<%BANNER%>

Saimiri sciureus and Cebus apella Mixed-Species Associations in Raleighvallen, Suriname

Permanent Link: http://ufdc.ufl.edu/UFE0024096/00001

Material Information

Title: Saimiri sciureus and Cebus apella Mixed-Species Associations in Raleighvallen, Suriname Ultimate Functions and Proximate Mechanisms
Physical Description: 1 online resource (72 p.)
Language: english
Creator: Phillips, Carson
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: association, capuchin, cebus, foraging, mixed, predator, saimiri, species, suriname
Anthropology -- Dissertations, Academic -- UF
Genre: Anthropology thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A mixed-species association (MSA) is two or more species traveling and foraging together as a cohesive group. Prior research suggests that MSAs are adaptive responses, which provide benefits to at least one of the taxa: reduce predation risk or foraging benefits. Ecological factors of predation and food yield are predictors for the function of MSAs. Ecological variables such as seasonality, type of habitat and food availability contribute assessment of MSA patterns among many species including New World primates. We looked at MSA patterns and processes between Saimiri sciureus and Cebus apella in Raleighvallen, Suriname. We tested three categories of Saimiri association against three MSA hypotheses: 1) Reduce predation risk only, 2) Foraging benefits only, 3) Reduce predation risk and foraging benefits. Our objective was to take our 5,211 15-minute scan sampling data that recorded ecological parameters (i.e., seasonality) and apply them to the mechanisms expected to be observed in each MSA function to help explain Saimiri-Cebus MSA function in Raleighvallen, Suriname. Our data showed that Saimiri associated with Cebus during seasons of low food availability to locate foods, in high-risk habitats (e.g., swamp), and to consume more foods in quantity and taxonomic diversity.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carson Phillips.
Thesis: Thesis (M.A.)--University of Florida, 2008.
Local: Adviser: Boinski, Sue.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0024096:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024096/00001

Material Information

Title: Saimiri sciureus and Cebus apella Mixed-Species Associations in Raleighvallen, Suriname Ultimate Functions and Proximate Mechanisms
Physical Description: 1 online resource (72 p.)
Language: english
Creator: Phillips, Carson
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: association, capuchin, cebus, foraging, mixed, predator, saimiri, species, suriname
Anthropology -- Dissertations, Academic -- UF
Genre: Anthropology thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: A mixed-species association (MSA) is two or more species traveling and foraging together as a cohesive group. Prior research suggests that MSAs are adaptive responses, which provide benefits to at least one of the taxa: reduce predation risk or foraging benefits. Ecological factors of predation and food yield are predictors for the function of MSAs. Ecological variables such as seasonality, type of habitat and food availability contribute assessment of MSA patterns among many species including New World primates. We looked at MSA patterns and processes between Saimiri sciureus and Cebus apella in Raleighvallen, Suriname. We tested three categories of Saimiri association against three MSA hypotheses: 1) Reduce predation risk only, 2) Foraging benefits only, 3) Reduce predation risk and foraging benefits. Our objective was to take our 5,211 15-minute scan sampling data that recorded ecological parameters (i.e., seasonality) and apply them to the mechanisms expected to be observed in each MSA function to help explain Saimiri-Cebus MSA function in Raleighvallen, Suriname. Our data showed that Saimiri associated with Cebus during seasons of low food availability to locate foods, in high-risk habitats (e.g., swamp), and to consume more foods in quantity and taxonomic diversity.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Carson Phillips.
Thesis: Thesis (M.A.)--University of Florida, 2008.
Local: Adviser: Boinski, Sue.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0024096:00001


This item has the following downloads:


Full Text

PAGE 1

1 Saimiri sciureus AND Cebus apella MIXED-SPECIES ASSOCIATIONS IN RALEIGHVALLEN, SURINAME: ULTIMA TE FUNCTIONS AND PROXIMATE MECHANISMS By CARSON RODHOLM PHILLIPS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Carson Rodholm Phillips

PAGE 3

3 To my Mom and Dad, family, Michael Dykes, Lee Ahsmann, Jackson Frechette and all of my friends.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Dr. Su e Boinski for making this thes is possible. Without her amazing data set, expertise and unwavering patience, this thesis would not be possible. I would also like to thank my supervisory committee: Dr. Sue Boinski and Dr. John Krigbaum. Finally, special thanks go to Jackson, and my family and friends, who have always supported me.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........7LIST OF FIGURES.........................................................................................................................8ABSTRACT.....................................................................................................................................9CHAPTER 1 INTRODUCTION..................................................................................................................10Ultimate Functions of Mixed-Species Associations............................................................... 11Reduce Predation Risk.......................................................................................................... ..11Safety in Numbers...........................................................................................................11Increased Predator Detection...........................................................................................12Confusion Effect..............................................................................................................14Reduce Vector-Borne Disease.........................................................................................14Group Retaliation on Predators.......................................................................................15Foraging Benefits....................................................................................................................15Increase Insect and Prey Capture....................................................................................16Locating Food Sites.........................................................................................................16Decrease Duplication Effort............................................................................................ 17Attain Unattainable Food................................................................................................18Proximate Mechanisms of Mixed-Species Associations........................................................18Hypotheses and Predictions.................................................................................................... 19General Hypothesis 1: Reduce Predation Risk............................................................... 21General Hypothesis 2: Foraging Benefits....................................................................... 21General Hypothesis 3: Reduce Predation Risk and Foraging Benefits.......................... 212 MATERIALS AND METHODS........................................................................................... 23Study Site and Subjects........................................................................................................ ..23Behavioral Sampling and Parameter Description................................................................... 24Data Analyses.........................................................................................................................243 RESULTS...............................................................................................................................28Frequency of Saimiri and Cebus Mixed-Species Associations.............................................. 28Habitat Use.................................................................................................................... .........28Seasonal Habitat Use.......................................................................................................29Mixed-Species Association Formation............................................................................ 30Mixed-Species Association Departure............................................................................30

PAGE 6

6 Season.....................................................................................................................................30Diet.........................................................................................................................................324 DISCUSSION.........................................................................................................................53Saimiri-Cebus Associations: Random or Intended?...............................................................53General Hypothesis 1: Reduce Predation Risk....................................................................... 54Mechanism 1: Saimiri Will Tolerate Riskier Habitats in an MSA and Will Be Intolerant of Risky Habitats Outside of an MSA......................................................... 54Mechanism 2: With Exception to Seasons with Low Foliage Cover, Saimiris Rate of Association Will Occur across All Seasons Equally............................................... 56Mechanism 3: When Cebus is Present, Foods Consumed Will Have Higher Feeding Rates than When Cebus is Not Present (MSA>No MSA + Post-MSA)..................... 57Mechanism 4: MSAs Will Form and Depart in a Safe Habitat....................................... 57Mechanism 5: Saimiri Will Tolerate Habitats with Microhabitats of Reduced Foliage Cover More Frequently during an MSA.........................................................58General Hypothesis 2: Foraging Benefits............................................................................... 59Mechanism 6: MSAs Will Occur During Periods of Low Food Availability to Locate Food and Reduce Duplication Efforts.............................................................. 59Mechanism 7: Foods Consumed During MSAs Will Have a Higher Feeding Rate than When Saimiri Are Not in an MSA (MSA > No MSA)........................................ 61Mechanism 8: Foods Consumed During Post-MSA Intervals Will Have a Higher Feeding Rate than When Not in an MSA (Post-MSA>No MSA)...............................61Mechanism 9: Foods Consumed in Post-MSA and MSA Intervals Combined Will Have a Higher Feeding Rate than When Not in an MSA (Post-MSA + MSA > No MSA)...................................................................................................................... 62Mechanism 10: Saimiri Will Have Greater Food Taxonom ic Diversity in an MSA...... 62General Hypothesis 3: Reduce Predation Risk and Foraging Benefits.................................. 64Mechanism 11: Mixed-species Groups Will Occur During Periods of Low Food Availability and in Riskier Habitats............................................................................. 64Mechanism 12: During Seasons with Low Food Availability, Saimiri Will Forage in Microhabitats with Low Foliage Cover in an MSA................................................ 65Mechanism 13: Feeding Rates Will be Highest with Cebus Present (MSA> No MSA and MSA> Post-MSA)....................................................................................... 65Mechanism 14: Saimiri Will Have Greater Good Ta xonomic Diversity in MSAs......... 66Conclusion..............................................................................................................................68LIST OF REFERENCES...............................................................................................................69BIOGRAPHICAL SKETCH.........................................................................................................72

PAGE 7

7 LIST OF TABLES Table page 2-1 Habitat description of Raleighvallen, Surinam e................................................................273-1 Proportional distribution (%) and raw count (n) of all 15-minute interval scan samples (N=5211) collected in 15 months acro ss four categories of habitats and three Saimiri associations...........................................................................................................333-2 Proportional distribution (%) and raw count (n) of Saimiri association categories in the bamboo habitat across five seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat across all seasons..................... 353-3 Proportional distribution (%) and raw count (n) of Saimiri association categories in the plateau forest habitat across five seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat across all seasons..................... 373-4 Proportional distributio n (%) and raw count (n) of Saimiri association categories in the liana habitat across five seasons of observation........................................................... 393-5 Proportional distribution (%) and raw count (n) of Saimiri association categories in the swamp habitat across fi ve seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat across all seasons..................... 413-6 Proportional distribution of 15minute scan samples of three Saimiri associations across five seasons (N=5211)............................................................................................433-7 Proportional distribution of Saimiris level of integration in each season in MSAs with Cebus .........................................................................................................................453-8 Genera feeding ra te observed in each Saimiri association category across all scan samples (N=5211). Genera ordered from most consumed to least consumed in all Saimiri associations...........................................................................................................473-9 Family feeding rate observed in three Saimiri associations throughout all scan sample intervals observed (N=5,211). Ranked from most consumed to least consumed...........................................................................................................................483-10 Genera foods consumed more than, less than or not at all in relation to each Saimiri association category........................................................................................................... 493-11 Family foods consumed more than, less than or not at all in relation to each Saimiri association category........................................................................................................... 503-12 Seasonal difference in quantity of food consumption in three Saimiri association categories and the overall distribution of foods consumed across seasons....................... 51

PAGE 8

8 LIST OF FIGURES Figure page 3-1 Proportional distribution of 15-minute inte rval scan sam ples (N=5211) collected in 15-months across four categories of hab itats and three Saimiri associations.................... 343-2 Bamboo habitat: seasona l distribution of three Saimiri associations................................. 363-3 Plateau forest habitat: seasonal distribution of three Saimiri associations........................383-4 Liana habitat: seasona l distribution of three Saimiri associations.................................... 403-5 Swamp habitat: season al distribution of three Saimiri associations................................. 423-6 Seasonal distribution of 15minute scan samples of three Saimiri associations................ 443-7 Proportional distribution of Saimiris level of integration in each season during MSAs with Cebus ..............................................................................................................463-8 Seasonal difference in quantity of foods consumed in three Saimiri association categories...........................................................................................................................52

PAGE 9

9 Abstract of Thesis Presente d to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts Saimiri sciureus AND Cebus apella MIXED-SPECIES ASSOCIATIONS IN RALEIGHVALLEN, SURINAME: ULTIMA TE FUNCTIONS AND PROXIMATE MECHANISMS By Carson Rodholm Phillips December 2008 Chair: Sue Boinski Major: Anthropology A mixed-species association (MSA) is tw o or more species traveling and foraging together as a cohesive group. Prior research suggests that MSAs are adaptive responses, which provide benefits to at least one of the taxa: reduce pred ation risk or foraging benefits. Ecological factors of predation and food yiel d are predictors for the functi on of MSAs. Ecological variables such as seasonality, type of habitat and food availability contribute a ssessment of MSA patterns among many species including New World primates We looked at MSA patterns and processes between Saimiri sciureus and Cebus apella in Raleighvallen, Suriname. We tested three categories of Saimiri association against three MSA hypotheses: 1) Reduce predation risk only, 2) Foraging benefits only, 3) Reduce predation risk and foraging benefits. Our objective was to take our 5,211 15-minute scan sampling data th at recorded ecological parameters (i.e., seasonality) and apply them to the mechanisms e xpected to be observed in each MSA function to help explain Saimiri-Cebus MSA function in Raleighvallen, Suri name. Our data showed that Saimiri associated with Cebus during seasons of low food availa bility to locate foods, in highrisk habitats (e.g., swamp), and to consume mo re foods in quantity and taxonomic diversity.

