Bird hierarchies and co-occurrences at urban garden in Venezuela

ANARTIA

Publicación del Museo de Biología de la Universidad del Zulia

ISSN 1315-642X (impresa) / ISSN 2665-0347 (digital)

Anartia, 33 (diciembre 2021): 55-65

Hierarchies and co-occurrences among bird species visiting

a feeder at an urban garden of Venezuela

Jerarquías y coincidencia entre especies de aves visitantes de un comedero en un

jardín urbano de Venezuela

Andrés E. Seijas

Universidad Nacional Experimental de los Llanos Occidentales “Ezequiel Zamora” (UNELLEZ). Guanare, Portuguesa, Venezuela.

Correspondence: aeseijas@gmail.com

(Received: 30-09-2021 / Accepted: 20-12-2021 / Online: 02-02-2022)

ABSTRACT

Compared to the vast literature that has been generated in developed countries, research on birds that attend feeders in

private gardens in the Neotropic are very scarce. is article analyzes the interactions among birds visiting a feeder in a

private garden in the city of Guanare, Venezuela. Bird activities were recorded in 336 videos for an eective recording

time (ERT) of 51.82 h, equally distributed between the dry and rainy seasons. Twenty species visited the feeder with the

Blue-gray Tanager (raupis episcopus) representing 40.3% of visits. Bird visits during the rainy season (3,975) were 59.5%

higher than those of the dry season (2,493). Taking the two seasons together, the feeder remained without birds 45.5% of

the ERT and of the 28.25 hours in which at least one bird was present, 80.2% was occupied by solitary individuals. Solitary

individuals of Cacicus cela, Turdus leucomelas, T. nudigenis and Mimus gilvus amounted to 50.5% of the time the feeder

was in use; these species also showed a high percentage of exclusive occupancy of the feeder (%Exc, not shared with other

species) reaching to 96.6% in the case of M. gilvus. e hierarchy of the species in their access to the feeder was positively

correlated with their size (Spearman’s rank, r = 0.87). Relatively large species (≥ 54g) won interspecic interactions in pro-

portions above the expected by chance. Of 55 possible pairings of species co-occurrences at the feeder, 7.3% had positive

associations, 30.9% negative associations, and 61.8% random associations. Negative associations always involved at least

one large species, whereas there were not negative associations between pairs of small bird species. e size of the species is

a determining factor in structuring the assemblage of birds visiting the feeder.

Key words: birds, co-occurrences, feeder, hierarchies, urban garden.

RESUMEN

Comparado con la vasta literature que se ha generado en países desarrollados, las investigaciones en el Neotrópico sobre

aves que visitan comederos en jardines privados son escasas. En este artículo se analizan las interacciones entre aves que

visitan un comedero abastecido con frutas en un jardín privado de la ciudad de Guanare, Venezuela. Las actividades de las

aves fueron registradas en 336 videos con un tiempo efectivo de grabación (TEG) de 51,82 h, repartido equitativamente

entre las estaciones seca y lluviosa. Veinte especies de aves visitaron el comedero, con el Azulejo (raupis episcopus) como la

más frecuente (40,3% de las visitas). El número de visitas durante la estación lluviosa (3.975) fue 59,5% mayor que las de la

estación seca (2.493). Tomando las dos estaciones en conjunto, el comedero permaneció sin aves presentes 45,5% del TGE.

De las 28,25 horas en las cuales al menos un ave estuvo presente en el comedero, 80,2% correspondió a individuos solita-

rios. La ocupación por individuos solitarios de cuatro de las especies de mayor tamaño (Cacicus cela, Turdus leucomelas, T.

nudigenis and Mimus gilvus) representó 50,5% del tiempo de uso del comedero. La Paraulata Llanera (M. gilvus) mostró un

porcentaje de uso en exclusividad del comedero (%Exc) de 96,6. Hubo una correlación positiva entre la posición jerárquica

Seijas

de las especies y sus tamaños (Spearman, r = 0,87). Las especies relativamente grandes (≥ 54g) desplazaron a individuos

de especies distintas a las suyas en proporciones superiores a las esperadas al azar. De los 55 pares de especies que podrían

coincidir en el uso de comedero, 7,3% mostraron asociaciones positivas, 30,9% asociaciones negativas y 61,8% asociaciones

al azar. Las asociaciones negativas entre pares de especies involucraron al menos una especie grande y no hubo asociaciones

negativas entre especies pequeñas. El tamaño de las especies es un factor determinante en la estructuración del ensamblaje

de aves que visitan el comedero.

Palabras clave: aves, coincidencias, comedero, jardín urbano, jerarquías.

INTRODUCTION

e urbanization process advances unstoppably on a

global scale (Grimm et al. 2008, Cohen 2015, United Na-

tions 2019) bringing drastic changes in the characteristics

of the aected lands and makes them less hospitable for

most of the species that occupied the former undisturbed

areas or, on the contrary, creates favorable conditions for

species adaptable to the new ecosystems (Chace & Walsh

2004, Shochat et al. 2006, Faeth et al. 2011, Belaire et al.

2014). People living in cities have few opportunities to ob-

serve “wild” animals other than those capable of occupying

public green areas and private gardens. One way to increase

the possibilities of observing these animals in their own

houses is by providing resources that attract them. e most

common of these attractants are birdfeeders (Goddard et al.

2009, Tryjanowski et al. 2015, Cox & Gaston 2018).

Compared to the vast literature that has been generated

in developed countries, research on birds that attend feed-

ers in private gardens in the tropics and in Latin America

in particular are very scarce (Echeverría & Vasallo 2008),

and in the case of Venezuela, there is just a handful of stud-

ies on this subject (Levin et al. 2000, Sainz-Borgo & Levin

2012, Seijas & Seijas-Falkenhagen 2020a, Seijas 2021).

