Two Invaders Instead of One: The True Identity of Species under the Name Ceracis cucullatus (Coleoptera: Ciidae)

The Neotropical obligate fungivorous beetle Ceracis cucullatus (Mellié) has attracted attention of coleopterists due to the increasing number of records of populations in Africa. Although its disjunct populations have been interpreted as a cohesive taxonomic unity, previous comparisons between African and Neotropical specimens revealed differences in their external morphology, causing uncertainty about the true unity of the species. Here, we compare the external morphology of specimens named Cer. cucullatus from several localities of the Neotropical, Palearctic, Afrotropical, Afrotemperate and Oriental regions. As results, we reverse three previous junior synonymies of Cer. cucullatus, proposing Cer. lamellatus (Pic) and Cer. tabellifer (Mellié), both reinstated status and new combinations, as separate species. We also propose Enn. bilamellatum Pic as a new synonym of Cer. tabellifer. In face of these taxonomic changes, we identify Cer. tabellifer as the actual invasive species on African lands, instead of Cer. cucullatus as was previously accepted. Then, through historical records gathered from scientific collections and literature, and through examination of recently collected specimens from South Africa and Brazil, we provide data on host fungi and geographic distribution of Cer. tabellifer. Based on these data, we discuss possible explanations to the successful invasion of Cer. tabellifer in Africa and elsewhere and its potential threat to native faunas of ciids. This study helps to fulfil an old gap in the literature on biological invasions, with considerably more studies on predatory species, disease vectors or potential pests of agricultural crops, than on non-pest fungivorous organisms.


Introduction
Ceracis cucullatus (Mellié) belongs to Ciidae (Coleoptera: Tenebrionoidea), a family of small obligate fungivorous beetles that live and breed in polypore basidiomes worldwidely. Ceracis cucullatus was originally described as Ennearthron cucullatum by Mellié in 1849 [1], based on specimens from Cayenne (French Guiana), Cape of Good Hope (South Africa) and Reunion Island, and subsequently transferred to Ceracis Mellié by Lawrence [2] (Fig. 1). It names the cucullatus species-group, which currently comprises Cer. bicornis (Mellié), Cer. cassumbensis Antunes-Carvalho & Lopes-Andrade, Cer. cucullatus and Cer. navarretei Antunes-Carvalho & Lopes-Andrade. It also encompasses the names Ennearthron tabelliferum Mellié, Enn. Bilamellatum Pic and Enn. lamellatum Pic, junior synonymies of Cer. cucullatus. These synonymies were proposed by Lawrence [2] who argued that they were described as new based primarily on differences in size and development degree of male pronotal projections. Such secondary sexual characteristics exhibit a wide variation in size and sometimes even in form, as they have allometric growth. It is common to find male ciids with either conspicuous or weak secondary characteristics coexisting in a single population. This phenotypic plasticity may hamper the delimitation and identification of species, and may lead to the proposition of synonymies.
As Enn. Bilamellatum was described from Madagascar, Enn. Tabelliferum from South Africa and Enn. lamellatum from Brazil ( Fig. 1), the synonymization of these with Cer. cucullatus led it to be considered a broadly distributed species. The records of populations in areas beyond its native range are old and have accumulated over the years, attracting the attention of coleopterists. Populations have been found in Italy, France and Britain [3,4,5], although they appear not to be free-living in these countries (note that another invasive species cited for Britain [5] is Cis bilamellatus Wood, a very distinct species with almost the same specific epithet of Ennearthron bilamellatum). Various records of populations in islands are also known, such as in Galapagos, in the Pacific Ocean, and Reunion, Mauritius, Seychelles and Aldabra, in the western Indian Ocean [2,6,7]. As most species of Ceracis inhabits the Neotropical region, and given the absence of other Ceracis species in Africa, non-Neotropical populations of Cer. cucullatus have been interpreted as introduced species [2,7].
The most reliable way to distinguish ciid species is combining traditional information of the external morphology with that of male abdominal terminalia. High morphological stability and specificity of male abdominal terminalia is observed not only in ciids, but in many other animals, especially arthropods [8] and it is largely explained by its rapid evolutionary divergence among phylogenetically related species [9]. For this reason, the taxonomic importance and use of the morphology of abdominal terminalia, mainly of genitalia, is widespread in animal taxonomy. Despite this, in the taxonomy of subtropical and tropical Ciidae the use of male abdominal terminalia to distinguish species was consolidated mostly over the last decade, during which several new ciid species have had their descriptions largely supported by the morphology of this structure [10,11,12,13,14,15]. Although the disjunct populations named Cer. cucullatus have been interpreted as a cohesive taxonomic unity, our preliminary comparisons between African and Neotropical specimens revealed inconsistencies in their external morphology, including male abdominal terminalia, leading us to doubt on the conspecificity of populations under this name.
