Importance of riparian habitats for small mammal and herpetofaunal communities in agricultural landscapes of southern Québec

Agriculture, Ecosystems and Environment 83 (2001) 165–175

Importance of riparian habitats for small mammal andherpetofaunal communities in agricultural

landscapes of southern Québec

Charles Maisonneuvea,∗, Stéphanie Riouxba Société de la Faune et des Parcs du Québec, Direction de la faune et des habitats, 675 René-Lévesque Est, Québec, Que., Canada G1R 5V7

b Société de la Faune et des Parcs du Québec, Direction régionale du Bas Saint-Laurent, 506 Lafontaine, PO Box 445,Rivière-du-Loup, Québec, Que., Canada G5R 3C4

Received 14 December 1999; received in revised form 18 July 2000; accepted 4 August 2000

Abstract

The presence of adequate riparian strips in agricultural landscapes is generally recognized to contribute to the reduction ofthe impacts of agricultural practices on the water quality of streams, to regularize water temperature and to help in the creationof important wildlife habitats. This study aimed at determining the importance of riparian strips in agricultural landscapes ofsouthern Québec for small mammal and herpetofaunal communities, and verifying farmers’ general belief that these habitatsbecome shelters for species considered as agricultural pests. Abundance, composition and diversity of communities werecompared between three types of riparian strips: herbaceous, shrubby and wooded. A total of 1460 small mammals belongingto 14 species and 329 amphibians and reptiles belonging to 11 species were captured with line trapping and drift fences. Thegeneralist speciesSorex cinereus, Zapus hudsonius, Blarina brevicauda, andBufo americanuswere abundant in all three typesof riparian strips.Peromyscus maniculatus, Sorex fumeus, Clethrionomys gapperi, andRana pipienswere associated moreclosely to wooded strips, whereasRana sylvaticawas captured mostly in shrubby strips. The abundance of small mammalsand herpetofauna increased with complexity of vegetation structure. Small mammal diversity was higher in herbaceous andwooded riparian strips, whereas the herpetofaunal community was more diverse in shrubby strips. Proportion and abundanceof pest species diminished with complexity of vegetation structure, whereas insectivores increased in abundance. Maintainingwoody vegetation in riparian strips should increase abundance and diversity of wildlife within agricultural landscapes whereincreasing development pressure is presently contributing to the conversion of such habitats to herbaceous strips. Such amanagement approach should also help reducing the risk of riparian strips becoming shelters for pest species. © 2001 ElsevierScience B.V. All rights reserved.

Keywords:Agricultural landscapes; Herpetofauna; Riparian strips; Small mammals; Québec

∗ Corresponding author. Tel.:+1-418-623-1650;fax: +1-418-623-0420.E-mail address:c [email protected] (C. Maisonneuve).

1. Introduction

Over the last few decades in most agriculturalregions of the province of Québec, there was a tran-sition from local, family subsistence farming to moreindustrial farming practices oriented towards regionaland national markets. The traditional dairy-oriented

0167-8809/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.PII: S0167-8809(00)00259-0

166 C. Maisonneuve, S. Rioux / Agriculture, Ecosystems and Environment 83 (2001) 165–175

agriculture that was previously characterized by a mo-saic of pastures, prairies and woodlots was replacedby a more specialized agriculture aimed at large-scaleproduction, with corn as (Zea maysL.) the dominantcrop in the south-west. The new agricultural practicesdeveloped for this large-scale production has led tothe expansion of cultivated areas, thus exerting an in-creasing pressure on uncultivated portions of the land.Forests in agricultural landscapes have been reducedand fragmented in numerous small woodlots, and ri-parian strips which often represent the only remainingcorridors for wildlife between these woodlots are alsobeing threatened.

Riparian strips are not only recognized as impor-tant wildlife habitats for a great diversity of species(Thomas et al., 1979; Small and Johnson, 1986; De-camps et al., 1987; Naiman et al., 1993), but they alsohelp reduce the impacts of agricultural practices on thewater quality of streams by controlling bank erosion,and by filtering fertilizers, pesticides and sedimentsfrom adjacent crops (Lowrance et al., 1985, 1986;Brenner et al., 1991; Gilliam, 1994; Vought et al.,1994) and they maintain quality of aquatic habitats byregularizing water temperature (Karr and Schlosser,1978). Thus, efforts to integrate the conservationof riparian strips in the management of agriculturallands should lead to both sound agroecosystems andenvironments.

However, many factors contribute to limit thevalue of riparian strips in agricultural landscapes ofQuébec. Under the provincial Protection Policy forLakeshores, Riverbanks, Littoral zones and Flood-plains, a buffer strip of 3 m is required for the pro-tection of riparian areas in agricultural landscapes,whereas 10–20 m are required in urban or forestedlandscapes. Moreover, the vegetation in these narrowriparian strips is often maintained at the herbaceousstage through mowing or burning. This practice stemsfrom the farmers’ belief that such management re-duces the risk of riparian strips becoming sheltersfor pest species (birds, rodents, weeds, insects). Ina recent study made for the Union des producteursagricoles (Lamarre et al., 1993), pesticide use forweed control was even recommended within riparianstrips. This perception of riparian habitats repre-sents a major obstacle for efforts to integrate wildlifehabitat needs in the management of agriculturallandscapes.

The objectives of this study were thus to determinethe importance of riparian strips for small mammal andherpetofaunal (amphibians and reptiles) communitiesin agricultural landscapes of southern Québec, and toverify if there is a basis to farmers’ belief concerningthe risk of riparian strips being used as shelters byrodent pest species.