From a conservation point of view, birdfeeders can have

favorable or unfavorable eects (Galbraith et al. 2017, Mc-

Burney et al. 2018, Deguines et al. 2020). Among the un-

favorable eects, it has been pointed out that feeders can

facilitate the transmission of diseases and the proliferation

of unwanted species; they also may contribute in the mal-

nutrition of birds due to the provision of inappropriate

food (Dunkley & Cattet 2003, Ishigame & Baxter 2007,

Orros et al. 2015). Birdfeeders have been identied as re-

sponsible for exerting an important eect on the structur-

ing of bird communities and to inuence several aspects of

bird ecologies, such as reduction of foraging time, increase

in body condition, changes in survival and reproductive

rates, changes abundance or density of species, among oth-

ers (Galbraith et al. 2015, Møller et al. 2015, Tryjanowski

et al. 2016). So, the use of feeders for conservation pur-

poses should be based on well-conducted research (Fuller

et al. 2008).

ere are many factors that could intervene in structur-

ing the assemblage of birds that visit a feeder, such as the

species pool in the region, the seasonal changes in their

respective abundances, the changes in nutritional require-

ments and in the behavior of birds in relation to their re-

productive phenology, or the feeder location with respect

to surrounding plant cover (Horn et al. 2002), but perhaps

the most important factors are the preferences of the dif-

ferent bird species for the food that is oered in the feeder

and the hierarchies that are stablished among them when

accessing the food (Robb et al. 2008, Wojczulanis-Jakubas

et al. 2015, Le Louarn et al. 2016, Deguines et al. 2020).

In this paper, I analize the interactions among birds

visiting a feeder provided with fruits in a private garden

in Guanare, Venezuela. e research is based on a data set

greater than the one used in previous studies in the same

garden (Seijas & Seijas-Falkenhagen 2020a, Seijas 2021)

but with an emphasis on the role that inter and intra-spe-

cic relationships play in dening the pattern of use of the

feeder and on the associations (or lack of them) between

the dierent bird species.

MATERIALS AND METHODS

e study was carried out in a private garden on the

outskirts of the city of Guanare, Portuguesa state, Ven-

ezuela. e birdfeeder consisted of a square cement block

(40cm x 40cm and 5cm thick) placed at ground level. On

it, pieces of fruit were placed and covered with a grid (5x5)

of plastic-coated wires whose function was to prevent the

birds from taking out or turning over the food, but which

also served as perch for the birds. More details on the gar-

den and feeder features can be found in Seijas & Seijas-

Falkenhagen (2020a, b).

e activities of the birds were recorded on video with a

cell phone placed on a tripod at a height of 30 cm and 1m

away from the feeder. e recordings were made during

days separated at irregular intervals of both the dry sea-

son (from December 14, 2019 to May 17, 2020) and the

rainy season (from May 20, 2020 to October 13, 2020).

e birdfeeder was provided with pieces of up to four fruit

types, selected from banana, plantain, papaya, mango and

Bird hierarchies and co-occurrences at urban garden in Venezuela

avocado. Each combination of fruits was set as a trial to

determine the feeding preferences of the dierent bird

species. Detailed analyses of these preferences will not

be attempted in this paper but partial results of some of

these trials were discussed in Seijas & Seijas-Falkenhagen

(2020a) and Seijas (2021).

Two to six sessions of variable duration (from 4 to

14min) were recorded during sampling days. e food was

supplied ad libitum, as the birds never fully consumed it

in the period from the start of the rst recording session

to the end of the last. All recordings were carried out dur-

ing the early hours of the morning, but in the rst 38 days

of recording (between December 14, 2019 and April 20,

2020) the rst session began 15 minutes before sunrise. As

the rainy season approached, the cloudy sky made those

minutes before dawn very dark. For that reason, begin-

ning on April 25 2020, the rst recording session of the

day began at sunrise. e eective recording time (ERT)

was taken as the duration of each video minus 30 seconds,

considering that the behavior of the birds in the rst 15

and last 15 seconds of each session could be conditioned

by the presence of the researcher placing and removing the

cell phone from the tripod where it was located.

For each bird visiting the feeder the following infor-

mation was taken: Species, time of arrival and departure,

interaction with other birds (in solitary, time shared with

individuals of its own or other species). When a bird le

the feeder, it was determined whether it was displaced by

another bird or le it for reasons that are not relevant for

this investigation. e individual who displaced another

one from the feeder was taken as the winner and the dis-

placed one as the loser, regardless of whether the displace-

ment involved some physical contact or just intimidation

(Wallace & Temple 1987).

Data were uploaded to an Excel spreadsheet designed

to calculate the duration of each visit, time shared with

other individuals, number of individuals (and species)

sharing their visits, and time that the feeder was not in

use (no birds present). e values and variables calculated

from the data are dened in Table 1.

To evaluate the intensity of the intra and interspecic

interactions, the quotient Ti/T was taken as the expected

proportion of time that individuals of each species should

spend at the feeder in solitary or sharing both with in-

dividuals of its own or other species if these possibilities

were dictated by chance. at expected proportion was

then substracted from the observed proportions Solitaryi/

All solitaries, Share-owni /All sharing own, and Share-

othersi /All sharing others. To facilitate comparisons, the

deviation from expected were expressed as Z-scores that

were calculated substracting the mean of each set of dif-

ferences and then dividing the results by their correspond-

ing standard deviation (McClave and Dietrich 1994). e

Z-score represents the distance between a given dierence

(expressed in standard deviation units) and zero, the ex-

pected value. e same procedure was followed to analyze

the dierences in the proportions of negative interactions

between individuals of the same species (Win-owni/All

own) or other species (W-othersi/All-others) in relation to

the expected proportion Vi/VT. For these analyses, pro-

Table 1. Names and denitions of values and variables used in the analyses. All values calculated taking into account the

entire study.

Symbol Definition

TiTime spent at the feeder by all individuals of the ith species.

TTime spent at the feeder by all visitors (ƩTi)

SolitaryiTime spent by solitary individuals of the ith species

ShareOwniTime shared by the ith species only with conespecifics.

