Citation
Post-fire regeneration of "Pinus albicaulis" in western Montana

Material Information

Title:
Post-fire regeneration of "Pinus albicaulis" in western Montana patterns of occurrence and site characteristics
Creator:
Sund, Sharren Kay
Publication Date:
Language:
English
Physical Description:
63 leaves : illustrations, map ; 28 cm

Subjects

Subjects / Keywords:
Whitebark pine ( lcsh )
Forest regeneration -- Case studies -- Bitterroot National Forest (Idaho and Mont.) ( lcsh )
Forest regeneration ( fast )
Whitebark pine ( fast )
United States -- Bitterroot National Forest ( fast )
Genre:
Case studies. ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )
Case studies ( fast )

Notes

Bibliography:
Includes bibliographical references (leaves 54-61).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Arts, Department of Integrative Biology.
Statement of Responsibility:
by Sharren Kay Sund.

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:
19817844 ( OCLC )
ocm19817844
Classification:
LD1190.L45 1988m .S86 ( lcc )

Downloads

This item has the following downloads:


Full Text
POST-FIRE REGENERATION OF Pinus albicaulis
IN WESTERN MONTANA:
PATTERNS OF OCCURRENCE
AND SITE CHARACTERISTICS
by
Sharren Kay Sund
B.A., University of Colorado, 1976
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Arts
Department of Biology
1988


This thesis for the Master of Arts degree by
Sharren Kay Sund
has been approved for the .
Department of
Biology
by
Emily L. Hartman
Janis W. Driscoll
Date


Sund, Sharren Kay (M.A., Biology)
Post-Fire Regeneration of Pinus albicaulis in Western
Montana: Patterns of Occurrence and Site Characteristics
Thesis directed by Associate Professor Diana F. Tomback.
In 1961, a severe fire in the Bitterroot
National Forest, western Montana, destroyed about 11,350
hectares of subalpine forest, which included Pinus
albicaulis (whitebark pine). The devastated area became
known as the Sleeping Child Burn. Because of the recent
decline of populations of P. albicaulis, a major
subalpine component throughout the northern Rocky
Mountains (Arno 1986), regeneration of P. albicaulis in
the burn is a general concern. Unlike most pines, P.
albicaulis relies* on a bird, Clark's nutcracker
(Nucifraga Columbiana), for seed dispersal. In 1987,
forest regeneration within the Sleeping Child Burn was
studied on 77 plots along two transects: 1) an east-west
ridge exhibiting a gradient for both elevation (range
2148 m to 2182 m) and distance from a P. albicaulis
seed source (range 50 m to 3650 m)
(n = 63 plots), and 2) a north-south transect of greater
distance from a seed source but relatively constant
elevation (n = 14 plots).
Results indicated that nutcrackers have
dispersed clusters of P. albicaulis seeds up to 8 km
into the burn. On the ridge transect, a significant


11
negative correlation was found between P. albicaulis
regeneration density and distance from seed source. A
significant positive correlation was also found between
density and elevation; it was not possible to separate
the effects of elevation and distance. Neither of these
correlations were significant on the north-south
transect.
Regeneration patterns show that Pinus albicaulis
and Pinus contorta are dominating the post-fire
succession on the east-west ridge. The regeneration
status of P. albicaulis in the Sleeping Child Burn
indicates an advantage over its associates, Abies
lasiocarpa and Picea engelmannii, and perhaps a return
to the successional upper subalpine forest communities
that were widespread prior to the fire suppression
policies of this century.
The form and content of this abstract are approved,
recommend its publication.
7
Signed
Diana F. Tomback


iii
ACKNOWLEDGEMENT
This work was supported by funds from the United
States Department of Agriculture, U.S. Forest Service to
Diana F. Tomback, whose guidance and encouragement are
much appreciated. I thank Steve Arno, Jim Brown and Bob
Keane of the Fire Sciences Laboratory, Forest and Range
Experiment Station, Missoula, for supplying equipment
and assistance in the field. Michael Reuss provided the
use of computer hardware and software, assisted with
programming, and was a constant source of encouragement.
A special thanks to Lyn Hoffmann for weeks of field
work, painstaking data recording, plant identifications,
and helpful critiques.


CONTENTS
CHAPTER
I. INTRODUCTION .............................. 1
Pinus albicaulis ........................ 3
Fire History in the Northern
Rocky Mountains ....................... 5
Seed Dispersal Mechanisms ............... 6
Sleeping Child Burn ..................... 9
Hypotheses ............................. 10
II. METHODS .................................. 11
Study Area ............................. 11
Study Site #1 Ridge .................. 12
Study Site #2 Road.................... 15
Unburned Forest ........................ 15
Field Methods .......................... 17
Statistical Methods .................... 19
III. RESULTS .................................. 21
Plot Descriptions (Elevation,
Distance, Slope, Area) ............... 21
Conifer Species and Site Counts ........ 23
Understory Plant Species Occurrence
on Plots ............................. 24
P. albicaulis Clusters and Microsite
Vegetation ........................... 27
Nearest Objects on P. albicaulis
Microsites
31


V
Vegetation Comparisons Between Plot
Series and P. albicaulis Microsites .. 31
P. albicaulis Site Density,
Elevation and Distance ...... 32
Regression Model for Site Density vs.
Distance ............................. 34
P. albicaulis Tree Density,
Elevation and Distance ...... 36
Regression Model for Tree Density vs.
Distance ............................. 38
P. albicaulis Tree Age, Height
and Distance ......................... 38
IV. DISCUSSION......................... 42
Regeneration: Consequences of Seed
Dispersal by Nutcrackers ............. 42
Variation in Density and Cluster Size .. 46
Post-Fire Disturbance .................. 48
Vegetation Comparisons Between Plots
and Microsites ....................... 49
Fire and P. albicaulis Regeneration .... 49
V. CONCLUSIONS ........................ 52
LITERATURE CITED ................................ 54
APPENDIX
A. Plant Species List for Sleeping
Child Burn ................................ 62


vi
TABLES
Table
1. Plot series characteristics: aspect,
elevation, distance, slope,
area............................. . ....... 22
2. Pinus albicaulis clusters: no. of
seedlings and frequency....................... 28
3. Pinus albicaulis regeneration by plot
series: density, age, height.................. 33
4. Pearson correlation analysis for Pinus
albicaulis: density vs. elevation, density
vs. distance from seed source, age vs.
height....................................... 35


vii
FIGURES
Figure
1. Study site #1: Sleeping Child Burn,
Ridge transect............................... 13
2. Study site #2: Sleeping Child Burn,
Road transect................................. 16
3. Five most frequently occurring plant
species on plots, by plot series.............. 25
4. Five most frequently occurring plant
species on P. albicaulis microsites, by
plot series................................... 30
5. Scatterplot of P. albicaulis tree density
vs. distance from seed source, for three
ridge plot series combined.................... 39
6. Scatterplot of P. albicaulis log10 tree
density vs. distance from seed source, for
three ridge plot series combined.............. 40


CHAPTER I
INTRODUCTION
Recently, there is concern that Pinus albicaulis
(whitebark pine) a major component of subalpine forests
in the northern Rocky Mountains, is declining in some
regions (Arno 1986). Reduction of the species results
from a combination of fire suppression policy, insects,
and disease. While not an important timber species, P.
albicaulis is now recognized as a valuable food source
for wildlife because of its nutritious seeds (Tomback
1978; Kendall 1980a,b; Hutchins and Lanner 1982; Arno
1986). Also, it is the only North American member of
the Cembrae subsection of the Strobus pines; species of
this subsection are characterized by indehiscent cones
and large, wingless seeds (Mirov 1967; Mirov and
Hasbrouck 1976). All Cembrae pines are dependent upon
birds of the genus Nucifraga ("nutcrackers") for seed
dispersal (Lanner 1980; Tomback 1983; Tomback and
Linhart, in prep.).
Arno (1986) describes the effect of fire
policies on P. albicaulis. Large fires have played an
important ecological role in the perpetuation of P.


2
albicaulis, with surface fires allowing the tree to
regenerate in small openings within mixed coniferous
forest and severe, stand-replacing fires giving P.
albicaulis regrowth an initial advantage over more
shade-tolerant species such as Abies lasiocarpa. Fire
suppression beginning in the early 1900's has severely
limited the size of wildfires and lengthened their
intervals, thus reducing the rate at which P. albicaulis
is rejuvenated. Two additional factors exacerbate the
problem: insects and disease. Infestations of mountain
pine beetle (Dentroctonus ponderosae) have destroyed
hundreds of thousands of hectares of P. albicaulis
forests in the northern Rockies since 1904. Older
stands of P. albicaulis and its associate Pinus contorta
are favored by mountain pine beetles, and older stands
become widespread with longer fire intervals. A disease
introduced from Europe, white pine blister rust
(Cronartium ribicola), is also a serious threat to P.
albicaulis in cool, moist mountain regions such as in
western Montana, Idaho, and Washington (Arno 1986, Arno
and Hoff 1989).
This study investigated the regeneration of P.
albicaulis after a severe fire in 1961 that resulted in
the devastation of over 11,000 hectares of subalpine
forest in the Sapphire Range, Bitterroot National
Forest, western Montana. The patterns of P. albicaulis


3
regeneration and competition from forest associates are
of interest to forest managers who are concerned about
losses of successional communities of P. albicaulis,
once abundant in this area.
Pinus albicaulis
The following information is summarized from
Arno (1986) and Arno and Hoff (1989). P. albicaulis is
a haploxylon (soft) pine that ranges from British
Columbia south to central California and east to western
Wyoming, occurring on exposed slopes and ridges of high
mountains. Found only at timberline in the north end of
its distribution, P. albicaulis increases in abundance
at both timberline and upper subalpine as latitude
decreases. In west-central Montana, it is a major
component of high elevation forests and the timberline
zone between 2130 and 2830 meters. The boundary between
the lower and upper subalpine is considered to be the
lower limit of P. albicaulis as a major forest species
(Pfister et al. 1977). This slow growing, long-lived
(as much as 1000 years) conifer reaches heights of 5 -
15 meters and often exhibits a multi-trunk growth form.
P. albicaulis occurs on coarse, rocky soils in climatic
zones characterized by short, cool summers and cold,
snowy winters. In mesic mountain ranges it favors warm
dry exposures, while in semi-arid mountains it is more


4
abundant on cool, moist slopes.
Principally a serai component of mixed
coniferous forest in the upper subalpine, and a pioneer
on recently burned or disturbed sites, P. albicaulis is
also a climax species on xeric sites and in severe
timberline zones (Fischer and Bradley 1987). However,
in the Rocky Mountains Pinus albicaulis usually
associates with Picea engelmannii, Tsuga mertensiana,
and Larix lyalli at high elevations bordering timberline
(Wright and Bailey 1982), and with Picea engelmannii,
Abies lasiocarpa, and Pinus contorta at lower
elevations. On some sites bordering timberline, P.
albicaulis, P. engelmannii, and A. lasiocarpa may end up
as climax co-dominants (Arno and Hoff 1989). But in
most subalpine areas, P. albicaulis is eventually
replaced by A. lasiocarpa with P. engelmannii and P.
contorta as minor components. P. albicaulis is
considered to be moderately shade-tolerant. It is less
tolerant than A. lasiocarpa or P. engelmannii but more
tolerant than P. contorta (Arno and Hoff 1989).
As a successional species in subalpine
communities, P. albicaulis is periodically regenerated
by fire. The order of fire-resistance or tolerance is
P. contorta > P. albicaulis > P. engelmannii > A.
lasiocarpa (Fischer and Bradley 1987). Dry exposed
sites and open stands offer P. albicaulis more


5
protection from fire than does its moderately thin bark
(Mirov 1967; Mirov and Hasbrouck 1976). P. albicaulis
trees shed their lower branches as they grow, preventing
vertical fuel continuity that can cause surface fires to
become crown fires (Romme 1982).
Fire History in the Northern
Rocky Mountains
In coniferous forests of the northern Rocky
Mountains, fire regimes vary from relatively frequent,
light surface fires that are localized in effect, to
severe crown and surface fires that burn large areas at
100 to 300 year intervals (Arno 1976, Heinselman 1978,
Romme 1982, Wright and Bailey 1982, Arno and Petersen
1983). Although the calculated lengths of fire
intervals vary somewhat with the method used to
determine them, researchers agree that major changes
have occurred in natural fire regimes due to human
intervention.
Fire history in the Bitterroot National Forest
is summarized by Arno and Petersen (1983). Depending
upon the severity of fire and the sample size unit,
estimates of fire recurrence intervals in the Bitterroot
National Forest during the 300 years prior to 1910 range
from a mean of 20 to 60 years on lower subalpine slopes,
and 40 to 100+ years on upper subalpine. In contrast,


6
the approximate present fire cycle may be greater than
1000 years in both zones. The longer fire intervals are
causing the dominant overstory tree species to be
affected. Prior to 1900, the dominant overstory in the
upper subalpine was Pinus albicaulis, followed by Pinus
contorta. However, with continued fire exclusion, Abies
lasiocarpa is predicted to dominate. In the lower
subalpine, P. contorta and Pseudotsuga menziesii are
being replaced by A. lasiocarpa.
Seed Dispersal Mechanisms
Several seed-dispersal mechanisms are found
among conifers. Some cones of Pinus contorta release
seeds upon ripening, while other cones are serotinous
and adhere to the tree for several years, until the heat
of fire opens cones and broadcasts seeds (Lotan 1975,
1976). All the cones of Abies lasiocarpa and Picea
engelmannii open when ripe, disseminating seeds. Pinus
contorta, Abies lasiocarpa, and Picea engelmannii are
anemochores, or wind-dispersed species. These three
species generally exhibit a U-shaped seed dispersal
distribution across forest openings, with the quantity
of seeds dispersed being greatest at the windward and
leeward edges of clearings and decreasing towards the
center. The dispersal curve shape for these three
species is a negative exponential curve (McCaughey et


7
al. 1986).
Pinus albicaulis, in contrast, is one of about
eight pines in the world known to be dependent upon
nutcrackers and their relatives (family, Corvidae) for
dissemination (Lanner 1980; Tomback 1983). The
indehiscent cones, i.e., cones that do not open upon
ripening, contain large, wingless seeds. Clark's
nutcracker (Nucifraga Columbiana) is the primary
dispersal agent of P. albicaulis (Tomback 1982, 1983;
Hutchins & Lanner 1982). Other western pines, P.
flexilis and the pinyon pines (P. monophylla and P.
edulis), are also dependent on nutcracker dispersal
(Vander Wall & Baida 1977; Lanner & Vander Wall 1980).
The pinyon pines (Pinus edulis) are also dispersed by
pinyon jays (Gymnorhinus cyanocephalus) (Ligon 1978).
This relationship has a European counterpart; four
species of Cembrae pines of Eurasia are disseminated by
the Eurasian nutcracker (Nucifraga caryocatactes)
(Turcek and Kelso 1968). Tomback (1983) emphasizes how
closely the ecology of the Eurasian Cembrae pines and P.
albicaulis are tied to the two species of nutcrackers;
these mutualistic associations are most likely co-
evolved.
After Pinus albicaulis seeds have ripened and
hardened, nutcrackers pry whole seeds from the cones
with their bills and transport the seeds in a sublingual


8
pouch for distances of up to 12.5 km (Tomback 1982).
Pouch contents can average 77 P. albicaulis seeds
(Tomback 1982). Vander Wall and Baida (1977) report
nutcrackers carrying pinyon seed distances of 22 km.
Nutcrackers disperse pine seeds by burying clusters or
"caches" of seeds in the soil, from 1-15 seeds per
site with a mean of 3.7; many of these seeds are
retrieved by nutcrackers during winter and spring when
other foods are scarce (Tomback 1982). Bird dispersal
affects the growth form of the trees and genetic
structure of P. albicaulis populations (Linhart and
Tomback 1985, Furnier et al. 1987). A multi-trunk or
clustered growth form is common in P. albicaulis and a
few other species dispersed by nutcrackers.
Electrophoresis shows multiple trunks in adult P.
albicaulis often to be composed of > 1 genotype, the
result of germination from seed caches (Linhart and
Tomback 1985; Furnier et al. 1987). Nutcrackers usually
harvest seeds from only few trees before caching, so
several of the seeds in a cache may be full or half-
siblings. Even though stems within a cluster may be
genetically distinct individuals, the degree of
relatedness within a cluster is greater than between
neighboring clusters, as caches in one area may be the
result of more than one bird caching (Furnier et al.
1987).


