Evaluation of Capture Techniques on Lesser Prairie-Chicken Trap Injury and Survival

Increased attention to the conservation of nongame species and species of conservation concern within academia and regulatory agencies has gained momentum over the past 15 y (Scalet 2010). The shift in conservation philosophy has focused research on the welfare of individuals as well as populations. This includes ethical treatment of research animals in the lab and during wildlife field studies. The Wildlife Society maintains a similar philosophy, as indicated by the society’s final position statement on animal rights that states capturing individuals for research is acceptable given individuals are treated ethically and humanely (http://wildlife.org/wp-content/uploads/2014/05/animal_rights_8.30.2011.pdf; accessed 5 February 2015).

Ethical treatment of lab research animals is required under the Animal Welfare Act of 1966. However, little attention has been given to wildlife research animals until recent years despite the fact that universities require Institutional Animal Care and Use Committees review of all animal research procedures to insure research protocols minimize pain and duress to research animals. According to The Wildlife Society’s Code of Ethics, professionals should “exercise high standards in the care and use of live vertebrate animals used for research” (http://joomla.wildlife.org/index.php?option=com_content&task=view&id=769; accessed 5 February 2015). High standard of care for all animals is important, but is essential for animals of special conservation concern because unnecessary pain and duress may indirectly influence vital rates that contribute to population persistence.

Lesser prairie-chicken Tympanuchus pallidicinctus populations have decreased across their distribution (Van Pelt et al. 2013) and the species was listed as threatened under the United States Endangered Species Act in May 2014 (ESA 1973, as amended; United States Fish and Wildlife Service 2014). Research and management efforts have increased to aid in conservation of the species. Although walk-in funnel traps, rocket nets, and drop nets are used to capture the species, knowledge of frequency or extent of injuries caused by the various trapping methods is lacking. This information is important for two reasons. First, productivity may be indirectly influenced if injuries to female lesser prairie-chickens are severe or frequent because females are solely responsible for nest and brood-rearing activities (Hagen et al. 2009). Second, as a species listed under the Endangered Species Act, any activity that would result in "take" as defined by the act (e.g., to harass, harm, pursue, hunt, shoot, wound, kill, trap, capture, or collect, or to attempt to engage in any such conduct) would require Endangered Species Act compliance (ESA 1973, as amended). Therefore, an investigation of the number and type of injuries and the effects of capture injury on survival is warranted for this species not only for ethical considerations, but for biological and regulatory purposes.

The objectives of our study were to 1) assess whether type of injury and injury location occurred proportionately across walk-in funnel traps, rocket nets, and drop nets; 2) assess whether injuries occurred proportionately between male and female lesser prairie-chickens; 3) assess whether injuries caused by walk-in funnel traps, rocket nets, and drop nets influenced postcapture survival of radiotagged individuals 7–65 d postcapture; and 4) assess whether postcapture survival differed between injured and uninjured birds (7–65 d postcapture). We hypothesized injuries would be more frequent with walk-in funnel traps. We hypothesized lesser prairie-chickens captured in walk-in funnel traps would be susceptible to cuts and scrapes, birds captured with rocket nets would have more severe injuries such as broken bones or decapitation, and birds captured in drop nets would be prone to broken feathers. In addition, we expected more injuries to occur in walk-in funnel traps on account of increased probability of recapturing non-target individuals, as well as increased probability of multiple individuals being captured in one trap. Finally, we hypothesized postcapture survival would differ among the three trapping techniques, but not between injured and uninjured individuals.

This study took place on privately and publicly owned lands in Cochran, Hockley, Terry, and Yoakum counties, Texas, and Roosevelt County, New Mexico, on the Southern High Plains (33°22′N, 102°49′W). The landscape consisted of a matrix of grassland, cropland, and gently undulating sandhills and was dominated by sand shinnery oak Quercus havardii and sand sagebrush Artemisia filifolia with mixed grasses and forbs. Temperatures averaged 7.6°C in winter and 25.0°C in summer, average yearly precipitation was 44.6 cm, and there was extreme variation in weather parameters within and among years on the study area (see Grisham et al. 2013, 2014). The study area experienced drought conditions in the winter and spring of 2009 and throughout the entirety of 2011 (Nielsen-Gammon 2012). The major land uses for this study area were intensive agriculture and cattle ranching. Common crops were cotton and grain sorghum (Grisham et al. 2014). Oil production also occurred in this area during the study. Detailed descriptions of the vegetation community on the study areas are in Zavaleta (2012), and Grisham et al. (2013).

