Review on the use of sexually dimorphic characters in the taxonomy of Diabroticites (Galerucinae, Luperini, Diabroticina)

Abstract Sexual dimorphism occurs frequently in Chrysomelidae Latreille, 1802 and is particularly variable in subfamily Galerucinae Latreille, 1802. This diversity has been early noted by authors a potential source of taxonomic characters. The section Diabroticites (Luperini Gistel, 1848) is one of the largest assemblies of chrysomelid genera with currently 823 valid species in 17 genera (12 based on dimorphic characteristics), being most diverse in the neotropical region. Apart from a revision work on the type specimens for the section, there are no general taxonomic studies for this group. The occurrence of sexually dimorphic characteristics in the section Diabroticites is revised and their practical taxonomic relevance evaluated. A total of 240 species was studied (145 species with males available), representing 15 out of the 17 genera included in Diabroticites. The analysis of characters was based on the study of specimens in south-american collections, literature and the aid of photos in online databases. Sexual dimorphism occurred in most species analyzed. Dimorphic features were divided in general (i. e., occur in higher taxa) and special characters (those that support the definition of species and genera). Special dimorphism was observed in every tagma, and most modifications occur in antennae. Characters used as diagnostic of genera often do not correspond to the modifications present in species included in them. Many modifications were considered by earlier authors as a single character, probably due to vague definitions. Most generic definitions are, therefore, inaccurate. The study of morphology and the homology assessment of characters are needed to increase understanding of the genera in Diabroticites.


Keywords
Taxonomy of Coleoptera, rootworms, review, leaf-beetles introduction Sexual dimorphism has always been a subject of great curiosity amongst naturalists. Since Darwin's suggestion of his Theory of Sexual Selection, many explanations have been proposed, and several have been successfully tested, for the existence of often peculiar modifications in males. Even though most beetles lack conspicuous sexual dimorphism (Kawano 2006), there are striking examples in all major Coleoptera taxonomic groups (Eberhard 2009). Most sexually dimorphic characters in beetles are described by strongly positive allometries (Kawano 2006), and also modifications found in antennae, tarsi, posterior legs and ventrites (Crowson 1981). Other less common characteristics include the reduction of the wings (Thayer 1992) and the presence of luminescent (Branham and Wenzel 2003) or stridulatory organs (Jansson and Selander 1977).
In Chrysomelidae Latreille, 1802, sexual dimorphism is thought to occur more frequently at the species level (Jolivet and Verma 2002). Common dimorphic features that are the body size (with females usually bigger than males) and the modification of tarsi, usually related to greater adhesion of males to the females' dorsal surface during copulation (Jolivet and Verma 2002, Hammack and French 2007, Voigt et al. 2008, Nardi et al. 2012. Within the subfamily Galerucinae Latreille, 1802, sexual dimorphism is particularly variable, as Mohamedsaid and Furth (2011) have illustrated and summarized. This diversity has been early noted by authors as a potential source of taxonomic characters. As Horn (1893) pointed out, many taxonomic issues related to this group could be resolved with the aid of "sexual peculiarities", which could be a useful guide for understanding the relationships between species. Blake (1958) also stated that the use of such characteristics could help the delimitation of genera in problematic groups such as the tribe Luperini Gistel, 1848.
The section Diabroticites Chapuis, 1875 (Luperini) is one of the largest assemblies of chrysomelid genera with over 900 recorded names in 17 genera (12 based on dimorphic characteristics), being most diverse in the neotropical region. Apart from a revision work on the type specimens for the section (Smith and Lawrence 1967), there are no general taxonomic studies for this group. The most recent catalogue mentions 793 species (Wilcox 1972), but a review of the subsequent literature reveals that the group has currently 823 valid species. Table 1 presents an overview on the current composition of Diabroticites. Prior to 1906, Diabroticites included Diabrotica Chevrolat, 1837 (almost half of the total species in the section), and three other genera, which were all monotypic: Ensiforma Jacoby, 1876, Pseudodiabrotica Jacoby, 1892, and Paratriarius Schaeffer, 1906. Barber (1947 was the first to investigate male genital characters to understand the relationships within the group and described two more genera in the section, Acalymma Barber, 1947 and Amphelasma Barber, 1947. In the subsequent years, the remaining 11 genera were erected, eight of which based on species formerly included in Diabrotica (Anisobrotica Bechyné & Bechyné, 1969, Aristobrotica Bechyné, 1956, Buckibrotica Bechyné & Bechyné, 1969, Cochabamba Bechyné, 1955, Cornubrotica Bechyné & Bechyné, 1969, Gynandrobrotica Bechyné, 1955, Synbrotica Bechyné, 1956, and Paranapiacaba Bechyné, 1958, and most supported primarily on dimorphic features (Table 1). Surprisingly enough, only one genus has a detailed description of genital characters (the most recent genus, Platybrotica Cabrera & Cabrera Walsh, 2004).
About 80% of the diabroticites species have been described prior to 1895, mainly by Joseph Sugar Baly, Charles J. Gahan and Martin Jacoby (Smith and Lawrence 1967). Most of those descriptions lack detailed morphological information, and usually depict characters relative to color pattern and, sometimes, punctuation. For many of the genera this is also true, with internal characters being almost completely ignored. General morphology has also been vaguely treated, described usually without any aid of illustrations whatsoever. Such scarcity of information and the apparent uniformity in morphology of some diabroticites resulted in a difficult taxonomic scenario.
The purpose of this study is to summarize the occurrence of sexually dimorphic characteristics in the section Diabroticites, as well as to review these characters chosen by earlier authors to support their definitions of genera and evaluate their practical taxonomic relevance. Systematic research is being conducted on Diabroticites by the author, and the first results are reported here.

