Parasitoid biology preserved in mineralized fossils

About 50% of all animal species are considered parasites. The linkage of species diversity to a parasitic lifestyle is especially evident in the insect order Hymenoptera. However, fossil evidence for host–parasitoid interactions is extremely rare, rendering hypotheses on the evolution of parasitism assumptive. Here, using high-throughput synchrotron X-ray microtomography, we examine 1510 phosphatized fly pupae from the Paleogene of France and identify 55 parasitation events by four wasp species, providing morphological and ecological data. All species developed as solitary endoparasitoids inside their hosts and exhibit different morphological adaptations for exploiting the same hosts in one habitat. Our results allow systematic and ecological placement of four distinct endoparasitoids in the Paleogene and highlight the need to investigate ecological data preserved in the fossil record.

P arasitic lifestyles are extremely successful among animals and evolved independently, perhaps hundreds of times 1 .
With an estimated 50% of species, parasites comprise a huge proportion of animal life on Earth 2 , and the arms races between parasites and their hosts are considered major driving forces for evolution 3 . In insects, parasitism is especially diverse in the order Hymenoptera, where many wasp species develop as parasitoids on or within an arthropod host, ultimately causing its death. In hymenopteran evolution, multiple transitions between host species, developmental stages and modes of parasitoidism are considered key events linked to enormous adaptive radiations [4][5][6][7][8] and an estimated 10-20% of all extant insects are parasitoid wasps 9,10 . Being antagonists of a wide variety of terrestrial arthropods, they have profound ecological and economic impact and many species are used as biological control agents 11,12 .
Evidence for parasitism in fossils is generally rare 13 , as it requires preserved information of interaction between both partners. As a consequence, the fossil record of parasitoid wasps is nearly exclusively restricted to isolated adults, with few examples of unidentified larvae trapped in amber next to their hosts [14][15][16][17] . Therefore, our understanding of parasitoid evolution is based on the inference that fossil organisms exhibited habits resembling those of their extant relatives. The only record of a putative fossil parasitoid wasp inside its preserved host derives from a thinsection of a mineralized fly pupa 18,19 from the later middle to late Eocene fissure fillings of the Quercy region in France, approximately 34-40 million years old 20 . The sectioned pupa was thought to comprise an adult braconid wasp, which was only traceable as faint silhouette lacking any diagnostic characters.
By employing robot-assisted synchrotron-based highthroughput X-ray microtomography, automated graphics processor unit (GPU)-based tomographic reconstruction and advanced semiautomated image segmentation algorithms, we investigate 1510 pupae of three different morphospecies sensu Handschin: 18 Eophora sp. (unavailable genus name 21 ) (1448), Megaselia sp. (55) and Spiniphora sp. (15). The high number of specimens allow morphological and ecological characterization as well as systematic placement of endoparasitoid wasps. The parasitoids are identified as four new species of the family Diapriidae, which we assign to three genera, two of them new.

Results
Preservation and occurrence of parasitoids. Externally, nearly all Quercy fly pupae were preserved as isolated endocasts, of which many were still covered by the puparium, the hardened skin of the last larval instar (Fig. 1a, j, Supplementary Fig. 1). Sometimes body parts of adult flies (especially legs) were recognizable through a partly translucent surface ( Supplementary  Fig. 1ay, be, Supplementary Table 1). Apart from legs and isolated bristles, remains of host flies (Fig. 1j-o) were rarly preserved and did not provide diagnostic characters. In 55 pupae (3.8%) of Eophora we identified parasitation events, which were mostly represented by adult wasps. Preservation of the parasitoids ranged from barely recognizable to well-preserved specimens (Supplementary Figs. 2-5, Supplementary Table 1). In most cases, sclerites were preserved as voids inside the mineralized matrix ( Fig. 1c, g-i). Nineteen wasps had folded wings and showed the symmetric posture of a late wasp pupa (Fig. 1e), while 20 specimens were evidently hatched, as indicated by unfolded wings and an asymmetric body posture (Fig. 1f, Supplementary Table 1).
