School of Biological Sciences, University of Manchester, 3.614 Stopford Building, Oxford Road, Manchester
M13 9PT, UK
2
Institut fur Biologie III, Albert-Ludwigs-Universitat Freiburg, Schanzlestrasse 1, D-79104 Freiburg, Germany
3
Botanischer Garten, Albert-Ludwigs-Universitat Freiburg, Schanzlestrasse 1, D-79104 Freiburg, Germany
Received 24 February 2000; Accepted 7 July 2000
Abstract
Introduction
The petiole of the banana (Figs 1, 2) has several characteristic morphological features. First, the petiole is not
To whom correspondence should be addressed. Fax: q44 161 275 3938. E-mail: Roland.Ennos@man.ac.uk
2086
Ennos et al.
Fig. 1. (A) Overall view of the shoot system of a banana, showing the petioles, which form the lower regions of the pseudostem bending away from
it higher up to support the leaves. (B) Transverse section of the stem, showing the U-shaped cross-section and unusual semi-hollow structure, with
longitudinal partitions splitting up the internal air space.
2087
Fig. 2. The anatomy of the banana petiole. (A) A transverse section, showing the longitudinal partitions and the lateral partitions composed of stellate
parenchyma between them. Scale bar equals 5 mm. (B) Longitudinal section showing the regularly arranged lateral partitions. Scale bar equals 5 mm.
(C) A view of the stellate parenchyma that makes up the lateral partitions. This is composed of a single layer of star-shaped cells. Scale bar equals
50 mm.
petiole and, though corrugated, hangs like a ag `downwind'. These movements would all reduce drag dramatically, and would help reduce the force acting on the stem
of the banana. When the force is removed the petiole
reverts to its original position and the leaf, if it has not
been torn, springs back up.
The results of previous studies suggest ways in which
the design of the petiole might be adapted to allow such
2088
Ennos et al.
Fig. 3. Diagram showing the deection of the banana petiole and leaf
lamina when they are subjected to lateral wind forces. (A) Top view in
calm conditions. (B) Deection under light winds. The downwardly
curved petiole twists away from the wind, so its ventral surface points
downwind. Meanwhile the upwind lobe of the lamina exes downwind at its joint with the petiole. Together these movements produce
a streamlined ag-like shape with the two lobes of the lamina pressed
together. A further increase in wind speed (C) causes the petiole to ex
further away from the wind to produce a still more streamlined shape.
Banana petioles
Complete petioles and leaves of the banana (Musa textilis Nee)
were removed, at the point at which they left the pseudostem,
from specimens growing in the heated glasshouses of the
Freiburg Botanical Garden. The base of the petioles were
immediately placed in water, and used within a day of removal.
In between tests individual parts of the petiole were kept in
water and, if they were to be kept overnight, placed in
a refrigerator at 4 8C. They were then subjected to a range
of mechanical tests.
Morphology and overall mechanical performance
To investigate the mechanical performance and morphology of
petioles all the way along their length, and to investigate
the twist-to-bend ratio, two petioles were subjected to a range
of measurements.
(1)
2089
Fig. 4. Apparatus used for the four point bending tests. Force was
applied at the outer points on the petiole by applying weights to the
outer contact points (arrowed), and the upward movement of the centre
was measured.
(iii) Torsion tests: The petiole sections were removed from the
bending apparatus and cut into 18 cm lengths whose centre
points were at 13, 32, 58, 77, 103, and 122 cm from the stem
end of the petiole. These lengths of petiole were then mounted
at each end in a torsion balance. One end was held rm, while
the other end was located into a mounting which was free to
rotate within an outer mounting but which was attached to it
via a spiral spring that had a stiffness of 0.0541 Nm rad 1. In
this way, rotation of the outer mounting applied a torsional
moment to the piece of petiole, tending to twist it. The stiffer
the petiole the less it would twist at any given torque. For
each stem segment the outer casing was moved by 20, 40, 60,
80, 100, and 1208 and the movement of the inner casing
recorded. The torsional rigidity, GK, of the petiole section was
then calculated using the following equation.
GK 0:0541L(uo
ui )=ui
(2)
(3)
(4)
(5)
2090
Ennos et al.
from a single petiole, six each from around points 15, 30, and
60 cm from the base of the petiole. Two of each six were kept
intact, two had the stellate parenchyma removed, and two had
all the inner tissue removed. One of each was then subjected to
a dorsoventral or lateral compression test. In dorsoventral
compression the segments were crushed between a at platten
and a cylindrical rod of radius 3 mm (Fig. 5B) which was
attached to a universal mechanical testing machine (Instron
Ltd). In lateral compression they were crushed between two at
plattens (Fig. 5C). The machine continuously recorded the force
required as compression proceeded and calculated the initial
slope of the force/displacement graph.
Results
Morphology, mass and density
Fig. 5. Apparatus used for the tests to investigate resistance to buckling of the petioles and attening of their sections. (A) Axial bending
test apparatus. The specimens are held at their ends by cups which are
free to rotate up and down. The ends of the sections are moved inwards
at a constant rate, while the force required, the distance between the two
ends, and the vertical position of the centre of the section are recorded.
(B, C) Crushing test apparatus for stem sections being subjected to
dorsoventral (B) and lateral (C) compression. The sections are crushed
at a constant rate in a universal testing machine, while the force required
is recorded.
Intact stalks: On being bent, the intact petioles maintained their cross-sectional shape until the curvature was
well over 2. Only late in the process of Euler buckling
did small creases appear in the lower surface of the
petioles, but the sections never completely attened out.
