The authors have responded to our criticisms (Gereau et al., 2013) in the following major areas: inadequate knowledge of the study area and its flora, lack of transparent or repeatable methods for data selection, and an inconsistent dataset. However, their misunderstanding of data sources has led to a number of serious errors, both in the original article (Yessoufou et al., 2012) and in their subsequent defense (Daru et al., 2014):
• The CEPF database used to compile the checklist of Red Listed flora within the Eastern Arc forest blocks (Table S2 in Yessoufou et al., 2012) can only be Appendix 1 in the [list of species outcomes for the Eastern Arc Mountains and Coastal Forests] (www.cepf.net/Documents/final.EasternArc.EPappendices.pdf). The authors failed to recognize that this database includes species from throughout the Eastern Arc Mountains and Coastal Forests, not only from the Eastern Arc. As a result, the 229 species in Table S2 includes 63 species (28% of the total) that do not occur anywhere within the Eastern Arc (list available on request from firstname.lastname@example.org).
• The authors claim to have analyzed only the 12 Tanzanian mountain blocs of the Eastern Arc, excluding the Taita Hills in Kenya. In fact, due to the above error, Table S2 includes three species endemic to the Taita Hills, as well as five species endemic to the coastal forests of Kenya.
• The Tanzanian plant data downloaded from the IUCN Red List (Table S1 in Yessoufou et al., 2012) includes species assessed in all Red List categories, threatened and non-threatened. The data from the CEPF database (Table S2) includes only species assessed in threatened categories. The differences between the tables regarding inclusiveness by category (all categories vs. threatened only) as well as selection by country (Tanzania only vs. Tanzania and Kenya) make the two datasets incompatible.
In addition to the concerns outlined previously (Gereau et al., 2013), we find that the authors have analyzed a dataset not belonging to their stated geographic study area and placed it in the context of an incompatible dataset from a larger study area that does not include all of the former. Our original conclusion, that the results of the study should be disregarded, remains unaltered.
Daru B, Yessoufou K, Davies J (2014). Speculation versus data-driven conclusions: A response to Gereau et al.’s "Phylogenetic patterns of extinction risk: the need for critical application of appropriate datasets". [PeerJ PrePrints 2:e323v1] (http://dx.doi.org/10.7287/peerj.preprints.323v1)
Gereau RE, Burgess ND, Fjeldså J, Hall J, Hemp A, Jump AS, Kajuni AR, Marchant RA, Marshall AR, Platts PJ, Taylor CM, Tibazarwa FI (2013). Phylogenetic patterns of extinction risk: the need for critical application of appropriate datasets. [PeerJ PrePrints 1:e55v1] (http://dx.doi.org/10.7287/peerj.preprints.55v1)
Yessoufou K, Daru BH, Davies TJ (2012). Phylogenetic patters of extinction risk in the Eastern Arc ecosystems, an African biodiversity hotspot. PLoS ONE 7(10): e47082. doi:10.1371/journal.pone.0047082