Raphidioptera fam. gen. sp. indet. A

Figs. 1, 2

Material examined. Specimen PIN 3387 /175, deposited in the Paleontological Institute, Moscow, Russia. An empty exoskeleton of a partly destroyed larva; only its head, prothorax, and one foreleg are preserved. Syninclusions include the head of a first instar cockroach larva, other crumpled insect fragments, and numerous fungal hyphae.

Locality and horizon. Russia: Sakhalin Island: Dolinsk District: the village of Starodubskoye; Sakhalinian amber, middle Eocene (Baranov et al. 2015).

Description. Head blackish, not narrowed posteriorly, short, ca. 2 mm long including mandibles, 1.4 mm wide (ca. 1.43 as long as wide). Ecdysial cleavage lines distinct, consisting of frontal and coronal sutures; frontal sutures diverge at obtuse angle. Antennal socket located at swollen, rounded projection (= antennal tubercle of MacLeod 1964). Antenna short; first and third antennomeres elongate, narrow; second antennomere not clearly discernible, probably very short; fourth antennomere poorly discernible, probably much shorter than third antennomere. Stemmata poorly discernible, three detected in dorsal view, one or two in ventral view. Mandibles not protruding, close to head capsule. Palpi not discernible.

Pronotum destroyed, blackish, apparently short, 1.45 mm wide, ca. 1.7 mm as preserved.

Foreleg crumpled. Coxa, trochanter not clearly discernible. Femur short, stout, covered with scarce setae. Tibia appears narrow, covered with relatively dense setae. Tarsus rather stout, clearly not conical, covered with relatively dense setae. Claw slightly curved, with basal dilation (Fig. 2C).

Remarks. Its one-segmented tarsus indicates that this insect fragment is a larva, certainly belonging to Raphidioptera. The micro-CT shows that it is probably represented by an empty exoskeleton lacking internal tissues and air bubbles. The larva somewhat resembles those of some Hydrophilidae and Gyrinidae (Coleoptera) (e.g., Michat et al. 2010; Minoshima & Hayashi 2012), but differs from these by some details, e.g., these coleopteran larvae possess the anterior margin of the clypeolabrum furnished with projections; or their antennae are relatively long; or their mandibles are long with the inner teeth. It differs from the larvae of Megaloptera in particular by the relatively long head (it is usually short and rounded in Megaloptera) and the very short coxa (relatively long in Megaloptera).

The head of larval Raphidioptera usually bears projecting mandibles, but this larva is unusual, similar to that of the extant Phaeostigma notatum (Fabricius, 1781), in which mandibles do not project (see Beutel & Ge 2008: Fig. 1C).

We follow the antennal segmentation interpretation of Beutel & Ge (2008). This approach seems reasonable. Indeed, as in their understanding, the antennal tubercle of this larva is obviously not a part of the antenna. We think that Haug et al. (2022) incorrectly interpreted the antennal segmentation in fossil Raphidioptera larvae. They considered the antennal tubercle as the first antennal antennomere (their ‘element’), and the very short second ‘element’ to be not an antennomere, although it is well discernible as such in some photographs (see e.g., Haug et al. 2022: Fig. 11a). According to Aspöck et al. (1991), the antenna of Raphidioptera larvae is three-segmented; they consider the antennal tubercle as a possible other, basal antennomere, and the second antennomere of Beutel & Ge (2008) as an inconspicuous sclerite between the two long basal antennomeres. According to Tauber (1991), the antenna is four-segmented, but she did not indicate if the fourth antennomere is the antennal tubercle or the second (shortest) antennomere of Beutel & Ge (2008).