Iowa State University Capstones, Theses and Graduate Theses and Dissertations Dissertations

1980 Size and spacing of sedentary guanaco family groups Robert Thomas Jefferson Jr. Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd Part of the Biology Commons

Recommended Citation Jefferson, Robert Thomas Jr., "Size and spacing of sedentary guanaco family groups" (1980). Graduate Theses and Dissertations. 16257. https://lib.dr.iastate.edu/etd/16257

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. - '(. Size and spacing of sedentary guanaco family groups .t~ --~ I '!/ ?.:J ~ :r~~ -9 by

Robert Thomas Jefferson, Jr.

A Thesis Submitted to the Graduate Faculty in Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE

Department of Ecology Major: Wildlife Biology

Signatures have been redacted for privacy Signatures have been redacted for privacy

Iowa State University Ames, Iowa 1980

1276116 i i

TABLE OF CONTENTS Page ABSTRACT iii

INTRODUCTION 1

Study Area 2a

METHODS 5 RESULTS 8 Distribution and Feeding Preferences of Vegetation Types 8

Feeding Territories 15a DISCUSSION 24 LITERATURE CITED 29 ACKNOWLEDGEMENTS 31 i i i

ABSTRACT

A socioecological study of the South American guanaco ( guanicoe) was conducted from July 1977 to June 1978 on Isla Grande, Tierra del Fuego, . The study area was situated in an ecotone between open Patagonian plains and closed austral forest habitats. Observations of guanaco family groups were made from a 2 story observation hut and from a motor vehicle positioned on the road . Twelve vegetation types were defined and analyzed by the point frame method to determine plant species composition and percent cover. Sites of territorial defenses by family group males were plotted; maps for neighboring males were then compared and territorial boundaries for each season were drawn. Feeding territories were defended year-round by the family group adult male; neighboring groups overlapped very little in their use of space. Shapes of feeding territories were approximately rectangular; territories averaged 29.5 ha in size. Territory quality was estimated using mean primary production per unit area of territory, percentage surface area occupied by the 4 most preferred vegetation types, percentage surface area occupied by the 2 least preferred vegetation types, and mean escape distance from a highly preferred vegetation type to the forest. Two of the 12 vegetation types (VT-4 and VT-5) were highly preferred by adult female guanacos; VT-4 exhibited a statistically significant higher primary production than 8 of the other 11 vegetation types .. Feeding territories dominated by the 2 least preferred vegetation types (VT-11 and VT~14) were nearly twice the size of those territories with a low percentage surface area of those types. Guanaco family groups averaged 7 ; however , the size of individual family iv groups fluctuated on a seasonal basis. There was no statistically significant difference in the number of adult females between groups. There also was no significant correlation between group size and total primary production per territory (r= -0.21, d.f.=6), but group size and production per ha of territory were close to the acceptable level of significance (r= 0.61). A negative correlation existed between summer and fall mean number of adult females and percentage surface area of the 2 least preferred vegetation types; no correlation existed between number of adult females and the percentage surface area of the 4 most preferred vegetation types. Territory size was positively correlated with the pe rcentage surface area of the 2 least preferred vegetation types and also with mean distance from VT-4 to the forest edge. No statistically significant differences were found between territories using the territory quality variables for analysis. Hm>~ever, the differences that did exist may have caused females to select one territory over another. Explanations for female selection of territories and the benefits accrued to the female are discussed. Adult females did not appear to select territories on the basis of preferred forage availability; however, the amount of least preferred forage may have acted as a selective factor. 1

INTRODUCTION

The guanaco (Lama guanicoe) is one of two wild members of the family inhabiting . Guanacos range from the highlands of , into parts of , the length of Chile onto the Patagonian plains and south to Tierra del Fuego (Franklin 1975). This paper is based upon results obtained during the second year of a 2-year project on the socioecology of the guanaco. Prior to this study no detailed investigation of guanaco social organization nor territorial behavior had been conducted. A short-term study of guanacos by Franklin (1975) on t he arid west-facing slope of the in northern Peru suggested that this species was territorial, and Raedeke (1978) reported that guanacos defended territories on Tierra del Fuego. Guanacos are organized socially into sedentary or migratory family groups, male groups, and solo males. Adult females are found within family groups, although occasionally adult female groups without an adult male form in the winter. ·Sedentary family groups and solo adult males live i.n permanent year-round territories. Only the adult male participates in defense of the territory. The exact nature of guanaco social organization and territor iality has been described elsewhere (W. Franklin and R. Jefferson, Department of Animal Ecology, Iowa State University, unpublished manuscript). Objectives of this phase of the study were to determine: (1) if territories of sedentary guanaco family groups were sharply separated 2a

and exclusively occupied, (2) if family groups occupied territories of differing resource quality and, if so, what was the effect on group size and degree of occupancy, and (3) if a relationship existed between territory si ze and family group size.

