The Dederiyeh 1 infant died at roughly a year and a half. Its pelvic girdle had not yet fused, so each hip bone was still three separate pieces waiting on adolescence to become one. The individual had been walking for a few months at most. And the shape was already, unambiguously, Neanderthal: iliac blades flaring outward, the front edge of the iliac crest sitting further back than it does in a human child, a sacrum tilted less steeply, the whole structure short from front to back relative to how wide it was.
That combination shows up in a principal component of shape variation that has nothing to do with age or size. Along the axis that tracks growth, the Dederiyeh infant sits at the upper end of the modern human range and looks like a slightly large human toddler. Along a separate, age-independent axis, it separates cleanly from all 34 human infants in the comparative sample, which spanned one month to eight and a half years. The Neanderthal pelvis was not something an individual grew into by loading it a particular way. It was there before the loading started, which fits earlier work showing that Neanderthal neonatal postcrania from Mezmaiskaya and Le Moustier already sit at the edge of human newborn variation.
The second fossil in the new PNAS1 study is a young adult female from Sima de las Palomas del Cabezo Gordo in Murcia, Spain, pulled out of breccia infill in 2006 and 2007. She came out fragmented and taphonomically distorted, which is the normal condition of fossil pelvises and the reason this part of the skeleton has generated so much argument and so little resolution. Zollikofer, Ponce de León and colleagues ran her through medical CT, stripped the sediment matrix digitally, generated 3D surfaces of every fragment, and reassembled them in virtual anatomical space using preserved contacts between adjacent pieces and the constraint of bilateral symmetry. Missing regions were filled with mirror images of the surviving side.
One detail of the method deserves attention, because it is where a lot of previous work has gone wrong. Every adult fossil in the analysis, Palomas included, was represented by three separate reconstructive variants rather than one preferred version. The uncertainty in the reassembly is carried through the statistics instead of being resolved by fiat at the start. Palomas and the previously reconstructed Tabun female from Israel and Kebara 2 male, also from Israel, were compared against 112 modern human adults aged 14 to 49, quantified with 123 anatomical landmarks apiece.