In the present study we aim to evaluate the status of disjunct populations under the name Cer. cucullatus. Thus, we analysed and compared the external morphology, including male abdominal terminalia, of populations from several localities of the Neotropical, Palearctic, Afrotropical, Afrotemperate and Oriental regions. As results, we reverse the synonymies of Enn. bilamellatum, Enn. Lamellatum and Enn. Tabelliferum with Cer. cucullatus, propose Cer. lamellatus (Pic) and Cer. tabellifer (Mellié), both reinstated status and new combinations, as separate species and propose Enn. bilamellatum as a new synonym of Cer. tabellifer. We also redescribe Cer. cucullatus, Cer. lamellatus and Cer. tabellifer. Additionally, based on data of host fungi and historical records of Cer. tabellifer, we conduct a wide discussion on its successful invasion in Africa and other regions.

Comparisons
We examined and compared the external morphology of a representative number of specimens originally named Cer. cucullatus ( Fig. 1) from localities in the Neotropical, Palearctic, Oriental, Afrotropical and Afrotemperate regions, including a number of islands of the western Indian Ocean. We also extracted and dissected the male abdominal terminalia of representative specimens from these regions (Fig. 2) and carefully compared their morphology. Comparison, examination, measurement and dissection of specimens were made under a Zeiss Stemi 2000-C stereomicroscope or a Zeiss Axiolab microscope. Terms used here for external morphology, including male abdominal terminalia, are explained and discussed by Lopes-Andrade and Lawrence [16].

Pictorial documentation
Digital photographs of adult specimens were taken with a Canon EOS 1000D digital camera attached to a Zeiss Stemi 2000-C stereomicroscope. Photographs showing the prosternum were taken under a Zeiss Axiolab A1 compound microscope equipped with the same camera. Final images were the result of joining 20 to 50 photomicrographs at different focal depths using the image stacking software Zerene Stacker (v1.04). The names Cer. cucullatus, Cer. lamellatus and Cer. tabellifer used from now on refer to these species in the senses and combinations here proposed ( Fig. 1), unless otherwise specified. For the sake of organization, we show images of specimens of the type series and labels of the three species consecutively (Figs. 3, 4, 5), a plate comparing sclerites of their male abdominal terminalia (Fig. 6), a comparison of development degree of pronotal projections in Cer. tabellifer (Fig. 7) and variation of morphology of the prosternum in the species (Fig. 8). Images and redescriptions of Cer. cucullatus and Cer. tabellifer are based on the respective male lectotype, here designated, and those of Cer. lamellatus on a male paralectotype. The specimen chosen as lectotype of Cer. lamellatus, here designated, was examined but could not be borrowed and was therefore not pictured. The syntypes of species treated in the present work were all labelled as lectotypes or paralectotypes by John F. Lawrence, but they were not officially designated in the literature. We preferred to maintain Lawrence's labels. Whole mount preparations followed the protocol described by Lopes-Andrade [14], and photographs were taken under a Zeiss Axiolab A1 compound microscope equipped with a Zeiss Axiocam Erc5S or a Canon EOS 1000D digital camera. Measurements Specimens of Cer. cucullatus (82 specimens), Cer. lamellatus (17) and Cer. tabellifer (90) were measured. The values provided for Cer. cucullatus are of the lectotype and from populations collected in Brazil (45), Panama (18) and Galapagos (18). For Cer. tabellifer, we measured the lectotype and specimens from Brazil (10), South Africa (26), Congo (40) and Gambia (13). Values for Cer. lamellatus are of paralectotypes (2) and topotypes (15). Topotypes, as used here, are specimens collected in the type locality but not labelled as paratypes [17]. Abbreviations used for measurements (in millimetres) and ratios are provided in Table 1. Range, mean and standard deviation are given for each measurement and ratio in the section on ''Variation''.

Maps and geographic distribution
Aiming to trace the geographical distribution of Cer. cucullatus, Cer. lamellatus and Cer. tabellifer (Fig. 9), we searched for records of these species by direct observation of labels in museum specimens and throughout literature. Dates for each record on the map were obtained from specimens' labels. For records obtained from scientific literature, we assigned the date of publication of the work providing the information. We estimated latitude and longitude coordinates by tracking localities in the online database Geo-Names [18] and plotted them in a map using the software ArcGis 9.3 (ESRI, Redlands, CA, USA).
Here, invasive species refers to non-native species that establish populations and spread widely beyond the site of initial introduction [19,20,21]. This is not necessarily associated to environmental impact.

