The role of nocturnal rodents in dispersal of southwestern white pine seeds

Material Information

The role of nocturnal rodents in dispersal of southwestern white pine seeds
Pruett, Elizabeth Linnea
Publication Date:
Physical Description:
ix, 47 leaves : ; 28 cm


Subjects / Keywords:
Southwestern white pine -- Seeds -- Dispersal ( lcsh )
Animal-plant relationships ( lcsh )
Rodents ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 43-47).
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Alizabeth Linnea Pruett.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
227994276 ( OCLC )
LD1193.L45 2007m P78 ( lcc )

Full Text
Elizabeth Linnea Pruett
B.S., University of Colorado at Denver, 2002
A thesis submitted to the
University of Colorado at Denver and Health Sciences Center
in partial fulfillment
of the requirements for the degree of
Master of Science

This thesis for the Master of Science
degree by
Elizabeth Linnea Pruett
has been approved
Diana F. Tomback
Leo P. Bruederle

Pruett, Elizabeth Linnea (M.S., Biology)
Thesis directed by Professor Diana F. Tomback
The seeds of southwestern white pine (Pinus strobiformis, Family Pinaceae,
Subgenus Strobus) are primarily dispersed by Clarks nutcracker (Nucifraga
columbiana) in the pines northernmost range. Southwestern white pine has traits
associated with bird-dispersallarge, wingless seeds, and asynchronous seed ripening
phenologies. Nocturnal rodents have previously been regarded as pine seed
predators. Recent studies indicate that deer mice (Peromyscus maniculatus) and other
rodents act as secondary seed dispersers for several pines with large wingless seeds,
including an outlying limber pine (P. flexilis) population and singleleaf pinyon pines
(P. monophylla), both previously believed to be primarily bird-dispersed.
Results from studies in the Chiricahua Mountains, southeastern Arizona, of
rodent seed-caching behavior, live-trapping to identify potential nocturnal rodent
seed-cachers, and greenhouse seed germination experiments, indicated that nocturnal
rodents do, in fact, disperse seeds of southwestern white pine in this region, and that
simulated rodent-caches, like those observed in the field, will germinate. These
results suggest that nocturnal rodents may contribute to the regeneration of
southwestern white pine, and that this bird-dispersed pine has a dual-strategy for seed
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Diana F. Tomback

I wish to dedicate this thesis to my parents, Jim and Barb and many dear friends who
generously supported my efforts to complete this degree. Without their never ending
love and guidance, I would not have been able to complete this chapter of my life.

I wish to thank to my advisor, Dr. Diana F. Tomback, for her contribution and patient
support of my research. Additional thanks to my research partner, Sheridan Samano,
as well as Theresa Ferg, Kristin Grompone, and Mario Perez for their invaluable help
in the field and lab. Thanks to Stephen Fisher, Dr. Karen Kafadar, and Dr. Kathe
Bjork for statistical consultation. I also wish to thank my committee for their
participation and insights. Special thanks to Dr. Cheri Jones for consultations,
insight, and editing; the Biology department for space in the greenhouse and the UCD
graduate school for a Graduate Council Award for Graduate Student Research.

1. Introduction........................................................... 1
1.1 Objectives............................................................. 4
2. Review of the Literature............................................... 5
2.1 Pine Seed Predation.................................................... 5
2.2 Pine Seed Dispersal.................................................... 7
2.3 Southwestern White Pine............................................... 12
2.4 Study Area............................................................ 15
3. Methods............................................................... 18
3.1 Seed Collection....................................................... 18
3.2 Seed Dispersal Tracking............................................... 18
3.3 Rodent Trapping....................................................... 20
3.4 Germination Experiments............................................... 20
3.5 Statistical Analysis.................................................. 23
4. Results............................................................... 25
4.1 Seed Dispersal Tracking............................................... 25
4.2 Rodent Trapping....................................................... 32
4.3 Germination Experiments............................................... 32
5. Discussion............................................................ 38


2.1 Literature accounts of rodent seed-predators and the seeds they consume... 6
4.1 Summary statistics from fluorescent pigment seed dispersal tracking
conducted in September and October 2004 and October 2005. Southwestern
white pine seeds cached by nocturnal rodents at the Chiricahua Mountains
field site, Arizona.................................................. 26
4.2 Statistical model: relationship of number of cached seeds as explained by
cache type, distance, and interaction between cache type and distance .... 31
4.3 Number of rodents trapped per night in rustler park area Coronado National
Forest, Chiricahua Mountains......................................... 32
4.4 Germination results from greenhouse study of simulated caches........ 34
4.5 Seed germination success by block from greenhouse study of simulated
caches............................................................... 35

2.1 Southwestern white pine and limber pine range map................. 14
2.2 Study site map rustler park area Coronado National Forest, Chiricahua
Mountains........................................................... 17
3.1 Diagram of block placement on greenhouse benches.................. 32
4.1 Average number of southwestern white pine seeds cached by nocturnal
rodents per cache type during the seed dispersal tracking study, with outlier
datum point removed................................................. 27
4.2 Average distance southwestern white pine seeds were transported by
nocturnal rodents from seed stations to caches, as determined from the seed
dispersal tracking study............................................ 28
4.3 Residual plots for Poisson log-linear regression models (Observations on the
x-axis and Model fit Residuals on the y-axis........................ 30

1. Introduction
Many nocturnal and diurnal forest rodents (e.g. Heteromyidae, Muridae,
Sciuridae) consume conifer seeds (Gashwiler, 1969; Reynolds, 1950). Seeds provide
high-caloric food for these small mammals, which have high metabolic rates and thus
high energy needs (Whitaker, 1996). When seeds are available, rodents consume
large quantities; therefore, rodents have been viewed as seed predators by forest
managers (Boyer, 1964; Brink & Dean, 1966). Mice and voles tend to eat and store
most of the seeds that are available to them (Abbott, 1961). There is considerable
documentation showing that nocturnal and diurnal forest rodents eat a variety of
conifer seeds, suggesting antagonistic association with the trees. However, recent
research has shown that nocturnal and diurnal rodents also have a mutualistic
association with many pine species (Vander Wall, 1992; Vander Wall, 1994; Vander
Wall, 1997; Vander Wall, Thayer, Hodge, Beck, & Roth, 2001; Vander Wall &
Longlond, 2004; Tomback, Schoettle, Chevalier, & Jones, 2005; Vander Wall, Kuhn,
& Gworek, 2005), serving as secondary seed dispersers.
Primary or initial seed dispersal is movement of seeds away from the parent
tree (Chambers & MacMahon, 1994; Vander Wall & Longlond, 2004). In most
pines, primary dispersal can occur as mature seeds are blown or fall out of opened
cones, or as bird-dispersers remove seeds from closed cones. Secondary seed
dispersal is any movement of seeds after primary dispersal, usually accomplished by

