VERTICAL STRATIFICATION OF THE SMALL MAMMAL COMMUNITY IN A NORTHERN HARDWOOD FOREST

7 Pages·1996·0.18 MB·

by  Taylor, Peter H

Selbyana 17: 15-21
VERTICAL STRATIFICATION OF THE SMALL
MAMMAL COMMUNITY IN A
NORTHERN HARDWOOD FOREST
PETER H. TAYLOR'
Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara,
Santa Barbara, CA 93106
MARGARET D. LoWMAN
The Marie Selby Botanical Gardens, 811 South Palm Avenue, Sarasota, FL 34236
ABsTRACI'. Small mammal research in forests is typically conducted with traps placed on the ground
or affixed to trees at heights up to only 2 meters. In a study conducted in northwestern Massachusetts,
USA, we compared these conventional methods with an innovative new technique of trapping high in the
forest canopy. Using an adaptation of Malcolm's (1991, 1995) pulley method and a canopy walkway, we
compared canopy-trapping to ground-trapping. Total captures per station and total number of individuals
captured per station differed between canopy and ground for both dominant small mammal genera: fiying
squirrels (Glaucomys volans) and mice (Peromyscus spp.). Ground traps were biased toward Peromyscus
spp., underestimated the abundance of G. volans, and misrepresented the species composition of the small
mammal community at the site.
We also compared canopy-trapping with understory-trapping and found that total captures per station
and total number of individuals captured per station differed between understory and canopy locations.
Peromyscus was more abundant in the understory. The small mammal community was represented quite
differently according to trap location, and vertical stratification of this temperate forest community was
evident. We discuss implications of these results for other species (e.g. gypsy moth).
INTRODUCTION Carey & Witt 1991), plastic-laced bait (Packer &
Layne 1991), and live-trapping (Harney & Dues
Forest canopies are inhabited by an extraor
er 1987, Carey et al. 1991, Witt 1991, Sonen
dinarily diverse assemblage of organisms (re shine et at. 1979, Kaufman et at. 1985).
viewed in Wilson 1992). Yet studies of forest
Of these, the two-dimensional, ground-level
ecology have primarily taken place on the forest
grid of live traps is the standard setup for as
floor, which has a community of organisms quite sessing temperate small mammal populations,
different from that of the canopy (Lowman &
but this technique does not adequately sample
Moffett 1993). Canopy studies are typically lim
populations of arboreal woodland mammals
ited by the inherent difficulty of access, but re (Witt 1991, Etheridge et al. 1989). Researchers
cently the canopy has become more accessible
in temperate forests have attached traps to tree
for research through innovative use of cranes trunks to account for vertical stratification of
(Parker et at. 1992), walkways (Lowman & Bour mammals in forests (Hamey & Dueser 1987,
icius 1993), and hot-air balloons (Halle & Pascal Kaufman et al. 1985, Carey et at. 1991, Hall
1992). Even so, despite the obvious importance 1991), yet their traps have been affixed within
of canopies to forest animals, little research has human reach, usually 1-2 m above the ground.
been conducted on mammals in the tropical for The fluorescent powder technique (Lemen &
est canopy and virtually none in the temperate Freeman 1985) has also been used to investigate
forest canopy (Malcolm 1991, 1995; Emmons arboreality of small mammals, but the practical
1995). difficulties of following a trail through the upper
Techniques used to study small woodland
layers ofa forest restricted these efforts to heights
mammals on the forest floor include radio telem of only a few meters above ground (McShea &
etry (Wolff & Hurlbutt 1982, Mikesic & Drick
Gilles 1992, Lemen & Freeman 1985, Etheridge
amer 1992), fluorescent powder tracking (Lemen et al. 1989, Mullican & Baccus 1990).
& Freeman 1985, McShea & Gilles 1992, Eth Recently, Malcolm (1991) designed a system
eridge et al. 1989, Mullican & Baccus 1990), employing pulley mechanisms that enabled him
smoked aluminum plates (Raphael et at. 1986, to trap high in the canopy of a neotropical forest.
