Rediscovery of a species feared to be extinct generates conservation optimism and positive media attention. The hope of evading extinction contrasts with the gloom of the biodiversity crisis, although the outlook for rediscovered animals is in fact grim, particularly if they are small and uncharismatic species as is increasingly the case (Fisher, 2011a). To be rediscovered a species must first be missing, but how do we know when a species is missing, possibly undetected or possibly extinct? Scientifically documented yearly distributions of sightings vary among species and depend on search effort (Fisher & Blomberg, 2011), location remoteness in relation to searchers (who are typically not local practitioners in biodiversity hotspots (Fisher, 2011b)), identification expertise available, species distinctiveness, and features that affect conspicuousness of the animal such as body size and microhabitat (Fisher & Blomberg, 2011). In 1994, the IUCN recognised that the appropriate time missing to detect extinction varies between taxa. Based on the work of the late Prof. Georgina Mace and colleagues, the IUCN Red List definitions changed to reflect evidence-based criteria (IUCN, 1994). Now Species Survival Commission Specialist Groups and their expert assessors judge if search effort has been adequate and appropriate for each species. Authors typically provide this rationale in the Red List species account. A component of ‘appropriate’ surveys is of course the number of years of effort, but time missing is not an explicit criterion. Martin et al. (2022) acknowledge this complexity but have taken a simple approach to define lost terrestrial vertebrates (mammals, birds, amphibians and reptiles) as species not recorded by science for at least fifty years, evaluated in the IUCN Red List of Threatened Species, and not designated as EX (extinct). They identify 562 lost species, more than 90% of which are classified as Critically Endangered or Data Deficient by the IUCN. More than 90% of these lost species are tropical, concentrated especially in countries with diverse habitats, islands and mountain-tops where distributions are restricted, and remote forests, such as in Indonesia, Mexico and Brazil. Nearly a third of lost terrestrial vertebrates are island endemics and a third mountain endemics. More than 60% of EX species are island endemics (Spatz et al., 2022). In mammals, missing island endemics are extremely unlikely to ever be rediscovered because their small ranges are searchable, and the particularly lethal driver of invasive predators is their main cause of extinction (Fisher & Blomberg, 2011, Clavero & Garcia-Berthou, 2005), which suggests that many of the ~170 lost island endemics are extinct, although they lack a ‘Possibly Extinct’ flag. These findings enhance our understanding of extinction and suggest ways to improve the value of information in Red List accounts by increasing the use of the ‘Possibly Extinct’ flag. Martin et al. (2022) suggest that the PE flag should be applied more widely when consistent with recently published guidelines. They also propose that a separate last sighting date field in Red List accounts should be routinely included with species classified below Critically Endangered. These measures would help to quantify the change in status of lost species in future. This study also contributes to our understanding of extinction because it provides an independent source of evidence of reliable correlates of species extinction and detectability. There is a striking agreement between the approach of Martin et al. (2022) and other work on the numbers and features of missing and extinct terrestrial vertebrates. For example, despite different methods and thresholds, the conclusions of Martin et al. (2022) are remarkably similar to those of Re:wild (2022) (non peer-reviewed research published online by NGO Re:wild in collaboration with the IUCN and others). Re:wild researchers contacted SSC Species Specialist Groups to collate a list of species judged to be missing by taxon experts based on their knowledge of survey effort, taxonomy, and other factors. The Re:wild database thus broadly uses IUCN criteria, and includes ‘NE (not evaluated)’ species and ‘EX’ species judged to be lost by experts. Re:wild (2022) identified 1008 lost terrestrial vertebrates, and a third of the total lost species were NE. The taxon expert approach thus concluded a similar number of lost terrestrial vertebrate species to the 50-year rule approach applied by Martin et al. (2022), if we omit NE and EX species (excluded from Martin et al. definition of ‘lost’) from the Re:wild list. Martin et al. (2022) emphasised the worrying disparity between the proportion of lost and extinct reptiles, and show graphically that reptiles had more species still missing after longer time periods than other vertebrate classes, including reptile species lost for more than 170 years. They conclude that reptile extinction is likely to be under-diagnosed. Re:wild (2022) also agreed that by far the most represented class of missing species is reptiles (a third of the missing species in their database are reptiles, 582 species). Both conclude that the least represented class of lost species is birds, probably because bird extinction is most often correctly diagnosed. Martin et al. (2022) and Re:wild (2022) both showed that tropical taxa dominate lost species lists, especially in Indonesia. Methods to quantify extinction and its detection in vertebrates based on trait modelling, sighting distribution modelling, chronological Red List changes and literature reviews corroborate these numbers and the predominance of tropical species in the ‘lost’ category (e.g. Fisher & Blomberg, 2011, Fisher, 2011b, Scheffers et al., 2011, Butchart et al., 2018). Martin et al. (2022) advocated a dynamic and accessible list of lost species: This is also the main purpose of Re:wild's (2022) online database. The suggestion by Martin et al. (2022) to address taxonomic uncertainty by focusing on priority data-deficient lost species may be problematic because the reason for data deficiency in lost species is typically the lack of specimens. For example, Fisher & Blomberg (2012) found that 70% of missing/extinct mammals had five or fewer records and 45 species had only one record. However, Martin et al.'s (2022) recommendation to fund targeted searches for priority missing species makes sense, and is in agreement with conservation NGOs including Re:Wild and EDGE. Martin et al. (2022) recommended models to prioritise species that are most likely to be rediscovered and sites where they may occur, using attributes of lost /extinct and rediscovered taxa. This is also potentially useful; several such studies have been published for birds and mammals, and also plants, but not herpetofauna (e.g. Fisher, 2011b, Fisher & Blomberg, 2012, Butchart et al., 2018, Humphreys et al. 2019). This study re-enforces the importance of using a species-by-species approach to extract IUCN Red List data. Martin et al. (2022) and Re:wild (2022) both began by omitting potential lost species that have been seen in the last ten years, then (as in Fisher & Blomberg, 2011; Butchart et al., 2018) Martin et al. (2022) extracted relevant data by reading individual species accounts. Species Specialist Groups do not monitor species, but rather taxon experts provide assessments. SSGs are run by volunteers, so although there are some annual updates, it may take several years to draft new versions of individual species accounts, which are asynchronously built on prior versions. The content of species accounts varies, often with the quantitative information in the main text of the justification section or another field. Although time consuming, to avoid bias and incorrect conclusions, it is crucial to read every species account carefully and to check the primary sources (Clavero & Garcia-Berthou, 2005), as Martin et al. (2022) have done. Open access publishing facilitated by The University of Queensland, as part of the Wiley - The University of Queensland agreement via the Council of Australian University Librarians. Open access publishing facilitated by The University of Queensland, as part of the Wiley - The University of Queensland agreement via the Council of Australian University Librarians.