Neonate survival is affected by predation (Nelson & Woolf 1987, Kunkel & Mech 1994, Benzon 1998, Ballard et al. 1999, Brinkman et al. 2004a, Swanson 2005), disease (Schulz 1982, Brinkman et al 2004b), maternal age (Ozoga & Verme 1986, Kunkel & Mech 1994), and condition of the dam (Porath 1980). In addition, survival of research animals may be affected by capture and marking activities (Hamlin et al. 1982). Capture techniques that minimize disturbance to females and neonates are needed to decrease the probability of capture-related mortality (White et al. 1972). Furthermore, studies should evaluate the indirect effects that marking techniques may have on survival (e.g. predation, starvation and disease) to validate their use (Murray & Fuller 2000).

Although female behaviour is a useful technique for finding neonate deer (Downing & McGinnes 1969, White et al. 1972, Garner et al. 1976, Huegel et al. 1985), capturing neonates in areas with dense cover may be difficult. However, vaginal-implant transmitters (VITs; Bowman & Jacobsen 1998, Carstensen et al. 2003, Johnstone-Yellin et al. 2006) expelled from females at parturition allow researchers to locate birth sites of neonates (Garrott & Bartmann 1984) independent of habitat characteristics (Bowman & Jacobson 1998). Furthermore, VITs allow researchers to sample newborn animals that may typically be missed (Seward et al. 2005).

Garrott & Bartmann (1984), Bowman & Jacobson (1998), Carstensen et al. (2003), and Johnstone-Yellin et al. (2006) used VITs with varying success. Their studies focused on efficacy of VITs to capture neonate deer but did not evaluate impact of VITs on neonate survival. Our objective was to compare survival rates of neonate white-tailed deer Odocoileus virginianus captured using VITs and traditional (i.e. nocturnal and diurnal) ground searches to determine if neonate survival was impacted by capture method. We hypothesized that the use of VITs would not affect neonate survival.

Study area

Southwest Minnesota (43°29′N to 45°16′N - 093°38′W to 096°27′W) is characterized by flat to rolling topography with elevation ranging within 229–608 m a.s.l. (Albert 1995). As a result of the Wisconsin Glaciation (10,000–100,000 years ago), lobes of glacial ice shaped the landscape in this region (Minnesota River Basin Data Center 2003) creating steeper topography along the Minnesota River and its tributaries (Voigtlander 1999) than in upland prairie habitats. Our study area covered 34,627 km² of the farmland region in Minnesota (Swanson 2005).

Tall and mixed prairie grasses (6.5%) accounted for the majority of native vegetation in southwest Minnesota. Of these, big bluestem Andropogon geradii, little bluestem Schizachyrium scoparium, Indiangrass Sorghastrum nutans, switchgrass Panicum virgatum, tall dropseed Sporobolus asper, and sideoats grama Bouteloua curtipendula characterized uplands (Johnson & Larson 1999, Voigtlander 1999). In low wet areas, prairie cordgrass Spartina pectinata, reedgrass Calamagrostis arundinacea, and sedges (Cyperaceae) were common (Voigtlander 1999). Forested areas (3.0%) were dominated by bur oak Quercus macrocarpa, basswood Tilia americana, green ash Fraxinus pennsylvanica, and eastern cottonwood Populus deltoides (Minnesota Association of Soil and Water Conservation Districts Forestry Committee 1986).

Methods

During January-February 2003, we captured female white-tailed deer using helicopter net-guns (Barrett et al. 1982). We immobilized captured deer with ketamine hydrochloride (5 mg/kg IM; Ketaset®; Fort Dodge Laboratories, Fort Dodge, Iowa, USA) and xylazine hydrochloride (1 mg/kg IM; Xyla-ject®, Phoenix Pharmaceutical Inc., St. Joseph, Missouri, USA) prior to transport (Mech et al. 1985, Kreeger et al. 2002). Adult female white-tailed deer were radio-collared (Advanced Telemetry System, Isanti, Minnesota USA), aged, ear-tagged, measured (chest and neck circumference), and given a broad spectrum antibiotic, after which they were fitted with VITs (N=14; Advanced Telemetry System, Isanti, Minnesota, USA). Each VIT had a unique frequency and was equipped with a temperature activated sensor that doubled in pulse rate when expelled from the female (Bowman & Jacobson 1998). To reduce vulvar trauma, ends of the antenna were covered with plastic and placed approximately 1–2 cm outside the vulva (Carstensen et al. 2003). Following Mech et al. (1985), we reversed chemical immobilizations using yohimbine hydrochloride (0.125 mg/kg IV; Yobine®, Ben Venue Laboratories, Inc., Bedford, Ohio, USA).

