ZOBODAT - www.zobodat.at

Zoologisch-Botanische Datenbank/Zoological-Botanical Database

Digitale Literatur/Digital Literature

Zeitschrift/Journal: Mammalian Biology (früher Zeitschrift für Säugetierkunde)

Jahr/Year: 2001

Band/Volume: 66

Autor(en)/Author(s): Caro T. M., Brock Rachel, Kelly Marcella

Artikel/Article: Original investigation Diversity of in the Bladen , , and factors affecting their trapping success 90-101 © Heritage Library, http://www.biodiversitylibrary.org/

Mamm. biol. 66 (2001) 90-101 Mammalidn Biology © Urban & Fischer Verlag http://www.urbanfischer.de/journals/mammbiol ^^^P" Zeitschrift für Säugetierkunde

Original investigation

Diversity of mammals in the Bladen Nature Reserve, Belize, and factors affecting their trapping success

By T. M. Caro, Rachel Brock, and Marcella Kelly

Department of Wildlife, Fish and Conservation BioLogy, University of Cahfornia, Davis, California, USA

Receipt of Ms. 12. 11. 1999 Acceptance of Ms. 20. 04. 2000

Abstract

The presence of 33 non-volant species was recorded in the BLaden Nature Reserve, an area of subtropicaL wet in south central Beh'ze, as determined from waLking transects and using Sherman and Tomahawk traps to capture mammals. Using trapgrids over 6 075 trapnights, the ef- fects of trap design, bait, moon phase, , elevation, and proximity to a river on three mea- sures of trapping success were examined systematicalLy. Open wire mesh traps yielded somewhat higher trapping success than Sherman traps; oats and molasses produced higher trapping success than other kinds of bait; and trapping success was higher in selectively logged than in unlogged forest, and marginally higher at lower elevations and dose to a river. Moon phase had no effect on trapping success. These results provide baseline data on mammal diversity in a relatively un- exploited area of central America and, though preliminary, indicate which aspects of trapping tech- nique need to be standardized when comparing species diversity and abundance across neotropical

Sites.

Key words: Mammals, trappability, diversity, Bladen, Belize

Introduction

Most ecological studies of neotropical 1997; Malcolm 1997), loss of top predators mammals require data on species diversity (Wright et al. 1994; Terborgh et al. 1997), and abundance. For example, questions and mammal exploitation (Dirzo and Mir- about Population dynamics (e. g., O'Con- ANDA 1991; Glanz 1991; Wright et al.

NELL 1989), Population demography (e. g., 2000) on communities of mammals. In the ToRRES-CoNTRERA et al. 1997), Community neotropics data on mammal communities structure (e. g., Asquith et al. 1997), and re- are growing (Voss and Emmons 1996) but gions of mammal abundance (e. g., Mares they still come from only a handful of loca- 1992) all require Information on the num- tions, the most notable of which are La Sel- ber of mammal species or their relative va in Costa Rica (Timm 1994 a), Los Tuxlas, abundance in an area. Conservation studies (Estrada et al. 1994), Barro Color- similarly require data on diversity and ado Island, Panama (Glanz 1990), Cosha abundance in order to understand the ef- Cashu, Peru (Janson and Emmons 1990) fects of (Lynam and near Manaus, Brazil (Malcolm 1990)

1616-5047/01/66/02-090 $ 15.00/0. © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammaLs in the Bladen Nature Reserve 91

which limits our ability to make generaliza- mal densities being lower (Eisenberg et al. tions about how mammalian communities 1979) and small mammal abundance being are organized. higher in logged sites (Kasenene 1984; Isa- As studies of mammals in the neotropics in- biyre-Basuta and Kasenene 1987; Mal- crease, and comparisons between areas be- colm 1995). Increased elevation has been come more feasible, researchers must stan- found to change species diversity in some dardize their techniques for estimating areas of the tropics (Rickart et al. 1991) mammal abundance, or at least be aware while proximity to seasonally flooded areas of the biases inherent in different methods close to rivers may alter species diversity of mammal estimation (Wilson et al. and abundance in complex ways (Janson 1996). We therefore investigated six differ- and Emmons 1990). Therefore the aim of ent factors that influenae trapping success this study is to investigate the diversity of outside the neotropics and might therefore species at a new site in the neotropics and be of importance elsewhere. Two of these to analyse factors affecting trapping success were methodological factors (type of small in this environment. mammal trap and characteristics of the bait) and four were ecological factors (phase of the moon, selective logging, ele- Material and methods vation, and proximity to rivers). Trap design can influence trapping success The study was conducted in and adjacent to the substantially in temperate regions (Sealan- Bladen Nature Reserve in the , this der and James 1958) and problem has , Belize (Fig. 1). The reserve en- been identified in at least one neotropical compasses 350 km^ of the watershed of the Bla- Site (WooDMAN et al. 1996). We therefore den River between latitudes 16° 36' 18" and compared measures of trapping success 16° 24' 34" N and longitudes 88° 42' 16" and using hand-made wire mesh small mammal 89° 04' 51" W. Elevation ranges from 50 to 1000 m. Rainfall averages around 3 000 per traps which could be seen through with mm annum or more: January to April are the driest Standard Sherman traps which obscure visi- months with starting from June onwards biUty. It is also well known that baits can af- through November. Mean monthly temperatures fect trapping success in temperate zones ränge from 16 ° to 33 °C in Belize. Much of the re-

