small mammals of maulino forest remnants, a vanishing ecosystem of south-central chile

337MAMMALIA • 2005 • 69 (3-4) © Publications Scientifiques du Muséum national d’Histoire naturelle, Paris.

Small mammals of Maulino forest remnants,a vanishing ecosystem of south-central Chile

Bárbara SAAVEDRAJavier A. SIMONETTI

Depto. Ciencias Ecológicas, Facultad de Ciencias, Universidad de Chile,Casilla 653, Santiago (Chili)

[email protected]

Saavedra B. & Simonetti J. A. 2005. — Small mammals of Maulino forest remnants, avanishing ecosystem of south-central Chile. Mammalia 69 (3-4) : 337-348.

ABSTRACTThe Maulino forest, located at the coastal range of south central Chile, hasbeen severely disrupted by intense human activities. Currently, landscape isdominated by large extensions of plantations of Monterrey pine, where rem-nants of native forest are immersed. Here, we assess consequences of Maulinoforest fragmentation and habitat replacement upon small mammal fauna. Wedescribe habitat characteristics, small mammal’s composition and abundancein native forests (fragmented and continuous), but also in Pinus plantation.Population and body condition were compared among habitats, along withmovement among native and Pinus forests. Higher species richness was foundin continuous forest comprising Abrothrix longipilis, A. olivaceus, Oligoryzomyslongicaudatus, Irenomys tarsalis, Geoxus valdivianus, Rattus rattus, Octodonbridgesi, Thylamys elegans, and Dromiciops gliroides. Higher abundance wasobserved in fragmented forest and Pinus plantations, where A. longipilis,A. olivaceus, O. longicaudatus, and T. elegans accounted for 80% of total cap-tures. Population structure and body size were not affected negatively by frag-mentation and habitat replacement, although habitat characteristics differedamong surveyed habitats. Fragmentation of Maulino forests seems to have asignificant impact on overall small mammal’s richness, threatening the per-sistence of rare taxa of small mammals. Conservation and management prac-tices are discussed to favor protection of this highly endangered habitat.

KEY WORDSForest fragmentation,

Microbiotheriidae,Rodents,

Didelphidae,conservation,

Chile,South America.

Brought to you by | Universitaetsbibliothek Frankfurt/MainAuthenticated | 10.248.254.158Download Date | 9/8/14 1:32 PM

Saavedra B. & Simonetti J. A.

338 MAMMALIA • 2005 • 69 (3-4)

INTRODUCTION

The Maulino forest is a unique deciduous forestlocated in the coastal range of south-central Chile(Fig. 1, 35°55’S-36°20’S). This forest harbors asuite of endemic and relict taxa, such as popula-tions of the tree Gomortega keule, the only livingspecies of the monotypic family Gomortegaceae(San Martín and Donoso 1995). Among smallmammals, the Maulino forest embraces popula-tions of Dromiciops gliroides, the single represen-tative of the order Microbiotheria; Irenomystarsalis and Geoxus valdivianus, the single speciesof sigmodontine Rodent genera endemic tothe temperate forests of southern Chile andArgentina (Redford & Eisenberg 1992),and Octodon bridgesi, the single Caviomorpha

species found in southern forests. These speciesare typical components of a more southernValdivian fauna (Osgood 1943), revealing thatthey attain their northernmost populations atthe more mesic Maulino forest (Saavedra &Simonetti 2000, 2001).Forests at the coastal range have been severelydisrupted by intense human activities, includingforest clearing for agricultural purposes, loggingand currently replacement of natural vegetationby commercial plantations of exotic tree species.These activities have resulted in a massive reduc-tion and fragmentation of the Maulino forest(Lara et al. 1995). Areas formerly covered byMaulino forests are currently dominated by largeextensions of plantations of Monterrey pine(Pinus radiata). Remnants of the native forest are

MOTS CLÉSFragmentation forestière,

Microbiotheriidae,Rongeurs,

Didelphidae,conservation,

Chili,Amérique du Sud.

RÉSUMÉPetits mammifères des résidus de forêt d’un écosystème en voie de disparition duChili sud-central.La forêt de Maulino, située dans la bordure côtière de la région sud-centraledu Chili, a été fortement fragmentée par une intense activité humaine.Aujourd’hui, le paysage est dominé par des plantations intensives de Pin deMonterrey, dans lesquelles subsistent des îlots de forêt native. Dans ce travail,nous évaluons les conséquences sur la faune de petits mammifères de la frag-mentation de la forêt de Maulino et le remplacement de cet habitat. Nousdécrivons les caractéristiques de l’habitat, la composition et l’abondance depetits mammifères dans les forêts originelles (fragmentées et continues) etdans les plantations de pin. La condition corporelle des individus et des popu-lations ont été comparées entre habitats, ainsi que les mouvements entre laforêt native et celle de Pinus. Nous avons observé une richesse spécifique plusélevée dans la forêt continue, en incluant Abrothrix longipilis, A. olivaceus,Oligoryzomys longicaudatus, Irenomys tarsalis, Geoxus valdivianus, Rattus rattus,Octodon bridgesi, Thylamys elegans et Dromiciops gliroides. Une plus grandeabondance des rongeurs est observée dans la forêt fragmentée et plantée dePinus, où 80 % des captures totales concernent A. longipilis, A. olivaceus,O. longicaudatus et T. elegans. Malgré la différence des habitats observés, lastructure de la population et la taille corporelle ne semblent pas affectées parla fragmentation et le remplacement de l’habitat. Cependant, la fragmenta-tion de la forêt de Maulino semble avoir un impact global important sur larichesse en petits mammifères, en menaçant les taxons les plus rares. Lesméthodes de conservation et de gestion sont discutées afin de permettre laprotection de cet habitat particulièrement menacé.


