AN EXPERIMENTAL TRAINING PROGRAM DESIGNED TO TEACH 1 TO 1 CORRESPONDENCE TO PRESCHOOL CHILDREN WAS TESTED TO SEE WHAT EFFECT IT MIGHT HAVE ON THEIR UNDERSTANDING OF NUMBER CONSERVATION. THIRTY-FIVE CHILDREN OF AGES 3, 4, AND 5 WERE RANDOMLY DIVIDED INTO 3 EXPERIMENTAL AND 3 CONTROL GROUPS. THE EXPERIMENTAL GROUPS WERE TRAINED TO PERFORM 1 TO 1 CORRESPONDENCE TASKS IN 8 DAILY HALF-HOUR SESSIONS. THE CONTROL GROUP RECEIVED A GENERAL PRESCHOOL PROGRAM INTENDED TO FACILITATE MATHEMATICAL UNDERSTANDING. A PRETEST INDICATED THAT THERE WAS NO SIGNIFICANT DIFFERENCE BETWEEN THE EXPERIMENTAL AND CONTROL GROUPS. POSTTESTS INDICATED THAT THE EXPERIMENTAL GROUP PERFORMED BETTER ON SPECIFICALLY 1 TO 1 TEST ITEMS, BUT WERE NO BETTER ABLE TO UNDERSTAND NUMBER CONSERVATION THAN BEFORE. AS WAS EXPECTED, THE OLDER CHILDREN PERFORMED SIGNIFICANTLY BETTER THAN THE YOUNGER CHILDREN, BUT THE INCREASE IN MEAN SCORES WAS APPROXIMATELY EQUAL FOR EACH AGE GROUP, INDICATING THAT EVEN THE 3-YEAR-OLDS HAD PROFITED FROM THE TRAINING. TWO EXTENSIVE APPENDICES CONTAIN A DETAILED DESCRIPTION OF THE TRAINING PROGRAM AND THE TEST USED TO EVALUATE CONCEPT UNDERSTANDING. (DR)