This study is based on arthropod inclusions preserved in a mid-Cretaceous piece of amber from Hukawng Valley, Kachin, northern Myanmar. The size of the piece is 2.4 × 1.5 cm (longest distances taken). All five specimens of larvae of both rhipiphorid species are numbered R1–R5 (Figs 1–3): R1 = conicocephalate larva, R2–R5 = Ripidiini larvae (R2 isolated specimen, R3 and R4 (the slightly curved and shrunken specimen) close to each other, and R5 (isolated specimen with partly missing tergites) preserved at the opposite edge of the amber). Syninclusions belong to Acari, Collembola, Psocodea and an unidentified nymph of a hemimetabolous insect. For all specimens we have adopted the combination of the amber piece code PMJ Pa 6100 and the specific subcode of the individual R1–R5, Ps, Ac1–Ac4, Ny or Co.

The barklouse (Figs 1, 3, Ps; PMJ Pa 6100-Ps) can be assigned to the family Compsocidae and genus †Burmacompsocus as defined by Nel and Waller (2007) and Álvarez-Parra et al. (2023) of the suborder Troctomorpha. Using the description and information provided by Nel and Waller (2007), Sroka and Nel (2017) and Ngô-Muller et al. (2020), we could identify it as belonging to either the species †Burmacompsocus perreaui Nel & Waller, 2007 or †Burmacompsocus pouilloni Ngô-Muller, Garrouste & Nel, 2020. The specimen displays several specific features of the forewing fitting with the descriptions of both species, the short cross-vein between Rs and M, the long M vein branches, and the separation of veins A1 and A2. The slightly lower area between R2+3 and R4+5 as well as between M2+M3, and the position of the apex of Rs distad of the half length of the pterostigma is in stronger agreement with the description of †B. perreaui. It differs distinctly from †Burmacompsocus conjugans Sroka & Nel, 2017 by the longer M veins and the cross vein rs-m on the forewing. Body length: 1.6 mm. Forewing length: 1.8 mm. Hindwing length: 1.5 mm. The length measurements of the wings are similar to the ones of †B. pouilloni, while the specimen is smaller than described for †B. perreaui. Due to a low amount of specimen sampling and fragmentary species descriptions, it is presently not clear, if both species, †B. perreaui and †B. pouilloni are two distinct species, as they mainly differ in two ratios of the forewing.

As previously discussed by Mockford et al. (2013), †B. perreaui shares a highly similar wing venation with †Burmacompsocus banksi (Cockerell, 1916). Due to the poor preservation of the holotype of †B. banksi (Cockerell, 1916) its taxonomic status is currently unclear. It might either represent a unique species, or †B. perreaui is a synonym of this species (Azar et al. 2016). Another species, †Burmacompsocus ojancano Álvarez-Parra & Nel, 2023, is described from Cretaceous Spanish amber (Álvarez-Parra et al. 2023).

There are also four mites included in the amber piece (Figs 1, 4, Ac1–Ac4). One individual likely belongs to the Oribatida (Figs 1, 4, Ac2; PMJ Pa 6100-Ac2), while two specimens are nymphs (possibly 2^nd instars) (Figs 1, 4, Ac3, Ac4; PMJ Pa 6100-Ac3, PMJ Pa 6100-Ac4) of Erythraeidae (J. Dunlop, pers. comm.). One single mite could not be identified but could possibly also belong to Erythraeidae (Fig. 1, 4, Ac1; PMJ Pa 6100-Ac1). Additionally, one specimen of Collembola (Figs 1, 4, Co*; PMJ Pa 6100-Co), presumably with entomobryomorphan affinities, is included, close to the posterior side of the psocid. Unfortunately, the inclusion is not well preserved, and it is not possible to observe any more details. Furthermore, one inclusion that likely represents a nymphal stage of a hemimetabolous insects (Figs 1, 4, Ny; PMJ Pa 6100-Ny) is located close to the conicocephalate larva.

The amber specimen is stored in the Phyletisches Museum, belonging to the Friedrich-Schiller-Universität Jena (Jena, Germany), under the entrance number 23/023, cataloging number PMJ Pa 6100 and the collection number of MW BuA_MW_0001.

The amber piece was polished in several successive steps, using wet Starcke abrasive paper (Starcke, Melle, Germany). Flattening and thinning the piece was done with abrasive paper ST400, and the surface was successively polished with ST1200, ST2000, ST3000, ST5000, and ST7000. Finally, lime powder was used to remove minute scratches.

All photographs were developed in Adobe Lightroom classic (v.11.5) (Adobe, San Jose, USA). Single images were denoised (option: standard) with Topaz DeNoize AI (Topaz Labs, Dallas, USA) and merged with Zerene stacker 1.04 (Zerene System LLC, Richland, USA). All plates were compiled using Adobe Photoshop (v. 24.1.0) and Adobe Illustrator (v. 27.2) (Adobe, San Jose, USA).

