine recumbent hairs and the post-coxal margin is lined with a row of punctures.

Anterior Metacoxal Process. In most cassidines, the process is flat or very weakly notched. In Elytrogona and Stoiba, it is deeply notched and receives the mesosternal process. These genera are flightless and this feature may be part of a syndrome of morphological changes in the pterothorax associated with flightlessness. Other changes include distortion of the metasternal surface, narrowing of the metasternal length (the lateral margin is twice as long as the medial length in Elytrogona), a deeply impressed medial longitudinal line, a long abdominal intercoxal process, an anterior lobe in the metepimeron that locks the elytra (Elytrogona), and wing reduction and loss in both genera.

Longitudinal compaction appears to be a trend in the cassidine thorax. This is achieved in two ways. First, the prosternal hypomeral lobe may overlap the mesosternum, reducing the longitudinal distance between the pro- and mesocoxa. In extreme cases, the hypomeral lobe may reach as far as the mesocoxal margins (e.g., Acromis). Second, the metasternal length varies, with the length along the medial line becoming shorter relative to the width. Thus, the distance between the meso- and metacoxae is shortened.

Elytra (Figs. 67–Figure 6869; Characters 128–156)

Cassidine elytra are fully developed and completely cover the abdomen in dorsal aspect. Generally, the elytra are strongly sclerotized, heavy, and rigid, but Promecotheca has thin, soft elytra (fig. 14N). Elytral coloration varies widely, as discussed under the Biology section (figs. 9–Figure 10Figure 11Figure 12Figure 1314). Most elytra are opaque, but some metallic-golden forms can have transparent margins (e.g., fig. 9F). Diversity in coloration, shape, proportions, and surface texture makes the cassidine elytra complex but rich with potential characters. No elytral characters were used in Reid (1995) or Borowiec (1995); I present 39 elytral characters for Cassidinae.

Figure 13

Cassidinae, adults, lateral aspect. A. Acromis sparsa, male. B. Agenysa caedemadens. C. Anepsiomorpha deplanata. D. Anisodera guerini. E. Asamangulia sp. F. Basiprionota quadrinodosa. G. Basipta glauca. H. Batonota lerouxi. I. Botryonopa foveicollis. J. Calliaspis rubra. K. Callistola speciosa. L. Calyptocephala nigrans. M. Canistra plagosa. N. Ceratispa palmicola. O. Chalepus sp. P. Chelymorpha cassidea. Q. Cyperispa thoracostayi kolombagara. R. Dorcathispa sp. S. Epistictina viridimaculata. T. Estigmena chinensis.

Shape. In dorsal aspect (figs. 9–Figure 1011), cassidine elytra have parallel, subparallel, wedge, or rounded shapes (Gressitt, 1963415; Würmli, 1975). In parallel and subparallel forms (e.g., fig. 9H), the width-length ratio is approximately 1:3 or longer, with the width consistent from the base up to the apex; at the apex, the margins are sharply rounded. Frequently, this shape is associated with a flattened lateral profile and the apex is gently curved or declivous (e.g., fig. 12G). In wedge-shaped forms (e.g., Chalepus), the beetle is widest at the apex, and the apical margin appears abruptly truncated. In rounded forms, the width-length ratio is 1:1 to 1:1.5, with the beetle widest at the basal margin, humeral line, or midline, and the apical margin is always rounded (e.g., figs. 10 A, B, E, G, I, K).

In lateral aspect, the elytral profile may be flattened, rounded or arcuate (figs. 12G–K, 13, 14). Elytrogona has inflated globose elytra (fig. 10J), which may be correlated with flightlessness (Chaboo, 2000). Arcuate forms are highest along the post-scutellar area (e.g., Psalidonota). This post-scutellar protuberance can be greatly extended, forming a spine (e.g., fig. 13I) or an umbo (e.g., fig. 14H). In addition to the post-scutellar spine, Dorynota has another distinctive elytral feature, a transverse ridge extending from the humerus across the margin to the anterolateral angle (fig. 8E). In Canistra, Dolichotoma and Oxynodera, there are paired flattened elytral discs adjacent to the mesoscutellum and extending to the post-scutellar protuberance. I term these parascutellary discs (fig. 67C).

Figure 67

Elytra of Cassidinae, dorsal aspect. A. Acromis sparsa, female. B. Acromis venosa, male. C. Oxynodera biplagiata. D. Anoplites inaequalis. E. Basipta glauca. F. Callistola speciosa.

Angles. Cassidine elytra have anterolateral, posterolateral, and sutural angles (figs. 67A–E). These angles vary in their extension and orientation and in the presence/absence of ridges, serration, spines and denticles. In most Coleoptera the humeral and anterolateral angles are usually coincident, but within Cassidinae, these angles are either separated (fig. 67A) or coincident (fig. 67D). In dorsoventrally compressed forms that lack explanate margins, they tend to be coincident. In forms with explanate margins, the humeral angle is produced at the epipleuron-disc junction. The humeral angle may be protuberant, forming a shoulder that obscures the anterolateral angle in dorsal aspect.

The separation between the explanate margin and the disc may be imperceptible, with little to distinguish the two areas dorsally. More commonly, their separation may be defined by marked changes in angulation between the disc and margin (e.g., Aspidimorpha), a protuberant humeral angle (e.g., Exothispa), and/or with a row of deeply impressed punctures (e.g., Basipta and Conchylotenia). The posterolateral angle is commonly rounded and not well defined. In a few genera this angle is drawn out as a single heavy spine (e.g., Cephalodonta).

Basal Margin. The elytral width at the base may be equal to the pronotum, producing a continuous lateral margin in dorsal aspect (e.g., fig. 7B, C, E, J–L; fig. 68C). The basal margin is wider than the pronotal base, in some cases up to twice as wide as the pronotal base (e.g., fig. 7A, D, 68I). The basal margin may be straight (fig. 68F), broadly arched (fig. 68J), bisinuate (fig. 68C, G), or variously sinuate. In straight forms, the margin may be oriented transversely (e.g., Delocrania) or gently anteriad. Arched basal margins frequently overlap the pronotal basal margin, partially obscuring the latter (e.g., Aproida and Callistola); in such cases, the elytral edge is inflected ventrad, and the pronotum forms a secondary dorsal edge with the elytra (fig. 67F). In some sinuate forms the elytral base at the epipleuron extends gently anteriad (e.g., Spaethiella) or sharply anteriad (e.g., Goniochenia). In bisinuate forms, the base is divided into two more or less equal sections with three angles, the anterolateral angle, a medial angle (which corresponds ventrally to the epipleural ridge) and a posterior angle (at the mesoscutellum) (fig. 67C, E). The two sections of the margin may be convex and tightly fitted with corresponding concavities in the pronotal basal margin (fig. 67C), or the outer section may be straight (fig. 67E).

Figure 68

Elytra of Cassidinae, dorsal aspect. A. Canistra plagosa. B. Physonota sp.C. Basiprionota quadriimpressa. D. Botryonopa sp. E. Chalepus sp. F. Dactylispa sp.G. Dorynota sp. H. Klitispa opacicollis. I. Scoloenopla sp. J. Undetermined sp.

The surface texture of the basal margin can be smooth (fig. 68A) or crenulate (fig. 68B). Crenulation can extend along the entire margin (e.g., Basiprionota, Concylotenia, Epistictina), over short sections (e.g., Aspidimorpha), or from the mesoscutellum and tapering off near the anterolateral angle (e.g., Asteriza). Elytral basal crenulation appears to be independent of pronotal posterior marginal crenulation; for example, Basiprionota has both the pronotal posterior and elytral basal margins crenulate. Dorcathispa has a row of long spines along the basal margin.

Lateral Margin. The development of the margin varies widely within Cassidinae. In many species (e.g., Trichispa) the lateral margin appears coincident with the edge of the elytral disc. In some species (e.g., Alurnus and Odontota) the margin forms a narrow gutter bordering the disc. Where the margin is coincident with the discal edge or it is very narrow, the epipleural ridge is coincident with the edge of the elytra ventrally.

Many cassidines have the margin explanate, and the extent of this margin varies widely. It may be as wide as the disc or wider than the disc (e.g., Acromis). Some genera, (e.g., Omaspides) exhibit interspecific variation in this relative ratio of disc-margin width. The marginal width may be consistent from anterior to posterior or it may be widened or narrowed posteriorly. Its orientation, punctation and setation are variable. When the margin is explanate, the external elytral edge and the epipleural ridge are separated, with the latter in contact with the pleuron. In some cases, a sub-elytral cavity is created, bounded by the ventrolaterally or ventrally oriented elytral margin, the epipleuron and epipleural ridge. Hemisphaerotines retract their legs into this cavity, and early instar larva may be hidden here in subsocial cassidines. Crowson (1981) suggested that this explanate body form of cassidines probably helps to break up the outline in dorsal aspect and makes the animal difficult to lift once the margin is flattened against the substrate.

Males in the genus Acromis have elytra that are unique among cassidines in two ways. The lateral margin has a thin anterior section that will often tear during male combat for females; many museum specimens have holes in the elytra as a result (fig. 5B). The anterolateral angle is also greatly extended anteriad, producing a spine-like lobe used in male combats (fig. 67B).

The surface texture of the lateral edge varies. In many cassidines, it is smooth. In Baliosus, Chalepus and Uroplata, the edge is entirely serrate and has a saw-like appearance. Lateral serration in Demotispa increases in density posteriorly. The lateral edge can also be spinose (e.g., Dorcathispa). Trichispa has a few short spines along the anterior section of the lateral edge, as well as sparse irregular serration in the posterior section. In Dactylispa the lateral edge is spinose from base to apex, with spines slender and twice as long as the interval between them. The lateral edge of Delocrania has a unique arrangement of pores with single bristles at regular intervals. In some Dactylispa and Platypria species the elytral texture is complicated with the margin highly sinuate, scalloped and spinose, and the dorsum has large porous punctation (fig. 68F).

Sutural Margin. The lateral profile of this margin may be flattened, arcuate, or rounded. In arched and rounded forms, the elytral dome is highest in the post-scutellar area, and the dome may also be extended into a spine or umbo (e.g., Batonota and Dorynota). In Elytrogana, the dome is highest at the midpoint of the suture. The sutural margin is commonly flat but it is convex and carinate posteriorly in Anisostena.

The elytral suture of Coleoptera may have a tongue-and-groove mechanism for locking the elytra together. This mechanism is very tight in Eytrogona and Stoiba, making the elytra difficult to open, and it may be related to flightlessness in these genera. The sutural margin at the apex may be continuous or indented, and rounded or drawn out into broad lobes or spines. In some genera, the suture terminates in a short denticle (e.g., Botryonopa). In Aproida, the apical margin is drawn out into a heavy, posteriorly directed spine, and the sutural margin has a small, ventrally directed denticle.

Apical Margin (fig. 69A–E). The apex of the elytra may be rounded (fig. 69A), truncated (fig. 69B–C), or indented (fig. 69E). In truncated forms, the margin can be entirely or partly (medially) truncated. Wedge-shaped forms have the apical margins as the widest part of the elytra, and the posterolateral angles are developed. The apical angle may be extended into a single, heavy spine (fig. 69D). The sutural margin may also terminate with a small denticle (fig. 69E).

Figure 69

Elytra of Cassidinae, dorsal aspect. A. Coptocycla dorsata. B. Callistola sp. C. Ceratispa sp. D. Eurispa sp. E. Botryonopa sp. Ventral aspect, scale bars  =  1 mm. F. Canistra plagosa. G. Paraselenis flava, scanning electron micrograph. H. Agroiconota bivittata. I. Aproida balyi. J. Asteriza flavicornis. K. Charidotella bicolor. L. Promecotheca papuana.

The texture of the apical margin may be smooth, serrate, or spinose. Chalepus and Xenochalepus have two distinct sizes of serration on the apical margin whereas Demotispa and Uroplata have one size of serration. Dorcathispa has a row of spines along the apical margin. Promecotheca has some long erect hairs on the apex.

Surface Texture. Cassidinae elytra may be smooth, or punctate, spinose and tuberculate. Texture patterns between the elytra and pronotum, and between the elytral disc and margin appear to be independent. Punctation varies in sizes, shapes, individuality, degree of impression, and arrangement. Punctures can be round or irregularly shaped, and distinct or confluent. Punctures may be barely perceptible (e.g., Alurnus), shallow (e.g., Hispodonta), deeply impressed (e.g., Sceloenopla), fenestrate (e.g., Cyperispa, Dactylispa), or even porous (e.g., Notosacantha). They may be arranged in a confused way (e.g., Alurnus), in small, rounded regular clusters (e.g., Agenysa), or, more commonly, in well-defined discrete longitudinal rows along striae (e.g., Dorcathispa). Punctation arrangements of the disc and margin are independent; for example, discal punctures of Calyptocephala are in rows, whereas marginal punctures are confused. In Delocrania, discal punctures are small and rows are difficult to distinguish whereas marginal punctures are large, deeply impressed and arranged in rows that are separated transversely by slightly carinate joints. In Arescus, punctation is in rows medially but these rows become confused in the posterolateral area; in Anisodera, puncture rows are well defined basally and medially, but become obsolete posteriorly with the insertion of additional rows; in Aspidimorpha, the external punctures are in rows but medial punctures are confused. In Conchylotenia, punctures are widely separated and rows are vaguely defined. In Prosopodonta, the anterior and central parts of the discal surface are impunctate but puncture outlines are discernible within the cuticle. These puncture rows emerge at the surface apically.

Some cassidine elytra are spinose, tuberculate, and denticulate, or they even have these textures combined together with punctation and pubescence. This produces much complexity to the general dorsal appearance. In Asmangulia, the basal and lateral margins are spinose whereas the dorsum is deeply punctate-striate and tuberculate. In Dactylispa, the disc has irregularly arranged, fenestrate punctures and spines in two sizes, long and short. The lateral and apical margins are also spinose. In Trichispa, punctures are fenestrate and irregularly arranged, and the disc and lateral and apical margins are spinose. Some punctures have a single pore with a short, stiff bristle; the bristle appears to be flattened with a slightly expanded apex. In spinose elytra, the spines are unbranched and vary in length and interval distance.

