Route of administration of various drugs is an area of study to which specific attention must be paid in study of different HIV risks of drug use by various routes. If changes in route are seen in individuals or within populations, then study of these transitions in route may identify new approaches which could be developed in HIV prevention. The consideration in this paper is based around ten questions: (i) What is a transition? (ii) Do routes of administration vary by time and place? (iii) Is choice of route influenced by availability of drug paraphernalia? (iv) How does the context influence initial choice of administration, and possible subsequent transitions? (v) Are lapse and relapse meaningful concepts? (vi) Transitions: how much of it is going on? (vii) How much does change of route (with the same drug) signify a change of drug effect, its significance, or its relationship with other risk behaviour? (viii) Is change of route of use of one drug always accompanied by the same change of route of other drugs? (ix) Injectors/non-injectors and sharers/non-sharers: do these behavioural characteristics exist as categories or are they distributed along a continuum? (x) Are transitions reversible? This paper is accompanied by two research reports which describe explorations into the extent and nature of transitions amongst heroin users.