Revision of the genus Mimesthes (Coleoptera: Meloidae: Mylabrini)

The species of the mylabrine genus Mimesthes are revised and a new species (M. karooensis) is described from the South African Karoo. The systematic position of the genus is also discussed and an identification key to the species is proposed. Geographical distribution and habitat preferences, as well as other bionomic features, are summarised.


INTRODUCTION
The meloid tribe Mylabrini (as defined by Bologna, 1991) is the largest tribe of the subfamily Meloinae, in cludes approximately 750 described taxa and occurs only in the Old World.Recently, I have started a project ad dressed to the review of Mylabrini, primarily focused on the revision of the smaller genera and of single subgenera or groups of species of the large genera Mylabris F., 1775 and Hycleus Latreille, 1829.
All Mylabrini were referred by early taxonomists to the "collective" genus Mylabris.Since the primitive attempt . of revision made by Billberg (1813), Marseul published two monographs on the Palaearctic (1870) and Old World (1872) species, and subdivided Mylabris into four sub genera, which are now considered as distinct genera (the nominate; two described in the first monograph: Lydoceras, Ceroctis; another described in the second mono graph: Mimesthes).Marseul also considered three other genera previously described by other authors on the basis of the antennal segment number and shape (Actenodia Castelnau, 1840;Coryna Billberg, 1813;Decatoma Cas telnau, 1840).The systematics of the tribe is still con fused, and several taxa have been considered to a generic or subgeneric level, using different synonyms.The classi fication proposed in my recent contributions (Bologna, 1978(Bologna, , 1990(Bologna, , 1991;;Bologna & Coco, 1990) is temporarily utilised in this paper.
The eleven genera included in the tribe (Bologna, 1991) are distributed throughout the Old World, particularly in Africa.The present arrangement of the genera (and syn onymies), based on adult characters, is partially in dis agreement with the more recent literature (Kaszab, 1969;Selander, 1991), but similar to that proposed by Pardo Al caide (1950, 1954), who used a different nomenclature.
As recently emphasised (Bologna, 1991), larval mor phology and biology greatly influenced the phylogenetic reconstruction of several groups of Meloidae.Recent re search on larval morphology of Mylabrini (Bologna & Pinto, in press) partially clarifies the possible relation ships among a few taxa.Unfortunately our knowledge on the mylabrine larval morphology and development is minimal: records of less than 60 species of 5 genera have been published, but only a few of these studies are ex haustive (e.g.Paoli, 1932Paoli, , 1937)).In the last ten years I obtained triungulins of about 25 additional species of Mediterranean and African Mylabris, Hycleus, Ceroctis, and Paractenodia Peringuey, 1904.Attempts to rear other genera (Mimesthes, Lydoceras, Croscherichia Pardo Al caide, 1950 andActenodia Castelnau, 1840) were not suc cessful (see "Bionomics" for an additional note).The first instar larvae of the two last genera have been described in literature (cf.Bologna & Coco, 1990 for Croscherichia; Cros, 1930 tor Actenodia), and are partially re-examined in the framework of a comparative study (Bologna & Pinto, in press).
The possibility of working on the revision of the tribe is greatly limited by the large number of species of some genera, especially Hycleus and Mylabris, and by the pres ence of numerous infraspecific taxa described on the ba sis of the very variable, elytral colouring.Consequently my studies have been primarily addressed to the smaller groups of Mylabrini: I previously revised the Saharo-Sindian genus Croscherichia (Bologna & Coco, 1990) and I am presently working on Lydoceras, Paractenodia, Actenodia, and on some groups of species or subgenera of the large genus Mylabris (the nominate subgenus Mylab ris and the subgenus Calydabris Kaszab, 1960).
The present paper is aimed at a revision of Mimesthes.This genus, endemic to southern Africa, and described by Marseul (1872) as a subgenus in his large revision of the genus Mylabris, originally included a single species, My labris (Mimesthes) maculicollis Marseul, 1872.Kaszab (1952Kaszab ( , 1981) ) added two more species (Mimesthes holgaticits (1952) and M. nigricollis (1981), and a fourth species (M karooensis sp.n.) is described in the present revision.After the original description, Mimesthes was considered at the generic level by Peringuey (1909) and by all other specialists of the family.The three species previously known were described in the literature more or less in de tail, therefore, in the present contribution only a brief di agnosis is given while additional morphological characters are discussed and figured.Some ecological and faunistic records are also added, as well as a key to spe cies.

