The present research addresses two questions and goals. The first goal was to create and validate an alternative assessment tool for Spanish-speaking and Spanish-English bilingual children. In Experiment I we developed and selected the appropriate nonwords for the Spanish language. Experiment 2 compared Spanish nonword repetition performance for Spanish- and Spanish-English-speaking children with or without language impairment (LI). Numerous studies have documented that monolingual English-speaking children with LI differ from their peers on nonword repetition performance and have significantly more difficulty repeating three- and four-syllable nonwords. It was predicted that Spanish-speaking children with LI will have significantly more difficulty at Spanish nonword repetition performance, both in general and with the longer nonwords. The second research question examined the nature of language impairment in emerging bilingual populations. The various models of working memory (WM) differ in the explanations given for the underlying cause of the working memory deficits present in this population of children. A general limitation will cause the children with LI to perform significantly worse on all measures, while a specific limitation will produce errors specific to the verbal domain only. Both general processing and verbal processing measures were used to examine the nature of language impairment. It was predicted that the children with LI will demonstrate deficits on all measures, verbal and nonverbal, supporting the unitary claim that children with LI possess a general processing limitation. Experiment 2 demonstrated that the children with typical language development (TLD) performed significantly better than the group with LI on the Spanish Nonword Repetition task. The findings from Experiment 3 supported the unitary theory of WM; the combination of significant intercorrelations among the dependent variables, and the fact that the group with LI performed significantly worse on both the verbal and nonverbal measures provided evidence for a general working memory limitation. These children possess a working memory deficit--not a language impairment. The clinical implications of the experiments are exciting. The Spanish Nonword Repetition task has the potential of screening for WM deficits avoiding time consuming language sampling and transcriptions. Moreover, support for a general WM deficit has significant implications for therapy.