Systematic palaeontology. In order to assess the size variation within the species, we measured the length between the anterior margin of the propleurae and the anterior tip of the median keel of the propodeum. The reference lengths for holo-and paratypes are listed in Supplementary Table 1 along with information on the preservation of hosts and parasitoid wasps. All other measurements refer to holotypes only and are included in the species descriptions. Differences in size and preservation are further documented by surface renderings of 30 parasitoid heads covering all species ( Supplementary Fig. 5 Apical flagellomeres bead-like, with few scattered setae. First flagellomere cylindrical, distinctly longer than subsequent flagellomeres. Subsequent flagellomeres short, hardly longer than wide. Multiple setal bases present as pores on individual antennomeres. Pronotum with distinct neck, dorsal pronotal surface short, and pronotum adjacent to mesoscutum. Posterior pronotal margin with elongate setae. Lateral panel of pronotum large and triangular, adjacent to mesopleuron. Pronotal depression for accommodation of profemur absent. Pronotal neck with irregular sculpture, dorsal and lateral pronotal margin with indistinct foveae, rest of pronotum smooth. Hind corner of pronotum reaching tegula. Mesothoracic spiracles positioned at lateral margin of pronotum, posteriorly enclosed by prepectal shelf, dorsally reaching tegula. Mesothoracic spiracles nearly completely enclosed by cuticle, omitting just small membranous stripe dorsally. Prosternum subrectangular, transversely divided by complete cross carina. Three large profurcal pits present, one median in anterior half and two submedian in posterior half of prosternum. Profurca u-shaped, profurcal arms completely fused with prosternum. Position of articulation point between propleuron and profurcal arm at posterior end of propectus. Propleural arms anteriorly pointed.
Wings: Forewing unfolded, venation not traceable, and outer wing margin with long pilosity.
Legs: Foreleg with elongate simple trochanter, protibial spur with distinct cleft. Midleg with elongate simple trochanter and two mesotibial spurs. Hind leg with two-segmented trochanter and two metatibial spurs.
Male (Figs. 1e, 4n-v). Measurements given for paratype: NRM-PZ Ar65720. Very similar to female but differs in following features. Reference length: 911 µm. Head height: 605 µm, head width: 599 µm, and head length: 481 µm. IOD:OOD: 0.70. Antenna: 14-segmented, but distinctly longer than in female. Scape, pedicel, and first flagellomere comparable to female, but subsequent flagellomeres cylindrical, i.e., distinctly longer than broad. Eyes 328 µm high and 260 µm wide. Malar space 82 µm. Petiole very similar in proportions (1.75 times as long as wide) and shape, but with more extensive pilosity, also extending to ventral surface. All legs with two-segmented trochanters.  Clypeus narrow, laterally with enlarged anterior tentorial pits, dorsally marked by distinct and straight epistomal sulcus. Mandibles covered by numerous scattered punctures. Mandibles narrow, two-toothed, leaving large, semicircular area from ventral clypeal margin. Area covered by membranous labrum. Toruli dorsally oriented, positioned on indistinct antennal shelf, about half-way upon face. Antennal shelf with few indistinct oblique wrinkles. Antennal shelf connected to epistomal sulcus by two submedian frontal sulci. Supraclypeal area between sulci slightly expanded. Antenna: 14-segmented, elbowed with elongate scape. Scape reaching mid-height of lateral ocellus. Apical flagellomeres bead-like, with few scattered setae. First flagellomere longer than subsequent flagellomeres. Subsequent flagellomeres distinctly to slightly longer than wide. Multiple setal bases present as pores on individual antennomeres. Pronotum with distinct neck, dorsal pronotal surface short, pronotum adjacent to mesoscutum. Posterior pronotal margin with elongate setae. Lateral panel of pronotum large and triangular, adjacent to mesopleuron. Pronotal depression for accommodation of profemur absent. Pronotal neck with irregular sculpture, dorsal and lateral pronotal margin with indistinct foveae, rest of pronotum smooth. Hind corner of pronotum reaching tegula. Mesothoracic spiracles positioned at lateral margin of pronotum, posteriorly enclosed by prepectal shelf, dorsally reaching tegula. Mesothoracic spiracles nearly completely enclosed by cuticle omitting just small membranous stripe dorsally. Prosternum subrectangular, transversely divided by complete cross carina. Three large profurcal pits present, one median in anterior half and two submedian in posterior half of prosternum. Profurca u-shaped. Only bases of profurcal arms preserved. Propleural arms incompletely preserved.