Study Area This investigation was conducted on Isla Grande, the principal and largest island of the archipelago of Tierra del Fuego, at the southern tip of South America (see Humphrey et al., 1971). Roughly triangular in shape, Isla Grande is bounded to the north and west by the Straits of Magellan, to the east by the Atlantic Ocean, and to the south by the Beagle Channel. The island is approximately 29,000 km 2 in area; the

2 eastern port ion (12, 500 km ) is part of the republic of and

2 the west part (16,500 km ) belongs to Chile (Talbot 1974). The study area, locally known as Campo Asseraderro, is on the Chilean side of Isla Grande on the sheep ranch Estancia Cameron (53° 40• S, 69° 55• W). It is located 35 km east of Cameron and 2 km west of the sub- ranch Russfin. Elevation is 488 m above sea level. The study site is a 152 ha open meadow bordered by a beech forest (Nothofagus pumilio) on the north, a road and dwarf beech forest (Nothofagus antarctica) on the south, and a wire fence on the east and west (Fig. 1). This region of the island is an ecotone between open Patagonian pampa and closed · austral forest habitats. The west end of the study area (lines 1 through 12, Fig. 1) is characterized by xeric plant communities and sloping topography, while the central and eastern portions are dominated by mesic to hydric plant communities and flat terrain. The Russfin River, varying 2b Figure 1. Campo Asseraderro study area showing observation hut, grid, and major land features. 3

.J..n lil ... ~ ., ::;) II lo'7 :z: II "II'! ~ z II z~ ,, 0 ojCII ~ j: 0 Ill 17 Ill ~ ~ 1111 Ill: Ill: Ill: "'N 0 Ill ,, ... Ill •I~ ! Ill ,, ~ c ~ ·'"':~ t Ill .,~ Ill Ill: II u Ill z ., ... Ill ~ ,, ... t Ill: "-i z "IN 0 llil 0 N I~ N c II "_, ... I I I II Cllll II II I I I I I I •= I I • '!/ I • .A,I ...I I • .,. I r1 11'1 .I ... I I o/N,,... ,., ... ,,I .,. • I lo 0 II ... •I I •ell I I I I ,.I •Ill I I .... '·'I I I• I •UI I 1 IJ ,,I •Ill • I I" .... I I .I I • •I'! II •N I /·'I I I" .... I I 4 in width from 3 to 5 m. flows from west to east across the study area. Data concerning the climate of Campo Asseraderro are reported elesewhere (Franklin and Jefferson, unpublished manuscript). 5

METHODS

Methodology was based on techniques developed by Franklin (1978). A grid of 1.5 km 2 was established on the meadow of Campo Asseraderro. Painted wooden stakes 1.5 m in height were spaced 100m apart in lines roughly perpendicular to the road (Fig. 1). By using the stakes I was able to estimate locations within the grid to the nearest 10 m. Obser vations of guanaco family groups were made from a 2 story observation hut and from a motor vehicle positioned along the road. Once every hour t he area was scanned from left to right and the following . information about each family group present was recorded on a map of the study area: location; total number of animals, including adult males, adult females, yearlings, and juveniles; numbers and age-class of individuals feeding on defined vegetation types; and the distance the adult male was from his family group. An estimate of a group•s continuous daily movement on the meadow was obtained from these hourly samples. Intergroup encounters and territorial defenses were recorded throughout the hour. Four to six groups were typically under observation at the same time. Guanaco family groups were observed for a total of 953 hours from Ju ly 1977 to June 1978. An average of 17 days per month (7 to 21 days) was spent observing. The length of the observation period varied with the season, depending upon available daylight and animal activity. Family groups were normal ly observed from their emergence out of the forest onto the meadow in the morning, until their return to the forest in the early evening. 6

Animal activity in the forest was undetermined. Seasons were divided into 3-month periods, with December, January, and February representing summer. Twelve vegetation types were defined based on their visual distinctness, dominant plant species, and substrate differences. Plant species composition and cover of vegetation types were measured by the point frame method (Mueller-Dombois and Ellenberg 1974). The point frame was 110 em long with 10 pins spaced 10 em apart. A homogeneous stand of each vegetation type was chosen by its visual uniformity. A line transect was randomly selected through the type and at least 5 stops (50 pin drops) were made per line. Twelve vegetation types were analyzed with 9,480 pin drops during the last 2 months (March and April) of the 1977 growing season.

To estimate primary productivity, 42 wire exclosures were constructed and placed in the different vegetation types prior to the 1977 growing season. At the completion of the growing season, 1 m2 plots within each

2 1.2 m · exclosure··were clipped to ground level, air-dried and weighed. For shrubs, only the current year's growth was clipped. The entire area of the 1.2 m2 plots was not sampled in order to minimize influences due to micro-habitat changes from the sides of the exclosure. Mean production (Kg/ha) of each vegetation type was statistically analyzed using a t-test. Spring and summer production per unit area of territory and total vegetative production per territory were tested using a paired sample t-test.

