I quote Cordain from the paleodiet list:

Human's living at northern latitudes preferentially hunted megafauna
because these beasts contained both absolutely and relatively more
fat.
Hence, at northern latitudes wherein carbohydrate (plant food) sources
are
seasonally restricted, the fat obtained from larger mammals was
sufficient
to dilute the lean protein from muscle tissues.  The fossil record
shows
that the worldwide extinction of animals that took place at the end of
the
Pleistocene occurred primarily in animals over 100 kg (220 lbs) (5).
Using
the Pitts and Bullard regression (2), a 220 lb mammal would be
expected to
have about 15% body fat.  Applying our cubic regressions (4) to this
value,
a 220 lb mammal would have 60 % of its total body energy as fat and
40% as
protein.  The protein value then is very close to maximal protein
ceiling
(also 40  % of energy) -- hence it is not surprising that the "cutoff"
values for megafauna extinction (100 kg) corresponds almost exactly to
the
value for the maximal physiological protein ceiling in humans.  In
animals
weighing less than 100 kg, the entire carcass cannot be consumed
unless
there is a carbohydrate source, whereas in animals weighing more than
100
kg, the entire carcass can be eaten with no worry about protein
toxicity and
with no need to find a carbohydrate source.