Why not be random to build a woody crown network?

We compared the decomposition, topology, and properties of 58 natural woody crown networks (WCN) from Cerrado and Caatinga vegetation with the same number of WCN built randomly at different sizes, from 10 to 248 nodes (the branching regions). The random distribution of nodes did not wrack the intrinsic relationships between WCN properties and created woody crowns similar to those found in Caatinga with relatively safe hydraulics to face severe water and other stresses. Nevertheless, growing by random node addition is an inflexible, unintentional procedure incapable of creating WCN-like Cerrado trees. These Cerrado trees showed WCN with less laterality and basitony, potentially achieving higher maximum crown height, but inherently had a riskier hydraulic architecture with lower navigability between nodes. Hence, random node distribution is unsuitable for building WCN in environments with mild stresses where trees typically attain considerable height. A sense of node numerosity (abundance), location of node insertion (rectilinear or lateral), node relative position (near or far from the initial node) along the woody axis, and anticipation behavior are essential to construct a suitable and adaptable WCN in each changing environment. Besides remarkable structural plasticity, the cognition-related features could be accomplished by some constitutional random traits of WCN, constructing safe hydraulic architecture even by haphazard node addition, and modulating the WCN decomposition and topology within limits through a blueprint project.