The interaction between Clark's nutcracker and ponderosa pine, a "wind-dispersed" pine

Material Information

The interaction between Clark's nutcracker and ponderosa pine, a "wind-dispersed" pine energy efficiency and multi-genet growth forms
Torick, Lisa L
Publication Date:
Physical Description:
ix, 44 leaves : illustrations ; 29 cm


Subjects / Keywords:
Corvidae -- Behavior -- Colorado ( lcsh )
Pine -- Seeds ( lcsh )
Seeds -- Dispersal ( lcsh )
Corvidae -- Behavior ( fast )
Pine -- Seeds ( fast )
Seeds -- Dispersal ( fast )
Colorado ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 42-44).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Arts, Biology.
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Lisa L. Torick.

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:
34234067 ( OCLC )
LD1190.L45 1995m .T67 ( lcc )

Full Text
Lisa L. Torick
B.A., Elmira College, 1984
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts

This thesis for the Master of Arts
degree by
Lisa L. Torick
has been approved for the
Graduate School
Diana F. Tomback
(/ 7 Date

Torick, Lisa L. (M.A., Biology)
The Interaction between Clarks Nutcracker and Ponderosa Pine, a "Wind-
Dispersed Pine: Energy Efficiency and Multi-Genet Growth Forms
Thesis directed by Associate Professor Diana F. Tomback
The Clark's Nutcracker {Nucifraga columbiana) disperses seeds for several
pine mutualists. It is also known to harvest and store the seeds of some wind-
dispersed pines. In the first part of this study I examine the timing of the
nutcracker's switch from limber pine seed (Pinus flexilis), a bird-dispersed pine, to
ponderosa pine seed (Pinus ponderosa), a wind-dispersed pine, to determine if
nutcrackers forage in an energetically optimal manner.
Nutcrackers harvested and cached limber pine seeds until the supply was
depleted. When seeds became scarce, they did not switch to ponderosa pine seed
even though the cones were open and seeds were abundant. They found enough
limber pine seeds to eat but not to store. When nutcrackers finally switched, they
harvested ponderosa pine seeds from the ground and made caches. Other studies
suggest that nutcrackers are optimal foragers. However, if nutcrackers are optimal
foragers, they would have switched to ponderosa pine seed earlier when the
energetic return for their efforts exceeded the return on limber pine seed. The
timing of switching is probably dependent on the relative abundance of cone crops.
In this season the cone crop for both pines was very large.
The nutcracker is an important seed disperser for several species of pines.
It caches the seeds and often stores more than it retrieves. Those seeds that are in

multi-seed caches may germinate and develop into multi-genet tree clusters.
However, single genet multi-trunk trees also occur and must be examined
genetically to determine which growth form they are. Tree clumps also occur in
wind-dispersed pine trees. For the second part of this study, I tested the hypothesis
that some of these tree clumps are multi-genet clusters. Using starch gel protein
electrophoresis, I examined tree clumps of three wind-dispersed pines, ponderosa,
brisdecone (Pinus aristata), and lodgepole pine (Pinus contorta). I found that 80%
of ponderosa, 20% of bristlecone, and 40% of lodgepole pine tree clumps were
multiple genets. Thus, tree clusters may commonly occur in wind-dispersed pines
and may result from seed dispersal by nutcrackers and rodents.
This abstract accurately represents the content of the candidate's thesis. I
recommend its publication.
Diana F. Tomback

Study Area..............................................3
Ripening Phenology...............................6
Handling Times...................................7
Limber Pine Seed Harvest.........................7
Ponderosa Pine Seed Density......................8
Caloric Return...................................9
Data Analysis....................................9
Ripening Phenology of Limber and Ponderosa Pine
Limber Pine Seed Harvest....................... 13
Caching Limber Pine Seeds.......................16
Ponderosa Pine Seed Harvest.....................16

Caching Ponderosa Pine Seeds....................20
Caloric Return..................................21
LITERATURE CITED...................................................42

1. Location of study areas,.........................................5
2. Limber pine ripening phenology at Beaver Ponds....................11
3. Limber pine ripening phenology at Camp Francis....................11
4. Ponderosa pine ripening phenology at Beaver Ponds................12
5. Ponderosa pine ripening phenology at St Vrain Canyon.............12
6. Nutcracker preference for limber pine cone.......................14
7. Ripening phenology and handling times...........................19
8. Handling times and rate of caloric return.......................22
9. Rate of caloric return through time.............................23

1. Genotypes at the PER and FE loci for each trunk of sampled
ponderosa and lodgepole pine tree clumps..........................39
2. Genotypes at the PGM locus for each trunk of sampled bristlecone
pine clumps.......................................................40

I would like to thank all those who assisted me in this research project. I
especially want to thank Diana Tomback, my major advisor, for her continual
guidance and encouragement I have learned a lot thanks to her. I wish to thank
my other committee members as well: Emily Hartman, for instructing me in
botany and for her kind inspiration, and Chuck Preston, for his instructive feedback
and for being on my committee. I also thank Mary Powell for her help in the field,
and in teaching me the techniques of electrophoresis. I am especially grateful to
Ron Espinoza for all his invaluable help in the lab. Thanks also to Karen Baud for
her help in the field, in the lab, and for her friendly support. Finally, I thank Dave
Stephens of Rocky Mountain National Park for allowing me the opportunity to
conduct some of my research in the park.

The Clark's Nutcracker (Nucifraga columbiana) interacts mutualistically
with several western pines, including whitebark (Pinus albicaulis) (Toraback 1978,
1981,1982, Lanner 1980, Hutchins and Lanner 1982), limber (Pinus flexilis)
(Lanner 1980, Lanner and Vander Wall 1980, Tomback and Kramer 1980),
Colorado pinyon pine (Pinus edulis) (Vander Wall and Baida 1977), and single-
leaf pinyon pine (Pinus monophylla) (Tomback 1978, Vander Wall 1988). From
this interaction, the nutcracker gains a nutritious food source, i.e. pine seeds, which
it stores in caches for future use, and the pines benefit when unretrieved caches
Nutcrackers are well-adapted for seed transport. They carry seeds in a
sublingual pouch estimated to hold between 35 to 150 whitebark pine seeds
(Tomback 1978). While some caching occurs in the vicinity of the seed harvest,
nutcrackers have been observed flying distances from 3 to 22 km to store seeds
(Vander Wall and Baida 1977, Tomback 1978). A single cache may consist of 1 -

15 whitebark pine seeds (Tomback 1978) and be buried at a depth of 2-3 cm
(Lanner and Vander Wall 1980).
The nutcracker's pine mutualists share certain characteristics that are
adaptive for this interaction. The cones are oriented perpendicular to vertically
growing branches which makes cones easily visible and accessible (Tomback
1978, Smith and Baida 1979). Seeds are relatively large and wingless, and thus
cannot be dispersed by wind, with thin seed coats (Vander Wall and Baida 1977,
Tomback 1978,, Smith and Baida 1979). This enables the nutcrackers to harvest
seeds efficiently, providing the nutcracker the greatest return on the amount of
energy expended (Tomback 1978). Since cone crops of these pines vary in size
from year to year, the nutcracker appears to rely on other conifers to supplement its
supply of seeds.
Nutcrackers also harvest and cache seeds of wind-dispersed conifers.
In the Sierra Nevadas, Tomback (1978) observed nutcrackers foraging on Jeffrey
pine seeds (Pinus jeffreyi) when the supply of whitebark pine seeds was virtually
exhausted. Vander Wall and Baida (1977) report that in the San Francisco
Mountains nutcrackers use ponderosa (Pinus ponderosa) and bristlecone pine
(Pinus aristata) in addition to pinyon pine. A study by Giuntoli and Mewaldt
(1978) found Douglas fir (Pseudotsuga menziesii) and ponderosa pine seeds in the
stomachs of nutcrackers. Such wind dispersed pines are characterized by small
seeds with wings, spines on their cone scales, cones positioned such that visibility
and accessibility are limited, and cones that open and release seeds in a short period
of time (Smith and Baida 1979).

Tomback and Kramer (1980) propose that nutcrackers forage on seeds
which can be found and handled in the shortest period of time. The nutcracker's
choice of seeds probably depends on foraging efficiency, which would relate in part
to the ripening phenology of the pines and seed handling times (Tomback and
Linhart 1990). I studied the foraging behavior of Clark's Nutcrackers in the
Colorado Front Range in fall, 1991, as they switched from limber pine seed, a bird-
dispersed pine, to ponderosa pine seed, a wind-dispersed pine. If nutcrackers are
"optimal foragers" (Tomback and Kramer 1980), then I predict that the timing of
the switch between limber and ponderosa pine should occur when ponderosa pine
seeds, though smaller in size, provide a greater energy return than limber for the
energy expended in foraging.
In this study I show that nutcrackers harvest and store ponderosa pine
seeds in the Colorado Front Range, but do so only after the supply of limber pine
seeds is depleted. I tested the hypothesis that nutcrackers are sensitive to changes in
available energy, by correlating handling times with cone ripening phenology and
rate of caloric intake by nutcrackers. I show that nutcrackers are efficient foragers
for limber and ponderosa pine seeds, but the timing of the switch is not
energetically optimal.
Study areas
I observed nutcrackers foraging on limber pine seeds from 20 August
to 27 October, 1991, in two locations in the Front Range of the Colorado Rocky

Mountains: Camp Francis in Roosevelt National Forest and Beaver Ponds in
Rocky Mountain National Park (Fig. 1). My study area at Camp Francis is a north-
facing slope at about 2780 m elevation with a mixed stand of limber pine and aspen
(Populus tremuloides). The Beaver Ponds study area is the south face of a ridge
running east-west, with the top of the ridge leveling off at about 2810 m elevation.
In addition to limber pine, there are ponderosa and lodgepole pine (P. contorta),
aspen, and Englemann spruce (Picea engelmannii) on the ridge and at the bottom
of the ridge by the beaver ponds.
I observed nutcrackers harvesting ponderosa pine seeds on 27 October
to 7 December at two sites: St. Vrain Canyon in Roosevelt National Forest, and
Deer Ridge in Rocky Mountain National Park (Fig. 1). The St. Vrain Canyon
study area is a south-facing slope at about 2450 m elevation. It supports a dense,
mixed forest mostly of ponderosa pine, with some Douglas fir, limber pine, and
Rocky Mountain juniper (Juniperus scopulorum). The study area at Deer Ridge is
south facing at about 2680 m elevation. A few Douglas fir occur, but the forest is
comprised mainly of ponderosa pines which are larger and less densely distributed
than at the other site. There was an abundant crop of both limber pine cones and
ponderosa pine cones throughout the Front Range in 1991.

Figure 1. Location of study areas

Ripening phenology
To examine the ripening sequence of limber and ponderosa pine, I
tagged 17 limber pine trees at Camp Francis on 2 September and 25 at Beaver
Ponds on 11 September, using surveyor's tape. I tagged 31 ponderosa pine trees at
Beaver Ponds on 22 September and 30 at St. Vrain Canyon on 2 October. Each
week I counted the number of cones per tree that I could see from one vantage
point, categorizing them as either "broken into", "closed", "almost open", or
"open", as a measure of their ripeness. The scales on "closed" cones were
completely sealed. When they just began to separate on at least some of the cone, I
classified the cones as "almost open". When all the scales had separated, I called
the cones "open". Cones broken into by nutcrackers were shredded on one or more
sides. I categorized cones weekly for each tagged tree until all the cones were open.
Over the course of this survey, many cones on the tagged trees were removed by
pine squirrels (Tamiasciurus hudsonicus), with some trees completely cleared of
cones. This was evident by the marks left on stems where cones were once
attached. If cones were still left in the tagged trees, I counted the cones as usual.
Only if nearly all, or all of the cones were removed did I eliminate the tree from the

Handling Times
As nutcrackers foraged in trees for limber pine seeds, I watched them
with binoculars and timed seed extraction using a stop watch. After a nutcracker
extracted a seed, it would either throw its head back slightly as it put the seed in its
sublingual pouch or eat it. I only timed nutcrackers pouching seeds. I started the
stop watch the moment a nutcracker jabbed its bill into the cone. I stopped timing
after it placed the last seed into its pouch and stopped harvesting seeds. I recorded
handling times at Beaver Ponds on 11,16, and 22 September (n = 19,24, and 12
extraction times, respectively), and at Camp Francis on 14 and 21 September (n =
10 and 40 extraction times, respectively).
When nutcrackers foraged for ponderosa pine seeds on the ground, I
timed them for 30 second runs as they found and pouched seeds. I observed this
behavior at St. Vrain Canyon on 6 and 7 November (combined n = 64 runs), and at
Deer Ridge on 13 and 14 November (combined n = 53 runs). I did not observe
nutcrackers collecting ponderosa pine seeds at Beaver Ponds where I recorded the
ripening sequence.
In all cases I tried to avoid collecting data from the same nutcracker
more than once. However, because the nutcrackers were not banded, it is possible
that not all rates are independent
Limber Pine Seed Harvest
On 11 and 16 September at Beaver Ponds, and 14,15, and 21
September at Camp Francis, as I collected foraging rate data for nutcrackers
harvesting limber pine seed, I also noted if the cones chosen by nutcrackers were

"closed", "almost open", or "open" (See method for ripening survey for
explanation of these terms). At Beaver Ponds, I collected 40 cone preference
observations for both dates combined, and at Camp Francis I collected 29
observations for all three dates combined.
Later in fall, as nutcrackers found fewer limber pine seeds left in the
cones, it became difficult to time them with a stop watch since they visited each
cone only briefly. Instead, I counted the number of seeds they pouched from each
cone visited. These data were collected at Beaver Ponds on 13,19, and 27 October
(n = 21,18, and 34 seeds, respectively). To determine the average number of
limber pine seeds left in each cone, I sampled 37 cones from 15 trees at Beaver
Ponds on 19 October and 33 cones from 15 trees at Camp Francis on 20 October.
On 13,16, and 17 October at Beaver Ponds and 12 and 20 October at
Camp Francis, for one to two hours each day I watched nutcrackers steal limber
pine cones from pine squirrel middens. They appeared to choose cones that were
partially hidden, rather than cones that lay on the surface. To see if there was a
difference in the number of seeds in these cones, I counted seeds in 31 cones from
the midden surface and seeds in 30 cones partially hidden by litter, other cones, or
common juniper (Juniperus communis). I did this at both study sites.
Ponderosa Pine Seed Density
To estimate density of shed ponderosa pine seeds on the ground, I
counted the number of seeds in one meter square plots. When nutcrackers pouched
the seeds, they removed and left behind the wings. Consequently, I found many
wings without seeds. Some seeds were separate from their wings in their original

position, and some of the seeds may have fallen below in the litter, or the
kinnikinnik (Arctostaphylos uva-ursi), leaving the wings behind. I counted wings
with seeds and wings without seeds separately and combined these counts for an
estimate of the total number of seeds per plot. I counted seeds and wings for a total
of 12 plots at St. Vrain Canyon on 7 and 9 November and 5 plots at Deer Ridge on
11 and 13 November. At both sites the snow cover was patchy. Since I did not
see nutcrackers forage in the snow, I did not sample these areas.
Caloric Return
I collected 2 open cones from 15 trees for a total of 30 cones for each
sample of limber and ponderosa pine cones. From each sample I randomly chose
300 seeds and weighed each one on an electric balance (Fischer Scientific XT-
400D) accurate to 0.00 lg. I then removed the seed coats by striking the seeds
gently with a pestle, and measured the thickness of the seed coats at the flattest part
with calipers (for limber, n = 179, and ponderosa, n = 300). I also weighed the
endosperm for limber (n = 298) and ponderosa pine seeds (n = 300). I used bomb
calorimetry (Shoemaker, et al. 1981) to determine the caloric content of both pine
seeds, running three samples for each.
Data analysis
After finding that data for handling times and ponderosa pine seed plot
counts were not normally distributed, I used non-parametric statistical tests. I used
the Kruskal-Wallis test to compare differences between means of handling time
data and seed counts in the plots. I tested the differences between distributions of

cone preference using the chi-square test for independence. I used Pearson
correlation analysis, a parametric test, to assess the relationship between the number
of limber pine cones per tree on the first survey date and the number of cones
broken into by nutcrackers. To compare the data between pine species (seed and
endosperm weight, seed coat thickness, and the arcsine transformed ratio of
endosperm weight to seed weight), I first used Hartley's F-max test to compare
variance between samples and then the appropriate two sample analysis for
independent measures t-test.
Ripening Phenology of Limber and Ponderosa Pine Cones
Nutcrackers opened green limber pine cones to feed on unripe seeds in
August. The opened scales on the cones appeared brown and shredded in contrast
to the green, undamaged cones. By 11 September at Beaver Ponds and 15
September at Camp Francis, the cones began opening. Shortly after 22 September,
all the cones at both sites were open (Fig. 2 & 3).
Between 5 and 13 October, the ponderosa pine cones at Beaver Ponds
began opening. At St. Vrain Canyon, the cones began opening by 6 October. No
more than 5 ponderosa pine cones at Beaver Ponds were damaged by nutcrackers
in contrast to about 205 at St. Vrain Canyon. By 19 and 20 October, respectively,
cones at both sites were all open (Fig. 4 & 5).

9-11 9-16 9-22
Figure 2. Limber pine ripening phenology at Beaver Ponds.
E3 Broken Into M Closed ES9 Almost Open III Open
100 t
9-2 9-15 9-21
Figure 3. Limber pine ripening phenology at Camp Francis.
EsJ Broken Into IM Closed ES Almost Open HU Open

9-22 9-28
Figure 4. Ponderosa pine ripening phenology at Beaver Ponds.
B = cones broken into by nutcrackers.
ESS Broken Into M Closed Almost Open K3 Open
10-2 10-6 10-12 10-20
Figure 5. Ponderosa pine ripening phenology at St. Vrain Canyon.
Broken Into Hi Closed ESI Almost Open HI Open

The ripening sequence at each of the four sites and the percentage of
unripe cones damaged by nutcrackers are shown in Figures 2-5. Although some of
the damaged cones were not entirely shredded and depleted of seeds, the degree of
ripening was not noted for these cones. The undamaged portions of the cones did,
however, open as the cones ripened.
The number of unripe limber pine cones damaged by nutcrackers and
cone crop size per tree were not significantly correlated, (r = 0.073; P = 0.730; df =
23 at Beaver Ponds, or r = 0.018; P = 0.944; df = 15 at Camp Francis).
Limber Pine Seed Harvest
Before limber pine cones were open, nutcrackers broke into unripe
cones and extracted seeds. A nutcracker usually stood on the cone it worked on, or
on an adjacent branch, and jabbed its bill in between the scales. If the cone was still
unripe, it may have taken two or more jabs before being able to access the seed. I
observed this on 20 and 31 August at Camp Francis. Although I did not see any
nutcrackers remove unripe cones from branches, I found many shredded cones on
the ground, characteristic of nutcracker harvesting activity. I also observed
nutcrackers breaking into cones that had already been removed from the branches.
They propped cones on the stump of a tree or in a branch fork, which they would
use as an anvil, and jabbed their bills between the scales. Pine squirrels also
removed unripe cones but took the cones to their middens.
After cones began opening, nutcrackers harvesting seeds at Beaver
Ponds chose open limber pine cones 83% of the time, almost open cones 17%, and
closed cones 0%. At Camp Francis, they chose open cones 76% of the time,

almost open cones 21%, and closed cones 3%. Data combined from both sites
showed 80% of the cones chosen were open, 19% almost open, and 1% closed
(Fig.6). Cone preference did not differ between study areas (X^ = 0.11, df = 2, P >
Figure 6. Nutcracker preference for limber pine cones.
9 Closed Eg] Almost Open m Open
The handling times at Beaver Ponds decreased as the cones ripened
(Fig. 7). Mean handling times recorded on 11,16, and 22 September were 10.90,
5.73, and 3.33 seconds per seed, respectively, and were significantly different (H =
34.11, df = 2, P < 0.001, Kruskal-Wallis test). There was also a significant

difference between the means on 11 and 16 September (H = 21.19, df = 1, P <
0.001), and on 16 and 22 September (H = 11.68, df = 1, P = 0.001). At Camp
Francis there was also a decrease in handling times (mean = 5.75 and 3.87 seconds
per seed on 14 and 21 September), and the difference was significant (H = 4.61, df
= 1, p = 0.032). A comparison of the last rates recorded at Beaver Ponds on 22
September, and Camp Francis on 21 September showed no significant difference
(H= 1.74, df=l,P =0.188).
By mid-October, limber pine seeds were clearly depleted from the open
cones. I observed 5 or fewer nutcrackers per day foraging in the study areas.
Twenty-one days after the last handling times were recorded at Beaver Ponds,
nutcrackers harvested an average of 1.14 limber pine seeds per cone (SD = 1.39).
On 19 October nutcrackers harvested a mean of 1.0 (SD = 0.91), and on 27
October they harvested a mean of 0.70 seeds per cone (SD = .09).
Based on my seed counts of randomly selected cones, the average
number of seeds in limber pine cones which were still in the trees was 5.0 seeds
per cone (SD = 5.43) on 19 October at Beaver Ponds, and 1.94 seeds per cone (SD
= 3.27) on 20 October at Camp Francis.
I observed nutcrackers robbing cones from a pine squirrel midden at
Beaver Ponds which had ponderosa and limber pine cones. I counted 9 limber pine
cones stolen by nutcrackers on 13 October, one on 16 October, and one on 27
October. Of the 11 cone thefts, 6 were cones partially hidden and 5 were cones on
the surface. To find the partially hidden cones, the nutcrackers poked around in the
ground first, pushing away the dirt with their bills before taking a cone. I was
unable to determine how many different nutcrackers visited the midden, since only

one came at a time and flew out of sight with a cone. The mean number of seeds in
cones partially hidden was 37.9 (SD = 29.5), and mean number of seeds in cones
on the surface of the midden was 8.4 (SD = 6.7).
At Camp Francis I found nutcrackers stealing cones from a midden on
12 October. I watched 7 thefts and 4 attempts thwarted by a protective squirrel.
On 20 October I saw 6 thefts (3 were cones partially hidden) and 2 attempts. The
mean number of seeds in cones partially hidden was 30.8 (SD = 24.0), and 3.4
(SD = 4.3) for seeds in cones on the surface of the midden. It is clear that
nutcrackers were obtaining greater numbers of seeds by stealing cones partly buried
than from the surface cones.
Caching Limber Pine Seeds
At Beaver Ponds, most of the nutcrackers with limber pine seeds in
their pouches flew off in the direction of Fall River Road (NW of Beaver Ponds) to
make caches. On 22 September, I saw one nutcracker make a cache in the tundra
along Fall River Road, at about 3485 m elevation (10 km from Beaver Ponds). I
also observed nutcrackers cache at Beaver Ponds at the bottom of the ridge beside a
stand of Englemann spruce. I saw 4 caches made on 13 October and 7 on 19
October. All but one cache, which was made in the top of a log, were made in the
forest litter about 1 meter from one of the spruces.
Ponderosa Pine Seed Harvest
Few unripe ponderosa pine cones were opened by nutcrackers.
Although I did not see nutcrackers at either study area breaking into the unripe

cones, at Beaver Ponds I found three trees with one or two shredded cones in each
(22 September). At St Vrain Canyon, among the six trees with shredded unripe
cones, two trees had nearly all of their cones broken into, but the majority of trees
had no damaged cones (2 October).
By 6 and 13 November, nutcrackers foraged on ponderosa pine seeds
that had been blown out of open cones at St Vrain Canyon and Deer Ridge,
respectively. Since ponderosa pine seeds have wings, a nutcracker had to remove
the wing before pouching a seed. It picked up a seed from the ground and then
with a sideways movement wiped the wing off the seed. Other seeds were loosely
attached to the wings or were already separated. When the nutcracker's pouch was
full, it flew off to cache the seeds. I saw a nutcracker take seeds directly from a
ponderosa pine cone at Beaver Ponds on 27 October and at Deer Ridge on 14
November. Each time the nutcracker carried the cone to a tree where it extracted
the seeds. I estimated that about 30 nutcrackers at St. Vrain Canyon were on the
ground foraging for ponderosa pine seeds on 6 November and at least 15 on 7
November. At Deer Ridge there were about 30 nutcrackers foraging together on
the ground on 13 November and between 30 to 40 on 14 November. When the
nutcrackers moved to a new location, across the road for example, they moved
The handling times for ponderosa pine seeds on the ground at Saint
Vrain Canyon on 6 and 7 November (mean = 3.1 seconds per seed) and Deer
Ridge on 13 and 14 November (mean = 3.0 seconds per seed) were not
significantly different (H = 0.42, df = 1, P = 0.515) (Fig. 7). A comparison of the
handling times on the last date nutcrackers harvested limber pine seeds from the

cones (22 September, Beaver Ponds) and the first timings recorded for ponderosa
pine seeds (6 November, Saint Vrain Canyon) showed no significant difference (H
= 0.07, df = 1,P = 0.797).
The estimated density of ponderosa pine seeds on the ground was as
follows: At St Vrain Canyon, the mean number of seeds with wings was 59.0 per
m^ (SD = 15.5), wings without seeds was 50.8 per m2 (SD = 29.6), and total
seeds based on the number of seeds with wings and wings alone was 109.8 per m2
(SD = 37.5). At Deer Ridge, the mean number of seeds with wings was 67.2 per
m^ (SD = 9.2), wings without seeds was 57.8 per m2 (SD = 14.0), and total seeds
was 125.0 per m2 (SD = 18.8). There was no significant difference among the
means of all three counts between the plots in St. Vrain Canyon and Deer Ridge (H
= 0.82, df = 1, P = 0.365 for seeds with wings; H = 0.80, df = 1, P = 0.370 for
wings without seeds; and H = 1.23, df = 1, P = 0.268 for total seeds).

1 00 T
T 15
Figure 7. Ripening phenology and handling times(seconds per seed).
Limber pine cones at Beaver Pond. Limber pine cones at Camp
Francis. Ponderosa pine cones at Beaver Pond. Ponderosa pine
cones at St. Vrain Canyon. Handling times for limber pine seed: Beaver
Pond, O Camp Francis; for ponderosa pine seed: St. Vrain Canyon,
Deer Ridge.

Caching Ponderosa Pines Seeds
On 28 September, before nutcrackers harvested ponderosa pine seeds at
St Vrain Canyon, I watched 2 nutcrackers make caches of undetermined seeds in a
small grassy area. On 6 and 7 November during the time that nutcrackers foraged
for ponderosa pine seeds on the ground, I saw 7 caches made, all about one meter
from the trunk of a tree. Most of the nutcrackers flew off with full pouches. From
Deer Ridge, nutcrackers with full pouches flew in the direction of Fall River Road
or Rainbow Curve (3 km west of Beaver Ponds at 3240 m elevation). On 13
November I watched a nutcracker cache seeds in the pine litter beside a tree. I
uncovered the cache and counted 31 ponderosa pine seeds. On 14 November I
watched 2 caches being made and what appeared to be 4 recoveries of caches.
These may have been temporary caches. Before caching seeds, nutcrackers pushed
around the litter. When it appeared that they found an acceptable site, they threw
their heads back in the same manner they pouch seeds, then pushed their bills into
the ground. All of these caches were made more than a meter from the trunk of a
On subsequent visits to St. Vrain Canyon between 7 November and 7
December, I saw nutcrackers fly overhead with full pouches but was unable to find
where they were foraging. On 7 December, I counted 12 nutcrackers overhead
with full pouches, and one harvesting ponderosa pine seeds on the ground. Thus, it
appeared that as long as seeds were available, nutcrackers continued to cache them.

Caloric Return
The average seed weight for limber and ponderosa pine was 0.085 g
(SD = 0.022) and 0.042 g (SD = 0.007), respectively. The average weights of the
endosperm were 0.042 g (SD = 0.012), and 0.020 g (SD = 0.004), respectively.
The averages for seed coat thickness were 0.323 mm (SD = 0.051), and 0.291 mm
(SD = 0.042), respectively. The differences in seed weight, endosperm weight, and
seed coat thickness were all significant (t = -32.9, df = 598, P < 0.001; t = -30.7, df
= 596, P < 0.001; and t = 7.1, df = 477, P < 0.001, respectively).
The averages of the ratios of endosperm weight to seed weight for
limber and ponderosa pine seeds were 0.514 (SD = 0.056) and 0.504 (SD =
0.070), respectively. These means did not differ significantly (t = 1.8, df = 596, P
= 0.069).
The mean caloric content, in calories per gram, for limber pine seed was
7197.3 c/g (SD = 43.0), and for ponderosa pine seed was 7046.8 c/g (SD = 85.2).
The average caloric content per seed of limber pine was 302.3 calories, with a range
of 93.6 to 575.8. Ponderosa pine seeds had an average of 140.9 calories per seed,
and a range of 63.4 to 218.45 calories. Thus, a seed of limber pine is more than
twice as rewarding energetically as a ponderosa pine seed. On 21 and 22
September at Beaver Ponds and Camp Francis, respectively (the last handling times
recorded for limber pine seed), it took nutcrackers more time to harvest one seed of
limber pine from cones than ponderosa pine seeds from the ground (6 and 13
November at St. Vrain Canyon and Deer Ridge, respectively) but less time to
harvest 100 calories of limber pine seed (Fig. 8). The change in caloric return
through time for both limber and ponderosa pine seeds is shown in Figure 9,

including an estimation of the increase in time nutcrackers took to harvest 100
calories of limber pine seeds after the last handling times were recorded.
Limber pine Ponderosa pine
Beaver Camp St. Vrain Deer
Pond Francis Canyon Ridge
Figure 8. Handling times and caloric return.
The number of seconds it took nutcrackers to harvest one pine
seed, or 100 calories of limber and ponderosa pine seed are
shown above. The handling time data shown here for Beaver
Ponds and Camp Francis are from the last dates that handling
times were recorded (September 21 and 22). The handling
time dates for St. Vrain Canyon and Deer Ridge are 6 & 7, 13
& 14 November, respectively.

.5 T
4 --
3 --
2 --
1 --
HH +
x- -
1--)i----1---1--1--1 1----1--1|---1---1--1t-
9/15 9/30 10/15 10/30 11/14
Figure 9. Caloric return through time.
The number of seconds nutcrackers take to harvest one hundred
calories of pine seed. X Limber pine seeds at Beaver Ponds,
x Limber pine seeds at Camp Francis. + Ponderosa pine seeds
at Deer Ridge. 1- Ponderosa pine seeds at St. Vrain Canyon.
The arrow represents the hypothetical increase in the time
nutcrackers required to harvest 100calories of limber pine
seeds after 21 & 22 September.

Nutcrackers began to forage on unripe limber pine seeds in August, or
possibly earlier, and on ripe seeds from September until late October. Tomback
and Taylor (1987) report that foraging for unripe limber pine seeds in Rocky
Mountain Park began the last week of July. In September nutcrackers were
harvesting unripe seeds from closed cones and became more efficient as the cones
opened. Early in October, ponderosa pine cones began opening, but nutcrackers
continued foraging for limber pine seeds until early November, which is when they
switched to ponderosa pine seeds. They harvested the ponderosa pine seeds from
the ground and flew off with full pouches to make caches. In early December
nutcrackers were still finding seeds and storing them.
Nutcrackers probably rely on limber pine seeds as their primary seed
source for several reasons. Relative to ponderosa pine seeds, limber pine seeds are
larger and have a higher caloric value. They ripen earlier in the season, which
means nutcrackers have more time to make caches before the first snow cover.
The trees occur in the subalpine, which is where nutcrackers appear to make many
if not most limber pine seed caches. My data support the hypothesis that foraging
nutcrackers are sensitive to changes in available energy. As the limber pine cones
began to open, nutcrackers stopped removing them from the trees to extract the
seeds. Vander Wall (1988) also observed this. Since the scales were separating,
they could access the seeds just as easily without the use of an "anvil". When

limber pine cones were in different stages of ripening, nutcrackers chose open
cones when closed ones were also present; this was similarly observed by
Tomback and Kramer (1980), Benkman et al. (1984), and Vander Wall (1988).
I predicted that nutcrackers would switch to ponderosa pine seeds when
it became energetically advantageous for them to do so. The timing of the switch
would occur when the caloric return for ponderosa pine seed was greater than it
was for limber pine seed. My data generally support this but do not suggest that
the timing of switching was optimal. My observations suggest that, at least in
1991, nutcrackers continued foraging for limber pine seeds as their rate of energetic
gain declined below what was available from harvesting ponderosa pine seeds from
the ground. They probably continued for as long as they did because they were
able to find enough to meet their energetic requirements, but not because the
energetic rewards exceeded those of ponderosa pine seed. When they switched
seeds, they were forced to find an alternate food source because the limber pine
seed supply was nearly exhausted. However, the cone crops of 1991 for both
limber and ponderosa pines were extremely abundant, and the timing of switching
may differ under other conditions. For example, if the limber pine cone crop were
smaller, nutcrackers may switch earlier to ponderosa pine seeds and forage in the
cones rather than wait for seeds to be dispersed. It is possible that nutcrackers
would be less efficient foraging in cones than on the ground. In this study, since
ponderosa pine cones at both study sites were almost all open by 13 October, by
switching earlier nutcrackers may still have been able to forage as efficiently on the

As limber pine cones ripened and the scales began separating,
nutcrackers were able to extract and pouch the seeds more quickly. Their fastest
handling times for limber pine seed were comparable to their ground foraging
times for ponderosa pine seed, which means that up until 21 September, limber
pine seed was more energetically rewarding because the number of calories per
gram of dry seed mass is greater. Between 22 September and 6 November,
nutcrackers were not finding many seeds left in the cones and began stealing cones
from squirrel middens. The energetic return for their efforts was decreasing. Also,
during this time I saw few nutcrackers fly off to store seeds, suggesting a great
decline in supply. If nutcrackers foraged in an optimal manner, they should have
switched to ponderosa pine seeds sometime during this month long interval, since
ponderosa pine seeds were not only available and abundant, but limber pine seeds
were becoming relatively scarce. When they did finally switch, they again began
making caches.
Nutcracker reliance on two different seed sources is also described by
Tomback (1978) and Vander Wall (1988). In the Sierra Nevada of California, the
nutcracker's primary seed source is whitebark pine (Tomback 1978) which, like
limber pine, occurs in the subalpine. When most of the whitebark pine seeds were
harvested, nutcrackers switched to the seeds of Jeffrey pine, which are wind-
dispersed, ripen later in the season, and occur at a lower elevation, like ponderosa
pine in this study. Extraction rates show that nutcrackers were able to harvest
whitebark pine seeds more efficiently than the smaller Jeffrey pine seeds. In the
Raft River Mountains of Utah, nutcrackers foraged for limber pine seeds and
single-leaf pinyon pine seeds (Vander Wall 1988). Although pinyon pine seeds are

larger and contain more calories, they have only 70.9% as much energy per gram
as limber pine seeds do. Since handling times were faster for limber pine cones,
and the cones ripened between 7-10 days earlier than pinyon pine cones, limber
pine seeds are the nutcracker's preferred seed source there as well.
Seed harvesting efficiency may in part depend on the synchrony of cone
ripening. Benkman et al. (1984) discuss the ripening pattern expected for wind-
dispersed and animal dispersed pines. If cones ripen synchronously both within
trees and among trees, predators and dispersers should be saturated. We would
expect this to occur in trees that rely on wind for dispersal A pattern in which
cones ripen synchronously within trees and asynchronously among trees would
benefit seed dispersers. I expected that limber pine would fit this second pattern as
noted by Benkman, et al (1984) and Tomback and Kramer (1980), and that
ponderosa pine would conform to the former (Baida 1981). My data, however,
show no appreciable difference in ripening patterns between limber and ponderosa
pine cones. Both pines ripened asynchronously within trees and synchronously
among trees.
Before limber pine cones began opening (early in September),
nutcrackers harvested unripe seeds from the green cones, even though the energetic
return was small relative to the work required to extract the seeds. According to
Vander Wall (1988), the caloric value of seeds increases as cones ripen. In his
study, nutcrackers began harvesting unripe limber pine seeds a month before the
cones were ripe. He speculates that although the energetic return may not exceed
the energetic output, the advantage in harvesting unripe seeds may be that
nutcrackers can assess the size and quality of the cone crop, so they can respond to

possible seed shortages before they occur. Tomback (1978) reported that
nutcrackers typically harvest unripe seeds in whitebark pine and often feed their
juveniles with these seeds. In this case it appears that there is no energy deficit.
Furthermore, this explanation does not explain why nutcrackers began foraging on
whitebark pine seeds as early as July in the Sierra Nevada. In this study
nutcrackers foraged extensively on unripe limber pine cones. While they may have
been testing the degree of ripeness as suggested by Tomback (1978) or assessing
the cone crop (Vander Wall 1988), they were probably also fulfilling their energetic
Few ponderosa pine cones showed damage from nutcrackers,
suggesting that there was little interest in assessing the cone crop while the cones
were unripe. This information may not have been critical in 1991 since the limber
pine cone crop was so great. Data from St. Vrain Canyon show that nutcrackers
did forage on unripe cones, but nearly all of the cones were on just two trees.
Almost all of the cones in these two trees were broken into by nutcrackers. Doing
this would tell nutcrackers less about the condition of the stand in general than if
they had sampled cones from a number of trees, as they did for limber pine; but, if
the cones on those two trees were more ripe relative to the rest of the stand,
nutcrackers may simply have been after the seeds. Again, this suggests that
nutcrackers forage in unripe cones to obtain food and not just to test ripeness of
cones or survey cone crop size. In this study area, there were few limber pine trees,
which may explain why nutcrackers used ponderosa pine seeds. One interesting
situation occurred as a result of this activity. Red crossbills (Loxia curvirostra),
which are predators of ponderosa pine seed and normally have to wait for the cones

to open before they are able to access the seeds, found the remains of seeds left in
the cones shredded by nutcrackers. I observed a flock of crossbills foraging in one
of these trees visited by nutcrackers weeks before they could forage on any of the
other cones.
Christensen et al. (1991) show that nutcrackers foraging on Colorado
pinyon pine prefer trees with more cones on them. They demonstrate this in the
field and in the laboratory. A tree with more cones is likely to attract more
dispersers, and a nutcracker may forage more efficiently in a tree with more cones
since less time and energy may be spent handling the seeds (Christensen et al.
1991). Vander Wall and Baida (1977) also report that nutcrackers increase
efficiency by choosing the most productive trees. My results, however, show no
evidence of discrimination based on cone crop size by nutcrackers foraging on
limber pine cones at either site. These data were collected during an exceptionally
good cone crop, which may possibly explain why I found no correlation, since the
need for nutcrackers to be discriminating was less.
Nutcrackers may have been discriminating among the cones they stole
from squirrel middens. Although the data in this study are insufficient to conclude
this, the fact that there was a significant difference in seed quantity between cones
hidden and exposed makes this an intriguing possibility. If more data were
collected which showed that a significant number of nutcrackers foraged in this
manner, this would be additional evidence that nutcrackers are optimal foragers.
One of the difficulties in acquiring these data is in determining how many different
nutcrackers visit the middens. Tomback (personal communication) observed

nutcrackers north of Yellowstone National Park stealing whitebark pine cones
from middens. They appeared to be discriminating in a similar manner.
The observations I made of nutcrackers foraging on the ground for
ponderosa pine seeds instead of in the trees suggest efficient foraging for a couple
of reasons. Ponderosa pine has the characteristics of wind-dispersed pines, which
serve to deter direct cone access by predators and potential dispersers. The cones
have spines and are oriented in such a way that facilitates wind-dispersal but
hampers accessibility and visibility from above. Foraging on the ground
eliminates these hindrances.
In addition, according to Vander Wall and Baida (1977), 63.6% of the
wings in a ponderosa pine cone are seedless. Predators can spend considerable
time and energy pulling out wings and examining them only to find that no seeds
are attached. Nutcrackers foraging on the ground may be more efficient because
they can see the seeds they harvest, and because seeds on the ground separate easily
from the wings, more easily than I expect they would when still in the cones.
There are, however, negative consequences of waiting until the seeds
have already blown from the trees. Fewer seeds will be available since seed
predators such as crossbills will have eaten many of the seeds in the cones. There
may also be more competition for seeds on the ground from rodents (Vander Wall
1992a and 1992b). Another consequence is that nutcrackers put themselves at
greater risk of predation. I counted up to 40 nutcrackers on the ground in a given
area, many more than any I observed when they foraged in the trees. This is most
likely a safety response to their increased vulnerability.

Vander Wall and Baida (1977) observed nutcrackers harvesting
ponderosa pine seeds but none were seen caching the seeds. They speculate that
these seeds may be stored in the event of a poor pinyon and limber pine cone crop,
which in the San Francisco Mountains are the nutcrackers primary seed sources.
Since ponderosa pine cones in Colorado's Front Range ripen later in the season than
limber pine cones, and cone crops for limber pine are not apt to get much better
than the 1991 crop, it seems unlikely that nutcrackers would not harvest and cache
ponderosa pine seeds if they were available in other years. Tomback (1978)
observed nutcrackers routinely extracting Jeffrey pine seeds directly from the cones
and caching the seeds. In other years, nutcrackers in the Front Range may start
earlier and harvest ponderosa pine seeds directly from the cones as well. Although
foraging on the ground may allow nutcrackers to harvest seeds more efficiently, the
difference in the productivity of limber pine is probably what determines the timing
of the ponderosa pine seed harvest.
My observations in 1991 suggest that nutcrackers acted as secondary
seed dispersers for ponderosa pine, similar to Vander Wall's (1992a) observations
of rodents. He shows that rodents may actually play a significant role in the
regeneration of Jeffrey pine, a close relative of ponderosa pine. After the seeds are
blown out of the cones, rodents harvest the seeds and make caches. Since
nutcrackers are known to forage on some winged conifer seeds (Douglas fir,
ponderosa, Jeffrey, and bristlecone pine), they and rodents may play a role in the
regeneration of these trees as well.

Recently, multi-genet tree clusters were shown to occur commonly in
several species of bird-dispersed pines, hi this growth form, two or more trunks of
different genotypes are either contiguous or fused; the fusion may be just at the
base or continue partway up the bole. Tomback and Linhart (1990) suggest that
bird-dispersed pines are tolerant of crowding because they are slow-growing and
To date, the bird-dispersed species that have been examined and found
to have multi-genet tree clusters include: whitebark pine (Pinus albicaulis) (Linhart
and Tomback 1985, Fumier et al. 1987); Swiss stone pine (Pinus cernbra)
(Tomback et al. 1993); and limber pine (Pinus flexilis) (Linhart and Tomback 1985,
Schuster and Mitton 1991, Carsey and Tomback 1994). Tomback et al. (1993)
estimated that tree clusters occurred at roughly 21% of the tree sites in a Swiss
stone pine stand in Upper Engadine Valley of Switzerland. Similarly, Carsey and
Tomback (1994) examined limber pine in four populations in the Front Range of
the Colorado Rocky Mountains and estimated that the percentage of tree clusters

varied from 19% to 23%. Working at the eastern periphery of the limber pine
range, however, Schuster and Mitton (1991) found tree clusters at only 5% of the
tree sites in their Pawnee Grassland stand. Growth form distribution and genetic
relationships within tree clusters of bird-dispersed pines have recently been
reviewed by Tomback and Schuster (1994).
The aforementioned pines are dispersed by nutcrackers (Nucifraga),
seed-storing corvids with a Holarctic distribution. Clark's Nutcrackers (Nucifraga
columbiana) of western North America bury pine seeds in shallow, subterranean
caches containing from 1 to 15 or more seeds, with a mean of 3 or 4 seeds per
cache (e.g., Tomback 1978,1982, Hutchins and Lanner 1982). Often, more than
one seed in a cache will germinate, resulting in a cluster of seedlings (Tomback
1982, Tomback and Linhart 1990). Each seedling may grow with others into a
mature tree cluster that resembles one tree with multiple stems. Pines, however,
may also produce a multi-stemmed growth form from a single genet as the result
of mechanical damage to the leader or other stressful conditions (e.g., Schuster and
Mitton 1991). Consequently, the multi-genet nature of tree clusters must be shown
by means of genetic analysis of cluster members.
Growth forms with multiple trunks also occur at low frequencies in
pines that are supposedly wind-dispersed (Lanner 1980, Tomback and Linhart
1990, Carsey and Tomback 1994). Recent studies by Vander Wall (1992a, 1992b,
1993) show that several rodent species, and particularly chipmunks (Tamias,
Sciuridae), may play an important role in the regeneration of Jeffrey pine (Pinus
jeffreyi), a "wind-dispersed" species. The Clark's Nutcracker also caches the
winged seeds of "wind-dispersed species, including Jeffrey pine (Tomback 1978),

ponderosa pine (Pirns ponderosa) (Torick, unpublished observations), Great Basin
brisllecone pine (Pinus longaeva) (Lanner 1988), and Rocky Mountain bristlecone
pine (Pinus aristata) (Baud 1993, Torick, unpublished observations). In this study
I test the simple hypothesis that tree clusters occur in several wind-dispersed pines.
These tree clusters may originate from the caches of Clark's Nutcrackers,
chipmunks, or other seed-storing animals, or possibly from the accumulation of
wind-dispersed seeds in depressions or against objects (Y. B. Linhart, personal
communication). If tree clusters do occur in wind-dispersed pines, their existence
provides some support for potential recruitment of mature trees from animal
The following growth form terminology is based on Tomback et al.
(1990) and will be used throughout this paper. A "multi-trunk tree" has more than
one trunk, but all comprise a single genet. A "tree cluster" is composed of more
than one genet as a consequence of more than one seedling from a seed cache
surviving to maturity. Tree clusters and multi-trunk trees cannot be distinguished
by morphology with certainty (Schuster and Mitton 1991, Tomback et al. 1993,
Carsey and Tomback 1994). Thus, I refer to those growth forms not genetically
analyzed collectively as "tree clumps" from Lanner (1980).
Foliage samples were taken from all trunks of 10 mature tree clumps
each of ponderosa, bristlecone, and lodgepole pine (Pinus contorta). Clumps were

sampled as encountered, without bias as to degree of fusion of trunks. The three
study areas were in the Front Range of the Colorado Rocky Mountains. I sampled
ponderosa pine from Saint Vrain Canyon, Roosevelt National Forest, at about
2,450 m elevation. There, ponderosa pine occurs on a south-facing slope in
association with Douglas fir (Pseudotsuga menziesii), Rocky Mountain juniper
(Juniperus scopulorum), and limber pine. The sampling area for bristlecone pine
was Goliath Peak, Arapaho National Forest, at 3,500 m elevation. The forest in
this area is composed primarily of bristlecone pine with scattered limber pine.
Lodgepole pine clumps were sampled from a dense stand on the north side of
Squaw Mountain, Arapaho National Forest, at about 3,150 m elevation.
The number of trunks per sampled clump ranged from 2 to 5 for a total
of 83 trunks, all species combined. Each clump was assigned an identification
number and each trunk an identification letter. Using liquid nitrogen, I ground each
sample to a coarse powder and mixed it with the grinding buffer of Mitton et al.
(1979); I stored all samples at -70C in small centrifuge tubes.
I examined the genotype of each sample by means of horizontal starch
gel protein electrophoresis, using the following allozymes: acid phosphatase
(ACP), alcohol dehydrogenase (ADH), fluorescent esterase (FE), glutamate
dehydrogenase (GDH), glutamate oxaloacetate (GOT), isocitrate dehydrogenase
(IDH), leucine aminopeptidase (LAP), malate dehydrogenase (MDH), peroxidase
(PER), 6-phosphogluconate dehydrogenase (6-PGD), phosphoglucose isomerase
(PGI), and phosphoglucomutase (PGM). Only the results for PER, FE, and PGM
were polymorphic for the species examined and thus useful for this study. For
PER I used a tris-citric acid gel buffer (pH 8.5) and a lithium hydroxide-boric acid

electrode buffer (pH 8.1); gels ran for 5 to 6 hr at 250 to 275 V (Shaw and Prasad
1970). For FE and PGM, respectively, I used the following gel and electrode
buffers cited by Schuster and Mitton (1991) and Carsey and Tomback (1994): tris-
citrate (pH 8.2) and NaOH-boric add (pH 8.6), 2 to 4 hr at 250 to 275 V; histidine
II (pH 8.0) and tris-citrate (pH 7.0), 4 to 5 hr at 170 V. References for allozyme
stains may be found in Carsey and Tomback (1994).
For polymorphic loci, gels were run at least twice at each locus for each
sample or until results were conclusively resolved. Samples were unresolvable for
four foliage samples.
The results were interpreted as follows: If two or more trunks within a
clump had genotypes differing at a minimum of one locus, the clump was
classified as a tree cluster. If the genotypes of all trunks within a clump were
identical, the clump was classified as a multi-trunk (single genet) tree. I expect the
number of tree clusters to be an underestimate, because nutcrackers often cache
sibling seeds together. Genetic relatives (half-sibs, full-sibs, or selfed individuals)
have a high probability of carrying similar genotypes (see Schuster and Mitton
1991, Carsey and Tomback 1994), which may not be distinguishable with the
number of loci used for this study. It is possible for clumps to contain trunks that
are different genets and trunks that are the same genet (Tomback et al. 1993, Carsey
and Tomback 1994). These growth forms are classified as tree clusters.

At the Saint Vrain Canyon sampling area, ponderosa pine tree clumps
occurred at about 22% of the tree sites surveyed; and, at the Goliath Peak sampling
area bristlecone pine tree clumps occurred at about 40% of the tree sites. Lodgepole
pine clumps at the Squaw Mountain sampling area occurred at about 12% of the
tree sites.
The mean number of trunks per clump for ponderosa pine was 3.2 (SD
= 1.4), for bristlecone pine 2.9 (SD = 0.7, and for lodgepole pine 2.3 (SD = 0.7).
Genetic analysis by means of protein starch gel electrophoresis indicated that tree
clusters occurred in all three species of wind-dispersed pines. Only three of the loci
used for analysis were polymorphic, and these varied with the species. The PER
and FE loci were polymorphic in ponderosa and in lodgepole pine, and PGM in
bristlecone pine.
Seven (70%) of the ponderosa pine clumps scored for PER and 8
(80%) of those scored for FE were tree clusters, i.e., consisted of more than one
genet (Table 1). For bristlecone pine, 2 clumps (20%) scored for PGM were tree
clusters (Table 2); and for lodgepole pine, 4 clumps (40%) scored for PER and 2
clumps (20%) scored for FE were tree clusters (Table 1). Thus, altogether 8 (80%)
of the ponderosa pine clumps were tree clusters, 2 (20%) of the bristlecone pine
clumps, and 4 (40%) of the lodgepole pine clumps.

Although these results are underestimates, the data imply that 20% of
the ponderosa pine tree clumps, 80% of the bristlecone pine clumps, and 60% of
the lodgepole pine clumps were multi-trunk, single genet trees.
The results of this genetic analysis indicate that tree clusters routinely
occur in ponderosa, bristlecone, and lodgepole pine. By determining that some
clumped growth forms are indeed tree clusters, I demonstrate 1) that wind-
dispersed pines may be sufficiently tolerant of crowding to occur in tree clusters
and 2) that mature trees of "wind-dispersed" pine species may be recruited from the
caches of seed-storing birds and mammals.
Based on the frequencies of tree clumps in our sampling areas, and
assuming that the percentage of clumps that I resolved as clusters is typical of these
sampling areas, I estimated that about 18% of the ponderosa pine tree sites, 8% of
the bristlecone pine tree sites, and 5% of the lodgepole pine tree sites are occupied
by tree clusters. For comparison, about 21% of Swiss stone pine tree sites were
estimated to support clusters in the Engadine Valley of Switzerland (Tomback et al.
1993), and 19 to 23% of limber pine tree sites in the Colorado Front Range (Carsey
and Tomback 1994). Thus, tree clusters may be less common in winged seed
pines than in bird-dependent pines, possibly because of lower crowding tolerance
or because a high proportion of recruited trees are from single seed caches or
dispersal by wind.

Table. 1. Genotypes at the PER and FE loci for each trunk of sampled ponderosa
and lodgepole pine tree clumps. Genotypes are designated by allozyme migration
speed on gel: 1 = fast allele, 2 = intermediate allele, 3 = slow allele.
Genotype of designated trunk
Ponderosa pine Lodgepole pine
clump A B C D E A B C D E A B C D A B C D
1 12 22 12 11 22 12 22 22
2 23 22 22 12 12 12 12 12
3 22 23 22 12 22 12 22 22 22 22 22 12 22 22
4 22 22 23 22 22 11 22 22 22 22 12 12 12 12
5 22 22 22 23 22 22 22 12 12 22 22 22 22
6 23 22 22 12 12 12 22 22
7 22 22 22 12 11 12 12 12 23 23
8 12 12 22 22 22 22 22 22
9 22 22 12 12 12 22 12 22 12 22 22
10 22 23 22 22 22 22 1211 12 22 22 22 23 22

Table 2. Genotypes at the PGM locus for each trunk of sampled bristlecone pine
clumps. Genotypes are designated by allozyme
migration speed on gel: 1 = fast allele, 2 = slow allele.
Genotype of designated trunk
Clump ________
number A B C D
1 22 22 22
2 22 22
3 22 22 12 12
4 22 12 22 22
5 12 12 -
6 22 22
7 22 22
8 12 12 12
9 22 22 22
10 22 22

If we assume that tree clusters are primarily recruited from animal
caches, I must caution the reader that the percentage occurrence of tree clusters may
greatly underestimate the contribution to tree recruitment by animal-mediated seed
dispersal for the following reasons: 1) as previously mentioned, electrophoretic
analysis underestimates the occurrence of tree clusters, 2) single seed caches, which
may give rise to single genet growth forms, are frequently made by nutcrackers
(Tomback 1982, Hutchins and Lanner 1982) and rodents (Vander Wall 1993), and
3) clusters of young trees may suffer attrition through time, resulting in a single
genet tree.
Baud (1993) has recently demonstrated that the winged seeds of
ponderosa and bristlecone pine germinate as readily in simulated nutcracker caches
as do the seeds of limber pine, a bird-dependent species. This suggests that the
seeds of bird-dependent pines have no special adaptations for the caching
conditions imposed by nutcrackers. With respect to her findings and the common
occurrence of the tree cluster growth form, I suggest that animals are contributing
to the ultimate seed distribution and recruitment of the Front Range "wind-
dispersed" pines.

Baida, R. P. 1981. Are seed caching systems co-evolved? Proceedings International
Ornithological Congress XVII.
Baud, K. S. 1993. Simulating Clark's Nutcracker caching behavior: germination
and predation of seed caches. Master's Thesis. University of Colorado at
Benkman, C. W., R. P. Baida, andC. C. Smith. 1984. Adaptation for seed
dispersal and the compromises due to seed predation in limber pine.
Ecology 65(2):632-642.
Carsey, K. S. and D. F. Tomback. 1994. Growth form distribution and genetic
relationships in tree clusters of Pinusflexilis, a bird-dispersed pine.
Oecologia 98:402-411.
Christensen, K. M., T. G. Whitham, and R. P. Baida. 1991. Discrimination among
pinyon pine trees by Clark's Nutcrackers: effects of cone crop size and cone
characteristics. Oecologia 86:402-407.
Fumier, 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 41:607-612.
Giuntoli, M. and L. R. Mewaldt. 1978. Stomach contents of Clark's Nutcracker
collected in western Montana. Auk 95:595-598.
Hutchins, H. E. and R. M. Lanner. 1982. The central role of Clark's Nutcracker in
the dispersal and establishment of whitebark pine. Oecologia 55:192-
Lanner, R. M. 1980. Avian seed dispersal as a factor in the ecology and evolution
of limber and whitebark pines. Pages 15-48 in Proceedings of the Sixth
North American Forest Biology Workshop, University of Alberta,
Edmonton, Alberta, Canada.

Lanner, R. M. 1988. Dependence of Great Basin bristlecone pine on Clark's
Nutcracker for regeneration at high elevations. Arctic and Alpine Research
Lanner, R. M. and S. B. Vander Wall. 1980. Dispersal of limber pine seed by
Clarks Nutcracker. J. Forest. 78:637-639.
Linhart, Y. B. and D. F. Tomback. 1985. Seed dispersal by nutcrackers causes
multi-trunk growth form in pines. Oecologia 67:107-110.
Mitton, J. B., Y. B. Linhart and J. L. Hamrick. 1979. Allozyme polymorphisms
detected in mature needle tissue of ponderosa pine. Journal of Heredity
Schuster, W. S. F. and J. B. Mitton. 1991. Relatedness within clusters of a bird-
dispersed pine and the potential for kin interactions. Heredity 67:41-48.
Shaw, C. R. and R. Prasad. 1970. Starch gel electrophoresis of enzymes a
compilation of recipes. Biochemical Genetics 4:297-320.
Shoemaker, D. P., C. W. Garland, J. I. Steinfeld, and J. W. Nibler. 1981.
Experiments in physical chemistry, 4th edition. McGraw-Hill, New York,
New York, U.S.A.
Smith, C. C. and R. P. Baida. 1979. Competition among insects, birds and
mammals for conifer seeds. Am. Zool. 19:1065-1083.
Tomback, D. F. 1978. Foraging strategies of Clark's Nutcracker. Living Bird
Tomback, D. F. 1981. Notes on cones and vertebrate-mediated dispersal of Pinus
albicaulis. Madrono 28:91-94.
Tomback, D. F. 1982. Dispersal of whitebark pine seeds by Clark's Nutcracker: a
mutualism hypothesis. Journal of Animal Ecology 51:451-467.
Tomback, D. F., L. A. Hoffmann and S. K. Sund. 1990. Coevolution of
whitebark pine and nutcrackers: implications for forest regeneration, in
Schmidt, W. C., McDonald, K. J. (compilers), Proceedings Symposium
on whitebark pine ecosystems: Ecology and management of a high-
mountain resource. USDA Inter-mountain Research Station, Ogden, Utah,
INT-270, pp. 119-129.

Tomback, D. F., Holtmeier, F. K., Mattes, H., Carsey, K. S., Powell, M. L. 1993.
Tree clusters and growth form distribution in Pinus cembra, a bird-
dispersed pine. Arctic and Alpine Research 25:374-381.
Tomback, D. F. and K. A. Kramer. 1980. Limber pine seed harvest by Clark's
Nutcracker in the Sierra Nevada: Timing and foraging behavior. Condor
Tomback, D. F. and Y. B. Linhart. 1990. The evolution of bird-dispersed pines.
Evol. Ecology 4:185-219.
Tomback, D. F., Schuster, W. F. 1994. Genetic population structure and growth
form distribution in bird-dispersed pines, in Schmidt, W. C., Holtmeier, F.
K. (compilers), Proceedings International workshop on subalpine stone
pines and their environment: the status of our knowledge. USDA
Intermountain Research Station, Ogden, Utah, INT-GTR-309, pp. 43-50.
Tomback, D.F. and C.L. Taylor. 1987. Tourist impact on Clark's Nutcracker
foraging activities in Rocky Mountain National Park. Proceedings of the
Conference on Science in die National Parks, Fort Collins, Colorado.
Vander Wall, S. B. 1988. Foraging of Clark's Nutcracker on rapidly changing pine
seed resources. The Condor 90:621-631.
Vander Wall, S. B. 1992a. The role of animals in dispersing a "wind-dispersed"
pine. Ecology 73:614-621.
Vander Wall, S. B. 1992b. Establishment of Jeffrey pine seedlings from animal
caches. Western Journal of Applied Forestry 7:14-20.
Vander Wall, S. B. 1993. Cache site selection by chipmunks (Tamias spp.) and
its influence on the effectiveness of seed dispersal in Jeffrey pine (Pinus
jeffreyi). Oecologia 96:246-252.
Vander Wall, S. B. and R. P. Baida. 1977. Coadaptations of the Clark's Nutcracker
and the pinyon pine for efficient seed harvest and dispersal. Ecol.
Monogr. 47:89-111.