2. Material and methods

2.1. Study area

The study was carried out in the Boyer River water-shed, located on the south shore of the St. LawrenceRiver near Québec City. Agriculture is the predom-inant land use in this watershed, covering more than60% of its 217 km2. Most of the agricultural land isdevoted to hay (43%) and cereal production (14%).The remaining area of the watershed is covered bywoodlots (28%), peatlands (4%), and edge-transitionhabitats (2%; abandoned farmland, riparian strips andhedgerows). The watershed comprises a total lengthof 345 km of streams, including ditches created to im-prove drainage. Most of the streams circulate in agri-cultural lands; 67% have agricultural fields on both oftheir banks, 25% woodlots on both banks, and the re-maining 8% both a wooded and an agricultural bank.

Most herbaceous and shrubby riparian strips arelocated in areas where topography permits cultiva-tion right up to the rim of the streambanks. Meanwidth of the strips is close to the minimum of 3 mrequired by the provincial policy: 3.2 ± 1.0 m forshrubby strips, and 3.7± 1.0 m for herbaceous strips.The wooded strips are located in stream sectionswhere the steepness of the banks impedes cultivation,and are 19.2 ± 14.0 m wide. Herbaceous riparianstrips are mostly covered by gramineous plants likePhalaris arundinacea, Bromus inermis, andCalama-grostis canadensis, and forbs likeImpatiens capensis,Eupatorium maculatum, Prunella vulgaris, andFra-garia virginiana. Shrubby riparian strips (<3 m high)are dominated byPrunus virginiana, Spirea latifo-lia, Rubus idaeus, Alnus rugosa, Crataegusspp., andCornus stolonifera. The most abundant tree species(>3 m) in wooded riparian strips areAcer negundo,Salixspp.,Fraxinusspp.,Acer saccharum, P. virgini-ana, andPopulus tremuloides.

C. Maisonneuve, S. Rioux / Agriculture, Ecosystems and Environment 83 (2001) 165–175 167

2.2. Field methods

Two methods were used to trap small mammals.The first one consisted of lines of traps installed par-allel to the streams. A total of 18 sites were selectedto cover a total of 3600 m in each of the habitat types.Each line had a length of 600 m. Museum specialsnap traps, Sherman live traps, and pitfall traps (2 l)were placed alternately every 10 m. Thus, each linecomprised 20 of each of these traps. Snap traps werebaited with peanut butter, and live traps with ap-ple pieces dipped in peanut butter. Pitfall traps werenot baited, but filled with enough water to rapidlydrown mammals. Half of these lines were operatedfor five consecutive nights during September 1995,the other half during September 1996. Total trappingeffort was thus 1800 night-traps in each of the habitattypes.

Drift fences (Corn, 1994; Kirkland and Sheppard,1994) were used as a second trapping method tosample amphibians and reptiles, and to make ad-ditional small mammal captures. Drift fences wereinstalled on the same 18 sites (six by habitat type)where line trapping had been carried out in the previ-ous year. Each of the arrays consisted of jute fences45 cm high and 30 m long installed parallel to thestreams. Four 25 l pitfall traps were installed flushto the ground at every 10 m and with enough waterto immediately drown mammal species. Six funneltraps were also placed alongside the fences, on bothsides, one set between each pitfall traps. Each of thearrays was operated for a total of 22 nights. In 1996and 1997, sampling was carried out during 14 nightsin May, four nights in June–July, and four nightsin September. A trapping effort of 1320 night-trapswas thus carried out in each of the three habitattypes.

Except for easily identified species (squirrels, chip-munks, weasels) captured in live traps, all mammalspecimens were sacrificed and kept frozen until lateridentification with the use of cranial and dental charac-ters. Amphibians were identified in the field, markedby cutting a toe in order to consider recaptures in eval-uation of abundance, and released.

Land use adjacent to each trapping station along linetransects was noted. These could be grouped into thefollowing four categories: cereals, pastures, prairiesand fallow lands.

2.3. Statistical analyses

Shannon’s index (Zar, 1984) was used to measurediversity within each of the three riparian habitats

H = n logn − ∑fi logfi

n(1)

wheren is the total number of individuals capturedfor all detected species combined, andfi the numberof captures for speciesi.

Hutcheson’s test (1970) was used to compare diver-sity indices between habitat types

t = H1 − H2

(S2H1

+ S2H2

)1/2(2)

whereS2H is the variance of the diversity index ob-

tained as follows:

S2H =

∑fi logfi − (

∑fi logfi)

2/n

n2(3)

Since habitat types with similar diversity indicesmay be inhabited by different communities, an overlapindex (Horn, 1966) was calculated

O =

∑(fi + gi) log(fi + gi)

− ∑fi logfi − ∑

gi loggi

(n1 + n2) log(n1 + n2)

−n1 logn1 − n2 logn2 (4)

wherefi is the number of captures of speciesi, gi thenumber of captures of speciesj, n1 the total numberof captures in habitat 1, andn2 the total number ofcaptures in habitat 2. This index varies from 0, whentwo communities have no species in common, to amaximum of 1 when all species and relative abundanceare the same in both habitats compared.

The reciprocal of Simpson’s index was used todetermine niche breadth (Levins, 1968; Colwell andFutuyma, 1971; Whittaker and Levin, 1975; Brownand Parker, 1982) for each species

W = 1∑

p2ij

(5)

wherepij is the occurrence rate of speciesi in habitatj. This rate is obtained as follows:

pij = Oij∑

Oij(6)


whereOij is the number of captures of speciesi inhabitatj. Since three habitat types were compared, aniche breadth value of 1 indicates that a species isonly present in one habitat type, whereas a maximumvalue of 3 indicates that a species is distributed evenlyin all three habitat types. This value of niche breadthcan be used as a tolerance index to habitat modifica-tions; species with great niche breadths are consideredtolerant and to modifications of their habitat and viceversa (Best et al., 1979; Stauffer and Best, 1980).

