EARLY FORMS OF NUMERICAL REASONING WERE INVESTIGATED BY EXAMINING THE ROLE OF A SET OF VARIABLES IN THE DEVELOPMENT OF CONSERVATION OF NUMBER. CONSERVATION OF NUMBER WAS DEFINED AS THE RECOGNITION THAT THE NUMBER OF OBJECTS CONSTITUTES AN AGGREGATE WHICH IS NOT CHANGED MERELY BY REARRANGING THEM. EXPERIMENT ONE WAS CONDUCTED WITH CHILDREN FROM DIVERSE SCHOOL BACKGROUNDS. THREE GROUPS OF BOYS FROM THE KINDERGARTEN AND FIRST-GRADE LEVELS WERE STUDIED. THERE WERE 24 SUBJECTS IN THE 3 KINDERGARTEN GROUPS WHO RANGED IN AGE FROM 5.3 TO 6.3 YEARS, AND THE FIRST-GRADE SUBJECT AGES WERE 6.3 TO 7.3 YEARS. EXPERIMENT 2 HAD 98 SUBJECTS FROM THE FIRST-GRADE CLASSES OF 2 PUBLIC SCHOOLS IN A MIDDLE-CLASS NEIGHBORHOOD CURRENTLY DETERIORATING. CONSERVATION MATERIALS CONSISTED OF MINIATURE TRUCKS AND BLOCKS OF VARYING SIZE, SHAPE, AND COLOR. THE NUMBER AND QUALITY OF THESE OBJECTS WERE DETERMINED BY THE EXPERIMENTAL CONDITIONS IN WHICH THEY WERE EMPLOYED. IT WAS HYPOTHESIZED THAT THE ABILITY TO DISTINGUISH BETWEEN SPATIAL AND NUMERICAL CUES IS SALIENT TO EFFECTIVE FUNCTIONING. IN GENERAL, EVIDENCE SEEMED TO INDICATE THESE EXPECTATIONS WERE CONFIRMED, BUT THE CORRELATIONS BETWEEN VARIOUS MEASURES OF DIFFERENTIATION AND CONSERVATION OF NUMBER TENDED TO BE LOW. THE REPORT CONCLUDED THAT IT IS THE ORGANIZING AND INTEGRATING ASPECT OF COGNITION WHICH NEEDS TO BE DELINEATED IF COMPLEX FORMS OF BEHAVIOR ARE TO BE UNDERSTOOD. (HB)