| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Anatomy and Morphology |
2Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts 02138 USA; 3Laboratoire de Physiologie Végétale, Forêts et Biodiversité, Université d'Antananarivo, Antananarivo (101), Madagascar
Received for publication February 9, 2006. Accepted for publication June 28, 2006.
ABSTRACT
The stems of large trees serve in transport, storage, and support; however, the degree to which these roles are reflected in their morphology is not always apparent. The large, water-filled stems of baobab trees (Adansonia spp.) are generally assumed to serve a water storage function, yet recent studies indicate limited use of stored water. Through an analysis of wood structure and composition, we examined whether baobab morphology reflects biomechanical constraints rather than water storage capacity in the six Madagascar baobab species. Baobab wood has a high water content (up to 79%), low wood density (0.09–0.17 g · cm–3), high parenchyma content (69–88%), and living cells beyond 35 cm into the xylem from the cambium. Volumetric construction cost of the wood is several times lower than in more typical trees, and the elastic modulus approaches that of parenchyma tissue. Safety factors calculated from estimated elastic buckling heights were low, indicating that baobabs are not more overbuilt than other temperate and tropical trees, yet the energy investment in stem material is comparable to that in temperate deciduous trees. Furthermore, the elastic modulus of the wood decreases with water content, such that excessive water withdrawal from the stem could affect mechanical stability.
Key Words: anatomy • biomechanics • Bombacaceae • construction cost • Madagascar • parenchyma • safety factor • water storage
Understanding plant structure and function requires a consideration of the many roles that a particular organ may fill. The stems of large trees, for example, serve as conduits for the transport of water and carbohydrates, provide mechanical support, and are storage sites for various resources. Although certain plant structures, such as spines and thorns, may be readily attributable to a particular function, others are deceiving in their apparent simplicity. For example, the stomatal plugs found in plants of the family Winteraceae were thought to reduce transpirational water loss, seemingly an advantage to plants that lack vessels in their wood. However, Feild et al. (1998)
report that stomatal plugs do not prevent water loss in Drimys winteri, but instead serve to maintain higher rates of gas exchange when the leaves have water on their surfaces.
Baobab trees (Adansonia L), which occur in seasonally dry regions of Africa, Madagascar, and Australia, have long been assumed to depend on water stored in their large, swollen stems to survive in these arid environments (Newton, 1974
; Owen, 1974
; Wickens, 1983
). Recent results, however, indicate only a limited use of stored water in baobab trees for physiological processes such as leaf flushing and buffering of daily water deficits (Chapotin et al., 2006a
, b
). In this study, we therefore address the possibility that swollen stems and high water content may reflect constraints other than water storage, and we consider alternative hypotheses to explain the distinctive morphology of the baobab tree.
In addressing the question of baobab morphology, we seek to reexamine the assumption that resource storage justifies the large size of baobab stems, and in doing so we approach the question from a biomechanical perspective. Baobab wood has a very low density and is often described as soft, spongy, and unsuitable for construction purposes (Guy, 1971
; Owen, 1974
; Wickens, 1983
; Dhillion and Gustad, 2004
). Wood of this description is expected to be quite weak, yet some species of baobab grow to heights of 30 m or more (Baum, 1995
). Based on general mechanics, a tree comprised of soft, weak wood will require a larger diameter than a tree having much stiffer wood, if they both are to grow to equal heights and support equal crown masses while maintaining comparable mechanical stability.
In this study, we complete a structural analysis of baobab trees to determine whether biomechanical constraints might explain their morphology. The wood in baobab stems is comprised mainly of parenchyma tissue, much of which remains alive for many years and at significant depths into the stem. Unlike highly lignified woody tissue, the stiffness of parenchymatous tissue is affected by moderate changes in water content and turgor pressure (Niklas, 1989
). Therefore, this structural analysis considers not only the degree to which stem diameter is explained by biomechanical considerations, but also whether the water contained in the stem of the baobab tree may play a structural, rather than a physiological, role.
To our knowledge, there have been no biomechanical studies of trees with stem parenchyma content comparable to that in the baobab. Many authors have considered biomechanical constraints on the size and form of trees (McMahon, 1973
; King and Loucks, 1978
; Holbrook and Putz, 1989
; Niklas, 1992
), some biomechanical studies have been completed on cacti (Niklas and Buchman, 1994
; Molina-Freaner et al., 1998
), and several authors have specifically addressed the biomechanics of parenchyma tissue and the degree to which biomechanical properties depend on tissue water content and turgor pressure. Niklas (1988)
demonstrated that the elastic modulus (E) of parenchyma tissue changes as a function of water potential in cylinders of potato tuber incubated in mannitol solutions of varying concentrations. Elastic modulus decreased six- to 12-fold as tissue water potential decreased from –0.4 to –1.4 MPa. In another study, the flexural rigidity of chive leaves similarly decreased as a function of water potential (Niklas and O'Rourke, 1987
). Spatz et al. (1998)
considered the case of Equisetum giganteum, a hollow-stemmed plant with a thin outer ring of strengthening tissue encircling a thick ring of parenchyma tissue. They determined that the outer ring of tissue is primarily responsible for mechanical stability but that the parenchyma tissue provides resistance against local buckling. As in the cases described previously, they also found that the elastic modulus of the Equisetum stem parenchyma tissue declined with turgor pressure. Molina-Freaner et al. (1998)
speculated that seasonal changes in water content may affect stem stiffness in cacti, another group of plants having a high parenchyma content.