Adult morphology
A member of Mylabrini sensu Bologna (1991).Small sized (length: 7-13.1 mm; width on elytra: 2.7-5.2mm), short and sub-rectangularly shaped (Fig. 1); body black without metallic reflections, antennae black, partially or entirely rufous; elytra yellow-orange with black spots and bands.Head transverse, temple shorter than eyes, which are large and bulging; antennae with 11 visible segments, the last 3 very compressed and approached, or with 8 visible segments due to the complete fusion of the last 4 segments (karooensis), short and clavate, distinctly longer and more slender in the male, more enlarged apically in female; male maxillae and palpi unmodified.Pronotum transverse, subquadrate, not distinctly narrowing anteri orly; mesosternum without a definite shield or oblique furrows, with a large anterior smooth and glabrous area well demarcated from the remaining surface which is densely setose; mesepisterna narrow and only slightly bordered; elytra flat and large, subrectangular, subtransversely truncate at apex with sutural angle subquadrate.Male pro-and mesotarsi with thickly sericeous yellow pads and slightly enlarged segments, only with a few modified setae at apex of posterior tarsi; female tarsi nor mally setose, slender and with isolated elongate setae on external margin; segment I of tarsi longer than II + III to gether; female fore tibiae externally acutely protruded at apex; spurs of metatibiae both similar, elongate, pointed; tarsal claws normal, ventral blade about as long as dorsal.Parameres apically slender, more or less depressed on side; aedeagus apically subquadrate in lateral view, with a single hook (see notes on M. nigricollis), extremely far from apex, similar and probably homologous to the proxi mal hook of other Mylabrini; aedeagus gibbose near the ventral opening of the ductus ejaculatorius; hook of endophallus very elongate, slender, slightly curved at apex.

Relationships
Bologna (1978,1991) and Bologna & Coco (1990) pro posed a phylogenetic arrangement of Mylabrini and dis tinguished at least three possible phyletic lineages.One of these included Mimesthes, Actenociia and Paractenodia (considered as possible synonym of Actenodia).The re cent study of numerous Paractenodia specimens and of the triungulin of P. parva Peringuey, 1904 (Bologna, unpubl.)confirmed the validity of this genus and clarified its true relationship with another lineage that includes the genera Hycletts and Ceroctis.
This phyletic lineage, containing only Mimesthes and ■Actenodia, is identified by the following characters: Mesosternum without a definite shield, mesepisterna scarcely bordered; aedeagal hooks positioned far from apex or reduced to a single hook; antennal apical seg ments compressed and progressively enlarged, completely or incompletely fused.
The genus Actenodia is well differentiated from other Mylabrini both by adult (pronotum, mesosternum, aedea gus, antennae) and larval features (leg structure).Some of the southern African species, included in this genus due to the presence of only 8 antennal segments, actually should be referred to Hycletts based on mesosternal and aedeagal structures (Bologna, unpubl.; cf. also Bologna, 1990 for other East African species with the same charac ter).Unfortunately, the Mimesthes triungulin is still un known (see "Bionomics" for an additional note); conse quently the true affinities of this genus remain unclarified, and only a few unverified synapomorphies with Acteno dia (pronotum transverse, mesosternum without a definite shield, aedeagus with proximal hook very far from apex) can be proposed.

Bionomics
Zoogeography.Mimesthes is endemic to the "Karoo-Kalahari" zoogeographical subregion, restricted to south ern Namibia and western South Africa.The range of the genus (Figs 2, 3) includes a coastal and subcoastal area in Namibia and Namaqualand, between about 26° on North and 33°30' to South, the Bushmanland and northern Ka roo (from Aggeneys to Putsonderwater, and from Vanrhynsdorp to Carnarvon and Prieska), the southern and central Karoo (from Stellenbosch to Willowmore and Victoria West) to East.This range is more or less paral leled by several other insect genera, including some Meloidae: Iselma Haag-Rutenberg, 1879, Paractenodia, and an undescribed lyttine genus (Bologna, unpubl.).It is generically referable to the distribution of the "Cape-Namaqua elements" (Holm, 1990; see also the "Capewestern extension" proposed by Endrody-Younga, 1978), or, more extensively, to the "Karoo-Namib regional cen tre of endemism" proposed by White (1983) for plants.