An overview image was taken through partially focused stacks with a Canon EOS R5 equipped with a Canon EF 100mm f/2.8L Macro IS USM (Canon, Krefeld, Germany). For stacks the focus bracketing software of the camera was used. The camera was mounted on a Kaiser copy stand. The specimen was illuminated with a Euromex LE.5211-230 cold light source (Euromex, Papenkamp, Netherlands) with three gooseneck lamps. The amber piece was placed in a clear glass petri dish filled with distilled water to prevent any damage to the object, as suggested by Sadowski et al. (2021).

Detail photos were taken at the Museum für Naturkunde Berlin with a Zeiss Axioscope 5 (Carl Zeiss AG, Oberkochen, Germany), mounted with a Canon EOS 80D (Canon, Krefeld, Germany) via a T2-T2 1,6× SLR tube. Stacks were taken manually.

Additional images of syninclusions were taken with a Canon EOS R5 (Canon, Krefeld, Germany) mounted with a Micro-Tube M26 (Novoflex, Memmingen, Germany) and equipped with a five times magnification microscope objective M Plan Apo (Mitutoyo, Kawasaki, Japan). Stack shots were realized with a StackShot macro rail (Cognisys, Traverse City, MI, USA) that were afterwards combined with Zerene Stacker ((Zerene Systems LLC, Richland, WA, USA). For stable positioning of the amber piece, it was placed in Balea Hygiene Handgel (Dalli GmbH & Co. KG, Stolberg (Rhld.), Nordrhein-Westfalen) (Septigel) within a petri dish (Weingardt et al. 2023), with a cover slip on top. To achieve an evenly illumination the whole petri dish was covered with a plastic dome (Rayher, Laupheim, Germany), which was grinded with abrasive paper (Starcke, Melle, Germany) and colored with white (RAL 9010) spray paint (Maston, Veikkola, Finland). The specimens were illuminated with two flashlights (Phottix Juno Li60 Flash, New York, USA) controlled with two Yongnuo RF603C II wireless triggers (Yongnou, Shenzen, China). After taking photographs the amber piece was rinsed and cleaned under running water.

Stacked UV-images of amber were made using the same settings as for other pictures. For images (Fig. 4A–D) a LepiLED with a wavelength of 368 nm (https://www.gunnarbrehm.de/lepi-led) was used. For fixation of the lamp and amber, we used white FIMO® as it does not fluoresce under UV light, to prevent artifacts in the photographs. For Fig. 5 (A–D) two paired UV-LEDs (368 nm) were held diagonally above the amber with a distance of ca. 30 cm. This generates an opaquer image with a better visibility of the resin layers (flows).

For comparative analysis and differential diagnosis of the Ripidiini species, we refer to descriptions of the following primary larvae of the tribe: Blattivorus inquirendus (Silvestri) (Silvestri 1906); †Paleoripiphorus Perrichot, Nel & Néraudeau, 2004 (Batelka et al. 2019); Ripidius quadriceps Abelle de Perrin (Besuchet 1956); and †Ripidius sp. from Eocene amber (Batelka et al. 2020).

A matrix with 19 characters (all unweighted and unordered) scored for 14 terminal taxa was used in the phylogenetic analysis. The parsimony analyses were carried out with NONA (ratchet, 1000 replicates) (Goloboff 1995), and also with TNT (traditional search) (Goloboff et al. 2008), also using the implied weighing option (K = 3.000). Branch support values (Bremer 1994) were calculated with NONA. Finally the trees were redrawn in Adobe Illustrator (v. 27.2) (Adobe, San Jose, USA).

Outgroup taxa: Ericmodes sp. (Protocucujidae), Tribolium sp. (Tenebrionidae); Ripiphoridae (12 taxa): Pelecotoma fennica (Pelecotominae), Ptilophorus dufourii (Ptilophorinae), †Paleoripiphorus sp., Cretaceous †Ripidiini Gen. sp., Blattivorus inquirendus, Ripidius quadriceps (Ripidiinae), Ripiphorus smithi, Macrosiagon ferruginea (Ripiphorinae), †“Ripilarva” parabolica, †“Ripilarva” kachinensis, †“Ripilarva” from Taymir, †“Ripilarva” from Manitoba (genus and subfamily incertae sedis). Data for Ericmodes were taken from Beutel and Ślipiński (2001) and Ślipiński and Beutel (2010).

All of the image data and the character matrix are available from MorphoBank via http://morphobank.org/permalink/?P5334.

Characters of primary larvae of Ripiphoridae with potential phylogenetic significance

