The acasta conundrum: Polymorphism and taxonomic confusion within the parasitoid genus Melittobia (Hymenoptera: Eulophidae) Author Matthews, Robert W. Author González, Jorge M. text Zootaxa 2008 1854 45 54 journal article 10.5281/zenodo.274436 34ba49f6-55b5-4414-9f26-e491df98bfa7 1175-5326 274436 Melittobia megachilis (Packard, 1864) Anthophorabia megachilis Packard, 1864: 134 . Pteromalus gerardi Hickok, 1875 : 131 . [not confirmed and possibly incorrect – see below] Chrysocharis aeneus Brues, 1909 : 161 . Melittobia femorata Dahms, 1984a : 293 . Syn. n. Support from morphological study. Dahms’ (1984a) female key, couplet 9, separates M. femorata from M. megachilis using three antennal characters: color of scape and pedicel (darkened reddish brown dorsally in M. femorata versus not darkened in M. megachilis ); relative length of the nipple on the terminal club segment (longer in M. femorata ); and the relative position of the nipple seta (more basal in M. megachilis ). The latter two characters cannot be assessed reliably on intact air-dried specimens; however, variability in these antennal characteristics was encompassed by the variability found in our slide-mounted samples of M. femorata females from numerous localities, with minor differences sometimes exaggerated depending upon specimen orientation. Subtle color differences also prove unreliable when larger samples are studied. When we compared our samples on a selected suite of other measurements given by Dahms (1984a) in his species descriptions for female M. megachilis and M. femorata , the variation encompassed that given in Dahms’ descriptions for each species. For example, setae on the forewing submarginal vein numbered between 3 and 5, most commonly 3, but there were always only two setae on each submedian lobe of the scutellum. We found that for both M. femorata and M. megachilis , as well as for Brues’ 13 Chrysocharis aeneus syntypes , female head shape in frontal aspect was relatively narrow (head length to genal width, 1.1– 1.4) with the genal-clypeal margin relatively straight. As noted above, males of M. megachilis were unknown. In his key to Melittobia males, couplet 10, Dahms (1984a) separates M. chalybii and M. femorata from M. evansi and M. scapata on antennal characters. However, our results noted above revealed that the male antennae are rather variable in the first-brood males of M. femorata , thus not very useful to distinguish members of this subgroup. Specifically, male M. femorata that eclosed with the first brood had antennae ( Fig. 2 a–c) virtually identical to those figured by Dahms for M. chalybii (his figs. 203–205) and M. evansi (his figs. 191–193). Second-brood males that emerged after diapause all had antennae ( Fig. 2 d) like those of Dahms’ M. femorata (his figs. 200–202). For a visual example, compare Figures 1 c and 1d, showing two males reared from the same laboratory culture; note the relatively broader wings and first funicular segment of the “ femorata ” morph (c) compared to the “ chalybii ” morph (d). FIGURE 2. Variation in M. megachilis from a single culture. a–c: antennae of different first-brood males, showing range of variation. d: antenna of second-brood male. e: forewing of first-brood male. f: forewing of second-brood male. g: middle leg of first-brood male. h: middle leg of second-brood male. In couplet 12, Dahms (1984a) distinguishes M. chalybii and M. femorata on relative forewing shape and costal cell length, relative form of the mid-femoral fringe, relative density of head setae, and relative head shape. However, our studies showed that all of these varied with eclosion order. For example, the forewings of our first-brood males ( Fig. 2 e) were longer and narrower than wings of second-brood males ( Fig. 2 f). Furthermore, in the first brood of males to eclose, the patterns and relative lengths of the mid-femoral setae ( Fig. 2 g) generally conformed to the variation that Dahms illustrated for M. scapata , M. evansi , and M. chalybii (his figs. 222–223, 226). However, for the second group to eclose, the mid-femoral setal pattern and relative lengths ( Fig. 2 h) matched that illustrated for M. femorata (Dahms’ fig. 225). In his M. chalybii redescription, but not his key, Dahms mentions a tuft of setae on the male fore trochanters that is supposedly present on M. chalybii and M. australica Girault ; in the M. femorata description he notes that this tuft is absent. However, this material is somehow confused, because the redescription for M. australica says the tuft is