Because of their high parenchyma content and low wood density, baobab stems are likely to face biomechanical constraints different from those of tree stems with more typical wood structure, yet it would be incorrect to compare them too closely to plant structures that are comprised entirely of storage parenchyma with very few interstitial air spaces, such as the potato tuber described previously. The mechanical properties of baobab stem wood may more closely resemble those of the rhizome of Arundo donax, a structure similarly comprised of unlignified parenchyma tissue with embedded vascular bundles (Speck and Spatz, 2003
). As an upright, aboveground structure, however, the baobab stem is subjected to mechanical forces additional to those faced by the potato tuber or the Arundo rhizome.
Our objectives in this study are to ascertain to what degree the stem morphology and allometry of baobab trees can be explained through a consideration of their structural properties and biomechanical constraints and to determine whether the water content of the stem plays a role in structural support. In support of the analysis, we measured the construction costs of baobab wood to compare the energy investment in baobab stems with that in stems from more typical trees. We also examined whether the abundant parenchyma cells in the stem could serve as sites for the storage of carbohydrates and other nutrients, which would be important in an environment where fluctuating water availability limits carbon uptake and nutrient acquisition. There is little information available on the biology of baobab trees, and many basic details of the anatomy and physical composition of the wood are lacking, particularly for the Malagasy species. Therefore, in addition to completing the biomechanical and carbohydrate analyses, we present detailed information on stem water content, wood density, anatomy, and cellular composition in six species of baobab trees from Madagascar.
MATERIALS AND METHODS
Study species and field sites
Six species of baobab were included in this study: Adansonia za Baillon, A. rubrostipa Jum. & H. Perrier, A. grandidieri Baillon, A. madagascariensis Baillon, A. suarezensis H. Perrier, and A. perrieri Capuron. The primary field site for this study was the Kirindy Forestry Reserve near the town of Morondava in western Madagascar (A. za, A. rubrostipa, and A. grandidieri). Additional field sites near the town of Antsiranana in northern Madagascar were the Special Reserve of Ankarana (A. madagascariensis and A. perrieri), the Beantely Forest (A. suarezensis), and Cap Diego (A. suarezensis). Mature study trees ranged from 1 to 2.6 m in diameter; additional measurements were made on younger trees less than 0.8 m in diameter. Access to the crown of the trees for branch sampling and allometry measurements was through the use of the single rope technique (Laman, 1995
).
Wood sampling
Stem wood was sampled by extracting cores 12 mm in diameter and 50 cm in length from the main stem of the tree between 1 and 1.5 m from the ground using an increment borer. The main branches in the crown (20–50 cm in diameter) were also sampled with the increment borer. Branch segments 5 cm in length were sampled from small branches (0.5–4 cm in diameter) using a handsaw and pruning shears.
Anatomy
Stem cores from all six species (4–5 trees per species) and small branch segments from A. rubrostipa and A. za (three trees per species) were preserved in 50% isopropanol for anatomical analyses. Samples were sectioned by hand and with a sliding microtome and stained with phluoroglucinol and HCl for lignin, Iodine-KI for starch granules, or toluidine blue as a general stain. An AxioCam HRc camera (Carl Zeiss, Göttingen, Germany) mounted on a microscope was used to take pictures and Photoshop (Adobe Systems, San Jose, California, USA) and ImageJ (U.S. National Institutes of Health, Bethesda, Maryland, USA) were used for image analysis. Vessel areas and densities were measured on transverse stem and branch sections. Vessel diameters were calculated from the equivalent circular area of each vessel area. Average and maximum vessel diameters from each sample were used to calculate the averages for each species. Percent parenchyma area was measured in transverse stem sections by subtracting the vessel lumen area and the area of lignified tissue (as determined through staining with phluoroglucinol and HCl) from the area of each section.
Wood density and water, wood, and air fractions
Stem cores were extracted from five mature trees for each of the six species for determination of wood density and wood, water, and air fractions. In A. rubrostipa and A. za, cores were extracted from two branches per tree and 12–18 small branch segments were collected from each tree for three trees per species. To obtain the density of the wood formed during the height extension stage of tree growth, stem cores were extracted from five young trees each of A. rubrostipa and A. za having a radius smaller than the length of the increment borer. For the cores from mature trees and the branches, the wood was separated from the bark and then divided into sections 5 cm in length. In the stem cores, the section immediately beneath the cambium was further divided into two 2.5-cm sections. The young tree cores were divided into 5-cm sections starting from the center of the tree, and the innermost section was further divided into two sections. The bark and all wood sections were then immediately sealed into preweighed plastic bags. The small branch segments were sealed into plastic bags without further subdivision after the bark was removed and discarded. Fresh mass and volume were determined before drying the sections at 75°C for 24 h to a constant mass. Wood density was calculated as the ratio of dry mass to fresh volume, and volumetric water, wood, and air fractions were calculated on a percentage of fresh volume basis as described by Domec and Gartner, where dry cell wall material is assumed to have a density of 1.53 g · cm–3 (2002).