Methods
The analysis of characters was based on the study of specimens, literature (original descriptions and revision works, when available) and the aid of photos in online databases. A total of 240 species was studied, representing 15 out of the 17 genera included in Diabroticites. Specimens were obtained from south-american collections listed in Table 2, always in comparison to original descriptions -and many types were available. Out of the total species available, only 145 species had males available or known. These taxa were listed in Appendix. The specimens were examined and illustrated using a Zeiss Discovery.V8 stereomicroscope with a camera lucida attached. Final art was done in Adobe Illustrator®. Photographs were taken using a Leica M205C stereomicroscope with an attached magnifying lens and Leica DFC 295 video camera. Image combination was performed with Leica Application Suite V3.6.0, and subsequent edition was done in Adobe Photoshop®.
Most taxonomic literature available for Diabroticites was reviewed. The original descriptions of monotypic genera Palmaria Bechyné, 1956, and Pseudodiabrotica, known only for their type-specimens, which could not be loaned, were the only source of characters for comparison. Revisionary works were available only for genera Acalymma (in  (Munroe and Smith 1980, Cabrera 1999, Cabrera and Durante 2003, Diabrotica (in part) , Christensen 1943, Krysan and Smith 1987, Cabrera 2000a, Cabrera 2000b and Synbrotica (in part) (Cabrera 1995). Due to the peculiarity of genus Isotes Weise, 1922, which was described based on a single species later found to be a senior synonym of the type-species of genus Synbrotica (at that time with over 100 species), the original description of the latter was also included in the analysis for comparative purposes. As a reference to their original descriptions, both names will be used interchangeably throughout the text, even though Isotes is the current valid name.
Characters mentioned in original descriptions and other taxonomic works, when available, were compiled and later compared to specimens. Those characters were then redefined, in order to fulfill uniform homology criteria. The broader studies of Mohamedsaid and Furth (2011) and Mohamedsaid (2004) were used for character comparison with other taxonomic groups.
Also, some taxa that had not enough specimens available in south-american collections were studied by the analysis of photos of type specimens available in the Museum of Comparative Zoology online Type Database, Harvard University, Cambridge, USA.

Results and discussion
Out of the 17 genera that comprise section Diabroticites, 12 have sexually dimorphic features as diagnostic characters mentioned in original descriptions by nine different authors (15 papers) in a period of more than 150 years (Table 1). Most genera were established based on a single or on few species, further taxa being later added, frequently not by their original authors. This information is relevant when checking whether the initial concept proposed for the genus was maintained or not.
Most species studied have some kind of sexual dimorphism. Those modifications were categorized in two distinct groups, concerning its level of taxonomic comprehension: general and special dimorphism.