ShareOthersiTime shared by the ith species with individuals of other species

%Exci

Percentage of time the ith species spent at the feeder without sharing it with individuals of other species:

(Solitaryi + ShareOwni)/Ti)*100.

ViNumber of visits of the ith species

VT Total number of visits to the feeder (ƩVi)

All-own Number of times that visiting birds were displaced by individuals of their own species

All-others Number of times that visiting birds of a certain species was displaced exclusively by individuals of species

other than their own.

W-owniNumber of interactions won by the ith species to individuals of its own species

W-othersiNumber of interactions won by the ith species to individuals of other species

Seijas

portions were arcsine squareroot transformed to stabilize

variances and normalize values.

A Winner-loser dominance matrix (Levin et al. 2000,

Seijas & Seijas-Falkenhagen 2020a) was used to calculate

the percentage of interactions with individuals of other

species a particular species won (%W). e Hierarchy (H)

of a species was determined counting the the number of

other species it displaced from the feeder in the majority of

their interactions. e hierarchies of the species were corre-

lated with their weight and their %Exci. Statistical analyzes

were performed with Past 4.02 (Hammer et al. 2001).

To evaluate the degree of association (positive, negative

or none) between species pairs co-occuring at the feeder, a

subset of the visits were selected. is was done by taking

in each recording session the h individual reaching the

feeder, and then the following individuals separated by ten

positions in their order of arrival (5, 15, 25, 35, etc). is

guarantees that each selected visit did not overlap with the

previous or subsequent selected one, since a preliminary

analysis of the data showed that rarely does a bird share

with 10 or more individuals during a visit (that only oc-

curred in 0.49% of the total visits to the feeder). Visits of

ve or less seconds, and individuals that remained with the

focal individual for such short period of time were also dis-

carded, because in many cases stays at the feeder for such

a short duration can hardly qualify as “time-sharing”, since

the displacement of one individual by another does not

occur instantaneously, but there may be some intimida-

tions, threats or even ghts, which can take a few seconds

to end up with the abandonment of the feeder of one of

the individuals involved. e selected visits (N) allowed

determining which species shares the feeder. To evaluate

whether there were positive, negative or random associa-

tion between the dierent pairs of species co-occurring

at the feeder, the probabilistic model described by Veech

(2013, 2014) was used. Said model uses presence/absence

data to calculate expected frequencies of occurrences be-

tween species pairs if they were distributed independently

of each other across the selected visits. at model is ex-

pressed mathematically as:





=(,) × ( − ,



− , ) × ( − 



,



− )

(,



) × (,



)

where N is the number of selected visits; N1 and N2 are the

number of times species 1 and 2 appear in N, respectively; j

is the number of visits where the same species appear shar-

ing the feeder; C(N, j) is the number of possible combina-

tions in which two species could appear simultaneously J

times in the N selected visits; C(N - j, N2 - j) is the number

of possibilities of locating species 1 among the remaining

places where species 2 is not present; C(N - N2, N1 - j) is the

number of ways in which species 2 could be located among

those that do not have species 1. e denominator repre-

sents all the possible combinations in which species 1 and

2 can be arranged in the N selected visits, without taking j

into account. e numerator is a subset of the numerator

and therefore the quotient Pj is always < 1 and represents

the probability that a given combination of two species

will appear sharing j times when N visits are selected. More

details and discussions about this model, its meaning and

its applicability can be read in Arita (2016). Grith et al.

(2016) show a simplied version of the model.

e birds visiting the feeder were not tagged, so with

the exception of a few individuals that have peculiar char-

acteristics that distinguish them, it was not possible to

individualize the visits. However, based on observations

made over the years (Seijas & Seijas-Falkenhagen 2020b) it

can be stated that the most abundant species in the garden

is the Saon Finch (Sicalis aveola) whose numbers in the

lawn of the garden can sometimes exceed several dozens.

Around 20 Yellow-rumped Cacique (Cacicus cela) are in

and out of the garden every day. e number of Blue-gray

tanagers (raupis episcopus) that daily visit the feeder has

been estimated between 20 and 30. In the case of thrushes

(Turdus leucomelas and T. nudigenis) it is possible to ob-

serve a maximum of 5-6 individuals foraging in the garden

along the day. Some species go down to the feeder in what

appear to be small family groups (Ramphocelus carbo, Mi-

mus gilvus, Euphonia laniirostris, Campylorhynchus nuch-

alis, Icterus nigrogularis). Other species (Stilpnia cayana,

raupis palmarum, Saltator coerulescens, and occasionally

Melanerpes rubricapillus) oen descend in pairs. Other

birds were sporadic visitors.

RESULTS

Two-hundred and ve videos were recorded during

the 2019-2020 dry season and 131 during the 2020 rainy

season. e average duration of the videos in the rainy

season was longer, but the ERT was very similar between

both seasons (25.90 and 25.92 hours, respectively). In

each climatic season, the feeder had no birds present for

a high percentage of the recordings time. ese visitor-

free lapses represented 51.0% of the ERT during the

dry season and 40.0% during the rainy season (Table 2).

However, in the dry season recording sessions that began

15 minutes before sunrise (N = 38), the feeder remained

59.5% of the time without birds, while in subsequent ses-

sions recorded on the same days, that time was 49.2%;

the dierences are signicant (Two-sample paired test,

t= 3.04, P = 0.014).