Triphenyltetrazolium chloride vital staining
Fresh stem wood sections from A. rubrostipa, A. za, and A. grandidieri were incubated with triphenyltetrazolium chloride (TTC) to determine the depth into the stem to which parenchyma cells remained alive (Ruf and Brunner, 2003
). One core was taken from each of five trees per species and cut into 3–4 mm wide sections. Core sections were sampled with greatest frequency near the vascular cambium. Ten sections were sampled from 0–3 cm from the cambium, then one section every centimeter from 3 to 6 cm, and then one section every 3 cm thereafter, up to 39 cm from the cambium. Wood sections were incubated in a 1% mass/volume solution of TTC in water for 24 h, after which time the outer tangential surface of each section was shaved clean and scored for color. Radial parenchyma, axial parenchyma, and tangential bands comprised exclusively of parenchyma were scored separately. Based on their color scores, sections were assigned to one of three categories: all parenchyma cells still living, some parenchyma cells dead, and all parenchyma cells dead.
Total nonstructural carbohydrate content (TNC)
Total nonstructural carbohydrate (TNC) content of stem wood sections was measured at five depths into the stem and at five times during the year in three species: A. rubrostipa, A. za, and A. grandideri. One core was taken from each of five trees per species on each of the following dates: 6 February, 12 March, 3 July, 15 September, and 15 December 2003. Sections of cores sampled were from depths of 0–2.5, 2.5–5, 5–10, 15–20, and 30–35 cm from the cambium. Oven-dried wood sections were ground to a fine powder in a Retsch MM200 mixer mill (Retsch, Haan, Germany) and then dried again at 75°C to a constant mass. The TNC analysis was performed as described in Bauer et al. (1997)
with modifications by Richer (2004)
. Sugars were extracted with ethanol, and an acid hydrolysis reaction was used to convert starch to glucose. The addition of the anthrone-sulfuric acid reagent to the extracted sugars causes a color change such that the absorbance of the solution at 730 nm is proportional to the concentration of glucose-equivalent sugars. Absorbance was measured with an ELx Microplate Reader (Biotek Instruments, Winooski, Vermont, USA), and each sample run was calibrated with a series of standard glucose solutions. The data were analyzed with a repeated-measures ANOVA model to test for significant effects of time, depth, and species on TNC concentration (Statistica 6.1; StatSoft, Tulsa, Oklahoma, USA).
Calorimetry and lignin content
Wood samples were combusted in an oxygen bomb calorimeter (1341 Plain Jacket Calorimeter; Parr Instrument, Moline, Illinois, USA) to determine the heat of combustion. Additional material from the same samples was analyzed by Rock River Laboratory (Watertown, Wisconsin, USA) for lignin content (72% sulfuric acid method). One stem core was extracted from each of five trees per species for all six species. Oven-dried sections of stem cores 15 cm in length (5–20 cm beneath the cambium) were ground to a fine powder in a Retsch MM200 mixer mill and then dried again at 75°C to a constant mass. Approximately 0.5 g of sample was combusted in the bomb calorimeter according to the manufacturer's instructions (Leith, 1975
; Anonymous, n.d.
). Ash content was determined by heating samples in a muffle furnace at 500°C for 6 h (Neenan and Steinbeck, 1979
). Heat of combustion is reported uncorrected as well as on an ash-free basis.
Bending tests
Segments of stem wood and bark were subjected to four-point bending tests to determine their elastic modulus (Holbrook and Putz, 1989
). The experimental setup is described in Fig. 1. Elastic modulus (E) was measured according to the following formula:
|
| (1) |
|
Water content and wood density were measured on the same or adjacent segments. Mean water content, density, and radial elastic modulus were compared between the two positions in each species by conducting paired t tests. To learn how water content and wood density may interact to affect elastic modulus, a factorial linear regression analysis was conducted. Because there were not sufficient samples to conduct this analysis for each species and there was only a small effect of species on elastic modulus, the outer core sections were combined into one data set, and the analysis was conducted on this data set (Statistica 6.1).
Tree biomass and buckling height calculations
Overall tree biomass was measured and a safety factor calculated for each of nine trees of A. rubrostipa and four trees of A. za. On each tree, the height at which major branches diverged from the trunk and the diameter at the base, midway up the trunk, and at the point of main branch divergence were measured. Values of fresh density were used to calculate total stem mass based on the above measurements, taking into account the decreasing fresh wood density (due to decreasing water content) with increasing depth into the stem. To obtain an estimate of crown mass, the diameter and length to the next major branching point of all the main branches on the tree were measured. These measurements were repeated on each successive order of branches until they measured approximately 20 cm in diameter or were no longer accessible. Using the values obtained for fresh branch wood density, the calculated mass of each measured branch was summed over all measured branches in the crown to obtain a lower limit on crown mass. These crown mass values are underestimates because they do not include the mass of the smaller branches or leaves.