The elytra often have striae and intervals. Striae are regarded as external manifestations of the columella (internal connections between dorsal and ventral surfaces of the elytra) (Spilman, 1971). Intervals may represent the ancestral wing veins (Spilman, 1971) but elytra are so modified that it is impossible to be conclusive about the derivation of intervals. Striae and intervals alternate and are generally numbered from the sutural stria laterad (e.g., Uhmann, 1954). In Cassidinae, when punctures are arranged in rows (punctate-striate), the striae and intervals can be detected. Intervals vary in number, development, convexity, and surface texture (smooth, tuberculate, or spinose). Interval widths are commonly more or less equal to those of striae, but they are almost twice as wide as the latter in Prenea.

Where striae are apparent, the first one is often less than half the length of a complete stria and has been referred to as the scutellary striole in Cassidinae (Würmli, 1975) (fig. 69D) and scutellar stria in other Coleopteran families (see Will, 2002). These terms are not precise since the mesoscutellum lacks striae. For precision and consistency of homology statements, Will (2002) argued for the use of the term “parascutellar stria”, which I follow here. The presence of the parascutellar stria has been noted in Anisoderini, Aproidini, Botryonopini, Callispini, Callohispini, Cryptonychini, Eurispini, Leptispini, and Oncocephalini (Würmli, 1975) and Prosopodontini (Maulik, 1931a), and its absence has been noted in Exothispini, Promecothecini, Coelaenomenoderini and Gonophorini (Würmli, 1975). Presence/absence of the parascutellar stria has not been commonly used in the taxonomy of Cassidinae s.str., althought it is mentioned in the description of Oocassida (Maulik, 1919). The number of punctures within the parascutellar stria varies among genera; for example, 2 found in Octotoma, 3 in Conchylotenia, 4 in Chalepus, 7 in Anisodera, 9 in Charidotella, and 13 in Demotispa, and 15 in Callistola. The number of punctures within the parascutellar stria can also vary between two elytra; for example, Estigmena has seven punctures on the left elytron and four on the right one.

Intervals. These may be smooth or coarse (dentate, serrate, spinose, or setose), and flat or convex. Patterns of surface texture and convexity can vary within the length of the interval and between intervals. In Lasiochila each interval is convex. In Anisostena, Anoplites, Baliosus, Callistola, Klitispa, Sceloenopla, and Uroplata, the intervals have an alternate arrangement of smooth and convex forms: intervals I, III and V are smooth and flat; intervals II, IV, and VI are smooth basally but convex beyond the humerus. These intervals may also be carinate posteriorly and connected at their apices by a single transverse carina. This transverse carina and intervals IV and VI are usually dentate; in Klitispa they are smooth. In Anisostena, interval III terminates immediately before this transverse carina. In Anisodera, only alternate intervals appear convex in their posterior sections. In Xenochalepus all intervals are smooth; the second interval ( =  the first convex one) is convex, prominent and shiny black; the sixth interval is slightly carinate basally but flat distally; the eight interval is slightly carinate from base to apex; all other intervals are flat; only the second interval is shiny black whereas all others are matte black. In Chalepus only the second interval ( =  the first convex one) has serration at its apex.

Some cassidines can have the disc with irregular convex lines and wrinkles, or with a net-like arrangement of ridges. Octotoma has carinate wrinkles. Anepsiomorpha has short, irregular, slightly convex lines, whereas Agenysa and Poecilaspis have a convex net-like pattern. Notosacantha also has a somewhat carinate, irregular lattice. These patterns are diverse and are not considered homologous. A second net-like pattern is also discernible within the elytral disc (e.g., Aspidimorpha) or within the elytral margin (e.g., Paratrikona).

Pubescence. Most cassidines are glabrous. In Dorcathispa, deep elytral punctures have a single long bristle, usually on the margin of the puncture. The bristle is stiff and erect, with a somewhat flattened appearance and a slightly inflated apex. Delocrania has single, long bristles at regular intervals along the lateral edge. In Demotispa, each tooth of the saw-like apical margin has a single long subapical bristle. The body of Trichispa has short, dense pubescence over the entire body, including the elytra. Fine pubescence occurs in Cyperispa, Oxynodera, and Promecotheca. In Cyperispa hairs are bent and form a sparse fringe along the lateral edge and are denser at the discal apex. In Promecotheca, pubescence is erect and concentrated at the discal apex.

Michalski (1995) first described the thickened transverse brace and the longitudinal carina often found on the internal side of the epipleural ridge structure in Stolaini. In the anterior third or midway along the venter of the elytra, a thickened transverse brace and a longitudinal carina are often found on the internal side of the epipleural ridge (fig. 69F–L). The carina varies in length, sometimes extending close to the brace, but not connecting with, the brace. In some cassidines, the carina joins the brace, and together they form a groove bordered internally by the carina, anteriorly by the brace, externally by the epipleural ridge, and open posteriorly. The elevation of both carina and brace varies, as does the distance between the carina and epipleural ridge. The carina height is short in Chirida but taller in Asteriza. Usually the carina is close to the epipleural ridge, positioned less than one fifth of the elytral width from the ridge. In Stenispa the carina is more distant from the ridge, positioned at about one fourth of the elytral width. Both the carina and brace may be entirely or partially covered with dense microtrichia.

Elytral Locking Mechanisms in Cassidinae. Morphological devices for locking the Coleopteran elytra to the body include a metepisternal muscle controlling the elytra (Breed, 1903), a tight sutural groove (Packard, 1898; Sharp, 1899; Breed, 1903; Breed and Ball, 1915), binding spicules (on the venter of the elytra) (Hammond, 1988; Samuelson, 1994, 1996; Bouchard and Gorb, 2000), insertion of the elytral anterior margin under the mesoscutellum (Breed and Ball, 1915), and insertion of the elytral epipleural ridge under or over ridges on the metepisternum (Breed and Ball, 1915).

Although a special metepisternal muscle has not been described in Cassidinae, many other elytral locks appear in the group. Morphology of these structures is described above. The sutural groove can be extremely tightly fitted in Elytrogona and some Stoiba species. A mesepimeral lobe lock holds the epipleural ridge in Aspidimorpha (fig. 66D). A metepisternal anterior lobe lock overlaps and holds the epipleural ridge (e.g., Basipta). The metepisternal lateral margin lock overlaps and holds the epipleural ridge in Basiprionota. The metasternal lateral marginal lock performs a similar function in Anisostena. The epipleural ridge lock inserts under or over ridges or cavities on the metepisternum (e.g., Dorynota). The longitudinal ventral carina (fig. 69F–L) near the epipleural ridge may lock the elytra to the metathoracic pleuron and/or to the pleuron of the first abdominal segment; the micro-trichiate surface may enhance the locking mechanism. Basolateral patches of binding microtrichia have been described in 27 species of Cassidinae (Samuelson, 1994, 1996). These frictional surfaces vary in their distribution, size, and general shapes and in the density and directionality of microtrichia (Hammond, 1979; Samuelson, 1994, 1996). Finally, the bisinuate basal margin in some genera (fig. 69C, E) may also check elytral opening.

Hindwings (Figs. 70–Figure 7172; Characters 157–159)

In the first broadly comparative study of hindwings in Chrysomeloidea, Jolivet (1954) sampled 24 species in 32 genera from 17 cassidine tribes. His illustrations show variation in wing development, coloration, vein development, cell development and sizes, and vein connections. The anal field development appears to vary, with reductions in Leptispa filiformis Germar and Octhispa (Hepthispa) limbata Baly. Suzuki (1969a, 1969b, 1970a, 1970b, 1994) has described additional taxa in his discussions of chrysomelid hindwing evolution. In Cassidinae, hindwings have been described for Alurnus spp. (Suzuki, 1994), Jonthonota mexicana (Sanderson and King, 1951), Thlaspidea cribrosa (Boheman) (Suzuki, 1994), Chrysochus auratus (Wilson, 1934), and Acromis sparsa (Chaboo, 2001). Chen (1940) proposed that Cassidinae s.str. and Hispinae s.str. share vein Cu1 continuous with anal cell 2, and he pointed out that the anterior cross vein is usually normal in Cassidinae s.str. but often atrophied in Hispinae s.str. Reid (1995) identified seven characters (chars. 18–24) from the hindwing for his higher-level phylogenetic analysis, and these were homogenous for Cassidinae.

The hindwing is commonly fully developed in Cassidinae. The veins Cu, Sc, RA, r3, MP, CuA, AA and AP₃₊₄ are usually present. The radial cell (rc) is developed and varies in shape (triangular or rounded) and size. The cell is relatively large in Arescus (fig. 70K), small in Asmangulia (fig. 70L), and barely discernible in Delocrania and Philaspis. The cell is commonly closed, but it may be open due to reduced sclerotization of RA₃₊₄. The former positions of RP₁ and RP₂ may be marked as dusky lines, while RP3+4 and r4 are absent. The presence or absence of trichiation on the veins has been used to identify species (Riley et al., 2001).

Figure 70

Wings of Chrysomelidae. A. Lamprosoma sp. 1. B. Lema trivittata. C. Ophraella sp. 1. D. Pachybrachis sp. 1. E. Plataumaris sp. 1. F. Sagra sp. 1. G, L. Cassidinae. G. Alurnus bipunctatus. H. Anisodera guerini. I. Anoplites inaequalis. J. Aproida balyi. K. Arescus sp. 1. L. Asmangulia sp. 1.

Cassidines show greatest heterogeneity in the development of veins MP4, CuA, and AA, and in cubital anal cells 1 and 2. These veins and structures are all present in Hemisphaerota palmarum (fig. 34B). Reductions of these veins results in open cubito-anal cells. CuA cell 2 is more commonly open than CuA cell 1, for example, Anisodera (fig. 70H) and Oediopalpa (fig. 72F). In some taxa, the veins are further reduced and both cells are lacking, for example, Anoplites (fig. 70I) and Calyptocephala (fig. 71I).

Figure 72

Hindwings of Cassidinae. A. Dolichotoma sp. 1. B. Imatidium capense. C. Ischyrosonyx oblonga. D. Jonthonota maculata. E. Odontota dorsalis. F. Oediopalpa guerini. G. Physonota sp. 1. H. Plautyauchenia deyrollei. I. Prenea strigata. J. Spilophora aequatoriensis. K. Syngambria bivulnerata.

Figure 71

Hindwings of Cassidinae. A. Aspidimorpha miliaris. B. Asteriza flavicornis. C. Baliosus californicus. D. Basiprionota chinensis. E. Batonota sp. 1. F. Botryonopa foveicollis. G. Calliaspis rubra. H. Callistola speciosa fasciata. I. Calyptocephala nigricans. J. Canistra plagosa. K. Charidotella bicolor. L. Coptocycla dorsoplagiata.

Jolivet and Hawkeswood (1995) listed Delocrania and Elytrogona as apterous, whereas Stoiba, Fornicassis, Cassida (Pilemostoma) and Cassida (Mionycha) are micropterous. Based on specimen examinations, I determined that Elytrogona species range from brachypterous to apterous (Chaboo, 2000), all Delocrania species have fully developed wing, and Stoiba species range from fully developed to micropterous wings.

Wing coloration (tinting) varies across the wing, with duskiness or stains marking original position of lost veins. The apical field is generally darker than the proximal portion, and the areas around the radial cell and central field can be particularly dark. The terminal portions of veins AA, AP, CuA, MP4, and the medial spur may taper off into duskiness.

Legs (Fig. 73; Characters 160–188)

In his key to cassidine tribes, Hincks (1952) defined four characters from the leg: claws divergent (commonly) or parallel (a synapomorphy for Dorynotini); tarsomere III with apical margin expanded and hiding the claw base (a synapomorphy for Eugenysini); claw base simple or appendiculate; and inner and outer basal margins of claws simple, striate, or appendiculate. Monophyly of Botryonopini is based on its femural tooth (Maulik, 1919; Würmli, 1975), and monophyly of Promecothecini is based on its long, toothed hindfemur (Würmli, 1975).

Reid (1995) considered chrysomelid leg morphology to be relatively unvariable, and he defined three characters (chars. 28–30): the presence/absence of the tarsal empodium, bifid tarsal setae, and spatulate tarsal setae. His tarsal empodium character is not discussed here since Cassidinae and my sampled outgroups lack this feature. In their analysis of galerucine-alticine relationships, Lingafelter and Konstantinov (2000) proposed eight leg characters: relative proportions of tarsomeres; mesocoxal position and separation; distinction of the trochantin; intermetacoxal distance; metafemural spring; and metatibial cross-section and metabial notch. These sources provide a starting pointing for searching for leg characters in Cassidinae.

Leg morphology appears heterogenous across Cassidinae, but all members have lost tarsomere IV (perhaps by fusion to the base of tarsomere V). The pro-, meso-, and metalegs do not differ greatly except in overall length, with the metalegs being slightly longer than the preceding legs. The femur is usually the longest segment, but the tibia is as long as or longer than the femur in Notosacanthini, Hemisphaerotini, Delocraniini, and Ischyrosonychini.

Coxae. The pro-, meso- and metacoxae vary in their relative sizes, shapes, prominence, and in intercoxal distance. Procoxal cavities are always closed behind by the hypomeral lobe and lateral extensions of the prosternal process, although the latter can appear very narrow (e.g., Arescus). Of the three pairs of legs, the inter-metacoxal distance is the shortest, separated by a narrowed abdominal intercoxal process. Intercoxal distances of the pro- and mesocoxae vary according to the development of the prosternal process. These coxae are proximate in Arescus due to the slender intercoxal processes, but they are more distant, almost laterally positioned in Chalepus and Cyperispa. Procoxae vary in shape and prominence. In most taxa, they are slightly elongate (pear-shaped) and well inserted into the coxal cavity. In a few taxa, they are globular and conical (e.g., Arescus and Exothispa). Margins of the procoxal cavities may be simple or emarginated, as discussed above. Mesocoxae may also be rounded or transversely elongate. In most cassidines, they are inserted into the mesocoxal cavities. In Arescus, the mesocoxae are rounded, appearing globose, and protuberant. Mesocoxal cavity margins may be simple or rimmed. Metacoxae appear homogenous across Cassidinae.

Trochanters are usually somewhat flattened and triangular and do not appear to vary between legs, except perhaps in relative proportions. Across Cassidinae, trochanteral morphology is also similar, but in Delocrania, trochanters are elongate (especially the pro-trochanter) and flattened, and they appear almost as long as the coxae. In Calyptocephala, the metatrochanter is slightly prolonged as a medial lobe beyond its margin with the metafemur. The trochanteral surface is commonly smooth or with sparse ventral setation. Mecistomela has tufts of setae on its trochanters (C. Staines, personal commun.).