(e. g., BucHALCZYK aud Olszewski 1971). serve is composed of Coban formation Although far less work has been carried and volcanic rock. The reserve contains subtropi- out on this problem in the neotropics, preli- cal wet forest (Hartshorn et al. 1984) with the minary evidence suggests that bait is not an lowest parts, along the main flow of the river important factor in mammal trapping where we worked, supporting alluvial soils and tall broad-leaved forest. Most of the Bladen Nat- (WooDMAN et al. 1996). Increased moon- ure Reserve is unlogged; selective logging has light lowers trapping success in open habi- been practiced immediately outside its eastern tats in temperate regions (Brown et al. border, however. The reserve is subject to un- small often show 1988) because mammals known hunting pressure by local people for game changes in activity in accordance with a re- meat. duction of predation risk from visually The study was conducted during four periods of hunting aerial and terrestrial predators fieldwork: June and July 1994, June through (Clarke 1983). Although moonlight is August 1995, March 1997 and July 1998. Some known to influence activity patterns of rain feil in each of these periods but trapping was suspended under extremely wet conditions some tropical mammals (Fenton et al. because we were concerned about hypothermia 1977; Emmons 1987; Alkon and Saltz of captured individuals. All the work was con- 1988), it has received little systematic atten- ducted adjacent to the eastern entrance of the re- tion in small neotropical mammal trapping serve, 1-3 km inside, and 0.5 km outside it. studies. Selective logging has many influ- Mammal diversity was assessed in three ways: ences on mammal Community structure in through night and daytime walks in which obser- tropical environments with arboreal mam- vations and calls of mammals were noted, by in- © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

92 T. M. CAROetaL

Fig. 1. Belize and the location of Bladen Nature Reserve (in black) in the country. Top right inset shows Location of Belize in central America. Bottom right inset shows the boundary of Bladen Nature Reserve (dashed line) and the Bladen Branch River and its tributaries (solid lines). The entrance to the reserve is at Forest Hill shown as a black dot. Three 2 km transects were located (i) just inside the reserve boundary running south west to north east terminating at Forest Hill; (ii) between Forest Hill and the Richardson Creek (fifth tributary upstream from the entrance) confluence south and parallel to the Bladen River; (iii) and outside the reserve from and due north east of Forest Hill. All trapgrids within the reserve were set south of the Bladen River between Forest Hill and the sec- ond tributary upstream from the entrance. Trapgrids outside the reserve were set within 500 m of the reserve en- trance. Black bar denotes 10 km; arrow shows north (adapted from Hartshorn et al. 1994). © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammaLs in the BLaden Nature Reserve 93

terpreting mammal tracks in the mud, and green or ripe bananas, raisins, or a fruit found on through live trapping. Abundances of certain the forest floor, Warre Cohune palm, Astrocar- mammal species were estimated through repeat ynm mexicamim. These traps were set approxi- trapping using Standard Sherman, custom-made mately 50-450 m from the river. In one set of tests small mammal traps, and three sizes of Toma- these traps were placed at a height of 1-2.5 m in field hawk traps. We used the guide of Emmons trees and baited with oats and molasses in order (1990) to identify mammals in the field. Two pre- to estimate squirrel abundance. vious surveys had been conducted in mammal We also set large Tomahawk traps the Bladen (Brokaw and Lloyd-Evans 1987; (65x22.5x22.5 cm) in a 6x8 grid with traps set The Nature Conservancy 1993). 50 m apart. These were baited with either green We set Up three 2 km transects that were walked or ripe bananas, Astrocarynm mexicamim fruit, at a pace of 1 km/hour. One was located within commercial cat food or fresh fish; these latter the reserve and followed an abandoned logging two baits were combined in bait analyses since track that serves as the main road into the re- they both contained animal protein and smelled serve; the second was further inside the reserve similar. In none of the trapping protocols were on the south side and parallel to the Bladen Riv- different baits run at the same time on either the er; the third was located adjacent to but outside same or different grids. the Reserve in a selectively logged area. Night Traps were usually set for 5 consecutive nights walks (N = were conducted between 18.30 18) although a small minority was set for fewer or and 21.00 h and animals were spotted using flash- more nights (ränge 1-7). Traps were set either in = lights; day walks (N 28) took place between unlogged forest with a tall thick canopy and rela- 05.30 and 07.30 h. Tracks of mammals were exam- tively open understory inside Bladen Nature Re- ined opportunistically and in specially prepared serve, or in the area east of the reserve where se- flats of mud, at the center at which was placed a lective logging allowed light to penetrate, commercial carnivore olfactory Iure. producing a thicker understory of Vegetation. Small mammals were trapped using Standard Traps were opened and baited between 16.00 and Sherman traps (23x8x8 cm), or traps of the 17.00 h and checked next morning between 06.00 same dimensions custom-made entirely out of and 09.00 h. Captured animals were individually galvanized wire mesh (by the late R. Schwab) marked by cutting small patches of für since we save for a galvanized aluminum plate door that were only interested in recaptures over a maxi- was part of the floor until the door swung up- mum of 7 days. Quarter of the moon was noted wards on closure (see also Emmons 1984). Traps during each sequence of trapping; in some ana- were usually set in an 8 x 8 or 7 x 7 grid with traps lyses traps set during the first and last quarter the second 10 m apart; infrequently they were set along a spanning the new moon, and then quarters the füll were transect precluding calculation of density. Grids and third spanning moon combined. of small mammal traps were baited either with a recorded number of species caught, percen- mixture of oats and molasses, with peanut butter, We tage capture success, individual mammals caught/ with green or ripe bananas, or with a medley of 100 trapnights, and densities when traps were set fruit consisting of apple, banana, and coconut in a grid Square. Percentage capture success was mixed together. Ripe and green bananas were the number of captures divided by the number of lumped together in bait analyses because green trapnights (i.e., number of traps multiplied by bananas quickly ripened in traps making it diffi- the number of nights on which they were set); in- cult to distinguish between the two. Traps were dividual mammals caught per 100 trapnights was usually set on the flat valley floor of the reserve the number of different individuals captured di- approximately 20-200 m from the Bladen River vided by the number of trapnights x 100; and den- bank. In another protocol, trapgrids were set at sities were calculated by dividing the number of higher elevations on the steep limestone slope individuals captured by the area covered by the bordering the approximately 50 and grid expressed as number of individuals/km"^. We valley floor 200-250 100 m above the and m from took this area to be the dimensions of the grid the river. plus 5 m either side (i. e., 70 m x 70 m or 4 900 m^ We set two types of Tomahawk traps in a 7 x 7 grid) because paucity of captures made