A detailed vegetation map (scale 1:5000) was made by overviewing the study area with binoculars and spotting scope from the observation hut and by ground reconnaissance. Total surface area for each vegetation type was estimated by planimetric measurements of the vegetation on the 7 vegetation map. Total primary production within territories was estimated by multiplyi ng the total surface area of each vegetation type within the territory by the mean production per unit area for that type. Sites of territorial defenses by family group males were plotted. Maps for ne ighboring males were then compared and territorial boundaries for each season were drawn. Feeding preference indices were calculated from guanaco feeding sam- ples. The percentage of total numbers of guanacos observed feeding on a vegetation type was divided by the percentage of the total hectarage that type represented on the study area and multiplied by 100. The smaller the number the lower the preference; 100 indicated no preference. Preference indices were calculated for adult males, adult females, juveniles, all animals combined, and for individual vegetation types within each territory. The criteria for measuring territorial quality were mean primary production per unit area of territory, percentage of territory surface area occupied by the 4 most preferred vegetation types, and mean escape distance from a highly preferred vegetation type along the river to the forest. The percentage surface area of the 2 least preferred vegetation types in each territory also was used as an indicator of territorial quality, In the statistical analysis of group size, numbers of individuals were equated into •'guanaco units ... Adults and yearlings each equaled 1 guanaco unit, while juveniles 1 to 3 months of age equaled .2 guanaco units, 4 to 7 months equaled .4 units, 7 to 9 months equaled ,6 guanaco units, and 10 to 12 months equaled .8 guanaco units . Total group size, therefore, refers to a group size based on guanaco units . 8

RESULTS

Distribution and Feeding Preferences of Vegetation Types

Information on plant species composition and production of vegetation types is given in Table 1. Points of special interest for certain vegetation types, in descending order of abundance, are presented below. VT-7 covered 35% of the study area but accounted for only 5% of total primary production (the second lowest of all vegetation types). VT-7 was dominated by a short coarse grass-like plant {tarex magellanica) and moss (Rhizogonium mnioides) . Adult female guanacos showed a low feeding preference for VT-7 despite its wide availability (Fig. 2). VT-11 occupied 21% of the area and was dominated by a short, spiny grass (Phleum commutatum) and low growing shrubs (Azorella lycopodioides and Pernettya pumila). VT-11 ranked low in primary production and produced 6% of the total primary production. It was the second least preferred vegetation type by feeding adult female guanacos (Fig. 2) and occurred in 3 of 5 territories. VT-5 had a low production and accounted for only 6% of the total plant production, yet was the second most preferred vegetation type and received a high percent use by adult female guanacos {Fig. 2). VT-5 was found in all territories and was dominated by shrubs (Fig. 1). VT-5 had the highest plant species diversity of all vegetation types (Fig. 1). VT -4 had a significantly higher primary production per unit area than 8 of the other 11 vegetation types (p s 0.01). All vegetation types were statist ically tested, yet, only VT-4 exhibited a statistically 9 Table 1. Characteristics of vegetation types found in Asseraderro study area, Isla Grande, Tierra del Fuego, Chile. Total surface area of vegetation types equals 151.8 ha. List of species includes those that contributed greater than 2 percent relative cover.

VEGETATION TYPES

7 11 13 9 4 RELATIVE SURFACE AREA 34.7% 20.9% 12.0% 9.6% 4.9% PRODUCTION 1946 2153 4212 2682 5525 ( Kg/ha) Number of 1 m2 plots 4 3 4 4 4 ABSOLUTE PERCENT COVER Plants 82.2% 100.0% 98.1% 96.6% 98.2% Water 2.4 0.0 0.0 1.5 0.0 Bare Ground 15.4 0.0 1.9 2.0 1.7 Dead Woody Material 0.0 0.0 0.0 0.0 0.1 Number of point frame pins 2000 1000 640 550 2000 RELATIVE PERCENT PLANT COVER Grasses Alopecurus magellanicus Carex magellanica 24.7 3.8 20.7 Deschampsia parvula 2.6 3.8 3.4. Deyeuxia neglecta 5.3 Festuca gracillima 2.5 3.7 Phleum commutatum 20.3 3.2 3.0 Poa annua Poa pratensis 4.7 29.7 Low Growing Shrubs Azorella lycopodioides Empetrum mnioides 18.9 21.5 Pernettya pumil a 2.5 16.4" 8.1 4.0 10

5 6 8 14 12 3 10

3.4% 3.4% 3.2% 3.1% 1.84% 1.6% 0.9% 2380 2626 4333 3030 3485 3530 455

3 3 3 3 2 3 3

100.0% 98.9% 98.4% 100.0% 98.1% 82.5% 3.61% 0.0 0.0 0.0 0.0 0.0 0.3 0.0 0.0 1.1 0.1 0.0 2.0 16.7 96.4 0.0 0.0 1.4 0.0 0.0 0.5 0.0

700 530 700 200 200 600 360

6.8 6.5

3.0 23.0 2.2 4.0 4.4 15.1 24.4

15.8 10.0 14.5 26.4 23.1 10.0 4.7 6.2 5.3 6.6 4.8 2.4 11 Table 1 (cont.)

