Aquatic Amphibian and Reptile Surveys. Fall 2006 Rotenone Treatment of Diamond Lake. Final 2008 Report

Aquatic Amphibian and Reptile Surveys during the second year after the Fall 26 Rotenone Treatment of Diamond Lake Final 28 Report Marc P. Hayes and Christopher J. Rombough submitted to Oregon Department of Fish and Wildlife Southwest Region 492 North Umpqua Highway Roseburg, Oregon 9747 3 December 28

Aquatic Amphibian and Reptile Surveys during the first year after the Fall 26 Rotenone Treatment of Diamond Lake Final 28 Report Marc P. Hayes and Christopher J. Rombough Executive Summary: During 28, we surveyed Diamond Lake, the areas of its inflow and outflow creeks, and two small nearby fishless lakes (Horse and Teal Lakes), which served as reference (control) sites, for aquatic amphibian and reptiles. The effort involved visual encounter surveys of all aforementioned areas, snorkel surveys along several -m segments of Diamond Lake shoreline, and dip net-assisted surveys in areas with emergent vegetation. As in 27, the ice and snowbound western margin of Diamond Lake was not surveyable during the early (post-thaw) survey, but we replicated all prerotenone surveys conducted in these same areas in 996. Notably, the post-thaw breeding surveys were delayed until June (conducted in May 27) because of the extensive levels of snow and delayed break-up of ice on Diamond Lake in 28. Collectively, we recorded all six aquatic amphibians (Northwestern salamander [Ambystoma gracile], Long-toed salamander [Ambystoma macrodactylum], Western toad [Anaxryus (formerly Bufo) boreas], Pacific chorus frog [Pseudacris (formerly Hyla) regilla], Cascade frog [Rana cascadae], and Roughskin newt [Taricha granulosa]) and the one aquatic reptile (Common garter snake [Thamnophis sirtalis]) recorded during both 996 and 27. Though we found most species in the same places and we recorded evidence of reproduction in many of the same areas as during surveys in 996 and 27, only two of the five differences that we recorded between the 996 and 27 surveys remained the same for surveys conducted in 28. These were: ) We failed to detect Cascade frog life stages in Diamond Lake proper; and 2) We found evidence of Western toad reproduction in Diamond Lake. The three remaining differences between 996 and 27 surveys did not remain the same for 28 surveys. In 28: ) We detected a some Northwestern salamanders in the portions of Diamond Lake accessible to fish and in fish stomachs (Northwestern salamanders were neither detected in the portions of Diamond Lake accessible to fish nor in fish stomachs in 27); 2) We detected moderate numbers of Common garter snakes or evidence of them [shed skins] along the northeast margin of Diamond Lake (Almost no common garter snakes observed in Diamond Lake); 3) Recruitment of Cascade frogs from Horse and Teal Lakes seemed extensive; in fact, recruitment of all three anuran species (Cascade frogs, Pacific chorus frog, and Western toad) appeared extensive in 28 (we found almost no evidence of recruitment of Cascade frogs from Horse and Teal Lakes in 27).

Redetection of Northwestern salamanders in the fish-accessible portions of Diamond Lake and in fish stomachs in 28 when they were not detected in 27 may reflect some level of recovery for this species in Diamond Lake proper. Our data cannot distinguish whether Northwestern salamanders were really extirpated from Diamond Lake during the rotenone treatment because it is unknown whether the snorkel surveys we used to detect them are capable of detecting them at very low densities, a condition that cannot be excluded for 27. Additionally, even though Northwestern salamanders in Diamond Lake are assumed to be neotenic (adults look like large larvae and are entirely aquatic), our data cannot exclude the possibility that this population includes non-neotenic animals capable of overland movement from nearby sources (e.g., the pond near the Diamond Lake outflow). Thus, we can conclude a rotenone effect on Northwestern salamanders, but our data cannot effectively gauge the magnitude of that effect for Diamond Lake because of potential low density detectability issues (i.e., extirpation versus reduction to a low population level in Diamond Lake proper). Increased evidence of garter snakes presence in 28 may reflect increased availability of the amphibian food resource (especially the three species of anurans) in 28. Female garter snakes mature in more than one year, so this response is likely not demographic. More likely it represents movement or immigrant from nearby areas, as garter snakes are capable of multi-kilometer movements seasonally. Lack of Cascade frog reproduction in Diamond Lake likely reflects loss of habitat due to succession first observed in 27 since shallow waters on the spit of the Short Creek peninsula extending into Diamond Lake, which was used by this species for oviposition in 996, no longer exist. Moreover, our failure to detect post-metamorphic Cascade frogs in Diamond Lake proper given their longevity (-5 years) complicates interpretation. We cannot exclude a rotenone-treatment effect contributing to our inability to detect Cascade frogs in Diamond Lake, but if there was such an effect exists, it is confounded with a fish effect since reference sites had extensive production of Cascade frogs in Diamond Lake in 28. This revises the suggestion we made in the 27 that this species may be in decline regionally, and the best interpretation of available data to date is that conditions in most of Diamond Lake proper may be unsuited to Cascade frog survival. However, the latter fails to explain the high production of both Pacific chorus frog and Western toad in the marsh along the northwest margin of Diamond Lake in 27. Lack of Cascade frogs at this location may reflect a localized extirpation event from which recolonization has not yet taken place, one of several classic patterns of local inter-year variability that requires more years of data to interpret. Results of the 28 surveys imply that recovery toward something approximating a prerotenone treatment level of amphibian distribution is occurring. Swab sampling for the amphibian chytrid fungus failed to reveal its presence in the Diamond Lake system in 28. INTRODUCTION Diamond Lake, a large (,87 ha), shallow, oblong-shaped lake of glacial origin at,58 m (5,84 ft) elevation, is nestled in the headwater region of the North Umpqua

