Heredity has been dismissed as an insignificant object in Claude Bernard's physiology, and the topic is usually ignored by historians. Yet, thirty years ago, Jean Gayon demonstrated that Bernard did elaborate on the subject. The present paper aims at reassessing the issue of heredity in Claude Bernard's project of a "general physiology". My first claim is that Bernard's interest in heredity was linked to his ambitious goal of redefining general physiology in relation to morphology. In 1867, not only was morphology included within experimental physiology, but it also theoretically grounded physiological investigations. By 1878, morphology and physiology were considered as completely independent sciences, and only the latter was perceived as suitable to experimentation. My second claim is that this reversal reflected the existence of two opposite attitudes towards heredity. In the late 1860s, Bernard was convinced that heredity would soon be accessible to experimental manipulation and that new species would be produced in the laboratory exactly like organic chemistry succeeded to do for raw bodies. Ten years later, he ascertained that this was impossible. My third claim is that Bernard was epistemologically ill-equipped to address the issue of heredity. Bernard was strongly committed to a general reasoning scheme that acknowledged only three categories: determining conditions, constant laws and phenomena. This scheme was a key factor in his successes as a physiologist who was able to capture new mechanisms in living bodies. Nonetheless, it also prevented him from understanding how time and history could be endowed with a causal action that cannot be reduced to timeless parameters.