Natural category learning is important in science education. One strategy that has been empirically supported for enhancing category learning is testing, which facilitates not only the learning of previously studied information (backward testing effect) but also the learning of newly studied information (forward testing effect). However, in category learning, such benefits of testing have mostly been examined without explicit instructions. This is not aligned with a real educational practice where teachers often provide students with explicit instructions that highlight the diagnostic features of the category. Thus, we investigated the effect of interim testing and feature highlighting on rock category learning and whether the provision of feature highlighting further enhances testing benefits. The participants learned 12 rock categories, which were divided into two sections (Sections A and B). They studied a series of rock images with or without feature highlighting and were given an interim test or not on Section A before proceeding to Section B. After Section B, all the participants took a final test in which they had to classify both old and new rock images of the studied categories. Three experiments demonstrated the benefits of interim testing (compared to restudy) for both previously and newly studied categories. However, feature highlighting did not further enhance learning and sometimes even impeded learning. The findings suggest that providing more information is not always better than providing less information in natural category learning.