Materials and Methods

Cats and setting—For participation in the study, free-roaming domestic cat groups were required to have regular and fixed feeding hours and a habitat representative of a typical feeding group site in the urban environment in Israel. All locations at each study site had to be accessible to investigators to facilitate observations.

Four defined cat feeding groups were selected (A, B, C, and D) and monitored from October 1999 to October 2000. Details regarding the cats and their environment are provided elsewhere.²² Cats in groups A and B were trapped, taken to a veterinary clinic, neutered, vaccinated against rabies, and returned to their site of capture (after 24 hours of recovery in the clinic). Cats in groups C and D were neither captured nor treated in any manner.

Each group had a particular feeder, and 3 of the 4 feeders had their preferred cats, which they handled routinely. All other cats were familiar with the feeder's habits, but no physical contact was made with the feeder. Throughout the study period, feeders in all 4 groups provided food only, without any medical treatment to the cats.

All cats were characterized as belonging to 1 of 2 age categories: kittens (up to the age of 6 months) and adults (older than 6 months). Cats older than 6 months are referred to as adults rather than juveniles^21,b in the present report because females can go through puberty and experience their first estrus as early as 5 months of age, and in free-roaming cats, this can occur even earlier.²³

Data collection—During a preliminary period, cats were individually identified in each feeding group by direct observation or by use of binoculars when needed. Details regarding the means of identification are provided elsewhere.²² Sex and neuter status were defined by the presence of testes in unneutered males, the scrotum in neutered males, enlarged mammary glands of lactating queens, enlarged cheek glands in males, and the overall body size (males are usually larger than females).^24,25

On the basis of results of preliminary observations of feeding duration in group A, a duration of 160 minutes was initially chosen as the observation period for all groups. However, after the study began, it became apparent that 3 of the 4 study groups did not require such a long observation period. Accordingly, the duration of subsequent observations was group specific, ranging usually between 30 and 100 minutes but no longer than 160 minutes. The study started with 5 to 7 weeks of observations, performed 2 to 3 times/wk (total of 15 initial observations), followed by 8 months of weekly observations, and at the end of the latter period, another 5 to 7 weeks of observations, performed 2 to 3 times/wk.

Each observation began when the cats started to gather at the feeding site, before the feeder initiated food delivery, and ended when most of the cats had left and only a few were still eating. A cat was counted as present if it arrived at any time during the observation period.

Data recorded for each cat consisted of the duration of appearance at the feeding site, time of feeding, and observed social interactions (Appendix).^24,26,27 Each display of social interaction between 2 or more cats, including agonistic, amicable, or male ritualized vocalization interactions, was recorded following the all-occurrences recording method.²⁸ For convenience of recording, the duration of appearance and feeding times during each observation period were recorded in continuous but discrete observation intervals (eg, identities of cats present 5 to 10 minutes from the beginning of the observation and identities of those that were feeding).

Statistical analysis—All analyses were performed with statistical software.^c To determine whether neutered cats (groups A and B) differed from unneutered cats (groups C and D) in their duration of appearance at various intervals throughout each observation period and in their feeding behaviors, a ratio was calculated as the number of times each cat appeared and fed during a specific observation interval throughout the study period divided by the total number of observations made during that specific interval throughout the same year. For example, throughout the year, observations were recorded on 56 occasions during the first 5 through 10 minutes of observation, yielding a yearly total of 56 observations in group A for that specific interval. One cat in that group was observed to feed 20 times during that interval throughout the year; therefore, the feeding ratio for that cat in that interval was 20/56 (0.4). Appearance was recorded every 10 minutes in group A and feeding every 5 minutes. Consequently, there were 15 observation intervals for appearance during a 160-minute observation period and 25 for feeding (the 16th and 17th observation intervals for appearance were omitted from the statistical analysis because there were only 2 observation periods in which cats ate during this interval). Similarly, appearance was recorded every 5 minutes in group B and feeding every 3 minutes. This way, there were 15 observation intervals for appearance and 12 for feeding during a 70-minute observation period (the 13th interval was omitted from the statistical analysis because there was only 1 observation in which a cat ate during this interval). For the statistical analysis, these 15 and 25 intervals were decreased to 5 and 5 averaged observation intervals for appearance and feeding in group A, and the 15 and 12 intervals were decreased to 3 and 3 averaged observation intervals for appearance and feeding in group B.

