Tests of grating acuity often underestimate amblyopia and underrefer esotropic infants with a fixation preference. To evaluate the effect of the large grating field used in preferential looking (PL) procedures, an eight-alternative, small field (about 1 degrees) grating acuity test was devised. Gratings contained at least eight cycles. Thirty-seven strabismic and/or anisometropic amblyopes, ages 3-13 yr, were tested. In most amblyopic eyes, grating acuities were better than recognition acuities; the difference was reduced, however, in the small field test compared to the large field (6 degrees test (means, 1 oct vs. 1.6 oct; paired-t = 5.5, P less than .001). Nevertheless, the same relation between grating and recognition acuities occurred for small as for large fields: an increased discrepancy between grating and recognition acuities accompanied poorer acuity. This larger discrepancy is attributed to increased probability summation of amblyopic eyes for low spatial frequencies. For preschool children who can be tested by both procedures, the eight-alternative grating acuity test may be preferable to operant PL because it is more easily administered and materials are simpler.