Previous research has shown that heavy-vehicle driver workload can be measured in a simulator with simple secondary tasks such as choice reaction time and immediate recall (Kantowitz, 1995). The present experiment replicates and extends these findings to complex tasks requiring drivers to use cellular phones and to interpret text message displays. Fourteen commercial drivers each drove eight simulator modules, each 100,000 feet in length. In-cab tasks were divided into cellular phone dialing, phone dialogue, text message reading, tachometer reading, clock reading, and manual radio tuning. These were evaluated in light and heavy traffic with and without pedestrian detection tasks. Complex tasks requiring message reading had the greatest impact upon driver performance. Results illustrate both advantages and limitations of driving simulators.