River basin on the west slope of the Oregon Cascade Mountains. Historically harboring among the most productive hatchery rainbow trout (Oncorhynchus mykiss) fisheries in Oregon, harvest (angling) and other biological data revealed a decline in the quality and quantity of this fishery that began to attract attention in the 98s (Loomis 995, ODFW 996). Tui chub (Gila bicolor), either accidentally or intentionally introduced to Diamond Lake in the 94s and removed via rotenone treatment during the 95s, were again detected in Diamond Lake in 992, and underwent an eruptive population expansion soon after their detection. Trout decline and chub expansion did not appear directly coupled, but responses of other systems containing these two fishes under similar conditions indicated that chub expansion likely threatened the trout fishery (ODFW 996). Hence, renewed rotenone application was viewed as a potentially important remediation effort, and following lengthy negotiation and agency review, application of rotenone was approved and occurred in fall 26. Using rotenone to scrub this system requires NEPA evaluation 2 of non-target fauna (ODFW 996). Baseline data was collected for pre-treatment evaluation of the aquatic amphibian and reptile fauna in 996 (Hayes 997). This report provides preliminary evaluation of the aquatic amphibian and reptile fauna based on surveys conducted during the second post-treatment year (28), an initial step in completing the NEPA evaluation; and contrast the results with the first post-treatment year surveys (27) and the original pre-treatment surveys (996). METHODS We surveyed aquatic habitats diurnally (between 9: and 8: hours) using visualencounter surveys (VES) designed to detect amphibians (Crump and Scott 994) within the intervals 4-5 June and 28-3 August 28. The early (June) survey dates followed break-up of ice on Diamond Lake, and were intended to assess amphibian reproduction; the late (August) survey dates assessed amphibian recruitment. The early survey in 28 was just over a month later than that in 27 (-2 May 27) because heavy snow accumulation from winter (27-28) and a substantially delayed break-up of ice. The late survey in 28 differed only by a few days from that in 27 (26-27 August 27). The 28 surveyed were conducted during the same diel interval as surveys conducted in both 27 and 996. We use a modified form of the basic survey for stillwater habitats (Thoms et al. 997) that focused on shallow-water habitats and aquatic edges because detection of most life stages of all aquatic amphibian and reptile species likely to be present is most likely to occur in such habitat. We surveyed: ) the edges of Diamond Lake; 2) the interface areas between Diamond Lake and its inflow creeks (i.e., Silent, Short, Camp, Porcupine, Rabbit, Spruce and Two Bear Creeks as well as three small unnamed creeks on its southeast side); this includes the marshy areas along Silent Creek and along the south margin of Diamond Lake. 3) the interface areas between Diamond Lake and its outflow creek (i.e., Lake Creek); this includes the closed channel, now comprising a pond, that was used to pump Loomis (995) gives the size as 2,932 ac (about 4.6 mi 2 ), the maximum depth as 5.8 m (52 ft; most (>67 percent) of the lake is 7.6 m (25 ft) deep, and the elevation as,58 m (5,83 ft). The elevation represents the lake level at full pool. 2 National Environmental Protection Act guidelines.

