S. Mahunka szerk.: Folia Entomologica Hungarica 49. (Budapest, 1988)

is almost exactly the opposite: there are "a few situations" that are clear, but our general image of systematic relations within the family remains nebulous. To my mind.COBOS (1980) has rather obscured the picture, complementing the faults of earlier workers (schematism, attributing too much weight to single features: mesosternal cavity, scutellum, sensorial pores of antennae, etc.) with those of his own (strong "splitting" inclination, similarly schematic use of "modern" characters like genitalia or wing venation). In my opinion, there are only four primary (of subfamilial rank) lineages within the Buprestidae Leach: the Schizopodinae Lec., Julodinae Lac., Buprestinae Leach and Agrilinae C.G. (my interpretation of the first two does not differ from that of earlier authors; Agrilinae C.G. include Cylindromorphinae and Trachynae of COBOS 1980 as well as "Galbellinae COBOS" of BELLAMY 1985; all the remaining groups belong to Buprestinae Leach). This is not a proper place for a detailed discussion of the high-level classification of jewel beetles (my aim being only to provide the necessary background for the placement of Maoraxia Obb.), so I confine my argumentation to some general remarks. The Schizopodinae Lec. and Julodinae Lac. are so well defined, and evidently so distinct from each other and from the rest of the family, that their subfamilial status has practically never been questioned (indeed, there were - and still reappear from time to time proposals to separate the former into a family of its own). The differences between Acmaeoderinae Kerr., Polycestinae Lac., Mastogeniinae L. H., Thrincopyginae Lec, Chalcophorinae Lac, Buprestinae Leach, Sphenopterinae Lac. and Chrysobothrinae C.G. (COBOS 1980, BELLAMY 1985) seem much less profound: these groups are defined by few relatively superficial and frequently inconsistent characters (scutellum, sternal cavity, antenna! sensory pores), and giving them equal rank to Schizopodinae Lec. or Julodinae Lac. evidently obscures the actual relationships. Similarly, the removal of the tribe Cylindromorphini Port, from Agrilinae C.G. with the creation of a separate subfamily (COBOS 1980) does not seem justified: the alleged differences in the structure of labrum or prosternai sutures - even if real - are by no means so fundamental as to warrant the separation at the subfamilial level; frequent atrophy of the tibial spurs in Agrilinae C.G. has been admitted by COBOS (1980) himself, the situation in Cylindromorphini Port, being but the extreme manifestation of this tendency; the larva of Paracylindro

morphus transversicollis (Rtt. ), as described by BILY (1983), seems - contrary to the COBOS' (1980) claim - to match rather well the general pattern of Agrilinae C.G. (more cylindrical body with not enlarged prothorax is a relatively simple adaptation to the life in stems of grasses); at last, the peculiarities of the sternal cavity have been obviously strongly exaggerated by the Spanish author. By the way, the structure of sternum, traditionally used - in terms of sternal cavity being formed entirely of mesosternum, of both mesoand metasternum, or of metasternum alone - as the character of primary importance in buprestid classification, is notoriously unreliable in this role. This should be expected on theoretical grounds (at the "cross-roads" of so many sclerites - prosternai process, mesosternum, metasternum, pro- and mesocoxae - the slightest modification of any of them must drastically alter the relations among the others), and indeed, this is exactly what we observe in reality: well-known exceptions from the "rule" are Neojulodis Kerr., Amblysterna Ths., and Belionota Esch., but exceptional is also Trachykele Mars., similar deviations can be also found among Julodis Esch., Acmaeodera Esch, (pers.obs.), or Coomaniella Bourg. (BILY 1974), while Polycestinae Lac. (COBOS 1980), Mastogenius Sol. (MANLEY 1986, cf. also TO YA MA 1984, fig. 1-4), Chrysochroa Sol. (pers.obs.), Coraebini Bed. (cf. BELLAMY 1986b, fig. 25-32), and probably many other groups show a great deal of variation in this character. The case for merging Trachydinae C.G. into Agrilinae C.G. seems less strong:their larvae, tarsal structure, location of coxal cavities, etc., are decidedly different. The similarities seem, however, also much too numerous to be considered simply convergent; even if not based on strict homology, they are obviously the result of parallel evolution rather than convergence, i. e. they represent similar manifestations of still similar - inherited from not too distant common ancestor - genetic background. Some of COBOS'

s (1980) subfamilies are undoubtedly natural taxa and should be considered tribes (Acmaeoderini Kerr., Mastogeniini L. H., Thrincopygini Lec, Chrysobothrini C.G.) or subtribes (Sphenopterina Lac.) within Buprestinae Leach. Others, however, seem either polyphyletic ("Polycestinae"), or at least wrongly defined (including subgroups that