On African Eupsilobiinae (Coleoptera: Endomychidae) with Descriptions of a New Genus and Species

Species of the South African genus Microxenus Wollaston are revised. Microxenus laticollis Wollaston is redescribed, and M. muelleri sp. nov. and M. krugeri sp. nov. are described. Natalinus gen. nov. and its single included species, N. klimaszewskii sp. nov. are described. All of these taxa are diagnosed and illustrated, and a key to the species of Microxenus is presented. Female genitalia of newly described species are discussed in terms of monophyly of Eupsilobiinae. Zoogeographical and biological data of African Eupsilobiinae are summarized.


Introduction
Casey established the tribe Eupsilobiini in 1895 for his new species Eupsilobius politus. Historically it has been difficult to place among the Cerylonid Series families. Sen Gupta and Crowson (1973) classified Eupsilobiini in Cerylonidae, and synonymized Eupsilobius with Eidoreus, established in Erotylidae by Sharp (1885) for E. minutus Sharp from Hawaii. Crowson (1981) suggested the relationship of Eidoreus with Coccinellidae, but Sasaji (1986Sasaji ( , 1987 placed it in Endomychidae and established Eidoreinae as a new subfamily, being unaware that the name Eupsilobiini of Casey was available. Pakaluk and lipi ski (1990) followed Sasaji in the family placement and reviewed the subfamily Eupsilobiinae at the genus and species level.
Phylogenetic studies of the family Endomychidae based on adult and adult and larval morphology combined (Tomaszewska 2000(Tomaszewska , 2005 confirmed Eupsilobiinae as members of Endomychidae. Based on these studies, Eupsilobiinae forms a distinct, monophyletic group based on adult synapomorphies like median lobe coiled apically and ovipositor with stiff, inflated, infundibulum-like structure between the bursa copulatrix and the sperm duct. However, they are also characterized by having antennal grooves on the head and a median lobe with a T-shaped basal capsule (similar to that in Coccinellidae). The single known larva of Eupsilobiinae (Evolocera Sharp 1902) has a rigid tooth-like prostheca like in Mycetaea Stephens 1829 (Mycetaeinae). These phylogenetic analyses have not demonstrated any clear relationships between Eupsilobiinae and any other subfamily of Endomychidae. Shockley et al. (2009b) listed 6 genera and 13 species of the subfamily. Five genera are restricted to small endemic areas of Central and South America (4 genera) and South Africa (1 genus), while Eidoreus is known from widely scattered islands like Cuba, Guadeloupe, Virgin Islands, Galapagos, Mascarene Islands, Seychelles, Sri Lanka, Fiji, French Polynesia, Solomon Islands, and Hawaii; E. politus (Casey 1895) was also collected in the Florida Keys.
So far, only the genus Microxenus was known from the Afrotropical region, established by Wollaston (1861) for M. laticollis from South Africa and placed in the family Mycetophagidae. Csiki (1905Csiki ( , 1910 placed this genus in the endomychid subfamily Mycetaeinae, and Strohecker (1953) followed this arrangement in his world catalogue of Endomychidae. Subsequently, Pakaluk and lipi ski (1990) placed Microxenus in the subfamily Eupsilobiinae, and this placement was later confirmed by phylogenetic analyses (Tomaszewska 2000(Tomaszewska , 2005.
While studying additional material of Endomychidae from South Africa, two new species of the genus Microxenus and a new genus of Eupsilobiinae were discovered. These taxa are described here as follows: M. muelleri sp. nov., M. krugeri sp. nov., and Natalinus gen. nov. with its single new species N. klimaszewskii sp. nov. This work raises the number of known eupsilobiine species from the Afrotropical region to 4 and the known genera to 2.
NHM -The Natural History Museum, London, England MIZ -Museum and Institute of Zoology PAS, Warszawa, Poland TMNH -Transvaal Museum of Natural History, Pretoria, South Africa Measurements were made using an ocular micrometer attached to an Olympus SZH-10 (www.olympus.com) dissecting microscope. Measurements recorded were as follows: total length from apical margin of clypeus to apex of elytra; pronotal length from the middle of anterior margin to base of pronotum; pronotal width at the widest part; elytral length along suture, including scutellum; and elytral width across both elytra at the widest part. Male and female genitalia were dissected, cleared in 10% solution of KOH, and placed in glycerine on slides for further study. Illustrations were made from slide preparations using a camera lucida attached to the same Olympus dissecting microscope.
Scanning electron micrographs photographs were made using a Hitachi S-3400N (www.hitachi.com), and digital photographs were made using a Leica digital camera (us.leica-camera.com) mounted on microscope and subsequently enhanced using Auto Montage software in the Electron Microscopy Laboratory of the MIZ.
Terminology used in this paper follows Tomaszewska (2010).

