The context for the study was a naturally occurring quasi-experiment in the core mathematics program in a large Australian university. Delivery of teaching was changed in a sequence of two initial core mathematics subjects taken by engineering and science students. The change replaced one of two face-to-face tutorial classes per week by an online tutorial. Tasks in the online tutorial were designed to lead the students through the week's topics, using initially simpler tasks as scaffolding for more complex tasks. This was the only change: syllabus and written materials were the same, as was students' access to help from staff and discussion with peers. The study compared learning outcomes among students in two adjacent years: Cohort 1, the last before the change, and Cohort 2, in the first implementation of the change to a blended learning environment. Learning outcomes were assessed by a method derived from the SOLO taxonomy, which used a common scale for scoring written answers to examination questions in the two cohorts. In the first mathematics subject students doing online tutorials had significantly higher scores than those studying before the change. In the second mathematics subject there were no significant differences. The conclusion was that the online tutorials gave an advantage to students beginning university study and gave adequate support to those in the subject taken a little later. It can be concluded that the use of an online teaching component in the delivery of university mathematics programs is not only justifiable but desirable, subject to careful design of the teaching material offered.