(40x13x13 cm and 40x17x17 cm) which we it difficult to calculate maximum distance be- termed middle-sized traps. These were usually tween captures of known individuals. We did not set in a 6x8 grid with traps 50m apart. Traps use mark-recapture techniques to estimate den- were baited either with oats and molasses, with sity because recapture rates were so low. © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

94 T. M. Caro etaL.

Data were analysed by comparing trapgrids non-volant species (n = 18 grids; 3 521 trap- although the number of trapnights that these re- nights). With the middle-sized Tomahawk presented is also presented for clarity. Non-para- traps, we caught only two species, Ototyl- metric statistics were used as number of species omys phyllotis and Tylomys nudicaudas. was an ordinal measure, and captures/trapnight Trap success was low at 0.7% (sd± 1.0), or and individual mammals caught/100 trapnights 0.6 individuals/100 trapnights (sd ± produced too many zeroes (no captures) to justify 0.8) normalizing the data required for parametric sta- (n = 9 grids; 1 354 trapnights). Density of tistics. The use of non-parametric statistics made these two species was 2/km^ and it difficult to control for confounding variables; in- 10/km^, respectively (n = 5 grids; 1 200 trap- stead we conducted a series of carefully controUed nights). With the large Tomahawk traps, comparisons among grids by excluding variables we caught Didelphis marsupialis, Didelphis that were found to be important in previous ana- virginianus, Dasypus novemcintus and a lyses even though these resulted in a reduction in Tylomys nudicaudus giving an average per- sample sizes. a was set at 0.05; nevertheless p va- centage trap success of 4.3% (sd±4.3), or lues lying between 0.1 and 0.05 are noted and dis- 3.9 individuals/100 trapnights cussed with appropriate caution. (sd ± 3.9) (n = 9 grids; 1 218 trapnights). Density of these species was 28/km^ 8/km^ 2/km^ and 2/km^, respectively (n = 5 grids; Results 1 152 trapnights). Excluding , we calcu- lated a Shannon-Wiener index of 2.021. Captures

In this study, twenty eight species of non- Factors affecting trapping success volant mammals were identified inside and outside but within 0.5 km of the Bladen Type of trap; Compared to Standard Sher- Nature Reserve, although two of these were man traps, custom-made wire mesh traps equivocal identifications (Tab. 1). In our of the same dimensions caught marginally study all of these species except five, Phi- more terrestrial mammal species (n = 6; lander Opossum, Urocyon cinereoargenteus, 20 grids, respectively; 618, 3 618 trapnights, Conepatus semistriatus, Leopardus sp., and Means (Xs) = 1.0 (sd±0.9), 2.1 (sd±1.4) Panthern onca were found inside the re- species, Mann-Whitney U test, z = -1.763, serve; in a previous study conducted by the P = 0.078), demonstrated marginally higher Rapid Ecological Assessment Team in 1993 percentage capture success (Xs - 2.2% a and a small fehd had been identi- (sd±3.0), 7.7% (sd±6.1) respectively, fied inside Bladen (The Nature Conser- z = -1.951, P = 0.051), and caught a margin- VANCY 1993). Our results, combined with ally greater number of individuals per trap- those of the two earUer surveys (Tab. 1), night (Xs-2 (sd±3), 7 (sd±6) indivi- show that the Bladen area holds a minimum duals/100 trapnights respectively, z = - of 33 non-volant species including large 1.830, P = 0.067). predators such as Felis concolor and Type of bait; For small mammal traps, there Panthera onca. were significant differences in the number Employing small mammal traps, we cap- of species caught (n = 26 grids; 4 236 trap- tured five non-volant species, Heteromys nights, Kruskal-WalHs test, H = 11.444, desmarestianus, Ototylomys phyllotis, Tyl- P = O.Ol), percentage trap success (H = omys nudicaudus, Marmosa robinsoni, Or- 8.464, P = 0.037), and individuals captured/ yzomys couesi and an unknown species of 100 trapnights (H = 7.888, P = 0.048) de- with an average percentage capture suc- pending upon the type of bait offered. On cess of 6.5% (sd±5.9), or 5.6 individuals/ each measure, oats and molasses were most 100 trapnights (sd ± 5.6) (n = 26 grids; successful followed by green and ripe bana- 4236 trapnights). These traps yielded re- nas combined, and then the fruit medley spective densities of 6836/km^, 270/km^, (Tab. 2). There were no significant differ- 183/km^ 925/km^ and 2 127/km^ for the five ences between baits on measures of density. © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammaLs in the BLaden Nature Reserve 95

Table 1. List of species of mammaLs in and immediateLy adjacent to BLaden Nature Reserve.

1994-1998 this study: Tp: trapped; 0: observed; Tr: tracks; H: hieard; I/O: inside or outside BLaden Nature Re- serve. 1. refers to species noted by tfie 1993 Rapid EcoLogicaL Assessment Team (The Nature Conservacy 1993);

2. refers to species noted by the 1987 Manomet survey (Brokaw and Lloyd-Evans 1987). * species may have been Marmoso mexicana; + species may have been Leopardus wiedii ().