PAGE 10

10 CHAPTER 1 INTRODUCTION A mixed-species association (MSA) is tw o or more species traveling and foraging together as single cohesive group (Morse 1977; Terborgh 1983). Other terms defining the same behavioral pattern include: interspecific association (Mwango mo et al. 2008), polyspecific association (Gartlan and Struhsaker 1972) and heterospecific associations (Morse 1977). In an MSA, one or all species may benefit from the a ssociation. If one species benefits, the MSA will have a passive and active player (Terborgh 1983). The active player usually initiates, maintains and ends the MSA. The duration of an MSA is va riable; some lasting only an hour while others lasting a couple of days (Terborgh 1983). An MS A can include animal species within the same genus (Norconk 1990), or family (Terborgh 1983) or in completely different classes all together (e.g., zebras and ostriches) (Grzimek et al. 1972). These asso ciations have been observed across various species such as fish (Ehrlich and Ehrlich 1973), birds (Beauchamp 2004) and nonhuman primates (Terborgh 1983; Cords 1987; Norconk 1990). Benefits of an MSA do not usually include access to potential mates or increasing fitness through kinship since the two groups are not of the same species (T erborgh 1983), so why do different animal species form social groups? Mixed-specie s associations occur in many ecosystems across the world. Although widely di spersed, this behavior is only observed in a scant number of species. Species living in c onspecific groups, as not ed by Alexander (1974), have an increase in resource competition due to more individuals occupying an area and competing for the same food source. When re sources are limited and two separate species groups with similar diets form an MSA, compe tition increases even more. Is an MSA a last ditch effort to adapt to a diffi cult situation such as food scarci ty, or do they form MSAs when foods are of high quality?

PAGE 11

11 Researchers look for proximate mechanisms and ultimate functions driving MSAs to investigate why species form these potentially maladaptive associations with another species. Proximate mechanisms are processes species use to maintain low competition and/or low predation risk during an MSA. Examples of proximate mechanisms include: maintaining heavy leaf cover to hide from aeri al predators, use of microhabi tat (e.g., lower canopy or higher canopy) to minimize competition or reduce predation risk, and seasonal associ ation to locate ripe foods. These mechanistic processes set the stage for defining and distinguishing the ultimate function of an MSA. The ultimate function is the motivation be hind the MSA: reduce predation risk and/or foraging benefits. The following pa ragraphs give more de tail and examples of proximate mechanisms and ultimate functions. Ultimate Functions of Mi xed-Species Associations Two broad functional explanations have been explored to explain the MSA behavioral pattern. The two main theories of MSA function are reduce predation risk and foraging benefits (Morse 1977). Reduce Predation Risk Some species exploit MSAs as a tactic to mi nimize predation risk. As group size swells during an MSA, a variety of predator avoidan ce strategies materialize: safety in numbers (Terborgh 1986), high predator de tection, confusion effect (Mor se 1977), reduction in vectorborne disease (Freeland 1977), and physical retaliation against a predator (Richardson and Bolen 1999). Safety in Numbers The safety in numbers strategy relies on th e large group size crea ting a dilution effect on predation risk (Terborgh 1990). As the number of individuals in a group increase, each individuals chance of becoming the target of a predator becomes diluted (Terborgh 1990;

PAGE 12

12 Terborgh 1986); however the effect of this strategy is purely statisti cal (Terborgh 1986). For instance, a solitary animal has almost no chance of surviving a predati on event. Meanwhile, each individual has a 25% chance of being killed in a group of four Terborgh (1986) argues the probability of getting killed gets smaller as the group size increases In a study of flocking birds found on islands with differing pred ation risk, the frequency of MSAs decreased with reduced predation pressure (Beauchamp 2004). During mock predator e xperiments the birds increased MSAs in reaction to a mock avian predator, sugg esting that birds join an MSA to dilute their chance of becoming prey (Beauchamp 2004). Red colobus ( Prolobus badius ) form MSAs with diana monkeys ( Cercopithecus diana ) to reduce predation risk; whereas diana monkeys were not observed to gain any benefits (Noe and Bshary 1997). Chimpanzees ( Pan troglodytes) hunt colobus monkeys exclusively (Chapman and Chapman 1999), thus colobus monkeys increa se frequency of associations with diana monkeys during the chimpanzee hunting seasons (Noe and Bshary 1997). During playback experiments of chimpanzee vocalization, colobus monkeys sought and maintained associations with diana monkeys (Noe and Bshary 1997). If colobus monkeys were already in an MSA, the duration of the association lasted longer comp ared to the circumstance when colobus monkeys were in an MSA with no playback experiments (Noe and Bshary 1997). Increased Predator Detection Mixed-species associations can increase th e detection of predators by increasing overall vigilance (Morse 1977). Increased predator detec tion can occur in a number of ways: increase in number of eyes looking out for predators (Terborgh 1983), one speci es may have better predator detection abilities due to microhabitat use, or useful anatomical attri butes (e.g., acute hearing) (Grzimek 1972). Some species seek MSAs to exploit these predator detection abilities (Stensland et al. 2003).

PAGE 13

13 Spinner dolphins (Stenella longirostris ), for instance, aggregat e with spotted dolphins ( Stenella attenuata ) to exploit the spotted do lphins high vigilance so th ey can rest in the murky ocean where predator detection is difficult (Norris and Dohl 1980; Stensland et al. 2003). Two species of gazelle come together because one spec ies is better at detecting predators than the other (Fitzgibbon 1990). Cowtail stingrays ( Pastinachus sephen ) associate with whiprays ( Himantura uarnak ) in water with little vi sibility (Semeniuk and Dill 2006). The cowtails exploit whiprays ability to detect the electric si gnals of predators, thus the cowtails can reduce vigilance while resting in association with whiprays (Semeniuk and Dill 2006). Associated species may be able to detect different predator types (e.g., raptor vs. felid) due to different microhabitats used while in an MSA (Gautier-Hion et al. 1983). Cercopithecus species receive predator avoida nce benefits by not only responding to congeners alarm calls, but associating species alarm calls as we ll (Gautier-Hion et al 1983). Both C. nictitans and C. pogaonias use the upper canopy and are ad ept at detecting aerial pr edators (Gautier-Hion et al 1983). Thus, C. nictitans and C. pogaonias associate with C. cephus, a species that uses lower canopy and detects terrestrial pred ators (Gautier-Hion et al. 1983) These species predator detection abilities complement each other while in association because of their differential microhabitat use. Increased predator detection while in an MSA can be the result of one species anatomical attributes (e.g., acute hearing). Ze bras and ostriches form MSAs to exploit the others anatomical qualities to their advantage (Grzimek et al. 1972). Ostriches have better vision than zebras, and zebras ha ve better hearing than ostriches (Grzimek et al. 1972). If a predator is detected, the detector emits an al arm call and both species are able to maximize predator detection by reacting e qually to each others alarm calls (Grzimek et al. 1972).

PAGE 14

14 Confusion Effect The confusion effect is when every indivi dual in an MSA responds to a predator by vocalizing or running in different directions, which causes th e predator to be confused and disoriented (Morse 1977). All indi viduals are in safe areas by the time the predator makes sense of the situation (Morse 1977). In many MSAS, the species will respond to one anothers alarm calls as seen between ostriches and zebras (Grzimek 1972), and guenons (Gautier-Hion et al. 1983). When a predator alarm is vocalized, many individuals will call in response to the original call or escape in various di rections (Morse 1977; Charnov 1976) This rapid movement and harmonious vocalizations is suggested to confuse the predator (Morse 1977). To test the confusion effect, Terborgh a nd Landeau (1990) performed an experiment on how successful the largemouth bass ( Micropterus salmoides ) was in capturing silvery minnows ( Hybognathus nuchalis ). When a bass attacked the mi nnows, the minnows would scatter in different directions making it difficult for the ba ss to single out specific prey (Terborgh and Landeau 1990). The bass has lower catch succe ss as the silvery minnow school size increases (Terborgh and Landeau 1990). To te st the success capture rate on a school of different looking fish, the authors added blue-dye to some of the minnows (T erborgh and Landeau 1990). The results show that if the varied colored school si ze is eight individuals the bass are successful at capturing both types of prey, as school size incr eases to 15 individuals, th e bass success rate goes down (Terborgh and Landeau 1990). This illustrate s how a predator may react to an MSA of different sized or colored species. Reduce Vector-Borne Disease Decreasing parasitic load per individual is another theory to why species from mixedspecies groups (Freeland 1977). Parasites and diseases are often overl ooked as playing the predator role similar to raptors and felids; how ever, parasites and diseas es reduce host mortality,

PAGE 15

15 fitness and reproductive success (Nunn and Altizer 2006). Mixed-species groups of Cercocebus albigena, Colobus badius and Cercopithecus ascanius were found to have lower vector-borne bites then conspecific groups (Freeland 1977). On the other hand, malaria research on New World primates show an increase in the rate of mosquito attacks as gr oup size increases (Davies et al. 1991). Reducing vectorborne disease by forming an MSA needs to be researched more extensively to get a better id ea of how large groups and parasitic load are related. Group Retaliation on Predators Some species form MSAs to exploit another sp ecies aggressive retaliation tactics against a predator (Richardson and Bolen 1999). Two me thods to combat a predator are aggressive tactics or the ability to overpower the predat or based on sheer number. Bullocks orioles ( Icterus galbula bullockii ) actively associate with yellow-billed magpies ( Pica nuttalli ) in nesting areas (Richardson and Bolen 1999). Although the yello w-billed magpie is known to be aggressive against predators around their nest sites, they are not aggressive towards the orioles (Richardson and Bolen 1999). Orioles likely in itiate an association to attain protection against predators, resulting in a high density of orioles nesting around magpie nest sites (Richardson and Bolen 1999). Foraging Benefits Sympatric species may have ecological parallel s such as diet and use of a microhabitat, which may increase competition during an MSA. If ecological parallels ar e present, then why do species form MSAs since competition could increase? As resources become limited, segregation should be selected for since competition is elevat ed (May 1974). But, species that form MSAs find ecological benefits to form MSAs (Terborgh 1983) such as: increased insect and other prey capture (Morse 1977; Peres 1992), increased abil ity to locate food sites (Morse 1977; Cords

PAGE 16

16 1986; Chapman and Chapman 2000), decreased dup lication effort (Morse 1977), and access to foods otherwise unattainable as a solitary individual (Morse 1977). Increase Insect and Prey Capture In an MSA, species flush out insects and other prey (Morse 1977; Peres 1992; Hankerson et al. 2006). This flushing effect is created by species moving through the forest and disturbing the leaf litter (Peres 1992). The disturbance forces insects and prey to retreat. As insects retreat, they are quickly eaten by an individual in the MSA. Moustached tamarins ( Saguinus mystax pileatus) have been observed flushing insects to the benefit of saddle-back tamarins ( Saguinus fuscicollis avilapiresi ) (Peres 1992). As moustached tamarins forage in the upper canopy they flush insects down to the saddle-back tamarins foraging in lower levels of the canopy (Peres 1992). Birds associate with primates for insect fora ging benefits as well (Hankerson et al. 2006). White-fronted nunbirds (Monasa morphoes ) and woodcreepers (Dendrocolaptidae) follow golden-headed lion tamarin ( Leontopithecus chrysomelas ) troops to capture flushed insects (Hankerson et al. 2006). Tawny-winged woodcreepers (Dendrocincla anabatina), doubletoothed kites (Harpagus bidentatus) and gray-headed tanagers ( Eucometis pencillata ) all associate with Saimiri oerstedii during seasons of low arthropod availability (Boinski and Scott 1988). A similar seasonal association is f ound between kites and buffy-headed marmosets (Ferrari 1990). Locating Food Sites Sympatric species can share similar ecologi cal food sources but have difficulty locating specific food sites, especially during periods of low food availability (Gartlan and Struhsaker 1972). The diet of Cercopithecus cephus, C. pogonias and C. nictitans is more diverse in fruit species in an MSA than when with their re spective conspecific gr oups (Cords 1983). Cords

