Mentoring is defined as a reciprocal relationship between a less experienced individual (protégé) and a more experienced individual (mentor) that has consistent, regular contact over a period of time and is intended to promote mutual growth, learning, and development within the career context (Haggard, Dougherty, Turban & Wilbanks, 2011; Kram, 1985; Ragins & Kram, 2007). Inherent in this definition of mentoring is that individuals learn, develop, and grow from their mentoring interactions. Despite this, limited research explores the learning that occurs from mentoring relationships. The purpose of this study is to examine what mentors learn from mentoring experiences and how these experiences relate to mentor outcomes. The outcomes include mentoring self-efficacy, mentor behavioral change intentions, and mentor learning. I draw on the relational mentoring perspective (Ragins, 2012) and social learning theory (Bandura, 1971, 1977) to ground my hypotheses. My hypothesized model addresses three broad research questions: 1) What do mentors learn from their experiences with their protégés? 2) How do mentoring experiences relate to learning? and 3) Under what conditions do mentoring experiences relate to learning? I conducted a mixed methods study in an academic setting. My population included professors in North American doctoral granting universities and the PhD students they mentored. After identifying interested professors (mentors), I asked mentors to send study information to their PhD students (protégés). Both quantitative and qualitative data were collected via electronic survey. Surveys were time-lagged with three waves over a four month period. I found that mentors gained mentoring-specific, occupational-specific, and relational knowledge from mentoring their protégés. Additionally, mentors identified specific changes they wished to make to their mentoring behavior from their experiences with their protégés. I found that protégé positive role behavior and mentor positive psychosocial behavior were directly and positively related to career mentoring self-efficacy. Mentor positive career behavior was positively related to behavioral change intentions and protégé positive role behavior was positively related to relationship quality. I found that relationship quality did not mediate the relationship between role behaviors and mentor learning outcomes. Finally, I did not find support for the moderating effects of internal attribution for relationship quality, growth mindset, and feedback seeking. This study makes four specific contributions to the management field. First, research in mentoring often confounds relationship quality with behavior (e.g., Eby, Butts, Lockwood, & Simon, 2004; Eby, Durley, Evans, & Ragins, 2008) and outcomes (e.g., Eby, Butts, Durley, & Ragins, 2010; Ragins, 2012). By studying role behaviors, relationship quality, and mentor learning outcomes as distinct constructs, I provide clarity and an avenue for future mentoring research. Second, this study contributes to the mentoring literature by demonstrating what and how mentors learn from mentoring experiences. A significant contribution of this study is the identification of three types of mentor learning and behavioral change intentions. Third, I examined the theoretical explanation for mentoring role behaviors and mentor learning outcomes. Whereas I found that relationship quality did not explain the relationship between role behaviors and mentor learning, leader-member exchange provides a promising avenue for future research. Finally, I introduced mentoring self-efficacy as an important outcome of positive mentoring relationships, with mentors experiencing increased self-efficacy through positive experiences with their protégés.