Only one species, M. maculicollis, has an extensive geographical range in Namib, Namaqualand and Karoo.This range overlaps completely with those of M. nigricollis (strictly endemic to the extreme southern Namib desert) and M. holgaticus (endemic to a narrow coastal strip along Namaqualand), while M karooensis is en demic to the Little Karoo and southern Great Karoo (Figs 2-3).
Except for M. maculicollis, other species are allopatric and perhaps represent primarily vicariant elements de rived from speciation events in original isolated xeric ar eas.The recent strong modifications of the original habi tats caused by grazing (see Acocks, 1988: maps 1 and 2) perhaps produced a recent extension of the range of the more euryoecious species (M maculicollis).M. nigricol lis is syntopic and synchronic only with M. maculicollis in a single locality (Obib dunes); M. holgaticus is syn topic and more or less completely synchronic with M maculicollis only in a few localities of the Namaqualand coast (Jakkalsputs, Holgat and environs, Oograbies, Porth Nolloth, Buffelsrivier, Dembergsdraai, Kotzesrus).
Habitat preferences.The range of Mimesthes includes different types of habitats, but particularly the Succulent-Karoo semidesert, and in part the Namib desert and Main Karoo deserts and semideserts (Little Karoo, Great Karoo, Central Upper Karoo, Northern Karoo, etc.).For habitats details see White, 1983;Acocks, 1988;Seehly, 1990;and Rutherford & Westfall, 1994.In the coastal southern Namib, Richtersveld and Namaqualand, these xeric habitats are characterised by very scarce rains, usu-Tabu: 1. Distribution of records of Mimesthes species for Habitats (South Africa) and Biomes (Namibia).Habitats and Biomes (no. of records) maculicollis karooensis holgaticus nigricollis In Table 1, the distribution of records for the different habitats (according to the South Africa veld types pro posed by Acocks, 1988) is represented.No similarly de tailed habitat division is available for Namibia; in Table 1, the distribution of Namibian records are summarised, separated into the biomes proposed by Ruthenford & Westfall (1994) and slightly modified according to J. Irish (unpubl.).
M. nigricollis Kaszab is particularly related to the Na mib Desert Biome and the Succulent Karoo Biome; the original labels indicate that type specimens were collected on dunes.M. holgaticus is totally restricted to coastal sand flat or subcoastal sand hills ("Strandveld"); labels of several specimens indicate as habitat red dunes (Holgat, Manganese mines), white dunes (Kotzerus, Zwartduine), generically "dunes" (Rooidam Farm), or also "sand hill" (Buffelsrivier, Dembergsdraai).Also some specimens of M. maculicollis have labels indicating "red dunes" (Hol gat; Roudabel Farm), but this is a more euryoecious spe cies, associated particularly with several "velds" types of the Succulent Karoo and Nama Karoo Biomes.Finally, M. karooensis also is associated with different "velds" types of the Succulent Karoo Biome All species are apparently restricted to low and middle elevations: M. holgaticus is distributed from sea level to 200 m a.s.l.; M. nigricollis perhaps reaches slightly more elevated areas; M. maculicollis is more euryzonal, from sea level to about 1,000 m a.s.l. of elevation in the  2. The preponder ance of Arctotis and Osteospermum records suggests that these genera are preferred over other Asteraceae, as well as Cephalophyllum vs. other Mesembryanthemaceae.Some M. maculicollis specimens were maintained by the author on Asteraceae flowers of different species.
As regards to other species, the museum label of one holgaticus specimen from Hoekbai River (TMSA) indi cates "on Mesembryanthemaceae".A single specimen of M. karooensis was personally collected near Ladismith on a yellow unidentified species of Osteospermum (Astera ceae); the single specimen from Victoria West is labelled "from flowers".No information is available on M. nigri collis.
Behaviour and general adult activity.No behavioural information is available; only defence behaviour with thanatosis and autohaemorrhagy, on flowers or on ground, was personally observed in M. maculicollis and M. ka rooensis.Adult Mimesthes are, from personal observa tions and the apparent complete lack of records at lights, strictly diurnal.
Records of seasonal adult activity are summarised in Fig. 4: all species have a very restricted period of activity during spring (particularly after the winter rains in Sep tember).