9
Sleeping Child Burn
In August 1961, a lightning-strike fire burned
about 11,350 hectares in the Sapphire Range of the
Bitterroot National Forest in western Montana. Within
a few days plant crowns were totally destroyed and the
organic layer reduced to mineral ash. The devastated
area became known as the Sleeping Child Burn (Lyon and
Stickney 1976). The Bitterroot National Forest, on the
eastern boundary of the Selway-Bitterroot Wilderness, is
divided by the Bitterroot Valley into the Sapphire Range
on the east and the Bitterroot Range on the west. It is
an area of steep, rugged mountains with dense forest.
Upper subalpine forest grows above 2130 m (Arno et al.
1985). Surface geologic formation is of granitic
origin, with medium to coarse-textured, shallow, rocky
soils; ancient wind-deposited ash and loess have
resulted in the leeward slopes (north and east aspects)
having deeper soil mantles (Arno et al. 1985).
The primary elevational range for Pinus
albicaulis in the Bitterroot National Forest is within
the upper subalpine forests from 2290 to 2620 m (Arno
and Petersen 1983). Forest composition at lower
subalpine elevations (from 2130 m up to 2290 m) is
mixed-conifer species, with occasional islands of P.
albicaulis. The nearest P. albicaulis seed source for


10
the Sleeping Child Burn is considered to be the
continuous stands of cone-bearing P. albicaulis that
occur primarily at elevations above 2257 m, since forest
composition at lower elevations is mixed conifers with
relatively few islands of P. albicaulis (Arno, personal
communication). The habitat type groups for potential
climax on upper subalpine slopes are: 1) Abies
lasiocarpa-P. albicaulis/Vaccinium scoparium, 2) A.
lasiocarpa/Luzula hitchcockii and 3) P. albicaulis-A.
lasiocarpa (Pfister et al. 1977).
Hypotheses
1) Because Pinus albicaulis is a pioneer species
and because of the 26 year time period following the
fire, regeneration of this conifer should be advanced in
the Sleeping Child Burn. 2) The patterns of P.
albicaulis regeneration should be consistent with
patterns of avian seed dispersal.


CHAPTER II
METHODS
Study Area
The Sleeping Child Burn covers an area about
14.5 km by 8 km in Ravalli County, approximately 80 km
south-southeast of Missoula, Montana. From Hamilton,
the burn is reached by travelling east on state highway
38 to Skalkahoe Rye Road. The northern boundary of the
burn extends to Township 4 N and the southern boundary
extends into the north half of Township 2 N. On the
east and west the boundaries extend, respectively, from
Range 18 W into Range 19 W. Located west of the
Continental Divide, this area has an inland maritime
climate. As elevation increases from valley (1220 -
1525 m) to mountaintop (2135 2745 m), the
precipitation increases, and temperatures become cooler.
Distinct differences occur between south and north
aspects (Pfister et al. 1977). The south slopes are
much warmer and drier, whereas north slopes experience
less solar insolation, greater snow accumulations and
deeper soils resulting in cooler, moister conditions.


12
Annual precipitation averages 33 cm and varies with
elevation and topography (Lyon 1976). Seventy percent
of the precipitation above 2135 m is from snowfall.
Early winter and late spring are the wettest seasons;
lightning activity and low relative humidity occur in
July and August, resulting in a fire season. Daily
temperature ranges are greatest at low elevations and
least on mountaintops. Monthly mean temperatures in the
area show a slight decrease on a south-north gradient.
Winds are predominantly from the west or southwest
(Finklin 1983).
Field work was conducted from 3 to 26 August on
two study sites in the burn. The first study was
located along an east-west ridge transect, 3.65 km long,
from Lost Tooth Cabin at the edge of the burn heading
west to Two Meadows, approximately in the northern
center of the burn. The second site was a 20.6 km
north-south road transect following Skalkahoe Rye Road
from the northern edge of the burn, south and then
southwest on Paint Creek Road approximately
through the center of the burn to its southern edge.
Study Site #1 Ridge
The first site (Figure 1) was selected to gather
data on forest regeneration patterns in the northern
center of the burn. Plots were sampled along a straight


13
x 2173 m
----R]DGE_ TR4A/S£cr
SLEEPING CHILD
BURN
0.5 km
---------
Fig. 1. Study site #1: Sleeping Child Burn, Ridge
Transect. The stippled area is the subalpine
forest bordering the burn. The hatchmarked
area indicates a subsection of the forest with
elevations above 2257 m where continuous
stands of P. albicaulis provide the seed
source for regeneration.


14
ridge of gradually decreasing elevation and increasing
distance from the unburned forest at the eastern edge of
the burn.
Plots were established 150 m apart on a 3.65 km
east-west transect. Each ridge plot was paralleled by
plots on the north- and south-facing slopes of the
ridge, thus creating three parallel series of plots:
north, south, and ridge. In some instances, a
representative north or south aspect was not available
or could only be obtained by walking several meters east
or west. A total of 63 plots were studied: 21 north, 19
south, and 23 ridge. All plots were oriented east-west
longitudinally, with belt width measured south from the
transect line. Size of the plots ranged from 50 x 1.25
m to 50 x 12 m. Belt width varied for purposes of
efficiency. It was important to avoid zero densities of
Pinus albicaulis regeneration and yet quickly process
plots where densities were high. On the west end of the
ridge, P. albicaulis regeneration was sparse, so belt
width was increased to 10 m or 12 m. Narrower widths
were used on the east end of the ridge were P.
albicaulis densities were much higher.
Belt widths were identical for all other
species, but were chosen according to density.


15
Study Site #2 Road
This second site was selected to measure forest
regeneration at fairly constant elevation and aspect but
increasingly greater distances from the Pinus albicaulis
seed source. A total of 14 plots were studied on north-
facing aspects along the Skalkahoe Rye Road and Paint
Creek Road (Figure 2). The plots were separated by road
distances of 644 m to 2574 m, and their locations
corresponded to aerial distances of 900 m to 7950 m from
the unburned forest. Plot sizes ranged from 30 x 1.5 m
to 30 x 10 m. Plot width was determined by both the
density of the vegetation and an effort to include at
least one P. albicaulis seedling per plot.
Unburned Forest
For comparison with the burned area on the ridge
study site, two plots in the adjacent unburned forest
were located on north and south aspects at a distance of
150 m east of the first ridge plot. The north and south
forest plots were at elevations of 2440 m and 2420 m,
respectively, and each plot was 500 m2 in area. All
trees within these plots were noted.


16
Fig. 2. Study site #2: Sleeping Child Burn, Road
Transect. The stippled area is the subalpine
forest bordering the burn. The hatchmarked
area indicates a subsection of the forest with
elevations above 2257 m where continuous
stands of P. albicaulis provide the seed
source for regeneration.
i


17
Field Methods
On each plot a list of up to 10 understory plant
species, following the checklist of Hitchcock and
Cronquist (1973) and Dorn (1984), was compiled by
inspection. The purpose was not necessarily to indicate
dominance but to list plants representative of the
habitat on each plot. In August when field work was
conducted, annuals and early-blooming perennials were
past flowering, and thus identification of flowering
plants was difficult. The percentage occurrence among
all plots in a series was calculated for the most common
plant species. Evidence of post-fire disturbance such
as grazing, thinning, woodcutting, etc. was noted.
Plot measurements included: elevation, slope,
aerial distance from Pinus albicaulis seed source,
aspect, and number of regeneration sites. Elevation was
measured by altimeter and USGS 7.5 minute series
topographic map (Kent Peak, and Jennings Camp Creek)
readings rounded to the nearest 5 meters. An estimate
of slope was made to the nearest 10 degrees, and aerial
distance from the boundary of the P. albicaulis seed
source was measured on the topographic map to the
nearest 10 meters. The seed source is considered to be
the continuous stands of P. albicaulis at elevations
above 2257 m within the forest; only isolated islands of


18
P. albicaulis are found in the mixed conifer forest
below this elevation. A compass was used to determine
aspect. Post-fire regeneration, which ranged in age
from recent germinants to saplings and young trees, is
collectively referred to as 'trees'. The number of P.
albicaulis sites on each plot were recorded. The term
'site' is used to refer to any location with either two
trees or a cluster of one or more P. albicaulis trees.
At each P. albicaulis site the following data were
gathered: slope, aspect, identification of 1 to 4
vascular plant species found within 15 cm of the site,
and the nearest object (rock, wood pieces or fallen
snag, etc.) located within 15 cm. Cluster size was
recorded as the number of individual trunks at a single
site and was determined by separating trunks to or below
ground level. Although we looked for evidence of
blister rust on the trees, none was seen.
The height of each individual tree was recorded,
and a number of trees of every representative height
category were aged. On the four plot series within the
burn, tree age was determined by severing the trunk at
ground level and counting tree rings with a hand lens.
Newly germinated trees were given a 0.5 year age value;
all other ages were in whole years. Tree age on the
unburned plots was determined by coring at ground level
and counting tree rings.


19
Statistical Methods
Although a ratio scale of measurement was used
for distance, elevation, and tree density data, non-
parametric statistics were used in some analyses because
underlying parametric assumptions concerning random
sampling, normal population distributions, and
homogeneity of variance were not always met. Our
transects were systematically laid out to obtain plots
at pre-determined orientations (east-west along an
elevational gradient from center of burn to edge of
burn, and north-south along a distance gradient from
edge of burn). Normal population distributions were not
indicated for P. albicaulis tree or site density on any
transect, or for elevation on the north, ridge or south
transects. Heterogeneity of variance was also found for
P. albicaulis site density on the ridge, north and south
plot series (F-max = 19.8). In addition to descriptive
statistics for all plot series, Kruskal-Wallis ANOVA was
used to compare results among the three ridge plot
series (north, ridge, south). Further comparisons were
made with Wilcoxon Rank Sum tests. The non-parametric
statistical tests were selected in accordance with
Seigel (1956).
The independent sampling, ratio measurement
scale, and sample sizes were considered adequate to


20
justify use of parametric analyses for correlation and
regression. Pearson Product Moment Correlation analyses
(Bailey 1981) were used for testing the following:
distance (aerial distance from seed source) vs. tree
density; distance vs. site density (a site may be one
tree or multiple trees in a single cluster); distance
vs. tree age; tree height vs. tree age; elevation vs.
site density, and elevation vs. tree density for all
plot series. North, south and ridge plot series were
combined, and scatterplots with simple linear regression
lines were calculated for tree density vs. distance,
site density vs. distance, log10 of tree density vs.
distance, and log10 of site density vs. distance. The
log10 of the density was plotted to see if linearity
would result. The regression formula was:
Y = bX + a
where, Y = dependent variable, b = slope, a = y
intercept, X = independent variable (Gravetter and
Wallnau 1985). The statistical significance of
regression coefficient r was determined from Appendix 4
in Bailey (1981). Analyses were run on an IBM-
compatible microcomputer with Microstat (version 4.1,
Ecosoft, Inc., Indianapolis, IN). In all tests, results
were considered significant only if P <0.05.


CHAPTER III
RESULTS
Plot Descriptions (Elevation,
Distance, Slope, Area)
For the ridge study site, there were a total of
63 plots; 21 on the north series; 23 on the ridge, and
19 on the south. The general aspect of plots on each
plot series was; north north aspect, range 290 20;
ridge west aspect, range 90 330; south south
aspect, range 170 230; road north aspect, range 330
- 90 (Table 1). All three ridge plot series were
similar in their ranges of elevation and distance to the
P. albicaulis seed source (Table 1). Elevation on the
north, south and ridge plots ranged from 2105 m to 2480
m with means respectively of 2305 m, 2295 m, and 2305 m.
The straight line distance from the eastern edge of
individual plots to the seed source ranged from 50 m to
3650 m. Mean distance for north plots was 1680 m, for
south, 1810 m, and for ridge, 1790 m. The steepest
slopes were on the north aspect (range 0 to 50, mean =
30), the south was moderately steep (range 0 to 40,
mean 20), and the ridge averaged much flatter (range 0


22
TABLE 1. Plot series characteristics: aspect,
elevation, distance, slope, area.
PLOT SERIES
NORTH RIDGE SOUTH ROAD
Aspect Range (deg) north 290-20 west 90-330 south 170-230 north 330-90
No. of plots 21 23 19 14
Elevation Mean Range (m) 2305 2150-2475 2305 2175-2480 2295 2165-2465 2200 2105-2290
Distance from seed source (m)
Mean 1680 1790 1810 3860
Range 50-3500 50-3650 50-3450 900-7950
Slope (deg)
Mean 30 10 20 20
Range 0-50 0-40 0-40 0-30
Area (m2)
Mean 154.8 201.1 228.9 84
Range 62. 5-500 62.5-500 62.5-750 12.5-500
No. of sites
P. albicaulis 271 119 65 45
P. contorta 115 186 135 80
A. lasiocarpa 33 21 6 123
P. engelmannii 19 16 2 38
Total sampling area (m2)
P. albicaulis 3500 5750 6225 2975
Other 3250 4625 4350 1176
(P. contorta,
A. lasiocarpa,
P. engelmannii)


23
to 30, plus one uncharacteristic plot with 40 slope,
mean = 10). For both the ridge and north series, plot
sizes ranged from 62.5 m2 to 500 m2, with a mean of 201.1
m2 and 154.8 m2, respectively. South series plots had a
wider size range (62.5 to 750 m2) and a mean of 228.9
m2.
For the 14 plots on the road transect, elevation
was generally lower than on the ridge series of plots,
between 2105 and 2290 m (mean = 2200 m) (Table 1). The
distances to seed source were much greater, with a range
of 900 to 7950 m, and a mean of 3860 m, over twice that
of the ridge series of plots. The slope on road plots
averaged 20. Road plots were smaller in area than the
ridge series because of shorter plot length and ranged
from 12.5 to 500 m2,
averaging 84 m2.
Conifer Species and Site Counts
Four species of conifers were found on each plot
series (north, south, ridge, road) (Table 1). The
number of sites for P. albicaulis was: north 271,
ridge 119, south 65, road 45. The number of sites
of P. contorta was higher than that of P. albicaulis on
every plot series but the north, despite the smaller P.
contorta sampling area; P. contorta site counts on the
ridge, south, north, and road were 186, 135, 115, and


24
80, respectively. In addition, the total number of
sites of P. contorta (n = 516) was greater than that for
each of the other conifers. The following comparison of
total sites per species is based on sampling areas of
the same size: Pinus contorta > Abies lasiocarpa > Picea
engelmannii. Only on the road plot series were A.
lasiocarpa site counts (n = 123) higher than counts of
P. contorta. P. engelmannii had the lowest number of
sites on every transect (ridge 16, north 19, south -
2, road - 38) relative to A. lasiocarpa and P.
contorta.
A single Pinus ponderosa was found on a south
plot, and one Pseudotsuga menziesii was found on a north
plot.
The two unburned forest plots supported
predominantly P. albicaulis (north 44, south 19) and
A. lasiocarpa (north 38, south 11), with a few Pinus
contorta and Picea engelmannii on the south.
Understory Plant Species Occurrence
on Plots
A comparison of understory plant composition
among plot series shows the north and south plot series
to be very different, reflecting relative occurrence of
xeric and mesic species, and the ridge to be
intermediate between north and south. Figure 3 shows


25
PLOT -
SERIES FREQUENCY OCCURRENCE (%)
10 20 30 40 SO 60 70 80 90 100
N
0
R
T
H
Vaccinium scoparium (95.3)
____ Luzula sp. (85.7)
Xeropyllum tenax (71.4)
Carex geyeri (66.7)

Polytrichum sp. (66.7)
Xerophyllum tenax (100)
R
I
D
G
E
S
0
u
T
H
___________________Carex seyeri (95,6)
Vaccinxum scoparium (78.3)_______
Luzula sp. (73.9)
Festuca sp. (65.2)__________
Carex geyeri (100)
Xerophyllum tenax (100)
Festuca sp. (78.9)

Phleum pratense (78.9)
Vaccinium scoparium (57.9)
Luzula sp. (57.9)
Vaccinium scoparium (78 6)

Polytrichum sp. (78 6)

Xerophyllum tenax (71.4)

Epilobium sp. (64.3)
Luzula sp. (50)
C. geyeri (50)
P. pratense (50)
Fig
3. Five most frequently occurring plant species
on plots, by plot series.


