Biomass-diversity relations of epiphytic bryophytes on small Quercus copeyensis stems in a Costa Rican montane cloud forest

Resumen: Se muestreó briófitas epífitas sobre tallos pequeños de Quercus copeyensis Müll. en un bosque montano nuboso maduro (Talamanca, Costa Rica). Se muestreó a 0.8-1.8 m en cinco tallos (d.a.p. 1.8 -2.8 cm), hallándose 22 especies de musgos y 22 de hepáticas. La biomasa de briófitas ("briomasa") se correlacionó con la frecuencia, contribuyendo los musgos 54-99%. Un 14% de las especies contribuyó casi el 90% de la briomasa, sobresaliendo Pilotrichella flexilis (Hedw.) Ángstr., Rigodium sp. 1, Porotrichodendron superbum (Taylor) Broth., Prionodon densus (Hedw.) Müll. HaI., Neckera chilensis Schimp. ex Mont. y Plagiochila guilleminiana Nees & Mont. Estas especies podrían jugar un papel importante en la regulación del flujo de agua.

Epiphytic bryophytes are an important structural feature of tropical montane cloud forests, where they play an important role in rainfall interception (pócs 1980,Veneldaas andVan Ek 1990).They also have a significant capability of nutrient retention (Nadkarni 1986, Hofstede et al. 1993), affecting nutrient cycling within forests (Coxson 1990).Furthermore, they pro vide shelter for numerous small invertebrates (Gradstein 1992).Epiphytic biomass in tropical montane forests has been estimated on several tons per hectare for tropical montane forests (e.g.Pócs 1980, Nadkarni 1984).Other studies give an indication of epiphyte abundance, and especially of epiphytic bryophytes, in biomass per unit area, usually in g dry weight per dm 2 (e.g.Frahm 1987, Veneldaas et al. 1990, Wolf 1993a, 1993b).
However, at present little information is available on biomass-diversity relations among epiphytic bryophytes on small stems in the understory of tropical montane cloud.forests (Wolf 1993a).Thus, pattems of bryophyte biomass (bryomass) distribution were studied among epiphytic mosses and liverworts growing on small stems of a common oak tree species in a Costa Rican montane cloud forest.For the present analysis a subset of five small stems with a dbh (diameter at breast height) between 1.8 and 2.8 cm were randomly selected; this was done in order to avoid differences in age of substrate.Epiphytic bryophyte vegetatíon was censused and biomass harvested.Species were collected and identified with help from specialists and temporarily stored at the herbarium of the University oí" Amsterdam (AMD).Bryomass samples were sorted at species level and dried for 24 hr at 70°C in order to determine dry weight.
On five small oak stems 44 epiphytic bryophyte species were recorded in 18 families, equally distributed over mosses and hepatics (Tables 1 and 2).A total of 24 were identified to species level (54.5%), whereas the remaining collections were recognized as morphospecies and determined at genus level.The biomass of mosses outweighs the value for liverworts in all five samples.This is indicative for a relatively dry forest, as humid    .g" '5 1; '" " " "" :::-;:t ;l: Fig. 2, Biomass-diversity diagram for epiphytic bryophytes in a mature montane c10ud forest, San Gerardo de Dota, Costa Rica.Bryophyte species are ranked from high to lowbiomass per unít area (g/dm2) vs, theírbiomass per uni! area (Y-axis).Datarepresent mean'+ 1 S.E, where,n =.5, with n isthe number of sampled oak teees.
Implications of bryomass-diversity relations for conservation and management of tropical montane cloud forests are discussed.This study was carried out in a mature, undisturbed upper montane Quercus copeyensis Müll.dominated cloud forest stand on the Pacific slope of the Cordillera de Talamanca in Costa Rica.A 0.05 ha forest plot (20 x 25 m; SES aspect) was sampled at ca. 3000 m elevation near San Gerardo de Dota in the eastem comer of the Los Santos Forest Reserve (9°35'40" N; 83°44'30").In Villa Milis, at a distance of 10 km W of the selected forest plot, the average annual temperature is 1O.9°C and the average annual rainfaIl ís 28 12 mm (Kappelle 1992).According to Kappelle (1996) this upper montane cloud forest is exclusively dominated by 30 to 35 m tan Quercus copeyensis and Q. costaricensis canopy trees, in association with Magnolia sororum, Schefflera rodriguesiana and Weinmannia pinnata.Numerous arborescent genera determine the aspect of the subcanopy, e.g.Ardisia, Comarostaphylis, Cornus, Ilex, Ocotea, Oreopanax, Styrax, Symplocos, Vaeeinium, Viburnum and Zanthoxylum.The understory is dominated by Chusquea bamboos.The fem Elaphoglossum was particularly abundant at the forest floor of the selected plOL In August 1992 a 2.5 m wide belt transect was laid out along the length axis of the the selected forest pIot using a rope.Epiphytic bryophytes were sampled on each Quercus eopeyensis stem within a distance of 1.25 m of the rope.Only stems of Q. eopeyensis were selected in order to avoid differences in host specificity and bark properties.Samples were taken from the bark of stems covering the height interval between 0.8 and 1.8 m above soil level.A lower height of 0.80 m was selected to avoid bottom-up water and nutrient fluxes as a result from splashing at soil leve} during heavy rain fallo Only the west-facing bark of stems was sampled in order to avoid the strong influence of polarity on epiphytic bryophytes, such as caused by wind and horizontal precipitation Ce! Barkman 1958: 40).