immersed in this exotic matrix, scattered in ahighly fragmented pattern (San Martín &Donoso 1995; Bustamante & Castor 1998).Similar to the Tumut Fragmentation Experimentin Australia (Lindenmayer et al. 1999), the land-scape in south-central Chile contains smallislands of native vegetation in a sea of Monterreypines. However, the consequences of forest frag-mentation and replacement for the biota havebeen scarcely addressed (e.g., Estades & Temple1999).Although small mammals are able to thrive inexotic plantations, few taxa use this habitat com-pared to native shrublands (Muñoz & Murúa1990; Muñoz-Pedreros et al. 1990). In fact,species richness is higher in native forest com-pared to Pinus plantations, despite the scarceexistence of native remnants (Muñoz & Murúa1990). Higher habitat complexity, determinedprimarily by the availability of refugia due to thepresence of a well developed understory in nativeforest compared to exotic plantations, is pre-sumed to explain this pattern (Muñoz &Murúa1990). The effect of forest fragmentationper se upon the diversity and abundance ofChilean small mammals has been assessed onlyfor Valdivian forest, where no association amongspecies richness and abundance to fragment sizeis found (Kelt 2000). Frequent disturbance, and

limited habitat diversity were advanced to explainthis lack of relationship (Kelt 2000).Unfortunately, small mammal’s presence in thesurrounding matrix, or its capacity to moveamong fragments was not assessed in Valdivianremnants. Moreover, generality of these findingsremains to be assessed.Here, we estimate the consequences of forestfragmentation upon small mammal fauna ofMaulino forest. First, we describe microhabitatcharacteristics of continuous and fragmentednative forest, and also of Pinus plantations. Then,we assess whether forest fragmentation impingesupon the richness of small mammals. Based onclassical biogeography theory, we expect to findricher assemblages in larger and continuousforests, and depauperate ones in small forest frag-ments, as well as in the pine matrix surroundingthem. With this information, we describe smallmammal’s assemblages, as well as habitat avail-ability in native forest (continuous and frag-mented) and in Pinus matrix. At the populationlevel, we assess potential changes in populationsize, and reproductive status comparing smallmamma ls ’ popu lations among continuousforests, fragments and Pinus plantation. If frag-ments and matrix are low quality or stressfulhabitats, populations ought to be comprised byindividuals of diminished biological conditions,

Small mammals of Maulino Forest, Chile

339MAMMALIA • 2005 • 69 (3-4)

FIG. 1. — Location (★) of fragments of Maulino forest surveyed for small mammals, VII Región, Chile. Shaded area corresponds tonative forest: LQNR, National Reserve Los Queules; F1, Fragment 1 (asterisk indicates location of Pine plantation surveyed); F3,Fragment 3; white areas correspond to Pinus radiata matrix (M).


less abundant and exhibiting lower reproductivesuccess (see Díaz et al. 1999). Furthermore, weexplored whether forest fragmentation is in factperceived as an habitat disruption by small mam-mals, testing whether they move from remnantsof native vegetation into the surrounding matrixof pines. If this “sea of pines” is an inhospitablehabitat, small mammals ought to be restricted topatches of native vegetation, suffering fromencroachment at the fragments compared tothose surviving in continuous forest.

MATERIAL AND METHODS

STUDY SITE

We performed our faunal survey in Los QueulesNational Reserve (35°59’S-72°41’W, 540-570 ma.s.l.), and surrounding forest fragments in theCoastal range of south central Chile (Fig. 1). Alarge tract of forest, the last 600 ha piece ofcontinuous Maulino forest shelters Los QueulesNational Reserve (LQNR hereafter), a 145 haconservation unit. LQNR is surrounded by Pinusradiata plantations, where several small remnantsof native forest are immersed (Fig. 1). Thecontinuous forest, along with forest remnants, aretypical Maulino native forest, with Nothofagusglauca (Fagacea), Aextoxicum punctatum(Aextoxicaceae), Cryptocarya alba (Lauraceae),Gevuina avellana (Proteaceae) and Persea lingue(Lauraceae) as dominant tree species (San Martín

& Donoso 1995). Diversity and abundance ofthe mammalian fauna was assessed at LQNR,and two nearby fragments: one of 3 ha (Fragment3, 35°59’S-72°41’W, 570 m a.s.l.), and anotherof 6 ha fragment (Fragment 1, 35°58’S-72°42’W,348 m a.s.l.). Additionally, we surveyed smallmammals in one neighboring P. radiata planta-tion to Fragment 3 (Fig. 1). Current fragmenta-tion of the Maulino forest was mainly caused byagricultural and forestry activities, that began 20-40 years ago, indicating that forest fragments areat least 20 years old.