26
the five most frequently occurring plant species by plot
series. These species are dominant species for the
Bitterroot region, and are classified as mesic or meso-
xeric (Lackschewitz 1986). The major vegetation
differences between aspects are the result of
differences in occurrence of common species (Figure 3):
Vaccinium scoparium (north 95.3%, ridge 78.3%, south
- 57.9%); Luzula sp., north 85.7%, ridge 73.9%, and
south 57.9%); Polytrichum sp. (north 66.7%, ridge -
21.7%, south 5.3%); Carex rossii, (north 61.9%,
ridge 52.2%, south 26.3%).
Grasses and other xeric vegetation were more
frequent on the south plot series, less on the ridge,
and even less on the north. Xerophy1lum tenax occurred
on all plots on both south and ridge plot series, and on
71.4% of north plots. Carex geyeri occurred 100%
frequency on south plots, 95.6% on ridge and 66.7% on
the north;. Phleum pratense frequency occurrence was;
north 42.9%, ridge 43.5%, south 78.9%; Dactylis
glomerata occurred on 4.8% of the north plot series,
19.9% of the ridge, and 52.6% of the south.
Road plots showed some similarities to both the
north and ridge plot series. The road plots resembled
the north in order of species frequencies on plots;
Vaccinium scoparium (78.6%) > Xerophyllum tenax (71.4%)
> Carex geyeri (50%) > Anaphalis margaritacea (35.7%) >


27
Salix sp. (28.6%). Frequency values for the road, V.
scoparium (78.6%), P. pratense (50%) and A. margaritacea
(35.7%), were similar to the ridge.
The two unburned forest plots exhibited
understory occurrences of X. tenax, V. scoparium, and
C. geyeri.
P. albicaulis Clusters and Microsite
Vegetation
Cluster sizes (number of individual trees per
P. albicaulis site) ranged overall from 1 to 9 (Table
2). The ridge and north plot series showed a range of
1 to 8, the south 1 to 5, and the road 1 to 9. Mean
cluster size, respectively, was 2.06, 1.92, 1.57, and
2.49. Kruskal-Wallis ANOVA indicated a significant
difference in cluster size among the three ridge plot
series (H = 6.475, df = 2, P = 0.039). Further analysis
with the Wilcoxon Rank Sum test showed south plots
differing significantly both from the north (Z = 1.951,
P = 0.026) and the ridge (Z = -2.543, P = 0.006).
On every transect, the largest percentage of
sites were solitary trees (south 66.2%, north 52%,
road 51.1%, ridge 45.4%). As cluster size
increased, the percentage of clusters in each size
category decreased. One exception was on the road plot
series, where clusters of 3 trees accounted for 22.2% of


28
TABLE 2
Pinus albicaulis clusters: size
and frequency.
PLOT SERIES
NORTH
Number of sites 271
Number of trees per site
Mean 1.92
Range 1-8
Cluster size
1 52.0
2 26.6
3 8.5
4 6.3
5 4.8
6 1.1
7 0.4
8 0.4
9
RIDGE SOUTH ROAD
119 65 45
2.06 1.57 2.49
1-8 1-5 1-9
Frequencies (%)
45.4 66.2 51.1
28.6 21.5 6.7
12.6 4.6 22.2
6.7 4.6 6.1
4.2 3.1 2.2
0.8 4.4
0.8 2.2
0.8 2.2
2.2


29
all sites, and clusters of 2 trees accounted for only
6.7%. A cluster of 17 trees was found near a road plot;
this is larger than the cluster size range of 1 to 15
previously reported by Tomback (1982).
Pinus albicaulis sites on the three different
ridge plot series showed dissimilar frequencies of
microsite vegetation species. Vaccinium scoparium was
most frequent on all plot series, and Polytrichum sp.
or Carex rossii was the second most frequent. Many P.
albicaulis trees were growing in the middle of a clump
of C. rossii. On north and ridge sites, V. scoparium
frequency was 86% and 69.7%, respectively; and all other
species accounted for 25% or less of occurrences. South
plot series P. albicaulis sites showed no microsite
plant species frequency greater than 25% (Figure 4).
Despite the frequent occurrence of Xerophyllum tenax on
many plots, no P. albicaulis trees were actually growing
within these formidable clumps.
On the road transect, V. scoparium and
Polytrichum sp. each occurred on more than 50% of the
P. albicaulis sites. The species richness was lower on
road plots, but this is probably due to the smaller plot
size and thus smaller number of P. albicaulis sites.


30
PLOT
SERIES FREQUENCY OCCURRENCE (%)
10
20
30
40
SO
60 70
80
90
Vaccinium scoparium (86)
N
0
R
T
H
POLYTRICHUM SP. (21.8)
1
Carex ROSSIi (17.7)
Lupinus sp. (14)
Luzula sp. (10)
R
I
D
e
E
Vaccinium scoparium (69.7)
Carex rossii (24.4)
Carex geyeri (8.4)
Xerophyllum tenax (7.6)
Epzlobzum sp. (5)
S
0
u
T
H
Vaccinium scoparium (24.7)
Carex rosszx (23)
Festuca sp. (13.9)
NEHZZESZA FERRtWXNEA (10.8)
Carex geyerz (10.8)
R
0
A
D
Vaccinium scoparium (62.2)
PoLYTRICHUM SP. (55.6)
Epilobium sp. (17.8)
Menziesia ferruginea (13.3)
Carex geyeri (8.9)
Fig. 4. Five most frequently occurring plant species
on P. albicaulis microsites, by plot series.


31
Nearest Objects on P. albicaulis
Microsites
The total number of P. albicaulis microsites on
all four plot series combined (ridge, north, south,
road) was 500. Five categories of objects were noted
on P. albicaulis microsites within 15 cm of the tree
trunk(s). Wood pieces were the nearest object on 122
sites (24.4%), a log or single large branch on 112
(22.4%), a fallen snag on 69 (13.8%), rocks on 68
(13.6%), and a standing snag or stump on 34 (6.8%).
Ninety-five (19%) of the microsites were categorized as
'open', because no object was present.
Vegetation Comparisons Between Plot Series
and P. albicaulis Microsites
Analysis of vegetation near P. albicaulis sites
shows some similarities to the vegetation frequencies on
plot series. Vaccinium scoparium was the most
frequently occurring species on tree microsites on all
plot series, with a frequency on ridge microsites
intermediate between north and south, and a frequency on
the road microsites most resembling that of the north.
Vaccinium scoparium showed the same frequency pattern on
plots, with the exception that the road plots mostly
resembled the ridge plots. Polytrichum, a mesic
species, was the second most frequently occurring


32
species on north and road microsites and showed high
occurrences on north and road plots.
Distinct differences between microsite
vegetation and plot vegetation are reflected in the
frequencies of C. rossii and X. tenax. Although C.
rossii is the second most frequently occurring species
on ridge and south microsites, it ranks much lower in
occurrence on ridge and south plots. Conversely, X.
tenax was present on all south and ridge plots and over
70% of the road and north plots but occurred on no north
microsites and on fewer than 8% of microsites on the
other three plot series.
P. albicaulis Site Density,
Elevation, and Distance
The total number and range per plot of P.
albicaulis sites on each plot series were: ridge total
119, range 0 to 24; north total 271, range 1 to 32;
and south total 65, range 1 to 11 (Table 3). The
ranges of P. albicaulis site density per m2 per plot for
the ridge, north and south plot series, respectively,
were: 0 to 0.192, 0.006 to 0.512, and 0.001 to 0.11.
Mean site density per plot (defined as P. albicaulis
sites per m2) were: north 0.14, ridge 0.04, south -
0.02. Road plots had a total of 45 P. albicaulis sites,
with a range of 0 to 10 per plot, a mean density of 0.02


33
TABLE 3. Pinus albicaulis regeneration by plot
series: density (sites or trees per m2),
age (yrs), height (cm). N = total number.
NORTH PLOT SERIES RIDGE SOUTH ROAD
Sites
N 271 119 65 45
Range per plot 1-32 0-24 1-11 0-10
Density Mean 0.14 0.04 0.02 0.02
Range 0.006-0.512 0-0.192 0.001-0.11 0-0.007
Trees
N 520 245 102 112
Range per plot 1-61 0-56 1-13 0-22
Density Mean 0.275 0.099 0.035 0.051
Range 0.006-0.976 0-0.'448 0.001-0.13 0-0.127
Age
N 334 149 79 47
Mean 7.7 8.3 8.1 6
Range 1-21 0.5-20 1-21 0.5-19
Height
Mean 29.7 36.2 30.6 19.2
Range 1.5-234 3-238 3-202.3 1-186


34
per m2, and a density range of 0 to 0.067 per m2.
Site densities for the north, ridge and south plot
series showed differences of high statistical
significance (Kruskal-Wallis ANOVA, H = 10.614, df = 2,
P <0.005). North plot series site densities differed
significantly from the ridge plot densities (Wilcoxon
Rank Sum, Z = 2.758, P <0.003) and from the south plot
densities (Z = -3.521, P <0.001).
Whether a significant relationship occurred
between P. albicaulis site density (sites per m2) and
elevation, and site density and distance from seed
source, was determined for each plot series (Pearson
correlation, Table 4). Significant negative
correlations with P <0.01 between site density (sites
per m2) and distance occurred on all series of plots
except the, road: north, r = -0.832; ridge, r = -0.618;
south, r = -0.714 (Table 4). Site density and elevation
were significantly correlated for the three ridge plot
series with P <0.01, but not for the road series: north,
r = 0.982; ridge, r = 0.632; south, r = 0.682 (Table 4).
Regression Model for Site Density
vs. Distance
Data from the north, south and ridge plot series
were combined in a scatterplot of Pinus albicaulis site
density (dependent variable) vs. distance from seed


35
Table 4. Pearson correlation analysis: Pinus albicaulis
density (sites or trees per m2) vs. elevation,
density vs. distance from seed source, age vs.
height. (NS = not significant, P >0.05).
Log = Log10. In all cases, df = no. of plots
minus two.
Plot series No. of plots Correlation coefficient(r) P
Site density North 21 0.982 <0.001
vs. elevation Ridge 22 0.632 <0.01
South 19 0.682 <0.01
Road 13 0.428 NS
Tree density North 21 0.877 <0.001
vs. elevation Ridge 22 0.647 <0.01
South 19 0.682 <0.01
Road 13 0.194 NS
Site density North 21 -0.832 <0.001
vs. distance Ridge 22 -0.618 <0.01
South 19 -0.714 <0.001
Road 13 -0.374 NS
Log site density North 21 -0.820 <0.001
vs. distance Ridge 22 -0.733 <0.001
South 19 -0.779 <0.001
Road 13 -0.500 NS
Tree density North 21 -0.822 <0.001
vs. distance Ridge 22 -0.627 <0.01
South 19 -0.703 <0.001
Road 13 -0.207 NS
Log tree density North 21 -0.788 <0.001
vs. distance Ridge 22 -0.751 <0.001
South 19 -0.678 <0.01
Road 13 -0.334 NS
Tree age North 334 0.818 <0.001
vs. height Ridge 149 0.783 <0.001
South 79 0.827 <0.001
Road 47 0.742 <0.001


36
source (independent variable). The plot produced a
negative exponential curve with the following regression
equation:
Y = -0.00006(X) + 0.175
(Df = 60, r = -0.590, P <0.001). When log10 density was
used, the scatterplot produced a straight line and the
regression equation was:
Y = -0.0005(X) 0.806,
with r = -0.692 and P <0.001.
Regression analysis of the road plots indicated
no correlations between P. albicaulis site density and
distance from seed source. Even when the log10 density
was used, the regression was not significant.
P. albicaulis Tree Density,
Elevation and Distance
A total of 1029 Pinus albicaulis trees (more
than one tree may occur at a site) were found on the two
study sites: 917 on the ridge (includes ridge, north and
south plot series) and 112 on the road plot series
(Table 3). Total P. albicaulis tree counts on the north
plot series were twice as large as on the other two
ridge plot series (520 versus 245 for the ridge series
and 102 for the south series), although area sampled on
the ridge was 1.64 times larger than on the north, and
the south was 1.78 times larger. Tree densities,


37
measured as trees per m2, ranged as follows: north
(0.006 to 0.976), ridge (0 to 0.448), south (0.001 to
0.13) (Table 3). Mean tree densities per mz were: north
- 0.275, ridge 0.099, and south 0.035. Kruskal-
Wallis ANOVA showed P. albicaulis tree densities per
plot were significantly different among the north,
ridge, and south plot series (H = 14.191, df = 2, P
<0.001). Wilcoxon Rank Sum results were significant for
comparisons of the north series vs. both the ridge
series (Z = 2.478, P <0.007) and the south series (Z =
3.616, P <0.001). Thus P. albicaulis is significantly
greater on the north than on the other two aspects.
The mean tree density for road plots (0.051
trees per m2) was intermediate between south and ridge
values, and the density range per plot was 0 to 0.127
trees per m2 (Table 3).
Significant negative correlations (P <0.01) were
found for tree density vs. distance from seed source for
all three ridge plot series with r values as follows:
north, r = -.822; ridge, r = -.627; and south, r = -.703
(Table 4). The correlation was not significant for the
road.


38
Regression Model for Tree Density vs
Distance
Combining north, ridge and south plot series,
the scatterplot for tree density against distance also
assumed a negative exponential curve (Figure 5). The
regression equation was:
Y = -0.0001(X) + 0.344,
with df = 60, r = 0.571, and P <0.001. With log10 of
tree density (Figure 6), the plot became linear and the
equation became
Y = -0.0005(X) 0.542
with r = -0.666 and P <0.001.
The same analysis for P. albicaulis trees on the
road plots failed to produce a significant correlation
between density and distance.
P. albicaulis Tree Age, Height,
Distance
Each separate stem within a Pinus albicaulis
cluster was considered a tree for aging purposes; the
number of trees aged on north plots was 334; on the
ridge, 149; and on the south, 79 (Table 3). Field data
show that the ages of trees within an individual cluster
were the same or similar, usually differing by only 1 or
2 years, but in no case more than 4 years. In virtually
all cases no stems were fused at the base, thus it must


DISTANCE (m) 3650
39
Fig. 5. Scatterplot of P. albicaulis tree density vs.
distance, for three ridge plot series
combined.
9Z6 0


Fig. 6. Scatterplot of P. albicaulis log10 tree
density vs. distance, for three ridge plot
series combined.
-0-011


41
be emphasized that each member of a cluster appeared to
be a distinct individual. The age differences observed
were probably the result of counting error or later
germination. Age ranges of trees were nearly identical
for the three ridge plot series: north and south, 1 to
21 yr; ridge 0.5 to 20 yr. Compared to the ridge plots
and south plots, the north trees had a slightly lower
mean (7.7 yr vs. 8.3 yr, and 7.7 yr vs. 8.1 yr). Forty-
seven trees were aged on road plots, with a range of 0.5
to 19 yr and a lower mean (6 yr) than any of the ridge
plot series. P. albicaulis trees on the two unburned
forest plots showed a mean age of 127.7 yr and a range
of 76 to 202 yr.
Tree heights ranged from 1 to 238 cm (Table 3).
The range and mean for the ridge plot series was the
highest (3 to 238 cm, 36.2 cm), followed by the north
(1.5 to 234 cm, 29.7 cm), south (3 to 202.3 cm, 30.6 cm)
and road (1 to 186 cm, 19.2 cm), the lowest of all.
Highly significant Pearson correlations (P <0.001) were
found for tree age vs. tree height for all plot series:
north (r = 0.818, df = 332), ridge (r = 0.783, df =
147), south (r = 0.827, df =77), and road (r = 0.742, df
= 45).


CHAPTER IV
DISCUSSION
Regeneration: Consequences of Seed
Dispersal by Nutcrackers
Regeneration patterns of Pinus albicaulis in the
Sleeping Child Burn demonstrate the importance of
nutcrackers as seed dispersers. A few nutcrackers were
seen in the study site during field work in 1987, but
the lack of a P. albicaulis cone crop that year
prevented direct observation of foraging or caching
behavior. However, the occurrence of tree clusters and
the great distances from regeneration to seed source
confirm previous suggestions that nutcrackers are the
primary dispersers of P. albicaulis and that their
caching behaviors effect establishment of the conifer
(Lanner 1980; Hutchins and Lanner 1982; Tomback 1982).
This study also confirms that squirrels and other
rodents play a very minor role at best in the
dissemination of Pinus albicaulis. As reported by
Hutchins and Lanner (1982), red squirrels (Tamiasciurus
hudsonicus), chipmunks and other mammals are not
reliable dispersers compared to nutcrackers. Their
limited home ranges indicate that a more mobile


43
disperser has buried seeds to account for the observed
distances between seed source and regeneration.
Although Stellar's jays (Cyanocitta stelleri) were seen
in the study area, they also establish small
territories, damage seeds upon removal from the cone,
and bury one-seed caches (Vander Wall and Baida 1981).
Nearly half of all Pinus aibicaulis sites were
composed of two or more trees, with clusters of up to 9
trees observed on the road plot series, and a cluster of
17 trees observed near that transect. The distinct
multiple stems and the similarity in tree ages within a
cluster are evidence of germination from buried seed
caches.
Interestingly, no P. aibicaulis older than 21
yr was found on the plots, even though the burn occurred
26 years prior. This may indicate a lack of
regeneration during the five years immediately following
the burn, either because of no cone crops or unfavorable
conditions for seed germination.
Dispersal distances of P. aibicaulis in the
Sleeping Child Burn are similar to the distances noted
by Tomback (1978) and Hutchins and Lanner (1982),
although their studies occurred in California and
Wyoming, respectively. Regeneration from the seed
source in the adjacent forest extended into the burn
westward for 3.65 km and southward nearly 8 km. Because


44
P. albicaulis trees were found on the study plots
furthest from the seed source, we assume regeneration
has occurred beyond these distances also. These
patterns lead us to conclude that nutcrackers are
responsible for nearly 100% of the observed P.
albicaulis regeneration in the burn.
An interesting pattern with potential
application was the significant linear relationship
observed between the density of regenerating P.
albicaulis on the three ridge plot series and the aerial
distance from the seed source (Figures 5 and 6). This
correlation reflects the dispersal pattern generated by
the nutcracker, but at this time it is not possible
statistically to separate the effect of elevation on
density from the effect of distance on density.
Elevation also correlated significantly with distance
from seed source on the three ridge plot series, and the
change in environmental conditions (moisture,
temperature) that accompany changes in elevation
conceivably may have an effect on seed germination and
seedling survival. However, P. albicaulis grows well in
western Montana from elevations of 2290 2620 m and
grows as far down as 2130 meters (Arno and Petersen
1983). All plots on the three ridge plot series were
above 2130 m and nearly half of the plots on each series
were within the 2290 2620 m range. Thus, the pattern


45
generated should primarily reflect differences in
distance and not elevation effects.
The regeneration patterns that result from
nutcracker seed dispersal along a distance gradient from
a known seed source suggest that the birds cache
relatively more seeds close to the parent trees and make
fewer caches per m2 at increasing distances. The result
is a negative exponential curve (tree density vs.
distance) (Figure 6), a shape similar to the seed shadow
curves generated by individual wind-dispersed conifers
(McCaughey et al. 1986), but the P. albicaulis
regeneration extended over much greater distances. As
discussed by Tomback et al. (in preparation), the
dispersal distances for P. albicaulis may typically be
greater than that for Abies lasiocarpa, a competitor in
succession, giving P. albicaulis the edge because of
nutcracker-mediated dispersal.
The negative exponential curve was confirmed as
well for P. albicaulis regeneration in the Saddle
Mountain Burn to the south of the Sleeping Child Burn
(Tomback et al. in prep.). For management purposes,
this curve may be useful for its predictive value (e.g.,
Keane et al. 1989). Potential application includes
models of P. albicaulis regeneration for prescribed
fires or for past fires such as suitable areas of
Yellowstone National Park burned in 1988.