Mesoscutum smooth, with numerous scattered elongate setae. Notauli present as very broad, curved sulci, which are slightly dilated posteriorly. Notauli anteriorly nearly reaching anterior mesoscutal margin and posteriorly nearly reaching transscutal articulation. Notauli internally marked by broad ridges. Transscutal articulation straight and complete. Scutoscutellar sulcus marked by two large semicircular pits, which are medially separated by straight ridge. Pits internally well marked. Axillae narrow and smooth. Axillulae with scattered short setae. Mesoscutellar disc laterally separated from axillula by short ridges. Hind margin of mesoscutellum distinctly foveolate. Mesopleuron smooth and glabrous, mesofemoral depression indistinct. Mesopleuron laterally divided by diagonal sulcus. Mesepisternum anteriorly with distinct procoxal depressions, which are medially separated by distinct carina. Acetabular carina only weakly indicated, mesotrochantinal carina distinct, both carinae meeting medially on ventral mesopleuron. Mesodiscrimen complete and foveolate. Single mesofurcal pit present between mesocoxal foramina. Mesocoxal foramina not completely enclosed by cuticle. Mesofurca not preserved.
Metascutellum with two raised lateral and one raised median carina, and one less distinct transverse carina, metascutellum posteriorly expanded. Lateral panel of metanotum mainly smooth with traces of reduced foveae. Metapleuron subrectangular, coarsely reticulate. Metepisternum with indistinct depressions for accommodating mesocoxae, and without transverse or median carina. Single metafurcal pit present anteromedially of metacoxal foramina. Metafurca not preserved.
Wings: Folded but hardly traceable. Legs: All legs with two-segmented trochanters. Protibial spur with distinct cleft. Midleg with two mesotibial spurs. Hind leg with two metatibial spurs.
Male ( Fig. 5n-v). Measurements given for paratype: NMB F2543. Very similar to female but differs in following features. Reference length: 911 µm. Head height: 596 µm, head width: 608 µm, head length: 478 µm. IOD:OOD: 0.74. Frons with scattered punctures and elongate setae. Antenna: 14-segmented, but distinctly longer than in female. Scape, pedicel and first flagellomere comparable to female, but subsequent flagellomeres cylindrical, i.e., distinctly longer than broad. Multiple setal bases present as pores on individual antennomeres and distinct setation preserved. Eyes 321 µm high and 243 µm wide. Malar space 93 µm, malar sulcus present. Mandibles smooth, without scattered punctures. Antennal shelf with few oblique wrinkles. Scutoscutellar sulcus marked by two large irregular pits, which are medially separated by two oblique ridges. Axillulae with distinct short setation. Acetabular carina distinctly indicated. Petiole 1.20 times as long as wide, laterally with scattered elongate seta and ventrally with dense short setation. Few complete longitudinal carinae present ventrally. scutoscutellar sulcus marked by two large, ovoid pits, which are medially separated by a straight ridge (not clearly separated in Chilomicrus). Xenomorphia appears morphologically very similar to Xenismarus but differs in having the upper tooth of the mandible shorter than the lower tooth and in the presence of plicae. As most shared characters seemingly represent symplesiomorphies, we here refrain from placing the fossil species in the extant genus Xenismarus based on the available morphological evidence, but it may turn out to be closely related or even congeneric.    