A genetics problem practice program and tutor on microcomputer was used by 135 undergraduate education majors enrolled in an introductory biology course at Purdue University. The program presented four genetics problems, two monohybrid and two dihybrid, and required the users to predict the number and type of each class of offspring. Student responses were recorded on diskette and analyzed for evidence of misconceptions and difficulties in the genetics problem-solving process. Three main areas of difficulty were identified: difficulties with computational skills, difficulties in the determination of gametes, and inappropriate application of previous learning to new problem situations.