Research was conducted in the lowland tropical moist-forest of La Chonta, Bolivia. Annual precipitation is 1580 mm year−1
with a dry season (<100 mm month−1
) from April until October. The forest is semi-evergreen, has a canopy height of 27 m, a basal area of 19·3 m2
and a stem density of 367 ha−1
(stems >10 cm diameter at breast height, DBH; Peña-Claros et al., 2008
). In this forest the Instituto Boliviano de Investigación Forestal (IBIF) administers 320 ha of permanent sample plots in which trees >10 cm DBH are monitored in a nested design.
Fifty-eight of the most abundant species that differed in adult stature and light requirements were selected (Table ). Adult stature (Hmax
) is defined as the potential height that a species can attain when it attains its maximal diameter. For each species this potential height was calculated using species-specific allometric relationships between height and diameter (Poorter et al., 2006
) and the diameter of the third-thickest tree in the permanent sample plots. The third-thickest rather than the maximal tree diameter was used, thus avoiding outliers. Species were subjectively classified into shade-tolerants that can establish and survive in the shade, and pioneers that need high light as found in gaps for successful survival (Poorter et al., 2006
). This two-group classification is merely used to give the reader a first feeling as to which strategy a species adopts; for the formal statistical analysis an objective and continuous measure of the regeneration light requirements of the species is used. In a separate study the regeneration light requirements were determined by analysing for each species the crown exposure in relation to the height of individual trees (Poorter et al., 2006
). An average number of 612 individuals (median 283, range 41–9319) per species were measured over their whole size range for their height and crown exposure (CE; Dawkins and Field, 1978
). For eight of the small (<8 m) or rare species, fewer than 100 individuals were used. The crown exposure is categorized as 1 if a tree does not receive any direct light, 2 if it receives lateral light, 3 if it receives overhead light on part of the crown, 4 if it receives full overhead light on the whole crown, and 5 if it has an emergent crown that receives light from all directions. The crown exposure can be measured repeatably (average difference between two independent observers is 0·1 ± 0·01 s.e.), and there is a good relation between CE and both canopy openness and incident radiation (Davies et al., 1998
; Clark et al., 1993
; Keeling and Phillips, 2007
). For each species the CE was related to tree height using a multinomial regression analysis (Poorter et al., 2006
; cf. Poorter et al., 2005
). Using the regression equation, the average population-level crown exposure at a standardized height of 2 m (juvenile crown exposure: CEjuv
) was calculated. In principle, 2-m tall saplings can be found under a wide range of crown exposures, but the value of CEjuv
indicates the average CE at the population level for this plant size and is a good indicator of the regeneration light requirements (i.e. the inverse of the shade tolerance) of the species.
Overview of the species included in the study, their juvenile crown exposure (CEjuv), maximum adult stature (Hmax), and stem volumetric fractions of wood, gas and water for saplings and adults
Hmax and CEjuv are positively correlated (r = 0·38, P < 0·05, n = 58), indicating that taller species are also more light demanding in the sapling stage. Yet, Hmax explains only 14 % of the variation in light requirements, which suggests that there is ample opportunity for independent effects of Hmax and CEjuv on stem traits.
Adult wood characteristics were determined for 58 species and sapling wood characteristics were determined for a subsample of 30 species (van Gelder et al., 2006
). All samples were taken in the wet season. For adult wood, samples (approx. 2 × 2 × 2 cm) were taken from three trees per species, just interior to the cambium, and contained mostly sapwood. The trees were between 20 and 50 cm diameter for the tall species, and close to their maximal sizes for small species. Samples were stored for a week in plastic bags in a refrigerator. Fresh mass was determined with a balance, and fresh wood volume was measured using the water displacement method, after which the wood was oven-dried for at least 48 h at 70 ºC and then weighed. For saplings, the terminal meter of stem was sampled for ten individuals per species. This stem sample included the bark, because the intention was to relate the properties of the whole stem to sapling growth and survival (see below). It is acknowledged that wood and bark have different material properties, and hence different relationships between density and strength, but given the small amount of bark and the large interspecific differences in wood properties, such a whole-stem measurement gives a good estimate of the stem strength and rigidity of the species (van Gelder et al., 2006
Saplings were on average 2·9 m tall (range 1·8–4). The green volume was calculated based on the length and diameters at the beginning and end of the sample, assuming the shape of a truncated cone. Fresh mass was measured on the day of collection, and dry mass after oven-drying for at least 48 h at 70 ºC. It should be noted that a direct comparison between sapling and adult stems is somewhat confounded by the difference in sampling methodology, as for the adults it strictly refers to the xylem, whereas for saplings it refers to the whole stem (including xylem, phloem and outer bark).
The volumetric percentages of wood, gas and water in the sample were calculated following Gartner et al. (2004)
. The percentage of stem volume in wood material was calculated as 100 × (dry mass/fresh volume) divided by the wood density of pure wood material (1·53 g cm−3
; Kellog and Wangaard, 1969
). The percentage of cell volume occupied by water was calculated as 100 × (fresh mass – dry mass)/cell volume. The percentage of cell volume of gas was calculated as (100 – percentage wood material – percentage water). The three stem volume components might be negatively correlated, simply because of the fact that they should add up to 100 %. The stem volume of wood sets an upper limit to the volume that can be occupied by the fillers water and gas. To separate out the role of gas versus water, the gas fraction of the filler was also calculated: 100 × gas percentage/(gas percentage + water percentage).
For the analysis it was presumed that all dry matter was in wood. However, some species also contain latex and resins, thereby potentially over-estimating the percentage wood and under-estimating the percentage water and gas. For the saplings there were nine species that produced latex or resins, and for the adults there were 11 species. Their wood fraction was not significantly higher than non-latex/resin-producing species (t-test: P = 0·29 for saplings and P = 0·60 for adults), suggesting that it did not bias the results to a significant extent.
For each species approx. 16 additional saplings between 0·5 and 2 m tall were sampled in the forest understorey, or under as shaded conditions as possible. Height and survival were monitored several times a year for a 2-year period: see Poorter and Bongers (2006)
for a description of the calculation of annual height growth and survival rates. It is acknowledged that the survival rates provide a first estimation of the population-level survival rates of the species, especially for the species with very low inherent mortality rates, as mortality is a stochastic process and the sample size is relatively small.
Wood characteristics were related to CEjuv
, and sapling performance [log10
(height growth), survival] using Pearson correlations. Height growth and survival rates were related to the percentage wood material and gas fraction of the filler using a forward multiple regression. All statistical analyses were carried out using SPSS 12 (SPSS Inc. Chicago, IL, USA). Ontogenetic consistency in characteristics of sapling and adult wood was evaluated with standard major axis regression using SMATR (Falster et al., 2006