1. Habitus: (0) non-triungulinid, approximately parallel-sided body (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae); (1) triungulinid, boat-shaped body (Ripiphorinae, Ripidiinae, †“ Ripilarva”).
2. Head: (0) subglobular (Ericmodes, Tribolium); (1) flattened, bell-shaped, rounded anteriorly (Ripiphorinae, Ripidiinae, †“ Ripilarva” parabolica); (2) triangular and anteriorly pointed (†“ Ripilarva” spp. excl. †“ R.” parabolica); (3) flattened, parallel-sided and elongated (Pelecotominae, Ptilophorinae). The triangular anteriorly pointed head is a highly unusual derived condition.
3. Dorsal ecdysial suture of head: (0) present (Ericmodes, Tribolium); (1) absent (Ripiphoridae). The absence is likely a groundplan apomorphy of the family (Švácha 1994; Lawrence et al. 2010).
4. Lateral cranial area: (0) with scale-like sculpture (Pelecotominae, Ptilophorinae); (1) smooth (all remaining genera).
5. Eyes: (0) several stemmata (Ericmodes, Tribolium, Pelecotominae, Ripiphorinae, extant Ripidiinae); (1) stemmata absent (Ptilophorinae, †“ Ripilarva”). Separate stemmata belong to the groundplan of the family (e.g., Lawrence et al. 2010). Eyes are indiscernible in fossil Ripidiinae as the inclusions are translucent.
6. Antennal insertion: (0) anterolaterally (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae); (1) posterolaterally (Ripiphorinae, Ripidiinae pars.). Not applicable in †“ Ripilarva” and † Paleoripiphorus. The posterior shift of the antennal insertion is an unusual apomorphic condition of a large subgroup of the family.
7. Antennal sensorium: (0) large, cone-shaped (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae); (1) short, spiniform (Ripiphorinae); (2) long, bristle-like (Ripidius, Blattivorus). Not applicable in †“ Ripilarva” and † Paleoripiphorus.
8. Ctenidia: (0) present (†“ Ripilarva”, † Paleoripiphorus); (1) absent (all remaining genera). The presence of ctenidia on the ventral side is a highly unusual apomorphic condition. This is unknown in larvae of extant species of the family.
9. Shape of meso- and metanotum: (0) parallel sided (Ericmodes, Tribolium, Pelecotominae, Ptilophor­inae); (1) distinctly widening posteriorly (all remaining genera).
10. Tibiotarsus: (0) short and robust (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae, Ripiphor­inae, †“ Ripilarva”); (1) distinctly elongated and extremely thin (Ripidiinae).
11. Claws: (0) absent (†“ Ripilarva”, † Paleoripiphorus, Cretaceous †Ripidiini gen. sp.); (1) present (all extant ripiphorids, Ericmodes, Tribolium). A single claw is present in the groundplan of the family (Švácha 1994) and also occurs within extant Ripidiinae and Ripiphorinae (e.g., Lawrence et al. 2010). The absence in all known Cretaceous larvae is clearly a derived condition.
12. Unpaired pretarsal adhesive lobe (bladder): (0) absent (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae); (1) present (Ripidiinae, Ripiphorinae, †“ Ripilarva”). This flexible, lobe-like attachment structure, very similar to the one of strepsipteran larvae (Pohl and Beutel 2004), is likely an apomorphy of Ripiphoridae excl. Pelecotominae and Ptilophorinae. The presence in two non-related parasitic groups shows that this is very likely related with the necessity to attach efficiently to the cuticle of an insect host.
13. Size of pretarsal lobe: (0) minute (Ripidiinae), (1) large (Ripiphorinae, †“ Ripilarva”). Not applic­able for Ericmodes, Tribolium, Pelecotominae and Ptilophorinae.
14. Tibiotarsus: (0) delimitation of tibia and tarsus discernible (†“ Ripilarva”); (1) border between tibia and tarsus indiscernible (all remaining taxa).
15. Shape of abdomen: (0) parallel-sided or subparallel (Ericmodes, Tribolium, Pelecotominae, Ptilophor­inae); (1) widest in the area of segment 2, then continuously narrowing towards the abdominal apex (all remaining taxa). A parallel-sided or subparallel abdomen belongs to the groundplan of the order (e.g., Lawrence et al. 2010). An abdomen with broad segments I–II and then narrowing continuously is likely synapomorphy of the respective taxa.
16. Abdominal spiracles: (0) eight (Ericmodes, Tribolium, Ptilophorinae); (1) seven (Ripiphorinae); (2) less than seven (Pelecotominae, Blattivorus, Ripidius). Unknown in fossil Ripidiinae and †“ Ripilarva”.
17. Concentration of very long setae on abdominal segments: (0) absent (Ericmodes, Tribolium, Pelecotominae, Ptilophorinae, Ripiphorinae); (1) present (Ripidiinae, †“ Ripilarva”).
18. Segment X (pygopodium): (0) short and broad sucker, used for locomotion (Ericmodes, Pelecotominae, Ptilophorinae, Ripiphorinae); (1) elongate and tube-like (Ripidius, Blattivorus, † “Ripilarva” spp. except the one from Taymir); (2) rounded († Paleoripiphorus, †“ Ripilarva” from Taymir, Cretaceous Ripidi­ini gen. sp.); (3) paired pygopods (Tribolium).
19. Urogomphi on segment IX: (0) present (Ericmodes, Tribolium); (1) absent (Ripiphoridae).

Ripiphoridae, subfamily incertae sedis

Ripidiinae, tentative placement

Holometabolan larva of triungulinid morphotype (Beutel et al. 2016), known only in Ripiphoridae and Strepsi­ptera. From triungulin larvae of some Meloidae (Coleo­ptera) it differs mainly by the presence of a pretarsal pad, from Strepsiptera by abdominal segments widening posteriorly and a different placement of the terminal bristles. For discussion of other characters see Beutel et al. (2016), Batelka et al. (2019, 2020), and Pohl et al. (2018).