After square-root and arcsine transformation of values, a 2-way ANOVA with repeated measures (sex [male vs female] and neuter status [yes vs no] as between-subject effects) was performed for each neutered group to compare duration of appearance with feeding frequency at specific observation intervals (A, 5 consecutive intervals/variable; B, 3 consecutive intervals/variable).

To compare the rates of aggressive and amicable interactions among all 4 feeding groups, the number of agonistic and amicable interaction events (excluding male ritualized encounters, which were analyzed separately) recorded for each observation point in each group was divided by the number of cats present at that observation point. Values were transformed via arcsine transformation, and 1-way ANOVA was used to test the differences among all study groups.

The recording of the cats' behavior was necessarily dependent on their appearance at the study site. Consequently, we incorporated the annual appearance of each cat into calculations of the frequencies of agonistic and amicable behavioral events. After square-root transformation of the behavior frequencies, a curve of the observed frequencies of social behaviors (y-axis) was graphically plotted as a function of the annual appearance (x-axis). Linear regression was then used to obtain formulae for the expected frequencies of social behaviors. For the agonistic behavior analysis, the formula for group A was y = 0.07 + 3.5x (R² = 0.3) and for group B was y = 0.7 + 1.8x (R² = 0.2), whereas for the amicable behavior analysis, the formula for group A was y = −0.3 + 2.3x (R² = 0.3) and for group B was y = −0.1 + 1.7x (R² = 0.3). In these formulae, y represents the square root of the expected frequency of the social behavior and x is the square root of the expected frequency of the annual appearance. The value of the expected frequency was subtracted from each observed value. Because the resulting values were normally distributed, a 3-way ANOVA (group × sex × neuter status) was performed to examine differences in social behavior between neutered and un-neutered male and female cats in groups A and B. Values of P < 0.05 were considered significant for all analyses.

Results

Cats—At the beginning of the study, after the TNR procedure had ended, 27 of 37 (73%) adult cats (13 males and 14 females) had been neutered in group A and 30 of 40 (75%) adult cats (15 males and 15 females) had been neutered in group B.

During the observation period, 89 cats were observed in group A (18 neutered females, 12 sexually intact females, 11 neutered males, 30 sexually intact males, and 18 kittens), 72 cats were observed in group B (21 neutered females, 7 sexually intact females, 16 neutered males, 16 sexually intact males, and 12 kittens), 61 cats were observed in group C (22 sexually intact females, 1 neutered male that had been neutered by the caretaker months before the study had begun, 14 sexually intact males, and 24 kittens), and 38 cats were observed in group D (7 sexually intact females, 1 neutered male that had immigrated into the group from the surroundings, 8 sexually intact males, and 22 kittens).

In general, cats in all feeding groups gathered and waited for the caretakers before feeding time. When the cats recognized the sound of the arriving caretaker, they started to show signs of restlessness and an increase in the frequency of agonistic behaviors was noted. This restless behavior continued for the first minutes of feeding. Cats that finished eating drew away from the food source and left the study site immediately or remained in the area for a while longer. The large food amounts that were provided in these groups enabled rotation of feeding cats present throughout the observations, and in most observations, food remained at the site after most cats had dispersed.

Durations of appearance and feeding behaviors— The durations of appearance and feeding patterns were similar among all 4 groups. The largest number of cats appeared and fed at the beginning of food delivery, gradually decreasing in number with time. Although the groups varied in cat numbers and feeding habits, these appearance and feeding patterns were typical in all the study groups, with only some of the cats that were present also observed feeding at each observation interval (Figure 1).

View Full Size

Mean ± SE number of cats that appeared (diamonds) and fed (squares) at specific intervals during a full observation period in free-roaming cats in feeding groups A (A) and B (B) that were neutered as part of a TNR procedure prior to the 1-year study period and those in groups C (C) and D (D), in which neutering was not performed. The arrow represents the time at which feed was provided to the cats. Groups differed in the duration of observation intervals. Total numbers of observations made throughout the year for each interval were as follows: group A, 57 for appearance and 56 for feeding; group B, 68 for appearance and 68 for feeding; group C, 54 for appearance and 54 for feeding; group D, 54 for appearance and 54 for feeding.

