The limits of olfactory perception in black vultures: a field experiment

New World vultures have been considerably studied regarding their sense of smell. Cathartes species present a remarkable development of their olfactory apparatus, and experiments conducted with the turkey vulture (Cathartes aura) demonstrated that they can locate carrion exclusively by smell at considerable distances. Black vultures (Coragyps atratus), because of their less developed olfactory apparatus, have been compared with the turkey vulture as a phylogenetically related species lacking the sense of smell. However, little evidence from behavioural studies supports the lack of olfaction of black vultures. We conducted a field experiment where we presented urban black vultures a binary choice of decaying fish and sand (serving as control) inside plastic bags. Birds chose the fish bait in 81% of the times they approached the plastic bags. The probability of reaction decreased significantly with the distance between the birds and the fish bait, being virtually null at 40–50 m away. Furthermore, birds clearly reacted more when the fish presented higher decaying time (7 compared to 2 days). These results show that olfaction is not only used by black vultures but is probably the most relevant sense when they search for food hidden in plastic bags and in piles of unsorted waste, which is the most common foraging environment for this species in urban areas where they are very abundant.


INTRODUCTION
Avian olfaction was kept as a controversial topic until the 1960s when unequivocal evidence that birds rely on the sense of smell in multiple ecological contexts started to accumulate (e.g. Stager 1964;Wenzel 1968;Grubb 1972;Papi et al. 1972;Snyder & Peterson 1979;Wurdinger 1979). Some New World vultures are able to use the sense of smell to locate carrion (Potier et al. 2019;Potier 2020). The turkey vulture (Cathartes aura) has been often used as model species in olfaction studies due to the remarkable adaptations of their olfactory apparatus, including an enlarged nasal cavity and a very large olfactory bulb compared to brain size (Stager 1964;Bang & Cobb 1968), although the remaining Cathartes species (C. burrovianus and C. melambrotus) show similar adaptations (Stager 1964).
Very often studies on olfaction capacity of turkey vultures draw contrasting comparisons with the black vulture (Coragyps atratus), a species of the same family, distributed nearly among the same range but presenting less obvious anatomical adaptations for the sense of smell (Bang 1964;Stager 1964;Houston 1986). In experiments where carrion was placed under dense forest cover, turkey vultures typically arrived first but became rapidly outnumbered by black vultures and eventually displaced from the feeding sites (Houston 1986(Houston , 1988Lemon 1992;Byrne et al. 2019). This suggests that while turkey vultures rely on the sense of smell and cover large areas to search for food, black vultures are opportunists, mostly tracking other scavengers to find food ). In addition, black vultures are typically observed in open habitats, suggesting that they mostly search for feeding opportunities through sight, while turkey vultures prefer to fly over forested areas, supposedly searching for odours of dead animals laying at the forest floor, where the competition with other scavengers is reduced (Houston 1988;Novaes & Cintra 2015;Holland et al. 2019). The different soaring behaviour of these two species also underline their contrasting foraging strategies. While black vultures tend to soar high using thermal uplift, from where they have a wide range of sight, turkey vultures are particularly prone to use contorted soaring, a low altitude soaring mode that relies on shear-induced turbulence (Mallon et al. 2015).
Although both species use urban areas, black vultures are very abundant in this environment, particularly in the neotropics, and seem to be particularly well adapted to exploit anthropogenic organic waste (Novaes & Cintra 2015). This species commonly patrol refuse containers and street markets and outnumber turkey vultures by far in landfill sites (Novaes & Cintra 2015;Araujo et al. 2018). Among the many challenges posed by the colonisation of the urban environment, we highlight the detection of food inside plastic bags piled in refuse dumps. We hypothesised that using the sense of smell would be highly advantageous for black vultures in such a context, allowing individuals to locate and ingest food before being displaced by conspecifics and other scavengers.
Although many studies have assumed that black vultures lack the sense of smell (e.g. Stager 1964;Houston 1986;Buckley 1996;Sazima 2007a;Novaes & Cintra 2015), to our best knowledge none have unequivocally proved this hypothesis. A recent experimental study has shown that the southern caracara (Caracara plancus), a species with similar foraging ecology as the black vulture (Sazima 2007b), was able to use olfaction to find hidden food (Potier et al. 2019).
We conducted a field experiment in order to quantify the olfactory sensitivity of black vultures to food hidden in plastic bags. In specific, we provided isolated black vultures in urban areas a binary choice between decomposing fish and sand (serving as control) inside plastic bags at distances ranging from 10 to 50 m. Bags were visible from the birds' locations, but were dropped in way to avoid calling their attention. Fish was aged for 7 days in one set of experiments and for 3 days in another, intending to evaluate how the putrefaction level of the food would influence the behaviour of the birds. We predicted that: (1) birds will be attracted by the plastic bags containing fish more often than to the ones containing sand, which would prove that they use the sense of smell; (2) they will react more to the bags with fish placed at closer distances, and will stop reacting beyond a certain distance threshold; (3) they will react more and up to larger distances when the fish bait was decomposed for longer time (i.e. for 7 rather than for 3 days).