SAMPLING

We carried out bimonthly trapping of smallmammals from January 1999 to February 2001(Table 1). Each captured specimen was individu-ally marked with an ear tag, measured andreleased. In Fragment 1, we operated one grid thatwas used to evaluate individual movementbetween forest remnants and neighboring pineplantation. Therefore, half of trapping grid wasset inside the forest fragment, while the other halflaid in adjacent Pinus plantation (Table 1). In thisgrid we developed a χ2 test to compare observedcaptures as a function of the expected captures ifsmall mammals did not show preferences amongnative and exotic habitat. In each grid, Shermann-type traps were located 10 m apart. Live trappingsessions lasted for at least five trapping nights. Alltraps were baited with rolled oats and checkeddaily at dawn, since all mammals are nocturnal.


340 MAMMALIA • 2005 • 69 (3-4)

TABLE 1. — Description of small mammal’s trapping schedule in Maulino forest, south-central Chile. Abbreviations: LQNR, LosQueules National Reserve; J, January; F, February; A, April.

Variable LQNR Fragment 3 Pinus matrixFragment 1

native forest Pinus matrix

Traps/grid 50 30 50 25 25Number of grids 2 1 1 1 1Grid size (m2) 3,600 2,000 3,600 1,800 1,800Period of survey J1999-F2001 A2000-F2001 A2000-F2001 J1999-F2001 J1999-F2001Number of nights 41 37 37 41 41Total trap-night 3,909 1,061 1,825 995 997Trapping effort (trap/night) 95,3 28,7 49,3 24,3 24,3


FAUNAL SURVEY

Richness were estimated as the total number oftaxa captured. Small mammal’s abundance wasestimated by trapping success and by direct enu-meration (Hilborn et al. 1976) trough the assess-ment of the minimum number of animals knownto be alive (MNA) (Simonetti 1986). Trappingsuccess was compared among habitats using amultiple proportion test for unequal sample size(Zar 1984). Body condition was assessed only fornon-reproductive adults of Abrothrix longipilisand A. olivaceus captured during December1999-February 2000, comparing the ratio bodymass/body size among individuals captured innative forest (continuous forest and forest frag-ments), and Pinus matrix. Reproductive statuswas assessed for females registering if vagina wasopen, or if there were signs of lactation such asenlarged nipples. For males, we recorded if testi-cles were scrotal or abdominal. Based on bodysize and weight, we distinguished among juvenileand adults. We compared the reproductive pro-portion and age structure among continuous for-est, forest remnants and exotic matrix.

HABITAT ANALYSIS

Habitat description was based on 28 habitat met-rics that were recorded at every trap station inOctober 1999. These measures followed Glanz(1970), and included variables aimed to describehabitat characteristics at ground, shrub and treelevel. At each trap station, four line transects weretaken, each 2 m long, by pushing a 2 m boardthrough the vegetation in 90° direction, from adistance of 1 m away from the trap placement(Glanz 1970). Percentage of cover at ground,shrub and tree level was recorded. At ground level(up to 50 cm) we estimated percentage of cover forlitter (GRLITTCOV), branch (GRBRACOV),stem (GRSTECOV), shrub (GRSHRUCOV),herb (GRHERCOV), pine litter (SHRPLICOV),and percentage of bare soil cover (GRSOILCOV).At shrub level (from 50 cm to 3 m) we measuredpercentage of cover for empty space (SHREMP-COV), juvenile trees (SHRJUVCOV), branch(SHRBRACOV), shrub (SHRSHRCOV) andherb (SHRHERCOV) cover. At tree (more than

3 m) level we estimated percentage of cover forbranch (TRBRACOV), native trees (TRNAT-COV), pine (TRPINCOV), and percentage with-out overstory cover (TRWOSCOV). The densityof the shrub foliage in three different levels0-10 cm (FODEN010), 11-25 cm(FODEN1125), and at 1-2 m (FOLDEN12) wasdetermined by placing the same board vertically ata distance of one meter from the observer, andestimating the percentage of the board obscuredby vegetation at each level. We also estimated themaximum herb (HERBHIGH) and shrub high(SHRUBHIGH) within three meters of the trap.Also within this ratio we estimated fungi and fern(FFRICH), herb (HERBRICH), shrub(SHRUBRICH), and tree richness (TREERICH),by direct counting of species. Finally, we measuredthe inclination soil surface (SLOPE) in angles,along with soil hardness (SOIL) estimated as thedepth (cm) a stick was able to penetrate the soilsurrounding the trap. We took four randomlymeasures of soil hardness within a ratio of threemeters of each trap location.To compare habitat among fragments we con-ducted an ANOVA among LQNR, Fragment 3,and Pinus matrix. To evaluate the association ofsmall mammals to habitat, we conducted a step-wise multiple regression among trapping successand habitat variable. A variable was allowed toenter the model when P ≤ 0.1. Trapping successwas calculated as the standardized total numberof captures for the total number of traps in eachsite. Regressions were made for the most com-mon trapped taxa (Abrothrix longipilis, Abrothrixolivaceus, Oligoryzomys longicaudatus, Octodonbridgesi, Thylamys elegans).