Relative abundance of mammal species differed
significantly between 0, 2, and 15 meter trap
heights. We hypothesized that a similar pattern
Corresponding author. exists in temperate forests, although with overall
I
15
16 SELBYANA [Volume 17
rum), with an understory of beech (Fagus gran
Quare us ru bra
60 difolia) (FIGURE 1). Trees reached heights ofap
55 II Fagus 9 ra ndif 0 Jia proximately 50 m and were aged at an average
Acerrubrum
50 of95 years by increment boring (Maxwell & Col
45 lier 1991).
40
:[
35 Canopy-Versus Ground-trapping
1:
,'9" ' 30 Five stations were established to compare
:c
25 trapping in the canopy with trapping at ground
20 level. Each station consisted of three Sherman
15 live traps in oak canopy (15-20 m) and three
10 traps on the ground approximately 10m east of
the canopy-trap tree. Stations were approxi
mately 20 m apart. Canopy-trapping at four of
0
0 10 15 20 25 30 35 40 45 50 the stations was accomplished using an adapta
Distance (m) tion of Malcolm's (1991) pulley system. Canopy
trapping at the fifth station was accessed by a
FiGURE 1. Site profile (50 m x 5 m) representative
of site and including one station (canopy walkway). canopy walkway bridge and platform system
Dominant canopy species was red oak (Quercus rubra). constructed as a research station (see Lowman
Understory was primarily beech (Fagus grandifalia). & Bouricius 1993, 1995).
To minimize the possibility ofa nimals present
at a station not being captured because of insuf
lower species diversity than in Malcolm's trop ficient traps, three traps per location (canopy and
ical Brazilian forest. ground) were used at each station. (In a pilot
Interest in the vertical distribution of small study, we repeatedly set out five traps per loca
mammals in temperate oak forests stemmed from tion; the largest number of animals captured in
our observation that, at the time of our study in a single night was three.)
northwest Massachusetts, USA, predation on Traps were baited with a mixture of peanut
gypsy moth (Lymantria disp ar) pupae increased butter, oatmeal, and bacon. They were checked
with height above ground, reaching a maximum at 0700 hr. Captured animals were identified to
in the canopy. We speculated that this discrep species and most were ear-tagged. Trapping was
ancy was caused by differences in the predator conducted for 9 days during February and March
community between ground and canopy. Our 1992.
preliminary data on predation have since been
supplemented, and made more complex, by fur
Canopy-Versus Understory-trapping
ther experiments, some of which yielded contra
dictory results. Nonetheless, we were prompted In this experiment, each station consisted of
to examine vertical stratification of the small three traps in oak canopy and three traps nailed
mammal predator community to understand the 1-2 m above the ground on a tree trunk ap
mechanism causing increased predation with proximately 10m east of the canopy traps. Trap
height. ping was conducted at five stations for 12 days
Our study examined the influence of trap height during March and April 1992 using the same
on relative abundance of small mammal species protocol as in the previous experiment.
in a temperate forest. By comparing both cano
py-trapping with ground-trapping, and canopy
Data Analysis and Statistics
trapping with understory-trapping (1-2 m on tree
trunks), we tested the hypothesis that trapping For statistical analysis of trapping data, ob
on the ground or in the understory may provide servations on 'number of individuals captured'
a representation of the small mammal commu are preferable as experimental units to obser
nity that is misleading compared to trapping in vations on 'number of captures' because the for
the canopy. mer have a greater degree of independence (i.e.
'number of captures' may include recaptures,
which likely are influenced by previous captures).
METHODS
However, in both of our experiments some an
The study was conducted at Williams College imals were not tagged, and ID numbers of some
in the Hopkins Memorial Forest, a northern recaptured animals were not obtainable. Con
hardwood tract covering 2250 acres in NW Mas sequently, analysis of the data based on 'number
sachusetts, USA. Our site was predominantly red of individuals captured' was hampered.
oak (Quercus rubra) and red maple (Acer rub- Therefore, to approach the question of how
1996] TAYLOR & LOWMAN: VERTICAL STRATIFICATION 17
TABLE 1. Number of small mammal captures per location (canopy, understory, ground) at each station for (a)
canopy vs. ground comparison (9 trap nights) and (b) canopy vs. understory (1-2 m) comparison (12 trap
nights).
a)
Glaucomys volans Peromyscus spp.
