Our dataset includes 193 fossil samples from the Lower Ordovician of the Cordillera Oriental, Argentina (Fig. 1). Part of the data set has been already analyzed by Balseiro and Waisfeld (2014), Balseiro et al., 2011a and Balseiro et al., 2011b, Meroi Arcerito et al. (2015), Waisfeld and Balseiro (2016), while the rest are unpublished data (Meroi Arcerito, 2016). All samples are housed at the “Centro de Investigaciones en Ciencias de la Tierra” (CICTERRA, CONICET, UNC). All trilobite sclerites were identified in each sample at least to genus, according to the fossils’ preservation, summing a total of 798 occurrences. It is important to remark that the taxonomy of asaphids has been recently revised (Meroi Arcerito, 2016 and Meroi Arcerito et al., 2015), while the taxonomic resolution of other families is variable (e.g., Harrington and Leanza, 1957 and Waisfeld and Vaccari, 2003), which would avoid a taxonomic bias in diversity estimation.

We considered six successive temporal bins for the Early Ordovician (Bergström et al., 2009): Tremadocian 1 (Tr1), early and late Tremadocian 2 (early Tr2 and late Tr2), Tremadocian 3 (Tr3), Floian 1 (Fl1) and Floian 2–3 (Fl2–Fl3) regionally recognized on the basis of graptolite, trilobite, and conodont biostratigraphy (Fig. 2). Time bins are here constrained upon the schemes proposed by Vaccari et al. (2010 and references therein) for the early Tremadocian and by Toro et al. (2015 and references therein) for the late Tremadocian–Floian.

The samples of the lower Tremadocian (Tr1) stratigraphic units correspond to the Jujuyaspis keideli, Kainella andina and Kainella meridionalis biozones, and the lowermost part of the Kainella teiichii biozone. We subdivided the Tremadocian 2 (Tr2) in two time intervals: early Tr2 consists of samples coming from the Bienvillia tetragonalis biozone and the Asaphellus–Leptoplastides association, and late Tr2 with samples coming from the Notopeltis orthometopa biozone (see Toro et al., 2015 for graptolite and conodont calibrations). The late Tremadocian (Tr3) is represented by a few sparse samples coming either from the uppermost part of Araneograptus victoriae or the Hunnegraptus copiosus biozones. Unfortunately, Floian 1 was excluded because it has scarce samples to run the analysis. The Floian 2–3 interval corresponds to the uppermost Floian 2 (upper part of the Baltograptus cf. deflexus biozone) and the lowermost Floian 3 (lower part of the Dydimograptus bifidus biozone).

The data set comes from 15 different localities that are representative of the Lower Ordovician outcrops of the Argentine Cordillera Oriental (Fig. 1). However, the geographic coverage of individual bins varies. The samples from the Tr1 interval are the most geographically widespread (see Fig. 1). The samples from the Tr2 largely come from eastern outcrops of the Cordillera Oriental whereas no stratigraphic record of this interval is preserved to the west due to a regional unconformity (Tumbaya break, Astini, 2003 and references therein). As well, the samples from the Tr3 are scarce because the successions comprising this time bin are poorly fossiliferous or partially complete due to local faulting. In contrast, the records of Floian 1 deposits are relatively complete and widespread in the basin; however, they are weakly fossiliferous or even devoid of benthic fauna. Finally, samples from the Fl2–3 are concentrated in two areas, one in the west and the other in the North of the Cordillera Oriental.

The similarity between the raw data and the estimated sample trends is remarkable, reinforcing the fact that our sampling effort was enough to estimate total diversity (S _T) per temporal bin of 25 samples.

In Fig. 4, represented by black circles, we can see the diversity (S _T) pattern for the time sequence considered. The highest diversity recorded within the time bins was obtained for the Tr1; its S _T value is 20.031. A significant decrease in diversity can be detected by early Tr2, where S _T = 15.511. During late Tr2, generic diversity rises slightly and reaches S _T = 15.772. Later, in the Tr3, S _T falls to 13 genera. Finally, in the Fl2–3, diversity exhibits an increase (S _T = 16.015). In sum, we can observe that S _T remains approximately stable, reaching its minimum values during the Tr3.

