Original article

Variation in leaf morphology and branching pattern of some tropical rain forest species from Guadeloupe (French ) under semi-controlled light conditions

M Ducrey

INRA, Laboratoire de Recherches Forestières Méditerranéennes, Avenue A Vivaldi, F-84000 Avignon, France

(Received 18 March 1992; accepted 7 July 1992)

Summary — Seedlings of 7 canopy species from the Guadeloupe tropical rain forest (Dacryodes excelsa, Amanoa caribaea, grandis, Simaruba amara, Symphonia globulifera, Byrsonima coriacea and Podocarpus coriaceus) were raised in full sunlight and under artifical neutral shade transmitting 6, 11, 19 and 54% light for 2 to 3 years. At the end of this period, the number of leaves and branches, leaf size, specific leaf area and stomatal density were observed for each . For all species, the maximum number of leaves was obtained in partial shade (11 or 19% sunlight). Branch- ing occurrence depended more on species type than on light conditions. Both individual leaf size and specific leaf area increased regularly with shade, but in a proportion which varied according to the species. Stomatal density was highly variable from one species to another and increased with greater light. The morphological plasticity of species response to light conditions was then analysed and related to shade tolerance. In order of decreasing plasticity, the first species found were R gran- dis, S amara and B coriacea, which were the most plastic and the most shade intolerant, followed by A caribaea and P coriaceus, less plastic but shade-tolerant species. Finally, D excelsa and S globu- lifera were found to be the least plastic species and highly or moderately shade-tolerant. tropical rain forest / leaf morphology / specific leaf area / branching pattern / shade tolerance

Résumé — Variations de la morphologie foliaire et branchaison de quelques espèces de la forêt tropicale humide de Guadeloupe en conditions semi-contrôlées d’éclairement. De jeunes semis de 7 espèces de la strate arborescente de la forêt tropicale humide de Guadeloupe (Da- cryodes excelsa, Amanoa caribaea, , Simaruba amara, Symphonia globulifera, Byr- sonima coriacea et Podocarpus coriaceus) ont été élevés pendant 2-3 ans en pleine lumière et sous ombrages artificiels neutres laissant passer 6%, 11%, 19% et 54% de la pleine lumière. À la fin de cette période on a observé sur chaque plant, le nombre de feuilles et de ramifications, la taille et la surface spécifique des feuilles ainsi que la densité stomatique. Pour toutes les espèces étudiées, le nombre de feuilles est maximal pour des ombrages moyens (11 ou 19% de la pleine lumière). La présence de ramifications dépend davantage des espèces que des conditions d’éclairement. La sur- face individuelle des feuilles ainsi que leur surface spécifique augmentent régulièrement avec l’om- brage mais dans des proportions variables selon les espèces. La densité stomatique, très variable d’une espèce à l’autre, augmente avec l’éclairement. La plasticité morphologique des espèces en ré- ponse aux conditions d’éclairement est ensuite analysée et interprétée en termes de tolérance à l’om- brage. Par ordre de plasticité décroissante, on trouve R grandis, S amara et B coriacea qui sont les espèces les plus plastiques et les plus intolérantes à l’ombrage. On trouve ensuite A caribaea et P coriaceus, moins plastiques mais tolérantes à l’ombrage. D excelsa, et S globulifera sont les moins plastiques et sont modérément ou fortement tolérantes à l’ombrage. forêt tropicale humide / morphologie foliaire / surface foliaire spécifique / blanchaison / tolé- rance à l’ombrage

INTRODUCTION studied under 2 different thinning intensi- ties. The variations in environmental condi- tions due to the different silvicultural treat- The reaction of trees to envi- varying light ments were then used as a means of ronments, particularly to shade, can be determining the range of requirements at different levels. First of at light compared all, in the species studied, from the most the we find which species level, species shade-intolerant to the most shade-tolerant. require full sunlight and others which are A silvicultural is not more or less shade-tolerant. On the indi- uniquely approach sufficient to understand the forest behavi- vidual level, within the same species or our of a and its relative genotype, we find trees which have grown given species in different light environments and have place in a forest succession. It therefore different phenotypes (shade phenotypes seemed of interest to further the know- on these mor- or sun phenotypes). Finally, within the ledge species by studying variations in leaves and branch- same individual, particularly within a stand, phological in response to light conditions sun and shade leaves are found, depend- ing