Canopy Ground Canopy Ground
Station Total Min.lMax. Total Min.lMax. Total Min.lMax. Total Min.lMax.
No. captures No. Indiv. captures No. Indiv. captures No. Indiv. captures No. Indiv.
1 2 212 0 0/0 0 0/0 2 111
2 6 212 1 111 0 0/0 1 111
3 4 112 0 0/0 0 0/0 3 2/3
4 10 717 0 0/0 1 111 2 112
Mean 5.5 3.0/3.25 0.25 0.25/0.25 0.25 0.25/0.25 2.0 1.25/1.75
SD 3.42 2.7112.5 0.50 0.50/0.50 0.50 0.50/0.50 0.82 0.50/0.96
b)
Glaucomys vo/ans Peromyscus spp.
Canopy Understory Canopy Understory
Station Total Min.lMax. Total Min.lMax. Total Min.lMax. Total Min.lMax.
No. captures No. Indiv. captures No. Indiv. captures No. Indiv. captures No. Indiv.
1 4 3/4 0 0/0 0 0/0 1 111
2 2 2/2 3 3/3 0 0/0 2 112
3 1 0/1 0 0/0 0 0/0 3 113
4 4 3/3 2 212 1 111 1 111
Mean 2.75 2.012.5 1.25 1.25/1.25 0.25 0.25/0.25 1.75 1.0/1. 75
SD 1.5 1.4111.29 1.5 1.5/1.5 0.5 0.5/0.5 0.96 0/0.96
'number of individuals captured' varied between ties: That the numerically dominant genus at
locations, we made note of recaptures and cal ground-level (Peromyscus) could remain domi
culated both the minimum and the maximum nant at all strata (although overall abundances
possible number of new individuals captured at might change), or that numerical dominance
each station for each experiment. Calculation of could shift depending on height to the other ge
minimum possible number of individuals as nus (Glaucomys).
sumed that all untagged animals were subse Distinguishing between these options could
quently recaptured in the experiment and there potentially help explain our observed pattern of
fore should not be counted as unique individuals. predation on gypsy moth pupae. Toward this
Calculation of the maximum possible number of end, we used the G test to consider the hypothesis
individuals assumed that all untagged individ that the dominant species (as measured by num
uals were not subsequently recaptured and there ber of captures) was independent of height.
fore always should be counted as unique indi No animals were captured at one of the five
viduals. To test for differences between heights stations throughout the duration of both exper
we calculated Mann-Whitney U test statistics for iments. Therefore, this station was eliminated
a) total number of captures (including recap from analysis because it provided no informa
tures), b) minimum possible number ofindivid tion on stratification of the small mammal com
uals captured, and c) maximum possible number munity.
of individuals captured. However, trapping data
are notorious for containing violations ofv arious
assumptions of statistical theory, particularlyas REsULTS
sumptions of independence, and our p-values Canopy-vs. Ground-trapping
accordingly should be examined with some cau
tion. In the canopy, we captured 1 Peromyscus leu
Along with examining variation in small copus and 22 Glaucomys volans, including 9 G.
mammal population sizes by height, we also were volans recaptures (TABLE 1). The minimum pos
interested in whether the species composition of sible number of G. volans individuals captured
the small mammal community as a whole varied was 12; the maximum was 13.
between heights. At our site, the community con At ground level, only 1 G. volans was captured,
sisted primarily of two genera (Peromyscus and while 8 Peromyscus captures occurred, including
Glaucomys).This implied Qne of two possibili- one confirmed recapture. Minimum and maxi-
18 SELBYANA [Volume 17
mum possible numbers of Peromyscus individ Spatial Patterns of Flying
uals captured were 5 and 7, respectively. (The Squirrel Recaptures
G. volans captured at ground level was not tagged
In the entire study, nine individual flying
and may have been subsequently recaptured in
squirrels (Glaucomys volans) were captured mul
the canopy.)
tiple times. All were captured most frequently in
The Mann-Whitney U test (with Z corrected
the canopy, including five individuals that were
for ties) showed a significant difference in total
always captured in the canopy. Of these five, two
G. volans captures between canopy and ground
individuals were each captured three times, and
level (p=O.O 18). Likewise, the difference in total
the other three were captured twice. The other
Peromyscus captures between canopy and ground
four recaptured animals were all captured most
level was significant (p=O.022). Results were also
frequently, but not exclusively, in the canopy.
significant (p<0.05) when minimum and max
No individuals were repeatedly captured in the
imum possible numbers ofi ndividuals were used
low (ground/understory) traps alone.
instead of total number of captures.