down Diamond Lake prior to the rotenone treatment, and located next to the Lake Creek outflow. 4) Horse and Teal Lakes, and a small accessory pond and drainage leading to Horse Lake; Horse and Teal Lakes are fishless, so they represent important reference (or control) lakes to distinguish potential changes in fish-occupied Diamond Lake. During VES surveys, two individuals surveyed aquatic habitats with a slow walk of aquatic edges. One person surveyed the shallow aquatic habitat (< m in depth) between 2.5 m to 7.5 m from the edge (assuming shallow aquatic habitat exists to distance 7.5 m from the edge), and a second person surveyed 2.5 m in both directions (aquatic and landward from the edge). The original survey protocol (Hayes 996) and survey effort (997) also used a third individual surveying 2.5 m to 7.5 m in the landward direction from the edge, but the 996 effort revealed that surveying from this position did not add value to the survey (i.e., number of aquatic amphibians and reptiles found in this survey band were zero); hence, we eliminated this survey position. The 996 baseline surveys also revealed that VES surveys were insufficient for detection of all species in Diamond Lake proper due to the concealment behavior of Northwestern salamander in the presence of fish (Hayes 997), so we also conducted snorkel surveys along -meter segments of Diamond Lake shoreline to a depth of.5 m in habitat containing benthic cover (mostly rocks). We did not snorkel survey the Horse and Teal Lake reference sites because of their shallow depth and/or simplified habitat conditions during the summer survey interval made it unnecessary. We recorded numbers or an estimate of numbers (if not easily counted) of each life stage observed for all amphibians and reptiles encountered. We examined debris, vegetation, and substrates for potentially attached egg masses. We focused on detecting amphibian egg masses or packets during the early survey as all amphibians present reproduce following at spring thaw, and egg masses of all amphibian species present except perhaps those of Northwestern salamanders would have long hatched and disintegrated by the August survey interval. We used a kick net to help sample larval amphibians in softbottomed substrates, and areas of submergent and low emergent vegetation. We also used a kick net to sample stillwater habitat where water was turbid. Eggs or larval stages too numerous to count were estimated by sub-sampling randomized samples of masses or egg packets. Whenever possible, post-metamorphic stages of aquatic amphibians and reptiles found were captured, measured (as snout-vent length [SVL]), sexed, weighed, and released at the point of capture. Where animals were numerous, we measured the first 3 of each life stage of each species and subsequently, simply estimated their size (to the nearest 5 millimeters) and regularly measured an additional individual to verify that our estimates were accurate. Common garter snakes (Thamnophis sirtalis), dependent on amphibians for food (Gregory 978, Kephart and Arnold 982) and recorded during the 996 surveys (Hayes 997), were palped for identifiable prey. If identifiable amphibian prey were palped from snakes, where possible, we measured their size and determined their gender. We also estimated the sizes of aquatic amphibians and reptiles not captured. To better identify the importance of aquatic amphibian and reptiles in this system, we recorded birds and fishes encountered that might prey on amphibian or reptile life stages. We also recorded the presence and relative abundance of aquatic macroinvertebrates indicative of conditions especially favorable for amphibians, for example, the bright red

copepod (Hesperodiaptomus kenai) and any species of fairy shrimp (Order Anostraca), frequent in amphibian-abundant systems lacking fish (see Hayes 995), were recorded. We also recorded selected additional data basic to habitat assessment for any aquatic amphibian or reptile at each site sampled, these include: ) water temperature, 2) an ocular estimate of the percentage of emergent, floating, and submergent vegetation present at intervals along aquatic margins that includes an area ca. 5 m from the evaluated edge, 3) character of the substrate as an ocular estimate of the percentage of the standard substrate size categories (i.e., mud/slit, sand, gravel, cobble, boulder, bedrock) and organic fraction (i.e., leaves and fine wood debris); 4) an estimate of the amount of downed large woody debris (minimum diameter cm) as submerged cover and as cover on the upland edge of the aquatic habitat; and 5) description of the upland edge vegetation cover types with an estimate of the amount of such cover. Holly Truemper (Oregon Department of Fish and Wildlife; ODFW) and her field personnel of the provided data from 93 stomachs of rainbow trout taken from Diamond Lake in the interval May-September 28. All fishes in these samples were taken by angling. Prior to removing their gastrointestinal (GI) tracts, size data was obtained during creel surveys from each fish, which was already dead, in inches to the nearest.25 inch; these data were then converted to lengths in millimeters). The mass (to the nearest. lb then converted to grams) was also obtained for most of these fish. Their GI tract contents were then removed and preserved in formalin (as diluted Paracide F). Following an interval ranging from -3 months, preserved GI tracts were gently rinsed in water using a fine screen and preserved in a 7% ethanol for storage in subsequent examination. For this report, we report only amphibian remains of any kind by examination of these tracts; details of GI tract contents will be provided at a later date. We swabbed 32 amphibians taken from various aquatic habitats from Diamond Lake using wooden stick-type cotton swabs using a sterile procedure where a separate set of latex gloves are used to obtain each individual sample. Swabs were preserved in 95% molecular analysis-grade alcohol prior to shipment to the laboratory for analysis. Analysis involved polymerase chain reaction amplification of pooled samples for detection of DNA of the amphibian chytrid fungus, Batrachochytrium dendrobatidis. Samples from eight amphibians were included in each of four pooled samples. We used descriptive statistics to describe amphibian population and habitat variables obtained for the sampling units with aquatic sites and summarized for each site (e.g., Diamond Lake, Teal Lake). We present descriptive data either as tally information by species or life stage or provide the mean ( x ), standard deviation (s), and the range of analyzed variables. Analyses follow standard statistical procedures (Zar 999). Where appropriate, we used association statistics to compare life-stage specific population data, and made between-site comparisons as needed. For comparative analyses, we used nonparametric statistics because the distributions of the variable were skewed or multi-modal and variances were not homoscedastic. Critical rejection level of α was.5. RESULTS Similar to the surveys conducted in 27, the dates we selected for the June and August 28 surveys were conducted under near ideal conditions. For both surveys, conditions were cloudless or nearly so, and had low levels of wind (Beaufort scale score 2). The