Material examined
Most examined specimens were obtained in loans authorized by the following researchers and respective institutions: Bert Viklund

Taxonomy
Based on patterns of the external morphology of adults, including male abdominal terminalia, here we (i) reinstate both Enn. tabelliferum and Enn. lamellatum as separate species, and propose the new combinations Cer. tabellifer and Cer. lamellatus, (ii) reverse the synonymy of Enn. bilamellatum with Cer. cucullatus and (iii) propose Enn. bilamellatum as a new synonym of Cer. tabellifer, as the populations from continental Africa (Cer. tabellifer) and Madagascar (described as Enn. bilamellatum) do not have consistent morphological differences. Therefore, the cucullatus species-group now includes the following six species: N Ceracis bicornis (Mellié, 1849).  Although specimens of Cer. cucullatus, Cer. lamellatus and Cer. tabellifer are very similar morphologically, we consider the differences in the morphology of their male abdominal terminalia, shown here for the first time, as sufficient to allow their recognition as separate species. Ceracis tabellifer has a comparatively smaller tegmen (Figs. 6I-J) with basal portion subtriangular or rounded and basolateral margins thick (Fig. 6J, large arrows). Additionally, its penis has the basal portion membranous (Figs. 6I-J, vertical small arrows) and apical portion forming a well sclerotized arc (Figs. 6I-J, horizontal small arrows). In Cer. cucullatus and Cer. lamellatus, the basal portion of tegmen is narrowed (Figs. 6B, F, large arrows), as well as the basolateral margins in Cer. cucullatus (Fig. 6B, medium-sized arrows). Both species are similar in the basal portion of penis (Figs. 6B, F, small arrows), but the apical portion is different. The penis in Cer. cucullatus is narrowed along the apical one-fourth of its length (Fig. 6A), and in Cer. lamellatus it is narrowed at the apical third (Fig. 6E). Moreover, in Cer. cucullatus the eighth sternite has the anterior margin distinctly produced at Two Invaders Instead of One PLOS ONE | www.plosone.org middle (Fig. 6D, arrow), while it is weakly projected or almost straight in Cer. lamellatus (Fig. 6H). The basal piece also differs between species. The one of Cer. tabellifer has round angles (Fig. 6K, arrows) and is smaller than those of Cer. cucullatus (Fig. 6C) and Cer. lamellatus (Fig. 6G).
The redescriptions of Cer. cucullatus, Cer. lamellatus and Cer. tabellifer are provided below, as well as an identification key for adult males of Ceracis of the cucullatus species-group.
Ceracis cucullatus (Mellié, 1849), new sense  (Fig. 6B, small arrows); lateral margins narrow, subparallel at the basal three-fourths, abruptly converging at the beginning of the apical one-fourth (Fig. 6A, arrows); apical portion weakly sclerotized. Basal piece (Fig. 6C) elongate, subtriangular, nearly one-third the length of the tegmen and 1.5 times as long as wide. Eighth sternite (Fig. 6D) with anterior margin distinctly produced forward and angulate at middle, its apex conspicuously beyond the anterolateral angles (Fig. 6D Body elongate, subcylindrical; dorsal and ventral surfaces mostly dark reddish brown; appendices yellowish brown. Head barely visible from above; dorsal surface smooth, with a small salience at middle; frontoclypeal ridge produced forward, transversely concave, its anterior margin weakly emarginated at middle, the anterior edge with a row of setae along it. Eyes coarsely facetted, with minute slender yellowish setae emerging from the intersection between ommatidia. Each antenna with nine antennomeres; length of antennomeres (in mm) as follows (from base to apex): Pronotum with subparallel sides; lateral margins narrow, visible from above only for the posterior corners; anterior edge projected forwards forming a quadrangular plate, slightly emarginated at middle; anterolateral angles obtuse, not produced; disc with relatively fine, single, uniformly distributed punctation; interstices in between punctures from one to two puncture-widths; vestiture consisting of yellowish decumbent minute seta. Scutellum small, subtriangular, glabrous; BW 0.08 mm, SL 0.05 mm. Elytra with humeral calli; lateral margins subparallel at the basal two-thirds, then abruptly converging to the apex; only the anterior angles visible from above; punctation single, very fine, confused, denser than pronotal punctation; vestiture consisting of minute decumbent yellowish setae; interstices in between punctures smooth and shiny, shallowly rugose. Ventral sclerites with most of their surfaces granulate. Prosternum in front of coxae shallowly concave longitudinally and transversely convex; surface beside coxae weakly concave; prosternal process laminate, almost as long as coxae. Metaventrite moderately convex, subglabrous, with sparse slender setae; punctation not observed; discrimen indiscernible. Each protibia with the apex expanded; outer apical angle rounded and bearing a row of spines. Abdominal ventrites bearing several slender setae, longer than those on dorsal surface; punctation shallow and sparse; lengths of abdominal ventrites (from base to apex, at the longitudinal midline) as follows (in mm): 0.19, 0.07, 0.08, 0.08, 0.07; first abdominal ventrite with a basal width of 0.48 mm and bearing a circular sex patch located postered of center, with a transverse diameter of 0.04 mm. Male abdominal terminalia (of specimens compared to the type; Figs 6A-D) as follows: Tegmen (Figs 6A-B) three to four times as long as wide, twice as wide as penis, and   (Fig. 6B, medium-sized arrows); lateral margins subparallel at the basal two-thirds of its length, and then slightly angulate at the beginning of the apical third; apex of each lateral margin narrow, with a few sensilla. Penis (Figs 6A-B) elongate, subcylindrical; basal portion opened ( = not membranous), leaving the lower limits of the lateral