VEGETATION TYPES

7 11 13 9 4 Shrubs - treeS: Chiliotrichium diffusum 2.5 Nothofagus antarctica Forbs Acaena ovalifolia 10.0 5.7 6.3 28.1 Blechnum panna-marina Caltha sagittata 5.4 Cotula scariosa 2.~ 2.6 Galium aparine 3.3 Geranium patagonicum 2.1 Mosses Bolex gummifera 10.5 Brachytecium sp. 4.4 21.1 Bryum sp. Polytrichium alpestre Rhizogonium mnioides 2.9 5.4 5.1

Lichens Pseudocyphellaria 10.9 6.2

~1i see 11 aneous

Unknown sp. 9.0 4.7 7.1 8.0 6.8 12a

VEGETATION TYPES

5 6 8 14 12 3 10

27.6 3.6 8.9 32.3

9.8 2.5 6.6 4. 1 2.6 7.3 19.3 6.9

2.4 4.8 24.4 54.2

23.9

4.4 3.8

3.1 50.4 12b V'l Q) u s::: Q) ~ Q) 4- Q) ~ 0. C'l .....s::: "'0 Q) Q) I..L.

..... I..L. 35 il

20

l.J..J (.!) ~ I I Percentaqe Vegetation Type Was of z: Total Study Area (Availability) l.J..J u 0::: 111111 Percentage of Total Feeding l.J..J 0... Obervations (Utilization) ...... w (n•10,959 Feeding Samples) 10

7 11 13 9 4 5 6 8 14 12 3 10 VEGETATION TYPE 45 20 79 76 477 557 108 201 17 65 200 595 PREFERENCE 14 significant difference. VT-4 occupied less than 5% of the surface area, but received the highest feeding use (23%) of any type. It had the third highest preference index, despite not being utilized in winter because of deep snow (VT -4 was the most preferred vegetation type in spring and second most in summer). VT-4 was a verdant grass (Poa pratensis) and forb (Acaena ovalifolia) belt along the river that varied in width from 25 to 100m and was found in all territories. VT -8 occurred at the forest edqe in all , territories and extended 25 to 50 m under the forest canopy. It was the second highest producing vegetation type and accounted for 12% of total plant production . Dominated by forbs and mosses, VT-8 was second highest in plant species diversity. VT -8 was the fourth most preferred vegetation type, receiving most of its feeding use in the morning when animals came out of the forest and in the evening when they returned. VT-3 occupied only 1.6% of the study area, but had high feeding use during the winter (46%) with little use during the other seasons (3.2%, 1.3%, and 2%}. It was dominated by low growing dwarf beech trees and occurred in all territories . VT - 14 was the least preferred vegetation type (Fig. 2). It had an intermedi ate vegetative production and was dominated by a low growing shrub (Empetrum mnioides), a forb (Galium aparina), and a sedge (Marsippospermum grandiflorum). Occurring in 2 of the territories, VT-14 was restricted to the high and dry slopes of the meadow (lines 1 through 12, Fig . 1). VT - 10 covered the smallest surface area of all vegetation types and had a significantly lower production than 8 of the other 11 vegetation 15a

types (p~ 0.01); however, it had the highest feeding preference index (595). VT-10 received only 5% feeding use by guanacos and accounted for only a small percent of total plant production (Table 1). VT-10 occurred in all territories and was dominated by bare ground (96%) with Kentucky blue grass and a low growing shrub (Azorella lycopoides) occasionally occurring. Despite the high preference index, VT-10 was not an important vegetation type due to its limited availability.

Feeding Territories

Neighboring family groups overlapped very little in their use of space (Fig. 3); territories were sharply separated. If a non-family group member entered a territory, the resident male would chase the intruder away. Details of guanaco territorial behavior have been described elsewhere (Franklin and Jefferson, unpublished manuscript). Adjacent family groups normally did not enter a neighbor•s territory when that neighboring group was absent. If such a violation did occur, it was only for a brief time. All guanaco family groups retreated to the forest each evening. Territorial defense in the forest by the family group males was observed, but how much additional area was being defended is not known. The forest probably served as a safe site for sleeping and for protection during inclement weather. It was assumed that because guanacos returned each day to the open plains to feed, the forest was not important for feeding. Guanaco feeding territories varied in size from 2.0 to 45.8 ha 15b Vl Q)

.,.... -o Q) Q) 4- 0 u ttl t:: ttl ~ O'l 4- 0

Q) 0. ttl ..t:: Vl -o t:: ttl

Q) N

Vl t::

Q) O'l t:: ttl ..t:: u .-- ttl t:: 0 Vl ttl Q) (./')

C"")

Q) ~ ~ .,....O'l l.L.. TERRITORY BOUNDARlES WINTER --- SPRING •••

SUMMER - FA.LL ---

,_. 0"1

- ::_-_-_-_-_-: .... _ ----- .... -- - - ~ 1 2 3 4 s 6 1 a • 10-- 1C12-1J:t4-1S...J6---- 17 18 1st- 2.J-21::22-~Q4-25-26-27-2B-2•-3o-3T~z------~ ------. . -- FENCE--¢- OBSERVATION HUT~

