A review of wood warbler (Parulidae) predation of vertebrates and descriptions of three new observations

Wood warblers (Parulidae) eat insects, spiders and other small arthropods, fruit, and nectar, but have also been doc-umented preying on vertebrates. We conducted a literature review to determine which species of Parulidae have been observed capturing or consuming vertebrate species, such as small lizards, amphibians, and fish. We also include information and discussion of three previously unpublished observations: 1) a Black-and-white Warbler ( Mniotilta varia ) photographed with a bark anole ( Anolis distichus ) in its mouth on San Salvador Island, The Bahamas, 2) a Black-and-white Warbler that had small lizard bones ( Anolis spp.) in its stomach contents in Jamaica, and 3) a Yellow-throated Warbler ( Setophaga dominica ) photographed preying on a brown anole ( Anolis sagrei ) in Florida, USA. The latter observation was particularly interesting as the brown anole shed its tail and briefly escaped before being recaptured by the warbler. Including these three unpublished events, our literature review found 24 separate records of 12 Parulidae species preying on vertebrates. The majority of predation records were of lizards (58%), with an additional 25% predation records of amphibians and 17% of fish. We hypothesize that vertebrate predation by wood warblers is opportunistic and potentially limited by the difficulty in capturing, handling, and consuming larger vertebrate prey. There is also non-mutually exclusive evidence to suggest birds exhibit plasticity in prey items, partially switching to larger, more difficult-to-handle vertebrate prey during taxing periods when invertebrate prey are less available. Our study synthesizes previous work and can assist future research in better understanding trophic interactions, foraging efficiency and prey size, evolution, and the natural history of both warblers and their vertebrate prey.

Despite their small size, wood warblers (Parulidae) have also been observed infrequently preying upon and consuming vertebrates (e.g., Eaton 1953), but to our knowledge, no study has previously synthesized literature on this topic. Wood warblers are a family of approximately 120 species that reside in North and South America and include many Nearctic-Neotropical migratory species that winter in the tropics and subtropics, as well as non-migratory resident species that live in the tropics yearround (Lovette et al. 2010, Billerman et al. 2020. Compiling records of wood warblers preying on vertebrates is valuable because much of the literature on the topic is scattered in historic, difficult-to-find references (e.g., Wetmore 1916), and some re-searchers have incorrectly stated that Parulidae have not been previously observed consuming vertebrates (e.g., White and Cove 2016). Moreover, determining which species of parulids have been observed preying on vertebrates, along with the specific prey taken, is useful to better understand food webs, trophic links, and the natural history and evolution of both warblers and the vertebrate prey (Winemiller and Polis 1996, Blomberg and Shine 2000, Devenish-Nelson et al. 2019. We conducted a literature search to create a compilation of all wood warblers that have been observed capturing or consuming vertebrates. We also provide new documentation of two Blackand-white Warblers (Mniotilta varia) and a Yellow-throated Warbler (Setophaga dominica) preying on anoles. Black-and-white Warblers and Yellow-throated Warblers are primarily insectivorous, although Black-and-white Warblers occasionally supplement their diet with fruit (Kricher 2020, McKay andHall 2020).
Mots clés autotomie caudale, comportement de prédation, efficacité de la recherche de nourriture, espèce non nicheuse, régime alimentaire des oiseaux, Squamata Although we conducted our search to the best of our ability, we note that we only searched for articles in English and Spanish (the latter with limited ability), and there could be more unpublished or published observations in the subtropics and tropics that we were unaware of (Konno et al. 2020).
For each publication or observation, we recorded the bird species, vertebrate prey, date, and location. Some references did not contain a date or the exact species of vertebrate prey, and we noted these omissions in our records. We included records from studies that examined bird stomach contents, as well as from observations of birds capturing or consuming vertebrates in the field. For three unpublished observations, we included detailed accounts of these records.

Literature Review
Out of approximately 120 species in Parulidae, we found evidence of 12 species preying on vertebrates (Table 1), including both tropical residents and Nearctic-Neotropical migratory species. A large proportion of the species (75%) were migrants that winter in the Caribbean, Central America, or South America. The compiled list includes wood warbler species that forage on the ground (e.g., Ovenbird [Seiurus aurocapilla], Northern Waterthrush [Parkesia noveboracensis]) as well as species that usually prey on insects in shrubs or trees (e.g., Black-and-white Warbler).