Femur (chars. 163–164). This is always the thickest segment of the cassidine leg, and is slightly inflated medially. In Aproida, it is protuberant medioventrally, appearing slightly tuberculate. Femural length is greater than or equal to tibial length. The cross-sectional profile may be rounded or triangular. In triangular forms, the ventral surface is somewhat flattened, with inner and outer margins weakly to strongly developed. These margins are rarely spinose (e.g., Asamangulia). In Botryonopa, the femur is compressed laterally so that the base is narrowed, and a narrow groove runs along the length. The margins of this groove are blade-like. Cassidine femora can be smooth, wrinkled or punctate, and sparsely to densely setose. Femural setation may be unevenly distributed. Dorcathispa is unusual in having both femural and tibial surfaces spinose throughout and with erect, flattened setae similar to those found on the rest of the body. Anisostena has a rounded femur, with spines unevenly distributed on the medioventral face; these spines are generally short but vary slightly in sizes, and they taper off distally. In Asmangulia, these femural spines are fewer, and are better developed on the pro- and meso-coxae. The tibial and femural surfaces in Callistola and Ceratispa are shallowly punctate. Botryonopa (Maulik, 1919; Würmli, 1975) and Ceratispa have an apical femural tooth.

Tibia (chars. 165–169). Tibial length was discussed above with femural length. Cassidine tibiae generally have a triangular profile with the ventral surface rounded or flattened and with two margins, and the dorsal surface is rounded, angular, and ridged. In Aproida, they have a quadrate appearance. Tibiae are narrowest at the base and evenly tapered distally, reaching twice the width of the base. Cassidine tibiae are commonly straight, but they are strongly bowed in a few genera (e.g., Botryonopa) (Würmli, 1975). In Anisostena and Prosopodonta the bowed tibia and the spinose femur together resemble a nutcracker. Asmangulia legs have straight pro- and metatibia, but a bowed mesotibia. The tibial apex is inflated in Prosopodonta. All tibial apical margins in Arescus are rimmed with a row of short spines. Notosacantha, Ceratispa, and Callistola have a single spine on the mesal side of the apical margin. In the latter two genera this spine is small and pointed whereas in Notosacantha it is longer and hook-like. Oediopalpa has the mesotibia with an apical spine. These apical spines are not considered homologous with the sub-apical lobe in Sagra.

The dorsoapical surface of the tibia may be flattened or notched. The tibial notch or excavation notch receives the tarsus; it varies in size and depth, and its margin appears variously carinate. The margins may be slightly carinate in various combinations; for example, in Asamangulia, the basal margin is developed and slightly hooklike. Arescus has an unusual tibial apex where the notch appears enlarged because of the expansion of the anterior margin of the proleg (in dorsal aspect) and the posterior margin of the meso- and metalegs. This expansion is thin and gives the apex an asymmetrical appearance. Additionally the margin has a row of short spines. The tibial notch surface is commonly smooth, but it is finely microreticulate in Botrynopa. In Ceratispa both medial and external margins have a deep notch, with the medial notch slightly longer than the external notch. The external notch receives the tarsus but the medial one does not correspond with any particular structure. Tibiae may be smooth (e.g., Arescus), but they are more commonly sparsely setose proximally and more densely setose distally, especially in the apicoventral area. In some species tibiae may be slightly transversely or longitudinally wrinkled.

Tarsomeres (chars. 170–188). Cassidines share the tarsal formula 4-4-4, because tarsomere IV is lost on all legs. Tarsomeres I, II, and III vary in shapes, lateral margins (expanded; rounded or straight), basal margin (rounded or sinuate), and apical margin (straight or bilobed and degree of lobation). Tarsomere I is usually the shortest and narrowest, almost half the width of tarsomere III, and the apex is usually slightly wider than the base. In Arescus, it is as long as tarsomere II. In Cyperispa, it is expanded and as long as tarsomere II. Its lateral margins are commonly not expanded and are straight. In Delocrania, they are expanded and tarsomere I is as wide as tarsomere II. In Eurispa, the overall shape varies between legs, being expanded on the proleg but not on the metaleg. In some cassidines, the lateral margins of tarsomere I are of different lengths, with the posterior margin longer than the anterior margin, so the segment has an asymmetrical appearance (e.g., Basipta and Cyperispa). Tarsomere I is large in Leptispa. This segment may be symmetrical or asymmetrical in dorsal aspect (e.g., Eurispa). Callistola has several unusual features on its tarsus: tarsomere I is expanded and slightly asymmetrical, tarsomere II has the apical margin slightly concave and lateral margins parallel, and tarsomere III has enlarged lobes that overlap tarsomere V. The basal margin is indented in Klitispa and it corresponds with an indentation of the tibial apex. Apical margins are usually straight or slightly sinuate. Tarsomere I is usually evenly rounded in lateral profile.

Tarsomere II is usually intermediate in size and shape between tarsomeres I and II. The apical margin may be straight, slightly concave, or deeply bilobed with lobes initiating medially. Lateral margins are commonly rounded or, more rarely, parallel-sided (e.g., Alurnus). The lateral profile is commonly evenly rounded, but it has a slight medial longitudinal groove in Alurnus.

Tarsomere III is commonly the widest tarsal segment, almost twice as wide as tarsomere I, and it has the apical margin deeply bilobed. It is rarely narrow and with the apex not bilobed (e.g., Stenopodius). Staines (1988) amended the key of Arnett (1973) with this feature. Tarsomere III may also be bilobed or simple in other chrysomelids (e.g., Donaciinae; Monrós, 1959). The lobes are commonly symmetrical but they are asymmetrical in Anisostena in which the posterior lobe is longer than the anterior lobe. In Ceratispa the dorsal medial margin of the lobes appears slightly expanded and fits closely around tarsomere V, unlike in other cassidines.

Tarsomere IV is fused to tarsomere V, and the suture is commonly lost. In a few taxa the sutural line is faintly discernible, for example, Basiprionota (Chen, 1973, 1985). Crowson (1981) suggested that this unique loss within Chrysomelidae might be related to the broadening of the other tarsomeres and the shortening of tarsomere V in cassidines.

Tarsomere V is elongate and either evenly cylindrical or gradually widened anteriad. Its length varies widely; it can be short (e.g, Notosacantha) with the apex and claws hidden between the lobes and setation of tarsomere III. When long, the apex extends well beyond tarsomere III (e.g., 73A–E) and the claws are exposed. In some long forms, tarsomere V is arched. In Eugenysini, the distal margin of tarsomere V is expanded dorsally and overlaps the claw bases that are therefore hidden from view (Viana, 1968). Tarsomere V may have some ventral projections. Baliosus californicus has a pair of apicobasal triangular projections whereas Uroplata girardi and Chalepus have a single process, approximately half the claw length, and positioned medially between the claws and the claw bases.

Tarsal Setation. All tarsomeres are generally sparsely setose dorsally, with a small concentration of bristles medially on the anterior margins where they overlap the following segment (e.g., Botryonopa). Dorsal setation is fine and soft and longer than ventral setation. Legs in Basipta generally have dorsal setation longer and denser than is common and its tarsomeres have a medium dense cover. Ventrally, tarsomeres I–III are flattened and densely setose, forming the tarsal pad (Stork, 1980; Mann and Crowson, 1981a; Chen, 1985; Verma, 1999; Schmitt, 1989). These setae may be bifid (fig. 73U) and adhesive. The presence of bifid tarsal setae has been argued as the fundamental character supporting monophyly of Bruchidae, Sagrinae, Criocerinae, Donaciinae, Cassidinae s.str. and Hispinae s.str. (Stork, 1980; Mann and Crowson, 1981a; Schmitt, 1989). Schmitt (1989) refined this concept by pointing out the variation among these tarsomeres. In Bruchidae, Sagrinae, Criocerinae, and Donaciinae bifid setae occur only on the third tarsomere, whereas Cassidinae has them on tarsomeres I–III. Reid (1995) also suggested variation among subfamilies. Bifid tarsal setae and the adhesive oils that they secrete are discussed under Biology.

Figure 73

Prolegs and protarsi of Cassidinae. A, B. Anoplites inaequalis. C. Aspidimorpha miliaris. D, E. Baliosus californicus. F, G. Calliaspis rubra. H. Callistola speciosa fasciata. I, J. Calyptocephala nigricans. K, L. Delocrania panamensis. M. Dorynota pugionota. N. Imatidium fasciatum. O, P. Notosacantha badia. Q. Spaethiella sp. 1. R. Basipta sp., proclaws. S. Ischyrosonyx oblonga, proclaws. T. Physonota sp., proclaws. U. Basipta sp., bifid tarsal setae.

Claws (chars. 185–188). Cassidinae commonly have two claws, which are usually divergent (100°–180°). Parallel claws have been argued as a synapomorphy for Dorynotini (Monrós and Viana, 1949; Hincks, 1952), but they appear within many other genera of Hispinae s.str. Some genera exhibit a single claw, for example, Hispellinus Weise, Acymenychus Weise (Lopatin, 1984), some gonophoriine genera (Würmli, 1975), Eurispini (Gressitt, 1963), and Hemisphaerotini (Monrós and Viana 1951). Bifid tarsal claws, known in other chrysomelids (e.g., Ophraella) have not been described in Cassidinae. Cassidine claws are usually separated but they may be basally connate (e.g., Monochirus; Maulik, 1919). Connate claws are also known in other chrysomelids (e.g., Lema; Monrós, 1959). Cassidine claw bases are commonly exposed, but they are hidden in eugenysines (described above).

Claws tend to be tapered distally and pointed, but they are broadened apically in Acymenychus (Maulik, 1919). Cassidine claws are commonly symmetrical, but asymmetrical ones have been reported in Asamangulia (Maulik, 1919), some genera in Dorynotini, Hemisphaerotini and Cassidini (Monrós, 1949; Riley, 1986), and in Dactylispa (personal obs.). The claw ventral margin may be blade-like, or flattened and with inner and outer margins. The ventral margin has a single tooth in genera of Eugenysini and Stolaini. In many Aspidimorphini, Cassidini, and Charidotini, the claw base may be pectinate (comb-like) (Riley, 1985), with pectens varying in length, size, general shape, and surface texture (Maulik, 1919). Some pectens are discernible under a dissecting scope, but SEM is best for unambiguous state determination. In Coptocycla both margins are finely pectinate, with pectens tapering off apically and basally. In Conchylotenia, both margins are also pectinate, with pectens reaching about one-third of the claw length.

Claw Appendages. The ventral margins of claws are commonly simple. Hincks (1952) defined Eugenysini and Stolaini as sister taxa on the basis of the single tooth on the ventral margin of their claws. Other cassidines may also have a single basal tooth (e.g., Oocassida cruenta; Maulik, 1919). Stoiba and Elytrogona show taxonomically useful variation in the tooth margin (Chaboo, 2000). Claw pectens, being rows of small teeth on the basal margins of claws, have been used taxonomically in Cassidini, Charidotini and Aspidimorphini. These pectens vary in their presence/absence on the inner and outer margins and in the numbers and lengths of individual teeth. Riley (1986) synonymized Charidotini and Cassidini based on pecten evidence. Pecten features are used to identify some Panamanian cassidine genera (Windsor et al., 1992). Some cassidines have pectens on both the inner and outer basal margins of each claw, and their numbers vary between the medial and external edges. Pectens vary in their presence or absence, numbers between mesal and external rows, and in sizes within rows. Sindia has claws with simple outer margins and five to six dens of variable sizes on the mesal margin; the dens decrease in size proximally (Maulik, 1919). Conchylotenia has four to five pectens mesally and two to three pectens externally. These different sizes of pectens may be described as macropectinate and micropectinate. Aspidimorpha, Sindiola, and Laccoptera have both margins pectinate (Maulik, 1919; Gressitt, 1952).

Trends in Leg Morphology. Compaction of the legs appears to be part of a general syndrome of compaction in the cassidine body. In Notosacanthini, Hemisphaerotini, and Delocraniini, legs are relatively short and the femur-tibial joints do not reach beyond the elytral margin. In other taxa, long legs extend beyond the elytra, even when the latter is strongly explanate. Within the leg, the loss of tarsomere IV and the tibial notch that receives the tarsus also point to leg compaction.

Internal Morphology

Ventral Nerve Cord. Kasap (1979) first described the nerve cord in Hispa testacea. In their general review for Chrysomeloidea, Mann and Crowson (1983b) found the number of free abdominal ganglia in Cassidinae varying between three and six.

Midgut. Shrivastava and Verma (1982) and Verma and Shrivastava (1989) studied this aspect in Cassidinae and discussed peritrophic membranes and enzyme functions in Aspidimorpha miliaris.

Abdomen (Fig. 39a; Characters 189–210)

Cassidines have five exposed sterna that correspond to abdominal segments III–VII. Sterna III and IV are commonly connate in Cassidinae but they are separated in genera such as Ceratispa and Plesispa. The sutural line may be apparent (e.g., Brontispa) or evanescent (e.g., some Hispodonta). Sterna III–VI are usually glabrous or with sparse setation, and they are usually impunctate; Callistola has shallow punctation. Pubescence may be concentrated along the posterior section of lateral margins. This lateral pubescence increases slightly posteriorly and sternum VII has dense pubescence along the posterior margin. Sternum III usually has the anterior margin simple, or which may be thickened in a few cases. In Alurnus, it is somewhat marginate. Abdominal intercoxal processes are usually similar across Cassidinae, but in Callispa the margin is convex and continues across the sternum as a carina to about one-third the length of the sternum. Sterna III–VI have the anterior and posterior margins parallel, and tergum VII has the anterior margin convex and the posterior margin straight. Sternum VII, the propygidium, is more heavily sclerotized than other sterna, and it is hemispherical in shape, with a rounded, distal posterior margin; it is densely pubescent in the posterior half. Sternum VII in the female is slightly indented in the extreme lateral area; this indentation is absent in the male sternum VIII. The terga are hidden by the elytra except the posterior portion of tergum VIII. All spiracles are present in the pleural fold and gradually decrease in size posteriorly.

Reproductive System (Figs. 74–Figure 7576; Characters 194–210)

Reproductive systems are a traditional source of phylogenetic characters in entomology. It is therefore remarkable that Cassidinae reproductive systems have been generally presumed to be unvarying; they are routinely omitted in species descriptions and are virtually ignored in their systematics. Higher level chrysomelid subfamily distinctions in genitalia have been described (e.g., Chen, 1940), but within Cassidinae, key revisionary works (e.g., Hincks, 1952; Viana, 19641122; Würmli, 1975), omit reproductive morphology altogether. Spaeth, the most prolific cassidine worker, never discussed these systems.

