ANNALES ZOOLOGICI (Warszawa), 2000, 50(4): 449-558 MORPHOLOGY, PHYLOGENY AND CLASSIFICATION OF ADULT ENDOMYCHIDAE (COLEOPTERA: CUCUJOIDEA) K. WIOLETTA TOMASZEWSKA Muzeum i Instytut Zoologii PAN; Wilcza 64, 00-679 Warszawa, Poland; e-mail: [email protected] Abstract. The cucujoid family Endomychidae is characterized. Detailed morphology of 36 exemplar genera representing 12 recently recognized subfamilies is presented. Cladistic analy- sis was accomplished based on 50 taxa (47 genera from Endomychidae and Coccinellidae the ingroup, and 3 genera from Helotidae, Brachypteridae and Protocucujidae the outgroups) coded for 40 adult, morphological characters. Based on the analysis, a phylogeny of the family is hypothesized. Results of the analysis suggest that the Endomychidae form a monophyletic taxon, which may be divided into 12 subfamilies: Danascelinae (subfam. nov.), Xenomycetinae, Endomychinae, Anamorphinae, Merophysiinae, Lycoperdininae, Stenotarsinae, Epipocinae, Eupsilobiinae, Pleganophorinae, Mycetaeinae and Leiestinae. Short diagnoses for the family and subfamilies, as well as a key to the subfamilies are provided. The following two new, sub- family synonyms are proposed: Anamorphinae Strohecker, 1953 (=Acritosomatinae Pakaluk and lipiñski, 1995) and Merophysiinae Seidlitz, 1872 (=Holoparamecinae Seidlitz, 1888). Adamia gen. nov. (type species: A. mexicana sp. nov.) and Erotendomychus lawrencei sp. nov., are described. Type species are designated for the following genera: Niteta Weise, 1890 (N. quatuordecimpunctata Weise, 1890), Oediarthrus Gerstaecker, 1858 (O. natalensis Gerstaecker, 1858) and Stenotarsoides Csiki, 1900 (S. quadrimaculatus Csiki, 1900). History of classification and known facts concerning the biology of the family are provided. Ë Key words. Coleoptera, Cucujoidea, Endomychidae, genera, morphology, phylogeny, clas- sification. TABLE OF CONTENTS Introductions . 450 1. History of classification . 450 2. Systematic position . 450 2.1. The cerylonid-group of the Cucujoidea . 450 2.2. The Endomychidae and their relatives relations with the family Coccinellidae . 450 3. Larvae of the Endomychidae their diversity and similarities . 451 4. The recent classification . 452 5. Biology . 452 5.1. Feeding habits and habitats . 452 5.2. Behaviour . 452 Material, methods and terminology . 453 Family Endomychidae . 453 1. Adult characteristics . 453 2. Morphology of investigated genera . 455 Phylogenetic analysis . 489 1. Methods . 489 2. Taxa examined for analysis . 489 3. Characters, discussion and polarity decisions . 489 4. Results . 493 Proposed classification of Endomychidae . 494 1. Classification of suprageneric taxa . 494 2. Key to the subfamilies of Endomychidae . 494 Definitions of the family and subfamilies of Endomychidae . 494 Acknowledgements . 496 References . 496 Index of generic names of Endomychidae . 499 Illustrations and cladograms . 500 450 K. W. TOMASZEWSKA INTRODUCTION aedeagus resting on the side when retracted, which is characteristic for cerylonid-group, also occurs in several 1. History of classification other cucujoid taxa (e.g. Sphindidae, Protocucujidae, Boganidae, Erotylidae and Languriidae). The first described genus of this family was Although long established as a group and generally Endomychus, established by Panzer (1795) for the considered monophyletic (although without clear rela- Linnaean species Chrysomela coccinea, described in tionships to any other group of Cucujoidea), the above 1758. A few additional genera were proposed by various exceptions to diagnostic characters and the lack of rig- authors during the early nineteenth century, but most of orous family definitions, which are based exclusively on the endomychid species described in that period were apomorphic characters, show that the relationships of referred to such genera as Galleruca, Erotylus, members within the cerylonid-group remain uncertain. Chrysomela and Tritoma. The family Endomychidae lipiñski and Pakaluk (1992) drew attention to many was proposed by Leach in 1815, for Endomychus coc- problems in classification of the cerylonid group and cineus (Linnaeus), Eumorphus kirbyanus Latreille emphasized that for a better understanding of relation- and Lycoperdina bovistae (Fabricius). ships a critical reappraisal of the taxa and a phylogenet- The first general survey of the family was accom- ic study was desperately needed. plished by Gerstaecker. His monograph of the The aims of the current study, apart from reviewing Endomychidae (1858) had been the basis for study by the classification and biology of the Endomychidae, was subsequent workers, for many years. Very few addition- to make a detailed morphological study of representative al works of broad scope have appeared since then. Most genera followed by cladistic analysis. The aim of the important of these include early catalogues of analysis being the resolution of three problems: the rela- Endomychidae by Gorham (1873a) and Csiki (1910), tionships between Endomychidae and Coccinellidae; the Arrows fauna of British India (1925), and Stroheckers limits of Endomychidae as a monophyletic taxon; and the (1953) generic review and world catalogue, the largest limits and relationships of endomychid suprageneric work on the Endomychidae of the 20th century. taxa, the last in order to provide a more soundly based 2. Systematic position subfamily or tribal classification. 