BIOTROPICA 39(1): 25–30 2007 10.1111/j.1744-7429.2006.00228.x Predicting the Uncertain Future of Tropical Forest Species in a Data Vacuum Toby A. Gardner 1,4 , Jos Barlow 1,2 , Luke W. Parry 1,3 , and Carlos A. Peres 1 1 School of Environmental Sciences, University of East Anglia, NR4 7TJ, Norwich, United Kingdom 2 Museu Paraense Em´ ılio Goeldi (MPEG), Av. Perimetral 1901, Bairro Terra Firme, Bel´ em - Par´ a, 66077-530, Brazil 3 Center for International Forestry Research (CIFOR), Embrapa Amaz ˆ onia Oriental, Trav. En´ eas Pinheiro s/n, CEP 66.095-100, Bel ´ em, Par´ a, Brazil Key words: biodiversity; predictions; research; secondary forests. WHEN SUPPORTED BY SUFFICIENT EVIDENCE, predictive analyses and scenario building can make important contributions towards de- vising conservation policy strategies. Wright and Muller-Landau (2006a, henceforth termed WMLa) provide a timely analysis that attempts to predict the future of tropical forest species based on hu- man population-driven projections of tropical deforestation. Citing evidence of slowing population growth and rapid urbanization they suggest that deforestation rates are expected to decrease, and natu- ral forest regeneration through secondary succession to accelerate. WMLa use their predictions of an increase in secondary forest area to paint an optimistic picture of the future, going so far as to suggest that the widely anticipated mass extinction of tropical forest species will be avoided (see also Aide & Grau 2004). Brook et al. (2006, henceforth termed BBKS) reject the op- timistic predictions of WMLa by reinforcing the importance of a number of original caveats made by WMLa, as well as raising ad- ditional objections. BBKS indicate that the relationship between rural and urban population growth and deforestation is too com- plex to be reliably predicted. They also argue that the negative effects of the heavy momentum set in motion by current patterns of deforestation and population growth (i.e., loss of irreplaceable primary source habitat, e.g., Brooks et al. 2002, Brook et al. 2003, and extinction debt, Tilman et al. 1994) are irreversible before fu- ture reductions in deforestation rates are able to rescue committed extinctions. Despite this criticism Wright and Muller-Landau (2006b, henceforth WMLb) retain confidence in their optimistic projec- tion, arguing that their simulations were robust to variability in the growth of urban versus rural human populations projected by the FAO, which already take into account existing biases in age structure. However, in addition to assumptions regarding human demography and the fate of undisturbed primary forests, the predic- tions of WMLa rely centrally upon net changes in forest cover and the (explicit) assumption that all types of forest cover can be treated equally—specifically that primary, secondary, and degraded forests can be considered to have similar ecological value. Citing a single key Received 27 April 2006; revision accepted 15 June 2006. 4 Corresponding author; e-mail: t.gardner@uea.ac.uk study dealing with faunal recovery following regeneration (Dunn 2004), and three dealing with floristic succession (Turner et al. 1997, Chazdon 2003, Lugo & Helmer 2004) WLMa conclude that sec- ondary forests (a) provide a suitable habitat for many species and (b) will provide the necessary haven for species currently restricted to small patches of native habitat. In contrast BBKS argue convinc- ingly that forest quality is more important than total forest area, and that degraded and secondary forests are indisputably depauperate, support significantly fewer old-growth specialists, and provide re- duced ecosystem services. WMLb acknowledge the importance of this distinction in stating that “a crucial difference between BBKS and WMLa [in interpreting the implications of WMLa] concerns the conservation value of degraded and secondary forests.” Secondary forests clearly are an increasingly important com- ponent of tropical forest landscapes (Perz & Skole 2003, Wright 2005, FAO 2005), and we welcome the fact that WMLa have highlighted the increasing conservation importance of natural for- est regeneration. However, we are concerned that a paper entitled “The Future of Tropical Forest Species” presents very little ecolog- ical data on the biodiversity value of naturally regenerating forests in the tropics. In order to evaluate the relative merits of the argu- ments made by WMLa and WMLb versus BBKS, it is vital that we understand the proportion of species that is completely dependent on old growth forest—a point duly acknowledged by WMLb. Our main purpose in this commentary is to shed some objective light on the strength of the optimistic predictions made by WMLa by assessing our current understanding of the role of secondary forests for the conservation of tropical forest species. To do this we sur- veyed scientific papers published in international journals which have reported on the value of regenerating lands for four main groups of tropical forest vertebrate: birds, amphibians, reptiles, and primates. Birds are perhaps the best-known tropical taxa, yet we identified only 15 studies that examined the value of second growth for tropical forest birds. Nine of these 15 studies examined small plots following subsistence agriculture; of these, all were smaller than 5 ha where plot sizes were reported. Only two studies (Borges & Stouffer 1999, Sodhi et al. 2005) were in areas of secondary forest larger than 1000 ha. Levels of replication were also very poor, and eight of the C  2006 The Author(s) Journal compilation C  2006 by The Association for Tropical Biology and Conservation 25