In previous papers (3, 5) in this series, the generally important role of vocabulary items in intelligence testing has been discussed. Vocabulary tests are used at the twoto five-year level, although probably not as often then as later. The 1937 Stanford-Binet (11) measures vocabulary in several ways at the preschool level. Tests include Identifying Objects by Name, Identifying Parts of the Body, Identifying Objects by Use, and Picture Vocabulary. The Minnesota Preschool Scale (9) contains such indirect verbal items as Pointing out Parts of the Body, Pointing out Objects in Pictures, and Giving Word Opposites. Considerable importance is given to verbal subtests in the Binet and Minnesota Preschool, and even the Cattell Infant Intelligence Scale (8) calls for pointing to the parts of a doll as named by the tester and identifying objects and pictures by name. In contrast to these tests where non-verbal items are also used are the Van Alstyne Picture Vocabulary Test for Young Children (12) and the Full-Range Picture Vocabulary Test (i). Both use only identification-type vocabulary items to measure intellectual ability. To be really useful, a test must have satisfactory standardization as well as suitable items. A survey of the tests mentioned above reveals a wide range of adequacy in their standardizations. The Minnesota Preschool Scale (9) was given to 900 children, ioo in each of 9 half-year age groups, eighteen months to six years. These children were equally divided as to sex, and their fathers were representative occupationally of the total adult male population of Minneapolis according to the 192o census. The 1937 Stanford-Binet standardization (i i) was based on tests given to 76 to Ioo children at each half-year age level two to five and one-half years. The sampling was made fairly representative of the white American-born population on the basis of such criteria as age, sex, geographical location, and parents' occupations relative to the 1930 census.