MATERIALS AND METHODS
This study was carried out at two sites, an urban park in the city of Belém (1°28'31"S, 48° 29'57"W), northern Brazil, and a public football field at a small town located 70 km from Belém (1°10'02"S, 48°27'33"W). Black vultures are very common urban birds in the Amazon region, gathering in large numbers at street markets and refuse dumps where they feed on organic waste (Novaes & Cintra 2015). They are also relatively common in urban parks and residential areas where open rubbish bins and piles of waste in plastic bags are commonly found (Araujo et al. 2018).
We investigated the range of olfactory perception of black vultures by examining their behaviour in the presence of decomposing fish placed at distances between 10 and 50 m. We chose experimental subjects among birds that were isolated from flocks and either standing at the ground level or perched on streetlight poles lower than 8 m. For each bird selected, we placed two small opaque plastic bags (36 × 23 cm), one with sand and the other with decomposing fish, at pre-determined distances (10, 20, 30, 40 or 50 m). Each subject was only tested in one of the five experimental distances. The observer wore a long raincoat, to conceal its movements while dropping the plastic bags, and moved in a straight path and constant speed, simulating the behaviour of transient people (see Video 1 in Supplemental Data). Both bags were dropped at equal distances from each focal bird, but in nearly opposing directions. Distances were previously measured with a rangefinder. The bag with sand was always the dropped first, and the fish bag was kept in a styrofoam hermetic case until being dropped. This prevented birds to be influenced by olfactory cues before the start of the experiment. Plastic bags were loosely knotted, leaving two openings on the top to allow the fish odour to be released, but not to look inside. Fish bait in each bag consisted in two king weakfish (Macrodon ancylodon) ca 20-cm long, which were bought in a local fish market and aged in a styrofoam box for 3 or 7 days, depending on the experimental set. Bird reactions to the bait were observed and video recorded from a distance higher than 30 m (see Video 2 in Supplemental Data). The experiment stopped after the subject approached and tried to open the bag or in case it showed no reaction to the bait after 30 min. The plastic bags were recovered at the end of each experiment. We aborted experiments when non-focal individuals got close and interacted with focal birds.
Experimental trials were carried out between April 2020 and March 2021, with a minimum of 2 weeks interval between sampling days and alternating between the two study areas to minimise the changes of habituation by the vultures. We chose days with no perceptible wind or rain, which could interfere with the dispersion of the fish odours. Experiments were conducted from 6:30 to 8:30 am, when birds were very active and the number of transient people was relatively low. We conducted a maximum of three trials each day using subjects that were at least 50 m apart, to reduce the chances of pseudoreplication and interactions between subjects. In addition, movements of subjects previously tested were observed to make sure they were not tested again. However, subjects were not individually marked, thus we cannot exclude pseudo-replication fully, particularly between different sampling days.
We tested the effects of fish bait distance and decay time on the olfactory detection by black vultures using two binomial Generalised Linear Models (GLM). The first model was exclusive for the experiment where we used fish bait aged for 7 days, for which we tested a wider range of distances (10 to 50 m). The second model tested the fish bait distance and decay time effects simultaneously, but we used only data recorded at distances of 10 and 20 m. In both models the response variable was assigned as 1 if the bait was detected and 0 otherwise. Models were fit in R (R Core Team 2018) through the function GLM of the package stats. Goodness-of-fit was evaluated through maximum likelihood pseudo R squared.
The experimental procedures of this study were approved by the Instituto Chico Mendes de Conservação da Biodiversidade (licence SISBIO 66269-1) and the Ethics Committee for Animal Use of the Federal University of Pará (licence CEUA 9117141118).

RESULTS
We tested 56 individuals at distances ranging between 10 and 50 m using fish aged for 7 days and 20 individuals at 10 and 20 m using fish aged for 3 days (Table 1). Considering all birds tested, 21 birds reacted to our experimental setup of which 17 reacted to the fish bait (Table 1). Subjects reacted significantly more towards the bag with fish than that with sand (chi-squared test: χ 2 = 7.96, df = 1, P-value = 0.0048). Subjects that reacted to our experimental setup took between 3 and 30 min to approach the plastic bags (mean = 11.3).
The probability of the fish bait detection was significantly affected by the distance between the bait and the bird and the decay time of the bait (Fig. 1, Table 2). Birds had ca 50% chances of finding 7-day-old bait at 10 m distance, but detection probability declined sharply towards larger distances (Fig. 1, Table 2). At 40 and 50 m the bait was found only once (Table 1). At 10 and 20 m, birds had significantly higher chances of finding the 7-day-old bait than the 3-day-old bait ( Fig. 1, Table 2). Birds were only attracted to the bag with sand 4 times among all trials, with no apparent relationship to distance or the bait decay time (Table 1).