†“Ripilarva” Batelka & Beutel, nov.

Note 1. The quotation marks for “Ripilarva” indicate here its status as collective group name.

Diagnosis. First instar of parasitic larva. Body small, 0.34–0.56 mm long (excluding terminal bristles), boat-shaped, strongly flattened. Head triangular or parabolic, as long as pronotum or shorter, with one or more ventral straight ctenidia; mandibles simple (indiscernible in some specimens), antennae and palpi in most specimens indiscernible, short and bristle like if visible, and stemmata indiscernible. Thoracic segments with posteroventral ctenidia (synapomorphy of the group), overlapping each other slightly, width of segments increasing from prothorax to metathorax; meso- and metathorax distinctly widening posteriorly. Length of head + pronotum versus mesonotum + metanotum about 0.7–0.9 : 1. Coxae distinctly protruding beyond thoracic segments, trochanter not recognizable; femur broad, tibia and tarsus distinguishable (in contrast to typical tibiotarsus in Polyphaga), not flagellate as in Ripidiinae; pretarsus with soft, unpaired, pad-like appendage, lacking claws. Legs in all known specimens distinctly bent towards head (taphonomic character likely caused by muscle contractions). Abdomen widest at segment II or III, remaining segments strongly tapering towards apex; all segments with concentration of long bristle-like setae, two caudal bristles are the longest (sometimes longer than abdomen).

Note 2. A collective group name (ICZN 1999) is established here to accommodate conicocephalate larvae, a term introduced by Kathirithamby et al. (2017) and adopted for instance by Batelka et al. (2019, 2020) for larvae of a similar morphotype reported so far from Cretaceous amber of Canada, Myanmar and Russia. As the larvae from Canada and Russia are so far unique for the respective deposits, assigning them a species name is not necessary for the time being. Here we provide species names only for both larvae from Kachin amber to distinguish them formally in this contribution.

No type species is designated in accordance with the article 13.3.2. of ICZN (1999). See also articles 42.2.1., 42.3.1 and 66 ibidem.

Etymology. Composed from Ripi – (first part of the family name) and larva (indicating its immature nature). Feminine gender. The name is registered under ZooBank LSID https://zoobank.org/02E6F710-27C8-40B4-8D8F-52B0700D0CF6.

†“Ripilarva“ kachinensis Batelka & Beutel, sp. nov.

https://zoobank.org/A27442B4-0F9D-4111-A85E-273BA6F891B3

Figs 1, 2, 5, 6, R1

Holotype. PMJ Pa 6100 (collection number of MW BuA_MW_0001) stored in the Phyletisches Museum, Friedrich-Schiller-Universität Jena (Jena, Germany), specimen R1.

Description. Habitus. Small, about 0.56 mm excluding terminal bristles. Well sclerotized on dorsal side, with partially translucent cuticle on head and thoracic nota. Coloration of anterior body from dark yellow to dark brown, but abdominal segments apparently very dark. Vestiture of long setae present, but largely restricted to abdomen. Head with distinctly visible row of apically rounded spines (ctenidium) on ventral posterior margin. Thorax distinctly longer than abdomen, with meso- and metathorax distinctly widening posteriorly. Legs strongly developed; unpaired pretarsal adhesive lobes unsclerotised and translucent; all six tibiotarsi oriented towards head. Abdomen 10-segmented, widest in anterior third and very distinctly narrowing towards apex; with pair of long terminal bristles and tubular terminal segment.

Head. Shape equilateral triangular; short, ca. 0.1 mm from mandibles to posterior margin in dorsal view and ca. 0.1 mm broad posteriorly. Mouth opening discernible. Mandibles scarcely visible. Possible terminal palpomere and bristle-like terminal seta of maxillary palpus projecting on both sides of head. Cephalic setae absent. Labrum and clypeus not recognisable as separate structures or regions. Epicranial sutures absent. Stemmata lacking or at least not discernible. Posterodorsal margin of head capsule recognisable by darker coloration. Head without any scale- or lobe-like posterolateral extensions. Mouthparts, sutures, ridges or gula not recognizable on ventral side. Tentorium or tentorial grooves not visible, presumably absent. Posteroventral edge with medially concave regular row of 12 or 14 flattened, short and apically rounded spines, forming ctenidia, each less than 10 μm long.