The temporal sequence for the estimated diversity for olenids (S _O) can be seen in Fig. 4 (white diamonds). In the Tr1, we registered the highest diversity value (S _O = 6.48). The number of genera in this family decreases drastically in the early Tr2 (S _O = 4), and their diversity continues declining by the late Tr2 (S _O = 2.993). By the Tr3, there are only two genera left (S _O = 2). Nevertheless, they exhibit a recovery during the Fl2–3, when their generic diversity doubles the previous record (S _O = 3.959).

Olenid diversity shows a very similar trend to the one observed in S _T described above (Fig. 4). Both S _T and S _O decline from Tr1 to early Tr2. The main difference is that S _O decreases steadily from Tr1 to Tr3 and, finally, an increase is recorded on the Fl2–3.

Estimating the asaphids’ diversity (S _A; see white circles in Fig. 4), we obtained a different pattern with respect to S _T or S _O. The temporal trend shows that S _A increases steadily from Tr1 (S _A = 1.927), passing through early Tr2 (S _A = 1.862) and late Tr2 (S _A = 2.999), reaching the highest diversity peak during Tr3 (S _A = 9) and then decreasing by Fl2–3 (S_A = 5.572).

The maximum difference observed between S _A and S _T is registered during the early Tr2 (Fig. 4). The lowest S _A value was recorded in the early Tr2 and it slightly increases from the early Tr2 to the late Tr2; then, during Tr3, asaphid diversity triples, reaching its highest proportion (69.3%) of S _T. Finally, in the Fl2–3, S _A is severely affected and reduced by almost one third. On the other hand, S _T shows it highest values on Tr1, and then its value remains relatively stable.

The occupancy rank plots show that the distribution of genera and families changed considerably throughout the Early Ordovician (see Fig. 5). In the Tr1, olenids (gray circles) dominate the communities with Parabolinella, Leptoplastides and Angelina occupying the 1^st, 2^nd and 4^th places in the rank, respectively with occupancy values of 0.5, 0.4 and 0.32 during the time bin. Asaphids (white circles) are only represented by two genera, Asaphellus and Golasaphus, located at the 5^th and 15^th places.

In the early Tr2, although olenids remain as major components of the fauna (e.g., Leptoplastides, Parabolinella and Bienvillia, ranked 1^st, 3^rd and 4^th), they register a loss in diversity, and asaphids show an evident increase in occupancy. In this time bin, Asaphellus is the second most recorded genus with an occupancy of 0.76.

During the late Tr2 asaphids show a slight increase in diversity, now recording three genera: Asaphellus, Notopeltis and Mekynophrys. Unlike this, olenids decrease in diversity and also in occupancy: only Parabolinella, Leptoplastides and Bienvillia are left, occupying the 2^nd, 8^th, and 9^th places in the rank, with occupancy values of 0.56, 0.15 and 0.09, respectively.

By the Tr3, asaphids clearly exhibit the highest scores both in diversity and occupancy. The first five positions in the rank are occupied by Thysanopyge, Kayseraspis, Asaphellus, with values above 0.4, and Notopeltis, Megistaspis (Ekeraspis) with scores of 0.29 and 0.24 respectively. During this time bin, olenids record the lowest diversity (2) and occupancy values.

In the Fl2–3, trinucleids (orange circles) make appearance with Famatinolithus in the first place of the rank. Unfortunately, the lack of data from Fl1 does not allow one to identify when they were initially recorded. During this time bin, asaphids show a decrease in diversity and occupancy, but they are still one of the main constituents of the Floian fauna. Moreover, olenids have a small increase in diversity and intermediate occupancy distribution, three genera ranked 9^th, 10^th, and 11^th.