RESULTS

SPECIES COMPOSITION AND ABUNDANCE

We captured 435 individuals of nine small mam-mal’s species, over a total of 8,747 trap-nights(5% overall trapping success) performed in allfragments. Six murids (Abrothrix longipilis,Abrothrix olivaceus, Oligoryzomys longicaudatus,Irenomys tarsalis, Geoxus valdivianus, and the


341MAMMALIA • 2005 • 69 (3-4)


introduced Rattus rattus), one caviomorph(Octodon bridgesi), one didelphid (Thylamys ele-gans), and one microbiotherid (Dromiciopsgliroides) comprised the fauna of Maulino forest(Table 2). Four taxa (A. longipilis, A. olivaceus,O. longicaudatus, T. elegans) accounted for 80%

of total captures, and were used in subsequentquantitative analysis.The continuous forest at the LQNR supportedthe highest species richness, where nine specieswere recorded. In contrast, the pine plantationswere the poorest habitats, embracing six species


342 MAMMALIA • 2005 • 69 (3-4)

TABLE 2. — Trapping success (TS), number of captures (NC) and minimum number of individuals known to be alive (MNA) for eachtaxon of small mammal captured in Maulino forest and adjacent Pinus radiata plantation, VII Región, Chile. Similar proportions areunderlined. a, Los Queules National Reserve; b, not captured; c, total number of captures; *, 0.001 < P < 0.05; **, P < 0.001.

Site

LQNRa FRAGMENT 3 Pinus radiataMATRIX

Taxon TS NC MNKA TS NC MNKA TS NC MNKA χ2

RODENTIAMuridae

Abrothrix longipilis 2.14 84 39 2.17 23 17 5.64 103 52 35.3**

A. olivaceus 0.59 23 17 3.20 34 23 3.40 62 34 70.6**

Oligoryzomys longicaudatus 1.92 75 62 0.66 7 7 0.93 17 12 13.7**

Irenomys tarsalis 0.15 6 4 0.09 1 1 ncb 0 0 1.1Geoxus valdivianus 0.03 1 1 nc 0 0 nc 0 0Rattus rattus 0.59 23 ne 0.09 1 ne 0.38 7 ne 11.4*

OctodontidaeOctodon bridgesi 0.03 1 1 1.32 14 12 0.88 16 7 24.0**

MARSUPIALIADidelphidae

Thylamys elegans 0.43 17 9 4.52 48 16 0.88 16 5 12.0**

MICROBIOTHERIAMicrobiotheriidae

Dromiciops gliroides 0.03 1 1 nc 0 0 nc 0 0

Total captures 5.91 231 134 11.97 127 76 12.11 221 110 81.7**

TABLE 3. — Number of total captures of small mammals in Fragment 1, and its adjacent Pinus radiata plantation. Minimum numberknown to be alive is shown in parenthesis.

Taxon Maulino Forest Pinus Matrix Total

RODENTIAMuridae

Abrothrix longipilis 27 (8) 72 (46) 99 (48)A. olivaceus 16 (5) 62 (24) 78 (25)Oligoryzomys longicaudatus 6 (5) 10 (9) 16 (14)Irenomys tarsalis 1 (1) 1 (1)Geoxus valdivianus 2 (1) 2 (1)Rattus rattus 9(0) 2... 11......

MARSUPIALIADidelphidae

Thylamys elegans 39 (11) 10 (5) 49 (14)

Total captures 89 (30) 156 (85) 245 (102)Total species 5 0 5.. 5.....Total traps 955........ 997...... 1,952


(Table 2). The higher richness at the continuousforest was accounted for by the simultaneouspresence of D. gliroides, I. tarsalis, G. valdivianusand O. bridgesi, albeit they were scarce (Table 2).In fragments, combination of rare taxa onlyincluded I. tarsalis and O. bridgesi in Fragment 3.Dromiciops gliroides was never captured outsidethe continuous forest.Overall abundance of small mammals was higherin fragments and the exotic matrix, where trap-ping success doubled the abundance attained bysma l l mamma ls at the continuous forest(Table 2). Abrothrix longipilis and A. olivaceuswere the most abundant taxa in fragments andthe surrounding matrix. Oligoryzomys longicauda-tus was abundant in continuous forests, showingan extremely seasonal pattern with 87.5% of thecaptures occurring during April, June and August(Fig. 2). Rare taxa like G. va ldivianus,D. gliroides, or I. tarsalis were more “abundant”in continuous forest (Table 2). Although theintroduced murid R. rattus was present in all sur-veyed habitats, it was more abundant in continu-ous forest (Tab le 2). Octodon bridgesi wassignificantly more abundant in Fragment 3 andin the Pinus plantation. Only one individual wascaptured in the continuous forest, and no capturewas recorded in Fragment 1 (Tables 2; 3).