Thorax. Large, about half of total body length. All three segments subequal in length; meso- and metathorax distinctly widening posteriorly, maximum width of segments increasing from anterior to posterior. Prothorax with well-developed pronotum with wide median ecdysial line and rounded lateral margin; maximum width at posterior third; one strongly developed and long seta (ca. 30 μm) inserted on posterolateral corner; ventral side without recognizable sternal plate; ctenidium similar to that on head, partially recognisable. Forelegs with relatively short parallel-sided coxa distinctly protruding beyond lateral pronotal margin, with two opposite bristle-like setae in distal corners; trochanter not recognizable like on middle and hind legs; femur moderately sized, ca. 50 μm long, fairly broad, with almost straight anterior margin and distinctly convex posterior margin; distal segment composed of tibia and tarsus (fused but of distinct widths, forming tibiotarsus) very slender; claws missing but apical unpaired pad–like adhesive device present (erroneously addressed as pulvillus in some previous studies). Mesothorax slightly longer than prothorax, with wide median ecdysial line, and with distinctly widened and rounded posterolateral corners and long posterolateral seta. Mesonotum simple, without any differentiation or recognisable subdivision; ctenidium partially recognisable, similar to that on head. Middle legs with elongated, cylindrical, apically slightly widening coxa extending far beyond lateral mesonotal margin; mesofemur longer and slightly narrower than profemur; distal elements very similar those of proleg. Metathorax very similar in shape to mesothorax, but slightly shorter and slightly broader. Metanotum simple, with pair of setae close to hind margin, inserted relatively distant from lateral margin; very long seta inserted posterolaterally on ventral side of segment; ctenidium not recognisable. Hind legs distinctly longer than forelegs and middle legs, very promi­nent; metacoxae also elongate and cylindrical; metafemur and distal elements slightly larger than those of anterior pairs of legs.

Abdomen. Ten-segmented, shorter than thorax (ca. 0.19 mm). All terga simple and sclerotized, those of segments I–VIII with straight posterior margin; greatest width at segment II, strongly tapering towards apex. Segments II–VIII bearing long lateral setae. Segment I relatively long, widening posteriorly. Segments II–IV shorter, almost parallel-sided, III distinctly narrower than II, and IV narrower than III. Segments V–VIII slightly longer, each of them narrower than the preceding one. Segment IX about twice as wide as long, with rounded posterior margin and posterolaterally inserted caudal bristles distinctly longer than entire abdomen (0.24 mm). Segment X narrow, tubular. Spiracles not recognizable.

Differential diagnosis: †“Ripilarva” kachinensis sp. nov. differs from †“Ripilarva” parabolica sp. nov. illustrated in Beutel et al. (2016) by the distinctly triangular head. Other differences are inconspicuous or uncertain. The putative absence of the metathoracic ctenidium and the longer abdominal segment X might be due to imperfect or different preservation of both fossils and observation constrains. †“Ripilarva” sp. from Manitoba and †“Ripilarva” sp. from Taimyr have two rows of cteni­dia on the head. The anterior ctenidium is not visible in †“Ripilarva” kachinensis sp. nov. and possibly missing.

Etymology. The species name refers to place of origin of the Kachin amber.

†“Ripilarva” parabolica Batelka & Beutel, sp. nov.

https://zoobank.org/2A586114-6527-4DEC-8043-A963DECA6BC2

Holotype. NO. BU-001009. Presently in the collection of the Institute of Zoology of the Chinese Academy of Sciences in Beijing (Prof. Dr. BAI Ming). The specimen will eventually be deposited in the Three Gorges Entomological Museum, Chongqing, China.

Description. Beutel et al. (2016).

Differential diagnosis. †“Ripilarva” parabolica differs from †“Ripilarva” kachinensis by its parabolic head which is not pointed anteriorly.

Etymology. The species name refers to the parabolic shape of the head.

List of conicocephalate larvae included in the †“Ripilarva” Batelka & Beutel, nov.

†“Ripilarva” kachinensis Batelka & Beutel, sp. nov. (Myanmar; this study)

†“Ripilarva” parabolica Batelka & Beutel, sp. nov. (Myanmar; Beutel et al. 2016)

†“Ripilarva” sp. (Manitoba, Canada; Grimaldi et al. 2005)

†“Ripilarva” sp. (Taimyr, Russia; Kathirithamby et al. 2017)

†Ripidiini gen. et sp. indet.

Figs 1, 3, 5, 6, R2–R5

Material. Primary larvae PMJ Pa 6100 (collection number of MW BuA_MW_0001), specimens R2–R5.

Description. Habitus. Body (depending on preservation) ca. 0.31–0.45 mm long excluding terminal setae, subparallel. Well sclerotised, with middle to dark brown cuticle. Head and thorax covered with short finger-shaped sensilla. Pair of long bristle-like setae inserted on posterior edge of thoracic segments and pair of long spiniform setae close to lateral edge of each postcephalic segment. Abdominal segments connected by soft membranous cuticle, resulting in varying length of abdomen; abdominal segments I–III as long as mesothorax and metathorax combined. All three thoracic segments widest at posterior edge. Legs extremely prolonged; coxae cylindrical with one finger-shaped-sensillum and two bristle-like setae, and one long preapical bristle-like seta present on each femur; strongly elongated and thin tibiotarsus 1.3–1.7 × as long as femur, with recognisable minute semierect spines, and with unpaired, flexible lobe-like pretarsal attachment device. Abdominal segments IV–X with conspicuous concentration of long setae.