2.2. The Endomychidae and their relatives rela- 2.1. The cerylonid-ggroup of the Cucujoidea. In his tions with the family Coccinellidae. The relationship Natural classification of the families of Coleoptera between Endomychidae and Coccinellidae has always Crowson (1955) placed Endomychidae within the super- been a source of interest for taxonomists. Because of family Cucujoidea. He (Crowson 1955, 1960) proposed the peculiar appearance of the tarsi (pseudotrimerous), the cerylonid-group to include the most derived families both these taxa were often combined in a taxon called of this superfamily as follows: Endomychidae (including Trimera, which had been placed at the end of the Alexiidae=Sphaerosomatidae), Coccinellidae, Corylo- coleopterous system. This unique character, common phidae, Cerylonidae, Discolomatidae, Merophysiidae and for almost all Coccinellidae (except Lithophilinae) and Latridiidae. More recently, Pal and Lawrence (1986) for- higher Endomychidae, is evidently an advanced condi- mally transferred Bothrideridae, as an independent tion, but it is perhaps not a synapomorphy derived from taxon, from Colydiidae (Tenebrionoidea) to Cucujoidea, a common ancestor of both families. Members of the next to Cerylonidae, as one more family of the cerylonid- endomychid subfamily Eupsilobiinae, at least partially group. combine characters historically considered unique for The members of this group are generally character- Endomychidae and Coccinellidae e.g., a characteristic ized by the following characters: adult with tarsi 4-4-4 or median lobe and the loss of a fronto-clypeal suture 3-3-3 in both sexes; wings without a closed radial cell and (Pakaluk and lipiñski 1990, lipiñski and Pakaluk with reduced number of anal veins (if more than one anal 1992). The other characters common for both families, vein, the first runs into a medial fleck); aedeagus resting including 5 pairs of functional abdominal spiracles and on one side when retracted, tegmen strongly reduced, middle coxal cavities open outwardly (both characters median lobe strongly curved, without median struts; lar- seem to be important in the classification of superfami- vae with tarsungulus unisetose, spiracles almost always ly Cucujoidea), are also present in some other families annular, and sensory appendage of second antennal seg- of the cerylonid-group. Although Endomychidae are ment usually as long as the third segment (lipiñski and generally considered closely related to Coccinellidae, Pakaluk 1992). lipiñski and Pakaluk (1992) indicated that There are a number of exceptions to these supposed- Endomychidae are probably polyphyletic, and ly diagnostic characters. Some Bothrideridae have a Coccinellidae have basal clades with uncertain relation- closed radial cell; some Bothrideridae and some ships, some of which may include certain endomychids. Coccinellidae have more than one anal vein and the first Moreover, some of the diagnostic characters for these does not run into the medial fleck; there are a number of two families are probably plesiomorphic and cannot be taxa with several anal veins but without a medial fleck; used to define monophyletic taxa in the sense of cladis- Coccinellidae have the tegmen well-developed; and an tic taxonomy. All of these factors suggest, that the limits MORPHOLOGY, PHYLOGENY AND CLASSIFICATION OF ADULT ENDOMYCHIDAE 451 and relationships between Endomychidae and theca distinguish Leiestinae and Merophysiinae from Coccinellidae, as well as the limits and relationships of members of these two families. endomychid subfamilies, are unclear and should be Too few larvae within each subfamily of Endomy- reconsidered in a broader context using phylogenetic chidae are known and described to be fundamentally methods. instrumental in improving the higher-level classifica- Feeding habits of both families seem to be very inter- tion. Fortunately this situation has been improving esting in the context of mutual relations between these recently with each additional description of a new larva, taxa. While most endomychids are fungivorous in both and with any comparative studies. Such comparative