HABITAT DESCRIPTION AND PREFERENCES

Habitat characteristics differed among continu-ous and fragmented forest, as well as with thePinus matrix (Tab. 4). Twenty-two out of27 habitat variables differed among surveyedhabitats. Forest fragment and LQNR differed in11 out of 27 variables, while pine plantation dif-fered from Forest fragment and LQNR in18 variables each. Pinus plantation differed fromFragment and LQNR in nine variables, showingfor example lower litter cover, higher pine shrubcover, lower without overstory cover, higher pinetree cover, lower shrub and herb richness, higherfoliage density at 0-10 cm and 10-25 cm, andlower herb height (Table 4).Microhabitat variables explained small percent-age of variance in trapping success for Maulinosmall mammals, although all regression models


343MAMMALIA • 2005 • 69 (3-4)

FIG. 2. — Minimum number of individuals known to be alive(MNA) surveyed in native Maulino forest (LQNR and Fragment 3)and Pinus matrix, during January 1998-February 2000, VIIRegión, Chile.


were highly significant (Tab le 5). ForA. olivaceus, microhabitat explained 30% of trap-ping success, while for O. bridgesi, microhabitatcharacteristics explained only 14.8% (Table 5).Nevertheless, it is noteworthy that only 14 vari-ables were important in determining trapping

success (Table 6), and that they differed amongsmall mammals. Analyzing relevant variables totrapping success (Table 6), along with habitatattributes for each fragment (Table 4), it is possi-ble to conclude that ground branch cover is themost important variable in explaining species


344 MAMMALIA • 2005 • 69 (3-4)

TABLE 4. — Results (F and P level) of ANOVA for each microhabitat variable, testing for differences among LQNR, forest Fragmentand exotic Pinus radiata matrix, VII Región, Chile. N, number of measurements used in each case. Results of Sheffé posteriori testare indicated by lower case letters adjacent to average, with similar letters indicating no differences among habitat. Acronyms intext. a, Los Queules National Reserve.

LQNRa Fragment 3 Pinus MatrixMicrohabitat Average SD N Average SD N Average SD N F Pvariable

GRLITTCOV 70.7a 9.1 97 72.8a 18.1 30 54.5b 24.0 50 12.1 0.001GRBRACOV 7.2b 8.6 97 10.7a 11.4 30 12.2a 22.1 50 2.3 0.108GRSTECOV 8.1b 7.8 97 3.5a 6.5 30 1.9a 3.5 50 16.4 < 0.001GRSHRUCOV 0.9 4.6 97 0 0 30 0 0 50 1.6 0.199GRHERCOV 8.9ab 15.6 97 0.6a 3.4 30 15.7b 23.8 50 7.3 < 0.001GRSOILCOV 4.0c 5.7 97 11.7a 15.1 30 0.5b 1.9 50 20.9 < 0.001SHREMPCOV 51.6b 25.2 96 31.2ab 24.2 30 44.2b 23.6 50 8.0 < 0.001SHRJUVCOV 20.0c 17.1 96 9.8a 12.8 30 1.0b 3.3 50 32.1 < 0.001SHRBRACOV 12.2 16.0 96 8.9 16.8 30 6.3 14.7 50 2.4 0.097SHRSHRCOV 7.0c 14.3 96 46.9a 28.5 30 31.7b 22.4 50 57.0 < 0.001SHRHERCOV 8.8b 16.3 96 2.6ab 6.6 30 0a 0 50 9.1 < 0.001SHRPLICOV 0a 0 100 0a 0 30 3.3b 11.2 50 5.8 0.003TRWOSCOV 1.8a 4.9 96 5.8a 6.7 30 15.6b 22.0 50 19.9 < 0.001TRBRACOV 0 0 100 0.2 1.1 30 0 0 50 2.5 0.081TRNATCOV 23.2c 4.8 100 18.7a 7.3 30 0b 0 50 421 0TRPINCOV 0a 0 100 0a 0 30 81.4b 23.5 50 782.5 < 0.001TREERICH 2.6c 1.1 97 1.7a 1.1 30 1.1b 0.5 50 41.0 < 0.001SHRUBRICH 3.3a 1.7 97 2.9a 1.5 30 2.0b 0.9 50 11.5 < 0.001HERBRICH 3.0a 1.6 97 5.7a 2.6 30 1.6b 1.6 50 48.5 < 0.001FFRICH 0.3c 0.5 96 1.3a 0.6 30 0.3bc 0.4 50 49.6 < 0.001FODEN010 28.9a 27.0 96 22.2a 24.0 30 49.4b 24.6 50 13.8 < 0.001FODEN1125 27.5a 25.4 96 19.4a 24.2 30 43.1b 25.1 50 9.9 < 0.001FOLDEN12 31.7 21.7 96 39.8 22.0 30 28.8 22.9 50 2.4 0.092HERBHIGHT 58.4a 58.4 97 71.3a 86.1 30 21.1b 20.5 50 9.9 < 0.001SHRUBHIGH 238.1b 94.6 97 294.7a 110.1 30 257.6ab 76.1 50 4.3 0.014SLOPE 11.8a 9.8 95 31.3b 11.8 30 10.7a 7.9 50 52.9 < 0.001SOIL 10.2c 2.8 95 7.0a 2.6 30 8.8b 2.8 50 15.4 < 0.001

TABLE 5. — Models for microhabitat association obtained by stepwise multiple regression, for Abrothrix longipilis, A. olivaceus,Oligoryzomys longicaudatus, Thylamys elegans and Octodon bridgesi, captured in remnants of Maulino forest, and adjacent Pinusradiata matrix, in Tregualemu, VII Región, Chile. R, coefficient of multiple correlation; R2, adjusted by dividing the error sum ofsquares and total sums of square by their respective degrees of freedom (d.f.). The F-value and resulting P-value is used as an ove-rall F test of the relationship between the dependent variable and the set in independent variables.

