There is now a wide choice of medical imaging to show both focal and diffuse pathologies in various organs. Conventional radiology with plain films, fluoroscopy and contrast medium have many advantages, being readily available with low-cost apparatus and a familiarity that almost leads to contempt. The use of plain films in chest disease and in trauma does not need emphasizing, yet there are still too many occasions when the answer obtainable from a plain radiograph has not been available. The film may have been mislaid, or the examination was not requested, or the radiograph had been misinterpreted. The converse is also quite common. Examinations are performed that add nothing to patient management, such as skull films when CT will in any case be requested or views of the internal auditory meatus and heal pad thickness in acromegaly, to quote some examples. Other issues are more complicated. Should the patient who clinically has gall-bladder disease have more than a plain film that shows gall-stones? If the answer is yes, then why request a plain film if sonography will in any case be required to 'exclude' other pathologies especially of the liver or pancreas? But then should cholecystography, CT or scintigraphy be added for confirmation? Quite clearly there will be individual circumstances to indicate further imaging after sonography but in the vast majority of patients little or no extra information will be added. Statistics on accuracy and specificity will, in the case of gall-bladder pathology, vary widely if adenomyomatosis is considered by some to be a cause of symptoms or if sonographic examinations 'after fatty meals' are performed. The arguments for or against routine contrast urography rather than sonography are similar but the possibility of contrast reactions and the need to limit ionizing radiation must be borne in mind. These diagnostic strategies are also being influenced by their cost and availability; purely pragmatic considerations are not infrequently the overriding factor. Non-invasive methods will be preferred, particularly sonography as it is far more acceptable by not being claustrophobic and totally free of any known untoward effects. There is another quite different but unrelated aspect. The imaging methods, apart from limited exceptions, cannot characterize tissues as benign or malignant, granulomatous or neoplastic; cytology or histology usually provides the answer. Sonography is most commonly used to locate the needle tip correctly for percutaneous sampling of tissues. Frequently sonography with fine needle aspiration cytology or biopsy is the least expensive, safest and most direct route to a definitive diagnosis. Abscesses can be similarly diagnosed but with needles or catheters through which the pus can be drained. The versatility and mobility of sonography has spawned other uses, particularly for the very ill and immobile, for the intensive therapy units and for the operating theatre, as well in endosonography. The appointment of more skilled sonographers to the National Health Service could make a substantial contribution to cost-effective management of hospital services. Just when contrast agents and angiography have become safe and are performed rapidly, they are being supplanted by scanning methods. They are now mainly used for interventional procedures or of pre-operative 'road maps' and may be required even less in the future as MRI angiography and Doppler techniques progress. MRI will almost certainly extent its role beyond the central nervous system (CNS) should the equipment become more freely available, especially to orthopaedics. Until then plain films, sonography or CT will have to suffice. Even in the CNS there are conditions where CT is more diagnostic, as in showing calculations in cerebral cysticercosis. Then, too, in most cases CT produces results comparable to MRI apart from areas close to bone, structures at the base of the brain, in the posterior fossa and in the spinal cord. Scintigraphy for pulmonary infarcts and bone metastases and in renal disease in children plays a prominent role and its scope has increased with new equipment and radionuclides. Radio-immunoscintigraphy in particular is likely to expand greatly not only in tumour diagnosis but also in metabolic and infective conditions. Whether the therapeutic implications will be realized is more problematic. The value of MRS and NM for metabolic studies in clinical practice is equally problematical, although the data from cerebral activity are extremely interesting. While scanning has replaced many radiographic examinations, endoscopy has had a similar effect on barium meals and to a lesser extent on barium enemas. The combined visual/sonographic endoscope is likely to accelerate this process. There is no doubt that over the last 2 decades medical imaging has changed the diagnostic process, but its influence on the outcome of disease other than infections is less certain and probably indefinable. Data concerning the comparative efficacy in terms of patient outcome for each of the imaging techniques would be of considerable interest and a great help in determining diagnostic strategies.