Trjapitzin (1963) established Propelma based on Propelma rohdendorfi (Hymenoptera: Eupelmidae), which he described from a single female in Eocene Baltic amber. Gibson (1995) later synonymized Propelma under Eupelmus Dalman 1820 in a revision of the world genera of Eupelminae Walker 1833. The synonymy was based primarily on the lateral habitus drawing of Propelma rohdendorfi given by Trjapitzin (1963, fig. 1) without examining the holotype.

When Trjapitzin described Propelma he stated that it was similar to Metapelma Westwood 1835 in general habitus and size of the body, head shape, antennal structure, thorax, and presence of a long ovipositor. Metapelma is one of seven genera currently classified in Neanastatinae Kalina 1984, which along with Calosotinae Bouček 1958 and Eupelminae comprise the three recognized subfamilies of Eupelmidae. Of the seven neanastatine genera, four are extant, including Eopelma Gibson 1989, Lambdobregma Gibson 1989, Metapelma and Neanastatus Girault 1913. The other three genera, Aspidopleura, Brevivula and Neanaperiallus, were all described by Gibson (1999) from Baltic amber inclusions and are extinct. Description of the three extinct genera greatly expanded morphological limits of Neanastatinae and the new knowledge led me to re-examine Trjapitzin’s (1963) illustration of Propelma rohdendorfi and question the validity of synonymizing Propelma under Eupelmus. Resulting study of the amber holotype of Propelma rohdendorfi subsequently showed that Trjapitzin (1963) was correct in comparing Propelma with Metapelma relative to subfamilial affinities, and that Propelma represents a fourth, extinct, valid genus of Neanastatinae from Baltic amber. The purpose of this paper is to correct my erroneous synonymy and redescribe and illustrate Propelma rohdendorfi so that its classification is better established in Eupelmidae.

The description and photomacrographs are based on the holotype female of Propelma rohdendorfi [Holotype no. 364/360, Orlov Museum of Paleontology (formerly, Paleontological Institute of Russian Academy of Sciences), Moscow, Russia]. The holotype is in a mostly dark orange-colored block of amber (Fig. 1). It is complete except that an unknown length of the ovipositor sheaths are missing, as is most of the femur and tibia of the left hind leg and the apices of the femora and bases of the tibiae of the right middle and hind legs. The missing parts are because these projected beyond the sides of the polished amber block. The right side of the specimen, in particular, is clearly visible (Fig. 1), but artefacts prevent a direct ventral view of the mesosoma or the dorsal surface of the body beyond about the posterior angles of the axillae, and thickness of the amber prevents clear observation of the face. Images were taken with a Leica DFC 425C, 5 megapixel digital camera attached to a Leica Z16 APO macroscope. Serial images were combined using Zerene Stacker and digitally manipulated using Camera Raw and Adobe Photoshop 4 to enhance clarity. Images taken of the right side of the holotype for the plates of illustrations were flipped so that they face in the normal direction for specimen observation. All images except for Fig. 1, which illustrates color of the amber, are published in greyscale because this better facilitates differentiation of structures.

Terminology follows Gibson (1989, 1995) except terms used for the metanotum follow Heraty et al. (2013). Abbreviations used on the plates to indicate morphological features are: acs = acropleural sulcus; amd = anterolateral mesoscutal depression; ams = anterior, transverse region of mesoscutum; car = carina; cer = cercus; cl_1–3 = first, second, third clavomere; fu_{1, 8} = first, eighth funicular; gsp = gastral spiracle; map = mesotibial apical pegs; mbl = membranous lobe; mdr = depressed triangular region of mesopectus; mlp = lateral panel of metanotum; msp = mesotibial spur; Mt₈ = eighth metasomal tergite; mtp = mesotarsal pegs; mts = mesotarsal setal line; mtsa = metanotal scutellar arm; mtt = metanotal trough; pre = lateral panel of prepectus; psp = propodeal spiracle; ptl = petiole; sc₂h = mesoscutellar hook; sc₃ = metascutellum; sp₂ = mesothoracic spiracle; syn = syntergum (Mt₈ + Mt₉); tgl = tegula. Measurements of the antennomeres, fore wing venation, and metasomal tergites were all taken at the same magnification; measurements between square brackets are repeated from the original description.

Propelma is assigned to Neanastatinae based on pronotal structure, mesopectus posteroventrally abutting the mesocoxal bases without a membranous region anterior to each coxa, and mesotarsal peg pattern, all of which are diagnostic of the subfamily (Gibson 1989, 2009). The scutoscutellar suture is also crenulate similar to extant and extinct species of Metapelma (Gibson 2009, fig. 13) and some species of Lambdobregma (Gibson 2009, fig. 15) (extant) as well as Aspidopleura (Gibson 2009, fig. 49) and Brevivula (Gibson 2009, fig. 39) (extinct), and the scutellum apically is curved ventrally into a small hook-like process over the metanotum similar to extant species of Lambdobregma (Gibson 2009, figs 15, 16) and Neanastatus (Gibson 2009, fig. 21) as well as Brevivula (Gibson 2009, fig. 40).

