Prioritize features that can be compared with the missing person's description and records, distinguishing a general biological profile from a distinctive match. The problem asks what to observe; it supplies no actual bones from which to invent an identification.
Begin by documenting recovery and association. Lay out and inventory the remains, confirm that they are human, and consider whether all the bones could belong to one individual. Note missing or duplicated elements and any damage from animals, burial or recovery. An apparent absence is informative only if the relevant part was actually recovered and examined. Keep the teeth and small fragments; they may contain the most useful identifying information.
Estimate age using several indicators. In the young, tooth development and eruption, growth and fusion of the epiphyses are useful. In adults, changes at the pubic symphysis, rib ends and teeth may help, but provide ranges rather than an exact birthday. Tooth wear depends on diet and habits, and cranial suture closure is variable. The human skeleton's developmental and pathological changes must be interpreted together rather than converted into an unjustifiably precise age.[ref]
Forensic Anthropology by the Smithsonian National Museum of Natural History. “Reading a Skeleton” and “Analyzing Human Remains”: growth of bones and teeth, skeletal standards and pathological change
For estimated sex, give particular attention to the adult pelvis: the greater sciatic notch, subpubic region and overall form are more informative than size alone. Features of the skull can assist, but robusticity is variable and overlapping. Immature remains often do not support the same degree of certainty. For stature, measure preserved long bones, especially the femur and tibia, or reconstruct skeletal length if sufficiently complete. State the method and uncertainty: bone length is not living height, and neither sex nor stature should be inferred solely from a large or small skull.[ref]
Skeleton Secrets by Arizona State University. Interview, discussion of age, adult pelvis and skull, and long-bone estimation of height
The most discriminating observations may be particular departures from the usual anatomy: an old healed fracture with shortening or angulation, an amputation, congenital deformity, fused anatomical joints, unusual disease changes, or a distinctive pattern of missing and repaired teeth. Record tooth positions, fillings, extractions and any appliance, distinguishing loss during life from loss after death. Compare these features with actual medical or dental records and reliable descriptions. A family's report of a healed broken leg can guide the comparison, but it does not predict which leg or what deformity must be present.
Record agreement and contradiction. Approximate age, sex and height can narrow a search but fit many people; a well-documented distinctive lesion or dental pattern can provide much stronger evidence. A mismatch must be evaluated for record error, estimate uncertainty or exclusion, not silently ignored. Do not claim that nineteenth-century racial categories or a facial reconstruction uniquely determine identity. Soft-tissue features such as eye colour, hair and skin markings cannot simply be read off a bare human skeleton.
The modern extension is a specialist comparison of ante-mortem records with post-mortem observations, potentially supplemented by DNA profiling from suitable material. This is later corroboration, not a technique available in 1866. The institutional identification guidance emphasizes reconciliation of the two records and the importance of dental or genetic evidence for conclusive identification.[ref]
Disaster Victim Identification by INTERPOL. “Four phases to identification”: post-mortem examination, ante-mortem records and reconciliation