=== Test 1: Kruskal and Prim must agree, on 2000 random graphs ===
mismatches: 0 out of 2000

=== Test 2: the BFS postcondition, on 2000 random graphs ===
edges spanning more than one BFS level: 0

=== Test 3: does it scale as the analysis says? ===
    zones     roads      BFS ms  Kruskal ms   BFS/(V+E) us Kruskal/(ElgE) us
     2000      8000       0.084       0.446         0.0047            0.0043
     4000     16000       0.158       0.921         0.0044            0.0041
     8000     32000       0.355       1.771         0.0049            0.0037
    16000     64000       0.742       3.731         0.0052            0.0037
    32000    128000       1.886       7.174         0.0065            0.0033
