Seedling Growth of Shorea (Dipterocarpaceae) Across an Elevational Range in Southwest Sri Lanka

Performance of seedlings of seven rain forest, canopy dominant Shorea species was studied in a transplant experiment in forest sites a t three different elevations (low, mid and high) M-ithin the humid zone of southlt-est Sri Lanka. Five species generally inhabit lo~vto mid-elevations, one at midand lo~ver montane elevations, and one exclusively at lower montane elevations. Temperature, rainfall and cloudiness varied M-ith elevation. For each site seedlings were grolvn in pots under partial shade conditions using similar soils and evermoist conditions. All gro~vth measures sho~ved differences among elevation sites, among species and in the interaction bet~veen species and elevation sites. Performances of species collecti\~ely sho~ved (i) decline in height and leaf number M-ith increase in elevation, (ii) higher dry mass at lo~vand mid-elevation sites compared to that a t highelevation and (iii) a higher mass of single leaves at the mid-elevation site than at the high-elevation site. Rank order of species changed across elevations for both height and dry mass. Dry mass declined with elevation in four of the seven species studied. S.gardlzeri, the only exclusi\~ely Ion-er montane species, increased dr)mass with elevation. Height declined with elevation for six of the species with only S. gardneri showing no change. Changes M-ith elevation in the rank order of species for total leaf number and mass of single leaves lvere small. Hom-ever, total leaf number and masses of single leaves differed among species and among elevations. S.megistofih~llaand S. disticha had a few leaves with high individual masses, lvhile S. gardizeri, S . a f i ~ z i s and S.trapeczfolia had many leaves lvith less mass per individual leaf. One group of species sho~ved relatively little change in leaf number per seedling and large changes in mass of single leaves. The other group varied more in leaf number but mass of individual leaves remained constant. Grom-th allocation ' Author for correspondence to leaf production versus individual leaf size appears related to the successional division of Shorea section Doona. Also all species grew better a t the lo~v-elevation site irrespective of their natural ranges except S. gardneri, whose natural range is restricted to high elevations, and exhibits markedly lo~ver gro~vth responsiveness than the other M-ider ranging species. m Y TVORDS: Asia, dipterocarps, rain forest, regeneration, seedling morpholog), Sinharaja. I N T R O D U C T I O N We aim to examine causes of the distinct ecological ranges manifested by suites of related tree species in tropical rain forests. One aspect is to clarify the role of competition by comparing ranges in nature ~ v i t h measures of performance in pot experiments, in ~vhich the species are groLvn across separate gradients of light, soil moisture, nutrients and elevatioil Lvith other factors held constant. Later, results will be cornpared Lvith trailsplant experiinents in forest gaps, including mixed-species plantings across gradients of combined en\~ironrnental factors in the forest. I11 this paper we present results from pot experirneilts in ~vhich seedliilgs of a series of related species have been grolvn at three elevations and temperature regimes, but with soil coilditions and light held constant. Several distinctive rain forest communities have been described for an elevation gradient that asceilds from moist coastal lo\vlands to the lobver rnontaile hill ranges of southwestern Sri Lailka (De Rosayro 1942, Gunatilleke & Ashton 1987). Across this gradient tree species in the family Dipterocarpaceae codominate the forest canopy along with species in the Clusiaceae, Bombacaceae, Sapotaceae and Myrtaceae (Greller et al. 1987, Gunatilleke & Ashton 1987, Gunatilleke & Gullatilleke 1985). A clade of nine partially sylnpatric tree species in the genus Shorea ~vithin the endemic section Doona (Dipterocarpaceae) is distributed across this elevation gradient (Ashton 1977), and occur at lobver and higher altitudes than other members of the genus. Species diversity is conceiltrated at middle elevations, perhaps because only a f e ~ v forest fragments nobv survive at l o~v elevation. Rain forests of the coastal plains, valleys and lolver slopes up to 300 m elevation belong to the