Faunal Composition

The linguloid brachiopod Orbiculoidea (79 occurrences) occurs most frequently in these collections (Fig. 3). It is present in every lithologic category and is the most frequent taxon in both siliciclastic and phosphorite lithologies (Fig. 4). Although frequent, Orbiculoidea is typically rare. Other frequent brachiopod taxa include the spiriferinid “Spiriferina” (26) and orthotetid Derbyia (20). Derbyia is typically more abundant than “Spiriferina.” Frequent bivalves include the nuculid Nuculopsis (38) and trigoniid Schizodus (20). Nuculopsis is typically abundant, similar to the productid brachiopod Echinauris (8). Ramose (53) and fenestrate (19) bryozoans are frequent. Gastropod taxa occur less frequently than taxa of other major groups. Bellerophontaceans (11) and the indeterminate low-spired gastropods (11) are the most frequent gastropod taxa in the dataset.

Figure 3. Taxa ranked by frequency of fossil collections in which they occur, with the semiquantitative categories (abundant, common, rare, and present) denoted.

Figure 4. Taxa ranked by frequency of fossil collections in which they occur, filtered to include only taxa with a minimum of three occurrences, and subdivided into a collections-associated lithology. The semiquantitative categories (abundant, common, rare, and present) are noted.

To observe changes in faunal occurrences and abundances between different lithologies, we subdivided occupancy by lithology (Fig. 4). Siliciclastic fossil collections frequently contain bivalve and gastropod taxa, with the exception of Orbiculoidea (36). Nuculopsis (23) is frequent and typically abundant. Occurrences sharply decline after Orbiculoidea and Nuculopsis. Other frequent bivalve taxa include Streblochondria (11) and Polidevcia (6); both are of common abundance. Low-spired gastropods (7) are the most frequent gastropod taxon and are typically abundant. Fenestrate bryozoans (4) are infrequent.

In contrast to siliciclastics, carbonates are more evenly distributed. Carbonate fossil collections frequently contain bryozoan and brachiopod taxa. Ramose bryozoans (28) are the most frequent taxon. Frequent brachiopod taxa include Orbiculoidea (15), “Spiriferina” (13), and Composita (11). These three taxa are typically rare. Less frequent brachiopod taxa include Bathymyonia (9), Derbyia (9), and Echinauris (6). Both Bathymyonia and Echinauris are typically abundant. Schizodus (7) and bellerophontaceans (7) are the most frequent bivalve and gastropod taxa, respectively.

Phosphorites are similar to siliciclastics in the dominance of Orbiculoidea (15) and Nuculopsis (9). However, phosphorites have fewer taxa overall than other lithologies. The majority of phosphorites occur in condensed, outer-ramp mudstones and shales, which rapidly weather in rock outcrop (Yochelson and Van Sickle Reference Yochelson and Van Sickle1968). For this reason, fossil collections recovered from phosphorites are typically sampled from mining trenches made available to the USGS during the 1950s. Our preparation of the data also excludes fossil collections from south-central Idaho, the depocenter of the Meade Peak Phosphatic Shale. Fossil collections from this area did not have sufficient lithologic descriptions for use in this study. The restricted diversity is in agreement with observations of phosphorites across the Phosphoria Rock Complex (Yochelson and Van Sickle Reference Yochelson and Van Sickle1968).

Cherts are similar to carbonates, as brachiopod and bryozoan taxa are frequent in both lithologies. Ramose bryozoans (21), Orbiculoidea (13), and “Spiriferina” (11) are the three most frequent taxa. These two lithologies differ in the occurrence and abundance of accessory brachiopod taxa. In contrast to carbonates, Derbyia (8) and Hustedia (4) are both more abundant and Composita (5) and Bathymyonia (3) are rarer in cherts. Schizodus (7) and low-spired gastropods (3) are the most frequent bivalve and gastropod taxa, respectively; both are typically rare.