Male Reproductive System (fig. 74; chars. 192–196). Cassidine male genitalia have been included in surveys by Sharp and Muir (1912), Powell (1941), Mann (1988a, 1988b), and Mann and Crowson (1981b, 1983a, 1984b). The genitalia comprise the aedeagus, the ejaculatory duct that connects to the testes and to the flagellum at the base of the internal sac, and an incomplete tegmen.

Figure 74

Aedeagus of Cassidinae. A. Demotispa sp. B. Palmispa parallela. C. Stenispa metallica. D. Notosacantha sp. E. Notosacantha sp., lateral view. F. Alurnus bipunctatus, confocal miscroscope image of ejaculatory guide.

The aedeagus is commonly curved and forms an angle of varying sizes with the basal piece. The aedeagal apex may be rounded or acute and drawn out into a small hood flanking the ostium (or apical orifice). The lower wall of this ostium is differentially sclerotized into single or paired lateral plates (Powell, 1941). Lateral lobes are lacking. Notosacantha (fig. 74D–E) uniquely has a pair of apicolateral articulated flanges. These have not been described before, probably because they are easily disconnected under KOH treatment and were probably not noticed. The ostium can vary in size, from an opening half the diameter of the aedeagus to as wide as the aedeagus.

Externally, the basal piece or tegmen is attached to the ventral or posterior edge of the median foramen, and it consists of a single median manubrium and a pair of lateral processes that have been called arms, apopophyses, or struts. The latter do not completely enclose the base of the penis, but they are connected by muscles around the latter.

The internal sac (fig. 74E, F) is the inverted apical portion of the aedeagus and is continuous with the apical orifice. It is eversible and enters the female's bursa during copulation. The walls may be membranous or lightly sclerotized, and the surface may be armed with spicules and/or pubescence (Mann and Crowson, 1996652). The sac appears to have two areas, a simple tubular section and a section where muscles envelop the flagellum (Mann and Crowson, 1996652). Cassidines have an endophallic sclerite or ejaculatory guide (fig. 74F) at the base of the sac (Powell, 1941). This sclerite may be a symmetrical structure or reduced to a single bar. It marks the point of connection between the proximal ejaculatory duct and the distal flagellum. The sclerite is notched and its surface may be spinose.

With the exception of Notosacantha, the aedeagus of most cassidines does not appear to vary greatly except in subtle aspects of the ostium and general angulation. However, the internal sac seems to offer more diversity in its length, internal folding, surface texture, and ejaculatory sclerite. Mann and Crowson (1996) suggested that there was “a polyphyletic development of the internal sac in the Old and New World species of Cassidinae”, and they identified variations in symmetry, presence or absence of folds of the internal walls, presence of a dorsal hook, chitizinization patterns, setation distribution, spicule presence and arrangement, and position of the ejaculatory guide. My confocal study of two species showed that the morphology of the latter sclerite may be very different (Chaboo, unpubl. data), with discrete variation in the number and arrangement of projections.

The flagellum is a semisclerotized, wire-like tube that is enveloped by hairy folds of the internal sac wall (Barber, 1946; Mann and Crowson, 1996652). Its length is highly variable, and Rodriguez (1994a, 1994b, 1995) revealed that once the internal sac is everted into the female, the flagellum enters the spermathecal duct and travels up the duct, and even into the spermatheca, to discharge sperm.

The ejaculatory duct may be short, up to three times longer than the aedeagus, or more commonly it is longer than the total body length. It is especially elongate and coiled in Eugenysa species (Chaboo, 2002189). The duct consists of a short proximal sclerotized rod that terminates at the paired testes, and a long flexible distal ejaculatory duct that enters the basal foramen, and connects with the flagellum. Duct length may be correlated with spermathecal duct length (Eberhard, 1996).

Cassidine testes may be single or bilobed, with the folliculae distinct or indistinct and varying in number (Virkki, 1957, 1969). An accessory gland may be present or absent, single or paired, with varying shapes and number of lobes. Spermatozoa studies along the lines of Virrki and Bruck (1994) will help determine variations within Cassidinae and among other subfamilies.

Apart from Rodriguez’ (1994a, 1994b, 1995; Rodriguez et al., 2004) study of the physiology and mechanics of cassidine copulation, the only other functional study has been Verma and Kumar (1972). They described the process of “retournement” in Aspidimorpha miliaris where the aedeagus rotates 180° during copulation. This has not been studied further, so it is unclear how widespread is the phenomenon or what is its significance.

Female Reproductive System (figs. 39, 40, 75–76; chars. 198–210). The first study of cassidine female genitalia appeared in Muir and Sharp (1912). Spett and Lewitt (1927, 1928) survey of chrysomelid spermathecae included seven cassidine species. Sanderson (1948) examined four Physonota species, while Mann and Crowson (1983)647 surveyed 20 cassidines. Suzuki (1977a, 1977b, 1983, 1988b, 1997)1026 summarized the literature on chrysomelid reproductive systems, and later discussed genitalic character correlation in Chrysomelidae (Suzuki, 1989b). Leonardi (1972) suggested that spermathecal variation might be useful in taxonomy. Rodriguez' (1994a, 1994b, 1995) study of the reproductive system in Chelymorpha alternans revealed the mechanics of the ejaculatory and spermathecal ducts. Borowiec and Skuza (2004) examined spermathecal morphology in 45 species of Chelymorpha and subdivided the genus into 10 subgroupings based on variation in the ampulla shape and length and the length and coiling of the spermathecal duct.

Figure 75

Spermathecae of Cassidinae. A. Acromis venosa. B. Alurnus bipunctatus. C. Amythra reticulata. D. Anepsiomorpha deplanata, distorted. E. Anisodera guerini. F. Arescus sp. G. Asamangulia sp. H. Baliosus californicus. I. Batonota lerouxii. J. Botryonopa foveiocollis. K. Calliaspis rubra. L. Calliaspis rubra, entry of gland and duct.

Figure 76

Spermathecae of Cassidinae. A. Calyptocephala nigricans. B. Entry of gland and duct. C. Cephaloleia sp. D. Entry of gland and duct. D. Ceratispa palmicola. E. Entry of gland and duct. F. Demotispa sp. G. Dorynota ballista. H. Dorynota pugionota, entry of gland and duct. I. Echoma quadristillata. J. Eugenysa martae.

The ovipositor has the spiculum ( =  sternum IX; Jeannel and Paulian, 1944) lightly sclerotized, and the coxites are generally shorter than tergum VII. Tergite IX ( =  paraproct; (Teotia, 1958);  =  hemitergite (Konstantinov, 1963)) may be a single plate in some cassidines (e.g., Promecotheca) or two separate plates that connect basally by membrane (e.g., Miocalaspis).

Cassidine spermathecae (figs. 39, 40, 75, 76) are heavily sclerotized and commonly falcate, with muscles on the inner margin. Two regions, the pump and receptacle, are identified by their position and differences in diameter. They are usually broadly joined but may be distinguished externally by an indentation between them (figs. 75C, D, 76L) or, more rarely, by an internal wall that may act as a valve (fig. 75G). The pump is commonly longer than the receptacle, curving over the latter. The receptacle is basal and usually inflated, with a larger diameter than that of the pump. Spermathecal muscles attach to the mesal surfaces of the pump and receptacle, and their contractions pull the two together and probably control sperm expulsion. The receptacle receives the spermathecal duct and spermathecal gland. Most cassidine spermathecae have a single receptacle but a few are two or three-chambered, (e.g., Eugenysa; Chaboo, 2002189), and the gland and duct enter separate chambers.

Cassidines display diverse arrangements in the entries of the spermathecal duct and gland into the spermatheca. In some, the gland and duct connect and enter the spermatheca as a single tube (figs. 39D). In most cases, the insertion is on the surface or is slightly elaborated on a thick section of the duct (fig. 75F); in a few cases the insertion is inverted into the spermatheca (fig. 75C). In most cassidines the gland and duct have separate insertions, and in cases of multiple receptacles, they enter different chambers. The insertions are usually basal and may be proximal or widely separated. Conchylotenia has an unusual elongate spermatheca in which the receptacle is inflated and the pump is about five times longer than the latter, greatly narrowed and sinuate. Examination of freshly killed specimens will help determine muscle attachments.

The pump usually has an apical appendix{ label needed for app[@id='app2'] }{ label needed for app[@id='app3'] }{ label needed for app[@id='app4'] }{ label needed for app[@id='app5'] } (e.g., figs. 75A, 76I), a flap-like projection in a supraapical or subapical position. This appendix{ label needed for app[@id='app2'] }{ label needed for app[@id='app3'] }{ label needed for app[@id='app4'] }{ label needed for app[@id='app5'] } may be oriented longitudinally or horizontally to the axis of the spermatheca, and it serves as an attachment for spermathecal muscles. A few cassidines have an additional medial receptacle appendix{ label needed for app[@id='app2'] }{ label needed for app[@id='app3'] }{ label needed for app[@id='app4'] }{ label needed for app[@id='app5'] } (fig. 75I), which provides additional attachment points and may enable greater development of spermathecal muscles.

The spermathecal or accessory gland of cassidines is a long, flattened ribbon-like gland that is held in situ by connective membranes. It is easily torn in routine dissections, making comparative study difficult. The spermathecal duct displays much variation in length, pattern of coiling or folding, the presence of sclerotized sacs, and the enlargement of the base before inserting into the spermatheca. The duct is always longer than the length of the spermatheca, but its length can range from 2–5 times longer to >50 times longer than the receptacle. Long ducts are folded in some way, either by spiral coiling (like an old-fashioned telephone cord; e.g., Omaspides) or broad looping folds (e.g., Asteriza). Folding patterns may be consistent throughout the duct's length. In some cases, the proximate or distal section is folded while the second half is more or less straight. In the spermathecal duct of Chelymorpha cribaria (Fabricius), Rodriguez (1994)876 distinguished a region, the ampulla, which he proposed as an entanglement trap; the male flagellum is inserted into the duct and eventually coils on itself within the ampulla. The ampulla is difficult to identify and therefore homologize across Cassidinae. In Polychalca and Eurypepla, I found a large sclerotized sac along the spermathecal duct, and in cleared specimens it appears that a long section of the duct is tightly packed within this sac.

Rodriguez (1994a, 1994b, 1995) revealed that the male's flagellum travels up the spermathecal duct and sometimes reaches into the spermatheca where it deposits sperm. At the physiological level, it is remarkable that these microscopic tubes are capable of traveling such distances. The mechanism driving the flagellum up the spermathecal duct toward the spermatheca remains unclear. At the behavioral level, the interplay of spermathecal duct length and flagellum length probably influences courtship, mate selection, and sperm selection (Eberhard, 1996).

The bursa copulatrix receives the internal sac of the penis during copulation (Lindroth, 1957). It is coated in thick muscle fibers in cassidines. Stammer (1936b) first described the presence of vaginal pouches in cassidines, and Mann and Crowson (1983)647 found vaginal pouches in all 20 cassidine species they surveyed. These pouches open to the vagina in Cassidinae and Eumolpinae and to the common oviduct in Sagrinae. Cassidine vaginal pouches are paired and open into the lower third of the vagina. They may also be called mycetomes as they house bacteria-like microorganisms in Eumolpinae, which are transmitted to the eggshell (Stammer, 1936a, 1936b; Kasap, 1975; Crowson, 1981; Becker 1994; Becker, and Ferronatto, 1990). In Eumolpinae, the pouch consists of a sac and a duct, with the duct associated with a pair of accessory glands. In addition to the vaginal pouch, chrysomelid colleterial glands may also house microorganisms that are transmitted to the egg (Becker, 1994; Suzuki, 19881029).

Colleterial glands in insects are modified accessory glands that open into the common oviduct (Selman, 1994). They produce a colorless secretion that hardens into distinctive surface sculpturations around the egg, called an extrachorion (Hartley, 1961). Such extrachorionic material is produced by several glandular areas in cassidines: ectodermal accessory glands that open adjacent to the vulva; glandular epithelia in the lateral and pedicels of ovarioles in Gratiana spadicea (Klug) (Becker and Romanowski, 1986), Aspidimorpha icterica Boheman ( =  A. puncticosta Paterson) and Aspidimorpha tecta Boheman (Atkins et al., 1966); and the vagina of A. icterica (Muir and Sharp, 1904) and A. tecta (Hinton, 1981).

In addition to the coating of cassidine eggs with microorganisms and an extrachorion, some species enclose their egg masses in an elaborate oothecal case. The case can comprise up to 200 membrane layers (e.g., Aspidimorpha puncticosta; Muir and Sharp, 1904). The production of so many membranes must require a high volume of oothecal secretions. Becker and Romanowski (1986) suggested that secretions must come from the extensive surface of the ovariolar pedicels and not from the smaller glandular epithelium that open into the cloaca.

Chrysomelid ovariole numbers have been studied (Mann and Singh, 1979; Robertson, 1961; Mann and Singh, 1979; Suzuki, 1974; Suzuki and Hara, 1975; Suzuki and Yamada, 1976). The number of cassidine ovarioles range from 3 to 21 (Mann and Singh, 1979; Robertson, 1961). In Gratiana spadicea there are usually one to two maturing oocytes at any time (Becker and Romanowski, 1986). Bhattacharya and Verma (1982) defined a period of diapause in summer by arrested ovarian development and hypertrophic body fat development in Aspidimorpha miliaris. The internal surface of lateral oviducts of Gratiana spadicea has a cuticular intima lining with spiniform projections that may help in controlling oviposition (Becker and Romanowski, 1986).

Alurnini Weise (6 Genera, 2 Sampled; 29 Species, 2 Sampled)

Alurnines were originally characterized by the large-sized adults, 11-segmented antennae, and pronotal angles with setal tufts (Fisher, 1935; Staines, 2002990), but these are homoplasies across Cassidinae. The tribe has been allied with Arescini and Prosopodontini in various catalogs (e.g., Seeno and Wilcox, 1982). Hsiao and Windsor (1999) presented the novel hypothesis (Alurnini + (Cryptonychini + Hispini)) based on molecular support. Duckett et al. (2004) resolved their single alurnine exemplar as derived in the odd relationship (Coptocycla + (Alurnus + Chelymorpha)). The two alurnines sampled here are recovered separately, with Alurnus placed basally at clade 8 (Arescini + (Alurnini + all other cassidines)) (Bremer support 15) and Platyauchenia placed in a more derived position at clade 43, (cryptonychines + (Platyauchenia + derived cassidines)) (Bremer support 1). The latter finding provides some support for a putative relationship between alurnines and cryptonychines. The lack of support for monophyly of Alurnini is unexpected. Immature stages are known for only a few species in Alurnus (Bondar, 1916; Merino and Vasquez, 1963; Villacis Santos, 1968; Strong, 19771011), Coraliomela (Bruch, 1939141), Mecistomela (Macêdo et al., 1994), and Platyauchenia (Maulik, 1933a); however, they exhibit similar morphology, development and behavior. A comparative morphology of known immatures and more characters from immature stages should inform the monophyly, position, and internal relationships of Alurnini.

