DEFECTS IN OBTAINING DATA UNDER SUCH CONDITIONS1. We early found that the description of the local extent of the lesion varied too greatly, and often was too meager to be of comparative value. P a r t of this difficulty was no doubt due to the fact that for private patients the hospital records were not complete as regards vaginal examinations.2. Similarly, pathological grouping, as regards grade of malignancy, was too varied (classified by different pathologists) to be of value statistically.3. The histories were brief, and while in general the chief symptoms of cancer were accurately stated, the history of precancerous irritations (childbirth lacerations, chronic discharges, etc.) and of previous medical treatmcnt, and statements as to time relation and sequence of events were not always complete. 4.The records of the technical application of radium and x-ray therapy were not sufficiently detailed to contrast the results obtained from different methods of treatment.