Five recaptured animals were always captured
The G test for independence with Williams'
at the same station. This included 1 individual
correction was highly significant (df= 1, p< <0.05)
captured 5 times at the same station, 2 individ
for number of captures of each species according
uals captured 3 times at the same station, and 2
to location (canopy vs. ground), indicating that
captured twice at the same station. Two other
the dominant genus of the small mammal com
individuals were captured at the same station 2
munity varied with trap location. Peromyscus
ofthe 3 times they were captured. One individual
dominated at ground level, whereas Glaucomys
was captured 4 times, 3 at the same station.
dominated in the canopy.
DISCUSSION
Canopy- vs. Understory-trapping
Our adaptation of Malcolm's (1991) pulley
Over the 12-day period, we captured 11 G. method for canopy-trapping revealed striking dif
volans and 1 Peromyscus maniculatus in the can ferences in the small mammal community across
opy(TABLE O. Minimum and maximum pos vertical strata of a temperate forest.
sible numbers of G. volans individuals captured The importance of trapping in the canopy was
were 8 and 10, respectively. apparent as the relative abundance of small
Only 5 G. volans captures occurred in the un mammals differed dramatically according to trap
derstory traps, whereas 7 Peromyscus captures height. Ground-level traps gave the misleading
occurred there. None of the G. volans captured impression that the small mammal community
in the understory went untagged. Possible num at the site was dominated by Peromyscus spp.,
bers of captured Peromyscus individuals were 4 with G. volans comprising only a minor fraction
(minimum) and 7 (maximum). of the community. Traps located in the canopy
Captures of G. volans in the canopy included showed that flying squirrels were in fact quite
one individual that previously had been captured abundant. In our canopy- vs. ground-trapping
in the understory traps and one individual ini comparison, captures of flying squirrels in the
tially captured in the understory and then recap canopy outnumbered those on the ground by a
tured twice in the canopy. factor of22. Flying squirrels present in this abun
More than twice as many squirrels were cap dance may play a significant ecological role in
tured in the canopy as in understory traps, but the forest. Their importance would have gone
the Mann-Whitney U test (with Z corrected for undetected with ground traps alone.
ties) showed no significant difference in Glau The low frequency of flying squirrel captures
comys captures between canopy and understory in our ground traps is consistent with prior re
traps (p=0.19). Difference in Peromyscus cap search (Witt 1991, Carey et al. 1991). A United
tures between canopy and understory traps was States Forest Service manual on methods for
significant (p=O.034). The same pattern of sig measuring populations of arboreal rodents rec
nificant differences resulted when either the min ommends that each station consist of one ar
imum or maximum possible number ofindivid boreal trap and one ground trap (Carey et al.
uals was analyzed instead of the total number of 1991). Yet, it suggests placing the arboreal trap
captures. only 1.5 m above ground.