Proportions of insectivores and rodents within eachhabitat type were compared withG-tests (Scherrer,1984: 484). When this test indicated heterogeneitybetween the three habitat types, multiple comparisontests (Scherrer, 1984: 488) were carried out to de-termine to which habitat this was due. TheG-testswere also used to compare proportions of pest speciesamong the small mammal communities. Two specieswere considered as potential pests:Microtus pennsyl-vanicus, andMus musculus.

Comparison of observed numbers of individualspecies with numbers expected according to adjacentland use availability was carried out withG-tests.This test was carried out for species for which at least20 captures were obtained in the trap lines.Microtuspennsylvanicusand M. musculuswere grouped aspests species for this analysis.

Table 1Mean, standard deviation (S.D.) and total numbers of individuals of each species of small mammals captured within each of three riparianhabitat types and number of sites on which they were detected in agricultural landscapes of southern Quebec, 1995–1997

Species Habitat

Herbaceous Shrubby Wooded

Mean S.D. Total Sites Mean S.D. Total Sites Mean S.D. Total Sites

S. cinereus 15.5 6.6 92 6 27.7 7.6 166 6 26.3 5.9 158 6S. fumeus 0.3 0.5 2 2 0.8 1.5 5 2 4.0 4.2 24 6Sorex hoyi 0.2 0.4 1 1 0.5 0.8 3 2 0.2 0.4 1 1B. brevicauda 6.3 2.6 42 6 9.5 4.2 57 6 8.8 3.9 53 6Condylura cristata 0.7 0.8 4 3 0 0 0.5 0.8 3 2M. pennsylvanicus 4.5 4.0 27 5 3.5 1.3 21 6 2.2 2.6 13 3C. gapperi 2.2 2.9 13 3 2.3 1.7 14 4 6.2 5.7 37 6M. musculus 1.2 1.3 7 3 0.3 0.5 2 2 0.8 0.7 5 4P. maniculatus 0 0 0.5 1.1 3 1 10.8 8.6 65 6Napaeozapus insignis 0 0 0.2 0.4 1 1 0.5 0.8 3 2Z. hudsonius 25.3 7.1 152 6 35.5 14.8 213 6 41.8 13.3 248 6Tamiasciurus hudsonicus 0.8 1.5 1 1 0.2 0.4 1 1 1.0 0.8 6 4Tamias striatus 0 0 0 0 0.3 0.5 1 1Mustela herminea 0.5 0.8 3 2 1.7 1.5 10 4 0.3 0.5 2 2

Total 344 496 620

3. Results

3.1. Small mammals

A total of 1460 individual small mammals belong-ing to 14 species were captured (Table 1). Total num-ber of captures increased gradually with increasingcomplexity of vegetation structure: 23.6% of indi-viduals were captured in herbaceous riparian strips,34.0% in shrubby strips, and 42.5% in wooded strips(Table 2). All 14 species were detected in woodedstrips, 11 species in herbaceous strips and 12 inshrubby strips. Diversity was significantly lower inshrubby strips than in wooded or herbaceous strips,which had similar diversity indices. Relatively highvalues were obtained for overlap indices betweenhabitat types: 0.976 between herbaceous and shrubbystrips, 0.918 between herbaceous and wooded strips,and 0.931 between shrubby and wooded strips, indi-cating that small mammal communities differed muchbetween herbaceous and wooded strips.

The most abundant species wereZapus hudsonius,Sorex cinereus, and Blarina brevicaudawith 42.1,28.5 and 10.4% of the captures, respectively (Table 1).These species were captured on all the sites and hadhighest niche breadth values (2.89, 2.84 and 2.95,respectively).M. pennsylvanicushad a niche breadth


Table 2Number of individuals, number of species, and diversity indices of small mammals caught in three riparian habitat types in agriculturallandscapes of southern Quebec, 1995–1997

Habitat


Total number of individuals 344 496 620Mean number of individuals/site 57.3 82.7 103.3Minimum number of individuals/site 36 73 75Maximum number of individuals/site 76 105 115Total number of species 11 12 14Mean number of species/site 6.5 6.8 9.3Minimum number of species/site 4 6 7Maximum number of species/site 9 9 12Shannon’s diversity indexa 0.665A 0.613B 0.744A

a Indices followed by identical letters did not differ significantly (Hutcheson’s test).

value of 2.78 and, contrarily to the three precedingspecies whose abundance increased with complexityof vegetation structure, occurred mainly (44.3%) inherbaceous strips. A niche breadth value of 2.36 wasobtained forClethrionomys gapperi. The species wasdetected on all six sites covered in wooded strips,where 58% of all individuals captured. Over 95% ofall Peromyscus maniculatuswere captured in woodedstrips, which led to a very low niche breadth value(1.09) for this species. A relatively low niche breadthvalue (1.59) was also obtained forS. fumeuswith 75%of all captures made in shrubby strips. Other speciesrepresented less than 1% of captures.

Proportions of rodents and insectivores differed sig-nificantly between the three habitat types (G = 9.58,d.f . = 2, P < 0.01). A multiple comparisons testindicated that rodents were significantly more abun-dant in wooded strips (61%) than in shrubby strips(52%), but failed to detect any difference in the propor-tion of rodents between herbaceous strips (59%) and

Table 3Observed and expected numbers of individual species according to land use adjacent to trapping stations in riparian strips located inagricultural landscapes of southern Quebec, 1995–1997a

Species Cereal Pastures Prairies G-test

Observed Expected Observed Expected Observed Expected d.f. Value P

S. cinereus 39 45 72 52 177 190 2 4.128 0.127B. brevicauda 5 14 27 16 54 57 2 7.360 0.025P. maniculatus 3 7 12 8 29 29 2 2.451 0.294Z. hudsonius 25 41 60 47 173 171 2 5.511 0.064Pest speciesb 10 8 10 9 29 32 2 0.423 0.809

a Fisher’s exact test.b Mus musculusand Microtus pennsylvanicuscombined.

the two other habitat types. Even though proportionsof rodents were similar in herbaceous and woodedstrips, species composition differed.M. pennsylvani-cus and M. musculuswere more abundant in herba-ceous strips,P. maniculatusandC. gapperibeing es-sentially present in wooded strips.

