DISCUSSION

Variations in Geographic Range between Taxon Groups

The patterns of differentiation in the geographic ranges observed in taxonomic groups of Cincinnatian brachiopods can be explained, at least partly, by the relationship between geographic range size and niche breadth. The geographic range inhabited by a species is a manifestation of a species ecological niche at two levels, the fundamental and realized niche (Lomolino et al. 2006). The largest geographic range that a species may potentially occupy is the area of occurrence of a species' fundamental niche, the sum of the ecological parameters under which a species is able to maintain a population (Hutchinson 1957). All species, however, actually occupy a smaller geographic region, which is referred to as the realized niche. The difference between a species' fundamental and realized niches is determined by biotic interactions such as competition and predation (Hutchinson 1957). Species geographic ranges, therefore, are directly tied to the breadth of their ecological niche. A positive correlation between niche breath and geographic range has been recovered in numerous studies (e.g., Jackson 1974; Jackson et al. 1985; Brown 1984; Thompson et al. 1999; Gaston and Spicer 2001; Fernández and Vrba 2005; but see Williams et al. 2006 for a counterexample). Species with broad environmental tolerances, known as generalist or eurytopic species, exhibit larger geographic ranges than specialist or stenotopic species, which are characterized by highly constrained niches (MacArthur 1972; Mayr 1963; Stanley 1979). Because specialists can utilize only a limited array of ecological parameters, the geographic region, which encompasses their niche, will necessarily be smaller than those of eurytopic species.

Based on the relationship between geographic range and niche breadth, the ecological characteristics of the four brachiopod groups can be considered. Both Maysvillian restricted taxa and new species that evolve in the Richmondian from native taxa exhibit small range sizes characteristic of ecological specialists (Figure 4, Figure 5.7). Extrabasinal invaders are characterized by intermediate geographic ranges, and carryover taxa exhibit large geographic ranges suggestive of ecological generalists (Figure 4, Figure 5.7). Statistical results further support the classification of the Maysvillian restricted and native descendant species into one category of narrower niche breadth versus the carryover and invasive taxa into a second category of broader niche breadth. The geographic ranges of both the Maysvillian restricted taxa and the native descendant species are statistically significantly smaller than the carryover taxa (Table 5.1, 5.4). Conversely, the geographic ranges of the invader taxa, although smaller on average than carryover taxa, are statistically indistinguishable from the carryover taxa in several comparisons, particularly those conducted at the species level (Table 5.3, Figure 5.7). This apparent disparity between the two generalist groups results from temporal shifts in the geographic ranges of these groups, discussed more fully below.

The assignment of these taxon groups to specialist versus generalist ecologies is also consistent with their macroevolutionary histories. In general, specialist and generalist species exhibit characteristic differences in speciation and extinction rates; namely, generalist species tend to be longer lived with lower rates of extinction and speciation (Jackson 1974; Stanley 1979; Foote et al. 2008). This macroevolutionary relationship is a derivative of niche breadth; the larger the geographic region occupied by a species, the more likely that at least some part of that range will remain habitable following events of environmental or biotic perturbation. Consequently, one would predict a priori that the longest lived species in the dataset, the carryover species, should exhibit larger geographic ranges indicative of generalist ecologies; a prediction congruent with the results. In addition, the two specialist groups, the Maysvillian restricted taxa and the newly evolved descendant species, would be predicted to exhibit shorter temporal durations, which is also congruent with the results. Presumably, the native descendant taxa radiated into open specialist niches vacated by the extinction of the Maysvillian restricted species. Furthermore, the interpretation of invader taxa as ecological generalists is consistent with data from the modern invasion biology literature; most of the introduced species that have successfully invaded new ecosystems exhibit broad ecological tolerances (Davis 2009).

The stark difference between the geographic range size of the species and genera that become extinct by the end of the Maysvillian versus the range size of the carryover taxa (Table 5.1) suggests that species with limited niche breadth were unable to succeed in the new invasive regime in contrast to species with broader niches. Because the carryover taxa are present both before and after the influx of the invaders, it is possible that these taxa would have experienced a reduction in their realized niche, and hence geographic range, due to increased biotic interactions with the interbasinal invaders. To test this hypothesis, the geographic ranges of carryover taxa were compared before and after the onset of the invasion. Although mean taxon range decreased slightly after the invasion (more noticeably at the species level than the generic level), these differences were not statistically significant in any analysis (Table 5.2). Prior analyses (e.g., Tyler and Leighton 2007) have noted shifts in morphology of incumbent genera when they co-occur with morphologically similar invaders; however, these differences appear to represent local effects that do not manifest in species or genus-level geographic patterns.

