Sedimentological interpretations

As briefly mentioned in a previous section, Bell et al. (Reference Bell, Fanti, Acorn and Sissons2013) and Kimitsuki et al. (Reference Kimitsuki, Rodriguez, Sullivan, Zonneveld, Sissons, Bell, Campione, Fanti and Gingras2024) interpreted the fossil-bearing interval at the Red Willow Falls site as overbank and crevasse splay deposits laid down in a meandering channel belt. Indeed, the overall fining-upwards sequence and presence of thick mud beds are common features of meandering river sedimentation (e.g., Jackson, Reference Jackson and Miall1977; Halfar et al., Reference Halfar, Riegel and Walther1998; Viseras et al., Reference Viseras, Henares, Yeste, Garcia-Garcia, Ghinassi, Colombera, Mountney, Reesink and Bateman2018). Vertically aggrading crevasse splays commonly form as a result of repeated deposition of normally graded sand-sheets through flooding events (e.g., Burns et al., Reference Burns, Mountney, Hodgson and Colombera2017; Gulliford et al., Reference Gulliford, Flint and Hodgson2017).

On the other hand, the general lack of clear indicators of pedogenic alteration, such as rooting, mottling, soil horizons, and soil structures (e.g., Retallack, Reference Retallack1988; Tabor et al., Reference Tabor, Myers, Michel, Zeigler and Parker2017), suggests that the depositional environment did not experience significant subaerial exposure. The absence of pedogenic alteration is commonly found in submerged channel levees (e.g., Fielding, Reference Fielding1984; Farrell, Reference Farrell1987) or crevasse deltas, where the splay elements prograde into a standing body of water (e.g., Farrell, Reference Farrell1987, Reference Farrell2001; Smith and Pérez-Arlucea, Reference Smith and Pérez-Arlucea1994; Plint et al., Reference Plint, Helm and Lockley2023). Notably, some sedimentologists treat crevasse deltas as synonymous with crevasse splays that form under subaerial conditions (e.g., Colombera and Mountney, Reference Colombera and Mountney2021) because both are geomorphologically similar and result from the same mode of sediment transportation to the floodplain (i.e., levee breach during flooding). However, we consider the distinction between crevasse deltas and crevasse splays to be crucial because paleohydrology has a strong influence on the distribution of plants and animals, and therefore on the distribution of their traces in continental environments (e.g., Hasiotis, Reference Hasiotis and Hasiotis2002, Reference Hasiotis and Miller2007; Gibert and Sáez, Reference Gibert and Sáez2009; Genise, Reference Genise2017).

The Cumberland Marshes in Saskatchewan, Canada, provide a modern analogue of the depositional setting that may have existed in the Red Willow Falls area in the Late Cretaceous. The Cumberland Marshes were drastically altered when the “Old Channel” of the Saskatchewan River was clogged by an ice jam in 1873, and the river’s discharge was forced into a smaller channel system running through a wetland north of the Old Channel (Smith et al., Reference Smith, Cross, Dufficy and Clough1989). Overwhelmed by the sudden increase in discharge, the northern channels experienced extensive levee breaching, which delivered sediment-laden water into the adjacent wetlands. Initially, sediment and water were transported through sheetfloods, which became progressively channelized as the splays evolved (Smith et al., Reference Smith, Cross, Dufficy and Clough1989; Pérez-Arlucea and Smith, Reference Pérez-Arlucea and Smith1999).

According to Smith and Pérez-Arlucea (Reference Smith and Pérez-Arlucea1994) and Farrell (Reference Farrell2001), heterolithic deposits with minor rooting are found on the levees of both trunk and crevasse channels, and in the “prodelta” (sensu Smith and Pérez-Arlucea, Reference Smith and Pérez-Arlucea1994) or “distal mouth bar complex” (sensu Farrell, Reference Farrell2001) of the crevasse splay/delta. The heterolithic strata represent a repeating fining-upwards sequence (described as rhythmites), with each bed measuring a few millimeters to 30 cm thick and having a sharp erosional base overlain by ripple-laminated intervals and graded laminates (Farrell, Reference Farrell2001). Heterolithic rhythmites form during flooding events, in which coarse sediments are deposited first, and grain size then decreases under waning flow conditions (Farrell, Reference Farrell1987; Smith et al., Reference Smith, Cross, Dufficy and Clough1989).

