Size-Dependent Trophic Patterns of Pallid Sturgeon and Shovelnose Sturgeon in a Large River System

Knowledge of ontogenetic shifts in consumer diets is important for understanding factors that influence energy flow in aquatic food webs (Michelsen et al. 1994; Takimoto et al. 2002). The position of fish species within a food web is of interest to managers because trophic position of a consumer often has consequences for population dynamics such as growth and survival. Gut content analysis has traditionally been the basis for constructing food webs but may provide an incomplete picture of trophic structure because it offers only a snapshot of a consumer's diet. Conversely, stable isotope analysis (SIA) offers a time-integrated method of examining trophic relationships between consumers and their prey (Peterson and Fry 1987).

Stable isotope analysis provides information regarding trophic position of a consumer by examining ratios of δ¹³C and δ¹⁵N isotopes incorporated into the consumer's tissue. Researchers commonly use stable δ¹³C ratios to determine energy sources in fishes. Depleted δ¹³C (i.e., more negative; lesser) values in fishes generally represent more pelagic (open water) or autochthonous (in-stream production) based food sources; whereas, more enriched (i.e., less negative; greater) values are characteristic of littoral/allochthonous (near-shore, out-of-stream production) based food sources (Peterson and Fry 1987). The stable δ¹⁵N signature can indicate the trophic position, with an approximately 3.4‰ increase in δ¹⁵N signature and a 1.0‰ increase in δ¹³C typically occurring between predators and their prey (Cabana and Rasmussen 1996; Vander Zanden et al. 1997).

Combining SIA with traditional diet studies provides a more complete picture of food web structure because it integrates consumer diet history over a long time interval. Additional benefits of SIA include determination of the primary energetic pathway incorporated into the tissue of a consumer from all fish sampled (Chipps and Garvey 2002). Empty stomachs significantly reduce sample size in gut content diet studies (Gerrity et al 2006; Wanner et al. 2007; Grohs et al. 2009) but are not an issue with SIA because nonlethal tissue samples are available from all captured fish (Tieszen et al. 1983). Gut content analysis requires either sacrificing the fish or undertaking stressful and labor-intensive field procedures such as gastric lavage (Wanner et al. 2007) or colonic flushing (Hoover et al. 2007). Researchers can collect nonlethal fin tissue samples for SIA quickly and with a minimal amount of stress to the fish, which is an important consideration for rare fishes (Andvik et al. 2010).

Understanding the trophic relationships and food web linkages of sturgeons is particularly important given the threatened or endangered status of the majority of sturgeon species worldwide (Bemis and Findeis 1994). Given the imperiled status of many sturgeons, and Scaphirhynchus species in particular, obtaining a detailed understanding of their positions in food webs has become a priority of fisheries biologists. The genus Scaphirhynchus contains two species native to the Missouri and Mississippi river systems: the pallid sturgeon S. albus and shovelnose sturgeon S. platorynchus. These two species are morphologically similar, use main channel habitat with fine substrates (Gerrity et al. 2008), and consume benthic macroinvertebrate prey (Gerrity et al. 2006; Hoover et al. 2007; Wanner et al. 2007). However, pallid sturgeon grow larger (max size ∼1,500 mm) than shovelnose sturgeon (max size ∼1,000 mm) and are generally piscivorous as adults (Pflieger 1997). The pallid sturgeon is a federally endangered species (U.S. Endangered Species Act [ESA 1973, as amended]; U.S. Fish and Wildlife Service [USFWS] 1990; Dryer and Sandvol 1993; Krentz 2004). Physical and thermal habitat alterations as well as introduced species are threats to species recovery (Bergman et al. 2008). Moreover, the absence of naturally produced juvenile pallid sturgeon in the Missouri River suggests that spawning, recruitment, or both are inadequate for local species survival (Bergman et al. 2008). There is also growing concern over the status of shovelnose sturgeon populations throughout their native range in the face of increasing commercial harvest and habitat alterations (Keenlyne 1997; Quist et al. 2002; Tripp et al. 2009). Documented harvest of pallid sturgeon by shovelnose sturgeon commercial fishers (Bettoli et al. 2009) led to closure of these fisheries where the species overlap due to their similarity of appearance (USFWS 2010).