Supplementary Fig. 4g-j, Supplementary Fig. 5ab, ac, Supplementary Data 5 and 6). Locality. As for X. resurrecta. Description. Female (Fig. 6a-n Mesonotum dorsally flattened. Mesoscutum wider than long, smooth, with very few scattered elongate setae arising from punctures. Notauli present as very broad, curved sulci, which are distinctly dilated posteriorly and deeply pitted anteriorly. Notauli anteriorly nearly reaching anterior mesoscutal margin and posteriorly nearly reaching transscutal articulation. Notauli internally rather weakly marked. Preaxilla smooth, deeply concave with two carinae on surface, arising from anterior and antemedian mesonotal wing processes. Anterior carina articulating with anterior margin of tegula. Tegula huge and rounded, posteriorly reaching posterior wing process. Tegula laterally expanded, with smooth and flattened anterior surface. Tegula with few elongate setae. Transscutal articulation straight and complete. Mesoscutellum with few scattered setae. Scutoscutellar sulcus marked by two large, ovoid pits, which are medially separated by broad straight ridge. Pits internally well defined. Axillae large, triangular, and smooth. Mesoscutellar disc laterally separated from axillula by distinct straight ridges, which are marked by lateral depressions. Hind margin of mesoscutellum distinctly foveolate. Posterior wing process short and broadened, blunt with smooth surface. Mesopleuron smooth and glabrous, mesofemoral depression absent. Mesepimeron with two longi-tudinal ridges connecting anterior and posterior mesopectal margin and separating long rectangular area. Epicnemial pit present with reduced, short pilosity. Sternaulus Metanotum anteriorly overlapped by mesoscutellum. metascutellum with two distinctly raised lateral and distinct median carina. Lateral panel of metanotum composed of foveae. Metapleuron subrectangular, mainly smooth, with ventral row of foveae. Metepisternum with distinct depressions for accommodating mesocoxae, and with distinct median carina (corresponding to metadiscrimen). Single metafurcal pit present anteromedially of metacoxal foramina, posterior to raised carina. Metacoxal foramina with lateral projections. Metafurca indistinct, u-shaped, completely fused to highly raised paracoxal ridge. Metadiscrimenal lamella reaching nearly to mid-level of metacoxal foramina.
Median keel on propodeum formed by v-shaped median carina pointing anteriorly. Anterior margin of propodeum deeply excavate and smooth. Dorsal surface of propodeum with two submedian carinae (instead of single median carina). Plicae developed, dorsal propodeal surface between plicae and submedian carinae smooth. Posterior margin of propodeum deeply excavate, posterolateral corners strongly projecting and broadly bifurcate. Hind margin of propodeum carinate.
Petiole cylindrical, laterally with long pilosity, 1.49 times as long as wide. Petiole with three dorsal, three lateral and two ventral carinae. Second tergum greatly enlarged, anterior margin with deep and narrow median incision. Second tergum overlapping petiole. Subsequent terga extremely shortened. Second sternum greatly enlarged, covering subsequent three sterna. Second sternum with deep grooves filled with micropilosity.
Wings: Not traceable. Legs: All legs with elongate simple trochanters. Protibial spur with distinct cleft. Midleg with two mesotibial spurs. Hind leg with two metatibial spurs.
Male (Fig. 6o- Supplementary Fig. 4k, Supplementary Fig. 5ad, Supplementary Data 7). Locality. As for X. resurrecta. Description. Female (Fig. 7)  Petiole cylindrical, laterally with long pilosity, 1.31 times as long as wide. Petiole with three dorsal, three lateral and one ventral carinae. Second tergum greatly enlarged, anterior margin with deep and broad median incision. Second tergum hardly overlapping petiole. Third and fourth tergum completely covered by second tergum, subsequent terga visible but extremely shortened. Second sternum greatly enlarged, covering at least one of subsequent sterna.
Wings: Unfolded but hardly traceable. Legs: All legs with elongate simple trochanters. Protibial spur with two apices but without distinct cleft. Midleg with two mesotibial spurs. Hind leg with two metatibial spurs.
