ECOGRAPHY 23: 349 – 359. Copenhagen 2000 A comparison of fine-scale distribution patterns of four plant groups in an Amazonian rainforest J. Vormisto, O. L. Phillips, K. Ruokolainen, H. Tuomisto and R. Va ´squez Vormisto, J., Phillips, O. L., Ruokolainen, K., Tuomisto, H. and Va ´squez, R. 2000. A comparison of fine-scale distribution patterns of four plant groups in an Amazo- nian rainforest. – Ecography 23: 349 – 359. We carried out a comparison among the floristic patterns of four different plant groups (palms, trees, melastomes and pteridophytes) in a lowland rainforest site in Peruvian Amazonia. The study site consisted of a mosaic of edaphic patches reflecting the different geological formations that can be found on the surface. We collected the data along a linear transect (500 m long, divided into 20 ×20 m or 5 ×20 m subplots), and recorded of the four plant groups all individuals that exceeded a minimum size limit predefined for each plant group. We also recorded the drainage conditions and soil type classes in each subplot of the transect. The results indicated that different plant groups can produce similar floristic patterns in local spatial scales, and that these patterns reflect similarities in edaphic conditions. All matrix correlations calculated between pairs of the four plant groups were positive and statistically significant. Floristic composition in all plant groups correlated with soil class, and to a somewhat lesser degree with drainage. These results imply that any one of the four plant groups could serve as a rough indicator of more general floristic patterns, and that even the inventory of a limited part of the flora can shed light on the floristic variation found in Amazonian forests. J. Vormisto ( jaaor@utu.fi), K. Ruokolainen and H. Tuomisto, Dept of Biology, Uni. of Turku, FIN-20014 Turku, Finland – O. L. Phillips, School of Geography, Uni. of Leeds, Leeds U.K. LS29JT.– R. Va ´squez, Jardı ´n Bota ´nico de Missouri, Apartado 020, Jae ´n, Cajamarca, Peru. Any study that aims at understanding the structure and function of tropical forest plant communities can only sample a part of the flora. The use of a limited set of species is a practical necessity because the extreme species richness, inadequate taxonomic knowledge and logistic difficulties make it prohibitively laborious to sample all plant species. Thus, floristic studies tradition- ally deal only with trees (e.g. Ashton 1969, Lieberman et al. 1985, Gentry 1988a, b, ter Steege et al. 1993, Duivenvoorden 1995, Clinebell et al. 1995). More re- cently, studies have also been done on ground herbs (Poulsen and Balslev 1991, Poulsen 1996), palms (Kahn and de Castro 1985, Kahn and de Granville 1992), pteridophytes (Young and Leo ´n 1989, van der Werff 1990, Tuomisto and Ruokolainen 1994, Tuomisto et al. 1995, Tuomisto and Poulsen 1996, Ruokolainen et al. 1997), Lauraceae (van der Werff 1992), and Melastom- ataceae (Tuomisto and Ruokolainen 1994, Tuomisto et al. 1995, Ruokolainen et al. 1997). However, if the results of studies on restricted components of the flora are to be used as models for making generalisations concerning the flora in general, and potentially for developing conservation strategies, then it is important to assess how well the studied part of the flora repre- sents the unstudied part. In other words: are the studied groups good ‘‘indicators’’ of floristic patterns in the rest of the flora? Although all plants need light, water and nutrients, different plant groups may have different responses to these resources (Cox and Larson 1993). An obvious way to study the variation in ecological behaviour among plant groups is to measure if the plant groups Accepted 27 September 1999 Copyright © ECOGRAPHY 2000 ISSN 0906-7590 Printed in Ireland – all rights reserved ECOGRAPHY 23:3 (2000) 349