The proportion of pest species in the three habitattypes varied significantly (G = 21.73, d.f . = 2, P <

0.001). A multiple comparisons test indicated that thiswas due to a significantly higher proportion of pestspecies in herbaceous strips (10.0%) than in shrubby(4.7%) or wooded (2.8%) strips. Even if the total num-ber of captures increased from herbaceous to woodedstrips, the abundance of pest species decreased withcomplexity of vegetation structure.

With the exception ofB. brevicauda, which wasalmost three times less abundant than expected atstations adjacent to cereals, land use had no ef-fect on species distribution within the buffer strips(Table 3).


Table 4Mean, standard deviation (S.D.), and total numbers of individuals of each species of herpetofauna captured in each of the three riparianhabitat types and number of sites in which they were detected in agricultural landscapes of southern Quebec, 1995–1997

Species Habitat


Mean S.D. Total Sites Mean S.D. Total Sites Mean S.D. Total Sites

Ambystoma maculatum 0 0 0.3 0.8 2 1 0 0Eurycea bislineata 0.3 0.7 2 1 0.2 0.4 1 1 0.3 0.5 2 2B. americanus 2.8 2.9 17 4 6.7 7.6 40 6 16.7 9.5 100 6Hyla cricifer 0.2 0.4 1 1 0.7 1.1 4 2 0 0Rana catesbeiana 0 0 0.2 0.4 1 1 0.2 0.4 1 1Rana septentrionalis 0.5 0.8 3 2 1.3 1.5 8 3 1.0 1.2 6 3Rana clamitans 0.5 0.8 3 2 0.2 0.4 1 1 0.3 0.5 1 1R. sylvatica 0.2 0.4 1 1 3.8 4.9 23 4 1.0 1.0 6 4R. pipiens 1.2 2.6 7 1 0.8 1.1 5 3 14.7 11.4 88 6Thamnophis sirtalis 0 0 0.5 0.8 3 2 0.2 0.4 1 1Storeria occipitomaculata 0.2 0.4 1 1 0.2 0.4 1 1 0 0

Total 35 89 205

3.2. Herpetofauna

A total of 329 individual amphibians and reptiles be-longing to 11 species were captured (Table 4), i.e., nineamphibians with more than 98% of all captures, andtwo reptiles. All occurred in shrubby strips, whereaseight species were detected in the other habitat types(Table 5). Diversity was significantly lower in herba-ceous and wooded strips than in shrubby strips. Rel-atively high values were obtained for overlap indicesbetween herpetofaunal communities of the differenthabitat types: 0.847 for herbaceous and shrubby strips,

Table 5Number of individuals, number of species, and diversity indices of herpetofauna caught in three riparian habitat types in agriculturallandscapes of southern Quebec, 1995–1997


Total number of individuals 35 89 205Mean number of individuals/site 5.8 14.8 34.2Minimum number of individuals/site 0 4 7Maximum number of individuals/site 12 36 55Total number of species 8 11 8Mean number of species/site 2.2 4.2 4.0Minimum number of species/site 0 2 3Maximum number of species/site 5 7 5Shannon’s diversity indexa 0.688A 0.715B 0.456A

a Indices followed by identical letters did not differ significantly (Hutcheson’s test).

0.862 for herbaceous and wooded strips, and 0.760 forshrubby and wooded strips.

Bufo americanus, Rana pipiensandRana sylvaticawere the dominant species with 47.7, 30.4 and 9.1%of captures, respectively (Table 4). The majority ofcaptures ofB. americanus(89%) andR. sylvatica(97%) were made in wooded and shrubby strips,giving niche breadth values of 2.10 and 1.67, respec-tively. In R. pipiens, 88% of captures were made inwooded strips, for a niche breadth value of 1.29.Ranaclamitans, with 5.2% of all captures, was encounteredmostly in shrubby and wooded strips. The other seven


species represented less than 2% of captures and nichebreadth values were not calculated.

4. Discussion

4.1. Small mammals

Abundance and richness of small mammal speciesclearly increased with increasing complexity of thevegetation structure in riparian strips. Similar resultswere obtained in shelterbelts by Yahner (1983) andDambach (1948), who contended that linear habitatswith all the vegetation strata shelter a greater abun-dance and diversity of small mammals. Vertical strat-ification of the vegetation reduces predation risk byproviding a better cover (Stamp and Ohmart, 1978).

Although similar small mammal diversities wereobserved in herbaceous and wooded riparian strips, theoverlap index of small mammal communities in thesehabitat types was lowest, indicative of differences incomposition. These differences were ascribed to thehigh presence of some specialist species in woodedstrips (P. maniculatus, Sorex fumeus) or low abun-dance in herbaceous strips (C. gapperi). Contradictoryresults obtained from other studies onP. maniculatuswere due to the existence of subspecies with extremehabitat requirements.P. maniculatus bairdiiis associ-ated with cultivated fields and prairies in the UnitedStates (Hooper, 1942; Hansen and Warnock, 1978;Fleharty and Navo, 1983), where it is considered as apest (Linduska, 1949), whereasP. maniculatus gracilisandP. maniculatus abietorummostly occupy forestedhabitats in the Maritimes and up to the St. LawrenceRiver (Hooper, 1942; Iverson et al., 1967; Banfield,1974).S. fumeusis considered more habitat selectivethan S. cinereusand is more abundant in humid de-ciduous or mixed forests with soils covered with deephumus (Hamilton, 1940; Wrigley, 1969), in agreementwith the present results.C. gapperiis generally con-sidered a specialist species characteristic of woodedhabitats (Iverson et al., 1967; Miller and Getz, 1977;Yahner, 1983); it has a relatively high water turn-overrate and poorly efficient kidneys, impeding its survivalin low humidity (Getz, 1968). It has been shown thatPeromyscus leucopus, another forest dwelling species(Getz, 1961b; Kaufman and Fleharty, 1974; Snyderand Best, 1988), can use riparian herbaceous strips as

corridors between woodlots (Cummings and Vessey,1994).

