Chamaecyparis obtusa var. formosana and Chamaecyparis formosensis are congener gymnosperm tree species native to Taiwan cloud forests; occupying different niches there. While the seedlings of C. formosensis occur predominantly under bright conditions in large forest gaps, seedlings of C. obtusa var. formosana are mainly found below the canopy of mature forests or in small gaps. It is well known that congener species occupying different niches typically differ in several ecophysiological and morphological traits, but the differences in growth dynamics of such species are still totally unclear, as the diurnal growth dynamics of gymnosperm leaves have not been investigated before. Modern methods of digital image sequence processing were used in this study to analyse the leaf growth dynamics of the two species. We found that both species show similar base–tip gradients and pronounced diurnal growth rhythms with maxima in the evening. Differences between the two species concerning their growth dynamics correlated closely with their ecological amplitudes and abundances. Chamaecyparis obtusa var. formosana grew faster than C. formosensis in low light intensity, typical for closed-canopy situations, and reacted quickly by increasing or decreasing growth rate when light intensity changed within a range typically found below small canopy gaps. In contrast to this, C. formosensis grew better in light intensities typical for open vegetation situations, but reacted slower towards changes of light intensity. Based on those results, the hypothesis can be developed that fluctuations of leaf growth dynamics reflect heterogeneities of the light environment within the niche occupied by a given species.
Investigations of growth reactions of species are often restricted to observations of the occurrence of a given species in a given natural light climate or to measurements of classical growth parameters such as leaf area ratio or relative growth rate measured at intervals of several days in shading experiments (Dalling, Lovelock & Hubbell 1999). Dynamic growth parameters – such as principal features of the diurnal course or immediate responses towards changes of light intensity – have not yet been investigated in most ecological studies. Furthermore, the growth dynamics of gymnosperm leaves have not been investigated before due to their typically small size and slow growth rate. It is unknown whether they show diurnally oscillating growth intensity, in a similar way to angiosperm leaves (Walter & Schurr 2005). Novel measurement techniques (Schmundt et al. 1998; Walter, Feil & Schurr 2002a; Walter et al. 2002b; van der Weele et al. 2003) now allow us to link growth traits with physiological parameters, both analysed in temporal ranges of minutes to hours and spatial scales of suborgan levels (Walter & Schurr 2005). Yet, it is still an open question, whether fluctuations of light intensity are correlated with fluctuations of plant growth, and if so, whether plants have evolved growth traits adapting them to a given niche.
For this study, we selected two species from the gymnosperm genus Chamaecyparis that are adapted to different ecological niches in montane areas of Taiwan, where fog frequently occurs (Su 1984). Chamaecyparis obtusa var. formosana and C. formosensis (Lee 1962; Liu 1975) are dominant species in these forests. They are distributed in overlapping areas and have similar life history and life form. Yet, plants of C. obtusa var. formosana mostly regenerate from the seedling bank of the forest floor or in small canopy gaps, whereas plants of C. formosensis regenerate in large gaps and disturbed open areas, showing a more pioneer-species-like behaviour (Chang 1963; Liu 1975; Lin & Lai 1999). This difference was suggested to correlate with the intensity of irradiation received by seedlings (Chen 1968; Ho 1988) and it was hypothesized that the two species have diverged to partialize the occupied niches in order to eliminate heavy intrageneric competition.
The first aim of this study was hence to meet the experimental challenge of analysing the diurnal pattern of gymnosperm leaf growth variation. The second aim was to compare the leaf growth dynamics of both species in different light regimes to get an indication about physiological differences that might lead to the different ecological amplitudes of the two species.
