preprint: https://doi.org/10.5061/dryad.cnp5hqcdg

peer-reviewed article: https://doi.org/10.1093/beheco/arae055

We investigated the interplay between visual information and olfaction in two experimental feeding contexts. In experiment 1 (pipe experiment), we provided Barbary macaques with food or non-food items in either a visible or non-visible condition (drainage pipes with transparent or opaque cans) to assess if visibility affected the propensity to sniff at the setup during any stage of exploring it, if visibility affected at what stage of the exploration monkeys used olfaction, and the interplay between visual condition, olfaction and the (attempted) retrieval of items from the setup. Pipe setups were mounted at the bottom of trees at 12 locations within the monkeys’ enclosure (in 3 blocks of 4 locations) for 7 full days (with data for some locations expanding until the morning of the 8th day) and were monitored continuously for any approaching monkeys while mounted.
In experiment 2 (popcorn experiment), we manipulated the visual appearance of a familiar food item, i.e. popcorn, with food dye to assess how the salience in visual information affected sniffing behavior. For this purpose 32 adult focal individuals of both sexes received a total of 8 pieces of popcorn in two trials à 4 pieces, half of which were dyed and the other half left unmanipulated.

Description of the data and file structure

Data are structured as two tables in csv format (one for each experiment) and contain the transcribed video data of all scored approaches to the setup (pipe experiment) and all pieces of popcorn offered to the monkeys (popcorn experiment).

Table ‘data_pipe_experiment’ contains 1556 scored approaches to a pipe setup, including approaches without any interactions with the setup and approaches in which the individual monkey could not be identified.

The table contains a continuous number for each approach (‘nr’), the experimental block in which the approach was recorded (1, 2 or 3, ‘block’) and the day of the experiment per block (1-8, ‘day’), the date as dd:mm:yyyy (‘date’), date as Julian day (starting with 1 on Jan. 1st 2021, ‘JulianDay’), the time of the approach as hh:mm (‘time’), the hour of the day (‘hour’), the location where the setup was mounted (4 locations per block, i.e. 12 in total, ‘location’) and whether or not all approaches to the setup at a given location have been coded (‘fully-coded’). Columns J, K, L and M comprise the identity (‘ID’), sex (‘sex’), birthyear (‘cohort’) and group membership (‘group’) of the approaching monkey. If any of this information could not be unambiguously determined, ‘NA’ was entered in the respective cells. The table further contains the human rater who scored the behaviors (‘coder’), the orientation of the pipe’s opening at the time of the approach (with 0 denoting the top, ‘orientation’), the experimental condition (transparent or opaque, ‘condition’), the content of the setup at the time of the approach (peanuts, popcorn, stones, feces or empty, ‘content’), the number of items in the setup at the time of the approach (‘n_content’), the time (in minutes) since the content was filled into the setup (‘delta_fill’), the time (in minutes) since the previous approach to the setup (‘delta other’), the time (in minutes) since the approaching individual last approached the setup (‘delta_self’) and how many times the approaching individual has already visited the setup (including the current approach, ‘nth_approach’). Note that ‘delta_fill’, ‘delta_other’ and ‘nth_approach’ could only be accurately assessed for fully-coded locations. The table further contains the behaviors shown during an approach: whether or not the monkey showed any kind of interaction with the setup (‘interact’), whether it looked, sniffed, grabbed or showed any other form of interaction (‘look’, ‘sniff’, ‘grab’, ‘other’), the exact behavioral sequence shown during the approach (l=look, s=sniff, g=grab, o=other, ‘sequence’), the first behavior shown at the setup (l=look, s=sniff, g=grab, o=other,’firstBehav’), whether or not the first behavior shown was a sniff (scored only for approaches in which sniffs occurred, ‘sniff_first’), the number of sniffs shown during the approach (‘n_sniffing’) and whether or not a monkey sniffed before the first grab into the setup (‘sniff_before1g’). ‘sniff_before1g’ is labelled as yes if the monkey sniffed before any grab occurred or if a sniff, but no grab occurred, and as no if the monkey did not sniff before any grab occurred or did not sniff at all. In columns W to AC, AE and AF, NAs denote cases in which the respective information could not be retrieved from the video. In column AD, NAs denote the approaches in which no sniffs occurred. Columns AG to AI denote with ‘yes’ which approaches have been used for models 1 and 4 (‘model1_4’), model 2 (‘model2’) and/or model 3  (‘model3’) in the statistical analyses.

Table ‘data_popcorn_experiment’ contains the interaction of the monkeys with all 256 pieces of popcorn presented to the monkeys. The table contains a continuous number for each piece of popcorn presented during the experiment (‘nr’), trial number (labelled consecutively across monkeys, ‘trial_nr’), the session per individual (i.e. first or second trial for a given individual, ‘session’), the date as dd:mm:yyyy (‘date’), the time of the trial as hh:mm (‘time’), the hour of the day (‘hour’) and the focal animal’s identity (‘ID’), sex (‘sex’), birthyear (‘cohort’) and group membership (‘group’). ‘popcorn_nr’ denotes whether it was the 1st, 2nd 3rd or 4th piece of popcorn retrieved within a given trial, and ‘total_nr’ the number retrieved across trials with the same individual (i.e. 1-8). The table further contains the color of the respective piece of popcorn (‘color’), whether or not the monkey sniffed the piece (‘sniff’) and whether or not it consumed the popcorn (‘consumed’).

Sharing/Access information

If required, data and further information about the data set can also be obtained from the author at
brigitte_schloegl@eva.mpg.de or brigitte.schloegl@uni-leipzig.de

Change log

Table ‘data_pipe_experiment’ has been uploaded anew because of issues with readability. To solve the issue the file was saved anew but the content has remained identical to the earlier version.