Propelma keys to couplet 6 (Brevivula and Lambdobregma) using the key to genera of Neanastatinae in Gibson (2009). It differs from both genera in having the fore wing uniformly setose (Fig. 10) rather than with a linea calva (cf. Gibson 2009, figs 6, 42) and by having the prepectus as an isosceles triangle, only about 1.3× as long as high with the dorsal and ventral margins convergent to a narrowly rounded apex (Fig. 8: pre), rather than being conspicuously elongate-triangular and tapered into an acute angle (cf. Gibson 2009, figs 21, 41). It also differs from Lambdobregma by the absence of a transepisternal sulcus (Fig. 8; cf. Gibson 1989, fig. 92), and from Brevivula by having the mesoscutal lateral lobes evenly convex (Figs 6, 7) rather than carinately margined (cf. Gibson 2009, fig. 39).

Trjapitzin (1963) interpreted Mt₈ and Mt₉ as separate tergites in Propelma rohdendorfi, but these are certainly fused into a syntergum (Fig. 13: syn). He also interpreted the bare triangular region at the apex of the syntergum as Mt₁₀. However, comparison with extant taxa suggests that this region is a membranous lobe (Fig. 13: mbl) homologous with the anal filament of some extant chalcidoids with comparatively long ovipositor sheaths (cf. Gibson 1989, figs 148, 153). Trjapitzin (1963) also noted that the malar sulcus appears bifurcate in the holotype (Fig. 4), though this and what appears like a continuous sulcus across the propodeum and metanotum (Fig. 9: arrows) may more likely represent artefacts of preservation rather than real structures.

The recognition of Propelma as a valid genus of Neanastatinae results in four extinct genera described from Baltic amber (Aspidopleura, Brevivula, Neanaperiallus and Propelma) and four extant genera, of which one (Lambdobregma) is restricted to the New World, one (Eopelma) is restricted to the Oriental region, one (Neanastatus) is Old World in distribution, and one (Metapelma) is more widely distributed throughout both the Old and New World. Metapelma is also the only extant genus with a described extinct species, Metapelma archetypon Gibson (2009), from Baltic amber. Pike (1995) and Perrichot et al. (2010) both questionably recorded a species of Eupelmidae from Grassy Lake Canadian Cretaceous and Charentese French Cretaceous amber, respectively. I have been unable to obtain these inclusions for examination, but the earliest known verifiable eupelmids are all Neanastatinae from Baltic amber. Extant Neanastatinae is far less diverse, measured either by number of described species (82) or genera (4), than either Calosotinae (152/8) or Eupelminae (729/33) (Noyes 2012). It is unknown whether the absence of Eupelminae and Calosotinae from Baltic amber is because these two clades diversified more recently than the Eocene age (55–34 mya) of Baltic amber (Weitschat and Wichard 2010) or because some biological factor such as host taxon, host stage or host habitat favoured fossilization of Neanastatinae over the other two groups in Baltic amber resin. Eupelminae are recorded from Dominican amber (20–30 mya), including one taxon identified as the extant genus Zaischnopsis Ashmead (Wu 1997, Poinar and Poinar 1999), but the diversity of eupelmids in Dominican amber has yet to be described or analyzed. Eupelminae have extremely diverse host biologies (Gibson 1997), though relatively few, including Zaischnopsis (Gibson 1995), are parasitoids of wood-boring beetles. Most Calosotinae are parasitoids of wood-boring Coleoptera, which might be thought to favour fossilization of individuals in amber unless the amber producing trees in Baltic forests lacked suitable hosts. Members of Metapelma are also parasitoids of wood-boring Coleoptera and, as noted above, one species has been described from Baltic amber. Members of Neanastatus are primary or secondary parasitoids of Cecidomyiidae (Diptera), and at least some members of Lambdobregma could be egg parasitoids based on a single putative rearing of Lambdobregma schwarzii (Ashmead) from cricket eggs (Orthoptera: Grylloidea) (Gibson 1989). Such a diverse host range among so few extant genera and the greater morphological diversity encompassed by the extinct and extant genera (Gibson 2009) compared to Calosotinae and Eupelminae could indicate Neanastatinae is a comparatively old lineage. Munro et al. (2011, Fig. 1) retrieved Neanastatinae as a relatively basal clade of Chalcidoidea with no close relationships to Calosotinae or Eupelminae using strictly molecular evidence. Using combined molecular and morphological evidence, Heraty et al. (2013) either retrieved Neanastatinae + Calosotinae as the sister-group of Eupelminae (likelihood analysis, fig. 10) or as a paraphyletic assemblage relative to Cynipencyrtidae and Tanaostigmatidae (parsimony analysis, fig. 9).