seed-caching birds or rodents (Chambers & MacMahon, 1994). Vander Wall &
Longlond (2004) state that secondary dispersal often moves seeds to microsites that
offer a disproportionately high probability of seedling establishment. Secondary
dispersal behavior varies with respect to the type of cache, number of seeds per cache,
and distance seeds are dispersed, depending upon the species dispersing seeds.
Scatter-hoarding is a method of dispersal where animals store food in many, small
caches, creating a reserve food supply. This behavior may evolve when animals
cannot defend a large concentration of stored food. Dispersing caches spatially
challenges competitors that must encounter them randomly while foraging. Known
scatter-hoarders include several diurnal and nocturnal rodents species in
Heteromyidae, Muridae, and Sciuridae (Smith & Reichman, 1984; Vander Wall,
1997; Vander Wall, et. al., 2001; Tomback, et. al., 2005; Vander Wall, et al., 2005),
as well as several birds in the family Corvidae (Vander Wall & Baida, 1977;
Tomback, 1982; Tomback, 1983; Tomback, 1994; Samano & Tomback, 2003).
Secondary dispersal is thought to be important for pines with large-winged
seeds, because seeds blow from the cones to the ground, making them readily
available to ground foraging rodents and jays which disperse them (Vander Wall,
1992). Tomback, et. al. (2005) reported that deer mice (Peromyscus maniculatus)
and Ords kangaroo rats (Dipodomys ordii) dispersed seeds of a peripheral population
of limber pine (P.flexilis) in the Pawnee National Grasslands. Limber pine has large

wingless seeds and was previously thought to be primarily dispersed by Clarks
nutcracker (Nucifraga Columbiana). Nutcrackers bury seeds in subsurface caches, 1
to 3 cm deep, with 1 to 15 seeds per cache (Tomback, 1998). Limber pine cones
ripen in late August and early September, and some of the seeds fall to the forest floor
making them available to ground foraging rodents.
A close relative of limber pine, southwestern white pine (Pinus strobiformis)
has similar cone ripening and seed dispersal mechanisms. Southwestern white pine
and its relatives (family Pinaceae, genus Pinus, subgenus Strobus) have needles that
occur in fascicles of five, with a deciduous sheath (Perry, 1991). Cone ripening
occurs in late August and early September in an asychronous pattern; some of the
seeds fall to the forest floor and are then available to ground foraging rodents
(Benkman et al, 1984; Samano & Tomback, 2003). Seeds of southwestern white pine
are attractive to seed-eaters, because they are large, highly-caloric, and like limber
pine (Benkman et al, 1984; Tomback et al, 2005), not retained in the cones. Seeds
fall onto the forest floor when cones open, potentially providing food for ground
foraging nocturnal rodents and jays.
Working in the San Juan Mountains of southwestern Colorado, Samano
(2001) and Samano & Tomback (2003) found that southwestern white pine cones
open asychronously among and within individual trees, and that Clarks nutcracker is
the primary seed disperser. Similarly, Benkman (1984) reported that Clarks

nutcracker is the primary seed disperser of southwestern white pine in northern
Arizona, and that the interaction is mutualistic. Southwestern white pine occurs in
several sky island ranges of New Mexico and Arizona (Critchfield & Little, 1966)
without dependable nutcracker visitation (Tomback, 1998).
1.1 Objectives
The primary goal of this study was to examine the role of nocturnal rodents in
seed dispersal of southwestern white pine in the Chiricahua Mountains of
southeastern Arizona, which is outside the dependable range of Clarks nutcracker
(Tomback, 1998). In other words, I wished to determine whether nocturnal rodents
contribute to southwestern white pine regeneration, as they do for limber pine at least
in the Pawnee Grasslands and for other pine species. To date, there have been no
studies examining the potential role of nocturnal rodents as dispersers of
southwestern white pine.
To determine the importance of rodents to southwestern white pine regeneration, I
first tested the hypothesis that nocturnal forest rodents cache southwestern white pine
seeds using fluorescent pigment tracking. 1 identified which rodent species may be
the seed dispersers with live mammal trapping; and, finally, I simulated cache-types
observed in the field in the greenhouse to test the hypothesis that they would

Review of the Literature
2.1 Pine Seed Predation
Seed predation by nocturnal and diurnal forest rodents has been well-
documented (Table 2.1). Additionally, a considerable amount of time and energy has
been devoted by forest managers to eliminating rodents viewed as predators or
competitors by managers (Reynolds, 1950; Abbott, 1961; Gashwiler, 1969). For
example, before the environmental effects were fully known, foresters would treat
areas where they wanted to eliminate vermin populations with compound 1080
(sodium fluoracetate) (Gashwiler, 1969). Nocturnal forest rodents in the families
Heteromyidae and Muridae are all known pine-seed predators, as are diurnal forest
rodents in Sciuridae (Vander Wall, 1990).
Tevis (1953) reported that 54% of the cones of sugar pine (P. lambertiana) in
California were lost to cutting by pine squirrels (Tamiasciurus), 34% were destroyed
by white-headed woodpeckers (Dendrocopos albolarvatus)', and, for the remainder of
cones that escaped initial predation, their shed seeds were later attacked by ground-
foraging Stellers jays (Cyanocitta stelleri) and a variety of other small songbirds and
chipmunks. In a year of light logging, when wildlife conditions were favorable, the
competition between the squirrels and woodpeckers led to complete loss of the sugar
pine cone crop.

Table 2.1 Literature accounts of rodent seed-predators and the seeds they consume
Family Predatory Rodent Species Seed Species Eaten Reference
Heteromyidae Merriams kangaroo rat (Dipodomys merriami) Many conifer species Reynolds 1950
Muridae Old-field mouse (Peromyscus polionotus) Cotton mouse (Peromyscus gossypinus) Golden mouse (Ochrotomys nuttalli) Eastern harvest mouse (Reithrodontomys humulis) Longleaf pine (Pinus palustri) Boyer 1964
Muridae Deer mouse (Peromyscus maniculatus) Douglas-fir (Pseudotsuga mezniesi) Gashwiler 1969
Muridae White-footed mouse (Peromyscus leucopus) Red-backed voles (Clethrionomys gapperi) Meadow voles (Microtus pennsylvcmicus) White spruce (Picea glauca) Radvanyi 1970
Muridae White-footed mouse (Peromyscus leucopus) Sugar pine (Pinus lambertiana) Tevis 1953
Sciuridae Red squirrels (Tamiasciurus hudsonicus) Limber pine (Pinus flexilis) Benkman et al., 1984
Sciuridae Red squirrels (Tamiasciurus hudsonicus) Southwestern white pine (Pinus strobiformis) Samano & Tomback 2003
Sciuridae Red squirrels (Tamiasciurus hudsonicus) White spruce (Picea glauca) Brink & Dean 1966 Radvanyi 1970
Sciuridae Chipmunks (Eutamias amoenus) White spruce (Picea glauca) Radvanyi 1970
Sciuridae Yellow pine chipmunks (Tamias amoenus) Lodgepole pine chipmunks (Tamias speciousus) Golden-mantled ground squirrels (Spermophilus lateralis) Jeffrey pine (Pinus jeffreyi) Vander Wall 1992
Sciuridae Colorado chipmunk (Tamias quadrivitatus) Southwestern white pine (Pinus strobiformis) Samano & Tomback 2003
Sciuridae Douglas squirrels (Tamaisciurus douglasii) Sequoia (Sequoiadendron giganteum) Shellhammer 1966

