Since they were introduced more than 5 decades ago, a variety of single- pass indicator, thermal and osmotic dilution approaches have been developed for detecting and measuring excess fluid in the lungs. This brief review discusses why studies of the extravascular lung water (EVLW) continue to intrigue physiologists and clinicians and the likelihood that they will become sufficiently reliable for more widespread use. Emphasis is placed on the basic assumptions that underlie these measurements and limitations imposed by the nature of the data that are collected. A distinction is made between approaches that are based upon compartmental models of solute and water exchange, and those that represent extensions of more conventional washout procedures, which have been utilized extensively for measurements of gas volumes in the lungs. Although the compartmental approach has been used to simplify indicator dilution studies by eliminating the need for a vascular indicator, it is based upon assumptions that may not be realistic. Early recirculation inevitably limits the period in which observations can be made and impairs detection of those portions of the lungs with decreased perfusion. These general principles are also used to develop a new method of analyzing osmotic transient studies. A short account is given of EVLW observations that have been made in animals and man. Both the sensitivity and specificity of extravascular lung water measurements in man are uncertain and the normal clinical range of EVLW remains in doubt.