Cretaceous amber inclusions illuminate the evolutionary origin of tardigrades

Tardigrades are microscopic panarthropods with a compact body plan and four pairs of typically claw-bearing lobopodous legs, known for cryptobiosis and wide ecological distribution across marine, freshwater, and terrestrial habitats. Despite their ubiquity, the fossil record is extremely scarce, limiting macroevolutionary inferences about their body plan origin, terrestrialization, and evolution of cryptobiosis. Four crown-group tardigrade fossils are known from amber, but only Milnesium swolenskyi (Turonian, 89.8–93.9 Ma) and Paradoryphoribius chronocaribbeus (Miocene, ~16 Ma) have well-resolved taxonomic placements. The earliest crown-group fossil Milnesium swolenskyi displays Milnesium-type claws and cephalic papillae, indicating placement within Milnesiidae (Apochela) and suggesting morphological stasis over ~90 Myr. Paradoryphoribius chronocaribbeus possesses Isohypsibius-type claws and distinctive foregut morphology, supporting placement within Isohypsibioidea (Parachela). Two additional Canadian Campanian (72.1–83.6 Ma) amber tardigrades—Beorn leggi and a second undescribed specimen—have uncertain affinities due to vague or poorly preserved diagnostic characters (notably claw morphology). Beorn leggi was originally placed in the extinct Beorniidae but has been speculated to belong within extant families such as Isohypsibiidae or Murrayidae, though details have remained problematic. Previous molecular dating studies have variably inferred crown-group tardigrade diversification from late Cryogenian–early Ediacaran to early Paleozoic, often using limited fossil constraints and sometimes controversial stem-group fossils (e.g., Siberian Orsten material). Milnesium swolenskyi has been used as a calibration, but Beorn leggi has not, due to uncertain affinities and its younger age. Here, high-quality confocal fluorescence microscopy is used to resolve the morphology and phylogenetic affinities of Beorn leggi and the undescribed second Canadian amber tardigrade, integrating these data into molecular clock analyses with comprehensive extant sampling to refine the timing of tardigrade origins.

• Fossil record: Few definitive crown-group tardigrades are known—Milnesium swolenskyi (Late Cretaceous, Turonian) placed in Milnesiidae (Apochela) and Paradoryphoribius chronocaribbeus (Miocene) placed within Isohypsibioidea (Parachela). Beorn leggi (Campanian) and a second Canadian amber specimen have historically ambiguous affinities due to limited morphological resolution, particularly of claws and cephalic structures.
• Prior molecular dating: Early studies using three nuclear genes suggested Ediacaran diversification of crown-group Tardigrada (627–691 Ma). Later ecdysozoan-focused analyses with sparse tardigrade sampling placed eutardigrade splits around the Carboniferous but relied on a stem-group Orsten fossil. Group-focused studies (echiniscids, milnesiids) often inferred Ediacaran crown splits; more recent ecdysozoan analyses using Milnesium swolenskyi calibrated crown Tardigrada near the late Cambrian. None incorporated Beorn leggi due to taxonomic uncertainty and younger age.
• Taxonomic context: Claw morphology (Hypsibius-type vs Isohypsibius-type vs Ramazzottius-type) is critical for higher-level placement within Eutardigrada. Recent taxonomic revisions propose changes to Hypsibiidae subfamily ranks, but consensus is lacking. The need for revised morphological diagnoses and integration with molecular data has been emphasized to refine calibrations for divergence time analyses.

Studied material and provenance: Two amber inclusions from secondary deposits collected in 1940 near the entrance of the Saskatchewan River into Cedar Lake, Manitoba, were examined. Specimens are housed in the MCZ, Harvard University (MCZ PALE-5213: Beorn leggi holotype; MCZ PALE-45862: Aerobius dactylus gen. et sp. nov. holotype).

Microscopy and imaging: Amber specimens were mounted on slides with dental wax and glycerin medium (Immersol G). Transmitted light imaging used a Zeiss Axioscope 5 with Axiocam 208 color camera; Z-stacked optical sections were reconstructed using Sum Slices. Confocal fluorescence microscopy used Zeiss LSM 980 with Airyscan 2 at 639 nm excitation to capture autofluorescence of cuticular structures. Color-coded and inverted grayscale projections were generated in Fiji (physics LUT; Max Intensity Z-projection). Lighting adjustments were made in Adobe Lightroom. Comparative slides of extant tardigrades were imaged similarly; image stacks combined using Photoshop auto-blend; figures assembled in Illustrator.

Morphometrics: Body length measured from anterior tip to caudal end (excluding hind legs). Claw measurements followed Beasley et al. to obtain primary/secondary branch and basal section lengths; br ratio (secondary/primary length) was computed. Measurements performed in FIJI.

