The original dataset was captured in May 2008 using auto exposure and auto white balance. It has not been modified since. After looking hard at the skin tones they seem to be natural and there are no colour casts elsewhere in the picture. The exercise was carried out using the skin tones and ignoring the effect of the red life jacket. The PSP tool used is called 'Automatic Colour Balance' which is a bit of a misnomer as both the colour temperature and the percentage strength of the adjustment can be set manually. Strength was set to 100% and the colour temperature slider adjusted until the adjusted and unadjusted skin tones were as nearly identical as possible. This gave an initial setting (6650K) from which adjustments were carried out. The colour temperature slider was thenmoved to 7160K and the strength retained at 100%. The image became cooler but is still within an acceptable range. Viewing the original and two modified versions side by side on the screen the neutrality of the original is confirmed. The warm image does look noticeably more red and the cool image noticeably more blue but neither are sufficiently far off to reject out of hand.