COMPARATIVE SURVIVAL AND REPRODUCTION OF SYMPATRIC EASTERN AND RIO GRANDE WILD TURKEY FEMALES IN NORTHEASTERN SOUTH DAKOTA Chad P. Lehman 1,2 Department of Wildlife and Fisheries Sciences, South Dakota State University, Brookings, SD 57007, USA Anthony P. Leif South Dakota Department of Game, Fish and Parks, P. O. Box 915, Huron, SD 57350, USA Lester D. Flake Department of Wildlife and Fisheries Sciences, South Dakota State University, Brookings, SD 57007, USA Roger D. Shields Department of Wildlife and Fisheries Sciences, South Dakota State University, Brookings, SD 57007, USA Abstract: Rio Grande wild turkeys ( Meleagris gallopavo intermedia ) were introduced into northeastern South Dakota in the early 1970s, flourished until the mid 1980s, and then declined to 200 remnant birds by 1996. Once indigenous to southeastern South Dakota, the eastern subspecies ( M. g. silvestris ) may be a suitable alternative in northeastern South Dakota. We compared reproduction and survival between Rio Grande ( n 32) and transplanted eastern wild turkeys ( n 41) in northeastern South Dakota. We monitored radiomarked females over a 3-year period along the Coteau des Prairie region of northeastern South Dakota. Eastern and Rio Grande females did not differ in rates of nesting ( P 0.73), nest success ( P 0.31), female success ( P 0.53), and clutch size ( P 0.96). Poults reared by eastern wild turkey females had slightly higher ( P 0.07) poult survival to 4 weeks posthatch, while the proportion of successful renests contributing to female success was higher ( P 0.01) for Rio Grandes compared to eastern females. Annual survival (Rio Grande, 0.77 0.005; eastern, 0.72 0.004) and seasonal survival (Rio Grande, 0.83; eastern, 0.89) of females did not differ ( P 0.29) among subspecies. Mammalian predators, primarily coyotes ( Canis latrans ) and red fox ( Vulpes vulpes ), accounted for the largest proportion (45%) of mortality of females. Seasonal cause-specific mortality sources (mammalian, avian, and weather) did not affect ( P 0.65) 1 subspecies more than the other. The cause of the severe Rio Grande population decline over the past decade remains unknown. Survival and reproduction rates of eastern and Rio Grande females exceed those necessary for population maintenance despite occurrences of severe winter weather that are characteristic of the northern interior plains. Proceedings of the National Wild Turkey Symposium 8:123–135 Key words: eastern, Meleagris gallopavo intermedia, Meleagris gallopavo silvestris, mortality, nest success, radiote- lemetry, reproduction, Rio Grande, South Dakota, survival, wild turkey. Determining the best match between habitat and sub- species should be accomplished before developing a wild turkey management plan (Wunz 1992). Trap-and-transplant programs have accelerated restoration of wild turkeys across the United States since the early 1950s (Kurzejeski and Vangilder 1992). Past programs in South Dakota have released Merriam’s ( M. g. merriami ) and Rio Grande sub- species in habitats considerably outside their indigenous 1 Email: turkeys@gwtc.net 2 Present address: P.O. Box 78, Pringle, SD 57747, USA. 123 range (unpublished records, South Dakota Department of Game, Fish and Parks [SDGFP]; Peterson and Richardson 1975). The eastern subspecies of wild turkey has also been released in South Dakota, but within or near its historic range (Grinnell 1910, Over and Thoms 1946, Mosby 1975). Eastern wild turkeys were indigenous to southeastern South Dakota (Over and Thoms 1946), while Rio Grande wild turkeys historically occupied arid brush – grassland habitats in the southcentral Great Plains (Glazener 1967). Previous Midwest Rio Grande introductions in Nebraska (Wunz 1992), South Dakota (L. Rice, SDGFP, personal communi- cation), and North Dakota (Jacobson 1963, Johnson and Knue 1989) have had limited success. However, Midwest- ern reintroductions of the eastern subspecies along the northern edge of their historic range have been highly successful in Wisconsin (Paisley et al. 1995) and Minnesota (Porter 1978); reintroductions along the James River in South Dakota have shown initial success (Leif 1997). Reproductive parameters and nesting ecology have been extensively studied for reintroduced eastern wild tur- key populations in the Midwest (Hecklau et al. 1982, Little and Varland 1981, Miller et al. 1985, Leif 1997). Repro- ductive parameters and nesting ecology of Rio Grande wild turkeys have been studied in their indigenous (Cook 1972, Beasom 1970, Ransom et al. 1987) and introduced ranges (Schmutz and Braun 1989, Keegan and Crawford 1993). Wild turkey nesting success and female success can be highly variable for both eastern (Glidden and Austin 1975, Porter et al. 1983, Vangilder et al. 1987) and Rio Grande subspecies (Beasom and Pattee 1980, Reagan and Morgan 1980, Ransom et al. 1987). Renesting is an important com- ponent in overall reproduction for easterns (Porter et al. 1983, Vangilder et al. 1987), and Rio Grandes (Keegan and Crawford 1993). Investigators have reported survival and mortality rates of eastern wild turkeys in several Midwestern states includ- ing Missouri (Kurzejeski et al. 1987, Vangilder 1995, Van- gilder and Kurzejeski 1995), Wisconsin (Paisley et al. 1995, Wright et al. 1996), Iowa (Little et al. 1990), and Minnesota (Porter 1978). Rio Grande survival and mortality rates in the Midwest are unknown, but some survival and mortality information is available for their native range (Ransom et al. 1987). Predominant causes of wild turkey mortality in the Midwest are predation, severe weather, starvation, illegal killing, and hunting (Porter 1978, Vangilder and Kurzejeski 1995, Wright et al. 1996). The Rio Grande wild turkey population in Marshall and Roberts counties has declined from an estimated peak of 1000 in 1986 to less than 200 in 1996 (M. Stage, SDGFP, personal communication). The Rio Grandes are progeny of a translocation from Oklahoma in 1972. The cause of the decline is unknown. Eastern wild turkeys were released in this area in 1996 to rejuvenate the existing population; this eastern release provided an opportunity to compare repro- duction and survival of the 2 subspecies. The primary ob- jective of this study was to determine which subspecies of wild turkey was most suited to habitats in northeastern South Dakota. Our speci fi c objectives were to compare and evaluate survival, cause-speci fi c mortality, nesting rates, nest success, female success, clutch size, and poult survival between translocated eastern wild turkeys and an estab- lished Rio Grande wild turkey population (includes parent, F 1, and F 2 generations for both subspecies as characterized by phenotype). STUDY AREA Two study areas, located in Marshall and Roberts counties in northeastern South Dakota, were selected based on the proximity of Rio Grande wild turkey populations. The northern Labelle Hollow study area (approx. 95 km 2 [37 mi 2]) had an existing Rio Grande fl ock of about 30 birds, whereas the southern Seiche Hollow study area (ap- proximately 106 km 2 [41 mi 2]) had an existing Rio Grande population of about 70 birds. The study areas were located along the northeastern edge of the Coteau des Prairie region and Minnesota River – Red River Lowlands of northeastern South Dakota (Johnson et al. 1995). The Prairie Coteau is a highland plateau, gently undulating to hilly with elevations ranging from 372 to 610 m (1,220 – 2,001 ft) above mean sea level. Soil composition consists primarily of Seiche loam (6 – 50% slopes) on the sides of wooded coulees or drainages with Forman clay loams on the prairie uplands (Schultz 1975). Study areas had a continental climate with mean annual precipitation of 48 cm (18.72 in) and mean annual temperature of 6 ° C (43 ° F) (National Climatological Data Center 1996 – 98). The woodland habitat is associated with moist east facing breaks of the Prairie Coteau where they descend into the Minnesota River – Red River Lowland. Forests are dom- inated by burr oak ( Quercus macrocarpa ) on the drier slopes. More mesic areas have intermixed associations of trembling aspen ( Populus tremuloides ), green ash ( Fraxinus pennsylvanica ), box elder ( Acer negundo ), elm ( Ulmus spp.), eastern cottonwood ( P. deltoides ), and sugar maple ( A. saccharum ) (Knupp Moore and Flake 1994). Common grasses include cool – season grasses such as smooth brome ( Bromus inermis ), bluegrass ( Poa pratensis ), needle and thread ( Stipa comata ), western wheatgrass ( Agropyron smithii ), and green needlegrass ( S. viridula ). Abundant warm – season grasses include sideoats grama ( Bouteloua curtipendula ), little bluestem ( Andropogon scoparius ), and big bluestem ( A. gerardii ). Shrub species such as smooth sumac ( Rhus glabra ), American plum ( Prunus americana ), and western snowberry ( Symphoricarpos occidentalis ) are found within grasslands and along the edges of woodlands. Row crops increase in abundance as the slopes descend into the Minnesota River – Red River Lowlands. 