Bird hierarchies and co-occurrences at urban garden in Venezuela

Taking both seasons together, 20 species of birds visited

the feeder. Visits for the rainy season (3,975) were 59.5%

greater than those of the dry season (2,493). ere was also

a similar dierence (62.7%) in the accumulated time of the

visits of all the birds (29.4h during the rainy season versus

18.1h during the dry one; Table 3). ese dierences are

partially explained by the 38 sessions that started before

sunrise in the dry season; if these sessions are excluded there

was still a 1.46 fold diference in the number of visits per

hour during the rainy season. e most noticeable dier-

ence in the number of visits between the two seasons were

those of the Silver-beaked Tanager (Ramphocelus carbo, 288

more visits), the ick-billed Euphonia (Euphonia laniiros-

tris, 370 more visits) and the Blue-gray Tanager (714 more

visits). e increase in the number of visits of these three

species represented 92.6% of the dierence between the two

seasons. Despite these dierences, there was a high corre-

lation in the relative frequency of the species (Vi) in both

seasons (Spearman Rank, n = 20, r = 0.753, P < 0.001). For

most of the following statistical analyses birds with less than

30 visits to the feeder were not taken into account.

e Blue-gray Tanager was the bird with the highest

number of visits (40.3%) but its time at the feeder was

only 26.7% of the time (T) spent by all bird species. As oc-

curred with all species, the %Exc of T. episcopus was higher

during the dry season (71.1%) than during the rainy sea-

son (46.6%), with a combined %Exc = 55.2%. Large birds

(≥54 g) in solitary amounted to 52.7% of the time that

the feeder was in use, but almost half of that time (24.1%)

was spent by the Pale-breasted rush (Turdus leucomelas)

alone. e time spent at the feeder as solitary individuals

by the Tropical Mockinbird (Mimus gilvus) was not as

high as other large birds, but the time sharing in exclu-

sivity with individuals of its own species was the highest

(%Exc= 93.2).

Figure 1 depicts how much each of the 11 most com-

mon species departed from the expected proportion of

time at the feeder as solitary, or sharing it with its own or

other species. e patterns for three species deserve to be

highlighted. e Pale-breasted rush showed the larg-

est deviation above expected (Z = 1.91) when comparing

solitary visitors. is species was also the one that shares

the least with individuals of its own species (Z = -1.78)

and showed an important deviation below expected when

sharing with individuals of other species (Z = -1.49). e

Mockingbird was the least prone to share with individuals

Table 2. Time allocation at the feeder. Ert is the recording time minus 30 seconds per each session (see text).

Recording time category Dry season sessions (205) Rainy season sessions (131)

Hours %  Hours %

Total recording time 27.61 27.15

Effective recording time (ERT) 25.90 100 25.92 100

Without birds 13.21 51.0 10.36 40.0

In use (birds present) 12.69 49.0 15.56 60.0

Solitary individuals 10.64 41.1 12.00 46.3

Shared by two or more birds 2.05 7.9 3.57 13.8

-2.5 -2 -1.5 -1 -0.5 00.5 11.5 22.5

C. cela (90)

T. leucomelas (62)

T. nudigenis (60)

S. coerulescens (55)

M. gilvus (54)

T. palmarum (36)

T. episcopus (35)

R. carbo (25)

S. flaveola (20)

S. cayana (19)

E. laniirostris (14)

Standard deviation

Solitary With own With others

Figure 1. Dierences (expressed as Z-scores) of observed mi-

nus expected proportions of time spent by birds as solitary in-

dividuals, or sharing the feeder with individuals of their own

or other species in a private garden in Guanare, Venezuela. e

mass (g) of each species is shown in parentheses.

Seijas

of other species (Z = -2.14) and shared more than expect-

ed with individuals of its own species (Z = 1.27). Finally,

the Blue-gray Tanager shared more than expected with

individuals of its own species (Z = 1.67) and stayed as soli-

tary less than expected (Z = -1.48).

ere were 3,196 occasions in which a bird was dis-

placed by another from the feeder. e winner-loser domi-

nance matrix in Table 4 summarizes 2,536 of those events

for the 14 species with the higher number of records.

When interactions between individuals of the same spe-

cies are excluded from the analysis, both the percentage of

encounters won by each species (%W) and their hierarchi-

cal position (H) are positively correlated (P < 0.001) with

their mass (Spearman´s rank r = 0.84 and 0.87, respective-

ly). e size of the species also correlated positively with

%Solitary (P<0.001, r =< 0.001) and %Exc (P = 0.02,

r= 0.60).

Relatively large birds (≥ 54g) won interactions with in-

dividuals of other species in proportions above the expct-

ed value; the Pale-breasted rush highlights among them

with a Z-score of 2.05 (Fig. 2). ree species won their

interactions with individuals of their own species above

the expected proportion but only two of them are worth

mentioning: the Pale-breasted rush (Z = 1.22) and,

particularly, the Blue-gray Tanager (Z = 2.18). is Tana-

ger won the lowest proportion of interactions with birds

dierent to its own species (Z = -1.61) followed by the

ick-billed Euphonia (Z = -1.12). Finally, the Tropical

Mockingbird (Z = -0.80) and the Burnished-bu Tanager

(Stilpnia cayana, Z = -1.26) were the species with the least

negative interactions with individuals of their own species.

For the analyses of co-occurrences, 518 visits were se-

lected including 15 of the 20 species that went down to

the feeder. e relative frequency of the species that ap-

Table 3. Number of visits and cumulative time spent by the dierent bird species at a feeder in a private garden in Guanare,

Venezuela. Total times (at the bottom) expressed in hours. See denitions of variables in Table 1.

Species / (body mass, g)

Dry season (25.90h)  Rainy season (25.92h)

Cumulative time min) Cumulative time (min)