pattern This is of value ing on their position in the tree crown. during growth. approach for 2 reasons. First of all, the use of mor- These facts are generally known for phological criteria to account for physiolog- most tree species growing in temperate cli- ical potentials under varying light condi- but have been less studied for mates, trop- tions appears to be possible using existing ical species. In particular, the shade re- relationships between physiological and of the main commercial in sponse species morphogenetic processes (Tsel’Niker, the rainforest of is tropical Guadeloupe 1977). Secondly, the range of morphologi- unknown. practically cal variations in the leaf system under ex- The experiments conducted (Ducrey, treme light conditions is a good means of 1982; Ducrey and Labbé, 1985) on stimu- determining the forest behaviour of a given lated and controlled natural regeneration species (Smith, 1982; Fetcher et al, 1983; in the Guadeloupe rainforest provided the Goulet and Bellefleur, 1986). first results (Ducrey and Labbé, 1986) on This article examines the morphological the forest behaviour of the main tree spe- variations in leaves and branching pattern cies favoured for natural regeneration. for 7 evergreen species subjected to 5 dif- Methods similar to the progressive felling ferent light conditions. The experiment also regeneration and the tropical shelterwood took into account photosynthetic response, system were adopted. Survival and growth growth and biomass production, which will of seedlings from different species were be discussed in further papers. MATERIALS AND METHODS full sunlight. The 4 tunnel shelters were 15 m long and 6 m wide and covered with reinforced transparent PVC as a protection against rainfall. Three of them were shaded with different black of Description seedlings neutral shade screens in order to obtain various of species studied shade conditions. Finally, global radiation meas- urements with Li-Cor pyranometers indicated 6.4% light under tunnel I, 11.4% under tunnel II, The used for the were seedlings experiment 18.8% under tunnel III and 54.3% under tunnel sampled from the tropical rainforest of Guade- IV. loupe, French West Indies. They came from the Table I summarizes climatic data under tun- "Débauchée" area (Ducrey, 1986) at an eleva- nel shelters. These were and oriented in tion of 250 m. Mean temperatures were 23 °C in opened January and 26 °C in July. Mean annual rainfall the direction of prevailing winds. The microcli- matic conditions under the tunnels were the was > 3 000 mm. There was a short dry season from January to April, but the monthly rainfall same as those in the open air treatment (meteo- data measured a weather was always > 100 mm. rological by station), except for tunnel IV whose maximum tempera- The 7 studied were domi- species evergreen tures were slightly higher than the others. This nant and co-dominant trees from the middle and could be explained, as the shade under this tun- late successional of the gradient Guadeloupe nel was only created by the reinforced transpar- rainforest: excelsa Vahl, Amanoa ca- Dacryodes ent plastic cover which caused a more signifi- ribaea Kr et Urb and Podocarpus coriaceus LC cant warming effect. Rich are late successional shade-tolerant spe- The was to all the cies; Simaruba amara Aubl and Richeria grandis protocol applied species P coriaceus and A caribaea. The P coria- Vahl are middle successional shade-intolerant except species; Byrsonima coriacea is present in mid- ceus seedlings were placed under the same dle and late succession, whereas Symphonia moderately shaded tunnel (tunnel III) in March 1981 and then to the different globulifera L, a wet soil specialist, is a late suc- subjected experi- mental conditions in 1982. The cessional species. However, their shade reac- January experi- tion is not well known. ment with A caribaea started in March 1982. In each were D excelsa and S amara have compound tunnel, grouped by spe- cies with a container of 16 m2. leaves, while the other species have simple density plants per All the were moved a in- leaves. All could be easily identified in the forest plant groups once week side each tunnel so that same understorey with the exception of B coriacea, they occupied the every 8 weeks. This was undertaken to which was difficult to differentiate when young place uniformize from 2 neighbouring forms, the "Patagonian" growth light conditions. At the begin- of the were Byrsonima and the "Coal wood" Byrsonima. ning shading experiment, there be- tween 30 and 40 plants per species and per treatment. The number of plants remaining at the end of the experiment is given in tables II Experimental treatments and III. Containers were watered twice a week. No fertilizer was used during the experiment.