Dipterocarh~~s community (De Rosayro 1942, Gunatilleke & Ashton 1987). S . a&nis (Th~v. ) Ashton, S . congestflora (Th~v.) Ashton, and S . cordfolia (Thbv.) Ashton occur in this community. The i14esua-Shorea community occupies steeper upper slopes and middle elevations from 300-900 m but also descends to low elevatio~ls on shallobv soils over siliceous rocks (De Rosayro 1942, Gunatilleke & Ashton 1987). In this elevation zone eight of the nine Shorea section Doona species can occur (h'lerritt & Ranatunge 1959; Gunatilleke & Gunatilleke 1981, 1985) as well as five other Shorea species, in two other sections. In lo~ver montane rail1 forests at 900-1600 m, S. gardneri (Th~v. ) Ashton dominates forests either singly or in mixture with species in other families (Ashton 1977, Greller et al. 1987). Both S . qy lan ica (Thbv.) Ashton and 5'. trapezfolia (Thbv.) Ashtoil extend into the lower elevations of this zone. 233 Seedling growth of Shorea (Dipterocarpaceae) in Sri Lanka The distribution of dipterocarp species in Sri Lanka follobvs general patterns that have been reported else~vhere in the Far East (Ashton 1964, Ashton & Hall 1992). Research in the mid-elevation zone of the .Vlesz~a-Shorea community suggests that Shorea section Doona has undergone speciation in relation to local topographic differences in available soil moisture (Ashton 1992, Ashton et al. 1995), soil nutrition (Gunatilleke et al. 1996, 1997), and in irradiance regimes at the forest groundstorey (Ashton 1992, Ashton & B e r l ~ n 1992, Ashton 1995, Ashton et al. 1995). Ho~vever, we are abvare of no ecological stud) that has compared the performance of closely-related tropical tree species within and outside their natural elevation ranges. S T U D Y S I T E S A N D SPECIIES The study Lvas made at three different elevations in south\vestern Sri Lanka, representing the Dipterocarpus cornmunity, the :Wesun-Shorea cornmunity, and the S.gardneri association. The sites selected lvere Indikada h'Iukulana (125 111) in Kalutara district, and Sinharaja field station (580 m) and Sooriyakanda (1060 m), both in Ratnapura district. Indikada hIukulana is located 60 km to the northbvest of the Sinharaja field station while Sooriyakanda is 25 km to the east. The climate at all three sites is aseasonal, receiving precipitation from the southwest monsoon from May to July, the northeast monsoons from October to December, and convectional rains betbveen the monsoon periods. The mean annual rainfall at Indikada 'Iukulana is 2370 mm with a mean annual temperature of 26.6 "C. During February Indikada Mukulana receives an average rainfall of 160 mm and is the only month that receives <200 mm. The Sinharaja field station receives an average annual rainfall of 3904 mm lvith no months receiving <200 mm. At Sinharaja average monthly temperatures range betlveen 25 and 27 "C with a mean of 26 "C. The high-elevation site at Sooriyakanda has a mean annual temperature of 18 "C and receives annual rainfall of 2200 mm, with only February receiving <200 mm. The experimental plantings at all three sites were in large clearings adjacent to rain forest because these conditions provided suitably uniform environments across elevation. These will be referred to as the lolv-, midand high-elevation sites hereafter. Poor fruiting in tlvo of the Shorea species, one of which is uncornmon (S. zeylanica), restricted this experiment to seven of the nine species. The species examined were S. aflnis, S. cordfolia, S. distichs, S. gardneri, S. megistofihylla, S. trapezfolia and S. worthingtonii. These species reach the canopy of mature phase forest with the exception of S. aznis and S. cordfolia which are usually subcanopy species. S. aflnis and S. trahegfolia are known locally as 'thiniya', S. gardneri is called 'ratu dun', and all three species are light hard~voods that produce highly resinous fruits. The remaining species are heavy hardbvoods kno\vn locally as 'beraliya' and produce larger edible fruits (Ashton 1977, Gunatilleke & Gunatilleke 1991, Trimen 1892). 234 C . 