A typical graph of bending moment/curvature is shown
in Fig. 7A. This shows that, unlike hollow tubes, the
bending moment does not fall dramatically either, as it
would do for a hollow tube such as the stem of Arundo
donax (Spatz et al., 1997), but after reaching a peak
slowly declined. The rigidity, EI, and strength, Mmax, of
the petioles rose with their cross-sectional area, being
2091
Fig. 7. (A) Results of a typical axial bending test for an intact petiole
section, showing a graph of bending moment, M, against curvature, C.
The exural rigidity, EI, is equal to the initial slope of the graph, the
strength Mmax, is equal to its maximum height, and the critical
curvature, Ccrit, is the curvature at which there is the rst signicant
drop in the moment minus the original curvature. (B) Results of a test on
a hollowed stem without partitions or stellate parenchyma. The graph is
similar except that the critical curvature is much lower, since the section
attens out earlier.
Manipulated petioles: The petioles that had been hollowed out showed similar behaviour to intact ones
(Fig. 7B). Their calculated material properties were also
very similar (Table 1). One way ANOVA showed that
there was no signicant difference between manipulated
and intact petioles in either the structural modulus of
elasticity, E, or the maximum compressive stress, smax.
2092
Ennos et al.
Table 1. Mechanical properties of the petioles of Musa textilis and the material properties of the outer tissue derived from axial
bending tests on intact or manipulated stem sections
Mean"standard deviations are given, along with the number of specimens tested. For critical curvature and strain, averages were calculated only
from the larger samples, in which I )100 mm4. Properties of manipulated specimens which were significantly different from intact ones are
marked by asterisks (***, P-0.001).
Intact petiole
Hollow petiole
0.63"0.51
n 22
5.52"2.83
n 21
6.15"1.31
n 13
3.36"0.87
n 13
0.65"0.44
n 10
5.47"2.20
n9
5.66"1.84
n6
2.73"0.99
n6
0.55"0.42
n 14
4.40"1.56
n 13
3.07"0.89
n 12***
1.78"0.24
n 12***
Resistance (N mm 1)
Dorsoventral compression
Lateral compression
9
9
6.30"4.14
4.29"2.29
15 cm along petiole
30 cm along petiole
60 cm along petiole
6
6
6
4.46"2.90
4.03"1.99
7.39"4.37
Intact petioles
Minus stellate parenchyma
Hollowed out
6
6
6
7.87"2.84
6.05"2.90
1.95"0.98
B
(but not the interactions) had a highly signicant
effect (Table 2). First, sections were on average 47%
more resistant to dorsoventral than lateral attening
(P-0.001). Second, sections 60 cm along the petiole were
on average 83% more resistant to attening than those
30 cm along, and 66% more resistant than those 15 cm
along (P-0.01). Third, removing the stellate parenchyma
reduced attening resistance by on average 25%, while
removing all the internal tissue reduced it by on average
75% (P-0.001).
Fig. 8. Log/Log graphs showing the relationship between (A) the
exural rigidity of petiole sections with their axial second moment of
area, I, and (B) the maximal moment and section modulus, I/xmax of
the sections. The relationship between both is more-or-less linear with
a slope not signicantly different from 1, suggesting that the material
properties of the petiole remain constant along its length. (j) Intact
petioles; (r) stellate parenchyma removed; (m), hollow petioles.
Discussion
The results of the mechanical tests in most ways conrmed the ideas about petiole design and function that
had been built up following morphological examination
and manipulation. First, the hollow design, and the longitudinal strengthening elements do produce a structure
with a low density but high exural rigidity. Undoubtedly
the hollowness of the petioles greatly increases their
exural rigidity and strength, because tissue is thereby
located well away from the neutral axis. The U-shaped
cross-section must also help it to resist downward bending
because on loading, the arms of the U will tend to move
2093
Acknowledgements
Some of the work was carried out while ARE was on sabbatical
in Freiburg. We would like to thank two anonymous referees for
extremely helpful comments.
References
Ennos AR. 1993. The mechanics of the ower stem of the sedge
Carex acutiformis. Annals of Botany 72, 123127.
Ennos AR. 1997. Flexible structures in biology. Comments on
Theoretical Biology 4, 133149.
Esau K. 1977. Anatomy of seed plants, 2nd edn. New York: John
Wiley.
Gordon JE. 1978. Structures or why things don't fall down.
London: Penguin Books.
Johnson AJ, Koehl MAR. 1994. Maintenance of dynamic strain
similarity and environmental stress factor in different ow
habitats: thallus allometry and material properties of a giant
kelp. Journal of Experimental Biology 195, 381410.
Roark RJ, Young WC. 1975. Formulas for stress and strain.
London: McGraw-Hill.
Spatz H-Ch, Beismann H, Bruchert F, Speck T. 1997.
Biomechanics of the giant reed Arundo donax. Philosophical
Transactions of the Royal Society of London B 352, 110.
Spatz H-Ch, Boomgarden C, Speck T. 1993. Contributions to
the biomechanics of plants. II. Experimental and theoretical
studies of local buckling. Botanica Acta 106, 254264.
Spatz H-Ch, Speck T. 1995. Mechanische Eigenschaften von
Hohlrohen am Beispiel von Grasern. In: Nachtigall W, ed.
Technische Biologie und Bionik 2 BIONA Report 9: 91132,
Suttgart (Fischer).
Tomlinson PB. 1969. Anatomy of the monocotyledons III
Commelinales-Zingiberales (CR Metcalfe ed.) Oxford:
Clarendon Press.
Vincent JFV, Owers P. 1986. Mechanical design of hedgehog
spines and porcupine quills. Journal of Zoology 210, 5575.
Vogel S. 1992. Twist-to-bend ratios and cross-sectional shapes
of petioles and stems. Journal of Experimental Botany 43,
15271532.
Wootton RJ. 1981. Support and deformability in insect wings.
Journal of Zoology 193, 447468.