124 Ecology of Western Subspecies 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License METHODS Capture and Radiotelemetry Wild turkey females were captured in winter using cannon nets (Dill and Thornsberry 1950, Austin et al. 1972) and rocket nets (Thompson and Delong 1967, Wunz, 1974). Adult eastern wild turkeys were trapped in northern Mis- souri and transported to the study areas during January through March of 1996. During that same winter period resident Rio Grande wild turkeys were trapped on the study areas. During the following 2 winters (1996 – 97, 1997 – 98) additional birds (including F 1 and F 2 generations) were captured to replace mortalities. Females captured after the fi rst winter were classi fi ed as eastern or Rio Grande based on phenotype (Schorger 1966, Pelham and Dickson 1992). Captured females were radiomarked with 55-g (2-oz) neck- lace-mounted transmitters equipped with a mercury switch mortality sensor set to activate after 4 hours of inactivity (Advanced Telemetry Systems, Isanti, Minnesota, USA). Signals were monitored daily using hand-held and vehicle- mounted antennae with portable receivers. Reproduction Radiotelemetry was used to determine nest initiation, initiation of incubation, and nest hatching 1 day. Repeated radio locations from localized movements identi fi ed laying behavior and nest initiation. Oscillating pitch in transmitter output identi fi ed movement, and steady pitch at live pulse rates or mortality signals indicated incubation. After a fe- male localized her movements, the female and nest were located using a hand-held yagi antenna. Using just the receiver, the immediate nest site area was located and marked with fl ags on 4 sides at a distance of 5 – 20 m depending on topography and signal strength. We attempted to avoid fl ushing females while marking nests; however, nests that were abandoned due to observer disturbance were not included in nest and female success calculations. Hatch- ing rates and the number of poults hatched were determined at each nest bowl area by investigating eggshells and un- hatched eggs. Nest fates were classi fi ed as successful ( 1 egg hatched), abandoned, fl ushed (disturbed by investigator and subsequently abandoned or destroyed), or depredated. We de fi ned nesting rate as the proportion of females alive on 1 April that attempted to nest, and nesting success as the proportion of all known nests in which at least 1 egg hatched (Cowardin et al. 1985). We de fi ned female success as the proportion of females successful (hatch 1 or more poults) in at least 1 nesting attempt, and renesting rate as the proportion of females unsuccessful on their fi rst nesting attempt that attempted to renest (Cowardin et al. 1985). The nest depredation rate was de fi ned as the percentage of nests lost to predators. We estimated poult survival from the initial number of poults that fl edged to 4 weeks posthatch. Poults were counted by observing each brood feeding or loa fi ng in open areas or by fl ushing the female and her brood (Speake et al. 1985, Vangilder et al. 1987, Flake and Day 1995, Hubbard et al. 1999). Radiomarked females that formed creches with other females were omitted from anal- yses unless amalgamated poults could be differentiated by age. Poult mortality was not estimated after 4 weeks, due to the tendency for females to form brood amalgamations. Herein, F 1 and F 2 generation poults (both subspecies) are referred to as Rio Grande or eastern subspecies based on the phenotype of that poult ’ s female parent due to possible hybridization with an opposing subspecies of male turkey. Maximum likelihood techniques were used to evaluate loglinear models within the CATMOD procedure of SAS (SAS Institute 1989). These models were used to test the null hypotheses that nesting rates, nest success, female success, renesting, and clutch size did not differ between subspecies. Loglinear model methodology has been devel- oped for analysis of categorical data (Goodman 1968, Bishop et al. 1975, Stokes et al. 1995). Study area, age class, year, and subspecies were the main effects in the model. Interactions of the main effects within the model were tested using the maximum likelihood analysis of variance with the chi-square statistic. Study area, year, and age class effects were removed for subspecies comparisons when P 0.10. We used 1-way analysis of variance (ANOVA) within SAS to test the null hypothesis that the mean number of days from 1 April to the initiation of incubation did not differ between subspecies. The Shapiro – Wilks statistic was used to test for validity of the assumptions of equal vari- ances and normality. If the ANOVA assumption of normal- ity was violated, we log-transformed the data. We estimated poult survival using a modi fi ed Kaplan – Meier model (Flint et al. 1995). This estimator does not require independence among brood members for survival estimates, and allows for staggered entry and censoring within a structured sampling regime (Flint et al. 1995). For each subspecies, we used the methods of Sauer