This article evaluates theoretical claims linking relational uncertainty to experiences of stress during interactions with a partner. Two observational studies were conducted to evaluate the association between relational uncertainty and salivary cortisol in the context of hurtful and supportive interactions. In Study 1, participants (N = 89) engaged in a conversation about core traits or values with a partner, who was trained to be hurtful. In Study 2, participants (N = 89) received supportive messages after completing a series of stressful tasks and receiving negative performance feedback. As predicted, partner uncertainty was associated with greater cortisol reactivity to the hurtful interaction in Study 1. Contrary to expectations, Study 1 results also indicated that self uncertainty was associated with less cortisol reactivity, when self, partner, and relationship uncertainty were tested in the same model. Study 2 revealed that relational uncertainty dampened cortisol reactions to performing poorly on tasks while the partner observed. As predicted, Study 2 also found that partner uncertainty was associated with less cortisol recovery after the supportive interaction, but neither self nor relationship uncertainty was associated with rate of cortisol change during the recovery period.