Several authors have reported that diet overlap between shovelnose sturgeon and pallid sturgeon is generally low, owing to the increased proportion of fish prey and Ephemeroptera in pallid sturgeon diets (Gerrity et al. 2006; Braaten et al. 2007; Wanner et al. 2007; Grohs et al. 2009). However, this low overlap may be influenced by how studies examined diet composition (e.g., number of prey vs. biomass). Quantifying the importance of prey composition to growth of shovelnose sturgeon and pallid sturgeon has been difficult using gut content diet data because of small sample sizes, no prey items present in sturgeon stomachs, and/or spatiotemporal variation in diet.

Estimated trophic positions and food habits of multiple life stages obtained through a mixture of SIA and stomach content techniques will allow managers to examine the importance of varying prey taxa to Scaphirhynchus and to identify habitats where important prey taxa may be limiting (Bergman et al. 2008; Rapp et al. 2011). Additionally, SIA will allow managers to monitor long-term broad-scale changes in riverine food webs and possible implications for long-lived riverine species. This study examined patterns of δ¹⁵N and δ¹³C to infer the effects of body size and sturgeon species on trophic position for pallid sturgeon and shovelnose sturgeon in three reaches of the Missouri River.

Remnant, free-flowing reaches of the Missouri River, called Recovery Priority Management Areas (RPMAs; Figure 1), encompass available riverine habitats for stocking hatchery-reared fish as part of short-term recovery efforts for pallid sturgeon (USFWS 2008). Originating in Montana, RPMA 2 incorporates the area downstream of Fort Peck Dam (river kilometer [rkm] 2,850) to the headwaters of Lake Sakakawea in North Dakota (rkm 2,496) and the lower 298 rkm of the Yellowstone River. Forming the border of South Dakota and Nebraska, RPMA 3 runs from Fort Randall Dam (rkm 1,416) downstream to the headwaters of Lewis and Clark Lake (rkm 1,332). Beginning just downstream of RPMA 3, RPMA 4 extends from Gavins Point Dam (rkm 1,305) downstream to the confluence of the Mississippi River (rkm 0).

State and federal agencies captured pallid sturgeon (n  =  108) and shovelnose sturgeon (n  =  95) as part of annual standardized sampling (Drobish 2008) in Nebraska, North Dakota, South Dakota, and Montana (Table 1; Data S1). Sturgeon sampling relied upon trotlines, trammel nets, gill nets, or benthic otter trawls from RPMAs 2, 3, and 4 of the Missouri River (Welker and Drobish 2010). Prior to releasing the fish, researchers obtained a section of pectoral fin from each fish and preserved it in 95% ethanol (RPMAs 2 and 3) or froze it (RPMA 4) until analysis. The agencies did not collect shovelnose sturgeon from RPMA 2 because emphasis was on targeting pallid sturgeon. The majority of the pallid sturgeon in this study were fish of hatchery origin and known age, but genetic analysis revealed that a few of the largest individuals (>800 mm) lacking tags were wild fish (Welker and Drobish 2010; Shuman et al. 2010). All shovelnose sturgeon were wild fish of unknown age.

RPMAs 2–4 contain stocked pallid sturgeon, creating the possibility that fish in the SIA analysis may have acquired signatures in the hatchery. Andvik et al. (2010) found pallid sturgeon reared on a hatchery diet had acquired the isotopic signature of nonhatchery invertebrate prey after 186 days. All pallid sturgeon captured from RPMA 4 had individual tags with passive integrated transponders (PIT), and sampling took place during this study after a minimum of 2 y poststocking (Kirk Steffensen, Nebraska Game and Parks. Lincoln, Nebraska, personal communication). Individual stocking records from PIT tags were also available for most of the pallid sturgeon from RPMAs 2 and 3. Using the results of Andvik et al. (2010) as a guide, we assumed pallid sturgeon sampled in this study before 186 days poststocking may have retained SIA signatures from hatchery feed. Therefore, we removed two pallid sturgeon from RPMA 2 and one pallid sturgeon from RPMA 3 from analysis for this reason. Five pallid sturgeon from RPMA 3 were in the Missouri River for 153–155 d before capture, and we retained those fish for analysis because they had growth rates (0.4 g/d) that were 70% higher than those observed by Andvik et al. (2010). This should result in a faster turnover of hatchery acquired signatures because growth rate has been shown to affect turnover rates of tissues (Trueman et al. 2005).

