Tanenbaum hurried off to run our prize through CPIC, the Canadian Police Information Center. Like the U.S.’s AFIS, the Automated Fingerprint Identification System, CPIC functions as a database for fingerprints and other information critical to police investigations.

LaManche resumed his external examination of the baby. Ryan headed upstairs to check for responses to queries he’d circulated about Amy Roberts/Alma Rogers/Alva Rodriguez and Ralph Trees.

When I returned to autopsy room four, Lisa had finished photographing both LSJML-49277 and LSJML-49278. She’d also popped X-rays of the former onto light boxes ringing the room.

I moved through the films, pessimistic. I was right. The window-seat baby was so tightly constricted that overlap of the bones made assessment difficult and measurement impossible. Frustrated, I crossed to the table where LSJML-49278 now lay on its unrolled towel beside the window-seat baby.

The attic baby had been reduced to a skeleton and fragments of dry ligament. Lisa had arranged some of the bones to form a miniature person. Most lay to one side on the grimy green terry cloth.

I wasn’t surprised she’d failed to identify more. In a newborn, the cranial bones are unfinished and unfused. The vertebral arches are separate from the little disk bodies. Each pelvic half is composed of three disconnected bits. The long bones are amorphous shafts lacking the anatomical detail and joint surfaces that make femora, tibiae, fibulae, humeri, radii, and ulnae distinct. Ditto the teeny bones of the hands and feet.

Bottom line: most people wouldn’t recognize a fetal skeleton if it hit them on the head. Even with training in juvenile osteology, classification of specific elements can be tough.

I checked the clock. Already it was going on eleven.

“This will be slow,” I said to Lisa. “If you have things to do, I’m good working alone.”

She seemed undecided, then, “Call if you need me.”

I began by arranging the cranium in a pattern that looked like an exploded rose blossom. The frontal, the parietals, the sphenoid, the temporal and occipital segments. While sorting, I processed detail.

The occipital bone contributes to the back and the base of the skull. In a fetus, it consists of four pieces. The pars squama is the upper, rounded portion. The paired pars lateralis and the single pars basilaris lie down under, surrounding the foreman magnum, the hole through which the spinal cord enters the brain.

I used sliding calipers to measure the chunky little pars basilaris. Its width exceeded its length, placing the baby’s gestational age at over seven months.

I positioned the needles on each end of the left pars lateralis and read the dial. Its length exceeded that of the pars basilaris. That nudged the gestational age to eight months.

I selected the flat, serrate-edged portion of the temporal, the part that had formed the right side of the baby’s skull. A delicate circle of bone, the tympanic ring, was fused to an opening I knew to be the auditory canal.

Ring fusion bumped the gestational age to nine months.