Cycling pesticides has been proposed as a means of retarding the evolution of resistance, but its efficacy has rarely been empirically tested. We evolved populations of Chlamydomonas reinhardtii in the presence of three herbicides: atrazine, glyphosate and carbetamide. Populations were exposed to a weekly, biweekly and triweekly cycling between all three pairwise combinations of herbicides and continuously to each of the three herbicides. We explored the impacts of herbicide cycling on the rate of resistance evolution, the level of resistance selected, the cost of resistance and the degree of generality (cross-resistance) observed. Herbicide cycling resulted in a diversity of outcomes: preventing evolution of resistance for some combinations of herbicides, having no impacts for others and increasing rates of resistance evolution in some instances. Weekly cycling of atrazine and carbetamide resulted in selection of a generalist population. This population had a higher level of resistance, and this generalist resistance was associated with a cost. The level of resistance selected did not vary amongst other regimes. Costs of resistance were generally highest when cycling was more frequent. Our data suggest that the effects of herbicide cycling on the evolution of resistance may be more complex and less favourable than generally assumed.