Radvanyi (1970) conducted an experimental study to determine the degree to
which small-mammal populations impeded the regeneration of white spruce (Picea
glauca). Pesticide-treated, radiotagged seeds were tracked and assigned a seed-fate
category during two springs and two subsequent artificial fall seedings. Fifty percent
of the seeds were taken by small mammals (mice, chipmunks, and shrews) within 3 to
4 months of seeding. However, only a third as many were destroyed in the fall spruce
seeding; 35% of the spring losses and 11% of the fall losses were attributed directly
to deer mice (Peromyscus maniculatus). Both small mammals and insects continued
to consume seeds even after they were treated with the pesticides.
2.2 Pine Seed Dispersal
The following is an overview of pine seed dispersal by Clarks nutcracker
based on Vander Wall and Baida (1977), Vander Wall and Baida (1981), Tomback
(1982, 1994, 1998, 2001),, and Samano and Tomback (2003). Clarks nutcracker is
an important seed disperser for at least five pines in Pinus subgenus Strobus:
whitebark, limber, Colorado pinyon (P. edulis), singleleaf pinyon, and southwestern
white pine. Clarks nutcracker harvests and stores pine seeds from late summer until
the seed crop is depleted.
Morphological characteristics of Clarks nutcracker that aid in seed extraction
and caching are a long, sharp bill and a sublingual pouch. Nutcrackers remove unripe
seeds by stabbing their bill in-between cone scales and tearing off scale pieces,

allowing them to extract pieces of unripe seeds. They begin extracting whole seeds
for caching when seed coats have formed. The nutcracker sublingual pouch is then
used to transport seeds to caching sites. Nutcrackers transport seeds from a few
meters up to 22 km, often to a communal seed storage location used by the local
population of nutcrackers. The caches are typically on steep, windswept, south-
facing slopes that have little snow accumulation but are also dispersed throughout the
forest. When caching whitebark pine seeds, nutcrackers typically place 1 to 15 seeds
per cache, per bird disperser, with an average of 3 to 5 seeds at a depth of 1 to 3 cm.
Nutcrackers are coadapted and secondarily coevolved with southwestern white and
limber pine, but primarily coevolved with whitebark pine (Tomback, 1983).
Benkman et al. (1984) considered Clarks nutcracker the primary seed
disperser of both southwestern white and limber pine. Their study was conducted in
north-central Arizona in habitats that exemplified mid-elevation forest and open,
subalpine parkland. The aim of the study was to examine cone and seed
characteristics that facilitated and hindered seed harvest by dispersers and predators,
respectively. Comparisons of the morphology of cones, seeds showed that the cones
of limber pine cones weighed less and contained fewer seeds than those of
southwestern white pine. Observations revealed that limber pine was more evenly
and harvested by nutcrackers than southwestern white pine, because in this area the
limber pine cones ripened synchronously within individual tree but asynchronously

among trees. In contrast, the southwestern white pine cones ripened synchronously
within and among trees. The study showed that red squirrels removed 80% of closed
cones from both limber and southwestern white pine. However, they spent more
mean-time per seed, obtaining and eating seeds from closed limber pine cones than
from closed southwestern white pine cones. These observations and differences in
cone morphology led to the prediction that limber pine (P. flexilis) is better-adapted
than southwestern white pine for seed dispersal by nutcrackers. Samano and
Tomback (2003) however, postulated that total asynchronous ripening of cones
promotes avian seed dispersal.
The possibility of seed dispersal by rodents has been contemplated since the
late 1950s (Abbott, 1961), but only recently has been studied. The previous
prevailing opinion was that seed predation by rodents outweighed any possible
beneficial dispersal; however, Janzen (1971) indicated that some animals act both as
seed dispersal agents and seed predators, and that seed predation is the cost for
reliable dispersal. Scatter-hoarding by nocturnal rodents has many advantages for the
seed-cacher and the seed. For example, competing seed-predator species are less
likely to expend the energy necessary to locate many distributed caches, as it becomes
energetically costly. In turn, the caching rodent does not have to spend energy
defending a large cache. According to Smith and Reichman (1984), nocturnal rodents
tend to scatter-hoard seeds away from the parent tree. Seed caches not destroyed in

winter or invaded in the following spring germinate and produce seedlings, according
to Abbott and Quink (1970).
Vander Wall et al. (2001) described in detail the scatter-hoarding behavior of
deer mice, including cache sizes, depths, and placement of caches in different
substrates. The caching study was conducted using radioactively labeled Jeffrey pine
(P. jeffreyi) seeds in experimentally replicated underground nests. The mice were
presented with seeds, and the fate of the seeds and/or subsequent seed coats was
determined using a Geiger counter. Deer mice did scatter-hoard seeds in small caches
of 1-2 seeds, and mostly at depths ranging from 2-12 mm. Caches occurred most
frequently in mineral soil and in light plant litter, and it appeared that rodents avoided
caching in open areas. Deer mice apparently aid in the regeneration of some pine
It is commonly accepted that tree or stem clusters are the result of seed
caches. West (1968) reported that a minimum of 15% of the ponderosa pine (P.
ponderosa) seedlings in Oregon occurred as clusters, and resulted from rodent caches.
Vander Wall (1992, 1994, 1997) demonstrated that rodents served as important seed
dispersers of wind-dispersed pines and, in some cases, provided higher-quality
dispersal than corvids.
Vander Wall (1992) studied how four species of seed-caching rodents,
including yellow pine chipmunks (Tamias amoenus), lodgepole pine chipmunks

(Tamias speciousus), golden-mantled ground squirrels (Spermophilus lateralis), and
deer mice, redistributed Jeffrey pine through secondary seed dispersal. Jeffrey pine is
a wind-dispersed pine, and, as such, seedlings occur in a shadow of the parent tree,
where the prevailing winds typically carries the seeds. He indicated that seedlings
occur in clumps, suggesting dispersal by scatter-hoarding animals. He reported that
between 95% and 99% of the radioactively tagged seeds in the study were removed
and cached by rodents. Vander Wall (1992) suggested that Jeffrey pine seed dispersal
occurred in two stages: (a) primary dispersal resulting from wind and (b)
subsequent secondary dispersal into soil caches by rodents.
Of most relevance to this study is the work reported by Tomback et al. (2005).
They examined whether nocturnal rodents were the primary seed-dispersers of a
peripheral population of limber pine in the Pawnee National Grasslands. At this
location, nutcrackers and other diurnal seed dispersers are absent. The authors
tracked nocturnal rodents using fluorescent pigment to gather data on seed dispersal
and caching (see methods). Tracking data showed that rodents cached seeds in two
basic ways: on the surface and in buried caches. Surface caches had an average of
1.6 seeds, while buried caches contained an average of 4.4 seeds. Mammal trapping
implicated deer mice as the major seed-disperser of limber pine. Successful
germination of seeds placed in artificial caches, like those made by rodents, suggested
that in this isolated population, nocturnal, seed-caching rodents effectively replaced

Clarks nutcracker as the major disperser of limber pine seeds. This suggests that
limber pine could be dispersed by both Clarks nutcracker and rodents, representing a
dual strategy for seed dispersal, which apparently varies geographically. Clearly,
nocturnal rodents are capable of positively contributing to the regeneration of various
2.3 Southwestern White Pine
Southwestern white pine was first described by George Engelmann in 1848
(Engelmann, 1848). Historically, southwestern white pine has been reportedly
referenced incorrectly as either a variety of P. flexilis or P. ayachahuite (Riser, 2003).
Currently, the debate continues as to whether southwestern white pine has
geographically variable traits or if it hybridizes with P. flexilis where their ranges
overlap (Steinhoff & Andresen, 1971; Riser 2003). According to Steinhoff and
Andresen (1971) limber and southwestern white pines differ in overall tree
architecture but also in cone, seed, and needle characteristics southwestern white
pine has larger cones, heavier seeds, and longer needles. Southwestern white pine is
found in sky island populations in southwestern Colorado, the mountains of western
Texas, New Mexico, and Arizona; it extends in sky island communities to central
Mexico (Critchfield & Little, 1966; Krugman, S.L. & J.L. Jenkinson, 1974) (Figure
2.1). Overall it ranges from 35N, 111W to 21S, 100E. However, in its southern
U.S., and Mexican range, neither the primary seed disperser, Clarks nutcracker, nor

the major seed predator the red squirrel occur (Hall & Kelson, 1959; Tomback, 1998).
Spofford (1973), however, reported seeing Clarks nutcracker in the Chiricahua
Mountains in the fall of 1972, but nutcrackers are considered an infrequent visitor in
the area. Thus, southwestern white pine is an interesting species for study, because
seed dispersal by nutcrackers is documented from its northern populations, but little is
known about seed dispersal in the southern populations.