Critical heights for the trees measured above were estimated assuming the tree stems to be columns subject to the limits imposed by elastic buckling. We applied two different formulas to calculate the critical buckling height (Hcrit) and modified them (description follows). The first formula, developed by Greenhill in 1881 (McMahon, 1973
), assumes a self-loaded, non-tapered column:
|
| (2) |
expanded on Greenhill's formula by taking into account the mass of the crown in addition to the self-loading of the stem and by introducing stem taper into the formula:
|
| (3) |
|
| (4) |
|
| (5) |
Baobab trees are comprised of a soft, low-density material (wood) ensheathed by a thin-walled cylinder of a stronger, higher density material (bark). As the material furthest from the center of a column or hollow cylinder imparts the greatest strength onto a structure (Niklas, 1991
), it was necessary to take into account the contributions of both the bark and the wood in calculating the critical buckling height. Therefore, we treated the baobab stem as a large solid cylinder nested inside a thin-walled hollow cylinder and modified the above formulas by summing the flexural stiffness calculated for each section of the stem, where flexural stiffness is the product of elastic modulus (E) and the second moment of area (I) (Speck and Rowe, 2003
). This results in the modified formula:
|
| (6) |
To determine the degree to which baobab trees are "overbuilt," a safety factor was calculated for each of the 13 trees using the four different formulas, where:
|
| (7) |
RESULTS
Anatomy
Mean vessel diameters ranged from 173 to 211 µm in the main stem and from 47 to 49 µm in the small branches, and mean maximum vessel diameters from 338 to 364 µm and from 104 to 119 µm, respectively. Vessel density and lumen area are very low in the main stem wood, ranging from 1 to 2 vessels/mm2 and from 3.9% to 6.7% lumen area (Table 1). Vessels in the main stem are present either singly or in clusters of 2–3 and occasionally more (Fig. 2). Vessel density is much higher in the small branches, ranging from 90 to 100 vessels/mm2 for the two species examined (A. rubrostipa and A. za) with vessel lumen areas of 18% (Table 1; Fig. 3). Total parenchyma content in the stem wood ranged from 67% in A. rubrostipa to 76% in A. grandidieri (Table 1). Starch granules are abundant in the ray parenchyma and to a lesser extent in the axial parenchyma. Several of the species have tangentially oriented bands comprised exclusively of parenchyma tissue that are separated by bands of xylem also interspersed with parenchyma cells. Rays are multiseriate and abundant. Fibers are few, scattered, and thin-walled in the stem wood, more abundant and with thicker walls in the branches. Tyloses occur frequently in vessels of both stems and branches.
|
|
|
|
|
|
|
|
|
|
; Panshin and de Zeeuw, 1970
; Neenan and Steinbeck, 1979
; Wullschleger et al., 1997
). Ash-corrected heats of combustion were therefore somewhat higher, ranging from 17.54 x 106 to 18.67 x 106 J · kg–1, yet still lower than most values reported for wood in the literature (Neenan and Steinbeck, 1979
; Smith and Tumey, 1982
; Gower et al., 1984
; Wullschleger et al., 1997
). Because wood densities are so low, volumetric construction costs are therefore extremely low, ranging from 1500 J · cm–3 in A. grandidieri to 2900 J · cm–3 in A. perrieri. Lignin content of the wood is also low, ranging from 5.8% to 8.7%; typical lignin contents in wood generally range from 15% to 35% (Panshin and de Zeeuw, 1970
).
Bending tests
From the four-point bending tests conducted on longitudinal sections from the main stem of A. rubrostipa, the calculated elastic modulus was 77 ± 20 MPa for the outer sapwood and 206 ± 5 MPa for the bark. Elastic moduli calculated from bending tests conducted on radial cores from five of the baobab species ranged from 11 to 26 MPa in the outer core sections (0–15 cm beneath the bark) and from 2 to 17 MPa in the inner core sections (20–35 cm beneath the bark) (Fig. 9). Elastic modulus and water content in the outer core sections were consistently greater than in the inner core sections across all species, suggesting a dependence of elastic modulus on water content. The results of paired student's t tests indicated significant differences (P < 0.05) in elastic modulus between outer and inner core sections for all species but A. suarezensis (P = 0.09), significant differences in water content for all species but A. za, and significant differences in wood density only for A. za.
|
), and the interaction between water content and wood density on elastic modulus (E). The model yielded the following regression equation with the whole model P value equal to 0.019:
|
| (8) |
Tree biomass and buckling height calculations
For the 13 trees measured, the mass of the main stem ranged from 8800 to 23 700 kg and that of the main branches of the crown from 900 to 8600 kg (Table 4). Whole crown mass in these trees should be greater than the values presented here because the methodology ignored the mass of branches smaller than 20 cm in diameter, and occasionally difficulties with canopy access did not allow for the measurement of some of the larger branches.
|
DISCUSSION
Baobab trees exhibit considerable diversity in morphology, allometry, and anatomy, both across the different species and between populations within species. Adansonia rubrostipa, for example, grows as a tall columnar tree in the Kirindy Forest, yet on a dry rocky outcrop only a few kilometers away, it is short, squat, and bottle-shaped. Because most of the measurements in this study were made on individuals from one or two similar populations per species, mean values presented here should not be taken as representative for each species. Our goals here are to provide a general understanding of the morphology, anatomy, and biomechanics of the Malagasy baobab species as a group. Therefore, in the discussion that follows, we place less emphasis on explaining the differences observed between the species and instead focus our attention on characteristics common to all species. Much further work could be done to understand the manner in which the different species and populations of baobabs respond to various environmental variables.
Wood anatomy and composition
Baobab wood contains a greater proportion of parenchyma cells in the main stem (69–88% of volume) than any other trees for which such data are available. Consequently, stem water content is very high, wood density is very low, and the volumetric solid fraction of the wood is as low as 5% in some species. Very few trees have wood densities approaching the values measured in this study (the lowest being 0.09 g · cm–3 in A. grandidieri). Balsa (Ochroma pyramidale), with a wood density that can be as low as 0.12, is perhaps the best known example, but a few others, such as Aeschynomene hispida (0.044; Fabaceae) and Alstonia spathulata (0.058; Apocynaceae) have lower density wood (Kanehira, 1933
; Wiemann and Williamson, 1989a
). Vessel densities in the main stem (1–2 vessels/mm2) are similarly at the extreme low end of the range reported for tree species (Panshin and de Zeeuw, 1970
; Gartner, 1995
), and maximum vessel diameters are very large. The anatomical data presented here are not meant to be exhaustive. Baobab wood presents many unusual features, including variability in axial parenchyma cell shapes, tangential bands of parenchyma cells, and variable patterns of parenchyma cell death, which warrant further study. For example, it appears that the parenchyma cells may undergo radial elongation or further cell division during later stages of stem development when they are no longer adjacent to the vascular cambium. In this respect, the wood may be similar to the wood of Jacaratia spp. and other trees in the Caricaceae (Holbrook and Putz, 1992
; Carlquist, 1998
). Observations to support this speculation include the fact that the thickness of parenchyma bands increases with distance from the cambium (to a distance of 3–4 cm into the stem) and the consistently greater wood density of the outer 2.5 cm of sapwood over that of the wood deeper in the stem (data not shown).