Fig. 1
Fig. 1 shows the species/area curve for cumulative numbers of species and samples on basis of results n-om ten random permutations of possible sample sequences.A TWINSPAN analysis (Hill 1979) did not reveal a particular subdivision into two distinct bryophyte communities (Van Dunné 1993).This suggests a high homogeneity among samples which probably belong to one single epiphytic bryophyte community.S�rensen's similarity coefficíents, ranging from 0.5 1 to 0.73, also indicate that samples were similar in species composition, and consequently, that the sampling strategy was adequate for the purpose of this study.Average dry weight values of bryomass in samples, as weU as for moss and liverwort portions separately are given in Table 2. Total dry weight ranges from 1.02 to 6.02 g per sample.Biomass per unit area ranged from 0.13 to 0.68 gldm2.Mosses make up 54 to 99% of the total bryomass per sample.Fíg. 2 shows the bíomass-diversity reJatíon as expressed in the average dry weight and ranking of the 21 most important specíes.Frequency and bíomass per unit area are positively correlated for al! individual samples (Pearson's correlation coeficient is 0.718, P<O.OOl).No significant correlation was found for biomass per unÍt area and species number of individual samples.

o
Fig. l.Species/area relations showing cumulative number of epiphytic bryophyte species and cumulative number of sampled Quercus copeyensis tree stems in a mature montane cloud forest, San Gerardo de Dota, Costa Rica.Graphs represent mosses (top) and liveIWorts (centre) seperately, and total number of bryophytes (bottom).Plotted data in ¡he diagrarn represen!ten permutations of possible sequences in which sampled tree stems may be plotted.Plotted values correspond to numbers of species (mean + 1 S.E.).
and perhumid conditiom: in upper montane cloud forest interiors are expressed by a high diversity and dominance of liverworts (Gradstein et al. 1989).The concave biomass diversity pattern as presented in Fig. 2 is in accordance with the dominance-diversity theory of epiphytic bryophyte communities (e! Wolf 1993a).Thís concave trend can be attributed to the Principie of Priority, which holds that the first species to colonise a site has the biggest chane e to form pure patches through growth and vegetative reproduction, excluding new arrivals (Benzing 1981, Wolf 1993a and 1995).Species suitable to grow in a particular environment-i.e.mature forest and west-facing bark of oak tree stems in this study-would have an equal chance to settle, develop and attain high cover or biomass values.The correlation which was found between biomass and frequency of species in our five samples is in contrast with calculatíons based on Wolf's (l993b) relative frequency and average cover data of tree base cornmunitíes.This discrepancy is probably due to a difference in Given the importance of epiphytic bryophytes for forest hydrology(Pócs 1974, 1980, Nadkarni 1986, Veneklaas et al. 1990)  and the results of the present study-i.e. the fact that epiphytic bryomass in the forest's understory appears to be determined by relatively few species -, more detailed information about the structure of the epiphytic biomass in tropical forests, and particularly montane cloud forests, is urgently needed.Conservation and wise management of these forests requires baseline information about species of possible key importance to maintaining hydrological properties of the forest such as rain fall interception and water retention.In the present study 90% of total bryomass per unit area can be attributed to only six specíes, namely Pilotrichella flexilis (Hedw.)Á ngstr., Rigodium sp. 1, Porotrichodendron superbum (Taylor) Broth., Prionodon densus (Hedw.)MüIl.Hal., Neckera chilensis Schimp ex Mont.and Plagiochila guilleminiana Mont.(Fig. 2).We assume that these species contribute largely to stem flow regulation on small understorey trees such as saplings, as they are able to retain water and nutrients.Their proportion of the total bryomass on small-sized oak stems may be an indicatíon of their key importance to the hydrological functioning of tropical montane oak dominated cloud forests.Further study is needed to support this theory.