According to museum labels, specimens of M. maculi collis were collected using "faeces bait 76 days traps" (Vredendal, TMSA), or "ground traps with meat bait" (Stinkfontein, TMSA); probably in both cases the bait hu midity attracted a few specimens.
Mimicry.Possible cases of Müllerian mimicry concern the sympatric populations of different Mimesthes species as well as three sympatric undescribed mylabrine species perhaps referable to a new genus (Bologna, unpubl.),dis tributed along the Namib desert.Mimicry is assumed on the basis of museum specimens sympatric with M. macu licollis, and the extreme phenotypical resemblance.One species of undescribed mylabrine was collected in syntopy with M. nigricollis and shows similar red colour of antenna and similar elytral colouring.All these mylabrine species have orange lateral spots on pronotum, similar elytral colouring, and subrectangular shape of elytra; the presence of red-orange colouring on pronotum is very un common in the Mylabrini.They clearly differ from Mi mesthes by the antennae, pronotum, and the aedeagus (distal hook only, at least in two of the three species).
Larval biology.No information on larval biology and morphology is available.Rearing attempts of M. maculi collis from three different Namaqualand localities (Sep tember 1994) have been unsuccessful.(Note: While this paper was in press, I obtained the triungulin of M. maculi collis, which will be described in a separate paper.)

Species taxonomy
The species of this genus are phenetically very similar, and fonn a well-isolated monophyletic group.Only possi ble autapomorphies but no synapomorphies among spe cies have been identified, and consequently no phylo genetic inferences are proposed here.Phenetic similarities or differences involve antennal, pronotal and setal colour- Probably two phenetic groups of species can be distin guished: The first includes M. maculicollis and M. karooensis; the second includes M. holgaticus and M. nigricollis, characterised by the pronotum more trans verse and the shape of parameres similarly depressed on the side.
M. karooensis is particularly differentiated by the fu sion of the last four antennal segments.Other Mimesthes species have the apical antennal segments very ap proached but not fused.The tendency of the fusion of the last segments, sometimes forming an apical club, is probably a homoplasic condition occuring in several other mylabrine genera.Actenodia and Paractenodia have fused segments (8-11 or 7-11) into a distinct club; Semenovilia has completely fused segments 10-11, forming an apical subcylindrical segment; several species of Hycleus show a complete or incomplete fusion of the last 2-4 seg ments, forming a more or less clubbed apex.On the other hand, only two and one species of Mylabris and Croscherichia have segments 10-11 fused.
Three species of Mimesthes have partially red pronotum, an uncommon condition in Mylabrini, which usually have black pronotum.The black colour of M nigricollis perhaps could be considered a primitive condi tion.Antennae are partially red in two species, another uncommon condition in other mylabrine species (except in several Hycleus and in a few species of other genera), particularly in the genus Actenodia, here considered as a taxon probably related to Mimesthes.
The species of Mimesthes are very similar in general shape and colouring, and were described in detail by Marseul (1872, M. maculicollis) and Kaszab (1952, M. holgaticus;1981, M. nigricollis).Here are presented only brief descriptions and figures, in comparison with M. maculicollis.the male to the pronotum base; last three antennal segments approached and compressed but clearly distinguishable (Fig. 10); antennal segments V-VI longer than wide; body seta tion sparse also on frons; ratio of mandible length to head length in lateral view (from the fore margin of eye to the vertex) = 0.9; gonoforceps and aedeagus as in Figs 6-8.
Description.Body black, antennal segments III-V in some specimens dark reddish, pronotum orange on sides and partially on the middle, elytra yellow-orange with variable black colouring including spots and transverse bands; setation yellow-grey, long on head, pronotum and ventral side, shorter on elytra.Body length: 8.7-13.1 mm; elytral width at widest point: 3-5 mm.
Head slightly transversally depressed on frons, punc tures approaching but not confluent, particularly on tempora, large and moderately deep.Mandible elongate, about half as long as the length of head: In lateral view, the ratio of mandibles length to head capsule length (from the eye fore margin to vertex) is 0.9.Labrum longitudi nally depressed in the middle, less punctured than head.Labial and maxillary palpi completely black, last segment of maxillary palpi elongate, oval, subtruncate at apex.Male antennae extending to the pronotum base (Fig. 10), I I -segmented, last three segments compressed but dis tinct and not fused; segments I-II long with erect setae, III-XI with very short and dense setae; segment I elon gate and slightly inflated apically, segment II short and subspherical, III cylindrical and three times as long as segment II, 1.5 times as long as IV, segments IV-V cylin drical, decreasing in length, VI similarly shaped to IV and V, about as long as IV, segments VII-IX subtrapezoidal, anteriorly enlarged and similar in length, progressively more transverse, segment X subrectangular, slightly shorter than IX, segment XI obtuse apically, basally as wide as segment X at apex.Female antennae shorter and subclavate, segments VIII-XI distinctly wider than in male.
