In a metacontrast masking paradigm, adult subjects were required to detect briefly presented target lines followed at various delays by a flanking mask varying in spatial frequency and wavelength. Detection accuracy, the dependent measure, was recorded as a function of the delay of the mask. The results showed that long wavelength masks produced maximum masking at a relatively short delay, while short wavelength masks produced maximum masking at a relatively long delay. Results are discussed within the framework of transient/sustained (magnocellular/parvocellular) theory of visual processing, and suggest that low spatial frequency channels respond with shorter latency and/or faster rise time to short wavelength stimuli, and high spatial frequency channels respond with greater sensitivity to long wavelength stimuli.