This thesis examines sound production of eastern gray whales (Eschrichtius robustus) in the wintering lagoons to determine whether sound use is a function of social context. Proportions of sounds used, parameters of each sound class, and rates of sounds were compared among social contexts. Data revealed the strong possibility of context-specific use for particular sound classes. Additionally, sound parameters and rates of production varied by social context. These results reflect similar variations in gray whale repertoire throughout their range that may be due to changes in social and behavioral contexts. Gray whale sounds are classified into several classes based on aural and visual characteristics. This study verifies the classification system determined in previous studies, with the exception of class 8, and supports the division of class 1 into subclasses 1a and 1b. Class 1 appeared to be critical during sexual contexts and all highly social contexts, regardless of age and sex class. Although highly recognizable, its parameters exhibited much variation among social contexts; therefore class 1 may communicate graded emotional states in short-range interactions. Other classes of sounds may be utilized for long-distance communication, as startle responses, or "precursors" to the adult repertoire. Frequency-related parameters of all sound classes showed variation among social contexts, but duration demonstrated very little variation. Calf-containing contexts exhibited greatest and most varied frequencies; this is to be expected if gray whale's sound mechanism is related to body and tracheal length. Variation also may indicate that physical maturity or learning play a role in the repertoire development. The lowest and least varied frequencies were observed in adult contexts. Sound production rates also varied by social context. Active adults produced sounds at high rates during short intervals; mixed/unknown contexts were often silent. Calf-containing contexts produced sounds at intermediate rates and were never silent. The correlations demonstrated here between social context and use of sounds will allow for acoustics to be an indicator of group composition, seasonal movements, and social patterns, thus relieving dependency on difficult visual observation. Additionally, such correlations provide preliminary information for determining sound functions.