Figure 2.1 Southwestern white pine and limber pine range map (Critchfield and Little

2.4 Study Area
To examine whether nocturnal rodents disperse southwestern white pine
seeds, I selected a sky island study area outside of the dependable range of Clarks
nutcracker: Coronado National Forest, Chiricahua Mountains in southeastern
Arizona. In the Rustler Park area, Onion Saddle (N 31.914, W109.141, 7,600 2316
m) (Figure 2.2) I conducted seed caching studies, seed germination experiments, and
live trapping of mammals in September and October of 2004 and September 2005.
The forest there is comprised of mature and young southwestern white pine,
ponderosa pine (P. ponderosa var. arizonica), white fir {Abies concolor), and
Douglas fir {Pseudotsuga menziesii var. glauca). The understory is comprised of
spiderwort (Tradescantia spp.), asters (Aster spp.), verbena (Verbena spp.), lupine
(Lupinus spp.), native thistle (Cirsium spp.), gilia (Gilia spp.), geranium (Geranium
spp.), and grasses (Personal observation; Southwest Parks and Monuments
Association, 1998).
Cone opening in southwestern white pine was asynchronous both within and
among trees (Samano, Tomback, Maddox, & Pruett, unpublished data). Cones began
opening in the Chiricahua Mountains in late August, with the majority open by 9
October. Almost all cones had open scales at the time of fieldwork, and cones were
very resinous. Cone position on trees in the San Juan Mountains, Colorado, was

described by Samano (2001) as horizontal, above horizontal, upright, below
horizontal, or pendulous, but most cones in the Chiricahua Mountains were directed
below horizontal or pendulous (Samano, Tomback, Maddox, & Pruett, unpublished
data. The average seed dimensions for the Chiricahua Mountains were as follows:
length (1.26 cm SE 0.024), width (0.862 cm SE 0.009), depth (0.586 cm SE
0.007), and mass (0.254 g SE 0.005) (Samano, Tomback, Maddox, & Pruett,
unpublished data). Southwestern white pine cone crops occur generally every two to
four years (Steele, 1990); in this area, cones were not produced in 2004 or 2005
(personal observation).

Figure 2.2 Study site map rustler park area Coronado National Forest, Chiricahua
Mountains (American Museum of Natural Flistory 2004)

3. Methods
3.1 Seed Collection
Southwestern white pine seeds were collected from cones for use in the
fluorescent pigment tracking studies and germination experiments. Permits for cone
collection were obtained from both the San Juan National Forest, Colorado, and
Coronado National Forest, Arizona. Cones were collected on September 6 and 7,
2003, in the San Juan National Forest, by means of a standard six-foot tree pruner. A
few additional cones were collected in the Coronado National Forest in the fall of
2004 and 2005, as they became available; however, there was not a large cone crop
either year. Cones were placed in plastic bags for transport, open-air dried in the lab,
and seeds were extracted from the cones. For the germination experiment, seeds were
frozen at 0 C to simulate cold-stratification.
3.2 Seed Dispersal Tracking
A Special Use research permit was obtained from the Coronado National
Forest, Douglas Ranger District, for the following studies. I studied seed caching by
nocturnal rodents using fluorescent-pigment tracking, as first introduced by Lemen
and Freeman (1985), to find and describe rodent seed caches (Longland & Clements,
1995). This technique uses a non-toxic fluorescent-pigment (Radiant Color,
Richmond, California) to coat seeds and cover seed stations, resulting in a trail of
rodent footprints to cache sites and burrows. Seed stations each provided 200

southwestern white pine seeds coated with one of the four fluorescent pigment colors
(pink, yellow, blue, or green). The seeds were placed in a glass Petri dish that was
positioned in the center of a foil tray (~ 30 x 50 cm), which was covered with
sandpaper impregnated with the same pigment color as the seed color. Seed stations
were set up in areas where rodents would naturally forage.
I timed seed caching experiments to coincide with natural availability of
southwestern white pine seeds in late summer and fall. Seed caching experiments
were conducted over seven nights: September 4 and 5, 2004; October 8 and 9, 2004;
and September 1, 2, and 3, 2005. Each night, seed stations were placed under
southwestern white pine canopies at three to ten locations, about 50m apart, by
Tracking and data collection began at approximately 0300h each night and
continued until dawn. Rodent trails were followed with the use of a commercial UV
light (SuperBright 2010LW, with 368 nm lamp, and battery pack by UV Systems
Inc., Renton, WA). Starting at the seed stations, the UV light was used to illuminate
the area around the seed station to detect rodent trails. Each trail was followed to its
terminus, often where a fluorescent smudge marked a cache. Located cache sites
were quickly examined in the dark, marked with surveyors flags, and revisited for
additional description after sunrise. Further description of caches included whether
caches were on the surface or buried, substrate type, depth of cache, number of seeds

per cache, and whether caches were recovered by rodents on succeeding field days.
Distance from seed station to each seed cache was measured as a straight-line
distance, using a 50-m transect tape.
3.3 Rodent Trapping
I conducted live trapping to identify potential nocturnal rodent seed
dispersers. Live trapping was conducted on October 8 and 9, 2004 and September 2
and 3, 2005. Two traplines were established each night near southwestern white pine
trees, but avoided overlap with the seed-dispersal study areas. Each trapline consisted
of 40 Sherman traps (3x3x9 in), two per trapsite, spaced at 10m intervals, and
baited with a mixture of rolled oats and peanut butter. Traps were baited and set at
1800h and checked at sunrise the following morning. Captured animals were
identified and released on-site.
3.4 Germination Experiments
I examined germination success of simulated seed-caches, which replicated
the most common types observed in the field, to determine if rodent caches contribute
to southwestern white pine regeneration. Germination experiments were conducted
in the field, as well as in the greenhouse. On October 9, 2004, two germination plots
were set-up in the Rustler park study area to replicate seed caches observed in the
area; seeds for this experiment were collected in the San Juan National Forest in
southwestern Colorado in 2003. The simulated caches were of the following type:

one-seed surface caches, two-seed surface caches, one-seed buried caches, and two-
seed buried caches. Their planting position was randomized, with respect to cache
type to aid any inference made to one or all types observed. The caches were covered
by a hardware cloth cage to minimize seed predation by rodents, and the sites were
revisited on July 9, 2005. The hardware cloth was removed and the seed caches were
examined; ungerminated seeds remained under the hardware cloth and were either
dead or the embryos were missing. Not all of the seeds that were planted were
accounted for during this examination, indicating that some seeds may have been
removed by insects and/or foraging animals although there were no obvious signs of
this disruption.
Since the field germination experiment did not succeed, germination
experiments were also conducted in the greenhouse using the same four cache types.
The following techniques were performed to maximize successful germination, using
methods described in Powell (1996). After cones were collected in September 2003,
seeds were extracted from cones and placed in the freezer. Some of the seeds were
used in the fluorescent pigment tracking studies and the rest remained frozen. On
January 27, 2006, seeds were soaked in water and floaters were discarded. It has
been shown that floaters are more likely to contain underdeveloped embryos and or
empty seed coats. Approximately 20% of the seeds tested in this way were discarded.
The seeds were then washed in a 10% bleach solution to minimize fungal growth;

regardless some fungal growth was observed during the greenhouse study. Seeds
were then rinsed in running water for 24 hours, and clipped at the radical end of each
seed to further aid germination. Prior to planting, playground sand and peatmoss
were mixed by hand and put into 6 plastic pots, filled to within 2 cm of the rim, and
watered to moisten the peat. The statistical software R ( R Foundation) was used to
generate the randomized complete block design (RCBD) used for planting. The
RCBD was used to counter the environmental effects of heterogeneity in the
greenhouse, and allowed for comparisons within blocks (Ramsey & Schafer, 2002).
Ten 6 x 8 blocks were established, and each pot within the block was labeled
according to treatment and replicate number. I planted a total of 730 seeds on
February 3, 2006, 73 seeds per block. Blocks 1-9 were place on one bench in the
center of the greenhouse from north to south, with the exception of block number ten
which was placed on a separate bench due to a limited amount of space (Figure 3.1).
Figure 3.1 Diagram of block placement on greenhouse benches

Surface-type cache treatments were not covered by any soil, and buried cache
treatments were placed within a depression approximately 4 cm deep and covered.
Immediately following planting, the pots were watered thoroughly. The pots were
watered with different regimes over time:
every day for the first two weeks,
every other day for the next two weeks,
every third day for 41 days,
once per week for 107 days,
for a total of 176 growing days. Data were collected each day that pots were watered.
3.5 Statistical Analysis
The statistical software R ( R Foundation) was used for all non-descriptive
statistics. Descriptive statistics were compiled from the seed dispersal study for:
number of caches located, depth of cache, distance from seed station, and substrate
type in which the cache was located. To compare number of seeds per cache, I used
the Wilcoxon signed-rank test with continuity for correction, hypothesizing that cache
type had an impact on the observed seed count. Wilcoxon signed-rank test with
continuity for correction was also used to compare distances per cache type. Poisson
log-linear regression modeling was used to test the relationship of distance from seed
station and cache type, with respect to models with the number of seeds cached, cache
type, and cache distance from seed station. Residual plots of four models: a null

model, cache type model, distance model and an interaction model of distance and
cache type were generated to examine if one variable was more influential on
observed seed count than another.
For analysis of the greenhouse germination data, chi-square analyses and
Pearson's chi-squared analysis with Yates' continuity correction were preformed to
examine independence among treatment blocks, cache types and number of seeds that
germinated. Additionally, multivariable logistic regression was performed using seed
germination success as a binary outcome to estimate the effect that treatment blocks,
cache types, and number of seeds had on seed germination success. Additionally an
odds ratio was calculated to measure the likelihood of seeds germinating in one block
as compared to other blocks.

4.1 Seed Dispersal Tracking
Before making inferences based on the following tests, the relatively small
sample size for caches should be taken into account; additional data are needed to
substantiate these findings. Descriptive statistics reveal that buried caches were found
under litter, soil, and plants; surface caches were found on litter, soil, and rocks
(Table 4.1). One outlier datum point contained 12 seeds within a single buried cache.
All other caches tended to contain 1-2 seeds. Although many tracks led to burrows,
the number of seed caches located each night ranged from zero to 14. Each night,
nocturnal rodents removed from 0-200 (x= 24.9 38.8SE) seeds from each seed

Table 4.1 Summary statistics from fluorescent pigment seed dispersal tracking
conducted in September and October 2004 and October 2005.
Southwestern white pine seeds cached by nocturnal rodents at the
Chiricahua Mountains field site, Arizona. Means are presented with
standard errors in parentheses.
Cache Type No of located caches Total no. of seeds cached Mean (SE) no. of seeds per cache Range of no. of seeds per cache Mean (SE) distance (m) from seed station Range distance (m) from seed station
Under litter 2 15 7.5 (1.6) 3-12 4.08 (3.48) 0.60- 7.56
Under soil 7 8 1.14(0.1) 1- 2 10.31 (1.57) 2.14-12.61
Under plant 2 4 2.0 (0.0) 2- 2 7.80 (0.80) 7.00- 8.60
On litter 8 10 1.3 (0.1) 1- 2 8.09(1.95 2.60-16.96
On rock 4 5 1.3 (0.2) 1- 2 6.08 (1.55) 2.74- 9.61
On soil 5 14 2.8 (0.5) 2- 6 6.78 (2.41) 0.20-13.20
Buried 11 27 2.5 (0.6) 1-12 8.72 (0.33) 0.60-12.61
Surface 17 29 1.7 (0.2) 1- 6 7.23 (0.11) 0.20-16.96
Combined 28 56 2.0 (0.3) 1-12 7.82 (0.87) 0.20-16.96
One unique cache was located under the bark of a fallen tree with a large,
indeterminate number of seeds visible. Caches were found on the surface (n = 17
caches) of rocks, litter, or mineral soil, or they were buried {n = 11 caches) under
litter, soil as deep as 3 cm, but with an average of 1.2 cm ( + 0.33 cm), or plants.
Average number of seeds per cache differed significantly between buried caches (x =
2.5 seeds 0.6SE) and surface caches (x = 1.7 seeds 0.2SE) (Figure 4.1)

(Wilcoxon signed-rank test, with correction for continuity and ties,/? = 0.03, two-
tailed). However, the average number of seeds per cache did not differ significantly
when the outlier datum point in the buried caches was removed (buried* = 1.5
seeds 0.6SE) and (surface jc = 1.7 seeds 0.2SE).
Average No. of Seeds/Cache
Figure 4.1 Average number of southwestern white pine seeds cached by nocturnal
rodents per cache type during the seed dispersal tracking study, with
outlier datum point removed.

Distance from seed station to caches ranged from 0.2 to 16.9 m, with the
average distance differing between buried caches (x = 8.72 m 0.33SE) and surface
caches (x = 7.23 m 0.11SE) (Figure 4.2), (Wilcoxon signed-rank test, with
correction for continuity and ties,/? = 0.08, two-tailed).
Average Distance from Seed Station
Buried Surface
Figure 4.2 Average distance southwestern white pine seeds were transported by
nocturnal rodents from seed stations to caches, as determined from the seed
dispersal tracking study.

