The Asynchronous $\pi$-calculus, proposed by Honda and Tokoro (1991) and, independently, by Boudol (1992), is a subset of the $\pi$-calculus \cite{Milner:92:IC} which contains no explicit operators for choice and output-prefixing. The communication mechanism of this calculus, however, is powerful enough to simulate output-prefixing, as shown by Honda and Tokoro (1991) and by Boudol (1992), and input-guarded choice, as shown by Nestmann and Pierce (2000). A natural question arises, then, whether or not it is as expressive as the full $\pi$-calculus. We show that this is not the case. More precisely, we show that there does not exist any uniform, fully distributed translation from the $\pi$-calculus into the asynchronous $\pi$-calculus, up to any ''reasonable'' notion of equivalence. This result is based on the incapability of the asynchronous $\pi$-calculus to break certain symmetries possibly present in the initial communication graph. By similar arguments, we prove a separation result between the $\pi$-calculus and CCS, and between the $\pi$-calculus and the $\pi$-calculus with internal mobility, a subset of the $\pi$-calculus proposed by Sangiorgi where the output actions can only transmit private names.