The regional scale analysis of diversity and occupancy of olenids and asaphids show an interesting pattern of turnover. While olenids diminish their generic richness through time, asaphids became the richest family since the Tr3. It is remarkable that both share the same diversity in the late Tr2, however, asaphids occupancy for the first time stepped up over olenids occupancy, rising then persistently for the rest of the studied interval. Hence, the late Tr2 seems to be an inflexion point in diversity and occupancy trajectories of both groups, highlighting with high precision this turning point of the Ibex II Fauna in the Cordillera Oriental basin.

The early history of asaphids in Cordillera Oriental can be traced back in the Furongian by the occurrences of Asaphellus, the first member of the family recorded in the basin. During the Furongian, Balseiro et al. (2011a) found that Asaphellus was a quite rare component of the assemblages, occurring in low abundance and also very low occupancy. Interestingly, it was only present in high diversity and even olenid-dominated assemblages were restricted to intermediate environments of the onshore–offshore gradient. Based on the analysis of abundance distributions of trilobite taxa, Meroi Arcerito et al. (2015, fig. 4) observed that although Asaphellus was the only asaphid present in the early Tr2, it became abundant or even dominant in local communities. Our regional results, at the same interval, show that only one asaphid (Asaphellus) had high levels of occupancy (Fig. 5), whereas the rest of the high occupancies correspond to olenids. During the late Tr2, asaphids occupancy increased significantly, even they became dominant (Fig. 5); however, their diversity remained low (Fig. 4). Olenids are no longer dominant, remaining only one taxon with intermediate occupancy. Remarkably, during the Tr3, both asaphids diversity and occupancy increased drastically. Thus, a temporal decoupling between asaphids diversity and occupancy is recognized. The significant rise in diversity might be a consequence of the increase in occupancy in the Tr3. The diversity of asaphids in the Fl2–3 was markedly lower than in the Tr3 interval, but occupancy remains high. Nonetheless, a trinucleid, representing the Whiterock Fauna, shows the highest value of occupancy.

In sum, local and regional scale analyses portray the successive steps in the paleoecological dynamics of the asaphids and, hence, of the Ibex II Fauna in the basin. In general, a single (or more infrequently two) rarely present asaphid occurred in olenid-dominated communities since the Furongian to the early Tr2 (Fig. 5). However, in the latter interval, it became locally abundant and widespread, attaining high occupancy regionally. It was not until the late Tr2 that asaphids reached the highest occupancy scores, yet diversity remained low. At last, diversity peaked later in the Tr3 accounting for a diachronic diversity-occupancy relationship in the expansion of the Ibex II Fauna.

It is worth noting that previous paleoecological studies analyzed the timing of ecological shifts in the Cordillera Oriental (e.g., Balseiro and Waisfeld, 2013, Mángano and Buatois, 2011 and Waisfeld and Balseiro, 2016) and correlated them either with those derived from global analyses or issued mainly from Laurentian North America, where intensive studies resulted in comparable data (e.g., Adrain et al., 2000, Finnegan and Droser, 2005, Peters, 2004, Peters, 2006 and Westrop and Adrain, 1998, among others). For example, Balseiro and Waisfeld (2013) examined local richness and evenness in trilobite assemblages from the Cordillera Oriental and demonstrated that the rise in these parameters in the early-late Tremadocian (Tr2) lagged behind similar changes in the Late Cambrian of Laurentia (see Peters, 2004 and Peters, 2006). As well, Waisfeld and Balseiro (2016) found that equitable assemblages developed since the Tr2 at the local scale and since the early Floian at the regional scale, whereas highly dominated and regionally equitable trilobite communities can be traced far back to the Middle Cambrian in Laurentia. In a similar vein, Mángano and Buatois (2011) argued that the occupation of the mid-tier infaunal ecospace took place in the Tr2 in the Cordillera Oriental and, hence, was delayed when compared with Laurentia or Baltica where the exploitation of the same tier began in the Early Cambrian. Despite same differences, these examples highlight the existence of diachronic ecological changes in the Cordillera Oriental occurring after those documented in Laurentia or globally.