MATERIALS AND METHODS
Seeds of C. obtusa var. formosana and C. formosensis were collected from an area in the cloud forest of north-east Taiwan. The seeds were sown in a nursery and the seedlings were well supplied with nutrients and water. The 3-year-old seedlings, about 25–40 cm high, were transplanted into 1.5 L pots. Thereafter, they were grown for several months in growth cabinets with a relative humidity of 60–70% and temperatures between 20 and 25 °C under extreme low, low, medium, high, and extreme high light regimes: 10, 30, 75–100, 150–300 and 400 µmol m−2 s−1 photosynthetic photon flux density (PPFD) with 12/12 h day/night period. These light regimes mimicked the specific conditions at the natural site, where light intensity is generally very low due to the prevailing clouds and accompanying fog. The PPFD of the extreme low light regime corresponded to the average light intensity on a forest floor covered by broad-leaved trees. The PPFD of the low light regime corresponded to the average diurnal light intensity in the understorey of a Chamaecyparis forest. The PPFD of the medium light regime corresponded to the average situation close to small gaps caused by tree fall or by the construction of logging roads. The high light regime mimicked the PPFD of a large disturbed area in the montane cloud zone in Taiwan, such as regeneration sites of landslides or forest fires. The extremely high light regime was comparable with the average diurnal value under totally clear conditions outside the cloud zone in the field (Lai, unpublished results). After 2 months, most of the plants under extreme low and extreme high light regimes looked unhealthy, thus we only used the healthy individuals that were grown under low, medium and high light regimes for this study. During the experiments of our study, all leaves that had partly developed before transplanting were fully emerged and all growing leaves were formed in the described light conditions.
Gas exchange measurements
The gas exchange rates of small seedling branches with totally 3–5 cm2 mature and healthy foliage were measured with a portable computerized open-system infrared gas analyser (Li-6400; Li-Cor, Inc., Lincoln, NE, USA). A cold light source (6400–02 red-blue LED) was mounted on the leaf chamber as the source of variable light. All measurements were made in the growth cabinet where the seedlings were raised with cuvette CO2 concentration set at 350 ppm and relative humidity above 60%. The foliage had been dark-incubated in the leaf chamber for at least 20 min before recording the data. CO2 assimilation rates were recorded at PPFD values of 0, 4, 8, 12, 16 and 20 µmol m−2 s−1. Then PPFD was switched to 100 µmol m−2 s−1. Measurement time was at least 20 min before light intensity was increased in order to assess the highest potential assimilation rate in each light condition. Gas exchange was recorded at PPFDs of 100, 200, 400, 600, 800, 1000, 1200, 1600, 2000 µmol m−2 s−1. Photosynthetic light response curves were produced from the gas exchange data of each seedling. At least three seedlings from each species were measured in each light regime. Apparent quantum yield (Qy), dark respiration rate (Rd), light compensation point (LCP), and maximum CO2 assimilation rate (Amax) were calculated from the photosynthetic light response curves.
Image acquisition and processing
Leaves of the second or third branch from the shoot apex of the seedlings were selected for the imaging growth measurements. Those leaves had emerged and developed in the chosen light regime. Apices of growing branches were mounted gently between two microscope slide glasses (Fig. 1a, b & c). Two near-infrared LED-arrays (wavelength 940 nm) were positioned next to the branch to ensure constant image brightness throughout night and day. Near-infrared images of the foliage at the branch apex were acquired every 5 min by a CCD camera (Sony XC-75; DBS; Bremen, Germany) from underneath the leaf. The distance between camera and foliage was about 6.5 mm (Fig. 1c); image size was 3.2 mm × 2.4 mm (Fig. 1d & e), and the CCD-resolution was 640 × 480 pixels. Using this magnification, the best results were obtained when in the first image of an image sequence, two to four layers of opposite leaves were visible. This was on the one hand enough foliage to allow reliable calculations of growth rates from velocities at different points within the foliage (see below) and on the other hand, there was still enough space between foliage apex and image borderline (at least one-third of the image area) for the foliage to grow within the chosen image frame. Typically, after 3 d, the foliage reached the border of the image. To acquire longer sequences, the camera was then shifted.