Pronotum rectangular and transverse, about 0.8 as long as wide; laterally parallel and anteriorly obtusely arcuated, with a short longitudinal depression in the mid dle and slightly depressed basally, punctures more sparse than on head; basal half, medial longitudinal strip, and the anterior and lateral margins black, anterior sides yelloworange.Prosternum posteriorly short.Mesonotum rounded at apex.Mesostemum (Fig. 5) only with an ante rior smooth and glabrous subtriangular area, slightly de-Figs 5-11: Mimesthes maculicollis Marseul. 5 -mesosternum; 6 -gonoforceps, ventral view; 7 -gonoforceps, lateral view; 8 -aedeagus, lateral view; 9 -spiculum gastrale; 10male antenna; 11 -elytral variability (from Kaszab, 1952Kaszab, , 1955, modified), modified).pressd in the middle.Metastemum obtusely angulate at apex.Legs elongate and slender, completely black Elytra distinctly wider than pronotum, more than three times as long as pronotum, with contiguous punctures, deeper on black colouring.The extent of black colouring is variable; basically three transverse zig-zag bands con nected along the suture and extended along the apical margin, isolating a yellow suboval apical spot; anterior black band more or less completely extended on humerus and connected to the black base.The elytral colour pat tern is very variable (Fig. 11) and includes more or less wide or incomplete bands, sometimes anteriorly reduced to isolated spots, or reduced to an extreme yellow form with black margins.Kaszab (1952Kaszab ( , 1955) ) described 20 in fraspecific forms.
Remarks.In the original description (Marseul, 1872) the antennal segment I is erroneously indicated as "court et renflé" and the pronotum is indicated about as wide as the head.
Type material.In the Marseul's collection (MNHN) two very damaged specimens that could be syntypes of this species are preserved.
Type locality."Cap de Bonne-Espérance" (Marseul, 1872).The type specimens are labelled "Cap B." (probably = Cape of Good Hope), but no species of Mimesthes occur in the Cape Peninsula.According to the range of the species, the original la bel probably generically indicates the western Cape Province (South Africa).
(from eye fore margin to vertex) is 0.7.Antennae in both sexes not extended to the pronotum base; apparently only 8-segmented because of the complete fusion of segments VIII-XI, forming a club, slender in the male (Fig. 16), and subclavate in female; segment VI as long as V; in some specimens a track of the suture between segments VIII and IX is slightly visible.
Pronotum in some specimens less parallel and slightly convergent in the basal half; usually with middle longitu dinal black strip.Mesosternum as in Fig. 12. Elytral col ouring less variable (Fig. 17) but usually with anterior black band fragmented, with an isolated spot near the su ture.
The holotype lacks posterior left tibia and tarsus.Six para types have both antennae broken after segment II; two other paratypes have the left antenna broken after segment II.
The last locality requires confirmation because of its ecologi cal characteristics and geographical position: The new species is apparently related to the Karoo habitats ("Little Karoo" and "Great Karoo") in the more eastern areas of the Cape Province; Stellenbosch is characterised by Mediterranean habitats ("Coastal scrub relics" or "Fynbos"; see Acocks, 1988).Perhaps the label generically indicates the region around Stellenbosch, which includes eastern close areas characterised by Karoo vege tation.
Etymology.The name originates from "Karoo", the typical biome of this species.
Distribution.Endemic to central and southern Karoo (Western and Northern Cape Province) (Fig. 3).The de tailed localities are listed above.
Head and pronotum with black setae; antennal seg ments III-XI red, only rarely dark reddish, several speci mens with segments III-V basally black.Fore half of pronotum usually red orange, black only on base; only a few specimens with middle longitudinal black strip con
T able 2. Food plant records of Mimesthes maculicollis.
ing, shape of antennal segments and pronotum, and body puncturation.The aedeagal and mesosternal features are only slightly diagnostic and the elytral colour pattern is similar.