After the Fall, Before the Fall, During the Fall

NOVEMBER 2013

Julie sat in her apartment, studying the graphs on her computer screen. Her desk was jammed against the living room window. Beyond the glass a few flakes of early snow drifted through darkness. Cars swooshed through street slush and swept ghostly patterns of light across the ceiling.

The kidnappings and mysterious store burglaries had followed an erratic path, but the rough outline was clear. The first abduction—the first they knew about and had included in the data, she corrected herself—had been in Sarasota a year ago. October 16 2011: Tommy Candless, age six. Parents divorced, and the child basically a Ping-Pong ball for power struggles between them. John Candless, who did not have custody, had grabbed his son and run before, but he’d easily been caught by state troopers since he hadn’t even had the sense to leave Florida. Heather Candless had one conviction for DWI. Julie wasn’t even positive that Tommy wasn’t squirreled away with some obscure relative or friend somewhere that Gordon’s task force had failed to discover. Or the child could have been killed and the body never found. Both parents seemed to her capable of even that in their intense hatred for each other.

Drop Tommy from the data? Would that help? No, her gut said to leave him in. She did.

The path of subsequent abductions moved roughly up the East Coast, sometimes swerving inland, sometimes back-tracking. The intervals weren’t even, coming in clusters that themselves weren’t even. Nine children after Tommy, including Kara and Jennifer Carter. The kids ranged from ten months old to Tommy’s six years. Seven girls, three boys. Two hysterical witnesses, one relatively calm one (but her baby hadn’t been taken; she’d fought off the young abductress). One hysterical witness had been pushed over the slippery edge of her already advancing illness into an institutional schizophrenia, which did not help the FBI Assistant Director to trust her credibility.

The big problem with the data, as Julie had told Gordon, were the store burglaries. Here, too, the MO was the same: no forced entry, no money taken, seemingly random grabs of diverse merchandise, including shopping carts. Why take shopping carts, both worthless and conspicuous? Why take an entire display of Reese’s Peanut Butter Cups? At Wal-Mart’s, why take pillows and leave untouched a display of diamond rings? And should the Baltimore Kohl’s burglary be included or not? How about that earlier one in Georgia? Or the one in the New Jersey convenience store? That store had had a broken back window, which might have taken it off the list of no-forced-entry burglaries. However, a neighbor walking his dog at 3:00 a.m. thought he heard glass breaking. He shouted and saw kids run away, empty-handed, so did that mean there were two incidents at the same store at the same night, a coincidence?

She played with the data, putting in one incident, taking out another, changing the patterns. The maddening thing was that the patterns were there, and not just in MOs. The numbers showed patterns, too: nonlinear, closer to fractals than to conventional graphs, but nonetheless there. And the numbers should have pointed to another abduction or burglary in Hingham, Massachusetts, on Thursday. Which hadn’t happened. So obviously she had included something erroneous, or left something out, or missed something altogether.

She sat far into the night, scrutinizing data.