Within the black/white image sequence, natural grey value structures of the foliage are shifted with time because of the movement and growth of the leaves. To calculate distributions of relative growth rates (RGR) within the foliage, displacement vector fields for each image have to be calculated first. Any structure with a suitable grey value contrast to its neighbourhood results in oriented grey value structures in image stacks. The slope and orientation of those trajectories in the spatio-temporal neighbourhood of a central pixel were analysed via an optical flow approach (structure–tensor algorithm) and lead to a velocity vector for the motion of the central pixel. After determination of velocities for all suitable pixels in each image and within the entire sequence, missing information is filled in by interpolation for foliage regions showing too little grey value structure. Areal relative growth rates are now calculated at each pixel by taking the divergence of the velocities of neighbouring pixels (for detailed information see Schmundt et al. 1998; Walter et al. 2002a, b). Results can be averaged spatially and temporally on demand and are displayed as colour-coded maps or as time-series of selected areas of interest (AOI; see Figs 3b, d & 4a). Sub-algorithms allow the choice of an AOI in the beginning of the sequence and to extract time-series for relative growth rates for the foliage part of the initial AOI as borderlines are growing along with the moving and expanding organ (Fig. 3; Scharr 2004). For most purposes of this study, 1-h average values of RGR were calculated. For representation of ‘normalized’ growth data, 1-h average values of RGR were normalized with respect to the average 24-h value of RGR (average RGR = 100%). For statistical analyses, the results of at least three plants per species and per light treatment were used.
Transfer of plants between low and medium light regimes
After growth conditions had been established and the basic features of the diurnal growth pattern had been investigated, we recorded the change of RGR time-series, when the seedlings were transferred from the light condition to which they were acclimated to a different one. Two seedlings of each species which had been grown for at least 8 weeks in low and medium light regimes, respectively, were transferred to the opposite light regime (low to medium and medium to low). RGR was monitored for 6 d before and after the transfer. The hourly RGR of the two replicates was averaged to represent the response of the species to a sudden change in light regime.
All parameters measured in the gas exchange experiment and hourly averaged RGR time series were examined with two-way analysis of variance (anova) with ‘species’ and ‘light regime’ as fixed factors (STATISTICA version 6.0, StatSoft, Inc., Tulsa, OK, USA). The difference of means was compared using Fisher's Least Significant Difference (LSD) test.
Basic characteristics of diurnal leaf growth pattern
To establish optimal laboratory growth conditions for the seedlings, some characteristic parameters of their photosynthesis were used in addition to the observations concerning their overall growth performance at different light regimes. Although LCP increased with light intensity in both species, highest Rd, Amax and Qy were found in both species when they were grown in the medium light regime (Fig. 2). Gas exchange of both species was inhibited in the ‘high’ light condition applied in this study. There were significant differences for each parameter between light treatments (P < 0.05), whereas interspecific comparisons showed relatively similar results with the exception of Amax in the medium light regime: There, C. formosensis had a markedly higher Amax, indicating that this species acclimates much better to changes in light intensity.
In optimal growth conditions (medium light regime), strong and highly repetitive diurnal RGR patterns were found for leaves of both species (Fig. 3). In both species, clear spikes of negative expansion were found daily. There was no immediate difference between the growth dynamics of both species. Analysis of spatial leaf growth distribution within single leaves and within the foliage region used for most figures of this paper showed a high degree of patchiness, but again no clear difference was found between both species (Fig. 4). Both species showed the highest RGR at the day–night transition, and had a relatively stable RGR during the last 6 h of the night (Figs 3, 4b & c). Clear base–tip gradients were present throughout the entire diurnal course with highest growth rates found at the leaf base (Fig. 4b & d). When negative expansion occurred transiently, the base–tip gradient changed its sign and highest negative expansion rates occurred at the leaf base (Fig. 4b).
Diurnal leaf growth patterns in different light regimes
The comparison between relative leaf growth rates in three different light intensities showed for both species, that growth was strongly retarded in high and low light conditions (Fig. 5). Optimal growth rate was found under medium light conditions. This corresponded to the results of gas exchange analysis. Growth rate increased almost monotonically throughout the day and decreased almost monotonically throughout the night. Negative expansion occurred in both species at the night/day transition for every light regime and was most obvious in the medium light regime (Fig. 5). Chamaecyparis obtusa var. formosana reached higher RGR-values than C. formosensis in the low light regime (peak values of 0.4 versus 0.2% h−1) whereas C. formosensis reached higher RGR values than C. obtusa var. formosana for medium and high light conditions (peak values of 2.0 versus 1.5% h−1 and of 0.6 versus 0.3% h−1 in medium and high light intensity, respectively).