Generally considered a typical species of prairiesand other open grass-dominated habitats (Iversonet al., 1967; Grant, 1971; Morse, 1973; Yahner, 1983;Manson et al., 1999),M. pennsylvanicuswas moreabundant in herbaceous riparian strips, but many in-dividuals were also caught in shrubby and woodedstrips, leading to the fourth highest niche breadthvalue obtained. Some studies foundM. pennsylvan-icus in forested habitats (Dowler et al., 1985), andothers attributed its presence to irregular incursionsin times of high population densities and suggestedthat it avoids forest habitats (Grant, 1971; Tamarinet al., 1984). In shelterbelts, the species is presentin openings and avoids the more wooded sections(Yahner, 1982). The shrubby and wooded riparianstrips studied here were generally regularly brokenby small openings, which probably was the reasonfor the presence ofM. pennsylvanicus.

Linduska (1949) suggested thatM. musculus, con-sidered a pest species, rarely inhabits linear herba-ceous habitats during the summer, but occupies thesehabitats after adjacent fields have been harvested. Agood part of the present sampling was carried out afterharvest, and only 14 individuals were captured, halfof these in herbaceous strips, indicating a probableselection for this habitat type.M. musculussometimesprefers linear marginal habitats to cultivated prairies,hayfields and corn fields (Clark et al., 1996; Kirsch,1997). The low abundance of this species in thisstudy could reflect its distribution in the agriculturallandscape, in agreement with Kaufman and Kaufman(1990) who suggested thatM. musculusis rarely, ifever, abundant in prairies and in cultivated fields.Linear habitats such as riparian strips could simplybe corridors between farm buildings (Merriam, 1988;Clark et al., 1996).

Z. hudsonius, S. cinereus, andB. brevicaudawerethe most abundant species and were present on all thestudy sites. The greatest niche breadth values werealso obtained for these three species. Many other stud-ies confirmed that these species are generalists in theirhabitat selection (Quimby, 1951; Iverson et al., 1967;Getz, 1961a,b; Rickard, 1960; Brown, 1967; Dueserand Shugart, 1978; Wrigley et al., 1979; Yahner, 1982;Adler, 1985; Dowler et al., 1985; Snyder and Best,1988). Humidity seems to have a greater influence than


vegetation on habitat selection byZ. hudsoniusandB.brevicauda(Quimby, 1951; Getz, 1961a; Choate andFleharty, 1973), and the proximity of streams couldexplain their preponderance in this study. In spite ofits generalist nature,B. brevicaudawas less abun-dant than expected in riparian strips adjacent to cerealfields. Although this species may avoid corn fields,this avoidance does not reduce its use of linear habi-tats adjacent to corn fields (Kirsch, 1997). Low avail-ability of prey species within buffer strips adjacentto cereal fields could probably have influenced ourresults.

4.2. Herpetofauna

As for small mammals, the abundance of the her-petofauna increased with increasing complexity invegetation structure, and the diversity of herpetofaunalcommunities of herbaceous and wooded strips weresimilar. However, contrarily to what was obtained forsmall mammals, richness and diversity were greaterin shrubby riparian strips. The lowest overlap indexobtained between riparian strip-types was betweenshrubby and wooded riparian strips. This differencewas mostly due to the presence of species in shrubbystrips which were absent from wooded strips (Am-bystoma maculatum, Hyla crucifer, Storeria occipit-omaculata) and to a greater abundance ofR. pipiensin wooded strips.

Only three out of 11 species were caught on morethan 20 occasions. This low sample size reduced thepossibility of any conclusion in terms of habitat se-lection. B. americanusclearly was most generalist,having the highest niche breadth value and beingdetected in all sites but two located in herbaceousriparian strips, in agreement with previous studies(Leclair, 1985; Dubé, 1994; deMaynadier and Hunter,1998).

R. pipiensis generally considered a species of openhabitats, fields and prairies (Dole, 1965; Cook, 1984),vegetation structure being the main factor for habi-tat selection with a preference for good gramineouscover (Beauregard and Leclair, 1988), no shrubs andreduced litter (Dubé, 1994). The majority (88%) of theindividuals were caught in wooded riparian strips, giv-ing this species a relatively low niche breadth value.The wooded riparian strips in this study often had arich herbaceous cover without any shrub layer, which

would provide adequate habitat forR. pipiens. It is,however, more difficult to interpret the absence of thisspecies in herbaceous riparian strips.

More than 75% ofR. sylvaticawere captured inshrubby riparian strips, leading to a relatively lowniche breadth value. This species is generally asso-ciated with wooded habitats (Heatwole, 1961; Bellis,1962) and is considered intolerant to habitat modi-fications (deMaynadier and Hunter, 1998). Air andsubstrate moisture are important factors in the pro-cess of habitat selection (Marshall and Buell, 1955;Roberts and Lewin, 1979), the dense vegetation inthe shrubby riparian strips studied probably explainedthe preponderance ofR. sylvaticathere.

5. Conclusions

This study clearly shows the importance ofmaintaining a diversity of riparian strips in order tomaintain high wildlife diversity within agriculturallandscapes. Shrubby riparian strips had a higher di-versity of herpetofauna, whereas a higher diversity ofsmall mammals was found in herbaceous and woodedriparian strips. Even though diversity of small mam-mals was similar in these last two habitats, smallmammal communities differed in composition andrelative abundance. Moreover, species considered ashabitat specialists and intolerant to habitat modifica-tions were present in all of the habitats studied. Thus,all three types of riparian strips were important fordifferent species or groups of species. This underlinesthe interest of having a fair ratio of each of thesehabitats to maintain the highest possible diversity inagricultural landscapes. In these highly fragmentedlandscapes, riparian strips often represent wildlifecorridors between remaining habitat islands (Weg-ner and Merriam, 1979; Fahrig and Merriam, 1985;Henderson et al., 1985; La Polla and Barrett, 1993;Burbrink et al., 1998).