PAGE 17

17 (1983) suggest that the elevated consumption of different fru it species is a product of the increased number of individuals with knowledge of fruiting sites. Similarly, Callimico goeldii has a higher quality diet, includi ng sugars (ripe fruit) and protei n (arthropods) when associating with Saguinus fuscicollis than in solitaire (Porter and Garber 2007). California sea lions ( Zalophus californianus ) associate with three species of dolphins: bottlenose dolphins ( Tursiops truncates ), long-beaked common dolphin ( Delphinus capensis ) and short-beaked common dolphin ( Delphinus delphis). Sea lions associate with dolphins to exploit the dolphins ability to locate prey using a specific vo calization (Bearzi 2006). Sea li ons aggregate with bottlenose dolphins to benefit from their sk illed detecting ability and, as a result, increase their dietary breadth (Bearzi 2006). Often, species with large home ranges seek a ssociations to exploit the knowledge of food sites from a species with a smaller home range (Gartlan and Struhsaker 1972). Terborgh (1983) proposes that Gartlan and Struhsakers (1972) theory may explain associations between Saimiri and Cebus Saimiri has an average home range of >250 ha and Cebus has a home range of ca. 80 ha (Terborgh 1983). Ter borgh (1983) argues that Saimiri associate with Cebus to access fruit during periods of low food availa bility. However, Terborgh (1983) notes that food location may not be the sole explanation since Saimiri continue associating with Cebus even after the food source is located. Decrease Duplication Effort During periods of low food availability, sp ecies may benefit from MSAs by avoiding areas where food sites are alrea dy occupied by another species or where food has been depleted (Morse 1977; Terborgh 1983). Th is strategy decreases the amount of energy lost in finding foods or competing with another species. Since MSAs have high toleration for their associated species, this will allow low energy costs during low food availability (Morse 1977).

PAGE 18

18 Attain Unattainable Food Animals may form MSAs to catch prey they would otherwise be unable to capture alone or in a smaller group (Morse 1977; Struhsaker 1981). Chimpanzees, for instance, form cooperative hunting groups to seize colobus monkey prey (Boesch 1994). This behavior evolved, in part, because solitary hunting was less effective. Proximate Mechanisms of Mixed-Species Associations Proximate mechanisms are processes species use to attain maximum MSA benefits while maintaining low competition and low predation risk. Two theories for how MSAs form are: association by chance (Waser 1982), and actively seeking and persisting associations (Morse 1977). Mixed-species associations, as defined by Waser (1982), occur when more than one species occupies a mutual area due to a shared food source or habitat. The other theory states that one or more species actively seeks and ma intains association as a predator avoidance strategy (Morse 1977) or to exploit another specie s food knowledge (e.g., location of ripe fruit) (Gartlan and Struhsaker 1972). To understand the proximate mechanisms of MSAs it is important to fi nd the initiator of the MSA by identifying the species that benefits the most from the association (Terborgh 1983). At night, many MSA species will sleep separately (Cords 2000). As a result, many species will coordinate MSA reformation with vocalizatio ns during the morning hours (Cords 2000). An MSA of two species of tamarins, Saguinus mystax and S. fuscicollis is initiated by S. mystax emitting calls in the morning (Norconk 1990). Among some guenons (Cercopithecus), males emit loud calls to reform the MSA (Cords 2000). Saimiri vocalize before leaving an MSA with Cebus (Terborgh 1983). Looking at th e behavior of both MSA partic ipants in and out of an association can pinpoint which species seeks and benefits from the association (Terborgh 1983). For example, Cebus is a passive participant because their be havior never changes in or out of an

PAGE 19

19 association (Terborgh 1983). If Cebus is stationary, Saimiri leaves the MSA in search of another Cebus troop (Terborgh 1983). If the Saimiri troop is unsuccessful, they tend to return to their default Cebus troop (Terborgh 1983). The changes in one species behavior in and out of an MSA can be a good indicator for which partic ipant is the MSA initiator (Terborgh 1983). Mechanisms of maintaining an MSA are important because species must optimize benefits and minimize competition. Participan ts in an MSA have increased competition (Alexander 1974). As a result, species must find methods to decrease competition by either foraging on different foods or using different microhabitats. For example, in MSAs of Saguinus mystax and Saguinus fuscicollis the two species travel at differe nt vertical heights and have > 1m of interindividual spacing to maintain a stable group (Norconk 1990). Cohesion of the MSA is maintained by continual vocalizations th roughout their daily ac tivities (Norconk 1990). Hypotheses and Predictions Our study examined the ultimate functions and proximate mechanisms driving MSAs between Saimiri sciureus and Cebus apella in Raleighvallen, Suriname. Saimiri and Cebus are known to form MSAs in Manu, Peru where they travel and forage together from brief to extended periods of time (Terborgh 1983). Saimiri associated with Cebus over 90% of the data collected (Terborgh 1983). In Corcovado, Costa Rica, Saimiri oerstedii avoid associating with Cebus even when Cebus groups are available (Boinski 1989). Instead of Saimiri seeking Cebus as seen in Manu, Cebus seeks and forms MSAs with Saimiri (Boinski 1989). Only 6.6% of the scan samples had Saimiri associated with Cebus (Boinski 1989). In Raleighvallen, Saimiri were observed to actively associate with Cebus during periods of low food av ailability, risky habitats and to attain specific foods. Before analyzing the data we created asso ciation categories to separate any influence Cebus had on predator avoidance or enhanced fo raging efficiency: MSA, No MSA and Post-

PAGE 20

20 MSA. The first category, MSA, in cluded all observed intervals that Saimiri were in association with Cebus ( 50m). The MSA category has the benefits of predator avoidance and enhanced foraging efficiency for Saimiri The second category, No MSA, is intervals Saimiri were observed without Cebus ( 50m). The No MSA category has no MSA benefits. The final association category, Post-MSA, is intervals Saimiri were observed not in an MSA but followed an MSA and/or between MSA intervals with Cebus Although Cebus is technically not near the Saimiri troop but they are within close proximity. No MSA is different because it has long periods of time between or until another MSA with Cebus (e.g. day), whereas Post-MSA had just associated with Cebus or an MSA is about to come together again. The Post-MSA category is important to distinguish any influence Cebus might have on Saimiris diet or habitat location. The benefit gained from a Post-MSA associ ation is enhanced fo raging efficiency. Cebus for instance, could lead Saimiri to a fruit tree. After Cebus leaves Saimiri, Saimiri remain foraging in the fruit tree. Although Saimiri is technically not in an MSA, they benefit by gaining food site information from Cebus Commonly, MSA studies assign one ultimate function. It is ve ry unlikely to definitively conclude what function explains MSA behavi oral patterns since ecosystems and their communities are ever changing. So why do MSA f unctions have to be static when ecological factors determining MSA functions are plastic? Food yield, temper ature and rainfall are only a few examples of ecological factors. These eco logical factors determine the MSA behavioral pattern, not necessarily the indivi duals involved in the MSA. In this study, our goal was to identify the Saimiri-Cebus MSA ultimate function by applying our data set of various ecological para meters (i.e. habitat, seasonality, diet) to the proximate mechanisms we expect to be observed in each function. If a mechanism is supported

PAGE 21

21 by our data set, the function became a str onger candidate for the possible function to SaimiriCebus associations in Raleighvallen, Suriname. The following are our general hypotheses and their mechanisms: General Hypothesis 1: Reduce Predation Risk Mechanism 1: Saimiri will to lerate riskier, more expos ed habitats in an MSA and be intolerant of risky habitats outside of MSAs. Mechanism 2: With exception to seasons with low foliage cover, Saimiris rate of association will occur across all seasons equally. Mechanism 3: When Cebus is present, foods consumed will have higher feeding rates than when Cebus is not present. (MSA>No MSA+Post-MSA) Mechanism 4: MSAs will form and depart in a safe habitat (e.g., liana forest). Mechanism 5: Saimiri will tolerate habitats with microhabitats of reduced foliage cover more frequently during an MSA. General Hypothesis 2: Foraging Benefits Mechanism 6: MSAs will o ccur during periods of low food availability to locate foods and reduce duplication efforts. Mechanism 7: Foods consumed during MSAs will ha ve a higher feeding rate than when Saimiri are not in an MSA. (MSA > No MSA) Mechanism 8: Foods consumed during Post-MSA intervals will have a higher feeding rate than when not in an MSA. (Post-MSA > No MSA) Mechanism 9: Foods consumed in Post-MSA and MSA intervals combined will have a higher feeding rate than when not in an MSA. (Post-MSA + MSA > No MSA) Mechanism 10: Saimiri will have greater food taxonomic diversity in an MSA. General Hypothesis 3: Reduce Predation Risk and Foraging Benefits Mechanism 11: MSAs will occu r during seasons of low food availability and in riskier habitats. Mechanism 12: During seasons with low food availability, Saimiri will forage in microhabitats with low foliage cover in an MSA. Mechanism 13: Feeding rates will be the highest with Cebus present. (MSA> No MSA and MSA> Post-MSA)

PAGE 22

22 Mechanism 14: Saimiri will have greater food taxonomic diversity in MSAs.

PAGE 23

23 CHAPTER 2 MATERIALS AND METHODS Study Site and Subjects Data were collected at a study site esta blished in January 1998 in Raleighvallen, Suriname (4 42 N, 56 12 W). Raleighvallen is located within the Central Suriname Nature Reserve (CSNR) and contains 1.6 million ha of pr imary tropical forest (Boi nski et al. 2005a). The forest structure is classified into four habitats: liana forest plateau forest, swamp forest and bamboo patches (Boinski et al. in prep.). Detailed descriptions of these habitats are provided in Table 2.1. In 2000-2005, Raleighvalle ns average annual rainfall is 1,967 mm with an average temperature of 23-28 Celsius (B oinski et al. in prep.). The annual year is separated into four seasons: Fruit (Feb ruary-April), Wet (May-July), Flower (August-October) and Trans ition (November-January). This study used four seasons of data with an additional fruit season. The second fruit season (F ebruary-April 2001) illustrated variability from the first fruit season and we fe lt it would be pertinent to investigate this ecological variability to come to a substantia l conclusion on MSA pattern according to seasons. Our study analyzed Saimiris part in MSAs. Data were collected on two Saimiri troops. The primate troops, both Saimiri and Cebus were fully habituated due to the ongoing research at the study site (Boinski et al. 2005a). Saimiri troop size varies betw een 25 and 35 individuals (Boinski et al. 2005a). Average home ranges are estimated to be 44-130ha and day ranges average 700-2300 meters (Robinson and Janson 1987). The Saimiri diet consists of fruit, leaves, flowers and arthropods (Boinski 2002). In Raleighvallen, abundant evid ence from 10 years of longitudinal data indicates an intact predator community (Boinski 2002). Examples of primate aerial and terrestrial predators in Raleighvallen include the harpy eagle ( Harpia harpyja ), crested eagle ( Morphnus guianensis )