We used walk-in funnel traps, rocket nets, and magnetic drop nets to capture lesser prairie-chickens on 21 leks from 1 March to 15 April, 2006–2012. Walk-in funnel traps consist of a cylindrical wire cage with a small conical opening that allowed the bird to walk inside the trap (Haukos et al. 1989; Schroeder and Braun 1991). We built each trap using a 2.4 m × 0.6 m segment of 14-gauge, welded cage wire. We cut out each wire segment, circled the wire back on itself, and bound the segment end-to-end using hog-ring clips and zip-ties. Traps were approximately 0.7 m in diameter and 0.7 m tall. We constructed funnels using 20-gauge chicken wire and each funnel was approximately 20.0 cm in length, approximately 22.0 cm in diameter at the opening, and 15.0 cm in diameter at the exit. We crimped together loose edges to reduce risk of injury. We used cloth netting as the ceiling of each trap; we fastened the cloth with plastic zip-ties to the tops of the traps to prevent escape of trapped birds. We used drift-fences made of chicken-wire in conjunction with walk-in funnel traps to divide male territories on the lek (Haukos et al. 1989) and to channel individuals toward the entrance of the funnel (Figure 1). We used 10–30 walk-in funnel traps on each lek, depending on the physical shape (i.e., square abandoned oil pad with defined edges as compared with natural lek in rangeland with natural transition between open patches and vegetation) and size of the lek. Our goal was to capture females, so we immediately removed females from traps; however, we monitored male incidental captures and removed each individual within 30 min of initial capture if the bird was not showing behavioral signs of stress (i.e., repeated jumping or flapping) or immediately if other individuals on leks or predators harassed the captured individual.

We used a 23 m × 13 m, 4-rocket net system (Figure 2) with 3.5 cm × 3.0 cm mesh equipped with Winn-Starr explosive-powered recoilless rockets (Winn-Star, Marion, IL) to capture lesser prairie-chickens on leks. Each rocket was propelled by a 40-grain black-powder Winn-Star explosive charge. Rockets were 26 cm × 5.1cm, weighed 3.6 kg, and attached to the net at four locations using a clevis fastener. We anchored the rocket net into the ground at five locations using anchor-ropes, bungie cord, and rebar. We propelled the rocket net over lekking lesser prairie-chickens by means of a closed-circuit electrical discharge disrupted by a Handi-Blaster HB10 blasting machine (Orica Explosives Technology, Melbourne, Victoria, Australia). We used rocket nets to target female lesser prairie-chickens, and we immediately removed individuals (including nontarget males) from the rocket net after declaring the area safe from explosive materials (within ≤30 s).

We used 7.5 m × 7.5 m drop nets (Wildlife Capture Services LLC, Flagstaff, AZ) with 7.5 cm × 7.0 cm diamond-shaped mesh attached to adjustable corner-poles (up to 2.4 m) to capture lesser prairie-chickens (Figure 3). Each net had four, 4-m corner poles, and one, 6-m center pole attached to the net via electromagnets. We used one, 12-V battery and corner-pole mounted wireless receiver to hold the electromagnetics on the pole in suspension. When a bird was within the capture area under the net, we used a wireless remote to disrupt the electrical current between the battery and magnets, subsequently dropping the net. Drop nets were used to target females, and we immediately removed captured birds (including nontarget males) from the drop net when we targeted females.

We determined sex by pinnae length, presence of eye combs, and other plumage characteristics upon capture (Copelin 1963). Subadults were identified by white spotting within 2.5 cm of the tip of the 9th and 10th primaries, whereas the absence of white spotting indicated an adult bird (Copelin 1963). We banded each captured bird with a New Mexico Game and Fish–issued or Texas Parks and Wildlife Department–issued uniquely numbered, aluminum leg band. We equipped each female and a sample of males with a 9-g necklace-style radiotransmitter (American Wildlife Enterprises, Monticello, FL) equipped with an 8-h mortality sensor, and then released each individual at the capture site. This research was part of a large-scale study aimed at assessing reproductive ecology of the species, so we opportunistically radiotagged males (subadults and adults) based on the number of radiotransmitters that remained after females stopped attending leks. We radiotagged every male (regardless of age class) that we captured post–female lek attendance, to deploy the remaining radiotransmitters and minimize the number of trapping days on leks. Individuals were not chosen at random because we opportunistically radiotagged captured lesser prairie-chickens. All methods were approved under Texas Tech University Institutional Animal Care and Use Protocol 1052-08.

