This study investigates how learners with distinct native dialects, Cuban (CS) and Peninsular Spanish (PS), produce the English vowels æ, Λ, a/. Experiment 1 compares the native vowels along several acoustic parameters to determine the extent of the cross-dialectal differences, and Experiment 2 assesses the differences in L2 production of the same speakers who are also advanced learners of English. More fronted and shorter vowels are predicted for PS versus CS. As such, L2 /æ, A, a/ are expected to be more fronted in the interlanguage of PS learners as compared to CS learners. Dialect-specific patterns of assimilation are highlighted. PS learners produce æ, Λ, a/ with good spectral differences but with no duration differences whereas for CS learners /Λ, a/ overlap spectrally but are realized with different durations. Differences found in L2 production are caused by the conjoint effect of the native dialect, the input, and the learning experience.