Male unknown. Comments: The fossils shares a number of morphological characters with the genus Ortona Masner and García 22 , which is restricted to the New World. It differs in the following characters: head not depressed, level of toruli lower than midpoint of eye, oral carina distinctly developed, labrum semicircular (not transverse), clava not long ovoid, and pronotal neck distinctly developed. These characters support placement in a new genus.
Tribal placement. The current subdivision of Diapriinae into three tribes (Spilomicrini, Psilini, and Diapriini) is supported by morphological data 22 . Xenomorphia resurrecta and X. handschini can be readily placed in the tribe Spilomicrini based on the presence of a syntergite on the metasoma, the presence of a distinct malar sulcus (faintly indicated only in female of X. handschini), the absence of spike-like, protruding mesothoracic spiracles, and the presence of complete notauli. Both species have retained a number of plesiomorphic characters, such as the high number of antennomeres (14) in both sexes, which is otherwise only known from the extant genera Xenismarus Oglobin and Chilomicrus Masner and García. These genera are characterized by a restricted, putatively relictual, Valdivian distribution 22 . A macrotergite (second tergum), instead of a syntergite (fused second and third tergum), and the presence of spike-like, protruding mesothoracic spiracles characterize Coptera anka and Palaeortona quercyensis as members of Psilini, a small tribe that comprises only four extant genera. Only the tribe Diapriini, which is considered the most derived lineage of Diapriinae 22 , is not represented in the Quercy fossil parasitoid fauna.

Discussion
The most common species was Xenomorphia resurrecta, of which we found 18 females and 24 males, followed by X. handschini with one female, four males and one pupa and Coptera anka with three females and one male. Palaeortona quercyensis was represented by one female only. Additionally, we identified a single unknown putative second instar wasp larva (Fig. 2, Supplementary Fig. 4l), and a set of last larval instar mandibles (Fig. 2, Supplementary  Fig. 4m), presumably left behind by the emerged parasitoid.
The varying quality of preserved wasps (Supplementary Table 1) suggests that not all specimens present at the time of the fossilization are still traceable. Strikingly, 52 out of 55 discovered parasitation events were recognized by the presence of adult wasps, preserved inside the puparia shortly before or after ecdysis. Though soft tissue preservation in the Quercy fossils is known from beetles 23 and amphibians 20,24,25 the preservation in parasitoid wasps seemed to favor sclerotized structures inside the puparia. We hypothesize that the exoskeleton of the adult wasps was more chemically resistant to early postmortal decay than their earlier developmental stages and the host flies. This could explain the higher representation of adult wasps. X-ray diffraction (Methods, Supplementary Fig. 6) confirms that the fossils are composed of phosphate minerals (apatite). It is assumed that the Quercy arthropods fossilized by a rapid fixation by phosphaterich water followed by encrustation and mineralization 26 . After decay of the cuticle, air-filled cavities were left 23 . In comparison to amber inclusions, which show representational bias towards arboreal taxa 27 , the fossil arthropods of the Quercy localities constitute a unique composition of forest floor communities associated with vertebrate carrion 18,23 . Large numbers of specimens from the same species may be found alongside each other 18 , offering the potential to obtain not only morphological but also ecological information from this particular ecosystem.
