It's interesting that it's apparently the problem that laid the foundation for modern graph theory, but the problem can't actually be specified as a graph under its usual modern definition... can it? At least not without some shenanigans that would make it more complicated than "node = landmass, edge = bridge", as it's depicted in the standard diagram shown on Wikipedia.

A graph is usually defined as a set of nodes and a set of edges, where an edge is a pair of nodes (an ordered pair if it's a directed graph). The "set" of edges (bridges) of the "graph" in this case contains duplicates -- there are two edges with the same pair of vertices, two bridges crossing between the same pairs of landmasses. And the two pairs of duplicated bridges are essential to the problem setup. So AFAICT it can't be fully specified as a set of edges; it would have to be a multiset.

(I've seen this problem before but only just noticed this when reading Wikipedia now -- let me know if I'm missing something.)