ViTiSolitaryiOwni%Exci ViTiSolitaryiOwni%Exci

Cacicus cela (90) 90 48.8 39.6 2.3 86.0 57 37.3 25.7 2.6 75.7

Coereba flaveola (9) 4 6.3 3.7 0.0 58.5 9 4.5 0.8 0.0 17.5

Campylorhynchus nuchalis (25) 38 13.0 8.5 3.0 87.7 34 9.4 5.5 1.0 68.7

Euphonia laniirostris (14) 75 29.1 13.3 1.2 49.6 445 230.7 41.2 41.6 35.9

Icterus nigrogularis (38) - - - - - 62 33.3 10.3 2.6 38.9

Mimus gilvus (54) 255 198.9 120.8 71.3 96.6 207 200.4 141.5 38.7 89.9

Melanerpes rubricapillus (48) 25 17.2 9.3 0.0 53.8 31 21 8.6 0 40.9

Psarocolius decumanus (180) 3 2.0 2.0 0.0 100 - - - - -

Ramphocelus carbo (25) 10 6 2.0 0.7 45.3 298 111.7 32.5 8.4 36.6

Saltator coerulescens (55) 40 18.7 14.0 0.1 75.8 118 59 20.8 3.1 40.5

Sicalis flaveola (20) 79 41.5 20.7 5.3 62.7 53 35.5 9.9 5.7 43.9

Sporophila intermedia (12) 16 14.8 7.0 0.0 46.9 8 4 0.9 0.0 21.1

Stilpnia cayana (19) 186 80.6 28.5 20.2 60.4 223 92.9 19.6 11.9 33.9

Thraupis episcopus (35) 946 266.7 99.0 90.7 71.1 1660 492.9 120.3 109.3 46.6

Thraupis glaucocolpa (33) 3 1.9 0.2 0.0 12.3 - - - - -

Turdus leucomelas (62) 515 247.7 223.6 0.1 90.3 410 251.3 185.2 13.3 79.0

Turdus nudigenis (60) 107 45.4 33.7 0.0 74.2 237 129 85.9 0.8 67.2

Thraupis palmarum (36) 101 45.2 12.7 10.1 50.4 121 50 11.4 4.0 30.8

Tachyphonus rufus (33) - - - - - 1 0.6 0.1 0 22.7

Icterus icterus (68) - - - - - 1 0.3 0.3 0 100

Totals 2,493 18.1h 10.6h 3.4h -  3,975 29.4h 12.0h 4.1h -

Bird hierarchies and co-occurrences at urban garden in Venezuela

pear in the selected group was highly correlated with the

frequency of bird species in the total visits (Pearson r =

0.983). e 105 possibilities of dierent pairings and the

number of times these pairings occurred in the selected

sample appear in Table 5.

e analyses of co-occurrences showed discrepancies if

the data came from the dry season (N = 203), the rainy

season (N = 315), or the whole selected visits (N = 518).

In the rst case, two pairs of species showed positive asso-

ciations: the Burnished-bu Tanager with the ick-billed

Euphonia (P = 0.012), and the Saron-nch with the

Palm Tanager (raupis palmarum) (P = 0.039) and ve

pairs showed negative associations, all of them involving

the Pale-breasted rush and the Mockinbird. With the

data from the rainy season there were three pairs of species

with positive associations, all of them dierent from those

of the dry season: the Red-crowned Woodpecker (Mel-

anerpes rubricapillus) with S. cayana (P = 0.046) and also

with the Yellow Oriole (Icterus nigrolularis) (P =0.051),

as well as R. carbo with T. palmarum (P = 0.036). e

negative associations for the rainy season rose to 11, with

the Mockinbird the Pale-breasted rush as the most fre-

quently involved. Finally, with the joint data, there were

4 pairs of positive and 17 negative associations, among

the former, that involving R. carbo with T. palmarum was

Table 4. Winner-loser dominance matrix for species that visited the feeder in a private garden of Guanare, Venezuela. e

diagonal (underlined) indicates the number of times that an individual was displaced by another of its own species. ese

last values were not included in the accounts of “wins” or “losses”.

Winners Losers %W H Wg

Cc Mg Tl Sco Mr Tn Cn Rc Te Tp In Sca Sf El

C. cela 27 20 36 3 4 7 4 1 44 10 2 9 2 7 97.4 13 90

M. gilvus 4 6 46 7 3 5 2 2 69 9 0 14 3 6 92.4 12 54

T. leucomelas 0 0 185 7 0 82 13 54 571 48 7 70 7 21 87.1 10 62

S. coerulescens 0 0 17 2 0 13 0 4 37 3 1 5 0 3 78.6 7 55

M. rubricapillus 0 0 15 0 0 3 0 3 6 1 0 1 0 2 77.5 7 48

T. nudigenis 0 0 1 0 1 6 2 12 90 5 4 6 1 16 50.9 7 60

C. nuchalis 0 0 8 0 1 4 0 0 4 0 0 4 2 0 51.1 5 25

R. carbo 0 0 2 0 0 0 0 3 26 3 0 3 2 9 36.9 5 25

T. episcopus 0 2 4 1 0 17 0 1 526 22 0 59 17 64 17.8 4 35

T. palmarum 0 0 0 0 0 2 0 0 14 2 0 6 1 3 24.3 3 36

I. nigrogularis 0 1 0 0 0 0 0 0 2 0 1 0 0 0 17.7 2 38

S. cayana 0 0 1 0 0 0 0 0 0 0 0 0 1 7 4.8 2 19

S. flaveola 0 0 0 0 0 0 1 0 0 0 0 0 8 1 5.1 1 20

E. laniirostris 0 0 0 0 0 0 0 0 0 0 0 0 1 23  0.7 0 14

%W: Percentage of encounters won to other species. H: Hierarchy. Wg: mass in grams (from Hilty 2003).

-2 -1.5 -1 -0.5 00.5 11.5 22.5

C. cela (90)

T. leucomelas (62)

T. nudigenis (60)

S. coerulescens (55)

M. gilvus (54)

T. palmarum (36)

T. episcopus (35)

R. carbo (25)

S. flaveola (20)

S. cayana (19)

E. laniitostris (14)

Standard deviation

W-own

W-others

Figure 2. Dierences (expressed as Z-scores) of observed minus

expected proportions of interaction wons by birds to individuals

of their own or other bird species in a private garden in Guana-

re, Venezuela. e size (grams) of each species is shown within

parentheses.

Seijas

maintained. In all cases, negative associations involved

pairs of a large species (≥ 54g) with a small one (≤ 48 g)

or pairs of large species. at is, there were no negative as-

sociations between small species.