margins disconnected (Fig. 6B, small arrows); lateral margins narrow, subparallel at the basal threefourths, abruptly converging at the beginning of the apical onefourth (Fig. 6A, arrows); apical portion narrow and weakly sclerotized. Basal piece (Fig. 6C) subtriangular, elongate, nearly one-third the length of the tegmen and 1.5 times as long as wide. Eighth sternite (Fig. 6D) with posterior margin slightly emarginated at middle; posterior corners rounded, bearing bristles; lateral margins diverging from the posterior to the anterior portion; anterior margin distinctly produced and angulate at middle, its apex conspicuously beyond the anterolateral angles (Fig. 6D Frontoclypeal ridge and anterior edge of pronotum varying from strongly projected in large males to weakly or not produced in small ones. In the last case, the frontoclypeal ridge and anterior margins of pronotum are rounded, making the smallest males morphologically similar to females. In males from Midwest Brazil, the pronotal projection is usually projected forward, not upward. The body size of specimens from northern South America and Galapagos Islands is greater than those of Midwest Brazil. Size difference is also noticeable in male abdominal terminalia, which is longer in insular populations. Host fungi. The following records of host fungi were taken from labels, mainly from specimens collected in Panama, and updated consulting the online database of Index Fungorum (http://www.indexfungorum.org): Ganoderma applanatum (Ganodermataceae), Rigidoporus sp. (Meripilaceae), Gloeoporus thelephoroides (Meruliaceae), Coriolopsis caperata, Earliella scabrosa, Hexagonia hydnoides, Pycnoporus sanguineus and Trichaptum sector (Polyporaceae).
Distribution (Fig. 9). From southern Mexico to Midwest Brazil. Also in the Galapagos Islands (possibly introduced).
Type material.  [handwritten].  (Fig. 6F, small arrows); lateral margins narrow, subparallel at the basal one third, converging gradually along the second third  of its length, and abruptly converging at the beginning of the apical third (Fig. 6E, arrows). Basal piece (Fig. 6G) subtriangular, nearly one-third the length of the tegmen and about as wide as long. Eighth sternite (Fig. 6H)  appendices yellowish brown. Head barely visible from above; dorsal surface smooth, with a small salience at middle; frontoclypeal ridge produced forward, transversely concave, its anterior margin weakly emarginated at middle, the anterior edge with a row of setae along it. Eyes coarsely facetted; some minute slender yellowish setae emerging from the intersection between ommatidia. Each antenna with nine antennomeres; length of antenno-meres (in mm) as follows (from base to apex): 0.06, 0.05, 0.04, 0.03, 0.03, 0.03, 0.05, 0.05, 0.06 (right antenna measured; FL 0.13 mm, CL 0.16 mm, CL/FL 1.23); each antennomere of the club bearing several sparse slender setae and four conspicuous sensillifers positioned at its upper portion. Pronotum with subparallel sides; lateral margins narrow, visible from above only for the anterior corners; anterior edge projected forwards forming a quadrangular plate, slightly emarginated at middle (see the section on ''Variation''); anterolateral angles obtuse, not produced; disc with relatively fine, single, uniformly distributed punctation; interstices in between punctures from 1.50 to three puncturewidths; vestiture consisting of yellowish decumbent minute seta. Scutellum small, triangular, glabrous; BW 0.13 mm, SL 0.08 mm. Elytra with humeral calli; lateral margins subparallel at the basal half, then gradually converging to the apex; only the anterior angles visible from above; punctation single, very fine, confused, denser than pronotal punctation; vestiture consisting of minute decumbent yellowish setae; interstices in between punc- tures smooth and shiny, shallowly rugose. Ventral sclerites with most of their surfaces granulate. Prosternum in front of coxae shallowly concave longitudinally and transversely convex; surface beside coxae weakly concave; prosternal process laminate, almost as long as coxae. Metaventrite moderately convex, subglabrous, with sparse slender setae; punctation indiscernible; discrimen not visible. Each protibia with the apex expanded; outer apical angle rounded and bearing a row of spines. Abdominal ventrites bearing several slender setae, longer than those on dorsal surface; punctation indiscernible; length of the abdominal ventrites (from base to apex, at the longitudinal midline) as follows (in mm): 0.20, 0.06, 0.06, 0.06, 0.09; first abdominal ventrite with a basal width of 0.56 mm and bearing a circular sex patch located postered of center, with a transverse diameter of 0.04 mm. Male abdominal terminalia (of specimens compared to the paralectotype; Figs 6E-H) as follows: Tegmen (Figs 6E-F) three to about four times as long as wide, about 1.6 to twice as wide as penis, and about 1.3 to 1.4 times as long as the greatest width of the eighth sternite; basal portion narrowed (Fig. 6F, large arrow); lateral margins converging gradually to apex; apex of each lateral margin narrow, with a few sensilla. Penis (Figs. 6E-F) elongate; basal portion opened ( = not membranous) (Fig. 6F, small arrows); lateral margins narrow, subparallel at the basal one third, converging gradually along the second third of its length, and abruptly converging at the beginning of the apical third (Fig 6E,  arrows); apical portion narrowed and moderately sclerotized. Basal piece (Fig. 6G) subtriangular, nearly one-third the length of the tegmen and about as wide as long. Eighth sternite (Fig. 6H) with posterior margin barely emarginated at middle; posterior corners somewhat angulate, bearing bristles; lateral margins diverging from the posterior to the anterior portion; anterior margin weakly projected at middle, almost straight. Host fungi. Unknown. Distribution (Fig. 9) 6I-L, 7, 8, 9).