Interestingly, unlike previous examples, the expansion of the Ibex II Fauna in the Cordillera Oriental predates the global signal. The reasons underlying this pattern are unclear. Paleogeographic constraints and the early phylogenetic history of the asaphids in the basin might have been influential in the observed pattern. The early phylogenetic history of asaphids should be traced far back in the Cambrian and it is far from being resolved. Besides, there are also limited studies in settings with paleolatitudes and sedimentary regimes, similar to those of the Cordillera Oriental, precluding to confirm how widespread is the pattern recognized herein.

Another interesting aspect of the analyzed transition is that the expansion of the Ibex II Fauna (late Tr2–Tr3) takes place after the remarkable change in the ecological structure (early Tr2) recognized by Balseiro and Waisfeld (2013) and Waisfeld and Balseiro (2016). These authors emphasized a shift toward more local and regional equitable communities that took place in the context of olenid-dominated communities. This change is coincident with the initial rise in abundance of Asaphellus in local communities as well as with its high regional dominance, accounting for the onset of the Ibex II Fauna in the basin. Therefore, it seems plausible that both patterns are somehow causally related. Further research will probably shed some light on the actual cause and consequence relationship of these coincident patterns.

Finally, the radiation of the Ibex II Fauna in the Cordillera Oriental does not match the global trend identified by Adrain et al. (2004). These authors indicated that the Ibex II Fauna radiated during their O2 interval (= late Tremadocian s.l.) and peaked in diversity during the O3 (= Floian 2–3). Even though the resolution of Adrain et al. (2004) global data is coarser than that derived from our regional analysis, a distinct diversity peak in the late Tremadocian (Tr3) is recognized, predating that reported by those authors. Thus, a decoupling in the development of the Ibex II Fauna results evident.

On the other hand, Adrian et al. (2004, fig. 24.12B) and Zhou et al. (2007, fig. 3) reported that in South China (Yangtze Block) the Ibex II Fauna (mainly represented by asaphids and nileids) began its radiation in the late Tremadocian (time slice 1d of these authors) reaching a distinct diversity peak during the middle Floian (time slice 2b). So, expansion of the Ibex II Fauna in South China is broadly coincident with the global signal, but took place later than in the Cordillera Oriental. This diachronism may suggest that the radiation of the TEF Ibex II was heterogeneous in time and space, resembling the patterns recognized worldwide for the Whiterock TEF (e.g., Adrain et al., 2004).

Little has yet been said about transitions between TEFs (Adrain et al., 2004 and Zhou et al., 2007), although much has been discussed regarding great evolutionary faunas (Finnegan and Droser, 2005, Finnegan and Droser, 2008, Miller, 1998 and Westrop et al., 1995). Our results could, however, shed some light on some aspects of the dynamics of TEF replacement at the regional scale.

Understanding the dynamics of faunal transitions is a challenge, mainly because underlying processes are very complex and multiple alternatives have been invoked (Alroy, 2004, Benton, 1987, Finnegan and Droser, 2005, Finnegan and Droser, 2008, Miller, 1998, Sepkoski, 1975, Sepkoski, 1996 and Westrop et al., 1995). In a simplified scenario, two models of replacement could be possible for transitions among evolutionary faunas or clades, namely dilution and displacement (Miller, 1998 and Westrop et al., 1995). On the one hand, displacement involves biotic interactions over geologic time, with one clade actually displacing a previous one (Miller, 1998, Sepkoski, 1996 and Westrop et al., 1995). Therefore, the rise in diversity of the diversifying clade goes hand in hand with the decrease in diversity of the displaced clade, resulting in the classic double-wedge pattern (Benton, 1987 and Sepkoski, 1996). However, the double-wedge pattern is not the most probable result in a displacement scenario, where more complicated patterns should be more frequent in a realistically variable world (Sepkoski, 1996). On the other hand, the dilution process occurs when a clade holds a constant diversity, nevertheless its relative contribution diminishes, this happens as a consequence of the diversification of a new clade but not due to interaction between them (Westrop et al., 1995). While displacement can be easily ruled out for clades that occupy contrasting portions of the ecospace (Westrop et al., 1995), it becomes a plausible scenario when clades have similar life habits (Sepkoski, 1996). The latter is the case for the replacement between olenid-dominated (Ibex I) and asaphid-dominated (Ibex II) faunas of the Cordillera Oriental, as both occupied a similar portion of the ecospace, inhabited a similar variety of environments, and probably possessing similar feeding strategies. Olenids and asaphids occur along most of the bathymetric gradient, from lower shoreface to offshore, and shelf settings, and thrived either in well-oxygenated or oxygen-poor environments (Balseiro et al., 2011a, Balseiro et al., 2011b, Waisfeld and Vaccari, 2008 and Waisfeld et al., 2006). Feeding habits are difficult to assess among trilobites. However, based on different lines of evidence detritus feeding is currently considered as a primary strategy for olenids (Fortey and Owens, 1999). On the other hand, a similar feeding habit might be assumed for early asaphids with conservative hypostomes (e.g., early isotelinids, ogygiocaridiines) upon evidence provided by exceptional preservation of soft parts (Gutiérrez-Marco et al., 2017) and trilobite–trace fossil associations (Mángano and Waisfeld, 2004).

