Two groups of patients were studied by double contrast arthrography of the knee, and the incidence of popliteal cysts was correlated with age and sex of the patients and the presence of intra-articular lesions. The cysts were visualized most clearly on the lateral roentgenogram made with the knee in 90 degrees of flexion. One group was composed of 247 patients (211 male and thirty-six female) with an average age of twenty-seven years. The other group included 202 patients (ninety-seven male and 105 female) with an average age of forty years. On the basis of these studies and a review of the literature the following conclusions were drawn: Although there may be an anatomical communication between the gastrocnemius-semimembranosus bursa and the knee joint in many individuals, the communication is demonstrable by arthrography in most cases only after some intra-articular lesion has caused a joint effusion for a significant length of time. Recent acute minor injuries in young people rarely are associated with identifiable cysts. When cysts are present in this age group they usually are small and during arthrography they frequently fill with air only. In the absence of a clinically obvious popliteal cyst it is most unusual to demonstrate a communication between the joint space and the gastrocnemius-semimembranosus bursa in anyone under the age of sixteen. There is no difference between the clinically apparent popliteal cysts seen in adults and the cysts which can be demonstrated only by arthrography. The arthrographic appearance is the same in both groups and in both groups there is an equally high incidence of intra-articular lesions, usually meniscal tears, associated with the cysts. The prevalence of popliteal cysts is directly proportional to the age of the patient. Although there tend to be fewer popliteal cysts in females, this difference appears to be related to a lower incidence of intra-articular lesions in females, as reported in numerous series, rather than to intrinsic sex differences. Although the most common lesion associated with a popliteal cyst is a tear of the posterior horn of the medial meniscus, this incidence is no greater than the usual ratio of medial tears to lateral tears reported in the literature. Therefore, there is nothing specific about a medial meniscus injury that makes it any more likely to cause a cyst to form than any other effusion-producing abnormality. With rare exceptions, a clinically apparent popliteal cyst in an adult will be shown to communicate with the joint space during arthrography provided the patient can flex the knee actively after injection of the contrast medium, an adequate amount of contrast material is used, no constricting bandage is placed over the popliteal space, and a lateral roentgenogram is obtained with the knee in 90 degrees of flexion. The wide differences in the prevalence of cysts demonstrable by arthrography, as reported in the literature, are the result of several factors including the different ages of the patients in the various series, the varying severity and duration of the intra-articular lesions, the number of patients with clinically evident cysts included in the different series, and variations in the technique of arthrography. Because the evidence overwhelmingly supports the concept that popliteal cysts in the adult are usually secondary to some intra-articular pathological process, simple surgical removal of the cyst is not warranted, except on rare occasions, without a pre-arthrotomy arthrogram to determine more precisely the etiology of the cyst and the nature of the pathological lesion or lesions which so often accompany it.