General dimorphism
Characters that are referred to as of general comprehension are those proposed several times in the literature as being important to species definition, but were actually found to be more generalized, i. e., they in fact occur in higher taxa. General characters are found in all, or most, male of diabroticites analyzed, and might also support the definition of larger taxonomic groups: 1) Smaller body size: considered general for Chrysomelidae, observed for most diabroticites; 2) Bigger eyes (relative to the total size of the head): cited often in species descriptions, but actually observed in most diabroticites; 3) Tarsal adhesive disks: structures present in most Chrysomelidae, with variation found among subfamilies and often among tribes, regarding the number of legs in which they occur and the proportion of the dorsal surface that they occupy (Stork 1980). For Diabroticites, the adhesive disks are present at the first tarsomere in pro-and mesothoracic legs; 4) Emargination on the posterior margin of ventrite V: this characters sometimes supports the definition of the subfamily Galerucinae, and its shape can be diagnostic of tribes, being rounded in Luperini and Galerucini, for example Bechyné 1962, Wilcox 1965). In diabroticites this emargination is, usually, round; and 5) Prothoracic tibiae with continuous apex, without spurs: probably constant in the whole tribe (Wilcox 1965).

Special dimorphism
Special characters are those used to describe lower taxonomic ranks, i. e., genera and species, and have been or not mentioned as diagnostic features in original descriptions.
With the exception of two genera, Amphelasma Barber, 1947 and Cochabamba all genera in Diabroticites are represented with special dimorphic characters. Of all species analyzed, only 47 were found to have special dimorphic characteristics. This is interesting, since the original definitions of diabroticites genera were usually based in sexual dimorphism. In accordance with the pattern observed for the subfamily (Mohamedsaid and Furth 2011), special dimorphism was observed in every tagma in the analyzed taxa of Diabroticites (Table 3).
Out of the 12 genera recognized by dimorphic characters, 4 are monotypic (Buckibrotica, Palmaria, Platybrotica and Pseudodiabrotica). The diagnostic characters and its validity will be discussed for each tagma.

Head
Most of the dimorphic characters were found in the head, mainly in the antennae, a proportion which agrees with the general pattern observed in the subfamily (Mohamedsaid and Furth 2011).
Gynandrobrotica was described as showing an "excavated clypeus", without any further details or illustrations. All four species studied have the same kind of modification, which is better described as the frons being elongated and with a shallow, smooth, round concavity (Figure 1), accompanied by sparse, large punctuation. It is interesting to compare this character with the differently excavated type of frons found in species of Cerotoma Chevrolat, 1837 (Figure 2), Eucerotoma Laboissiere, 1939 and Neobrotica Jacoby, 1887-all of which are usually placed in sister-section Cerotomites Chapuis, 1875. Gynandrobrotica has been suggested to be more related to these taxa in some phylogenetic analyses (Eben andMonteros 2004, Gillespie et al. 2008). Other characteristics common to these genera, such as the small eyes, the elongated frons and different shapes of antennomeres I-III should be further studied in order to verify if they are indeed homologues.
There are also three other species with distinctive head features in the male: Isotes onira (Bechyné & Bechyné), 1961 has an enlarged head from the vertex up to the antennal insertions - Figure 3), Diabrotica serroazulensis Bechyné & Bechyné, 1962 has an enlarged labrum, which is as large as half of the frontal length, and Acalymma cornutum (Baly, 1886), has a deep cleft in the frons, with lateral projections, and also a labrum with an acute anterior projection that reaches the frons (illustrated in Munroe and Smith 1980).

Antennae
Galerucines commonly display filiform antennae, which can show numerous dimorphic variations (Jolivet andVerma 2002, Mohamedsaid 2004). This is also true for table 3. Location of diagnostic, sexually dimorphic characters mentioned in original descriptions.  Diabroticites, with the main antenna type being filiform and antenommeres mostly subequal in size and shape ( Figure 4A). Non-dimorphic modifications are often related to the length of some antennomeres and sometimes support generic definitions. For instance, the two largest genera in Diabroticites, Diabrotica and Synbrotica are essentially distinguished by the length of antennomere III, described as being subequal to II in the first (same as observed in genus Cochabamba in Figure 4B), and subequal to IV (that is, almost twice as longer as II, as in Figure 4C) in the latter (Bechyné 1956).