Total nonstructural carbohydrate content (TNC) in the outer sapwood (15–24% of dry mass) is much higher than values reported in the literature for temperate tree stems (Barbaroux and Bréda, 2002
; Hoch et al., 2002
; Landhausser and Lieffers, 2003
; Wong et al., 2003
), but generally within the range of values reported for tropical tree species (Bullock, 1992
; Mooney et al., 1992
; Richer, 2004
). The TNC content decreases towards the center of the stem, paralleling the pattern of radial and tangential band parenchyma death, yet even at 30–35 cm from the bark, percent TNC remains quite high (8–13%). Although there were seasonal changes in TNC content, at no time was TNC depleted by more than 25% of the maximum value at each position in the tree. We made no measurements of TNC content in roots or branches, although there is likely to be significant carbohydrate storage in these organs as well (Bullock, 1992
; Mooney et al., 1992
). The seasonal drop in TNC content in the stem precedes leaf flush by up to several months, and patterns of carbohydrate reallocation between different parts of the tree may also be important. With regard to whether the stem plays an important role in resource storage, it is clear that stem carbohydrates do fluctuate significantly on a seasonal basis and that parenchyma cells deep in the stem contain significant quantities of carbohydrates. It is not clear whether carbohydrate storage can be considered justification for the large stem size in baobab trees, yet the contribution of stem storage to overall carbohydrate dynamics in baobab trees cannot be ignored.
Wood structural properties and critical buckling heights
Further consequences of the high parenchyma content in baobab wood are the low lignin content (6–9% of dry mass), the low elastic modulus, and the dependence of elastic modulus on water content. At 77 MPa, the elastic modulus measured in the longitudinal direction for the outer sapwood of A. rubrostipa is much lower than any values reported for trees. It is still higher than values reported for pure parenchyma tissue (1.08 MPa, saguaro; 19 MPa, potato) but within the range of values reported for rhizomes of Arundo donax (67–112 MPa), plant organs having a similarly structured tissue comprised of parenchyma with embedded fibers and vascular strands (Niklas, 1988
; Speck and Spatz, 2003
). Baobab wood, like all other wood, is anisotropic in a manner that pure parenchyma tissue is not; the ratio of elastic modulus in the longitudinal direction to the radial direction (El/Er) is equal to seven in A. rubrostipa. We found that water content had a significant effect on the radial elastic modulus, and for the purposes of this analysis we make the assumption that changes in water content would similarly affect the longitudinal elastic modulus. We also assume that the anisotropy ratio measured in A. rubrostipa is constant between species so that we can estimate a longitudinal elastic modulus in A. za, for which only the radial elastic modulus was obtained. Obtaining a longitudinal elastic modulus for all species would clearly have been preferable, but as baobab trees have limited ranges and are considered threatened in Madagascar, it was simply not possible to conduct the destructive sampling that would have been necessary to obtain these data.
Before proceeding with a discussion of the results obtained from the elastic buckling analysis, it should be noted that the assumption of elastic behavior generally requires several conditions to be met (Niklas, 1992
). One condition is that the construction material must be isotropic, such that elastic modulus is uniform throughout the structure. This condition is not met in this study or any other study of tree biomechanics, since wood is anisotropic and there is a difference in E between wood and bark. Nevertheless, this is often ignored when elastic buckling formulas are applied to plant structures, and these formulas have in many cases been shown to be empirically valid. A second condition is that the proportional limit of the material cannot be exceeded through loading or it will no longer behave as an elastic material and will not return to its original dimensions after the load is removed. To evaluate this condition, we calculated a slenderness ratio for the tree (height/radius) and compared it to the minimum slenderness ratio allowable given the elastic modulus (E) and the critical compressive strength (
) for the material, according to the following formula (Niklas, 1992
):
|
| (9) |
Evaluating this formula with the value measured for E (77.5 MPa) and a value of 3.87 MN · m–2 reported for the proportional limit measured under compression for balsa wood (Ochroma pyramidale, a related tree having a wood density similar to baobab; Niklas, 1992
) indicates that a baobab tree must have a slenderness ratio (h/r) larger than 7.0 for the elastic buckling formula to be applicable. The slenderness ratios for the trees in this study ranged from 11 to 27. Although these values are greater than the calculated minimum slenderness ratio, they only exceed the limit by a small margin. Thus, elastic buckling may not be the best model with which to evaluate the structural mechanics of baobab trees.
