Material & Methods


During this study, I observed a group of 10 mandrills (Mandrillus sphinx) living in ARTIS royal zoo of Amsterdam between July 2022 and February 2023. The group of mandrills consists of one adult male (21 years old), six adult females (5-19 years old), one adolescent male (4 years old), one adolescent female (3 years old), and one juvenile male (2 years old). Male mandrills reach sexual maturity between the age of four and nine, and female mandrills reach sexual maturity when they are on average 3.6 years old. During the study, mandrills had day and night access to an indoor enclosure and an outside enclosure of 157 m2.

An artificial rock formation is in the middle of the outside enclosure and gates in this rock formation are used by the mandrills to enter their inside enclosure. During observation, the mandrills could only enter their outside enclosure using one gate, ensuring that every mandrill would enter the outside enclosure in the same place.

The mandrills are fed three times per day, (between 9 AM and 10 AM, between 1 PM and 2 PM, and between 4 PM and 5 PM) on a mainly vegetable based that was supplemented with monkey pellets (Primacon), fruit, insects, honey, peanut butter, boiled eggs, nuts, fresh branches, and wildflowers. Caretakers would provide some of this food in an enrichment circa 5 days per week. 

Synchrony cue experiment

Schematic overview of mandrille enclosure. C represents carrot locations and G represents grape locations. The red square represents the gate between the inside and outside enclosure

To investigate mandrills’ use of synchrony cues, fixed foraging locations for two food types were introduced in the enclosure, and mandrills were presented with one of these food types as a cue. To create fixed foraging locations where mandrills could always expect to find the same type of food, there were six locations in the mandrill enclosure where either 60 grams of grapes or 60 grams of carrots were buried in the soil every morning of a test day. These locations did not change throughout the study, and at each location, the same type of food was always buried. To help the mandrills to locate the food and to simulate a natural foraging situation, wherein mandrills associate the visual characteristics of a plant with the fruit it carries, willow branches were set upright in the ground at each location after the food was buried.

To provide a synchrony cue that could motivate a foraging choice, information about the food that would be buried in the enclosure on that testing day was given at the entrance to the outside enclosure. This cue would be either carrots or grapes. The carrots or grapes were scattered around the gate between the inside enclosure and the outside enclosure. The mandrills could only enter their outside enclosure through this gate, and all mandrills always entered the outside enclosure after they were done feeding in the inside enclosure, ensuring that every mandrill noticed the food when they entered their enclosure. Eating, seeing, or interacting with this food could serve as a cue that could motivate the mandrills to visit the location in their enclosure corresponding to that food type. 

A testing day was either a “carrot day” or a “grape day”. On a “carrot day”, 60 grams of carrots were buried at carrot locations and there was no food present at grape locations. On a “grape day”, 60 grams of grapes were buried at grape locations and there would be no food present at carrot locations.

On the morning of a testing day, after everything was set up in the outside enclosure, the observer would leave the enclosure, and a zookeeper gave the mandrills access to their outside enclosure. The observer then noted which individuals visited which locations first, and the order a mandrill visited all locations. Upon visiting a location, a mandrill would usually remove the willow branch and start digging. A visit to a location was recorded when an individual started digging. Once a location was visited, further visits at that location were not recorded during that testing day. If an individual had already visited a location, no further visits of this individual were scored during that testing day. This is because I was only interested in the first visit of an individual. The use of only first choices excluded possible effects of individuals following other individuals instead of associating a synchrony cue with a location in its enclosure. A correct foraging choice was defined as:

  1. A mandrill makes a first visit at a carrot location when carrots are provided as cue.
  2. A mandrill makes a first visit at a grape location when grapes are provided as cue. 

Olfaction Experiment

Puzzle feeders used in the study

To study how mandrills use olfaction to locate food from a distance, four identical puzzle feeders were placed at 4 of 6 possible random locations in the outside enclosure on each testing day. The puzzle feeder had two purposes: to spread an olfactory cue, just like ripe fruit would in a rainforest, and to challenge the mandrills to have to manipulate the feeder to get to the reward, this is intended to be analogous to a mandrill manipulating fruit from a tree to eat it in the natural habitat. Each puzzle feeder had a food compartment and an olfactory compartment.

The feeder was attached to the outer wall of the outside enclosure, approximately 50 centimeters high, with a screw so the feeder could spin on its axis. If the mandrills turned the feeder around, food fell out of the holes in the top lid. .

Two of the 4 puzzle feeders placed in the outside enclosure would have an olfactory cue and a food reward (olfactory feeder), and the two other puzzle feeders would be completely empty (control feeder), this was decided using a randomizer. The food reward was dried mango. The olfactory cue was two sheets of paper towel soaked in mango juice. To a human observer, the feeder would smell strongly of mango after adding the olfactory cue to the feeder. There were six possible locations in the outside enclosure where feeders could be placed. 

Schematic overview of the mandrill enclosure. Each P represents one of six positions where a puzzle feeder could be placed. The red square represents the gate mandrills use to enter their outside enclosure.

On each morning of a testing day between 9 AM and 10 AM, the puzzle feeders were placed in the outside enclosure while the mandrills would be voluntarily herded to the inside enclosure to receive food. When the mandrills entered the outside enclosure, they were recorded for an hour. To investigate the effect of olfactory cues on foraging choices, first visits to a puzzle feeder were recorded. A visit was defined as an individual approaching the feeder and either spinning or looking. Spinning is defined as touching a stationary puzzle feeder and manipulating the puzzle feeder to have it spin on its axis. Looking is defined as standing on one’s hind leg to reach the top of the puzzle feeder and peering into the puzzle feeder. To record only foraging choices made based on sensing olfactory cues and not based on observing other mandrills, only the first visits at a puzzle feeder were recorded. So, when a puzzle feeder had been visited by a mandrill, no further visits to this puzzle feeder were recorded for that day. So, each recorded visit to a puzzle feeder could be either the first visit by an individual on that day or a subsequent visit by an individual who had already visited at least one puzzle feeder earlier that day. The foraging choice was either correct or incorrect. A correct foraging choice is defined as:

1) An individual makes a first visit at an olfactory feeder.

2) An individual that has previously only visited an olfactory feeder subsequently visits the second olfactory feeder.

3) An individual that has visited both olfactory feeders subsequently visits a control feeder. 

An incorrect foraging choice is defined as:

1) An individual makes a first visit at a control feeder.

2) An individual that has previously visited a control feeder subsequently visits an olfactory feeder.

3) An individual that has previously visited an olfactory feeder subsequently visits a control feeder while there is still an olfactory feeder that remains unvisited.


I used generalized linear mixed models for all analyses. See my thesis for more details!

Go to Results