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

When examining the differences among appearance data obtained at the 5 observation intervals in group A by use of 2-way ANOVA with repeated measures, the Mauchly test revealed that the assumption of sphericity had been violated (P < 0.001). Therefore, degrees of freedom were corrected by use of Greenhouse-Geisser estimates of sphericity (ϵ = 0.4).²⁹ Results indicated a significant (P < 0.001) difference between observation intervals in the number of cats present (Greenhouse-Geisser estimate of sphericity, 2-way ANOVA with repeated measures; Figure 2). When examining differences between neutered and unneutered cats and between males and females, a 2-way ANOVA with repeated measures revealed that neutered cats were present significantly (P < 0.001) more often than were unneutered cats during each observation interval throughout the study, with the highest appearance frequencies during the second and third observation intervals, which represented the beginning of feeding time. No significant (P = 0.431) difference was detected between males and females in the frequency of appearance at a feeding site.

View Full Size

Mean ± SE appearance (A) and feeding (B) frequencies at specific intervals during an observation period for free-roaming cats in feeding group A that were neutered as part of a TNR procedure prior to the 1-year study period. Neutered males (n = 11) are represented by black triangles, neutered females (18) are represented by black circles, unneutered males (30) are represented by gray triangles, and unneutered females (12) are represented by gray circles. In panel A, the first interval (1) represents the time of food delivery. Each interval in panel A represents 32 minutes of observation and each in panel B represents 24 minutes. Mean appearance and feeding frequencies differed significantly between neutered and unneutered cats (P < 0.003 and P < 0.042, respectively; 2-way ANOVA with repeated measures).

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

When examining differences among the 5 observation intervals in group A for feeding cats, the Mauchly test again revealed that the assumption of sphericity had been violated (P < 0.001). Therefore, degrees of freedom were corrected by use of Greenhouse-Geisser estimates of sphericity (ϵ = 0.4). Findings indicated a significant (P < 0.003) difference between observation intervals in the number of cats feeding (Greenhouse-Geisser estimate of sphericity, 2-way ANOVA with repeated measures; Figure 2). Neutered cats in group A also spent significantly (P < 0.042) more time feeding during each observation interval than did unneutered cats (ANOVA with repeated measures). In general, neutered cats fed more during the first intervals and gradually ceased to feed toward the fifth interval. There was no difference in feeding time between males and females (P = 0.422; 2-way ANOVA with repeated measures). In group B, 2-way ANOVA with repeated measures revealed no differences between males and females and between neutered and unneutered cats in frequency of appearance (P = 0.929 and P = 0.348, respectively) and feeding (P = 0.401 and P = 0.637, respectively; Figure 3).

View Full Size

Mean ± SE appearance (A) and feeding (B) frequencies at specific intervals during an observation period for free-roaming cats in feeding group B that were neutered as part of a TNR procedure prior to the 1-year study period. Neutered males (n = 16) are represented by black triangles, neutered females (21) are represented by black circles, unneutered males (16) are represented by gray triangles, and unneutered females (7) are represented by gray circles. In panel A, the first interval (1) represents the beginning of food delivery. Each interval in panel A represents 23.33 minutes of observation and each in panel B represents 23.33 minutes.

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

Social behaviors—Two hundred sixty-six agonistic interactions were observed in group A, 166 in group B, 135 in group C, and 141 in group D. A significant (P < 0.001) difference in the rate of agonistic interactions between the study groups was evident (1-way ANOVA). A post hoc Tukey honestly significant difference test revealed that the rate of agonistic interactions in groups A and B was significantly lower than that in groups C and D (Figure 4).

View Full Size

Mean ± SE agonistic interaction rates (number of agonistic interactions observed per cat per observation) among the cats in Figure 1. *Values differ significantly (P < 0.001) between groups indicated by the brackets. †Values differ significantly (P < 0.05) between groups indicated by the brackets.

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

In groups A and B, no significant differences in agonistic behaviors were evident between neutered and unneutered cats, between males and females, or between the groups themselves (Figure 5). Males appeared to have a higher rate of agonistic behaviors (0.2 ± 0.2 agonistic interactions/cat/observation) than females (–0.4 ± 0.2 agonistic interactions/cat/observation), but the difference was not significant (P = 0.058). Unneutered males had a higher rate of agonistic behavior (0.5 ± 0.2 agonistic interactions/cat/observation) than neutered males (–0.1 ± 0.3 agonistic interactions/cat/observation), but again, the difference was not significant (P = 0.063).

