The structure of the Atlantic Forest (AF) has been studied for almost 70 years. However, the related existing knowledge is spread over hundreds of documents, many of them unpublished and/or difficult to access. Synthesis initiatives are available, but they are restricted to only a few parts or types of the AF or are focused on species occurrence. Here, we conducted an extensive review to compile quantitative tree community surveys on all types of the AF until 2013 and to study where and how these surveys were conducted. We found 1157 relevant references, containing 2441 forest surveys published since 1945. These surveys corresponded to 2.24 million trees and 1817 ha of forests sampled. This total sampled area represents only 0.01 % of the AF remnants, showing how limited our knowledge is on AF structure. For Paraguay and the Brazilian states of Bahia and Mato Grosso do Sul this proportion was much smaller. The same was true for evergreen rainforests, Brejos de altitude and deciduous forests and most probably for the rare cloud, swamp, Caxetal and Mussununga forests for which no accurate remnant estimates were Communicated by Jefferson Prado, Pedro V. Eisenlohr and Ary T. de Oliveira-Filho.Electronic supplementary material The online version of this article (found. Since the 1980s, the amount of AF area sampled each year has increased continuously, but approximately 100 years will be necessary to sample at least 1 % of the AF. Thus, we urgently need an enormous amount of high-quality quantitative data to overcome our limited knowledge of the AF and to support conservation programs aiming to safeguard this threatened biodiversity hotspot.

Welcome small patches; beware of the risks of changing conservation prioritiesRecently, Riva and Fahrig revisited the SLOSS question with fresh data from Chase et al. (2019) and concluded that small patches (SP) are disproportionately important for biodiversity conservation compared to large patches (LP). They recommended that conservation priorities, canonically oriented toward LP, should shift to recognize the value of SP. We appreciate their new analysis, agree that SP have outstanding importance for maintaining diversity, and understand they do not suggest neglecting LP. However, we want to clarify two points. First, the political reality is that conservation decisions are rarely as clean cut as choosing to conserve one large patch versus several SP. Second, time lags and spatial dependencies in extinction dynamics severely complicate the interpretation that SP are better for biodiversity conservation.Conservation biology does not happen in a void, and studies that seek to influence decision-making need to address political and institutional reality. Many countries, especially in the tropics, face recurring threats to biodiversity, including inside protected areas, and these threats are not distributed evenly across patch sizes, occurring more severely on SP (e.g., Reid et al., 2019). Even large reserves with highly restrictive human use face multiple threats including deforestation, changes in protection status, and/or area reduction. Moreover, the dismantling of environmental laws and institutions in charge of environment protection, currently happening in several countries (e.g., Vale et al., 2021), put at risk the role current reserves of any size have in protecting highly sensitive species and those that require large habitats.Furthermore, the detection of high species richness and globally endangered species in SP, even the older ones, does not guarantee the local presence of viable populations. These patterns may result from past landscape connectivity, especially in highly dynamic landscapes (e.g., Atlantic Forest). Although fragmented for decades, landscapes may have only recently crossed extinction thresholds (Figure S2), sheltering declining populations andThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.