Taxon R R2 Adjusted R2 S.E. F d.f. 1 d.f. 2 P

A. longipilis 0.477 0.228 0.195 1.930 6.921 9 211 < 0.000A. olivaceus 0.547 0.300 0.263 1.301 8.146 11 209 < 0.000O. longicaudatus 0.401 0.161 0.104 0.764 2.819 14 206 0.001T. elegans 0.494 0.244 0.204 0.928 6.125 11 209 < 0.000O. bridgesi 0.385 0.148 0.108 0.532 3.668 10 210 < 0.000


trappability for A. longipilis, since this value ishigher in Pinus plantation, where species abun-dance was higher. Also, foliage density in the first10 cm above ground seems to be important forthe trappability of this species. Elevated shrubhigh seems to explain capture pattern for A. oli-vaceus, that was higher in Fragment 3 and Pinusplantation. For O. longicaudatus the presence ofpines seems to affect negatively its presence.Higher abundance of this species was observed inLQNR (Table 2) where lower pine abundanceexisted (Table 4). It is significant to notice thespatial heterogeneity that has abundance in thisspecies, since it was captured more frequently inPinus matrix than native forest in Fragment 1.Unfortunately habitat use was not analyzed inthis fragment. Higher abundance of T. elegansobserved in Fragment 3 is associated to elevatedshrub cover and higher soil hardness. Finally, the

distribution of O. bridgesi, which was more abun-dant in Fragment 3 and the Pinus plantation,seems to be associated to higher shrub cover.

POPULATION STRUCTURE AND BODY CONDITION

Proportion of females, and juveniles, and alsoproportion of reproductive individuals did notvary among fragments (Table 7). There were nodifferences in body condition (BC) among sexesfor A. longipilis (F = 0.9; 32 d.f.; P = 0.35), orA. olivaceus (F = 0.28; 13 d.f.; P = 0.61). BC didnot vary among native forest and Pinus planta-tion for A. longipilis (F = 0.57; 30 d.f.; P = 0.64;Native forest-BC 124.1 ± 7.3, N = 15; Pinusplantation-BC 120.5 ± 11.4, N = 19). On thecontrary, A. olivaceus achieved higher BC (F =8.6; 13 d.f.; P = 0.011) in Pinus matrix (BC =99.3 ± 4.6, N = 7) compared to native forest(BC = 87.3 ± 9.9, N = 8).


345MAMMALIA • 2005 • 69 (3-4)

TABLE 6. — Standardized coefficients for stepwise multiple regression among microhabitat variables for Abrothrix longipilis,A. olivaceus, Oligoryzomys longicaudatus, Thylamys elegans and Octodon bridgesi, captured in remnants of Maulino forest, andadjacent Pinus radiata plantations, in Tregualemu, VII Región, Chile. Only significant variables are shown. Acronyms in text.

Microhabitat A. longipilis A. olivaceus O. longicaudatus T. elegans O. bridgesivariable

GRLITTCOV 0.202GRBRACOV 0.148 0.269GRSTECOV – 0.121SHRJUVCOV – 0.156SHRSHRCOV 0.218SHRBRACOV 0.135 0.162 0.165TRNATCOV 0.181 – 0.192TRPINCOV – 0.217FODEN010 0.199FOLDEN12 0.127 – 0.158SHRUBHIGH 0.159 0.142SOIL – 0.158FFRICH 0.264

TABLE 7. — Proportion of females (F), juveniles (J), and reproductive (R) out of total individuals captured in Maulino forest, andadjacent Pinus radiata matrix. χ2 and P were non significant for each comparison and are not shown. Abbreviations: LQNR, LosQueules National Reserve; Al, Abrothrix longipilis; Ao, A. olivaceus; Ol, Oligoryzomys longicaudatus; Te, Thylamys elegans.

Taxon LQNR Fragment 3 Pinus MatrixF J R F J R F J R

Al 0.56 0.15 0.59 0.47 0.06 0.88 0.35 0.13 0.73Ao 0.71 0.18 0.23 0.56 0.17 0.74 0.41 0.09 0.67Ol 0.61 0.62 0.06 0.14 0.42 0.00 0.42 0.08 0.00Te 0.67 0.33 0.11 0.69 0.31 0.31 0.40 0.40 0.40


MOVEMENT

A total of 102 individuals was captured inFragment 1, and they were used to test for habi-tat preferences among native and exotic vegeta-tion. Here, 245 captures were achieved,comprising a total of seven species (Table 3). ForA. longipilis, only six individuals (10.7% of thetotal) moved between native forest and Pinusplantations. All were adult males, except for onejuvenile. For A. olivaceus, four (13.8%) individu-als moved between native and Pinus matrix, allbeing adults, three of them males. For T. elegans,two individuals (11.8%) moved among adjacenthabitats. These were one male and one female,both adults.

