Food Web Interactions Associated With a Lahontan Cutthroat Trout Reintroduction Effort in an Alpine Lake

Introduced species can have a significant effect on aquatic and terrestrial food webs. In western North America most of the native Cutthroat Trout Oncorhynchus clarkii subspecies have been negatively affected by the introduction of nonnative salmonids. In addition to competition and hybridization threats, predation can be responsible for substantial declines in native trout populations. For example, nonnative Lake Trout Salvelinus namaycush have consumed large numbers of Yellowstone Cutthroat Trout O. c. bouveri in Yellowstone Lake (Ruzycki et al. 2003), which has caused cascading effects through the aquatic and terrestrial food webs (Koel et al. 2005).

Lahontan Cutthroat Trout O. c. henshawi was the apex predator throughout their historic range in the Lahontan hydrographic basin of northern Nevada, southeastern Oregon, and northeastern California, which includes the Lake Tahoe Basin of California–Nevada (Behnke 1992). By the mid-20th century Lahontan Cutthroat Trout had been extirpated from the Lake Tahoe Basin due to a suite of anthropogenic perturbations, which included overfishing, timber harvest, loss of access to spawning habitats, and introduction of nonnative species. Currently listed as threatened under the U.S. Endangered Species Act (ESA 1973, as amended), Lahontan Cutthroat Trout are no longer present in >90% of their historic range and >99% of historic lacustrine habitat (Coffin and Cowan 1995). Consequently, the U.S. Fish and Wildlife Service (USFWS) recovery plan for this subspecies, which identifies actions that are considered necessary for recovery and protection of federally listed species, has prioritized reestablishing lacustrine populations (Coffin and Cowan 1995).

Fallen Leaf Lake, California, is the site of an ongoing effort to reestablish a lacustrine population of Lahontan Cutthroat Trout (hereafter Cutthroat Trout). Fisheries managers have stocked Cutthroat Trout into Fallen Leaf Lake since 2002 on an annual basis, with a few exceptions (Table 1), with the goals of 1) identifying factors limiting Cutthroat Trout reintroduction and 2) establishing a naturally reproducing population of Cutthroat Trout (Al-Chokhachy et al. 2009). However, the Fallen Leaf Lake food web has been altered by the introduction of various nonnative species, which may represent significant challenges to Cutthroat Trout reintroduction and may necessitate an adaptive management strategy. For example, empirical observation (Allen et al. 2009) and putative relationships based on bioenergetic modeling (Al-Chokhachy et al. 2009) suggest that nonnative Lake Trout can consume large numbers of stocked Cutthroat Trout in Fallen Leaf Lake.

The aim of this study was to provide information on the current characteristics of the Fallen Leaf Lake food web. Although researchers have previously studied the influence of Lake Trout on Cutthroat Trout in Fallen Leaf Lake (i.e., Allen et al. 2006; Al-Chokhachy et al. 2009), there is limited information on interactions between recently stocked Cutthroat Trout and other species within the Fallen Leaf Lake food web, and no information on prey use by recently stocked Cutthroat Trout. Therefore, we sought to provide more inclusive information on trophic characteristics of the aquatic species assemblage in Fallen Leaf Lake. Specific objectives of this study were to 1) quantify the general dietary characteristics of recently stocked Cutthroat Trout and other fishes in Fallen Leaf Lake, 2) quantify dietary overlap between Cutthroat Trout and other fishes in Fallen Leaf Lake, 3) quantify changes in the trophic characteristics of fishes in Fallen Leaf Lake following Cutthroat Trout stocking, and 4) provide a descriptive characterization of the contemporary food web in Fallen Leaf Lake. Results of this study will provide information on prey use by recently stocked Cutthroat Trout, may elucidate potential competitive or predatory interactions between Cutthroat Trout and other fishes that have not been previously identified, and will provide data that may be used for comparative purposes following ongoing or future management actions.

Fallen Leaf Lake is a glacially formed, small (567 ha), deep (about 125 m maximum depth), monomictic, oligotrophic lake that was historically occupied by Lahontan Cutthroat Trout and which is connected to the larger Lake Tahoe by a single tributary (Taylor Creek). The contemporary native fish assemblage in Fallen Leaf Lake includes Mountain Whitefish Prosopium williamsoni, Tahoe Sucker Catostomus tahoensis, Mountain Sucker Catostomus platyrhynchus, Lahontan Redside Richardsonius egregius, Speckled Dace Rhinichthys osculus, Tui Chub Gila bicolor, and Paiute Sculpin Cottus beldingii. Fallen Leaf Lake has been the site of numerous nonnative aquatic species introductions including mysid shrimp Mysis diluviana, signal crayfish Pacificus leniusculus, Brown Trout Salmo trutta, Kokanee Oncorhynchus nerka, Lake Trout, and Rainbow Trout Oncorhynchus mykiss. During 2009 and 2010 Fallen Leaf Lake was thermally stratified by the beginning of July, the thermocline varied from about 10–20 m; thermal stratification persisted into early October, and the lake destratified near the end of October (Figure 1).

We sampled the fish assemblage in Fallen Leaf Lake using gill nets (Table S1, Supplemental Material) and minnow traps. Gill nets were constructed of monofilament, and had a bottom lead line and a top float line, five panels (38-, 51-, 76-, 102-, and 127-mm stretch measure), and a total length of 38 m and a depth of 1.8 m. We set all gill nets benthic and perpendicular to the lake shoreline at depths from 1.8 to 47.5 m. Our spatial distribution of gill nets followed a systematic sampling design around the perimeter of Fallen Leaf Lake, with the exception that we did not sample areas near urban development (e.g., recreational boat docks and marinas). We set gill nets during five distinct sample events in 2009 and 2010 and stratified gill net sets by depth (the average depth of shallow gill nets was 9.9 m and the average depth of deep gill nets was 30.2 m; Table 2). Sample events generally consisted of up to six consecutive days of sampling and up to five gill nets set per day. We set most gill nets (n = 185) 2.31 ± 0.75 h (mean ± SD) before sunrise, allowed them to soak for 4.38 ± 0.40 h, and pulled them after the morning crepuscular period. During sample event 2 and sample event 3 in 2009 (n = 20) we set gill nets 1.29 ± 0.31 h before sunset, allowed them to soak overnight for 12.97 ± 0.57 h, and pulled them the next day after the morning crepuscular period.

