Soil in the Mt. Longevity, Taiwan, eaten by Formosan macaques, was analyzed to determine the possible stimulus or stimuli for geophagy. I attempted to test the mineral supplementation hypothesis and to document macaques using particle size. I studied macaques soil-eating behavior from July 2000 to April 2002 in the Mt. Longevity, Taiwan, and analyzed soil samples eaten and uneaten by Formosan macaque for physical-chemical properties and geochemistry in the study area. The results show that samples of soil eaten contained relatively high iron ( 3.6% ) and aluminum ( 16.0% ). However, the concentration of samarium was significantly lower in soil from samples eaten than in the random samples. No difference in concentration of the remaining fifteen elements ( magnesium¡Bcalcium¡B chromium¡Bmanganese¡Bcopper¡Barsenic¡Bstrontium¡Bbarium¡Blanthanum¡Bcerium¡Bneodymium¡Bytterbium¡Bphosphorous¡Bsulfate¡Bchlorides ) and Nitrate nitrogen was found between these two groups of soil samples. Geophagy occurred at a high rate of 0.31 monkey per hour with an elevated frequency in the reproduction season. The density of geophagy has been estimated as 1.6¡Ñ103 individuals per km2 . Frequency of male geophagy were recorded highest in September. Frequency of Females eating soil was recorded mainly between February and April. The ingested soils were significantly richer in clay than control soils. This study supports the hypothesis that mineral supplementation is a major factor for Formosan macaque engaging in geophagy.

Adult female Formosan macaques were observed to collect data on the acquisition and stabilization of ranking, troop fission and to compare the female ranking system and nepotistic hierarchy. Behavioral observations were recorded from 1 October 2006 to 30 April 2008. Behavioral sampling methods included scan sampling of macaque troop members, all occurrences sampling of adult females¡¦ affiliative behaviors, and focal animal sampling of aggressive behaviors. In addition, demographic and ranking records were collected from 1998 until 2008. When females attained four years of age, they gained adult female hierarchy, and 88 % (n = 43) of them were middle or low ranking. Individual traits, ageing and mother¡¦s relative rank have significant effects on the ranking of adult females who were 5 to 11 years of age (P < 0.05), and the ranking of 47 % (n = 58) mothers were close to their daughters. When females were 5 to 11 years of age, their relative ranks had positive linear relationships to their mother¡¦s relative rank (P < 0.05). Maternal hierarchy affected the ranking of reminders in troop C, and the dominance matrilineal females stabilized high ranking. But, most subordinate matrilineal females were middle or low ranking. The ranking of 86.4 % (n = 22) of females who immigrated to branch troops was middle or low. Matrilineal members and the ranking before troop fission had effects on the female ranking after troop fission. The ranking of 59.4 % (n = 32) of mothers was higher than that of their daughter. When females reached 9 years old or older while their mothers were alive, mean relative rank of other matrilineal female and mother¡¦s relative rank had significant effects on their ranking (P < 0.05). However, individual traits, mean relative rank of other matrilineal female, ageing, and the number of adult daughters have significant iv effects on female ranking (P < 0.05). When females were 9 to 15 years of age, the rank maintaining ratio of female with mature daughter was 0.82 (¡Ó 0.12), which is significant higher than the ratio of female without mature daughter (P < 0.05). About 50 % (n = 18) of younger sisters outranked their older sisters. The ranking relationships of sister dyads had positive linear relationships (P < 0.05). The proximity index of mother and younger sister was significantly higher than the proximity index of mother and older sister (P < 0.05), but that is independent of whether females outrank their sister or not. Only 4 % (n = 423) female aggression data were recorded that 13 supporters helped attackers to attack victims. The kin and non-kin supporter were 46.1 % and 53.9 % respectively (n = 13). Maternal hierarchy affected the adult female ranking and half of the females outranked their older sisters. However, daughters could also outrank mothers. Therefore, female ranking system of Formosan macaques follows a weakly nepotistic hierarchy. Sterck EHM, Watts DP, vanSchaik CP (1997) The evolution of female social relationships in nonhuman primates. Behav Ecol Sociobio 41:291-309 Su HH (2003) Acquirement of social ranks of females in one group of Taiwanese macaques (Macaca cyclopis) at Fushan Experimental Forest, Taiwan. Am J Phys Anthropol:203-203 Su HH, Birky WA (2007) Within-group female-female agonistic interactions in Taiwanese macaques (Macaca cyclopis). Am J Primatol 69:199-211 Su HH, Lee LL (2001) Food habits of Formosan rock macaques (Macaca cyclopis) in Jentse, northeastern Taiwan, assessed by fecal analysis and behavioral observation. Int J Primatol 22:359-377 Suzuki S, Hill DA, Sprague DS (1998) Intertroop transfer and dominance rank structure of nonnatal male Japanese macaques in Yakushima, Japan. Int J Primatol 19:703-722 Thierry B (1990) Feedback loop between kinship and dominance: the macaque model. J Theor Biol 145:511-521 Wu HY, Lin JF (1992) Life history variables of wild troop of Formosan macaques (Macaca cyclopis) in Kenting, Taiwan. Primates 33:85-97