DISCUSSION

Fragmentation of Maulino forest, along withhabitat replacement, affected small mammal’scomposition and abundance. As expected, speciesrichness was higher in continuous forests, exceed-ing richness in fragments and Pinus plantations.Continuous forest shelters a particularly interest-ing small mammal assemblage, due to the coexis-tence of relict populations of Valdivian fauna likeD. gliroides, G. valdivianus, I. tarsalis, andO. bridgesi, a combination not observed in anyother forest type today (Reise & Venegas 1987;Saavedra & Simonetti 2000, 2001). Since cen-tral-south Chile probably acted as a forest refugeeduring Quaternary glaciations (Villagrán et al.1995), mammals now inhabiting Maulino forestprobably constitute the oldest representatives ofthe fauna currently thriving in the more southernValdivian district. In Maulino forest rare speciesare concentrated in continuous forest, while frag-ments harbor mostly common taxa. Fragmen-tation of Mau lino forests seems to have asignificant impact on overall small mammal’srichness, menacing the persistence of rare taxa ofsmall mammals. The fact that rare species areconcentrated in continuous forest indicates thatthis native spot has a significant value for the per-sistence of these populations. Despite largerextensions occupied by exotic P. radiata planta-

tions, this was the habitat with lowest speciesrichness, like Burca site located southern of ourstudy site (Muñoz & Murúa 1989). Nevertheless,total abundance was higher in Pinus plantations,and adjacent Maulino forest remnants, a patternnot described before for this fragmented forest.Here, as has been observed in other forests ofsouthern Chile (Muñoz & Murúa 1989; Kelt2000), A. longipilis, A. olivaceus, O. longicaudatusand T. elegans are the principal components ofsmall mammal assemblages, becoming extremelyimportant to define the observed abundance pat-tern in Maulino forest. All these taxa use and/orare more abundant in the Pinus plantations, indi-cating that this constitutes a suitable habitat forthem. This fact is reflected not only in abundancepattern, but also at the individual level, wherebody condition is even better in Pinus matrixcompared to native forest at least for A. olivaceus.Although negative effects to fragmentation aretheoretically expected, particularly for habitatspecialists, a growing amount of evidence isshowing that small mammals are able to over-come this disturbance, achieving elevated popula-tion or body size (e.g., Mahan & Yahner 1998;Díaz et al. 1999). Our findings indicate thatPinus plantations do fulfil at least partial habitatrequirements for some small mammals thatinhabit Maulino forest, mostly due to the exis-tence of some ground structures, like shrubs, thatare probably used as refuges by these small mam-mals. Habitat suitability of Pinus plantations maybe increased by the existence of significant foodresources like insects that show similar abun-dance in pine matrix compared to forest frag-ments (Grez et al. 2002). Actually, A. longipilis,A. olivaceus, and T. elegans that were the mostabundant taxa in Pinus plantations, includeinsects in their diet (Muñoz-Pedreros et al.1990).Nevertheless, small mammal’s populations inpine plantations and even in fragments can berelict ones or will be metapopulations, withoutstable numbers. Habitat suitability in theseconditions should be low, no offering suitableconditions particularly for habitat specialists, thatcan use these habitats as corridors. Although this


346 MAMMALIA • 2005 • 69 (3-4)


proposition remains to be tested for small mam-mals, Acosta (2001) showed that Monterrey pineis used as a corridors by other species likePseuda lopex cu lpaeus and Oncifelis guigna.Neither of these carnivores prefers Pinus planta-tions to native forest, but nevertheless they areable to use this habitat as food source, or corridoramong native forest fragments (Acosta 2001).Also Pinus plantations that surrounded Maulinoforest fragments do not constitute a barrier forseveral bird species, including understory birds(Estades et al. 1999; Vergara 2002). Dependingon species idiosyncrasies, these taxa are able touse Pinus plantations by developing nesting activ-ities and achieving important abundance. Allthese evidence suggest that forest fragments andPinus plantations are not alternative habitatsfor some species, and they are used as part of acontinuum where forest fragments are articulatedthrough pine plantations.Our findings have important consequences forthe conservation and management of the lastremnants of Maulino forest that persist immersedin this very modified habitat. Preservation of thelarge tract of native forest seems to be indispensa-ble for the maintenance of elevated species rich-ness at the landscape level, preventing thedisappearance of rare and endangered taxa. Theyseem to find a refuge in this last spot of Maulinoforest that is not supported by smaller forest frag-ments. At the same time, the maintenance of for-est remnants interspersed in a well developedPinus plantation, seem to be an adequate man-agement practice for ensuring persistence ofcommon species that are able to use this foreignhabitat, along with increasing the connectivity offorest remnants for uncommon taxa. This land-scape approach to the conservation of this seri-ously modified habitat seems to be a realisticalternative to conserve native fauna that stillinhabits the area.

AcknowledgementsFondecyt 2990120 to BS, 1981050 to JAS,1010852 to JAS, and PG/15/99 fromUniversidad de Chile to BS supported this work.We thank Forestal Millalemu S.A. and CONAF

VII Región for granting permits to work on theirlands. J. Arroyo, C. Veloso and T. Lasnibat werehelpful field assistants. Nicolás and Julie madethe Résumé. BS is a doctoral fellow of CONI-CYT.

REFERENCES

ACOSTA G. 2001. — Efecto de la fragmentación delbosque nativo en la conservación de Oncifelis guigna yPseudolapex culpaeus en Chile central. Master Thesis,Universidad de Chile, Santiago, Chile.