Scientific name Common name Tp 0 Tr H I/o

MarsupiLia

Dideiphis marsupiatis Common opossum X I

Dideiphis virgim'anus 2 Virginia opossum X I Philander opossum Common gray four-eyed opossum X 0

Chiron ectes minimus 1 Water opossum X I

Micoureus alstoni 1 ALston's woLLy mouse opossum Marmosa robinsoni* Robinson's mouse opossum X I/o

Xenarthra

Tamanduo mexicana Northern tamandua X I Dasypus novemcinctus Nine-banded Long-nosed armadiLLo X I/o

Chiroptera

• Balaniopteryx io 1 Least sac-winged bat V T Noctibo leponnus Greater fishing bat X i Y T Unidentified species A 1

Primates

Alouatto pigro 1 Mexican bLack howLer monkey X I/o

Ateles geojfroyi 1, 2 CentraL American spider monkey X I

Carnivora

Urocyon dnereoargenteus Gray fox X 0 V T Nasua narica 1 White-nosed coati A i Potosflavus Kinkajou X I Mustelafrenata 1 Long-taiLed weaseL

Eiro barbaro 2 X I Conepotus semistriatus Striped hog-nosed skunk X 0 Lontra tongicoudus 2 NeotropicaL otter X I Leopardus pardabs + 1 X 0 Puma concolor 2 Puma X I/o

runLnera onco i Jaguar X 0

PerissodactyLa

Tapirus bairdii 1, 2 Baird's tapir X I/o

ArtiodactyLa

Toyassu tojacu 1 CoLLared

Tayassu pecari 1, 2 White-Lipped peccary X I/o Y T /n Mozoma omericona 1, 2 Red brocked deer A i/U

Odocoileus virginionus 1, 2 White-taiLed deer

Rodentia

Sciurus yucatanensis Yucatan squirreL X I

Sdurus deppei 1, 2 Deppe's squirreL X I Heteromys desmarestianus Forest spiny pocket mouse X I/O Oryzomys couesi Coues' rice rat X I/O

Tylomys nudicaudus 1 Naked-taiLed cLimbing rat X I/O Ototybmys phyllotis Big-eared cLimbing rat X I/O

Agouti poco 1, 2 Paca X I/O

Dosyprodia punctata 1, 2 CentraL American agouti © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

96 T. M. Caro etaL

Table 2. Mean (X) and Standard deviation (sd) measures of trapping success in aU smaLL mammal traps separated by type of bait; round brackets refer to numbers of trapgrids or traplines, Square brackets to the number of trapnights.

Oats and Peanut butter Bananas Fruit medLey moLasses

(7) (3) (14) (2)

[1 973] [502] [1461] [300]

Kliimhpr nf cnorioc X 3.1 0.7 1.6 1l.Un sd 0.9 0.6 1.2 1.4

rerceniage trap suclcss Ay iU. J U.D O.D i.U sd 1.0 0.6 6.8 1.4

Individuals/100 trapnights X 7.5 0.6 6.5 0.4 sd 1.3 0.6 6.8 0.6

(7) (1) (10)

[1 973] [245] [1 303]

IndividuaLs/km^ X 8 600 6100 12 200 sd 14 300 0 12 400

* traps were set in a Line, rather than grid, so estimates of density are unavaiLable. however. When analyses were restricted found that percentage capture success and only to small mammal wire mesh traps, the individuals/trapnight were significantly number of species captured still differed greater in logged forest outside Bladen than significantly by type of bait (n = 20 grids; in unlogged forest inside (n = 7,13 grids^re- 3 618 trapnights, H = 7.792, P = 0.02) alt- spectively; 741, 2 877 trapnights, Xs = hough this was no longer the case for mea- 12.2% (sd±5.4), 5.3% (sd±5.1), z = 2.260, sures of percentage trap success and indivi- P = 0.024; Xs = 12 (sd ± 4), 4/100 trapnights duals/100 trapnights. There were no effects (sd ± 6), z = 2.539, P = 0.011). In the Toma- of bait for any measure in the medium-sized hawk traps, a slightly greater number of or large traps. species was captured in the unlogged than Ivloon phase: Considering either small, in the logged forest in the medium-sized medium-sized or large traps, there was no (n = 7,2_grids, respectively; 1 312, 42 trap- effect of moon phase on number of species nights, Xs = 0.7 (sd ± 0.5), 0 (sd ± 0) species, captured, percentage capture success, num- z = 1.690, P = 0.091) and in the large traps ber of individuals caught/100 trapnights, or (n = 7,2_grids, respectively; 1 200, 18 trap- density either when quarters were analyzed nights, Xs = 2.0 (sd ± 1.0), 0.5 (sd ± 0.7) spe- separately or when quarters respectively cies, z = 1.869, P = 0.062). spanning the new and füll moons were com- Elevation: We obtained a marginally greater bined. number of species, percentage capture suc- Selective logging: Somewhat more indivi- cess and number of individuals/100 trap- dual mammals were captured in logged for- nights on the Valley floor than at higher ele- est than in unlogged forest using small vations on the slope above the Bladen River mammal traps (n = 11,15 grids, respectively; (n = 24,2 grids, respectively; 3 876, 360 trap- 956, 3 280 trapnights, n = 8 (sd±7), 4 nights, Xs = 2.0 (sd ± 1.3), 0.5 (sd ± 0.7) spe- (sd ± 4) per 100 trapnights, IVlann-Whitney cies, Mann-Whitney U test, z = 1.642, U test, z = 1,664, P = 0.096) but there were P = 0.1; Xs = 0.7% (sd ± 0.6), _ 0.3% no significant differences on the three other (sd±0.4), z = 1.832, P = 0.067; Xs = 6.1 measures. Restricting analyses to the cus- (sd±5.6), 0.3 (sd±0.4), z = 1.832, P = tom-made traps that were somewhat more 0.067). Two of these three results still held effective in catching small mammals, we after analyses were restricted to custom- © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammaLs in the Bladen Nature Reserve 97

