The general purpose of this investigation was to study the relationship of various kinds of vocabulary functioning to mental deterioration in the senile dements, and to determine whether the vocabulary tests devised by Capps offered a valid and reliable index of deterioration. Five types of vocabulary tests were used: (1) a synonym test, (2) an antonym test, (3) a categorization test involving the classification of several "species" words with a common "genus" word (4) a list of 12 homographs with a minimum of 5 different meanings, and in which the subject was to state as many as he could within 2 minutes, and (5) a free-association naming test where the subject named as many words as possible in 3 min. The subjects were 50 patients, [male] and [female] of the type senile dementia, simplex, ranging in age from 60 to 85. Hospital records were so incomplete that only a fair amt. of control was maintained in the matching of mental capacity, education and background. The subjects were divided on the basis of hospital reports into 3 groups: least, mild, and most deteriorated. The more deteriorated the subject the lower was his score on all 5 tests. The mildly deteriorated were reliably inferior to the least deteriorated only on the synonym and antonym tests, but the most deteriorated group was reliably inferior to the other 2 groups on all 5 tests. Previous findings seem to indicate that vocabulary tests remain relatively constant and unchanged in deterioration as opposed to non-vocabulary tests. Indeed, that premise is the basis for the currently used "Efficiency Index" of Babcock and of the Hunt-Minnesota test for Organic Brain Damage. The present findings run contrary to this premise, and actually imply that tests involving vocabulary functioning serve as a good means of discriminating among deteriorated patients. The synonym and antonym tests are offered as the most reliable measures in this respect.