Total-evidence phylogenetic analysis: A concatenated dataset combined a 36-character morphological matrix (body surface, claws, bucco-pharyngeal apparatus, egg morphology) with 18S rRNA sequences (final alignment 1774 bp). Morphological coding based on type species (with exceptions where sequences were missing). Alignment used MAFFT I-INS-i; trimming in AliView. Bayesian inference in MrBayes 3.2 with Mk+Gamma (variable coding) for morphology and GTR+I+G (PartitionFinder-selected) for 18S; two runs, four chains each, 2,000,000 generations, sampling every 500, 25% burn-in; convergence assessed by split frequencies <0.01, ESS>200, PSRF≈1; 50% majority-rule consensus tree obtained.

Divergence time estimation: Two datasets used: (1) phylogenomic (nine tardigrades + Drosophila outgroup; 335 orthologs, 139,117 aa sites; ASTRAL species tree from gene trees inferred with IQ-TREE). MCMCTree (PAML 4.9) with independent clock and Birth-Death priors; uniform node age priors; two runs; convergence verified by R²=1 across runs. (2) 18S/28S rRNA dataset (139 taxa: 18S for all, 28S partial region for 80; concatenated 3210 nt; ML in IQ-TREE with two partitions and GTR+I+G; BI in MrBayes as above). BEAST 2.6 used for relaxed lognormal clock and BD tree model; partitioned site models selected via bModelTest; three runs per calibration strategy; logs combined with LogCombiner (resampling every 50,000; 25% burn-in), ESS>200 confirmed in Tracer; MCC tree with CA heights via TreeAnnotator. Calibration strategies: 0fossil (root only), 1fossil (Milnesium swolenskyi for crown Tardigrada), 2fossils_SupF (Milnesium swolenskyi for crown Eutardigrada and Beorn leggi for Hypsibioidea), and 2fossils_Fam (as SupF but Beorn leggi at Hypsibiidae). Posterior age distributions were compared via density plots; for wide distributions (2fossils_Fam), posterior trees (n=4503) were binned by geological period to visualize relative frequencies; Kolmogorov–Smirnov tests compared posterior vs prior-only distributions.

Data availability: Alignments, matrices, trees, MCMC outputs, scripts, and assembled genome referenced via Dryad and MorphoBank; ZooBank registration provided for Aerobius dactylus.

• High-resolution confocal imaging resolved diagnostic external morphology of two Cretaceous Canadian amber tardigrades.
• Beorn leggi (MCZ PALE-5213): possesses Hypsibius-type claws with clear flexible connection between primary branch and basal section; external/posterior claws with continuous curve secondary branch, internal/anterior claws more robust; accessory points present, pseudolunules absent. Redescribed as a member of Hypsibioidea and reallocated to family Hypsibiidae; extinct family Beornidae rejected due to lack of synapomorphies.
• Aerobius dactylus gen. et sp. nov. (MCZ PALE-45862): unique claw combination. Anterior leg pairs (I–III) show a modified Isohypsibius-type external claw (secondary branch at right angle to basal section) but with an evident flexible connection (curved base of primary branch). Hind leg (IV) claws differ markedly in shape/size; posterior claw primary branch notably longer than secondary; accessory points present; pseudolunules at least on internal claw. Placed within Hypsibioidea but not assigned to Hypsibiidae.
• Phylogenetic placement: Total-evidence Bayesian analysis recovered both fossils within Hypsibioidea; Aerobius dactylus in a polytomy with Beorn leggi and Hypsibius dujardini.
• Divergence times (BEAST, 18S/28S, 2fossils_Fam strategy using Beorn leggi at Hypsibiidae): • Crown Tardigrada (Eutardigrada–Heterotardigrada) mean 498.86 Ma (95% HPD: 613.66–380.5 Ma). • Crown Heterotardigrada mean 370.94 Ma (95% HPD: 481.47–261.14 Ma). • Apochela–Parachela (crown Eutardigrada) mean 315.69 Ma (95% HPD: 450.25–192.23 Ma). • Crown Echiniscoidea (marine vs limnoterrestrial split) mean 270.26 Ma (95% HPD: 370.39–181.56 Ma). • Crown Parachela mean 199.2 Ma (95% HPD: 279.08–126.49 Ma). • Crown Hypsibioidea mean 138.41 Ma (95% HPD: 190.99–92.38 Ma). • Crown Isohypsibioidea mean 104.61 Ma (95% HPD: 149.74–65.06 Ma). • Crown Macrobiotoidea mean 106.02 Ma (95% HPD: 151.89–68.52 Ma). • Crown Eohypsibioidea mean 66.83 Ma (95% HPD: 124.27–19.01 Ma). • Crown Apochela mean 67.98 Ma (95% HPD: 119.64–23.06 Ma). • Crown Echiniscidae mean 146.01 Ma (95% HPD: 209.59–89.93 Ma).
• Calibration impact: Incorporating Beorn leggi as a Hypsibiidae calibration shifts estimates for shallower nodes older compared to analyses without it or using only outgroup/fewer fossil calibrations; posterior distributions for deep splits overlap across datasets and strategies, but family-level calibration notably affects shallow divergences.
• Evolutionary inferences: Independent terrestrialization events inferred for eutardigrades (around end-Carboniferous) and heterotardigrades (Lower Jurassic). Estimates provide minimum ages for convergent acquisition of cryptobiosis. Evidence of morphological stasis in Beorn leggi claws contrasts with unique claw morphology in Aerobius dactylus, potentially indicating transitional forms between Isohypsibius- and Hypsibius-type claws and functional differentiation of hind-leg claws.