June interval had daytime high air temperatures of 8.3-2.7 C (65-7 F) and nonfreezing nights (low air temperatures 3.3-6. C [38-43 F]); the August interval had daytime highs of 25.6-3.6 C (78-87 F) and also had non-freezing nights (low air temperatures 6.7-8.9 C [44-48 F]). We recorded five aquatic amphibians (Northwestern salamander [Ambystoma gracile], Long-toed salamander [Ambystoma macrodactylum], Western toad [Anaxryus (formerly Bufo) boreas], Pacific chorus frog [Pseudacris (formerly Hyla) regilla], and Cascade frog [Rana cascadae]) and one aquatic reptile, the Common garter snake during 28 surveys (TABLE ). Though not typically aquatic, we also recorded northern alligator lizards (Elgaria coerulea) along the edge of Diamond Lake. Except for the failure to record Roughskin Newt (Taricha granulosa), this species assemblage is identical to that which we recorded in 996 and 27. Because roughskin newt was recorded in extremely low numbers in both 997 and 27, we cannot effectively evaluate whether failure to record it in 28 represents a detectability problem or a local extirpation. These alternatives are not distinguishable without additional survey effort. We recorded a higher species richness (by one species) at most survey areas with the exception of the Outflow Pond/Lake Creek outlet in 28 when compared to 27 (TABLE ). Species richness in Outflow Pond declined by one in 28 (TABLE ). As Outflow Pond is such a small site, this pattern may reflect a combination of its opportunistic use for reproduction and/or the opportunistic appearance of different species over our relatively short-interval surveys. For example, Western toads observed at the Outflow Pond were always a few juveniles that we suspect were the result of reproduction at the marsh along the northwest margin of Diamond Lake that is located roughly.4 kilometers distant. The greater species richness in Diamond Lake between 27 and 28 was the result of recording Northwestern salamander there as we had in 996. Failure to record Northwestern salamanders in Diamond Lake in 27 may also represent a detectability problem. Though our snorkel surveys covered a fair amount of area, some likelihood exists that they would not detect Northwestern salamanders if densities of the latter were extremely low. Northwestern salamanders also not being recorded from fish stomachs in 27, but being found in fish stomachs in 28 suggests a consistent pattern. However, failure to record the species as a prey item would also be expected to occur if prey densities were very low and predator sampling (in this case rainbow trout) was not extremely extensive. Hence, failure to record Northwestern salamanders in 27 cannot distinguish between densities so low that our methods could not detect them and actual extirpation of the species from Diamond Lake. We have little doubt that the failure to detect Northwestern salamanders in 27 reflects the rotenone treatment, but whether Northwestern salamanders were briefly extirpated following that treatment is a question that cannot be answered. Failure to record Common garter snakes in all units except for Diamond Lake was an important contributor to the reduction in species richness in 27 and was one contributor to the local increases in species richness in 28. Our re-recording Western toads along Silent Creek and Long-toed salamander in Teal Lake also contributed to local increases in species richness during the 28 surveys. In sum, the species richness pattern observed in 28 probably reflects some combination of recovery from the rotenone treatment in Diamond Lake coupled with the positive effects of a high precipitation year.

TABLE. Aquatic amphibian and reptile species recorded in surveyed units in and around Diamond Lake, 996, 27, and 28. Species Diamond Lake Outflow Pond/ Lake Creek 2 Silent Creek 3 996 27 28 996 27 28 996 27 28 Ambystoma gracile + + + + + Ambystoma macrodactylum Anaxryus (=Bufo) boreas + + + + + + + Pseudacris (= Hyla) regilla + + + + + + + + Rana cascadae + + + + + Taricha granulosa + + Thamnophis sirtalis + + + + + Species Richness 6 4 5 3 3 2 4 2 3 Diamond Lake refers to Diamond Lake proper below the mouth of its inflow creeks and not extending into its outflow creek (Lake Creek). 2 Outflow Pond/Lake Creek refers to the pond at the north end of Diamond Lake created by the closure of the outflow channel that was open during the pump down of Diamond Lake prior to fall 26 rotenone treatment, and the portion of Lake Creek between its Diamond Lake outflow and the NFC 4795 culvert at Lake Creek. 3 Silent Creek refers to the areas between the mouth of Silent Creek at Diamond Lake and the bridge crossing of Silent Creek at NFC 4795 about kilometer upstream from Diamond Lake. Silent Creek also includes a few snowmelt ponds on either side of the creek and a marshy expanse on its east side at the south end of Diamond Lake.