Lithology and Facies Categorization

We sorted fossil collections into three facies based upon their lithologic descriptions (Sheldon Reference Sheldon1963; Cressman and Swanson Reference Cressman and Swanson1964) (Table 2). Facies categories are: inner ramp, mid-ramp, and outer ramp. Mid-ramp is the most frequent facies in these data. Of the three depositional cycles, the Grandeur and Franson cycles contrast with the Ervay cycle. In the Grandeur and Franson cycles, the majority of fossil collections are sampled from the mid-ramp facies. In the Ervay cycle, the majority of fossil collections are sampled from the outer-ramp facies. The most sampled formations are the Franson Member and the Retort Phosphatic Shale Member. The Franson is primarily composed of inner- to mid-ramp carbonates, while the Retort is primarily composed of outer-ramp siliciclastics and phosphorites.

Two styles of chert preservation are present in rock outcrops of the Phosphoria Rock Complex: bedded cherts and nodular cherts. Bedded cherts are rocks in which the entirety of the stratum is composed of chert, whereas nodular cherts are those in which chert nodules are frequent constituents within a host stratum of a different lithology. In total, 50 fossil collections are classified as chert. Of those, 33 collections are of the nodular type, and 15 collections are of the bedded type; 2 collections have lithologic descriptions in which both types are present. Nodular chert is more frequent in both the Franson (16 occurrences) and Ervay (10), and bedded cherts are more frequent in the Tosi Chert (3), Rex Chert (4), Grandeur (3), and Lower Chert (3). With regard to nodular cherts, the host rock is most frequently carbonate (23) (e.g., carbonate rock, dolomite, or limestone). Other host rocks are sandstone (5) and mudstone (5). Nodular cherts have 25 collections that are not singletons, and bedded cherts have only 2 collections. This indicates the chert signal in these data favors the nodular chert type.

Faunal composition varies between the two chert types. In bedded cherts, Orbiculoidea (6 occurrence), Schizodus (3), and nuculids (1) are most frequent, and both Orbiculoidea and Schizodus are abundant. In nodular cherts, Ramose bryozoans (20), “Spiriferina” (10), Derbyia (8), and fenestrate bryozoans (8) are most frequent, and Derbyia, Hustedia, and Echinauris are abundant.

It is important to further elaborate the connection made between chert and substrate. Chert is diagenetic, but it originates primarily from the burial of biosiliceous skeletal elements (e.g., tests, frustules, and spicules) (Knauth Reference Knauth1979, Reference Knauth, Heaney, Prewitt and Gibbs1994; Schubert et al. Reference Schubert, Kidder and Erwin1997; Kidder and Erwin Reference Kidder and Erwin2001; Kidder and Tomescu Reference Kidder and Tomescu2016; Maldonado et al. Reference Maldonado, López-Acosta, Sitjà, García-Puig, Galobart, Ercilla and Leynaert2019). In these analyses, we make the assumption that more biogenic silica in the substrate, predominantly sponge spicules in this region of the Phosphoria Rock Complex (Murchey Reference Murchey, Hein and Hale2004), is equivalent to chert, whether it is nodular, bedded, or both. Chert is thought to alter and precipitate early on in diagenesis, exemplified by its association with both body and trace fossils (Bromley and Ekdale Reference Bromley and Ekdale1984; Maliva and Siever Reference Maliva and Siever1989; Wistort et al. Reference Wistort, Ritterbush and Ekdale2019). We draw the connection that biosiliceous deposition likely has an impact on substrate due to the abundance of spicular cherts in the Phosphoria Rock Complex, and the presence of chert is the best available data to track this.

Ordination Results

Multivariate ordinations were utilized to explore the relationships among formation, facies, and lithology with respect to faunal composition of fossil collections. Previous paleoecologic studies have demonstrated that biotic gradients in community data can elucidate the environmental parameters that influence them. (Holland et al. Reference S. M., Miller, Meyer and Dattilo2001; Holland Reference Holland2005; Holland and Patzkowsky Reference Holland and Patzkowsky2007; Clapham and James Reference Clapham and James2008; Patzkowsky and Holland Reference Patzkowsky and Holland2012). The environment influences both recruitment and survivorship of marine taxa and can control community structure via a variety of parameters (hydrodynamic energy, oxygenation, photic energy, and substrate consistency, etc.) (Holland and Patzkowsky Reference Holland and Patzkowsky2007; Patzkowsky and Holland Reference Patzkowsky and Holland2012). Previous studies have suggested water depth as a primary control on community composition; however, water depth covaries with environmental parameters and thus is not necessarily the cause of biotic gradients. For a thorough review of biotic and environmental gradients in marine paleoecology, please see Patzkowsky and Holland (Reference Patzkowsky and Holland2012).