Prior to analysis, we rinsed samples with deionized water, placed them in individual aluminum trays, and dried them at 60°C for 72 h. After drying, we homogenized samples into a fine powder using a mortar and pestle and placed them into individually labeled glass scintillation vials. Approximately 0.08 mg (±0.002) of the homogenized tissue went into individual tin capsules. We sent samples to the Cornell Isotope Laboratory (COIL, Cornell University, Ithaca, New York; http://www.cobsil.com/) for carbon (δ¹³C) and nitrogen (δ¹⁵N) analysis using a Thermo Delta V Isotope Ratio Mass Spectrometer interfaced to a Carlo Erba NC2500 elemental analyzer.

As a preservation medium, ethanol can affect δ¹³C and δ¹⁵N signatures in tissues of fish, primarily due to interactions with lipids within the tissues (Sweeting et al. 2004). As lipid content can vary considerably between species, species-specific correction factors are necessary to reduce sampling error due to preservation effects. We adjusted the pallid sturgeon δ¹⁵N values from samples preserved in ethanol (RPMAs 2, 3) for preservation effects using a regression developed from captive pallid sturgeon (Andvik et al. 2010). Because the regression model to correct for ethanol preservation effects on δ¹³C signatures in pallid sturgeon had a poor fit (R²  =  0.16; Andvik et al. 2010), we did not adjust samples from RPMAs 2 and 3. We also did not adjust ethanol-preserved shovelnose sturgeon samples from RPMA 3 for either element because no correction factor was available for shovelnose sturgeon. We only analyzed shovelnose sturgeon values from ethanol-preserved tissue in RPMA 3, where pallid sturgeon tissue also was preserved in ethanol.

Lipids are depleted in δ¹³C and have the potential to bias analyses in tissues with high lipid content. There are varying methods for dealing with lipids ranging from removal before analysis to mathematical correction models applied postanalysis. However, if lipid content is low it may have minimal effect on δ¹³C signatures and no corrections may be necessary. Many of these mathematical corrections rely on the C∶N ratio of the sample as a surrogate for lipid content (Kiljunen et al. 2006; Sweeting et al. 2006; Post et al. 2007). Post (2002) found lipid correction models minimally affected results from tissues with a C∶N ratio of less than 3.5. The mean C∶N ratio of the fin tissue samples in this study was 3.13 (±0.06 SE), so correcting for lipid removal was not necessary.

We used ANCOVA to determine if an interaction between FL and sturgeon species had an effect on δ¹⁵N and δ¹³C enrichment in RPMAs 3 and 4 (Table S1, Supplemental Material). The ANCOVA model was as follows: N or C  =  species (2 levels) + FL (covariate) + species*FL interaction. If we found significant interactions within an RPMA, then we applied separate linear regression models to each species. We used linear regression to examine size effects (FL) on δ¹⁵N and δ¹³C enrichment in pallid sturgeon in RPMA 2, and separately for pallid sturgeon and shovelnose sturgeon in RPMAs 3 and 4. In situations with curvilinear data, we compared R² values of linear and curvilinear models and used the model with the highest R² value.

We used ANOVA (ANOVA N or C  =  size [three levels]) to compare trophic position between three groups of sturgeon: small pallid, large pallid, and shovelnose sturgeon. We grouped pallid sturgeon into small (<500 mm) or large (>500 mm) size categories based on the approximate size observed in field studies when pallid sturgeon became mostly piscivorous. Fish composed >50% of dry weight of pallid sturgeon diets for fish of >500 mm FL (Grohs et al. 2009). In contrast, we analyzed shovelnose sturgeon as a single group regardless of body size since benthic macroinvertebrates were predominant in their diets at all length classes (Held 1969; Modde and Schmulbach 1977; Berry 2002; Braaten et al. 2007).

If we detected differences among sturgeon groups using ANOVA, we used a pairwise t test with Bonferroni's adjustment for multiple comparisons to compare trophic position between groups (Bonferroni 1936). We performed all analyses in SAS (2009) with an α  =  0.05. We only compared mean δ¹⁵N and δ¹³C values for large and small pallid sturgeon and shovelnose sturgeon graphically for RPMA 3 because tissue samples were preserved in ethanol.