RIVER ROAD ::: ._..... 100m 17

and averaged 29.5 (n=8 territories, s .d.=11 . 1). The maximum number of territories in the study area was 6 in the fall of 1978, 5 during all other seasons. Eight total family groups were observed over the year because of the replacement of some territorial males. Territories located at the west end of the study area (Fig. 2) on the xeric substrate were much larger than those on the flat, hydric east end (x=42.2 ha, s.d.=2.3; vs x=21.7 ha, s .d.=3.8) . Territory size changed little for all groups

between seasons (Table 2) with the exception of 208~ 208's territory changed in size due to a territorial changeover with the 210 male. The largest ter ritory (46 ha) was occupied by a solo adult male (107); it contained 86% of VT -14 and 80% of VT-11 found in the study area, the 2 least preferred vegetation types. The shapes of feeding territories were approximately rectangular (Fig. 3). If not disturbed, guanacos would come to the river each day to feed on VT-4. The high preference for VT-4 was apparently the factor dictating contact with the river, as guanacos were not observed drinking from the river. VT-8 probably also exerted a similar influence at the forest edge. Defending a strip with access to the forest for an escape route also appeared to be important. Spring and summer mean vegetative primary production per unit area

of territory did not differ significantly among territories (t=0.96, p ~ 0. 50, d.f.=1). There also was no significant difference in total spring

and summer primary production among territories (t=0.35, p ~ 0.90). Guanaco family groups averaged 7 animals (925 observations, s.d.=2.8); however, the size of individual family groups fluctuated on a seasonal basis. The number of adult males in any particular family group never exceeded Table 2. Seasonal chanqes in size (ha) and standard error of ouanaco family group territories. NP indicates family group not present for that season.

Family group Winter Spring Surruner Fall x

107 NP 45.8 42.2 NP 44.0 ± 1.8 111 NP NP NP 44.0 201 23.8 25.4 23.0 22.7 23.7 ± 0.6 202 27.0 NP NP NP 203 20.1 20.6 19.4 23.4 20.9 ± 0.9 204 39.9 38.3 40.5 40.6 39.8 ± 0.5 ...... co 208 NP 22.5 26.5 2.1 17.0 ± 7.6 210 NP NP NP 18.8

Total 27.7 ± 4.3 30.5 ± 4.9 30.3 t 4.6 25.3 ± 6.3 29.0 ± 2.5 19

1; family group males were present nearly every observation period. Numbers of adult females, yearlings, and juveniles fluctuated the greatest, with numbers being most constant in summer and most variable in winter (Table 3). Grouo 208 and 111 had a wide variation in numbers of individuals. In the spring, male 208 was establishing a territory, and in the fall he lost 90% of his territory to an invading male. Similarly, male 111 was establishing a territory during the fall of 1978. During the fall, some adult females and their juveniles voluntarily begin leaving family groups . In the winter, family group size is at a minimum (Table 3). l·!ith the arrival of spring, family group size again begins to expand as females and near-yearlings join the adult territorial males . Summer group size is the yearly maximum because offspring are born and more adult females join the groups. Data on mean number of adult females and total group size were analyzed only for those family groups present all 4 seasons of the study (201, 203, 204). The 3 family groups exhibited a significant difference in the mean number of adult females between seasons (f=55 .6, 46.3, 31.1, p ~ 0.0001, d.f.=3,3) . However, there was no significant difference in the mean number of adult females (f=2.61, 0.10 ~ p ~ 0.25, d.f.=3,2) (Table 4}, or total group size (f=2.08, 0.25 < p < 0.50) among groups among seasons (Table 3).

Group size was influenced by yearling dispersal. Yearlings were expelled by the family group adult males during the summer. By the end of February all but 1 yearling were gone from their family groups. Data on yearling expulsion and dispersal are reported in Franklin and Jefferson {unpublished manuscript) , Table 3. Mean group size and standard error of guanaco family groups for winter, spring, summer, and fall. NP indicates fami ly group not present for that season.

Family group Winter Spring Summer Fall x

107 NP 2.3 ± 0.4 (40)A 1. 0 ± 0 . 0 ( 42) NP 1.6 ± 0.6

111 NP NP NP 2.0 ± 0.3 (47)

201 1.7 ± 0.5 (11 ) 3.9 ± 0.4 (45) 9.5 ± 0.4 (42 ) 9.2 ± 0.3 (51) 6.1 ± 2.0

202 1.8 ± 0.5 ( 9) NP NP NP

203 2.9 ± 0.2 (21) 9.2 ± 0.4 (45) 12.8 ± 0.8 (39) 9.5 ± 0.5 (50) 8.3 ± 2.1

204 6. 3 ± 0.9 (15) 11.9 ± 0.7 (47) 11.7 ± 0.2 (41) 7.6 ± 0.4 (51) 9.4 ± 1.0 N 0 208 NP 15.0 ± 1.2 (26) 7.5 ± 0.4 (42) 2.4 ± 0.2 (50) 8.3 ± 3.7

210 NP NP 2.5 ± 0.7 (19 ) 6.5 ± 0.3 (52)

a · Number of observations. Table 4. Mean and standard error of the number of adult female guanacos in family groups. NP indicates family group was not present for that season.