absent and his table (p. 296) mentions not fore but middle trochanters. Our examination of the fore trochanters of both broods of M. femorata males failed to reveal any evidence of a tuft of setae; such a tuft is also lacking in Ashmead’s eight syntype males and the two slide-mounted specimens upon which the male description was presumably based. Interestingly, in a related paper, Dahms (1984b) briefly addressed the morphological variability introduced by Melittobia polymorphism, particularly in the M. acasta group, and illustrated it with three examples of “second form” male antennae from Washington, District of Columbia and West Virginia (his figs. 11, 17, 18). It is thus ironic that he inadvertently added to the taxonomic confusion by failing to carefully compare his presumed M. chalybii males with those of Ashmead’s type series. In short, M. megachilis males are polymorphic and variable, and many of the characteristics Dahms used to distinguish between males of M. megachilis , M. femorata , M. chalybii , M. scapata , and M. evansi are thus open to question pending better knowledge. The demonstrated variation in midleg and antennal characters for M. megachilis has the result that using Dahms’ key to identify male Melittobia may yield erroneous identifications for species in the acasta group. Further study of males of both broods of each species reared from a variety of different host species will be required in order to determine the range of variability in diagnostic characters and to revise the key. There is a distinct possibility that upon further study, including use of molecular techniques, the acasta group may prove to contain only three species, M. acasta , M. digitata , and a morphologically variable M. megachilis . TABLE 1. Selected biological attributes that differ for three sympatric Melittobia species in the acasta group that attack Trypoxylon politum in the southeastern United States.

Trait	M. acasta	M. digitata	M. megachilis
Lethal male fighting	yes	yes	no
Female tolerance of conspecifics	high	high	low
Oviposition latency after host-feeding	ca.2 days	ca. 2 days	ca. 4–5 days
Summer brood development*	continuous	continuous	discontinuous**
Adult eclosion	synchronous to asynchronous	synchronous	asynchronous
Accept Diptera pupae as hosts	yes	yes	no

*All temperate-zone Melittobia that have been studied diapause during the winter. **Obligatory diapause before MF eclosion. Support from biological study. While M. megachilis is clearly a member of the acasta group, biologically it is unusual ( Table 1 ). Whereas other acasta species develop more or less continuously during the warmer months of the year, for M. megachilis an obligatory diapause separates the emergence of the brachypterous and macropterous broods on the same host ( Matthews et al. 2005 ). Whereas all other acasta group species can develop on Diptera, M. megachilis apparently cannot (unpubl. observ.). Male fights are less intense and rarely lethal, but females sharing a host are highly intolerant of one another, often physically damaging competitors (Matthews & Deyrup 2007). Finally, unlike the loose prepupal distribution that occurs in other acasta group species, diapausing M. megachilis prepupae in laboratory cultures characteristically cluster tightly beneath the dried host remains as though under a blanket (unpubl. observ.). Additional biological and behavioral aspects (under the name M. femorata ) are presented in Matthews et al. (1985) , González et al. (1985), González et al. (1996) , Lapp (1994) , Matthews et al. (2005) , Deyrup & Matthews (2007a), Matthews & Deyrup (2007), and Molumby (1995) . Given how common and widespread M. megachilis is, it seems odd that a major comparative courtship study ( Assem et al. 1982 ) failed to include it; however, that study did include M. evansi (“species 3”). The original population of the latter species was also studied by Evans & Matthews (1976) as M. species A; its courtship duration was reported not to be significantly different from that of M. femorata ( González et al. 1996 ) . Although we are not yet prepared to formally synonymize it, we suspect M. evansi may ultimately prove to be a variant of M. megachilis . The comparative courtship behavior of the two M. megachilis morphs has been reported ( González et al. 1996 ; Lapp 1994 ). González et al. (1996) found courtship duration to be longer for BF courtships. In contrast, Lapp (1994) reported male display in BF courtship was shorter and seemed to truncate some components of that observed in MF pairs.