DCA and NMDS analyses were performed and fossil collections were coded for their membership in three categories: formation, facies, and lithology (Fig. 5). Formation centroids do not shift considerably between plots of the two ordination methods, with the exception of the Meade Peak Phosphatic Shale Member and Tosi Chert Member. In both plots, the Franson Member, Ervay Member, and Lower Shedhorn plot closely in ordination space at low scores of axis 1. These three formations are composed of shallow-water carbonates and siliciclastics. The Retort Phosphatic Shale Member and the Upper Shedhorn member plot closely in ordination space at high scores of axis 1 in both plots. The Retort is commonly composed of offshore siliciclastics and phosphorites, but the Upper Shedhorn is composed of nearshore siliciclastics. In the DCA ordination space, the Upper Shedhorn plots at more intermediate values proximal to the Grandeur Member. In the NMDS ordination space, the Upper Shedhorn plots at high scores of axis 1 proximal to the Retort. This indicates fauna of the Upper Shedhorn must be transitional between the mid-ramp, carbonate-dominant Grandeur fauna and the outer-ramp, siliciclastic- dominant Retort fauna. The Tosi Chert Member centroid varies in relative position in ordination space between these two plots. In the DCA ordination space, the Tosi Chert plots well within the carbonate-dominant shallow-water fossil collections at low axis 1 scores. In the NMDS ordination space, the Tosi Chert plots at intermediate axis 1 scores, the transitional zone between the inner/mid- and outer-ramp collections, similar to the Upper Shedhorn.

Figure 5. Detrended correspondence analysis (DCA) and nonmetric multidimensional scaling (NMDS) (k = 3) of all fossil collections in the filtered dataset (“all”; 105 collections). Collections are coded to the categories of facies, lithology, and taxa. The letter corresponds to the formation code (see Table 2). The shape of collections corresponds to facies. The color of collections corresponds to lithology. The black bold text in each plot corresponds to that plot's category centroids. Formations are labeled by formation code; facies by facies name: inner ramp, mid-ramp, or outer ramp; and lithology by lithologic code: Ch, chert; Ca, carbonate; Si, siliciclastic; and Ph, phosphorite. Taxon scores are plotted as points with their associated taxon code (see Table 1).

Facies centroids separate strongly into two end-members along axis 1 of both plots. Fossil collections from outer-ramp facies have high scores along axis 1 in contrast with collections from inner- and mid-ramp facies that have low scores along axis 1. Centroids of the inner-ramp and mid-ramp facies plot close together in ordination space, indicating very similar faunal compositions. This is unexpected, as commonly in paleoecologic studies there is more significant separation along a nearshore to offshore gradient corresponding with the primary axis of variance. Little segregation between the taxa of the inner-ramp and mid-ramp facies may reflect homogeneous environmental conditions. Also, the position of the Upper Shedhorn Member is suspect. This unit is entirely composed of inner-ramp facies, but it plots proximal to the outer-ramp centroid. Axis 2 appears to have little impact on the segregation of facies groupings in ordination space.

Similar to facies, lithology centroids separate strongly into two end-members along axis 1 of both plots. Siliciclastic and phosphorite lithologies are more frequent at high axis 1 values, while carbonate and chert lithologies are more frequent at low axis 1 values. This separation between lithologies is weaker than the separation between outer-ramp and inner/mid-ramp facies. Although more densely concentrated at these end-members, all lithologies are present across axis 1 in both plots. Along axis 2, phosphorites tend to occur at higher scores, with the exception of a lone outlier at low values of axis 2. This outlier skews the phosphorite centroid toward intermediate axis 2 values, particularly in the NMDS plot.