The 1-yr-old seedlings were sampled from the for- est margin in January 1981, transplanted in 9-I containers filled with surface forest soil, and Plant observations and measurements placed under the forest canopy to ensure better After 3 were recovery. months, the containers At the end of the experiment (between March transferred to tunnel shelters covered with shade 1983 and January 1984 depending on the spe- cloths to obtain the required amount of shade. cies) when the plants were approximately 1.00- Seedlings were then between 10 and 20 cm 1.50 m in height, several observations were height. made: counting leaves on the main stem and on The seedlings were separated into 5 different branches, dry weight and surface area of 2 ran- treatment groups: 4 treatments under plastic domly selected leaves from the stem and 2 tunnels and one treatment in the open air and leaves from the branches on each plant. The

RESULTS data were used to calculate the specific leaf area (cm2 g-1) of each species for each light condition. The leaf stomatal density (number of stomata Leaf counting per leaf area unit) was determined during the last quarter of 1982 via leaf prints. A thin collod- Table II summarizes data concerning the ion film was spread on the leaf surface to pre- mean number of leaves per seedling for pare a print of epidermic and stomatal cells that The mean number could be observed by optical microscopy. These simple-leaved species. of leaf prints were taken for 2-6 leaves per species leaves varied from one species to an- and per tunnel and were made systematically on other: 22 on average for R grandis, 54 for the lower and upper side of the leaves. B coriacea, 95 for A caribaea, 140 for S globulifera and 317 for P coriaceus. For Study of branching pattern each species, the maximum number of leaves was observed either in tunnel II or All the seedlings studied were very young. III and some statistical differences might It was thus interesting to note the appear- have occurred among tunnels. The distri- ance of branches and their variations un- bution of leaves on the main axis or on the der different light conditions (tables II, III). branches was related to the percentages of branched seedlings and to the number The compound-leaved species D excel- of branches per branched seedling. sa and S amara had no branches. These R grandis leaves were almost entirely situ- only appeared under natural forest condi- ated on the main axis while those of A ca- tions in larger and older trees. ribaea, B coriacea and P coriaceus were The simple-leaved species had different mainly located on the branches. degrees of branching. R grandis had only Table III provides the same information just begun to ramify and had very few for compound-leaved species. D excelsa branches. All the S globulifera seedlings had an average of 11 leaves per plant, but were highly branched and had between 15 the number of leaflets per leaf increased and 17 branches per seedling. The other with increasing shade from 3 in the open species also had a high percentage of air to 5 in the darkest tunnel. S amara had branched plants, often close to 100%. This between 5-10 leaves. It would appear that percentage was maximum under low light the number of leaflets per leaf increased conditions for A caribaea and P coriaceus with exposure to shade, but the repeated and under sunlight conditions for B coria- attacks of phyllophagous caterpillars typi- cea. However, it appeared that branching cal of this species made the results difficult occurrence was more species-dependent to interpret. than light regime-dependent. Leaf characteristics leaved species) for each species in rela- tion to relative light intensity which they re- Figure 1 indicates the variations in area of ceived during growth. First of all, there was individual leaves or leaflets (for compound- a high variability in leaf size from one spe- cies to another. Taking all the tunnels to- was close to 50 cm2 g-1. S amara was the gether, the average leaf areas increased most affected by increasing shade: 149% from 10 cm2 for P coriaceus to nearly 200 increase in specific leaf area when going cm2 for R grandis. from full sunlight to shadiest tunnel. It was followed R B coriacea and P co- There was also a regular decrease in by grandis, riaceus with = 100% then D ex- leaf area for all species when relative light increase, by increased. Some species reacted strongly celsa and A caribaea with = 75% increase, to shade and the area of individual leaves and finally by S globulifera which had < 50% increase. As mentioned for in- more than doubled when going from full already dividual leaf an decrease sunlight to 6% sunlight. This was the case area, exponential for R grandis (150% increase), B coriacea in specific leaf area in relation to relative (120% increase) and S amara (100% in- light intensity was found