The evaluation of relative trophic position and energy flow within a system is straightforward using stable isotope values (Post 2002). However, comparisons across systems require baseline corrections to isotope values to account for considerable variation in δ¹⁵N and δ¹³C at the base of the food web. Without information on isotope values of lower trophic levels from each river segment, we could not evaluate differences in trophic position for pallid sturgeon and shovelnose sturgeon among RPMAs. As a result, we limited our analysis to general patterns of trophic structure as related to body size and between pallid sturgeon and shovelnose sturgeon within each RPMA.

Enrichment of δ¹⁵N for pallid sturgeon and shovelnose sturgeon in RPMA 4 increased with increasing body size (Figure 2). FL significantly influenced differences in δ¹⁵N signatures of pallid sturgeon and shovelnose sturgeon from RPMA 4 (ANCOVA: F_1,70  =  28.17, P < 0.01) but not by species (ANCOVA: F_1,70  =  0.84, P  =  0.36). Pallid sturgeon δ¹⁵N values (range: 12.8–16.7‰) increased with FL (range: 325–1,057 mm), showing that small fish transitioned from more depleted δ¹⁵N (13–14‰) to a more enriched signature (15–17‰) as they grew. Shovelnose sturgeon δ¹⁵N values (δ¹⁵N range: 14.6–17.2‰) were also positively correlated with FL (range: 341–719 mm) but spanned a narrower range than those of pallid sturgeon.

Comparison of δ¹³C signatures for pallid sturgeon and shovelnose sturgeon in RPMA 4 revealed a significant interaction between FL and species (ANCOVA: F_1,70  =  10.66, P < 0.01). Therefore, we fit separate regressions for each species: pallid sturgeon, R²  =  0.34, N  =  48, P < 0.01; shovelnose sturgeon R²  =  0.21, N  =  23, P < 0.05. Pallid sturgeon δ¹³C values (δ¹³C range: −23.4 to −18.0) in RPMA 4 increased with FL, whereas shovelnose sturgeon δ¹³C values (δ¹³C range  =  −23.7 to −20.1) decreased with FL.

Analysis of δ¹⁵N signatures of pallid sturgeon and shovelnose sturgeon in RPMA 3 revealed a significant interaction between FL and species (ANCOVA: F_1,110  =  15.09, P < 0.01); whereas, the interaction for δ¹³C signatures was insignificant. We fit separate regressions to δ¹⁵N signature by FL for each species: pallid sturgeon R²  =  0.34, N  =  42, P < 0.01; shovelnose sturgeon R²  =  0.01, N  =  72, P  =  0.40 (Figure 3). Pallid sturgeon δ¹⁵N signatures in RPMA 3 increased with FL, while shovelnose sturgeon δ¹⁵N signatures showed no relationship with FL. Comparisons of δ¹³C signatures of pallid sturgeon and shovelnose sturgeon in RPMA 3 showed significant main effects of FL (ANCOVA: F_1,110  =  9.8, P < 0.01) and species (ANCOVA: F_1,110  =  5.6, P  =  0.02).

The relation of pallid sturgeon body size to δ¹⁵N and δ¹³C in RPMA 2 was similar to RPMA 3, regression analysis showed pallid sturgeon δ¹⁵N signature in RPMA 2 increased with FL: R²  =  0.48, N  =  18, P < 0.01 (Figure 4). There was no significant relationship between δ¹³C signatures and FL of pallid sturgeon in RPMA 2.

There were significant differences in δ¹³C (F_2,70  =  26.51, P < 0.001) and δ¹⁵N (F_2,70  =  29.94, P < 0.001) among sturgeon groups in RPMA 4 (Figure 5, Table 2). Small pallid sturgeon and shovelnose sturgeon had similar δ¹³C signatures. Large pallid sturgeon had more enriched δ¹³C signatures than small pallid sturgeon and shovelnose sturgeon. Small pallid sturgeon and all shovelnose sturgeon had similar δ¹³C signatures, but small pallid sturgeon had less enriched δ¹⁵N signatures than all shovelnose sturgeon. Large pallid sturgeon δ¹³C signatures were significantly more enriched than all shovelnose sturgeon and small pallid sturgeon. Large pallid sturgeon δ¹⁵N signatures were similar to all shovelnose sturgeon but significantly more enriched than small pallid sturgeon. Graphically, the relative positions of shovelnose sturgeon and the two pallid sturgeon size classes in RPMA 3 were similar to RPMA 4.

