Behaviour of two species of psittacine birds at wild bird feeding sites in Australia

This study investigated how Crimson Rosellas Platycercus elegans (CR) and Australian King-Parrots Alisterus scapularis (AKP) used provisioned seed at two public bird feeding sites in Australia. A total of 197 CR and 72 AKP were trapped and colour-banded. Observational data was collected every 10mins between 08:00-16:00 for three consecutive days during autumn and spring. Foraging effort was described using five metrics that quantified the birds’ visiting frequency and foraging duration over each day and observation period. Seed selection (over 5mins) and intake (over 10mins) were determined, and the energy intake was calculated. Total counts and population estimates were calculated for each species. Individual, species, seasonal and geographic variation in the use of provisioned seed was demonstrated by the metric summaries and Restricted Maximum Likelihood Modelling. Both species fed as part of large mixed species flocks that would not naturally congregate together to forage. Overall, CR were found to have higher foraging effort and feed in greater numbers than AKP, but a spectrum of use was observed for both species. Individuals were observed using the provisioned seed between 0-3 days/observation period. When birds used the provisioned seed, they were found to make between 1-8 visits/day, with most lasting 10-30mins. Few daily durations lasted longer than 50mins. Within a 10-minute interval, it was possible for a CR and AKP to obtain between 1.73-62.91% and 6.84-88.54% of their daily energy requirements, respectively. In a visit, either species could fill their crop and meet most, if not all, of their daily energy requirements. A small percentage of birds (6.5%) were found to use the feeding sites daily and for long durations (up to 160mins). It is likely that a proportion of the birds using the provisioned seed at both sites were dependent on the food source and would be at risk if the seed supply were suddenly reduced. The study also provided evidence that wild bird feeding provided an advantage to one or more species, as well as evidence that the food source did not affect the study species’ seasonal dispersal patterns or juveniles’ ability to forage on natural food sources.

08:00-16:00 for three consecutive days during autumn and spring. Foraging effort was described using five 23 metrics that quantified the birds' visiting frequency and foraging duration over each day and observation 24 period. Seed selection (over 5mins) and intake (over 10mins) were determined, and the energy intake was 25 calculated. Total counts and population estimates were calculated for each species. Individual, species, 26 seasonal and geographic variation in the use of provisioned seed was demonstrated by the metric 27 summaries and Restricted Maximum Likelihood Modelling. Both species fed as part of large mixed species 28 flocks that would not naturally congregate together to forage. Overall, CR were found to have higher 29 foraging effort and feed in greater numbers than AKP, but a spectrum of use was observed for both 30 species. Individuals were observed using the provisioned seed between 0-3 days/observation period. When 31 birds used the provisioned seed, they were found to make between 1-8 visits/day, with most lasting 10-32 30mins. Few daily durations lasted longer than 50mins. Within a 10-minute interval, it was possible for a 33 CR and AKP to obtain between 1.73-62.91% and 6.84-88.54% of their daily energy requirements, 34 respectively. In a visit, either species could fill their crop and meet most, if not all, of their daily energy 1. Introduction 44 Wild bird feeding is one of the most common forms of wildlife interaction in the world (1-4). It is particularly 45 popular in developed countries with birds being fed at up to 75% of people's homes (2,(5)(6)(7)(8)(9)(10)(11)(12). The primary 46 reasons people engage in wild bird feeding are to experience the pleasure the activity provides and to compensate for habitat modifcation or harsh climatic conditions (6,9,(13)(14)(15). Despite its popularity, wild 48 bird feeding has elicited extensive criticism (16)(17)(18)(19). Authors argue that wild bird feeding activities threaten 49 to alter natural foraging behaviour, resulting in habitutation (17), reduced vigilance (16), impaired foraging 50 skills (16,20,21), dependency on the suppled food (4,16,17,19,22), distruption of daily activity patterns 51 (16,17) and changes in species movement patterns (16,20). Wild bird feeding may advantage dominant 52 individuals or species (5,9,16,23) and provide support to invasive species (24)(25)(26)(27). There are also health 53 concerns that include, but are not limited to, increased disease transmission between birds congregating at 54 fixed point feeding sites and the spread of zoonotic diseases to humans (27,28). Individually, or as a result 55 of their interactions, these impacts could then contribute to carry-over effects on recruitment, reproduction 56 and survival (29,30), ultimately disrupting natural selection processes (22), particularly if the foraging 57 resources provide support for less competitive individuals (30). Despite these often-raised concerns, 58 empircal in situ data to validate them has been limited (18).