Despite having similar sedimentological attributes, submerged levee deposits and crevasse deltas can be differentiated based on geomorphology. The former are ribbon-like and occur parallel to the channel with which they’re associated, whereas the latter are sediment lobes extending lateral to the trunk channel (e.g., Fielding, Reference Fielding1984; Pérez-Arlucea and Smith, Reference Pérez-Arlucea and Smith1999; Farrell, Reference Farrell2001). Unfortunately, determining the depositional setting for the Red Willow Falls strata through geomorphology is challenging because the exposure is restricted to a narrow area (< 50 m wide) that has been eroded by the present-day Redwillow River. On the other hand, we interpret the homolithic trough cross bedded sand to the south of the interbedded sandstone and mudstone to be crevasse channel fill (cf., Farrell, Reference Farrell2001), while the interspersed mud lenses throughout the outcrop section most likely represent mud plugs of abandoned distributary channels (cf., Tye and Coleman, Reference Tye and Coleman1989). These are a sedimentological constituent of crevasse splays and deltas (e.g., Smith et al., Reference Smith, Cross, Dufficy and Clough1989; Farrell, Reference Farrell2001); thus, we favor the interpretation that the Red Willow Falls strata formed as crevasse deltas rather than as levee deposits.

Fossil interpretations

Our reanalysis has identified the following notable characteristics of the specimens described here: (1) the fossils protrude about 1.5–2 mm from the base of the sandstone; (2) all of the fossils are in life orientation, with their ventral sides facing down; (3) all of the fossils have their anterior ends pointing into the current, mostly within 30° of the mean current-flow direction; and (4) the limb impressions exhibit features expected from live insect larvae and do not indicate disarticulation, dislocation, or abnormal positioning of the limbs (Figs. 2, 3). We are unaware of abiotic processes that can orient 16 insect bodies in the same direction and body position (i.e., ventral side down), and subsequently force them into the underlying substrate. In a recent experiment, Storari et al. (Reference Storari, Salles, Fonseca, Saraiva and Rodrigues2025) studied the postmortem processes that operate on modern mayfly and dragonfly larvae in a tank setup. According to them, only 47% of the deceased mayflies were found with the dorsal side up, whereas 36% of the dragonfly larvae were in dorsal-side-up position.

On the other hand, according to Vallon et al. (Reference Vallon, Schweigert, Bromley, Röper and Ebert2015), elongate objects (they gave fish carcasses and logs as examples) that are subjected to a current have a tendency to become oriented parallel to the current direction, with whichever end is closer to the center of gravity pointing upstream. However, when the objects that have settled into this rheotactically stable position are disturbed, they will regain stability by flipping over (Vallon et al., Reference Vallon, Schweigert, Bromley, Röper and Ebert2015). Therefore, while it is physically possible for a set of insect carcasses to become oriented in the same direction under a unidirectional flow regime, they would be unlikely to all adopt the same body position in this situation, with each individual’s ventral side facing down. It is also unlikely that overlapping pairs of larval carcasses would have been preserved in three dimensions as seen in UALVP 53483-4 (Fig. 3.7). Instead, a more parsimonious interpretation is that the specimens are not body fossils as suggested by Bell et al. (Reference Bell, Fanti, Acorn and Sissons2013) but rather represent traces of insects that shallowly buried themselves in the substrate, while orienting their heads into the current flow. The two overlapping body impressions in UALVP 53483-4 can be explained as traces successively produced by a single live insect repositioning itself within the substrate.