View Full Size

Mean ± SE agonistic interaction rates (number of agonistic interactions observed per cat per observation) in groups A and B of free-roaming cats at different feeding sites, by sex and neuter status.

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

One hundred fourteen amicable interactions were observed in group A, 69 in group B, 52 in group C, and 90 in group D. The rate of amicable interactions differed significantly (P < 0.001) among the groups (1-way ANOVA). A post hoc Tukey honestly significant difference test revealed that the rate of amicable interactions was significantly higher in group D than in groups A, B, and C (Figure 6). In groups A and B, there were no differences in amicable behavior rate between neutered and unneutered cats, between males and females, or between the groups themselves.

View Full Size

Mean ± SE amicable interaction rates (number of amicable interactions observed per cat per observation) for the cats in Figure 1. *Values differ significantly (P < 0.01) between groups indicated by the brackets. †Values differ significantly (P < 0.001) between groups indicated by the brackets. ‡Values differ significantly (P < 0.05) between groups indicated by the brackets.

Citation: Journal of the American Veterinary Medical Association 238, 9; 10.2460/javma.238.9.1141

• Download Figure
• Download figure as PowerPoint slide

Of the 30 unneutered and 11 neutered males observed in group A throughout the study period, 12 unneutered and 3 neutered males engaged in a total of 165 ritualized agonistic male encounters. All encounters in this group involved pairs of unneutered-unneutered or unneutered-neutered cats; no encounters between pairs of neutered-neutered cats were observed. Of these 165 encounters, 101 (61.2%) involved the same 2 pairs of cats (an unneutered male vs another and an unneutered male vs a neutered male).

Of the 16 unneutered and 16 neutered males observed in group B throughout the study period, 10 unneutered and 4 neutered males engaged in a total of 34 ritualized agonistic male encounters. Only 1 encounter between a pair of 2 neutered cats was observed.

Of the 14 unneutered males and 1 neutered male observed in group C, 4 unneutered males and the 1 neutered male were involved in 6 agonistic encounters. Of the 8 unneutered males and 1 neutered male observed in group D, 5 unneutered males and the 1 neutered male were involved in 5 agonistic encounters.

Discussion

Feeding behavior in groups of free-roaming domestic cats has been mainly evaluated in the context of competition over access to food sources and appears to be affected by many factors, including cat sex, body size, age, and social status.^4,30,31 To add to this information, the study reported here showed that neutered cats can have different patterns of appearance at feeding sites and of feeding, compared with unneutered cats.

In general, the free-roaming cats from the 4 feeding groups (2 neutered [A and B] and 2 unneutered [C and D]) in the urban study setting had distinctive behaviors related to food delivery, which have been elucidated in house cats³² and in free-roaming domestic cats in rural feeding groups.¹⁶ The cats gathered around the feeding area as feeding time approached and dispersed gradually as time elapsed. Although in all groups not all cats were feeding during each observation interval, during the observation periods, all the cats were observed to eat. This gathering of the cats was accompanied by an increase in the frequency of agonistic behaviors, as has been reported elsewhere.¹⁶ In that other report,¹⁶ this increase was attributed to the occasional breakdown in the so-called personal space of the cats, which in turn increased tension between conspecifics and resulted in an increase in aggressive behaviors.

The research observers of all 4 groups in the present study also visited the feeding sites during the study period at hours other than feeding hours but were able to observe at those times only a few cats at most, all of which belonged to these feeding groups. This is in contrast to the more stable colonies in farms and rural areas with lower cat population turnover rates.^1,9,12,24 The observed pattern of behavior in urban free-roaming cats supports previous suggestions that these feeding groups are time and place dependent.¹⁶ Such multi-male and multifemale cat groups exist mainly because food is delivered at regular times.¹ Yet, such cats also engage in social interactions and individual familiarity displays among themselves, as was demonstrated in the amicable interactions and in the uneven distribution of ritualized agonistic male encounters between cat pairs in our study.