Considering only the 11 species with the highest num-

ber of occurrences in the 518 selected records (Table5),

ve of the ten possible pairs among the large species

showed negative associations (Fig. 3). In contrast, all 15

possible pairs between the smallest species (upper le cor-

ner) showed neutral or positive associations. e Tropical

Mockingbird and the Pale-breasted rush were the spe-

cies with the highest number of negative associations with

other species (8 and 7, respectively). Globally, the species

pairs included in gure 3 show 7.3% positive associations;

30.9% negative associations and 61.8% random associa-

tions.

e ick-billed Euphonia was the bird with highest

number of positive associations (three), two of them with

relatively large species such as the Grayish Saltator (Sal-

tator coerulescens) and the Bare-eyed rush (Turdus nu-

digenis). is last species (the third largest bird) did not

show a negative association with any of the small species.

A species pair not included in the gure but that showed

a signicant positive association was that of the Saron-

nch and the Gray seedeater (Sporophila intermedia)

(P<0.001).

Table 5. e number of co-occurrences of individuals of dierent species at a feeder in a private garden in Guanare (Vene-

zuela) is indicated where rows and colunms intercept. e underlined values show the number of times individuals of the

same species shared the feeder.

C. cel C. nuc E. lan I. nig M. gil M. rub R. car S. coe S. fla S. int S. cay T. epi T. leu T. nud T. pal 

2300111100015200C. cel

70000000001010C. nuc

90 0 0 3 12 10 9 0 17 57 12 12 11 E. lan

1221110014000I. nig

560000035012M. gil

13201049200M. rub

48 6 1 0 7 27 8 4 8 R. car

32 3 0 6 17 2 0 0 S. coe

31 4 4 12 5 2 5 S. fla

601201S. int

74 39 11 3 8 S. cay

260 26 23 25 T. epi

119 2 1 T. leu

42 3 T. nud

46 T. pal

Figure 3. Association between species pairs at a birdfeeder in a

private garden in Guanare, Venezuela. e abbreviated scientic

names of the birds indicate both rows and columns. e species

are ordered according to their size with the smallest (Euphonia

laniirostris) in the upper le corner and the larger (Cacicus cela)

in the lower right corner. e numbers inside the red or green

boxes indicate the P values calculated according to the proba-

bilistic model of co-occurrences (Veech 2013). Boxes included

in the square delimited by a broken line represent interrelation-

ships between small species (≤ 48g) and large ones (≥ 54g).

Boxes above that rectangle represent interrelationships between

small species and those to the right represent interrelationships

between large species.

Bird hierarchies and co-occurrences at urban garden in Venezuela

DISCUSSION

Although the time spent videoing bird activities dur-

ing dry or rainy seasons was nearly the same, during the

dry season there was a lower number of visits to the feeder.

Not all species changed the frequency of their visits in the

same proportion and three of them (R. carbo, E. laniiros-

tris and T. episcopus) accounted for almost 93% on the in-

crease in numbers of visits between the two seasons. ese

dierences are probably related to changes in the phenol-

ogy of birds and fruiting plants alike. Although there is

no detailed information available on the reproductive bi-

ology of the rst two species, according to Hilty (2003)

they reproduce mostly during the dry season, and maybe

the time devoted to nesting and raising chicks (and look-

ing for other food types, for example) keeps them partially

away from the feeder. On the other hand, the Blue-gray

Tanager reduced drastically its number of visits from late

January to early March 2020, which coincided with the

copious fruiting of a Chrysophillum sp tree just 20 meters

from the feeder, where individuals of all bird species that

visited the feeder were observed eating its fruits, and the

most numerous of them belonged to T. episcopus (Seijas &

Seijas-Falkenhagen 2020b).

ere was a clear hierarchy among the birds accesing the

feeder, which was essentially determined by the size of the

species: the largest ones dominated the smallest, as has been

observed in many other studies in birdfeeders (Wojczula-

nis-Jakubas et al. 2015, Galbraith et al. 2017, Deguines et

al. 2020, among others). A notorious discrepancy in this as-

pect was oered by the Bared-eyed rush, a bird with a hi-

erarchical position below what would be expected accord-

ing to its weight (60 gr). is contrasts with the comments

on Verea et al. (2016) who pointed out that T. nudigenis

dominates over several species, including T. leucomelas,

something that occurred only once in the 83 confronta-

tions between these species recorded in the present study.

e Bared-eyed rush shares very little with individuals

of its own species, an indication of strong intra-specic an-

tagonistic relationships, but compared to T. leucomelas it

shares a little more with species smaller than itself.

A high fraction of the time that the feeder was in use

(with birds present) it was mostly occupied by solitary

individuals of some of the largest species, highlighting

among them the Pale-breasted rush and the Tropical

Mockingbird. ese two species dominated the feeder by

excluding other species, particularly the smaller ones.

Larger species (≥54g) displaced individuals of other

species more frequently than expected by chance. at

dierence was particularly high in the case of the Pale-

breasted rush. is species totaled a number of inter-

actions won greater than its number of visits. is means

that in some visits it displaced more than one individual

at a time. Most of these antagonistic interactions (17.4%)

were against individuals of its own species. Verea et al.

(2016) point out that T. leucomelas can form groups of up

to eight individuals, but it seems that these groups are not

very cohesive. is thrush is a very territorial species and in

the reproductive season ghts occur (presumably between

males) that can lead to the death of one of the contend-

ers (Seijas & Seijas-Falkenhagen 2020b). e Tropical

Mockingbird showed a much lower number of antagonis-

tic encounters with other species than the Pale-breasted

rush, but the confrontations between these two species

were overwhelmingly won by M. gilvus (49 to zero). e

dominance that the Mockinbird exerts over the feeder oc-

curs even when it is not occupying it. One or more Mok-

ingbirds remain vigilant in the vicinity of the feeder and

attack other birds that approach it, which was not always

registered in the videos.

Most small birds (<54g) displaced individuals of other

species less than expected by chance; that is, they used

the feeder sharing it with individuals of their own species

or with individuals of other relatively small species. is

can be interpreted as a response of the small species to the

monolopy of the feeder by the larger species; that is, they

need to share with other species the relative little time that

larger birds leave them available. One exception to this

pattern is oered by the Blue-gray Tanager that interacted

negatively much more than expected with individuals of

its own species and much less than expected with individu-

als of other species. e fact that T. episcopus shares a high

proportion of its time at the feeder with individuals of its

own species is probably a consequence of its abundance,

with more than 40% of the total visits. However, it seems

to be a very tolerant species toward individuals of other

species.