46
The correlation between density and distance did
not hold for the road plot series. This could be due to
the effects of elevation on P. albicaulis seed
germination and seeding survival. All of the road
series plots were at elevations of 2290 m or less, and
thus were below the range within which P. albicaulis
grows well in this region. The forest bordering the
southern part of the burn is composed of mixed conifers,
and thus may be a better source of regeneration for
Abies lasiocarpa, Picea engelmannii, and Pinus contorta
than for P. albicaulis.
Variation in Density and Cluster Size
The data show significantly higher densities of
both Pinus albicaulis sites and trees on the north
aspect of the ridge than on the ridge or south (at least
7 times greater mean density on the north). This is
consistent with the prevalence of P. albicaulis on cool
exposures and moist sites in semiarid ranges (Arno and
Hoff 1987). North slopes, with deeper soil mantles,
more moisture and cooler temperatures, may provide more
suitable germination conditions. Leadem (1986) examined
the dormancy-breaking mechanisms of P. albicaulis seeds,
and found that cold temperatures were necessary to
overcome physiological barriers to growth.
Alternatively, nutcrackers may cache more on north


47
slopes, but Tomback (1978, 1982) and Vander Wall and
Baida (1977) suggest that south slopes are preferred.
Regeneration density must be a result of both 1)
nutcracker caching preferences, and 2) site quality.
Competition from other plant species could also be a
factor in density variation among plot series.
Xerophyllum tenax tufts were widespread on south slopes,
as were grass species; both are potential competitors
for young trees.
Another interesting pattern was the relatively
large cluster sizes found on the road plot series, where
the largest percentages of cluster sizes greater than 5
occurred, and a 17-stem cluster was found nearby.
Distances from road plots to the seed source averaged
over twice as far as ridge plots. This poses an
interesting question: do nutcrackers bury larger
clusters of seeds when they fly greater distances? Or
perhaps germination and seedling survival conditions
were better on the road plot series despite the lower
elevations. Although highly tolerant of crowding, P.
albicaulis seedlings within a cluster may compete to
some degree as they mature. The lower height values
observed for trees on the road plot series could be
indicative of intra-cluster competition.


48
Post-Fire Disturbance
Successional processes have been complicated by
a history of disturbances in the study area. Lyon
(1976) summarized several years of post-fire management
activities in the Sleeping Child burn, including
chemical thinning of lodgepole pine, pole cutting,
cattle grazing, Christmas tree cutting, firewood
gathering, road building, aerial sowing of annual
grasses, and recreational uses such as hunting and
camping. He estimates little long-term effect from
these management activities, since competition would
have produced substantial losses in dense stands anyhow.
However, some of these activities, especially chemical
thinning of Pinus contorta, grazing, and sowing of
grasses, may affect the type and density of regrowth
within some areas of the burn. A case in point is the
road plot series, with an elevation range stated to be
dominated by P. contorta prior to the fire (Lyon and
Stickney 1976), but with regeneration density of P.
contorta lower than Abies lasiocarpa. Evidence of
grazing and browsing was seen often on south slopes,
which tended to be less steep in both study sites.
Range cattle were encountered during field work on south
aspects as well as on the dirt roads that access the
burn.


49
Vegetation Comparisons Between Plots
and Microsites
Much consistency was seen in the understory
species on P. albicaulis microsites overall, despite
differences in characteristic vegetation found on the
four plot series. The predominance of species such as
Vaccinium scoparium, Polytrichum sp., and Carex rossii
at tree sites reflects the preference of P. albicaulis
for mesic to meso-xeric habitat in these mountains
(Lackschewitz 1986). This contrasts with the high
frequency of X. tenax which grew as large, dense clumps
on all the plot series. The growth form of X. tenax
undoubtedly makes it a superior competitor compared to
young P. albicaulis, and nutcrackers probably do not
cache seeds near these plants.
Fire and P. albicaulis Regeneration
Twenty-six years after a severe fire, the
Sleeping Child Burn shows regeneration of Pinus
albicaulis to be well under way. Tree counts of the
four conifer species on the higher elevation study site
(the three ridge plot series) support the observation
that stand-replacing fires favor P. albicaulis in
relation to its more shade-tolerant subalpine
associates, Abies lasiocarpa and Picea engelmannii (Arno
1986, Arno and Hoff 1987). This pattern is complicated


50
by a higher density of Pinus contorta on the plots with
south exposures and on the few plots at the west end of
the ridge that are within lower subalpine elevations.
Our two plots in the unburned forest at the upper edge
of the burn show species composition dominated by P.
albicaulis and A. lasiocarpa, the potential climax
species for the upper subalpine habitat type (Pfister
et. al 1977). The striking abundance of P. contorta
regeneration on the lower subalpine road plots may be
due to the serai P. contorta forest existing at the time
of the fire. Large numbers of seeds were probably
released as the serotinous cones opened from the heat
(Fischer and Bradley 1987).
Tree counts for the road plot series show Pinus
albicaulis regeneration to be a relatively minor
component in relation to Abies lasiocarpa and Pinus
contorta. Regeneration patterns along this plot series,
which extends into the upper limits of the lower
subalpine, is most likely the consequence of the great
distance from P. albicaulis seed crops, lower elevation,
and closer proximity to seed crops of the other
conifers.
Clearly, the result of severe fire in the
Sleeping Child Burn is regeneration of P. albicaulis at
higher elevations. The post-fire dominance of this
species is typical of the forest overstory composition


51
for much of the northern Rocky Mountains that existed
prior to 20th century fire suppression policies.


CHAPTER V
CONCLUSIONS
Regeneration of Pinus albicaulis, a major
component of northern Rocky Mountain subalpine forests
and an important wildlife food source, is well underway
in an area where it had been declining. Clark's
nutcrackers are responsible for the Pinus albicaulis
regeneration patterns observed in the Sleeping Child
Burn. This severely burned area now has P. albicaulis
trees growing at distances of at least 8 km from the
nearest seed source, distances that can only be
accounted for by the seed caching behavior of the birds.
The even-aged tree clusters that were present on both
study sites are evidence of seed caches buried by
nutcrackers. P. albicaulis regrowth on the ridge study
site along a distance gradient from a known seed source
corresponds to a linear relationship between
regeneration density and distance. This indicates that
nutcracker caching behavior is most concentrated nearest
the seed source but extends for distances of several
kilometers. Birds may also bury larger numbers of seeds
per cache when they fly greater distances from the
harvest site.


53
The regeneration densities for four conifer
species indicate that avian dispersal results in earlier
establishment of P. albicaulis during post-fire
succession, compared to its upper subalpine associates
Abies lasiocarpa and Picea engelmannii. The
regeneration species composition at high subalpine
elevations resembles that of regional forests prior to
the fire suppression policies initiated in the early
1900's.


54
LITERATURE CITED
Arno, S. F. 1976. The historical role of fire on the
Bitterroot National Forest. USDA Forest Service Res.
Paper INT-187, p 29. Intermountain Forest and Range
Experiment Station, Ogden, Utah.
------. 1986. Whitebark pine cone crops a diminishing
source of wildlife food? Western Journal of Applied
Forestry l(3):92-94.
Arno, S. F., and T. D. Petersen. 1983. Variation in
estimates of fire intervals: a closer look at fire
history on the Bitterroot National Forest. USDA Forest
Service Research Paper INT-301, p 8. Intermountain
Forest and Range Experiment Station, Ogden, Utah.
Arno, S. F., D. G. Simmerman, and R. E. Keane. 1985.
Forest succession on four habitat types in western
Montana. USDA Forest Service Gen. Tech. Rep. INT-177,
p 74. Intermountain Forest and Range Experiment Station,
Ogden, Utah.
Arno, S. F., and R. J. Hoff. 1989. Whitebark pine. In:


55
Silvics of North America, Volume 1, Conifers. USDA
Forest Service, Washington, D.C. (In press).
Bailey, N. T. J. 1981. Statistical Methods in Biology,
Second Edition. John Wiley and Sons, New York.
Dorn, R. D. 1984. Vascular Plants of Montana. Mountain
West Publishing, Cheyenne, Wyoming.
Finklin, A. I. 1983. Weather and Climate of the Selway-
Bitterroot Wilderness. University Press of Idaho.
Fischer, W. C., and A. F. Bradley. 1987. Fire Ecology
of western Montana forest habitat types. USDA Forest
Service General Technical Report INT-223. Intermountain
Research Station, Ogden, Utah.
Furnier, G. R., P. Knowles, M. A. Clyde, and B. P.
Dancik. 1987. Effects of avian seed dispersal on the
genetic structure of whitebark pine populations.
Evolution 4(13):607-612.
Gravetter, F. J., and L. B. Wallnau. 1985. Statistics
for the Behavioral Sciences. West Publishing Company,
St. Paul, Minnesota.


56
Heinselman, M. L. 1978. Fire intensity and frequency as
factors in the distribution and structure of Northern
ecosystems. In: Fire Regimes and Ecosystem Properties,
proceedings of the 1978 Conference. USDA Forest Service
General Technical Report WO-26, p 594.
Hitchcock, C. L., and A. Cronquist. 1973. Flora of the
Pacific Northwest. University of Washington Press,
Seattle, Washington.
Hutchins, H. E., and R. M. Lanner. 1982. The central
role of Clark's nutcracker in the dispersal and
establishment of whitebark pine. Oecologia (Berl)(1982)
55:192-201.
Keane, R. E., S. F. Arno, J. K. Brown, and D. F.
Tomback. 1989. Modeling stand dynamics in whitebark pine
forests. Ecological Modeling (in press).
Kendall, K. C. 1980a. Bear-squirrel-pine nut
interaction. In: Yellowstone grizzly bear
investigations. Annual Report 1978-79, USDI-National
Park Service, pp 51-60.
Kendall, K. C. 1980b. Food habits of Yellowstone
grizzly bears, 1978-79. In: Yellowstone grizzly bear


57
investigations. Annual Report 1978-79, USDI-National
Park Service, pp 24-34.
Lackschewitz, K. 1986. Plants of west-central Montana -
Identification and ecology: annotated checklist. USDA
Forest Service Gen. Tech. Rep. INT-217, p 128.
Intermountain Research Station, Ogden, Utah.
Lanner, R. M. 1980. Avian seed dispersal as a factor in
the ecology and evolution of limber and whitebark pines,
pp. 15-48. In: B.P. Dancik and K.O. Higginbotham (eds.),
Proceedings of the Sixth North American Forest Biology
Workshop. University of Alberta, Edmonton, Alberta,
Canada.
Lanner, R. M., and S. B. Vander Wall. 1980. Dispersal
of limber pine seed by Clark's nutcracker. Journal of
Forestry 78: 638-640.
Leadem, C. L. 1986. Seed dormancy in three Pinus species
of the inland mountain west. In: Proceedings Conifer
Tree Seed in the Inland Mountain West Symposium. USDA
Forest Service General Technical Report INT-203, p 289.
Intermountain Research Station, Ogden, Utah.
Ligon, J. D. 1978. Reproductive interdependence in


58
pinyon jays and pinyon pines. Ecological Monographs
48:111-126.
Linhart, Y. B., and D. F. Tomback. 1985. Seed dispersal
by nutcrackers causes multi-trunk growth form in pines.
Oecologia (Berlin) 67:107-110.
Lotan, J. E. 1975. The role of cone serotiny in
lodgepole pine forests. In: Management of Lodgepole
Pine Ecosystems Symposium Proceedings. Washington State
University Cooperative Extension Service, Pullman,
Washington.
------. 1976. Cone Serotiny Fire relationships in
lodgepole pine. In: Proceedings, Tall Timbers Fire
Ecology Conference No. 14 and Intermountain Fire
Research Council Fire and Land Management Symposium.
Lyon, L. J. 1976. Vegetal development on the Sleeping
Child burn in western Montana 1961 to 1973. USDA Forest
Service Research Paper INT-184, p 24. Intermountain
Forest and Range Experiment Station, Ogden, Utah.
Lyon, L. J., and P. F. Stickney. 1976. Early vegetal
succession following large Northern Rocky Mountain
wildfires. Proc. Montana Tall Timbers Fire Ecology


59
Conference and Fire and Land Management Symposium, No.
14, 1974:355-375. Tall Timbers Research Station,
Tallahassee, Florida.
McCaughey, W. W., W. C. Schmidt, and R. C. Shearer.
1986. Seed-dispersal characteristics of conifers in the
inland mountainous west. In: Proceedings Conifer Tree
Seed in the Inland Mountain West Symposium. USDA Forest
Service General Technical Report INT-203, p 289.
Intermountain Research Station, Ogden, Utah.
Mirov, N. T. 1967. The Genus Pinus. Ronald Press
Company, New York.
Mirov, N. T., and J. Hasbrouck. 1976. The Story of
Pines. Indiana University Press, Bloomington, Indiana.
Pfister, R. D., B. L. Kovalchik, S. F. Arno, and R. C.
Presby. 1977. Forest habitat types of Montana. USDA
Forest Service Gen. Tech. Rep. INT-34, p 174.
Intermountain Forest and Range Experiment Station,
Ogden, Utah.
Romme, W. H. 1982. Fire and landscape diversity in
subalpine forests of Yellowstone National Park.
Ecological Monographs 52(2):199-221.


60
Siegel, S. 1956. Nonparametric Statistics for the
Behavioral Sciences. New York:McGraw-Hill.
Tomback, D. F. 1978. Foraging strategies of Clark's
nutcracker. The Living Bird (16th Annual, 1977):123-161
1982. Dispersal of whitebark pine seeds by
Clark's nutcracker: a mutualism hypothesis. Journal of
Animal Ecology 51:451-467
1983. Nutcrackers and pines: coevolution or
coadaptation? In: M.H. Niteck,editor. Coevolution.
University of Chicago press, pp. 179-223.
1986. Post-fire regeneration of krummholz
whitebark pine: a consequence of nutcracker seed
caching. Madrona, 33(2):100-110.
Tomback, D. F., L. Hoffman, and S. Sund. Co-evolution
of whitebark pine and nutcrackers: implications for
forest regeneration. (In preparation).
Tomback, D. F., and Y. Linhart. The coevolution of
bird-dispersed pines. (In preparation).


61
Turcek, F. J,. and L. Kelso. 1968. Ecological aspects
of food transportation and storage in the Corvidae.
Communications in Behavioral Biology, Part A, 1(4):277-
297.
Vander Wall, S. B., and R. P. Baida. 1977. Coadaptations
of the Clark's nutcracker and the pinon pine for
efficient seed harvest and dispersal. Ecological
Monographs, 47(1):89-111.
1981. Ecology and evolution of food-storage
behavior in conifer-seed-caching corvids. Z.
Tierpsychol. 56:217-242.
Wright, H. A., and A. W. Bailey. 1982. Fire Ecology:
United States and Southern Canada. John Wiley and Sons,
New York.


62
APPENDIX A.
Plant Species List for Sleeping
Child Burn
COMMON NAME SCIENTIFIC NAME
Subalpine fir Yarrow Pearly-everlasting Pussytoes Sandwort Arnica Mountain arnica Aster Smooth brome-grass Elk sedge Abies lasiocarpa Achilles millefolium v. alpicola Anaphalis margaritacea Antennaria Arenaria aculeata Arnica cordifolia Arnica latifolia v. latifolia Family Asteraceae Bromus inermis ssp. inermis Carex geyeri
Ross sedge Carex rossii
Composite (unknown) Unknown composite
Orchard-grass Fireweed Buckwheat Fleabane Forb (unknown) White hawkweed Western hawkweed Labrador tea Patridge foot Lichen (unknown) Honeysuckle Lupine Woodrush Fools huckleberry Alpine fernleaf Penstemon Common timothy Engelmann spruce Whitebark pine Lodgepole pine Ponderosa pine Bluegrass Bistort Moss Jacobs ladder Dactylis glomerata Epilobium sp. Eriogonum sp. Erigeron sp. Unknown forb Hierachium albiflorum Hierachium albertinum Ledum glandulosum Leutkea pectinata Unknown lichen Lonicera utahensis Lupinus sp. Luzula sp. Mensiesia ferruginea v. glabella Pedicularis contorta Penstemon procerus Phleum pratense Picea engelmannii Pinus albicaulis Pinus contorta v. latifolia Pinus ponderosa v. scopulorum Poa sp. Polygonum bistortoides Polytrichum sp. Polemonium sp.


APPENDIX A. (cont.)
COMMON NAME SCIENTIFIC NAME
Douglas-fir Gooseberry Sheepsorrel Catchfly Green needlegrass Needlegrass Snowberry Dandelion Downy oatgrass Huckleberry Grouseberry False hellebore Beargrass Pseudotsuga menziesii Ribes sp. Rumex acetosella Silene spaldingi Stipa nelsonii Stipa occidentalis Syraphoricorpus sp. Taraxacum officinales Trisetum spicatum Vaccinium globulare Vaccinium scoparium Veratrum viride Xerophyllum tenax


Full Text

PAGE 1

POST-FIRE REGENERATION OF Pinus albicaulis IN WESTERN MONTANA: PATTERNS OF OCCURRENCE AND SITE CHARACTERISTICS by Sharren Kay Sund B.A., University of Colorado, 1976 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Department of Biology 1988

PAGE 2

This for the Master of Arts degree by Sharren Kay Sund has been approved for the Department of Biology by Emily L. Hartman Date ------------------------

PAGE 3

Sund, Sharren Kay (M.A., Biology) Post-Fire Regeneration of Pinus albicaulis in Western Montana: Patterns of Occurrence and Site Characteristics Thesis directed by Associate Professor Diana F. Tomback. In 1961, a severe fire in the Bitterroot National Forest, western Montana, destroyed about 11,350 hectares of subalpine forest, which included Pinus albicaulis (whitebark pine). The devastated area became known as the Sleeping Child Burn. Because of the recent decline of populations of P. albicaulis, a major subalpine component throughout the northern Rocky Mountains (Arno 1986), regeneration of P. albicaulis in the burn is a general concern. Unlike most pines, P. albicaulis reliesA on a bird, Clark's nutcracker (Nucifraga columbiana), for seed dispersal. In 1987, forest regeneration within the Sleeping Child Burn was studied on 77 plots along two transects: 1) an east-west ridge exhibiting a gradient for both elevation (range 2148 m to 2182 m) and distance from a P. albicaulis seed source (range 50 m to 3650 m) (n = 63 plots), and 2) a north-south transect of greater distance from. a seed source but relatively constant elevation (n = 14 plots). Results indicated that nutcrackers have dispersed clusters of albicaulis seeds up to 8 km into the burn. On the ridge transect, a significant

PAGE 4

ii negative correlation was found between P. albicaulis regeneration density and distance from seed source. A significant positive correlation was also found between density and elevation; it was not possible to separate the effects of elevation and distance. Neither of these correlations were significant on the north-south transect. Regeneration patterns show that Pinus albicaulis and Pinus contorta are dominating the post-fire succession on the east-west ridge. The regeneration status of P. albicaulis in the Sleeping Child Burn indicates an advantage over its associates, Abies lasiocarpa and Picea engelmannii, and perhaps a return to the successional upper subalpine forest communities that were widespread prior to the fire suppression policies of this century. The form and content of this abstract are approved. I recommend its publication. Signed Diana F. Tomback

PAGE 5

iii ACKNOWLEDGEMENT This work was supported by funds from the United States Department of Agriculture, u.s. Forest Service to Diana F. Tomback, whose guidance and encouragement are much appreciated. I thank Steve Arno, Jim Brown and Bob Keane of the Fire Sciences Laboratory, Forest and Range Experiment Station, Missoula, for supplying equipment and assistance in the field. Michael Reuss provided the use of computer hardware and software, assisted with programming, and was a constant source of encouragement. A special thanks to Lyn Hoffmann for weeks of field work, painstaking data recording, plant identifications, and helpful critiques.

PAGE 6

CONTENTS CHAPTER I. INTRODUCTION 1 Pinus albicaulis ......... 3 Fire History in the Northern Rocky Mountains ......... 5 Seed Dispersal Mechanisms ..... 6 Sleeping Child Burn .... 9 Hypotheses . . . . . . . . . . . . . . 10 II. METHODS 11 Study Area . . . . . . . . . . . . . . 11 Study Site #1-Ridge ............. 12 Study Site #2 -Road . . . . . . 15 Unburned Forest . . . . . . . . . . . . 15 Field Methods . . . . . . . . . . . . . 17 Statistical Methods ........ 19 III. RESULTS . . . . . . 21 Plot Descriptions (Elevation, Distance, Siope, Area) ......... 21 Conifer Species and Site Counts ....... 23 Understory Plant Species Occurrence on Plots . . . . . . . . . . . . . . 24 P. albicaulis Clusters and Microsite Vegetation . . . . . . . . . . . . . 27 Nearest Objects on albicaulis Microsi tes . . . . . . . . . . . . . 31

PAGE 7

Vegetation Comparisons Between Plot Series and albicaulis Microsites . 31 P. albicaulis Site Density, -Elevation and Distance ........ 32 Regression Model for Site Density vs. Distance . . . . . . . . . . . . . . 34 P. albicaulis Tree Density, -Elevation and Distance .......... 36 Regression Model for Tree Density vs. Distance . . . . . . . . . . . . . . 38 albicaulis Tree Age, Height -and Distance . . . . . . . . . . . . 38 IV. DISCUSSION . . . . . . . . . . . . . . . 42 Regeneration: Consequences of Seed Dispersal by Nutcrackers ......... 42 Variation in Density and Cluster Size .. 46 Post-Fire Disturbance ......... 48 Vegetation Comparisons Between Plots and Microsites ....................... 49 Fire and P. albicaulis Regeneration ... 49 V. CONCLUSIONS . . . . . . . . . . . . . 52 LITERATURE CITED . . . . 54 APPENDIX A. Plant Species List for Sleeping Child Burn . . . . . . . . . . . . . . . 62 v

PAGE 8

TABLES Table 1. Plot series characteristics: aspect, elevation, distance; slope, vi area. . . . . . . . . . . . . . . . . . . . 22 2. Pinus albicaulis clusters: no. of seedlings and frequency ........... 28 3. Pinus albicaulis regeneration by plot series: density, age, height ....... 33 4. Pearson correlation analysis for Pinus albicaulis: density vs. elevation, density vs. distance from seed source, age vs. height. . . . . . . . . . . . . . . . . . . 3 5

PAGE 9

vii FIGURES Figure 1. Study site #1: Sleeping Child Burn, Ridge transect .. ........... 13 2. Study site #2: Sleeping Child Burn, Road transect ................ 16 3. Five most frequently occurring plant species on plots, by plot series ........ 25 4. Five most frequently occurring plant species on P. albicaulis microsites, by plot series: ............... 30 5. Scatterplot of P. albicaulis tree density vs. distance from seed source, for three ridge plot series combined ........... 39 6. Scatterplot of albicaulis log10 tree density vs. distance from seed source, for tnree ridge plot series combined ..... 40

PAGE 10

CHAPTER I INTRODUCTION Recently, there is concern that Pinus albicaulis (whitebark.pine) a major component of subalpine forests in the northern Rocky Mountains, is declining in some regions (Arne 1986). Reduction of the species results from a combination of fire suppression policy, insects, and disease. While not an important timber species, albicaulis is now recognized as a valuab1e food source for wildlife because of its nutritious seeds (Tomback 1978; Kendall 1980a,b; Hutchins and Lanner 1982; Arne 1986). Also, it is-the only North American member of the Cembrae subsection of the Strobus pines; species of this subsection are characterized by indehiscent cones and large,: wingless seeds (Mirov 1967; Mirov and Hasbrouck 1976). All Cembrae pines are dependent upon birds of the genus Nucifraga ("nutcrackers") for seed dispersal (Lanner 1980; Tomback 1983; Tomback and Linhart, in prep.). Arne ( 1986) describes -the effect of .fire policies on albicaulis. Large fires have played an important ecological role in the perpetuation of P.

PAGE 11

2 albicaulis, with surface fires allowing the tree to regenerate in small openings within mixed coniferous forest and severe, stand-replacing fires giving P. albicaulis regrowth an initial advantage over more shade-tolerant species such as Abies lasiocarpa. Fire suppression beginning in the early 1900's has severely limited the size of wildfires and lengthened their intervals, thus reducing the rate at which P. albicaulis is rejuvenated. Two additional factors exacerbate the problem: insects and disease. Infestations of mountain pine beetle ( Dentroctonus ponderosae) have destroyed hundreds of thousands of hectares of P. albicaulis forests in the northern Rockies since 1904. Older stands of P. albicaulis and its associate Pinus contorta are favored by mountain pine beetles, and older stands become widespread with longer fire intervals. A disease introduced from Europe, white pine blister rust (Cronartium ribicola), is also a serious threat to P. albicaulis in cool, moist mountain regions such as in western Montana, Idaho, and Washington (Arno 1986, Arno and Hoff 1989). This study investigated the regeneration of P. albicaulis after a severe fire in 1961 that resulted in the devastation of over 11, 000 hectares of subalpine forest in the Sapphire Range, Bitterroot National Forest, western Montana. The patterns of P. albicaulis

PAGE 12

3 regeneration and from forest associates are of interest to forest managers who are concerned about losses of successional communi ties of P. al bicaulis, once abundant in this area. Pinus albicaulis The following information is summarized from Arno (1986) and Arno and Hoff (1989). P. albicaulis is a haploxylon (soft) pine that __ ranges from British Columbia south to central California and east to western Wyoming, occurring on exposed slopes and ridges of high mountains. Found only at timberline in the north end of its distribution, albicaulis increases in abundance at both timberline and upper subalpine as latitude decreases. In west-central Montana, it is a major component of high elevation forests and the timberline zone between 2130 and 2830 meters. The boundary between the lower and upper subalpine is considered to be the lower limit of albicaulis as a major forest species (Pfister et al. 1977). This slow growing, long-lived (as much as 1000 years) conifer reaches heights of 5 -15 meters and often exhibits a multi-trunk growth form. P. albicaulis occurs on coarse, rocky soils in climatic zones characterized by short, cool summers and cold, snowy winters. In mesic mountain ranges it favors warm dry exposures, while in semi-arid mountains it is more

PAGE 13

4 abundant on cool, moist slopes. Principally a seral component of mixed coniferous forest in the upper subalpine, and a pioneer on recently burned or disturbed sites, albicaulis is also a climax species on xeric sites and in severe timberline zones (Fischer and Bradley 1987). However, in the Rocky Mountains Pinus albicaulis usually associates with Picea engelmannii, Tsuga mertensiana, and Larix lyalli at high elevations bordering timberline (Wright and Bailey 1982), and with Picea engelmannii, Abies lasiocarpa, and Pinus contorta at lower elevations. On some sites bordering timberline, P. albicaulis, P. engelmannii, lasiocarpa may end up as climax co-dominants (Arno and Hoff 1989). But in most subalpine areas, P. albicaulis is eventually replaced by lasiocarpa with engelmannii and P. contorta as minor components. P. albicaulis is considered to be moderately shade-tolerant. It is less tolerant than A. lasiocarpa or engelmannii but more tolerant than P. contorta (Arno and Hoff 1989). As a successional species in subalpine communities; P. albicaulis is periodically regenerated by fire. The order of fire-resistance or tolerance is P. contorta > P. albicaulis > P. engelmannii > A. lasiocarpa (Fischer and Bradley 1987). Dry exposed sites and open stands offer P. albicaulis more

PAGE 14

5 protection from fire than does its moderately thin bark (Mirov 1967; Mirov and Hasbrouck 1976). P. albicaulis trees shed their lower branches as they grow, preventing vertical fuel continuity that can cause surface fires to become crown fires (Remme 1982). Fire History in the Northern Rocky Mountains In coniferous forests of the northern Rocky Mountains, fire regimes vary from relatively frequent, light surface fires that are localized in effect, to severe 'crown and surface fires that burn large areas at 100 to 300 year intervals (Arno 1976, Heinselman 1978, Remme 1982, Wright and Bailey 1982, Arno and Petersen 1983). Although the calculated lengths of fire intervals vary somewhat with the method used to determine them, researchers agree that major changes have occurred in natural fire regimes due to human intervention. Fire history in the Bitterroot National Forest is summarized by Arno and Petersen (1983). Depending upon the severity of fire and the sample size unit, estimates of fire recurrence intervals in the Bitterroot National during the 300 years prior to 1910 range from a mean of 20.to 60 years on lower subalpine slopes, and 40 to 100+ years on upper subalpine. In contrast,

PAGE 15

6 the approximate present fire cycle may be greater than 1000 years in both zones. The longer fire intervals are causing the dominant overstory tree species to be affected. Prior to 1900, the dominant overstory in the upper subalpine was Pinus albicaulis, followed by Pinus contort a. However, with continued fire exclusion, Abies lasiocarpa is predicted to dominate. In the lower subalpine, P. contorta and Pseudotsuga menziesii are being replaced by A. lasiocarpa. Seed Dispersal Mechanisms Several seed-dispersal mechanisms are found among conifers. Some cones of Pinus contorta release seeds upon ripening, while other cones are serotinous and adhere to the tree for several years, until the heat of fire opens cones and broadcasts seeds (Lotan 1975, 1976). All the cones of Abies lasiocarpa and Picea engelmannii open when ripe, disseminating seeds. Pinus contorta, Abies lasiocarpa, and Picea engelmannii are anemochores, or wind-dispersed species. These three species generally exhibit a U-shaped seed dispersal distribution across forest openings, with the quantity of seeds dispersed being greatest at the windward and leeward edges of clearings and decreasing towards the center. The dispersal curve shape for these three species is a negative exponential curve (McCaughey et

PAGE 16

7 al. 1986). Pinus albicaulis, in contrast, is one of about eight pines in the world known to be dependent upon nutcrackers and their relatives (family, Corvidae) for dissemination (Lanner 1980; Tomback 1983). The indehiscent cones, i.e. cones that do not open upon ripening, contain large, wingless seeds. Clark's nutcracker (Nucifraga columbiana) is the primary dispersal agent of albicaulis (Tomback 1982, 1983; Hutchins & Lanner 1982). Other western pines, P. flexilis and the pinyon pines monophylla and P. edulis), are also dependent on nutcracker dispersal (Vander Wall & Balda 1977; Lanner & Vander Wall 1980). The pinyon pines {Pinus edulis) are also dispersed by pinyon jays (Gymnorhinus cyanocephalus) (Ligon 1978). This relationship has a European counterpart; four species of Cembrae pines of Eurasia are disseminated by the Eurasian nutcracker (Nucifraga caryocatactes) {Turcek and Kelso 1968). Tomback (1983) emphasizes how closely the ecology of the Eurasian Cembrae pines and albicaulis are tied to the two species of nutcrackers; these mutualistic associations are most likely coevolved. After Pinus albicaulis seeds have ripened and hardened, nutcrackers pry whole seeds from the cones with their bills and transport the seeds in a sublingual

PAGE 17

8 pouch for distances of up to 12.5 km (Tomback 1982). Pouch contents can average 77 P. a1bicau1is seeds ( Tomback 1982). Vander Wall and Balda (1977) report nutcrackers carrying pinyon seed distances of 22 km. Nutcrackers disperse pine seeds by burying clusters or "caches" of seeds in the soil, from 1 -15 seeds per site with a mean of 3. 7; many of these seeds are retrieved by nutcrackers during winter and spring when other foods are scarce (Tomback 1982). Bird dispersal affects the growth form of the trees and genetic structure of P. albicaulis populations (Linhart and Tomback 1985, Furnier et al. 1987). A multi-trunk or clustered growth form is common in P. albicaulis and a few other species dispersed by nutcrackers. Electrophoresis shows multiple trunks in adult P. albicaulis often to be composed of > 1 genotype, the result of germination from seed caches (Linhart and Tomback 1985; Furnier et al. 1987). Nutcrackers usually harvest seeds from only few trees before caching, so several of the seeds in a cache may be full or halfsiblings. Even though stems within a cluster may be genetically distinct individuals, the degree of relatedness within a cluster is greater than between neighboring clusters, as caches in one area may be the result of more than one bird caching (Furnier et al. 1987).

PAGE 18

9 Sleeping Child Burn In August 1961, a lightning-strike fire burned about 11,350 hectares in the Sapphire Range of the Bitterroot National Forest in western Montana. Within a few days plant crowns were totally destroyed and the organic layer reduced to mineral ash. The devastated area became known as the Sleeping Child Burn (Lyon and Stickney 1976). The Bitterroot National Forest, on the eastern boundary of the Selway-Bitterroot Wilderness, is divided by the Bitterroot Valley into the Sapphire Range on the east and the Bitterroot Range on the west. It is an area of steep, rugged mountains with dense forest. Upper subalpine forest grows above 2130 m (Arno et al. 1985). Surface geologic formation is of granitic origin, with medium to coarse-textured, shallow, rocky soils; ancient wind-deposited ash and loess have resulted in the leeward slopes (north and east aspects) having deeper soil mantles (Arno et al. 1985). The primary elevational range for Pinus albicaulis in the Bitterroot National Forest is within the upper subalpine forests from 2290 to 2620 m (Arno and Petersen 1983). Forest composition at lower subalpine elevations (from 2130 m up to 2290 m) is mixed-conifer species, with occasional islands of P. albicaulis. The nearest P. albicaulis seed source for

PAGE 19

10 the Sleeping Child Burn is considered to be the continuous stands of cone-bearing P. albicaulis that occur primarily at elevations above 2257 m, since forest composition at lower elevations is mixed conifers with relatively few islands of P. albicaulis (Arno, personal communication). The habitat type groups for potential climax on upper subalpine slopes are: 1) Abies albicaulis/Vaccinium scoparium, 2) A. lasiocarpa/Luzula hitchcockii and 3) P. albicaulis-A. lasiocarpa (Pfister et al. 1977). Hypotheses 1) Because Pinus albicaulis is a pioneer species and because of the 26 year time period following the fire, regeneration of this conifer should be advanced in the Sleeping Child Burn. 2) The patterns of P. albicaulis regeneration should be consistent with patterns of avian seed dispersal.

PAGE 20

CHAPTER II METHODS Study Area The Sleeping Child Burn covers an area about 14.5 km by 8 km in Ravalli County, approximately 80 km south-southeast of Missoula, Montana. From Hamilton, the burn is reached by travelling east on state highway 38 to Skalkahoe Rye Road. The northern boundary of the burn extends to Township 4 N and the southern boundary extends into the north half of Township 2 N. On the east and west the boundaries extend, respectively, from Range 18 W into Range 19 W. Located west of the Continental Divide, this area has an inland maritime climate. As elevation increases from valley ( 1220 -1525 m) to mountaintop (2135 2745 m), the precipitation increases, and temperatures become cooler. Distinct differences occur between south and north aspects (Pfister et al. 1977). The south slopes are much warmer and drier, whereas north slopes experience less solar insolation, greater snow accumulations and deeper soils -resulting in cooler, moister conditions.

PAGE 21

12 Annual precipitation averages 33 ern and varies with elevation and topography (Lyon 1976). Seventy percent of the precipitation above 2135 rn is from snowfall. Early winter and late spring are the wettest seasons; lightning activity and low relative hurnidi ty occur in July and August, resulting in a fire season. Daily temperature ranges are greatest at low elevations and least on mountaintops. Monthly mean temperatures in the area show a slight decrease on a south-north gradient. Winds are predominantly from the west or southwest (Finklin 1983). Field work was conducted from 3 to 26 August on two study sites in the burn. The first study was located along an east-west ridge transect, 3.65 krn long, from Lost Tooth Cabin at the edge of the burn heading west to Two Meadows, approximately in the northern center of the burn. The second site was a 20. 6 krn north-south road transect following Skalkahoe Rye Road from the northern edge of the burn, south and then southwest on Paint Creek Road approximately through the center of the burn to its southern edge. Study Site #1 -Ridge The first site (Figure 1) was selected to gather data on forest regeneration patterns in the northern center of the burn. Plots were sampled along a straight

PAGE 22

SLEEPING CHILD BURN 0.5 km 13 Fig. 1. Study site #1: Sleeping Child Burn, Ridge Transect. The stippled area is the subalpine forest bordering the burn. The hatchmarked area indicates a subsection of the forest with elevations above.2257 m where continuous stands of P. albicaulis provide the seed source for-regeneration.

PAGE 23

14 ridge of gradually decreasing elevation and increasing distance from the unburned forest at the eastern edge of the burn. Plots were established 150 m apart on a 3.65 km east-west transect. Each ridge plot was paralleled by plots on the north-and south-facing slopes of the ridge, thus creating three parallel series of plots: north, south, and ridge. In some instances, a representative north or south aspect was not available or could only be obtained by walking several meters east or west. A total of 63 plots were studied: 21 north, 19 south, and 23 ridge. All plots were oriented east-west longitudinally, with belt width measured south from the transect line. Size of the plots ranged from 50 x 1.25 m to 50 x 12 m. Belt width varied for purposes of efficiency. It was important to avoid zero densities of Pinus albicaulis regeneration and yet quickly process plots where densities were high. On the west end of the ridge, albicaulis.regeneration was sparse, so belt width was increased to 10 m or 12 m. Narrower widths were used on the east end of the ridge were P. albicaulis densities were much higher. Belt widths were identical for all other species, but were chosen according to density.

PAGE 24

15 Study Site #2 -Road This second site was selected to measure forest regeneration at fairly constant elevation and aspect but increasingly greater distances from the Pinus albicaulis seed source. A total of 14 plots were studied on northfacing aspects along the Skalkahoe Rye Road and Paint Creek Road (Figure 2) The plots were separated by road distances of 644 m to 2574 m, and their locations corresponded to aerial distances of 900 m to 7950 m from the unburned forest. Plot sizes ranged from 30 x 1.5 m to 30 x 10 m. Plot width was determined by both the density of the vegetation and an effort to include at least one P. albicaulis seedling per plot. Unburned Forest For comparison with the burned area on the ridge study site, two plots in the adjacent unburned forest were located on north and south aspects at a distance of 150 m east of the first ridge plot. The north and south forest plots were at elevations of 2440 m and 2420 m, respectively, and each plot was 500 m2 in area. All trees within these plots were noted.

PAGE 25

16 SLEEPING CHILD BURN : . : : '.1. ::. Fig. 2. Study site #2: Sleeping Child Burn, Road Transect. The stippled area is the subalpine forest bordering the burn. The hatchmarked area indicates a subsection of the forest with elevations above 2257 m where continuous stands of P. albicaulis .provide the seed source for-regeneration.

PAGE 26

17 Field Methods on each plot a list of up to 10 understory plant species, following the checklist of Hitchcock and Cronquist (1973) and Dorn (1984), was compiled by inspection. The purpose was not necessarily to indicate dominance but to list plants representative of the habitat on each plot. In August when field work was conducted, annuals and early-blooming perennials were past flowering, and thus identification of flowering plants was difficult. The percentage occurrence among all plots in a series was calculated for the most common plant species. Evidence of post-fire disturbance such as grazing, thinning, woodcutting, etc. was noted. Plot measurements included: elevation, slope, aerial distance from Pinus albicaulis seed source, aspect, and number of regeneration sites. Elevation was measured by altimeter and USGS 7.5 minute series topographic map (Kent Peak, and Jennings Camp Creek) readings rounded to the nearest 5 meters. An estimate of slope was made to the nearest 10 degrees, and aerial distance from the boundary of the P. albicaulis seed source was measured on the topographic map to the nearest 10 meters. The seed source is considered to be the continuous stands of P. albicaulis at elevations above 2257 m within the forest; only isolated islands of

PAGE 27

18 P. albicaulis are found in the mixed conifer forest below this elevation. A compass was used to determine aspect. Post-fire regeneration, which ranged in age from recent germinants to saplings and young trees, is collectively referred to as 'trees'. The number of. P. albicaulis sites on each plot were recorded. The term 'site' is used to refer to any location with either two trees or a cluster of one or more P. albicaulis trees. At each P. albicaulis site the following data were gathered: slope, aspect, identification of 1 to 4 vascular plant species found within 15 em of the site, and the nearest object (rock, wood pieces or fallen snag, etc. ) located within 15 em. Cluster size was recorded as the number of individual trunks at a single site and was determined by separating trunks to or below ground level. Although we looked for evidence of blister rust on the trees, none was seen. The height of each individual tree was recorded, and a number of trees of every representative height category were aged. On the four plot series within the burn, tree age was determined by severing the trunk at ground level and counting tree rings with a hand lens. Newly germinated trees were given a 0.5 year age value; all other ages were in whole years. Tree age on the unburned plots was determined by coring at ground level and counting tree rings.

PAGE 28

19 Statistical Methods Although a ratio scale of measurement was used for distance, elevation, and tree density data, nonparametric statistics were used in some analyses because underlying parametric assumptions concerning random sampling, normal population distributions, and homogeneity of variance were not always met. Our transects were systematically laid out to obtain plots at pre-determined orientations (east-west along an elevational gradient from center of burn to edge of burn, and north-south along a distance gradient from edge of burn) Normal population distributions were not indicated for P. albicaulis tree or site density on any transect, or for elevation on the north, ridge or south transects. Heterogeneity of variance was also found for albicaulis site density on the ridge, north and south plot series (F-max = 19.8). In addition to descriptive statistics for all plot series, Kruskal-Wallis ANOVA was used to compare results among the three ridge plot series (north, ridge, south). Further comparisons were made with Wilcoxon Rank Sum tests. The non-parametric statistical tests were selected in accordance with Seigel (1956) . The independent sampling, ratio measurement scale, and sample sizes were considered adequate to

PAGE 29

20 justify use of parametric analyses for correlation and regression. Pearson Product Moment Correlation analyses (Bailey 1981) were used for testing the following: distance (aerial distance from seed source) vs. tree density; distance vs. site density (a site may be one tree or multiple trees in a single cluster); distance vs. tree age; tree height vs. tree age; elevation vs. site density, and elevation vs. tree density for all plot series. North, south and ridge plot series were combined, and scatterplots with simple linear regression 1 ines were calculated for tree density vs. distance, site density vs. distance, log10 of tree density vs. distance, and log10 of site density vs. distance. The log10 of the density was plotted to see if linearity would result. The regression formula was: Y = bX + a where, Y = dependent variable, b = slope, a = y intercept, X = independent variable ( Gravetter and Wallnau 1985). The statistical significance of regression coefficient r was determined from Appendix 4 in Bailey (1981). Analyses were run on an IBMcompatible microcomputer with Microstat (version 4.1, Ecosoft, Inc., Indianapolis, IN). In all tests, results were considered significant only if P

PAGE 30

CHAPTER III RESULTS Plot Descriptions (Elevation, Distance, Slope, Area) For the ridge study site, there were a total of 63 plots; 21 on the north series; 23 on the ridge, and 19 on the south. The general aspect of plots on each plot series was: north -north aspect, range 290 20; ridge -west aspect, range 90 330; south south aspect, range 170 230; road -north aspect, range 330 90 (Table 1) All three ridge plot series were similar in their ranges of elevation and distance to the P. albicaulis seed source (Table 1). Elevation on the north, south and ridge plots ranged from 2105 m to 2480 m with means respectively of 2305 m, 2295 m, and 2305 m. The straight line distance from the eastern edge of individual plots to the seed source ranged from 50 m to 3650 m. Mean distance for north plots was 1680 m, for south, 1810 m, and for ridge, 1790 m. The steepest slopes were on the north aspect (range 0 to 50, mean = 30 )., the south was moderately steep (range 0 to 40, mean 20), and the ridge averaged much flatter (range 0

PAGE 31

22 TABLE 1. Plot series characteristics: aspect, elevation, distance, slope, area. NORTH Aspect north Range (deg) 290-20 No. of plots 21 Elevation (m) Mean 2305 Range 2150-2475 Distance from seed source (m) Mean 1680 Range 50-3500 Slope (deg) Mean 30 Range 0-50. Area (m2) Mean 154.8 Range 62.5-500 No. of sites P. albicaulis 271 P. contqrta 115 A. lasiocarpa 33 P. eng:elmannii 19 Total sampling area (m 2 ) P. albicaulis 3500 Other 3250 contorta, A. lasiocarpa, P. engelmannii) PLOT SERIES RIDGE west 90-330 23 2305 2175-2480 1790 50-3650 10 0-40 201.1 62.5-500 119 186 21 16 5750 4625 SOUTH south 170-230 19 2295 2165-2465 1810 50-3450 20 0-40 228.9 62.5-750 65 135 6 2 6225 4350 ROAD north 330-90 14 2200 2105-2290 3860 900-7950 20 0-30 84 12.5-500 45 80 123 38 2975 1176

PAGE 32

23 to 30, plus one uncharacteristic plot with 40 slope, mean= 10). For both the ridge and north series, plot sizes ranged from 62.5 m2 to 500m2 with a mean of 201.1 m2 and 154.8 m2 respectively. South series plots had a wider size range (62.5 to 750 m2 ) and a mean of 228.9 For the 14 plots on the road transect, elevation was generally lower than on the ridge series of plots, between 2105 and 2290 m (mean = 2200 m) (Table 1). The distances to seed source were much greater, with a range of 900 to 7950 m, and a mean of 3860 m, over twice that of the ridge series of plots. The slope on road plots averaged 20. Road. plots were smaller in area than the ridge series because of shorter plot length and ranged from 12.5 to 500 m2 averaging 84 m2 Conifer Species and Site Counts Four species of conifers were found on each plot series (north, south, ridge, road) (Table 1). The number of sites for P. albicaulis was: north -271, ridge -119, south -65, road -45. The number of sites of P. contorta was higher than that of P. albicaulis on every plot series but the north, despite the smaller P. contorta sampling area; contorta site counts on the ridge, south, north, and road were 186, 135, 115, and

PAGE 33

24 80, respectively. In addition, the total number of sites of P. contorta (n = 516) was greater than that for each of the other conifers. The following comparison of total sites per species is based on sampling areas of the same size: Pinus contorta > Abies lasiocarpa > Picea engelmannii. Only on the road plot series were A. lasiocarpa site counts. (n = 123) higher than counts of P. contorta. P. engelmannii had the lowest number of sites on every transect (ridge -16, north -19, south -2, road 38) relative to A. lasiocarpa and P. contorta. A single Pinus ponderosa was found on a south plot, and one Pseudotsuga menziesii was found on a north plot. The two unburned forest plots supported albicaulis (north-44, south-19) and lasiocarpa (north-38, south-11), with a few Pinus contorta and Picea engelmannii on the south. Understory Plant Species Occurrence on Plots A comparison of understory plant composition among plot series shows the north and south plot series to be very different, reflecting relative occurrence of xeric and mesic species, and the ridge to be intermediate between north and south. Figure 3 shows

PAGE 34

25 PLOT SERIES FREQUENCY OCCURRENCE (%) ----r----------------------------------------------------------10 20 30 40 50 60 70 :8o 90 100 VACCINIUM SCOPARIUM (95.3) I N LUZULA SP. (85. 7) J 0 R XEROPYLLUM TEN AX (71. 4) I T H CAREX GEYERI (66. 7) I POLYTRICHUM SP. (66. 7) I --------------------------------------------------------------XEROPHYLLUM TENAX (100) J R CAREX GEYERI (95. 6) J I 0 V.t.CCINIUM SCOPARIUH (78. 3) I G E LUZULA SP. (73. 9 ) FESTUCA SP. ( 65. 2) I ---------------------------------------------------------------CAREX GEYERI (100) -:=1 s XEROPHYLLUH TENAX (100) I 0 u FESTUCA SP. (78.9) I T H PHLEUH PRATEHSE (78. 9) I VACCINIUM SCOPARIUM (57. 9) I LUZULA SP. (57. 9) --------------------------------------------------------------VACCINIUM SCOPARIUH (78 6) I R POLYTRICHUM SP. (78. 6) I 0 I A XEROPHYLLUM TENAX (71. 4) D EPILOBIUM SP. (64.3) J LUZULA SP. (50) c. GEYERI (50) P. PRATEHSE (50) Fig. 3. Five most frequently occurring plant species on plots, by plot series.

PAGE 35

26 the five most frequently occurring plant species by plot series. These species are dominant species for the Bitterroot region, and are classified as mesic or mesoxeric (Lackschewitz 1986). The major vegetation differences between aspects are the result of differences in occurrence of common species (Figure 3): Vaccinium scoparium (north -95.3%, ridge -78.3%, south -57.9%); Luzula sp., north-85.7%, ridge -73.9%, and south -57.9%); Polytrichum sp. (north -66.7%, ridge -21.7%, south -5 Carex rossii, (north -61.9%, ridge -52.2%, south-26.3%). Grasses and other xeric vegetation were more frequent on the south plot series, less on the ridge, and even less on the north. Xerophyllum tenax occurred on all plots on both south and ridge plot series, and on 71.4% of north plots. Carex geyeri occurred 100% frequency on south plots; 95.6% on ridge and 66.7% on the north;. Phleum pratense. frequency occurrence was: north -42.9%, ridge -43.5%, south -78.9%; Dactylis glomerata occurred on 4. 8% of the north plot series, 19.9% of the ridge, and 52.6% of the south. Road plots showed some similarities to both the north and ridge plot series. The road plots resembled the north in order of species frequencies on plots: Vaccinium scoparium (78.6%) > Xerophyllum tenax (71.4%) > Carex geyeri (50%) > Anaphalis margaritacea (35.7%) >

PAGE 36

27 Salix sp. (28.6%). Frequency values for the road, V. scoparium (78.6%), pratense (50%) and A. margaritacea (35.7%), were similar to the ridge. The two unburned forest plots exhibited understory occurrences of X. tenax, V. scopar!um, and C. geyeri. P. albicaulis Clusters and Microsite Vegetation Cluster sizes (number of individual trees per P. albicaulis site) ranged overall from 1 to 9 (Table 2). The ridge and north plot series showed a range of 1 to 8, the south 1 to 5, and the road 1 to 9. Mean cluster size, respectively, was 2.06, 1.57, and 2.49. Kruskal-Wallis ANOVA indicated a significant difference in cluster size among the three ridge plot . . series (H = 6.475, df 2, P = 0.039). Further analysis with the Wilcoxon Rank Sum test showed south plots differing significantly both from the north (Z = 1.951, P = 0.026) and the ridge (Z = -2.543, P = 0.006). On every transect, the largest percentage of sites were solitary trees (south -66.2%, north 52%, road 51.1%, ridge 45.4%). As cluster size increased, the percentage of clusters in each size category decreased. One exception was on the road plot series, where clusters of 3 trees accounted for 22.2% of

PAGE 37

28 TABL. E 2. Pinus albicaulis clusters: size and frequency. PLOT SERIES NORTH RIDGE SOUTH ROAD Number of sites 271 119 65 45 Number of trees per site Mean 1.92 2.06 1.57 2.49 Range 1-8 1-8 1-5 1-9 Cluster size Frequencies (%) 1 52.0 45.4 66.2 51.1 2 26.6 28.6 21.5 6.7 3 8.5-12.6 4.6 22.2 4 6.3 6.7 4.6 6.1 5 4.8 4.2 3.1 2.2 6 1.1 0.8 4.4 7 0.4 0.8 2.2 8 0.4 0.8 2.2 9 2.2

PAGE 38

29 all sites, and clusters of 2 trees accounted for only 6.7%. A of 17 trees was found near a road plot; this is larger than the cluster size range of 1 to 15 previously reported by Tomback (1982). Pinus albicaulis sites on the three different ridge plot series showed dissimilar frequencies of microsite vegetation species. Vaccinium scoparium was most frequent on all plot series, and Polytrichum sp. or Carex rossii was the second most frequent. Many albicaulis' trees were growing in the middle of a clump of C. rossii. On north and ridge sites, v. scoparium frequency was 86% and 69.7%, respectively; and all other species accounted for 25% or less of occurrences. South plot series albicaulis sites showed no microsi te plant species frequency greater than 25% (Figure 4). Despite the frequent occurrence of Xerophyllum tenax on many plots, no albicaulis trees were actually growing within formidable clumps. On the road transect, V. scoparium and Polytrichum each occurred on more than 50% of the P. albicaulis sites. The species richness was lower on road but this is probably due to the smaller plot size and thus smaller number of P. albicaulis sites.

PAGE 39

30 PLOT SERIES FREQUENCY OCCURRENCE (%) --------------------------------------------------------------. 10 20 30 40 50 60 70 80 90 (86) I N I POLYTRICNUM SP. (21.8) 0 R J CAREX ROSSll <17. 7) T H I (14) LUZULA SP. (10) -------------------------------------------------------------VACCINIUM SCOPARIUM (69.7) I R J CAREX ROSSll (24. 4) I B D CAREX GEYERI (8.4) E XEROPHYLLUM TENAX (7. 6) p EJtiLOBIUM SP. (5) 1---I VACCIHIUM SCOPARIUM (24. 7) s I CAREX ROSSII (23) 0 u I Fr.sTUCA SP. (13.9) T p "' lltozi!.SIA FERitUGIMEA (10.8) t=J CAREX GEYERI uo. a> 1---1----------------------------------------------------------VACCINIUM SCOPARIUM (62 . 2) I R f'QLYTRICHUM SP. (55. 6) I -0 A IEPILOBIUM SP. <17. 8) D I MENZIESIA FERRUGINEA (13.3) w I:AIIIEX GEYEJIII ca. !I> Fig. 4 Five most frequently occurring plant species on P. albicaulis microsites, by plot series.

PAGE 40

Narest Objects on P. albicaulis Microsites 31 The total number of P. albicaulis microsites on all four plot series combined (ridge,. north, south, road) was 500. Five categories of objects were noted on P. albicaulis microsites within 15 em of the tree trunk( s). Wood pieces were the nearest object on 122 sites ( 24.4%), a log or single large branch on 112 ( 22.4%), a fallen snag on 69 ( 13.8%), rocks on 68 ( 13. 6%) and a standing snag or stump on 34 ( 6. 8%) Ninety-five (19%) of the microsites were categorized as 'open', because no object was present. Vegetation Comparisons Between Plot Series and P. albicaulis Microsites Analysis of vegetation near albicaulis sites shows some similarities to the vegetation frequencies on plot series. Vaccinium scoparium was the most frequently occurring species on tree microsites on all plot series, with a frequency on ridge microsites intermediate between north and south, and a frequency on the road microsites most resembling that of the north. Vaccinium scoparium showed the same frequency pattern on plots, with the exception that the road plots mostly resembled the ridge plots. Polytrichum, a mesic species, was the second most frequently occurring

PAGE 41

32. species on north and road microsi tes and showed high occurrences on north and road plots. Distinct differences between microsite vegetation and plot vegetation are reflected in the frequencies of C. rossii and X. tenax. Although _g_. rossii is the second most frequently occurring species on ridge and south microsites, it ranks much lower in occurrence on ridge and south plots. Conversely, X. tenax was present on all south and ridge plots and over 70% of the road and north plots but occurred on no north microsites and on fewer than 8% of microsites on the other three plot series. P. albicaulis Site Density, Elevation, and Distance The total number and range per plot of P. albicaulis sites on each plot series were: ridge -total 119, range 0 to 24; north -total 271, range 1 to 32; and south -total 65, range 1 to 11 (Table 3). The ranges of albicauiis site density per m2 per plot for the ridge, north and south plot series, respectively, were: 0 to 0.192, 0.006 to 0.512, and 0.001 to 0.11. Mean site density per plot (defined as albicaulis sites per m2 ) were: north 0.14, ridge 0.04, south 0.02. Road plots had a total of 45 albicaulis sites, with a range of 0 to 10 per plot, a mean density of 0.02

PAGE 42

33 TABLE 3. Pinus albicaulis regeneration by plot series: density (sites or trees per m2>, age (yrs), height (em). N =total number. PLOT SERIES NORTH RIDGE SOUTH ROAD Sites N 271 119 65 45 Range per plot 1-32 0-24 1-11 0-10 Density Mean 0.14 0.04 0.02 0.02 Range 0.006-0.512 0-0.192 0.001-0.11 0-0.007 Trees N 520 Range per plot 1-61 Density Mean 0.275 Range 0.006-0.976 N Mean Range Mean Range 334 7.7 1-21 29.7 1.5-234 245 0-56 0.099 0-0:448 149 8.3 0.5-20 36.2 3-238 102 112 1-13 0-22 0.035 0.051 0.001-0.13 0-0.127 Age 79 8.1 1-21 Height 30.6 3-202.3 47 6 0.5-19 19.2 1-186

PAGE 43

34 per m2 and a density range of 0 to Oo067 per m2o Site densities for the north, ridge and south plot series showed differences of high statistical significance (Kruskal-Wallis ANOVA, H = 10o614, df = 2, p and elevation, and site density and distance from seed source, was determined for each plot series (Pearson correlation, Table 4) 0 Significant negative correlations with P <0.01 between site density per m2> and distance occurred on all series of plots except the. road: north, r = -Oo832; r = -0o618; south, r = -oo714 (Table 4). Site density and elevation were significantly correlated for the three ridge plot series with P
PAGE 44

35 Table 4. Pearson correlation analysis: Pinus albicaulis density (sites or trees per m2l vs. elevation, density vs. distance from seed source, age vs. height. (NS = not significant, P >0.05). Log = Log10 In all cases, df = no. of plots minus two. Site density vs. elevation Tree density vs. elevation Site density vs. distance Log site density vs. distance Tree density vs. distance Log tree density vs. distance Tree age vs. height Plot series North Ridge South Road North Ridge South Road North Ridge South Road North Ridge South Road North Ridge South Road North Ridge South Road North Ridge South Road No. of plots 21 22 19 13 21 22 19 13 21 22 19 13 21 22 19 13 21 22 19 13 21 22 19 13 334 149 79 47 Correlation coefficient(r) 0.982 0.632 0.682 0.428 0.877 0.647 0.682 0.194 -0.832 -0.618 -0.714 -0.374 -0.820 -0.733 -0.779 -0.500 -0.822 -0.627 -0.703 -0.207 -0.788 -0.751 -0.678 -0.334 0.818 0.783 0.827 0.742 p <0.001 <0.01 <0.01 NS <0.001 <0.01 <0.01 NS <0.001 <0.01 <0.001 NS <0.001 <0.001 <0.001 NS <0.001 <0.01 <0.001 NS <0.001 <0.001 <0.01 NS <0.001 <0.001 <0.001 <0.001

PAGE 45

36 source (independent variable) The plot produced a negative exponential curve with the following regression equation: Y = -0.00006(X) + 0.175 (Df = 60, r = -0.590, P <0.001). When log10 density was used, the scatterplot produced a straight line and the regression equation was: Y = -0.0005(X) 0.806, with r = and P <0.001. Regression analysis of the road plots indicated no correlations between P. albicaulis site density and distance from seed source. Even when the log10 density was used, the regression was not significant. P. albicaulis Tree Density, Elevation and Distance A total of 1029 Pinus albicaulis trees (more than one tree may occur at a site) were found on the two study sites: 917 on the ridge (includes ridge, north and south plot series) and 112 on the road plot series (Table 3) Total P. albicaulis tree counts on the north plot series were twice as large as on the other two ridge plot series (520 versus 245 for the ridge series and 102 for the south series), although area sampled on the ridge was 1.64 times larger than on the north, and the south was 1. 78 times larger. Tree densities,

PAGE 46

measured as trees per ranged as follows: 37 north (0.006 to 0.976), ridge (0 to 0.448), south (0.001 to 0. 13) (Table 3) Mean tree densities per m2 were: north 0.275, ridge 0.099, and south 0.035. KruskalWallis ANOVA showed P. albicaulis tree densities per plot were significantly different among the north, ridge, and south plot series (H = 14.191, df = 2, P
PAGE 47

Regression Model for Tree Density vs Distance 38 Combining north, ridge and south plot series, the scatterplot for tree density against distance also assumed a negative exponential curve (Figure 5). The regression equation was: Y = -0.0001(X) + 0.344, with df = r = -0.571, and P <0.001. With logw of tree density (Figure 6), the plot became linear and the equation became Y = -0.0005(X) 0.542 with r = -0.666 and P <0.001. The same analysis for albicaulis trees on the road plots failed to produce a significant correlation between density and distance. P. albicaulis Tree Age, Height, Distance Each separate stem within a Pinus albicaulis cluster was considered a tree for aging purposes; the number of trees aged on north plots was 334; on the ridge, 149; and on the south, 79 (Table 3). Field data show that the ages of trees within an individual cluster were the same or similar, usually differing by only 1 or 2 years, but in no case more than 4 years. In virtually all cases no stems were fused at the base, thus it must

PAGE 48

39 9 U) TREES/m2 -..J en 0 0 u; -i l> z (') ITI 3 4 01 9 .en (J1 0 Fig. 5. Scatterplot of P. albicaulis tree density vs. distance, for three ridge plot series combined.

PAGE 49

40 I 0 b TREES/m2 (log) 0 CJ) z (') I'll -3 ..... 01 m (JI 0 Fig. 6. Scatterplot of albicaulis log10 tree derisity vs. distance, for three plot series combined.

PAGE 50

41 be emphasized that each member of a cluster appeared to be a distinct individual. The age differences observed were probably the resu1 t of counting error or later germination. Age ranges of trees were nearly identical for the three ridge plot series: north and south, 1 to 21 yr; ridge 0.5 to 20 yr. Compared to the ridge plots and south plots, the north trees had a Slightly lower mean (7.7 yr vs. 8.3 yr, and 7.7 yr vs. 8.1 yr). Fortyseven trees were aged on road plots, with a range of 0.5 to 19 yr and a lower mean (6 yr) than any of the ridge plot series. P. albicaulis trees on the two unburned forest plots showed a mean age of 127.7 yr and a range of 76 to 202 yr. Tree heights ranged from 1 to 238 em (Table 3). The range and mean for the ridge plot series was the highest (3 to 238 em, 36.2 em), followed by the north (1.5 to 234 em, 29.7 em), south (3 to 202.3 em, 30.6 em) and road (1 to 186 em, 19.2 em), the lowest of all. Highly significant Pearson correlations (P <0.001) were found for tree age vs. tree height for all plot series: north (r = 0.818, df = 332), ridge (r = 0.783, df = 147), south (r = 0.827, df =77), and road (r = 0.742, df = 45).

PAGE 51

CHAPTER IV DISCUSSION Regeneration: Consequences of Seed Dispersal by Nutcrackers Regeneration patterns of Pinus albicaulis in the Sleeping Child Burn demonstrate the importance of nutcrackers as seed dispersers. A few nutcrackers were seen in the study site during field work in 1987, but the lack of a P. albicaulis cone crop that year prevented direct observation of foraging or caching behavior. However, the occurrence of tree clusters and the great distances from regeneration to seed source confirm previous suggestions that nutcrackers are the primary dispersers of P. albicaulis and that their caching behaviors effect establishment of the conifer (Lanner 1980; Hutchins and Lanner 1982; Tomback 1982). This study also confirms that squirrels and other rodents play a very minor role at best in the dissemination of Pinus albicaulis. As reported by Hutchins and Lanner (1982), red squirrels (Tamiasciurus hudsonicus), chipmunks and other mammals are not reliable dispersers compared to nutcrackers. Their limited home ranges indicate that a more mobile

PAGE 52

43 disperser has buried seeds to account for the observed distances between seed source and regeneration. Although Stellar's jays (Cyanocitta stelleri) were seen in the study area, they also establish small territories, damage seeds upon removal from the cone, and bury one-seed caches (Vander Wall and Balda 1981). Nearly half of all Pinus aibicaulis sites were composed of two or more trees, with clusters of up to 9 trees observed on the road plot series, and a cluster of 17 trees observed near that transect. The distinct multiple stems and the similarity in tree ages within a cluster are evidence of germination from buried seed caches. Interestingly, noP. albicaulis older than 21 yr was found on the plots, even though the burn occurred 26 years prior. This may indicate a lack of regeneration during the five years immediately following the burn, either because of no cone crops or unfavorable conditions for seed germina.tion. Dispersal distances of P. albicaulis in the Sleeping Child Burn are similar to the distances noted by Tomback (1978) and Hutchins and Lanner (1982), although their studies occurred in California and Wyoming, respectively. Regeneration from the seed source in the adjacent forest extended into the burn westward for 3. 65 km and southward nearly 8 km. Because

PAGE 53

44 P. albicaulis trees were found on the study plots furthest from the seed source, we assume regeneration has occurred beyond these distances also. These patterns lead us to conclude that nutcrackers are responsible for nearly 100% of the observed P. albicaulis regeneration in the burn. An interesting pattern with potential application was the significant linear relationship observed between the density of regenerating P. albicaulis on the three ridge plot series and the aerial distance from the seed source (Figures 5 and 6). This correlation reflects the dispersal pattern generated by the nutcracker, but at this time it is not possible statistically to separate the effect of elevation on density from the effect of distance on density. Elevation also correlated significantly with distance from seed source on the three ridge plot series, and the change in environmental conditions (moisture, temperature) that accompany changes in elevation conceivably may have an effect on seed germination and seedling survival. However, albicaulis grows well in western Montana from elevations of 2290 2620 m and grows as far down as 2130 meters (Arno and Petersen 1983). All plots on the three ridge plot series were above 2130 m and nearly half of the plots on each series were within the 2290 2620 m range. Thus, the pattern

PAGE 54

45 generated should primarily reflect differences in distance and not elevation effects. The regeneration patterns that result from nutcracker seed dispersal along a distance gradient from a known seed source suggest that the birds cache relatively more seeds close to the parent trees and make fewer caches per m2 at increasing distances. The result is a negative exponential curve (tree density vs. distance) (Figure 6), a shape similar to the seed shadow curves generated by individual wind-dispersed conifers (McCaughey et al. 1986), but the P. albicaulis regeneration extended over much greater distances. As discussed by Tomback et al. (in preparation), the dispersal distances for P. albicaulis may typically be greater than that for Abies lasiocarpa, a competitor in succession, giving albicaulis the edge because of nutcracker-mediated dispersal. The negative exponential curve was confirmed as well for P. albicaulis regeneration in the Saddle Mountain Burn to the south of the Sleeping Child Burn (Tom back et al. in prep. ) For management purposes, this curve may be useful for its predictive value (e.g., Keane et al. 1989). Potential application includes models of P. albicaulis regeneration for prescribed fires or for past fires such as sui table areas of Yellowstone National Park burned in 1988.

PAGE 55

46 The correlation between density and distance did not hold for the road plot series. This could be due to the effects of elevation on P. albicaulis seed germination and seeding survival. All of the road series plots were at elevations of 2290 m or less, and thus were below the range within which P. albicaulis grows well in this region. The forest bordering the southern part of the burn is composed of mixed conifers, and thus may be a better source of regeneration for .Abies lasiocarpa, Picea engelmannii, and Pinus contorta than for P. albicaulis. Variation in Density and Cluster Size The data show significantly higher densities of both Pinus albicaulis sites and trees on the north aspect of the ridge than on the ridge or south (at least 7 times greater mean density on the north) This is consistent with the prevalence of P. albicaulis on cool exposures and moist sites in semiarid ranges (Arno and Hoff 1987). North slopes, with deeper soil mantles, more moisture and cooler temperatures, may provide more sui table germination conditions. Leadem ( 1986) examined the dormancy-breaking mechanisms of albicaulis seeds, and found that cold temperatures were necessary to overcome physiological barriers to growth. Alternatively, nutcrackers may cache more on north

PAGE 56

47 slopes, Tomback (1978, 1982) and Vander Wall and Balda (1977) suggest that south slopes are preferred. Regeneration density must be a result of both 1) nutcracker caching preferences, and 2) site quality. Competition from other plant species could also be a factor in density variation among plot series. Xerophyllum tenax tufts were widespread on south slopes, as were grass species; both are potential competitors for young trees. Another interesting pattern was the relatively large cluster sizes found on the road plot series, where the largest percentages of cluster sizes greater than 5 occurred, and a 17-stem cluster was found nearby. Distances from road plots to the seed source averaged over twice as far as ridge plots. This poses an interesting question: do nutcrackers bury larger clusters of seeds when they fly greater distances? Or perhaps germination and seedling survival conditions were better on the road plot series despite the lower elevations. Although highly tolerant of crowding, P. albicaulis seedlings within a cluster may compete to some degree as they mature. The lower height values observed for trees on the road plot series could be indicative of intra-cluster competition.

PAGE 57

48 Post-Fire Disturbance Successional processes have been complicated by a history of disturbances in the study area. Lyon (1976) summarized several years of post-fire management activities in the Sleeping Child burn, including chemical thinning of lodgepole pine, pole cutting, cattle grazing, Christmas gathering, road building, grasses, and recreational tree cutting, firewood aerial sowing of annual uses such as hunting and camping. He estimates little long-term effect from these management activities, since competition would have produced substantial losses in dense stands anyhow. However, some of these activities, especially chemical thinning of Pinus contorta, grazing, and sowing of grasses, may affect the.type and density of regrowth within some areas of the burn. A case in point is the road plot series, with an elevation range stated to be dominated by contorta prior to the fire (Lyon and Stickney 1976), but with regeneration density of P. contorta lower than Abies lasiocarpa. Evidence of grazing and browsing was seen often on south slopes, which tended to be less steep in both study sites. Range cattle were encountered during field work on south aspects as well as on the dirt roads that access the burn.

PAGE 58

Vegetation Comparisons Between Plots and Microsites 49 Much consistency was se.en in the understory species on P. albicaulis microsi tes overall, despite differences in characteristic vegetation found on the four plot series. The predominance of species such as Vaccinium scoparium, Polytrichum sp., and Carex rossii at tree sites reflects the preference of P. albicaulis for mesic to meso-xeric habitat in these mountains (Lackschewitz 1986). This contrasts with the high frequency of tenax which grew as large, dense clumps on all the plot series. The growth form of tenax undoubtedly makes it a superior competitor compared to young P. albicaulis, and nutcrackers probably do not cache seeds near these plants. Fire and P. albicaulis Regeneration Twenty-six years after a severe fire, the Sleeping Child Burn shows regeneration of Pinus albicaulis to be well under way. Tree counts of the four conifer species on the higher elevation study site (the three ridge plot series) support the observation that stand-replacing fires favor P. albicaulis in relation to its more shade-tolerant subalpine associates, Abies lasiocarpa and Picea engelmannii (Arno 1986, Arno and Hoff 1987). This pattern is complicated

PAGE 59

50 by a higher density of Pinus contorta on the plots with south exposures and on the few plots at the west end of the ridge that are within lower subalpine elevations. Our two plots in the unburned forest at the upper edge of the burn show species composition dominated by albicaulis and A. lasiocarpa, the potential climax species for the upper subalpine habitat type (Pfister et. al 1977). The striking abundance of contorta regeneration on the lower subalpine road plots may be due to the seral contorta forest existing at the time of the fire. Large numbers of seeds were probably released as the serotinous cones opened from the heat (Fischer and Bradley 1987). Tree counts for the road plot series show Pinus albicaulis regeneration to be a relatively minor component in relation to Abies lasiocarpa and Pinus contort a. Regeneration patterns along this plot series, which extends into the upper limits of the lower subalpine, is most likely the consequence of the great distance from P. albicaulis seed crops, lower elevation, and closer proximity to seed crops of the other conifers. Clearly, the result of severe fire in the Sleeping Child Burn is regeneration of albicaulis at higher elevations. The post-fire dominance of this species is typical of the forest overstory composition

PAGE 60

51 for much of the northern Rocky Mountains that existed prior to 20th century fire suppression policies.

PAGE 61

CHAPTER V CONCLUSIONS Regeneration of Pinus albicaulis, a major component of northern Rocky Mountain subalpine forests and an important wildlife food source, is well underway in an area where it had been declining. Clark's nutcrackers are responsible for the Pinus albicaulis regeneration patterns observed in the Sleeping Child Burn. This severely burned area now has P. albicaulis trees growing at distances of at least 8 km from the nearest seed source, distances that can only be accounted for by the seed caching behavior of the birds. The even-aged tree clusters that were present on both study sites are evidence of seed caches buried by nutcrackers. P. albicaulis regrowth on the ridge study site along a distance gradient from a known seed source corresponds to a linear relationship between regeneration density and distance. This indicates that nutcracker caching behavior is most concentrated nearest the seed source but extends for distances of several kilometers. Birds may also bury larger numbers of seeds per cache when they fly greater distances from the harvest site.

PAGE 62

53 The regeneration densities for four conifer species indicate that avian dispersal results in earlier establishment of P. albicaulis during post-fire succession, compared to its upper subalpine associates Abies lasiocarpa and Picea engelmannii. The regeneration species composition at high subalpine elevations resembles that of regional .forests prior to the fire suppression policies initiated in the early 1900's.

PAGE 63

54 LITERATURE CITED Arno, S. F. 1976. The historical role of fire on the Bitterroot National Forest. USDA Forest Service Res. Paper INT-187, p 29. Intermountain Forest and Range Experiment Station, Ogden, Utah. ------1986. Whitebark pine cone crops -a diminishing source of wildlife food? Western Journal of Applied Forestry 1(3):92-94. Arno, S. F., and T. D. Petersen. Variation in estimates of fire intervals: a closer look at fire history on the Bitterroot National Forest. USDA Forest Service Research Paper INT-301, p 8. Intermountain Forest and Range Experiment Station, Ogden, Utah. Arno, S. F., D. G. Simmerman, and R. E. Keane. 1985. Forest succession on four habitat types in western Montana. USDA Forest Service Gen. Tech. Rep. INT-177, p 74. Intermountain Forest and Range Experiment Station, Ogden, Utah. Arno, s. F., and R. J. Hoff. 1989. Whitebark pine. In:

PAGE 64

55 Silvics of North America, Volume 1, Conifers. USDA Forest Service, Washington, D.C. (In press). Bailey, N. T. J. 1981. Statistical Methods in Biology, Second Edition. John Wiley and Sons, New York. Darn, R. D. 1984. Vascular Plants of Montana. Mountain West Publishing, Cheyenne, Wyoming. Finklin, A. I. 1983. Weather and Climate of the SelwayBitterroot Wilderness. University Press of Idaho. Fischer, W. c., and A. F. Bradley. 1987. Fire Ecology of western Montana forest habitat types. USDA Forest Service General Technical Report INT-223. Intermountain Research Station, Ogden, Utah. Furnier, G. R., P. Knowles, M. A. Clyde, and B. P. Dancik. 1987. Effects of avian seed dispersal on the genetic structure of whitebark pine populations. Evolution 4(13):607-612. Gravetter, F. J., and L. B. Wallnau. 1985. Statistics for the Behavioral Sciences. West Publishing Company, St. Paul, Minnesota.

PAGE 65

56 Heinselman, M. L. 1978. Fire intensity and frequency as factqrs in the distribution and structure of Northern ecosystems. In: Fire Regimes and Ecosystem Properties, proceedings of the 1978 Conference. USDA Forest Service General Technical Report W0-26, p 594. Hitchcock, C. L., and A. Cronquist. 1973. Flora of the Pacific Northwest. University of Washington Press, Seattle, Washington. Hutchins, H. E., and R. M. Lanner. 1982. The central role of Clark's nutcracker establishment of whitebark pine. 55:192-201. in the dispersal and Oecologia (Berl)(1982) Keane, R. E., S. F. Arne, J. K. Brown, and D. F. Tomback. 1989. Modeling stand dynamics in whitebark pine forests. Ecological Modeling (in press}. Kendall, K. C. 1980a. Bear-squirrel-pine nut interaction. In: Yellowstone grizzly bear investigations. Annual Report 1978-79, USDI-National Park Service, pp 51-60. Kendall, K. C. 1980b. grizzly bears, 1978-79. Food habits of Yellowstone In: Yellowstone grizzly bear

PAGE 66

57 investigations. Annual Report 1978-79, USDI-National Park Service, pp 24-34. Lackschewitz, K. 1986. Plants of west-central Montana -Identification and ecology: annotated checklist. USDA Forest Service Gen. Tech. Rep. INT-217, p 128. Intermountain Research Station, Ogden, Utah. Lanner, R. M. 1980. Avian seed dispersal as a factor in the ecology and evolution of limber and whitebark pines, pp. 15-48. In: B.P. Dancik and K.O. Higginbotham (eds.), Proceedings of the Sixth North American Forest Biology Workshop. University of Alberta, Edmonton, Alberta, Canada. Lanner, R. M., and S. B. Vander Wall. 1980. Dispersal of limber pine seed by Clark's nutcracker. Journal of Forestry 78: 638-640. Leadem, C. L. 1986. Seed dormancy in three Pinus species of the inland mountain west. In: Proceedings -Conifer Tree Seed in the Inland Mountain West Symposium. USDA Forest Service General Technical Report INT-203, p 289. Intermountain Research Station, Ogden, Utah. Ligon, J. D. 1978. Reproductive interdependence in

PAGE 67

58 pinyon jays and pinyon pines. Ecological Monographs 48:111-126. Linhart, Y. B., and D. F. Tomback. 1985. Seed dispersal by nutcrackers causes multi-trunk growth form in pines. Oecologia (Berlin) 67:107-110. Lotan, J. E. 1975. The role of cone serotiny in lodgepole pine forests. In: Management of Lodgepole Pine Ecosystems Symposium Proceedings. Washington State University Cooperative Extension Service, Pullman, Washington. ------1976. Cone Serotiny -Fire relationships in lodgepole pine. In: Proceedings, Tall Timbers Fire Ecology Conference No. 14 and Intermountain Fire Research Council Fire and Land Management Symposium. Lyon, L. J. 1976. Vegetal development on the Sleeping Child burn in western Montana 1961 to 1973. USDA Forest Service Research Paper INT-184, p 24. Intermountain Forest and Range Experiment Station, Ogden, Utah. Lyon, L. J., and P. F. Stickney. 1976. Early vegetal succession following large Northern Rocky Mountain wildfires. Proc. Montana Tall Timbers Fire Ecology

PAGE 68

59 Conference and Fire and Land Management Symposium, No. 14, 1974:355-375. Tall Timbers Research Station, Tallahassee, Florida. McCaughey, W. w., w. C. Schmidt, and R. C. Shearer. 1986. Seed-dispersal characteristics of conifers in the inland mountainous west. In: Proceedings -Conifer Tree Seed in the Inland Mountain West Symposium. USDA Forest Service General Technical Report INT-203, p 289. Intermountain Research Station, Ogden, Utah. Mirov, N. T. 1967. The Genus Pinus. Ronald Press Company, New York. Mirov, N. T., and J. Hasbrouck. 1976. The Story of Pines. Indiana University Press, Bloomington, Indiana. Pfister, R. D., B. L. Kovalchik, S. F. Arno, and R. C. Presby. 1977. Forest habitat types of Montana. USDA Forest Service Gen. Tech. Rep. INT-34, p 174. Intermountain Forest and Range Experiment Station, Ogden, Utah. Romme, W. H. 1982. Fire and landscape diversity in subalpine forests of Yellowstone National Park. Ecological Monographs 52(2):199-221.

PAGE 69

60 Siegel, S. 1956. Nonparametric Statistics for the Behavioral Sciences. New York:McGraw-Hill. Tomback, D. F. 1978. Foraging strategies of Clark's nutcracker. The Living Bird (16th Annual, 1977):123-161 1982. Dispersal of whitebark pine seeds by Clark's nutcracker: a mutualism hypothesis. Journal of Animal Ec6logy 51:451-467 ------1983. Nutcrackers and pines: coevolution or coadaptation? In: M.H. Niteck,editor. Coevolution. University of Chicago press. pp. 179-223. ------1986. Post-fire regeneration of krummholz whitebark pine: a consequence of nutcracker seed caching. Madrona, 33(2):100-110. Tomback, D. F., L. Hoffman, and S. Sund. Co-evolution of whi tebark pine and nutcrackers: implications for forest regeneration. (In preparation). Tomback, D. F., and Y. Linhart. The coevolution of bird-dispersed pines. (In preparation).

PAGE 70

61 Turcek, F. J,. and L . Kelso. 1968. Ecological of food transportation and storage in the Corvidae. Communications in Biology, Part A, 1(4):277-297. Vander Wall, S. B., and R. P. Balda. 1977. Coadaptations of the Clark's nutcracker and the pinon pine for efficient seed harvest and dispersal. Ecological Monographs, 47(1):89-111. 1981. Ecology and evolution of behavior in conifer-seed-caching Tierpsychol. 56:217-242. food-storage corvids. z. Wright, H. A., and A. W. Bailey. 1982. Fire Ecology: United States and Southern Canada. John Wiley and Sons, New York.

PAGE 71

APPENDIX A. Plant Species List for Sleeping Child Burn COMMON NAME SCIENTIFIC NAME Subalpine fir Abies lasiocarpa Yarrow Achilles millefolium v. alpicola Pearly-everlasting Anaphalis margaritacea Pussytoes Antennaria Sandwort Arenaria aculeata Arnica Arnica cordifolia Mountain arnica Arnica latifolia v. latifolia Aster Family Asteracea e Smooth brome-grass Bromus inermis ssp. inermis Elk sedge Carex geyeri Ross sedge Carex rossii Composite (.unknown)Unknown composite Orchard-grass Dactylis glomerata Fireweed Epilobium sp. Buckwheat Eriogonum sp. Fleabane Erigeron sp. Forb (unknown) Unknown forb White hawkweed Hierachium albiflorum Western hawkweed Hierachium albertinum Labrador te.a Ledum glandulosum Patridge foot Leutkea pectinata Lichen ( unk.nown) unknown lichen Honeysuckle Lonicera utahensis Lupine Lupinus sp. Woodrush Luzula sp. Fools huckleberry Mensiesia ferruginea v. glabella Alpine fernleaf Pedicularis contorta Penstemon Penstemon procerus Common timothy Pnleum pratense Engelmann spruce Picea engelmannii Whitebark pine Pinus albicaulis Lodgepole pine Pinus contorta v. latifolia Ponderosa pine Pinus ponderosa v. scopulorum Bluegrass Poa sp. Bistort Polygonum bistortoides Moss Polytrichum sp. Jacobs ladder Polemonium sp. 62

PAGE 72

APPENDIX A. (cont.) COMMON NAME Douglas-fir Gooseberry Sheepsorrel Catchfly Green needlegrass Needlegrass Snow berry Dandelion Downy oatgrass Huckleberry Grouse berry False hellebore Beargrass SCIENTIFIC NAME Pseudotsuga menziesii Ribes sp. Rumex acetosella Siiene spaldingi Stipa nelsonii Stipa occidentalis Symphoricorpus sp. Taraxacum officinales Trisetum spicatum Vaccinium globulare Vacciriium scoparium Veratrum viride Xerophyllum tenax 63