You're describing a clustering problem. If you know how many clumps you want to end up with, you can run a k-means clustering algorithm, which is just a mathematical formulation of Notromda's suggestion. Because you have a desired maximum distance, you can choose the number of clusters (K) by starting with K=1, and running k-means clustering. If the distance between elements in any cluster is too big, increment K and recluster. Iterate until until all clusters meet the maximum distance criterion.

Brian