Dataset DOI: 10.5061/dryad.cc2fqz6kh

Description of the data and file structure

We combined reanalysed field observations with laboratory experiments to test how cleaner fish (Labroides dimidiatus) service quality varies with hunger state. Field data from Ras Mohammed National Park, Egypt, included 16 adult cleaners observed for 3–4 hours each, recording client species, interaction duration, and jolts (a proxy for cheating). Cleaners were classified as “satiated” when interacting with residents immediately after high feeding sequences of visitor clients, and as “hungry” otherwise. Jolt rates per client species were then compared between these states.

In laboratory experiments at Lizard Island (2010, 2017, 2018), adult female cleaners interacted with client fishes or Plexiglas plates offering preferred prawn vs less-preferred flake food. Treatments manipulated hunger (recently fed vs food-deprived), and interactions were video-recorded and scored blind for duration and jolts, or for food choice.

Files and variables

Please refer to the manuscript for context, data collection methods, and statistical analyses.

For questions or to notify the authors of any errors, please contact Zegni TRIKI (zegni.triki@gmail.com).

data_field.csv

Summary data table from field observations of natural cleaning interactions between cleaner fish and their clients.

• species: Client fish scientific species name

• site: reef site

• cleaner_state: the satiation level of cleaner fish

• jolts: total number of client jolts recorded during the observations

• interaction_duration: the duration of cleaning interactions in seconds

• jolt_1s: jolt frequency as the number of jolts per 1 sec of interaction

data_all_client.csv

Summary data table from laboratory experiments with real client fish. Missing data: NA

• experiment: the experiment label

• year: the year of the experiment

• site: the reef site from which fish (cleaner and clients

• session: morning and afternoon test sessions

• cleaner_ID: cleaner fish identity

• client_ID: client fish identity

• cleaner_id_figure: cleaner fish identity used to generate the corresponding figure

• client_species: the Latin name of the client fish species

• treatment: hungry or satiation treatment

• total_interaction_time: total of cleaner-client interaction duration in seconds

• nb_jolt: number of client fish body jolts

• jolt_1s: frequency of client fish body jolt per 1sec of cleaning interactions

data_all_plate.csv

Summary data table from laboratory experiments using Plexiglas plates as surrogates for client fish.

• experiment: the experiment label

• year: the year of the experiment

• site: the reef site from which cleaner fish were captured for lab experiments

• cleaner_ID: cleaner fish identity

• flake_content: the percentage of fish flakes in the flakes-prawn mix (empty cells- lack of flake content).

• treatment: the manipulation of the cleaner fish satiation level

• total_trial: total number  of test trials

• flake_item: total number of flake items consumed by the cleaner fish over the test trials

• prawn_item: total number of prawn items consumed by the cleaner fish over the test trials

ratio: total flakes items divided by total prawn items

Further details on the experiements, dates and sample sizes (the same information is reported as Table 1 in the main text):

Experiments with Real Client Fish

• Exp. 1 (2010): 16 Labroides dimidiatus each paired with Acanthurus curacao, Scarus bilineatus, and Chaetodon striatus; 2 × 10 min sessions/day for 6 days; tested hungry vs. satiated, client species counterbalanced.
• Exp. 2 (2017): 20 L. dimidiatus + 20 C. striatus (20 pairs); 2 × 15 min sessions/fish; hungry vs. satiated.
• Exp. 3 (2018): 20 L. dimidiatus + 20 C. striatus (20 pairs); 2 × 15 min sessions/fish; hungry vs. satiated.

Experiments with Plexiglas Plates (Client Surrogates)

• Exp. 4 (2004): 16 L. dimidiatus; 14 trials over 2 days (7 per treatment/day); hungry vs. satiated.
• Exp. 5 (2017): 18 L. dimidiatus; 20 trials over 2 days (10 per treatment/day); hungry vs. satiated.
• Exp. 6 (2018): 18 L. dimidiatus; 16 trials over 2 days; hungry vs. satiated.
• Exp. 7 (2018): 20 L. dimidiatus; 40 trials over 4 days (10 per treatment × flake mix/day); hungry vs. satiated × flake concentration (10% vs. 40%).

code_stats_figures.R:

The code displays summary statistics and the first few rows of data_plate and data_client to inspect the data. It then converts treatment and flake_content in data_plate to factors for categorical analysis, preserving the original level order for flake_content