Lectotype ( Body elongate, reasonably robust, subcylindrical; pronotum and ventral surfaces mostly dark reddish brown; elytra reddish brown; appendices yellowish brown. Head barely visible from above; dorsal surface smooth, with a small salience at middle; frontoclypeal ridge produced forward, transversely concave, its anterior margin weakly emarginated at middle, the anterior edge with a row of setae along it. Eyes coarsely facetted; minute slender yellowish setae emerging from the intersection between ommatidia. Each antenna with nine antennomeres; length of antennomeres (in mm) as follows (from base to apex): 0.06, 0.04, 0.03, 0.02, 0.02, 0.01, 0.04, 0.04, 0.05 (left antenna measured; FL 0.08 mm, CL 0.13 mm, CL/FL 1.62); each antennomere of the club bearing several sparse slender setae and four conspicuous sensillifers positioned at its upper portion. Pronotum with subparallel sides; lateral margins narrow, visible from above only for the anterior corners; anterior edge projected forwards forming a plate, slightly emarginated at middle (see the section on ''Variation''); anterolateral angles obtuse, not produced; disc with relatively fine, single, uniformly distributed punctation; interstices in between punctures from 1.25 to two puncture-widths; vestiture consisting of yellowish decumbent minute seta. Scutellum small, subtriangular, punctate, glabrous; BW 0.08 mm, SL 0.04 mm. Elytra with humeral calli; lateral margins subparallel at the basal second third, then abruptly converging to apex; only the anterior angles visible from above; punctation single, very fine, confused, denser than pronotal punctation; vestiture consisting of minute decumbent yellowish setae; interstices in between punctures smooth and shiny, shallowly rugose. Ventral sclerites with most of their surfaces granulate. Prosternum in front of coxae shallowly concave longitudinally and transversely convex; surface beside coxae weakly concave; prosternal process laminate, almost as long as coxae. Metaventrite moderately convex, subglabrous, with sparse slender setae; punctation very shallow and sparse, almost imperceptible; discrimen indiscernible. Each protibia with the apex expanded; outer apical angle rounded and bearing a row of spines. Abdominal ventrites bearing several slender setae, longer than those on dorsal surface; punctation shallow and sparse; length of the abdominal ventrites (from base to apex, at the longitudinal midline) as follows (in mm): 0.17, 0.06, 0.07, 0.08, 0.08; first abdominal ventrite with a basal width of 0.48 mm and bearing a circular sex patch located postered of center, with a transverse diameter of 0.04 mm. Male abdominal terminalia (of specimens compared to the type; Figs. 6I-L) as follows: Tegmen (Figs. 6I-J) about three times as long as wide, about twice as wide as penis and 1.2 times as long as the greatest width of the eighth sternite; basal portion subtriangular; basolateral margins thick (Fig. 6J, large arrows); lateral margins subparallel at their basal two-thirds, angulate at the beginning of the apical third and then slightly curved to apex; apex of each lateral margin converging inwards, bearing several sensilla. Penis (Figs. 6I-J) elongate, subcylindrical; basal portion closed, membranous (Figs 6I-J, vertical small arrows); apical portion forming well sclerotized arcs (Figs. 6I-J, horizontal small arrows). Basal piece (Fig. 6K) small, semicircular or subtriangular and with rounded angles, one-fourth the length of the tegmen and about 1.2 times as wide as long. Eighth sternite (Fig. 6L) with posterior margin slightly emarginated at middle; posterior angles rounded, bearing bristles; lateral margins diverging; anterior margin weakly projected at middle, but apex at the same height of the anterolateral angles.