Also, there was no indication that, unless regu-larly burned or mowed, riparian strips can becomeshelters for agricultural pests. Even if the abundanceof small mammals increased from herbaceous, toshrubby and wooded riparian strips, this was essen-tially due to insectivore or rodent species restricted towooded habitats and not considered as pests. Abun-dance of potential pest species was rather reduced


by the presence of shrubs and trees in riparian strips,in agreement with similar studies in which the mostabundant small mammal species generally inhabitedforested habitats or pastures and were not consideredas pest species (Dambach, 1948; Yahner, 1983). Theabundance of species likely to be pests was relativelylow in herbaceous riparian strips, and about twiceas high in shrubby and wooded strips. Dambach(1948) concluded that herbaceous field borders havemore rodent pest species than shrubby borders, andthere are indications that increases inM. musculusnumbers are induced by the removal of woody vege-tation (Geier, 1978; Best et al., 1979; Geier and Best,1980).

Not only are pest species less abundant in ripar-ian strips with woody vegetation, the abundance ofinsectivore species is also greater. Total mammalianand amphibian insectivore numbers in wooded ri-parian strips were 2.4 times greater than in herba-ceous strips. Moreover, some rodent species alsocontribute in the control of insect species (Parmenterand MacMahon, 1988): a small mammal commu-nity of shrews and rodents may consume up to 6800preys/ha/day (Churchfield and Brown, 1987). Thepresence of linear habitats like wooded riparian stripsmay also favour bats (Verboom and Huitema, 1997).The diversification of the vegetation structure maytherefore contribute to integrated control of pest in-sects spending part of their annual cycle in riparianhabitats.

Acknowledgements

We are particularly grateful to the numerouslandowners who granted us access to their fields. Wewould also like to thank A. Desrosiers, M. Leclerc,R. Mc Nicoll and S. St-Onge who participated inthe identification of small mammals and providedassistance in the field with L. Choinière, A. Cossette,G. Couture, S. Gagnon, M. Gosselin, L. Lessard, M.Paquin, and É. Sénécal. We would also like to thankL. Bélanger for his collaboration to the project and forgiving us access to the data on vegetation characteris-tics and riparian strip widths. Funding for the projectwas provided by the Minsitère de l’Environnement etde la Faune du Québec.

References

Adler, G.H., 1985. Habitat selection and species interactions: anexperimental analysis with small mammal populations. Oikos45, 380–390.

Banfield, A.W.F., 1974. Les mammifères du Canada. Presses del’Université Laval and University of Toronto Press, Canada.

Beauregard, N., Leclair, R., 1988. Multivariate analysis of thesummer habitat structure ofR. pipiens Schreber in LacSaint-Pierre (Quebec, Canada). In: Szaro, R.C., Severson, K.E.,Patton, D.R. (Coord.), Management of Amphibians, Reptilesand Small Mammals in North America. USDA Forestry Service,General Technical Report RM-166, pp. 129–141.

Bellis, E.D., 1962. The influence of humidity on wood frog activity.Am. Midl. Nat. 68, 139–148.

Best, L.B., Stauffer, D.F., Geier, A.R., 1979. Evaluating theeffects of habitat alteration on birds and small mammalsoccupying riparian communities. In: Strategies for protectionand management of floodplain wetlands and other riparianecosystems. USDA Forestry Service, General Technical ReportWO-12, pp. 117–124.

Brenner, F.J., Mondok, J.J., McDonald, R.J., 1991. Impact ofriparian areas and land use on four non-point source pollutionparameters in Pennsylvania. J. Penn. Acad. Sci. 65, 65–78.

Brown, L.N., 1967. Ecological distribution of six species of shrewsand comparison of sampling methods in the Central RockyMountains. J. Mamm. 48, 617–623.

Brown, W.S., Parker, W.S., 1982. Niche dimensions and resourcepartitioning in a Great Basin desert snake community. In: Scott,N.J. (Ed.), Herpetological Communities. US Department ofInternational Fisheries and Wildlife Service, Wildlife ResearchReport 13, pp. 59–81.

Burbrink, F.T., Phillips, C.A., Heske, E.J., 1998. A riparian zonein central Illinois as a potential dispersal corridor for reptilesand amphibians. Biol. Conserv. 86, 107–115.

Choate, J.R., Fleharty, E.D., 1973. Habitat preference and spatialrelations of shrews in a mixed grassland in Kansas. Southwest.Nat. 18, 110–112.

Churchfield, S., Brown, V.K., 1987. The impact of small mammalsin successional grasslands. Biol. J. Linnean Soc. 31, 273–290.

Clark, B.K., Clark, B.S., Munsterman, W.E., Homerding, T.R.,1996. Differential use of roadside fencerows and contiguouspastures by small mammals in southeastern Oklahoma.Southwest. Nat. 41, 54–59.

Colwell, R.K., Futuyma, D.J., 1971. On the measurement of nichebreadth and overlap. Ecology 52, 567–576.

Cook, F.R., 1984. Introduction aux amphibiens et reptiles duCanada. Musée national de Sciences Naturelles, MuséesNationaux du Canada, Ottawa.

Corn, P.S., 1994. Straight-line drift fences and pitfall traps.In: Heyer, W.R., Donnelly, M.A., McDiarmid, R.W., Hayek,L.-A.C., Foster, M.S. (Eds.), Measuring and MonitoringBiological Diversity: Standard Methods for Amphibians.Smithsonian Institute Press, Washington, DC.