Notably, Knecht et al. (Reference Knecht, Benner, Dunlop and Renczkowski2024) established four criteria for identifying body impressions made by living organisms, and the specimen described here meets three of them. The three criteria observed are that the impressions are exclusively of the ventral surface, the limbs represented in the impressions are those expected to make contact with the substrate, and the limb impressions do not indicate postures that exceed the range of motion likely present in the living animal. The fourth criterion, which these fossils do not meet, is that body impressions may be accompanied by trackways approaching and/or departing from them. This fourth criterion would of course not be met if the living organism flew or swam to the location in which the body impression occurs or walked to the location on a surface that was unsuitable for preserving trackways and/or subsequently eroded away. Although the taphonomic study by Storari et al. (Reference Storari, Salles, Fonseca, Saraiva and Rodrigues2025) revealed that undisturbed larval bodies may long remain in articulation, the current flow experienced by the larvae described herein would likely have sufficed to cause major amputations or disarticulations if the larvae had been deceased. The legs are among the first parts of the body to disarticulate under mechanical disturbance, along with antennae, gills, and caudal filaments in the case of mayfly larvae, and labia in the case of dragonfly larvae (Storari et al., Reference Storari, Salles, Fonseca, Saraiva and Rodrigues2025).

As an alternative to interpreting the body impressions as resulting from active self-burial, one might argue that insect larvae clinging to the sediment surface at a specific orientation against a rapid current could have been pressed into the substrate through the weight of the sediment that piled up on top of them. However, insects are less dense than the surrounding sand/porewater medium, and the relatively broad ventral surfaces of the larvae described here would have prevented their bodies from sinking into the underlying mud. The lack of load structures (e.g., Owen, Reference Owen2003) also indicates that the substrate was not soft enough to be deformed under the weight of the larvae and/or that of the sediment itself. Thus, active self-burial by the larvae is the only mechanism available to explain why these fossils are found under the bedding surface.

Shallow self-burial behaviors are observed in modern dragonfly and mayfly larvae, both of which have similar body plans to the fossils described here. For example, some stream-dwelling dragonfly naiads, particularly those in the family Gomphidae, are known to burrow shallowly in the substrate to ambush prey and hide from predators (e.g., Chamberlain, Reference Chamberlain and Frey1975; Tennessen, Reference Tennessen2019). Huggins and DuBois (Reference Huggins and DuBois1982) observed the behaviors of two burrowing dragonfly naiads (Progomphus obscurus (Rambur, Reference Rambur1842) and Gomphus externus Hagen in Selys, Reference Selys1858) and found that the larvae always directed their heads upstream to stabilize themselves before burrowing. These larvae burrow in sand- to silt-sized sediments, although P. obscurus prefers coarser-grained substrates than G. externus. They also found that the burrowing depths were shallow, with P. obscurus larvae “never observed to burrow deeper than 2 cm”, and G. externus “normally burrowing only 3–5 mm deep” (Huggins and DuBois, Reference Huggins and DuBois1982, p. 5–6). Notably, Huggins and DuBois (Reference Huggins and DuBois1982) observed that the larvae of P. obscurus would move around within the substrate after burial, leaving an irregular trail. Such behavior was not observed in G. externus larvae.

On the other hand, Soluk and Craig (Reference Soluk and Craig1988) documented suspension-feeding mayfly naiads (Ametropus neavei McDonnough, Reference McDonnough1928) that bury themselves in the sediment in a ventral-down position and extend their forelegs to capture food particles (mainly detritus) or remove unwanted sediment particles from around their mouths (Fig. 6). Burial depth is extremely shallow, and the larvae keep their eyes projecting above the sediment surface (Clifford and Barton, Reference Clifford and Barton1979). The larvae also orient themselves facing into the current to trap suspended food particles near their mouthparts and to allow water to flow over their abdominal gills. Ametropus neavei occur natively in the bed of the Athabasca River in Alberta, typically on very fine-grained sand (average grain size of ~100 μm) substrate on the downstream side of sandbars and on mud–sand substrates in deeper waters (Clifford and Barton, Reference Clifford and Barton1979). These mayfly nymphs are documented to be “excellent swimmers”, although they spend most of their time stationary in the substrate under lab conditions (Clifford and Barton, Reference Clifford and Barton1979, p. 856–857). A few other mayfly species have been reported to bury themselves in the sediment (Riek, Reference Riek1973; Courtney, Reference Courtney and Giberson2018), although details about their behavior are not available. It is also noteworthy that some “burrowing” mayfly naiads sometimes produce poorly defined burrows with unstable walls (Bae and McCafferty, Reference Bae, McCafferty, Corkum and Ciborowski1995), which may collapse over the larvae through current action.

