One technique that is often used for a sparse graph, say , uses linked lists--one for each vertex. The linked list for vertex contains the elements of , the set of nodes adjacent to . As a result, the lists are called adjacency lists .
Figure shows the adjacency lists for the directed graph of Figure and the directed graph of Figure . Notice that the total number of list elements used to represent a directed graph is but the number of lists elements used to represent an undirected graph is . Therefore, the space required for the adjacency lists is .
By definition, a sparse graph has . Hence the space required to represent a sparse graph using adjacency lists is . Clearly this is asymptotically better than using adjacency matrices which require space.