Diagnosis
Microxenus is closely related to Natalinus. It differs from Natalinus by having the metaventrite with postcoxal lines, the mesoand metaventrite without postcoxal pits and the scutellum distinctly much more transverse, with at least weakly emarginate hind margin. The scutellum, which is at least 3 times wider than it is long, with weakly emarginate hind margin, is unique for Microxenus within Eupsilobiinae.
Pronotum (Figures 19,22,30) Pakaluk and lipi ski (1990), based on specimens from South Africa and Mexico, is a mixture of features of two genera. After many years of extensive study on Endomychidae, including 2 large phylogenetic papers illustrating many different taxa (Tomaszewska 2000(Tomaszewska , 2005, it is obvious that the Mexican specimens from Pakaluk and lipi ski (1990) do not belong to Microxenus, but instead most likely belong to Evolocera, a genus transferred from Merophysiinae to Eupsilobiinae by Tomaszewska (2005).

Diagnosis
This species is most similar to M. krugeri by body shape, but can be separated by the following combination of characters: short lateral sulci present on pronotum, prosternal process distinctly wider than procoxal diameter, abdominal postcoxal lines shallow and extending posteriorly less than half the length of ventrite I, and male abdominal ventrite V emarginate.

Distribution
South Africa.

Etymology
The specific epithet refers to Kruger National Park in the Republic of South Africa, the type locality of this new species.

Diagnosis
Microxenus krugeri resembles M. laticollis in body shape, but can be separated by the following combination of characters: lateral sulci absent from pronotum, prosternal process as wide as procoxal diameter, abdominal postcoxal lines deep and extending posteriorly more than half the length of ventrite I, and male abdominal ventrite V subtruncate.

Etymology
The name of this new species is dedicated to Dr. Ruth Müller, a curator of the Coleoptera collection, Transvaal Museum, Pretoria (Republic of South Africa).

Diagnosis
This is a very distinctive species of Microxenus by its regularly long-oval body, very long lateral sulci on the pronotum extending at least length of pronotum, prosternal process with longitudinal, median carina, and male abdominal ventrite V deeply emarginate with additional submarginal groove running parallel to the emargination.

Description
Length 1.20-1.45 mm. Body (Figures 3, 6, 26) 2.0-2.2 times as long as wide; pronotum 0.58-0.63 times as long as wide; elytra 1.32-1.38 times as long as wide; 2.52-2.60 times longer than pronotum, 1.13-1.19 times wider than pronotum. Color brown with appendages slightly lighter; vestiture pale, about as dense ventrally as on dorsum. Antenna (Figures 8,  12, 30) with terminal segment shorter than penultimate one. Pronotum (Figure 30) widest near of apical length, with base distinctly narrower than base of elytra; hind angles blunt; lateral sulci on pronotum very long, extending of pronotal length; prosternal process ( Figure 27) weakly bordered laterally, with weak, longitudinal median carina; nearly parallel-sided, truncate at apex; 0.8-0.9 times as wide as procoxal diameter. Mesoventral process (Figure 28) about as wide as mesocoxal diameter; elytra with lateral margins visible from above along nearly half of their basal length ( Figure

Etymology
The genus name refers to the Natal province in Republic of South Africa, the type locality of the type species. Gender masculine.