and Wil- liams (1989) within the program CONTRAST (Hines and Sauer 1989) to determine if there were among – year and study area differences. We used the Z -test described by Pollock et al. (1989) to compare end point survival between subspecies. Poult survival was pooled by age (adult and juvenile female) because the sample size of juvenile females ( n 6 eastern, n 6 Rio Grande) with poults was small. Poult survival is presented standard error. The signi fi - cance level was set at P 0.05 for all tests. Comparison of Female Survival We estimated seasonal and annual survival distribu- tions using the Kaplan – Meier product limit method (Kaplan and Meier 1958) modi fi ed for staggered entry (Pollock et al. Eastern and Rio Grande Wild Turkeys Lehman et al. 125 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 1989). Seasonal survival was divided into 3 time periods: (1) winter (1 Dec – 31 Mar), (2) spring – summer (1 Apr – 31 Aug), and (3) fall (1 Sep – 30 Nov). We estimated annual survival from 1 April to 31 March of the following year. We classi fi ed juveniles as adults on 31 March of the year following capture. Within the survival model we assumed (1) survival times were unique for each individual and independent of others, (2) a random and unbiased sample of wild turkeys was obtained, (3) transported Missouri eastern females had enough time to adapt to northeastern South Dakota before being incorporated into the survival model (e.g., survival distributions do not include data from the fi rst winter [1995 – 96] following release), and (4) radiotransmit- ters did not affect survival. We tested hypotheses concerning differences in sur- vival among years via the generalized chi-square hypothesis testing procedures described by Sauer and Williams (1989) using the program CONTRAST (Hines and Sauer 1989). Age class (juvenile or adult) within each subspecies and subspecies end-point survival distributions were compared using a Z -test described by Pollock et al. (1989). Age class, study areas, and years were pooled for end-point survival analysis when P 0.10. Signi fi cant differences occurred when P 0.05. Observations were censored the day after radiotransmitters failed, or when radiotransmitter contact was lost. All means are presented standard errors. Mortality We estimated seasonal cause-speci fi c mortality using the Kaplan – Meier product limit method (Kaplan and Meier 1958) modi fi ed for staggered entry (Pollock et al. 1989). We compared subspecies seasonal mortality using a Z -test de- scribed by Pollock et al. (1989). We de fi ned cause-speci fi c mortality as the probability of a wild turkey dying during a given interval from a mortality source in the presence of other competing mortality sources. Necropsy of mortalities determined causes of death, which were classi fi ed as mam- malian predation, avian predation, weather (emaciation or freezing), or unknown. Death was attributed to predation when evidence of hemorrhaging accompanied by puncture wounds occurred on the wild turkey carcass. Species of mammalian predators were identi fi ed from sign such as tracks, scat, blunt puncture wounds, and den holes in the vicinity of the carcass, or caching of the carcass (Hazard 1982). Sign such as sharp puncture wounds, and removal of the head and neck region from the carcass identi fi ed avian predators (most likely great horned owls [ Bubo virginia- nus ]) (Miller and Leopold 1992). During winter, carcasses with emaciated breast muscles without hemorrhaging were classi fi ed as weather-related deaths due to starvation. Car- casses without hemorrhaging or emaciation were classi fi ed as weather-related deaths due to freezing. Specimens with puncture wounds but no hemorrhaging were assumed to have died prior to the puncture wounds. If the cause of death could not be ascertained, it was classi fi ed as unknown. Weather We collected temperature and snow depth data daily for winters (1 Dec – 31 Mar) of 1996 – 97 and 1997 – 98. Minimum and maximum temperatures were obtained from the Britton weather station of the National Climatic Data Center for winters 1996 – 97 and 1997 – 98 (National Cli- matic Data Center 1996 – 98). We averaged minimum and maximum temperatures for yearly comparisons. Snow depths were recorded at Veblen Field Headquarters in Ve- blen, South Dakota (South Dakota State University). Snow depth was averaged from 3 randomly selected sites to the nearest centimeter. Unpaired t -tests within SAS (SAS Insti- tute 1989) were used to compare winter snow depths among years. RESULTS Data were pooled for subspecies comparisons if tests showed nonsigni fi cant differences ( P 0.10) among study areas, years, and age classes. Capture