1. In 13 dogs used repeatedly as donors for plasmapheresis and plasma injection experiments and fed a kennel diet that for a period of over a year consisted chiefly of bones with much adherent fat, cirrhosis of the liver occurred in 10. Marked fatty change without definite fibrosis occurred in 2 of the dogs.
2. In control dogs fed the same kennel diet during the same period but not subjected to repeated bleedings, no instance of cirrhosis was observed.
3. In previous plasmapheresis and plasma injection experiments "donor dogs" subjected to similar bleedings over comparable periods but maintained on a different kennel diet did not develop cirrhosis.
4. This series of events suggests that in dogs maintained on a relatively high fat diet repeated bleedings predispose them to cirrhosis of the liver. No data are available to decide the more fundamental question: Do the repeated bleedings remove something (other than hemoglobin) necessary for continued integrity of the liver, or is the cirrhosis the result of relative anoxia or increased lipemia occasioned by the repeated bleedings or perhaps of both combined?