Adenovirus is a robust vector for therapeutic applications, but its use is limited by our understanding of its complex in vivo pharmacology. In this review we describe the necessity of identifying its natural, widespread, and multifaceted interactions with the host since this information will be crucial for efficiently redirecting virus into target cells. In the rational design of vectors, the notion of overcoming a sequence of viral "sinks" must be combined with re-targeting to target populations with capsid as well as shielding the vectors from pre-existing or toxic immune responses. It must also be noted that most known adenoviral pharmacology is deduced from the most commonly used serotypes, Ad5 and Ad2. However, these serotypes may not represent all adenoviruses, and may not even represent the most useful vectors for all purposes. Chimeras between Ad serotypes may become useful in engineering vectors that can selectively evade substantial viral traps, such as Kupffer cells, while retaining the robust qualities of Ad5. Similarly, vectorizing other Ad serotypes may become useful in avoiding immunity against Ad5 altogether. Taken together, this research on basic adenovirus biology will be necessary in developing vectors that interact more strategically with the host for the most optimal therapeutic effect.