made small mammal traps placed in un- tal period; Timm (1994 b) documented logged areas only (n = 11 low and 2 higher 50 species for La Selva in Costa Rica over elevation grids; 3 258, 360 trapnights, 20 years; and Glanz (1990) reported Xs = 2A (sd±1.6), 0.5 (sd + 0.7) species re- 39 species on Barro Colorado Island, Pana- spectively, z = 1.610, P = 0.107; Xs = 6.3% ma, which had been studied for 13 years at (sd±5.0), 0.3% (sd±0.4) respectively, the time. z = 1.781, P = 0.075; Xs = 4.7 (sd ± 3.5), 0.3/ Small mammal trapping success in Bladen 100 trapnights (sd ± 0.4) respectively, (6.5%) was comparable to that in other z = 1.781, P = 0.075). Tomahawk traps were neotropical wet such as Cockscomb. not placed above the valley floor. For example, trapping success in the Gi- Proximity to the river: Finally, we compared gante Peninsula, Panama was 4.2% for the traps Set within < 50 m of the river bank wet and 7.3% for the dry season with those placed further away (50-200 m). (McClearn et al. 1994). In contrast to mea-

Results showed that the number of species sures of mammal diversity, it is difficult to captured was marginally higher close to the make many direct comparisons of species' river than further away from it densities with other sites as data for many (n = 11,11 grids, _ respectively; 2 239, of the same species are unavailable. Didel- 1 349 trapnights, Xs = 2.4 (sd + 1.6), 1.2 phis and Dasypiis densities appeared low (sd ± 0.9) species, Mann-Whitney U test, compared to Barro Colorado Island and z = 1.793, P = 0.073) but there was no effect south American sites (Glanz 1990), on percentage trap success or individuals/ whereas Marmosa densities were higher 100 trapnights. When analyses were re- than either at Barro Colorado or even Gua- stricted to custom-made traps set on the topo, Venezuela (Eisenberg et al. 1979). Valley floor in unlogged areas, however, Oryzyomys densities were extremely high proximity to the river resulted in greater in comparison to Barro Colorado, Guatopo, numbers of mammals caught on all three Cosha Cashu and Cabassou, French Guiana measures (n = 8,3 grids next to and away (Charles-Dominique et al. 1981; Glanz from the river, respectively; 2 103, 414 trap- 1990) possibly because some traps were set nights, Xs = 2.9 (sd ± 1.5), 1.0 (sd ±1.0) spe- in selectively logged habitats. cies, z = 1.763, P = 0.078; Xs = 7.9% (sd±4.6), 1.9% (sd±2.8) trap success re- Factors affecting trapping success spectively, z = 2.041, P = 0.041; Xs = 5.7 (sd ± 3.2), 1.9 individuals/100 trapnights There is a substantial hterature on the ef- (sd ± 2.8) respectively, z = 1.837, P = 0.066). fects of trap type on trapping success in tem-

perate regions (e. g.. Sealander and James

1958; Slade et al. 1993). The few studies that Discussion have been conducted in the neotropics have compared live traps to snap traps and found the latter to catch more species and indivi- Captures duals (PizziMENTi 1979; Woodman et al. Bladen Nature Reserve and land immedi- 1996). The only studies to compare wire ately adjacent to it held a minimum of mesh and Sherman traps were conducted in

33 species of non-volant mammals which is temperate climates (Holdenreid 1954; comparable to other central American sites. O'Farrell et al. 1994). In both cases a great- Rabinowitz and Nottingham (1989) docu- er Proportion of captures was made using mented 39 species of non-volant mammals wire mesh traps and heteromyid in in the Cockscomb basin which is almost ad- particular were captured in mesh traps. jacent to the Bladen; Medelon (1994) re- O'Farrell et al. (1994) found that custom- ported 48 species in Selva Lacondona, made wire mesh traps captured two to three Chiapas, Mexico but these were compiled times more individuals than Sherman traps. Over 10 years as opposed to our 6 month to- Our results from the neotropics replicated © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

98 T. M. CAROetaL

these findings in that they showed margin- eles, principally marsupials, outside the re- ally greater percentage success and indivi- serve. There may be many reasons for these duals caught/100 trapnights in mesli traps. associations including a more abundant and Clearly, comparisons of small mammal den- predictable insect prey base (Malcolm sities in the neotropics must take into ac- 1995), increased seed abundance stemming count of whether traps are of mesh or box from increased vertical Vegetation density design. (MoNADJEM 1997), or even reduced threat

There is also a considerable literature on of predation (daFonesca and Robinson the effects of baits on trap success but again 1990) but these were not investigated. mostly from temperate regions (e. g., Beer Studies that have looked into the effects of

1964; Slade et al. 1993; but see Laurance elevation on small mammals have often

1992) and it is well known that rolled oats found different results. For example, abun- and peanut butter is a very effective bait in dance increased with elevation in a tropical capturing terrestrial mammals in this part in the Philippines but Speeles of the World. Although very few compari- richness did not change (Heaney et al. sons have been reported for the neotropics, 1989). In contrast, in a temperate rainforest WooDMAN et al. (1996) found no differences in Chile, number of Speeles, number of in- in captures using suet or peanut butter in dividuals, and Speeles diversity all declined tropical forest in south-eastern Peru. In with increased elevation (Patterson et al. contrast, we found that oats and molasses 1989). We found a marginally reduced caught more Speeles, generated greater trap number of Speeles, trapping success, and success, and captured more individuals/ number of individuals at higher elevations 100 trapnights than other baits, with peanut although trapping effort was relatively low butter producing poorest results. In addi- off the Valley floor. In addition, we caught tion, there are many other baits that include a somewhat greater number of Speeles, animal protein that we did not use. It there- had somewhat greater success, and cap- fore appears premature to suggest that bait tured somewhat more individuals per trap- has little influence on trapping in the neo- night near the river than farther from it. tropics. Taking these two results together, there ap- In contrast to many studies in deserts of peared to be a gradient of decreasing small

North America (e. g., Price et al. 1984), mammal diversity and abundance as one phase of the moon had no effect on mea- progressed away from the Bladen River sures of trapping in this study. Possibly, the and Up the slope. Whether these findings thick canopy obscured the moon to such an reflect differences in humidity, soll drai- extent that httle hght penetrated to the for- nage or type, or habitat structure remains est floor. unresolved.