Patterns of pallid sturgeon and shovelnose sturgeon isotope composition corroborate previous studies based on field-derived diet information that found use of invertebrate prey by small, young pallid sturgeon and an ontogenetic diet shift to increased piscivory in large juveniles and adults (Carlson et al. 1985; Hoover et al. 2007; Wanner et al. 2007; Grohs et al. 2009). The relationship between δ¹⁵N signatures and fish size in the current study suggests that large pallid sturgeon fed at a higher trophic level than small pallid sturgeon. Moreover, δ¹⁵N signatures of pallid sturgeon > 800 mm were separated by a full trophic level (i.e., >3.4‰) from smaller pallid sturgeon (<400 mm). However, intermediate δ¹⁵N signatures for 500 to 800 mm pallid sturgeon suggested that this transition in diet was gradual. This transition was most evident in RPMAs 3 and 4, where our samples included the largest size range of pallid sturgeon. Wanner et al. (2007) and Grohs et al. (2009) found invertebrate as well as fish prey in the stomach contents of pallid sturgeon of >500-mm FL, also indicating a gradual transition to piscivory.

Pallid sturgeon δ¹³C signatures in RPMA 4 varied with increasing FL, indicating the energy source changed as sturgeon grew, possibly resulting from a shift in diet. We observed a similar pattern in pallid sturgeon δ¹³C signatures in RPMA 3; however, pallid sturgeon δ¹³C signatures in RPMA 2 showed no significant relationship with FL. This may have been due to the smaller sample size from RPMA 2 and (or) the smaller size range of pallid sturgeon from this reach compared with RPMAs 3 and 4. While our lack of basal isotope signatures for each RPMA prevents direct comparisons across river segments, pallid sturgeon in all RPMAs displayed an enrichment of δ¹⁵N with increasing FL. Additionally, pallid sturgeon in RPMAs 3 and 4 displayed enrichment in δ¹³C with increasing FL. These observed patterns of δ¹⁵N and δ¹³C enrichment in each RPMA are consistent with the expectation of an ontogenetic diet shift to piscivory.

For shovelnose sturgeon in RPMA 4 the relationship between δ¹⁵N and FL spanned a narrower range than that in pallid sturgeon and suggested most foraging occurred at a single trophic level for the majority of sampled fish. The similar enrichment of δ¹⁵N signatures between pallid sturgeon and shovelnose sturgeon was unexpected because field studies of shovelnose sturgeon diets in the Missouri and Platte rivers found benthic macroinvertebrates were the dominant prey in all length classes of fish, with piscivory being rare (Held 1969; Modde and Schmulbach 1977; Berry 2002; Braaten et al. 2007; Wanner et al. 2007; Rapp et al. 2011). There was no significant relationship between FL and δ¹⁵N for shovelnose sturgeon in RPMA 3, which was likely due to the truncated size distribution of our samples. In RPMA 3, no capture of shovelnose sturgeon < 475 mm occurred from 2003 to 2010 (Shuman et al. 2010), indicating an aging population with minimal recruitment.

Sampling of pallid sturgeon and shovelnose sturgeon occurred in RPMA 3 from 2005 to 2008, while sampling in the other RPMAs only occurred during 2008. If Missouri River food webs changed during this time, there is the possibility that the broader temporal scale for RPMA 3 may have had some effect on δ¹⁵N and δ¹³C enrichment, although the similar patterns in each RPMA suggest that any effects of sampling period were minimal. Additionally, since we treated each RPMA as a separate system in our analysis and avoided direct comparisons between RPMAs, the temporal synchronization of sampling in each RPMA is not critical to our interpretations. We believe the similarity of patterns in RPMA 3 from 2005 to 2008 to those from RPMAs 2 and 4 in 2008 may indicate food web stability in the highly regulated Missouri River (Pegg et al. 2003).