We set minnow traps during the spring and autumn of 2009 and the summer and autumn of 2010 to sample taxa that were not susceptible to gill nets (e.g., Lahontan Redside, Speckled Dace, signal crayfish). Minnow traps were 40 cm long with a diameter of 20 cm and were constructed of steel mesh (5-mm bar measure). We set minnow traps in the morning and allowed them to soak for about 24 h; we added a small amount of bait to each minnow trap. We used a tow net or a Wisconsin-style plankton net to sample zooplankton and mysids. We conducted net tows during the night in 2009 and 2010. We collected freshwater snails by hand from littoral habitat during the autumn of 2010.

We identified fish sampled using gill nets and minnow traps to species, enumerated them, and measured them for length (mm total length; Table S2, Supplemental Material). We constructed length-frequency histograms to characterize the size structure of fishes sampled using gill nets and to provide context for potential predatory interactions related to gape limitations. We calculated catch per unit effort (C/f; fish/h) for each gill net by species. We used a mixed linear model and least-square means to estimate the overall C/f by species and C/f by species, depth strata, and season (PROC MIXED; SAS version 9.4; SAS Institute Inc., Cary, NC). We categorized gill net data from sample event 1 in 2009 and sample events 1 and 2 in 2010 as spring (i.e., prior to development of a distinct thermocline; Table 2; Figure 1); gill net data from sample event 4 in 2009 and sample events 3 and 4 in 2010 as summer (i.e., during the period of thermal stratification; Table2; Figure 1); and gill net data from sample event 5 in 2009 and sample event 5 in 2010 as autumn (i.e., during the period of thermal destratification; Table 2; Figure 1). We did not include gill net data from sample event 2 and sample event 3 in 2009 in this analysis because we only sampled shallow depths during these sample periods and because nets were allowed to soak overnight as opposed to only during the morning crepuscular period (see above). The linear mixed model included three fixed effects (species, depth strata, season), a three-way interaction term (species × depth strata × season effect), and a random effect (year). We used least-square means estimate statements (LSMESTIMATE; SAS version 9.4; SAS Institute Inc.) to estimate overall C/f by species and to estimate C/f by species, depth strata, and season.

We removed whole stomachs from Cutthroat Trout (n = 24), Mountain Whitefish (n = 66), Tui Chub (n= 6), Lake Trout (n = 303), Brown Trout (n = 22), Rainbow Trout (n = 18), and Kokanee (n = 18) that died during gill net sampling. The range of individual lengths of fishes used for stomach content analysis was similar to the range of individual lengths of all fishes captured (Table 3). We preserved whole stomachs in 10% formalin in 2009 and 70% ethanol in 2010 and transported them to the laboratory for stomach content analysis. We examined stomach contents using a dissecting microscope (×6 to ×40 magnification) and categorized them as 1) zooplankton, 2) mysids, 3) aquatic insects, 4) terrestrial arthropods, 4) crayfish, and 5) fish (Table S3, Supplemental Material). It was generally not possible for us to identify fish with high taxonomic precision due to stage of digestion; however, we noted fish species when possible. We dried stomach contents for each fish to a constant mass at 60°C (Bowen 1996) and weighed them to the nearest 0.1 mg (sensitivity of balance) by diet category (Table S3, Supplemental Material); we assigned a value of 0.09 mg to stomach contents by diet category that were present, but weighed less than the sensitivity of the balance (i.e., 0.1 mg).

We separated Lake Trout into two size groups for stomach content and stable isotope analyses based on estimated length at switch to piscivory. Lake Trout may exhibit a size-based switch from a primarily nonpiscivorous diet to a primarily piscivorous diet (Guy et al. 2011). We used a logistic regression model to estimate the length at switch to piscivory for Lake Trout in Fallen Leaf Lake. This model used data from Lake Trout with nonempty stomachs sampled during this study (n = 190; see below). The presence (1) or absence (0) of fish in the diet of Lake Trout was modeled as a function of Lake Trout total length. We assumed that the switch to piscivory occurred at the length where the estimated probability of fish in the diet was 0.5. Based on this analysis, we separated Lake Trout into one group of Lake Trout <499 mm (hereafter small Lake Trout) and one group of Lake Trout ≥499 mm (hereafter large Lake Trout).

We calculated frequency of occurrence (Bowen 1996) of empty stomachs by species and species by size group for Lake Trout. We calculated percentage of composition by dry mass (Bowen 1996) of each diet category for nonempty stomachs by species and species by size group for Lake Trout to quantify the general dietary characteristics of recently stocked Cutthroat Trout and other fishes in Fallen Leaf Lake (objective 1). We did not analyze stomach content by sample period because 1) we were primarily interested in general diet composition, 2) general diet composition patterns are likely more complimentary to stable isotope analysis, which integrate trophic characteristics over periods of months to years (see Church et al. 2009; Martínez del Rio et al. 2009), and 3) sample sizes would have precluded precise diet composition estimates over short time periods. We calculated percentage of overlap (C_xy) as a measure of diet overlap between Cutthroat Trout and all other fishes sampled for stomach contents (objective 2). Percentage of overlap was calculated as

where p_x,i and p_y,i are the proportions by dry mass of diet category i in the stomach contents of species x and y (Schoener 1970; Hurlbert 1978; Bowen 1996).

