Introduction

Many animals, including echinoderms, are able to autotomize parts of their body usually as a defense strategy against predators1. It has been argued that shed appendages, which sometimes display vigorous post-autotomy movements, reduce the animal’s mortality in two major ways: (1) they enable the animal to break away from predators that have grasped it, and (2) divert the attention of the predators away from the vulnerable body parts2.

Crinoids commonly referred to as sea lilies, a group of echinoderms that is subject to a high predation pressure, have remarkable ability to autotomize and regenerate their appendages3,4,5,6. It has been shown that arm autotomy in these echinoderms is achieved through the nervously mediated (L-Glutamate invoked) destabilisation of collagenous ligamentary fibres at specialized autotomy planes, namely (crypto)syzygial articulations7. Although considerable effort has been devoted to study autotomy in crinoids8,9,10,11, little attention has been paid to post-autotomy thrashing behaviour. Wilkie et al.7 in their study on living stalkless comatulids (feather stars) only briefly reported that: “after autotomy induced in both intact animals and isolated arms, the detached distal portion of the arm showed rhythmical cycles of flexion and extension in the oro–aboral plane”.

To further examine thrashing behaviour in extant sea lilies, and explore its ichnological potential we conducted neoichnological aquarium experiments using the stalked crinoid Metacrinus rotundus Carpenter, 1884. We then analyzed samples with articulated arm fragments of the oldest (Early Triassic) stem-group isocrinids (holocrinids) in order to identify evidence of this behaviour in the fossil record.

Aquarium experiment

Specimens of living stalked crinoids of Metacrinus rotundus were used in this study. Sampling and handling of these specimens followed described procedure12,13,14. Crinoids were dredged from Suruga Bay (near the town of Numazu, Shizuoka Prefecture; Japan; ~ 35° 3′ N, ~ 138° 48′ E, ~ 140 m depth) from the sea bottom using a 90-cm-wide naturalist dredge with a net. Then, the living specimens were transferred to an experimental seawater tank in the Nagoya University Museum. The aquarium was maintained at a constant seawater temperature (~ 16 °C) in darkness and under circulation provided by a water pump. After few weeks of acclimatizing the crinoids, a box (~ 40 × 30 cm) floored with fine-grained sand was placed, which was smoothed before an autotomized arm was introduced. Autotomy of crinoid arm was induced following described procedure15,16. Neoichnological experiments were repeated several times using different arms from different individuals (electronic supplementary movies S1S4).

For comparison purposes, a series of aquarium experiments was made on production of transport-induced sole markings left on the sediment surface (fine-grained sand) by isolated dead arms being dragged (with the pinnules facing upstream or downstream; electronic supplementary movies S5, S6) by a current (0.35 cm/s) adjusted by a water pump that was placed at the bottom of the aquarium. At the lower or higher velocity, the arms (which were placed at a distance of ~ 25 cm from the pump) were not moving or were lifted above the sediment surface, respectively, leaving no traces. These experiments were conducted at the Laboratory of Experimental Taphonomy at the Faculty of Natural Sciences of the University of Silesia in Katowice, Sosnowiec, Poland.

Gypsum casts from both experiments are deposited at the Institute of Paleobiology, Polish Academy of Sciences (ZPALV.42ICH).

Fossil samples

A Lower Triassic (lowest Spathian) slab of thin-bedded silty limestone of the Thaynes Group (west of Paris, Idaho, USA) preserving five isolated fragments of arms belonging to one of the oldest (if not the oldest) post-Paleozoic holocrinid taxon (Holocrinus sp.) deposited in the collection of the Université de Bourgogne, Géologie Dijon, France (UBGD 30564) was investigated17. These arm fragments are of different lengths and lack distal arm tips and some distal pinnules (see figs. 2 and 3 in17).