All caches were revisited the day immediately following original identification, again
1-2 days later, and finally in succeeding field seasons in 2005 and 2006; however, no
seeds remained, which was likely the result of rodent recaching or removal.
The residual plots from the Poisson log-linear regression models indicated a
small but statistically significant relationship between distance from seed station with
respect to number of seeds cached, and showed a small amount of exponential
behavior with high positive values attributed to observations with high seed counts
(e.g., Observation #1 had 12 seeds, #3 had 5 seeds) as opposed to a mean of 2.3 seeds
per observation (Figure 4.3).

Cache Type
* CN

o *
* *
* d
o *
i i 0 s o-

Distance & Cache Type

d *
ID *
O ~ * *
Figure 4.3 Residual plots for Poisson log-linear regression models (Observations on
the x-axis and Model fit Residuals on the y-axis

Cache type was the more influential variable on its own in the models.
However, the interaction model containing both distance and cache type yielded the
least residual deviation (Table 4.2), indicating predictive value from the number of
seeds in each cache. The residual deviation, as a measure of model fit, was 25.8 with
21 degrees of freedom (DF). The model yielded a Chi-squared p-value of 0.79,
indicating poor overall performance. Given only poor performance of this model a
further look in to the validity of the assumptions behind Poisson analysis should be
considered. An analysis of variance of the model showed that cache type as a
coefficient of the model has a p-value = 0.006 while the p- value for distance as a
coefficient of the model is = 0.07. This additional analysis reinforces the position that
cache type is the most influential coefficient as a predictor of number of seeds per
Table 4.2 Statistical model: relationship of number of cached seeds as explained by
____________cache type, distance, and interaction between cache type and distance
Model Residual Df Residual Deviation Deviance
1 Null 23 36.7
2 Cache Type 22 29.1 7.5
3 Distance 22 32.4 4.3
4 Distance & Cache Type 21 25.8 10.9

4.2 Rodent Trapping
Twenty-three deer mice were trapped over four nights, with an average of
7.2% trapping success over all traps for all nights (range 0.-15% per night). These
data suggest that deer mice are common and may be dispersers of southwestern white
pine seeds at this site (Table 4.3). Visible fluorescent tracks were comparable in size
and shape to deer mice fore and hind foot tracks. Whereas only deer mice were
caught in traps, the one outlier cache size observation (12 seeds in one-buried cache)
suggests that other larger nocturnal rodent seed cachers were present in the area as
Table 4.3 Number of rodents trapped per night in rustler park area Coronado National Forest, Chiricahua Mountains
Trapline 10/09/2004 N Traps % Success 10/10/2004 N Traps % Success
# 1 5 40 12.5 1 40 2.5
#2 6 40 15.0 3 40 7.5
Trapline 9/2/2005 N Traps % Success 9/3/2005 N Traps % Success
#1 5 40 12.5 0 40 0.0
#2 3 40 7.5 0 40 0.0
4.3 Germination Experiments
On July 9, 2005, after nine months, the field germination experiment was
revisited; one seed had germinated under the hardware cloth. The remaining
relocated sown-seeds were either dead or the hulls had holes and were empty,

suggesting that insects raided the experiment or had been present in the developing
On February 3, 2006 a total of 730 seeds were planted in the greenhouse: 350
surface caches and 380 buried caches. Of the 45 total two-seed surface caches that
germinated, in only one instance did both seeds that were planted in the same pot
germinate (Table 4.4). Likewise, of the 53 two-seed buried caches that germinated,
in only two instances did both seeds that were planted in the same pot germinate.
There was no significant difference in seed germination success between total surface
and total buried caches, 18.6% and 20.0% respectively (Table 4.4). This analysis
treats all seeds as statistically independent, although they are not. Pearson's Chi-
squared analysis with Yates' continuity correction also indicated that cache type
(surface versus buried) was not a predictor of seed germination success (j2 = 0.008,
df = 1, p-value = 0.929). Additionally, there was no significant difference in seed
germination success between total one-seed and total two-seed caches, Pearson's Chi-
squared analysis with Yates' continuity correction (j2 = 0.001, df = l,/?-value =

Table 4.4 Germination results from greenhouse study of simulated caches
Cache Type No. Caches No. of Seeds Planted No. of Seeds Germinated Percent Germination
One-seed 110 110 20 18.2%
Two-seed 120 240 45 18.8%
Total Surface 230 350 65 18.6%
One-seed 120 120 23 19.2%
Two-seed 130 260 53 20.4%
Total Buried 250 380 76 20.0%
Overall 480 730 141 19.3%
Table 4.5 shows a linear trend in seed germination success by block number.
Chi-square analysis also indicates that block as a treatment had a significant effect on
seed germination success (j2 = 48.560, df = 9,/?-value = 2.007e-07). This implies
that the placement of the blocks within the greenhouse had a linear effect on seed
germination success, suggesting that there was a gradient in environmental

Table 4.5 Seed germination success by block from greenhouse study of simulated caches
Block No. No. Seeds/Block No. of Seeds Germinated Percent Germination
1 73 5 6.8%
2 73 9 12.3%
3 73 11 15.1%
4 73 12 16.4%
5 73 8 11.0%
6 73 14 19.2%
7 73 16 21.9%
8 73 18 24.7%
9 73 19 26.0%
10 73 29 39.7%
Total 730 141 19.3%
Blocks 1-9 were placed on the same bench in the center of the greenhouse,
arranged in a north to south orientation, with block number nine being the furthest
south. Block number ten was placed on a separate bench west of the main bench
where blocks 1-9 were placed (Figure 3.1). This block was the furthest away from
the greenhouse door. As a anecdotal observation, the location furthest from the door
provided the most stable temperature conditions, which is consistent with a higher
number of seeds germinating in block number ten (Table 4.5).
Logistic regression analysis of the block order as a predictor of seed
germination success indicates that the additive effect on the log odds ratio of seed

germination success increases by 0.22 with each one unit increase in block number.
The odds ratio (eABeta) is the ratio of odds for a 1 -unit change in block number, or
the multiplicative change in odds for each 1 -unit change in block number. Thus, the
odds ratio of seed germination success was 1.25 for each unit increase in block
number, and 9.03 over all ten blocks. This clearly indicates that the closer the blocks
were to the southern end of the greenhouse, the more seeds germinated. Anecdotal
observations showed that on 16 February 2006, 13 days after the seeds were planted,
the temperature in the greenhouse dropped to approximately 25 C. For ten days
following this temperature drop, mold developed in pots within blocks on the north
end of the greenhouse. By the end of February 2006, the temperature returned to 31
C, which was the greenhouse normal. On 5 June 2006 the temperature increased to
55 C, and did not return to 30 C until the first week in July. From 5 June to 1 July
2006, no seeds germinated. It was also observed that there appeared to be a cool spot
in the middle of the greenhouse, which is consistent with the low number of seeds
that germinated in block number five (Table 4.5). Its my opinion that the
temperature variation I observed negatively impacted seed germination success in this
To investigate seed germination success further by cache type, logistic
regression analysis was performed and showed that cache type (surface versus buried)
and number of seeds planted (1 versus 2) showed no significant predictive value for

seed germination success. However cache type (surface versus buried) and number
of seeds planted (1 versus 2) appeared to be significantly correlated (Pearsons
product-moment correlation, r = -47.507, df = 725,/7-value < 2.2e-16, 95 %
confidence interval {-0.9 -0.9}). The negative correlation value reflects a negative
linear relationship, such that as germination success by cache type increased, the
number of seeds germinating decreases. To further analyze these results, multivariate
logistic regression analysis was conducted but still indicated that only block treatment
had an effect on seed germination success. It is likely that the block treatment effect
is too dominant to discern any effects of cache type and seed number. Modeling and
the two predictors (surface versus buried) and number of seeds planted (1 versus 2)
showed high correlation or confounding effects.