Stable isotope analysis provides an indirect, time-integrated method for examining trophic relationships. The δ¹⁵N value of a consumer is generally greater than that of its diet due to isotopic discrimination (e.g., +3.4‰; Post 2002; Martínez del Rio et al. 2009); therefore, δ¹⁵N is useful for examining the trophic position of species in a food web. Differences in δ¹³C between a consumer and its diet are generally very small (e.g., +0.4‰; France and Peters 1997; Post 2002; McCutchan et al. 2003; Martínez del Rio et al. 2009), and δ¹³C values are often greater for benthic–littoral primary producers and consumers than for planktonic–pelagic primary producers and consumers (France 1995; Vander Zanden et al. 2003). Therefore, δ¹³C may be used to infer where or on what group of species a consumer is feeding (Vander Zanden and Rasmussen 1999).

We used a soft-tissue biopsy punch (5 mm) to remove a muscle sample from the dorsal musculature from a subsample of Cutthroat Trout (n = 64), Mountain Whitefish (n = 39), sucker species (n = 45), Tui Chub (n = 17), Lake Trout (n = 93), Brown Trout (n = 31), Rainbow Trout (n = 26), and Kokanee (n = 9). The range of lengths of fishes used for stable isotope analysis was similar to the range of lengths of all fishes captured (Table 3). We sacrificed a subsample of Lahontan Redside (n = 13) and Speckled Dace (n = 13) and used a scalpel to remove a muscle sample from their dorsal musculature for stable isotope analysis. We sacrificed a subsample of signal crayfish (n = 13) and used a scalpel to remove abdominal muscle for stable isotope analysis. We pooled whole-body tissue samples from multiple individual zooplankton, mysids, and freshwater snails to create composite samples: zooplankton (n = 5), mysids (n = 10), and freshwater snails (n = 5). We froze tissue samples collected for stable isotope analysis and transported them to the laboratory for further preparation and analysis. We dried tissue samples to a constant mass at 60°C, ground them with a mortar and pestle, and placed 1.6 to 1.8 mg (balance sensitivity = 0.0001 mg) of the dried and ground sample in a tin capsule (Jardine et al. 2003). Prepared samples were analyzed for δ¹³C and δ¹⁵N (Table S4) at the Nevada Stable Isotope Lab (University of Nevada, Reno).

We used analysis of variance (ANOVA; PROC GLM, SAS version 9.4, SAS Institute Inc.) to compare isotopic values of fishes, crayfish, and mysids in Fallen Leaf Lake before Cutthroat Trout stocking (fishes sampled in 2009) to after Cutthroat Trout stocking (fishes sampled in 2010). We used this analysis to provide inference on changes in the food web in Fallen Leaf Lake following Cutthroat Trout stocking (objective 3) and these data may be used as a baseline for future studies in Fallen Leaf Lake. Isotopic incorporation into muscle tissue is relatively slow (Martínez del Rio et al. 2009). For example, laboratory studies have shown that isotopic changes in muscle tissue following a diet shift can take from 3 to 4 mo (e.g., Steelhead Trout O. mykiss; Church et al. 2009) to greater than a year (e.g., Broad Whitefish Coregonus nasus; Hesslein et al. 1993). Therefore, based on the timing of Cutthroat Trout stocking (July 2009; Table 1) and the timing of our sampling (Table 2) we assumed that isotopic changes of fishes in Fallen Leaf Lake associated with Cutthroat Trout stocking in 2009 would likely not be detectable until 2010. Although Cutthroat Trout had previously been stocked in Fallen Leaf Lake, the estimated biomass of Cutthroat Trout stocked during the most recent previous stocking events (i.e., 2006 and 2008; Table 1) were each over an order of magnitude less than that of Cutthroat Trout stocked in 2009. Therefore, we assumed that the effect of these stocking events on the isotopic characteristics of the Fallen Leaf Lake food web would be relatively small or nondetectable compared to that of effects associated with the 2009 stocking.

We performed ANOVA separately for each species and species by size group for Lake Trout (as above) and sucker species (see below) by stable isotope (i.e., δ¹³C and δ¹⁵N). Each ANOVA model had one qualitative factor (sample year) with two levels (2009 [corresponding to prestocking isotope value] and 2010 [corresponding to poststocking isotope value]); we included individual length (total length for fishes and tip of the rostrum to the end of the telson for crayfish) as a quantitative factor (Kuehl 1994) to account for differences in isotope values associated with individual size, and the response variable was either δ¹³C or δ¹⁵N. We mathematically normalized values for δ¹³C prior to analysis to account for differences in lipid concentrations (Post et al. 2007:equation 3; Table S4, Supplemental Material). We initially fit each model with an interaction term to test for interactions between sample year and individual length. We observed a significant interaction (α = 0.05) for sucker species for both δ¹³C and δ¹⁵N. None of the suckers that we sampled were between about 240 and 280 mm total length (Figure 2) and a preliminary evaluation of the interactions indicated that suckers less than about 240 mm differed in δ¹³C and δ¹⁵N between sample years, but that suckers greater than about 280 mm did not differ in δ¹³C or δ¹⁵N between sample years. Consequently, we separated sucker species into two groups (sucker species <250 mm [hereafter small sucker species] and sucker species ≥250 mm [hereafter large sucker species]) for further analyses.

We did not observe significant interactions between sample year and individual length for any of the other species or length group by species for either δ¹³C or δ¹⁵N; therefore, we refit ANOVA models without the interaction term. For each stable isotope, we used ANOVA models to obtain P values for the main effect of sample year for each species or length group for Lake Trout and sucker species. We assumed differences between sample years in δ¹³C, or δ¹⁵N, or both to be significant only if the null hypothesis of no difference in isotope values between sample years could be rejected following a sequentially rejective Bonferroni test (Holm 1979).

