Little consensus exists on how best to analyze natural fracture spacings and their sequences. Field measurements and analyses published in geotechnical literature imply fracture processes radically different from those assumed by theoretical structural geologists. The approach adopted in this paper recognizes that disruption of rock layers by layer-parallel extension results in two spacing distributions, one representing layer-fragment lengths and another separation distances between fragments. These two distributions and their sequences reflect mechanics and history of fracture and separation. Such distributions and sequences, represented by a 2 ?? n matrix of lengthsL, can be analyzed using a method that is history sensitive and which yields also a scalar estimate of bulk extension, e (L). The method is illustrated by a series of Monte Carlo experiments representing a variety of fracture-and-separation processes, each with distinct implications for extension history. Resulting distributions of e (L)are process-specific, suggesting that the inverse problem of deducing fracture-and-separation history from final structure may be tractable. ?? 1985 Plenum Publishing Corporation.