Head. Prognathous, elongated, rounded anteriorly, nearly parallel sided, very slightly widening posteriorly, with greatest width at hind margin; about half as long as pronotum, ca. 50 μm from anterior to posterolateral margin in dorsal view [R2], slightly less than 50 μm wide posteriorly [R2]. Stemmata or eye spots indiscernible. Transverse frontoclypeal strengthening ridge and dorsal ecdysial sutures (frontal sutures and coronal suture) not visible, probably absent. Posterodorsal edge slightly concave, overlapped by distinctly convex anterior pronotal margin. Scale- or lobe-like posterolateral extensions of head capsule absent. Posteroventral cephalic margin adjacent with anterior prosternal edge. Tentorial grooves not recognizable. Labrum not recognizable as a separate structure, apparently fused with clypeofrons. Antennae at least two-segmented, situated in posterior half of head, consisting of one or two robust basal segments and bristle-like terminal seta. Mandibles robust, apparently movable, protruding in some specimens [R2], with curved outer margin; apices not overlapping each other when closed, mesal edges approximately parallel. Maxillary palpus long, with two segments visible, with one lateral and one terminal bristle-like seta [R4]. At least two pairs of finger-shaped sensilla present dorsally on anterior part of head, well projecting beyond cephalic edge [R2, R4].

Thorax. Prothorax longest and narrowest, metathorax shortest and widest [R2]. Pronotum well developed, moderately widening posteriorly, with rounded posterolateral corners; anterior margin slightly convex, very slightly broader than posterior cephalic margin; distinct lateral edge absent; several pairs of erect finger-shaped sensilla placed close to lateral pronotal margin; two pairs of bristle-like setae placed at posterior edge of pronotum. Prosternal ctenidia absent. Procoxae prominent, elongate, cone-shaped, inserted posterolaterally, very distinctly protruding beyond lateral pronotal margin, with pair of bristle like seta and one finger-shaped sensillum [R2]; trochanter indiscernible; femur strongly elongated (ca. 80 µm), slender, strongly narrowed basally, with very slightly convex hind margin, and one long preapical seta; tibiotarsus elongated (ca. 0.14 mm), about 1.75 × longer than femur, very thin, strongly tapering distally, with several extremely small spines [R4]; pretarsal pad 1.5 × longer than diameter of base of respective tibiotarsus, without other lateral projections or claws. Mesothorax distinctly shorter than prothorax, with anterior mesonotal corner overlapped by broadly rounded posterolateral pronotal edge and posterior corners widely rounded. Mesonotum simple, undivided, without median ecdysial line; finger-shaped sensilla placed close to posterior mesonotal edge; several pairs of bristle-like setae and finger-shaped sensilla placed along posterior margin. Sternal region without recognizable surface structures. Middle legs very similar to forelegs but even more elongated, total length 0.25 mm; mesocoxa similar in shape and size to procoxa, and similar preapical femoral seta and pretarsal pad present [R4]. Metathorax similar in shape to mesothorax. Metanotum slightly shorter than meso­notum. Hind legs very similar to middle legs.

Abdomen. Ten-segmented (ca. 0.2 mm [R2]), subparallel, slightly rounded laterally. Segments I–IV distinctly wider than remaining segments. Each segment with two pairs of erected long bristle-like setae at posterolateral corners and several setae on posterior edge, overlapping with each other (precise counting difficult). Ultimate segment rounded, without visible sucker or pygopod.

Measurements (mm). R2: length = 0.45, head and thorax combined = 0.25, head = 0.05, pronotum = 0.075, mesonotum = 0.07, metanotum = 0.55, abdomen = 0.2. R3: length = 0.42. R4 (shrunken and bent): length = 0.31.

Differential diagnosis. Species differs from Blattivorus, Ripidius and †Paleoripiphorus by distinctly prolonged pretarsal pads. Like Blattivorus and †Paleoripiphorus with slender head and thorax (much wider in Ripidius); like Blattivorus and Ripidius without prosternoctenidium (present in †Paleoripiphorus); antennae present (absent in †Paleoripiphorus). In its general morphology the larva is very similar to that of Blattivorus, also sharing long and thickened thoracic spiniform sensilla.

At least 5 different resin flows are recognizable observing the amber piece from the flat side (Figs 5, 6A, B). The long sides indicate a relatively complicated structure of the amber, with several closely spaced resin flows (Figs 5, 6C, D). The results show that larvae R1–R4 are located in the same flow, while R5 is located in a different one. We define the specimens in the layer containing larvae R1–R4 as eusyninclusions, while R5 is a parasyninclusion.

Analyses with NONA (Figs 7, 8) yielded between 4 and 24 trees of 30 steps (consistency index 0.83, retention index 0.91). TNT yielded 4 trees with the same length, and the same number was obtained after using the implied weighing option. In the strict consensus tree Ptilophorus and Pelecotoma were placed closest to the root of Ripiphoridae. The rest of the genera form a well-supported clade (branch support [bs] 5). Within this monophyletic unit all genera excluding Ripiphorus and Macrosiagon form a monophylum (bs 2). †“Ripilarva” is also confirmed as a clade (bs 2). Apomorphies and branch support values are shown in Fig. 7.

