In 2015, an ambiguous picture of a Dress triggered popular and scientific interest as different people perceive it having widely different colours. To help understand this phenomenon, we performed two studies. First, few days after the Dress appeared on the Internet, we projected the Dress to 240 students on a large screen (Experiment 1). The distribution of different colour perceivers partially confirms previous reports with the majority reporting seeing the Dress in blue and black (B&B; 48%) or white and gold (W&G; 38%). The third group of blue and brown (B&Br; 7%) viewers also emerged, and is worth further consideration. In a second, laboratory study, 57 participants used a computerized colour picker to precisely match colours of small squares (a) and vertical strips (b) cut out of the Dress picture, and the original Dress picture (c) (Experiment 2). Results indicated that lightness and chroma values of matched colours changed from B&B to B&Br to W&G viewers in contextualized images (b, c) only. Irrespective of viewer type or context conditions, the hues of the matched colours were either blue-purple (light parts of the Dress) or orange-yellow (dark parts). Finally, colours matched by B&B viewers were the closest to the colours displayed on the ambiguous Dress picture as compared to other viewer types. Our research supports the notion that the Dress effect emerges from differences in perceived lightness and chroma, but not hue. This phenomenon only occurs when all aspects of the picture are integrated, highlighting the importance of contextual information within the ambiguous Dress image.