Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
619
2014 NORTHEASTERN NATURALIST 21(4):619–629
American Chestnut Seed Dispersal and Regeneration
Bernd Heinrich*
Abstract - Wild stock Castanea dentate (American Chestnut) were planted at the edge of
a forest clearing in western Maine in 1982, and four had grown to a height of 13 m and
diameter breast high of 39 cm by 2014. They showed no sign of disease and were producing
annual seed crops. The seeds had been spread in the surrounding forest by animal-made
seed caches placed up to at least 300 m from the source; 238 surviving offspring (119 of
them one-year old) of the planted trees were located in 139 groupings. Of the surviving
seedlings, 110 were growing singly, although groupings ranged from two to 20. At least six
of the animal-spread seeds had, by 2014, produced 2–6.5-m-tall saplings that were growing
up to 70 cm per year. Experiments of seed survival showed high seed predation at or near
the ground surface, but fresh seeds placed 10 cm underground in the fall had 80% survival
to the seedling stage after one year. Comparative behavior of seed foragers, as well as the
distance and pattern of seed dispersal, implicate Cyanocitta cristata (Blue Jay) as the most
likely primary dispersers and planters of the chestnut seeds.
Introduction
Castanea dentate (Marsh.) Borkh. (American Chestnut) was largely eliminated
as a reproducing tree from the American continent by the introduced chestnut blight
near 1900. Vegetative regrowth continues to occur by root suckers, but most of the
young trees die before fruiting (Paillet 1988, 2002). Much effort is in progress breeding
blight-resistant trees to restore the species to its former range (Anagnostakis
2009). To my knowledge, there is no published data of seed dispersal and reproduction
of wild American Chestnut trees in forests, although Thoreau (1906) in his
journals speculated seed dispersal by birds because he found no young chestnut trees
in chestnut groves but only outside them in adjacent fields and Pinus (pine) stands.
Chestnut seeds are eaten and could potentially have been spread by a variety of
wildlife, including Meleagris gallopavo L. (Wild Turkey), deer, bear, Procyon lotor
(L.) (Raccoon), squirrel, mice, Cyanocitta cristata (L.) (Blue Jay) and crow, and
formerly also Ectopistes migratorius (L.) (Passenger Pigeon). Currently American
Chestnut reproduction relies primarily on regrowth from root-stocks because
the blight kills maturing trees to the roots, which may then re-sprout. It has been
proposed that the American Chestnut is a species that is not much dependent on
seed dispersal, but rather that its re-sprouting habit and the fact that the seedlings
are “essential immortal” because they can persist in shade point to an alternate
reproductive strategy (Paillett 1988). However, we know essentially nothing about
American Chestnut seed dispersal in the wild.
I here report on observations of a reintroduction of four “pure” American Chestnut
trees to a location in a Maine forest where, as of 2014, 32-year-old trees were
*PO Box 153, Weld, ME 04285; bheinrich153@gmail.com.
Manuscript Editor: Kent McFarland
Northeastern Naturalist
620
B. Heinrich
2014 Vol. 21, No. 4
annually producing fruit and spreading viable seed through the surrounding forest.
Since the forest surrounding these trees do not contain pre-existing chestnut trees,
any seedlings/new trees can be traced to the four planted trees. The locations of
the offspring relative to the parents can in turn be correlated with behavior of the
animals that are the likely the seed-dispersal agents.
Methods
The original stock of four approximately one-year-old trees was purchased in
1982 from the Wexford County Conservation Unit of the USDA in Cadillac, MI,
where they originated from an apparent group of relict trees of the main largely contiguous
population of American Chestnut trees in northeastern North America. The
seedlings were planted at an elevation of 430 m adjacent to Mount Blue State Park
in Franklin County, ME, that same year. In 2014, the trees ranged in height from
11 to 13 m and had diameter at breast height of 33.0 to 39.5 cm. They had by then
been producing viable seed for at least ten years. Observations of seed predators
were made in the autumns of 2010–2013 from a “blind” (my residence 20 m from
the trees).
Only fresh seeds derived directly from fruit about to shed the seeds (in October)
were used in tests for seed survival. Unless indicated otherwise, seeds were tested
in the forest near the site of the planted trees (during the winters of 2012–2013 and
2013–2014) for effects of physical environment, seed predators, and conditions approximating
those potentially experienced by nuts left in caches made by Blue Jays.
In the first winter, three separate treatments of 50 nuts each were divided into
10 groups of 5 each and placed into a circle of approximately 4 m diameter. Each of
the 10 “caches” per circle of seeds was flagged for aid in relocating them in following
year. In one treatment, the 50 seeds were placed directly onto the leaf-covered ground.
In the second treatment, three separate circles each with 50 seeds were buried 2 cm into
the soil beneath the leaf litter. In the third treatment, 50 seeds were buried 10 cm beneath
the leaves. Seed survival (sprouting) was checked (removing soil to expose the
seeds) six months later, on 28 April, right after snow-melt. The soil was replaced and
the presence of seedlings was then checked on 4 October.
The following winter (2013–2014) a variation of the above tests was repeated.
In one treatment, 30 seeds were placed at individually separate locations directly
below the freshly fallen leaf liter in a maple grove. Near the same site, two folded
metal window screens (to exclude seed predators) each containing 30 seeds were
placed (one on the leaves and the other under them) to test winter seed survival from
7 November 7 April.
Distribution of naturally planted seedlings and young trees were determined in
October after most of the forest trees had turned color and/or lost their leaves. At that
time, the chestnut seedlings still retained green leaves and were visibly conspicuous
from about 10 m. Searches were made approximately equally in all directions and distances
around the clearing of the parent trees. Locations of seedlings and small trees
were noted for several years and marked with flagging tape. The cache locations of
remaining trees/seedlings were recorded with a GPS in the fall of 2013.
Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
621
Results
Seed survival
Seeds left exposed to the air indoors dried out in two weeks, and none of 50
seeds left in a refrigerator on moist peat moss in an airtight plastic container
survived over a winter, due either to low temperature or mold. Another 50 seeds
under the same conditions but placed outdoors under a roof, also failed to survive.
Similarly, all thirty screened-in seeds left on the leaf litter failed to survive over
the winter, and only seven of thirty that were also screened but protected by the
leaf litter sprouted. Since the latter seeds were not molded, the major mortality was
likely due to low temperature.
Although all exposed seeds in the forest were apparently killed by freezing,
predators also accounted for high levels of seed mortality. Thirty seeds that were
individually “planted” directly under the leaf litter were all were missing by 7 May.
There were usually small-mammal tunnels in the soil where the seeds had been
deposited. Similarly, of the 50 seeds left on the surface of fallen leaves in October,
none remained the following spring after snow melt (on 28 April) and no seedlings
appeared there later. On the other hand, 21.2% of the 150 seeds buried at 2 cm had
left seedlings by 4 October the following fall, whereas 80.0% of the 50 buried to
10 cm became seedlings.
Behavior of seed predators
The four parent chestnut trees were visited by two squirrel species, and both
foraged for the nuts about a week before they were shed by the trees. Tamiasciurus
hudsonicus (Erxleben) (Red Squirrel) came at first light in the morning and
spent about an hour in the tree snipping off fruit (“burrs”) that fell to the ground.
Red Squirrels are highly territorial, and there was never more than one squirrel to
a tree. These squirrels carried individual burrs one at a time to a perch under the
tree, and there chewed through the fruit’s spines and outer shell to then consume
the seeds. Sciurus carolinensis Gmelin (Eastern Gray Squirrel) were also occasionally
present and always only one at a time. Each fed only in the upper crown
of the chestnut tree. Neither squirrel species were seen carrying off either burrs or
seeds. Tamias striatus (L.) (Eastern Chipmunk), which have cheek pouches and
could potentially carry off two chestnuts at a time to cache, are likely already in
hibernation when the chestnut seeds are shed. None were seen during the heavy
chestnut seed year of 2012.
Blue Jays were never seen on the ground under the chestnut trees, and they never
harvested whole fruits. Instead, they procured seeds by extracting them directly from
the opening fruits that were still on the tree. They then flew into or over the forest,
carrying nuts in their gular pouches. They travelled to and from the chestnut trees in
apparently preferred flight directions as single birds, pairs, and sometimes in small
groups of up to a half dozen. Cache-making was not observed since the birds usually
flew over the forest and out of sight; they did not cache nearby in the clearing. However,
cache-retrieval by them was inferred in early winter from their tracks on the first
fresh snow in the forest where leaf-litter had been pulled up.
Northeastern Naturalist
622
B. Heinrich
2014 Vol. 21, No. 4
Whether or not Blue Jays foraged on the annual October fruiting of American
Chestnuts appeared to depend on the episodic September–October masting events of
Fagus grandifolia Ehrh. (American Beech). In 2010, there was no beech seed crop,
and the Blue Jays were numerous on the chestnuts. In 2011, with a huge amount of
beech mast produced, the Blue Jays concentrated on those seeds and ignored the
chestnuts. There were again no beechnuts but a heavy crop of chestnuts in 2012, and
numerous Blue Jays were again an almost constant presence on the chestnut trees
throughout their fruiting. In 2013, there were both modest beechnut and modest
chestnut crops, and a flock of Blue Jays collected beechnuts for at least two weeks in
a grove located 100 m from the chestnuts, but no jays collected chestnuts (which ripened
a week after the beechnuts). I had not followed masting events prior to 2010.
Seed choice was affected by previous conditioning. Blue Jays coming to a feeder
and given a choice between chestnuts and sunflower seeds (in 2012) always chose
the chestnuts. During their first feeder visits they took one nut at a time, to fly to an
adjacent tree and hold the nut in their feet against a branch and then hack it open
and consume the contents in about one minute. They then returned to the feeder and
picked up one or two chestnuts to pack into their gular pouch, and to sometimes
hold a third in their bill, before flying off with them. The jays never left the feeder
with more than three chestnuts at a time. Given the same choice in the next season,
the jays ignored chestnuts for a day in preference to t sunflower seeds, which the
feeder had been stocked with up until that time, but switched over to the chestnuts
within a day.
Distribution of chestnut offspring
A hundred and ten of the total of 238 chestnut seedlings found were growing
singly. The rest of the seedlings occurred in groups of 2 or more and up to 20 individuals
(Fig. 1). The numbers of groups with two, three, four, five, ten, eighteen
and twenty seedlings (or small trees) were 13, 3, 8, 2, 1, 1, and 1, respectively. In
many of the groups (and in all those with ten or more seedlings), the nutshells were
still attached to the seedlings, which were usually in contact with each other.
The offspring of the four chestnut trees at the edge of the clearing were, when
surveyed in the fall of 2013, spread in all directions over an area of forest approximately
22 acres, and up to 300 m from the source of the seeds. This spread includes
that of a single-age cohort (Fig. 2). However, no seedlings or trees had spread
naturally into the clearing itself, the site of the nut-bearing trees. No seedlings were
found >300 m beyond the parent trees, despite routine searches of the area within
over 1 km over several years. The 139 total number of seedling-groupings/caches
(Fig. 3) and total seedlings (Fig. 4) that were discovered were not only sometimes
clumped but also located at distances that preclude them having arrived there passively
by falling from the tree.
The seedling-groupings were not associated with any apparently consistent
feature, such as natural depressions, the base of trees, etc., nor with any particular
kind of cover such as Abies balsamea (L.) Mill. (Balsam Fir) thicket, under pine
trees, or in deciduous woods. The only seeming consistency was their absence in
the clearing.
Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
623
Figure 1. A “cache” showing a group of over ten one-year-old American Chestnut seedlings
(exposed in situ in July 2013), which are still attached to their seed coats.
Northeastern Naturalist
624
B. Heinrich
2014 Vol. 21, No. 4
The seedlings per any one group were always of the same apparent age and
similar size. The size-distribution of seedlings/trees by group were strongly skewed
toward small size (height), with 37 at 7–10 cm, 69 at 11– 50 cm, 24 at 51 cm–1 m,
7 at 1–2 m, and 2 over 6 m. The latter had reached access to sunlight due to a hole
Figure 2. Distribution of all one-year-old American Chestnus seed caches located in October
2013 (superimposed on a summer Google Earth photo showing the clearing where the four
parent trees are located).
Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
625
in the tree canopy, and showed twig growth of up to 70 cm per year. Seedlings from
under conifer cover (nearly continuous shade) were all less than 15 cm tall. Age
of seedlings was strongly skewed toward young plants: 119 of the total of the 239
were one-year olds (i.e., planted in October 2012).
Figure 3. Distribution of all found groupings of young chestnut trees and seed caches
per GPS location, including where a number of caches/groupings of trees were situated
within 3 m of one another.
Northeastern Naturalist
626
B. Heinrich
2014 Vol. 21, No. 4
Discussion
American Chestnut fruit (“burrs”) are spiny balls of about 5 cm diameter that
each contain three seeds. The burrs’ sharp spines appear to be a design protecting
the seeds from some predators. However, perhaps the fruit saves the seeds/nuts for
Figure 4. Distribution of numbers of seedlings situated within approximately 3 m of one
another per location.
Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
627
specific seed dispersers (since when the fruit ripen, three flanges of the burr curl to
the sides to present and release the nuts).
In this study in western Maine, the spiny chestnut burrs did not exclude squirrels
from feeding on the seeds; both Red and Eastern Gray Squirrels fed on the seeds
days before the burrs opened. Blue Jays, who came only after the fruits released
their seeds, were the only birds seen to carry off seeds.
Blue Jays are “scatter hoarders” (Brodin 2010) that are well-known cachers of individual
acorns (Bossema 1979, Crowley and Long 1995, Crowley-Hill and Johnson
1981, Johnson et al. 1997, Moore and Swihart 2006, Richardson et al. 2013). Yet in
the present study, up to 20 chestnuts were closely packed together into one group, and
another apparent cache of almost the same number of nuts next to it along with numerous
other smaller groups of seeds nearby, indicate that multiple caching trips had
been made to the same area and also to the same caching site in it. Jays presumably
use visual landmarks for orientation and for remembering their caches (Balda and
Kamil 1989, Gomez 2003, Kelly et al. 2010, Thayer and Vander Wall 2005). However,
there was no consistent and obviously site-specific distinguishing characteristic at
this site relative to any other site except for the locally specific landmarks at the cache
itself. Therefore, perhaps the clumped caching, resulting by repeatedly frequenting
the same direction and/or location, is a mnemonic strategy that enables the jays to remember
distant cache sites distributed over a wide area.
Those seeds that are not recovered may sprout and aid in the trees’ reproduction
through dispersal and colonization of new habitat. The question remained, however,
to what extent chestnut seeds might survive on their own in the environment of the
woods without the intervention of the seed dispersers. The present study indicates
that in western Maine, at the edge of the American Chestnuts’ former range, seed
survival could be problematical because of high mortality on the ground due to both
freezing-intolerance and seed predation. Wildlife could therefore be important for
American Chestnut reproduction not only by seed dispersal but also by providing
an environment for the seeds that enhances their survival.
The distribution of seedlings through the forest indicated that they were spread
by winged dispersal. Squirrels could not have carried more than one seed at a
time, although when they open a fruit they can get three. Eastern Gray Squirrels
were never observed to have left the trees, and Red Squirrels are highly territorial,
so squirrels, although they cannot be excluded from consideration for burying
seeds, seemed unlikely candidates responsible for the distribution of the seedlings
here described. The jays were never seen to deposit seeds within the clearing,
and no seedlings sprouted there. They were the only birds collecting seeds, and
a “highway” or highways of them flew in specific directions and could explain
some of the clumped distribution of the chestnut seed caches. Based on these observations
and considerations, I inferred that the jays, rather than squirrels, are the
likely the trees’ original seed dispersers. However, subsequent observations made
after this paper was originally written have shed new light on the role of squirrels
in dispersing American Chestnuts and will be presented in a follow-up paper.
Northeastern Naturalist
628
B. Heinrich
2014 Vol. 21, No. 4
Since seedling mortality is unknown, the number of observed seedlings per
cache represents the minimum number per fresh seed cache. Secondly, although
evidence of cache recovery by Blue Jays was inferred from tracks on fresh snow,
cache recovery rates are also unknown. Single-growing seedlings could therefore
be the result of caches where a jay recovered some but not all of the seeds originally
placed there, and/or because of seed mortality such as freezing and seed predation
by rodents and seedling mortality due to browsing by deer or hare.
Ironically the American Chestnut fruit appear to be best defended against the
jays, which are the main dispersers of acorns and beechnuts (Crowley et al 1981,
Richardson et al 2013). Squirrels, who penetrated the fruit to get at the seeds, were
not seen to disperse them. Paillet (2002) proposed that American Chestnuts have
little or no selective pressure to deter seed predators because they exhibit few of
the mechanisms usually associated with tradeoffs between dispersal and predation
(Howe and Smallwood 1982, Vander Wall 2001).
The present observations of American Chestnut reproduction, where both the
source and the distribution of the chestnut offspring from seeds are known, suggests
that there may be a balance between potential squirrel predation versus bird
dispersal. This balance likely varies from one year to the next depending on the
respective relative abundance of the animals and the number and kinds of seeds
available. In the absence of squirrels, a huge crop of nuts would be available “suddenly”
due to the opening of the fruit, and the birds would have a surplus of seeds
to cache. Furthermore, the balance of seed predators and dispersers depends on
competition between mast tree species. In this study, the jays were not harvesting
chestnuts when there were beechnuts. When Blue Jays are absent, then the chestnut
seeds not harvested fall to the ground and are available for squirrels to cache. With
both squirrels and jays present, the potential for distant seed dispersal could be reduced,
but would be compensated for by a larger seed crop that would swamp the
effect of the predation of the fruit by the squirrels. The different dynamics of one
potential seed disperser versus another and other mast trees has not been clarified
and suggests future study.
Acknowledgments
I thank Annie W. Rosenbauer and Gus Goodwin for their generous help in GPS data
processing.
Literature Cited
Anagnostakis, S.L. 2009. American Chestnuts in the 21st century. Arnoldia 66(4):22–31.
Balda, R.P., and A. C. Kamil. 1989. A comparative study of cache recovery by three corvid
species. Animal Behavior 38:486–495.
Bossema, I. 1979. Jays and oaks: An eco-ethological study of a symbiosis. Behaviour
70:1–117.
Brodin, A. 2010. The history of scatter-hoarding studies. Philosophical Transactions of the
Royal Society B 365:869–881.
Northeastern Naturalist Vol. 21, No. 4
B. Heinrich
2014
629
Crowley, M.J., and C.R. Long. 1995. Alternate bearing, predator satiation, and seeding
recruitment in Quercus robur L. Journal of Ecology 83:683–696.
Crowley-Hill, S., and W.C. Johnson. 1981. Acorn dispersal by the Blue Jay (Cyanocitta
cristata). Oecologia 50:231–232.
Gomez, J.M. 2003. Spatial patterns of long-distance dispersal of Quercus ilex acorns by
jays in a heterogeneous landscape. Ecography 26:573–584.
Howe, H.F., and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology
and Systematics 13:201–228.
Johnson, W.C., C.S. Adkisson, T.R. Crow, and M.D. Dixon. 1997. Nut caching by Blue Jays
(Cyanocitta cristata L.): Implications for tree demography. American Midland Naturalist
138:357–370.
Kelly, D.M., A. Kamil, and K. Cheng. 2010. Landmark use by Clark’s Nutcrackers (Nucifraga
columbiana): Influence of disorientation and cue rotation on distance and direction
estimates. Animal Cognition 13:175–188.
Moore, J.E., and R.K. Swihart. 2006. Nut selection by captive Blue Jays: Importance of
availability and implications for seed dispersal. The Condor 108 (2):377–388.
Paillett, F.L. 1988. Character and distribution of American Chestnut sprouts in southern
New England. Bulletin of the Torrey Botanical Club 115:322–344.
Paillet, F.L. 2002. Chestnut: History and ecology of a transformed species. Journal of
Biogeography 29:1517–1530.
Thoreau, H.D. 1906. Journal, Vol. XIV. Houghton Mifflin Co., Boston, MA .
Richardson, K.A., N.I. Lichti, and R.K. Swihart. 2013. Acorn foraging preferences of four
species of free-ranging avian seed predators in eastern deciduous forest. The Condor
115:863–873.
Thayer, T.C., and S.B. Vander Wall. 2005. Interactions between Steller's Jays and Yellow
Pine Chipmunks over scatter-hoarded Sugar Pine seeds. Journal of Animal Ecology
74:365–374.
Vander Wall, S.B. 2001. The evolutionary ecology of nut dispersal. Botanical Review
67:74–117.