Lizards, including anoles, were some of the most depredated taxa, but small amphibians and fish were also captured by warblers. Of the predation events with known dates, most were during the non-breeding season in either the fall, winter, or early spring. A few observations were during the breeding season, notably for Northern Waterthrush and Louisiana Waterthrush (Parkesia motacilla). Most records (83%) were in the southeastern United States and in the Caribbean. Over half (63%) were from records of prey found in stomach contents, while the remainder were from field observations. In addition to our Yellow-throated Warbler observation (see below), two other studies noted warblers beating their prey on branches, presumably to maim or kill their prey.

New Observations
Black-and-white Warbler, The Bahamas.-One of the new Black-and-white Warbler records was an observation on the grounds of the Gerace Research Centre (GRC; 24°07'03"N, 74°27'52"W), on San Salvador Island, The Bahamas (see Akresh and King 2015 for the GRC location and a map of the island). Black-and-white Warblers are an uncommon to common wintering migrant that use a variety of habitats on San Salvador (Sordahl 1996, Murphy et al. 2001. On the morning of 6 March 2020, MEA, SL, and JMCS were observing and photographing birds in primarily non-native vegetation at the GRC (see Jones et al. 2013, Akresh et al. 2019a for detailed vegetation composition and structure). While birding, JMCS photographed a female Black-and-white Warbler that had a bark anole (~4 cm total length) in its bill at 0808 (Fig. 1). We could not determine if the anole was alive. The bird was perched 5-6 m above ground in a non-native almond tree (Terminalia catappa). After photographing the bird, we unfortunately did not follow the warbler to determine if the bird consumed its prey.
Black-and-white Warbler, Jamaica.-The second new record of a Black-and-white Warbler occurred at Copse Hill on the Kew Park Estate near Bethel Town in Westmoreland Parish, Jamaica (18°15'00"N, 77°56'24"W), as part of a study quantifying dietary overlap and specialization in wintering warbler communities . The study site was in wet limestone forest habitat, in one of the few old-growth wet limestone forest fragments in the region. Black-and-white Warblers are consistently present at this site in low numbers throughout the winter (CMK unpubl. data). During the Kent and Sherry (2020) study, CMK captured a Second Year (< 1 year old) female Black-andwhite Warbler (Pyle 1997) in a mist net on 25 February 2017 at 0957 and administered an emetic to induce regurgitation (Johnson et al. 2002). Stomach contents were placed in 70% EtOH and were later examined in the lab using an Olympus SZH binocular dissecting microscope at 75-650× magnification.
Alongside typical arthropod fragments, the stomach contents of this Black-and-white Warbler contained the bones and parts of a small lizard (Fig. 2), including seven claws, several vertebrae, fragments of the pelvic and shoulder girdle, bones of the lower jaw, and teeth. The lizard was identified as an Anolis species based on a tricuspid tooth (J. Daza pers. comm.). In addition to the lizard, the individual had been foraging on ants (n = 46, including 11 Camponotus capperi and 34 Pheidole jamaicensis), bark-lice (n = 7), caterpillars (n = 4), spiders (n = 3), Orthoptera (n = 1), and Auchenorrhyncha (n = 1). During the entire Kent and Sherry (2020)  was birding and noticed a Yellow-throated Warbler 1-1.5 m above the ground on horizontal branches, within mangrove and buttonwood (Conocarpus erectus) vegetation. Yellow-throated Warblers are uncommon in southwestern Florida, and LS proceeded to photograph the warbler between 1328 and 1330. LS took 40 photos in total, with the best photos of the predation sequence presented here (Fig. 3). The warbler appeared to be a Second Year individual (< 1 year old), and possibly a male, based on plumage characteristics and molt limits (Pyle 1997). The warbler was first seen without any prey, and then grabbed a small brown anole by the head. The warbler then switched its grip to holding the anole by the tail, and the anole was able to briefly escape. The warbler looked around and was able to recapture the anole, now holding it by the legs. In this latter photo and in subsequent photos, the anole's tail was shorter, and the end was clearly broken off (caudal autotomy), which occurred when the anole briefly escaped (Fig. 3).
The warbler then proceeded to process its prey by biting on the head of the anole, as well as holding onto the legs while bashing Table 1. Literature records of Parulidae preying on vertebrates. For each bird species, vertebrate prey species, date and location of observation, record type (field observation or prey found in bird stomach contents), interesting additional information, and the reference is given.  the anole on the branch. The behavior appeared to be very similar to prey-handling behavior of Parulidae when processing large invertebrate prey, such as large caterpillars (e.g., Nolan 1978:491). The last photos were taken with the anole in the warbler's bill. We are unsure if the warbler consumed its prey or how this may have occurred, as the warbler proceeded to move out of view.

