A common position taken in recent literature on the passive is that by-phrases are adjuncts (Bruening 2013, Legate 2014). This paper shows that passive by-phrases differ from adjuncts in terms of binding phenomena. From these facts, I conclude that by-phrases are not adjuncts. I show that the theory of the passive in Collins 2005 accounts for the data without stipulation, unlike other recent theories of the passive.