I have investigated the social grooming in kinship, rank, age and seasonal change among adult female Formosan macaques (Macaca cyclopis) that inhabit Mt. Longevity, Kaohsiung. The major study groups were C and Cd groups. Field observations were conducted from August 2000 to February 2003 covering three mating seasons and two non-mating seasons. The observations covered a total of 188 work days including 1248.8 hours. I actually had recorded C group for 660.6 hours, and Cd group for 244.5 hours. During my study, C group consisted of 8-13 adult males and 14-15 adult females, while group Cd had 1-3 adult males and 2-4 adult females. In order to analyze grooming data, I divided 15 adult females into sub-groups such as dominant/ submissive groups, old (>13) /young age (5-12), relative higher/lower ranking and related/unrelated. I have also divided adult males into troop-males and periphery males. Adult female allo-grooming activities accounted for 37.62% ¡Ó 13.59 (n = 15) of the behaviors in the daytime. I also found that adult females grooming infants and juveniles were greater than received from them (p < 0.001). The social grooming among adult females occurred mainly during non-mating seasons and its frequency was 2.12 times of mating seasons. Regardless of mating or non-mating seasons, the frequency of grooming among related females was significantly higher than among unrelated females (both p < 0.001). In addition, kinship affected the grooming frequency among female macaques with relative lower or higher ranking females during mating seasons (p < 0.05), while the dominant rank did not have the effect. However, the highest grooming frequency occurred in the high-ranking females who groomed relative lower ranking females within relatives (0.38 ¡Ó 0.40 bouts / 100 scans, n = 7). During non-mating seasons, kinship affected the grooming frequency among female macaques with relative lower ranking females (p < 0.05); the highest grooming frequency occurred in the old females who groomed relative lower ranking females within relatives (1.57 ¡Ó 1.74 bouts / 100 scans, n = 8). Moreover, seasons (mating or non-mating) and kinship relationship had significant effects on grooming frequencies among female macaques (both giving and receiving p < 0.01), as well as on the grooming frequency of females groomed with relative lower ranking females (p < 0.01). The highest grooming frequency occurred on females groomed with relative lower ranking females within relatives during non-mating seasons (1.25 ¡Ó 1.48 bouts / 100 scans, n = 11). Without kinship relationship, low-ranking females groomed relative higher ranking females more frequent than high-ranking females did (p < 0.05). Within adult females, 65% of social grooming was among relatives. However, 40% of adult females groomed equally with related and unrelated females, while 20% disproportionately groomed more with unrelated females than with related females. The grooming was kin-biased for 40%. About 8.68% of social grooming among unrelated females was being reciprocated. On the other hand, the ratio of related grooming female partners to the total number of available related females was higher than that with unrelated females (p < 0.05). It also indicated that the ratio of each female received grooming from high-ranking grooming partners was higher than that from low-ranking females (p < 0.01).On the other hand, the ratio of the number of old or young grooming partners of adult females had similar values. The grooming frequency of adult females gave or received from adult males during mating seasons was higher than non-mating seasons (both p < 0.05) while high-rank females groomed adult males more than low-rank females did. The socionomic sex ratios of these two social groups were similar during mating seasons (AM:AF = 1:1.8). Both troop and periphery males had significantly higher frequency of social grooming with adult females in mating seasons than in non-mating seasons. The types of males and social groups had significant effects on the allogrooming frequency among adult males in mating seasons (p < 0.05) but not in non-mating seasons (p > 0.1). Moreover, troop males had higher grooming partners than periphery males. Major grooming partners of troop males were adult females regardless of the seasons. Subordinate males were mostly the receivers in the grooming dyads with dominant males in the mating seasons, but the relationships changed during non-mating seasons. Agonistic interactions occurred mainly during mating seasons and its frequency among periphery males was 1.8 times of troop males (p < 0.05). The preference grooming sites between allo-grooming and auto-grooming of C and Cd groups had varied significantly (p < 0.001). The back region was the preferred grooming site in allo-grooming of C and Cd group and the ano-genital region was the least groomed site. In auto-grooming, monkeys paid much attention to the legs ignoring the back and face. The result indicated that when the adult individuals groomed the head, back and face which showed significant difference in the frequency among AM and AF (p < 0.05). The related female adults groomed head more frequently than unrelated female adults (p < 0.05), but dominance rank and age-class did not apparently affect the corresponding values for the frequency of grooming sites among adult females (both p > 0.1). The results indicated that social grooming among adult females took place more often during non-mating seasons, and more often in kin-related females than unrelated females. Moreover, the dominant females were likely to groom related females. Therefore social grooming among kin-related females may reinforce relationships while reciprocal grooming of unrelated females may serve to form alliance or ranking promotion in the social group. On the other hand, social grooming between adult male and female macaques more frequent in mating seasons than in non-mating seasons. This showed that adult male Formosan macaques employed complex strategies to achieve reproductive success. Nonetheless, the periphery males had more male grooming partners than troop males did which seemingly to enhance male coalitions.