Figure 6.

Schematic drawing of modern suspension-feeding mayfly larvae, Ametropus neavei. Soluk and Craig (Reference Soluk and Craig1988) observed captive larvae of this species burrowing in the substrate with their heads oriented into the current (shown by blue arrows), as they captured suspended particles in pits excavated near their heads. A similar burrowing strategy may have been used by the producers of the body impressions from the Wapiti Formation. Image redrawn and modified from Soluk and Craig (Reference Soluk and Craig1988, p. 641, fig. 2) with permission number 6307301195953.

Although we interpret the specimens described here as trace rather than body fossils, we agree with Bell et al. (Reference Bell, Fanti, Acorn and Sissons2013) that they are attributable to mayfly naiads rather than to dragonfly naiads, which have a broadly similar overall body plan. This is based on the fact that the impressions do not show the prominent labium found in dragonfly naiads (e.g., Kukalová-Peck, Reference Kukalová-Peck2009; Schädel et al., Reference Schädel, Müller and Haug2020; Storari et al., Reference Storari, Salles, Fonseca, Saraiva and Rodrigues2025) but do show lobe-like structures in the posterior lateral region resembling the tracheal gills of mayfly naiads (e.g., Sroka et al., Reference Sroka, Godunko and Prokop2023) (Fig. 3.5, 3.6). Furthermore, the presence of only two pairs of leg impressions in most of the fossils described here is consistent with the tendency of some mayfly naiads, but not dragonfly naiads, to hold the forelegs clear of the substrate and use them for feeding purposes (Fig. 6).

While we are less certain about the correct identification of the tracemakers at lower taxonomic levels, the suspension-feeding behavior of modern Ametropus naiads hints that the tracemakers could have been referable to Ametropodidae rather than to cf. Heptageniidae as suggested by Bell et al. (Reference Bell, Fanti, Acorn and Sissons2013). Extant Ametropodidae have highly reduced forelimbs and seven pairs of ovate tracheal gills, and the impressions described herein indicate similar features in the tracemakers (e.g., Allen and Edmunds, Reference Allen and Edmunds1976; Wang et al., Reference Wang, Zhou and Zhou2013; Memishishi et al., Reference Memishishi, Bikashvili, Japoshvili and Mumladze2023). However, the impressions do not capture the morphology of the tracemakers completely enough for them to be definitively identifiable as ametropodids.

Based on our sedimentological interpretations and our observations of the fossils found in association with tractional sedimentary structures (i.e., tool marks), we propose the following taphonomic scenario: (1) the depositional event that preserved the fossils took place in a flooded overbank setting adjacent to a breach in the levee of a river’s trunk channel; (2) outflowing water from the trunk channel created hydrologic conditions optimal for suspension-feeding mayfly larvae (likely belonging to Ametropodidae or a similar group), prompting them to bury themselves in the muddy substrate with their bodies oriented against the current (cf., Soluk and Craig, Reference Soluk and Craig1988); (3) during seasonal flooding and increased fluvial discharge, sand-sized sediment was transported to the study area; (4) rapid deposition of sand preserved the impressions of the mayfly larvae as part of a crevasse delta deposit, possibly on the lee side of a migrating mouth bar complex (cf., Clifford and Barton, Reference Clifford and Barton1979).

Type ichnospecies

Kalyptichnus belli n. isp. by monotypy.

Diagnosis

A small, bilaterally symmetrical, trace fossil preserved in positive hyporelief and oriented subparallel to the bedding plane. The trace fossil shallowly projects out of the bottom of the bedding surface with sheer, locally overhanging edges. It consists of a small subcircular to oval anterior region, a broad middle region with two pairs of elongate transverse imprints projecting laterally at oblique angles to the fossil’s long axis, and a posterior region that gradually tapers towards the end. An elongate terminal impression is commonly present.

