Substrate type and palaeodepth do not affect the Middle Jurassic taxonomic diversity of crinoids

Crinoids are largely considered as good indicators for determining environmental conditions. They are robust proxies for inferring changes in salinity and sedimentation rate and for inferring substrate type. Some crinoid groups (e.g., certain comatulids, cyrtocrinids, millericrinids) have a depth preference, thus, making them useful for palaeodepth estimation. The hypotheses that crinoid distribution is substrate-dependent (rock type) or palaeodepth-dependent is tested here based on (a) archival Bathonian-Callovian (Middle Jurassic) crinoid occurrences from Poland and (b) newer finds from five boreholes from eastern Poland. Qualitative data suggests that isocrinids and cyclocrinids occur in both carbonate and siliciclastic rocks. The cyrtocrinids and roveacrinids occur within carbonate rocks, whereas the comatulids are exclusive to siliciclastics. In terms of palaeodepth, most crinoid groups dominate in shallow environments with the sole exception of cyrtocrinids, that are ubiquitous and occur in both shallow (near shore and shallow marine) and slightly deeper (deeper sublittoral to open shelf) settings. The occurrences of the cosmopolitan taxa, Chariocrinus andreae and Balanocrinus subteres (isocrinids), is independent of both substrate type and palaeodepth. Quantitative analyses (Analysis Of Variance; ANOVA) based on substrate type, i.e., substrate-dependency (claystones, sandstones and limestones), and palaeodepth i.e., palaeodepth-dependency (near shore, shallow-marine, mid-ramp and offshore), corroborate qualitative results. Statistical analysis suggest that the distribution of crinoids shows a strong substrate-dependency but not for palaeodepth, although very weak significance (low p value) is noted for near shore and shallow marine settings and crinoid distribution.


INTRODUCTION
The Middle Jurassic (especially Bajocian and Callovian) strata of the epicratonic Poland (Central European province = Submediterranean province) are well-known for their

Previous records
The Middle Jurassic (Callovian) crinoids of the epicratonic basins of Poland have been recorded from the Polish Jura Chain ('A' on Fig. 1) and the Mesozoic margin of the Holy Cross Mountains ('B' on Fig. 1). Early records only mentioned crinoid localities (e.g., Wójcik, 1910;Makowski, 1952;Różycki, 1953). Later, Dayczak-Calikowska (1980) listed several taxa but did not mention the exact locations from where the crinoid specimens were collected or where they were kept (sample repository). Furthermore, neither description, nor illustration, or even basic taxonomic group assignment was provided; most, however, belong to balanocrinids. Radwańska & Radwański (2003) from northern Poland recorded Callovian columnals of Cyclocrinus macrocephalus and Oxfordian cyclocrinid of Cyclocrinus couiavianus. Salamon & Zatoń (2006) from the Callovian marly limestones of the Zalas Quarry in southern Poland ('A' in Fig. 1; see also Table 2) erected a new species of balanocrinid (Balanocrinus hessi); this was later synonymized with B. pentagonalis by Krajewski, Olchowy & Salamon (2019). Salamon & Gorzelak (2007), from the same locality (Zalas Quarry in southern Poland; 'A' in Fig. 1) described few isolated remains of indeterminable cyrtocrinids associated with the cyrtocrinid cup of Dolichocrinus cf. aberrans (see Table 2).  from a collection of ∼1,500 remains of columnals, pluricolumnals, brachials and cups, illustrated 11 crinoid taxa, including some isocrinids, comatulids, millecrinids and cyclocrinids (see Table 2). These specimens came from three Bajocian, four Bathonian and two Callovian localities of the southern part of the Polish Jura Chain and the Mesozoic margin of the Holy Cross Mountains (see . Current re-examination of this material led to the assignment of the columnals described as Millericrinina to the cyrtocrinids, Cyrtocrinida indet. (see Salamon & Feldman-Olszewska, 2018) (Table 2). Salamon (2008a) and Salamon (2008b) Table 2). This collection contained a relatively small number of complete or nearly complete individuals, as a side-effect of the maceration process. Before the maceration process, the samples formed a typical encrinite and consisted largely of crinoids, as is also the case in the present study, yielding only fragmentary crinoids (see Table 2). Additionally, the Middle Jurassic (Bathonian-Callovian carbonates) exposures in southern Poland (Tethyan province, 'E' and 'F' in Fig. 1) have also yielded isocrinids, cyrtocrinids, roveacrinids, and cyclocrinids (see Głuchowski, 1987; see Table 2).