and Radiotelemetry We captured and radiomarked 93 wild turkey females. Of those, 73 females (41 eastern and 32 Rio Grande) were monitored long enough for comparative analyses. All east- ern wild turkey progeny (4 F 1 and 11 F 2 generation females) investigated originated from 26 adult eastern females (par- ent generation, all radiomarked) that survived the fi rst win- ter after being translocated from northern Missouri and released in northeastern South Dakota in winter of 1995 – 96. Of the 32 Rio Grande females used for comparing survival and reproduction, 8 F 1 and 5 F 2 generation females were determined Rio Grande by phenotype. Of the 93 females, 68 (73%) were captured with rocket nets and 25 (27%) with cannon nets. We captured 20 (63%) of the Rio Grande females with cannon nets and 12 (37%) with rocket nets, whereas 56 (92%) of the eastern females were captured with rocket nets (includes females captured in Missouri) and 5 (8%) were captured with cannon nets. Nesting Chronology and Reproduction Nest initiation dates for fi rst nests ranged from 14 April to 5 June for eastern females, and from 14 April to 14 July for Rio Grande females. Chronology of nesting was similar 126 Ecology of Western Subspecies 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License between subspecies (Fig. 1). Years combined, most females (72% of eastern females, 60% of Rio Grande females) began incubation of fi rst nests in May. We observed a difference in the initiation of incubation timing among years ( F 36.74, df 2, 94, P 0.001). Females (both subspecies) initiated incubation earlier in 1998 (median date 5 May) than 1996 (median date 31 May) ( t 6.40, P 0.001), and approached earlier initiation than in 1997 (median date 26 May) ( t 1.78, P 0.08). Dates of initiation of incubation did not differ by subspecies ( F 1.15, df 1, 95, P 0.29). Over the 3-year study period, 31 of 40 eastern females nested, and nesting females initiated 60 nests (70% nest success). Twenty-six of 32 Rio Grande females nested, and those females initiated 64 nests (59% nest success) (Table 1). Nesting rates did not differ ( 2 1.58, df 1, P 0.73) between subspecies in any year. Nesting success be- tween eastern and Rio Grande females also did not differ ( 2 1.05, df 1, P 0.31). However, the year effect was signi fi cant as nest success was lower ( 2 6.79, df 1, P 0.03) in 1997 than either 1996 or 1998. Adult Rio Grande females and juvenile Rio Grande females were not pooled for female success, because adult female success showed a tendency to be higher than juve- niles ( 2 3.41, df 1, P 0.06). Female success did not differ ( 2 0.39, df 1, P 0.53) between adult Rio Grande and eastern females. However, female success was lower ( 2 6.08, df 2, P 0.05) in 1997 than 1996 or 1998. Increased nest depredation in 1997 decreased nest success and female success for both subspecies. The nest depredation rate was 24% in 1996, 49% in 1997, and 28% in 1998. Clutch size did not differ ( 2 0.00, df 1, P 0.96) between subspecies in any year. Years and age class combined, 7 of 40 (18%) successful eastern nests were the result of renests, whereas 17 of 39 (44%) successful Rio Grande nests were the result of renests. Years combined, 88% (14 of 16) of unsuccessful nesting (initial attempt) adult Rio Grande females renested, whereas 50% (3 of 6) of unsuccessful nesting juvenile Rio Grande females renested. Years combined, 64% (9 of 14) of unsuccessful nesting (initial attempt) adult eastern females renested, whereas 33% (1 of 3) of unsuccessful nesting juvenile eastern fe- males renested. Years and age class combined, females did not renest at different rates ( 2 0.81, df 1, P 0.37) between subspecies; however, the proportion of successful Figure 1. Cumulative percentage of wild turkey females initiating nesting in northeastern South Dakota, 1996 – 98. Table 1. Number of females monitored, nests located, and reproduction variables (nesting rate, nesting success, female success, renesting, and mean clutch size [SE]) for eastern females and Rio Grande females in northeastern South Dakota, 1996 – 98. Year No. of Nests Nest Nest Female Renest Clutch b Subspecies females a located rate success success rate size 1996 Eastern 26 (0) 25 0.81 0.68 0.81 0.67 9.75 (0.39) Rio Grande 15 (6) 16 0.93 0.69 1.00 c 0.50 10.17 (0.31) 1997 Eastern 16 (2) 18 0.88 0.45 0.56 0.50 10.44 (0.32) Rio Grande 20 (6) 23 0.80 0.52 0.80 c 0.70 10.23 (0.28) 1998 Eastern 16 (7) 17 0.94 0.82 0.93 0.67 11.00 (0.49) Rio Grande 16 (1) 25 1.00 0.60 0.93 c 1.00 11.57 (0.56) 1996 – 98 Eastern 40 60 0.86 0.70 0.78 0.59 10.45 (0.22) Rio Grande 32 64 0.90 0.59 0.91 c 0.77 10.62 (0.34) aNumbers within parentheses represent proportion