Family group Winter Spring Summer Fall x

107 NP 0.9 ± 0.3 (40)a 0.0 (42) NP

111 NP NP NP 0.7 ± 0.2 (47)

201 0.6 ± 0.4 (11) 2.0 ± 0.3 (45) 7.8 ± 0.2 (42) 6.4 ± 0.3 (51) 4.2 ± 1.7

202 0.6 ± 0.3 (9) NP NP NP

203 1.4 ± 0.1 (21) 5.1 ± 0.3 (45) 9.5 ± 0.6 (39) 6.9 ± 0.2 (50) 5.7 ± 1.7

204 ± ± ± ± ± N 4.3 0.7 (15) 7.1 0.5 (47) 6.9 0.2 (41) 4.9 0.3 (51) 5.8 0.7 ...... 208 NP 10.5 ± 0.9 (26) 4.2 ± 0.4 (42) 0.4 ± 0.2 (50) 4.2 ± 3.1

210 NP NP 1.2 ± 0.4 (19) 4.2 ± 0.2 (52) 2.7 ± 1.5

a NUmber of observattons. 22

Mean group size was positively correlated (r= 0.84, 0.05 S p S 0.10, d.f.=3) with winter territory size but there was no significant relationship for the other 3 seasons (Spring r= -0.40, Summer r= -0.55, Fall r= a 0.65 0.20 s p s 0.50). There was no significant correlation between group size and total primary production per territory (r= -0.21, 0.50 S p S 0.60, d.f.=6), but group size and production per ha of territory were close to the acceptable level of acceptance (r= 0.61, 0.10 S p ~ 0.20). No correlation was found between group size and the percentage surface area of the 4 most preferred vegetation types (r= 0.26, 0.50 ~ p ~ 0.60,); nor of group size and the percentage surface area of the 2 least preferred vegetation types (r= -0.36, 0.20 ~ p ~ 0.50). A negative correlation existed between summer and fall mean number of adult females and percentage surface area of the 2 least preferred vegetation types, VT-11 and VT-14 (r= -0.82 and r= -0.91, 0.05 ~ p ~ 0.10, d.f.=3), but not for winter or spring (r= -0.45 and r= -0.53, 0.20 ~ p ~ 0.50). There was no significant relationship between mean number of adult females and percentage surface area of the 4 most preferred vegetation types for any season (Spring r= 0.45, Summer r= 0.62, Fall r= 0.52, Winter r= 0.41; 0.20 ~ p ~ 0.50). Group size was negatively correlated with mean distance from VT-4 at the river•s edge to the forest in each territory

(r= -0.80, 0.05 S p ~ 0.10). Territory size was postively correlated with percentage surface area of the 2 least preferred vegetation types (Spring r= 0.89, Summer r= 0.95, Fall r= 0.91, Winter r= 0.91; 0.02 S p s 0.05) for all seasons and with the percentage surface area of VT-4 (Spring r= -0.93, Summer r= -0.90, 23

Fall r= -0.89, 0.02 ~ p ~ 0.05) except for winter (r= -0.85, 0.05 s p s 0.10). Territory size was also positively correlated with mean distance from VT-4 to the forest edge (r= 0.96, 0.005 ~ p ~ 0.01). A significant negative correlation exis·ted between territory size and vegetative production per ha of territory for summer and fa 11 ( r= -0.83 and - 0.87,

0.05 < p ~ 0.10) and nearly so for winter and spring (r= -0.79 and -0.82,

0.10 < p ~ 0.20). Total primary production was positively correlated with territory size for all s·easons (Spring r= 0.93, Summer r= 0.98~ Fall r= 0.94,

Winter r= 0.96; 0,005 ~ p ~ 0.01). 24

DISCUSSION

When comparing the percentage surface area of the least preferred vegetation types, mean primary production per unit area of territory, and territory size, no statistically significant differences were found among territories. Yet, differences did exist between territories that may have caused females to select one territory over another. For example, the largest territory in the study area contained the greatest percentage surface area of the least preferred vegetation types, had the lowest primary production per unit area, and contained the fewest number of adult females. Larger territories also could be more often infringed upon by outside intruders that fed within the territory before being detected by the territorial male. Such feeding intrusions could decrease the available forage for the resident females. Gosling (1974 ) found that the territory size for the Coke•s (Alcelaphus buselaphus cokei) was correlated with the dominant vegetation type in which the territory was located; the larger territories were in the least preferred vegetation types. Avian investigators also have found territory size to vary in relation to the quality and

ava i 1abi 1i ty of resources (Gi 11 and Wo 1f 1975; ~~atson and Moss 1971) . The lack of a relationship between guanaco territory size and group size is perplexing. If territory size increased as territory quality decreased, then one would expect group size to decrease; but no such statistically significant relationship was found. Franklin (1978) found that the size of vicuna ( vicuana) family groups in the central 25

Andes was correlated with the size of feeding territories and the production of forage within the territories. However, these relationships did not exist for guanaco family groups on Tierra del Fuego. The influence that available browse within the forest may have had on group size within the territory was not measured for reasons already mentioned. It is possible that if the adjoining forest is defended as part of the feeding territory, a relationship between group size and territory size might be more apparent to an observer.

