Several visual functions were examined before and during self-application of therapeutic doses of digitoxin. The ERG recordings showed a reduction of the critical flicker fusion frequency from 70 Hz to 35-40 Hz. Recovery of color vision after macular photostress was examined with Nagel's anomaloscope and Jaeger's tritanomaloscope. After digitoxin intake the matching range at the tritanomaloscope was enlarged. Following macular photostress the recovery time of both the Rayleigh and the Trendelenburg matches was significantly prolonged. The mesoptometer readings did not show any changes during the application of digitalis. This demonstrates that the process of neural adaptation is intact during digitoxin therapy. The error score in the Lanthony desaturated 15 panel test was slightly increased after digitalis application. The Farnsworth-Munsell 100-hue test showed a marked increase in the total error score. There were no changes in the standard panel D-15. The Ishihara pseudoisochromatic plates were read without mistakes after digitalis medication. The chromatic visual acuity for red-green and blue was examined by recording the maximum reading distance for the Velhagen pseudoisochromatic plates "65" (Jaeger's test) and "49". For both plates the reading distance was reduced by 50% during digitoxin therapy. Computerized perimetry by colored stimuli (Tübingen Automatic Perimeter by Aulhorn and Durst) did not reveal any definitive changes in sensitivity thresholds with therapeutic serum levels of digitalis. Spectral increment sensitivity for the isolated blue cone system (Wald-Marré approach) was not influenced by therapeutic doses of digitoxin. Topical digitoxin application (0.02 mg/10 ml) did not interfere with the pupillary light reflex (infrared pupillography) or with the accommodative amplitude or intraocular pressure. It did, however, result in toxic keratopathy with swelling of endothelial cells and edema of the corneal stroma and epithelium. All changes disappeared after withdrawal of digitalis.