Diagnosis
Natalinus is apparently closely related to Microxenus. It can be separated by having the metaventrite without postcoxal lines, the meso-and metaventrite with large, setose postcoxal pits and the scutellum distinctly less transverse with regularly rounded hind margin. The lack of postcoxal lines within Eupsilobiinae occurs only in Ibicarella from Brazil, but it has 11-segmented antennae (10segmented in Natalinus) and hypomeron with long, deep antennal grooves (absent in Natalinus). Postcoxal pits on meso-and metaventrite is unique for Natalinus.
Pronotum (Figures 42, 46) transverse, widest from base to about mid-length with base about as wide as base of elytra; pronotal surface finely and sparsely punctate; lateral margins visible throughout; basal sulcus rather distinct, lateral sulci absent; anterior angles rounded; posterior angles nearly right-angled. Pronotal disc convex. Prosternal process ( Figure 50) moderately wide, bordered along lateral margins, with weak longitudinal, median carina extending anterior of procoxae nearly to anterior margin; very weakly expanded apically, with apex rounded; extending posteriorly well beyond procoxae. Procoxae circular in outline, procoxal cavity externally open, internally closed; trochantin concealed. Scutellum ( Figure 42) very small, transverse, rounded posteriorly. Mesoventrite ( Figure 50) with a pair of large, setose pits anterolaterally, just posterior of procoxal cavities; intercoxal process bordered, somewhat trapezoidal in shape with weak longitudinal, median carina, slightly narrower than mesocoxal diameter, not extending beyond them posteriorly. Mesocoxae circular in outline, mesocoxal cavity outwardly open; trochantin exposed. Meso-metaventral junction of straight-line type. Elytron ( Figure  42) regularly elongate-oval, with rounded apex, abruptly convex, with lateral margin visible from above along half of basal length; punctation irregular; sutural stria visible from mid-length to apex; epipleuron incomplete apically (Figures 43, 49). Metaventrite (Figures 43, 50) strongly transverse, weakly convex, with two pairs of large, setose postcoxal pits; medial pit directly posterior to midpoint of mesocoxal cavity, lateral pit directly posterior to the lateral margin of the mesocoxal cavity; femoral lines absent; discrimen absent. Metacoxae transverse, widely separated. Metendosternite like that in Microxenus, with very short stalk and widely separated anterior arms and tendons (see Tomaszewska 2000: 526). Hind wing absent.

Etymology
The name of this new species is dedicated to Jan Klimaszewski, a Polish entomologist and one of the collectors of the type series.

Systematics
Eupsilobiinae is a well-supported clade in cladistic analyses based on adult characters (Tomaszewska 2000) and on adult and larval characters combined (Tomaszewska 2005). The following two adult synapomorphies have supported monophyly of this subfamily: median lobe long, thin with T-shaped capsule at base and coiled apically, and female genitalia with infundibulum-like structure (stiff, inflated structure between bursa copulatrix and a proper sperm duct).
The present study confirmed both of these characters, although with slight modification regarding the structure of the female genitalia. The term "infundibulum-like" structure used previously by Tomaszewska (2000Tomaszewska ( , 2005 is outdated and inaccurate. The term suggests that this modified structure is located exactly between the bursa copulatrix and the sperm duct, or is formed from an apical part of the bursa fused with the outlet of the sperm duct. However, female genitalia of M. muelleri suggest that this structure may be entirely a modification of the proper sperm duct; the genitalia of N. klimaszewski show less inflation in this structure and indicate that it may be formed from a sclerotized part of the sperm duct forming a twisted, rigid spiral.
The African Eupsilobiinae now includes two endemic genera: Microxenus and Natalinus gen. nov. The presence of large, setose pits on meso-and metaventrite is so far unique for Natalinus, while the scutellum at least three times wider than it is long, with weakly emarginate hind margin is unique for Microxenus within Eupsilobiinae.

Zoogeography
New taxa described and collection localities increase the known distribution of African Eupsilobiinae from the Western Cape region to the Kruger National Park in the northeast and Natal Province in the southeast.
The range of Microxenus includes a western coastal area of Western Cape and Kruger National Park in northeastern South Africa. All three species appear to be allopatric. Microxenus laticollis is known only from the Cape of Good Hope; M. muelleri collected from a few costal localities of Western Cape between 31˚ 39 N and 32˚49 S; M. krugeri collected at two localities, close to each other: 24˚ 13 S, 31˚ 56 E and 24˚ 59 S, 31˚ 39 E in Kruger National Park. Natalinus is known so far from a single locality in southern part of the Kwazulu-Natal Province (30˚ 34 S, 29˚ 43 E).

Biology
Microxenus laticollis has been collected from or in the vicinity of ant nests (Pakaluk and lipi ski 1990), and like several other species of Eupsilobiinae is probably a social insect inquiline. However, it is most probably mycophagous, feeding on a combination of spores and hyphae of microfungi (Skelley and Leschen 2002). The nature of the relationship between Eupsilobiinae and their social insect hosts remains unclear (Shockley et al. 2009a).
The specimens of M. krugeri and M. muelleri were collected from groundtraps with meat and banana bait. Additionally, M. muelleri was collected from groundtraps with feces bait. Attraction to meat and feces is unusual, and so far an unknown habit for endomychids.

I am indebted to Ruth Müller (TMNH) and
Roger Booth (NHM) for the loan of specimens used in this study. Thanks to Floyd Shockley and Adam lipi ski for reading a draft of this paper and making many helpful suggestions.
Magdalena Kowalewska-Groszkowska (MIZ) helped with SEM and habitus photographs. Two anonymous reviewers are acknowledged for valuable and critical comments on the manuscript.

Editor's note
Paper copies of this article will be deposited in the following libraries.