of juvenile females for that year. bMean clutch size for fi rst nests. cRepresents adult Rio Grande females only. Eastern and Rio Grande Wild Turkeys Lehman et al. 127 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License renests contributing to female success was higher ( 2 6.02, df 1, P 0.01) for Rio Grandes compared to eastern females. Poult Survival We monitored 595 poults for survival ( n 333 east- ern poults, n 262 Rio Grande poults). In the fi rst period (0 – 2 weeks), eastern poult survival showed a tendency to be higher than Rio Grande poult survival (Table 2), but did not differ. In the second period (2 – 4 weeks), eastern poult survival was nearly identical to Rio Grande poult survival (Table 2). Combined (0 – 4 weeks) eastern poult survival again was marginally although not statistically higher than Rio Grande poult survival (Table 2). Comparison of Female Survival In winter 1995 – 96, we initially released 46 adult east- ern females transported from Missouri. Only 26 of 46 fe- males survived the fi rst winter, and 14 out of 20 (70%) mortalities were attributed to emaciation or freezing due to extreme cold and deep snow. Survival of females from the 2 releases ( fi rst release: 13 Jan 1996, second release: 17 Feb 1996) were compared, and females from the second release (0.95, SE 0.05) had higher survival ( 2 45.80, df 1, P 0.001) than those from the fi rst release (0.27, SE 0.09). Seasonal survival differed little among study areas and subspecies (Table 3). The only difference occurred in winter when Labelle Hollow eastern females had lower ( 2 5.60, df 1, P 0.02) survival than Seiche Hollow eastern females. Labelle Hollow eastern females also had lower ( 2 4.09, df 1, P 0.04) survival than Rio Grande females in winter. Estimates of annual survival for years combined (Fig. 2) did not differ between subspecies (Ta- ble 4). Mortality Mammalian predators (primarily coyote or red fox) caused 45% of the mortality, avian predators (great-horned owls) caused 9%, weather (emaciation and freezing) killed 14% of the females (all eastern females), and 31% of the mortalities were unknown. Most unknown mortalities (5 of 7) occurred in winter during blizzards. This made locating birds daily dif fi cult, and we were sometimes unable to discern weather from predation on those mortalities when much of the carcass had been consumed. Cause-speci fi c Table 2. Poult survival S ˆ ( t ) at 2 and 4 week posthatch intervals from observations of eastern broods and Rio Grande broods during spring – summer in northeastern South Dakota, 1996 – 98. Interval weeks Eastern wild turkeys Rio Grande wild turkeys Year S ˆ ( t ) SE n S ˆ ( t ) SE n Z -value P 1996 0 – 2 0.639 0.101 144 a 0.538 0.172 65 0.676 0.500 2 – 4 0.823 0.052 62 b 0.840 0.076 25 0.327 0.748 0 – 4 0.486 0.095 51 c 0.396 0.167 21 0.725 0.469 1997 0 – 2 0.578 0.133 90 0.449 0.110 89 1.051 0.294 2 – 4 0.789 0.058 52 0.750 0.165 32 0.257 0.802 0 – 4 0.456 0.119 41 0.343 0.121 24 1.055 0.292 1998 0 – 2 0.636 0.115 99 0.546 0.103 108 0.736 0.231 2 – 4 0.912 0.033 57 0.911 0.053 45 0.017 0.990 0 – 4 0.598 0.129 52 0.441 0.097 41 1.320 0.187 Pooled 0 – 2 0.622 0.064 333 0.512 0.068 262 1.633 0.103 2 – 4 0.842 0.058 171 0.843 0.061 102 0.013 0.992 0 – 4 0.511 0.064 144 0.398 0.070 86 1.792 0.073 aInitial number of poults alive that left nest. bNumber of poults alive at 2 weeks of age. cNumber of poults alive at 4 weeks of age. Table 3. Kaplan – Meier seasonal survival estimates ( S ˆ ) and statistical comparisons during winter (1 Dec – 31 Mar), spring – summer (1 Apr – 31 Aug), and fall (1 Sep – 30 Nov) for eastern and Rio Grande females in Marshall and Roberts counties, South Dakota, 1996 – 98. Estimates for both subspecies were pooled by study area, age, and year. Eastern Rio Grande Interval n S ˆ SE n S ˆ SE Z- value P Winter a 23 1.000 0.000 43 0.974 0.024 1.045 0.294 Spring – Summer 65 0.889 0.041 54 0.830 0.052 0.900 0.368 Fall 34 0.938 0.058 27 1.000 0.000 1.072 0.285 aRepresents only Seiche Hollow eastern and Rio Grande females. Labelle Hollow eastern females had lower survival compared to Seiche Hollow eastern females, and Labelle Hollow Rio Grandes had a small sample size ( n 3). 