Because it appears that the snows of Kilimanjaro are retreating for reasons other than Global Warming:
[Philip] Mote and Georg Kaser, a glaciologist at the University of Innsbruck in Austria, write in American Scientist that the decline in Kilimanjaro's ice has been going on for more than a century and that most of it occurred before 1953, while evidence of atmospheric warming there before 1970 is inconclusive.
They attribute the ice decline primarily to complex interacting factors, including the vertical shape of the ice's edge, which allows it to shrink but not expand. They also cite decreased snowfall, which reduces ice buildup and determines how much energy the ice absorbs -- because the whiteness of new snow reflects more sunlight, the lack of new snow allows the ice to absorb more of the sun's energy.
The researchers in this case are legit--Mote has chronicled the retreating ice on the Olympic Mountains here--but that is okay. Science is essentially self-correcting, and here is an example of the fact.
And what we are seeing with the theory of global warming is something that has occurred again and again in the history of science (Newtonian physics might be another example), which is that a theory that has provided an exemplary explanation in one area is extended to more and more adjacent phenomenon, sometimes successfully, sometimes not. So for example the idea that global warming is creating more/more intense hurricanes is certainly plausible, and physical mechanisms by which it might occur have been suggested. Nevertheless the IPCC puts the odds of this theoretic extension's being correct at a little over 50%, and it may yet turn out to be false. Should this prove to be the case, it will not really shake the core of the AGW theory, but only show that it does not explain one particular aspect of climatological reality.