During May-June 2003, neonates were captured and radio-collared during traditional ground searches in Redwood and Renville counties, Minnesota (Swanson 2005). Nocturnal searches were conducted with spotlights from vehicles on roads near potential fawning areas by 2-person teams. Neonates were located by observing behavioural changes exhibited by postpartum females (Downing & McGinnes 1969, Huegel et al. 1985, Benzon 1998). When a neonate was sighted or if the female's behaviour was suspect, a fast, noisy, approach was made to attempt to invoke the ‘drop’ response of neonates (Downing & McGinnes 1969). Neonates able to run were pursued on foot and captured using long-handled landing nets (Frabill Inc., Jackson, Wisconsin, USA). Ground searches were systematically conducted if an observed adult female remained in the vicinity after being flushed or used vocalizations (e.g. snorting) indicating the possible presence of a neonate (Lund 1975). Additionally, diurnal ground searches (Lund 1975) were conducted in probable fawning habitats using multiple-person crews (i.e. 3–8 people). Crews searched potential fawning areas in a linearly-spaced format looking for hiding neonates or solitary females.

Beginning in mid-May 2003, we monitored VIT signals three times daily (i.e. every eight hours) using a vehicle-mounted radiotelemetry antenna system equipped with a military grade compass (Brinkman et al. 2002, Brinkman 2003) to determine location of birth sites (area with vegetation consumed by female and evidence of afterbirth or neonate present) and to capture neonates. When the temperature activated switch indicated an implant had been expelled, we located the transmitter using hand-held telemetry and secured neonates for processing. If neonates were not immediately located at the birth site, a 300-m radius area surrounding the VIT and the location of radio-collared females was searched (Carstensen et al. 2003). Expelled VITs were located regardless of time of day to reduce chance of neonates moving away from birth sites.

Captured neonates were handled for an average of 4.7 minutes, during which sex, age and weight were recorded. Neonates were aged based on hoof growth measurements (Haugen & Speake 1958, Brinkman et al. 2004c), placed in a drawstring sac that had been stored in natural vegetation to reduce scent contamination and weighed using a digital hanging scale (Extech Instruments, Melrose, Massachusetts, USA). Neonates were fitted with radio-collars (Advanced Telemetry Systems, Isanti, Minnesota, USA and Telonics Inc., Mesa, Arizona, USA) designed to expand and break away six months post capture. To reduce scent contamination, radio-collars were stored in natural vegetation common to searched habitats. Additionally, vinyl gloves were worn by all crew members involved with capture to minimize transfer of human scent. A Global Positioning System (Garmin International Inc., Olathe, Kansas, USA) was used to mark capture locations.

Radio-collared neonates were monitored for survival 1–3 times daily until three months of age. When radio-collars remained motionless for four hours, a motion sensitive mortality switch activated doubling the pulse rate of the radio-collar, indicating a potential mortality. Field necropsies were conducted at the site of death and <24 hours from a previous live signal. Evidence of predators (e.g. tracks, scat, hair or carcass burial) along with condition of carcass was examined. When cause of neonate death was not identifiable at the death site, the carcass was sent for further testing to the Animal Disease and Diagnostic Lab at South Dakota State University. Mortalities caused by coyote Canis latrans or domestic dog Canis familiaris were considered canid predation based on evidence at the death site. All field methods complied with animal care and use guidelines of the American Society of Mammalogists (1998) and the study was approved by the Animal Care and Use Committee (approval number: 02-A043) at South Dakota State University.

We used the Kaplan-Meier method (Kaplan & Meier 1958) modified for staggered entry (Pollock et al. 1989) to calculate monthly survival rates from June through August 2003. Survival estimates were compared between neonates captured using VITs and traditional searches using Program CONTRAST (Hines & Sauer 1989). We conducted t-tests (alpha set at P≤0.05) to determine if retention of VITs was influenced by deer weight or chest girth. Also, we used t-tests to test for differences in age and handling time between capture methods and χ² test to determine if the percentage of neonates that died differed between capture methods. A Bonferroni correction factor was used to maintain alpha when multiple χ² tests or t-tests were performed. We censored neonates if collars failed or prematurely dropped off neonates. Sibling neonates captured at the same birth site via VITs were considered to be independent in the survival analysis because neonates were usually located >0.4 km apart one week post capture.