Cummings, J.R., Vessey, S.H., 1994. Agricultural influenceson movement patterns of white-footed mice (Peromyscusleucopus). Am. Midl. Nat. 132, 209–218.


Dambach, C.A., 1948. A study of the ecology and economic valueof crop field borders. Ohio State Univ. Biol. Sci. 2, 1–205.

Decamps, H., Joachim, J., Lauga, J., 1987. The importance forbirds of the riparian woodlands within the alluvial corridor ofthe River Garonne, s.w. France. Reg. Rivers Res. Mgmt. 1,301–316.

deMaynadier, P.G., Hunter, M.L., 1998. Effects of silviculturaledges on the distribution and abundance of amphibians inMaine. Conserv. Biol. 12, 340–352.

Dole, J.W., 1965. Spatial relations in natural populations of theleopard frog,Rana pipiensSchreber, in northern Michigan. Am.Midl. Nat. 74, 464–478.

Dowler, R.C., Katz, H.M., Katz, A.H., 1985. Comparison of livetrapping methods for surveying small mammal populations.Northeast. Environ. Sci. 4, 165–171.

Dubé, C., 1994. Inventaire de l’herpétofaune de la plained’inondation de quatre bassins de la région de Montréal: lelac Saint-Louis, le lac des Deux Montagnes, le lac Saint-Pierreet le Haut-Richelieu en 1992 et 1993. Pour le ministère del’Environnement et de la Faune du Québec, Direction dela faune et des habitats et Service de l’aménagement et del’exploitation de la faune de Montréal.

Dueser, R.D., Shugart, H.H., 1978. Microhabitats in a forest-floorsmall mammal fauna. Ecology 59, 89–98.

Fahrig, L., Merriam, G., 1985. Habitat patch connectivity andpopulation survival. Ecology 66, 1762–1768.

Fleharty, E.D., Navo, K.W., 1983. Irrigated cornfields as habitatfor small mammals in the sandsage prairie region of westernKansas. J. Mamm. 64, 367–379.

Geier, A.R., 1978. Habitat selection by small mammals of ripariancommunities: evaluating the effects of habitat alterations. M.Sc.Thesis. Iowa State University, Ames, IA.

Geier, A.R., Best, L.B., 1980. Habitat selection by small mammalsof riparian communities: evaluating effects of habitat alterations.J. Wildl. Mgmt. 44, 16–24.

Getz, L.L., 1961a. Factors influencing the local distribution ofshrews. Am. Midl. Nat. 65, 67–88.

Getz, L.L., 1961b. Notes on the local distribution ofPeromyscusleucopusand Zapus hudsonius. Am. Midl. Nat. 65, 486–500.

Getz, L.L., 1968. Influence of water balance and microclimate onthe local distribution of the red-backed vole and white-footedmouse. Ecology 49, 276–286.

Gilliam, J.W., 1994. Riparian wetlands and water quality. J.Environ. Qual. 23, 896–900.

Grant, P.R., 1971. The habitat preference ofMicrotuspennsylvanicus, and its relevance to the distribution of thisspecies on islands. J. Mamm. 52, 351–361.

Hamilton Jr., W.J., 1940. The biology of the Smoky Shrew (Sorexf. fumeusMiller). Zoologica 25, 473–492.

Hansen, L.P., Warnock, J.E., 1978. Response of two species ofPeromyscusto vegetation succession on land strip mined forcoal. Am. Midl. Nat. 100, 416–423.

Heatwole, H., 1961. Habitat selection and activity of the woodfrog, Rana sylvaticaLe Conte. Am. Midl. Nat. 66, 301–313.

Henderson, M.T., Merriam, G, Wegner, J., 1985. Patchyenvironments and species survival: Chipmunks in an agriculturalmosaic. Biol. Conserv. 31, 95–105.

Hooper, E.T., 1942. An effect on thePeromyscus maniculatusRassenkreis of land utilization in Michigan. J. Mamm. 23, 193–196.

Horn, H.S., 1966. Measurement of overlap in comparativeecological studies. Am. Nat. 100, 419–424.

Hutcheson, K., 1970. A test for comparing diversities based onthe Shannon formula. J. Theor. Biol. 29, 151–154.

Iverson, S.I., Seabloom, R.W., Hnatiuk, J.M., 1967. Small-mammaldistributions across the prairie-forest transition of Minnesotaand North Dakota. Am. Midl. Nat. 78, 188–197.

Karr, J.R., Schlosser, I.J., 1978. Water resources and the land-waterinterface. Science 201, 229–234.

Kaufman, D.W., Fleharty, E.D., 1974. Habitat selection by ninespecies of rodents in north-central Kansas. Southwest. Nat. 18,443–452.

Kaufman, D.W., Kaufman, G.A., 1990. House mice (Musmusculus) in natural and disturbed habitats in Kansas. J. Mamm.71, 428–432.

Kirkland, G.L., Sheppard, P.K., 1994. Proposed standard protocolfor sampling of small mammal communities. In: Merritt, J.F.,Kirkland, G.L., Rose, R.K. (Eds.), Advances in the Biologyof Shrews, Vol 18. Carnegie Mus. Nat. Hist., Spec. Publ.,pp. 277–283.

Kirsch, E.M., 1997. Small mammal community composition incornfields, roadside ditches, and prairies in eastern Nebraska.Nat. Areas J. 17, 204–211.

Lamarre, G., Laroche, R., Barrington, S., Madramootoo, C., 1993.Sites de démonstration pour évaluer l’impact d’une banderiveraine en milieu agricole. Projet no. 22-13765-580-042.Ministère de l’Agriculture, des Pêcheries et de l’Alimentationdu Québec, Faculté des sciences agricoles et environnementalesde l’Université McGill, pour le compte de la Fédération del’Union des producteurs agricoles de Saint-Jean Valleyfield.

