Changing relationships with nonhuman animals have led to important modifications in animal welfare legislations, including the protection of animal life. However, animal research regulations are largely based on welfarist assumptions, neglecting the idea that death can constitute a harm to animals. In this article, four different cases of killing animals in research contexts are identified and discussed against the background of philosophical, societal, and scientific-practical discourses: 1. Animals killed during experimentation, 2. Animals killed before research, 3. “Surplus” animals and 4. (...) “Leftover” animals. The scientific community and, accordingly, animal research regulations such as the internationally acknowledged framework 3R (“Replace”, “Reduce”, “Refine”) tend to aim at the reduction of “surplus” and, to some extent, “leftover” animals, whereas the first two classes are rather neglected. However, the perspective that animal death matters morally is supported by both societal moral intuitions and certain theoretical accounts in animal ethics. Therefore, we suggest the implementation of the 3Rs in regulations, so that they: 1. Make their underlying philosophical position transparent; 2. Are based on a weighing account of animal death; 3. Are applicable to procedures on living and dead animals; 4. Apply the “reduction” principle to procedures on dead animals; 5. Entail that methods using (parts of) dead animals need to be replaced by animal free methods, if possible; 6. Do not suggest replacing research on living animals by research on killed animals; 7. Include all kinds of animals, depending on the respective harm of death; 8. Are applied to the broader context of experimentation, including breeding and the fate of the animals after the experiment. (shrink)