There are other models that could be considered to evaluate the structural safety factors of baobab stems. Some plant structures are comprised of a thin strengthening ring encircling a ring or core of parenchyma tissue; the Equisetum stem and the chive leaf are two structures of this composition (Niklas and O'Rourke, 1987
; Speck and Spatz, 2003
). Thin-walled hollow cylinders are subject to brazier buckling through ovalization of the stem cross section, and a core of relatively incompressible parenchyma tissue can afford some protection against brazier buckling. In baobab trees, the parenchymatous wood could similarly reinforce the thin shell of bark against local buckling. Alternatively, the baobab stem could be treated as a plant hydrostat, where the inner core of parenchyma tissue maintains a high water content and, through turgor pressure, stiffens the outer strengthening ring. These two models are not unrelated, and an analysis of baobab structure taking into account the hydrostatic nature of the parenchymatous wood and the strengthening nature of the thick, fibrous bark would certainly shed light on the biomechanical reliance on stem water and the degree to which baobab trees are overbuilt, if they are at all.
While the elastic buckling model may not be the best approach to consider in the case of the baobab, as described previously, its overall simplicity is appealing, and it is the only model that has been extensively studied and for which formulas have been developed. We will, therefore, tentatively discuss the results obtained from this model. Safety factors (Hactual/Hcrit) calculated for the 13 trees in this portion of the study ranged from 1.4 to 2.8, and there was no significant difference between the Greenhill model (Niklas and Spatz, 2004
) and the King and Loucks model (1978) when we considered the trees as solid wood cylinders of low elastic modulus nested inside thin-walled hollow cylinders of bark having a greater elastic modulus. In a study of a broad range of tree species in the United States, McMahon (1973)
found that the safety factor of large trees was, on average, equal to four, a value twice as great as the mean safety factor calculated here. The trees from McMahon's study may have had overly large safety factors, however, because many of the trees were open grown and among the largest and oldest representatives for each species. Several studies of forest-grown trees have obtained safety factors between one and three (King, 1981
; Holbrook and Putz, 1989
; Sterck and Bongers, 1998
). The results indicate, therefore, that baobab trees may be significantly less overbuilt than they would seem at first glance. The low elastic modulus of their wood is compensated for by the large diameter of their stems, such that they grow to comparable heights and have safety factors within the range of other trees that have been studied.
Stem construction costs
Another approach in considering baobab morphology is to examine whether, given their large diameter, baobabs stems are more costly to construct than stems of more typical trees. In other words, does the baobab tree invest more resources into its stem than a more typical tree of equal height? On a dry mass basis, baobab wood has a construction cost somewhat lower than most trees, but on a volumetric basis the average construction cost for the six baobab species is over five times lower than the average construction cost calculated from values reported for 12 North American tree species (Neenan and Steinbeck, 1979
; Smith and Tumey, 1982
). Using A. rubrostipa, we evaluated the energy costs of building a baobab stem vs. another tree stem. It would be preferable to compare baobab trees with coexisting dry forest species, or even other tropical trees, but as the required data are not available, we present this comparison as a first approximation. We use the average slenderness ratios for A. rubrostipa and large temperate dicotyledonous trees from North America (Niklas, 1992
), as well as the volumetric construction costs measured in this study and calculated from the literature (Neenan and Steinbeck, 1979
; Smith and Tumey, 1982
), and take the simplest form of a non-tapering columnar stem. Because the bark thickness is substantial and bark density is more than three times greater than wood density in A. rubrostipa, the estimated construction costs of the bark are also included in the calculation.
Surprisingly, the two cases result in almost identical stem construction costs (Table 5). The slenderness ratios for the temperate trees are perhaps slightly high, considering the trees were among the largest for their species, both in height and diameter, but the baobab trees selected for the allometric measurements were also large. The fact that agreement was so close suggests that the large stems of baobab trees may not represent a significant resource investment over other trees. The construction costs of a living structure do not take into account the maintenance costs of the cells that comprise it, however, and in the case of a stem comprised largely of parenchyma tissue, these could also be significant.
|
, b
) contradict the often-stated assumption, both in the popular and the scientific literature, that baobab trees rely heavily on the water stored in their stems to survive in arid climates. In this study, we have demonstrated that baobab trees are not more overbuilt than other trees, nor does the construction of their stem require a larger energy investment. We propose, therefore, that the large stem diameter of the baobab tree is necessary to prevent the stem from collapsing under its own mass and the mass of the extensive crown. Furthermore, the thick, fibrous bark appears to contribute significantly to structural support; the inclusion of bark in the formula increases the calculated buckling height by 6–19% for the trees in this study. The bark likely serves a protective function, but its role as a strengthening rind around the softer wood core could also be strategic from the perspective of stem water usage. Although the water release properties of the parenchyma cells in the wood are such that significant water withdrawal from the wood leads to turgor loss, with the associated negative consequences for cell water relations, baobab trees do withdraw 10–12% of their stem water during the leaf flushing period (Chapotin et al., 2006a
). Given the relationship between elastic modulus and water content, use of stored water should decrease the elastic modulus of the stem and could potentially affect its safety factor. The water content of the bark, however, remains constant throughout the year, and its contribution to stem support should not change. Small decreases in elastic modulus of the wood resulting from moderate stem water usage will therefore have little effect on the stem safety factor. Because the stem safety factors of the baobab are generally small compared with other trees, however, the withdrawal of a large amount of stem water could still affect mechanical stability, and there are reports of baobab trees succumbing mechanically to extended drought (Guy, 1971
; Wickens, 1983
). Baobab bark has an elastic modulus several times greater than baobab wood in the longitudinal direction, yet as a structural material is considerably less rigid than the wood from most species of trees. Because of the high tensile strength and netlike arrangement of its fibers, however, it should also be very strong in the circumferential direction, lending support to the idea that the baobab tree could be modeled as a hydrostat structure. If the tree were functioning as a hydrostat, or even as a dense rind with a parenchyma core, rather than an elastic column, then its structural rigidity could actually be much more susceptible to loss of turgor pressure through water withdrawal than as described previously.