Results

Traces of recent crinoid arms

Following autotomy of the arm, within a few seconds, its activity increased, i.e., it started to thrash mostly in the oral–aboral plane. Both frequency and style of flexions (sluggish writhing and violent flicks; electronic supplementary movies S1S4) and the duration of movement varied (~ several hours to up to about 7 days). However, only in the first hours after autotomy, detached arms showed the highest activity. The frequency and amplitude of motions decreased with time; i.e., after several hours autotomized arms remained in place; their activity could still be recorded but they did not produce any traces. In the first hours after autotomy, each autotomized arm produced two major types of traces on the sediment surface, which can be closely associated with each other (Fig. 1):

  1. 1.

    straight to slightly arched grooves [1.5–3 cm long (mean: 2.1 cm); 0.2–0.6 cm wide (mean: 0.3 cm); ~ 0.1 cm deep) usually arranged in radiating groups (Fig. 1a–c, g–i); the grooves may be inclined at different angles to each other (the angles between adjacent grooves ranging from 9° to 28°) and are separated from each other by a distance of 0.9–2.8 cm (measured in the widest distance between two ends of neighbouring grooves). The length-to-width ratio of these grooves ranges from 3.8 to 10.9 (mean: 6.8). These traces were left by a rotating and more or less rhythmic movement of the flicking arm that placed the most pressure on the substrate with its median-distal arm part.

  2. 2.

    sets of thin and parallel (locally curving) furrows [0.3–1.7 cm long (mean: 1.1 cm), 0.04–0.17 cm wide (mean: 0.1 cm), ~ 0.03 cm deep) (Fig. 1d–f, j–l)]; these sets may be inclined at different angles to each other, locally forming a herringbone pattern. The length-to-width ratio of these furrows ranges from 4.1 to 33.6 (mean: 11.4). These traces were made by pulling the pinnules along the substrate.

Figure 1
figure 1

Traces produced by autotomized arms of Metacrinus rotundus. (ac) Straight deep grooves arranged in radiating group; (df) a few sets of straight parallel grooves and furrows inclined at different angles to each other; (gi) two large arched grooves; (jl) small short parallel furrows. (a, d, g, j) Photographs of sediment surface; (b, e, h, k) photographs of gypsum casts; (c, f, i, l) false-color depth maps of gypsum casts (for source data see the Open Science Framework, https://osf.io/b8zu2/project ”3D models of crinoid traces”, files SOM7_ichno_3D, SOM8_ichno_3D, SOM9_ichno_3D); (m) color scale of elevation. Large and small arrows indicate deep grooves left by the arm and shallow furrows made by pulling the pinnules along the substrate, respectively. Scale bars = 1 cm.

Current-produced sole markings, resulting from the more or less continuous contact of the arm with the sediment surface (electronic supplementary movies S5, S6), display a very different morphology (Fig. 2). These grooves are continuous, long (4–6 cm), generally linear, deep (~ 0.2 to 0.4 cm) and run parallel to the flow direction.

Figure 2
figure 2

Current-induced sole markings (arrows) left on the sediment surface. (a) Photograph of sediment surface; (b) photograph of gypsum cast.

Trace fossils

Close examination of Lower Triassic slabs from the Thaynes Group (Spathian, west of the city of Paris, Idaho, USA18) revealed similar traces associated with isolated articulated fragment of arm of Holocrinus sp. (Fig. 3). Given their size, morphology and close association with the crinoid arm, it appears that these traces could have been produced by thrashing movements of shed arm. However, the scarcity and imperfect state of preservation make the erection of the new ichnospecies impossible.

Figure 3
figure 3

Possible trace fossils associated with articulated piece of arm of Holocrinus sp. (UBGD 30564; Thaynes Group, Lower Triassic; collection of Université de Bourgogne, Géologie Dijon, France). (a) Slab preserving articulated fragment of arm and associated traces (arrows); a photograph taken using a camera fixed to a tripod. (b) Slab preserving articulated fragment of arm and arched grooves (curved red arrows) arranged in radiating group and a few sets of indistinct straight parallel furrows (straight small red arrows with a question mark) and analogical traces (in dotted circle) produced by autotomized arms of Recent isocrinid; a photograph taken using a camera attached to an optical microscope. (c) False-color depth map of a slab; (d) false-color depth map of a slab acquired from photographs taken from binocular microscope (elevation has been increased (× 2) to enhance depth contrast); (e) color scale of elevation. For source 3D data see the Open Science Framework, https://osf.io/b8zu2/project ”3D models of crinoid traces”, files SOM10_old, files SOM11_newest. Scale bars = 1 mm; scale bar in dotted circle = 1 cm.

