Minor mist-net injuries do not affect apparent annual survival in Cliff Swallows (Petrochelidon pyrrhonota)

Abstract

Researchers commonly use mist nets to study avian populations. Although studies have concluded that the risk of mortality from netting is low, few have examined the effect of minor injuries on long term survival. We used data on Cliff Swallows (Petrochelidon pyrrhonota) consisting of 206 birds with leg injuries and 206 non-injured birds caught at the same time during mark-recapture in southwestern Nebraska, USA, 1986–2013. The injuries were primarily cuts from mist-net tangling around the legs and toes and often produced a small amount of bleeding. Mark-recapture analysis showed that the probability of apparent annual survival of injured birds was 0.52, which was similar to that of 0.51 for non-injured birds. Our results suggest that small cuts to the legs and toes of Cliff Swallows during mist-netting do not result in long-term harm and support the widely held assumption that mist-netting is in general safe even when minor injuries occur.

INTRODUCTION

Researchers commonly use mist nets to study wild birds in the field (Stamm et al. 1960; Karr 1981; Ralph et al. 1993) for studies on demography, nest site fidelity, annual reproductive success, survival, and life history characteristics (Peach et al. 1996; DeSante et al. 1999; Dunn and Ralph 2004). While birds will learn to avoid recapture (Roche et al. 2013), the common understanding in the scientific community is that mist-netting is a safe practice if done by experienced handlers (Ralph et al. 2004; Wilson and McMahon 2006; Spotswood et al. 2012) and does not affect reproductive success or survival (Jennings et al. 2009).

Most scientists assume that the benefits of research outweigh the costs to individual animals, in part because mortality during mist-netting is generally low (O’Brien et al. 2016; Clewly et al. 2018). Documented mortality during mist-netting can fluctuate based on species, season, age of the bird, time spent in the net, height of the net, predation, incidents of severe tangling, and habitat (Recher et al. 1985; Melo et al. 2018; de Moura et al. 2023), with most known mortality occurring at the time of mist-netting. We know far less about the long-term consequences for a bird that experiences a small injury (e.g., a minor cut) during mist-netting (Spotswood et al. 2012).

Here, we use 26 years of data from a mark-recapture study to analyze the effect of injuries to the legs during mist-netting on the annual survival of Cliff Swallows (Petrochelidon pyrrhonota) in western Nebraska. These injuries typically involved some slight blood loss. Cliff Swallows were reported to suffer long-term survival disadvantages of blood collection for research purposes (Brown and Brown 2009). Our objective is to investigate whether more minor, unintentional blood loss during mist-netting has a similar effect.

METHODS

Study system. –

Cliff Swallows are migratory aerial insectivores that breed in North America during the summer months (May–August) and spend the winter in southern South America. Cliff Swallows have been studied since 1982 near the Cedar Point Biological Station (41°13’N, 101°39’W) in Keith County, southwestern Nebraska, USA, along the North Platte and South Platte rivers. The study area also includes colony sites located in Deuel, Garden, and Lincoln counties. Cliff Swallows construct gourd-shaped mud nests in dense breeding colonies. Although some birds still nest under natural cliffs, most Cliff Swallows in our study area use the sides of bridges or the walls of road culverts (Brown et al. 2013a). The study area contains ~220 colony sites, and about two-thirds are used annually. Colony size varies between 2 to 6,000 nests (mean ± SE = 393 ± 15; n = 1812 colonies), with some birds nesting solitarily. The study site is described in detail by Brown and Brown (1996) and Brown et al. (2013a).

Mark-recapture. –

Cliff Swallows were routinely mist-netted at 27–40 colonies beginning in 1986 and continuing through the 2013 field season. We mist netted (for this study) at 107 total colonies throughout the breeding season, starting in early May and usually continuing through mid-July (occasionally into late July or early August). We typically netted at a colony site multiple times per year (Brown 1998; Brown and Brown 1996; Roche et al. 2013; Brown et al. 2016). Mist nets were placed at the entrance to culverts, parallel with nests, or hung over the side of a bridge if the colony was over water (Roche et al. 2013). Nets were generally level with the nests at a given site. We routinely taped the loops around the pole to hold them in place and maintained more slack in the net than in most other studies. At most sites, nets were continually monitored and Cliff Swallows removed from mist nets within 20 minutes of initial capture. All birds were banded with a U.S. Geological Survey numbered aluminum leg band upon first capture, sexed by the presence of a cloacal protuberance or brood patch, and weighed with a Pesola scale by placing the bird in a cloth bag. Some birds had additional procedures done, such as body measurements taken or blood collected. During the course of the study, we assembled about 407,000 total mist-net captures of Cliff Swallows.

Mist net injury. –

During mist-netting, sometimes birds incurred an injury in the form of a cut to the legs or toes that produced a small amount of bleeding. These injuries usually occurred from the mist net tightly tangling and getting caught around the scutes and reticulas on the bird’s legs and toes. If the bird struggled in the net while tangled, this could cause cuts to the legs or toes. We routinely banded the right leg, but if an unbanded bird sustained the injury on the right leg, we banded its left leg. Rare injuries happening during banding after removal from a net (e.g., while closing a band, pinching a leg and drawing blood) were also included in the study and are considered part of the mist-netting process. Our analysis only examines injuries to the legs and feet and does not include injuries to any other parts of the body.

We recorded 206 mist-net injuries at 107 colonies, 1986–2013. For a comparison to non-injured birds captured under the same conditions at the same time and handled the same way, we used the next bird processed after each injured bird. This yielded a dataset of 412 birds for survival analysis, 206 injured and 206 non-injured. The two groups were balanced in regard to colony site, colony size, date, time handled, and capture method.

Statistical estimation of apparent survival. –

Survival and recapture models were fit to the 412-bird dataset using Program MARK (White and Burnham 1999), and from these, annual survival (ϕ) and recapture (p) parameters were estimated (Lebreton et al. 1992). Some models used a group designation (g) based on whether a bird was injured or not, whereas others treated injured and non-injured birds as the same. Time-dependence (t), in which a parameter was estimated for each year differently, was included in some models. Because of the relatively small size of our dataset and the absence of strong sex effects on mark-recapture in Cliff Swallows in earlier analyses (Brown et al. 2013b, Roche et al. 2013), we did not include sex in the survival and recapture models. The data were also insufficient to model age effects, as age was unknown for most of the 412 birds (beyond knowing all were at least one-year old).

We compared models using the Quasi-Akaike’s Information Criterion corrected for small sample size (QAICc; Burnham and Anderson 2002). QAICc was used because we applied a slight correction using the overdispersion parameter (ĉ) as estimated by MARK for our dataset (1.037), although our data still closely met the variance assumptions inherent in the binomial distribution used in mark-recapture analysis (Burnham and Anderson 2002, Pradel et al. 2005). Annual survival and recapture parameters were derived from the 4 models with the lowest QAICc values (Table 1). We model-averaged the estimated parameters based on those models’ QAICc weights (Dormann et al. 2018). As in most mist-netting studies done in a finite area, we were estimating apparent survival only, and we did not account for birds dispersing outside of our study area.

Table 1.

Models fit to recapture data for 412 individuals to investigate the effect of minor mist-netting injuries on Cliff Swallow annual survival and likelihood of recapture. Lowest QAICc = 805.4. Model notation: (c), no grouping by injury status; (g), two groups of injured and non-injured birds; (t), recapture time (year)-dependent.

Model	ΔQAICc	QAICc weight	No. estimable parameters	Description
(1) ϕ (c), p(c)	0.00	0.451	2	Survival was constant and recapture was constant among years and neither included a group effect of injured or not injured.
(2) ϕ (c), p(g)	1.44	0.220	3	Same as model (1) except recapture included a group effect of injured versus not injured.
(3) ϕ (g), p(c)	1.83	0.181	3	Same as model (1) except survival included a group effect of injured versus not injured.
(4) ϕ (g), p(g)	2.23	0.148	4	Same as model (1) except both survival and recapture included a group effect of injured versus not-injured.
(5) ϕ (c), p(t)	31.27	0	26	Same as model (1) except recapture probability varied by year.
(6) ϕ (g), p(t)	33.29	0	27	Same as model (5) except survival included a group effect of injured versus not injured.

RESULTS

The top model (lowest QAICc) was one that considered both survival and recapture to be constant across years and treated injured and non-injured birds the same (Table 1). Models that considered a group (injury) effect on either survival or recapture alone were within 2.0 QAICc, and one with a group effect on both parameters was slightly greater than 2.0 in QAICc (Table 1). Models with time-dependent recapture were well below all others in fit (Table 1), probably due to data sparseness (only 412 birds across 26 years).

Model-averaged annual survival probability (ϕ ± SE) from the 4 models with the lowest QAICc (Table 1) was 0.52 (± 0.036) for injured birds and 0.51 (± 0.036) for non-injured birds. Model-averaged annual recapture probability (p ± SE) was 0.43 (± 0.051) for injured birds and 0.46 (± 0.057) for non-injured birds.

DISCUSSION

With the top model being one without an injury effect (Table 1) and the numerical similarity of the survival and recapture estimates, there is little to no evidence that minor injuries to the legs and toes of Cliff Swallows during mist-netting have any effect on annual survival, as compared to a “control” group of birds caught and handled at the same time as the injured birds. We also note that these minor injuries were quite rare in our study: only about 0.05% of the ~407,000 Cliff Swallow captures resulted in a toe or leg injury that drew blood. That leg injuries may be survivable was also suggested by our recapturing an individual about two months after a severe injury to its leg in a banding accident, resulting in the complete loss of the left foot.

The blood loss to birds from toe and leg injuries was not quantified but was likely much less than even the category of a “small” amount of blood (< 0.70 μL) taken in the study of how blood collection for research purposes affected annual survival in Cliff Swallows (Brown and Brown 2009). That analysis found negative effects of any amount of blood (≥ 0.70 μL) collected from the brachial vein. In aerial birds such as swallows, injury to the brachial vein may have a greater effect than to the legs or toes, in addition to the much smaller amount of blood lost during leg and toe injuries. Our results support the widely accepted assumption that mist-netting is a generally safe procedure that, when done appropriately, does not result in long-term harm to birds. However, we acknowledge that injuries to the legs or toes could be more important in species that spend much of their time on the ground or perched, or use their toes to hold food.