The acoustic properties that listeners may rely on in both the production and perception of ethnolinguistic variation are an important yet poorly understood topic in modern sociolinguistics. Though several studies (Purnell, Idsardi and Baugh 1999, Tarone 1973, Walton and Orlikoff 1994) have found that individuals generally make accurate and reliable judgments of speaker ethnicity, scholars have had difficulty identifying the specific features that listeners react to in making judgments (Thomas 2015). There is also little research about the production side of these types of ethnic identification tasks and these studies have also often overlooked the potential role of intra-speaker variation in use of suprasegmental features. This study addresses this gap in the literature by focusing on two aspects of production that have been observed to differ between Mainstream U.S. English (MUSE) and African American Language (AAL): the use of the H* versus L+H* contours and length of peak delay interval. This analysis is based on a sample of eight male speakers with one black parent and one white parent and it specifically explores how these intonational features may be useful in the construction and performance of complex racial identities. Results of regression models indicate that these speakers do not employ pitch accent type in intraspeaker variation, but that they do differ in their use of peak delay, employing longer delay intervals in conversations with black interlocutors than in conversations with white interlocutors. Understanding how speakers employ these and other intonational variables in both intra- and interspeaker variation is an important step in further describing ethnolinguistic varieties as well as addressing the phonetic features that may contribute to linguistic racial profiling.