Materials and methods Directory UMM :Data Elmu:jurnal:A:Applied Animal Behaviour Science:Vol69.Issue2.Sept2000:

2. Materials and methods

2.1. Subjects In order to form a uniform group of nonstressed animals, with no a priori changed sensitivity to the taste of food items, we sought domestic cats which had — in principle Ž . — to comply with the following criteria: 1 be of known origin and not separated from their mothers before the age of 8 weeks, since early separation may change the Ž . Ž . sensitivity of the stress system Liu et al., 1997 ; 2 be older than 1 year to discard any Ž . effects due to maturation; 3 be neutered, since, e.g. the oestrus cycle may affect Ž . Ž . Ž . palatability Cabanac, 1979 ; 4 be able to go outside regularly; 5 not have had a history of serious illnesses, allergies or problem behaviour such as overgrooming, Ž aggression or overeating, which may indicate a history of chronic stress Bolka, 1984; . Ž . Willemse et al., 1994 ; 6 not being a pedigree cat, since some breeds may have an Ž . altered sensitivity to stressors and rewards Van den Bos, in preparation . Originally, 27 cats were selected. However, four cats were rejected after or during testing. Of one cat, it proved to be very likely that it had had a history of overgrooming. The other three cats proved to be too nervous of the observer or would not eat at all Ž during the observations. The remaining 23 cats 16 castrated males, six spayed females . and one intact female were recruited from 18 households, varying between students’ houses and single-family homes, located in urban settings in the Netherlands. Of these cats, 10 were the only cat in the household, 12 were from two-cat households and one was from a three-cat household. Ten cats were living in households together with other animals, e.g. dogs and rodents. Four cats did not comply to all criteria: one cat was taken out of a cat shelter at the age of one. According to the owner, direct examination by the veterinarian had not revealed any signs of stress, e.g. due to early separation of the mother. One cat had only free access to the roof of the block of houses where it lived and was allowed to go outside on the street about one night a week. One 2-year old female cat was not spayed. Ž . She was administered anticonception ‘cat pill’ the first year and had not been in oestrus the second year. One cat was born in Africa and brought to Europe as the owners moved. Its mother was a European shorthair cat, the father was of unknown origin and may therefore have been a wild cat. However, it has been found that male cats may pass on the tendency to be friendly to people to their offspring without ever coming into Ž . social contact with them Turner et al., 1986 . This particular cat was very sociable Ž . towards humans suggesting that its father was a domesticated nonwild cat. In order to be sure that the data of these cats would not confound the data of their respective Ž . groups, an outlier test was performed per group see below . Ž . The cats were divided over three groups Table 1 . Because four cats were rejected Ž . afterrduring testing, the groups differed in size 5, 8 and 10 cats, respectively . Groups Ž A and B consisted of cats that were fed one to three meals a day dry food or canned . food with no other food supplied during the rest of the day. Group C consisted of cats that had ad libitum access to dry cat food with most cats also fed canned cat food one or two times a day. Table 1 Characteristics of group A, B and C Group No. of Cats No. of mealsrday First food Second food A 5 1–3 Normal Experimental B 8 1–3 Experimental Normal C 10 Ad libitum Experimental Normal 2.2. Test procedure All cats were tested twice at their owners homes. During each test session, the cats were presented two types of food: the food they were usually fed at that time of day Ž . Ž normal food: NF and one of the two experimental foods one food being more Ž . Ž .. flavourful than average MFF , the other one less flavourful than average LFF . Cats of group A were presented NF as the first food and an experimental food as the second food. Cats of groups B and C were presented an experimental food as the first food and NF as the second food. Each cat was fed MFF during one test session and LFF during the other. The intersession interval was at least 2 weeks. To be sure that the cats’ response to the different types of experimental food was not influenced by the order in Ž which they were presented, each group was divided