It is well documented that small mammal Our findings are necessarily preliminary be- abundance is greater in selectively logged cause we chose to examine a large number habitats in temperate regions (e. g., Month- of factors which reduced our sample size. EY and Southiere 1995) as well as in the Nevertheless, they highlight the importance tropics (Delany 1971; Struhsaker 1997). of carefully selecting the type of trap and For example, Malcolm (1995) showed that type of bait in trapping studies of neotropi- terrestrial mammal abundance, richness cal mammals. They also point to the differ- and diversity were all greater in pasture ences that may be expected in estimating and young secondary forest than in continu- mammal abundance and diversity in areas ous forest north of Manaus. Our results are with different logging regimes, elevations consistent with his findings in that percen- and proximity to rivers in neotropical habi- tage success and number of individuals cap- tats. As such, they reinforce the necessity tured was greater in areas that had been of standardizing techniques when compar- logged outside the reserve than inside it. In ing Speeles abundance and diversity across addition, we caught a greater variety of Spe- neotropical sites. © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammaLs in the Bladen Nature Reserve 99

Acknowledgements San Francisco State University Work was con- ducted under UC Davis Animal Care and Use We thank The Forestry Department and Belize Protocols. We thank Deirdre Doherty, Dave - Audubon Society for permission to work in the Drouillard, Chris Gregory, Beth Johnstone, Bladen Nature Reserve and Raphael Manza- Tom McNamara, Jeremy Mull, Megan Temple- NERO and John Pinello for important assistance. TON, and Susan Woofsy for logistical help in the We thank Jake and Kelly Marlin for permission field; Ron Cole, Jake and Kelly Marlin, and to work in the Belize Foundation for Research Dirk Van Vuren for loaning and storing equip- and Environmental Education (BFREE). This ment; Steven Brewer, Doug Kelt, Marcel Rej- work was supported by two University of Califor- MANEk, and Dirk Van Vuren for discussions; and nia faculty research grants and a President's un- Sabine Schmid-Holmes for translating the sum- dergraduate fellowship as well as funding from mary into German.

Zusammenfassung

Diversität von Säugern im Bladen Naturreservat, Belize, und Faktoren, die einen Fangerfolg beeinflussen

Im BLaden Naturreservat einem subtropischen Feuchtwaldgebiet, wurde durch Zählungen an Tran- sekten und unter Einsatz von Sherman- und Tomahawk-LebendfaUen das Vorkommen von 33 Säuge- tierarten festgestellt. In insgesamt 6 075 Fallennächten wurden Einfluß von Fallendesign, Köder, Mondphase, Holzeinschlag, Höhenlage und Nähe eines Flusses auf Fangerfolg systematisch unter- sucht. Drahtgitterfallen hatten etwas größeren Fangerfolg als Sherman-Fallen, Haferflocken und Molasse erzielten größeren Fangerfolg als andere Köder, Fangerfolg in Wald mit selektivem Holz- einschlag war größer als in Wald ohne Einschlag und er war etwas größer in höher gelegenen Gebie- ten und näher an einem Fluß. Die Mondphase hatte keinen Einfluß auf Fangerfolg. Die Resultate lie- fern Basisdaten über die Säugetiervielfalt in einem relativ unerforschten Gebiet Zentralamerikas und geben an, wenn auch nur vorläufig, welche Aspekte im Fangdesign standardisiert werden soll- ten, um einen Vergleich der Artenvielfalt zwischen verschiedenen neotropischen Studiengebieten zu ermöglichen.

References

Alkon, P. U; Saltz, D. (1988): Influence of sea- Charles-Dominique, R; Atramentowicz, M.; son and moonhght on temporal activity pat- Charles-Dominique, M.; Gerard, H.; Hla- terns of Indian crested porcupines {Hystrix in- DiK, A.; Hladik, C. M.; Prevost, M. F. (1981):

dica Kerr). J. Mammalogy 69, 71-81. Les mammiferes frugivores arboricoles noc-

Asquith, N. M.; Wright, S. J.; Clauss, M. J. turnes d'une foret guyanaise: inter-relations (1997): Does mammal Community composition plantes-animaux. Revue Ecologie (Terre et control recruitment in neotropical forests? Vie) 35, 341-435.

Evidence from Panama. Ecology 78, 941-946. Clarke, J. A. (1983): Moonhght's influence on

Beer, J. R. (1964): Bait preferences of some small predator/prey interactions between short-

mammals. J. Mammalogy 45, 632-634. eared owls {Asio flammeus) and deermice Brokaw, N. V. L.; Lloyd-Evans, T. L. (1987): The {Peromyscus maniculatus). Behav. Ecol. and Bladen Brauch wilderness. Manomet Sociobiol. 13, 205-209.

Observatory, Mass. Delany, M. J. (1971): The biology of small ro-

Brov^n, J. S.; Kotler, B. P; Smith, R. J.; Wirtz, dents in Mayanja forest, Uganda. J. Zool.