An alternate explanation for the limited range size of invader and native descendant species relates to the relationship between geographic range size and taxon age. All taxa necessarily occupy small geographic ranges immediately following speciation, expand their geographic ranges as population sizes increase through time, and contract their geographic ranges as population size decreases prior to extinction (Vrba and DeGusta 2004; Foote et al. 2008). Consequently, young species may be expected to occupy smaller geographic ranges than older species with large population sizes (Webb and Gaston 2000). Studies of the timing of range expansion in the fossil record have determined that geographic range size increases rapidly following speciation (Liow and Stenseth 2007), and species typically attain full range size within one million years or less (Vrba and DeGusta 2004; Webb and Gaston 2000). Since the scale of temporal resolution of this study is approximately one million years per time slice, it is unlikely that the estimates of geographic range size analyzed herein capture the early post-speciation phase prior to full range size establishment. This possibility, however, cannot be excluded completely and may contribute to the observed increase in average geographic range of native descendant species from the C4 to C6 sequence or invader species between the C4 and C5 sequences.

Temporal Variations in Geographic Range

Biogeographic patterns shift dramatically across the six Cincinnatian sequences (e.g., Figure 5). From C1 to C3 sequences, carryover taxa exhibit larger mean geographic range size than Maysvillian restricted taxa. However, this difference is statistically significant only during the C3 sequence (Table 6). Apparently, it is during the final Maysvillian sequence only when establishing a larger geographic range is a predictor for taxon success across the invasion interval. At the onset of the Richmondian Invasion in the C4 sequence, carryover taxa are characterized by large mean geographic range sizes, and both new descendant species and invader taxa exhibit smaller ranges than carryover taxa (significant in three of four analyses for either coding scheme). New speciation is limited to only one to three species, but these are narrowly distributed, specialist style species. By the C5 sequence, a very diverse set of taxa is present. The most broadly distributed taxa are still the carryover taxa, but invaders exhibit larger ranges than their C4 sequence counterparts. Newly evolved species continue to occupy restricted ranges characteristic of ecological specialists. Although statistical differences remain at the genus level, the overlap between carryover and invader taxa at the species level renders these groups biogeographically indistinguishable in the C5 sequence. Finally, during the C6 sequence class differentiation breaks down and carryover, new species, and invader taxa have overlapping mean geographic ranges.

Holland and Patzkowsky (2007) interpreted three phases of community organization from their generic-level analysis of paleocommunity structure: stability from C1 to C3, reorganization in C4, and stability in C5 and C6 sequences. Our results are largely congruent with their pattern. Similar biogeographic patterns appear in C1 through C3 sequences, but each of C4, C5, and C6 exhibit different organizational patterns (Table 6, Figure 5).

The observed shifts in Richmondian biogeographic ranges, particularly the increase in invader range between the C4 and C5 sequence and the reduction in range size of both carryover and invader taxa between the C5 and C6 sequences, could potentially be attributed to several causes: response to invasive regime, niche shift of species due to environmental rather than biotic changes, or sampling bias. In this case, sampling bias can be ruled out as a primary driver of the observed pattern. The C4 sequence comprises only two of the four Cincinnatian depositional environments (deep and shallow subtidal), whereas the C5 sequence includes deposits of all four environments (also including offshore and peritidal) (Brett and Algeo 2001; Holland 2001). Based on ordination analyses, invader taxa most commonly occur in the deep and shallow subtidal communities, which are preserved in both the C4 and C5 sequences (Holland and Patzkowsky 2007). Because the percent area colonized by invaders increases with additional facies, range increase observed in the C5 sequence cannot be attributed to undersampling of facies. Furthermore, biogeographic patterns recovered when geographic range is normalized by number of localities (a proxy for sampling intensity) are congruent with those recovered from the outcrop area normalization method (compare Table 6 and Appendix 4). This finding indicates that observed biogeographic patterns are not controlled by sampling bias. The C6 sequence exhibits potential for sampling bias as this sequence is both geographically restricted (Figure 1) and limited to only the shallow subtidal and peritidal facies (Hay 2001), and this limitation should be considered when interpreting C6 patterns. Paleoenvironmental changes can be ruled out on similar grounds. All sequences record shallowing upward (Holland 1993); symmetrical biogeographic patterns, therefore, should be expected between sequences if range sizes are strongly affected by changes in relative sea level. Aside from the oceanographic shift in the early Richmondian—which affected all subsequent sequences equally, no significant paleoenvironmental changes are recorded sedimentologically. The biotic impact of the extrabasinal invaders, therefore, emerges as the most likely cause of the observed biogeographic shift.

Characteristics of Richmondian Invaders

A significant body of research has emerged within the modern invasion biology literature aimed at delimiting the characteristics of species that successfully invade new ecosystems (see review in Davis 2009). For a species to successfully invade a new habitat, four basic stages must occur: transport, establishment, spread, and impact (Lockwood et al. 2007). Modern invasive species must exhibit traits that facilitate both the transport and establishment stages. Certain characteristics, such as access to human transport devices (ship ballast water, fruit cargo containers) are matters of contingency; whereas other characteristics, such as the ability to exploit a variety food sources in the new ecosystem relate directly to the biological parameters of a species.