Similarity in δ¹⁵N signatures between pallid sturgeon and shovelnose sturgeon was unexpected given the piscivorous nature of adult pallid sturgeon (Carlson et al. 1985). One explanation for this may involve the observed differences in invertebrate diet composition between shovelnose and pallid sturgeon. Grohs et al. (2009), for example, showed that Ephemeroptera represented a significant proportion of pallid sturgeon diets (∼25% by weight), whereas net-spinning caddisflies (Trichoptera: Hydropsychidae) only represented about 0.3%. In contrast, a detailed study of shovelnose sturgeon diets in the Missouri River showed that Hydropsychidae represented over 50% of the diet by wet weight (Modde and Schmulbach 1977). Analysis of the feeding ecology of Hydropsychidae found animal remains constituted a significantly higher proportion of diet in spring and summer, especially for later larval developmental stages (Fuller and Mackay 1980). In a previous study, omnivorous or predatory aquatic invertebrates had elevated trophic levels and δ¹⁵N signatures in relation to herbivorous invertebrates in the same locations; for example, carnivorous opossum shrimp Mysis relicta had a δ¹⁵N signature more than 3‰ greater than herbivorous opossum shrimp (Branstrator et al. 2000). Thus, a higher proportion of invertebrates (i.e., Hydropsychidae) with potentially enriched δ¹⁵N values may have contributed to enriched δ¹⁵N values we observed for shovelnose sturgeon. Similarly, fish eggs can be part of the diets of shovelnose sturgeon in RPMA 3 (Dane Shuman, U.S. Fish and Wildlife Service. Pierre, South Dakota, personal communication) and could potentially increase δ¹⁵N signatures if consumed in significant amounts. However, the overall significance of fish eggs to the diet of shovelnose sturgeon is unknown.

Changes in invertebrate prey composition for shovelnose sturgeon may also explain the observed decrease in δ¹³C signatures with increasing FL in RPMA 4; although, this decrease was less pronounced than the increase in pallid sturgeon δ¹³C signatures. Previous researchers have documented variation in δ¹³C signatures within a site, even among closely related species. Trichopterans in a New Zealand stream had mean δ¹³C values vary by over 2‰ between species (Rounick et al. 1982). The decrease in δ¹³C signatures we observed potentially could have been caused by large shovelnose sturgeon foraging on different invertebrate prey as they grew and also potentially influenced pallid sturgeon stable isotope signatures. Detailed taxonomic resolution (i.e., genus level) is generally lacking for diet studies of both species (Carlson et al. 1985; Berry 2002; Gerrity et al. 2006; Hoover et al. 2007; Wanner et al. 2007; Grohs et al. 2009). However, Rapp et al. (2011) found four genera of Chironomidae comprised over 80% of ingested diet items by shovelnose sturgeon in the Platte River in Nebraska.

Alternatively, the variability in δ¹⁵N signatures of shovelnose sturgeon may be partially due to variation in age because δ¹⁵N signatures can become enriched in older fish, independent of diet. Overman and Parrish (2001) found 81% of variation in walleye Sander vitreus δ¹⁵N signatures (range: 15.3–19.2‰) was due to age (range: 2–27 y). ¹⁵N accumulation over the lifespan of the fish may explain this enrichment. The majority of pallid sturgeon in this study were juveniles and early adults of known age from PIT tagging records. Additionally, the size of the few unknown pallid sturgeon suggests a similar age range (<850-mm FL and age <12 y; Keenlyne and Jenkins 1993), whereas shovelnose sturgeon were not PIT tagged and of unknown age. Everet et al. (2003) found ages ranged from 12 to 30 y for 500 to 600 mm FL shovelnose sturgeon in the Missouri River downstream of Garrison Dam in North Dakota. Thus, the enrichment in shovelnose sturgeon δ¹⁵N signatures relative to those of pallid sturgeon may have been partially due to differences in ages of fish in our samples. Regardless, the unexpectedly enriched δ¹⁵N signatures of shovelnose sturgeon warrant further investigation because any number of factors (including shifts in diet) may have caused enrichment similar to that in pallid sturgeon.