into two subgroups group A: three . versus two; group B: five versus three; group C: five versus five . One subgroup was fed MFF during the first session and LFF during the second session; the other subgroup was fed MFF and LFF in the reverse order. All food items were presented at room temperature. NF was presented in the cat’s own feeding bowl, MFF or LFF was presented in a feeding bowl that was brought by the observer. It should be noted that cats of group A and C were fed the experimental food while Ž . Ž . they were supposed to be more or less satiated created through different conditions , whereas cats of group B were fed the experimental food while they were supposed to be Ž . more or less hungry. 2.3. Food items NF was either dry or canned cat food of a brand and taste the cat was used to be fed Ž . all brands were commercially available . The normal amount of food was presented. If the cat was only fed dry food ad libitum, a full feeding bowl was supplied. MFF was Ž ; 50 g of Sheba Trout a commercially available premium quality product on a meat . and fish base . This product proved in other experiments to be highly appreciated by cats Ž . Van den Bos et al., unpublished data . LFF was ; 50 g of a cheap canned cat food that Ž . was mixed with ; 30 drops of orange essence Genfarma, Houten, the Netherlands . Ž . Pilot experiments had shown this to be a useful LFF as it turned out to be less or un- flavourful, but not uneatable. Furthermore, it was a completely harmless product. Ž It should be noted that in contrast to experiments in rats Berridge, 1996; Berridge . and Robinson, 1998; Grill and Berridge, 1985 , neither sucrose solutions nor quinine solutions were used to arrive at hedonic and aversive taste reactivity patterns. The reasons for this are as follows. First, cats appear to be rather insensitive to sucrose’s Ž . taste Bartoshuk et al., 1971; Beauchamp et al., 1977; Bradshaw, 1992 . Second, Ž . although cats do show an aversion to quinine Bradshaw et al., 1996 , the use of quinine would have discouraged the owners to allow their cats to participate in experiments. The Ž same applies to the use of sweet- and bitter-tasting amino acids White and Boudreau, . 1975 . 2.4. Recording procedure All observations were done by the second author and were arranged in advance. The observations were done on weekdays between 1500 and 2100 hours, around the time the Ž . cats were usually fed. Cats of group C with dry food ad libitum were observed at the time that they were fed canned food. If only fed dry food ad libitum, no special time point was chosen. Ž Before observations started, the owners defined as the persons normally feeding the . cats were instructed to present or take away the feeding bowls at a signal of the observer, and were asked to try not to talk loudly or make sudden noises and not to stroke the cats during the observations. The cats were not able to leave the room during the observations. Cats of multi-cat households were fed separately to prevent cat–cat interactions. The test sessions were composed of two parts: presentation of the first food item at Ž . t s 0 min maximal presentation duration: 10 min and presentation of the second food Ž . item starting at t s 15 min latest maximal presentation duration: 10 min . The be- haviour of the cats was recorded throughout the session. Observations started when the first food item was presented to the cat and the cat was aware of it. For both food items, if necessary, the cat’s attention was drawn by calling the cat and showing the food item; in a few cases, when the cat was reluctant to come, it was picked up by its owner and placed next to the feeding bowl. In practice, this never happened for the first food item, and only a few times for the second. For both food items, 5 min of post-meal behaviour Ž . period cf. Bradshaw and Cook, 1996 were sampled as follows: either when cats Ž stopped eating and did not return to the food bowl within 5 min which was subse- . quently taken away after 5 min of recording , or in case cats ate nothing at all, from the Ž time-point on that they had sniffed the food and went away again the food bowl was . taken away after 5 min . When the cats did not finish eating after 10 min of food Ž . presentation see above , or kept returning to the bowl before 5 min of recording passed and after 10 min of food presentation had passed, the bowl was taken away and 5 min were sampled from thereon. The latter happened only once. The maximal test session duration