59
Authors have recommended that wild bird feeding studies investigating these concerns should begin by 60 assessing the birds' use of the provisioned foraging resources (31). To date, research on foraging effort at 61 wild bird feeding sites has been dominated by studies of passerines in the United Kingdom and the United 62 States of America (1,30,[32][33][34][35][36][37], and a small number of studies have quantified the foraging effort of 63 carnivorous species (38) and waterbirds (39). Foraging effort has been found to be influneced by the 64 species studied (8, 23,36), time of day (1,33,34,38), location and the number of feeding opportunities then transferred to a calico bag until sampled and colour-banded. Each observation day ran from 08:00 to 16:00, with data collected every 10 minutes. Preliminary fieldwork 50 conducted pre-dawn to dusk indicated that it was rare for study species to be observed outside these times.

51
During each 10-minute interval (observation interval), the entire feeding area was scan sampled (56) by the 52 same observer, who followed a circuit transect, maintaining a line of sight <5m from the birds under 53 observation. In Victoria, a second observer was required to assist counting during the holiday period, due to 54 the larger area to be covered and the number of people and birds present.

55
All colour-banded birds observed foraging during each observation interval-including those that flew in-56 were recorded in the field data sheets. Birds were considered to be foraging if they were ingesting, 57 manipulating or reaching for food (54), or located at a point where seed was available The data recorded 58 included the time, species and unique colour-band set. Data recording for each observation interval took 59 between one and eight minutes; there were no occasions when there were so many birds that the recording 60 of one interval interfered with the start of the next. were included in the dataset if they met the following criteria: 1. The unique colour-band combination noted 89 in the field data sheet cross-referenced to an individual in the capture and banding data, positively 90 identifying the bird; 2. The bird was recorded foraging during any observation period after capture (including 91 birds recorded foraging during additional post-study observations at site 1) and was therefore known to be 92 alive. The data for three birds was discarded due to a band falling off (n=2) and a recording error in the field 93 (n=1).

94
The following metrics (demonstrated in S1 Fig.)

Queensland and Victoria (2a)
66 Over the course of the entire study, a total of 269 birds were colour-banded ( Parrots-QLD am=8, af=3, au=2, jm=0, jf=7, ju=3; VIC am=2, af=2, jm=1, jf=2. What happened to the 74 banded birds that were never observed again is not known. They may have still been using the feeding 75 sites at times other than the observation periods, or they may have died or emigrated.

12
Given the small percentage of birds recorded making a maximum of 7 and 8 visits in a day, these outcomes were pooled.

Visit duration
14 The percentage of birds that fed at a site for a maximum visit duration of 10 to 120 minutes during each 15 three-day observation period is shown in Fig. 3. Overall, most individual visits lasted 20 minutes or less.

26
The percentage of birds that recorded a daily duration between 0 and 160 minutes over each three-day 27 observation period is shown in Fig. 4. As a result of the birds that were not observed using the site on one 28 or more days during an observation period, the most frequently recorded daily foraging duration was zero.   Rosellas than Australian King-Parrots and higher for adults than juveniles, as listed in Table 3. There was 55 no significant difference for gender, holiday/not holiday, or fly-off events.

Seed selection
82 Preferential seed selection was observed during each of the 5-minute intervals that birds were offered the 83 seed mix. Birds favoured sunflower and safflower seeds, followed by hulled oats. Preferential seed 84 selection was consistent for Australian King-Parrots, but Crimson Rosellas would intermittently prehend 85 smaller seeds. In each interval, all of the sunflower seeds were consumed and in six intervals all of the 86 safflower seeds were consumed. In the two intervals that safflower seeds were not depleted, fewer than 87 10% remained. In each interval between 50-80% of the oats were consumed. Small seeds were always 88 consumed, but at less than 10% of the mass offered. Within the 5-minute intervals, between 1-11 birds 89 would feed from the tray. Competition was observed between birds of the same species and between 90 Crimson Rosellas and Australian King-Parrots; this regularly resulted in feeding being disrupted and birds 91 being displaced from the tray.

92
Two occasions when visitors brought their own supply of sunflower seeds provided the opportunity to 93 observe birds feeding on this seed type. On one occasion, a person hand fed an Australian King-Parrot for 94 five minutes. A feed intake rate of 1 prehended seed per 3 seconds was calculated, resulting in prehension 95 of approximately 200 seeds. On another occasion, a person was observed hand feeding black sunflower 96 seeds to two Australian King-Parrots. One bird fed for one minute and the other fed for fifty seconds before 97 they were displaced. The first bird consumed 28 seeds (one seed/2.1secs) and the second bird consumed 98 14 seeds (one seed/3.5secs).