TABLE (CONTINUED). Aquatic amphibian and reptile species in surveyed units in and around Diamond Lake, 996, 27, and 28. Species Horse Lake Teal Lake 2 996 27 28 996 27 28 Ambystoma gracile Ambystoma macrodactylum + + + + + Anaxryus (=Bufo) boreas + + + Pseudacris (= Hyla) regilla + + + + + + Rana cascadae + + + + + + Taricha granulosa Thamnophis sirtalis + + + Species Richness 5 4 5 4 2 3 Data for Horse Lake includes a small accessory pond (6 m 35 m) about 3 meters south of and 4 meters higher in elevation than Horse Lake proper, and was connected to Horse Lake during the early survey in all three years by small shallow channel that was dry during the August survey in all three years. 2 Teal Lake was an unambiguously separate aquatic unit during all years surveyed.

When our data were standardized as animals observed per unit effort, we recorded more individuals for the equivalent effort for most species and most units in 28, though there were scattered exceptions (TABLE 2). However, different units revealed substantially different patterns, as indicated in the brief discussion of each that follows. Diamond Lake: We found most amphibians in one area of Diamond Lake proper, the marsh located along the northwest lake margin, in all survey years. This marsh, centered roughly kilometer north of the entry road to Theilsen View Campground, is about 7 meters long and 5 to meters wide. Much of the marsh is sedge (Carex)-dominated, but its lakeside fringe has a significant development of cattails (Typha), and a trough at least 2 meters in length exists on its landward side that is now isolated from the lake by a combination of somewhat higher ground grown to mostly lodgepole pines (Pinus contorta), hardhack (Spirea douglasii), and a mix of sedges. Water depth in this trough is variable depend on annual precipitation, but the maximum depth is close to a meter. In 28, water depths here varied from. to.2 meters. We have observed three amphibian species in this sedge-dominated trough: Pacific chorus frogs, Roughskin newts, and Western toads. Though this is the only area of Diamond Lake in which Roughskin newts have been observed, we observed none here in 28. However, production of Western toad from this site, which has only been observed during the post-rotenone surveys, was particularly high in 28 (TABLE 2). Our VES surveys in this area likely substantially underestimate the actual numbers present even at the time of the survey because dense emergent sedges obstructed our visual field nearly everywhere and nearly all individuals we observed were either on floating lodgepole pine logs or dip-netted from the water column amongst sedges. Juvenile toads produced at this location are probably responsible for the few observations of juvenile Western toads observed near Outflow Pond and the Lake Creek outlet during both post-treatment years. Contrary to the 27 effort, we recorded Northwestern salamanders during both snorkel surveys (TABLE 2) and from trout stomachs (next paragraph) in 28, indicating that northwestern salamanders were present in the fish accessible areas of Diamond Lake in 28. Though our snorkel survey effort was slightly greater in 27 than 28 (six versus five sites were surveyed), we turned up seven larval Northwestern salamanders during the 28 surveys (TABLE 2). Additionally, unverified observation by ODFW personnel suggests that Northwestern salamanders may have been observed during electroshocking efforts on the west side of Diamond Lake (A. Darr, pers. comm.). Native signal crayfish (Pacifastacus leniusculus) with carapace lengths of to 7 mm were the only other macrobiota consistently observed during snorkel surveys. Gastrointestinal tracts from 93 rainbow trout original collected by angling over the interval June-September 28 revealed two Northwestern salamanders, a 95 mm SVL neotene and 6 mm larvae. Both salamanders came from fish exceeding 457 mm (8 in) in length. This pattern differs from the 27 effort, during which no amphibians were found in fish stomachs. Northwestern salamanders were also found in trout GI tracts during the pre-rotenone surveys in 996 (Hayes 997).

TABLE 2. Aquatic amphibian and reptile species by unit and survey in and around Diamond Lake, 996, 27, and 28. Data are the number of each life stage and in parentheses, life stage numbers standardized for effort (i.e., the numbers of a particular life stage observed divided by the number of hours of effort for the indicated survey periods or combinations of sites). Life stages are indicated by a lower case letter following the number: egg mass or packet (e), larvae (), neotene (n), juvenile (j), and adult (a). Surveys units and species names are as in TABLE. Diamond Lake Species Early VES Late VES Snorkel Surveys 996 27 28 996 27 28 996 27 28 Ambystoma gracile 2n (.3) 4n (.37) 7l (2.33) Ambystoma macrodactylum Anaxryus (=Bufo) boreas 6j (.) j (.2) 26j (8.8) a (.29) 228j (65.43) j 2a (.4) a (.83) Pseudacris (Hyla) regilla 5j (.8) 69j (48.25) 4j (4.) 8,63e (363.44) Rana cascadae 3j (.6) j (.7) 3j (.5) Taricha granulosa l (.2) 2l (.4) Thamnophis sirtalis 4a (.8) 3j (.6) 9a (.5) 9j (.32) j (.2) 2a (.57) 6j (.7) a (.9) Amphibians (all) 8,68 (363.54) 8 (3.) 2 (.2) 97 (.64) 4 (.37) Amphibians (aquatic stages) 8,63 (363.44) 3 (.5) 2 (.24) 4 (.37) Garter snakes (all) 7 (.4) 28 (.47) (.2) 8 (2.29) (.9)