To compare the growth dynamics of both species at different light regimes, 1-h average RGR values were normalized (Fig. 6). Both species had an almost identical diurnal growth rate cycle in low light conditions. In medium and high light conditions, C. obtusa var. formosana showed a much more ‘disturbed’ growth rate cycle than C. formosensis with values extending into the negative region for a longer time and with greater relative intensity. In high light conditions, leaves of C. obtusa var. formosana shrank throughout the morning hours (0600–1100 h). In contrast to the behaviour of C. obtusa var. formosana, the diurnal growth rate patterns of C. formosensis showed an almost identical shape for all three light regimes.
Differences in general growth behaviour were most apparent when RGR was averaged throughout the entire day (n = 30 or more, Fig. 7). Chamaecyparis obtusa var. formosana grew at a higher rate than C. formosensis in the low light regime (P < 0.001), whereas C. formosensis showed higher growth rates than C. obtusa var. formosana in the medium light regime (P < 0.001) and also in the high light regime, although the difference was not significant (P = 0.288). Comparison of the average growth rates between light intensities showed that growth rates in the medium light intensity were at least a factor of two higher than in low or high light conditions for both species. For C. obtusa var. formosana, mean growth rates did not differ between low and high light intensity, whereas C. formosensis showed significantly higher growth rates in the high light regime compared to the low light regime (P < 0.001).
Transfer of plants between light regimes
Foliage growth was followed throughout 12 d in this experiment with plants growing for the first 6 d in conditions to which they were acclimated and for the last 6 d in altered light conditions (Fig. 8; n = 2). This experiment revealed another difference between the diurnal growth behaviour of the two species: C. obtusa var. formosana showed normal, co-ordinated growth rhythms in both light conditions with a much lower amplitude in low light. When light conditions were switched, the amplitude of the plant formerly in medium light conditions decreased within 1 d, but the growth rhythm remained. The plants that were transferred to medium light conditions increased their growth amplitude within 24 h.
Chamaecyparis formosensis leaves grown in low light conditions showed very small amplitudes of diurnal growth fluctuations, and medium light conditions resulted in a normal diurnal cycle. Upon transfer of the plants to the other condition, the plants from medium light conditions showed fading growth intensity throughout the entire 6 d, reaching values comparable to those that were reached by C. obtusa var. formosana at day 2, only at day 6. The plants that were switched to medium light conditions showed continuously increasing growth intensity and also needed 6 d to reach comparable growth cycles as C. obtusa var. formosana at day 2 after switching.
The results of this study show that the investigated congener species differ in leaf growth dynamics and that those differences in their growth kinetics reflect the most important features of their natural light environment. The species adapted to grow under closed canopies (C. obtusa var. formosana) is much more dynamic (Fig. 8) and grows more successful in low light conditions (Figs 5 & 7). In constantly high light, its leaves shrink throughout an extended period of the day (Figs 5 & 6). Hence, this species seems to be adapted to highly fluctuating but overall shady conditions, as is the case in the understorey of a forest. In contrast to this, the species adapted to open vegetation situations, C. formosensis, has lower growth rates in low light conditions (Figs 5 & 7), and reacts much slower towards changes of light intensity (Fig. 8). Chamaecyparis formosensis seems to tolerate higher light conditions better than C. obtusa var. formosana, as can be concluded from the finding that it does not show negative growth rates under high light conditions. The sudden onset of artificial light in the morning will surely cause changes in leaf water relations, such as stomatal conductance, water potential, and turgor. The kinetics with which those plant traits can recover from suddenly imposed, stressful conditions might be closely related to the kinetics of sudden leaf growth responses and hence to the overall growth performance of a plant. Growth rates of C. formosensis were also higher in the medium light regime in comparison to C. obtusa var. formosana. Hence, C. formosensis is better adapted to an open vegetation situation with overall moderate light intensities, which corresponds to the distribution of C. formosensis seedlings in their natural site.