00
A summary of data for individual birds' foraging for a 10-minute interval is presented in

15
The foraging opportunity was influenced by the type and quantity of whole seeds available at each feeding 16 point (hand/ground). As seeds were hulled, there was an increasing ratio of husks to whole seeds. A bird's 17 feeding efficiency would be affected by the number of times it prehended husks. The ratio of whole seeds 18 to husks at the hand would remain reasonably high as people continually replenished the supply. An 19 examination of waste (seeds and husks collecting on the ground) from a 10cm square revealed more than 20 half of the material was husks and small seeds represented the bulk (>95%) of the whole seeds, but these 21 quantities would constantly be changing. Both of the study species fed from peoples' hands and on the 22 ground. Taking into consideration all of the foraging observations of Crimson Rosellas (n=1671) and

23
Australian King-Parrots (n=235), Crimson Rosellas and Australian King-Parrots were observed feeding at 24 the hand 8.98% and 85.96% of the time, respectively. vigilance, the amount of time a bird spent walking or flying between feeding opportunities, the level of 28 disturbance, the time a bird spent taking shelter in response to a threat (often preceded by alarm calls 29 triggered by loud noises or predatory birds), the number of birds feeding and the level of competition.

30
Seed availability and competition was also influenced by other species of birds that used the provisioned 31 seed. In Queensland, the primary incidental species were Australian Brush-turkey (Alectura lathami) (max.

59
From the numbers of studied species counted using the feeding site (banded and unbanded birds), the 60 maximum recorded each observation period is shown in Table 5   Victorian sites at peak use (in spring) was 577 and 492, respectively.

89
To understand the potential impacts of birds being offered provisioned seed at wild bird feeding sites, we 90 sought to verify if birds foraged at the sites every day, to determine how often and how long they foraged 91 on visit days, and to estimate the proportion of birds' daily energy requirements being met by the seed 92 consumed. If birds met most or all of their energy requirements by foraging at wild bird feeding sites, they 93 would be more likely to be negatively impacted by sudden changes in seed availability. It has also been 94 postulated that a constant source of provisioned feed could alter birds natural movement patterns and 95 result in juveniles not learning natural foraging skills (16,20).
96 using it at other times, birds that foraged at a feeding site one or two days out of three, and birds that 00 foraged at a feeding site every day. In Queensland over both seasons, the number of birds that did not use 01 the feeding site during the observation periods was small: <15%. Thirty to 45% used the feeding site one or 02 two days out of three and approximately half used the feeding site every day. Given that the feeding site in 03 Queensland was surrounded by wilderness, on days that Crimson Rosellas did not use the feeding site 04 they would have been obtaining their diet from natural food sources. In Victoria, the percentage of Crimson 05 Rosellas that did not use the feeding site or used it only one or two days was higher. These birds could 06 potentially have been using the time away from the feeding site to feed on natural food sources; however, 07 they could have fed at a nearby feeding site or backyard feeders-that were not monitored.

08
On days that Crimson Rosellas used the provisioned seed, the number of times a day they visited also 09 varied between individuals. Just under 30% of the Queensland Crimson Rosellas in both seasons only 10 visited the feeding site once a day, between 20% (spring) and 40% (autumn) visited the feeding site twice a 11 day and 35% (autumn) and 55% (spring) visited the feeding site three of more times a day. Overall, 12 frequency of visits to the feeding site in Victoria was comparatively lower in both seasons; this finding is 13 consistent with the birds' opportunity to make use of other feeding sites. The visit duration of the majority of 14 Crimson Rosellas using a feeding site in both seasons was 10 to 30 minutes. Maximum daily durations at 15 the feeding sites were commonly 50 minutes or less. In one ten-minute observation interval, it was possible 16 for a Crimson Rosella to consume between 2-63% of their daily energy requirements. of a range of resource patches. When foraging naturally, Crimson Rosellas regularly move from one 21 foraging opportunity to another (52). This natural behaviour may influence seed intake at a feeding site,  (23), seed preferences and 53 how the feed is presented (32, 66). Australian King-Parrots have been described as wary and while these 54 parrots are known to forage on the ground they are more commonly observed foraging in trees (52). At the 55 feeding sites, Australian King-Parrots were found to have a preference to feed directly from peoples' hands, 56 where competition was higher.

57
In addition to the discovery that fewer Australian King-Parrots used the feeding sites, they were also 58 generally less likely to visit the feeding sites as compared to the Crimson Rosellas, and when they did, they 59 visited fewer numbers of days, and averaged fewer visits per day. Despite this, Australian King-Parrots 60 potentially gained a similar percentage of their daily energy requirements from the provisioned seed.

61
Feeding from a persons' hand, provided the advantage of preferential seed selection and the availability of 62 a higher proportion of whole seeds. The observations of Australian King-Parrots' seed intake confirmed this 63 species could gain a half to full crop of seed in a ten-minute interval, providing between 50 and 89% of their