MATERIAL AND METHODS
The archival cores (72 core samples) drilled in eastern Poland (G1-G6 on Figs. 1 and 2) were investigated for crinoids. These cores are stored in the Polish Geological Institute National Research Institute, Warsaw (Poland). The samples selected for the maceration process were selected from the following boreholes at respective depths: (1) Kock IG 2 (core depth: 859.0 m; age: Callovian?); (2) Maciejowice IG 1 (core depth: 1,431.5 m; age: late Bathonian);     Table 2). The present paper also includes a large collection of Callovian crinoids from Poland collected by Mariusz Salamon and Bartosz Płachno (2006-present) and others, mentioned in the Acknowledgements section.
The first step for the current analysis consisted of examining the drill cores under a binocular microscope for crinoids. Thereafter, the carbonate samples were soaked with Glauber's salt. Based on sample cohesion, these were then successively boiled and freezed. The residue was then washed under running tap water and sieved (mesh sizes: Ø1.0, 0.315 and 0.1 mm, respectively). The final step consisted of drying the screened residue at 170-180 • C (following the methodology outlined by Krajewski, Olchowy & Salamon, 2019). Both the sandstone sample Maciejowice IG 1 (core depth: 1,431.5 m; age: late Bathonian) and the carbonate sample Magnuszew IG 1 (core depth: 1,505.7 m; age: late Callovian) were naturally macerated and left outside for 30 days (from January 22nd to February 21st, 2021). As a result of this natural maceration, the samples were partially disintegrated. In addition to this, several thin sections (TS) and polished slabs (PS) were also made. All crinoids were hand-picked from the maceration residue and photographed using a Canon Eos 350D digital camera, LeicaWildM10 coupled with a NikCamPro1 microscope and a Scanning Electron Microscope Philips XL-20. All specimens are housed in the Institute of Earth Sciences of the University of Silesia in Katowice, Poland, and catalogued under registration number GIUS 8-3734. Other Middle Jurassic specimens used in the current study are also housed at the Institute of Earth Sciences of the University of Silesia in Katowice, Poland, and catalogued under: GIUS 8-2510, 8-2569, 8-2571

Statistical methods
The taxonomically standardized dataset was subjected to Analysis Of Variance (ANOVA).
To check for substrate-dependency (Hypothesis 1), the dataset was categorized under claystones, sandstones and limestones (the former two represent siliciclastics, whereas the latter, carbonates). For palaeodepth-dependency (Hypothesis 2), the dataset was categorized under near shore, shallow-marine, mid-ramp and offshore categories. For pairwise comparisons, the Tukey's HSD (honestly significant difference) test was applied.

Taxonomic standardisation
The cyrtocrinids are identified at the specific level based on their cups. In some cases, when the cups are fragmentary or some typical features are not visible, the samples were assigned to the generic level, only. Most of the disarticulated remains of cyrtocrinids (holdfasts, columnals, radials, basals, brachials, etc.) are assigned under Cyrtocrinida indet.
As for comatulid centrodorsals, where cups with basals and radials were not available for additional information, they were also classified at the generic level. The brachial with muscular articulation on the proximal side and syzygial one on the distal side are classified as Paracomatulidae sp. et gen. indet.
The isolated remains of isocrinids, consisting of columnals, pluricolumnals, cirrals, cup plates, and brachials, are classified at the specific level. If they were found in different levels, they are described as Isocrinida indet.
A complete individual of cyclocrinid has not been found so far. Despite this, the uniqueness of their remains (e.g., large cylindrical columnals with peculiar tuberculate facets) allows us to identify them at the specific level (see detailed discussion in Radwańska & Radwański (2003). Głuchowski (1987) assigned some remains from the Pieniny Klippen Belt to Cyclocrinus sp. due to the incompleteness of columnals. Głuchowski (1987) classified all roveacrinids (saccocomids) as Saccocoma sp.; these were recorded either as isolated cup remains or observed in thin sections (TS).
Thus, the dataset used in the present study is taxonomically standardised. The crinoids classified in the present study are given in Tables 1-4. A summary of all data with their respective inferred palaeodepth and species diversity (number of taxa) is given in Table 3 (see also Fig. 3).