We thoroughly inspected each captured individual for injuries prior to release and documented each injury with a photograph (Figure 4). We classified injuries into five categories: scrapes (Figure 4a, 4b), cuts (Figure 4c), feather damage, bone damage, and miscellaneous injuries. We defined scrapes as abrasions to the upper layer of the epidermis. Feather loss and minor abrasions were common with scrapes and these criteria helped us assess differences between scrapes and cuts (Figure 4a vs. Figure 4c). We defined cuts as a jagged tear or wound to the epidermis. Feather damage was the dislocation, removal, or fracture to the shaft of primary, secondary, and tertiary flight feathers, pinnae, or rectrices. Contour feather loss was common on the torso and backs of lesser prairie-chickens because of lek behavior, so we limited our definition of feather damage to flight feathers only. We defined bone damage as an injury that caused a visible break or dislocation to a bone. We defined miscellaneous injuries as those that we were unable to classify into the aforementioned categories. We considered an injury to be caused by traps based on the presence of fresh blood, wounds, partially lacerated feathers, visible bones, dislocated appendages, or other indicators of a recent injury.

We used PROC FREQ in SAS 9.3 (SAS Institute, Cary, NC) to calculate the total number of individuals captured with each trap type, the number and injuries caused by each trap type, location of the injury, and type and location of injury for each sex (female, male). We used a log-linear model to assess whether lesser prairie-chicken injuries (scrapes, cuts, feather damage, bone damage, and miscellaneous injuries) occurred in equal proportion among interactions of trap type (walk-in funnel, rocket nets, drop nets), and sex (female, male) using PROC CATMOD in SAS 9.3. We used PROC CATMOD to fit linear models to functions of response frequencies (i.e., numerical transformation of cells in a categorical contingency table) and estimate parameters using maximum likelihood estimators. We replaced zero values in the data set (e.g., we observed no cuts on lesser prairie-chickens captured with rocket nets) with 0.01 to ensure that the sampling zeros were not treated as structural zeros (Bishop et al. 1975). Sampling zeros can have nonzero expected values and be calculated via maximum likelihood estimators, whereas structural zeros always have an expected value of zero and do not produce chi-squared test statistics or P-values. Injuries occurred disproportionately among interactions of trap type, injury type, and sex if the chi-squared test statistic for the analysis of response functions was significant (P < 0.05). If the three-way interaction was nonsignificant (P > 0.05), we assessed whether injuries occurred in equal proportion across 2-way interactions of independent variables as well as single factors.

We used a hand-held 3-element Yagi antenna and an Advanced Telemetry Systems R-2000 receiver (Advanced Telemetry Systems, Ashanti, MN) to obtain telemetry relocations. When we received a mortality signal from a transmitter, we used homing (White and Garrott 1990) to locate the bird or radio, assess evidence for mortality (as opposed to a bird slipping the radio off), and identify the cause of death. The earliest date we captured an individual in one of our traps was on 17 March and the latest was 18 May.

We used the nest survival model (Dinsmore et al. 2002) in Program MARK (White and Burnham 1999) to assess postcapture survival rates of injured lesser prairie-chickens using the logit link function. First, we developed five a priori models to assess postcapture survival probabilities among trap type (Table 1). We grouped captured and marked individuals by trap type and assessed whether postcapture survival was constant or varied among trap type. Second, we assessed whether postcapture survival probabilities differed between injured and uninjured birds. For this assessment, we randomly selected the same number of uninjured individuals within our data set as injured individuals to maintain a balanced study design. We did not develop any a priori models for this stage of our analysis; rather, we assessed whether survival was either 1) constant, or 2) varied between groups (injured, uninjured). We monitored injured and uninjured lesser prairie-chickens for 7–65 d postcapture, resulting in total exposure period for all birds of 127 d (17 March–21 July) for our study.

We used Akaike’s Information Criterion for small sample sizes (AIC_c), ΔAIC_c values, and model probability (w_i) to select the best approximating model for each analysis (Burnham and Anderson 2002). We considered any model that had ΔAIC_c ≤ 2 as competing and calculated variable weights to assess the relative variable importance by summing the Akaike weights of models that included the variable of interest (Burnham and Anderson 2002). We model-averaged the parameter estimates across the candidate set of models in instances of multimodel support (no single model with w_i > 0.90) and used the delta method to calculate variance and standard errors for model-averaged estimates.