except for A cari- crease), followed by A caribaea (65% in- baea, R grandis and S amara which were less affected crease), D excelsa, S globulifera and P co- by deep shading. riaceus (50% increase for each species) which reacted less strongly to variations in light conditions. The right side of figure 1 Stomatal density shows that for most species there was a quasi-linear decrease in individual leaf The leaf prints showed that for all the stud- area in relation to the logarithm of relative ied species, stomata were present only on light intensity. This demonstrated an expo- the lower side of the leaves. The stomata nential variation in relation to relative light as well as the epidermic cells had a large intensity, a relationship which has fre- variety of forms and sizes, as shown in fig- quently been found for similar phenomena. ure 3. This variability was demonstrated by means of stomatal Specific leaf area (leaf area recorded by comparisons density unit of dry leaf biomass) is shown in figure (number of stomata per mm2) for each 2. Leaves of all species in full sunlight had species in each light treatment (table IV). a specific area close to 100 cm2 g-1 except Stomatal density for full sunlight condi- for P coriaceus, whose leaves were thicker tions showed the highest values for D ex- and tougher and whose specific leaf area celsa (661 stomata per mm2) and A cari- baea (325 stomata per mm2). The 5 other Stomatal density was highest under full species had a stomatal density close to sunlight conditions and decreased as light 150 stomata per mm2. intensity diminished. All the species did not react in the same manner. R grandis was B coriacea (58%), A caribaea (43%), P co- the most affected species with 67% de- riaceus (38%), and D excelsa (35%). In creased from full sunlight to the shadiest contrast, S globulifera, with only 3% de- environment. The decrease in stomatal crease, did not appear to be affected by density was smaller for S amara (59%), shading. Morphological plasticity Some authors (Logan and Krotkov, 1969; Loach, 1970) found that with temper- ate species the number of leaves reached Species plasticity for a given trait - leaf a maximum in full sunlight. In many decid- size, specific leaf area or stomatal density uous leaved the number of leaves - species, may be calculated as the range of this is set from budbreak, while in trait from full sunlight to the shadiest condi- evergreen is more or less tion, divided data under tropical species growth by corresponding continuous and the number of leaves full sunlight conditions. For each trait, spe- present at a given moment is more highly cies was calculated, and plasticity species related to environmental conditions. ranked from the most plastic to the least plastic species (table V). Then a mean According to Smith (1982), the branch- ranking was calculated which gave an ing ability could be considered as a criteri- overall appreciation of the morphological on for adaptation to shade. This hypothe- plasticity for each species. sis agrees with that of Bazzaz and Pickett who found that trees in the first Ranked by decreasing order of plastici- (1980), successional stages ramify little and have ty, it was found that R grandis, S amara weak branches. Such a lack of branches and B coriacea were the most plastic spe- was observed from cies, A caribaea and P coriaceus the medi- young-aged pioneer in um-plastic species, and finally D excelsa species present Guadeloupe: Cecropia and to a lesser Miconia mi- and S the least peltata degree globulifera plastic species. rabilis. For the species studied, a lack of DISCUSSION AND CONCLUSION branches was also the case for R grandis and S amara which appear in the middle successional stages of species during col- differences were observed Large regarding onization of open areas by forest. On the leaf and be- morphology branching pattern other hand, D excelsa, a final species in tween the on con- species depending light the succession, was not branched either. ditions. The of these differ- interpretation There are thus species-specific differences ences in terms of light behaviour could independent of adaptation to shade. The on the of differ- improve knowledge ability other species were all more or less ent to under determined species grow light branched. Except for B coriacea, more conditions. seedlings were branched in the shadiest tunnels than in the open air. These results show a tendency towards a greater occu- Counting of leaves and ramifications pation of available space for better energy capture by plants grown in the shade. In general for all the studied species there were more leaves in tunnels II and III than in the others. The decrease under strong Individual leaf area light conditions could be due to a more rapid aging which brought about premature For all the species studied, shade in- leaf fall. The