Fig. 3. Reconstruction of the cranium of Australopithecus afarensis, assembled 
from fragments of several individuals. Scale is 1 cm. 
consists of much of a 3.2 myr old skeleton, and the members of the 
‘First Family,’ fragmentary bones of a group of at least 13 indi- 
viduals who may have perished together about 3.4 myr ago. Beyond 
these Ethiopian fossils, it seems most likely that the famous 3.5 myr 
old footprint trails of Laetoli, in Tanzania, were also made by mem- 
bers of this species, whose upright-walking behavior has thereby 
literally been fossilized. This unique insight into the locomotion of 
A. afarensis is confirmed by examination of the fossils themselves, 
but such scrutiny also reveals that these rather small-bodied crea- 
tures did not walk upright quite as we do (e.g., Johanson and Edgar, 
1996). Descendants of tree-living ancestors, they retained a variety 
of features that would have helped them to exploit their ancestral 
habitat even as they moved more freely beyond it than ever before. 
This have-it-both-ways adaptation made the early hominids neither 
as agile in the trees as apes are, nor as efficient on the ground as 
we are; but it served them well, remaining essentially unaltered for 
over 2 million years, even as new hominid species came and went. 
Over this period early hominids seem to have been largely confined 
to the forest fringes, where true forest grades into grassy woodland; 
and indeed, 4 million years ago, true Serengeti-style savannas lay 
very far in the future, even as increased seasonality and climatic 
drying steadily shrank the African forests. 
‘‘Why bipedalism?” is a complicated and as yet incompletely 
resolved story for which there is no time here. However, it’s im- 