RESULTS
The present study identified several crinoid remains of cups, columnals, cirrals and brachials from the 5 boreholes drilled in eastern Poland (Fig. 1). Only two pluricolumnals were recorded. The first one consists of three or four columnals observable on the surface of a rock fragment from the Magnuszew IG 1 core (depth: 1,505.7 m). The second one consists  1 1 1 1 1 1  1 2 1 1 1 1 1 1 1 1 4     of two columnals and comes from the sample of Siedliska IG 1 (depth: 828.77 m). The other echinoderms observed on slab surfaces include numerous partly preserved cidaroid spines, cidaroid interambulacral plates and asteroid marginal plates. The echinoderm remains were also observed in thin sections and on the surface of polished slabs. In addition to these echinoderm remains, bivalves, gastropods and ostracods were observed both on the slab surfaces and in the macerated residues. In the late Bathonian sample Maciejowice IG 1, the remains of carbonized plant remains were also observed (for a summary of the above data see Table 4). In general, the Bathonian and Callovian crinoid assemblages are almost identical (see also Table 1). The other investigated Middle Jurassic crinoids were represented by numerous cups, cup remains, brachials, columnals, pluricolumnals, cirri, cirrals, and holdfasts (for details see Głuchowski, 1987;Salamon & Gorzelak, 2007;Salamon, 2008a;Salamon, 2008b;Salamon, 2008c;Salamon & Feldman-Olszewska, 2018).

State of preservation
The preservation of crinoids was examined only on the non-macerated surfaces of fresh core fragments. These were screened for signs of abrasion, bioerosion traces, chemical alteration of ossicle structure, evidence of epibionts and predation traces. Most of the crinoids display a high degree of disarticulation (almost 100%). However, they are all well preserved and do not show any traces of lateral surface abrasion. Excellent preservation and lack of abrasion are also noticed for columnal articular surfaces; the crenulae, petal floors, perilumens and lumens are complete. The two pluricolumnals also show no signs of abrasion. Additionally, the studied crinoid remains are very rarely bioeroded (bioerosion was only noticed in case of three columnals in the form of small and rounded holes). These may be ascribed to acrothoracican cirripedes, algae, fungi, polychaetes, sipunculans, or to sponge activities (for more details see Salamon & Zatoń, 2006;Zatoń, Villier & Salamon, 2007;. Donovan (1991) suggested that under normal oxygenated conditions, complete crinoid disarticulation takes place within two weeks. However, Brett, Moffat & Taylor (1997) argued that under certain circumstances, complete disarticulation may occur even after a year. In living comatulids, disarticulation starts just after death and by the end of the first week, only isolated calyx and a few arm fragments remain articulated (Ausich, 2001). Under high water temperatures and with increased physical disturbance, the disarticulation process significantly speeds up (see Ausich, 2001;Hess, 2006). Baumiller & Ausich (1992) noted that after 19 days, the isocrinid stem disarticulates into noditaxes and after 22 days, the column segments get disarticulated into isolated columnals. Gorzelak & Salamon (2013) noted that under high-energy conditions (constant transportation), and at room temperature, disarticulation of the crinoid skeleton is nearly complete after 17 days, where only a few articulated arms and cirral ossicles remain intact.
In the present study, the crinoid remains are preserved as cups, isolated columnals, brachials, and cups. Therefore, these remains can confidently be classified as taphonomic types 2 and 3 sensu Brett, Moffat & Taylor (1997). Type 2 includes isocrinids with some remains of cyrtocrinid that undergo rapid postmortem disarticulation, on account of weak sutures. Type 3 comprises of cyrtocrinid cups in which major portions of the skeleton are resistant to disarticulation.
Thus, for the present study, all the data suggest that the recorded crinoids are autochthonous with negligible or no post-mortem transportation. Possibly, after death, they remained for some time on the sea bottom, but were not subjected to significant transportation or reworking before burial, or they were shortly covered by sediment, as suggested by the lack of abrasion surfaces. Bearing this in mind we assume little timeaveraging that might alter the palaeoecological and palaeodepth information. Additionally, no trace of chemical alteration, predation traces and epibionts were noticed, suggesting that the inferred palaeoecological and palaeodepth signals are primary.

Substrate dependency (Hypothesis 1)
The ANOVA test (Table 5a) yielded statistically significant differences among the fauna of the different substrate types (claystones, sandstones and limestones) at p < .001. For pairwise comparisons, the Tukey's HSD (honestly significant difference) test was applied (Table 5b). This yielded a significant difference between both claystones-sandstones and limestones but no significance difference between claystones and sandstones (both being siliciclastics; Table 5b).