decrease under lower light creased leaf size. Some species reacted conditions could be due to a decrease in very strongly: R grandis, B coriacea and morphogenetic activity following a nutrition- S amara; other species reacted less: P co- al and energetic deficiency. riaceus, S globulifera and D excelsa; A caribaea fell between the 2 groups. The of specific leaf area in response to light en- results from various reports in the litera- vironment. in those of and Krot- ture, particular Logan S amara was the most plastic species. It kov Loach (1969), Logan (1970), (1970), was followed in decreasing order of leaf McClendon and McMillan (1982) showed plasticity by R grandis, B coriacea and P that shade does not always have the same coriaceus, then by D excelsa and A cari- effect. baea and finally by S globulifera. From these it that authors, appears This ranking is basically the same as some species, such as Populus deltoides, the typical forest ranking for increasing tremuloides or Prunus american- Populus shade adaptation as found previously (Du- us react to shade. Others such negatively crey and Labbé, 1986). Similar results as Quercus rubra or Acer saccharum bare- have been reported by Fetcher et al (1983) show reaction. Still others such as ly any who found that in very shady conditions, Morus Fraxinus or Li- alba, pennsylvanica Heliocarpus appendiculatus, a pioneer or riodendron react tulipifera very positively large gap species, was twice as plastic as to shade twice the In the lat- (leaves size). Dipteryx panamensis, a small gap species ter too much shade can species, however, (see table VI). Among the temperate spe- have a depressive effect. This is what was cies, Loach (1967) found results along the also observed in R grandis. same lines: Liriodendron tulipifera, a shade-intolerant species, was more plastic than Fagus grandifolia, a shade-tolerant leaf area Specific species. However, Populus tremuloides, a highly intolerant shade species, does not Shade has the most noticeable and con- conform to this rule (table VI). sistent effect on the specific leaf area. Results obtained for other species Leaves of equal dry weight always had a (table VI) show the regular increase in spe- larger surface area in shade than in sun- cific leaf area as light decreases. It is al- light. The effect of shade, however, dif- ways hazardous to compare results ob- fered depending on the species, as illus- tained under different experimental trated in table V which shows the plasticity conditions. Nevertheless, looking at results obtained for conditions ranging from 13- tion to light can be found in Platanus occi- 20% light, it can be seen that shade- dentalis (Duba and Carpenter, 1980), and tolerant species have a specific leaf area Quercus robur (Tronchet and Grandgirard, close to 1.4-fold greater than those in full 1956), as well as in Quercus sessiliflora sunlight, while shade-intolerant species and Fagus silvatica (Aussenac and Du- have values from 1.8-2.0-fold greater. crey, 1977). Increase in specific leaf area in shade is Studies on non-woody plants (Schoch generally accompanied by a decrease in et al, 1980) showed that the stomatal in- leaf thickness. Leaves exposed to full sun- dex, ie the number of stomata related to light could be twice as thick as leaves in the total number of epidermic cells, de- the shade, as shown by Tronchet and pends on light conditions. The stomatal in- Grangirard (1956), Aussenac and Ducrey dex increases when light increases during (1977), Duba and Carpenter (1980), Fet- the ontogenic phase of the leaf. Regarding cher et al (1983) and Nygren and Kelloma- our results this could indicate that shade ki (1983). These modifications are accom- has a doubly negative effect on stomatal panied by variations in the relative impor- density: a), by increasing cell size; and b), tance of the lacunose parenchyma and the by decreasing the percentage of stomata palisade parenchyma of the leaf (Star- in relation to epidermic cells. This is obvi- zecki, 1974) which may cause changes in ously important to the physiological func- the diffusion of carbon dioxide within the tions of the leaf, particularly to their stoma- leaf and thus in photosynthetic processes. tal conductance. Differences in stomatal plasticity among species occurred, as shown in table V. Stomatal density Variations in stomatal density from sunlight to shade environments were greater for Different stomatal densities were observed R grandis, S amara and B coriacea (more than for A cari- from one species to another: in general, shade-intolerant species) P coriaceus and D excelsa many small stomata or few large-sized baea, (more stomata were found. Our results agree shade-tolerant species). S globulifera, an- with those of Carpenter and Smith (1975) other shade-tolerant species, had no stom- at all. who found a stomatal density for some 50 