The wasps diagnosed herein all belong to the single family Diapriidae, although various hymenopteran lineages are known to exploit fly hosts in decaying substrates 10 . The extant diapriid fauna comprises more than 2000 described species 28 . Diapriid wasps develop as solitary or gregarious endoparasitoids of fly pupae, while some species develop at the expense of beetles or ants 22 . Only in one specimen did we recover fly legs alongside the parasitoid (Supplementary Fig. 4h). This either means that the respective species (C. anka) had the potential to parasitize not only early stages of fly pupae but also almost fully grown adults or that the development of parasitized fly pupae was not immediately paused after oviposition. Unfortunately, the lack of developmental data for extant diapriids hinders a comparison of fossil and extant biology. The observed differences in wing condition, body position and posture indicate a resting period of hatched wasps inside the pupae, putatively resulting from the need for synchronized emergence, a known strategy for insect parasitoids 29 . Based on morphology, four distinct wasp species can be characterized, which were seemingly able to coexist within a forest floor community exploiting fly hosts associated with vertebrate carrion. There is no data on extant diapriid species communities within the same host group in the same habitat but there are many examples from other parasitoid wasp lineages, including the occurrence of four sympatric species of the genus Nasonia parasitizing fly puparia in birds' nests 30,31 . All species described herein are fully winged, indicating the need for dispersal, while several extant diapriid species have their wings reduced in one or both sexes 22 . It seems plausible that there have been differences between the ecological niches of the four species indicated by the striking morphological disparity of species of the tribes Spilomicrini and Psilini. Coptera anka and Palaeortona quercyensis (both Psilini) are characterized by numerous putatively derived cuticular expansions on the articulation points leading to antennae (modification of scape), wings (enlargement of tegula) and petiole (lateral expansion of propodeum) serving as protections for the concerned articulation points. With these characters they would be better equipped than the two Xenomorphia species (Spilomicrini) for a ground-dwelling lifestyle as an adaptation to more concealed hosts. The head expansions of C. anka further facilitate such a forward-directed movement through leaf litter and other ground associated material.
The evolutionary history of Diapriinae is largely unresolved due to the scarcity of well-preserved fossils [32][33][34] , and the absence of a robust phylogenetic hypothesis for the whole family. From a phylogenetic perspective it is relevant to note that the morphological body plan represented by the fossil C. anka remained largely unaltered over a period of about 30 million years, while the other three fossil species represent morphological concepts that have either been significantly modified among extant descendants, or that represent evolutionary dead ends. This implies that the extant diapriine wasp fauna comprises species that exhibit varying degrees of ecological niche conservation, i.e., the retention of ecological characters over evolutionary time scales 35,36 . The ecological and morphological data preserved in the fossil host-parasitoid complex described herein will provide the basis for future comparative studies. It also highlights the need for closing the existing knowledge gap of the morphological and biological diversity of extant parasitoid wasps.

Methods
Samples. The fossils originate from the collections of the Natural History Museum of Basel (NMB) and the Swedish Museum of Natural History (NRM), where all holo-and paratypes of the current study have been deposited. They were collected in the phosphorite mines of the Paleogene fissure fillings of the Quercy region in South-Central France, but the exact locality, collection date and the original collector are unknown. Given the information provided by Handschin 18 , it is likely that the samples housed in Basel were acquired mainly by the fossil collector Rossignol around 1900, who sold them to the Natural History Museum of Basel. This collection was extended by specimens picked by Stehlin and Helbing. The Eophora-specimens were discovered near Bach, which is also true for Spiniphoraspecimens, whereas Megaselia-specimens were also collected near Caylux 18  Photography. Z-stack photographs of the puparia were acquired with a Zeiss Axio Zoom.V16 microscope (Carl Zeiss AG, Oberkochen, Germany) equipped with a PlanApo Z 1.0 × /0.25 FWD 60 Objective, a CL 6000 LED Illumination and an AxioCam HRc Camera. Images were processed with the software Zen 2 and Photoshop (Adobe Systems Incorporated, San José, USA) by cropping, contrast enhancement and the removal of the background.