This pattern of appearance according to time of food delivery was further demonstrated in group A by the neutered cats, compared with the unneutered cats. We suggest that the neutered cats, being present more consistently at the study site²² and therefore more familiar with a feeder's routine, were thus better able to predict the actual timing of food delivery to their group. Cats would particularly benefit from such an ability when competing for access to the more appealing foods, as was the situation in group A. Cats in group A fed from 1 large pile consisting of a mix of leftovers such as fish and chicken, rice, pasta, and bread, and the cats consequently needed to arrive in advance of actual food delivery to approach it as soon as it arrived lest they miss their chance to obtain the most appealing components. Neutered cats may place a higher priority on feeding than do unneutered cats because their breeding and territorial behaviors are fewer, which might explain the higher feeding rates found for neutered versus unneutered cats in group A. It is also possible that neutered free-roaming cats simply eat more, as is true for house cats in which, after neutering, there is an increase in food consumption, decrease in basal metabolic rate, and increase in body weight.^33–35 In group B, arrival and feeding times were less important for the cats because the feeder supplied each cat she knew with its individual dish.

Although there is general agreement that a prominent behavioral effect of neutering is that of reduction of aggressive behavior in domestic male cats,^{20,21,32,36–40} the neutering effect on female aggressive behavior remains controversial. For example, whereas some investigators claim that sterilization does not alter female behavior,³² others have suggested decreased aggression in neutered versus unneutered females.^21,37,41 Moreover, only a few studies^20,21 were conducted to examine the behavioral effect of neutering on free-roaming cats rather than on household cats. Although these studies revealed a decrease in aggressive behavior, they were small in scale (5 to 10 cats), without comparative control cats.

The multimale and multifemale groups that underwent TNR in the present study had fewer displays of aggression than the groups that did not undergo TNR. This decreased aggression was concomitantly observed with increased immigration into the neutered groups.²² Such possible immigration in the presence of decreased aggression has been suggested by others,³² who posit that neutered cats that are not guarded by their caretakers while eating may be usurped by visiting sexually intact cats.

Given the findings in group A, we suggest that neutered cats in such groups may engage in a compensatory mechanism to minimize competition (eg, over food) from the immigrating sexually intact cats. In our study, they showed up earlier, remained longer at the feeding site, and thus gained access to the choicest food portions and larger quantities of food than the unneutered cats.

Although unneutered and neutered cats in the present study did not differ in the frequency of aggressive behaviors, a difference, albeit not significant, in aggressive male behavior between the 2 types of cats was noticed. This difference in male aggressive behavior, together with the finding that only 1 ritualized encounter was engaged in by 2 neutered males, might suggest a behavioral effect of neutering. It may be that a bigger sample size of cats or examination of individual cat behavioral patterns before and after neutering is needed to obtain a better understanding of the effect of TNR programs on aggressive behavior of free-roaming cats. Females were less aggressive than males in the present study, as is the situation with household cats.³⁷

Reports of changes in affiliative behavior in free-roaming cats after neutering are scarce, and to our knowledge, only 2 studies^20,21 have been conducted to evaluate this. One study²⁰ found that allogrooming, a common affiliative behavior observed between females in a breeding group, was absent from the behavioral repertoire of a neutered feeding group, whereas the other²¹ found an increase in social amicable interactions after neutering.

Data for the neutered cat groups used in the present study reflected the existence of a core of cats that were consistently present at the feeding site throughout the study period.²² Amicable interactions purportedly serve to strengthen the social bonds between core members.¹¹ If this supposition is correct, then one would expect the neutered groups to have a higher rate of amicable interactions. However, unneutered group D had a significantly higher rate of amicable interactions per cat than the neutered groups. This might have been attributable to several of the females belonging to 1 female lineage in group D. Female lineages in farm groups and other rural settings have strong social bonds with extensive affiliative interactions, communal kitten care, and nursing.^11,32,42 In unneutered group D, we recorded 1 instance of a female licking the kittens of another female. These female collaborations necessarily strengthen the social bonds between the cats involved, particularly considering the benefits emerging from this behavior, such as success in rearing kittens, which reportedly was higher for females rearing their kittens communally and close to a food source than for females that raised their kittens far from the group in another study.¹⁰ However, no additional communal kitten care events were noted in group D in the present study, suggesting that although rearing kittens communally is beneficial to female reproductive success, females in this urban group might have had weaker social bonds than females in the rural colonies or that signs of communal kitten care appear less frequently during feeding times.