atal plasticity shade-grown shrub and arborescent spe- cies ranging from 65-900 stomata per mm2 and also with those species reviewed Species plasticity and shade adaptation by Willmer (1983). In particular an increase was observed The species studied all reacted to shade in stomatal density with increase in light by increasing individual leaf area and spe- conditions. Similar results were obtained cific leaf area and by decreasing stomatal by Fetcher et al (1983) for H appendicula- density. tus whose stomatal density more than Variations in specific leaf area, which is doubled when exposed to between 2 and generally accompanied by variations in 100% light. This species also has the par- leaf thickness, demonstrate an adaptation ticular trait of possessing stomata on the to shade by decreasing the distance trav- upper side of leaves when in full sunlight, elled by photons to carboxylation sites and which are absent in strong shade. The by decreasing resistance to the diffusion of same increase in stomatal density in rela- carbon dioxide in the mesophile. More

generally, reducing leaf biomass per unit least plastic species are D excelsa which area in shade leaves is a plant strategy was found to be more shade-tolerant than used to reduce leaf cost under limiting light A caribaea, and S globulifera, another spe- environment. cies with a shade-tolerant reputation. In the same manner, the increase in the The agreement between these aspects amount of stomata in full sunlight shows is thus not perfect and morphological crite- that the leaf must have a better control of ria alone are insufficient. In fact, many oth- temperature as seen through an increase er characters should be examined to inves- in stomatal conductance and thus transpi- tigate tree plasticity in response to light ration. environment. In particular, plasticity should The morphological plasticity of leaves be analysed at a leaf level for photosyn- thetic anato- differs from one species to another. Many light response, biochemistry, ultrastructure and at a authors have attempted to link this mor- my, morphology, phological plasticity to species shade be- plant level and at a canopy level (Board- haviour. For temperate species whose for- man, 1977; Bjorkman, 1981; Givnish, est behaviour is fairly well known, it is 1988). possible to rank species from most to least From a forester’s point of view forest be- shade tolerant (Baker, 1949). There is a haviour is not a well-defined concept, as good agreement between degree of leaf shown by the following examples. Shade- plasticity and shade tolerance where the tolerant versus shade-intolerant behaviour most plastic species are the most shade- support the assumption that a full sunlight intolerant (see Specific leaf area). environment is the standard reference. In tropical species, empirical and silvi- The light-demander species notion implies that some need more than cultural knowledge is basically non- species light most existent and forest behaviour can only be others, although species may grow deduced from morphological variations. under full sunlight environments. Other means of explaining differences in tree In the species we have studied in light response are to consider their place in Guadeloupe, initial insight into their forest a forest successional cycle (Bazzaz and behaviour was obtained studies through Pickett, 1980) from pioneer species to late on natural regeneration. The results re- successional species, or to emphasize garding the morphological plasticity of growth response to gap size in the forest these species are in approximate agree- canopy 1978; Denslow, 1980, ment with the results. (Whitmore, preceding 1987). Other authors (Grime, 1979; Kolb et In order of decreasing plasticity, the first al, 1990) consider that competition or plant species found is R grandis, followed imme- tolerance strategy in response to stresses diately by S amara and then by B coria- should include all stress factors and not cea. From growth studies in experimental only light stress. conditions of natural regeneration (Ducrey and Labbé, 1986), S amara was found to be slightly less shade-tolerant than R gran- REFERENCES dis. No information was obtained for B co- riacea. The following species, in decreas- Aussenac G, Ducrey M (1977) Étude bioclima- order of are A caribaea ing plasticity, tique d’une futaie feuillue (Fagus silvatica L which was found to be shade-tolerant and et Quercus sessiliflora Salisb) de l’Est de la P coriaceus which usually had a reputation France. I. Analyse des profils microclima- of being very shade-tolerant. Finally, the tiques et des caractéristiques anatomiques et morphologiques de l’appareil foliaire. Ann Sci two species of tropical trees. Oecologia (Berl) For 34 (4), 265-284 58, 314-319 Baker FS (1949) A revised tolerance table. J For Givnish TJ (1988) Adaptation to sun and shade: 