PAGE 24

24 and the jaguar (Panthera onca). Although predation events are rare, it is a constant threat throughout the year. Behavioral Sampling and Parameter Description The Saimiri data consists of 5,211 15 -m inute interval group-scan samples from February 2000 through April 2001. This time frame was chosen because it marked the beginning of Cebus data collection, which will be the focus of a separate MSA study (i.e., the role of Cebus in MSAs with Saimiri ). We selected the 15-minute interval group-scan sampling method because it resembles data from other MSA research and, as a result, we are able to compare our study to other sites. The group-scans re corded: date, time, troop identit y, troop dispersal (troop length and width), troop activity, troop speed, season, habita t type, habitat cover (0-3 scale, 0 being the least covered and 3 being fully covered), mini mum and maximum height of troop, observability index (number of individual monkeys seen in a one-minute scan), presence of any animal species observed 50 meters of Saimiri troop and plant foods consumed at any time within the 15minute scan sample. When Cebus were observed 50 meters from a Saimiri troop, it was considered a mixedspecies association (MSA). Cebus troop identity and category of overlap with the Saimiri troop (overlap, partial overlap, no ove rlap) were recorded. If Cebus were 50 meters from the Saimiri troop, Cebus and Saimiri were considered not to be in an MSA. Data Analyses Data from both Saimiri troops were pooled for analysis and divided into three categories of association: No MSA, MSA and Post-MSA. No MSA association occurred when the Saimiri troop was not accompanied by Cebus and had no contact prior to th e scan sample interval. MSA association occurred when Cebus were 50 meters of a Saimiri troop. Post-MSA association

PAGE 25

25 occurred when Cebus were 50 meters of a Saimiri troop but, importantly, had been in an MSA prior to the noted scan sample. The Post-MSA ca tegory was formulated to eliminate any bias of Cebus influence from the No MSA category. Th e No MSA category is assumed to have no Cebus influence and a long period of time until the next MSA; whereas the Post-MSA category could have influenced the activity, type of food consumed or behavior of the Saimiri troop because Saimiri had just had or between an MSA with Cebus Three main parameters used for analysis we re habitat, season, troop foliage cover, and plant foods consumed. Habitat type was tested for evidence of associ ation against the three categories of association, MSA formation and de parture (includes interval before No MSA and interval in No MSA). Given the categorical natu re of the data, analysis was performed using a Chi-Squared test. Seasonal significance was tested against ca tegory of association, use of habitat in different association categories a nd type of overlap during MSA. The statistical method used for finding evidence of association was a Chi-Square test. Troo p foliage cover in different association categories was tested using ANOVA b ecause we were testing the difference between multiple means. Food consumption in different association categories was assigned to genus and family separately (e.g., In Ed was entered as Inga and Mimosaceae). Arthropods were not included in this study. The plant foods were tallied into the three categories of association: MSA, No MSA and Post-MSA. The proportion of food consumed was calculated by taking the raw count of the food genus or family, and dividing it from the to tal 15-minute interval scan samples (N=5,211). The genus and family foods were presented in a table indicating if a specific genus or family were consumed more, less than, equal to or not at all in each associati on category. This method

PAGE 26

26 will allow us to illustrate specific foods Saimiri were eating with Cebus or alone. In the discussion, these tables will support our conclusions.

PAGE 27

27 Table 2-1. Habitat description of Raleighvallen, Suriname Habitat TypeDescription Liana forestDense forest consisting of lianas and vines. Plateau forestPrimary forest' with tall trees, few lianas and location of most dietar y fruit. Swamp forestSeasonally flooded forest with structure similar to plateau forest but with palm species and large understory. Bamboo patchesDense homogenous patches of bamboo ( Guadua lati f olia )

PAGE 28

28 CHAPTER 3 RESULTS Our objective was to take our data set of 5,211 15-minute interval scan samples that recorded various ecological parameters (i.e., hab itat, seasonality and diet ) and apply it to the proximate mechanisms of Saimiri-Cebus associations in Raleighvallen, Suriname. After testing the mechanisms, we can identify the ultimate f unction of the MSA in this dynamic ecosystem. Frequency of Saimiri and Cebus Mixed-Species Associations Saimiri were recorded in association (50 m) with Cebus 55.6% of the 5211 15-minute scan samples. No association (50 m) with Cebus constituted 22% of all scans. Post-MSA (when Cebus were >50m from Saimiri but followed an MSA and/or between intervals of an MSA) was 22.3% of all scans. Saimiri were observed in contact with five different Cebus troops during the course of the observa tions but were pooled during da ta analysis. Anecdotal data showed Saimiri actively sought and joined Cebus troops. We conclude Saimiri was the active player and initiator of MSA formations with Cebus Habitat Use Chi-Square analysis rendered strong ev idence supporting the asso ciation between the three categories of Saimiri associations and habitat use (2= 247, df = 6, p0.001). In the bamboo and swamp habitats, Saimiri were more likely to be in MSA than expected (adjusted residual = 10.802, 8.305 respectively). In the liana forest, Saimiri were less likely to be in an MSA than expected (adjusted residual = -12. 7), but plateau forest yielded no significance. The liana forest was used proportionally higher in all Saimiri associations, as illustrated in Figure 3.1 and Table 3.1, whereas the swamp and bamboo had the smallest proportional distribution.

PAGE 29

29 Seasonal Habitat Use Analysis of seasonal habitat use in three categories of Saimiri associations, Chi-Square revealed strong evidence supporting associa tion between MSA and habitat use by season. During the first fruit season (February to April 2000) (2=26.5, df=6, p0.001), Saimiri were more likely to be in an MSA in the bamboo and sw amp habitats than expected (adjusted residual = 5.337, 5.86 respectively), and less likely than expected to be in the liana forest (adjusted residual = -6.14). In the wet season (May to July 2000) (2=94.2, df = 6, p0.001), Saimiri were more likely to be in an MSA in the plateau forest, bamboo and swamp forest th an expected (adjusted residuals = 2.47, 5.474, 4.06 respectively). In the liana forest, an MSA was less likely than would be expected (adjusted residual = -7.67). During the flower season (August to October 2000) (2= 81.1, df=6, p <0.001), Saimiri were more likely to be in an MSA in the ba mboo and swamp forest habitats than expected (adjusted residual = 5.337, 5.86 respectively). Plat eau forest and liana forest habitats were less likely than expected to be occupied duri ng an MSA (adjusted residual = -2.28, -6.14 respectively). The transition season (November to January 2001) (2= 42.6, df=6, p <0.001) yielded that the bamboo and plateau forest habitats more likely than exp ected to be used in an MSA (adjusted residual = 4.552, 2.47 respectively). The liana forest habitat was the least likely than expected for Saimiri to be in an MSA (adjusted residual = -4.96). Lastly, during the second fruit season (February to April 2001) (2= 41.5, df=6, p <0.001) Saimiri were more likely to be in an MSA in the bamboo and swamp forest habitats than

PAGE 30

30 expected (adjusted residual =3. 03, 2.4 respectively). The liana forest was less likely to be used during an MSA (adjusted residual = -3.45 respectively). Tables 3-2 through 3-5 and Figur es 3-2 through 3-5 illustrate si milar results. Across five seasons of data, the bamboo and swamp habitat ha d an increase of use during the seasons when Saimiri had low food availability and associated with Cebus proportionally more. Mixed-Species Association Formation Chi-Square analysis reveal ed no evidence supporting an a ssociation between habitat and Saimiri MSA formation (2= 2.75, df = 9, p = 0.97). The test showed the site of MSA formation was random 97% of the time. Mixed-Species Association Departure Chi-Square tested associati on between departing from an MSA and the habitat the event occurred. The test yielded a small significance (2= 15.3, df = 9, p = 0.08). The liana forest was the habitat Saimiri were more likely than expected to de part from an MSA (adjusted residual = 2.085). The swamp forest was less likely than expected for Saimiri to digress from Cebus (adjusted residual = -2.62). Season The seasonal proportions of Saimiri associations indicated th ey spent the most scans, overall, in an MSA in all five seasons (Figure 36). The MSA rate increased from 5.2% in the first fruit season (February to April 2000) to 15.5% in the flower season (August to October 2000). The season Saimiri were recorded alone the most wa s the second fruit season (February to April 2001) (Table 3-6). Post-MSA peaked at 7.6% during second fr uit season (February to April 2001). Chi-Square analysis rendered strong evid ence supporting associati on between season and category of Saimiri association (2= 305, df = 8, p0.001). During the first fruit season

PAGE 31

31 (February to April 2000), Saimiri were less likely than expected to be in an MSA than expected (adjusted residual = -3.952). In the flower season (August to October 2000), Saimiri were more likely to be in an MSA than exp ected (adjusted residual = 11.52). Saimiri were less likely than expected to be in an MSA during the transi tion season (November to January 2001) (adjusted residual = -7.64). During the second fruit season (February to April 2001), Saimiri were less likely to be in MSA than expected and more lik ely to be in a post-MSA (adjusted residual = 2.03, 3.661 respectively). The degree of integration was te sted to show if there was an association between type of integration (overlap, partial ove rlap and no overlap) and season. Chi-Square results revealed strong evidence supporting this association (2= 557, df = 8, p0.001). During the first fruit season (February to April 2000), Cebus and Saimiri were more likely to have no overlap than expected (adjusted residual = 2. 2791). In the wet season (May to July 2000), they were more likely to be partially overlappe d than expected and less likely to be overlapped or have no overlap (adjusted residual = 10.3, -6.11, -6.44 respectively). Saimiri and Cebus were more likely to have no overl ap than expected in the flower season (August to October 2000) (adjusted resi dual = 10.38). During the transition season (November to January 2001), MSAs were less likely to be completely overlapped than expected (adjusted residual = -4.78). The second fruit se ason (February to April 2001) had very strong evidence showing Saimiri and Cebus were more likely to be overlapped than expected (adjusted residual = 20.8). Similar results were shown in Table 3-7 and Figure 3-7. No overlap peaked during the flower season (August to October 2000) when foods were the scarcest, whereas overlap with Cebus was observed peaking during the second fruit season (February to April 2001) when foods were most abundant.

PAGE 32

32 Diet In all 5,211 15-minute scan sample data Saimiris diet consisted of sixty-one plant genera and thirty-nine plant families. In all food s can samples collected, 59.6% of the food scan samples were consumed during an MSA, 17.9% were eaten in Post-MSA and 22.4% were consumed when Saimiri were No MSA. Table 3-8 and Table 3-9 illustrate the relativ e feeding rates of each genus and family food in the three association categories. In both fami ly and genera, MSA has the highest total feeding rate, followed by No MSA. Further analysis wa s performed to find what plant food taxa were eaten more, less, or not at all (n=0) in each association category. This method allowed us to investigate the three hypotheses more thoroughly. Table 3-10 and Table 3-11 map the genus and family plant foods that were eaten in each associa tion category. Foods were listed from most to least consumed. Of the 61 total plant genera (Table 3-10), 43 were eaten more in MSAs than No MSAs. Of the 43 genera, 22 were not eaten (n=0) in No MSA. Of the 39 plant families (Table 3-11), 31 families were eaten more in MSAs than No MSAs. Of the 31 families, 16 were not eaten (n=0) in the absence of Cebus

PAGE 33

33 Table 3-1. Proportional distribution (%) and raw count (n) of all 15 -minute interval scan samples (N=5211) collected in 15 months across four categories of habitats and three Saimiri associations Bamboo Plateau forest Liana Swamp % n % n % n % n MSA 6.8 35513.1 68328.5 14856.9 359 No MSA 0.4 22 4.8 25214.9 7761.3 66 Post-MSA 1.1 59 4.2 21714.1 7331.1 55 Total 8.4 43622.1 115257.5 29949.2 480

PAGE 34

34 Figure 3-1. Proportional di stribution of 15-minute interval sc an samples (N=5211) collected in 15-months across four categories of hab itats and three Saimiri associations 0 5 10 15 20 25 30 BambooPlateau ForestLianaSwamp Habitat Type MSA No MSA Post-MSA

PAGE 35

35 Table 3-2. Proportional distribut ion (%) and raw count (n) of Saimiri association categories in the bamboo habitat across five seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat across all seasons. Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 % n % n % n % n % n MSA 0.3 16 1.1 573.4 1771.0 521.0 53 No MSA 0.03 20.1 70.2 90.07 4 Post-MSA0.1 6 0.07 40.4 220.2 110.3 16 Total 0.4 22 1.2 633.9 2061.4 721.4 73

PAGE 36

36 Figure 3-2. Bamboo habitat: se asonal distribution of three Saimiri associations 0 0.5 1 1.5 2 2.5 3 3.5 4 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 37