Results

We fitted 14 adult females with VITs and seven of the implants (50%) were recovered at birth sites, leading to the capture of 14 neonates; two sets of triplets, three sets of twins, and two single neonates. During 20 May-3 June 2003, we captured and radio-collared 21 neonates (14 from VITs and seven from traditional ground searches; 10 males and 11 females) in Redwood and Renville counties, Minnesota. Mean date of birth was 25 May (SE=1.3; range: 13 May-3 June). Mean neonate age averaged <1 and 6 days (range: 2–11) for neonates captured via VITs and traditional ground searches (Table 1), respectively. Age at capture differed (t=3.182, df=7, P=0.015) between neonates captured using VITs versus those captured via traditional ground searches (see Table 1). However, age at death did not differ (t=1.538, df=5, P=0.185) between capture techniques. During summer 2003, survival of neonates captured using VITs one, two and three months post capture was 0.76 (SE=0.05, N=14), 0.64 (SE=0.07, N=11) and 0.64 (SE=0.08, N=9), respectively. Neonate survival one, two and three months post capture for neonates captured using traditional ground searches was 0.71 (SE=0.11, N=7), 0.71 (SE=0.15, N=5) and 0.71 (SE=0.15, N=5), respectively. Survival did not differ one month (χ²₁=0.349, P=0.554), two months (χ²₁=0.186, P=0.666) or three months (χ²₁=0.177, P=0.674) post capture for neonates captured using VITs compared to those captured using traditional ground searches. Canid predation accounted for all (N=7) neonate mortalities during the first three months post capture (see Table 1). Canids killed five of seven (71%) neonates captured via VITs. However, the percentage of neonates that died did not differ between capture methods (χ²₁=0.107, P=0.743). Further, capture weight (see Table 1) of neonates did not differ (t=0.911, df=19, P=0.374) by capture method. Additionally, neonates captured using VITs that died did not differ (t=1.628, df=17, P=0.122) in weight at capture from those that survived. Neonates captured using traditional ground searches were handled longer (t=3.700, df=19, P=0.002) than neonates captured using VITs (see Table 1). However, handling time did not differ (t=1.826, df=17, P=0.086) between neonates that died and were captured using VITs versus neonates that survived. During our study, no instances of abandonment occurred.

Table 1.

Capture and mortality data of neonate white-tailed deer captured using traditional ground searches and vaginal-implant transmitters in Redwood and Renville counties, Minnesota during spring 2003.

Adult female white-tailed deer that expelled implants prior to parturition had average weight and chest girths of 67.1 kg (SE=1.4) and 103.9 (SE=1.7), respectively. For deer that maintained VITs until parturition, average weight and chest girth was 67.8 kg (SE=1.3) and 104.7 cm (SE=1.8), respectively, so there was no difference in deer weight (t=0.361, df=12, P=0.724) or chest girth (t=0.361, df=12, P=0.724) for deer that prematurely expelled VITs versus those that maintained VITs until parturition.

We estimated the cost per neonate between capture methods. Cost to deploy one VIT used to capture neonates was $1176.19. The cost included $400 to capture each adult female (N=14) by helicopter net-gun (Barrett et al. 1982), $200 for each adult radio-collar (N=14), $25 for immobilizing drugs (Ketamine, Xylazine and Yohimbine) for each adult captured (N=14), $200 for each VIT (N=14), $10/hour/person (N=10) cost during eight hours' adult capture (during winter), $400 lodging expense, +$10/hour/person (N=1) cost during 81 hours' neonate capture and locating expelled VITs (during spring) +$106.60 for 209.3 liters of fuel to monitor for expelled implants, $200 for neonate radio-collars (N=14), and miscellaneous equipment (e.g. nets, scales and calipers). Cost to capture one neonate using traditional ground searches was $702.86. The cost included $200 for neonate radio-collars (N=7), $10/hour/person (N=2) cost during 116 hours' neonate capture (during spring), +$800 lodging expense, $400 for 757.1 liters of fuel to search for neonates during traditional ground searches, and miscellaneous equipment (e.g. spotlights, scales and calipers).

Discussion

During summer 2003, all documented neonate mortalities (N=7) were attributed to canid predation, which accounted for 36 and 29% of mortality of neonates captured using VITs and during ground searches, respectively. Berger et al. (2001) suggested that a lack of vigilance and failure to detect dangerous predators may contribute to mortality in regions where young are primary targets of predation. During our study, specific measurements and observations of neonates were conducted at birth sites. This disturbance could have weakened the female-neonate bond because of the association with danger at the birth site. Complete isolation is necessary during parturition for proper female-neonate bonding (Ozoga et al. 1982), and disturbance during the first hours of life may weaken this bond. Similarly, Cook et al. (1971) suggested that neonates may eventually die due to the breakdown of female-neonate bonds when a disturbance occurs during the first two days of life. As there was no difference in survival between capture methods during our study, disturbance to the female-neonate bond was not impacted by our activities at the birth site.

