Abstract
In a classic study, Bornstein (1975) showed 4-month-old infants many pairings of monochromatic, isoluminant stimuli. The infants showed spontaneous looking preferences for red (630nm) and blue (460nm) over yellow (580nm) and green (520nm). We previously tested and rejected the possibility that infants' spontaneous “hue” preferences are due solely to variations in perceived brightness (Teller et al., VSS 2003). However, to adults, luminance matched reds and blues look more saturated than do yellows and greens. If infants see the same pattern of perceived saturation, and prefer more saturated stimuli, then Bornstein's spontaneous “hue” preferences could be due solely to differences in perceived saturation. To test this hypothesis, 3-month-old infants were tested with 10 deg, 4.5 cd/m2 test disks embedded in a 68 × 42 deg, 0.45 cd/m2 white surround. Twenty three chromatic stimuli varying in wavelength and purity were each tested against white (CIE x,y=.33,.33). In general confirmation of Bornstein's results, blue, purple, and red were preferred to yellow, green, and blue-green. In each case, reduced purities yielded reduced preferences. We generated iso-preference contours for infants. These contours were compared to adult iso-saturation contours derived from published data (Indow & Stevens, 1966) and adult iso-saturation contours derived from adult saturation matches made in our laboratory. The shape of infants' iso-preference contours differed markedly from adult iso-saturation contours. Adult iso-saturation contours are roughly centered at white while the centers of infants' iso-preference contours are shifted towards the green region of color space. Equating the saturation of the stimuli reduced infants' differential preferences. However, some residual preference differences remained, especially for blue stimuli. We conclude that infants' spontaneous hue preferences are not due solely to adult-like variations in perceived saturation.