37 Table 3-3. Proportional distribut ion (%) and raw count (n) of Saimiri association categories in the plateau forest habitat across five seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat across all seasons. Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 % n % n % n% n % n MSA1.0 543.1 1612.9 151 2.1 108 4.0 209 No MSA0.5 290.9 50 0.8 41 0.8 43 1.7 89 Post-MSA0.9 210.7 40 0.8 42 0.9 4811.3 66 Total2.0 1044.8 2514.5 2343.8 199 7.0 364

PAGE 38

38 Figure 3-3. Plateau forest habita t: seasonal distribution of three Saimiri associations 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 39

39 Table 3-4. Proportional distribution (%) and raw count (n) of Saimiri association categories in the liana habitat across five seasons of observation Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 % n % n % n % n % n MSA 3.2 166 4.6 243 6.6 346 4.5 235 9.5 495 No MSA 2.5 132 4.3 224 1.9 100 2.4 124 3.7 196 Post-MSA1.6 82 1.1 58 2.1 111 4.3 222 5.0 260 Total 7.3 38010.1 52510.7 55711.1 58118.2 951

PAGE 40

40 Figure 3-4. Liana habitat: seasonal distribution of three Saimiri associations 0 1 2 3 4 5 6 7 8 9 10 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 41

41 Table 3-5. Proportional distributi on (%) and raw count (n) of Saim iri association categories in the swamp habitat across five seasons of observation. The total represents the proportion of scan samples Saimiri spent in each habitat ac ross all seasons. Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 % n% n % n % n % n MSA 0.6 341.8 982.5 1300.5 241.4 73 No MSA0.09 50.5 270.05 30.05 30.5 28 Post-MSA0.1 80.1 80.2 90.4 210.2 9 Total 0.9 472.5 1332.7 1420.9 482.1 110

PAGE 42

42 Figure 3-5. Swamp habitat: seasonal distri bution of three Saimiri associations 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 43

43 Table 3-6. Proportional distribution of 15-minute scan sample s of three Saimiri associations across five seasons (N=5211) Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 Total % n % n % n % n % n % n MSA 5.2 27110.8 56715.5 811 8.0 41916.0 83355.6 2901 No MSA 3.2 171 5.8 306 2.9 153 3.7 193 6.3 32622.0 1149 Post-MSA 2.3 124 2.4 125 3.5 185 6.3 330 7.6 39722.2 1161 Total10.8 56619.1 99822.0 114918.1 94229.8 1556

PAGE 44

44 Figure 3-6. Seasonal distribution of 15-minute scan samples of three Saimiri associations 0 2 4 6 8 10 12 14 16 18 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 45

45 Table 3-7. Proportio nal distribution of Saimiris level of integration in each season in MSAs with Cebus Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 Total % n % n % n % n % n % n Overlap 13.3 36 5.3 30 1.8 14 6.2 2634.5 287 13.6 393 Partial Overlap43.9 11969.3 39346.7 37354.7 22939.4 328 49.9 1442 No Overlap 42.8 11625.4 14451.5 41139.1 16426.2 218 36.5 1053

PAGE 46

46 Figure 3-7. Proportio nal distribution of Saimiris level of integration in each season during MSAs with Cebus 0 10 20 30 40 50 60 70 80 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season Overlap Partial Overlap No Overlap

PAGE 47

47 Table 3-8. Genera feeding rate observed in each Saimiri association category across all scan samples (N=5211). Genera ordered from most consumed to least consumed in all Saimiri associations. MSANo MSAPost-MSATotal Inga 4.11.31.16.4 Pourouma 2.41.71.35.4 Mendocia 0.80.30.31.5 Paullinia 0.80.10.31.2 Tetragastris 0.60.20.10.9 Mucuna 0.40.10.30.8 Cecropia 0.40.10.20.7 Guadua 0.50.040.10.6 Passiflora 0.30.10.10.6 Maripa 0.30.20.020.5 Zygia 0.20.30.040.5 Protium 0.20.10.020.4 Macuna 0.10.20.10.3 Pinzona 0.20.020.020.2 Micropholis 0.10.2 0.2 Cayaponia 0.20.020.020.2 Abuta 0.10.10.020.2 Astrocaryum 0.20.02 0.2 Piper 0.1 0.10.2 Trichilia 0.2 0.2 Laetia 0.1 0.020.1 Calycorectes 0.10.04 0.1 Prieurella 0.10.1 0.1 Syngonium 0.10.040.020.1 Guatteria 0.1 0.1 Pouteria 0.1 0.040.1 Virola 0.10.02 0.1 Amanoa 0.04 0.040.1 Cordia 0.1 0.020.1 Couepia 0.020.020.040.1 Eugenia 0.040.04 0.1 Moutabea 0.1 0.020.1 Strychnos 0.10.02 0.1 Cupania 0.1 0.1 Euterpe 0.02 0.040.1 Faramea 0.1 0.1 Ficus 0.1 0.1 Heteropsis 0.020.020.020.1 Manilkara 0.020.04 0.1 Talisia 0.1 0.1 Amaioua 0.04 0.04 Bonafousia 0.04 0.04 Coussapoua 0.020.020.04 Eschweilera 0.020.02 0.04 Myrcia 0.020.02 0.04 Oenocarpus 0.04 0.04 Petrea 0.04 0.04 Spondias 0.04 0.04 Aegiphila 0.02 0.02 Carapa 0.02 0.02 Clusia 0.02 0.02 Chrysophyllum 0.020.02 Eperua 0.02 0.02 Gnetum 0.02 0.02 Gustavia 0.02 0.02 Helicostylis 0.02 0.02 Iryanthera 0.02 0.02 Siparuna 0.02 0.02 Socratea 0.02 0.02 Symphonia 0.02 0.02 Tabebuia 0.02 0.02 Total 13.75.54.323.6

PAGE 48

48 Table 3-9. Family feeding rate observed in three Saimiri associations throughout all scan sample intervals observed (N=5,211). Ranked fr om most consumed to least consumed. MSANo MSAPost-MSATotal Mimosaceae 4.41.61.17.1 Cecropiaceae 2.81.81.56.1 Acanthaceae 0.80.30.31.5 Sapindaceae 0.90.20.31.4 Burseraceae 0.80.30.11.2 Papilionaceae 0.40.10.30.8 Arecaceae 0.50.20.10.8 Graminaceae 0.50.040.10.6 Sapotaceae 0.20.20.10.6 Passifloraceae 0.30.10.10.6 Leguminosae 0.10.20.10.3 Araceae 0.10.10.040.2 Dillenaceae 0.20.020.020.2 Myristicaceae 0.20.1 0.2 Cucurbitaceae0.20.020.020.2 Meliaceae0.2 0.2 Piperaceae0.1 0.10.2 Menispermaceae0.10.10.020.2 Flacourtiaceae0.1 0.020.1 Annonaceae0.1 0.1 Rubiaceae0.10.04 0.1 Bignonaceae0.10.02 0.1 Boraginaceae0.1 0.020.1 Chrysobalanaceae0.020.020.040.1 Euphorbiaceae0.04 0.040.1 Lecythidaceae0.040.04 0.1 Myrtaceae0.040.04 0.1 Polygalaceae0.1 0.020.1 Clusiaceae0.1 0.1 Moraceae0.1 0.1 Verbenaceae0.020.04 0.1 Anacardiaceae0.04 0.04 Apocynaceae0.04 0.04 Monimiaceae0.04 0.04 Viscaceae0.04 0.04 Caesalpinaceae 0.02 0.02 Concolvulaceae0.02 0.02 Gnetaceae0.02 0.02 Solanaceae0.02 0.02 Total13.75.54.323.6

PAGE 49

49 Table 3-10. Genera foods consumed more than, less than or not at all in relation to each Saimiri association category MSA > No MSAMSA > Post-MSANo MSA > MSANo MSA > Post-MSAPost-MSA > MSAPost-MSA >No MSAMSA=0No MSA=0 Post-MSA=0 IngaIngaInga Pourouma Pourouma Pourouma Mendocia Mendocia Mendocia Paullinia Paullinia Paullinia Tetragastris Tetragastris Tetragastris Mucuna Mucuna Mucuna Cecropia Cecropia Cecropia Guadua Guadua Guadua Passiflora Passiflora Passiflora Maripa Maripa Maripa Zygia Zygia Zygia Protium Protium Protium Macuna Macuna Macuna Pinzona Pinzona Micropholis Micropholis Micropholis Cayaponia Cayaponia Abuta Abuta Astrocaryum Astrocaryum Astrocaryum Astrocaryum Piper Piper Piper Piper Trichilia Trichilia Trichilia Trichilia Laetia Laetia Laetia Laetia Calycorectes Calycorectes Calycorectes Calycorectes Prieurella Prieurella Syngonium Syngonium Syngonium Guatteria Guatteria GuatteriaGuatteria Pouteria Pouteria Pouteria Pouteria Virola Virola Virola Virola Amanoa Amanoa Amanoa Cordia Cordia Cordia Cordia Couepia Couepia Eugenia Eugenia Eugenia Moutabea Moutabea Moutabea Moutabea Strychnos Strychnos Strychnos Strychnos Cupania Cupania Cupania Cupania Euterpe Euterpe Euterpe Euterpe Faramea Faramea Faramea Faramea Ficus Ficus Ficus Ficus Manilkara Manilkara Manilkara Manilkara Talisia Talisia Talisia Talisia Amaioua Amaioua Amaioua Amaioua Bonafousia Bonafousia BonafousiaBonafousia Coussapoua Coussapoua Coussapoua Eschweilera Eschweilera Eschweilera Myrcia Myrcia Myrcia Oenocarpus Oenocarpus OenocarpusOenocarpus Petrea Petrea Petrea Petrea Spondias Spondias SpondiasSpondias Aegiphila Aegiphila AegiphilaAegiphila Carapa Carapa Carapa Carapa Clusia Clusia Clusia Clusia ChrysophyllumChrysophyllumChrysophyllumChrysophyllum Eperua Eperua Eperua Eperua Gnetum Gnetum Gnetum Gnetum Gustavia Gustavia Gustavia Gustavia Helicostylis Helicostylis HelicostylisHelicostylis Iryanthera Iryanthera Iryanthera Iryanthera Siparuna Siparuna Siparuna Siparuna Socratea Socratea Socratea Socratea Symphonia Symphonia SymphoniaSymphonia Tabebuia Tabebuia TabebuiaTabebuia

PAGE 50

50Table 3-11. Family foods consumed more than, less than or not at all in relation to each Saimiri association category MSA > No MSAMSA > Post-MSANo MSA > MSANo MSA > Post-MSAPost-MSA > MSAPost-MSA >No MSAMSA=0No MSA = 0Post-MSA=0 MimosaceaeMimosaceae Mimosaceae CecropiaceaeCecropiaceae Cecropiaceae AcanthaceaeAcanthaceae Acanthaceae SapindaceaeSapindaceae Sapindaceae Burseraceae Burseraceae Burseraceae PapilionaceaePapilionaceae Papilionaceae Arecaceae Arecaceae Arecaceae GraminaceaeGraminaceae Graminaceae Sapotaceae Sapotaceae PassifloraceaePassifloraceae Passifloraceae LeguminosaeLeguminosae Leguminosae Araceae Araceae Araceae Dillenaceae Dillenaceae Myristiceae Myristiceae Myristiceae Myristicaceae CucurbitaceaeCucurbitaceae Meliaceae Meliaceae MeliaceaeMeliaceae Piperaceae Piperaceae Piperaceae Piperaceae MenispermaceaeMenispermaceaeMenispermaceae Flacourtiaceae FlacourtiaceaeFlacourtiaceae Flacourtiaceae Annonaceae Annonaceae AnnonaceaeAnnonaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae BignonaceaeBignonaceae Bignonaceae Bignonaceae BoraginaceaeBoraginaceae Boraginaceae Boraginaceae ChrysobalanaceaeChrysobalanaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Lecythidaceae Lecythidaceae Lecythidaceae Myrtaceae Myrtaceae Myrtaceae PolygalaceaePolygalaceae Polygalaceae Polygalaceae Clusiaceae Clusiaceae ClusiaceaeClusiaceae Moraceae Moraceae MoraceaeMoraceae VerbenaceaeVerbenaceae Verbenaceae Verbenaceae AnacardiaceaeAnacardiaceae AnacardiaceaeAnacardiaceae ApocynaceaeApocynaceae ApocynaceaeApocynaceae MonimiaceaeMonimiaceae MonimiaceaeMonimiaceae Viscaceae Viscaceae ViscaceaeViscaceae CaesalpinaceaeCaesalpinaceae Caesalpinaceae Caesalpinaceae ConcolvulaceaeConcolvulaceae ConcolvulaceaeConcolvulaceae Gnetaceae Gnetaceae GnetaceaeGnetaceae Solanaceae Solanaceae SolanaceaeSolanaceae