Palaeodepth-dependency (Hypothesis 2)
The comparison between Palaeodepth parameters (near shore, shallow-marine, mid-ramp and offshore) did not yield any statistically significant results (Table 5c). The pairwise comparisons, Tukey's HSD test (Table 5d) retained a statistically significant difference only between the first two variables (Near shore and Shallow marine); although, the p value is very low (p > 0.01).

Depositional environment of the eastern part of the epicontinental Polish Basin
Upper Bathonian and Callovian of the eastern part of Polish Basin are represented by organodetritic, mostly crinoidal limestones with abundant limonite. There are grainstones and packstones which are dominated by echinoderm plates, less frequently appear echinoid spines, bryozoans, bivalves and brachiopods, occasionally gastropods, foraminifera, ammonites and belemnites are present. Quartz grains are frequently an admixture. In places also ferruginous ooids are present. Sedimentary environment is interpreted as a mid-ramp located between normal and storm wave base (Feldman-Olszewska, 2018).

Depositional environment of the Łuków glacial draft
The Łuków glacial drift exposed in eastern Poland ('C' in Fig. 1), is a 30 m-thick Callovian unit containing black clays with carbonate concretions (see also Mizerski & Szamałek, 1985), deposited in an offshore setting (e.g., Olempska & Błaszyk, 2001;Kaim, 2004). However, recently, based on the presence of a high diversity asteroid assemblage, a varied environment (ranging from coastal to deep-shelf) was proposed (Villier, 2008). This unusually high crinoid diversity is likely a reflection of ossicles being transported by currents from shallow-to deep-water offshore environments (Villier, 2008). However, according to Salamon (2008a), based on the ecology of crinoids, the Łuków clays was deposited in a relatively shallow environment and contrary to asteroids, they were not transported. Based on the presence of macro-and micro-remains of land flora, a nearshore environment was also inferred (see also Brand, 1986;Marynowski et al., 2008). Similarly Gedl (2008) using dinoflagellate cysts concluded that these originated presumably from shallow marine environments. Sedimentological investigations of the Aalenian-Bathonian (Middle Jurassic) claystones and shales from central Poland indicate that their sedimentation took place in offshore (below storm wave base) and transition (between normal and storm wave base) zones of the epeiric sea (see Feldman-Olszewska, 2007;Feldman-Olszewska, 2008;Feldman-Olszewska, 2012a;Feldman-Olszewska, 2012b;Pieńkowski et al., 2008). Water depth corresponds to the Cruziana ichnofacies but shallower than the Zoophycus ichnofacies (see Knaust & Bromley, 2012). The same depositional environment is expected for the Callovian clays from Łuków.