Etymology

Derived from the Greek words “Kalyptos (Καλυπτός)” meaning “covered”, “concealed” or “hidden”, and “Ichnos (Ίχνος)” meaning “trace”.

Remarks

Kalyptichnus n. igen. is a distinct ichnogenus for shallow excavations of arthropods, representing three-dimensional structures that closely reflect the form of much of the tracemaker’s body. Excavations of this kind can be differentiated from many arthropod (e.g., insect or chelicerate) resting traces, such as Alacranichnus, Cardinichnus, Chliceratichnus, Hedriumichnus, Inmontibusichnus, Narragansettichnus, and Rotterodichnium, which are primarily surface traces of tracemakers that did not fully dig themselves into the substrate (e.g., Braddy and Briggs, Reference Braddy and Briggs2002; Lucas et al., Reference Lucas, Lerner and Voigt2013; Benner et al., Reference Benner, Knecht and Engel2015; Dalman and Lucas, Reference Dalman and Lucas2015; Getty et al., Reference Getty, Sproule, Stimson and Lyons2017; Knecht et al., Reference Knecht, Benner, Dunlop and Renczkowski2024). These trace fossils are expressed as shallow, poorly defined impressions of a portion of the ventral surface of the tracemakers, often limited to parts of the abdomen and the appendages. Furthermore, these ichnogenera typically differ from Kalyptichnus n. igen. in lacking a well-defined circular anterior region. Tonganoxichnus is another shallow arthropod resting trace fossil that sometimes features exceptional preservation of a ventral body impression of the tracemakers (Mángano et al., Reference Mángano, Buatois, Maples and Lanier1997, Reference Mángano, Labandeira, Kvale and Buatois2001). However, Tonganoxichnus is diagnosed based on the presence of anteriormost appendage imprints that are parallel to the main axis and common abdominal appendage imprints (Mángano et al., Reference Mángano, Buatois, Maples and Lanier1997; Benner et al., Reference Benner, Knecht and Engel2015), both of which are absent in Kalyptichnus n. igen. Conversely, Tonganoxichnus often lacks a well-defined subcircular anterior region, and the three pairs of anteromedian appendage imprints (which is positionally equivalent to the two pairs of transverse imprints in the type ichnospecies Kalyptichnus belli n. igen. n. isp.) are short and ellipsoidal. On the other hand, Tripartichnus (particularly the type ichnospecies T. triassicus Vallon and Röper, Reference Vallon and Röper2006) has a similar gross overall appearance to Kalyptichnus n. igen., but Tripartichnus is diagnosed based on its three-part morphological division into three parts, with the anterior part broadest, the middle part medium in width, and the posterior part narrow (Vallon and Röper, Reference Vallon and Röper2006). By contrast, Kalyptichnus has a relatively small anterior region that is typically narrower than, or at most similar in width to, the posterior region. Tripartichnus also lacks distinct transverse imprints as in Kalyptichnus n. igen.

Some examples of arthropod cubichnia that constitute shallow excavations include Rusophycus, Selenichnites, and Surculichnus, all of which differ distinctly in outline from Kalyptichnus n. igen. (overall coffee-bean shape in Rusophycus; broad, crescentic anterior region in Selenichnites and Surculichnus) (e.g., Gand et al., Reference Gand, Garric, Schneider, Walter, Lapeyrie, Martin and Thiéry2008; Knecht et al., Reference Knecht, Benner, Rogers and Ridge2009; Buhler and Grey, Reference Buhler and Grey2017). Kalyptichnus n. igen. is similar to Solusichnium, a shallow, suboval excavation that represents the impression of a marine decapod tracemaker (O’Brien et al., Reference O’Brien, Braddy and Radley2009). However, Kalyptichnus n. igen. differs from Solusichnium in not being suboval in outline and in having a divergent overall body plan: the anterior imprint of Solusichnium is circular to arrow-shaped with two pairs of linear (antennal) imprints, whereas that of Kalyptichnus is consistently rounded and lacks linear imprints; Kalyptichnus n. igen. lacks the narrow medial imprint seen in Solusichnium; and Solusichnium has only one pair of posteriorly directed lateral imprints with bifurcating ends, whereas Kalyptichnus has two pairs of non-bifurcating transverse imprints that form an x-shaped configuration. In addition, Solusichnium has a posterior region preserved in convex epirelief reflecting the use of the uropod to push clear of the substrate, which is absent in our specimen.