The minute primary larvae of Ripiphoridae are not only of great interest as highly specialised life forms, but are also a valuable source of phylogenetically relevant data complementing adult morphology (e.g., Švácha 1994; Batelka et al. 2022). The first instars of †“Ripilarva” are morphologically very distinctive. They share several distinctive features with primary larvae of Ripiphorinae and Ripidiinae, such as unsclerotised pretarsal pads and a cuticular surface covered by long bristle-like setae (notably on abdominal segments VIII and IX). However, they differ distinctly from larvae of all known extant Ripiphoridae by the presence of ctenidia on the head and thorax. The only exception is a similar prosternoctenidium in Cretaceous †Paleoripiphorus, but it is unclear whether these structures are homologous (Batelka et al. 2019). While the head of †“Ripilarva” kachinensis sp. nov., †“Ripilarva” sp. from Manitoba, and †“Ripilarva” sp. from Taimyr is conspicuously triangular, it is parabolic in †“R.” parabolica sp. nov. as illustrated in Beutel et al. (2016), similar to the condition found in Ripiphorinae. The bristle-like seta on both ultimate maxillary palpomeres of †“R.” kachinensis sp. nov. is similar to those observed in Ripidiini including †Paleoripiphorus (Batelka et al. 2019). Their “absence” in some †“Ripilarva” is likely due to observation constraints, as it is likely the case with the putative absence of trochanters in all †“Ripilarva”, sternal ctenidia in †“Ripilarva” from Manitoba, or metasternal ctenidia in †“Ripilarva” kachinensis sp. nov. Some structures may change their shape artificially in fossil larvae due to weak sclerotization. Abdominal segment X of †“Ripilarva” sp. from Taimyr, for instance, displays a knob-like shape, in contrast to the tubular one in all three remaining species. However, it is conceivable that this weakly sclerotised segment changes its shape when used for locomotion. The abdominal sucker in Ripiphorinae has similar characteristic and function (Tomlin and Miller 1989; Batelka and Straka 2011). All specimens of †“Ripilarva” display legs bent towards the head in almost identical orientation, a possible artifact caused by post-mortem contractions of muscles adapted for phoresy (Batelka et al. 2018).

Cretaceous invasive primary larvae of Ripidiini evolved a remarkable set of derived features. They developed various specialized structures involved in locomotion, such as lobe-like pretarsal attachment devices, microscopic spines on the prolonged tibiotarsus, and specialized setae and ctenidia on the body surface. It is likely that these unusual morphological features have evolved in the context of evolutionary arms races with their hosts. Considering that only larvae of four out of 24 extinct and extant genera are reliably documented (Batelka et al. 2019), it is premature to make conclusions about their evolutionary history. However, considering the fact that they are available in amber of various ages (sometimes with presumed conspecific adults) and larvae of extant species can be reared from eggs (Silvestri 1906; Riek 1955; Besuchet 1956), the group has a great potential to be among the best understood subgroups of Ripiphoridae. The larva of Ripidiini described in this study shares several features with some extant relatives. This includes similar arrangements of notal and abdominal spiniform setae, the configuration of the legs including unusual pretarsal pads, the shape of the mandibles, antennae and maxillary palpi, and a conspicuous body shape similar to Blattivorus. This Cretaceous morphotype thus provides evidence, that larvae of some extant lineages have acquired specialised features related with parasitism already 100 My ago. Barbier (1947) described sedentary behaviour for the female of Ripidius pectinicornis, an extant species of Ripidiini, allowing this species to spread worldwide (Falin 2001). The sedentary habit is likely for the Cretaceous species of Ripidiini due to their preserved morphological adaptations for this lifestyle. In contrast, a distinctive prosternoctenidium and some other derived features of the larva of †Paleoripiphorus remain unknown in post Cretaceous taxa. The larva we described here thus represents another remarkable example of morphological stasis of some derived characters in this subfamily. This was documented previously based on uniflabellate antennae, reduced mouthparts and wing venation, enlarged compound eyes, and scale-like elytra in adult males of Cretaceous †Amberocula Batelka, Engel & Prokop, 2018, †Cretaceoripidius Falin & Engel, 2010, and †Ripidinelia Batelka & Prokop, 2019. The same applies to larviform bodies of females of †Paleoripiphorus (Batelka et al. 2021a).

Our preliminary phylogenetic analyses of larval morphological characters reflect the results of a recent molecular study (Batelka et al. 2016) surprisingly well. Pelecotoma and Ptilophorus are placed closest to the root of the family. The sister group relationship between the two genera (Batelka et al. 2022) is not unambiguously confirmed, but suggested by an unusual potential synapomorphy, a lateral cephalic area with a scale-like surface structure (char. 4.1.). The monophyly of Ripiphoridae excl. Pelecotoma and Ptilophorus is strongly supported (branch support 5) by five unambiguous and unusual synapomorphies, notably a boat-shaped body (char. 1.1), a posterolateral antennal insertion (char. 6.1), posteriorly widening meso- and metanota (9.1), unpaired pretarsal adhesive lobes (char. 12.1), and a characteristically shaped abdomen (char. 18.1). Ripidiini and †“Ripilarva” share three non-homoplasious synapomorphies, lacking claws (char. 11.0), a concentration of very long setae on the abdominal segments (char. 17.1), and an elongate and tube-like pygopodium (char. 18.1). †“Ripilarva” is supported by one non-homoplasious apomorphy, a tibiotarsus with a (secondarily) visible intrinsic border (char. 14.0), and two additional homoplasious derived features. A triangular and apically pointed head is a conspicuous apomorphy of †“Ripilarva” excl. †“Ripilarva” parabolica. Our data strongly suggest that †“Ripilarva” should be included in Ripidiinae.