TABLE 2 (CONTINUED). Aquatic amphibian and reptile species by unit and survey in and around Diamond Lake, 996, 27, and 28. See beginning of table for data description and life stage codes. Surveys units and species names are as in TABLE. Outflow Pond/Lake Creek Species Early VES Late VES Ambystoma gracile Ambystoma macrodactylum Anaxryus (=Bufo) boreas Pseudacris (=Hyla) regilla Rana cascadae Taricha granulosa Thamnophis sirtalis 996 27 28 996 27 28 2e (24.) 2a (4.46) 4j (4.82) a (.2) 2j (2.4) 3e (22.5) 3e (8.) 6l (.7) 2a (.8) 3j (.2) j (.) 3l (4.44) l (4.) a (4.) Amphibians (all) 36 (4.32) 3 (22.5) 3 (8.) 6 (2.43) 4 (5.56) 2 (8.) Amphibians (aquatic stages) 2 (2.4) 3 (22.5) 3 (8.) 6 (2.43) 3 (4.44) (4.) Garter snakes (all) 3 (.36) 5 (2.6)

TABLE 2 (CONTINUED). Aquatic amphibian and reptile species by unit and survey in and around Diamond Lake, 996, 27, and 28. See beginning of table for data description and life stage codes. Surveys units and species names are as in TABLE. Species Ambystoma gracile Ambystoma macrodactylum Anaxryus (=Bufo) boreas Pseudacris (=Hyla) regilla Rana cascadae Taricha granulosa Thamnophis sirtalis Amphibians (all) Amphibians (aquatic stages) Garter snakes (all) Horse Lake Early VES Late VES 996 27 28 996 27 28 2a a a (.) (.3) (.3) 2l (.25) 4e 2e 6e (.78) (.26) (4.8) 8a 45a 9a (4.44) (5.74) (5.7) (.2) 22j j ~,65j 5j 535j (.22) (.3) (47.47) (5.78) (8.89) 3e 23e 94e (.7) (2.94) (28.23) 8a 3a 45a (.44) (.38) (3.5) 3j j ~76j 24j 2,54j (.7) (.33) (89.37) (9.23) (564.67) 6e 2e (.89) (3.63) a 6a a a a (.6) (.77) (.33) (.3) (.22) j 37j j 3j (.33) (4.68) (.38) (29.) 6e 3e 32e l (.6) (3.96) (.22) a a 2a a (.6) (.33) (.25) (.22) j (.3) 55 2 34,9 4 3,29 (8.6) (5.36) (94.29) (24.9) (5.38) (73.) 39 56 247 2 (2.7) (7.5) (74.7) (.25) (.22) (.6) (.3) 3 (.38) (.22)

TABLE 2 (CONTINUED). Aquatic amphibian and reptile species by unit and survey in and around Diamond Lake, 996, 27, and 28. See beginning of table for data description and life stage codes. Surveys units and species names are as in TABLE. All Sites Combined Species Early VES Late VES 996 27 28 996 27 28 Ambystoma gracile 2 (.26) 3 (.2) 8 (.) 8 (.) Ambystoma macrodactylum 8 (.24) 3 (.2) 2 (.3) 2 (.3) Anaxryus (=Bufo) boreas 8 (.54) 68 (2.76) ~,66 (6.6) ~,66 (6.6) 42 (5.27) 42 (5.27) Pseudacris (=Hyla) regilla 8,9 (236.84) 3 (.2) 78 (9.89) 78 (9.89) 8 (3.56) 8 (3.56) Rana cascadae 2 (.26) 4 (.67) 5 (.7) 5 (.7) 2 (.25) 2 (.25) Taricha granulosa (.) (.) 2 (.25) 2 (.25) Thamnophis sirtalis 2 (.6) 37 (.5) 37 (.5) (.3) (.3) Amphibians (all) 8,295 (239.28) 2 (4.87),955 (26.93),955 (26.93) 53 (9.2) 53 (9.2) Amphibians (aquatic stages) 8,23 (237.3) 63 (2.56) (.5) (.5) 5 (.88) 5 (.88) Garter snakes (all) 2 (.6) 37 (.5) 37 (.5) (.3) (.3)