In territorial species the quality of the territory and the capability of the male defending the territory may influence female mate choice (Orians 1969; Verner 1964)'. Many investigators have proposed that in most species females choose their mates, while males are less discriminating. This hypothesis is supported by the greater initial investment of the female in her offspring (Trivers 1972; Williams 1966). Avian researchers have been most active in attempting to correlate mating success and territory quality, using food availability and type, and territory size as measures of quality. Mayfield (1960) reported that female Kirtland's warblers (Dendroica kirtlandii) showed a greater tendency to return to the same territory to breed than to return to the same male, even if the male was in close physical proximity. Verner (1964) working with marsh wrens (Telmatodytes palustris), Willson (1966) with yellow-headed blackbirds (Xanthocephalus xanothocephalus), and \~ittenberger (1976) with bobolinks (Dolichanyx oryzivorus) also have shown correlations between territory quali~y and mating success of the owner. Searcy (1979) argued that mate choice in female red-winged blackbirds (Agelaius phoenicus) is 26

influenced by availability of food in the territory. Bromley (1969), David (1973), Franklin (1978), Leuthold (1966) and Spinage (1974) have used the amount of time family groups or territorial males spend away from their territories as an index of territory quality. As the quality of the territory increases, the time the individual or group spends away from the territory should decrease. It was not possible to examine this relationship during my study because guanaco family groups not present in their feeding territories were out of view in the adjoining forest . If a family group was not present on a particular day, it did not necessarily mean that the group was away from its territory. It is possible that the quality of habitat on the territories of guanaco males is such that the expected reproductive success of a newly arriving female guanaco is higher if she attempts to mate with a mated male but on a superior quality territory, rather than mating with a male with fewer females on poorer habitat. This assumes that female guanacos have the ability to assess some index of food quality and/or availability. If adult females were selecting one territory over another, then selection should be most evident during late spring and early summer when females are searching for a safe and secure site to give birth and raise their offspring. Availability of high quality forage would be important in order to provide sufficient resources for the female and her offspring. Group 203's territory had the highest primary production per unit area and also had the largest summer mean number of adult females. Guanaco adult females did not appear to select territories on the basis of availability of preferred vegetation types, since those territories with greatest surface area of preferred vegetation types did not have a 27 significantly larger number of adult females. However, the amount of least preferred forage may have acted as a selective factor because the territory with the largest percentage surface area of least preferred vegetation types had a significantly lower number of adult females than any other territory. The family group male has an important effect on the number of adult females within the group . Females may use behavioral cues of the adult male as an indicator of territory quality, since the male either accepted or rejected females attempting to enter his group. For example, male 208 invaded 202's territory in the spr ing of 1977 and successfully expelled the 202 male. Within a week the number of adult females in the old 202 territory increased from 5 to 15. This was not in response to a territory quality change, but rather in response to a change in the quality of the territorial male. The territory quality may have been sufficient to support more females than the 202 male was capable of defending, but the 208 male apparently had the ability to do so. Seven weeks after the territorial changeover (mid-summer) the number of females in 208's territory stabilized at 7, far below the spri ng average, but still more than what 202 possessed. It is interesting to compare the 202 -208 sequence of events with another territorial changeover by the 210 male. The 210 male arrived in summer and took over 90% of 208•s territory. However, far fewer females joined the 210 male because most already were established in family groups, contrary to the spring season when many females were on the move and not yet ~stablished with a family group .

Guanaco and vicuna social organization and territorial behavior 28 have much in common, but they sharply differ in stability of the family group. The mean number of adult females per guanaco family group fluctuated greatly on a seasonal and occasionally on a daily basis. The number of adult females in a vicuna family group is relatively constant throughout the year (Franklin 1978). The reason for the instability of the number of adult females in a guanaco family group may be related to food availability. Deep winter snows forced guanacos to turn to browse, as indicated by the shrub community VT-3 being the most preferred winter forage. With the lack of sufficient forage, adult females left the family group in search of adequate food, or in a few cases, the adult male expelled some adult females. Family group size increased in the spring as the same or new females joined the group. The number of adult females then increased until a maximum was reached during the summer. 29

LITERATURE CITED

Bromley, P. 1969. Territoriality in prong~1orn bucks on the National Range, Moiese, Montana. J. Mammal. 50: 81-89. Brown, J. 1964. The evolution of diversity in avian terrestrial systems. Wilson Bull. 76: 160-169. Brown, J. and G, Orians. 1969 . Spacing patterns in mobile animals. Ann. Rev. Ecol. Syst. 1: 239-262. David, J. 1973. The behavior of the bontebok with special reference to territorial behavior. Z. Tierpsychol. 33: 38-107.

Franklin, W. 1974. The social behavior of the vicuna. Pages 417-487 ~ V. Geist and F. Walthers, eds. The behavior of ungulates and its relation to management. I.U.C.N . , Morges. 940 pp.