Alurnines are the largest cassidines and are primarily specialists of palms (Villacis Santos, 1968; Macêdo et al., 1994; Mariau, 2004). The life cycle lasts over 1 year, with up to nine larval instars (Villacis Santos, 1968; table 5). Larvae of Alurnini, Arescini, and Cephaloleiini have been grouped into one ecological guild, rolled-leaf and sheath feeders (Staines, 2004); however, alurnine larval morphology is distinct from the latter tribes. They are heavily sclerotized with bladelike margins that may permit cutting thick fibrous tissues of unopened palm leaves. Both larval morphology and host plants suggest that alurnines are distinct from rolled-leaf feeders.

Anisoderini Weise (3 Genera, All Sampled; 75 Species, 3 Sampled)

Würmli (1975a) diagnosed this moderately speciose tribe by the convex eye facets. It was not included in Borowiec (1995) and Hsiao and Windsor (1999). All three anisoderine genera are represented here and are resolved within clade 11 as sister to Promecotheca. They are united in the monophyletic subclade 12 by the autapomorphy of having eye facets convex (char. 43) and by homoplasies in characters 23, 94, 101, 145, and 175 under unambiguous transformation. This finding concurs with Würmli's (1975a) diagnosis of Anisoderini. The tribal phylogeny is resolved as (Lasiochila + (Anisodera + Estigmena)).

Anisoderines are among the largest cassidines, with elongate, heavily sclerotized bodies. Lasiochila has leaf-mining larvae (Kimoto and Takizawa, 1997) whereas Anisodera and Estigmena both have stem-mining larvae (Beeson, 1941). The clade topology indicates that stem mining was a single derivation from a leaf-mining ancestor. The functional significance of the convex eye margins is unclear.

Aproidini Weise (1 Genus, 3 Species; 1 Sampled)

Würmli (1975) considered this monotypic tribe aberrant among Hispinae s.str., with adults resembling Donaciines more than other hispines. In lateral aspect, the body is rounded ventrally and is slightly upwardly curved at the head and pronotum and at the elytral apex. The mode of pupation is also unique in Cassidinae; the fifth larva becomes affixed by the abdomen to the plant, and as the pupa emerges, the larval exoskeleton is peeled backward, and the tracheoles become distended (Monteith, 1970). The pupa is thus suspended from the leaf by the tracheoles and the larval exuviae (Monteith, 1970). Metasternal distortion in aproidines suggests flightlessness (Samuelson, 1989).

Aproidini was not considered in previous cladistic assessments of Cassidinae. The single exemplar, supported by several autapomorphies from the adults and immatures, here is placed in clade 36 (Bremer support 1) with the relationship (Eurispa + (Aproida + Exothispa)) based on homoplasious support. Previous hypotheses of a relationship with Eurispa (Chapuis, 1875; Chen, 1973; Würmli, 1975) are somewhat supported here. Other suggested relationships with Anisoderini (Weise, 19111153; Würmli, 1975; Samuelson, 1989) and an intermediate placement between Hispinae s.str. and Cassidinae s.str. (Lawrence and Britton, 1994) are not supported.

Arescini Weise (4 Genera, 17 Species; 1 Sampled)

Staines (2002)990 diagnosed Arescini by the lack of the head keel and the enlarged pro- and mesocoxae. The latter feature is redefined here as characters 160 and 161. Arescini was not included in Borowiec (1995) and Hsiao and Windsor (1999). Wilf et al. (2002) inserted Arescini as sister to Cephaloleiini a posteriori on the topology of Hsiao and Windsor (1999). Arescus appears to form a distinct terminal within Cassidinae, and its branch is robustly supported here on all located MPTs as sister to all other Cassidinae (clade 7, Bremer support 14), with the latter supported by two autapomorphies: larval mining and adult abdominal sterna I and II fused. This is a novel hypothesis for Cassidinae and for the position of this tribe. Variation in cassidine thorax morphology appears to arise from compaction in the longitudinal line and to changes in the coxal articulation; characters 160 and 161 may be correlated with each other and also with changes in the surrounding segments and muscles.

If Arescini is indeed the most plesiomorphic cassidine, this has significant implications for our understanding of evolutionary patterns in Cassidinae. Arescines are considered rolled-leaf feeders (char. 4, state 0), and this behavior appears as convergent within Cassidinae.

Aspidimorphini hincks (7 Genera, 3 Sampled; 281 Species, 3 Sampled)

Hincks (1952) diagnosed this Old World tribe by its pectinate claws (char. 187), but this feature is widespread among Cassidini, Charidotini, and Aspidimorphini. Consequently, Cassidini and Charidotini were synonymized by Riley (1986) and then amalgamated with Aspidimorphini by Borowiec (1995). A single aspidimorphine exemplar, Laccoptera, was sampled by Hsiao and Windsor (1999) and was resolved among genera of Cassidini. Of the three aspidimorphine genera sampled here, two are united as clade 60, Aspidimorpha + Conchylotenia, while the third falls out in clade 69 as Laccoptera + Psalidonota. No apomorphic characters support either clade. Pectinate claws (char. 187, state 0) have two origins under all optimizations. These results support the previous synonymies (Riley, 1986; Borowiec, 1995). Aspidimorphini was poorly diagnosed from its inception, and its boundaries collapse within the context of derived cassidines. Some sampled genera of Aspidimorphini, Cassidini, and Charidotini are united in the monophyletic clade 68 (Bremer support 3). Further investigation of the Aspidimorpha + Conchylotenia clade will determine if they form a reliable monophyletic clade. Borowiec's (1994, 1997b) and Świętojańska's (2001) treatment of taxonomic aspects of aspidimorphines provide steps for their exploring systematics and biology.

Asterizini Hincks (1 Genus, 2 Species; 1 Sampled)

Hincks (1952) diagnosed this small monotypic Neotropical tribe by its narrowed and thickened lateral elytral margins, but this feature is homoplasious across Cassidinae. Borowiec (1995) hypothesized a close relationship among the genera of Asterizini, Ischyrosonychini and Physonotini and synonymized these three tribes under the group name Ischyrosonychini. The diagnostic features listed by Borowiec (1995) are also homoplasious across Cassidinae. Asteriza is recovered here as sister to Aspidimorphini in clade 59. This relationship has not been considered previously and is weakly supported here (Bremer support 3) by three characters (49, 90, and 193) under unambiguous and delayed optimizations, as well as by characters 56, 68, 155, 175, 187 under accelerated optimization. The Asterizini lineage (node 59, fig. 78) is not supported by any apomorphies. New data on immature stages and host plants collected recently by H. Matsuzawa (personal commun.) should help resolve this tribe's status.

Basiprionotini Hincks (7 Genera, 81 Species; 1 Sampled)

Tribal monophyly has been argued on the basis of the almost-completely hidden mouth and the triangular shape of the clypeus (Hincks, 1952). A close relationship with Epistictini was argued on the basis of antennal striations (Hincks, 1952). Borowiec (1995) united Basiprionotini and Epistictini under the older group name, Basiprionotini, and resolved this as a plesiomorphic branch in one of his two major subclades. Zaitzev and Medvedev (1982) also regarded Basiprionotini and Epistictini as closely related and probably the most primitive of cassidine tribes. Hsiao and Windsor (1999) did not sample Basiprionotini or Epistictini. The single basiprionotine sampled here does not test tribal monophyly; however, previous hypotheses could be examined. Basiprionotini and Epistictini are recovered here as sister taxa in a robustly supported clade 64 (Bremer support 16) (fig. 89). This finding supports hypotheses of Hincks (1952), Zaitzev and Medvedev (1982), and Borowiec (1999). The classical autapomorphy, striate antennae, is found to have three origins in Cassidinae and is not unique to clade 65. Synapomorphies of clade 64 include antennomere I punctate, mouth anterior margin between eyes, elytra lacking both a transverse internal ridge, and the longitudinal internal carina under unambiguous and fast optimizations, and additionally by the partial exposure of the head under slow optimizations. None of these character states is unique to the Basiprionotini + Epistictini clade; however, the strong Bremer support suggests that this is a distinct clade among the derived Cassidinae. The almost completely hidden mouth provides an autapomorphy for the branch Basiprionotini.

The clade Basiprionotini + Epistictini comprises eight genera and ca. 100 species and forms a well-circumscribed focus taxon for future study. Basiprionotine are found in parts of Asia and Madagascar, and only four host plant species of Verbenaceae are known (Shultze, 1908; Gressitt, 1952; Medvedev and Eroshkina, 1988; Borowiec, 1999). The pest status of Craspedonta leayana was summarized from Garthwaite (1939) and Beeson (1941) under Biology above. Immature stages are described for Basiprionota (Schultze, 1908; Gressitt, 1952; Gressitt and Kimoto, 1963; Zaitsev and Medvedev, 1982; Chen et al., 1986) and Craspedonta (Weise, 1901; Garthwaite, 1939; Beeson, 1941; Zaitsev and Medvedev, 1982; Cox, 1988; Zaitsev, 1992; Świętojańska and Ghate, 2003), collected but undescribed for Megapyga (appendix 2), and are unknown for the remaining basiprionotine genera, Androya Spaeth, Cassidopsis Fairmaire, Metriopepla (Waterhouse), and Pseudoandroya Spaeth. Borowiec's (1994) review of afrotropical basiprionotines provides a firm basis for further systematic work on this group.

Basiptini Hincks (1 Genus, 5 Species; 1 Sampled)

Hincks (1952) diagnosed this monotypic African tribe by the metepisternal inflation and a trapezoidal clypeus. Borowiec (1995) hypothesized a close relationship with genera of Cassidini and synonymized it under this name. The single Basipta species sampled here does indeed appear among derived cassidines of clade 54 in initial tree searches, but it was subsequently filtered from analyses due to high levels of missing information. Diagnostic characters proposed by Hincks (1952) are homoplasious in the context of Cassidinae. However, the independent findings of a paraphyletic Cassidini by Hsiao and Windsor (1999) and in the present study suggest that Boroweic's (1995) synonymy may be premature.

Recent collection of the life cycle of one Basipta species revealed a complex ootheca, five instars with exuvio-fecal shields, solitary pupation, and no parental care (Chaboo et al., unpublished data).

Botryonopini Weise (2 Genera, 40 Species; 1 Sampled)

Würmli (1975a) circumscribed botryonopines by the absence of pronotal punctation and the absence of pronotal sensory tufts. This particular character combination is not unusual among cassidines. Furthermore, the first character is incorrect since Botryonopa has pronotal punctation and the second feature applies to many cassidines. Gressitt's (1950) character of the proximity of the mouth cavity to the antennal insertions in Botryonopa will require a major comparison across Cassidinae to demarcate unambiguous states.

The single exemplar included in my analysis is resolved in clade 19 as ((Botryonopa + Callistola) + ((Coelaenomenoderini + Gonophorini) + Hispini))) and is supported by characters 106 and 203. Subclade 20, Botryonopa + Callistola, is weakly supported (Bremer value 2) by the synapomorphic characters 106 and 203 under unambiguous transformation, by characters 9, 64, 99, 147, and 152 under fast optimization, and by characters 64, 99, 107, 137, and 146 under slow optimization. Brief descriptions by Gressitt (1950) suggest that a thorough analysis of head morphology may reveal mouth character support for this obscure tribe. Maulik (1949b) described the larvae and pupa of Botryonopa sanguinea Guérin that were collected from the unopened buds of the palm, Metroxylon; Mariau (2004) indicated a coconut palm host for B. sanguinea Guérin.

Callispini Weise (6 Genera, 2 Sampled; 173 Species, 2 Sampled)

Würmli (1975) circumscribed this tribe by the first five puncture rows curving outward basally, a feature that is difficult to distinguish across Cassidinae. Borowiec (1995) placed Callispini within a large subclade of Cassidinae. Callispa and Hispodonta were sampled here and appear at clades 44 and 45 (fig. 84) respectively, indicating nonmonophyly of Callispini and under all optimizations.

Callispines occur in the Old World and specialize on Arecaceae, Musacaeae, Orchidaceae, Poaceae, and Zingiberaceae (Gressitt, 1960a, 1960b, 1963b; Gressitt and Kimoto 1963; Jolivet and Hawkewood, 1995). Immature stages are known in the genera Hispodonta (Gressitt, 1960a, 1960b, 1963a; Gressitt and Kimoto, 1963) and Callispa (Uhmann, 1949b; Kalshoven, 1951; Gressitt and Kimoto, 1963; Chen et al., 1986; Zaitsev, 2001) but are unknown for Amblispa Baly, Pseudocallispa Uhmann, and Spilispa Chapuis. The extremely thin and flattened larval body appears to resemble Imatidiine and Cephaloleiine larvae, as well as those in the coleopteran families Psephenidae (water pennies) and Colydiidae (W. Shepard, personal commun.), but detailed comparative study is needed.

Cassidini Hincks (76 Genera, 3 Sampled; 964 Species, 3 Sampled)

This is the largest cassidine tribe, accounting for 16% of subfamily diversity (fig. 2). Members include the quintessential cassidines with well-developed pronotal and elytral extensions, disc-like profiles in dorsal view, and metallic golden colors. Many species are relatively well studied. Cassidini have been difficult to circumscribe in a satisfactory manner. Hincks' (1952) diagnosis was based on negative characters (his couplet 42): absence of inflation of the metepisternum, and metepisternum not projected beyond metepimeron. Seeno and Wilcox (1982) recognized Cassidini, Aspidimorphini, and Charidotini as three separate tribes. Riley (1986) conducted a detailed study of the pectinate claws and concluded that charidotines belong in Cassidini. Borowiec (1994a) agreed with this synonymy and redefined Cassidini to include aspidimorphines. Hsiao and Windsor (1999) treated Cassidini + Charidontini and Aspidimorphini as two distinct groups, and they found a polyphyletic Cassidini with genera resolving in five different locations on their topologies.