The G test for independence with Williams' We found that traps placed low on tree trunks
correction was significant (df=l, p<0.05) for did not adequately sample the flying squirrel
number of captures of each species according to population. In the comparison between canopy
location (canopy vs. understory), once again and understory-trapping, more than twice as
showing that the composition of the small mam many flying squirrel captures occurred in the can
mal community varied with location. opy as in the understory. Furthermore, all re-
1996] TAYLOR & LOWMAN: VERTICAL STRATIFICATION 19
captured squirrels were captured more frequent known to consume gypsy moths and other insects
ly in the canopy than in the understory traps. (Smith & Lautenschlager 1978, Harlow & Doyle
Five individuals that were recaptured a total of 1990), may have been an important predator on
twelve times would not have been captured at gypsy moth pupae experimentally placed at the
all by an experimental design that did not include site. Predation is believed to be an important
canopy traps. Thus, in order to achieve compre factor affecting dynamics of sparse gypsy moth
hensive sampling of arboreal mammal popula populations (Bess 1961, Bess et al. 1947, Camp
tions, height is a crucial factor. Canopy traps are bell & Sloan 1977a, 19 77b). Peromyscushas been
necessary in addition to ground and understory widely implicated as the most important pred
traps. ator of gypsy moth pupae, but this is based on
The effect of vertical stratification is equally indirect evidence such as (1) scraps ofd ead pupae
true for species such as Peromyscus spp., which purported to have been killed by Peromyscus
are generally thought of as semi-arboreal rather (Campbell & Sloan 1976) and (2) abundance of
than arboreal. To our knowledge, our observa Peromyscus leucopus as assessed by conventional
tions of 2 Peromyscus spp. individuals at ap ground-level grids of live traps (e.g. Elkinton et
proximately 20 m high in the forest canopy are al. 1988). Our research indicates that acceptance
the first published accounts of Peromyscus oc ofP eromyscus as the most important gypsy moth
curring in the canopy, although their climbing predator may be premature, however, because
ability is well-known (Homer 1954). scraps of dead pupae may have been left by G.
Our research showed that, as expected, more volans, not Peromyscus. Ground-level trapping
Peromyscus were captured on the ground and in provides misleading information on the com
the understory than in the canopy, but it is clear position of the woodland small mammal pred
that to adequately study habits and populations ator community, meaning that large relative
of even a "semi-arboreal" small mammal such abundances of Peromyscus at ground-level may
as Peromyscus, the full extent of forest strata not necessarily implicate it as the most important
must be considered, from ground-level to can small mammal predator of the gypsy moth.
opy. This is especially true for P. leucopus and Often it is suggested that predation on gypsy
P. maniculatus, because they nest above ground moths by birds in the canopy is much more com
(Stah 1980, Wolff & Hurlbutt 1982) and much mon in Europe than in North America. Camp
of their traveling distances may be in trees (Eth bell & Sloan (1976) invoked this as an expla
eridge et al. 1989). nation for evolution ofg round-level pupation by
Beyond simply censusing small mammal pop gypsy moths in their native Europe and proposed
ulations, the pulley method facilitates study of that the same behavior in North American for
animal behavior as well. For example, with fur ests exposes individuals to higher mortality than
ther use of this method, and without having to ift hey remained in the canopy. Campbell & Sloan
overcome limitations of the fluorescent powder (1976) suggested that, rather than acting as a ref
or plastic-laced bait techniques, it would be pos uge from canopy predation, this behavior is a
sible to determine whether Peromyscus travel counterproductive evolutionary relict. Yet, our
regularly between ground and canopy or tend to data show that density of flying squirrels, a po
remain in a particular stratum. The pulley meth tential gypsy moth predator, and amount of pre
od would also be extremely useful for studying dation may, in fact, be greatest in the canopy,
the relative arboreality of various species. As ex making ground-level pupation an effective strat
pected, our experiments established that Glau egy for gypsy moth survival even in North Amer
comys volans was far more arboreal than Pero ica.
myscus, but the pulley technique could also be Our adaptation of Malcolm's (1991) pulley
used to distinguish differences in the arboreality system was a valuable means for studying ar
of sympatric Peromyscus species. Other methods boreal small mammals in a temperate forest.
have been used for this purpose, but none has More appropriate for the study of climbing an
extended more than a few meters above the imals than either ground or trunk traps and more
ground (Stah 1980, Harney & Dueser 1987, practical than non-trapping methods such as flu
Packer & Layne 1991). Our captures of 2 Pero orescent dye trails and plastic-laced bait, canopy
myscus (believed to be 1 P. leucopus and 1 P. trapping may reveal behavioral characteristics
maniculatus) in the forest canopy indicate that and ecological interactions currently unknown.
it would be possible to determine which species
uses the upper strata more frequently.
