It is well known in practice and research that e-mails evoke information overload. Existing contributions of e-mail overload often investigate the volume of e-mails but not the e-mail channel itself. In order to cope with the phenomenon, it is also necessary to know how much each component of a single e-mail contributes to the overall phenomenon. We look at one of the most promising components: the format of e-mail threads as a history of appended e-mail conversations. We propose cognitive load theory to explain how much e-mail threads contribute to the phenomenon of e-mail overload. We aim towards the detection of a split attention effect as a suboptimal presentation format, which engages a human brain in searching rather than in understanding. The experimental results support our hypothesis by showing that reading time of e-mail threads in comparison to conventional e-mails is longer and the amount of correct answers per second is lower.