We plotted bivariate isotope data to provide a descriptive characterization of the Fallen Leaf Lake food web (objective 4). For isotope by species combinations that did not differ between sample years and that where not influenced by individual length, we calculated the arithmetic mean (±SE) of the isotope value. For isotope by species combinations that differed between sample year, or that were influenced by individual length, or both, we calculated the mean (±SE) isotope value using least square means (LSMEANS, SAS version 9.4, SAS Institute Inc.).

Lake Trout and Brown Trout were the largest fishes sampled using gill nets, followed by Rainbow Trout, Mountain Whitefish, sucker species, Tui Chub, Cutthroat Trout, and Kokanee (Figure 2). Lake Trout and Brown Trout varied in length from about 190 to 800 mm, Rainbow Trout varied in length from about 160 to 550 mm, and Kokanee varied in length from about 210 to 330 mm. The length of Mountain Whitefish, sucker species, and Tui Chub sampled using gill nets varied from about 80 to 420 mm. The mean ± SD length of Cutthroat Trout stocked in 2009 was 272 ± 30 mm and the mean length of Cutthroat Trout stocked in 2010 was 248 ± 26 mm. Additionally, two larger Cutthroat Trout (343 mm and 381 mm) were sampled in 2010. We assumed that these individuals had been stocked in 2009 and had overwintered in Fallen Leaf Lake based on their size relative to Cutthroat Trout stocked in 2010.

Overall, Lake Trout were the most abundant fish sampled using gill nets, followed by Kokanee, Mountain Whitefish, sucker species and Cutthroat Trout, Tui Chub, and Rainbow Trout and Brown Trout (Figure 3). Catch per unit effort of Lake Trout was three or more times greater than any other species sampled overall (Figure 3) and we detected them among all seasons and depth strata (Figure 4). Mountain Whitefish, sucker species, Tui Chub, and Cutthroat Trout were generally more abundant than Brown Trout, Rainbow Trout, and Kokanee by season and depth strata (Figure 4). Kokanee abundance was generally low among gill nets with the exception of two individual gill nets that sampled greater than 8 Kokanee/h; these gill nets were both set in putative Kokanee spawning habitat during the autumn.

Lake Trout and Brown Trout both preyed on fish including Cutthroat Trout, and Cutthroat Trout and Mountain Whitefish preyed on a variety of taxa and had relatively high diet overlap. Forty-two percent of recently stocked Cutthroat Trout stomachs that we examined were empty. Cutthroat trout consumed mysids (7.3%), aquatic insects (14.6%), terrestrial arthropods (32.5%), crayfish (24.2%), and fish (21.4%) in relatively similar proportions (Figure 5). Forty-two percent of Mountain Whitefish stomachs that we examined were empty. Aquatic insects (39.9%) were the most abundant dietary category that we observed in the Mountain Whitefish diet; however, Mountain Whitefish also preyed on mysids (21.0%), terrestrial arthropods (20.7%), and crayfish (18.4%). Thirty-seven percent of small Lake Trout had empty stomachs. Small Lake Trout primarily preyed on mysids (45.6%), but aquatic insects (20.8%), crayfish (18.3%), and fish (14.8%) were present in their diet. Terrestrial arthropods (0.5%) made up only a small proportion of the diet for small Lake Trout. Fishes that we could positively identify from the stomach contents of small Lake Trout included one Cutthroat Trout, one Kokanee, one Lahontan Redside, four sucker species, and two sculpin species Thirty-five percent of large Lake Trout had empty stomachs. Large Lake Trout primarily preyed on fish (67.8%) and also preyed on mysids (7.5%), aquatic insects (15.6%), and crayfish (9.1%) in relatively similar proportions. We could positively identify one Cutthroat Trout from the stomach contents of a large Lake Trout. Eighteen percent of the Brown Trout stomachs examined were empty. Brown Trout primarily preyed on fish (65.5%), and also preyed on terrestrial arthropods (12.3%) and crayfish (22.2%). We could positively identify three Cutthroat Trout, two sucker species, and one sculpin species from the stomach contents of Brown Trout. Rainbow Trout preyed primarily on crayfish (74.9%), but also preyed on aquatic insects (2.8%), terrestrial arthropods (16.7%), and fish (5.6%); no Rainbow Trout stomachs that we examined were empty. Sixty-seven percent of the Kokanee stomachs that we examined were empty and Kokanee preyed on aquatic insects (17.4%) and terrestrial arthropods (82.6%). All Tui Chub stomachs that we examined (n = 6) contained unidentifiable material. Diet overlap was greatest between Cutthroat Trout and Mountain Whitefish (C_xy = 0.61), intermediate between Cutthroat Trout and Brown Trout (C_xy = 0.56), Lake Trout <499 mm (C_xy = 0.56), and Lake Trout ≥499 mm (C_xy = 0.53), and least between Cutthroat Trout and Rainbow Trout (C_xy = 0.49) and Kokanee (C_xy = 0.47).