To interpret the distribution of fossil collections in ordination space, we plotted taxon scores. Taxa that are farther from the center are most abundant in the collections in that area of the ordination space. In both ordination spaces, the bivalves Nuculopsis and Streblochondria, as well as the gastropod Babylonites, plot at high scores of axis 1, corresponding with the outer-ramp facies and the phosphorite and siliciclastic lithologies. A mix of gastropod and bivalve taxa plot at intermediate values of axis 1 in both plots. Brachiopods, bryozoans, and crinoids plot at low to intermediate values of axis 1 corresponding with the inner/mid-ramp facies and the carbonate and chert lithologies. Important taxa in this part of the ordination space include the brachiopods Sphenosteges, Dielasma, Liosotella, and Echinauris; the bivalve Myalina; and ramose bryozoans. Interestingly, in the DCA plot, ramose bryozoans and the brachiopod “Spiriferina” plot at more intermediate axis 1 scores in contrast with the NMDS plot, in which these taxa plot at negative axis 1 scores. These two taxa more strongly contribute to the arrangement of fossil collections in the NMDS analysis.

Along axis 2 in both ordination spaces, gastropods Knightites and Euphemites and large, shallow-water bivalves Aviculopecten and Schizodus plot at low scores proximal to siliciclastic, carbonate, and chert lithologies. Plotting at high axis 2 scores, particularly in the NMDS space, proximal to phosphorite and outer-ramp centroids are the brachiopods Orbiculoidea and Lingula, the bivalves Nuculopsis and Streblochondria, and the gastropod Babylonites and low-spired gastropods.

A PERMOVA test was performed to assess the influence of coded categories on patterns in the arrangement of fossil collections across ordination space (Table 3). The coded categories of formation, facies, and lithology were compared using the “all” community matrix (this corresponds with the data presented in Fig. 5). The formation category contributed to 22% of variance of the community matrix, while facies and lithology contributed to 3% each. This statistical test is supported by the proximity of formation centroids from different depositional cycles, such as Franson, Ervay, and Lower Shedhorn or Retort and Upper Shedhorn. A single category contributing to more than 20% of variance is quite large when dealing with ecologic data. This result indicates that membership in a particular formation, and probably in a particular systems tract, significantly affects the arrangement of fossil collections in ordination space to the point that formations of similar character plot more closely together, overcoming any signal of lithology or facies. Because of this, it is difficult to disentangle the potential effects of these latter categories on community data.

To mitigate the effect of formation membership on the ordination space, we separated the community matrix into the Franson cycle and Ervay cycle and performed NMDS analyses on both subsets (Fig. 6). Ordination analysis was not performed on the Grandeur cycle (e.g., Grandeur Member and Lower Chert), as there are too few fossil collections to draw any meaningful conclusions. Fossil collections were again coded for their membership in three categories: formation, facies, and lithology.

Figure 6. Nonmetric multidimensional scaling (NMDS) ordination of Franson cycle (49 collections) and Ervay cycle (56 collections) data. Collections are coded to the categories of facies, lithology, and taxa. The letter corresponds to the formation code (see Table 2). The shape of collections corresponds to facies. The color of collections corresponds to lithology. The black bold text in each plot corresponds to that plot's category centroids. Formations are labeled by formation code; facies by facies name: inner ramp, mid-ramp, or outer ramp; and lithology by lithologic code: Ch, chert; Ca, carbonate; Si, siliciclastic; and Ph, phosphorite. Taxon scores are plotted as points with their associated taxon code (see Table 1).

Formation centroids in the Franson cycle ordination space separate along MDS1. The Franson Member centroid plots at low MDS1 scores, the Lower Shedhorn plots at intermediate scores, and the Meade Peak plots at high scores. The Meade Peak is represented by only two fossil collections in the ordination space, one of which occurs at intermediate MDS1 scores. Fossil collections from the Franson Member plot most frequently in low and intermediate scores of MDS1; however, some samples do plot in high scores. Fossil collections from the Lower Shedhorn plot most frequently in intermediate scores of MDS1, with the exception of a lone outlier that plots in negative values of MDS1.