The study resolves the long-standing uncertainty in the affinities of two Campanian Canadian amber tardigrades, placing Beorn leggi securely within Hypsibiidae (Hypsibioidea) and describing Aerobius dactylus gen. et sp. nov. within Hypsibioidea but outside Hypsibiidae based on claw morphology. This provides a critical, taxonomically secure Cretaceous calibration within Parachela. Using Beorn leggi as a family-level calibration demonstrably influences divergence time estimates at shallow nodes, refining the temporal framework for diversification of speciose eutardigrade lineages. Deep-node estimates (crown Tardigrada) converge across datasets and calibration strategies, supporting a Cambrian origin of crown tardigrades, consistent with hypotheses of miniaturization from macroscopic lobopodian ancestors. The unique claw configuration in Aerobius dactylus suggests an evolutionary transition between Isohypsibius- and Hypsibius-type claws and highlights differential evolutionary trajectories and functional roles of hind legs versus anterior legs, congruent with developmental and behavioral data. The temporal concordance between diversification of limnoterrestrial tardigrades and the expansion of bryophytes and macrolichens during the Jurassic–Cretaceous suggests ecological drivers (increased substrates) facilitating tardigrade radiation. Independent evolution of cryptobiosis in eutardigrades and echiniscoidean heterotardigrades is supported by distinct genetic toolkits and divergence timing that spans major Phanerozoic crises, implying cryptobiosis may have buffered extinction risk. The work underscores the importance of detailed fossil redescription and accurate placement for robust molecular clock calibration and cautions against overreliance on controversial or insufficiently imaged fossils for genus-level calibrations.

By applying confocal fluorescence microscopy and total-evidence phylogenetics, this study clarifies the taxonomy of two Cretaceous tardigrade fossils: Beorn leggi is reinterpreted as a hypsibiid, and Aerobius dactylus gen. et sp. nov. is introduced within Hypsibioidea. Incorporating Beorn leggi as a calibration point refines divergence time estimates, indicating a Cambrian origin of crown tardigrades, Carboniferous diversification within Eutardigrada, and Lower Jurassic diversification in Heterotardigrada, with independent terrestrialization and convergent acquisition of cryptobiosis. The findings reconcile morphological stasis (Beorn leggi) with evidence for morphological innovation (Aerobius dactylus), illuminating claw evolution and limb functional differentiation. Future work should prioritize discovery and high-resolution imaging of additional amber inclusions, expanded genomic/transcriptomic sampling across Hypsibioidea and other clades to enable finer-scale calibrations, and integrative analyses to reconstruct the tempo and mode of cryptobiosis evolution and body plan miniaturization.

• Preservation constraints: Internal structures, especially the buccal apparatus, were not visualized in both fossils; Beorn leggi lacks informative buccal characters needed for subfamily or genus-level placement; Aerobius dactylus hind-leg claw orientation and incomplete extension limited confident determination of exact claw type and precise measurements.
• Calibration sensitivity and priors: Deep-node posterior age distributions overlapped with priors, making it difficult to disentangle data-driven signal from prior influence; wide HPD intervals persist for several nodes.
• Taxon and data sampling: Phylogenomic analyses lacked sufficient hypsibiid transcriptomes to apply family-level calibrations; reliance on 18S/28S rRNA (with partial 28S coverage) may limit resolution at shallow nodes.
• Generalizability: Only two fossils from a single amber deposit were analyzed; conclusions about morphological transitions (e.g., claw evolution) would benefit from additional specimens.
• Comparative fossil constraints: Caution remains regarding the use of Milnesium swolenskyi for genus-level calibration due to missing internal morphological imagery, potentially affecting specific node placements in future studies.