Thermogravimetric analysis is often used as a means to measure properties of nanotubes, for example the amount of chemical modification of tubes. We will present results that show conclusively that, unless very slow heating rates are used (e.g. less than 1 K/min), the measured degradation curve depends on the density of the nanotubes. Rate effects in TGA are well-known, what was surprising to us was how slow the rate had to be in order to eliminate these effects.