Formation centroids in the Ervay cycle ordination space also separate along MDS1. The Ervay Member and Tosi centroids plot at high scores on MDS1, and the Retort and Upper Shedhorn centroids plot at low to intermediate scores. The Retort centroid would likely plot at lower MDS1 scores if not for an outlier collection at very high scores. Removal of this collection would probably increase the distance between the Retort and Upper Shedhorn centroids. The Ervay Member, Tosi, and Upper Shedhorn all loosely cluster around their respective centroids. In contrast, fossil collections of the Retort are spread across ordination space with a strong clustering at low MDS1 scores. Along the MDS2 axis, almost all collections from the Tosi and Ervay Member plot at low scores. The Retort and Upper Shedhorn plot at intermediate to high MDS2 scores.

Similar to facies centroids of the “all” dataset (Fig. 5), facies centroids in the Franson cycle ordination space separate into two end-members along MDS1. Fossil collections from the outer-ramp facies have high MDS1 scores in contrast with collections from the inner- and mid-ramp facies, which have low MDS1 scores. Centroids of the inner-ramp and mid-ramp facies plot close together in ordination space. The lack of fossil collections from the Meade Peak in this analysis makes it difficult to say with confidence that the outer-ramp facies are really distinct form the inner/mid-ramp. Similar to the “all” data, there is little segregation between the taxa of the inner-ramp and mid-ramp facies, which may reflect homogeneous environmental conditions. The MDS2 axis seems to have little impact on the segregation of facies groupings in ordination space.

Facies centroids of the Ervay cycle show a different arrangement, contrasting with those of the “all” data and the Franson cycle. The outer-ramp centroid plots at low MDS1 scores, the inner-ramp centroid plots at intermediate scores, and the mid-ramp centroid plots at intermediate to high scores. This is the only ordination space in which the inner- and mid-ramp fossil collections have distinct centroids, and of particular note is the inner-ramp centroid plotting between the mid-ramp and outer ramp along MDS1. Inner-ramp fossil collections are evenly split between the Ervay Member (6) and the Upper Shedhorn (6). The Upper Shedhorn inner-ramp collections plot at intermediate MDS1 scores proximal to the outer-ramp centroid, whereas the Ervay Member inner-ramp collections plot at high MDS1 scores. This is odd and further illustrates the similarity in faunal composition of the Upper Shedhorn and the Retort. Without this strong similarity in faunas, the inner-ramp centroid would most likely plot closer to the mid-ramp centroid.

Lithology centroids in the Franson cycle separate similarly to those of the “all” data ordinations, in which the carbonate centroid plots proximal to the chert and the siliciclastic centroid plots proximal to the phosphorite. A single sample is categorized as a phosphorite lithology and is a member of the Meade Peak. Lithology centroids in the Ervay cycle separate similarly to those of the “all” data and the Franson cycle ordinations; however, the separation of fossil collections belonging to the carbonate/chert and siliciclastic/phosphorite end-members is much stronger. There is very little overlap in intermediate MDS1 scores of these lithologies compared with the other ordinations. Chert lithologies seem to form a tighter grouping in the Ervay cycle than in any of the other ordination spaces. Similar to the pattern observed in the “all” data, phosphorite lithologies plot at intermediate to high MDS2 scores, while chert and carbonate collections plot at intermediate to low MDS2 scores.

Taxon scores were plotted for both the Franson cycle and Ervay cycle ordination spaces. In the Franson cycle, the brachiopods Bathymyonia, Liosotella, “Spiriferina,” Sphenosteges, and Echinauris, along with ramose bryozoans, plot at low MDS1 and high MDS2 scores. The bivalves Nuculopsis, Schizodus, and Polidevcia, along with the bellerophontacean gastropods, plot at high MDS1 and high MDS2 scores. The brachiopods Orbiculoidea, Lingula, Neospirifer, and Squamaria, the bivalves Permophorus, parallelodontid, and Pseudomonotis, and the gastropod Knightites form a loose cluster at intermediate MDS1 scores and intermediate to low MDS2 scores. In the Ervay cycle, the relative position of taxon scores is similar to the Franson cycle. The brachiopods “Spiriferina,” Sphenosteges, Dielasma, and Derbyia, along with ramose bryozoans and fenestrate bryozoans, plot at high MDS1 scores. Similar to the “all” data ordinations, bivalves Nuculopsis, Streblochondria, and Babylonites plot near one another at low MDS1. Orbiculoidea, productids, and spiriferids plot close together at high MDS2 scores.