The physical properties of the wood in baobab trees are developmentally plastic. The wood in young trees and small branches is denser and has a higher fiber content and lower parenchyma content than the wood in mature stems and large branches (Figs. 2, 3, 5, and 6). The density of the stem wood declines only until the tree reaches a certain size, at which point stem wood density remains generally constant with continued diameter expansion (Fig. 9, bottom panel). This pattern of producing the densest wood at the center of the stem is in direct contrast with that of balsa, a related species, and many other tropical tree species, in which the wood increases in density with distance from the pith (Wiemann and Williamson, 1989a
, b
; Rueda and Williamson, 1992
). In these trees, it is thought that stem construction of lower density wood allows trees to quickly reach the canopy with minimum energy expenditure in order to exploit canopy gaps. Baobab trees, on the other hand, grow up into the forest canopy as they produce their densest wood (though still less dense than many species). Height extension occurs before extensive stem thickening, and younger trees tend to have relatively small crowns with branches distributed evenly along the length of the main stem. Adult baobab trees shed these early branches and maintain a crown consisting of several large branches emerging from the top of a main stem otherwise devoid of branches. It would be interesting to learn whether the transition from the juvenile architecture to the mature architecture occurs at the same time that the stem begins to thicken and produce less dense wood and a higher proportion of parenchyma tissue relative to strengthening fibers, as is characteristic in the mature tree.
In this study, we have shown that in adult baobab trees large diameter stems are necessary for mechanical support and are not more costly to construct relative to stems in other trees. The adaptive significance of building a stem out of weak, low-density wood is not immediately apparent, particularly as extensive use of stored water does not appear possible under this strategy. Baobab trees, however, have very thick bark, a design feature that contributes significantly to overall structural stability of the stem and may compensate for the reductions in stem stiffness that would otherwise occur through moderate use of stem water. More detailed biomechanical and energetic analyses may demonstrate that baobab trees can actually achieve greater strength for a given energy investment than other trees, and the possibility that the maintenance of a large quantity of living parenchyma cells is somehow advantageous, whether for carbohydrate storage or recovery from traumatic injury, cannot be discounted.
Baobab trees are not unique in having large stems and branches constructed of low-density, water-filled wood. A large group of trees, commonly called stem-succulent trees and with representatives from a number of families, have converged on this morphology. Similarly deciduous and growing in warm, seasonally dry environments, stem succulents are generally described as water-storing trees (Franco-Vizcaíno et al., 1990
; Nilsen et al., 1990
; Borchert, 1994
). The results we have obtained for baobab trees may be more widely applicable to stem-succulent trees, and although water storage could still play an important role in some of these species, biomechanical considerations should be taken into account when attempting to attribute a functional role to their large, swollen stems.
FOOTNOTES
1 The authors thank L. Andriamahefarivo and other Missouri Botanical Garden staff in Antananarivo for research permits and logistical support and the German Primate Center (Deutches Primatenzentrum) for permission to work at the Kirindy Research Station. This study was made possible through funding by the Sinclair Kennedy Traveling Fellowship, the Arnold Arboretum and the Department of Organismic and Evolutionary Biology at Harvard University, the Garden Club of America, and the National Science Foundation. R. Rasoloarivony, H. Patt, and A. Toly provided field assistance. The authors also thank two anonymous reviewers for helpful suggestions and comments. 
4 Author for correspondence (chapotin{at}post.harvard.edu
)
LITERATURE CITED
Anonymous.. [n.d.]. Instructions for the 1341 plain jacket oxygen bomb calorimeter Parr Instrument, Moline, Illinois, USA.
Barbaroux C. Bréda N.. 2002. Contrasting distribution and seasonal dynamics of carbohydrate reserves in stem wood of adult ring-porous sessile oak and diffuse-porous beech trees. Tree Physiology 22: 1201-1210.[ISI][Medline]
Bauer G. Schulze E.-D. Mund M.. 1997. Nutrient contents and concentrations in relation to growth of Picea abies and Fagus sylvatica along a European transect. Tree Physiology 17: 777-786.[ISI][Medline]
Baum D. A.. 1995. A systematic revision of Adansonia (Bombacaceae). Annals of the Missouri Botanical Garden 82: 440-470.[CrossRef][ISI]
Borchert R.. 1994. Soil and stem water storage determine phenology and distribution of tropical dry forest trees. Ecology 75: 1437-1449.[CrossRef][ISI]
Bullock S. H.. 1992. Seasonal differences in nonstructural carbohydrates in two dioecious monsoon-climate trees. Biotropica 24: 140-145.[CrossRef][ISI]
Carlquist S.. 1998. Wood and bark anatomy of Caricaceae; correlations with systematics and habit. IAWA Journal 19: 191-206.[ISI]
Chapotin S. Razanameharizaka J. H. Holbrook N. M.. 2006a. Baobab trees (Adansonia) in Madagascar use stored water to flush new leaves but not to support stomatal opening prior to the rainy season. New Phytologist 169: 549-559.[CrossRef][ISI][Medline]
Chapotin S. Razanameharizaka J. H. Holbrook N. M.. 2006b. Water relations of baobab trees (Adansonia L.) during the rainy season: Does stem water buffer daily water deficits?. Plant, Cell & Environment 29: 1021-1032.