Description

One arm fragment is associated with slightly arched grooves [~ 1.9 to 3.2 mm long (mean: 2.3), ~ 0.5 to 0.8 mm wide (mean: 0.55)] arranged in a radiating group (the angles between adjacent grooves range from 12° to 24°). These grooves are separated from each other by a distance of 1.9–2.6 mm (measured in the widest distance between two ends of neighbouring grooves). Dimensionally, the mean length-to-width ratio of these grooves (4.3) falls well within the range of this ratio for the grooves produced by the median-distal arm part during our neoichnological experiments. These grooves are associated with a few sets of indistinct thin and parallel (locally slightly curving) furrows [1.8–3.2 mm long (mean: 2.4), 0.07–0.13 mm wide (mean: 0.1)] of uncertain origin. Notably, however, the mean length-to-width ratio of these furrows (25.5) falls within the range of this ratio reported in our neoichnological experiments.

Remarks

As no ethological term exists for the trace fossils produced by post-autotomy thrashing movements of shed appendages we herein propose a new ethological category, autotomichnia (from the Greek auto- "self-" and tome "severing", αὐτοτομία). Although thrashing behaviour occurs in a wide range of vertebrate and invertebrate taxa1,2,19,20,21,22, autotomichnia are probably rare in the fossil record, and its identification is difficult if the traces are not found in association with the “producer”. Such cases, however, are not likely to be common given that shed appendages displaying movements attract attention of the predators which attempt to consume them. The only ichnospecies ascribed to activity of crinoids described so far is Krinodromos bentou23. This ichnotaxon, however, in contrast to the presently described traces, refers to unusual locomotion trail, and shows different morphology in the form of two bordering grooves combined with pushed sediment piles, and a central flat area or a narrow winding furrow. Possible explanation for the fossil traces reported herein is that they simply represent sole markings produced by a dead arm moved along by a current. However, these Triassic traces bear no morphological resemblance to the experimentally induced abiotic sedimentary structures (Fig. 1 vs Fig. 2).

Discussion

It has been argued that ichnology can provide valuable insights on the evolutionary origins of behaviors and on many of their functional and adaptive aspects24,25,26. Our experiments showed that autotomized arms of stalked crinoids (isocrinids) display vigorous movements similar to that observed in shed arms of brittlestars21. They produce distinct traces on the sediment surface, which may have some potential to be preserved as trace fossils. This finding opened new perspectives for tracing the origin of thrashing behaviour in crinoids. Indeed, our examination of fossil material from the Thaynes Group suggests that similar traces are associated with isolated articulated fragment of arm of the oldest (~ 250.6 Ma) stem-group isocrinids (Holocrinus). The fact that this arm fragment lacks distal arm tip and some distal pinnules is suggestive that it was autotomized due to predatory attack prior to rapid burial rather than another type of disturbance (e.g., storm). It must be emphasized that although abiotic factors (such as high current velocities induced by storms or other environmental trauma) may lead to arm loss, most autotomy in natural settings results from biotic interactions27. This is consistent with recent data28 which showed that the thruster-produced extreme flow or suction force did not lead to arm loss or any other type of injury in deep-water, stalked crinoid Democrinus sp.

Trace fossils described herein suggest that the origin of thrashing behaviour in crinoids could be traced back to at least the Early Triassic. The question of whether this behaviour could have appeared in the Paleozoic crinoids (independently in several clades or in one lineage and was then inherited by the post-Paleozoic descendants) is presently unclear. However, it should be emphasized that, in contrast to the highly flexible muscular arms of recent and post-Paleozoic crinoids in which syzygial or cryptosyzygial ligamentary articulations specifically designed for autotomy are localized, many Paleozoic crinoids possessed primitive arms with limited flexibility6. Although regeneration is frequently documented in Paleozoic crinoids29, most of them, if not all, have not yet developed autotomy planes6,11,30. This suggests that they lacked autotomy abilities, or at least that they were less specialized in that regard30.