According to Minter et al. (Reference Minter, Braddy and Davis2007) and Bertling et al. (Reference Bertling, Buatois, Knaust, Laing and Mángano2022), an ichnotaxobase should reflect the morphology of the trace fossil, which in turn is a function of tracemaker anatomy, tracemaker behavior, the nature of the substrate, and arrangement of subunits. The differences between Kalyptichnus n. igen. and Solusichnium mainly arise from the contrasting body plans of the tracemakers (hexapods and decapods, respectively) and the inclusion of an escape structure produced by action of the uropod in Solusichnium. In addition, although both ichnogenera represent shallow excavations, Solusichnium is primarily an escape trace related to a storm event with rapidly shifting sediments, whereas Kalyptichnus n. igen. is a dwelling trace related to less dynamic bottom conditions. We deem these differences significant enough to justify separating the two types of trace fossil at the ichnogenus level. The diagnosis for Kalyptichnus n. igen. is broadly formulated to accommodate similar trace fossils that may be produced by other insect larvae.

Kalyptichnus belli new ichnospecies

Figures 2, 3

Holotype

UALVP 55222-1.

Paratypes

UALVP 53483-1, UALVP 53483-7, UALVP 55221-1.

Materials

Sixteen trace fossils preserved on sandstone slabs UALVP 53483, UALVP 55221, and UALVP 55222.

Diagnosis

Kalyptichnus n. igen. with two pairs of transverse imprints in the middle region. The anterior pair of transverse imprints are directed anterolaterally and the posterior pair posterolaterally, forming an X shape. The posterior region can be subdividedinto a single posterior-middle region and paired posterior-lateral regions, all of which may exhibit segmentation in exceptionally preserved examples.

Occurrence

The type specimens occur in the Upper Cretaceous (upper Campanian) Wapiti Formation of northwestern Alberta, Canada.

Description

The trace fossils are preserved in positive hyporelief, and protrude out of the bedding surface to depths of 1.5–2.0 mm. The trace fossils measure ~14 mm in length, not including the terminal imprint, and 4–6 mm in width excluding the lateral imprints. The anterior region is generally wider than long, with a length of 2–3 mm and a width of ~3 mm. The lengths of the transverse imprints are highly variable, ranging from 2.5–8.5 mm. The anterior transverse imprints diverge from the midline at 35–80° (usually ~45°), whereas the posterior transverse imprints diverge at between 140–160°. The posterior region is antero-posteriorly elongated, with a length of 7–9 mm and a width of 4–5 mm. The posterior-lateral regions are ~1 mm in width. Where present, the terminal imprint can be as long as 7 mm.

Etymology

Named after Dr. Phil R. Bell, now of the University of New England in Armidale, NSW, Australia, in recognition of his having led the original description of the fossils redescribed herein.

Remarks

Kalyptichnus belli n. igen. n. isp. is a distinctive trace fossil that preserves three body segments and two pairs of appendages, interpreted to be a shallow burial trace created by an insect larva. Although insects have three pairs of walking legs, only two pairs are interpreted as having been used to dig into the substrate during formation of K. belli n. igen. n. isp. traces, whereas the anteriormost pair was raised dorsally. Such self-burial behavior is observed in modern mayfly naiads belonging to Ametropodidae (Clifford and Barton, Reference Clifford and Barton1979; Soluk and Craig, Reference Soluk and Craig1988). Where multiple individuals are preserved in association with current indicators, all of them are similarly oriented, with their heads pointing in the inferred upstream direction.