Hierarchical Clustering Analysis

To further explore the relationships of taxa and fossil collections in the Franson cycle and Ervay cycle, we generated heat maps to observe how fossil collections clustered with respect to taxon abundance. In both cycles, distinct clusters of fossil collections (biofacies) were formed. Biofacies are defined by significant changes in both taxa and/or abundance.

The Franson cycle clustered into three biofacies that are labeled Franson 1, Franson 2, and Franson 3 (Fig. 7). The Franson 1 biofacies is defined by the brachiopods Derbyia and “Spiriferina” and ramose bryozoans. These taxa are frequent in samples of Franson 1 but are typically rare. The brachiopods Bathymyonia and Echinauris are less frequent than the defining taxa but, when present, are typically common to abundant. These taxa both have a weakly concave dorsal valve and a strongly convex ventral valve with dense spines. This biofacies occurs primarily in inner-ramp to mid-ramp carbonates with minor cherts and siliciclastics. This convergence of functional characters may indicate a desirable body plan for living in the inner- to mid-ramp depositional environment, characterized by ample carbonate and biosiliceous sedimentation (frequent carbonate and chert lithologies).

Figure 7. Heat map of Q-mode and R-mode hierarchical clustering of taxa vs. fossil collections in the Franson cycle. Shapes of points correspond to depositional environment and pattern fill of points corresponds to lithology. A, C, R, and X correspond to the semiquantitative categories abundant, common, rare, and present, respectively. Results are separated into three biofacies based upon changes in faunal composition and abundance. These biofacies are separated by the three colored outlines and are labeled Franson 1, Franson 2, and Franson 3. The letter corresponds to the formation code (see Table 2).

The Franson 2 biofacies is defined by the brachiopods Orbiculoidea, Lingula, and Neospirifer. Orbiculoidea and Lingula are typically common to abundant and occur in only carbonate and siliciclastic lithologies. Neospirifer is less frequent and typically common. It occurs in chert and carbonate lithologies. The brachiopods Liosotella and Composita are infrequent but, when present, are abundant. Ramose bryozoans occur in Franson 2 less frequently than in Franson 1. Fossil collections in this biofacies are a mix of both inner- and mid-ramp facies and carbonate, chert, and siliciclastic lithologies. Franson 2 shares some similar taxa with Franson 3. These include the bivalves Permophorus, Polidevcia, and parallelodontids and the gastropod Bellerophon. With no distinct facies or lithology and sharing similarities with both the Franson 1 and Franson 3 biofacies, Franson 2 is transitional between these two end-member biofacies of the Franson cycle.

The Franson 3 biofacies is defined by the bivalve Schizodus and the gastropod bellerophontaceans. Both of these taxa are frequent and typically common to abundant. Fossil collections in this biofacies are from every facies and every lithology, with the exception of chert. Carbonate is the dominant lithology. This biofacies also contains both samples of the Meade Peak. No brachiopod or bryozoan taxa occur in this biofacies. These two major groups are frequent and abundant elsewhere in the Franson cycle. This indicates that fossil collections that contain brachiopods and bryozoans are significant to the underlying structure of the Phosphoria Rock Complex community.