These studies of southwestern white pine seed dispersal in the Chiricahua
Mountains of southeastern Arizona suggest that nocturnal rodents may be important
seed dispersers and frequently contribute to southwestern white pine regeneration in
this region.
As stated in the results, the majority of seed caches located in this study
contained 1-2 seeds, and buried caches were placed under soil at an average depth of
1.2 cm (0.33 cm), which is similar to and supports the findings of Vander Wall et
al. (2001), who reported that deer mice scatter-hoard seeds in small caches of 1-2
seeds at depths ranging from 1.2 to 12 mm. This study found on average 1.7 seeds in
surface caches and 1.5 seeds in buried caches with the outlier datum removed,
somewhat similar to the findings of Tomback et al. (2005) who reported that
nocturnal rodent caches contain on average 1.6 seeds in surface caches and 4.4 seeds
in buried caches. Whereas the results from Tomback et al. (2005) and this study are
very similar, it is important to report that the sample sizes in this study were
somewhat less robust.
Seeds collected in 2003 from the San Juan National Forest were not planted
until 2005; some seeds may have lost viability in storage. However, simulated
rodent-caches, like those observed in the field, did germinate in the greenhouse
experiment. Buried caches in this study germinated slightly but not significantly

more often than surface caches, which support the findings of Tomback et al. (2005),
who reported that buried caches had significantly higher germination success in their
field experiments. However, the germination results from this study, albeit conducted
in the greenhouse, had somewhat different outcomes compared to Tomback et al.
The proportion of caches that germinated did not differ significantly between
surface and buried caches or between one-seed and two-seed caches. These results
differ from those of Powell (1996), who reported mixed germination success and
seedling survival; caches with eight seeds had higher success rates than single-seed or
four-seed caches. Comparison to previous germination experiments is difficult,
because field-germination experiments in this study failed, and greenhouse
germination experiments had non-significant results. Additionally the linear block
effect was so pronounced that it dominated any other more subtle effects. Future
studies would benefit from conducting germination experiments in the field, which
could aid in making inference about seed germination success.
Additional seed dispersal studies were conducted on September 20 and 21,
2003, at the Williams Creek trailhead, San Juan National Forest, southwestern
Colorado, in an old growth southwestern white pine forest community. Rodents did
not take any seed from seed stations, although there was a southwestern white pine
cone crop that year in the area. Samano (2001) reported that Clarks nutcracker is the

primary seed disperser in that area, but that Stellers Jays (Cyanocitta stelleri) and
diurnal ground foraging rodents play an important role in secondary seed dispersal.
There does not appear to be any logical conclusion explaining the lack of results I
experienced, unless nocturnal rodents were absent in that forest community.
Studies examining the fate of secondarily dispersed seeds have been criticized
for being too narrow in scope (Wang & Smith, 2002) or for misinterpreting results
(Vander Wall, et. al., 2005). Studies of secondary seed dispersal, sometimes referred
to as Phase II dispersal (Vander Wall & Longland, 2004), greatly contribute to the
body of knowledge being compiled on seed dynamics, plant fitness, and forest
regeneration (Vander Wall, Kuhn, & Gworek, 2005). Ideally, secondary dispersal
moves a seed to a microsite that offers a disproportionately high probability of seed
germination and seedling establishment. To fully understand the regeneration
processes of southwestern white pine, future studies should focus on tracking the fate
of seeds from primary dispersal through establishment. Only then would we
understand how some seeds escape predation.
Seeds of southwestern white pine are anatomically suited to secondary
dispersal, as they are large, wingless, and drop to the ground once cones open. The
seed dispersal mechanism may, however, vary geographically depending on the
availability of the primary disperser, Clarks nutcracker. Clarks nutcracker and
Stellers Jays, to a lesser extent, were observed by Samano and Tomback (2003)

caching seeds of southwestern white pine in the San Juan Mountains; however,
Clarks nutcracker is a very irregular visitor to the Chiricahua Mountains. Seed
dispersal interactions between Clarks nutcracker and limber and southwestern white
pine are secondarily coevolved (Tomback, 1983). Nutcrackers expanded their range
upon arrival in North America and encountered these pines (Tomback, 1983). Given
the geographic dynamics of the primary disperser, it is likely that in Chiricahua
Mountains, and possibly range-wide, southwestern white pine has adapted a dual
dispersal strategy. Differences in cone morphology that I observed in the two areas
were evident; the cones from the San Juan Mountains were larger and more resinous
than the cones from the Chiricahua Mountains. The San Juan Mountains population
may have adapted to heavy seed predation by the red squirrel and seed dispersal by
Clarks nutcracker. Also, it is probable that the cone-to-seed mass ratio is higher, and
that the cones produce more resin to protect against squirrels (Benkman et al, 1984).
Alternatively, the San Juan Mountains population could be hybridizing with limber
pine, which is absent in the Chiricahua Mountains (Samano and Tomback, 2003).
Thompson (1994) has developed a model showing that coevolutionary
relationships between species may vary geographically (The Geographical Mosaic
Theory of Coevolution) with one species of disperser replacing another. This study
supports his model: deer mice and other nocturnal rodents may replace Clarks
nutcracker as the major seed dispersers in the Chiricahua Mountains, and, possibly in

most of the range of southwestern white pine, which is south of the reliable range of
Clarks nutcracker.
In conclusion, the results from this study suggest that nocturnal rodents
actively contribute to the regeneration of southwestern white pine in the Chiricahua
Mountains, and that this bird-dispersed pine has a dual-strategy for seed dispersal.
Future studies could benefit from expanded seed cache sample sizes, remote operated
camera observations, and in the specific case of southwestern white pine,
comparisons with limber pine studies.

Abbott, H. G. (1961). White pine seed consumption by small mammals. Journal of
Forestry, 59, 197-201.
Abbott, H. G., & Quink, T. F. ) 1970). Ecology of eastern white pine seed caches
made by small forest mammals. Ecology, 51(2), 271-278.
Benkman, C. W., Baida, R. P., & Smith, C. C. (1984). Adaptations for seed dispersal
and the compromises due to seed predation in limber pine. Ecology, 65(2),
Boyer, W. D. (1964). Longleaf pine seed predators in southwest Alabama. Journal
of Forestry, 62, 481-484.
Brink, C. H., & Dean, F. C. (1966). Spruce seed as a food of red squirrels and flying
squirrels in interior Alaska. Journal of Wildlife Management, 30(3), 503-512.
Engelmann, G. (1848). Botanical Index. In F. A. Wislizenus, Memoir of a tour to
northern Mexico, connected with Col. Doniphans expedition, 1846 (pp. 87-
115). Washington, DC: Tippin and Streeper.
Chambers, J.C. & J. A. MacMahon. (1994). A day in the life of a seed: movements
and fates of seeds and their implications for natural and managed systems.
Annual Review of Ecological Systems, 25:263-292.
Critchfield W. B., & Little, E. L., Jr. (1966). Geographic distribution of the pines of
the world. Washington, DC: United States Department of Agriculture,
Miscellaneous Publication 991,
Gashwiler, J. S. (1969). Deer mouse repopulation of a poisoned Douglas-fir clearcut.
Journal of Forestry, 67, 494-497.
Hall, E. R., & Kelson, K. R. (1959). The mammals of North America. New York:
Ronald Press.
Janzen, H. D. (1971). Seed predation by animals. Annual Review of Ecological
System, 2, 456-492.