We captured 401 individuals (127 females, 274 males) from 2006 to 2012. We captured 217 (54%) in walk-in funnel traps, 40 (10%) in rocket nets, and 144 (36%) in drop nets. Recaptures were excluded from this analysis. Sixty-nine (17%) individuals exhibited evidence of a trap injury and we observed no direct mortalities due to trapping. Overall, we witnessed 52 (75%) injuries to birds captured using walk-in funnel traps, 2 (3%) injuries to birds captured using rocket nets, and 15 (22%) injuries to birds captured with drop nets. Injuries occurred disproportionately among trap type, injury type, and sex interaction (χ²₈  =  35.39, P < 0.001). The predominant injuries were superficial cuts to the wings (19%) and head (16%) of male lesser prairie-chickens captured in walk-in funnel traps (χ²₈  =  31.06, P < 0.001). We witnessed 15 injuries using drop nets and 47% (7 of 15) of injuries were broken feathers on females. Rocket net injuries were uncommon (n  =  2) but serious (e.g., broken toe bones). We observed 40 cuts, 11 scrapes, 8 instances of feather damage, 6 broken toe bones, and 4 miscellaneous injuries. Multiple injuries to individual birds were more common in walk-in funnel traps (n  =  14) compared with drop or rocket nets (n  =  0) and there were more injuries females compared with males (χ²₁  =  9.71, P < 0.001).

Trap type assessment. We radiotagged 35 of 69 injured (51%) birds (13 drop net, 3 rocket net, 19 walk-in funnel traps). The top four candidate models had ΔAIC_c ≤ 2.00 and model S. (survival was constant) had the greatest model probability (AIC_wi  =  0.42). Models that incorporated variation in survival probability across trap type received less support (models 2–5; Table 1). Daily survival probability calculated using model-averaged parameters was 0.99 (SE  =  0.004, 95% CI  =  0.97–1.00) for birds injured in walk-in funnel traps, 0.99 (SE  =  0.006, 95% CI  =  0.95–1.00) for birds injured in rocket nets, and 1.00 (SE  =  0.00, 95% CI  =  1.00–1.00) for birds injured in drop nets. The probability of an injured bird captured in a walk-in funnel trap or rocket net surviving 7, 14, 21, 35, and 65 d postcapture was 0.93 (SE  =  0.003, 95% CI  =  0.86–1.00), 0.87 (SE  =  0.003, 95% CI  =  0.75–1.00), 0.81 (SE  =  0.003, 95% CI  =  0.65–1.00), 0.70 (SE  =  0.003, 95% CI  =  0.49–1.00), and 0.52 (SE  =  0.003, 95% CI  =  0.27–1.00), respectively. The probability of an injured bird captured in a drop net surviving to 65 d postcapture was 1.00 (SE  =  0.00, 95% CI  =  1.00–1.00).

Injured vs. uninjured assessment. We selected 35 uninjured individuals from our data set of uninjured individuals that received radiotransmitters (N  =  110). The model that incorporated constant survival between injured and uninjured birds received more support (AIC_wi  =  0.71) compared with the model that incorporated variation in daily survival rates (AIC_wi  =  0.28). We model-averaged between our two candidate models and the daily survival probability for injured birds was 0.99 (SE  = 0.0003; 95% CI  =  0.989–997) and 0.99 for uninjured birds (SE  =  0.009; 95% CI  =  0.994–0.998).The probability of an injured and uninjured bird surviving 7, 14, 21, 35, and 65 d postcapture was 0.93 (SE  =  0.003, 95% CI  =  0.86–1.00), 0.87 (SE  =  0.003, 95% CI  =  0.75–1.00), 0.81 (SE  =  0.003, 95% CI  =  0.65–1.00), 0.70 (SE  =  0.003, 95% CI  =  0.49–1.00), and 0.52 (SE  =  0.003, 95% CI  =  0.27–1.00), respectively.

Our results indicated injuries were minor and postcapture survival (7–65 d postcapture) did not vary across trapping technique or injured and uninjured lesser prairie-chickens. We attribute our findings to the passive nature of the walk-in funnel traps and the mesh size of the drop nets. Lesser prairie-chickens on the breeding ground harassed individuals trapped in walk-in funnel traps, which ultimately caused cuts and scrapes to the wings and head, especially to males. Our findings are explained by the territorial nature of lekking males (Haukos et al. 1989; Behney et al. 2012). Behney et al. (2012) found ≤2 males were responsible for ≥80% of all copulations on two leks in Texas, and these males were less idle and more active during courtship and typically had the smallest territories near the center of the lek. Males within the geographic center of the lek were more likely to be recaptured when we repeatedly placed traps near the center of activity on the lek. Walk-in funnel traps are an efficient way to capture lesser prairie-chickens, particularly territorial males in the spring (Haukos et al. 1989; Salter and Robel 2000); and our results suggest that, in context of stress and potential for injury, this method is ethically justifiable if birds are immediately removed from the traps.

