Much of the recent discussion surrounding the second language acquisition of morphology has centered on the question of whether learners can acquire new formal features. Lardiere's (2008, 2009) Feature Reassembly approach offers a new direction for research in this area by emphasizing the challenges presented by crosslinguistic differences in the overt expression of formal features. In this study, I examine the acquisition of number and gender in Swahili by speakers of English and explore how the data can be described by a number of current approaches, including the Full Transfer Full Access Hypothesis (Schwartz and Sprouse, 1996), the Representational Deficit Hypothesis (e.g. Hawkins and Chan, 1997), and the Feature Reassembly approach. The results of an elicited production task and a written gender-assignment task indicate that learners have difficulty detecting the number feature on Swahili noun prefixes, and because of this they are initially unsuccessful at marking plurals. The findings are best described under a Feature Reassembly approach. I suggest some directions for expanding the Feature Reassembly approach in future research.