Unlike 27, when we recorded on one Common garter snake on Diamond Lake proper, we recorded evidence of at last eight Common garter snakes in 28, all during the late VES survey (TABLE 2). Four of these records were represented by shed skins and the remaining live animals lacked food items. Based on standardized effort, garter snakes numbers observed in 28 were more similar to pre-rotenone surveys than numbers recorded in 27. Horse Lake: Despite its far smaller size, Horse Lake exceeded Diamond Lake in species richness in 28 (TABLE ). We had not recorded Common garter snakes at Horse Lake in 27, but recorded them their again in 28, which matches the pattern observed during the pre-rotenone surveys in 996 (TABLE ). Horse Lake appears to have the greatest abundance of amphibians among all the sites surveyed and both reproduction and recruitment appeared particularly robust in 28, which contrasts sharply with 27 (TABLE 2). We observed the highest water levels recorded during any previous survey years at Horse Lake in 28. Some water was present in the typically dry southeast lobe as well as the small pond connected to Horse Lake to the south; both were dry during the late survey in 27. Silent Creek: We recorded evidence of all three amphibians along Silent Creek in 28 (TABLE ). High numbers of post-metamorphic individuals of all three species occupying this area in this area during the late VES in 28 suggests that Silent Creek may represent a refuge for post-metamorphic anurans during the warmer summer season (TABLE 2). Except for limited evidence of reproduction by Cascade frogs at one location (recorded in both years of post-rotenone surveys), we found no evidence of reproduction along Silent Creek. However, large numbers of juvenile Pacific chorus frogs could also indicate that chorus frog breeding occurs after the early VES. Observation from the August 27 survey revealed that the margin of the Silent Creek channel may offer the only significant hydric refuge for amphibians in dry years. Teal Lake: We recorded three amphibian species in Teal Lake in 28, including Longtoed salamander that we had not recorded in 27, which is a pattern parallel to that observed for the pre-rotenone surveys in 996 (TABLE ). Only Common garter snake, which was recorded during the pre-rotenone surveys, was not recorded in 28. The striking pattern of the 28 data for Teal Lake is the high numbers of juvenile Pacific chorus frogs and Cascade frogs observed during the late VES despite lack of evidence of reproduction during the early VES. Two alternatives, which we cannot distinguish may be responsible for this pattern: ) reproduction may have occurred after the early VES; or 2) juveniles observed represent immigrants from a nearby natal site, like Horse Lake, where reproduction is known to have occurred. Similar to Horse Lake, we observed the highest water levels recorded in any previous survey year at Teal Lake in 28. On 3 August 28, the water level in Teal Lake was 26.4 cm (.86 feet) higher than it was during the survey on 27 August 27. DISCUSSION Though the species we recorded in Diamond Lake and its vicinity in 28 were the same species we have recorded in both previous survey years (996 and 27) with the

exception of Roughskin newt and we found them in generally the same places, we recorded more species and generally greater number in most units than in 27 and evidence of substantial recruitment was extensive. Some patterns suggest recovery from a post-rotenone treatment effect in Diamond Lake proper, but the generally increase in numbers across sites suggest an effect that results from the high precipitation winter of 27-28. We briefly discuss important differences that merit discussion. Several lines of evidence indicate that amphibians are either not present or in ecologically low numbers in Diamond Lake proper. In 996, we removed Northwestern salamander larvae, neotenes and even egg mass fragments from the digestive tracts of rainbow trout taken from Diamond Lake, and found Northwestern salamanders invariably concealed beneath rocks during snorkel surveys (Hayes 997), implying that the salamanders were under predation pressure. Such predation pressure was expected to be primarily from fish, as Rainbow trout were the dominant aquatic predator in Diamond Lake proper. At the time, we also removed Northwestern salamanders from Common garter snakes, found moderate numbers of Common garter snakes, and encountered several common garter snake engaged in diving behavior in a manner that indicated foraging beneath shoreline rocks, presumably for Northwestern salamander (common garter snakes are an ineffective fish predator), the only amphibian species we detected in the non-vegetated areas of Diamond Lake in 996 (Hayes 997). In contrast, in 27 we were unable to detect Northwestern salamanders in fish accessible areas of Diamond Lake, no Northwestern salamanders were taken from the digestive tract of rainbow trout, and we found only one young juvenile common garter snake. The fact that Northwestern salamander was present in Diamond Lake in 996, a heavily fishstocked lake for over 8 years, attested to the fact that Northwestern salamander is the amphibian species best able to survive under conditions of fish predation, a pattern well known elsewhere. Moreover, the fact that Northwestern salamanders in Diamond Lake may be neotenic, an aquatic form that reproduces as a larviform (i.e., gilled) adult would have forced individuals to remain in the water pool during drawdown. Hence, failure to record the species in the Diamond Lake proper following the rotenone treatment suggests a treatment effect. The fact that amphibians were found reproducing in areas around Diamond Lake (northwest marsh and pond created by the closed off channel used to pump down Diamond lake) that were isolated from the rotenone treatment pool is consistent with a treatment hypothesis. However, the 28 data again reveal Northwestern salamander both in Diamond Lake proper and in trout GI tracts. This pattern is consistent with recovery from the rotenone treatment, but this recovery may have occurred through two possible paths. One is that Northwestern salamanders were actually extirpated from the Diamond Lake drawdown pool and we are observing the results of recolonization by immigrants from nearby sources (such as the Outflow Pond near Lake Creek). If nearby sources consist of neotenic populations, immigration would have to occur via a water pathway; if nearby sources also have non-neotenic individuals, a terrestrial pathway would also be possible. A second possibility is that Northwestern salamanders were not actually extirpated from the drawdown pool, but a few individuals survived. These few individuals may have represented densities too low to be effectively detected by snorkel surveys and prey too rare to appear in fish stomachs. Both alternatives are plausible and cannot be distinguished on the basis of available data.