PAGE 51

51 Table 3-12. Seasonal difference in quantity of food consumption in th ree Saimiri association categories and the overall distribution of foods consumed across seasons Fruit 1 Feb-Apr 00 Wet May-Jul 00 Flower Aug-Oct 00 Transition Nov Jan 01 Fruit 2 Feb-Apr 01 Total per Association % n % n % n % n % n % n MSA 31.7 3958.8 57 87.5 2153.6 11962.7 47858.1 714 No MSA 57.7 7123.7 23 4.2 125.6 6017.2 13323.4 288 Post-MSA 10.6 1317.5 17 8.3 218.4 4319.8 15118.4 226 Total per Season 9.8 123 7.7 97 1.9 2418.1 22262.0 762

PAGE 52

52 Figure 3-8. Seasonal difference in quantity of foods consumed in three Saimiri association categories 0 10 20 30 40 50 60 70 80 90 100 Fruit 1 (Feb-Apr 00) Wet (May-Jul 00) Flower (Aug-Oct 00) Transition (Nov Jan 01) Fruit 2 (Feb-Apr 01) Season No MSA MSA Post-MSA

PAGE 53

53 CHAPTER 4 DISCUSSION In this study, we approached each genera l hypothesis of MSA function by applying our interval scan sampling data with the functions predicted mechanisms. If our data supported a functions mechanism, the hypothesized functi on became a stronger candidate as a viable explanation for Saimiri-Cebus associations. Here, we built an ecologically realistic picture using a long-term study that represents a dynamic system. MSA studies using empirical data commonly apply only one explanation to an MSA. Applying one explanatory functi on is appropriate for MSAs that are permanent to semipermanent such as MSAs between two Saguinus species (Terborgh 1983; Norconk 1990). However, MSAs that are dynamic should not be given the same inclusive approach since the driving force for these associations are ever changing ecological factors: stringent food availability, competition or predat or pressure. Due to this dynamic, we approach this study by analyzing the patterns and processe s of MSAs to test for more th an one possible function. As a result, we predict dynamic MSAs such as Saimiri-Cebus associations, would have variable associations according to th e local ecological factors. Saimiri-Cebus Associations: Random or Intended? Saimiri associated with Cebus for 2,901 (55.6%) 15-minute interval scan samples of the 5,211 scan sample data collected across five seasons. Saimiri-Cebus association in Manu, Peru associates up to 90% of the scan sample da ta (Terborgh 1983) and only 6.6% in Corcovado, Costa Rica (Boinski 1989). In this study, 1,149 (20.9%) 15-minute interval scan samples were taken when Saimiri had completely no contact or influence of Cebus We predict if Saimiri and Cebus associations were random events (Waser 1982), the proportion of MSA scan samples

PAGE 54

54 would resemble Saimiri Cebus associations in Corcovado, Costa Rica. Instead, our results correspond to Manus Saimiri-Cebus association pattern. To further support this n on-random association between Saimiri-Cebus differences in seasonal and habitat use in Saimiri-Cebus association categories were observed (Figure 3-3 and Table 3-3). Saimiri associated with Cebus during the flower seas on (August October 2000) when food is more limited. Saimiri s feeding rate, as shown in Figure 3-8, increased during the flower season (August October 2000) when associated with Cebus Risky habitat use, bamboo and swamp habitat, were rarely exploited when Saimiri were out of association (Table 3-1 and Figure 3-1). Habitats ar e classified as risky according to how exposed an arboreal primate is to aerial predators (i.e., leaf cover), and how low the forest substrates are for terrestrial predators to easily capture their arboreal prey Food and habitat use are two ecological variab les contributing to association patterns between Saimiri and Cebus Risky habitats re quire the presence of Cebus to reduce predation risk, wh ere low food seasons require Cebus for their foraging knowledge (e.g., fruit location sites). General Hypothesis 1: Reduce Predation Risk Mechanism 1: Saimiri Will Tolera te Riskier Habitats in an MSA and Will Be Intolerant of Risky Habitats Outside of an MSA Background. Risky habitats have low foliage cove r and limited options for escape. In Raleighvallen, swamp and bamboo habitats are cons idered risky. The swamp habitat has little food, low foliage cover and a large open unde rstory. A large open understory creates a microhabitat that has minimal options for arbor eal primates to travel. The only pathways Saimiri can travel are the top of the canopy, the ground, or jumping from one vertical substrate to another. This type of habitat structure makes it easy for predators to capture prey. Raptors can fly at ease in the open understory and terrestrial predators can easily detect and locate their prey.

PAGE 55

55 The bamboo habitat has little to no fruit, zero visibility due to dense vegetation and substrates are very low to the ground (0-5m). In the bamboo, Saimiri forage on bamboo shoots and arthropods (Boinski et al. in prep.). Predators that fre quent bamboo are felids and snakes (Boinski at al. in prep.). Dense vegeta tion makes it an ideal place for an ambush. The liana habitat is classified as a non-risky habitat because it has heavy foliage cover and high fruit availability (Boinski et al. in prep.). Due to heavy foliage cover, visibility for aerial predators is low. The habitats microha bitat characteristics create a safe area for many arboreal species. Prediction. If Saimiri associated with Cebus for reduced predation risk, then Saimiri will be observed to be more tolerant of risky habitats in an MSA and less tolerant of risky habitats outside of an MSA. Saimiri will access the risky habitats more frequently when Cebus is present so Saimiri can use a number of antipredatory st rategies such as: dilution effect (Terborgh 1986) and increased vigilance (Morse 1977). Results. According to our Chi-Square Test results, Saimiri associated with Cebus more likely than expected in the bamboo and swamp habitats. Saimiri rarely ventured into the swamp and bamboo habitats without Cebus (Table 3-1 and Figure 3-1). These results support the prediction that Saimiri needed the presence of Cebus to access these risky habitats. This pattern suggests that Saimiri associated with Cebus preferentially in riskier ha bitats as a means to reduce the inherent predation risk. Saimiri associated less frequently with Cebus in liana habitat. Liana habitat is the habitat most used by Saimiri especially when not associated with Cebus (Table 3-1 and Figure 3-1). Saimiri appears to avoid risky habitats more without Cebus present. Saimiri may not need Cebus for antipredatory strategies such as increased vigilance (Morse 1977).

PAGE 56

56 Mechanism 2: With Exception to Seasons with Low Foliage Cover, Saimiris Rate of Association Will Occur a cross All Seasons Equally Background. Predation events are rare but occur continually throughout the year (Fragaszy et al. 2004). In Raleighvallen, Saimiri are prey for terrestria l and aerial predators. Aerial predators include rapt ors such as harpy eagles ( Harpia harpyja ) and ornate hawk eagles ( Spizaetus ornatus). Terrestrial predators are jaguars ( Panthera onca ), pumas ( Puma concolor) and constricting or venomous snakes --e.g., Boa constrictor -(Fragaszy et al. 2004). Although predation is a continual obstacle, aerial predator s may have an advantage in detecting prey during periods of low foliage cover. The lower the leaf cover, the less areas prey can hide. In Raleighvallen, the highest percentage of foliage c over occurs during the wet season (May-July) and flower season (August October) (Boinski et al. in prep.). The lowest foliage cover takes place in the fruit season (Februar y April) (Boinski et al. in prep.). Prediction. If predation were an ongoing threat, then no seasonal difference in association would be present if Saimiri associated with Cebus for antipredatory strategies. Results. Saimiri associated with Cebus at different rates across all five seasons (Table 33 and Figure 3-3). If Saimiri formed an MSA for predator av oidance strategies, Figure 3-3 would not have changes in association frequency. Instead, Saimiri actively associated more during the wet season (May Ju ly 2000), flower season (Augu st October 2000) and second fruit season (February April 2001). Saimiri associated less during the first fruit season (February April 200) and tran sition season (November January 2001). According to ChiSquare Test results, Saimiri associated significantly more during the flower season (August October 2000) with less associa tions expected in the transition season (November January 2001) and both fruit seasons (February April 2000/2001).

PAGE 57

57 In regards to foliage cover, Saimiri associated with Cebus during periods of high foliage cover, therefore, predation a voidance may not be the only func tion driving the association. Saimiri associated significantly mo re during seasons with the hi ghest foliage cover: the wet season (May July 2000) and the flower season (August October 2000). Mechanism 3: When Cebus is Present, Foods Consumed Will Have Higher Feeding Rates than When Cebus is Not Present (MSA>No MSA + Post-MSA) Background. When feeding in a tree, individu als in an MSA can increase their food intake because they spend less time looking out fo r predators (Morse 1977) as well as benefiting from other antipredatory benefits such as th e dilution effect (Terbor gh 1986) and confusion effect (Morse 1977) discussed above. When Saimiri is not in an MSA they have lower food intake since they have to increase their vigila nce or avoid open microhab itats that put them at risk of an aerial predation ev ent (Boinski et al. 2003). Prediction. We predict Saimiri would have a higher feeding rate in MSA than both PostMSA and No MSA associations combined. The absence of Cebus would have caused Saimiri to feed less since Saimiri had to make up for the absence of antipredatory strategies. Results. Table 3-5 and Table 3-6 lists all f ood genera and family feeding rates by Saimiri in all association categories. The family and genera feeding rate for MSA was 13.7, whereas No MSA and Post-MSA combined had a feeding rate of 9.8 (13.7 > 9.8). This mechanism supporting predator avoidance function for Saimiri Cebus MSA is supported by our data. An additional explanation for this result could be that Saimiri are exploiting more foods due to Cebus knowledge of food locations. Mechanism 4: MSAs Will Form and Depart in a Safe Habitat Background. Saimiri would form an MSA with Cebus in a non-risky habitat since the habitat Saimiri occupies alone must be safe from pr edators (e.g., high fo liage cover). As

PAGE 58

58 indicated in Table 3-1 and Figure 3-1, the liana habitat is exploited the most when Saimiri is in a conspecific group. Prediction. We would predict the liana habitat would be th e site of MSA formation since it has a microhabitat with high foliage cover and poses as a safe zone for pick-up and drop off. Results. Saimiri Cebus formation location was found to be insignificant. In short, Saimiri met up with Cebus in random habitats. This result could be explained by looking at Saimiri Cebus associations in Manu, Peru. In Manu, Saimiri would depart from a Cebus troop if there was another Cebus troop near by or if the Cebus troop was stationary (Terborgh 1983). The randomness of the formation could be the result of breaking from another Cebus nearby and Saimiri risking entering into anot her habitat where the other Cebus is located. MSA departure, on the other hand, was not a random event. Statistical analysis shows Saimiri did not depart from an MSA in a swamp hab itat. Not departing in the swamp habitat shows Saimiri perceived this habitat or the microhabitat to be risky without Cebus The liana forest was the habitat most likely for Saimiri to depart from an MSA. MSA formation and departure mechanism is not fully supported by our data. In future MSA investigations, observers s hould make detailed notation of habitat type, direction of Saimiri troop travel and duration of stay duri ng formation and departure of MSAs. Mechanism 5: Saimiri Will Tolerate Habitats with Micro habitats of Reduced Foliage Cover More Frequently during an MSA Background. Low foliage cover allows aerial predat ors such as raptors, to spot their prey easily. Foliage cover is the overall leaf co ver of a tree species. The greater the foliage cover, the more an arboreal species is hidden in the microhabitat. Open microhabitats with low

PAGE 59

59 foliage cover create vulnerability to arboreal sp ecies. The openness fails to hide the arboreal species and provides an easy acc ess for the aerial predator. Prediction. If Saimiri used Cebus for antipredatory strategies, then Saimiri would tolerate low foliage habitats in MSAs. Results. The foliage cover data was inconclusi ve. The total data set contained 5,211 15minute scan sample intervals. Due to the large sample size, the results were automatically significant. The immense number of samples crea ted the statistical analysis to be highly significant, not the data it was an alyzing. In future MSA research, analysis of foliage cover should not include a high sample size. Instead, the research er should do random foliage cover sampling or use finer grain microhabitat pheno logy data to assess the foliage cover. Protocol for future MSA studies: Our data does not support some of the mechanisms suggesting Saimiri associated with Cebus for only reducing in predation risk. Suggestions for future research should: record MSA reactions to predation even ts (e.g., duration, microhabitat use before and after), between species microhabitat use (i.e., are both species using open microhabitats?), collect scan sample data from heterospecific associating species and collect finer grain microhabitat phenology. General Hypothesis 2: Foraging Benefits Mechanism 6: MSAs Will Occur Du ring Periods of Low Food A vailability to Locate Food and Reduce Duplication Efforts Background. Enhanced food efficiency is attain ed through various ways in an MSA: location of food sites (Gartlan a nd Struhsaker 1972), flushed insects or prey (Peres 1992) and decrease in duplication effort (Morse 1977). Species use MSAs during periods of low food availability or to expand their food taxonom ic diversity such as that observed with Cercopithecus monkeys (Cords 1986).