DISCUSSION
The results of our experiment clearly indicate that black vultures do have a sense of smell, contrary to the assumptions of earlier studies (e.g. Stager 1964;Houston 1986;Buckley 1996;Sazima 2007a;Novaes & Cintra 2015). Following our first prediction, subjects were attracted to the bag containing fish in 81% of the times they approached the plastic bags, systematically ignoring the bag with sand that served as control. Given that both plastic bags were visible to the birds, the systematic reactions towards the bag with fish proved that the plastic bag per se did not represent a strong stimulus to search  Table 1 for details). Shading areas (left plot) and error bars (right plot) represent 95% confidence intervals. Table 2.
Summary of binomial GLMs testing the effects of fish bait distance and decay time on the olfactory detection by black vultures. The response variable was assigned as 1 if the bait was detected and 0 otherwise. The first model was exclusive for bait with 7 days of decay and had higher discrimination of the detection distance. The second model tested the effects of bait distance and decay time but only included data recoded up to 20 m of distance. The maximum likelihood pseudo R 2 is presented as indicator of goodness of fit. Significant relationships are shown in bold and are plotted in Fig. 1 for food, contrasting to the observations of Sazima (2007a). This was not totally surprising as in both our study sites dumped garbage and particularly plastic bags are very abundant, likely reducing the stimulus for inspection. However, we cannot ignore that some birds did react to the bag with sand, showing that visual cues were used to some extend in addition to olfactory cues. We should also emphasise that a large proportion of the subjects did not respond to our experiment. Even among subjects tested at 10 and 20 m with the 7-day-old bait, nearly a half showed no reaction to our experimental setup. This might be explained by the large number and variety of stimulus in the field sites, such as the behaviour of conspecifics (that may indicate feeding opportunities and agonistic interactions), people (that may discard organic waste) or noise of different sorts. We also admit that individuals with low activity, such as those showing a roosting posture or preening, could be less receptive to olfaction stimulus. Although the chances of subject pseudo-replication in our experiments were low, as birds were very abundant in the study sites and the interval between trials in the same site was over a month, they cannot be ruled out, as birds were not individually marked. If present in our data, subject pseudoreplication would bias the error estimates of our models, but it is highly unlikely that this problem would change our main results and conclusions (Zuur et al. 2009). We were also able to prove our second prediction that the probability of bird reactions decreased as the distance to the fish bait increased, being virtually null at 40-50 m away (Fig. 1). Obviously, baits of other sorts and sizes could deliver different results as these variables influence the amount of volatile odorants that are emitted (Smith & Paselk 1986). Still, it is extraordinary that a species thought to not use the sense of smell were able to locate food by scent at least up to 30 m far away. The lack of studies with a similar experimental setup for other vulture species prevents us to draw clear comparisons on olfaction sensitivity, but it seems clear from previous studies that the turkey vulture is able to locate carrion at by scent at much larger distances (Stager 1964;Smith et al. 2002). In an experimental study, Stager (1964) showed that turkey vultures responded to the release of ethanethiol (a by-product of carrion decomposition) at more than 180 m. Nevertheless, being able to smell food at 30 m makes it possible for black vultures to locate carrion at the tropical forest floor by flying over the canopy, a niche described as exclusive to Cathartes species (Houston 1986(Houston , 1988. However, it becomes clear from the literature that black vultures are more successful to find foraging opportunities by tracking other scavengers then by locating food themselves (Houston 1988;Buckley 1997).
Our third prediction also held, as we were able to show that birds were more attracted to fish aged for 7 days than for 3 days (Fig. 1). The release of volatile odorants increases with the decomposition process likely increasing the olfaction stimulus for the birds (Smith & Paselk 1986), although birds may have specific sensitivity to particular compounds that are produced only at some stages of the decomposition process. Similar experiments with turkey vultures delivered opposing results, with birds showing better performance in finding 1-day-old than 4-day-old carrion (Houston 1986). However, this result was interpreted as a preference by fresher food rather than olfactory discrimination (Owre & Nothington 1961;Houston 1986). In our experiment, there was no evidence indicating a preference for fresher or older bait, as the birds were always engaged in ingesting the fish once found.
Overall, our results allow us to refute the common assumption that black vultures lack the sense of smell. The use of olfaction may be particularly relevant to urban black vultures, which are commonly challenged to locate food inside plastic bags piled in refuse dumps.