In the Cordillera Oriental, while asaphids underwent significant changes in diversity towards the latest Tremadocian (Tr3), regional diversity showed a stable trend through most of the Tremadocian to the Floian. Although this pattern is not a perfect double-wedge (Benton, 1997), it falls within the expected outcomes of competitive displacements (Sepkoski, 1996). These results suggest that asaphids gained diversity at the expense of other trilobites, what is expected in a displacement scenario. Moreover, the constant decline in diversity of olenids through the Tremadocian reaches a minimum during the Tr3, reinforcing the idea of displacement.

In a displacement scenario, occupancies of the displacing clade should increase before its diversity. This would happen because interactions should drive a rise in regional abundance of the displacing clade, until it becomes the dominant taxon, even before its diversity rises. However, if dilution was the main underlying dynamics, occupancies should increase only after, or together with, a rise in diversity. This is because regional abundance should increase hand in hand with the diversification of the clade as a diffusion process (Westrop et al., 1995). In the Cordillera Oriental, asaphids show a sharp increase in occupancy during the early Tr2, definitively reaching regional dominance during the late Tr2, but before their maximum diversity during the Tr3. Mirroring this rise in asaphids occupancy, olenids show a trend of decline in occupancy during the early Tr2–late Tr2. All in all, our results suggest, once again, a replacement driven mainly by displacement.

Moreover, integrating previously published local-scale relative abundance patterns during the Tremadocian of the Cordillera Oriental (Meroi Arcerito et al., 2015) reinforces the idea of the displacement scenario. At local scale, displacement should cause an increase in the relative abundance of the displacing taxon before the rise in regional occupancy. This is because regional occupancy could be delayed due to dispersal limitation or environmental preferences. Meroi Arcerito et al. (2015) showed that in the Cordillera Oriental, asaphids became the dominant taxon of local communities during the early Tr2, that is, before reaching regional dominance during the late Tr2 (Fig. 5).

Taking all three patterns into consideration, displacement stands as the most probable scenario for the regional replacement of the Ibex I Fauna by the Ibex II Fauna during the Tremadocian. Two things, however, need to be mentioned. First, the evidence for displacement blurs towards the Fl2–3 with the sudden appearance of Whiterockian taxa as dominant members coupled with a slight rise in olenid diversity and a clear drop in asaphid diversity (Fig. 4). Unfortunately, the lack of abundant data for the Fl 1 hinders the possibility of understanding the processes involved, but it could be possible that the dynamics driving the replacement between Ibex faunas changed with the rise of the Whiterock Fauna. Second, our analysis is restricted to the Cordillera Oriental, and therefore it is not easily to extrapolate them to other regions, even less globally. Indeed, it could be very possible that, in a fashion similar to the timing of the Ibex II fauna radiation, the processes involved in the replacement of the different faunas differ among regions across the globe, configuring a complex global mosaic of regional patterns similar to that observed during the Ordovician Radiation (Miller, 2012, Miller and Mao, 1995 and Waisfeld et al., 2003).