Two complementary studies were conducted. In Study 1 (categorical perception of alerts), a series of experiments examined the key parameters that result in listeners perceiving a given sound as an urgent warning versus other less urgent message categories. Initial perceptual sorting experiments identified the most significant parameters and subsequent driving simulator experiments confirmed and refined the findings. Study 1 found that four auditory characteristics accounted for most of the variance in classification of auditory alerts as urgent warnings. Sounds were classified as alarms over 90 percent of the time when they had a peak-to-total-time-ratio (ratio of the time that the sound is at full intensity to the entire pulse duration including onset and offset) of greater than or equal to .7, an interburst interval of less than or equal to 125 ms, at least 3 harmonics, and a base frequency of greater than or equal to 1000 Hz. These results were observed initially in laboratory studies, and replicated during simulated driving. In Study 2 (warning perception in ambient noise environments), an experiment was conducted to investigate the effects of different in-vehicle ambient noise conditions on auditory signal detection and perception. Participants driving on a freeway experienced three ambient noise conditions (windows closed and no music, music playing, front windows open). Fifteen auditory alerts, presented at 65 or 75 decibels (A-weighted), occasionally occurred. Participants pressed a button as soon as they detected the sound, then provided ratings of the noticeability, urgency, and intended meaning of the sound. Study 2 found significant main effects for ambient noise condition and for alert sound for perceived noticeability, urgency, and response time to alert. Detection was impaired by the presence of music, and even more so with the front windows open. Even when auditory signals were heard, noise conditions modified their perceived urgency and meaning. There were also interactions between ambient noise condition and sound, indicating differences in how well sounds of similar loudness tolerated interference from noise. Results also demonstrate that the perceived urgency and meaning of auditory messages can change under noisier ambient conditions, and some features of more noise-resistant signals were suggested by the data. The findings of this research may help interface designers to create auditory signals that indicate the appropriate type and urgency of message.