A pound of flesh...

Or, more accurately, a pound of fat. One of the common things one hears as one begins to research diet and exercise as relating to weight loss is that one pound of fat contains 3500 calories, and therefore to lose a pound of fat you need to increase your calorie deficit by 3500 calories to "burn off" a pound of fat, and therefore be one pound lighter.

I've been searching across the interwebs and haven't yet been able to find any concrete studies to back this up, but something about it seems fishy to me. It seems over-simplified, as if any calories we eat go into the same bank, regardless of the "quality" of the calories, or the source of the calories, and then any calories we burn, regardless of the way we burn them (weight training, cardio fitness training, breathing, thinking, etc.) all withdraw the calories from this same bank. I don't think this is necessarily the case.

When I started back at running in January of 2008 after slacking off after my first marathon, I weighed in at 225 lbs. Within a week I was down to 219. According to the "3500 calorie" theory, I created myself a calorie deficit in that ONE WEEK (?!?!) of 21,000 calories???!! Um, no. The running, totaling maybe 10 miles, may have added 1000-2000 calories to my "deficit" total, but nowhere near the 21,000 that the theory would suggest. I believe metabolism is more complicated than the theory suggests.

Does anyone have information or peer-reviewed studies, etc., that can back up the theory one way or another?