In vivo toxicological testing designed to assess acute oral toxicities of substances is still required by regulatory authorities for purposes such as classification and labelling, risk assessment and management, in support of public health protection. Traditionally, these tests were based on determination of the LD50 (lethal dose to 50% of animals) to give an indication of potency, based on a simple point estimate numerical comparison between substances. Today there is a choice of three test guidelines for acute oral toxicity studies (TGs 420, 423 and 425 of the Organisation for Economic Co-operation and Development (OECD)). Whilst TG 423 and TG 425 use death as an endpoint, TG 420 replaces death with 'evident toxicity', defined as clear signs that exposure to a higher dose would result in death. Thus, TG 420 avoids the need to carry out testing at higher doses. However, the subjective nature of 'evident toxicity' appears to be preventing wider uptake of TG 420 in a regulatory context. The NC3Rs and the European Partnership for Alternative Approaches to Animal Testing (EPAA) have collaborated to provide evidence and guidance on the recognition of 'evident toxicity', to make the decision more objective. Historical data on individual animal clinical signs recorded for acute oral studies were collected for 250 substances from 11 different companies. Inclusion criteria were applied (e.g., at least two doses per substance, no death at lower dose, death at higher dose) and pairs of studies (males or female rats) with at least a six-fold ratio between doses (higher, lower dose) were analysed to determine if any signs at the lower dose could have predicted death at a higher dose. The results from 80 different substances show that signs such as ataxia, laboured respiration, and tremors, if seen at least once in at least one animal, are highly predictive of death at the next highest dose (Positive Predictive Value (PPV) = 100%). Signs such as lethargy (PPV >90%), decreased respiration, nasal-ocular lacrimation and hunched posture were also very predictive (PPVs >85%). Combinations of signs such as nasal-ocular discharge and/or laboured respiration and/or ataxia and/or tremors were also highly predictive (PPV = 100%). The data have been used to develop guidance on the recognition of 'evident toxicity' for integration more broadly as the study endpoint, to reduce the suffering and numbers of animals used when in vivo acute oral toxicity studies are required.