Three Cassidini genera (Charidotella, Jonthonota and Orexita) are sampled here, along with all three aspidimorphine genera (Aspidimorpha, Conchylotenia and Laccoptera) and two charidotine genera (Metrionella and Psalidonota). Altogether 15 genera of Cassidini were examined, but this sample does not represent an inadequate grasp of diversity and cannot circumscribe tribal boundaries. Nevertheless, the current sampling regime allows some testing of previous hypotheses of monophyly and relationships among these three problematic tribes. Most of these genera are placed in clade 68 (Bremer support 3) and are fully resolved in subclade 70 (fig. 86). Also placed in clade 68 are two stolaine genera, Amythra and Anepsiomorpha. These placements support the previous synonymies of Riley (1986) and Borowiec (1994a) but contrast with the chaotic polyphyly of Cassidini found by Hsiao and Windsor (1999).

Clade 68 is monophyletic and characters 90, 132, and 145 provide synapomorphies under unambiguous and slow character optimizations. Characters 59 and 201 also support this node under accelerated transformation. There are no unambiguous autapomorphies within this clade, and its subclades have reasonable support (Bremer values range from 4 to 15). Synonymies with Charidotini (Riley, 1986), Aspidimorphini (Borowiec, 1996), and Basiptini (Borowiec, 1999) have resulted in the tribe Cassidini comprising more than 1500 species.

An extensive literature has accumulated on the biology of many Cassidini species. These immatures are the most commonly represented in specimen collections. Detailed comparative study of these species and genera is feasible and should promote resolution of relationships.

Cephaloleiini Weise (9 Genera, 3 Sampled; 311 Species, 3 Sampled)

Three features have been used to distinguish Cephaloleiini: finely punctate elytra, elytra lacking costae, ridges and tubercles, and presence of the anterior pronotal seta (Sanderson, 1967). These are all homoplasies. Borowiec (1995) treated Cephaloleiini as monophyletic and sister to Callispini. Hsiao and Windsor (1995) sampled a single species of Cephaloleiini and three species of Imatidiini, which were placed within a single monophyletic clade. In their densely sampled molecular phylogeny of Cephaloleia species (98 ingroup taxa and 7 outgroups [six other cassidines and one criocerine]), McKenna and Farrell (2005) found the relationship (Cephaloleia + (Cephaloleia + Imatidiini) + ((Alurnini + Prosopodontini) + Arescini)) + Criocerinae. Although the taxon sampling at the outgroup node was not a focus of their research, the placement of imatidiines within the Cephaloleia clade is intriguing and concurs with the finding of Hsiao and Windsor (1995).

Cephaloleiini have been considered synonymous with Imatiidini (Borowiec, 1995). The known cephaloleine larvae are highly flattened and, like arescines, live uniquely in rolled leaves (“rolled-leaf cassidines”). The only two known imatidiine larvae, Imatidium neivai Bondar (Bondar, 19409293) and Imatidum sp. (Chaboo, unpubl. data) also have similar morphology.

My sampling of Cephaloleiini, Arescini, and Imatidiini includes one arescine species, three species in three cephaloleiine genera that appear in three different locations on the consensus, and the imatidiines Calliaspis and Imatidium that are resolved as (plesiomorphic cassidines + Calliaspis + (Demotispa + (Imatidium + derived cassidines)))).

Character transitions on my topology suggest a correlation between extreme flattened larval morphology and rolled-leaf feeding. Rolled-leaf feeding occurs in several unrelated clades of cassidines, suggesting convergent evolution. The correlation of extremely flattened larval and pupal forms with rolled-leaf feeding is not a surprise. Such morphology probably aids or is constrained by movement in the restricted spaces between layers of a rolled leaf.

Chalepini Weise (25 Genera, 6 Sampled; 342 Species, 6 Sampled)

Chalepini ranks as the fifth largest cassidine tribe (fig. 2). It was circumscribed by Staines (2002)990 by adults with a long clypeus and reduced labrum, a tubercle and trichobothrium at the anterolateral pronotal angles, and by denticulate lateral elytral margins. Hsiao and Windsor (1999) sampled the chalepines Anisostena and Sumitrosis, and resolved these in a monophyletic clade and sister to a monophyletic Uroplatini (Microrhopala and Probaenia were sampled). Under the wider generic sample considered here, Chalepini was recovered as paraphyletic (clade 26); Baliosus and Anoplites grouped with the uroplatines, Microrhopala and Octotoma (clades 27 and 28), and Odontota, Xenochalepus and Chalepus grouped with another uroplatine, Uroplata (clades 32–34). This finding suggests that both chalepine and uroplatine taxa may share a close relationship, as proposed earlier by Hsiao and Windsor (1999). However, neither tribe forms natural groups here.

Biological data include some 17 plant family records (table 3) and immature stage reports in 9 genera, Anisostena, Baliosus, Chalepus, Craspedonispa, Odontota, Sternostena, Stethispa, Sumitrosis, and Xenochalepus (appendix 2). However, these offer an inadequate account of the biology of this moderately speciose clade.

Charidotini Hincks (6 Genera, 2 Sampled; 312 Species, 2 Sampled)

Charidotine adults are medium sized and rounded, and have metallic golden colors. Their historical treatment has been indicated in the above discussion of Cassidini. Riley's (1986) synonymy of charidotines with Cassidini, on the basis of pectinate claws in adults of both tribes, has been accepted (Borowiec, 1994a; Hsiao and Windsor, 1999). The single charidotine sampled by Hsiao and Windsor (1999) was resolved among genera of Cassidini. Metrionella and Psalidonota are sampled here and are resolved separately within clade 68, subclade 69 with (Psalidonota + Laccoptera) (Aspidimorphini) (Bremer support 4), and in subclade 73 with (Metrionella + (Amythra + Anepsimorpha)) (Stolaini) (Bremer support 5). These relationships are not robustly supported, and the lack of resolution among the 10 tribes situated in clade 54 is a complex problem for further investigation.

Despite their moderate diversity, charidotines are known from a few families of host plants (e.g., compare to Chalepini) (table 3). Immatures are superficially known in three genera (appendix 2).

Coelaenomenoderini Weise (8 Genera, 2 Sampled; 73 Species, 2 Sampled)

Würmli (1975a) diagnosed the tribe on the basis of large domed eyes and an elongate body with widened posterior in adults. The eye feature was not scored here due to the wide variation in eye shape. The other two characters, body elongation and posterior wideness, are homoplasious in the context of Cassidinae.

Pharangispa was filtered due to high levels of missing data. Cyperispa, the only coelaenomenoderine genus included in the analysis, is united with Gonophorini by four homoplasious characters 29, 144, 147, and 171 (clade 22, fig. 82). Further investigations must be directed at capturing more of coelaenomenoderine diversity (as suggested by recognition of eight genera) and must encompass more biological information. Hosts are known in four plant families (table 3), and immature stages are known for species of Balyana, Cyperispa, Coelaenomenodera, and Enischnispa (appendix 2).

Cryptonychini Weise (24 Genera, 3 Sampled; 136 Species, 3 Sampled)

Uhmann (1958) diagnosed this large Old World tribe by the rostrum shape; however, Würmli (1975) indicated that tribal circumscription was problematic due to sexual dimorphism and wide variability in morphology. Cryptonychini was treated as related to Eurispini and Callohispini (Seeno and Wilcox, 1982). Mariau (2004) treated Callispa (Callispini) as a member of Cryptonychini. The two cryptonychines sampled here are resolved in clade 20 (Callistola + Botryonopa) and clade 41 (Ceratispa + Palmispa). The “rostrum” is morphologically an extension of the vertex of the head beyond the eye margin; it is sexually dimorphic but the functions are unknown. This feature is treated as character 33 and originates four times on the consensus topology. Lack of monophyly is not unexpected given previous doubts about the tribal circumscription. The 24 cryptonychine genera encompass about 134 species (compare to the six charidotine genera encompassing more than 300 species) perhaps reflecting great intratribal morphological diversity or exuberant taxonomy.

Data on immature stages are known in 10 genera: Aulostyrax (Gressitt, 1960a), Brontispa (Böving and Craighead, 1931; Maulik, 1938; Lange, 1950; Kalshoven, 1951; Gressitt, 1955, 1960a, 1960b, 1963b; Gressitt and Kimoto, 1963; Cox, 1988; Kimoto and Takizawa, 1994, 1997); Caledonispa (Gressitt, 1960b, 1963); Callistola (Gressitt, 1960a, 1963b; Gressitt and Kimoto, 1963a); Cryptonychus (Maulik, 1932); Enischnispa (Gressitt, 1963b; Gressitt and Kimoto, 1963); Isopedispa (Gressitt, 1960b); Octodonta (Gressitt, 1960a, 1963b; Gressitt and Kimoto, 1963; Kogan and Kogan, 1979); Oxycephala (Gressitt, 1960a, 1955); and Stephanispa Gressitt (Gressitt, 1960b). Most of these were collected, reared, and described by J.L. Gressitt, a remarkable feat. Mariau (2004) summarized cryptonychine biology indicating that larvae are mostly sheath feeders between unopened spears of palms. Some also mine stems and leaves (Maulik, 1932; Lepesme, 1947). Eleven plant families have been recorded as hosts (table 3).

Delocraniini Hincks (1 Genus, 3 Species; 1 Sampled, All Examined)

Delocraniini is distinct for morphology, membership as a “transitional” group (Maulik, 1919), and is allied with Notosacanthini and Spilophorini (Hincks, 1952; Borowiec, 1995). Several autapomorphies support Delocraniini: a longitudinally grooved fronto-clypeus, an arrangement of cells on the elytral explanate margin, and the occurrence of single trichobothria at regular intervals along the elytral edge and that correspond with carina between cells on the explanate margin.

Hsiao and Windsor (1999) did not sample Delocraniini. In the present study, the three known species and two putative new species (Chaboo, unpubl. data) are scored the same. Delocraniine autapomorphies were pruned from the dataset prior to analysis. The single terminal on cladograms represents the entire tribe. Delocraniini falls out in clade 49 (Bremer support 3) at the root of the polytomy Notosacanthini + (Delocraniini + Hemisphaerotini) + ((Spilophorini + Cassidinae s.str.)). Maulik's (1919) hypothesis of Delocraniini positioned between Hispinae s.str. and Cassidinae s.str. is supported here. Immature stages of Delocrania cossyphoides are known (Bondar, 19409293; Zenner, 1968; Genty et al., 1978; Buzzi, 1988; Mariau, 2004) and a generic revision with its descriptions is under way (Chaboo, unpubl. data).

Dorynotini Hincks (6 Genera, 4 Sampled; 47 Species, 4 Sampled)

The relatively small Neotropical tribe was defined on the basis of parallel claws (Hincks, 1952), but this feature is widespread across Cassidinae. Borowiec (1995) resolved Dorynotini as a member of a derived polytomy with Cassidini and Ischyrosonychini. The single species considered by Hsiao and Windsor (1999) was placed among genera of Cassidini. The four genera sampled here are dispersed among various cassidines, and tribal monophyly is not recovered. Two genera, Dorynota and Paratrikona, are resolved here as sister taxa (Clade 58). The other two genera, Polychalca and Oxynodera, form clade 66 with Canistra (Omocerini) (fig. 89).

Only Bignoniaceae hosts are known (Jolivet and Hawkewsood, 1995; Borowiec and Świętojańska, 2005). Exophagous immatures have been described in Dorynota and Paranota (appendix 2) and are collected and await description in Paratrikona (Chaboo, unpubl. data).

Epistictini Hincks (3 Genera, 11 Species; 1 Sampled)

Hincks (1952) diagnosed Epistictini by its free mouthparts and short clypeus but these are neither unique to the tribe nor do they represent a unique combination. Monophyly was not tested here but the single exemplar is recovered among derived cassidines in clade 64, Epistictina + Basiprionota, supporting previous hypotheses (Hincks, 1952; Zaitzev and Medvedev, 1982; Borowiec, 1995). This finding is discussed above under Basiprionotini. The placement of (Epistictina + Basiprionota) as sister to the omocerine clade 65 (Prenea + (Polychalca + (Canistra + Oxynodera))) is identical in all MPTs and is supported by the sinuation of the elytral basal margin (fig. 89). Borowiec's (1994) account of afrotropical epistictines is a starting point for further research. Few immatures are known for these Asian species: in Epistictia (Gressitt, 1952; Takizawa, 1983) and in Epistictina (Zaitsev and Medvedev, 1982; Chen et al., 1986; Ghate and Rane, 2002). Their hosts are in Bignoniaceae and Meliaceae.

Eugenysini Hincks (3 Genera, All Sampled; 35 Species, 4 Sampled)

Eugenysines are considered a natural group within Cassidini based on the apical expansion of tarsomere V (which hides the base of the claws) (Hincks, 1952). Eugenysines and stolaines have been regarded as each other's closest relatives on the basis of a single ventral process of the claw (Hincks, 1952). Borowiec (1995) amalgamated Eugenysini, Omocerini and Stolaini under the group name Mesomphaliini and recovered this terminal in the relationship (Basiprionotini + (Omocerini + (Mesomphaliini + (Dorynotini + Cassidini + Ischyrosonychini)))). Hsiao and Windsor (1999) sampled a single species that was recovered in the most derived cassidine clade as (Eugenysa + (Stolas + Echoma)).

In the present study, multiple species in the three eugenysine genera were examined. Monophyly is supported by two synapomorphies: the apical expansion of tarsomere V (char. 185) (which hides the base of the claws) (supporting Hinck's [1952] character hypothesis); and the very long, tightly coiled ejaculatory duct (char. 206) (defined in Chaboo, 2002189). Another diagnostic character, the single basal tooth of the claws, also appears in all Stolaini and some Cassidini. Generic relationships are resolved in clade 81 (fig. 90), (Agenysa + (Miocalaspis + Eugenysa), with the sister group relationship of Miocalaspis + Eugenysa supported by a three-chambered spermathecal receptacle (described in Chaboo, 2002189).

The crown placement of Eugenysini is germane to discussions of maternal care in cassidines. The two sampled Eugenysa species exhibit maternal care (Windsor and Choe, 1994; Chaboo, 2002189; table 6). These relationships suggest that maternal care originated once within the tribe. Eugenysines are so little known that new records of care should be expected.