La Polla, V.N., Barrett, G.W., 1993. Effects of corridor widthand presence on the population dynamics of the meadow vole(Microtus pennsylvanicus). Landscape Ecol. 8, 25–37.

Leclair, R., 1985. Les amphibiens du Québec: biologie desespèces et problématique de conservation des habitats.Dép. Chimie-Biologie, Univ. du Québec à Trois-Rivières, pourle ministère du Loisir, de la Chasse et de la Pêche.

Levins, R., 1968. Evolution in Changing Environments: SomeTheoretical Explorations. Princeton University Press, Princeton,NJ.

Linduska, J.P., 1949. Ecology and land-use relationships of smallmammals on a Michigan farm. Ph.D. Thesis. Michigan StateCollege of Agriculture and Applied Science, 244 pp.

Lowrance, R., Leonard, R., Sheridan, J., 1985. Managing riparianecosystems to control nonpoint pollution. J. Soil Water Conserv.40, 87–92.

Lowrance, R., Sharpe, J.K., Sheridan, J.M., 1986. Long-termsediment deposition in the riparian zone of a coastal plainwatershed. J. Soil Water Conserv. 41, 266–271.

Manson, R.H., Ostfeld, R.S., Canham, C.D., 1999. Responsesof small mammal community to heterogeneity along forest-old-field edges. Landscape Ecol. 14, 355–367.

Marshall, W.H., Buell, M.F., 1955. A study of the occurrence ofamphibians in relation to a bog succession, Ithaska State Park,Minnesota. Ecology 36, 381–387.


Merriam, G., 1988. Landscape dynamics in farmland. Trends Ecol.Evol. 3, 16–20.

Miller, D.H., Getz, L.L., 1977. Factors influencing localdistribution and species diversity of forest small mammals inNew England. Can. J. Zool. 55, 806–814.

Morse, D.H., 1973. Habitat utilization by meadow voles on smallislands. J. Mamm. 54, 792–794.

Naiman, R.J., Decamps, H., Pollock, M., 1993. The role of ripariancorridors in maintaining regional biodiversity. Ecol. Appl. 3,209–212.

Parmenter, R.R., MacMahon, J.A., 1988. Factors limiting popula-tions of arid-land darkling beetles (Coleoptera: Tenebrionidae):predation by rodents. Environ. Entomol. 17, 280–286.

Quimby, D.C., 1951. The life history and ecology of the jumpingmouse,Zapus hudsonius. Ecol. Monogr. 21, 61–95.

Rickard, W.H., 1960. The distribution of small mammals in relationto the climax vegetation mosaic in eastern Washington andNorthern Idaho. Ecology 41, 99–106.

Roberts, W, Lewin, V., 1979. Habitat utilization and populationdensities of the amphibians of northeastern Alberta. Can.Field-Nat. 93, 144–154.

Scherrer, B., 1984. Biostatistique. Gaëtan Morin éditeur,Chicoutimi, Québec, Canada.

Small, M.F., Johnson Jr., W.N. 1986. Wildlife management inriparian habitats. In: Bissonette, J.A. (Ed.), Is Good ForestryGood Wildlife Management. Maine Agriculture ExperimentStation, Miscellaneous Publication No. 689, Orono, Maine,pp. 69–79.

Snyder, E.J., Best, L.B., 1988. Dynamics of habitat use by smallmammals in prairie communities. Am. Midl. Nat. 119, 128–136.

Stamp, N.E., Ohmart, D., 1978. Resource utilization by desertrodents in the Lower Sonoran Desert. Ecology 59, 700–707.

Stauffer, D.F., Best, L.B., 1980. Habitat selection by birds ofriparian communities: evaluating effects of habitat alterations.J. Wildl. Mgmt. 44, 1–15.

Tamarin, R.H., Reich, L.M., Moyer, C.M., 1984. Meadow volecycles within fences. Can. J. Zool. 62, 1796–1802.

Thomas, J.W., Maser, C, Rodiek, J.E., 1979. Riparian zones. In:Thomas, J.W. (Ed.), Wildlife Habitats in Managed Forests —The Blue Mountains of Oregon and Washington. USDA, ForestService, Agriculture Handbook No. 553, pp. 40–47.

Verboom, B, Huitema, H., 1997. The importance of linearlandscape elements for the pipistrellePipistrellus pipistrellusand the serotine batEptesicus serotinus. Landscape Ecol. 12,117–125.

Vought, L.B.-M., Dahl, J., Pedersen, C.L., Lacoursière, J.O., 1994.Nutrient retention in riparian ecotones. Ambio 23, 363–366.

Wegner, J.F., Merriam, G., 1979. Movement by birds and smallmammals between a wood and adjoining farmland habitat. J.Appl. Ecol. 16, 349–358.

Whittaker, R.H., Levin, S.A., 1975. Niche: theory and application.Benchmark Papers in Ecology, Vol. 3. Halsted Press,Stroudsburg, PA.

Wrigley, R.E., 1969. Ecological notes on the mammals of southernQuebec. Can. Field-Nat. 83, 201–211.

Wrigley, R.E., Dubois, J.E., Copland, H.W.R., 1979. Habitat,abundance, and distribution of six species of shrews inManitoba. J. Mamm. 60, 505–520.

Yahner, R.H., 1982. Microhabitat use by small mammals infarmstead shelterbelts. J. Mamm. 63, 440–445.

Yahner, R.H., 1983. Small mammals in farmstead shelterbelts:habitat correlates of seasonal abundance and communitystructure. J. Wildl. Mgmt. 47, 74–84.

Zar, J.H., 1984. Biostatistical Analysis. Prentice-Hall, EnglewoodCliffs, NJ.

Documents

Importance of riparian habitats for small mammal and herpetofaunal communities in agricultural landscapes of southern Québec