Agonistic male-male encounters generally consist of ritualized threat postures and vocalizations, usually ending in affected cats reducing the degree of threat by 1 cat, not necessarily the defeated one, moving very slowly away²⁴; more rarely, they can end in a fight.²⁷ In an attempt to determine the form of social hierarchy existing in a group of free-roaming domestic cats, investigators in another study² recorded the number of victories and defeats of each cat in such ritualized encounters. However, some claim it is not possible to establish the social hierarchy of males according to such encounters,²⁴ suggesting that when one begins to observe a group of cats that are already socially organized, encounters will be observed between males that are of equal social status, whereas males that rank differently from each other will avoid each other or one will behave submissively. Our findings in group A support this claim, with 2 male-male pairs engaging in > 60% of the total ritualized agonistic encounters in that group. These cats may have thus had a similar dominance rank, which they constantly tested.

Because aggressive behaviors in household male cats decrease after neutering,^36,38 male agonistic encounters of those cats are also expected to decrease in frequency. To date, there have been no reports of such a decrease. In the present study, however, we found that of the 199 ritualized agonistic encounters observed in the neutered cat groups, only 1 was performed by 2 neutered males. It appears, therefore, that when such ritualized encounters occur, at least one of the participants is likely an unneutered cat. Taking into consideration the decrease in aggressive behavior among neutered male household cats, we suggest that in encounters between neutered and unneutered males, the neutered cats are merely responding to ritualized threats by the unneutered cats.

Indeed, some males in the present study that had been neutered after reaching sexual maturity still responded to agonistic displays by another unneutered cat. For example, a neutered cat in group B always responded to displays of ritualized agonistic threats directed at it by a particular unneutered cat. Another event observed was the single incident of an encounter between 2 neutered males, both of which had been neutered after reaching sexual maturity. Because the behavior and social rank of some neutered cats is hypothesized to become fixed at the time of neutering,³² it is possible that these males, neutered after sexual maturity, had retained their social status and continued to fight for it.

In group A, 2 pairs of males (unneutered vs unneutered and neutered vs unneutered) engaged in 10-fold as many encounters as other male pairs, and in each of these 2 pairs, there was an orange-furred male (1 unneutered and 1 neutered, respectively). One study⁴³ showed that males carrying the orange sex-linked genotype are more aggressive and therefore may initiate more encounters than cats without the gene. It is therefore possible that because castration took place after the neutered orange cat had already reached sexual maturity, the gene effect negated the effect of neutering on aggressive behavior.

As has been suggested,^6,19,44 neutering has the potential to improve cat welfare because it reduces the frequency of negative physical interactions between cats and thus also their chances of being injured or contracting contagious diseases. Male-male encounters are of specific importance because their aggressive intensity is higher than that of other agonistic encounters (eg, male-female) and thus can sometimes end in physical fighting^24,27 and potential injury or disease. In the present study, neutered males participated in male-male encounters less frequently than unneutered males. Natoli et al⁴⁵ previously showed that the more aggressive males were more likely to become infected with FIV, an infectious disease transmitted through bites. The neutered males in the present study, participating in fewer male-male encounters, may benefit from improved health as a consequence.

In the study reported here, we found that in situations of competition over food (as in group A), the neutered cats in the present study were better adapted to the feeding times and thus gained better access to the food than did the unneutered cats. Not only did we find lower rates of aggressive interactions among cats in the neutered groups, but we also found that almost no neutered-neutered agonistic male encounters took place, which constitutes the first report of such a finding. It is possible that the lower aggressive interactions in the neutered groups enhanced the ability of visiting cats to integrate into the feeding groups, which might explain the higher proportion of immigrant cats detected in a concurrent study²² involving the same cats. Although these multimale and multifemale feeding groups were time and place dependent, the affiliative interactions observed and the existence of certain fixed pairs of agonistic male encounters indicated that familiarity exists between the members of the cat groups. The decreased agonistic behavior of the neutered male cats relative to unneutered males may result in decreased cat fighting and vocalizations and thus may lead to fewer injuries and decreased disease transmission among cats and decreased noise disturbance for their human neighbors.