According to the co-occurrences analyses, it is clear that

larger species associate negatively among them and with

the relatively smaller species. Although the smaller species

were limited to use the feeder only a small part of the time,

the positive associations between them were rather rare.

Overall, the percentages of positive, negative and random

associations found in this study are very similar to those

quantied by Galbraith et al. (2017) in birdfeeders in

Auckland, New Zealand: 7.3% vs 8.3%, 30.9% vs 27.8%,

and 61.8% vs 63.9%, respectively.

e positive association between the Silver-beaked

Tanager and the Palm Tanager is probably a fortuitous

event. e rst species of this pair visited the feeder very

few times in the dry season, but there was a considerably

increased in its number of visits during the rainy season,

Seijas

so the positive association is produced by their co-occur-

rences in that season. As has been mentioned by Blanchet

et al. (2020), co-occurrence is not evidence of ecological

interactions, although these authors were referring to asso-

ciation of species in natural communities. Co-occurrences

can occur due to various causes, such as coincidences in

their biological cycles and visits to the feeder at the same

stages of their cycles (whatever these are) or also because

the species share the preferences for the type of fruit of-

fered. A clear example of the latter case is the positive as-

sociation between the Saron Finch and the Gray Seed-

eater; both species only visited the feeder on the occasions

that mango was oered (Seijas & Seijas-Falkenhagen

2020a, Seijas 2021). e ick-billed Euphonia was the

species with the highest number of positive associations,

including two with relatively large species (S. coerulescens

and T. nudigenis). is bird is the smallest among the birds

that visited the feeder. Perhaps due to its small size it is

not seen as serius contender to care about by other birds,

which tolerate its presence at the feeder.

All species that visited the feeder during this study are

native. at is probably a consequence of the exclusive

use of fruits as attractants. If other types of food had been

used, such as grains (corn, rice, sunower) or cooked food

(pasta, cooked rice, bread) surely other species would have

appeared, as I have observed in some birdfeeders in the

city. ese species include the exotic domestic pigeon (Co-

lumba livia) and other species that, although not exotic,

are very synanthropic and uninteresting from a conser-

vation point of view, such as the Carib Grackle (Quisca-

lus lugubris) and several small pigeons such as the Eared

Dove (Zenaida auriculata), the Scale Dove (Columbina

squammata) and the Ruddy Dove (Columbina talpacoti).

Likewise, the presence of a granivore such as Saron Finch

should have been much more noticeable, as it is the most

synanthropic species in Guanare (Seijas et al. 2011).

ACKNOWLEDGEMENTS

Andrés E. Falkenhagen and Sara F. Seijas-Falkenhagen

helped write this article in English. Two anonymous re-

viewers provided highly pertinent and helpful comments

to improve the writing and technical quality of this article.

REFERENCES

Arita, H. T. 2016. Species co-occurrence analysis: pairwise ver-

sus matrix-level approaches. Global Ecology and Biogeography

25: 1397–1400.

Belaire, J. A., C. J. Whelan & E. S. Minor. 2014. Having our

yards and sharing them too: the collective eects of yards on

native bird species in an urban landscape. Ecological Applica-

tions 24(8): 2132–2143.

Blanchet, F. G., K. Cazelles & D. Gravel. 2020. Co-occurrence

is not evidence of ecological interactions. Ecology Letters 23:

1050–1063.

Chace, J. F. & J. J. Walsh. 2004. Urban eects on native avifauna:

a review. Landscape and Urban Planning 74: 46–79.

Cohen, B. 2015. Urbanization, city growth, and the new United

Nations Development Agenda. Cornerstone 3(2): 4–11.

Cox, D. T. C. & K. J. Gaston. 2018. Human–nature interac-

tions and the consequences and drivers of provisioning wild-

life. Philosophical Transacttions of the Royal Society B 373:

20170092.

Deguines, N., R. Lorrillière, A. Dozières, C. Bessa-Gomes

& F. Chiron. 2020. Any despot at my table? Competi-

tion among native and introduced bird species at garden

birdfeeders in winter. Science of the Total Environment

139263.139210.131016.

Dunkley, L. & M. R. L. Cattet. 2003. A comprehensive review

of the ecological and human social eects of articial feed-

ing and baiting of wildlife. Canadian Cooperative Wildlife

Health Centre: Newsletters & Publications 21: 1–68.

Echeverría, A. I. & A. I. Vassallo. 2008. Novelty responses in a bird

assemblage inhabiting an urban area. Ethology 114: 616–624.

Faeth, S. H., C. Bang & S. Saari. 2011. Urban biodiversity: pat-

terns and mechanisms. Annals of the New York Academy of

Sciences 1223: 69–81.

Fuller, R. A., P. H Warren, P. R. Armsworth, O. Barbosa & K.

J. Gaston. 2008. Garden bird feeding predicts the structure

of urban avian assemblages. Diversity and Distribution 14:

131–137.

Galbraith, J. A., J. R. Beggs, D.N. Jones & M. C. Stanley. 2015.

Supplementary feeding restructures urban bird communi-

ties. Proceedings of the National Academy of Sciences of the

United States of America 112(20): E2648–E2657.

Galbraith, J. A., D. N. Jones, J. R Beggs, K. Parry & M. C. Stan-

ley. 2017. Urban bird feeders dominated by a few species and

individuals. Frontiers in Ecology and Evolution 5(81): 1–15.

Goddard, M. A., A. J. Dougill & T. G. Benton. 2009. Scaling

up from gardens: biodiversity conservation in urban environ-

ments. Trends in Ecology and Evolution 25(2): 90–98.

Grith, D. M., J. A. Veech & C. J. Marsh. 2016. Cooccur: prob-

abilistic species co-occurrence analysis in R. Journal of Statis-

tical Soware 69(Code Snippet 2): 1–17.