Eurispini Weise (3 Genera, 14 Species; 1 Sampled)

As presently circumscribed, this Indo-Australian tribe comprises the genera Eurispa, Leucispa, and Squamispa (Würmli, 1975a; Seeno and Wilcox, 1982). The tribe was previously allied with Leptispini and Cryptonychini (Weise, 19111153; Würmli, 1975a; Seeno and Wilcox, 1982). Samuelson (1968) suggested that Aproidini and Anisoderini were more closely related, especially because of the granular eye facets. Neither Borowiec (1995) nor Hsiao and Windsor (1999) investigated this tribe. A single exemplar, Eurispa vittata, is scored here and is resolved in clade 36 as (Eurispini + (Aproidini + Exothispini)) (fig. 84; Bremer support 1) and is supported by three synapomorphies, characters 29, 32 and 83 under unambiguous optimization. A sinuous spermathecal duct (char. 203 state 2) may be an autapomorphy for Eurispa. Results here appear to support Samuelson's (1968) hypothesis of eurispine affiliations with Aproidini.

Adult eurispines have a very distinct narrowly elongate body form, cylindrical pronotum with posterior sensory tufts, relatively short legs, and body with deep punctures that contain erect scales (Würmli, 1975a). The body form and punctures with scales appear to be autapomorphies that will support tribal monophly when more eurispines are sampled. Adult eurispines are free-living while larvae are sheath feeders (e.g., of the Cyperaceae Gahnia). Both adults and larvae can overwinter in the tight basal axils of their hosts (Chaboo, personal obs.).

Exothispini Weise (1 Genus, 1 Species; Sampled)

Exothispini was previously considered as related to Callohispini and Coelaenomenoderini (Würmli, 1975a), but this is not supported here with its placement in clade 37, Aproida + Exothispa (Bremer support 1). This sister group relationship is supported by characters 26, 47, 51 and 101 under unambiguous (clade 37, fig. 84) and slow optimizations (appendix 5), and additionally by character 136 under fast optimization (appendix 5). Tarsomere V has a ventral projection positioned between the claws (char. 182, state 0) (Würmli, 1975a), but this appears to be homoplastic. The biology of Exothispa reimeri Kolbe is unknown and this impedes a decisive resolution of its position.

Goniocheniini Hincks (5 Genera, 13 Species; 1 Sampled)

This Neotropical tribe was defined on the basis of changes in antennal pubescence (Spaeth, 1942; Viana, 1964b), partial coverage of the head by the pronotum, and prosternal process shape (Hincks, 1952). It was not included in Borowiec (1995) and was represented by a single species in Hsiao and Windsor (1999), where it was resolved as sister to some stolaine genera. This tribal concept and its placement were not tested here since the single exemplar of Batonota was pruned from the final analysis due to ambiguity.

The sole detailed description of immature stages of Goniocheniini is that of a fifth instar of Chlamydocassis cribripennis (Boheman) (Buzzi, 1988; Świętojańska et al., 2005). Fiebrig (1910) provided limited description of immatures of two other Chlamydocassis species. Immatures of a Polychalma species have been collected and will be described (Riley, unpubl. data). Adults and larvae are exophagous leaf feeders of hosts in Ehretiaceae, Lamiaceae, and Sterculiaceae (table 3). Larvae carry a large shield mass with exuviae deeply embedded within the fecal matrix.

Gonophorini Weise (10 Genera, 290 Species; 1 Sampled)

Würmli (1975) circumscribed this Old World tribe by a suite of features; however, he acknowledged that this was a difficult group to define. Monophyly is still very much questionable. A single genus, Klitispa, was sampled here and is resolved in clade 22 (fig. 82; Bremer support 4) with the relationship (Coelaenomenoderini + Gonophorini). This sister relationship is moderately supported by secondary loss of the midcranial suture, presence of the parascutellary striole, presence of fenestrate elytral punctures, and expanded tarsomere I. Gonophorine larvae are leaf miners of Araceae, Arecaceae, Costaceae, Marantaceae, Musaceae, Orchidaceae, Pandanaceae, Poaceae, and Zingiberaceae (Kalshoven, 1951; Gressitt, 1963415; Jolivet and Hawkeswood, 1995).

Hemisphaerotini Hincks (2 Genera, All Sampled; 42 Species, 2 Sampled)

Borowiec (1995) located this tribe as derived within Cassidinae and sister to Hispinae s.str. Hsiao and Windsor (1999) sampled a single species that was united as sister to Prosopodontini. In the present study, Hemisphaerotini is recovered by all analyses with unambiguous support (clade 50, fig. 85) based on the membranous egg mass covering (autapomorphy), larvae with four caudal processes (autapomorphy), basket-like (or bird's nest) exuvio-fecal shield architecture (autapomorphy), immobile shield (autapomorphy), elongate prothoracic spiracles (autapomorphy), adults with broad angular prosternal processes, and a scutellum that overlaps the pronotal posterior angle (autapomorphy). Pupal caudal processes and retention of the exuvio-fecal shield is shared with other derived cassidines. Hemisphaerotine biology and aspects of immature morphology and shields were recently reviewed (Chaboo and Nguyen, 2004).

Hispini Weise (23 Genera, 4 Sampled; 611 Species, 4 Sampled)

This is the second largest tribe of Cassidinae, accounting for 10% of cassidine diversity (fig. 2). Members are found worldwide and adults typically have the dorsum spinose. This spininess has been considerably valued, so much so that Würmli (1975a) spent little time covering the morphology since the tribe appeared to be so well supported. The single exemplar sampled by Hsiao and Windsor (1999) did not test tribal monophyly, but it was placed in the relationship (Alurnini + (Hispini + Cryptonychini)). Four genera are sampled here and are recovered as the resolved monophyletic clade 23 (fig. 82) with subclades 24 and 25, in the relationship (Trichispa + (Dactylispa + (Asamangulia + Dorcathispa))). Four synapomorphies (i.e., chars. 28, 70, 113 and 134; Bremer support 6) support tribal monophyly. The spinose lateral elytral edges (char. 134) are an exclusive feature of Hispini, and this supports Würmli's (1975a) view.

The function of dorsal spininess has been historically attributed as a defensive mechanism. Although untested, this is a reasonable hypothesis, as adults are conspicuous on host plants and are slow flyers. The well-circumscribed tribal boundaries of Hispini provide a firm context for investigating its biology and systematics.

Given their rank as the second most speciose clade in Cassidinae, it is not a surprise that Hispini use the highest number of hosts—30 families of monocots and eudicots (table 3). What is surprising is how little known are immatures stages, which are unrecorded for Acymenychus Weise, Callanispa Uhmann, Cassidispa Gestro, Chrysispa Weise, Dorcathispa Weise, Jambhala Wurmli, Phidodontina Uhmann, Pleurispa Weise, Polyconia Weise, Pseudispella Kraatz, Rhadinosa Weise, Sinispa Uhmann, Thomispa Wurmli, Thoracispa Chapuis, Trichispa Chapuis, and Unguispa Uhmann.

Hispoleptini Uhmann (1 Genus, 4 Species; Not Sampled)

This small South American tribe has been diagnosed by a combination of adult features that appear throughout Cassidinae. Its placement near Sceloenoplini and Chalepini (Seeno and Wilcox, 1982) remains untested because it has not been represented in any systematic studies to date. Individuals are rarely collected, and few specimens exist in collections.

Mariau (2004) summarized what little is known of hispoleptine biology and indicated that some species have four instars and a prolonged developmental period from egg to adult. Immatures are undescribed.

Imatidiini Hincks (8 Genera, 2 Sampled; 78 Species, 2 Sampled)

The monophyly, relationships and inclusions of Imatidiini have always been ambiguous, often discussed but with little resolution. As early as 1910, Weise suggested that it was allied with Cephaloleiini and should be classified in Hispinae s.str. because of life-history features. Bondar (1940a, 1940b) also considered Imatidiini and Cephaloleiini closely related and transferred Imatidiini to Hispinae s.str., a move supported by hispine researchers (Monrós and Viana, 1947, 1951; Papp, 1953). However, cassidine researchers usually placed Imatidiini in Cassidinae s.str. (Spaeth, 1914, 1938; Blackwelder, 1946; Aslam, 1965; Windsor et al., 1992) where it has been allied with Hemisphaerotini and Spilophorini (Seeno and Wilcox, 1982).

Borowiec's (1995) synonymy of Imatidiini under the name Cephaloleiini has been supported (Borowiec, 1999; Hsiao and Windsor, 1999). However, Borowiec (2000) later considered Imatidiini valid, including Demotispa Baly, one of nine cephaloleine genera. Staines (2002)990 again synonymized Imatidiini with Cephaloleiini.

Bondar's (1940a, 1940b) illustrations of the larva of Imatidium neivei Bondar are poor, but they clearly show an extremely flattened form similar to described cephaloleine, arescine and callispine larvae. Immatures of another Imatidium species studied here were similarly flattened. Imatidiine larval biology is of extreme interest since some species are known to be open external leaf feeders (Bondar, 1940a, 1940b, 1941, 1942, 1943). At least two species have adults that construct leaf shelters in which they feed cryptically (Gilbert et al., 2001).

Hsiao and Windsor (1999) found Cephaloleiini (including Imatidiini) as the sister clade to all other Cassidinae. Wilf et al. (2000) inserted a terminal to represent Arescini, although it was not originally analyzed in Hsiao and Windsor (1999). Duckett et al. (2004) sampled a single Imatidium exemplar, which was resolved in the relationship Imatidium + (Cryptocephalinae + other cassidines). Farrell and Sequeira (2004) also found Imatidium falling out separately from other cassidines. The results of Duckett et al. (2004) and Farrell and Sequeira (2004) are real surprises and may be due to the different evolutionary histories of the sampled genes.

Remarkably, although Imatidiini appears to have enjoyed much attention relative to other cassidine tribes, there has been little discussion of actual morphological data to support various opinions and hypotheses of imatidiine monophyly, affiliations, or various synonymies. Synonymy with Cephaloleiini may be necessary because of the similar scale-like larvae (L. Borowiec, personal commun.) and similar adult morphology (C. Staines, personal commun.), but detailed comparative morphology is lacking.

In the present study, Imatidiini is represented by two generic exemplars, Calliaspis and Imatidium. These are recovered closely together at nodes 45 and 47, in a completely resolved area of topologies but with poor support (Bremer support 1). Imatidiine and cephaloleiine monophyly and sister relationships are not supported. Arescini is recovered as as sister to remaining cassidines, in constrast to Wilf et al. (2000). However no one (Borowiec, 1995; Hsiao and Windsor, 1995; McKenna and Farrell, 2005; present study) has examined imatidiine relationships adequately (i.e., taxon sampling).

Ischyrosonychini Hincks (2 Genera, 18 Species; 1 Sampled)

This New World tribe was diagnosed by the free mouthparts and short clypeus of the adult head. It was allied with Basiprionotini based on the longitudinally striate antennomeres (Hincks, 1952). In catalog classifications, it appears close to Asterizini and Stolaini (Seeno and Wilcox, 1982). These three tribes have been synonymized under two different group names, Physonotini by Boroweic (1999) and Ischyrosonychini by Riley et al. (2002); however, no unambiguous diagnostic characters have been proposed for this group concept. Hsiao and Windsor (1999) found support for the sister relationship Ischyrosonychini + Physonotini, nested among derived cassidines. They did not sample Asterizini. My sample of Asteriza (Asterizini), Ischyrosonyx (Ischyrosonchyini), and Eurypepla and Physonota (Physonotini) tests these previous hypotheses. No support is found for a close relationship between Ischyrosonychini with Asterizini or Physonotini. Ischyrosonychini is resolved at node 53 as sister to all other derived cassidines (Bremer support 14). Synapomorphies supporting node 54 include larvae with anus in a terminal position (char. 12, state 0), partial coverage of the head by the pronotum (an autapomorphy) (char. 26, state 0), adult head with midfrontal sulcus extended to frons (char. 31, state 0), mesosternum notched and receiving the prosternal process (char. 118, state 0), the elytral basal margin crenulate (an autapomorphy) (char. 139, state 1), elytral epipleura with an internal longitudinal carina (char. 154, state 0), and spermathecal gland and duct entries separated but closely positioned (char. 197, state 1). The quality and quantity of characters and the robustness of support for clade 53 are powerful arguments favoring this grouping of derived cassidines.

Leptispini Weise (1 Genus, 67 Species; 1 Sampled)

This tribe was previously defined by its triangular hairy clypeus (Würmli, 1975a). The single exemplar here is recovered between Stenispa (Cephaloleiini) and Ceratispa (Cryptonychini) based on characters 67, 98, 137, and 173 (clade 38). It is distinguished here by the unique cryptic feeding of larvae within a leaf shelter. Stenispa larvae are exophagous (Riley and Enns, 1979; Ford and Cavey, 1985) whereas Ceratispa larvae are miners (Gressitt, 1960a, 1963).

The unusual leaf-shelter building behavior of Leptispa larvae (Maulik, 1919) is complex and needs further study. The pubescent venter of adults described by Maulik (1919) and their ability to withstand periods of submersion suggest additional characters. The single long caudal process of larvae is a source of missing information since the available simple line illustrations do not show sclerite attachment. Homologizing this process with other caudal processes of cassidine immatures is difficult without examination of specimens.

Notosacanthini Hincks (2 Genera, 254 Species; 1 Sampled)

This well-defined tribe has been problematic for separating classical cassidines and hispines. It has historically been treated within Cassidinae s.str. (Zaitzev and Medvedev, 1982; Borowiec, 1999; Borowiec and Świętojańska, 2004) or within Hispinae s.str. When classified in Cassidinae s.str., Notosacanthini has been variously allied with Epistictini and Basiprionotini (Zaitzev and Medvedev, 1982), with Aspidimorpha (Aspidimorphini) (Hawkeswood, 1989; Monteith, 1991), and even with Oncocephalini (Medvedev and Eroshkina, 1988). Borowiec (1995) resolved Notosacanthini as sister to one cassidine subclade where other members included so-called transitional tribes, Delocraniini, Spilophorini, Hemisphaerotini, Cephaloleiini, and Callispini.

Notosacanthine monophyly is strongly supported by several autapomorphies that were removed prior to analysis. Putative affiliations with Aspidimorphini, Epistictini and Basiprionotini are rejected here. Oncocephalini was not included here so this hypothesis remains to be tested. Notosacanthines are firmly ensconced in an intermediate area between Cassidinae s.str. and Hispinae s.str., along with Delocraniini, Hemisphaerotini, Imatidiini and Spilophorini (clade 47). These tribes all have exophagous larvae; however, known Notosacantha larvae are leaf miners (Medvedev and Eroshkina, 1988; Reid, 1995; H. Ghate, personal commun.; M. de Baar and G. Monteith, personal commun; Chaboo, personal obs.).

