Experimental study of social signaling through delayed plumage maturation in a colony-nesting seabird
Data files
Mar 18, 2026 version files 105.57 KB
-
analysis.r
13.42 KB
-
behavior_dictionary.csv
341 B
-
nest_data.csv
9.56 KB
-
README.md
10.59 KB
-
trial_data.csv
60.99 KB
-
trial_dates.csv
10.68 KB
Abstract
Delayed development is a widespread evolutionary strategy that can reduce competition among social animals. Many seabirds exhibit delayed plumage maturation, in which young birds spend years in visually distinct predefinitive plumages before attaining the definitive plumage of adults. Previous work hypothesized that predefinitive plumages may function to reduce aggression towards young seabirds at breeding colonies, an idea known from other lineages of birds as the status signaling hypothesis. We tested this hypothesis with visual stimulus experiments at a breeding colony of American Herring Gulls (Larus smithsonianus). We presented painted models of four different plumage classes (first-cycle predefinitive gull plumage, third-cycle predefinitive gull plumage, definitive gull plumage, and a Canada Goose as a control) and measured the aggressive responses of breeding adults at their nests. Breeding gulls responded with significantly less frequent, lower, and slower aggression toward the first-cycle plumage and control stimuli compared to the definitive plumage. There were no significant differences in response towards the third-cycle plumage compared to definitive plumage. These results provide evidence for a status signaling function of the extensive brown plumage characteristic of first-cycle gulls, which may reduce aggression in foraging flocks, on the wintering ground, or during young birds’ brief visits to colonies. Given the lack of differences in aggressive responses to third-cycle versus definitive plumage, our data do not settle the issue of why third-cycle gulls, which spend more time at colonies than younger birds, also molt in brown, predefinitive plumages. Future research can investigate how immature seabirds—including third-cycle gulls—may combine plumage, posture, and behavior to shift the dynamics of social behavior at breeding colonies.
This repository contains the data and scripts needed to replicate the analyses in our manuscript.
Data file descriptions
trial_data.csv
Each row is a single experimental trial (one treatment at one nest)
- Trials_ID: (int) The unique identifier for the trials (1-120)
- Trial_Number: (int) Is this the first, second, third, or fourth treatment presented at a nest?
- Stimulus: (chr) The stimulus presented at the nest
- [Format: PlumageTreatment-VisualExemplar-AudioExemplar]
- Unique_Nest_ID: (chr) Year plus nest number combined for unique identifier
- Nest: (chr) Nest number (some numbered from other studies, may not be unique across years)
- Year: (int) Year of trial (2023 or 2024)
- Time: (hh:mm:ss) The time of day that the trial was conducted
- Weather: (chr) The weather during the trial
- NOTE trials were only conducted during 'regular' daily weather at the site, including clear, overcast, fog, and light rain
- After_Trial: (chr) Anecdotal notes of events immediately following the trial
- Events, multiple columns: (chr) The details of any events occurring during a trial
- Each event is inserted in one column
- [Format: Parent#;Behavior;DistanceFromModel;TimeToResponse]
- Parent# indicates whether this was parent 1 (P1) or parent 2 (P2) during a trial. If only one parent was present, only P1 is noted
- If both parents were present, we sometimes assigned a sex (e.g., P1=M, P2=F) based on visual discrimination of size. However, this cannot be done when only one parent was present during a trial. We did not analyze sex differences
- Because all models were placed ~0.75 m from the nest, incubation is always ~0.75 m from the model. Contact is always 0 m from the model, etc. We did not analyze distance of behaviors from the model
- Each event is inserted in one column
nest_data.csv
Each row is a nest
- Unique_Nest_ID: (chr) Year plus nest number combined for unique identifier
- Nest: (chr) Nest number (some numbered from other studies, may not be unique across years)
- Year: (int) Year of trial (2023 or 2024)
- Date_Floated: (mm/dd/yyyy) The date the egg was floated, to test whether incubated eggs were close to hatching. Generally ~24 hr before the first trial
- Egg_Float_Diameter_mm: (num) The diameter of the portion of the egg floating above the surface of the water. This can be used to more precisely age the eggs
- NOTE these were not analyzed
- Latitude: (dd.dddd) Latitude location of the nest on Kent Island
- Longitude: (dd.dddd) Longitude location of the nest on Kent Island
- Nearest_Neighbor_Distance_m (chr) The distance to the nearest neighboring nest
- NOTE These were not analyzed in the model. We include nest as a random effect.