The Ervay cycle clustered into three biofacies that are labeled Ervay 1, Ervay 2, and Ervay 3 (Fig. 8). The Ervay 1 biofacies is defined by bivalve Nuculopsis, which is both frequent and abundant. The bivalve Streblochondria is frequent and common. The bivalves Polidevcia, Aviculopecten, Schizodus, and Astartella, the gastropod Babylonites, and low-spired gastropods are infrequent and rare to abundant. Similar to the Franson 1 biofacies, brachiopod taxa are very limited. In Ervay 1, Bathymyonia and “Chonetes” are present in a single fossil collection each, but they are abundant in each of those collections. This biofacies corresponds strongly with the Retort Phosphatic Shale unit, the outer-ramp facies, and the siliciclastic lithology. It illuminates a potential cause of the proximity of the Upper Shedhorn and Retort centroids in Figure 6. Five Upper Shedhorn fossil collections cluster tightly together in the Ervay 1 biofacies. These collections contain the shallow-water bivalves Aviculopecten and Schizodus, which are also present in the Ervay 3 biofacies. However, these collections also contain abundant Nuculopsis. The strength of the abundant Nuculopsis when calculating the distance matrix for ordination is what likely pulls the Upper Shedhorn collections toward the Retort centroid in ordination space. If not for the abundant category of Nuculopsis, these collections would probably plot in Ervay biofacies 3. This exemplifies one of the limitations of using semiquantitative abundance data.

Figure 8. Heat map of Q-mode and R-mode hierarchical clustering of taxa vs. fossil collections in the Ervay cycle. Shapes of points correspond to depositional environment and fill of points corresponds to lithology. A, C, R, and X correspond to the semiquantitative categories abundant, common, rare, and present, respectively. Results are separated into three biofacies based upon changes in faunal composition and abundance. These biofacies are separated by the three colored outlines and are labeled Ervay 1, Ervay 2, and Ervay 3. The letter corresponds to the formation code (see Table 2).

The Ervay 2 biofacies is defined by the brachiopod Orbiculoidea. This brachiopod occurs in all fossil collections of the biofacies and is rare to abundant. This biofacies corresponds strongly with the Retort Phosphatic Shale unit, the outer-ramp facies, and the siliciclastic and phosphorite lithologies. The Ervay 2 biofacies is similar to the Franson 2 biofacies in that Orbiculoidea is a defining taxon. However, the Ervay 2 biofacies has a much lower diversity of taxa. It also has a strong affiliation with the phosphorite lithology in contrast to the Franson 2 biofacies, which has more carbonate and chert collections.

The Ervay 3 biofacies is defined by the brachiopods Composita, Hustedia, and Echinauris and ramose bryozoans. These taxa are frequent and abundant. Derbyia and Bathymyonia are less frequent and rare to abundant. Fenestrate bryozoans and crinoid columnals are also present. This biofacies corresponds with the Ervay Member and Tosi units, the inner- and mid-ramp facies, and the chert and carbonate lithologies. The Ervay 3 biofacies is similar to the Franson 1 biofacies in that it is defined by a diverse fauna of brachiopods and frequent ramose bryozoans. These two biofacies contrast in specific brachiopod taxa and abundances. The most frequent brachiopods in the Franson 1 biofacies are “Spiriferina” and Derbyia. These two taxa are frequent only in Ervay 3, but typically less abundant than the brachiopod taxa that define the biofacies. Echinauris and Bathymyonia are abundant in both biofacies. The defining fauna of Ervay 3, Composita and Hustedia, are infrequent and rare or not present in Franson 1. These biofacies also contrast in dominant fossil collection lithology. In Franson 1, carbonates are the dominant lithology with minor cherts, whereas in Ervay 3, cherts are the dominant lithology with minor carbonates. The differences between the constituent taxa of these two biofacies exemplify the differences between carbonate and biosiliceous depositional regimes.