ACKNOWLEDGMENTS
Finally, along with population censuses and
behavioral studies, the pulley method may prove We thank Jay Malcolm and Steve Gaines for
useful for exploring ecological interactions with their valuable comments on an earlier draft of
in the forest community. Our previous research, the manuscript; Steve Gaines for facilitating one
for example, suggested that the flying squirrel, of us (PT) to attend the First International Forest
20 SELBYANA [Volume 17
Canopy Research Symposium, Sarasota, FL, at Sagehen Creek, California. J. Mamm. 72: 615-
1994; Bart Bouricius (Canopy Construction; 617.
Amherst, Massachusetts) for installing pulley HALLE F. AND O. PASCAL, EDS. 1992. Biologie d'une
cano¢e de foret equatoriale II (Rapport de mis
equipment; Michele Dionne (Wells National Es
sion "Radeau des Cimes", Septembre-Decembre
tuarine Research Reserve; Wells, Maine) for sev
1991, Campo, Cameroun). Institut botanique, FR.
eral useful discussions; Jeff Boettner (University 200 pp.
of Massachusetts, Amherst) for teaching the dis HARLOw RF. AND A.T. DoYLE. 1990. Food habits
tinction between Peromyscus and Glaucomys of southern flying squirrels (Glaucomys volans)
species and sexes; Nadine Block, Robert Jeng collected from red-cockaded woodpecker (Pi
and Alex Smith for assistance with field work; coides borealis) colonies in South Carolina. Am.
and Lori Foote and Ellen Baskerville for assis MidI. Nat. 124: 187-191.
liARNEYB.A.ANDR.D.DuEsER. 1987. Verticalstrat
tance with the manuscript. Field research was
ification of activity of two Peromyscus species: an
funded by the Williams College Department of
experimental analysis. Ecology 68: 1084-1091.
Biology. Preparation of the manuscript was sup
HORNER B.E. 1954. Arboreal adaptations of Pero
ported by the Andrew W. Mellon Foundation myscus, with special reference to use of the tail.
along with a National Science Foundation Grad Contrib. Lab. Vert. BioI., Univ. Mich., No. 61.
uate Research Fellowship and a University of KAUFMAN D.W., M.E. PEAK. AND G.A. KAUFMAN.
California Regents Fellowship (both to PT). 1985. Peromyscus leucopus in riparian wood
lands: Use oft rees and shrubs. J. Mamm. 66: 139-
REFERENCES 143.
LEMEN C.A. AND P.W. FREEMAN. 1985. Tracking
BESS H. A. 1961. Population ecology of the gypsy mammals with fluorescent pigments: a new tech
moth, Porthetria dis par (L.) (Lepidoptera: Lyman nique. J. Mamm. 66: 134-136.
triidae). Conn. Agric. Exp. Stn. Bull. 646: 43. LoWMAN M.D. AND B. BoURIOUS.- T993. Canopy
BESS H.A., S.H. SPURR AND E.W. LTITLEFIELD. 1947. walkways-Techniques for their design and con
Forest site conditions and the gypsy moth. Harv. struction. Selbyana 14: 4.
For. Bull. 22: 56. LoWMAN M.D. AND B. BOURIOUS. 1995. The con
CAMPBELL RW. AND RJ. SLOAN. 1976. Influence of struction of platforms and bridges for forest can
behavioral evolution on gypsy moth pupal sur opy access. Selbyana 16(2): 179-184.
vival in sparse populations. Envir. Entom. 5: 1211- LoWMAN M.D. AND M. MOFFETT. 1993. The ecology
1217. of tropical rain forest canopies. TREE 8: 104-107.
CAMPBELL R.W. AND S.H. SLOAN. 1977a. Natural MALcoLM J.R 1991. Comparative abundances of
regulation of innocuous gypsy moth populations. neotropical small mammals by trap height. J.
Envir. Entom. 6: 315-322. Mamm. 72: 188-192.
CAMPBEll. R.W. AND S.H. SLOAN. 1977b. Release of MALcoLM J.R 1995. Forest structure and the abun
gypsy moth populations from innocuous levels. dance and diversity of neotropical small mam
Envir. Entom. 6: 323-330. mals. In: Forest canopies, Academic Press (eds.