Problems with isolated isocrinid classification
The isocrinids are a dominant crinoid group in the Middle Jurassic sediments of Poland. However, their correct classification into families and genera is very complicated. The ideal situation is when complete crinoid findings are available (e.g., Simms, 1989;Hess & Messing, 2011), but these are rare due to the unique structure of their skeletons (see State of preservation). There are of course also isocrinids that have a unique structure of columnals and it is easier to identify them from others. A classic example is that of Balanocrinus subteres or B. pentagonalis (see Klikushin, 1992;Hess, 2014a;Hess, 2014b;Krajewski, Olchowy & Felisiak, 2016;Krajewski, Olchowy & Salamon, 2019;Krajewski, Ferré & Salamon, 2020). Hess (2014a) and Hess (2014b) stated that the columnals of B. subteres could be associated with its isolated calyx elements. Likewise, with remains of Pentacrinites dargniesi or Isocrinus nicoleti display a unique shape and ornamentation of their articular facets, as also P. dargniesi with its characteristic, ellipsoidal cirrals. However, there is an issue with non-balanocrinid isocrinids, that have columnals of a similar morphology, and were erected on the basis of isolated skeletal remains. There are many such forms in the Middle and Upper Jurassic strata of Poland. According to De Loriol (1877-1879), the most similar taxon to Isocrinus amblyscalaris is I. munieri. But there are some noticeable differences between them, also i.e., the height of columnals and the ornamentation of the latera. Hess (2014a) added that Pentacrinus pellati with low columnals alternating with higher ones, and ridges occurring on its latera, is also a very similar form. He concluded that both species are of middle Oxfordian age and therefore may be conspecific. In turn, Bather (1898) suggested that I. amblyscalaris currently from the Oxfordian and Kimmeridgian of Poland is a synonym of I. pendulus. Hess (1975) added that both these species originated from the same middle Oxfordian strata (Bärschwil Fm, NW Switzerland). On the other hand, Klikushin (1992) stated that there is a series of morphological differences between these two species that supersedes Bather's (1898) opinion. The latter author, elsewhere in his monograph, however, proposed to consider I. pendulus, Oxfordian I. amblyscalaris, and Callovian-Oxfordian I. oxyscalaris as junior synonyms of the Oxfordian I. desori. Krajewski, Olchowy & Salamon (2019) differed from this view and emphasized that there are some differences in the facet morphology of the mentioned taxa. Apart from the differences in the articular facets (i.e., the presence of deep furrows between petal floors), the shape of columnals may also be different (compare Hess, 1975, p. 55, text-fig. 8, pl. 19, fig. 4;Klikushin, 1992, pl. 11, fig. 1-7 vs. pl. 13, figs. 8-10; Radwańska, 2005, fig. 8/4;Salamon, 2009, fig. 2H).
Taking into account all of the above, it is difficult to say unequivocally whether there are one or more biological species within the Middle/Upper Jurassic strata of Poland. Therefore, such forms, whose taxonomic affiliation raises serious doubts (Balanocrinus berchteni, Isocrinus bajocensis, I. bathonicus, and I. pendulus), must be evaluated with great caution in any analysis.
One of the previously investigated Callovian sections (the abandoned Wysoka Quarry placed within the Polish Jura Chain; 'A' in Fig. 1), exposes sandy limestones where numerous cyclocrinid columnals were collected from a narrow level. Remains of other crinoids were not found there. Along with these columnals, several ''millericrinid''-like brachial (see Fig. 3) plates were found, which probably belong to the same crinoid as the columnals. Brachials display a proximal synostosis and a muscular distal side, sometimes with a pinnule socket. Both facets are parallel. This seems to confirm Hess's (2008) idea of linking cyclocrinids to millericrinids. On the other hand, the ''bourgueticrinid''-like, strongly abraded cup was found by us in the Czerwieniec locality by one of the authors (BJP). This cup is very similar in shape and size to bourgueticrinid cup that may suggest that cyclocrinids really belong to bourgueticrinids as suggested by Radwańska & Radwański (2003). Wójcik (1910) first listed the occurrence of Cyclocrinus macrocephalus in Poland and the occurrence of columnals in the sands of Filipowice, Polish Jura Chain ('A' in Fig. 1). Głuchowski (1987) noted the occurrence of Cyclocrinus rugosus and Cyclocrinus sp. within the Bathonian and/or Callovian crinoidal limestones of the Tatra Mountains and the Pieniny Klippen Belt ('E' and 'F' in Fig. 1). Radwańska & Radwański (2003) questioned this assignment, claiming that the sketches and photographs are not of sufficient quality, specimens are poorly preserved, and that the 'cyclocrind' were small-sized. Re-examinations of Głuchowski specimens (1987, fig. 13/1, 13/4 and pl. 1, fig. 1-6), stored at the Institute of Earth Sciences of the University of Silesia in Katowice, Poland, prove that they are significantly smaller than the columnals collected from other areas. Additionally, they are cylindrical, low and with a slightly convex latera, suggesting that they may well belong to cyclocrinids. Therefore, the original designations of Głuchowski (1987) are currently upheld. Moreover, Radwańska & Radwański (2003) recognized four cyclocrinid species within the Middle and Upper Jurassic strata of Poland, namely: Cyclocrinus rugosus from the Bajocian, C. macrocephalus from the Callovian, C. areolatus from the Oxfordian, and C. couiavianus from the upper Oxfordian. They stated that all mentioned taxa (with the exception of C. couiavianus) lie within the variability of C. rugosus (Radwańska & Radwański, 2003). Though describing columnals with articular surfaces covered by numerous and irregular tubercles,  included these remains under C. macrocephalus. Reobservation of columnals coming from the same Callovian levels of the Polish Jura Chain, both ornamented by and devoid of tubercles, confirm the suggestion of Radwańska & Radwański (2003) that all these cyclocrinids actually belong to a single species, Cyclocrinus rugosus.