The divergence of δ¹³C signatures of large (>500 mm) pallid sturgeon from those of small pallid sturgeon and all shovelnose sturgeon supported previous findings of an ontogenetic shift to piscivory among pallid sturgeon (Grohs et al. 2009). Large pallid sturgeon used different energy resources than shovelnose sturgeon or small pallid sturgeon (Figure 5). Small differences in δ¹³C signature can represent important changes in food web structuring. Within RPMA 4, the differences in δ¹³C signature between large pallid sturgeon and shovelnose sturgeon (mean difference  =  1.9‰) and between large and small pallid sturgeon (mean difference  =  1.5‰) are comparable to the differences in δ¹³C signatures from the literature after significant changes in food web dynamics. For example, lake trout Salvelinus namaycush δ¹³C had a mean difference of 1.8‰ after a shift from a littoral fish diet to pelagic zooplankton (Vander Zanden et al. 1999). Cucherousset et al. (2007) observed a mean difference of 1.6‰ for brown trout Salmo trutta after the introduction of nonnative brook trout Salvelinus fontinalis. Similar shifts in δ¹³C signatures occurred for age-0 largemouth bass Micropterus salmoides (mean difference  =  0.8‰; Yasuno et al. 2012) during a shift to piscivory and for walleye after an introduction of smallmouth bass (mean difference  =  1.7–2.7‰; Galster et al. 2012).

The δ¹³C signatures of small pallid sturgeon and shovelnose sturgeon show that they used similar energy resources in RPMAs 3 and 4. In contrast, previous studies found low diet overlap between juvenile pallid sturgeon and shovelnose sturgeon (Gerrity et al. 2006; Wanner et al. 2007). However, the majority of pallid sturgeon in these studies were larger than 500 mm, and the low degree of overlap was primarily driven by fish prey in pallid sturgeon diets. Gerrity et al. (2006) sampled pallid sturgeon (mean FL  =  538 mm) from the upper Missouri River in Montana (upstream of RPMA 2), where they found diets dominated by fish, primarily native cyprinids of the genus Machrybopsis. Machrybopsis spp. are less abundant in the middle and lower Missouri River than in the upper Missouri River (Dieterman and Galat 2004). Wanner et al. (2007) found johnny darters Etheostoma nigrum (n  =  3) and small channel catfish Ictalurus punctatus (n  =  4) composed the majority of fish in the diet of pallid sturgeon (n  =  28, mean FL  =  585 mm) from the middle Missouri River (RPMA 3), suggesting pallid sturgeon may be able to forage on a variety of potential prey, given variations in local abundances.

Preservation techniques affect stable isotope signatures (Kelly et al. 2006; Sweeting et al. 2004); however, samples preserved in ethanol provide reasonable estimates of isotopic signatures and the development of species-specific correction factors allow researchers to minimize bias from preservation technique (Edwards et al. 2002; Schmidt et al. 2009). We applied existing correction factors for pallid sturgeon δ¹⁵N (Andvik et al. 2010) to our ethanol-preserved samples in RPMAs 2 and 3. Despite different preservation techniques, the relative patterns of increased δ¹⁵N enrichment with pallid sturgeon size were consistent in the three Missouri River RPMAs. The relative relation of shovelnose sturgeon δ¹³C and δ¹⁵N signatures to small and large pallid sturgeon was also similar in RPMAs 3 and 4.

In this study we compare δ¹⁵N and δ¹³C enrichment patterns of pallid sturgeon and shovelnose sturgeon within systems (i.e., within rather than between RPMAs). Post (2002) summarized why stable isotope comparisons within a system are relatively straightforward, although baseline isotopic values are necessary to address ecological questions between systems. In such situations, baseline isotopic signatures act as a scalar and allow comparisons between systems by weighting isotopic signatures of the organisms of interest against a baseline organism of known trophic position. In the case of comparisons within a system, however, the baseline scalar will have no effect on observed patterns or relative trophic position of the organisms of interest because it will affect all isotopic signatures in exactly the same way. Thus, our within system comparisons are valid in the absence of baseline isotopic values because the application of baseline scalars would not differentially affect relative δ¹⁵N enrichment of pallid sturgeon and shovelnose sturgeon.