Formosan macaques were observed to collect data on intergroup encounters, the strategies of male and female and the factors which influence inter-group dominance relationships. Four hypotheses exist to explain the strategies of male and female in inter-group encounter, including female resource defense, male mate defense, male resource defense and male mate attraction via infanticide. Behavioral observations were recorded from March 2008 to March 2009. Data collections included scan sampling and focal sampling of troop members which participated in encounter, containing their sex, age, rank and behavior. Duration of inter-group interaction and distances between two troops were also recorded. There is a significant positive relationship between chance of winning and troop number (AM+AF+SAM), and this chance of winning is higher than 50% when troop size exceeded 30. There is also a significant linear regression between weight win¡]main group and branch group¡^and troop number , but chance of weight win in some troops is higher than 90% when troop size is only between 20 to 30. Approach, line-up, displace and be-displaced have significant correlation with troop size type. Large troops displayed higher displace (65%), line-up (6%) and approach (28%) than other type of troops, and small troop were easily displaced by other type of troops (60%). Aggressive behavior also has significant correlation with mating season significantly; troops of Formosan macaques display more aggressive behavior in mating season (70%) than in non-mating season (41%). Troops of Formosan macaques display higher approach¡]33%¡^ and displace (93%) when troops encounter with all male troop than with bisexual troop. Inter-group dominance relationships exist among main group and branch group which newly split of Formosan macaques. But there is uncertain inter-group dominance relationships between main group and branch group which split early for the difference of maternal dominance hierarchies, variation of troop numbers, and the character of central males. Adult male and female Formosan macaques adopt different strategy in inter-group encounter. Adult female Formosan macaques participated (8.62%, n=8) and displayed aggressive behavior (3.66%, n=8) in more food-related encounter than in non-food-related encounter, so female resource defense hypothesis is supported. Adult male Formosan macaques display higher aggressive behavior (7.92%, n=23) significantly than adult female and sub-adult male in inter-group encounter, and the target of aggression is higher for males than for females. Adult and sub-adult male Formosan macaques display higher aggressive behavior in mating season significantly than in non-mating season. So male mate defense hypothesis is also supported. This study provides support for the male resource defense hypothesis. Adult male Formosan macaques would defend food resource directly and indirectly. Besides, frequencies of participation and aggression in low ranking adult male Formosan macaques are higher than high ranking ones to exchange mating opportunity.

Thesis (Ph. D.)--University of Washington, 1994. / Vita. Includes bibliographical references (leaves [77]-83).

Thesis (M.S.)--University of Michigan, 1992.

Thesis (Ph. D.)--University of Wisconsin--Madison, 1958. / eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (p. 92-94).

Thesis (Ph. D.)--University of Wisconsin, 1958. / eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references.

Thesis (M.A.)--University of Wisconsin--Madison, 1969. / eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (p. 138-144).

Thesis (Ph. D.)--University of Wisconsin--Madison, 1963. / Typescript. eContent provider-neutral record in process. Description based on print version record. Includes bibliographical references (leaves 149-156).