Distribution of crinoids within particular facies. Does it really work?
Hans Hess in several of his papers pointed out that the occurrence of certain species of crinoids may be closely related to or strictly defined by substrate (sediment type) and palaeodepth. He claimed that Chariocrinus andreae inhabited muddy bottoms and its dense colonies occurred in somewhat deeper waters, below fair-weather wave base, i.e., at depths around 10-15 m (Hess, 1972;Hess, 1999a). In this study, Chariocrinus andreae occurs in large amount in both carbonate and siliciclastic rocks. Balanocrinus subteres is one of the most cosmopolitan crinoid ever recorded in the Mesozoic and is known from both shallow and deeper environments (compare Table 3 and Fig. 2). Hess (1999a) claimed that another very common Middle Jurassic crinoid taxon, Isocrinus nicoleti, frequently occurring in Poland, may also occur within both ooid shoals of the carbonate shelf and in slightly deeper environments (intercalating marls and bioclastic limestones). Tang, Bottjer & Simms (2000) recorded I. nicoleti from the shallow-water tidal facies of the Middle Jurassic Carmel Formation in Mount Carmel Junction and also from the open shallow subtidal (carbonate) facies of the Gypsum Spring and Twin Creek Formations of Wyoming (USA). Hunter & Underwood (2009) further noticed that I. nicoleti from the Bathonian of England and France, occurred only in shallow-water sediments, represented by low-energy, inner muddy shelf and outer lagoon, oolitic high-energy shoal system, carbonate/muddy low-energy, marine lagoon complex and in the shallow subtidal zone. All these are actually consistent with current observations that indicates I. nicoleti (Tables 2 and 3) occurred in both carbonate and siliciclastic rocks, although it preferred shallow-water carbonates. Hunter & Underwood (2009) concluded that the distribution of crinoids corresponds quite nicely to particular facies (but see also the comment on this paper by Salamon, Gorzelak & Zatoń (2010) and reply to it by Hunter & Underwood (2010)). This substrate-dependence is also corroborated by the ANOVA test (although at the generic-level) that the crinoid distribution is highly sensitive for carbonates (limestones) as opposed to siliciclastics (Table  5a-5b).
Of interest is the observation on Pentacrinites dargniesi. Hunter & Underwood (2009) also collected P. cf. dargniesi from shallow, muddy and carbonate habitats. P. dargniesi, despite being predominantly found in carbonates, also occurs in offshore claystones of Poland (see Tables 2 and 3). Contextually, it must be keep in mind that P. dargniesi was a pseudo-planktonic crinoid that transported on the lower surfaces of driftwood fragments, and hence, is likely to occur in varied facies (see Simms, 1999;Hess, 1999a).
The most numerous isocrinids within the Middle Jurassic of Poland are balanocrinids (Balanocrinus subteres and B. pentagonalis). Hess (2014a) and Hess (2014b) argued that balanocrinids settled mostly on hardgrounds but they are also be found these in mudstonedominated settings, possibly on swells of the fine-grained bottom. The largest species of the slender crinoid genus (viz., B. subteres) was reported from below storm wave-base (Hess & Spichiger, 2001), where it thrived due to ample nutrient availability and well-oxygenated bottom waters (Hess, 2014a;Hess, 2014b). Klikushin's (1992) noted, as also in the present study, that Balanocrrinus subteres is another most cosmopolitan Mesozoic crinoid species. It occurs in both carbonate and siliciclastic facies and in environments of varying depth, from near shore (close to a beach) to the open shelf, but usually below normal wave base (Table 3 and Fig. 2). This lack of palaeodepth-dependency (although at the generic-level) is also reflected in the ANOVA test that suggests that palaeodepth played no or minor role in the distribution of crinoids (Tables 5c-5d), although very weak (low p value) significance is noted for near shore and shallow marine settings, as discussed above, also (see Fig. 2 and Table 5).
On the other hand, Hunter & Underwood (2009) concluded that the occurrences of balanocrinids (Balanocrinus cf. subteres) were restricted to neritic mudstone facies and brachiopod-rich limestones (i.e., within the offshore, low-energy shelf setting). Hunter & Underwood (2009) also noted that the comatulids show a wide range of distribution from offshore, low-energy shelf up to oolitic high-energy shoal system. In the present study, corroborating previous observations (e.g., (Hunter & Underwood, 2009) (see also Table  5c-5d), the comatulids are restricted to siliciclastic rocks of near shore, below wave-base settings (Table 3 and Fig. 2).
The roveacrinids represented by the genus Saccocoma are exclusive to the shallow marine Bathonian and/or Callovian carbonates of the Tatra Mountains and the Pieniny Klippen Belt (Tethyan province). Kowal-Kasprzyk, Krajewski & Gedl (2020) mentioned saccocomids from the Oxfordian-Kimmeridgian exotic clasts of the Outer Carpathians in southern Poland. Matyszkiewicz (1996) and Matyszkiewicz (1997) mentioned Saccocomacalciturbidites resting on the slope beds of Oxfordian cyanobacterial-sponge carbonate buildups formed in the Polish epicontinental basin (for summary see Fig. 2).
The cyclocrinids are not considered in the present study due to their controversial taxonomic assignment (e.g., Hess, 2008).
Thus, qualitatively (see Fig. 2) isocrinids and cyclocrinids occur in both carbonate and siliciclastic rocks. The cyrtocrinids and roveacrinids occur within carbonate rocks, whereas the comatulids are exclusive to siliciclastics. In terms of palaeodepth, most crinoid groups dominate in shallow environments with the sole exception of cyrtocrinids, that are ubiquitous and occur in both shallow (near shore and shallow marine) and slightly deeper (deeper sublittoral to open shelf) settings. The occurrences of the cosmopolitan taxa, Chariocrinus andreae and Balanocrinus subteres (isocrinids), is independent of both substrate type and palaeodepth. This qualitative inference (Fig. 2) is largely mirrored quantitatively, also (Table 5).
The quantitative analysis ( Fig. 2 and Table 5) suggests that the distribution of crinoids show a strong substrate-dependency (Hypothesis 1) (see Fig. 2 and Table 5) corroborating results by Simms (1989) and Hunter & Underwood (2009). The palaeodepth-dependency (Hypothesis 2) did not yield significant results, although very weak (low p value) significance is noted for near shore and shallow marine settings (Table 5), as also reflected in qualitative analysis (see Fig. 2). Although it must be accepted that the quantitative analysis (Table 5) is genus-based. Nevertheless, and until more species-based datasets are available, present quantitative analysis (Table 5) should prove useful in better understanding crinoid diversity trends. These results have important bearing on crinoid diversity studies and underline the importance of incorporating substrate details (facies analysis) while inferring distribution patterns of crinoid species diversity and palaeobiogeography. A more rigorous and quantitative species-based analysis may enable to better understand their varied distribution patterns.

