Conservation of number was assessed in 56 kindergarten boys and girls. In the standard condition children responsed to the usual verbal questions about number. In the modified condition the ability to conserve number was inferred from the child''s judgments concerning which operations did or did not produce a promised increment in his collection of candies. In comparison to the standard tests, the latter procedure was intended to be simpler verbally, more motivationally engaging and more similar to the real-life situations in which children make judgments about quantity. Children performed significantly better in the modified condition than in the standard condition, a finding which suggests that the usual tests may not fully capture the child''s understanding of number. Performance was also significantly better on tests of identity conservation than on tests of equivalence conservation. This identity-equivalence difference was limited to the standard condition; the 2 concepts were of equal difficulty when assessed by the modified procedure.