It is a basic tenet that good osmoregulation starts with an impermeable barrier to the outside world, but some skins are more watertight than others. Unlike reptiles, poor amphibians must spend their lives in damp places to prevent themselves from drying out through their leaky skins, while the waxy cuticle of small terrestrial insects has allowed them to survive some of the planet's most arid conditions. It is thus perhaps surprising to find a water channel in skin that is essential for skin's barrier properties.Many cells are more water permeable than lipid bilayers allow. This led to the discovery of the aquaporins, or water channels, which are members of the major intrinsic protein (MIP) family. Some aquaporins play predictable roles -for example, a class of diabetes insipidus is due to a defect in aquaporin 2 in the kidney, preventing concentration of the primary urine. Others let through small organic solutes, such as aquaporin 3 (AQP3) which is permeable to water and glycerol (the archetypal `aquaglyceroporin'). This unusual aquaporin is abundantly expressed in the keratinocytes of the mammalian epidermis. Hara and Verkman's paper shows that in mammalian skin it is glycerol permeability, rather than water, that is important to keep skin watertight.They had shown previously that an AQP3-knockout mouse was normal, except that it had dry skin - specifically, the overlying stratum corneum had only a third of the normal water content, compromising its elasticity and barrier properties. So a battery of tests was used to delineate the defect. Firstly,the mice were crossed into a hairless background, to give easier access to the skin. Then, they measured skin conductance and 3H2O permeability to find out how watertight the skin was in the absence of AQP3. They found that by administering glycerol to the mutant mice, they could restore their skin to normal levels of watertightness. Interestingly, glycerol was effective not just topically but also intraperitoneally or even orally. Also, the improvement was specific to glycerol and was associated with an increase in glycerol levels in the stratum corneum to near normal levels. They also used a skin elasticity meter before and after glycerol treatment, showing that glycerol doubled the skin's elasticity.The integrity of skin is compromised daily by minor abrasions. This was simulated by applying sellotape to the skin of the mutant mice and peeling it off, so removing most of the stratum corneum layer and so compromising the barrier function. Glycerol restored the rate of recovery of water barrier function to near wild-type levels and so was useful in the recovery of barrier integrity. Finally, they demonstrated in vitro that the defect in the dermis of AQP3-knockout mice was due to twofold-reduced glycerol uptake. The various experimental glycerol supplementations were therefore acting to bypass the defective transport system.The take-home message is thus that a member of the aquaporin family can actually reduce water loss through the skin by increasing water content of the SC and thus maximizing the integrity of the barrier. However, it does this not by the obvious way but by increasing glycerol content. Why should glycerol be so important? It seems to act in two ways; firstly, it inserts into and directly increases the fluidity of lipid bilayers and, secondly, it is hygroscopic and helps to hold water in the SC. The glycerol-transporting properties of AQP3 must thus be seen as a feature, rather than an irksome complication of the intuitively simple `aquaporin=water channel' story.As well as its physiological impact, the authors point out a relevance to our everyday lives; glycerol has been used in cosmetic preparations for over two hundred years as a humectant (moisturizer) and to improve the elasticity of skin. Now we have a scientific explanation for the practice.