- NOTE Maximum distance indicated by >10 m
- Notes: (chr) Observer notes about the nest
- Clutch_Trial_1: (chr) The number and status of clutch during the first trial
- NOTE If only a number is given, indicates number of eggs
- NOTE If codes are given (1C+2E) indicates number of small chicks hatched from eggs (still in nest) plus number of unhatched eggs
- NOTE Also includes details on whether eggs appeared to be actually hatching
- NOTE Daily status was not analyzed in the model. No chicks fledged during the trials
trial_dates.csv
Each row is a series of four trials, indicating the randomized treatment/exemplar orders presented at a single nest
- Trials_ID: (int) The unique identifier for the trials (1-120)
- Stimulus_1: (chr) The stimulus presented at the first trial
- [Format: PlumageTreatment-VisualExemplar-AudioExemplar]
- Stimulus_2: (chr) The stimulus presented at the second trial
- [Format: PlumageTreatment-VisualExemplar-AudioExemplar]
- Stimulus_3: (chr) The stimulus presented at the third trial
- [Format: PlumageTreatment-VisualExemplar-AudioExemplar]
- Stimulus_4: (chr) The stimulus presented at the fourth trial
- [Format: PlumageTreatment-VisualExemplar-AudioExemplar]
- Unique_Nest_ID: (chr) Year plus nest number combined for unique identifier
- Nest: (chr) Nest number (some numbered from other studies, may not be unique across years)
- Year: (int) Year of trial (2023 or 2024)
- Date_Trial_1: (mm/dd/yyyy) The date the first trial was conducted
- Date_Trial_2: (mm/dd/yyyy) The date the second trial was conducted
- Date_Trial_3: (mm/dd/yyyy) The date the third trial was conducted
- Date_Trial_4: (mm/dd/yyyy) The date the fourth trial was conducted
behavior_dictionary.csv
Each row is a behavioral response
- Behavior: (chr) The simple classification used for the behavior in the trial datasheet
- NOTE Scans of original datasheets are available upon request. These include additional and potentially subjective details on posture (e.g., neck raised), attitude (e.g., nervous), direction of attention for vocal behaviors (e.g., "yeow at observer" which is more correctly included here as just "yeow") etc., that were not analyzed in our behavioral categories here
- Observer interaction = direct physical interaction with the observer (e.g., swooping at an observer's head) rather than the experimental stimuli
- Flushed = fled the nest completely, not returning for the duration of the trial
- Conflict with neighbor = direct physical interaction or symbolic display conflict with a neighboring bird rather than the experimental stimuli
- Standing = standing in place. Includes a variety of neck and body postures, but no additional behaviors
- Incubating = sitting on the nest to warm eggs. Includes a variety of neck and body postures but no additional behaviors
- Preening = tending feathers
- Brooding = sitting or standing on the nest to cover very young chicks still in the nest. Includes a variety of neck and body postures but no additional behaviors
- Regurgitation = bringing up swallowed food, which was subsequently eaten or delivered to very young chicks
- Eating = consuming food in or near the territory
- Sitting = fully reclined outside of the nest
- Yawn = yawning, which also tends to make the observers yawn
- Kek = an alarm call ("kek-kek-kek"), performed often in cases of nest or territory threats that indicates defense or retreat. Also called the 'alarm', 'gakkering', or "eh-eh' call
- Nervous flight = a short, fluttering, backwards or upwards butterfly flight performed as a uncertain startle response
- Yeow = an alarm call ("yeow!"), performed often in cases of nest or territory threats. Also called the 'yelp,' 'keow', or 'long-call note'
- Approach = a quick walk towards the experimental stimuli without direct indication of aggression, retreat, or alarm
- Pacing = short, quick walks back and forth around the territory or near the experimental stimuli, often accompanied by alarm calls (yeow or kek)
- Circling flight = short or long circling flights above the territory at a variety of heights, often accompanied by alarm calls (yeow or kek) or nervous flights, swoops, or attacks towards the experimental stimuli