Stable isotope values varied by individual length for 11 isotope-by-species combinations, but the direction of the effect and the isotope varied considerably among species and size groups. Individual length was positively associated with δ¹³C for large Lake Trout (F_1,29 = 53.93, P = 0.0031). Individual length was negatively associated with δ¹³C for crayfish (F_1,10 = 40.92, P = 0.0034) and Cutthroat Trout (F_1,61 = 3.34, P < 0.0001). Individual length was positively associated with δ¹⁵N for crayfish (F_1,10 = 3.81, P = 0.0002), small Lake Trout (F_1,58 = 4.06, P = 0.0001), large Lake Trout (F_1,29 = 46.01, P < 0.0001), Mountain Whitefish (F_1,36 = 5.45, P = 0.0007), Rainbow Trout (F_1,23 = 2.22, P = 0.0025), and Speckled Dace (F_1,8 = 3.05, P = 0.0075). Individual length was negatively associated with δ¹⁵N for Cutthroat Trout (F_1,61 = 9.38, P < 0.0001) and Tui Chub (F_1,14 = 1.37, P = 0.0263). Stable isotope values did not differ between years for most species or species by size groups after controlling for familywise error rate with the following exceptions. Mysids differed significantly in δ¹³C (F_1,8 = 17.51, P < 0.0001), small sucker species differed significantly in δ¹³C (F_1,25 = 86.06, P = 0.0032), and large Lake Trout differed significantly in δ¹⁵N (F_1,29 = 30.18, P = 0.0006).

Mysids and zooplankton had relatively low δ¹³C and δ¹⁵N values, whereas freshwater snails and crayfish had relatively low δ¹⁵N values and relatively high δ¹³C values (Figure 6). Cutthroat trout δ¹⁵N was greater than that of any other species examined in Fallen Leaf Lake and Cutthroat Trout δ¹³C was also relatively high. With the exception of Cutthroat Trout, large Lake Trout had the highest δ¹⁵N values observed in Fallen Leaf Lake, and δ¹⁵N values of large Lake Trout were significantly greater in 2010 than in 2009 (see above). Small Lake Trout, Brown Trout, and Tui Chub had relatively high δ¹⁵N values. The majority of other fishes had intermediate δ¹⁵N values, but δ¹³C values varied considerably.

We documented use of a variety of prey resources by recently stocked Cutthroat Trout in Fallen Leaf Lake. Although Cutthroat Trout reintroduction efforts have been ongoing since 2002 in Fallen Leaf Lake, no studies have evaluated prey use by Cutthroat Trout following stocking. Results of this study show that captive-bred Cutthroat Trout are capable of locating and acquiring local prey resources, but that dietary overlap exists between Cutthroat Trout and other species in Fallen Leaf Lake. We also observed predation on Cutthroat Trout by Lake Trout. Allen et al. (2006) showed that Cutthroat Trout stocked into Fallen Leaf Lake were a substantial part of the diet of Lake Trout and Al-Chokhachy et al. (2009) suggested that the Lake Trout population in Fallen Leaf Lake has the capacity to consume large numbers of Cutthroat Trout. Although we observed relatively few Cutthroat Trout in the stomach contents of Lake Trout in the present study, piscivorous Lake Trout did exhibit a significant increase in δ¹⁵N and an increase in δ¹³C from 2009 to 2010. These changes occurred following Cutthroat Trout stocking and the direction of these changes are consistent with a diet shift to include a prey resource with δ¹³C and δ¹⁵N values similar to those exhibited by the Cutthroat Trout stocked in Fallen Leaf Lake during 2009 and 2010. Consequently, our findings generally support those of previous studies (e.g., Allen et al. 2006; Al-Chokhachy et al. 2009). Additionally, we observed predation on Cutthroat Trout by Brown Trout, a predatory interaction that had not been previously documented.

The presence of Lake Trout in high relative abundance compared to other fishes in Fallen Leaf Lake poses a challenge to reintroduction of Cutthroat Trout. Consequently, Lake Trout control efforts have been suggested by Al-Chokhachy et al. (2009) as a means of increasing the efficacy of Cutthroat Trout reintroduction. Decreasing the abundance of Lake Trout may reduce predation pressure on Cutthroat Trout, but may also have unanticipated ecosystem-level effects. For example, mysids were the dominant dietary item for small Lake Trout in this study; therefore, decreasing the abundance of Lake Trout could reduce predation pressure on mysids and result in an increase in mysid abundance, which could result in a decrease in native zooplankton abundance (Spencer et al. 1991; Vander Zanden et al. 2003) or other changes to the Fallen Leaf Lake food web. However, both Mountain Whitefish and recently stocked Cutthroat Trout preyed on mysids in this study, and Cutthroat Trout subspecies have been shown to prey on mysids in other lacustrine ecosystems (Clarke et al. 2005). Additionally, Lake Trout are the numerically dominant species of fish in Fallen Leaf Lake based on gill net surveys and they prey extensively on mysids, suggesting that the mysid population in Fallen Leaf Lake is capable of supporting a relatively large biomass of fish. Therefore, the presence of, or increase in abundance of mysids may benefit Cutthroat Trout reintroductions and other native fishes in Fallen Leaf Lake if they have similar abilities to prey on mysids as Lake Trout.

In addition to behavioral deficiencies associated with predator avoidance, captive-bred individuals may not recognize natural prey items or may be unable to forage efficiently (Snyder et al. 1996). Stomach contents of Cutthroat Trout stocked into Fallen Leaf Lake had previously not been quantified. Forty-two percent of stocked Cutthroat Trout that we sampled in this study had empty stomachs. This value is substantially higher than the mean percentage of empty stomachs observed for O. clarkii (26.4%) in a meta-analysis of 402 species of fish (Vinson and Angradi 2011). However, the percentage of Cutthroat Trout with empty stomachs in Fallen Leaf Lake was similar to some other fishes sampled (e.g., Mountain Whitefish and Lake Trout). Cutthroat Trout with nonempty stomachs exhibited the ability to locate and prey on a variety of wild prey resources including mysids, aquatic insects, terrestrial arthropods, crayfish, and fish. Cutthroat Trout had the greatest degree of diet overlap with Mountain Whitefish, which also preyed on a variety of taxa (Figure 5). Although relatively high diet overlap was observed between Cutthroat Trout and Mountain Whitefish in Fallen Leaf Lake, data are currently unavailable to determine whether competition for prey resources is negatively affecting Cutthroat Trout survival in Fallen Leaf Lake. For example, no data are available to evaluate prey limitation in Fallen Leaf Lake and competitive exclusion theory would require that prey resources are limited (Hardin 1960), and the generalized use of prey types exhibited by Cutthroat Trout and Mountain Whitefish may allow them to partition prey resources or make use of seasonally variable prey resources to minimize competition (Schoener 1971). Additionally, the present study provides stomach content data for a variety of fishes that have not been examined in Fallen Leaf Lake; however, small sample sizes associated with some species may limit the precision of this study to evaluate diet overlap within the Fallen Leaf Lake food web. We limited our stomach content analyses to fishes that died during sampling; however, future studies aimed at fine-scale analysis of diet overlap among fishes in Fallen Leaf Lake would benefit from larger sample sizes.