Injuries to females could be more detrimental to the population by potentially impairing nesting and brood activities (Hagen et al. 2009), and injuries to females were more common than injuries to males, but were rarely severe. We observed multiple instances of broken flight feathers on females when we used drop nets, but we did not observe any instances where captured individuals were unable to fly or navigate upon release. We removed from the drop nets multiple females that had their wings tangled within the mesh. This occurred as a result of females flushing when the nest was dropped, and likely increased the number of broken flight feathers we observed. Injuries were less common, but more severe, with use of rocket nets, which had much smaller mesh size. Therefore, we suspect using drop nets with mesh ≤5 cm × 5 cm and with trained staff will reduce the chance of an individual being encased within the mesh and damaging feathers or suffering a life-threatening injury. The drop nets we used for this assessment were designed to capture large mammals, so a smaller mesh size with the default manufactured magnets are sufficient to trap lesser prairie-chickens if the birds are removed immediately from under the net.

This study was a side objective of a larger scale project with the ultimate goal of assessing the reproductive ecology of the species, and our low sample size among trap types for the survival analysis was the result of our targeting females and indiscriminately capturing and radiotagging males when females stopped attending leks. We did not include injury type or injury location as covariates in this analysis to maintain sample size across trapping groups. Nevertheless, our results indicated that postcapture survival was similar among walk-in funnel traps, rocket nets, and drop nets and daily survival probabilities for injured and uninjured birds was ≥99%. Postcapture survival rates decreased as we monitored birds into the breeding season, but female mortality is typically highest during incubation and brood rearing (Hagen et al. 2007; Wolfe et al. 2007) and radiotagged males in our research program had their lowest survival probabilities in late summer and not within 30–45 d postlekking (Grisham and Boal, 2015). Our combined results suggest the survival probabilities reported herein are more likely related to other biological mechanisms (predation and drought; Wolfe et al. 2007; Hagen et al. 2009; Grisham and Boal, 2015) and not directly linked to capture or ensuing injuries.

Hagen et al. (2006) evaluated the impact of very high frequency, necklace-style transmitters on survival rates of captured male lesser prairie-chickens. Survival estimates of radiomarked males in Hagen et al. (2006) were equal to those of banded-only birds (S_radio  =  0.670, SE  =  0.061; S_band  = 0.671, SE  =  0.075). The results provided in this study, combined with data provided by Hagen et al. (2006), suggest researchers can trap and handle this threatened species without measureable negative effects to individuals if birds are immediately removed from traps, the number of consecutive trapping days on an individual lek are minimized (7–10 d), the position and location of the traps on the lek of capture is rotated every 2–4 d, and researchers use active trapping techniques to target individuals. These methods will allow achievement of sex- or age-specific research and conservation goals while adhering to the spirit of the Animal Welfare Act and existing guidelines and will aid in the formulation of biological opinions by regulatory agencies and institutional animal care and use committees.

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

Reference S1. Copelin FF 1963. The lesser prairie-chicken in Oklahoma. Oklahoma City, Oklahoma: Oklahoma Department of Wildlife Technical Bulletin 6.

Found at DOI: 10.3996/032015-JFWM-022.s1 (2090 KB PDF).

Reference S2. Zavaleta JC. 2012. Community response to use of prescribed grazing and herbicide for restoration of sand shinnery oak grasslands. Master’s thesis. Lubbock: Texas Tech University.

Found at DOI: 10.3996/032015-JFWM-022.s2 (4780 KB PDF).

We thank a multitude of private land owners in Texas for private land access. We thank A. Behney, N. Pirius, A. Wood, and a plethora of field technicians for field data collection and photos. We thank S. Fritts and several anonymous reviewers for reviewing earlier drafts of this manuscript.

Financial and logistical support was provided by The Grasslans Charitable Foundation, Texas Tech Department of Natural Resources Management, U.S. Geological Survey, Texas Parks and Wildlife Department, Great Plains Landscape Conservation Cooperative, and The Nature Conservancy.

This manuscript is Texas Tech College of Agricultural Sciences and Natural Resources T-9-1269.

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