In this paper we provide an overview of dorsovelar articulations and acoustics in several Australian Aboriginal languages, and we compare these results with data from English. We examine languages that have a single dorsal, as well as languages that have two dorsal places of articulation. Using direct palatography and F2 transition measures, we show that Australian languages appear to have a distinct velar target for each of the three major vowel contexts, with a high degree of coarticulation between each velar allophone and its following vowel target, whilst English has only two velar targets--back and non-back, with a lower degree of coarticulation between velar allophones and their corresponding vowel targets. Thus, whilst the range of allophonic variation for velars extends further back in the Australian Aboriginal languages than in English, there appears to be no difference in the articulation of velars in the front vowel context. Drawing on results from the biomechanics, language acquisition, speech perception and acoustics literatures, we suggest that this result may be due to conflict between systemic constraints imposed by the place-rich consonant systems of Aboriginal languages and universal acoustic constraints on the identity of front-velar sounds, which may contribute to the instability of such articulations and the relative infrequency of velar + high vowel combinations in the world's languages.