We built by hand a crane base with motors for rotating the arm and pulling the
hook. The evolved crane arm needed to attach to this base, so the experimental
setup had 5 ``fixed bricks'' where the rest of the structure was allowed to
attach. The fitness value is the horizontal length x of the arm, but
if the maximum load M supported at the tip is less than 250 g then xis multiplied by M/250, thus reducing the fitness (table 2.11).
The arm has to go up and away at the same time, because the space is restricted
to the diagonal semiplane -x>y.
We observed a curious phenomenon about the way a triangular solution appeared.
Soon the crane evolved a counterbalance, located at the top, to act against
the massive external load. This counterbalance took the shape of a bar going
back from the tip of the crane -- where the load is located, with a pile of
bricks at the other end to act against the weight. This counter-weight could
not be too heavy, for the crane is required to support its own weight before
adding the external load (fig.2.27).
With the tendency to reuse useful parts that had already been observed in the long bridge experiment, the counterbalance at the tip, with its ``J'' shape, reappeared at a lower position, creating an arm counterbalanced at two points. The fact that these two counterbalances ended up connecting to each other and touching down at the base was a fortuitous event that created a new stronger synthesis: a triangular shape which supports more weight than the original counterbalancing bar, by tensioning the vertical column. At the same time the triangle supports all this counterbalancing apparatus (by compression) when the crane is not lifting a load. This is a change of use, as discussed in section 2.6.2, only in a much larger scale.