This paper presents a preliminary description and analysis of the intonational patterns and phrasal domains involved in simple declarative and interrogative sentences in Cahuilla, an understudied Uto-Aztecan language of Southern California. To my knowledge, no such description or analysis on the topic of Cahuilla intonation has been proposed for any of its dialects or varieties. Following the basic principles and assumptions of the autosegmental-metrical (AM) framework for intonation originally set forth by Liberman (1975), Bruce (1977), Pierrehumbert (1980), and later rearticulated by Ladd (2008) and others, I describe and analyze intonational data from a single, elderly fluent speaker of the Torres-Martinez variety of the Desert Cahuilla dialect, which I then use to propose a tentative intonational tonal inventory and model for Cahuilla. Since this study is focused primarily on the annotation of intonational pitch accents and contours for the language, I leave aside the issue of developing a break indices system for future research. Thus, this study can be considered a first step in the development of a ToBI-like system (as outlined in Silverman et al. (1992); Beckman and Hirschberg (1994); Beckman and Ayers (1997); and Beckman et al. (2005)) for annotating intonational patterns in Cahuilla. Furthermore, given the paucity of intonational research on American Indian languages (e.g., Chickasaw; Gordon 2005, Lakota; Mirzayan 2010), this paper contributes to a building body of cross-linguistic research, being spearheaded by Jun (2005, 2014) and others, on the ability of an AM-based system of annotation like ToBI to model a broad range of intonation systems.