- Mew = a prolonged, whining call associated with social interactions, usually performed with a distinct posture featuring a downturned head, stretched neck, and open bill. In broader social contexts the mew call is associated with all of courtship, nest care, and aggressive responses to social threats (e.g., territory neighbors), but in the context of our trials mew calls were not performed towards partners or chicks
- Choking = a prolonged display that resembles regurgitation, including a posture with a downturned head and bill and a distinct 'choking' vocalization. In broader social contexts the choking display is associated with all of courtship, nest care, and aggressive responses to social threats (e.g., territory neighbors), but in the context of our trials choking displays were not performed towards partners
- Trumpet = a long call performed with a distinct posture that begins with the head facing down and ends with the neck bent backwards over the body. Also called the 'long call.' Often performed by birds asserting their presence at territories
- Grass pull = an agonistic display involving the pulling or tossing of grasses, sticks, or wrack, resembling the collection of nesting material. Often performed by birds establishing territorial boundaries
- Swoop = a fast, direct, aggressive approach (no direct contact) to the experimental stimulus by air. Often accompanied by outstretched foot or bill
- Lunge = a fast, direct, aggressive approach (no direct contact) to the experimental stimulus by ground. Often accompanied by short jumps, an outstretched neck and slightly spreading of wings
- Feint = a fast, direct, aggressive peck (no direct contact) towards the experimental stimuli performed from a standing position
- Attack = direct physical contact with the experimental stimuli, including both by air or by ground, generally involving contact through either the foot (by air) or the bill (by ground)
- Aggression_Level: (num) The aggression category for each behavior
- -1 = Excluded from analysis, including birds that fled the nest completely, direct aggression towards observer or neighbors rather than the experimental stimuli
- 0 = Relaxed behaviors, including baseline behaviors such as incubating, standing, or walking around that do not indicate any additional alarm or aggression towards the model
- 1 = Alarm behaviors, including defensive, fear, or surprise reactions that indicate response to a nest or territory threat without performing direct aggressive or offensive behaviors
- 2 = Social aggression, including symbolic behaviors used in the context of territorial conflict, defense of courtship partners, or exclusion from nest sites
- 3 = Physical threat, including lunges or direct threat to the model without making final physical contact
- 4 = Physical attack, indicating direct physical contact with the model
Running the code
After installing the necessary packages (see below), execute the analysis.r script.
Figures will be saved as FIGURE_X.png images in the working directory.
Model summaries and descriptive statistics will be saved as results_log.txt in the working directory.
Software requirements
Scripts are written for R v4.4.1.
Required R packages can be installed in R with:
install.packages(c("tidyverse",
"ordinal",
"lme4",
"nlme",
"patchwork"))
Hand-painted, plastic model decoys resembling four plumage treatments (first-cycle gull plumage, third-cycle gull plumage, and definitive gull plumage, plus Canada Goose control) placed near the nest of breeding American Herring Gulls (Larus smithsonianus) during the incubation season on Kent Island, New Bruswick, Canada. Rate of aggression, maximum level of aggressive response, and time to maximum aggression recorded.
All experimental procedures were approved by the Bowdoin Scientific Station, Bowdoin College (Protocol #2023-0008), and Yale University (Protocol #2024-20453). Birds were not captured or handled for this experiment. We designed the experiment to minimize the length of exposure of the animals to signals that could be perceived as aggressive.
