The comment by Hall and Sanders raises some issues we are able to clarify. These authors also criticize our study for its omissions, something we regard as inevitable in a one-off study of this nature. We contend that neither of these concerns invalidate our study's conclusions. Hall and Sanders's first criticism concerns sampling and interpretation and is prefaced by the rhetorical question "of whom can the authors legitimately speak?" At the time of the study, both authors had interacted with both communities in the Rewa Delta for more than 20 years, each author speaking one of their two vernacular languages (Fiji Hindi and Bauan-Fijian) and being intimate with their cultural mores, attributes that allowed us privileged access to these communities for the purpose of the study. Interviewee selection was not "haphazard". In both study sites, we were constrained in this by gender, age, language, relatedness, status, and religious affiliation, all of which affected our ability to freely speak to those we might have targeted had we not been so encumbered yet we are satisfied that the 64 people we interviewed (selected by age, gender and residence time) represented a cross-section of the target population in each community. Full details of interviewees are given in Lata's MSc thesis (Lata 2010), which was referenced several times in our paper (Lata and Nunn 2012). We are indeed somewhat "startled" that Hall and Sanders overlooked this. A requirement that potential interviewees must have been continuously resident in our sample locations for 30 years is not "troubling" to us. We used a 30-year figure because (1) our experience of gathering environmental-change data from Pacific Island communities suggests that this was the optimal period needed to comprehensively exclude persons who might give misleading information2 and (2) flood data in particular suggest that this is the period within which recent climate-change effects are detectable, something on which we sought to allow our informants to comment. As we explain in our paper, "although the specific data analysed in this paper were obtained from individual interviews, these were supplemented by focus-group discussions in appropriate cultural situations for the purposes of understanding both the broader context and canvassing group views" (Lata and Nunn 2012: 174).