Discussion
Our study presents a comprehensive list of Parulidae species that have been observed preying upon or consuming vertebrates. We found both historic, obscure records (e.g., Wayne 1910) as well as more recent observations in journal articles (e.g., Davis and Komar 2003) and social media (e.g., Twitter). We also describe the first observation of a Yellow-throated Warbler preying on vertebrates, and the fourth and fifth vertebrate prey records for Black-and-white Warblers. Our compilation is important for better understanding the complexity of food webs, trophic interactions, and predator-prey dynamics (Polis 1991, Winemiller andPolis 1996). Simplistic food webs that leave out uncommon predation events, such as warblers consuming vertebrates, may not embody complete, community-wide relationships (Goldwas-ser and Roughgarden 1993, Poulin et al. 2001). Additionally, acknowledging warbler predation on vertebrates may be important for studies using stable isotopes to determine dietary preferences (Kelly 2000).
Warblers consuming vertebrates may be opportunistic foraging events, including our observations in The Bahamas, Jamaica, and Florida. In The Bahamas, both Black-and-white Warblers and bark anoles often forage along tree trunks (Hillbrand et al. 2011, Kricher 2020, thus it is possible that the bird we observed had encountered the bark anole on the trunk of the almond tree while foraging for insects. Given the shared habitat, we suspect the anole in The Bahamas was alive and not found dead on the ground when captured. Yellow-throated Warblers similarly often forage along tree trunks in their non-breeding grounds (LS and MEA pers. obs.), and the observed Yellow-throated Warbler may also have opportunistically found the anole in Florida. We did not follow the photographed Black-and-white Warbler or Yellow-throated Warbler to determine if they consumed their respective anoles. It is possible the birds removed the anoles from the trees to reduce interspecific competition for prey (i.e., insects). However, this  hypothesis seems extremely unlikely, as the records of vertebrates found in stomach contents document the consumption of lizards, including our new record in Jamaica (Table 1, Fig. 2), and we observed the Yellow-throated Warbler exhibiting prey-handling behavior.
Opportunistic predation of vertebrates may rarely occur because of the difficulty for warblers to successfully find, capture, and consume vertebrate prey (Pérez-Rivera 1997). Small fish and amphibians may only be accessible prey for warblers residing in wetter habitats, including areas with small streams (Mulvihill et al. 2008). Furthermore, the most common lizard prey, anoles, are generally fast-moving and can elude avian predation by moving up, down, or to the opposite side of the trunk (Schneider et al. 2000, Hillbrand et al. 2011, Antley et al. 2016). However, anoles may be more active in the dry season while they search for invertebrate prey (Wunderle 1981), and thus may be more likely to be captured by predators during this season. Warblers may also have difficulty in handling and killing relatively large vertebrate prey, as we witnessed with the Yellow-throated Warbler. Indeed, other studies have also observed warblers having difficulty handling and swallowing their vertebrate prey (Davis andKomar 2003, Sykes et al. 2007). One apparent option is to swallow the entire vertebrate head-first (macaulaylibrary.org/ asset/47723181), while another method involves tearing up the prey and ingesting it incrementally (Davis and Komar 2003).
An alternative, non-mutually exclusive hypothesis is that these predation events represent behavioral plasticity in warblers' normal foraging tactics and prey items in response to underlying environmental conditions (Aborn and Froehlich 1995). Both of our new Black-and-white Warbler observations occurred during the late winter dry season in the Caribbean (Sealey 2006, Akresh et al. 2019a. The observation in The Bahamas occurred during a winter that was especially dry, with very little precipitation in January and February (T. Dexter pers. comm.), and vegetation on the island appeared to be particularly water-stressed at the time of the observation, even in comparison to previous dry seasons. Likewise, during the period of the Black-and-white Warbler capture in Jamaica, an arthropod sampling study documented a scarcity of high-value arthropod prey at the field site (Razeng andWatson 2015, Kent et al. 2019). Indeed, the Blackand-white Warbler's stomach sample was dominated by relatively low-value ants. Given the likely shortage of typical avian food resources, such as insects and fruit, during the time of our observations at both sites, the Black-and-white Warblers could have partially shifted to feeding on alternative prey items, such as lizards, to compensate (Poulin et al. 2001, Latta and Faaborg 2002, Wunderle et al. 2014. This latter hypothesis is consistent with previous studies that have shown wintering, migratory warblers to be especially stressed during late winter in very dry years and thus have difficulty maintaining overwinter body condition and adequately preparing fat and muscle stores for spring migration (Studds andMarra 2007, Akresh et al. 2019a). Additionally, events and processes on the wintering grounds can subsequently influence birds both during migration and on the breeding grounds (e.g., carry-over effects; Paxton andMoore 2015, Akresh et al. 2019b). The behavioral plasticity hypothesis is also consistent with these migratory warblers' life history and their capacity to exploit different food resources depending on availability (Bell 2011).