Neonates captured using traditional ground searches were handled longer than neonates captured using VITs (see Table 1). Similar to the results of Carstensen Powell et al. (2005) and White et al. (1972) handling time did not influence neonate survival. Age at capture differed between capture methods (see Table 1), whereas neonate age at death did not. Neonates captured using VITs survived as long as neonates captured using traditional ground searches and had no apparent survival disadvantage three months post capture. Furthermore, neonates captured using VITs did not seem to have any physiological disadvantages, as capture weight did not differ between capture methods (see Table 1). Weight at capture may have been similar between capture methods due to the condition of the female and the variability of neonate weight at birth. Additionally, neonates that died and were captured using VITs did not differ in weight at time of capture from neonates that survived. Because there was no significant difference in weight at capture between methods, we would expect that survival was not impacted by use of VITs.

Of the predated neonates captured using VITs, 60% died within one month post capture. Similarly, Carstensen Powell et al. (2005) documented high neonate mortality in animals captured within 24 hours of birth. It could be possible that some females react differently to disturbances within the first 24 hours after birth. However, we did not document any negative impact to neonate survival because of the disturbance associated with radio-collaring neonates captured using VITs at the birth site.

Using the predictive equation for adult female weights provided by Weckerly et al. (1987), we calculated an average live weight at capture for adult female deer fitted with VITs of 63.7 kg. Although this weight is slightly lower than that of deer (69.3 kg) captured by Brinkman (2003), we considered these deer to be in excellent condition. Therefore, based on the physical condition of the adult females and the results of other researchers (Bowman & Jacobson 1998 and Carstensen et al. 2003), we did not suspect any reproductive problems associated with using VITs. However, 50% (N=7) of VITs were prematurely expelled prior to parturition. Rate of premature expulsion (50%) was higher than documented by Carstensen et al. (2003; 11%) and Bowman & Jacobsen (1998; 31%). Seward et al. (2005) suggested that implants in elk Cervus elaphus can be prematurely expelled by early contractions, stillborn passage (Andersen & Linnell 1998), or an extruding antenna being pulled out by the animal. Because we could not determine if adult females that prematurely expelled VITs continued pregnancy to successful parturition, we were unable to document reproductive problems that may have been caused by VITs. Nevertheless, no VITs were retained past the parturition season.

Because cost is important when planning research (Conner et al. 1987), researchers must utilize efficient techniques that meet desired goals. During our study, the monetary investment of capturing neonates with VITs was approximately $473 higher per neonate than capturing with traditional ground searches. However, we required 151 less person hours to capture twice the number of neonates during spring neonate capture using VITs compared to traditional searches. Furthermore, we used approximately 552.7 fewer liters of fuel to monitor VITs versus traditional searches to capture neonates. We realize that the cost to capture adult females for the benefit of capturing neonates using VITs is high. However, there are economic benefits to deploying VITs when research objectives include monitoring adult females. Specifically, when adults are factored into the overall cost per adult female/VIT, the cost to capture neonates using VITs was approximately $115 less expensive than capturing neonates using traditional searches.

Conclusions

During our study, we did not document any negative effects associated with finding and capturing neonates with the assistance of VITs. Although, the majority (71%) of neonates that died were captured using VITs, this disparity was due to chance. We realize that our sample size of neonates was low, and that this may have had an influence on the outcomes of our statistical tests. Nevertheless, our findings are generally supportive of other studies that employed VITs to locate neonate cervids. Future research is needed to determine if the effects of radio-collaring adult females and implanting VITs during pregnancy is an effective tool for safely capturing neonates. Our results indicate that VITs are a means to study neonate white-tailed deer when other capture methods are not practical. VITs allowed for the capture of single to multiple (N=2 or 3) neonates within hours of parturition. Capturing neonates immediately following parturition ensures the detection of all mortality sources that may otherwise be unknown with traditional capture methods. Researchers must evaluate the success of similar studies to determine if the cost per animal is beneficial using VITs compared to traditional capture methods (Seward et al. 2005). Nevertheless, when capture of neonates is difficult using female behaviour (Downing & McGinnes 1969, Huegel et al. 1985, Benzon 1998) via random searches, VITs could provide the only means for capturing neonates and locating birth sites that would otherwise be missed in dense habitats.