High-throughput synchrotron X-ray microtomography. Tomographic scans were performed at the UFO imaging station of the KIT light source. A parallel polychromatic X-ray beam produced by a 1.5 T bending magnet was spectrally filtered by 0.2 mm Al to remove low energy components from the beam. The resulting spectrum had a peak at about 15 keV, and a full-width at half maximum bandwidth of about 10 keV. The beamline was equipped with an automated sample change robot (Advanced Design Consulting USA, Inc.) and a fast indirect detector system consisting of a 12 µm LSO:Tb scintillator 37 , diffraction limited optical microscope (Optique Peter) and 12 bit pco.dimax high speed camera with 2016 × 2016 pixels resolution 38 . We used the control system concert 39 for automated data acquisition. Fast screening of all samples was performed with an optical magnification of 5×, resulting in an effective pixel size of 2.44 µm. For screening, we took 1500 or 2000 equiangularly spaced radiographic projections in a range of 180°at 70 fps. Screening data were evaluated online during radiographic data acquisition and after reconstruction of the respective tomographic volumes. Samples containing presumptive parasitoids or hosts were selected for additional highresolution scans. These were done by taking 3000 projections at 70 fps and an optical magnification of 10× (1.22 µm effective pixel size). Due to a smaller field of view at high magnification, anterior and posterior portions of each puparium were scanned separately. Both tomograms were subsequently registered and stitched in Amira 5.6. Tomographic reconstruction was performed with a GPU-accelerated filtered back projection algorithm implemented in the UFO software framework 40 . With the exception of semiautomated segmentation (see below), postprocessing of tomographic data was largely performed using the ASTOR virtual analysis infrastructure at KIT 41 .
Volume rendering. The stitched high-resolution tomograms were aligned and cropped using Amira 5.6. For visualization of the inclusions (Fig. 1d-f, m; Supplementary Figs. 2-4), they were subsequently inverted and the puparia were isolated from the background using the software's segmentation editor. The labelfields were saved as TIF stacks and served for masking the background of the inverted datasets in Fiji 42 . Volume rendering of all postprocessed datasets was performed in Drishti 2.5.1 43 .
Semiautomated segmentation and creation of polygon meshes. High-resolution tomograms were imported into Amira 5.6. Important structures were presegmented in the software's segmentation editor by manually labeling every tenth slice. Distinct morphological structures were assigned to different "materials". The labels served as input for automated segmentation, which was done using the web application Biomedisa (https://biomedisa.de), developed by one of the authors (P.D.L.). Its segmentation process is based on a highly scalable diffusion method, which is free of hyperparameters and thus eliminates an elaborate and tedious configuration 44 . Segmentation results were imported in Amira 5.6 and all individual parts were converted into polygon meshes by employing the "SurfaceGen" tool. The meshes were exported as OBJ files and reassembled in CINEMA 4D R18, which was employed for polygon reduction and smoothing. The data were saved in the DAE format, imported into Deep Exploration 6 and converted into U3D files. The latter were embedded into PDF documents using Adobe Acrobat 9 Pro Extended 45 .
Illustration. The true-to-life impression (Fig. 3, Supplementary Movie 2) was created by rearranging the original mesh of the Xenomorphia resurrecta holotype (NMB F2875) and placing it on top of the mesh of a puparium (NRM-PZ Ar65767). Rearrangement, animation and rendering were done in CINEMA 4D 18. Coloration was finalized using Adobe Photoshop.
Micro X-ray diffraction. Five randomly selected specimens (NMB F2459, NMB F2460, NMB F2531, NMB F2557, and NMB F2915) were characterized by micro Xray diffraction (µ-XRD) at the SUL-X beamline of the KIT light source. Puparia were placed in glass capillaries with a diameter of 1.5 mm. Diffraction data measurements were recorded with a charge-coupled device (CCD) detector (Photonic Science XDI VHR-2 150) in symmetric transmission at 17 keV in focused beam (ca. 100 µm vertical and 250 µm horizontal at sample position). A sample-detector distance of about 110 mm and a detector area of 80 × 120 mm resulted in a maximum diffraction angle of ca. 35°(lattice spacing d ca. 1.2 Å), sufficient for mineral identification. CCD frames were analyzed using Fit2D 46,47 . Instrumental parameters were determined with a LaB6 NIST Standard (660b) and subsequently employed for integration of the sample frames into 1D diffractograms (Supplementary Fig. 6). For mineral identification the ICDD database (www.icdd.com; PDF-2) has been used.
Data availability. CT-raw data were generated at the imaging cluster of the KIT light source. Derived data supporting the findings of this study are available from the corresponding authors upon reasonable request. High-resolution datasets showing parasitation events are deposited at http://www.fossils.kit.edu.