This chapter considers the likelihood that a wide spectrum of infection-provoked arthritis exists, ranging from overt sepsis to apparently aseptic chronic arthritis in which very small numbers of causal bacteria can be detected only by using highly sensitive techniques. It asks whether joints are, as conventionally held, normally devoid of micro-organisms and how to judge the significance of bacteria detected within apparently sterile joints. Through a consideration of known septic, probably infective and apparently aseptic forms of arthritis, a set of criteria for attributing causality to putative arthritogenic micro-organisms is proposed.