BUSTAMANTE R. O. & CASTOR C. 1998. — Thedecline of an endangered temperate ecosystem: theruil (Nothofagus alessandrii) forest in central Chile.Biodiversity and Conservation 7: 1607-1626.

DÍAZ M., SANTOS T. & TELLERÍA J. L. 1999. —Effects of forest fragmentation on the winter bodycondition and population parameters of an habitatgeneralist, the wood mouse Apodemus sylvaticus: atest of hypotheses. Acta Oecologica 20: 39-49.

ESTADES C. F. & TEMPLE S. A. 1999. — Deciduous-forest bird communities in a fragmented landscapedominated by exotic pine plantations. EcologicalApplications 9: 573-585.

GLANZ W. E. 1970. — Comparative Ecology of SmallMammal Communities in California and Chile. Ph.D.Thesis, University of California, Berkeley, USA.

GREZ A., MORENO P. & ELGUETA M. 2002. —Coleópteros (Insecta: Coleoptera) epígeos asociadosal bosque Maulino y plantaciones de pino aledañas.Revista Chilena de Entomología 29: 9-18.

HILBORN R., REDFIELD J. A. & KREBS J. 1976. — Onthe reliability of enumeration for mark and recap-ture census of voles. Canadian Journal of Zoology 54:1019-1024.

KELT D. A. 2000. — Small mammal communities inrainforest fragments in Central Southern Chile.Biological Conservation 92: 345-358.

LARA A., DONOSO C. & ARAVENA J. C. 1995. — Laconservación del bosque nativo en Chile: problemasy desafíos, in ARMESTO J. J., VILLAGRÁN C. &ARROYO M. T. K. (eds), Ecología de los bosquesnativos de Chile. Editorial Universitaria, Santiago:335-362.

LINDENMAYER D. B., CUNNINGHAM R. B., POPE M. L.& DONNELLY C. F. 1999. — A large scale “experi-ment” to examine the effects of landscape contextand habitat fragmentation on mammals. BiologicalConservation 88: 387-403.

MAHAN C. & YAHNER R. H. 1998. — Lack of popula-tion response by eastern chipmunks (Tamias stria-tus) to forest fragmentation. The American MidlandNaturalist 140: 382-386.

MUÑOZ A. & MURUA R. 1989. — Efectos de la refor-estación con Pinus radiata sobre la diversidad y


347MAMMALIA • 2005 • 69 (3-4)


abundancia de los micromamíferos en un agroeco-sistema de Chile central. Turrialba 39: 143-150.

MUÑOZ A. & MURUA R. 1990. — Control of smallmammals in a pine plantation (Central Chile) bymodification of the habitat of predators (Tyto alba,Strigiforme and Pseudolapex sp., Canidae). ActaOecologica 11: 251-261.

MUÑOZ-PEDREROS A., MURUA R. & GONZALEZ L.1990. — Nicho ecológico de micromamíferos enun agroecosistema forestal de Chile central. RevistaChilena de Historia Natural 63: 267-277.

OSGOOD W. H. 1943. — The Mammals of Chile.Zoological Series Field Museum of Natural History,Volume 30, Chicago.

REDFORD K . H . & EISENBERG J . F . 1992. —Mammals of the Neotropics: the Southern Cone.Volume 2. Chile, Argentina, Uruguay, Paraguay.The University of Chicago Press, Chicago.

REISE D. & VENEGAS W. 1987. — Catalogue records,localities and biotopes from research work on smallmammals in Chile and Argentina. Gayana, Zoología51: 103-130.

SAAVEDRA B. & SIMONETTI J. A. 2000. — A northernand threatened population of Irenomys tarsalis(Mamma lia: Rodentia) from centra l Chile.Zeitschrift für Säugetierkunde 65: 243-245.

SAAVEDRA B. &. SIMONETTI J. A. 2001. — Newrecords of Dromiciops gliroides (Microbiotheria:Microbiotheriidae) and Geoxus va ldivianus(Rodentia: Muridae) in central Chile: their implica-tions for biogeography and conservation.Mammalia 65: 96-100.

SAN MARTIN J. & DONOSO C. 1995. — Estructuraflorística e impacto antrópico en el Bosque Maulinode Chile, in ARMESTO J. J., VILLAGRÁN C. &ARROYO M. T. K. (eds), Ecología de los bosquesnativos de Chile. Editorial Universitaria, Santiago:153-168.

SIMONETTI J. A. 1986. — Heterogeneity of recapturesin Chilean Small mammals. Revista Chilena deHistoria Natural 59: 59-63.

VERGARA P. 2001. — Fragmentación de bosques y usodel hábitat por Rinocríptidos . Master Thesis,Universidad de Chile, Santiago, Chile.

VILLAGRÁN C., MORENO P. & VILLA R. 1995. —Antecedentes palinológicos acerca de la historia delos bosques chilenos, in ARMESTO J. J., VILLAGRÁNC. & ARROYO M. T. K. (eds), Ecología de los bosquesnativos de Chile. Editorial Universitaria, Santiago:51-69.

ZAR J. H. 1984. — Biostatistical analysis. Prentice-Hall, New Jersey.

Submitted on 9 April 2003;accepted on 11 June 2004.


348 MAMMALIA • 2005 • 69 (3-4)


small mammals of maulino forest remnants, a vanishing ecosystem of south-central chile

Documents