Since lymphangiography has been advocated and employed in man as a means of demonstrating lymph vessels and nodes (2, 4–12, 14), one might justifiably question the timing of an animal study. Realizing the difference in lymphatic distribution in dogs (1), we nevertheless believe that observations on these animals serve to emphasize the potential hazards of the procedure. Some published accounts lead one to believe that lymphangiography is innocuous. There have been reports of the visualization of droplets of contrast medium in the lungs following its performance, but the danger of pulmonary embolization has been minimized (5, 9, 13). To the authors, the unpleasant pulmonary side-reactions frequently observed clinically indicate that the examination may not be without risk. As demonstration of the lymph nodes in a variety of pathologic states is highly desirable, an attempt has been made to promote the safety of the procedure by modification of the vehicles for introduction of the contrast substance. In this report, the hazards as found in dogs will be amplified, and preliminary attempts to produce a satisfactory medium will be discussed. Animal Study For the first part of this investigation, accepted methods of injecting water-soluble (4, 10) and oil-base media into dogs were employed. The injections were made into the lymphatics of the hindpaw according to Kinmonth's technic (6). Interval and twenty-four-hour lymphangiograms of the extremities, abdomen, and chest were obtained routinely. Water-soluble media were injected initially in the opposite paws of many of the experimental animals. Records were kept of untoward pulmonary and/or central nervous system effects. After the twenty-four-hour radiographs, the dogs were sacrificed and autopsies were performed. Postmortem radiographs of the brain, lungs, spleen, liver, and kidneys were made to supplement the usual autopsy procedures. In all, 15 dogs were used. In some, both oil-base and water-soluble media were injected on different occasions. An analysis of the autopsy findings disclosed no evidence of infarction, embolization, or radiographically demonstrable oil droplets in the brain, liver, spleen, or kidneys. In contrast, the lungs displayed marked changes in proportion to the volume of oily contrast medium injected. Multiple dark, ecchymotic, airless zones, ranging in diameter from a few millimeters to several centimeters, were found, with the gross and microscopic characteristics of pulmonary infarcts (Fig. 1). In every instance, oil dispersed throughout the lung fields within the pulmonary vessels (Figs. 2 and 3) was evident radiographically. The second part of the study was directed to modification of the oily medium so that its radiopaque properties would be retained while at the same time it would become more soluble in the blood stream, thereby obviating the problem of embolization. The oil was emulsified in a mixture utilizing the following formula: