Mechanical Abrasion and Intercrown Spacing
FRANCIS E. PUTZ
Department of Botany, University of Florida, Gainesville 32611
GEOFFREY G. PARKER
Institute of Ecology, University of Georgia, Athens 30602
and
RUTH M. ARCHIBALD
College of Forest Resources, University of Washington, Seattle 98195
ABSTRACT: The crowns of neighboring trees of similar height do not interdigitate
but rather are generally separated by spaces called "crown shyness" gaps. In a black
mangrove (Avicennia germinans) forest in Costa Rica, the width of crown shyness gaps
was positively correlated with the distance pairs of trees or branches adjacent to the gap
swayed in the wind (n = 22; p< .01). Abrasion of buds, leaves and branches due to
trees knocking into one another seems to create and maintain the spaces around each
tree crown.
INTRODUCTION
Though recent studies of forest structure and dynamics have focused on large gaps
caused by single and multiple treefalls, most openings in forest canopies are extremely
small (e.g.) Hartshorn, 1978; Runkle, 1982). Small gaps usually remain open for a
short time, being temporarily closed by branch movements during winds and per?
manently closed by lateral growth (e.g.) Runkle, 1982). However, in some forests light
penetrating into the canopy forms a persistent and striking border around individual
crowns that are visible from the ground (Fig. 1). The term "crown shyness" (from
Lane-Poole, 1927-1944, cited in Jacobs, 1955) refers to these leafless regions between
adjacent crowns of similar height. Much of the light reaching the understory as
sunflecks passes through crown shyness gaps. These openings may also be important
insofar as they restrict intercrown movements of arboreal animals and vines (Putz,
1982). A similar sort of defense by mechanical abrasion of competitors and dislodge?
ment of herbivores has been observed in wave-swept algae (Velimirov and Griffiths,
1979).
Two theories have been proposed to explain how crown shyness gaps are formed
and maintained. One view is that abrasion of buds on wind-blown branches leads to
open spaces between tree crowns (Tarbox and Reed, 1924; Jacobs, 1955; Richards et
al.) 1962). Ng (1977), on the other hand, proposed that crown shyness in Malaysian
dipterocarps (Dryobalanops aromatica and various species of Shorea) is caused by reduction
in lateral growth due to mutual shading. We tested the hypothesis th,at crown shyness is
caused and maintained by the mechanical abrasion of branches swaying in the wind.
Specifically, we predicted that if mechanical abrasion causes crown shyness, then flexi?
ble trees and branches should be more widely separated than rigid trees and branches.
STUDY SITE AND METHODS
We studied crown shyness in an approximately 50-ha stand dominated by Avicennia
germinans in the Parque Nacional de Santa Rosa, Guanacaste Province, Costa Rica
(10?35 'N, 85?85 'W). The forest floor is inundated with sea water only when the tides
24
1984 PUTZ ET AL. : CROWN SHYNESS 25
are especially high. Suppressed and codominant individuals of A. tonduzii Moldenke
are scattered through the forest but most of the canopy is occupied by A. germinans.
Voucher specimens were deposited in the Urtiversity of Florida Herbarium. The upper
surface of the forest canopy is level and individual crowns do not cast dense shade (Fig.
1). Wind velocities during the dry season, when the study was conducted, were high
but variable, with periods of calm interrupted by gusts to more than 11 m/sec.
If crown shyness gaps in Avicennia germinans are due to mechanical abrasion of
branches, intercrown spaces should be proportional to the potential of adjacent?
branches to rub against one another. To test this we arbitrarily selected 22 pairs of trees
of similar height, measured the distance between their crowns in still air, and measured
the maximum horizontal distance moved by branches of each tree into the crown
shyness gap when driven by gusts of wind. In other words, we compared the sum of the
lateral displacements of adjacent branches towards each other with the distance be?
tween the branches in still air. The widths of openings were measured by sighting the
positions of the crown edges from the ground relative to a measuring tape stretched
horizontally 4 m above the ground (Fig. 2). Actual distances between adjacent crowns
in still and gusty conditions were then calculated trigonometrically; heights of the
crown shyness gaps were measured directly by climbing the trees. The sum of the max?
imum horizontal displacements of tree crowns towards each other during windy
periods relative to the distance between them at rest is a measure of potential crown in?
teractions. Wind speed variability increased the variance of our estimates of maximum
horizontal distances swayed and thus introduced a conservative bias into estimates of
potential intercrown collisions.
26 THE AMERICAN MIDLAND NATURALIST 112(1)
Fig. 3). During calm periods, adjacent crowns interlocked like loosely fitting pieces of a
jigsaw puzzle (Fig. 1), but when the wind blew, they often touched and interdigitated.
These results suggest that intercrown spacing is in part determined by the potential of
adjacent canopies to knock into one another.
Branches bordering crown shyness gaps generally had broken twigs and few leaves
while branches in more protected parts of the crown had little such damage. The
absence of small twigs on many of the branches bordering crown shyness gaps makes
the proportion of damaged twigs an underestimate of the amount of damage incurred.
Dead branches, however, can be due to either abrasion or shading; hence, they cannot
be used for distinguishing between the two causes of crown shyness.
DISCUSSION
Crown shyness is most easily observed in monospecific stands of similar-size trees
but seems to be common in forests of all types. Closely spaced trees in plantations often
------gap
\ I
, I
, I
, I
, ,
, I
, I
\ /
, I
, I
1 I
\ I
\ I
\ I
, I
, I
\ I
, I
rtr--+--+--+--+----tl__.....__-+--+-fl--+--+-""""I"'f""'l- - meas UrI ng tap e
, I
, I
, I
, I
\ I
, I
, I
, I
, I
, I
\ I
, /
, /
,/
,/
- - - - 1. - - - - - - - eyele\ el of observer
Fig. 2. - Sketch showing method for measuring gap widths
1984 PUTZ ET AL.: CROWN SHYNESS 27
show distinct crown shyness (Richards et at. J 1962), but this may be because plantations
have little understory to obscure observations of crown shyness as well as because the
trees are even-aged. Crown abrasion apparently leads to reduced yields in overstocked
pine plantations (Tarbox and Reed, 1924). We expect crown shyness to be present in
all forests but to be prominent in (1) forests in windy areas; (2) monolayered forests
(where the gaps are not obscured by overlapping crowns); (3) stands of flexible trees,
and (4) early successional forests, comprised of flexible trees with limited abilities to
produce branches in lateral openings (De Castro e Santos, 1980). We observed crown
shyness between independently swaying branches on the same tree as well as between
trees. Indeed, branch, rather than whole tree, flexibility may sometimes determine
crown shyness gap size.
Crown shyness gaps may in some cases be due to reciprocal shading of adjacent
trees with similar light compensation points. We have no data on this alternative
hypothesis, but our observations of damaged leaves and twigs cannot be explained on
the basis of shading effects alone. While crown shyness may have other causes, our
study supports the importance of reciprocal pruning by mechanical abrasion in form?
ing crown shyness gaps in an Avicennia germinans forest.
Acknowledgments. - We appreciate the hospitality extended to us by the Costa Rican National
Park Service. One of us (F.E.P.) thanks A. P. Smith for rekindling interest in canopy interac?
tions and F. S. P. Ng for the original encouragement to look at the forest in three dimensions.
Contribution number 3850 to the Journal Series of the University of Florida Agricultural Experi?
ment Station.
?
1.0
E
~
?co
~
.~
c:::
0
.5~~
~
en
c:::
~~
U
0
? ? ?
? ?
?
?
? ? ??
?
? ?, ? ?
?
.5 1.0
?
r= .69
sum of displacements during winds, m
Fig. 3. - Intercrown distance in still air is significantly positively correlated with the sum of
the distances adjacent branches swayed towards each other when blown by the wind (r = .69,
p< .01)
28 THE AMERICAN MIDLAND NATURALIST 112(1)
LITERATURE CITED
DE CASTRO E SANTOS, A. 1980. Essai de classification des arbres selon leur capacite de
reiteration. Biotropica) 12: 187-194.
HARTSHORN, G. S. 1978. Treefalls and tropical forest dynamics, p. 617-638. In: P. B. Tomlinson
and M. H. Zimmermann (eds.). Tropical trees as living systems. Cambridge University
Press, Cambridge, England.
JACOBS, M. R. 1955. Growth habits of the eucalypts. Forestry and Timber Bureau, Common?
wealth of Australia. 262 p.
NG, F. S. P. 1977. Shyness in trees. Nat. Malays.) 2:34-37.
PUTZ, F. E. 1982. Natural history of lianas and their influences on tropical forest dynamics.
Ph.D. Thesis, Cornell University, Ithaca, New York. 104 p.
RICHARDS, N. A., R. R. MORROW AND E. L. STONE. 1962. Influence of soil and site on red pine
plantations in New York. I. Stand development and site index curves. Cornell Univ. Agric.
Exp. Stn. Bull. 977. 24 p.
RUNKLE, J. R. 1982. Patterns of disturbance in some old growth mesic forests of eastern United
States. Ecology) 63: 1533-1546.
TARBOX, E. E. AND P. M. REED. 1924. Quality and growth of white pine as influenced by
density, site, and associated species. Harv. For. Bull. 7. 38 p.
VELIMIROV, B. AND C. L. GRIFFITHS. 1979. Wave induced kelp movements and its importance
for community structure. Bot. Mar.) 22: 169-172.
SUBMITTED 4 MARCH 1983 ACCEPTED 8 SEPTEMBER 1983