was 30 min. Total duration, frequencies and sequences of behaviour patterns were recorded during Ž the entire period. Recording was done using a handheld computer Psion Workabout; . Ž Psion, London, England loaded with The Observer, version 3.0 Noldus Information . Technology, Wageningen, the Netherlands . Afterwards data were exported to a normal PC where they were edited and analysed. The behavioural patterns which were recorded are shown in Table 2. A rough estimate of the amount of food eaten was made after each Table 2 Ž . Ethogram Van den Bos, 1998a; Bradshaw and Cook, 1996; UK Cat Behaviour Working Group, 1995 Element name Code Description Stand St Cat stands on all four paws, is usually alert but not looking round or grooming Walk tail up Wu Cat walks with its tail held in a vertical position Walk tail down Wd Cat walks with its tail approximately horizontal Flick tail Ft Cat twitches its tail to the side, upwards or downwards Rubbing Rub Cat rubs any part of its body, usually head, flank or tail, on the experimenter, its feeder or an object Look at owner Lo Cat looks directly at its feeder Look at observer Lm Cat looks directly at the experimenter Look around La Cat looks behind or to the side by moving its head regularly Ž Sitrlie down SirLd Cat is sitting on hind paws and front paws the weight of the body . is supported by all four paws , or is lying on its side or back, the weight of the body supported by the body rather than the paws Sit up Su Cat is sitting upright, the weight of the body is supported by the hind paws Jump onroff furniture Jo Cat jumps on or off an item of furniture Sniff floor Sf Cat sniffs at the floor anywhere in the room Sniff object So Cat sniffs at any object except the floor, its feeding bowl and the food it received Lickrsniff feeding bowl Lb Cat licks its feeding bowl or sniffs at it Lickrsniff food Lf Cat licks the food it received or sniffs at it Eat Ea Cat eats from the food it received Lick nose Ln Cat flicks its tongue quickly over its nose Lick lip Ll Cat flicks its tongue around the outside of its mouth Lick paw Lp Cat licks one of its front paws without passing it over its head Groom facial area Gf Cat grooms its facial area by passing a previously licked paw over its head Groom chestrgroom body GcrGb Cat grooms its chest or any part of its body except front paws and face Other behaviour Ot Everything the cat does that is not described in this table presentation. It was denoted whether the cat had eaten all, a part or nothing of the food items. 2.5. Data analysis Ž For every cat, 5 min of post-meal behaviour after each of the four food items two . times experimental food, once MFF and once LFF, and two times normal food were Ž . analysed. Data of different cats were combined in Excel 97 Microsoft and were tested Ž . for significant differences p F 0.05; two-tailed unless otherwise stated in SPSS Ž . standard version, release 7.5 using Mann–Whitney U-test for independent samples Ž . data of different cats and Wilcoxon matched-pairs signed-ranks test in case of Ž . Ž . correlated samples data derived of one and the same cat Ferguson, 1981 . Nonpara- Ž . metric statistics over parametric statistics were chosen since 1 group A contained only Ž . Ž . a small number of animals n s 5 favouring nonparametric statistics and 2 at least some data were not expected to be normally distributed, e.g. durations of behaviour Ž . patterns are limited to 300 s 5 min , favouring nonparametric statistics. Therefore, all data are presented as medians instead of means. Total duration and frequencies of post-meal behaviour were tested for outliers using the ‘Explore: Statistics’ function of SPSS. Sequences of events of 5 min post-meal behaviour of MFF or LFF were obtained by generating transition matrices for each group of cats. Matrices were edited and tested for Ž significant transitions in Matman version 1.0 Noldus Information Technology, Wa- . Ž . geningen, the Netherlands . Significant transitions p F 0.05–0.001 were detected Ž . using the chi-square x 2 test followed by the adjusted residuals procedure in Matman. Ž . Sign tests Ferguson, 1981, pp. 401–402 were used to test for differences between Ž . the amounts of MFF and LFF, and normal food before or after MFF NF-MFF and Ž . normal food before or after LFF NF-LFF eaten per cat within the three groups. Per group, it was scored in how many cases a cat ate more of MFF than of LFF and more of NF-MFF than of NF-LFF. Between groups x 2 tests were used.

3. Results