A TEST OF SCIENCE APTITUDE WAS CONSTRUCTED AND PRELIMINARY EVIDENCE OF ITS VALIDITY OBTAINED BY INDIVIDUAL ADMINISTRATION TO 58 GIFTED, 6- AND 7-YEAR-OLD CHILDREN. TWO FORMS OF A 91-ITEM HUNTER SCIENCE APTITUDE TEST WERE CONSTRUCTED. ITEMS INCLUDED IN THESE TESTS WERE DESIGNED TO MEASURE THE ABILITY--(1) TO RECALL SCIENTIFIC INFORMATION, (2) TO ASSIGN MEANINGS TO OBSERVATIONS, (3) TO APPLY SCIENTIFIC PRINCIPLES IN MAKING PREDICTIONS, AND (4) TO USE THE SCIENTIFIC METHOD. THE TESTS WERE ADMINISTERED AT THE BEGINNING OF THE SCHOOL YEAR AND DURING THE YEAR AT THE COMPLETION OF SEVEN DIFFERENT SCIENCE UNITS. A SINGLE-WEIGHTED, COMPOSITE SCIENCE ACHIEVEMENT SCORE WAS THUS OBTAINED FROM EACH SUBJECT. THE PARALLEL-FORMS RELIABILITY COEFFICIENT OF THE HUNTER SCIENCE APTITUDE TEST AND THE RELIABILITY OF THE WEIGHTED COMPOSITE SCIENCE ACHIEVEMENT SCORE WERE FOUND TO BE SOMEWHAT LOWER THAN EXPECTED. HOWEVER, THE PREDICTIVE VALIDITY CORRELATION COEFFICIENTS WERE EXTREMELY HIGH, BECAUSE THE APTITUDE TESTS AND THE SCIENCE ACHIEVEMENT CRITERION THAT WERE USED OVERLAPPED THE KINDS OF ABILITY AND CONTENT MEASURED. BOTH WERE MODELED UPON THE CONTENT OF THE THIRD-GRADE SCIENCE CURRICULUM. RESEARCH EXTENSIONS WERE RECOMMENDED IN TIME AND SCOPE OF THE CRITERIA, AS WELL AS IN ADDITIONAL VALIDATION STUDIES. (JH)