II, W. O. (1988): The effects of owl predation (London) 165, 85-129. on the foraging behavior of heteromyid ro- DiRZO, R.: MiRANDA, A. (1991): Altered patterns dents. Oecologia 76, 408^15. of herbivory and diversity in the forest un-

Buchalczyk, T; Olszewski, J. L. (1971): Behav- derstory: a case study of the possible conse- ioural response of forest rodents against trap quences of contemporary defaunation. In: and bait. Acta Theriol. 18, 277-292. Plant-Animal Interactions: Evolutionary © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

100 T. M.CAROetal.

Ecology in Tropical and Temperate Regions. HoLDENREiD, R. (1954): A new live-catch

Ed. by P. W. Price, T. M. Lewinsohn, trap and comparison with two other traps. J. G. W. Fernandez, and W. W. Benson. New Mammalogy 35, 267-268.

York: John Wiley and Sons. Pp. 273-287. Isabirye-Basuta, G.; Kasenene, J. M. (1987):

Eisenberg, J. F.; O'Connell, M.; August, P. V. Small rodent populations in selectively felled (1979): Density, productivity, and distribution and mature tracts of Kibale forest, Uganda. of mammals in two Venezuelan habitats. In: Biotropica 19, 260-266. Vertebrate Ecology in the Northern Neotro- Janson, C. H.; Emmons, L. H. (1990): Ecological

pics. Ed. by J. F. Eisenberg. Washington, DC: structure of the nonflying mammal Community Smithsonian Institution Press. Pp. 187-207. at Cosha Cashu Biological Station, Manu Na- Emmons, L. H. (1984): Geographie Variation in tional Park, Peru. In: Four Neotropical Rain- densities and diversities of non-flying mam- forests. Ed. by A. W. Gentry. New Häven, mals in Amazonia. Biotropica 16, 210-222. Connecticut: Yale Univ Press. Pp. 314-338.

Emmons, L. H. (1987): Comparative feeding ecol- Kasenene, J. M. (1984): The influence of selective ogy of felids in a neotropical forest. Behav. logging on rodent populations and the regen- Ecol. and Sociobiol. 20, 271-283. eration of selected tree species in the Kibale Emmons, L. H. (1990): Neotropical Rainforest forest, Uganda. Trop. Ecol. 25, 179-195. Mammals: a Field Guide. Chicago: Univ. Chi- Laurance, W. f. (1992): Abundance estimates of cago Press. small mammals in Australian tropical rainfor- EstradA, A.; Coates-Estrada, R.; Meritt Jr, D. est: a comparison of four trapping methods. (1994): Non Aying mammals and landscape Wildlife Res. 19, 651-655.

changes in the tropical rain forest region of Los Lynam, A. J. (1997): Rapid decline of small mam- Tuxlas, Mexico. Ecogeography 17, 229-241. mal diversity in monsoon evergreen forest Fenton, M. B.; Boyle, N. G. H.; Harrison, T. M.; fragments in Thailand. In: Tropical Forest

Oxley, D. J. (1977): Activity patterns, habitat Remnants: Ecology, Management, and Conser- use, and prey selection by some African insec- vation of Fragmented Communities. Ed. by

tivorous bats. Biotropica 9, 73-85. W. M. Laurance and R. O. Bierregaard, Jr.

FoNESCA, G. A. B. da; Robinson, J. G. (1990): For- Chicago: Univ. Chicago Press. Pp. 222-240.

est size and structure: competitive and preda- Malcolm, J. R. (1990): Estimation of mammahan tory effects on small mammal communities. densities in continuous forest north of Man- Biol. Conserv 53, 265-294. aus. In: Four Neotropical . Ed. by Fleming, T. H. (1973): The number of rodent Spe- A. W. Gentry. New Häven, Connecticut: Yale

eles in two Costa Rican forests. J. Mammalogy Univ Press. Pp. 339-357.

54, 518-521. Malcolm, J. R. (1995): Forest structure and the Glanz, W. E. (1990): Neotropical mammal densi- abundance and diversity of neotropical small ties: how unusual is the Community on Barro mammals. In: Forest Canopies. Ed. by Colorado Island, Panama? In: Four Neotropi- M. D. LowMAN and N. M. Nadkarni. San cal Rainforests. Ed. by A. W. Gentry. New Diego: Academic Press. Pp. 179-197.

Häven, Connecticut: Yale Univ. Press. Pp. Malcolm, J. R. (1997): and diversity of 287-313. small mammals in Amazonian forest frag- Glanz, W. E. (1991): Mammahan densities at pro- ments. In: Tropical Forest Remnants: Ecology, tected versus hunted sites in central Panama. Management, and Conservation of Fragmen- In: Neotropical Wildlife Use and Conservation. ted Communities. Ed. by W. M. Laurance Ed. by J. G. Robinson and K. H. Redford. and R. O. Bierregaard, Jr. Chicago: Univ. Chicago: Univ. Chicago Press. Pp. 163-173. Chicago Press. Pp. 207-221. Hartshorn, G. S.; Nicolait, L.; Hartshorn, L.; Marls, M. A. (1992): Neotropical mammals and Bevier, G; Brightman, R.; Cal, J.; Ca- the myth of Amazonian biodiversity. Science witch, A.; Davidson, W; Dubois, R.; 255, 976-979. Dyer, C; Gibson, I; Hawley, W; Leo- McClearn, D.; Kohler, J.; McGowan, K. J.; Ce- nard, J.; Nicolait, R.; Weyer, D.; White, H.; deno, E.; Carbone, L. G; Miller, D. (1994): White, C. (1984): Belize: Country Environ- Arboreal and terrestrial mammal trapping on ment Profile. A Field Study. Belize City, Gigante peninsula, Barro Colorado Nature Belize: Robert Nicolait and Associates. Monument, Panama. Biotropica 26, 208-213. Heaney, L. R.; Heideman, P. D.; Rickart, E. A.; Medellin, R. A. (1994): Mammal diversity and Utzurrum, R. B.; Klompen, J. S. H. (1989): Elevational zonation of mammals in the cen- conservation in the Selva Lacondona, Chia-

tral Philippines. J. Trop. Ecol. 5, 259-280. pas, Mexico. Conserv Biol. 8, 780-799. © Biodiversity Heritage Library, http://www.biodiversitylibrary.org/