To understand how these biofacies fit along a gradient, we coded fossil collections in the ordination spaces of the Franson cycle and Ervay cycle with their respective biofacies memberships and plotted their taxon scores (Fig. 9). Taxon scores are slightly jittered for the sake of readability; this movement is small and does not affect the interpretation of the relative position of scores. The Franson cycle ordination space separates into two end-members that correspond with facies in Figure 6. Plotting at low MDS1 scores are the Franson 1 and Franson 2 biofacies, and at high MDS1 scores is the Franson 3 biofacies. These correspond with the facies centroids in which inner and mid-ramp plot more closely together at low scores and the outer ramp plots at higher scores. These two end-members segregate remarkably cleanly in the ordination space, with the lone exception of a single Franson 1 collection that plots within Franson 3. The Franson 1 and Franson 2 biofacies share many similar taxa; however, Franson 1 forms a tighter cluster within the broader Franson 2. This result indicates a compositional difference between the two communities that may correspond with inner-ramp and mid-ramp communities. We interpret that the Franson 1 biofacies likely correlates with the mid-ramp and Franson 2 with the inner ramp. The presence of a diverse brachiopod fauna along with ramose bryozoans and a generally carbonate lithology fit the expectation of a mid-ramp depositional setting (Blomeier et al. Reference Blomeier, Dustira, Forke and Scheibner2013). The brachiopods Lingula and Orbiculoidea occur along with a more diverse gastropod fauna and the only horn corals in the dataset, indicating a likely nearshore depositional setting.

Figure 9. Nonmetric multidimensional scaling (NMDS) ordination of the Franson cycle and Ervay cycle (from Fig. 6) coded to biofacies (from Figs. 7, 8). Taxon scores plotted and slightly jittered for readability; this does not significantly change their relative positioning. Convex hulls colored by their associated biofacies.

The Ervay cycle's three biofacies separate cleanly in ordination space with no overlap of fossil collections. Along MDS1, Ervay 1 plots at low to intermediate scores, Ervay 2 plots at intermediate scores, and Ervay 3 plots at intermediate to high scores. Along MDS2, Ervay 3 plots at low to intermediate scores, Ervay 1 plots at intermediate scores, and Ervay 3 plots at high scores. The Ervay cycle ordination space is unique, in that the inner-ramp centroid plots at intermediate MDS1 and MDS2 scores, between the outer-ramp and mid-ramp centroids. Taxa plotting at intermediate scores along MDS1 include the brachiopods Orbiculoidea, Lingula, and productids, the bivalves Cyrtorostra, Aviculopecten, and nuculids, and the gastropods bellerophontaceans and pleurotomariaceans. These taxa all separate along the MDS2 axis in which Orbiculoidea plots at high scores, Lingula plots at intermediate scores, and Aviculopecten plots at low scores. This separation along the MDS2 axis corresponds to broad changes in lithology in the Ervay cycle ordination space. Plotting at high MDS2 scores are a mix of siliciclastic and phosphorite collections from the Retort, and at low MDS2 scores are a mix of chert and carbonate collections from the Ervay Member and Tosi. Similar to the Franson 2 biofacies, the presence of Lingula and Aviculopecten could indicate a more nearshore depositional environment; however, in the case of Ervay 3, Orbiculoidea plots more closely with Retort outer-ramp than inner-ramp collections. This confounding result may be related to the stark contrast in lithologies. Orbiculoidea plots closely to phosphorite collections. Lithology also seems to influence Ervay 3 biofacies along MDS2. The Ervay 3 cluster is closely associated with chert collections with minimal contributions from other lithologies.

A PERMOVA test was performed to evaluate the influence of the coded categories of formation, facies, lithology, and biofacies on patterns in the arrangement of fossil collections in both the Franson cycle and the Ervay cycle ordinations (Table 3). In the Franson cycle, the biofacies category contributed to 24% of the variance in the community matrix. Lithology and formation contributed 4%, and facies 3%. In the Ervay cycle, the biofacies category contributed 23% of the variance in the community matrix, facies 18%, lithology 12%, and formation 8%. In both cycles, the percent of variance attributable to biofacies is high, as biofacies is defined by faunal composition. In the Franson cycle, little variance is explained by lithology, formation, or facies, in contrast with the Ervay cycle, in which each of these factors contributes significantly. This may be due to the overwhelming sampling intensity of the Franson compared with the Lower Shedhorn and the Meade Peak. More fossil collections from these two formations would likely contribute facies in the inner ramp and the outer ramp, which would produce a more well-defined gradient between facies along MDS1. The Ervay cycle is more balanced with regard to sampling intensity. In the Ervay cycle, a significant amount of variance is explained by each of the coded categories. Facies contributes the most to the organization of collections along MDS1. Lithology contributes to the separation of collections along MDS2.