Genus Antennae Legs Head (except antennae) Elytra
Eight genera have diagnostic characters based in their dimorphism of antennae (listed in Table 3). Examples of modified antennae are showed in Figure 4C-F. Because of their variability, dimorphic antennae are the most used structure in descriptions, but their modifications have been scarcely detailed. As a result, there are differ-   (Bechyné & Bechyné, 1961), dorsal view, detail of pronotum and head (female, left, male, right). ent genera proposed on characters depicted in sentences such as "apical antennomeres modified" (Cornubrotica), "antennomeres VII and IX of complicated shape" (Buckibrotica - Figure 4E), "antennomeres V-VII modified" (Paratriarius).
The absence of unified criteria in the understanding of what a "modified" antennomere is has lead, several times, to the establishment of artificial grouping of species, simply because a single "aberrant" antennomere can display an assembly of four different aspects of its morphology. Modifications include change in (in quoting marks, expressions used on original descriptions): length -antennomeres considered "elongated" or "shortened" when compared to the usually fixed antennomeres I and III; width -antennomeres described as "swollen", "inflated" (homogeneous modification), "distally expanded" (heterogeneous modification), and dorsoventrally "flattened"; shape  (Bechyné & Bechyné, 1961) D Aristobrotica angulicollis (Erichson 1878) e Buckibrotica cinctipennis (Baly, 1886) (detail in ventral view) F Ensiforma caerulea Jacoby, 1876. Scale bar = 1 mm.
-those described as having "lateral projections" or "ventral excavations"; structure -antennomeres with "rough punctuation" and different amounts of hairs/sensillae. As variations observed in the species do not always correspond to the diagnosis defined for genera they have been included in, these characteristics do not provide an accurate guide to the identification of taxa in Diabroticites.
For the non-monotypical genera based on antennal dimorphic features, most original definitions do not correspond to their actual characters. In Anisobrotica, for example, the "widened" apical antennomeres do not always appear -A. binisculpta Bechyné & Bechyné, 1969 only has in common with the other taxa the excavation present in glabrous ventral surface of antennomeres IX-XI (such as excavations observed in apical antennomeres of A. donckieri (Baly, 1889) in Figure 5). The same happens with Paratriarius, which includes several species that do not show "modified antennomeres V-VII" present in type-species P. dorsata (Say, 1824) (illustrated in Wilcox 1965) such as P. batesi (Baly, 1859), P. falvolimbata (Erichson, 1847), P. verrucosa (Jacoby, 1880), P. alternans (Weise, 1916), P. nigrotibialis (Bowditch, 1911), P. castanea (Bowditch, 1911), and also other four species studied by Mohamedsaid and Furth (2011). Instead, these taxa show antennae very similar to the general pattern seen in Diabrotica. The two species included in Cornubrotica do not show identical antennomeres VIII and IX, although both always have ventral excavations (illustrated in Bechyné 1969 andMoura 2005). In Paranapiacaba, the antennal character chosen was, unfortunately, a general one: male antennomeres III-XI uniformly "thickened" (in contrast with slightly slender antennae of females). Nevertheless, antennae do seem to vary uniformly in one genus. In Aristobrotica, the pattern of antennomeres III-V "thickened" is constantly repeated, followed by an unmentioned presence of larger punctuation ( Figure 4D).
Although Maulik (1936) suggested that, for indo-asian galerucines, the basal antennomeres are more frequently the altered ones, a result that has been corroborated by Mohamedsaid (2004), that feature does not apply to diabroticites analyzed. Also, no obvious topological pattern is seen in the variation of antennomeres (Table 4).
The number of modified antennomeres oscillated between 1 to 6. Antennomere II was recorded as dimorphic only in species of Aristobrotica, such as A. angulicollis (Erichson, 1878) ( Figure 4D) and in Isotes onira ( Figure 4C). Few modifications occur in antennomeres II and XI. The most affected are antennomeres V to IX. However, there is no indication of an explicit dependency of occurrence between any pair of modified antennomeres. This is the opposite of what has been observed for asian Galerucinae species (Mohamedsaid 2004). Although no pattern is observed, some variation can occur in blocks, i. e., one modified antennomere occurs with one or two adjacent antennomeres also modified.
It seems that most, if not all, antennal variations could be regarded as the result of the presence of punctuation and setae in greater number, either for the production and/or reception of chemical compounds (i. e., pheromones) (Jolivet 2007). A study on Diabrotica virgifera Leconte, 1858, for instance, showed that male antennae have a much greater number of sensilla than females, and numerous glandular points linked to the production of chemical compounds allegedly to be attractive to females (Newman Jr et al. 1993). Unfortunately, there is little knowledge on the biology or even on the anatomy of Diabroticites to support this view as a more generalized tendency.
Only metathoracic legs lack alterations in male diabroticites. In accordance with the more general pattern, morphological differences in the pro-and mesothoracic legs  usually are connected to the augmentation of absolute size in femora and tarsomeres I, the latter which are directly linked to the partial or total covering of the ventral surface by adhesive setae (Figure 7).
Tibiae and femora can also be modified, being greatly enlarged (such as in Zischkaita serrana Moura, 2003 -Figure 8), and frequently with internal margins concave or bearing tubercles, forming the "prehensile organ" (Bechyné 1956). Bechyné's concept of such structure is based on a combination of multiple adaptations and should be used with caution, since the homology of the "prehensile organs" can be difficult to assess. Aristobrotica, for instance, has been described as with one diagnostic feature: the "special build of the median tibiae in male". The detailed analysis of species included, however, indications that there are at least two distinct types of "prehensile organs" being treated as the same modification. While type-species A. angulicollis (Figure 9) bears only a concave mesotibiae with laterally flattened apex, A. mirapeua Moura, 1997 andA. capillosa Moura, 2011 (both illustrated in their original descriptions) display small projections of the ventral margin of the mesofemora (apical in A. mirapeua and basal in A. capillosa), and differently shaped tibiae (with a basal concavity in A. mirapeua and slightly concave tibiae with apical flattening in A. capillosa).
Moreover, general characters have been used to support definition of genera such as Cornubrotica and Synbrotica, a genus which is now a synonym of Isotes. The former  Moura, 2003, ventral view, male. was supposed to be distinguished by pro-and mesothoracic legs of males without emargination, which is rather common in the tribe, and the latter is characterized by antennomere III elongated and a "uniform pilosity covering the ventral surface of tarsomeres in both sexes" (freely translated from the original, in German) -something that does not accurately identify the males in this group, as they normally have distinctive adhesive disks in their tarsomeres.