Our findings show that stable isotope analysis is an effective tool for investigating relative trophic position of sturgeons in large rivers. Although there have been several diet studies of pallid sturgeon and shovelnose sturgeon (Berry 2002; Gerrity et al. 2006; Braaten et al. 2007; Hoover et al. 2007; Wanner et al. 2007; Grohs et al. 2009), the limitations of stomach content studies (i.e., absence of prey items in predator stomachs, small sample sizes, stressful handling procedures) make SIA an attractive technique when working with rare or endangered species such as sturgeons. The incorporation of stable isotope techniques will aid researchers in rapidly obtaining a more accurate picture of sturgeon trophic interactions by eliminating many of the limitations inherent in stomach content analysis. Additionally, SIA will provide a long-term picture of diet composition rather than the brief snapshot obtained via stomach contents. This long-term approach not only complements existing stomach content data, but may be of greater value for long-lived fishes such as sturgeons. Our study supports the conclusions of field-based studies of diet composition (Modde and Schmulbach 1977; Hoover et al. 2007; Wanner et al. 2007) that shovelnose sturgeon do not occupy the same niche as pallid sturgeon as a result of their distinct foraging habits. The establishment of a baseline isotope composition of sturgeon allows fisheries researchers to monitor the impacts of future disturbances (i.e., changes in the aquatic community, hydrologic variation, introduction of exotic species) within the ecosystem and better predict their effects on sturgeon populations.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Sturgeon were collected from RPMAs 2–4 of the Missouri River between 2005 and 2008. Excel spreadsheet containing all data used in the analyses presented in this manuscript. Sturgeon species (pallid sturgeon Scaphirhynchus albus and shovelnose sturgeon Scaphirhynchus platorynchus), fork length (mm), Recovery Priority Management Areas (RPMAs), and ¹⁵N and ¹³C signatures are provided.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S1 (41 KB XLS).

Table S1. Analyses performed for pallid sturgeon Scaphirhynchus albus and shovelnose sturgeon Scaphirhynchus platorynchus collected from Recovery Priority Management Areas (RPMA) 2-4 of the Missouri River between August 2005 and March 2008.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S2 (28 KB DOCX).

Reference S1. Bergman HL et al. Research needs and management strategies for pallid sturgeon recovery. Report to the U.S. Army Corps of Engineers. Proceedings of a workshop held July 31–August 2, 2007, St. Louis, Missouri.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S3 (2.9 MB PDF).

Reference S2. Berry CR Jr. 2002. Diet of shovelnose sturgeon downstream from Gavins Point Dam. Final report.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S4 (916 KB PDF).

Reference S3. Drobish MR. 2008. Pallid sturgeon population assessment program.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S5 (324 KB PDF).

Reference S4. Dryer MP, Sandvol AJ. 1993. Recovery plan for the pallid sturgeon.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S6 (1.0 MB PDF).

Reference S5. Welker TL, Drobish MR (eds). 2010. Missouri River Standard Operating Procedures for Fish Sampling and Data Collection, Volume 1.5.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S7 (2.6 MB PDF).

Reference S6. U.S. Fish and Wildlife Service. 2008: Pallid sturgeon (Scaphirhynchus albus) range-wide stocking and augmentation plan.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S8 (457 KB PDF).

Reference S7. Shuman DA, Klumb RA, Wanner GA, Grohs KL. 2010. 2009 Annual report pallid sturgeon population assessment and associated fish community monitoring for the Missouri River: Segments 5 and 6.

Found at DOI: http://dx.doi.org/10.3996/022012-JFWM-013.S9 (858 KB PDF).

We thank the following for assistance in the field: Ryan Wilson, Everett Nelson, and Zac Sandness (U.S. Fish and Wildlife Service, Missouri River Fish and Wildlife Conservation Office, Bismarck, North Dakota), Dane Shuman, Greg Wanner, and Steve Heutmaker (U.S. Fish and Wildlife Service, Great Plains Fish and Wildlife Conservation Office, Pierre, South Dakota), and Kirk Steffensen and Marty Hamel (Nebraska Game and Parks Commission, Lincoln, Nebraska). We also thank the Journal Subject Editor and three anonymous reviewers for comments on a previous version of this manuscript. All animals used in this study were handled according to animal use and care guidelines established by South Dakota State University (Animal Welfare Assurance no. A3958-01). Funding for this project was provided by the Missouri River Recovery Program, Omaha and Kansas City Districts of the U.S. Army Corp of Engineers, Northwestern Division. Andrew Wuestewald helped with sample preparation in the laboratory.

Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.