Variation Body varying from dark brown to dark reddish brown, sometimes almost black. Frontoclypeal ridge and anterior edge of pronotum strongly projected in the largest males and weakly or not produced in the smallest ones (Fig. 7). In the smallest males, the frontoclypeal ridge and anterior margins of pronotum are rounded, making them morphologically similar to females. Specimens from invasive populations usually have body and male abdominal terminalia larger than the ones from autochthonous Neotropical populations. However, despite also having a longer male abdominal terminalia, some specimens that inhabit the tropical African savannah have a body size similar to the Neotropical ones. We have also identified two curious morphological variations in specimens from Hanoi, Vietnam: the lack of prosternal process ( Host fungi. Neser [22] provided a fantastic dataset on the South African Ciidae, their host fungi and parasitoids. In that work, a total of 19 Ciidae species were reported from South Africa: 19 Lateral margins of pronotum subparallel. First abdominal ventrite bearing a circular margined sex patch. Tegmen with apex devoid of excavations...2. 2 (19) Apex of pronotum projected forward, with its anterior margin deeply emarginate forming two diverging horns, which may be circular in cross-section and narrow at apex. From Mexico to southern Brazil... Cer. bicornis (Mellié).
49 Tegmen with basal portion narrowed (Figs 6B, F, large arrows). Penis with basal portion opened, leaving the lower limits of the lateral margins disconnected (Figs 6B, F, small arrows); apical portion narrowed and weakly sclerotized, not forming an arc (Figs 6A, E, arrows). Basal piece comparatively larger, subtriangular...5. 5 (49) Eighth sternite with anterior margin biconcave, producing a remarkable subtriangular protuberance that extends beyond the lower limits of the lateral margins (Fig. 6D, arrow). Tegmen with lateral margins subparallel at the basal two-thirds of its length, and then slightly angulate at the beginning of the apical third (Figs 6A-B). Penis with lateral margins subparallel at the basal three-fourths and abruptly converging at the beginning of the apical one-fourth (Fig. 6A, arrows). Basal piece elongate, nearly 1.5 times as long as wide (Fig. 6C). Galapagos Islands and from Mexico to southern Brazil... Cer. cucullatus (Mellié).
59 Eighth sternite with anterior margin weakly projected at middle, barely biconcave to almost straight (Fig. 6H). Tegmen with lateral margins converging gradually to apex (Figs 6E-F). Penis with lateral margins subparallel at the basal one-third, converging gradually along the second-third of its length, and abruptly converging at the beginning of the apical third (Fig. 6E,  arrows). Basal piece about as wide as long (Fig. 6G). Known only from Nova Teutônia, southern Brazil... Cer. lamellatus (Pic).

Discussion
The amazing morphological similarity between Cer. cucullatus, Cer. lamellatus, Cer. navarretei and Cer. tabellifer challenges the skills of any taxonomist. They are not only very similar, but there is a high intraspecific morphological variability, notably in secondary sexual characteristics expressed in males, as the pronotal projections (Fig. 7). The confused taxonomic history involving Cer. cucullatus and species previously synonymized with it appears to be a reflex of such scenario. And it reinforces the potential and relevant role of the comparative morphology of male abdominal terminalia for solving taxonomic problems between morphologically similar Ciidae species.
Has Ceracis tabellifer a Neotropical or an Afrotropical origin?
Ceracis tabellifer was described in 1849 based on specimens from the Cape of Good Hope, at the southern tip of the African continent [1]. Populations may be collected throughout subsaharan Africa and we report here dozens of historical records outside the neotropics and a few records from South America (Fig. 9). At first, and considering that the Neotropical specimens is not quite typical in form compared to African populations, one might suggest that Cer. tabellifer is an autochthonous Afrotropical species. However, considering the absence of any other Ceracis on Africa and the great diversity of the genus in the New World, including all the species morphologically similar to Cer. tabellifer, we support the alternative explanation to the remarkable presence of Cer. tabellifer in subsaharan Africa and other regions: a successful invasion from the neotropics.
Among the six species that currently comprises the cucullatus group, three are known to be polyphagous and have a wide geographic distribution: Cer. bicornis, Cer. cucullatus and Cer. tabellifer. The latter two are successful invaders and individuals of their invasive populations have the largest body size and male abdominal terminalia (see sections on ''Variation'' of Cer. cucullatus and Cer. tabellifer), and are more frequently found than their conspecific native populations. By responses to selective pressures in new habitats or due to stochastic events [20,24,25], phenotypic variation between native and invasive populations is quite common and usually mentioned in works on biological invasion [26,27,28]. Species introduced into new habitats often face foundation events, which entails in genetically based shifts in phenotypic traits, among other effects [20,29]. For instance, it is possible that the reduction of the number of antennomeres and drastic reduction (or even loss) of the prosternal process in the population from Hanoi, Vietnam (Fig. 8B, arrow) were a consequence of foundation effect and genetic drift. These specimens were collected almost a century ago and they are possibly a subsample of the non-indigenous fauna of Cer. tabellifer that occupies the African mainland. Therefore, the curious lack of prosternal process seems to be an adverse effect of inbreeding. To our knowledge, this is the first record of a ciid without that structure.