Franklin, l~. 1975 . Guanacos in Peru. 13: 191-202. Franklin, W. 1978. The socioecology of the vicuna. PhD. dissertation. Utah State University, Logan. 172 pp. Gill, F. and L. Wolf. 1975. Economics of feeding territoriality in the golden-winged sunbird. Ecol. 56: 333-345. Gosling, L. 1974. The social behavior of the coke•s hartebeest. Pages 488-511 in V. Geist and F. Walthers, eds. The behavior of ungulates and its relation to management. I.U.C.N., Morges. 940 pp. Horn, H. 1968. The adaptive significance of colonial nesting in the Brewer's blackbird. Ecology 49: 682-694. Humphrey, P., P. Reynolds, and R. Peterson. 1970. Birds of Isla Grande (Tierra del Fuego). Kansas Museum of Natural History, Lawrence. 411 pp. Leuthold, W. 1966. Variations in territorial behavior of Uganda kob. Behaviour 27: 215-258. Mayfield, H. 1960. The Kirtland's warbler. Bull. Cranbrook Inst. Sci. 40: 1-242. Mueller-Dombois, D. and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York. 547 pp. Orians, G. 1969. On the evolution of mating systems in birds and . Am. Nat. 103: 589-603. 30

Raedeke, K. 1978. El guanaco de Magallanes, Chile: Su_distribu~ion Y biologia. Corparacion Nacional Forestal. Publ1cacc1on Techn1ca no. 4. 181 pp.

Searcy, W. 1979. Female choice of mates: A general model for birds and its application to red-winged blackbirds (Agelaius phoenicus). Am. Nat. 114: 77-100. Spinage, C. 1974. Territoriality and population regulation in the Uganda defasa waterbuck. Pages 635-643 in V. Geist and F. Walther, eds. The behavior of ungulates and its-relation to management. I.U.C.N., Morges. 940 pp. Talbot, R. 1974 . A history of the Chilean boundaries. Iowa State lln ·i vers i tv Press . Ames . 134 pp.

Trivers, R. 1972. Parental investment and sexual selection. Pages 136- 179 in B. Campbell, ed. Sexual selection and the descent of man, 1871~971. Academic Press, New York. 511 pp.

Verner, J. 1964 . Ev olution of polygamy in the long-billed marsh wren. Evolution 18: 252-261. Watson, A. and R. Moss. 1971. Spacing as affected by territorial behavior, habitat and nutrition in red grouse (Lagopus l· scoticus). Pages 92-111 in A. Esser, eds. Behavior and environment: Use of space by animals-and man. Plenum Press, New York. 411 pp. Wilson, E. 1975. Sociobiology, the new synthesis. Harvard University Press, Cambridge. 677 pp.

Williams, G. 1966. Adaptation and natural selection: a critique of some current evolutionary thought. Princeton Univ. Press, Princeton. 307 pp.

Will son, ~1. 1966. Breeding ecology of the yellow-headed blackbird. Ecol. Mongr. 36: 51-77. Wittenberger. J. 1976. Habitat selection and the evolution of polygyny in bobolinks (Dolichonyx oryzivorous). Ph.D. dissertation. Univ. Calif., Davis. 31

AKNOWLEDGH1ENTS

The study was financed through grants from the Institute of International Education, The National Geographic Society, and Iowa State University . Funding during thesis preparation was generously provided by the Graduate College of Iowa State University and the Department of Animal Ecology at Iowa State University. Th e Agrarian Land Reform provided housing and facilities at Estancia Cameron and allowed me to use Campo Asseraderro as my study area. Appreciat ion is expressed to Eduardo and Luis Barria and Angel and Nora Pawaulla of seccion Russfin, who contr ibuted their friendship and assistance during my stay in Chile. I,M. Ortega assisted in the field during the summer months. Appreciation also goes toM. Rosenfeld and C. Cunazza of the Corparacion Nacional Forestal and to C. Venegas and W. Sielfeld of the Institute of for their consultation on various phases of the study . Dr . E. Pisano painstakingly identified the plants of the study area. Special thanks are due to C, Estrada of CONAF and E. Tafra of SAG for their logistical assistance, for their aid with government and administrative paperwork, and for their comradeship. T. Haindfield helped with data reduction and D. Zimmerman helped immensely with computer programm i ng. The College of Agriculture of Iowa State University and W. Franklin generously provided funds for computer time .

I would like to thank fellow grad students Pat Brown, Bob Frederick, . 32

Mary Henry, Ellen Johnson, Pat McCraw, Mark Ryan, Joe Schaefer, and Adrian and Paula Wydeven for their friendship, hours of productive discussion, and for their assistance and consultation. I express my appreciation to my committee members at Iowa State University, Dr. Marilyn Bachmann, Dr. Louis Best, Dr. William Franklin (chairman), and Dr . Kenneth Shaw, who generously provided their counsel and time. Special gratitude and thanks are expressed to my advisor, Dr. W. Franklin, for his friendship, guidance, and unselfish support during my study and for introducing me to the exciting field of sociobiology. I wou l d also l ike to thank my family for their correspondence and moral support. Deepest appreciation and thankfulness goes to my wife Christine, for her profound dedication to the study and her companionship which provided immeasureable support during our stay in Chile. Last but not least, warmest thanks and sincere appreciation go to Jennifer Riggs for helping to ease the burden of graduate school, making life more enjoyable, and for her assistance with typing and editing.