In the test I used 8 meerkats (Suricata suricatta) 4 males and 4 females that all lived at Kolmården Wildlife Park in Norrköping, Sweden. The group consisted of two unrelated adult pair (it is unknown if the male is the father to the offspring due to the fact that the year before they had one more adult male that died which also could be the father) and 6 offsprings (figure 1). The animals were tested individually during the whole experiment and were recognized either by distinct natural markings or fur clippings. The meerkats was kept in both an inside and an outside enclosure in the beginning of the experiment, however when the weather got colder they were mainly kept in the inside enclosure. The meerkats were normally feed with a standard dry cat feed and some carrots and apples. In the summer they may catch insects when they have access to the outdoor enclosure.

I preformed a 3-chooise food test on the meerkats food preferences. I tested 9 different food items: 3 typs of animal matter (cricket, zophoba larva and boiled egg white), 3 types of vegetables (carrot, tomato and cucumber), 3 types of fruit (banana, apple and orange). The food items were prepared into equally sized pieces. If a zophoba larva or a cricket was presented the other items was cut to mach them. If both a zophoba larva and a cricket was presented the third item was cut to something in between.

The food items were presented 3 at the time and all 9 food items was combined in all possible ways, making 84 combinations (thereby at one presentation it could for example be 2 fruit and 1 animal matter or 1 from each category). Every combination was presented ones for every individual so that one food item was presented to the meerkats totally 28 times. The order of the combinations and in which bowl (left, middle or right) the food item were presented had been pseudo-randomized before the start of the experiment. This ensured that no food item appeared in two tests in a row and so that it was as long as possible before the food item appeared at the same position again.

Two tests were performed each day between 24^th of August and 25^th of October. The tests were performed in a plywood test arena (figure 1) which was netted to prevent escape of the testing meerkat. The corridor were the tests were performed was 40x100cm and the 3 bowls were the food items were presented were ø12cm. A sluice was placed in front of the entrance door to make sure that only one meerkat at the time was let in to the arena. At the beginning of each session (the test for one meerkat) one piece of food was placed in each bowl.

The food choices were scored; 3 points to the food item that was picked first, 2 points to the food item picked second, 1 point for the food item picked last and 0 points for a food item not picked. For each food item the total group score and the individual score for each of the eight meerkats were calculated (the maximum score for a food item in the whole group would be 672 and for the individuals 84). A control for side preferences was also performed by summarizing the scores and making a mean for each bowl (left, middle and right), regardless of the food item, for each date (8/24/2010 to 10/25/2010). The nutritional value of all the food items was collected from tables (Swedish Livsmedelsverket for the fruit, vegetables and egg and from (Grubco.com)[1] for the crickets and the zophobas).