The end-Permian mass extinction profoundly influenced the evolutionary history of crinoids, not only through the demise of major Paleozoic crinoid groups but also through changes in their functional morphology31,32. Notably, post-Paleozoic crinoids are considered to have descended from only a single survivor (i.e., a cladid ancestor—ampelocrinids)33,34. In the Middle Triassic, however, they rebounded and underwent a major radiation resulting in the appearance of several motile taxa showing many anti-predatory morphological and behavioral innovations to increased predation pressure during the Mesozoic Marine Revolution35,36,37,38,39. Holocrinids are among the first crinoids to appear in the aftermath of the end-Permian mass extinction40. These crinoids display many adaptations to benthic and nektonic predators. For instance, they developed specialized rupture points at the distal nodal facets in their stalk, allowing them to free themselves of the sea bottom, crawl with the aid of highly flexible muscular arms and re-attach41,42,43. Furthermore, they elaborated localized autotomy planes (cryptosyzygies) in the arms38,44, which in Recent isocrinid descendants greatly reduce mortality and arm damage16. Herein it is suggested that holocrinids might have also displayed post-autotomy arm thrashing. Interestingly, in some lizards, it has been observed that their wildly thrashing autotomized tails (in contrast to taxa having a much lower rate of tail thrashing) consistently distract attention of predator away from the escaping lizard45,46 and increase predator-handling time, providing additional opportunity to escape47,48. Intriguingly, however, in some lacertid lizards, duration of thrashing between different species, shows little variation, either because it is less costly, or because the underlying physiological pathways are evolutionary conservative49. Notwithstanding, the high numbers of shed lizard tails are commonly found in predator stomachs50,51. All these indicate that caudal autotomy and thrashing behaviour in these vertebrates is an effective antipredatory strategy. Accordingly, Wilkie21 argued that thrashing behaviour in ophiuroids is also likely to be an effective decoy against fish predators. Consequently, we hypothesize that this behaviour in crinoids also probably acts as a defense strategy to distract predator attention, and/or to increase predator handling time, providing additional getaway chances to crinoids. A moving autotomized arm might itself be sufficient attraction as a food source. Accordingly, acquisition of thrashing behaviour in the Early Triassic crinoids may underscore the magnitude of the anti-predatory traits that occurred during the Mesozoic Marine Revolution, which had already started soon after the end-Permian extinction52,53. The thrashing behaviour along with the specific arm construction consisting of two types of articulations (muscular ones playing a great role in locomotion and postural changes by maximizing arm flexibility, and strictly ligamentary (crypto-)syzygial ones, specifically designed for autotomy) have been maintained in isocrinid/comatulid descendants up to the present.

Given the scarcity and imperfect state of preservation of the trace fossils described herein, further ichnological findings are needed to fully test the above hypotheses. Results of neoichnological experiment presented herein highlight the preservation potential of thrashing behaviour of crinoid arms, thus rock slabs preserving detached crinoid arms are worthy of in-depth investigation.

Methods

Movements of the autotomized arms of crinoids were observed using a self-made underwater camera that was placed into a small transparent container with a power bank. A camera was constructed on the basis of small single-board computer—Raspberry Pi 3 device with connected V2 Camera Module and combined with the source of red LED light. Both continuous (~ 30 to 35 min) and time-lapse movies (~ 10 to 11 h; at the frequency: one photograph taken per 1 min) were captured (electronic supplementary movies S1S4). The traces on the sediment were photographed under different light conditions, and then gypsum casts were made. 3D models were acquired with Shining 3D EinScan Pro 2X 3D scanner fixed on a tripod, EXScan Pro 3.2.0.2 software, and then processed with Meshlab 1.3.3, Blender 2.82 and ParaView 5.80 to obtain the false-color depth maps.

Two sets of photographs of the trace fossils associated with of the holocrinid arm fragment were taken under different light conditions [by mean of a camera fixed to a tripod (Fig. 3a) and with a camera attached to an optical microscope (Fig. 3b)]. Two 3D models and false-color depth maps for each set were acquired by means of photogrammetric technique using Visual SFM 0.5.26 and the MeshLab 1.3.3 or Agisoft Photoscan 1.2.0.

Ethics

All experiments on live stalked crinoids were performed and approved by the University of Nagoya, where no special ethics approval is required for the maintenance and handling of this particular species. Nevertheless, our research conformed to the ethical principles of replacement, reduction, refinement and minimization of animal suffering following the guidelines reported in the European Directive 2010/63/EU.