Dhillion S. S. Gustad G.. 2004. Local management practices influence the viability of the baobab (Adansonia digitata Linn.) in different land use types, Cinzana, Mali. Agriculture, Ecosystems and Environment 101: 85-103.[CrossRef]
Domec J.-C. Gartner B. L.. 2002. How do water transport and water storage differ in coniferous earlywood and latewood?. Journal of Experimental Botany 53: 2369-2379.
Feild T. S. Zwieniecki M. A. Donoghue M. J. Holbrook N. M.. 1998. Stomatal plugs of Drimys winteri (Winteraceae) protect leaves from mist but not drought. Proceedings of the National Academy of Sciences, USA 95: 14256-14259.
Franco-Vizcaíno E. Goldstein G. Ting I. P.. 1990. Comparative gas exchange of leaves and bark in three stem succulents of Baja California. American Journal of Botany 77: 1272-1278.[CrossRef][ISI]
Gartner B. L.. 1995. Patterns of xylem variation within a tree and their hydraulic and mechanical consequences. In B. L. Gartner [ed.] Plant stems: physiology and functional morphology 125-149 Academic Press, San Diego, California, USA.
Gower S. T. Frederick D. J. Clark A. I.. 1984. Distribution of energy in different-aged southeastern bottomland forests. Forest Ecology and Management 9: 127-146.[CrossRef][ISI]
Guy G. L.. 1971. The baobabs: Adansonia spp. (Bombacaceae). Journal of the Botanical Society of South Africa 57: 30-37.
Hoch G. Popp M. Körner C.. 2002. Altitudinal increase of mobile carbon pools in Pinus cembra suggests sink limitation of growth at the Swiss treeline. Oikos 98: 361-374.[CrossRef][ISI]
Holbrook N. M. Putz F. E.. 1989. Influence of neighbors on tree form: effects of lateral shade and prevention of sway on the allometry of Liquidambar styraciflua (sweet gum). American Journal of Botany 76: 1740-1749.[CrossRef][ISI]
Holbrook N. M. Putz F. E.. 1992. Secondary thickening in stems of Jacaratia (Caricaceae). American Journal of Botany 79: S34-S35 (abstract).
Kanehira R.. 1933. On light-weight woods. Journal of the Society of Forestry of Japan 15: 601-615.
King D. A.. 1981. Tree dimensions: maximizing the rate of height growth in dense stands. Oecologia 51: 351-356.[CrossRef][ISI]
King D. A. Loucks O. L.. 1978. The theory of tree bole and branch form. Radiation and Environmental Biophysics 15: 141-165.[CrossRef][ISI][Medline]
Kurth E.. 1946. The extraneous components of wood. In L. E. Wise [ed.] Wood chemistry 385-445 Reinhold, New York, New York, USA.
Laman T. G.. 1995. Safety recommendations for climbing rain forest trees with "single rope technique.". Biotropica 27: 406-409.[CrossRef][ISI]
Landhausser S. M. Lieffers V. J.. 2003. Seasonal changes in carbohydrate reserves in mature northern Populus tremuloides clones. Trees 17: 471-476.
Leith H.. 1975. Measurement of caloric values. In H. Lieth and R. Whittaker [eds.] Primary productivity of the biosphere 120-129 Springer-Verlag, New York, New York, USA.
McMahon T.. 1973. Size and shape in biology. Science 179: 1201-1204.
Molina-Freaner F. Tinoco-Ojanguren C. Niklas K. J.. 1998. Stem biomechanics of three columnar cacti from the Sonoran desert. American Journal of Botany 85: 1082-1090.[Abstract]
Mooney H. A. Chu C. Bullock S. H. Robichaux R. H.. 1992. Carbohydrate, water and nitrogen storage in vines of a tropical deciduous forest. Biotropica 24: 134-139.[CrossRef][ISI]
Neenan M. Steinbeck K.. 1979. Caloric values for young sprouts of hardwood species. Forest Science 25: 455-461.[ISI]
Newton L.. 1974. Is the baobab tree succulent?. Cactus and Succulents Journal 36: 57-58.
Niklas K. J.. 1988. Dependency of the tensile modulus on transverse dimensions, water potential, and cell number of pith parenchyma. American Journal of Botany 75: 1286-1292.[CrossRef][ISI]
Niklas K. J.. 1989. Mechanical behavior of plant tissues as inferred from the theory of pressurized cellular solids. American Journal of Botany 76: 929-937.[CrossRef][ISI]
Niklas K. J.. 1991. Effects of tissue volume and location on the mechanical consequences of dehydration of petioles. American Journal of Botany 78: 361-369.[CrossRef][ISI]
Niklas K. J.. 1992. Plant biomechanics: an engineering approach to plant form and function University of Chicago Press, Chicago, Illinois, USA.
Niklas K. J. Buchman S. L.. 1994. The allometry of saguaro height. American Journal of Botany 81: 1161-1168.[CrossRef][ISI]
Niklas K. J. O'Rourke T. D.. 1987. Flexural rigidity of chive and its response to water potential. American Journal of Botany 74: 1033-1044.[CrossRef][ISI]
Niklas K. J. Spatz H. C.. 2004. Growth and hydraulic (not mechanical) constraints govern the scaling of tree height and mass. Proceedings of the National Academy of Sciences, USA 101: 15661-15663.