Diversity of mammals in the BLaden Nature Reserve 101

MoNADJEM, A. (1997): Habitat preferences and nants: Ecology, Management and Conserva- of small mammals in Swaziland. tion of Fragmented Communities. Ed. by

Afr. J. Ecol. 35, 64-72. W. F. Laurance and R. O. Bierregaard, Jr. MoNTHEY, R. W.; SouTHiERE, E. C. (1995): Re- Chicago: Univ Chicago Press. Pp. 256-274. sponses of small mammals to forest harvesting The Nature Conservancy (1993): Rapid ecologi- in northern Maine. Canad. Field Nat. 99, 13-18. cal assessment of the Bladen Nature Reserve, O'CoNNELL, M. A. (1989): Population dynamics October 1993. of neotropical small mammals in seasonal ha- Timm, R. M. (1994 a): The mammal fauna. In: La

bitats. J. Mammalogy 70, 532-548. Selva: Ecology and Natural History of a Neo- OTarrell, M. I; Clark, W. A.; Emmerson, F. H.; tropical Rainforest Ed. by L. A. McDade, JuAREZ, S. M.; Kay, F. R.; O'Farrell, T. M.; K. S. Bawa, H. A. Hespenheide, and Goodlett, T. Y. (1994): Use of a mesh live G. S. Hartshorn. Chicago: Univ. Chicago trap for small mammals: are results from Press. Pp. 229-237.

Sherman live traps deceptive? J. Mammalogy Timm, R. M. (1994 b): Mammals. In: La Selva: 75, 692-699. Ecology and Natural History of a Neotropical

Patterson, B. D.; Meserve, P. L.; Lang, B. K. Rainforest. Ed. by L. A. McDade, K. S. Bawa, (1989): Distribution and abundance of small H. A. Hespenheide, and G. S. Hartshorn. mammals along an elevational transect in Chicago: Univ Chicago Press. Pp. 394-398.

temperate rainforests of Chile. J. Mammalogy Torres-Contreras, H.; Silva-Aranguiz, E.; Mar- 70, 67-78. QUET, P A.; Camus, P A.; Jaksic, F M. (1997):

Pizzimenti, J. J. (1979): The relative effectiveness Spatiotemporal variability of rodent subpopu-

of three types of traps for small mammals in lations at a semiarid neotropical locahty. J. some Peruvian rodent communities. Acta Mammalogy 78, 505-513.

Theriol. 24, 351-361. Voss, R. S.; Emmons, L. H. (1996): Mammalian di- Price, M. V.; Waser, N. M.; Bass, T. A. (1984): Ef- versity in neotropical lowland rainforests: a fects of moonlight on microhabitat use by de- preliminary assessment. Bull. Amer. Mus.

sert rodents. J. Mammalogy 65, 353-356. Nat. Hist. 230, 1-115.

Rabinowitz, A.; Nottingham, Jr, B. G. (1989): Wilson, D.E.; Cole,F. R.; Nichols, J. D.; Ru- Mammal Speeles richness and relative abun- DRAN, R.; FosTER, M. S. (1996): Measuring dance of small mammals in a subtropical wet and Monitoring Biological Diversity: Stan- forest of Central America. Mammalia 53, dard Methods for Mammals. Washington DC: 217-226. Smithsonian Institution Press. Rickart, E. A.; Heaney, L. R.; Utzurrum, R. C. B. WooDMAN, N; Timm, R. M.; Slade, N. A.; Doo-

(1991): Distribution and ecology of small NAN, T. J. (1996): Comparison of traps and mammals along an elevational transect in baits for censusing small mammals in neotro-

southeastern Luzon, Phillipines. J. Mammal- pical lowlands. I Mammalogy 77, 274-281.

ogy 72, 458-469. Wright, S. J.; Gompper, M. E.; DeLeon, B. (1994):

Sealander, J. A.; James, D. (1958): Relative effi- Are large predators keystone Speeles in neo-

ciency of different small mammal traps. J. tropical forests? The evidence from Barro Mammalogy 39, 215-233. Colorado Island. Oikos 71, 279-294.

Slade, N. A.; Eitler, M. A.; Gruenhagen, N. M.; Wright, S. J.; Zeballos, H.; Dominguez, L; Gal- Deavelos, A. L. (1993): Differential effec- LARDO, M. M.; MoRENO, M. C; Ibanez, R. tiveness of Standard and long Sherman live- (2000): Poachers alter mammal abundance,

traps in capturing small mammals. J. Mam- seed dispersal, and seed predation in a neo- malogy 74, 156-161. tropical forest. Conserv. Biol. 14, 227-239. Struhsaker, T. T. (1997): Ecology of an African rainforest: logging in Kibale and the conflict Authors' address: between conservation and exploitation. Gai- T. M. Caro, Rachel Brock, and Marcella nesville: Univ. Florida Press. Kelly, Department of Wildlife, Fish and Conser- Terborgh, I; Lopez, L.; Tello, I; Yu, D.; Bru- vation Biology, 1 Shields Avenue, University of Ni, A. R. (1997): Transitory states in relaxing California, Davis, CA 95616, USA landbridge Islands. In: Tropical Forest Rem- (e-mail: [email protected])