PAGE 60

60 In Raleighvallen, the period of the lowe st food consumption was the flower season (August October 2000) and the highest food availabi lity was the second fr uit season (February to April 2001). Saimiri ate the most in the second fruit season (February April 2001) followed by Transition (November January 2001), first fruit season (Febru ary April 2000), wet season (May July 2000) and flower season (August October 2000) be ing the lowest (Table 3-9). Prediction. If Saimiri associated with Cebus for locating food during periods of food scarcity (Gartlan and Struhsaker 1972), then Saimiri Cebus associations woul d increase during the flower season (August Oc tober 2000), first fruit season (F ebruary April 2000) and wet season (May July 2000). The purpose of the asso ciations would be to locate food sites (1972) and possibly to avoid duplic ation efforts (Morse 1977). Results. Saimiri associated with Cebus more than expected during the flower season (August October 2000) and significantly less duri ng the transition season (November January 2001) and both fruit seasons (February Apri l 2000/2001). This resu lt correlates with food availability (Figure 3-9). The flower season (August Oc tober 2000) had the lowest food availability and analysis showed significantly higher rates of associati on (Figure 3-9 and Table 3-3). Proportion of association increased from the first fruit season (February April 2000) to the wet season (May July 2000), and peaked du ring the flower season (August October 2000). Competition could be high during low food seasons but Saimiri refrained from overlapping with Cebus to minimize competition (Figure 3-4 and Table 3-4). Overlap increased in seasons of high food availabil ity (Figure 3-4). This showed Saimiri s awareness in food scarcity and the mechanisms they used to decrease competition.

PAGE 61

61 Mechanism 7: Foods Consumed During MSAs Will Have a Higher Feeding Rate than When Saimiri Are Not in an MSA (MSA > No MSA) Background. Associating with a symp atric species can increase food intake due to their acute food location knowledge (Morse 1977). Speci es with large home ranges, as suggested by Gartlan and Struhsaker (1972), may unite with another species with a smaller home range to benefit from their knowledge of lo cal food sites. A species with a smaller home range has more detailed recollection of where the food is in comparison to a species with a large home range (Gartlan and Struhsaker 1972). Saimiri has a home range of 44-130ha (Robinson and Janson 1987) and Cebus has a home range of ca. 80ha (Terborgh 1983). Prediction. Saimiri s feeding rates are expected to increase during an MSA as a result of exploiting Cebus food knowledge. If Saimiri were in a conspecific gr oup, these foods could not be located. Results. The feeding rate during an MSA was 13. 5 and 5.5 in No MSA. (13.5>5.5). This prediction is supported by the data. Mechanism 8: Foods Consumed During Post-M SA Intervals Will Have a Higher Feeding Rate than When Not in an MSA (Post-MSA>No MSA) Background. Following a species to a food loca tion can increase their food intake without the actual presen ce of the other MSA species. The Po st-MSA association category, as stated earlier, occurred when Saimiri were associated with Cebus in previous or between MSA intervals. Post-MSA includes Cebus -influenced incidences such as Cebus traveled away from a fruit tree while Saimiri remained in the tree happily fora ging alone. This example shows that Cebus could lead them to a food source whil e increasing their f eeding rate without Cebus being present. This strategy could be a mechanism to keep competition low. Prediction. If Saimiri s sole function in associating with Cebus were enhancing foraging efficiency, then the genera and family f eeding rate would be greater than when Saimiri was not

PAGE 62

62 associated. Saimiri were able to explo it the food knowledge of Cebus and, as a result, should access more food than without any influence from Cebus Results. Post-MSA feeding rate was 4.3 and No MSA had 5.5 ( 4.3>5.5). This prediction is not supported by the data because Saimiri had a higher feeding rate during No MSA. Mechanism 9: Foods Consumed in Post-MSA and MSA Intervals Combined Will Have a Higher Feeding Rate than When Not in an MSA (Post-MSA + MSA > No MSA) Background. Attaining food knowledge from an MSA species may increase the associating species diet. Post-MSA and MS A associations are categories that include Cebus influencing or enhancing Saimiri s foraging efficiency. Whether Cebus led Saimiri to a fruiting tree or Saimiri ate discarded partly eaten fruit, these all influenced the diet of Saimiri Prediction. If Saimiri seek Cebus for foraging benefits, then Post-MSA and MSA feeding rate would be higher than no MSA. Results. Combined Post-MSA and MSA has 17.8 feeding rate, whereas No MSA has 5.5 feeding rate (17.8 > 5.5). The data supported this prediction. Mechanism 10: Saimiri Will Have Greater Food Taxon omic Diversity in an MSA Background. In an MSA, a species can gain access to wider variety of fruits due to the smaller home ranged species acu te knowledge of what fruits are ripe and still abundant (Terborgh 1983). Saimiri s diet was recorded during each 15-minute scan sample interval (N=5211) and each food item was assigned into their family and genus (e.g ., InEd: Inga and Mimosaceae). Across five seasons of data, 39 fa milies and 61 genera were recorded and analyzed. In Tables 3-7 and 3-8, each family and genera were documented as having been consumed more, less or not at all in each Saimiri association category. The Family and Genus data are ranked from most to least c onsumed across all association categories.

PAGE 63

63 Prediction. If Saimiri associated with Cebus for foraging benefits, then Saimiri would have greater diversity in food types when they were with Cebus or had direct influence of Cebus Therefore, Saimiri would have greater diversity in food types in MSA than in No MSA, and greater diversity in Post-MSA than No MSA. Results. Diversity in consumed food type s varied between Post-MSA and MSA categories. Of the 39 food families, 31 families we re eaten more in MSA then in No MSA. Of the 31 families eaten more in MSA, 16 of the families were not consumed (n=0) when Saimiri was in conspecific groups. In food genera, 43 we re eaten more in MSA than No MSA. Of the 43 genera, 22 genera were not eaten (n=0) durin g No MSA. These data support the contention that increased of foods are consumed during MSA than in conspecific groups. Of the 39 food families, nine families were eaten more than in No MSA but 17 families were eaten more in No MSA than Post-MSA. Of the 61 genera, 14 were eaten more in PostMSA then No MSA. But, No MSA had 25 genera eaten more than in Post-MSA. If food were a primary factor for MSAs, then having knowledge from Cebus would produce a greater yield of food consumption then if they were alone. These results fail to support this prediction. Protocol for future MSA studies: Some suggestions for future research is to record all arthropods consumed to get a more accurate di et of the MSA species. Document distance between conspecific and heterosp ecific during feeding bouts in all seasons to see how all species involved are minimizing competition. Collect scan samples on heterospecific associating species and note behaviors of heterosp ecific during low food periods.

PAGE 64

64 General Hypothesis 3: Reduce Predation Risk and Foraging Benefits Mechanism 11: Mixed-species Groups Will Occur During Periods of Low Food Availability and in Riskier Habitats Background. The function of sympatric associations, according to Gartlan and Struhsaker (1972), is to locate f oods during seasons of low food av ailability. In Cochu Cashu, locality of food is said to be part of the explanation for Saimiri Cebus associations (Terborgh 1983). Terborgh (1983) argues food locality is not the only explanation since associations continue well after Saimiri and Cebus left a fruiting tree. A species may form MSAs for attaining fruits in periods of f ood scarcity while taking bigger risk s since they have the presence of another species to use antipredatory strategies. Prediction. If the function for Saimiri Cebus associations were for locating food with an added perk of anti-predatory strategies, Saimiri would associate with Cebus during low food availability and in risky habitats. Results. Saimiri associated with Cebus in risky habitats during low food availability periods. One risky habitat, the bamboo, was accessed significantly more than expected during MSAs across all five seasons. This result supported Saimiri not entering the bamboo patches without the presence of Cebus Saimiri does not find this habitat to be a vital source for food or safe haven from predators. The next risky habitat, the swamp, had similar results across four of the five seasons, transition season yielding no significance. Saimiri were significantly more likely to be in MSAs with Cebus Similar to the bamboo, results show the interpretation that Saimiri did not enter the swamp without Cebus Saimiri may have perceived this habitat or the microhabitat to be both unsafe and containing no food. Seasonal distribution of habitat use in Saimiri associations is presen ted to further support this prediction (Table 3-2 and Figure 3-2). The bamboo and swamp habitats both exhibit peaks in

PAGE 65

65 habitat use during the low food av ailability season, the flower season (August October 2000). Both swamp and bamboo habitats, as illustrated in 3-1, are not exploited without the presence of Cebus Season-habitat association distributions and statistical analyses showed Saimiri occupied the riskiest habitats in MSAs not only in low food availability periods but also in high food abundance periods. The seasons of high food availa bility are: second fr uit season (February April 2001), transition season (November Januar y 2001) and the first fruit season (February April 2000) as shown in Table 3-9. Saimiri associated with Cebus significantly more during these seasons in both bamboo and swamp hab itats. We can postulate that food foraging efficiency may not be the only factor driving Saimiri to associate with Cebus Mechanism 12: During Seasons with Low Food Availability, Saimiri Will Forage in Microhabitats with Low Foliage Cover in an MSA Results. As stated earlier, the data set had a large sample size making the statistical result to have an automatic significant association. Therefore, no conclusion was postulated. Mechanism 13: Feeding Rates Will be Highest with Cebus Present (MSA> No MSA and MSA> Post-MSA) Background. In an MSA, species have access to both antipredatory strategies and foraging benefits. The species being exploited not only gives the other species abundant food source knowledge (e.g., fruit location), but the enla rged group size provides species to use risky microhabitats. Prediction. If the function for Saimiri associating with Cebus were to benefit from enhanced foraging and predator avoidance, the fe eding rates would be th e highest during MSA. With the protection of a larger group, Saimiri is able to spend more time feeding and less time looking out for predators.