47, 179-181 a whole-plant perspective. Aust J Plant Phys- Bazzaz FA, Pickett STA (1980) Physiological iol 15, 63-92 ecology of tropical succession: a comparative Gordon JC (1969) Effect of shade on photosyn- review. Annu Rev Ecol Syst 11, 287-310 thesis and dry weight distribution in yellow Bjorkman O (1981) Response to different quan- birch (Betula alleghaniensis Britton) seed- tum flux densities. In: Physiological Plant lings. Ecology 50 (5), 924-927 Ecology. I. Responses to the Physical Envir- Goulet F, Bellefleur P (1986) Leaf morphology onment OL et Ver- (Lange al, eds) Springer plasticity in response to light environment in lag, New York, 57-107 deciduous tree species and its implication on Boardman NK (1977) Comparative photosynthe- forest succession. Can J For Res 16, 1192- sis of sun and shade plants. Annu Rev Plant 1195 Physiol 28, 355-377 Grime JP (1979) Plant Strategies and Vegetation Carpenter SB, Smith ND (1975) Stomatal distri- Processes. John Wiley & Sons, New York bution and size in Southern Appalachian Jarvis PG (1964) The adaptability to light intensi- hardwoods. Can J Bot 53, 1153-1156 ty of seedlings of Quercus petraea (Matt) Denslow JS (1980) Gap partitioning among trop- Liebl. J Ecol 52, 545-571 ical rainforest trees. 12 Biotropica (Tropical Jones RH, McLeod KW (1990) Growth and pho- succession), 47-55 tosynthetic responses to a range of light envi- Denslow JS (1987) Tropical rain forest gaps and ronments in Chinese tallowtree and Carolina tree species diversity. Annu Rev Ecol Syst ash seedlings. For Sci 36 (4), 851-862 18, 431-451 Kolb TE, Steiner KC, McCormick LH, Bowersox Duba SE, Carpenter SB (1980) Effect of shade TW (1990) Growth response of northern red- on the growth, leaf morphology and photo- oak and yellow-poplar seedlings to light, soil synthetic capacity of an American sycamore moisture and nutrients in relation to ecologi- clone. Castanea 45 (4), 219-227 cal strategy. For Ecol Manage 38, 65-78 Ducrey M (1982) Study of stimulated natural re- Loach K (1967) Shade tolerance in tree seed- generation in the wet tropical rain forest of lings. I. Leaf photosynthesis and respiration Guadeloupe (FWI). In: Forestry in the Carib- in plants raised under artificial shade. New bean. US MAB rep No 7, 130-133 Phytol 66, 607-621 M Croissance de Ducrey (1986) juvénile Loach K (1970) Shade tolerance in tree seed- introduites dans quelques espèces lings. II. Growth analysis of plants raised un- l’arboretum de Débauchée (Guadeloupe). der artificial shade. New Phytol 69, 273-286 Rev For Fr 38 (5), 451-456 Logan KT (1970) Adaptation of the photosyn- Ducrey M, Labbé P (1985) Étude de la régéné- thetic apparatus of sun- and shade-grown ration naturelle contrôlée en forêt tropicale birch Can humide de I. Revue yellow (Betula alleghaniensis Britt). Guadeloupe. bibliogra- J Bot 48 (9), 1681-1688 phique, milieu naturel et élaboration d’un pro- tocole expérimental. Ann Sci For 42 (3), 297- Logan KT, Krotkov G (1969) Adaptation of the 322 photosynthetic mechanism of sugar maple in various Ducrey M, Labbé P (1986) Étude de la régéné- (Acer saccharum) seedlings grown light intensities. Physiol Plant 22, 104-116 ration naturelle contrôlée en forêt tropicale humide de Guadeloupe. II. Installation et McClendon JH, McMillan GG (1982) The control croissance des semis après les coupes of leaf morphology and the tolerance of shade d’ensemencement. Ann Sci For 43 (3), 299- by woody plants. Bot Gaz 143 (1), 79-83 326 Nygren M, Kellomaki S (1983) Effect of shading Fetcher N, Strain BR, Oberbauer SF (1983) Ef- on leaf structure and photosynthesis in young fect of light regime on the growth, leaf mor- birches, Betula pendula Roth and B pubes- phology and water relations of seedlings of cens Ehrh. For Ecol Manage 7, 119-132 Schoch PG, Zinsou C, Sibi M (1980) Depen- Tronchet A, Grandgirard A (1956) L’analyse his- dence of the stomatal index on environmen- tométrique et son application à l’écologie fo- tal factors during stomatal differentiation in restière. Ann Sci Univ Besançone 2 Sér Bot leaves of Vigna sinensis L. I. Effect of light 8, 3-30 intensity. J Exp Bot 31, 1211-1216 Tsel’Niker YL (1977) Regulation of processes of Smith H (1982) Light quality, photoperception exchange and morphogenesis in seed- and Annu Rev Plant CO2 plant strategy. Physiol lings of forest trees under conditions of shad- 33, 481-518 ing. Sov Plant Physiol 24, 43-48 Starzecki W (1974) of Topography photosynthe- Whitmore TC in the forest sis in the crowns of Tilia cordata and Acer (1978) Gaps canopy. pseudoplatanus trees. Bull Acad Pol Sci Ser In: Tropical Trees as Living System (Tomlin- Sci Biol 22 (9), 597-602 son PB, Zimmermann MH, eds) Cambridge University Press, 639-655 Takahara H (1986) Effects of shading on the growth of some Quercus seedlings. J Jpn Willmer CM (1983) Stomata. Longman Inc, New For Soc 68 (7), 289-292 York, 166 pp