Considering (i) that the greatest morphological variability of species in the cucullatus group is found in the neotropics, (ii) the easternmost record of Cer. tabellifer (Vietnam) is a morphologically impoverished population, so we assume it might be a subsample of an African population and (iii) the absence of other Ceracis species in Africa, we sustain here that Cer. tabellifer is a Neotropical species introduced elsewhere, and not the inverse. Our arguments are strong enough to sustain it as the most robust explanation at the moment.

Ceracis tabellifer and the general concept of species as metapopulation lineages
In the general concept of species, as it was treated and reinterpreted by de Queiroz [30] with the goal of creating a unified concept, species may be considered as separately evolving metapopulation lineages. According to de Queiroz [30] ''metapopulation lineages do not have to be phenetically distinguishable, or diagnosable, or monophyletic, or reproductively isolated, or ecologically divergent, to be species. They only have to be evolving separately from other such lineages''. This concept has serious implications to the taxonomy of invasive species, which frequently have disjunct distributions, as is the case of Cer. tabellifer. It shall be understood and applied with care. Adopting it without careful prior morphological and molecular analyses would lead to the separation of several invasive species into two or more species, which may be unviable and, more importantly, incorrect. In the case of our study, geographic isolation does not necessarily means that the Neotropical and Afrotropical populations of Cer. tabellifer, for instance, are evolving separately. We judge it is early to separate disjunct populations of Cer. tabellifer in different species, even considering the few morphological differences of the South American populations. Further studies incorporating molecular techniques may throw light on this issue.

Spread and current distribution of Ceracis tabellifer outside the Neotropical region
During the nineteenth and twentieth century, the collection of Cer. tabellifer varied substantially over time and space (Fig. 9). For this reason, they do not depict the chronological advance of populations on the space, but merely a variation in collection effort. Nevertheless, the historical records we provide here may be informative and reveal how old is the colonization of non-Neotropical areas by Cer. tabellifer.
The first Cer. tabellifer recorded in African lands date back to the beginning of the nineteenth century (Fig. 9, red circles), although it is possible that its arrival occurred in the previous century or even before. Until the first half of the twentieth century, populations had been recorded at various localities from the Democratic Republic of Congo and neighbouring countries. Currently Cer. tabellifer is widely distributed in subsaharan Africa and possibly one of the most frequently collected ciid species in the continent (Fig. 9). Several islands at the western Indian Ocean are also inhabited by Cer. tabellifer [2,7,10] and its presence in Sri Lanka and India suggests a possible route of introduction through these islands. Considering it has reached Vietnam in the first half of the twentieth century, if not before, it would not be a surprise to find the species at the remaining Oriental region, or even at the Australotropical, Neoguinean and Neozelandic regions. In con-trast to the success of this species in tropical zones and marginally tropical areas, there are few records of Cer. tabellifer in temperate zones. Each record in Britain and France were noted as introductions, but populations are not free-living in there [3,31], and that is possibly the case for the records in Italy either. The supposed record from Britain was, indeed, in a fungus from Zambia [31]. Low temperatures may be an impediment to successful establishment of populations in the Palaearctic region. Besides Cer. tabellifer, the morphologically related and also polyphagous Cer. cucullatus and Cer. bicornis are restricted to tropical lands too.

Causes for success and potential threat
Examining the invasiveness of a species and predicting whether it would succeed in new habitats is still a rather complex task [32,33,34,35]. However, comparisons between autochthonous and invasive species in a given ecosystem may be a way to identify features determining invasion success [19]. Despite information on life history and reproductive behaviour of tropical ciids are scarce, we hypothesized below some characteristics that may underlay, at least partially, the successful establishment and spread of Cer. tabellifer during the invasion process, and that continue allowing its dissemination.
Most ciids are mycetobionts, with some species exploiting several host fungi, while others are restricted to a few or even one host fungus [36]. We show that Cer. tabellifer is a remarkably polyphagous species in South Africa and in two surveyed localities of the Brazilian Cerrado. The available data suggest that this ciid is able to exploit more host species than native ciids in South Africa, and even more than other species of Ciidae as a whole. For comparison, see data available for host fungi of Nearctic ciids [37,38], Neotropical [23], and for ciids from Britain, Germany, North America and Japan [39]. It is noteworthy that polyphagy is common among widely distributed species of Ciidae, as in Cer. bicornis [23], which also belongs to the cucullatus species-group, and other ciids such as Cer. thoracicornis (Ziegler) [37], Cis bilamellatus Wood [5] and Cis creberrimus Mellié [37], the latter two introduced in Britain [5] and Galapagos Islands [40], respectively. The diet breadth has been seen as a characteristic involved in the invasion success of animals [41], including beetles [5,42,43]. Polyphagy has probably contributed to the invasion success of Cer. tabellifer.