The most important feature for determining species' success during the transport stage is propagule pressure, both the number of individuals arriving at a location and the frequency of their introduction (Lockwood et al. 2005). In the fossil record, transport occurs through natural, in this case oceanographic, processes. In the Richmondian, successful invaders would have exhibited high larval or adult dispersal in order to establish new populations at such large distances from their ancestral populations. Richmondian propagule pressure was likely much lower than that observed with invaders in modern ecosystems as tens of thousands of years were available for transport and establishment of the Late Ordovician taxa.

Establishment relates to initial colonization of the new habitat by a small number of individuals, whereas spread refers to the stage in which the invader species is both abundant and becoming widespread. Species with broad ecological tolerances are more likely to succeed in these stages than ecological specialists (Duncan et al. 2003; Lockwood et al. 2007). Due to the short duration and small population sizes during this interval, these stages are unlikely to be preserved in the fossil record. Instead, the first appearance of Richmondian Invaders in the fossil record most likely reflects the impact stage, which refers to the interval when the invader taxon is fully established and naturalized or integrated within the community (Lockwood et al. 2007).

Within the rhynchonelliform brachiopods, the Richmondian invaders appear to conform to the basic set of expectations developed from modern invasion biology: high dispersal ability and broad ecological tolerances. Although adult rhynchonelliform brachiopods lack locomotor ability, their larvae are free swimming. The duration of the free-swimming larval stage may persist several weeks (Rudwick 1970, Peck and Robinson 1994), and recent rhynchonelliform brachiopod species have been observed to expand their ranges as much as 3000 to 4000 km in only 10,000 years (Curry and Endo 1991). If Ordovician brachiopods exhibited similar larval development, then propagule pressure would be sufficiently high in these species to produce the observed interbasinal invasions rapidly once the oceanographic patterns shifted in the early Richmondian Age. Furthermore, the geographic ranges of the invader species statistically overlap those of the generalist carryover species (Table 5.3). The Richmondian invaders, therefore, also exhibit the large geographic ranges characteristic of broadly adapted species.

Although the geographic ranges of invader taxa are statistically similar to the carryover taxa at the species level, differences emerge at the genus level (Table 5.3). This pattern is partly attributable to the fact that most invasive genera are monotypic within the Cincinnati region whereas some native genera include multiple species. Since the monophyly of these genera is also questionable, the species-level comparison may be a more robust comparison of carryover versus invasive taxa across the aggregated sequences. However, the sequence level details also suggest a more complex pattern than the species level aggregate analysis. In the C4 sequence, geographic range for invader taxa is significantly smaller (in all variations of analyses) than in the C5 sequence. This apparent shift in range size may be a sampling artifact. There are only three (or five) C4 invaders versus 12 (or 17) C5 invaders, and two of the three species exhibit C4 range sizes close to their C5 range size. Even so, the observed lag in either range establishment or expansion of range size among the invaders potentially indicates that the full biotic impact of the invasion was spread across the C4 and into the C5 sequences, an interval of at least one million years.

Impact of the Richmondian Invaders

The recovered biogeographic patterns suggest that during the early stages of the biotic invasion, in the transition from the C3 to the C4 sequence, narrowly adapted native species did not adjust to the invasive regime and become extinct, while broadly adapted native species persisted and flourished. The native generalists, in fact, appear to be more successful at establishing broad geographic ranges than the newly invading taxa in the first two million years after the initial invasion. Meanwhile, new species that evolved in the Richmondian from Cincinnati natives occupied smaller geographic ranges than either the carryover or invader taxa in both the C4 and C5 sequences (Figure 5, Table 6). The mean range size of the new species overlaps with the range size vacated by the Maysvillian restricted species; this suggests that species evolving after the initial C4 reorganization radiated into open specialist niches. Speciation is very limited in the C4 sequence, which may be due to the introduction of invaders. A pattern of reduced speciation rate associated with intervals of high interbasinal invasion has also been reported for Late Devonian brachiopods (Stigall 2008, 2010).

As noted by Patzkowsky and Holland (2007), the overall effect of the introduction of the invader taxa was to increase local diversity. Studies of modern coastal ecosystems have also documented increases rather than decreases in total diversity following the introduction of invasive species (Reise et al. 2006). Although ecologically specialized Maysvillian species become extinct by the end of the C3 sequence as discussed above, this did not result in an abrupt extinction peak because these extinctions are spread out over two sequences (Table 1; Holland 1997). In modern marine ecosystems, no compelling evidence exists for species extinction caused by competition with invasive species (Davis 2003). Based on the results of the Cincinnatian analysis, it appears that the long-term impacts of species invasions, at least in the case of the Richmondian Invasion, also do not result in extinction from direct competition for resources between native and invasive taxa. In fact, based on biogeographic data, it appears that the influx of invaders had no impact on reducing the realized niche of native carryover taxa. Instead, invasive species were limited by incumbent taxa to smaller realized niches during the C4 sequence compared with their C5 niches. This suggests that native communities, at least the generalist taxa within these communities, are more resilient to invader domination that suggested by short-term ecological observations.