128 Ecology of Western Subspecies 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License mortality rates did not differ ( Z 0.38, P 0.65) between subspecies within any seasonal period (Table 5). Weather The winter of 1996 – 97 had colder temperatures and deeper snow depths than 1997 – 98. Average ambient tem- perature and snow depth (Fig. 3) differed by year for each month. Snow depth signi fi cantly differed by year for each month: December ( t 46.18, P 0.001), January ( t 24.35, P 0.001), February ( t 18.20, P 0.001), and March ( t 17.90, P 0.001). DISCUSSION In northeastern South Dakota, nesting chronology for eastern and Rio Grande females was similar (Fig. 1). Me- dian initiation of incubation dates were variable depending on year, ranging between 5 – 31 May. This range is compa- rable to dates found elsewhere near the same latitude in the Midwest (Porter 1978, Leif 1997). Earlier nesting chronol- ogy in 1998 may be associated with earlier warming tem- peratures as mean temperatures for the month of March increased each year (1996, – 6.5 ° C [20.3 ° F], 1997, – 4.39 ° C [24.1 ° F], 1998, – 3.92 ° C [24.9 ° F]; National Climatological Data Center 1996 – 98). Vangilder et al. (1987) also reported that median dates for initiation of incubation varied with spring phenology over a 4-year period for a stable northern Missouri population. Reproductive success of both subspecies was excellent and would not limit wild turkey populations. Nesting rates for eastern (86%) and Rio Grande females (90%) in this study were similar or slightly lower compared to eastern females introduced along the James River (91%, Leif 1997). These rates were higher than for adult Merriam ’ s females from an established population in southcentral South Da- kota (77%, Flake and Day 1995). In 1997, nesting success and female success were signi fi cantly lower than in 1996 or 1998 due to increased nest depredation. We are uncertain as to why nests were more heavily preyed upon in 1997. A fl uctuating predator population along the Coteau may be a possible explanation (J. Reidel, SDGFP, personal commu- nication). However, even the lowest nesting success (1997, 45%) observed on our study areas was higher than nesting success found elsewhere in South Dakota (44%, Flake and Day 1995; 41%, Leif 1997), and northern Missouri (35%, Vangilder and Kurzejeski 1995). Adult Rio Grande females tended to have higher fe- male success than juvenile females, which is consistent with other studies (Vander Haegen et al. 1988, Vangilder and Kurzejeski 1995). Lowest female success (1997, 56%) found on our study areas was higher than that reported for Rio Grandes (29%, Reagan and Morgan 1980; 25%, Ran- som et al. 1987), and easterns (40%, Vangilder and Kur- zejeski 1995; 47%, Leif 1997) found elsewhere. Mean clutch sizes (eastern, 10.5; Rio Grande, 10.6) were compa- rable to eastern females from northern Missouri (10.6, Vangilder et al. 1987), and South Dakota (10.4, Leif 1997). Clutch sizes were also comparable to Rio Grandes found in Texas (11.1, Ransom et al. 1987). Renesting from both subspecies (Rio Grande, 77%; eastern, 59%) was higher than rates reported for eastern females in South Dakota (26%, Leif 1997), and northern Missouri (41%, Vangilder and Kurzejeski 1995). Rio Grande renesting was higher than for Rio Grandes in Texas (6%, Ransom et al. 1987) and Rio Grandes from Oregon (62%, Keegan and Crawford 1993). Similar to other studies, most poult mortality occurred by 2 weeks posthatch for both subspecies (Speake et al. 1985, Flake and Day 1995, Vangilder and Kurzejeski Figure 2. Pooled (study areas, age class, and years) annual Kaplan – Meier survival estimates ( S ˆ ) of eastern and Rio Grande wild turkeys in Marshall and Roberts counties, South Dakota, for 1996 – 98. End-point eastern sur- vival was 0.72 (SE 0.004), and Rio Grande survival was 0.77 (SE 0.005). Table 4. Kaplan – Meier annual survival estimates ( S ˆ ) and statistical comparisons for eastern and Rio Grande females in Marshall and Roberts counties, South Dakota, 1996 – 98. Estimates for both subspecies were pooled by study area, age, and year. Eastern Rio Grande Year n S ˆ SE n S ˆ SE Z- value P 1996 – 97 29 0.734 0.008 24 0.741 0.009 0.053 0.958 1997 – 98 24 0.698 0.007 20 0.794 0.010 0.734 0.461 1996 – 98 (pooled years) 53 0.721 0.004 44 0.765 0.005 0.485 0.628 Eastern and Rio Grande Wild Turkeys Lehman et al. 129 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 1995). Easterns (i.e., poults of eastern females) had higher poult survival rates (51% at 4 weeks), and Rio Grandes (i.e., poults of Rio Grande females) had lower poult survival rates (40% at 4 weeks) (Table 2) compared with eastern poults from a northern Missouri population (46% at 4 weeks, Vangilder and Kurzejeski 1995). Easterns had higher poult survival rates, and Rio Grandes had similar poult survival rates compared with Merriam ’ s poults at 4 weeks posthatch in southcentral South Dakota (43%, Flake and Day 1995). Annual survival and seasonal survival for females, with the exception of winter seasonal survival, did not differ between age classes, study areas or subspecies. Analysis of winter survival indicated Labelle Hollow