The 28 data are inconsistent with our former speculation that Cascade frog is experiencing a regional decline. Recruitment of Cascade frog was dramatic in 28, though one juvenile was detected on Diamond Lake proper. Moreover, we found no evidence of chytrid presence of that could be associated with a decline pattern. Hence, failure to detect Cascade frog in Diamond Lake is likely attributable to conditions in the lake proper and likely results from more that one factor. Our observation in 27 indicates that the shallow water areas that Cascade frogs may have previously used for breeding had disappeared due to vegetation succession on the small peninsula at the mouth of Short Creek may be one contribution. However, while this explanation may address breeding, it does not address inability to detect the relatively long-lived adults. As the post-metamorphic stages of Cascades frogs are closely tied to water, these life stages may have followed the Diamond Lake pool during drawdown and were exposed during rotenone application. Additionally, unlikely Northwestern salamander, Cascade frogs are not particularly resistant to fish predation, a pattern implied by the favorable recruitment in Horse and Teal Lakes. These hypotheses are not mutually exclusive. Similarly, evidence of favorable recruitment in Western toad in the marsh on the Northwestern margin of Diamond Lake may be attributable to the high precipitation year in 28. This area lacking Western toad reproduction and recruitment in 996 may indicate that the current breeding area was not sufficiently isolated from Diamond Lake proper and fish to allow Western toad reproduction. One additional survey year of Diamond Lake will allow identification whether the observed apparent pattern of post-treatment recovery is real or simply reflective a potential more important background pattern of differences in precipitation between years. ACKNOWLEDGMENTS Holly Truemper and her ODFW personnel provided data and/or answered questions regarding Diamond Lake, and were the individuals involved in fish sampling in 28. LITERATURE CITED Crump, M.L., and N.J. Scott, Jr. 994. 2. Visual encounter surveys, pp. 84-92 in Chapter 6: Standard techniques for inventory and monitoring. In: W.R. Heyer et al. (editors), Measuring and monitoring biological diversity: Standard methods for amphibians. Smithsonian Institution Press, Washington, D.C. Gregory, P.T. 978. Feeding habits and diet overlap of three species of garter snakes (Thamnophis sirtalis) on Vancouver Island. Canadian Journal of Zoology 56:967-974. Hayes, M.P. 995. The amphibian fauna of the vicinity of the proposed Cherry Creek Research Natural Area. Final report to The Nature Conservancy, sponsored by the Winema National Forest. Hayes, M.P. 996. Proposal for assessment of the aquatic amphibian and reptile fauna of Diamond Lake. Proposal submitted to the Oregon Department of Fish and Wildlife, Southwest Region, 492 North Umpqua Highway, Roseburg, Oregon 9747.

Hayes, M.P. 997. Assessment of the aquatic amphibian and reptile fauna of Diamond Lake. Final report submitted to the Oregon Department of Fish and Wildlife, Southwest Region, 492 North Umpqua Highway, Roseburg, Oregon 9747. Jones, L.L.C., W.P. Leonard, and D.H. Olson (editors). 25. Amphibians of the Pacific Northwest. Seattle Audubon Society, Seattle, Washington. Kephart, D.G., and S.J. Arnold. 982. Garter snakes diets in a fluctuating environment: A seven-year study. Ecology 63(5):232-236. Loomis, D.W. 995. 994 Diamond Lake Angler Survey. Report from the Oregon Department of Fish and Wildlife, Southwest Region, 492 North Umpqua Highway, Roseburg, Oregon 9747. [August] Oregon Department of Fish and Wildlife. 996. Diamond Lake Fish Management Issues. Draft document white paper. [March] Nussbaum, R.A., E.D. Brodie, Jr., and R.M. Storm. 983. Amphibians and reptiles of the Pacific Northwest. University of Idaho Press, Moscow, Idaho. Thoms, C., C.C. Corkran, and D.H. Olson. 997. Chapter 3 Basic amphibian survey for inventory and monitoring in lentic habitats. In: D.H. Olson et al. (editors), Sampling amphibians in lentic habitats. Society for Northwestern Vertebrate Biology, Northwest Fauna Number 4. Zar, J.H. 999. Biostatistical analysis, 3 rd edition. John Wiley and Sons, New York, New York.