Several studies have shown that adding seductive details to instructional materials has a detrimental effect on learning. However, other studies have shown non-significant findings. The present study uses cognitive load theory as a theoretical framework to explain these controversial results in seductive details research. Using a 2 × 2 experimental design we asked a group of high-school students (N = 100) to learn about biology with a multimedia environment that manipulated the presence of seductive details (with vs. without) and the modality of the verbal information (high load, on-screen text vs. low load, narration). The findings showed that students' learning performance was significantly higher when seductive details were presented under the low load condition (narration) as compared to all other conditions. The theoretical implications for understanding the effects of non-redundant and interesting, but irrelevant learning material are discussed and future research directions are presented.