Abdomen
The most common abdominal modification seen in some galerucines is the presence of processes with different shapes. Although no abdominal characters aid the definition of diabroticites genera, one character was observed for a single species in the group: a central triangular projection, postero-ventrally oriented, in the posterior margin of the ventrite I, in Zischkaita serrana (Figure 10). A similar alteration is observable in Hemygascelis longicollis Jacoby, 1896, an asian species that belongs to section Phyllobroticites in subtribe Luperina, a group thought to be a sister group of Diabroticina (illustrated in Mohamedsaid and Furth 2011).

Conclusions
The taxonomic history of taxa in section Diabroticites, which dates to over 150 years of specimen sampling and species descriptions, has been supported primarily by scarcely detailed descriptions of morphological features, specially coloration and striking "sexual aberrations". Sexually dimorphic characters have been the foundation for the creation of many genera in this group, albeit with vague definitions and a general disregard for comparative morphology. The use of inappropriate terminology by some authors has caused further confusion, since the habit of inferring homology from homonymy is common in the taxonomy of Chrysomelidae (Schmitt 1996). This might have been the origin of the combination of many species into these scarcely defined genera in Diabroticites. As a result, most generic original definitions, based on dimorphic characteristics, are not correspondent to their actual assembly of species.
Nevertheless, the establishment of homology theories is far from being trivial, and some variations might even occur in patterns that can actually help the definition of certain taxa, although it is clear that many sexually dimorphic features found in this section are possibly singular and autapomorfic. In this case, a broader morphological study is necessary.
The comparison of the dimorphic characters in Diabroticites with their related taxa, such as the Asiatic Aulacophorites Chapuis, 1875 (Luperini) and remaining Galerucinae is desired, in order to understand the evolution of such characters. There are striking similarities among many modifications found in these groups and thus it should be useful in the evaluation of homology as well. Parallelism might be the more parsimonious choice in many cases, but that is yet to be tested. Understanding the morphology is critical for better character definitions. Also, genital characters, which have been generally ignored, should provide important characters, as well as the account of several non-dimorphic characters, usually regarded as too uninformative, without detailed consideration.