CONCLUSION
Almost all Bathonian and Callovian crinoids recorded from Poland dominate in relatively shallow environments. The only exception are cyrtocrinids, that are equally common in both shallow and deeper settings. The most commonly recorded Isocrinids, occur frequently in both carbonate and siliciclastic rocks. Cyrtocrinids also occur likewise, but dominate in carbonates. Roveacrinids are exclusive to carbonates, whereas cyclocrinids occur in both carbonates and siliciclastics. The comatulids occur exclusively in siliciclastics. The most cosmopolitan crinoids are the isocrinids, Chariocrinus andreae and Balanocrinus subtrees, that occur both in carbonate and siliciclastic rocks, and in shallow and open-shelf environments. Statistical analyses (ANOVA) corroborate these qualitative inferences, and suggests that the distribution of crinoids show strong substrate-dependency (Hypothesis 1) but not palaeodepth-dependency (Hypothesis 2), although statistical significance (albeit very weak; low p value) is noted for near shore and shallow marine settings. A more exhaustive species-based analysis may enable to better understand their varied distribution patterns.

ACKNOWLEDGEMENTS
Many friends, colleagues and students from the University of Silesia in Katowice (Tomasz Brachaniec, Przemysław Gorzelak, Jakub Jakubski, Rafał Lach, Mateusz Syncerz, Michał Zatoń, and some more whose names we do not remember) had been contributing through help during fieldworks and donating some crinoid specimens in years 2006-2020. Our reviewers (Aaron Hunter, Pedro Monarrez and an anonymous one) are acknowledged for critically reviewing the manuscript and providing constructive comments that considerably improved it. We also would like to thank Tomasz Wrzołek (University of Silesia in Katowice) for preparing polished slabs involved in the current study. We are also grateful to Alfred Uchman (Jagiellonian University) and Michał Zatoń for sharing references and providing subsequent literature. SJ thanks Ahmed Awad Abdelhady (Egypt) for confirming statistical results. Last but not least thanks go to Marcin Krajewski (AGH University of Science and Technology) for useful comments on the early draft of this MS.