PAGE 66

66 Results. The presence of Cebus MSA, will give Saimiri increased predator avoidance needed to invest more time feeding. In MSA, Saimiri had a feeding rate of 13.7. In No MSA and Post-MSA, Saimiri had feeding rates of 5.5 and 4.3, respectively. The data supported the presence of this mechanism (13.7 > 5.5 and 13.7 > 4.3). Although the Post-MSA association gave Saimiri foraging efficiency (i.e., location of food) it fails to provide the increased vigilance required to reach fruits in open microhabitats Analyzing feeding rates in the different association categories gave us a detailed picture of how Saimiri perceived the benefits of MSA. Mechanism 14: Saimiri Will Have Greater Good T axonomic Diversity in MSAs Background. An MSA could give a species the advantage of a greater variety of taxonomic diet while maintaining relatively low pr edation risk to access those foods. In Manu, Cebus has better knowledge of food locations because they have a smaller home range in comparison to Saimiri (Terborgh 1983). Saimiri exploits Cebus argues Terborgh (1983), for the foraging knowledge and antipredatory strategies th at may result in their increasing diet breadth when they are in an MSA. In Raleighvallen, any food Saimiri consumed was recorded within each 15-minute interval scan sample (N=5,211). In this data set, every food item was grouped into family and genera (e.g., InEd: Inga and Mimosaceae). In five seasons of data, 39 families and 61 genera were noted. Figures 3-7 and 3-8 presents which family or genera were consumed more, less than or not at all in each Saimiri association category. The families and genera are ranked from most to least consumed across all association categories. Prediction. If Saimiri used Cebus for foraging efficiency and reduce predation risk, then Saimiri s diet will have greater taxonomic diversity in MSA. Results. Data showed Saimiri s diet had greater diversity in MSA than other association categories. In Figure 3-8, 36 families were eaten more in MSA than in both Post-MSA and No

PAGE 67

67 MSA. Of the 39 families, only four were consumed more in No MSA and two were consumed more in Post-MSA. In Figure 3-7, 51 genera we re consumed more in MSA. Of the 61 genera, 11 were consumed more in No MSA and 5 were consumed more in Post-MSA than MSA. Food families and genera not consumed during Post-MSA and No MSA gave an interesting outlook on what foods Saimiri may be unable to locate alone or the microhabitat is too risky without Cebus In Figure 3-8 and 3-7, 22 families and 41 genera were not consumed (n=0) in No MSA and Post-MSA. These ta xonomic foods were unconsumed without the presence of Cebus This illustrates Saimiri relied on an MSA to access these fruits for both predator avoidance and foraging knowledge. Assuming after Saimiri is led to a food source by Cebus they would be able to locate this food again; why wouldnt they want to attain these foods without Cebus ? The answer could be predat or deterrence and detection. Saimiri s diet recorded across five seas ons supports the prediction that Saimiri had greater diversity in diet when th ey were in MSA or influenced by Cebus Without the presence of Cebus Saimiri were not able to locate many foods while maintaining low predation risk. The food data supported the presence of this mechanism. Protocol for future MSA studies: Future MSA research should analyze more than a year or season of MSA patterns. If ecological variables asse ss the MSA behavioral pattern, looking at more than one year will give a better picture of what predicts the associations: food, predator pressure or both. A nother suggestion includes looking at the individual level of an MSA. Group level MSA patterns have been resear ched thoroughly, the next logical step would be identifying individual MSA patte rns and processes. Data should include: heterospecific scan sample data, individual heights or distance to heterospecific, predation event reactions, foliage cover and detailed dietary intake.

PAGE 68

68 Conclusion 1. MSA formation had random associations with habitat type, but the maintenance of MSAs was not random. Association patterns of habitat use and seasonal distribution strongly correlate with periods of low food abundance an d use of habitats in MSA where predation risk was high. 2. Saimiri had an increase in foraging e fficiency that was seen through Saimiri s: increased food diversity in MSA, high feeding rate in MSA, accessing risky habitats during food scarcity in MSA and high asso ciation frequency during low f ood availability seasons. 3. Strong evidence supported the combined reduce predation risk-forage efficiency theory through habitat data, season data and the food family and genera Saimiri ate with Cebus Both genera and family foods were eaten more with the presence of Cebus If Saimiri only wanted foraging benefits, Post-MSA would have a higher feeding rate. On the other hand, if Saimiri only needed Cebus for predator avoidance, MSAs would occur across all seasons equally. But, results show associations were more likely to occur during periods of low food availability. 4. Reduce predation risk and increased foragi ng efficiency theories are both functional explanations for Saimiri Cebus MSAs in Raleighvallen, Suriname. Applying only one function to an MSA doesnt take into acc ount the ecological variables impacting the ecosystem. Every year will have different f ood yield, rainfall and temperature. These factors impact the ecosystem and, as a result the community may alter their behavior to adapt to these ecological variables acro ss days or years. In this study, Saimiri illustrated how they adapted to food scarcity and pred ation pressure by asso ciating with another species, Cebus Ecological variables are th e driving forces behind Saimiri Cebus associations in Raleighvallen. Food and habitat drive Saimiri associations with Cebus, as seen in this study. Since ecological factors are not static, we should not expect MSAs to always operate under one explanatory function. More re search on group and indi vidual level MSA pa tterns and processes will give a more detailed picture of the variation in the proximate mechanisms and ultimate functions across years and seasons.

PAGE 69

69 LIST OF REFERENCES Alexander R D (1974) The evolution of social behavior. Ann Rev Ec ol Syst 5: 325-83 Bearzi M (2006) California sea lions use dolphins to locate food. J Mammol 87(3): 606-617 Beauchamp G (2004) Reduced flocking by birds on is lands with relaxed predation. Proc R Soc Lond B 271: 1039-1042 Boesch C (1994) Cooperative hunting in wild chimpanzees. Anim Behav 48: 653-667 Boinski S, Scott PE (1988) Association of bird s with monkeys in Costa Rica. Biotropica 20: 136-143 Boinski S (1989) Why dont Saimiri oerstedii and Cebus capucinus form mixed-species groups? Int J Primatol 10(2): 103-114 Boinski S (1999) The social organizations of squirrel monkeys: Implications for ecological models of social evolution. Evol Anthropol 8(3): 101-112 Boinski S (2002) Does group size reflect a tradeoff between predation risk and within-group competition? Am J Phys Anthrop 34: 46 Boinski S, Kauffman L, Ehmke E, Schet S, Vreedzaam A (2005a) Dispersal patterns among three species of squirrel monkeys ( Saimiri oerstedii, S. boliviensis and S. sciureus ): I. Divergent costs and benefits. Behaviour 142: 525-632 Boinski S, Kauffman L, Westoll A, Stickler C.M., Cropp S, Ehmke E (2003) Are vigilance, risk from avian predators and group size consequences of habitat structure? A comparison of three species of squirrel monkey (Saimiri oerstedii, S. boliviensis and S. sciureus). Behaviour 140: 1421-1467 Chapman CA, Chapman LJ (2000) Interdemic varia tion in mixed-species a ssociation patterns: common diurnal primates of Kibale National Park, Uganda. Behav Ecol Sociobiol 47: 129-139 Charnov EL, Orians GH, Hyatt, K (1976) Ecological implications of resource depression. Am Nat 110: 247-259 Cords M (1987) Mixed-sp ecies association of Cercopithecus monkeys in the Kakamega Forest, Kenya. University of California Publications in Zoology 117: 1-109 Cords M (2000) Mixed species association and gr oup movement. In: Boinski, S and Garber PA (eds) On the Move: How and why animals travel in groups, 1st edn. Chicago and London, pp 73-99 Davies CR, Ayres JM, Dye C, Deane LM (1991) Malaria infection rate of Amazonian primates increases with body weight and grou p size. Functional Ecol 5: 655-662

PAGE 70

70 Ehrlich PR, Ehrlich AH (1973) Coev olution: Heterotypic schooli ng of Caribbean reef fishes. Am Nat 107: 157-60 Ferrari SF (1990) A foraging associa tion between two kite species ( Ictinea plumbea and Leptodon cayanensis ) and buffy headed marmosets ( Callithrix flaviceps ) in Southeastern Brazil. Condor 92: 781-783 Fitzgibbon CD (1990) Antipredator strategies of immature Thom son gazelles hiding and the prone response. Anim Behav 40(5): 846-855 Freeland WJ (1977) Blood-sucking flies and prim ate polyspecific associations. Nature 269: 801802 Gartlan JS, Struhsaker TT (1972) Polyspecific associations and niche separation of rain-forest anthropoids in Cameroon, West Af rica. J. Zool. Lond 168: 221-226 Gautier-Hion A, Quris R, Gautier J (1983) Mono specific vs polyspecific life: A comparative study of foraging and antipredator y tactics in a community of Cercopithecus monkeys. Behav Ecol Sociobiol 12: 325-335 Grzimek B, Schlager N, Ole ndorf D (1972) Grzimeks Animal Life Encyclopedia. Gale Cengage, New York Hankerson SJ, Dietz JM, Raboy BE (2006) Associations between golden-headed lion tamarins and the bird community in the Atlantic forest of southern Bahia. Int J Primatol 27 (2): 487-495 Haugaasen T, Peres CA (2008) As sociations between primates a nd other mammals in a central Amazonian forest landscape. Primates 49(3): 219-222 May RM (1974) On the theory of niche overlap. Theor Popul Biol 5: 297-332 Morse DH (1977) Feeding behavior and predator avoidance in hete rospecific groups. Bioscience 27: 332-339 Mwangomo EA, Hardesty LH, Sinclair ARE, Mduma SAR, Metzger KL (2008) Habitat selection, diet and interspecific associations of the rufous-tailed weaver and Fischers lovebird. Afr J Ecol 46(3): 267-275 Noe R, Bshary R (1997) The formation of re d colobus-diana monkey associations under predation pressure from chimpanzees. Proc R Soc Lond B. 264: 253-259 Norconk MA (1990) Mechanisms promoting stability in mixed Saguinus mystax and S. fuscicollis troops. Am J of Primatol 21: 159-170 Norris KS, Dohl TP (1980) Behavior of the Hawaiian spinner dolphin, Stenella longirostris Fishery Bulletin 77: 821-849

PAGE 71

71 Nunn CL, Altzier S (2006) Infect ious Diseases in Primates: Be havior, Ecology and Evolution. Oxford University Press Inc., New York Peres CA (1992) Prey-capture benefits in a mixed-species group of Amazonian tamarins, Saguinus fuscicollis and S. mystax Behav Ecol So ciobiol 31: 339-347 Porter LM, Garber PA (2007) Niche expansion of a cryptic primate, Callimico goeldii while in mixed species troops. Am J Primatol 69: 1340-1353 Richardson DS, Bolen GM (1999) A nesting association between se mi-colonial Bullocks orioles and yellow-billed magpies: evidence for the predator protection hypot hesis. Behav Ecol Sociobiol 46(6): 373-380 Robinson JG, Janson JG (1987) Capuchins, squirre l monkeys, and atelines: socioecological convergence of Old World Primates. In: Smut s BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker TT (eds) Primate Soci eties. University of Chicago Press, Chicago and London, pp. 69-82 Semeniuk CAD, Dill LM (2006) Anti-predator be nefits of mixed-species groups of cowtail stingrays ( Pastinachus sephen ) and Whiprays ( Himantura uarnak ) at rest. Ethol 11: 33-43 Stensland E, Angerbjorn A, Berggren P ( 2003) Mixed species groups in mammals. Mammal Rev. 33(3): 205-223 Struhsaker TT (1981) Polyspecific associations among tropical ra in-forest primates. J Comp Ethol 57(3-4): 268-304 Terborgh J (1983) Five New Wo rld Primates: A Study in Comparative Ecology. Princeton University Press, Princeton Terborgh J, Landeau L (1986) Oddity and the conf usion effect in predation. Anim Behav 34: 1372-1380 Terborgh J (1990) Mixed flocks and polyspecific associations: costs and benefits of mixed groups to birds and monkeys. Am J Primatol 21: 87-100 Waser PM (1982) Primate polyspecific associati ons: Do they occur by chance? Anim Behav 30: 1-8 Waser PM (1984) Chance and mixed-species a ssociations. Behav Ecol Sociobiol 15: 197-202

PAGE 72

72 BIOGRAPHICAL SKETCH Carson Phillips was born and raised in Chicago, Illinois. She received a B.A. in anthropology from University of Arizona, Tu cson. After graduation, Carson was employed for one year as a field assistant studying brown capuchins for Dr. Sue Boinski in Raleighvallen, Suriname. Carson entered graduate school at the University of Florida in fall 2006.