The piece of amber treated here raises the question of the circumstances of the embedding event. How accidental was the occurrence of two types of primary larvae of parasitic beetles in such a tiny space in a tropical Cretaceous forest? Ross (2024) reported 2,781 species from Kachin amber up to the end of 2023, and definitely many additional thousands of amber pieces were examined for biological inclusions since that time. Despite these intensive efforts, primary larvae of Ripiphoridae in any Cretaceous amber are only rarely reported (see Introduction). Moreover, previous studies have shown that the occurrence of †Paleoripiphorus is always subject to a certain pattern: they either appear in clusters up to several dozens of specimens or they co-occur with males or with both sexes of the same species (Batelka et al. 2019, 2021a). In contrast, conicocephalate larvae were always found isolated so far. Scientists usually handle polished and trimmed pieces, which means that most of them likely do not contain the entire fauna that they could have held originally. Therefore, the accidental occurrence of these short-lived life forms very close to each other appears highly unlikely.

It is evident from the state of preservation of the four larvae of Ripidiini that at least some of them were already dead (incomplete R5 and likely desiccated R4), while at least R2 was alive when rapidly covered by a flow of fresh resin. Fluid extruding from the mouth is visible, a phenomenon also observed in freshly killed primary larvae of another rhipiphorid, Ptilophorus dufourii (Batelka et al. 2022, fig. 4A). A different state of preservation in short-lived free instars (several days suggested by Besuchet (1956) for Ripidius quadriceps) is not surprising, as the same was reported for larvae and adults of †Paleoripiphorus in previous studies (e.g., Batelka et al. 2019).

Available data indicate that the mode of preservation of rhipiphorid first instars in amber is not accidental and that their taphonomy is affected by their bionomics. All known females of Ripidiini including Cretaceous †Paleoripiphorus are flightless and larviform. Pheromone communication of females of Ripidius quadriceps with restricted mobility, observed by Batelka et al. (2021b), is very likely, similar to other extant and extinct species of this tribe. The cluster of four Ripidiini larvae in the piece of amber treated here is likely the result of a behaviour of females similar to that documented for †Paleoripiphorus (Grimaldi et al. 2005; Batelka et al. 2019, 2021a). In clear contrast, the presence of a single specimen of †“Ripilarva” kachinensis sp. nov. corresponds with other isolated findings of similar larvae in Kachin amber, or amber from Manitoba and Taimyr. A likely mechanism explaining this rare co-occurrence with larvae of Ripidiini is a different reproductive behaviour of females. We suggest that more mobile females of †“Ripilarva” distributed relatively few eggs in a wide area, whereas sedentary females of contemporaneous Ripidiini deposited numerous eggs in a very confined site. This may have been related with the smaller size of females of †“Ripilarva” and the presence of wings. The shared preference for the same microhabitat and oviposition season might be explained by the same hosts used by both groups.

By comparing the amber piece under UV-light, we were able to distinguish several resin flows (Figs 5, 6A, B). It is likely that these occurred shortly after each other, as no noteworthy debris or air bubbles were trapped between the layers. It is evident that the larvae R1–R4 were trapped in the same flow of resin (Figs 5, 6A, B), but the larva R5 in another one. The placement of the specimens strongly suggests that these events occurred within a very limited time span. This conforms with our interpretation that the conicocephalate larvae and the larvae of Ripidiini occurred in the same microhabitat within a very narrow time window. Additionally, the psocid specimen is located at the transition zone between two flows and Schlauben are present around this insect. The preservation of the entire body is very uniform, without any noticeable differences between the anterior and posterior part, which are apparently trapped in two different flows. All ripiphorid larvae could therefore arguably be considered as eusyninclusions. However, as we use these terms as defined by Solórzano-Kraemer et al. (2023), we characterize the ripiphorid larvae R1–R4 as eusyninclusions, and larva R5 as a parasyninclusion. We consider the psocid (Fig. 1, Ps) to be part of the eusyninclusions as the main part of the body is trapped in the same resin flow as larvae R1–R4 (Figs 5, 6). Additionally, one erythraeid immature (Fig. 1, Ac4), the collembolan (Fig. 1, Co*), as well as the im­mature hemimetabolan (Fig. 1, Ny) are eusyninclusions. The oribatid mite (Fig. 1, Ac2), one mite of uncertain identity (Fig. 1, Ac1), and the second erythraeid immature (Fig. 1, Ac3) apparently belong to a single, different resin flow and are considered parasyninclusions (Figs 5, 6).

The entrapment of the psocid together with the ripi­phorid larvae is likely accidental, as extant Compsocidae are known to inhabit tree trunks and bark microhabitats in rain forests (Mockford 1967; Mockford 2018). While the biology of the mites we observed in the amber is unknown, it is possible that both the ripiphorid larvae and the mite nymphs shared the same host. At least some erythraeid nymphs are ectoparasites of cockroaches (e.g., Southcott 1999; Muñoz-Cárdenas et al. 2015). This is tentatively suggested by the occurrence of both types of immatures in the same microhabitat in the same time period.