CAMPBEll. R.W., S.H. SLOAN AND C.E. BIAZAK. 1977. M.D. Lowman and N. Nadkarni, 1995) pp. 179-
Sources of mortality among late instar gypsy moth 194.
larvae in sparse populations. Envir. Entom. 6: 865- MAxWELL K AND C. COllIER. 1991. Species specific
87l. effects of climate on tree ring growth in Hopkins
CAREY A.B. AND J.W. WITT. 1991. Track counts as Forest. In: The ecology of Hopkins forest (ed. M.
indices to abundance ofa rboreal rodents. J. Mamm. D. Lowman). Williams College Bicentennial Pub
72: 192-194. lication pp. 1-1 to 1-13.
CAREY A. B., B.L. BISWELL AND J.W. WITT. 1991. McSHEA W.J. AND A.B. GIll.ES. 1992. A comparison
Methods for measuring populations of arboreal of traps and fluorescent powder to describe for
rodents. Gen. Tech. Rep. PNW-GTR-273. Port aging for mast by Peromyscus ieucopus. J. Mamm.
land, OR: U.S. Dept. of Agriculture, Forest Ser 73: 218-222.
vice, Pacific Northwest Research Station, 24 pp. MlKEsIC D.G. AND L.c. DRICKAMER. 1992. Effects of
ELKINTON J.S., J.R. GOULD, A.M. LIEBHOLD, H.R. radiotransmitters and fluorescent powders on ac
SMITH AND W.E. WALLNER. 1988. Are gypsy moth tivity of wild house mice (Mus musculus). J.
populations in North America regulated at low Mamm. 73: 663-667.
density? In: W.E. Wallner and KA. McManus, MULLICAN T.R. AND J.T. BACCUS. 1990. Horizontal
eds., pp. 233-249. and vertical movements oft he white-ankled mouse
EMMONS L. 1995. Mammals of rain forest canopies. (Peromyscus pectoralis) in Central Texas. J. Mamm.
In: Forest canopies, Academic Press (eds. M. D. 71: 378-381.
Lowman and N. Nadkarni, 1995) pp. 199-221. PACKER W.C. AND J.N. LAYNE. 1991. Foraging site
ETHERIDGE D.R., M.D. ENGSTROM AND RC. STONE, preferences and relative arboreality of small ro
JR. 1989. Habitat discrimination between sym dents in Florida. Am. MidI. Nat. 125: 187-194.
patric populations of Peromyscus attwateri and PARKER G.G., A.P. SMITH, KP. HOGAN. 1992. Ac
Peromyscus pectoralis in West-Central Texas. J. cess to the upper forest canopy with a large tower
Mamm. 70: 300-307. crane. Bioscience 42: 664-670.
HAu.. D.S. 1991. Diet of the northern flying squirrel RAPHAEL M.G., C.A. TAYLOR AND R.H. BARRETT.
1996] TAYLOR & LOWMAN: VERTICAL STRATIFICATION 21
1986. Smoked aluminum track stations record WALLNER W.E. AND K.A. McMANUS, EDS. 1988. The
flying squirrel occurrence. Res. Note PSW-384. Lymantriidae: a comparison offeatures ofn ew and
Berkeley, CA: Pacific Southwest Forest and Range old world tussock moths. USDA For. Servo Gen.
Experiment Station, Forest Service, US Dept. of Tech. Rep. NE-123. 554 pp.
Agriculture, 3 pp. WILSON E.O. 1992. The diversity of life. Harvard
SMITH H.R. AND R.A. LAUTENSCHLAGER. 1978. Pred University Press. Cambridge, Mass.
ators of the gypsy moth. USDA Agric. Handb. WITT J.W. 1991. Fluctuations in the weight and trap
534: 72. response for Glaucomys sabrinus in western Or
SONENSHINE D.E., D.M. LAUER, T.e. WALKER, AND egon. J. Mamm. 72: 612-615.
B.L. EUSBERG. 1979. The ecology of Glaucomys WOLFF J.O. AND B. HURLBUTT. 1982. Day refuges of
volans (Linnaeus, 1758) in Virginia. Acta Ther Peromyscus leucopus and P. maniculatus. J.
iologica 24: 363-377. Mamm. 63: 666-668.
STAR C.D. 1980. Vertical nesting distribution of 2
species of Peromyscus under experimental con
ditions. J. Mamm. 61: 141-143.