Krugman, S.L. & J.L. Jenkinson. (1974). Pinaceaepine family. In: Schopmeyer,
C. S., technical coordinator. Seeds of woody plants in the United States.
Agric. Handb. 450. Washington, DC: U.S. Department of Agriculture, Forest
Service: 598-637.
Lanner, R. M. (1996). Made for each other: a symbiosis of birds and pines. New
York, NY. Oxford University Press.
Lemen, C. A., & Freeman, P. W. (1985). Tracking mammals with fluorescent
pigments: A new technique. Journal of Mammalogy, 66(1), 134-136.
Longland, W. S., & C. Clements. (1995). Use of fluorescent pigments in studies of
seed caching by rodents. Journal of Mamma logy 76(4): 1260-1266.
Perry, Jesse P., Jr. (1991). The pines of Mexico and Central America. Portland, OR:
Timber Press. 231 p.
Powell, M. L. (1996). A greenhouse study of the effects of cache size and relatedness
on seed germination, seedling survival, robustness, and root grafting in
whitebarkpine. Unpublished masters thesis, University of Colorado at
Radvanyi, A. (1970). Small mammals and regeneration of white spruce forests in
western Alberta. Ecology, 51(6), 1102-1105.
Ramsey, F. L., & D. W. Schafer. (2002). The statistical sleuth: a course in methods
of data analysis, 2nd ed. Pacific Grove, CA. Duxbury.
Reynolds, H. G. (1950). Relation of Merriam kangaroo rats to range vegetation in
southern Arizona. Ecology, 31(3), 456-463.
Riser, J. P. (2003). Seed dispersal and stand structure in southwestern white pine
(pinus strobiformis Engelmann). Unpublished masters thesis, University of
Colorado at Denver.
Samano, S., Tomback, D.F., Maddox, D.N., & Pruett, E.L.. (unpublished).
Southwestern White Pine seed characteristics. University of Colorado at
Denver and Health Sciences Center.

Samano, S. (2001). Seed dispersal of southwestern white pine (Pinus strobiformis).
Unpublished masters thesis, University of Colorado at Denver.
Samano, S., & Tomback, D. F. (2003). Cone opening phenology, seed dispersal, and
seed predation in southwestern white pine (Pinus strobiformis) in southern
Colorado. Ecoscience, 10(3), 319-326.
Shellhammer, H. S. (1966). Cone-cutting activities of Douglas squirrels in sequoia
groves. Journal of Mammalogy, 47(3), 525-526.
Smith, C. C., & Reichman, O. J. (1984). The evolution of food caching by birds and
mammals. Annual Review of Ecological Systems, 15,329-351.
Southwest Parks and Monuments Association. (1998). A Checklist of the Trees and
Shrubs of Chiricahua National Monument, Arizona. Tucson, AZ.
Spofford, S. H. (1973). Some notes on Clarks nutcracker at a banding station in
southeastern Arizona. Western Bird Bander, 48:57-58.
Steele, R. (1990). Pinus flexilis James. In R.M. Bums, & B.H. Honkala [tech.
coords.], Silvics of North America: Vol. 1. Conifers. Agriculture Handbook
654. U.S. Department of Agriculture, Forest Service, Washington, DC. vol.2,
877 p.
Steinhoff, R.J. & J. W. Andresen. (1971). Geographic variation in Pinus flexilis and
Pinus strobiformis and its bearing on their taxonomic status. Silvae Genetica,
Tevis, L., Jr. (1953). Effects of vertebrate animals on seed crop of sugar pine.
Journal of Wildlife Management, 17(2), 128-131.
Thompson, J. N. (1994). The revolutionary preocess. Chicago: University of
Chicago Press.
Tomback, D. F. (1982). Dispersal of whitebark pine seeds by Clarks nutcracker: A
mutualism hypothesis. Journal of Animal Ecology, 51(2), 451-467.

Tomback, D. F. (1983). Nutcrackers and pines: coevolution or coadaptation. InN.
H. nitecki (Ed.), Coevolution (pp. 179-223). Chicago: University of Chicago
Tomback, D. F. (1992). Ecological relationship between Clarks nutcracker and four
wingless-seed Strobus pines of western north America. In W. C. Schmidt &
F. K. Holtmeier (Eds.), Proceedings of the international workshop on
subalpine stoen pines and their environment: The staqte of our knowledge
(pp._221-224). Ogden: United States Department of Agriculture Forest
Service Intermountain Research Station INT-6TR-309.
Tomback, D. F. (1998). Clarks Nutcracker (Nutfraga Columbiana). The Birds of
North America (Number 331). Washington, DC: Academy of Natural
Sciences, Philadelphia, PA, and the American Ornithologist Union.
Tomback, D. F. (2001). Clarks Nutcrakcer: agent of regeneration. In: D.F.
Tomback, S. F. Amo, & R. E. Keane (Eds.), Whitebackpine communities,
ecology and restoration. Washington, DC: Island Press.
Tomback, D. F., Schoettle, A.W., Chevalier, K. E., & Jones, C. A. (2005). Life on
the edge: Collapse of a seed dispersal mutualism, (for submission to
Vander Wall, S. B. (1990). Food hoarding in animals. Chicago: University of
Chicago Press.
Vander Wall, S. B. (1992). The role of animals in dispersing a Wind-Dispersed
pine. Ecology, 73(2), 614-621.
Vander Wall, S. B. (1994). Removal of wind-dispersed pine seeds by ground-
foraging vertebrates. Oikos, 69, 125-132.
Vander Wall, S. B. (1997). Dispersal of singleleaf pinon pine (Pinus monophylla) by
seed-caching rodents. Journal of Mammalogy, 75(1), 181-191.
Vander Wall, S. B., Kuhn, K. M., & Beck, M. J. (2005). Seed removal, seed
predation, and secondary dispersal. Ecology, 86(3), 801-806.

Vander Wall, S. B., Kuhn, K. M., & Gworek, J. R. (2005). Two-phase seed
dispersal: Linking the effects of frugivorous birds and seed-caching rodents.
Oecologia, 145,282-287.
Vander Wall, S. B., & Longlond, W. S. (2004). Diplochory: Are two seed
dispersers better than one? TRENDS in Ecology and Evolution, 19(3), 155-
Vander Wall, S.B., Thayer, T. C., Hodge, J. S., Beck, M. J., & Roth, J. K. (2001).
Scatter-hoarding behavior of deer mice (Peromyscus maniculatus). Western
North American Naturalist, 61( 1), 109-113.
Vander Wall, S. B., & Baida, R. P. (1981). Ecology and evolution of food-storage
behavior in conifer-seed-caching corvids. Zeitschrift tur Tierpsycology, 56,
Wang, B. C., & Smith, T. B. (2002). Closing the seed dispersal loop. TRENDS in
Ecology and Evolution, 17(8), 379-385.
Whitaker, J. O., Jr. (1996). National Audubon Society field guide to North American
mammals. New York: Chanticleer Press.