Oediopalpini Monrós and Viana (1 Genus, 37 Species; 1 Sampled)

This monogeneric tribe was previously allied with Cephaloleiini (Seeno and Wilcox, 1982). The single species sampled by Hsiao and Windsor (1999) was resolved as sister to Spilophorini. Staines (2002)990 synonymized Oediopalpini with Spilophorini. Wilf et al. (2000) revised the topology of Hsiao and Windsor (1999), relocating Oediopalpa as sister to Delocraniini, although no additional analyses were performed. The present study recognizes a sister relationship between oediopalpines and Cephaloleia (clade 15).

The placement of Oediopalpa suggests that cassidine larval shields and the attendant morphological apparatus (telescoped anus, urogomphi) originated twice, with Oediopalpa and with Cassidinae s.str.

Omocerini Hincks (7 Genera, 4 Sampled; 139 Species, 4 Sampled)

Goniocheniini and Omocerini were separated on the basis of an antennal feature, the demarcation of the distal section of the antenna beginning with antennomeres IV and V in goniocheniines and with VI and VII in omocerines (Spaeth, 1942; Viana, 1964a). Hincks (1952) noted that a change in pubescence demarcated distal and proximal antennal sections. Viana (1968) viewed this as a weak character since the pubescence feature is not easily defined. Seeno and Wilcox (1982) treated omocerines with Spilophorini and Goniocheniini. Borowiec (1995) synonymized Omocerini with Eugenysini and Stolaini under Mesomphaliini on the basis of unclear morphological distinctions. Two genera, Omocera and Discomorpha, were sampled by Hsiao and Windsor (1999) and were recovered as sister taxa, and sister to derived cassidines. Four genera, Canistra, Oxynodera, Prenea, and Polychalca, including many species, were sampled in the present study and these are recovered in the monophyletic clade 65 (fig. 89) based on three synapomorphies: epipleural brace and internal carina connected (char. 155), straight or slightly sinuate apical margin of tarsomere III (char. 172), and duct and gland insertions located on an expanded section of the proximal spermathecal duct (an autapomorphy) (char. 199). However, the original character diagnosis based on distributions of proximal and distal antennomeres fails when compared among a wider sample of cassidines. The identification of a genitalic novelty is not surprising, as genitalia were ignored in cassidine systematics until recently.

Oncocephalini Weise (3 Genera, 72 Species)

Oncocephalines were previously allied with Gonophorini and Hispini (Seeno and Wilcox, 1982). This Old World tribe has not been represented in any systematic studies to date, and little is known about their biology and morphology. A few species are known to be leaf miners of Convolvulaceae, Dioscoraceae, Orchidaceae, Poaceae, and Zingiberaceae (table 3).

Physonotini Hincks (6 Genera, 2 Sampled; 46 Species, 2 Sampled)

Physonotini was diagnosed by narrowed elytral lateral margins, thickened elytral margins, a broad prosternum that is expanded posteriorly, and dorsum lacking opalescence (Hincks, 1952). Borowiec (1995) synonymized this tribe with Ischyrosonychini and Asterizini because of ambiguous boundaries, and he resolved this terminal in the subclade of derived cassidines as (Basiprionotini + (Omocerini + (Mesomphaliini) + (Dorynotini + Cassidini + Ischyrosonychini)))). Head visibility separated (Omocerini + derived cassidines) from Basiprionotini. Hsiao and Windsor (1999) sampled a single species of Physonota that was resolved as Physonotini + Ischyrosonychini. Borowiec (1999) synonymized these two tribes plus Asterizini, under the name Physonotini; however, Riley et al. (2002) called the same group Ischyrosonychini because it is the oldest available name.

Physonotine systematics has been touched on in the above treatments of Asterizini and Ischyrosonychini. Six physonotine species in four genera were examined here, and characters were scored for Eurypepla and Physonota. These are resolved separately in clade 54 with Eurypepla in clade 57 as sister to Dorynotini (fig. 86). Ischyrosonyx (Ischrysonychini) and Asteriza (Asterizini) also appear separated, with the former at clade 54 as sister to derived cassidines and the latter ensconced with aspidimorphine genera. The general lack of resolution at node 54 restricts discussion of these tribal relationships.

Physonotines have several unusual features: gregarious larvae that range widely in shield behavior (naked, wet, or hard shields) (Chaboo and Gómez, unpubl. data); petallophagous larvae (Chaboo, 2004); a fixed quiescent fifth instar (pre-pupa) (Gómez, 2004), and adult color polymorphism in Physonota (Caulfield, 1887; Sanderson, 1948; Kirk, 1971; Britten et al., 2003) and Eurypepla (Chaboo, 2004).

Petalophagy by Eurypepla larvae (Chaboo, 2004) is a flexible behavior, in contrast to that reported in two stolaine Echoma species (Windsor et al., 1995). This rare feeding pattern has two independent origins among crown cassidines.

Promecothecinini Weise (2 Genera, 34 Species; 1 Sampled)

Würmli (1975) circumscribed this small Indo-Australian tribe on the basis of a short adult head retracted into the prothorax and a long, toothed hindfemur. The head shortening and retraction may have appeared distinct when considered only among Hispinae s.str. but these appear to be common across Cassidinae. Promecothecinini were previously allied with Colaenomenoderini and Gonophorini (Seeno and Wilcox, 1982). The single sampled species is recovered as sister to Anisoderini (fig. 81) based on characters on 107, 170, 172, and 193 under unambiguous and slow optimizations, and 170, 172 and 193 under fast optimizations. This sister group relationship is a novel hypothesis. Promecothecines are found in Asia, Australia, and Madagascar where they feed on Arecaceae, Flagellariaceae, Heliconiaceae, Pandanaceae, Poaceae, Musaceae, and Zingiberaceae (Jolivet and Hawkeswood, 1995; Staines, 2004; Mariau, 2004).

Prosopodontini Weise (1 Genus, 26 Species; 1 Sampled)

Staines (2002)990 diagnosed Prosopodontini by the small head, clypeus extended to antennal bases, setae on posterior pronotal angles, scutellum shape, and elytra with 10 puncture rows and a parascutellary striole. None of these features is unique to this tribe. The single species sampled by Hsiao and Windsor (1999) was recovered as sister to Hemisphaerotini. In contrast, the single species sampled by McKenna and Farrell (2005) was resolved as sister to Alurnus. Neither hypothesis is supported here.

Prosopodonta was recovered here in a weakly supported basal branch (clade 10, Bremer support 3; fig. 81) with the relationship (Prosopodontini + (Promecothecini + Anisoderini)). Support for this group included discrete clypeus (char. 66) under unambiguous and fast optimizations, distal medial emargination of the labrum (char. 78), shovel-shaped terminal segments of larvae (char. 9), and spermathecal gland separated but closely situated, (under slow optimization; char. 197). All of these are homoplasious.

Prosopodonta larve have been reported as miners (Jolivet and Hawkeswood, 1995) and external leaf feeders (McCoy, 1984, 1985) of Neotropical Heliconia (Jolivet and Hawkeswood, 1995) and Arecaceae (Staines, 2004). There is some ambiguity about larval identification (Staines, 2004); nevertheless, if Prosopodonta species do range from mining to exophagy, this diversity has implications for understanding transitions in larval feeding and microhabitat.

Sceloenoplini Uhmann (5 Genera, 2 Sampled; 219 Species, 2 Sampled)

Seeno and Wilcox (1982) previously placed this moderately large Neotropical tribe near Prosopodontini and Hispoleptini. Staines (2002)990 diagnosed Sceloenoplini by proximal and distal sections of antenna distinguished at antennomeres IV, anterior pronotal seta, presence of parascutellary stria, and number of elytral puncture rows (Staines, 2002990). All of these characters are plesiomorphic when compared across Cassidinae. Sceloenoplini was not included in previous cladistic analyses and is investigated here by sampling two genera, Cephalodonta and Sceloenopla. Sceloenopline monophyly is not supported; instead, Sceloenopla is resolved basally in clade 17 as sister to seven other tribes (fig. 82), and Cephalodonta is placed among chalepines in clade 30 (fig. 83). The lack of monophyly is not unexpected given historically problematic tribal circumscription.

Members have been recorded on six plant families (table 3), and some species have larvae that are scrapers (Mantovani et al., 2005)

Spilophorini Uhmann (2 Genera, Both Sampled; 30 Species, 2 Sampled)

Spilophorini was defined on the basis of presence of the anterior and posterior trichobothria and carinate labrum (Hincks, 1952). These are symplesiomorphies in the context of all Cassidinae. Borowiec (1995) presumed monophyly of Spilophorini and recovered it in a clade with Delocraniini, Hemisphaerotini, Notosacanthini, Cephaloleiini, and Callispini. Hsiao and Windsor (1999) sampled the two recognized genera and supported tribal monophyly and a sister group relationship with Oediopalpini. Wilf et al. (2000) used the results of Hsiao and Windsor (1999); however, Spilophorini was omitted without explanation on their figure 2. The inclusion of this taxon would have conflicted with their evolutionary hypothesis.

The present study recovered a reasonably supported (Bremer support 9) spilophorine branch (clade 52), Calyptocephala + Spilophora, within the transitional zone between Cassidinae s.str. and Hispinae s.str., close to Delocraniini, Hemisphaerotini, and Notosacanthini. Under unambiguous character optimization (fig. 85), support comprises the following five homoplasious states: discontinuous dorsal body profile (char. 22), posterior pronotal trichobothria present (char. 101), posterior margin of mesoscutellum truncate (char. 116), elytral basal margin wider than pronotal posterior margin (char. 136), and CuA cell 1 open in hindwings (char. 157). Under slow optimization, this character list included the presence of antennal ventro-marginal grooves (char. 59). Under fast optimization, characters 22, 72, 77, 82, 99, 101, 116, 136, 157 and 200 provided additional support.

Spilophorines form a small Neotropical tribe. The known larvae (Moura, 1985; Buzzi and Miyazaki, 1992; Windsor et al., 1992) make an exuviae-only shield, with the exuviae longitudinally compressed and the thick paired caudal processes prominently displaced. This design is very similar to the shield constructed by the larva of Oediopalpa guerini (Chaboo, unpubl. data). These particular shields are different from other exuviae-only shields in their pattern of exuvial compression.

Stolaini Hincks (16 Genera, All Sampled; 528 Species, 20 Sampled)

This Neotropical tribe accounts for 9% of cassidine diversity (fig. 2) and was originally diagnosed on the basis of a single basal tooth on the claws and confused elytral punctation (Hincks, 1952). This tooth feature also occurs in Eugenysini and in some cassidine genera. The elytral character is widespread across Cassidinae. Morphology of Stolaini was previously treated in Michalski (1995).

In previous cladistic analyses, Borowiec (1995) synonymized Eugenysini, Omocerini and Stolaini under the group name Mesomphaliini and presumed this terminal monophyletic prior to analysis. Hsiao and Windsor (1999) failed to find monophyly of Stolaiini, with genera appearing among derived cassidines with the relationship—((((Stolaini + Eugenysini) + Stolaini) + (Cassidini + (Goniocheniini + Ischyrosonychini))) + Stolaini).

All 16 stolaine genera are sampled here to test tribal monophyly and the tribe was not recovered as monophyletic. Stolaine genera were grouped in a large derived polytomy along with genera of Aspidimorphini, Asterizini, Basiprionotini, Cassidini, Charidotini, Dorynotini, Epistictini, Eugenysini, Omocerini, and Physonotini. A few stolaine genera are united as sister taxa; for example, clade 75 Amythra + Anepsiomorpha; clade 56 Elytrogona + Stoiba; clade 55 Chelymorpha + Phytodectoidea; clade 61 Zatrephina + (Poecilaspis + Stolas), and finally, clade 75 is a polytomy involving the remaining genera grouped with a monophyletic Eugenysini. This finding is not a surprise given the previous ambiguous support. Clearly presence or absence of a claw tooth is a weak phylogenetic signal.

Monophyly of Echoma, Eugenysa, and Omaspides is supported, but not of Stolas. Stolas was also previously found to be paraphyletic (Hsiao and Windsor, 1999). The boundaries of Acromis (fig. 90) and Elytrogona (fig. 87) were previously redefined (Chaboo, 2000, 2001) and hold up under the present analysis. The position of Eugenysini among Stolaine genera is discussed above under Eugenysini. Maternal care in these tribes (fig. 19, table 6) is discussed below.

The sister group relationship Elytrogona + Stoiba is not surprising since this was proposed before based on their shared development of a suite of morphological modifications related to flightlessness (Chaboo, 2000). They occur only in the Greater Antillean Caribbean islands.

Stolaine larvae are typically exophagous leaf feeders, but petallophagy has been reported as obligate in two species of Echoma (Windsor et al., 1995). This rare feeding pattern has two origins, in Echoma and in Eurypepla (Physonotini) among derived cassidines.

Uroplatini Weise (32 Genera, 3 Sampled; 396 Species, 3 Sampled)

Uroplatines form the fourth largest cassidine tribe, accounting for 7% of known species diversity (fig. 2). Uroplatini and Chalepini were differentiated by fusion of their terminal antennomeres (Maulik, 1932) and a close relationship between them has been proposed in various catalogs. Riley et al. (2001) argued that the antennal feature is continuous among these genera and synonymized Uroplatini under Chalepini. Hsiao and Windsor (1999) sampled two uroplatines and these were resolved in a basal monophyletic clade along with the sampled chalepine taxa. Uroplatines and chalepines sampled here appear in the fully resolved monophyletic clade 26 that also includes the exemplar sceloenopline, Cephalodonta. This clade is recovered on all MPTs, and appears in the consensuses of both complete (fig. 77) and restricted matrix analyses (fig. 91). Two homoplasious characters support clade 26, depression of the scutellum and a truncate posterior shape of the scutellum. The placement of the uroplatines approximates to previous hypotheses, as follows (fig. 83): clade 27 as (chalepines + (Microrhopala + Octotoma). Microrhopala + Octotoma is supported by two synapomorphies: fusion of distal antennomeres and frontoclypeus distinctly protuberant from antennal bases. Characters 152 and 182 support clade 32 (chalepines + Uroplata). Uroplatine monophyly is not supported but there is clearly some relationship between Uroplatini and Chalepini.

The character, number of antennomeres, was removed in the final analyses due to scoring difficulties (missing or incomplete antennae). Determining the exact number of antennomeres is difficult because fusion of sections can be complete or partial; sutures may be developed, reduced or absent. Because this feature has been fundamental in diagnosing Uroplatini and Chalepini, it must be investigated in additional taxa and in more detail.