1'. S . G U K A T I L L E K E E T A L . Mature fruits were collected from several populations for each of the six species from the mid-elevation site and for S. gardneri from the high-elevation site. Insufficient seed was available from lobv-elevation sites because little mature rain forest no\\. exists that is not in a degraded state due to past land clearance and logging that has preferentially removed most mature Shorea trees. Shorea a f ln i s , S . cordfoolin, S . distichs, S . megistohhylla and S . worfhingtonii fruited during January-February 1989, while S . trahezzfolia and S . gardneri fruited May-June 1989. As all these species have recalcitrant seeds, they \\.ere immediately germinated in nursery beds after collection. Seedlings lvere grolvn under partial shade in a nursery at the Sinharaja mid-elevation site until the experiment commenced in September 1989. At the start of the experiment, seedlings from the nursery lvere transplanted into black polythene bags, each 15 cm in diameter and 30 cm in height containing 2.5 kg of soil (5.3 1 volume). Large polythene bags lvere used to pot seedlings from the start to avoid transplanting into larger pots during the experimental period. The topsoil lvas taken from 0-25 cm depth of a 10-y old Pinus caribaea plantation adjacent to the boundary of the Sinharaja rain forest and near the mid-elevation site. The plantation lvas established on abandoned slvidden land that \\.as originally cleared of rain forest 30 y previously. In a separate experiment, the growth of these Shorea seedlings, in soil from this Pinus plantation, lvas shobvn to be comparable to that in their original rain forest soils of valley and midslope (Ashton et al. 1997; Gunatilleke et al. 1996, 1997). h'lacro-nutrient analyses of these soils have been given in Gunatilleke et al. (1996, 1997). A handful of soil taken from a composite soil sample from natural forests in the lo\\.-, midand high-elevation sites \\.as also added to each pot to serve as rnycorrhizal inoculum. Previous work on four of the species (Ashton & Berlyn 1992, Ashton et al. 1995) indicated that, although maximum grobvth rates for each species \\.ere attained at different irradiance levels, the level closest to the maximum for all species was 50% of full sunlight. The shade houses lvere therefore constructed in open areas free of surrounding vegetation and receiving direct light for > 8 h per day. Shading lvas provided by over-laying a coir (coconut fibre) mesh material with 1-cm' mesh-size on the roof and sides of the houses. This material allowed approximately 50% of sunlight of normal red: far red ratio to filter through as measured at noon on sunny days in October 1989 using PPFD sensors attached to a data logger (LI 190SZ sensors, LI 1000 data logger; LICOR Inc., Lincoln, Nebraska, USA). All seedlings in shade houses received natural precipitation through the coir material during rainy periods, but during relatively dry periods they were irrigated. Seedlings were not fertilized during the period of the experiment. The experiment comprised planting seedlings of seven different Shorea spp. lvithin shade houses at low-, midand high-elevation sites. Twelve potted seedlings of each species lvere randomly arranged in one of four groups within each 235 Seedling growth of Shorea (Dipterocarpaceae) in Sri Lanka shade house. In total 1,008 seedlings (48 seedlings x 7 spp. x 3 sites) liere monitored over a 24-mo period from September 1989 to October 1991. Seedling height from the soil surface to the apical leader, to the nearest 0.5 cm, and leaf nurnber per seedling were measured 1 mo after transplanting and then at intervals of 3 mo for 2 y. At the end of the experiment, dry mass of the lihole plant and that of its constituent parts (stem, leaves, tap root and fine roots) of a randornly selected subsarnple of three seedlings per group per species per elevation site Lias recorded after harvesting, washing them free of soil particles, and oven-drying them at 85 O C for 48 h. No record Lias made of leaves and other stem and root tissues that abscised over the experimental period. Because of some seedling mortality over the extended period of the study, data were subjected to an unbalanced AILOVA using the GLM procedure of SAS (Ray 1982). ANOVAs liere done using logarithmically transformed data for each of the attributes measured to compare performailce of species grolin at the three different elevations. For analyses, groups within each shade house liere pooled and a tlio-way ANOVA performed comparing species, elevation and their interaction. Multiple cornparisons arnong means of different species liere carried out using Tukey's hoilestly significant difference method (P 5 0.05).