Photosynthetic strategies also adapt C. formosensis to the moderate light intensities in cloud forests (Fig. 2). Amax, Rd and Qy are inhibited in the high light conditions applied in this study; leaves reach their optimal physiological state at moderate PPFD-conditions. The fact that – in the medium light regime –C. formosensis has a higher Amax compared with C. obtusa var. formosana might indicate that C. formosensis can build up more reserves from photosynthesis, enabling a slower decay of growth rate after transition to a lower light regime. Although the applied light regimes with a constant illumination throughout 12 h were different from natural light conditions, they mimicked average light intensities at typical growing sites and the results confirmed findings from nursery management. There, it is well known that planted seedlings need a generally shady environment (Lin & Lin 2000). Vegetation surveys show clearly that seedlings of C. formosensis are predominantly found in open gaps or severely disturbed areas (Chang 1963; Liu 1975; Lin & Lai 1999). There, the prevailing fog will typically lead to moderate PPFD and so the optimal growth conditions for this species will most probably be on foggy and not on clear days as has been expected previously.
In summary, the present findings indicate that a detailed analysis of leaf growth dynamics can be a useful tool to verify the ecological amplitude of a given species in much the same way as this can be done via measurements of photosynthetic acclimation of the investigated species (Murchie & Horton 1998). However, causal explanations for the ecological amplitude of a given species cannot be expected from analysis of a single trait alone, since the capacity to adjust to a given situation is always connected to a wide range of traits. Moreover, it has to be noted that the results obtained here for two congener gymnosperm species do not allow general conclusions for other species. Yet, those results might be a first step towards the extraction of ‘growth acclimation strategies’ that are similar to photosynthetic acclimation strategies that were elucidated for a wider range of angiosperms (Murchie & Horton 1997).
The basic rhythm of diurnal growth activity does not differ between the two species investigated here, indicating that the same mechanisms govern the growth dynamics in both species. However, the diurnal rhythmicity reported here shows antiphasic behaviour in comparison with the results from leaf growth in Ricinus communis (Walter, Feil & Schurr 2002a) or Nicotiana tabacum (Walter & Schurr 2005) that show maximal growth intensity in the late night. This supports the view that a network of mechanisms is responsible for the regulation of growth intensity throughout 24 h, with different weights of certain mechanisms in different species.
Biophysically, growth is regulated by the interaction between internal plant pressure (turgor) as the driving force and the rigidity (or extensibility) of the cell wall as the retarding force (Lockhart 1965; Cosgrove 1986). The controlling mechanisms have to alter these biophysical properties in a diurnally regulated manner. Carbohydrate metabolism is one of those mechanisms affecting diurnal growth patterns (Kehr et al. 1998; Walter & Schurr 2005). Other mechanisms include changes in the activity of cell-wall loosening factors such as expansin and xyloglucan-endo-transglycosylase (Cosgrove 1999). Moreover, sudden changes in light intensity do exert fast effects on, for example, cell wall pH (Mühling et al. 1995) and thus may affect cell wall extensibility. They also affect stomatal conductance (Mott & Buckley 2000) and may thus influence turgor. Yet, the connection between all these elements of the network regulating diurnal growth variations is still unclear. This study shows that comparison of growth patterns between congener species occupying different ecological niches is another means of elucidating connections within the regulatory network as differences in growth dynamics are connected to differences in physiology of these species. In future studies, it will be important to clarify at the level of metabolites, enzymes and gene activities, why shade-tolerant species can make faster use of their growth potential in dynamically changing light conditions, why they are shrinking in the early morning hours at high light and why shade-avoiding species adapted to open vegetation situations can use more of their growth potential in constant situations of moderate PPFD.
The differences observed in growth dynamics of the two species correlate with the ecological amplitude in light intensity that each species experiences under natural conditions. This leads to the conclusion that a detailed analysis of growth dynamics is a helpful tool for investigating physiological differences between species and that it can also serve as a helpful tool in predicting plant competitiveness under given environmental situations by clarifying the variance of the ecological amplitude of a given species.
We are grateful for the financial support by a grant to I-L.L. from the international co-operation ‘Sandwich Program’ for Ph.D. candidates by German Academic Exchange Service (DAAD) and the National Science Council of Taiwan (NSC). We acknowledge provision of seedlings by the ‘Forest Protection Department under the Veterans Affairs Commission of the Executive Yuan of the Republic of China’. We thank Lisa Ainsworth and Shizue Matsubara for helpful comments on the manuscript, Michael Schilbach and Beate Uhlig for help with plant cultivation, and Maja Christ for assistance with the imaging software.