Ceracis tabellifer is a multivoltine species (as well as Cer. cucullatus) and may be easily collected throughout the year in South Africa, where it is the most frequently collected ciid species within its broad host range. Populations have been encountered from areas near the coast to heights of up to 1700 meters, and from urban areas and cultivated lands to endemic forest and savannah [7,22]. The polyphagous feeding habit and the wide range of habitats where Cer. tabellifer is able to self-sustaining suggest a strong adaptive potential. In the Cerrado biome of Brazil, which we consider to be its native habitat, Cer. tabellifer was the only species of Ciidae found in all surveyed types of vegetation, from open areas to riparian forests inside the biome (Lopes-Andrade, pers. obs.). Such amplitude of supported conditions is maintained in invasive populations. Allied to it, the notable occurrence of populations in islands separated by long stretches of ocean indicate the high ability to transpose long distances, either by active migration or through human intervention. However, it is still unclear which are the actual adverse effects derived from the outstanding presence of Cer. tabellifer in invaded habitats, or even if it offers any threat to autochthonous species.

Other introduced or invasive ciids
Other invasive ciid species have been recorded in several regions. The Autralasian fungivore Cis bilamellatus was accidently introduced in England during the nineteenth century and is currently distributed across England, Wales and southern Scotland, with further records from Ireland, Channel Island and France [5]. Cis creberrimus, widespread in the neotropics, is considered an introduced species in Galapagos Islands. Together with Cer. cucullatus, they are the unique ciid species that occupy these islands [40]. Interestingly, in the population of Cer. cucullatus we reared in laboratory, we have observed contamination by Cis creberrimus in a few basidiomes. These species coexisted in the same basidiome for a long time, where they appear to have partitioned the resource temporally. We noted that Cer. cucullatus had greater activity in the morning and afternoon, while Cis creberrimus was more active in the late afternoon. Moreover, we observed that Cis creberrimus is able to exploit the fungus in more advanced stages of decomposition. Temporal partitioning of a host fungus by ciids was already observed [36] and would not be a surprise to report this behaviour in the wild. Cis fuscipes Mellié, an abundant and widely distributed ciid in North America, has been interpreted as an invasive species in Cuba, Hawaii, Madeira, Australia and New Zealand, with several parthenogenetic populations [44,45]. Hadreule elongatula (Gyllenhal) is recorded throughout the Old World and northern Africa, and the presence of this minute beetle in North America has been cited as an invasion from Europe [38,40], although it is still uncertain. Cis chinensis Lawrence was described based on specimens found in commercial dried fungi exported to USA [46,47]. Nowadays this species is free-living in Europe (Germany, Italy, France and Malta), China and Thailand (possibly autochthonous in eastern Asia), southeastern Brazil and Reunion, in the western Indian Ocean [10,48]. Other species are possibly also invasive, as Ennearthron victori Lopes-Andrade & Zacaro, the unique Ennearthron species found in the South Hemisphere.

Conclusion
We refute an ancient scenario in which the broadly distributed populations under the name Cer. cucullatus were interpreted as a cohesive taxonomic unity, and demonstrate that this ciid in fact hid for a long time two additional similar species, which we have reinstated: Cer. lamellatus and Cer. tabellifer, the latter being the true invasive Ceracis in Africa. Due to the morphological similarity, Cer. cucullatus and Cer. tabellifer were historically confused in the literature since they were described [6,7], and even the descriptor was ''deceived'' by the similarity of these small beetles (see ''Introduction''). We show that the morphology of male abdominal terminalia is the most effective way to distinguish these species.
Since the first record of Cer. tabellifer in Africa, and based on the number of records in the continent and on information drawn from recent field collections in South Africa, it becomes evident that this ciid overcame the sequential transitions of the invasion process successfully [19], at least in African continental lands and islands of the western Indian Ocean. However, although its biological characteristics seem to confer a great competitive ability and suggest a potential threat to native fauna, it is yet unclear whether the remarkably presence of Cer. tabellifer in Africa implies any ecological impact. Due to its potential to colonize new habitats and ability to transpose long distances, it is possible that Cer. tabellifer will spread across tropical lands, but be limited by the low temperatures of temperate zones.
Future researches should evaluate the relatedness of invasive and native populations of Cer. tabellifer through molecular approaches. Ceracis tabellifer offer an exciting opportunity to study the effects of a non-pest mycetobiont organism to native communities. Our study helps to fulfil a gap in the literature on biological invasions, with considerably more studies on predatory species, disease vectors or potential pests of agricultural crops, than on non-pest fungivorous organisms.