Isotopic characteristics of taxa within the Fallen Leaf Lake food web were generally consistent with stomach content data, general patterns associated with stable isotope ecology, and other food web studies. Mysids and zooplankton had relatively low δ¹⁵N and δ¹³C values, which is consistent with taxa occupying a low trophic level and pelagic habitats (France 1995; Vander Zanden and Rasmussen 1999; Vander Zanden et al. 1999, 2003; Martínez del Rio et al. 2009). Freshwater snails had low δ¹⁵N values and high δ¹³C values, which is consistent with taxa occupying a low trophic level in littoral habitats. With the exception of recently stocked Cutthroat Trout, Lake Trout and Brown Trout had the highest δ¹⁵N values suggesting that they are top-level predators in this system; this finding is consistent with stomach content analyses (see above). Tui chub had a relatively high δ¹⁵N value; this result was unexpected because some studies have shown that Tui Chub often feed on zooplankton (Marrin and Erman 1982; Galat and Vucinich 1983a, 1983b; Cooper 1985). The majority of other fishes had intermediate δ¹⁵N values, but varied substantially in δ¹³C. These data suggest that many fishes in Fallen Leaf Lake occupy a relatively similar trophic position, but use different prey items or use similar prey items in different proportions.

Mountain Whitefish and Kokanee had relatively low δ¹³C values suggesting a planktonic or pelagic dietary carbon source; similar trends have been observed for these species in other oligotrophic lakes (Vander Zanden et al. 2003; Meeuwig et al. 2011). Terrestrial arthropods were present in the diets of both Mountain Whitefish and Kokanee, which suggests that the isotopic composition of one or both of these species may be a result of terrestrial sources. The isotopic composition of terrestrial arthropods was not evaluated in this study; however, δ¹³C values of terrestrial arthropods may be intermediate between planktonic–pelagic and benthic–littoral sources (France 1995). Sucker species varied substantially in δ¹³C. Large suckers had relatively low δ¹³C values, which were similar to those of Mountain Whitefish and Kokanee; whereas small suckers had higher, but variable, δ¹³C values. A study in nearby Lake Tahoe (Vander Zanden et al. 2003) showed that suckers had high δ¹³C values relative to other fishes examined; however, the size range of suckers used in that study is unknown. It is possible that suckers in Fallen Leaf Lake exhibit an ontogenetic diet switch, which could explain the differences in δ¹³C values between large and small suckers. Furthermore, some other studies have shown adult suckers to have low δ¹³C values relative to other fishes (Keough et al. 1996).

We sampled Speckled Dace and Lahontan Redside exclusively with minnow traps in littoral areas of Fallen Leaf Lake. Speckled Dace had relatively high δ¹³C values indicative of a benthic–littoral dietary carbon source; similar trends have been observed in Lake Tahoe (Vander Zanden et al. 2003) and oligotrophic lakes in Montana (Meeuwig et al. 2011). Conversely, δ¹³C values for Lahontan Redside were relatively low. Lahontan Redside exhibited a relative low δ¹³C value compared to other fishes in a stable isotope analysis of Lake Tahoe fishes (Vander Zanden et al. 2003), and Redside Shiner Richardsonius balteatus, a congeneric species, has been shown by Meeuwig et al. (2011) to vary in δ¹³C among lakes. Additionally, Lahontan Redside are opportunistic predators and may derive carbon from a variety of aquatic and terrestrial sources.

Zooplankton was not identified in the stomach contents of any fish evaluated in this study. Kokanee are generally considered to prey extensively on pelagic zooplankton (Scott and Crossman 1973); however, the presence of zooplankton in the diets of Kokanee is seasonally variable in some lakes. For example, Kokanee diets were dominated by zooplankton species during the spring and autumn and chironomid pupae during the summer in Nicola Lake, British Columbia (Northcote and Lorz 1966). Additionally, in the present study we primarily sampled Kokanee stomachs during the late summer (n = 3) and autumn (n = 15) when sexually mature Kokanee may exhibit hypophagia prior to spawning. In the present study, 67% of the Kokanee stomachs that we examined were empty. Cutthroat Trout may also prey on zooplankton, but this may be variable among size classes, seasons, and spatial location. For example, Nowak et al. (2004) examined the diets of four size classes of Cutthroat Trout, during four seasons, and in two lake zones in Lake Washington, Washington, and zooplankton made up only 0–10% of the Cutthroat Trout diets among 75% of the size class by season by lake-zone combinations. Therefore, the lack of zooplankton in the diets of fishes examined in this study may be a function of the species evaluated or the timing of sampling. Alternatively, it is possible that zooplankton were more readily decomposed in the stomach samples than other prey items. However, Bosmina sp. and Daphnia sp. are present in Fallen Leaf Lake (Allen et al. 2006) and these species are relatively resistant to digestion (Sutela and Huusko 2000), and we readily identified other species of invertebrates (e.g., mysids, chironomid pupae). Finally, it is possible that zooplankton abundance is currently depressed in Fallen Leaf Lake due to competitive or predatory interaction with mysids. For example, the introduction of mysids has been shown to alter the zooplankton assemblage in other lakes (Rieman and Falter 1981; Spencer et al. 1991, 1999; Vander Zanden et al. 2003); therefore, future research should evaluate interactions between mysids and zooplankton and how those interactions may influence other taxa.

