It's something I've mentioned over the years, an assertion that my exercise physiology professor had made regarding fat loss.

Note that under normal conditions, fat cells contain ~90% triglycerides and ~10% other stuff where other stuff includes some water, the cellular machinery that makes all the stuff that fat cells make and a couple of other things that I'm forgetting right now. Basically, fat cells do not normally contain much water.

He told us that, after triglycerides were removed from the cell, that the fat cells refilled with water in the short-term, eventually the body dropped that water and the fat loss 'became evident' (a goofy way for me to try to describe when the fat loss actually shows up on calipers, one of those dumb Tanita scales, or visually).

If nothing else, this gives a plausible mechanism for the non-linear fat loss that is so often seen. Folks will do everything right for weeks with no results. then overnight, something happens and the scale drops a bunch. Many diet newsgroups and forums refer to this as a 'whoosh' which often follows a stall.

A couple of empirical data points in support of this: people who use tanita scales have often reported that it will tell them that their BF has gone up right before a 'whoosh' occurs and a big drop. This suggests something goofy is going on with water balance.

Another is that fat often gets squishy (suggesting a change in what's in there) prior to a drop in skinfolds/ improvement in appearance.

I looked for research on the topic for a decade to no success. I made up my own plausible mechanism having to do with glycerol levels in the fat cell (glycerol is hydrophilic); if fatty acids were being lost at a greater proportion than glycerol, this mght explain how water is attracted into the fat cell. Except that, usually, glycerol and fatty acid are released in about the proportion you'd expect (3:1 FFA:glycerol).

edit: For what very little it's worth, Colgan mentions something similar in OPtimum Sports Nutrition, something about the body 'tracking' glycerol to keep track of fat stores. It's possible that the research on this is just pre-medline. Or he and my teacher just pulled it out of the old ass.

A couple of years back, a paper came out showing an increase in water content of visceral fat with dieting. First semi-direct data I've seen. I don't recall the mechanism being mentioned but I may not have ever read the full paper.