Preble1 reports 4 new cases of fatal gastrointestinal hemorrhage, and adds to this a study based on 56 from the literature, or 60 in all. The conclusions are to the effect that, though infrequent, fatal gastrointestinal hemorrhage is not a rare complication of hepatic cirrhosis, which may be atrophic or much more rarely hypertrophic. It is of interest to learn that in one-third of the cases the first hemorrhage is fatal, and undoubtedly many deaths from hemorrhages of this kind are attributed to other causes; in the other two-thirds the hemorrhages continue at varying intervals over a period of months and years, the maximum being eleven years.
In one-third of the cases the diagnosis can be made before or at the time of the first hemorrhage. In others it can not be made at all, or only after months or years during which other and more definite symptoms of