We observed relatively few differences in isotopic values between 2009 (pre–Cutthroat Trout stocking) and 2010 (post–Cutthroat Trout stocking) by species or length group. We observed an increase in δ¹³C for mysids and a decrease in δ¹³C for small sucker species. The isotopic turnover rate of small, rapidly growing species may be quick; consequently these species may exhibit significant temporal variability in isotope values (e.g., Post 2002). Mysids can exhibit a variety of feeding strategies, but it is unknown whether they are omnivorous, herbivorous, or carnivorous in Fallen Leaf Lake. It is possible that the change in mysid δ¹³C from 2009 to 2010 is a result of temporal variability in δ¹³C of either primary producers or consumers. However, the observed change in mysid δ¹³C does not appear to have been propagated to larger, slower-growing organisms in the Fallen Leaf Lake food web. For example, little variability in δ¹³C was observed for small Lake Trout even though 45.6% of their diet was composed of mysids (Figure 6); additionally, Mountain Whitefish varied little in δ¹³C despite preying on mysids (21.0% of their diet). A significant decrease in δ¹³C was observed from 2009 to 2010 for small suckers; this is a change that we did not anticipate. Small suckers sampled in 2009 and 2010 had generally similar length distributions and were sampled using similar techniques; therefore, it is unlikely that the observed pattern is an artifact of sampling. Small suckers in Fallen Leaf Lake may feed heavily on algae, small aquatic insects, or both, and these taxa may exhibit significant temporal variability in stable isotope values (e.g., Post 2002). Consequently, future studies of the Fallen Leaf Lake food web should consider quantifying the stomach contents and temporal variability in the stomach contents of sucker species

A significant increase in δ¹⁵N from 2009 to 2010 (2.04‰ increase) was observed for large Lake Trout; large Lake Trout also increased in δ¹³C from 2009 to 2010 (1.67‰ increase). These changes may be associated with direct predation on recently stocked Cutthroat Trout. The size structure of Lake Trout in Fallen Leaf Lake is currently dominated by individuals less than about 500 mm. However, we sampled Lake Trout in excess of 500 mm and observed that Lake Trout in Fallen Leaf Lake switched to a primarily piscivorous diet at about 499 mm. Studies have shown that Lake Trout can consume relatively large prey species; for example, Lake Trout in Yellowstone Lake, Wyoming, consumed Yellowstone Cutthroat Trout that varied from 11 to 57% of their own length (Ruzycki et al. 2003). Therefore, piscivorous Lake Trout in Fallen Leaf Lake are capable of preying on a large proportion of the Cutthroat Trout, Mountain Whitefish, sucker species, Tui Chub, and Kokanee that we sampled during this study (Figure 2).

The isotopic composition of muscle samples from Cutthroat Trout stocked into Fallen Leaf Lake was distinct from other species examined and was likely a result of hatchery diet as the isotopic composition of a consumer generally reflects that of its diet and muscle tissue has a relatively long turnover rate (Hesslein et al. 1993; Phillips and Eldridge 2006; Church et al. 2009; Martínez del Rio et al. 2009). The high δ¹³C and δ¹⁵N values of Cutthroat Trout may have provided an inadvertent isotopic tracer for evaluating trophic shifts. The direction of the change in δ¹³C and δ¹⁵N of large Lake Trout is consistent with a diet shift between 2009 and 2010 to include more isotopically enriched prey, such as hatchery-reared Cutthroat Trout (Figure 6). Additionally, two Cutthroat Trout that we sampled in 2010 were assumed to be individuals that had been stocked in 2009 and that had overwintered in Fallen Leaf Lake based on their size relative to Cutthroat Trout stocked in 2010. Isotopic values for these two Cutthroat Trout suggest that from 2009 to 2010 this cohort of Cutthroat Trout decreased in δ¹³C by about 1.8‰ and decreased in δ¹⁵N by about 2.3‰ (Figure 6). Although these two individual Cutthroat Trout still had elevated δ¹³C and δ¹⁵N values, the directional change in isotopic values suggests they are becoming more isotopically similar to other species in the Fallen Leaf Lake food web and that isotopic changes following a diet shift are relatively slow for this population of stocked Cutthroat Trout.

Overall, results from this study show that recently stocked Cutthroat Trout were capable of preying on a variety of taxa; however, we observed predation on Cutthroat Trout by Lake Trout and Brown Trout. This supports previous research that showed high predation rates on stocked Cutthroat Trout by Lake Trout in Fallen Leaf Lake (Allen et al. 2006) and provides direct evidence for predation by another top-level predator. Therefore, the combined effects of Lake Trout and Brown Trout predation on Cutthroat Trout should be considered in the context of predator management. A previous study suggested that that initiating active predator (i.e., Lake Trout) removal efforts may facilitate Cutthroat Trout reintroduction efforts in Fallen Leaf Lake (Al-Chokhachy et al. 2009). We concur, but caution that without sufficient exploitation levels a predator reduction effort and its potential benefit to Cutthroat Trout reintroduction may not be realized. For example, research indicates that annual mortality rates of about 50% per year are needed to achieve a population level decline in Lake Trout abundance (Healey 1978; Hansen et al. 2008; Cox 2010; Cox et al. 2013; Syslo et al. 2011), and such exploitation levels can be costly (Syslo et al. 2013). Additionally, we believe that Cutthroat Trout reintroduction efforts should have quantifiable objectives and follow an active adaptive management framework (e.g., see Walters and Hilborn 1978). Within this framework, results from this study and others on Fallen Leaf Lake (e.g., Allen et al. 2006; Al-Chokhachy et al. 2009) may be used to develop hypotheses and quantify changes associated with implementing management activities. Finally, stable isotope tracer experiments have been used to evaluate food web characteristics, energy flow pathways, and dispersal of a variety of organisms (Hall 1995; Mulholland et al. 2000; Raikow and Hamilton 2001; Macneale et al. 2005). However, to our knowledge no studies have experimentally manipulated stable isotope values of hatchery-reared fish with the express purpose of evaluating predation on stocked fish. In the present study large Lake Trout exhibited a measurable, directional change in isotopic composition following Cutthroat Trout stocking consistent with feeding on isotopically enriched prey, such as the stocked Cutthroat Trout; this shift may not have been observed if stocked Cutthroat Trout were isotopically similar to other prey items in Fallen Leaf Lake. Therefore, we believe that the use of stable isotope tracers as a tool for evaluating stocking efforts and the fate of Cutthroat Trout in Fallen Leaf Lake and other systems warrants further investigation.

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.

Table S1. Gill net set and pull data for gill nets set during 2009 and 2010 to evaluate characteristics of the Fallen Leaf Lake, California, food web. We set gill nets benthic and perpendicular to the lake shoreline. We set gill nets during five distinct sample events in 2009 and 2010 and stratified gill net sets by depth (the average depth of shallow gill nets was 9.9 m and the average depth of deep gill nets was 30.2 m).

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S1 (73 KB PDF).

Table S2. Fish total length (mm) by year, gill net ID, and species for individual fish sampled during 2009 and 2010 to evaluate characteristics of the Fallen Leaf Lake, California, food web. A dot (“.”) denotes no length data. Species that we sampled included Lahontan Cutthroat Trout Oncorhynchus clarkii henshawi, Mountain Whitefish Prosopium williamsoni, sucker species, Tui Chub Gila bicolor, Lake Trout Salvelinus namaycush, Brown Trout Salmo trutta, Rainbow Trout Oncorhynchus mykiss, and Kokanee Oncorhynchus nerka.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S2 (83 KB PDF).

Table S3. Species, total length (mm), and dry mass (mg) of six dietary categories for individual fish that we sampled for stomach content analysis from Fallen Leaf Lake, California, during 2009 and 2010. Dietary categories included 1) zooplankton, 2) mysid shrimp Mysis diluviana, 3) aquatic insects, 4) terrestrial arthropods, 5) signal crayfish Pacificus leniusculus, and 6) fish. Fish species that we sampled for stomach content analysis included Brown Trout Salmo trutta, Kokanee Oncorhynchus nerka, Lahontan Cutthroat Trout Oncorhynchus clarkii henshawi, Lake Trout Salvelinus namaycush, Mountain Whitefish Prosopium williamsoni, Rainbow Trout Oncorhynchus mykiss, and Tui Chub Gila bicolor.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S3 (101 KB PDF).

Table S4. Sample year, species identification, total length, and stable carbon (δ¹³C) and nitrogen (δ¹⁵N) isotope values for individuals sampled in Fallen Leaf Lake, California, during 2009 and 2010. Stable carbon isotope data were normalized following equation 3 in Post et al. (2007). Species sampled for stable isotope analysis included Brown Trout Salmo trutta, freshwater snails, Kokanee Oncorhynchus nerka, Lahontan Cutthroat Trout Oncorhynchus clarkii henshawi, Lahontan Redside Richardsonius egregius, Lake Trout Salvelinus namaycush, Mountain Whitefish Prosopium williamsoni, mysid shrimp Mysis diluviana, Rainbow Trout Oncorhynchus mykiss, signal crayfish Pacificus leniusculus, Speckled Dace Rhinichthys osculus, sucker species, Tui Chub Gila bicolor, and zooplankton.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S4 (103 KB PDF).

Reference S1. Allen BC, Chandra S, Atwell L, Vander Zanden MJ, Reuter JE. 2006. Evaluation of the re-introduction of native Lahontan cutthroat trout, Onchorhynchus clarki henshawi, in Fallen Leaf Lake, California, and development of management strategies for recovery. University of California, Davis, Tahoe Environmental Research Center to U.S. Fish and Wildlife Service, Nevada Office, Reno, Nevada.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S5 (1768 KB PDF).

Reference S2. Coffin PD, Cowan WF. 1995. Lahontan cutthroat trout (Oncorhynchus clarki henshawi) recovery plan. Portland, Oregon: U.S. Fish and Wildlife Service.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S6; also available at https://www.fws.gov/oregonfwo/documents/RecoveryPlans/Lahontan_Cutthroat_Trout_RP.pdf . (9766 KB PDF).

Reference S3. Weagley S. 2010. REGION 8: Lahontan cutthroat trout stocked into Fallen Leaf Lake. U.S. Fish and Wildlife Service Field Notes.

Found at DOI: http://dx.doi.org/10.3996/092016-JFWM-073.S7; also available at https://www.fws.gov/FieldNotes/regmap.cfm?arskey=28191. (858 KB PDF).

Funding for this project was provided by the USFWS Lahontan National Fish Hatchery Complex (USFWS F12AC00129; UNR 1320-114-23BJ). Logistical support was provided by D. Bloomquist, S. Byers, and L. Heki. Field assistance was provided by J. Smith, L. Tennant, and D. Zangari. Laboratory facilities were provided by W. Cross and C. Guy. Assistance with stable isotope analysis was provided by S. Poulson. L. Tennant and the Journal of Fish and Wildlife Management review team (the Associate Editor, and two anonymous reviewers) provided comments on an earlier draft of this manuscript.

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