eastern females had lower survival than both Seiche Hollow eastern females and Rio Grande females. Seiche Hollow eastern females and Rio Grande females had an emergency food source (e.g., farmstead grains), whereas Labelle Hollow eastern females probably had dif fi culty fi nding food because of deep snow (Fig. 3). Rio Grandes, which primarily remained within or near farmstead feedlots, may have enhanced eastern sur- vival at Seiche Hollow. The larger Rio Grande fl ock may have attracted the Seiche Hollow eastern females to an emergency food source because home range movements of Seiche Hollow eastern females were near farmstead sites during periods of deep snow and cold temperatures (Leh- man 1998). In contrast, Labelle Hollow eastern females avoided farmstead sites during winter, and relied more on residual grain from fi elds for winter survival. In Minnesota, eastern wild turkeys that had access to winter food sources (e.g., food plots) had higher survival than wild turkeys without winter food sources (0.90 versus 0.35) during severe winter years (Porter et al. 1980). In Massachusetts (Vander Haegen et al. 1989) and Iowa (Little et al. 1990), survival was high ( 0.72) where turkeys had access to dairy farms or grain fi elds. Unlike Labelle Hollow eastern females, Seiche Hollow eastern and Rio Grande female survival rates were lowest during spring – summer, and most (71%) spring – summer mortality was caused by predation. All spring – summer pre- dations occurred during incubation or within 1 week after hatching. High predation rates, especially during nesting and brood rearing, have also been reported in Minnesota (Porter 1978), Massachusetts (Vander Haegen et al. 1988) and Iowa (Little et al. 1990). Spring – summer survival rates Table 5. Seasonal cause-speci fi c mortality rates (standard error), and deaths ( n ) of eastern and Rio Grande females (study areas and age class combined) during winter (1 Dec – 31 Mar), spring – summer (1 Apr – 31 Aug), and fall (1 Sep – 30 Nov) in Marshall and Roberts counties, South Dakota, 1996 – 98. Year Spring-Summer Fall a Winter a Mortality source Rate (SE) n Rate (SE) n Rate (SE) n Eastern 1996 Mammalian 0.000 (0.000) 0 0.091 (0.058) 2 0.042 (0.041) 1 Avian 0.000 (0.000) 0 0.000 (0.000) 0 0.000 (0.000) 0 Weather 0.000 (0.000) 0 0.000 (0.000) 0 0.042 (0.041) 1 Unknown 0.040 (0.039) 1 0.000 (0.000) 0 0.083 (0.055) 2 1997 Mammalian 0.211 (0.091) 4 0.000 (0.000) 0 0.000 (0.000) 0 Avian 0.000 (0.000) 0 0.000 (0.000) 0 0.000 (0.000) 0 Weather 0.105 (0.071) 2 0.000 (0.000) 0 0.000 (0.000) 0 Unknown 0.000 (0.000) 0 0.000 (0.000) 0 0.083 (0.055) 1 1998 Mammalian 0.000 (0.000) 0 Avian 0.000 (0.000) 0 Weather 0.000 (0.000) 0 Unknown 0.000 (0.000) 0 Rio Grande 1996 Mammalian 0.133 (0.085) 2 0.000 (0.000) 0 0.000 (0.000) 0 Avian 0.000 (0.000) 0 0.000 (0.000) 0 0.000 (0.000) 0 Weather 0.000 (0.000) 0 0.000 (0.000) 0 0.000 (0.000) 0 Unknown 0.067 (0.064) 1 0.000 (0.000) 0 0.050 (0.048) 1 1997 Mammalian 0.100 (0.067) 2 0.000 (0.000) 0 0.000 (0.000) 0 Avian 0.050 (0.049) 1 0.000 (0.000) 0 0.000 (0.000) 0 Weather 0.000 (0.000) 0 0.000 (0.000) 0 0.000 (0.000) 0 Unknown 0.050 (0.049) 1 0.000 (0.000) 0 0.000 (0.000) 0 1998 Mammalian 0.053 (0.051) 1 Avian 0.053 (0.051) 1 Weather 0.000 (0.000) 0 Unknown 0.000 (0.000) 0 aSeasonal cause-speci fi c mortality not monitored during 1998 – 99. 130 Ecology of Western Subspecies 23285540, 2000, S1, Downloaded from https://wildlife.onlinelibrary.wiley.com/doi/10.1002/j.2328-5540.2000.tb00262.x by University Of Florida, Wiley Online Library on [15/10/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License for eastern females (0.89) and Rio Grande females (0.83) were above rates reported for Wisconsin (0.72, Wright et al. 1996), and equal or below rates reported for Missouri (0.88, Vangilder and Kurzejeski 1995). Fall survival rates for eastern females (0.94) and Rio Grande females (1.0) were above rates reported for Wisconsin (0.81, Wright et al. 1996) and Missouri (0.88, Vangilder and Kurzejeski 1995). Pooled annual survival rates of females in northeastern South Dakota (eastern, 0.72; Rio Grande, 0.77) were higher than rates reported for Missouri (0.56, Vangilder and Kur- zejeski 1995), Wisconsin (0.53, Wright et al. 1996), Texas (0.62, Campo et al. 1984), Iowa (0.64, Little et al. 1990), and Indiana (0.50, Miller et al. 1985). Eastern females had lower annual survival rates, and Rio Grande females had similar annual survival rates, when compared with eastern females along the James River, South Dakota (0.78, Leif 1997). Most mortality resulted primarily from mammalian predation (45%) for both subspecies, and weather (14%) only affected survival for eastern females. How