Grimm, N. B., S. H. Faeth, N. E. Golubiewski, C. L. Redman, J.

Wu, X. Bai & J. M. Briggs. 2008. Global change and the ecol-

ogy of cities. Science 319: 756–760.

Hammer, O. 2020. PAST: Paleontological Statistics. Version 4.02

reference manual. Oslo, Norway: Natural History Museum,

University of Oslo.

Hilty, S. L. 2003. Birds of Venezuela Princeton: Princeton Uni-

versity Press, 776 pp.

Horn, D. J., S. E. Fairbairn & R. J. Hollis. 2002. Factors inuenc-

ing the occurrence of birds that use feeders in Iowa. Journal of

the Iowa Academy of Sciences 109(1-2): 8–18.

Bird hierarchies and co-occurrences at urban garden in Venezuela

Ishigame, G. & G. S. Baxter. 2007. Practice and attitudes of sub-

urban and rural dwellers to feeding wild birds in Southeast

Queensland, Australia. Ornithological Science 6: 11–19.

Le Louarn, M., B. Couillens, M. Deschamps-Cottin & P. Cler-

geau. 2016. Interference competition between an invasive

parakeet and native bird species at feeding sites. Journal of

Ethology 34: 291–298.

Levin, L., L. Fajardo & N. Ceballos. 2000. Orden de llegada y

agresiones en aves urbanas en una fuente de alimento contro-

lada. Ecotrópicos 13(2): 75–80.

McBurney, S., C. Stephen, D. Jeerson & P. Zimmer. 2018.

Strategies to prevent and control bird-feeder associated diseases

and threats: Canada: Canadian Wildlife Health Coopera-

tive, 12 pp.

McClave, J. T. & F. H. Dietrich II. 1994. Statistics. 6th ed. Engle-

wood Clis, NJ: Macmillan Publishing Company, 967 pp.

Møller, A. P., P. Tryjanowski, M. Díaz, Z. Kwieciński, P. Indy-

kiewicz, C. Mitrus, A Goławski & M. Polakowski. 2015.

Urban habitats and feeders both contribute to ight initia-

tion distance reduction in birds. Behavioral Ecology 26(3):

861–865.

Orros, M. E., R. L. omas, G. J. Holloway & M. D. E. Fellowes.

2015. Supplementary feeding of wild birds indirectly aects

ground beetle populations in suburban gardens. Urban Eco-

systems 18: 465–475.

Robb, G. N., A. S. McDonald, D. E. Chamberlain & S. Bearhop.

2008. Food for thought: supplementary feeding as a driver of

ecological change in avian populations. Ecology and Environ-

ment 6(9): 476–484.

Sainz-Borgo, C. & L. E. Levín. 2012. Análisis experimental de la

función antidepredadora del agrupamiento en aves que visi-

tan una fuente de alimento. Ecotrópicos 25(1): 15–21.

Seijas, A. E. 2021. Eating the forbidden fruit? avocado con-

sumption by neotropical birds at an urban garden Boletín del

Centro de Investigaciones Biológicas 55(1): 57–69.

Seijas, A. E., A. Araujo Quintero, J. J. Salazar Gil & D. Pérez

Aranguren. 2011. Aves de la ciudad de Guanare, Portugue-

sa, Venezuela. Boletín del Centro de Investigaciones Biológicas

45(1):55–76.

Seijas, A. E., & S. F. Seijas-Falkenhagen. 2020a. Birds at a feeder

in an urban garden in Venezuela: abundances, interactions

and fruit preferences. Anartia 31: 63–77.

Seijas, A. E., & S. F. Seijas-Falkenhagen. 2020b. Fauna de mi

casa. Columbia, S.C. (USA): Kindle Direct Publishing

(KDP), 136 pp.

Shochat, E., P. S. Warren, S. H. Faeth, N. E. McIntyre & D.

Hope. 2006. From patterns to emerging processes in mecha-

nistic urban ecology. Trends in Ecology and Evolution 21(4):

186–191.

Sokal, R. R. & F. J. Rohlf. 1969. Biometry. San Francisco: W. H.

Freeman and Company, xii + 776 pp.

Tryjanowski, P., A. P. Møller, F. Morelli, W. Biaduń, T. Brauze,

M. Ciach, P. Czechowski, S. Czyż, B. Dulisz, A. Goławski, T.

Hetmański, P. Indykiewicz, C. Mitrus, Ł. Myczko, J. J. Nowa-

kowski, M. Polakowski, V. Takacs, D. Wysocki & P. Zduniak.

2016. Urbanization aects neophilia and risk-taking at bird-

feeders. Scientic Reports 6: 28575.

Tryjanowski, P., F. Morelli, P. Skórka, A. Goławski, P. Indykie-

wicz, A. P. Møller, C. Mitrus, D. Wysocki & P. Zduniak.

2015. Who started rst? Bird species visiting novel birdfeed-

ers. Scientic Reports 5: 11858.

United Nations. 2019. United Nations: World urbanization

prospects. e 2018 revisión: New York: Department of Eco-

nomics and Social Aairs. Population Division (ST/ESA/

SER.A/420), xx + 103 pp.

Veech, J. A. 2013. A probabilistic model for analysing species co-

occurrence. Global Ecology and Biogeography 22: 252–260.

Veech, J. A. 2014. e pairwise approach to analysing species co-

occurrence. Journal of Biogeography 41: 1029–1035.

Verea, C., N. Espósito & M. Lentino. 2016. Paraulatas de Ven-

ezuela. Columbia, S. C.: Kindle Direct Publishing, 336 pp.

Wallace, M. P. & S. A. Temple. 1987. Competitive interactions

within and between species in a guild of avian scavengers. e

Auk 104: 290–295.

Wojczulanis-Jakubas, K., M. Kulpinska & P. Minias. 2015. Who

bullies whom at a garden feeder? Interspecic agonistic in-

teractions of small passerines during a cold winter. Journal of

Ethology 33: 159–163.