Conceptual learning was assessed in a partially online physical geology course using the Geoscience Concept Inventory (GCI), developed by Libarkin and Anderson. Eight sections of the course were taught between spring 2013 and spring 2016 at San Diego Mesa College, a large, urban community college in southern California. A majority of students (86%) taking the partially online course were non-STEM majors seeking to satisfy graduation/transfer requirements. Online course elements included an interactive, web-based textbook, weekly discussions, chapter quizzes, and homework assignments. Face to face elements included a field trip, office hours, exam reviews, and exams. To maintain academic standards, students took exams on campus, in a proctored setting. The GCI, a multiple choice test bank applicable to entry-level college geoscience courses, has been used extensively to assess student conceptual knowledge across various courses. As an assessment tool, the GCI incorporates Rasch model analysis and has undergone extensive reliability and validation studies. Pairing pre- and post-course GCI scores for individual students reveals a mean, pre-course scaled GCI score of 46.5 ± 14.2 for the combined 8-section data set (n=93) and a post-course GCI score of 57.7 ± 13.9, an increase (Δ GCI) of 11.2%. Δ GCI values for individual sections range from 9 to 18%. These increases are substantially higher than the national average gain of 4% reported by others. As a group, partially online students showed improvement on 14 of 15 GCI test questions. Student conceptual knowledge decreased or showed minimal improvement on two questions focused on the early Earth, a result consistent with previous studies. Student feedback survey results indicate that 82% agree/strongly agree that the course was as challenging as a face to face course. Over two thirds of survey respondents (68%) would recommend/highly recommend the course to others. These results suggest that learning geoscience partially online can be a pedagogically sound and satisfying experience for students. Further work is necessary to better understand why student conceptions regarding the early Earth appear resistant to instruction.