A cross-sectional study with children between the ages of 8 and 19 yr. showed the occurrence of loss in originality mean scores as measured by Onomatopoeia and Images (Khatena, 1971a) especially between the ages of 9 and 10 yr. with gains occurring at 11 yr. (Khatena, 1971b). The purpose of this study is to examine this tendency longitudinally. Forty-eight 8-yr.-old children (26 boys and 22 girls) attending Cammack Elementary School in Huntington, West Virginia were initially administered Forms 1A and 1B of Onomatopoeia and Images (Torrance, Khatena, & Cunnington, 1973) as a retest package. The group was administered the same measures at four consecutive yearly intervals, with complete data available for only 18 boys and 17 girls. Their responses to onomatopoeic stimuli were scored for originality on a scale of credits from 0 to 4, with 0 awarded for responses that occurred in 5 % or more of the cases, 1 for 3 to 4.99% of the cases, 2 for 2 to 2.99% of the cases, 3 for 1 to 1.99% of the cases, and 4 for those responses that occurred less than 1% of the time. The scoring of all response sheets was done by the second author. Addttional data on the rationale, reliability and validity of the measure may be found In the Norms-Technical Manual (Khatena & Torrance. 1973) . Means and standard deviations on Forms 1A and 1B of Onomatopoeia and Images for these children at ages, 8, 9, 10 and 11 yr., respectively are: Form 1A-Ms = 28.54, 28.03, 32.11, and 38.09; SDs = 13.88, 10.72, 10.25, and 12.76; and Form 1B-Ms = 30.31, 28.94, 24.51, and 32.40; SDs = 15.45, 8.31, 16.58, and 12.38. The loss in mean originality scores at age 9 yr. is negligible o n Form 1A; on Form lB, the loss in mean originality scores at age 9 yr. is more noticeable and continues to an even lower level at age 10 yr. Mean score gains begin to show at age 10 yr. and continue to rise at age 11 yr. on Form lA, whereas mean scores on Form 1B show gains at age 11 yr. When the originality scores on Forms 1A and 1B of each S at each of the four age levels were combined and averaged, the means and standard deviations of these averaged originality scores were as follows: Forms 1A and 1B combined-Ms = 29.71, 28.30, 28.79, and 36.03; SDs = 13.84, 8.51, 8.27, and 11.18. The loss in mean originality score which appears at age 9 yr. changes to some small gain at age 10 yr. to rise to an even higher mean originality score at age 11 yr. Some fluctuation in the variance relative to the loss and gain in mean originality scores is apparent. When the data were subjected to analysis of variance for repeated measures design (Edwards, 1960), the patterns of rise and fall in mean originality scores at the four age levels a peared significant (Form 1A-F = 7.03, df = 3/102, f i < .01; Form 1l3-F = 3.20, a$ = 3/102, fi < .O5; Forms 1A and 1B combined-F = 4.89, df = 3/102, p < .01). This gives some support to the earlier observation that some loss appears to occur for children at ages 9 and 10 yr. and gain at 11 yr.