Observations of other small passerines preying on vertebrates have also occurred during stressful environmental conditions or taxing periods during the avian life cycle, further supporting the behavioral plasticity hypothesis. For instance, a Prothonotary Warbler (Protonotaria citrea) consumed a lizard 2 days after a tropical storm (Davis and Komar 2003), and Red-eyed and Yellow-throated Vireos (Vireo olivaceus and V. flavifrons) were observed eating green anoles (Anolis carolinensis) during migration (Sykes et al. 2007). Mulvihill et al. (2008) observed that salamanders were fed to Louisiana Waterthrush nestlings more often in acidified, poorer-quality streams compared to circumneutral streams on the breeding grounds. Optimal foraging theory predicts that predators should try to maximize food intake and energy gain per unit of foraging time (Krebs 1978). Therefore, larger, more difficult, time-consuming prey such as vertebrates may only be efficient prey items for warblers during taxing periods or in environments with a lack of available invertebrate prey (Hasegawa 1990, Hirvonen andRanta 1996, Tremblay et al. 2005). Although a single anole has more weight and calories than individual invertebrate prey, it may have similar caloric content per gram compared to certain insect taxa (Vitt 1978, Ramos-Elorduy et al. 1997.
Many of the prey items recorded in our literature review were lizards and amphibians, which are both abundant in the non-breeding grounds of migratory Parulidae and small enough to be captured (Schoener and Schoener 1978, McLaughlin and Roughgarden 1989, Poulin et al. 2001, Rosamond et al. 2020. Moreover, lizard abundance may be increasing in these regions; regional and international trade has resulted in the human-mediated introduction of many new anole species in North America and the Caribbean (Losos 2009). For instance, the brown anole is Fig. 3. Sequence of photographs of the Yellow-throated Warbler (Setophaga dominica) handling a brown anole (Anolis sagrei) in southwestern Florida, USA, on 8 February 2010. Times given as hr:min:sec. a) Bird is without prey, 13:28:13, b) catches anole, 13:28:38, c) still has anole, 13:29:05, d) has prey by leg and tail, but anole begins to escape, 13:29:06, e) without the anole, 13:29:13, f) begins to recapture anole, 13:29:15, g) recaptures anole, 13:29:18, h) anole is immobile on branch, 13:29:20, i-l) further handling of anole, including bashing on branch, 13:29:37-13:29:59. Photos by Lillian Stokes. an introduced species now found throughout Florida, with population densities as high as 12,000 individuals/ha (Krysko et al. 2019). Interestingly, besides being a newly abundant potential prey item, introduced lizards could also impact and potentially reduce invertebrate prey density and availability for warblers (Schoener et al. 2002). Determining if warbler consumption of anoles increases in the future would improve our understanding of how trophic interactions are shifting in changing environments (David et al. 2017, Calizza et al. 2021.
Further research would help elucidate why only certain species of wood warblers have been observed preying on vertebrates, and if determinants such as bill size, foraging behavior, habitat type, and geographic range explain variation of vertebrate predation among warbler species (Poulin et al. 2001, Rosamond et al. 2020. Incidentally, wood warblers that have preyed on vertebrates appear to have relatively longer and larger bills (Davis andKomar 2003, Kricher 2020); based on the 80 stomach content samples of five species examined in Kent and Sherry (2020), only the longer-billed Black-and-white Warbler consumed a vertebrate. Further studies should also examine how avian predation, even from warblers, may influence herpetofauna's behavior and evolution (Wunderle 1981, McLaughlin and Roughgarden 1989, Poulin et al. 2001, Acosta and Morún 2014. The ability of anoles to lose their tails (caudal autotomy) can give them a chance to escape predation (Clause and Capaldi 2006, Bateman and Fleming 2009, Hoefer and Robinson 2020, as we observed in Florida. We recommend researchers and citizen scientists photograph and report on more foraging observations of warblers and other birds, especially in tropical and sub-tropical regions where less is known about predation behavior (Boal 2008, Devenish-Nelson et al. 2019, Naude et al. 2019, Berryman and Kirwan 2021.