No Witnesses

“Gus.”

“Gus?”

“That’s right.”

Gus sat down alongside Boldt. He worked a computer keyboard as fast and as delicately as Boldt’s grandmother used to knit. “Your request was easy,” he bragged. “You told us which aisle and what to look for. Without that, it might have taken us a day or more. I think I may have your offender, although I’m not familiar with this particular game-placing product onto a shelf. What’s your interest in this anyway?”

“Corporate espionage,” Boldt lied, making it up on the spot and feeling self-conscious until the technofreak grinned enthusiastically as if he’d been let in on something.

“Cool,” he said. “What I’ve done is catalogue the images I have found so far and placed them in chronological order. Here’s the first image in the progression. This is the entrance door to Foodland as caught by one of our cameras.”

On the screen was displayed a slightly fuzzy black-and-white image that showed a pair of automatic doors. The left door swung open admitting a person wearing a gimme cap and a dark jacket. Medium height and weight. He (she?) turned into the store and walked off the screen.

“That’s our first look,” Gus muttered. “Not much.”

The bottom right of the screen was date and time-coded. The suspect had entered Foodland at 5:02 P.M. on June 21. Clearly a busy time of day for the store. And late in the day, when the shelves were more likely to have room for the killer’s substitution. Boldt experienced a pang of anxiety: Was this the Tin Man?

“Our next decent hit is three minutes later. And you should know something here, Lieutenant.” Boldt didn’t bother to correct the mistaken rank. “Your average shopper-your innocent shopper-ends up all over these videos. But except for a flash here and there, this mark has avoided the cameras for nearly three minutes. And that’s not easy. Granted, Shop-Alert didn’t install the Foodland system-we only analyze their images, and it isn’t the greatest system, but even so, to avoid these cameras is something of an art form. It requires prolonged study of the facility, and even at that, a hell of a lot of luck. Of course, dress has a lot to do with it. You’ll note the dark clothing and the hat. Dark clothing in saturated black-and-white video-in this kind of light, as you can see-tends to absorb too much light, throwing off the gray-scale balance on the areas immediately around it, causing a graininess like a shadow that renders the image difficult to evaluate. The dark clothing makes it difficult to see her face.”

“Her?” Boldt asked. “She looks a little androgynous to me.”

“A woman, I think so, yes.”

Gus consulted a time log on a clipboard in front of him, then keyed in a set of numbers. “For the time being, we’re going to jump ahead two minutes and fifty seconds to show you this.” He hit the ENTER key. A new image appeared, ran for only half a second, and then, as he struck another key, was freeze-framed. It was this same person in another area of an aisle. The person’s head turned slightly, which was where Gus stopped it. “I’m going to zoom and enhance now. It takes a second or two for the screen to refresh at each phase.” Using a computer mouse, Gus dragged a box around the face. This box then filled the entire screen. Box by box the electronically enhanced enlargements continued, and the suspect’s head grew ever larger. The tighter the image, the fuzzier it grew, because “enhancement can’t keep pace with enlargement,” as Gus explained. By the time the process was completed, much of what was on the screen was only made discernible by Boldt’s imagination and the images that had come before. He wasn’t even sure what he was looking at.

“Lower head and neck,” Boldt guessed.

“Exactly right, Lieutenant.” The boy sounded impressed. He typed additional instructions into the machine and sat back. “Now let’s run that again.” He ran it several times, like instant replay, before Boldt saw it.

“The bounce to the hat?” Boldt asked.

“It’s oversize. And the way it bounces means there’s a lot of hair up inside there.”

“You’re good at this,” Boldt complimented.

“We spend enough time at it.” Gus drew a box around the woman’s ear, and the computer began a series of enhancements. At the same time, the sequence played in slow motion, backed up, and played again repeatedly. Gus slowed the motion even further. “There!” he declared excitedly-and a little too loudly for Boldt’s ear. “It’s our only real chance to see it.” He pointed to the earlobe, where a square black mark winked at them.

Boldt studied the repetition for several passes, and Gus had the good sense to keep quiet and let the detective have some room. Boldt finally tested, “A freckle? A mole? I’m not sure I see the importance.”

“Lower earlobe,” the boy hinted. This was a contest.

“Pierced ears!” Boldt said loudly, briefly drawing the attention of the other video attendants in the room. “No earring, but that’s a hole in the ear! Even so, that hardly indicates a woman.”

“Added to the height of the individual and the apparent weight of hair inside that hat-”

“It may be a woman!” Boldt stated. “I’ll give you that,” although this shattered his image of the Tin Man, whom he had assumed to be male. “I need to follow her every movement.”

The technician showed him all the images in which the female suspect was captured by the cameras. At no point did she reveal her face. “Here’s where we vote her All-Pro,” Gus said. “There are only a few shoplifters as good. Note her position to the camera. She’s in aisle four: soup and vegetables. Positioned this way, she fully blocks any chance we might have of seeing her specific actions. She checks her watch-see that?” He replayed the moment. “And now she’s gone from our view for over seven seconds. By the time we pick her up again, she has moved quickly down the aisle. She bumps into that man with the cart, there-see that? — and by the time we pick her up again, she is paying cash, head still down, for a candy bar, and she’s gone. The thing is, checking her watch: She had the cameras timed.”

“A woman?” Boldt asked uneasily. The Tin Man? he wondered.

“Now check this out. This is beautiful!” the technician said enthusiastically. The screen blanked to a deep blue. When an image reappeared again, it showed aisle 4. The technician blocked a segment using a white box, tripped a key, and leaned back. The area zoomed and enlarged several times, the shelves moving increasingly closer, the products-soup cans-more easily identified: Adler soup cans. “This is where the woman was standing,” he explained. He split the screen into two similar images and said: “Before and after. See the difference? She’s not a lifter after all; we’ve got nothing on her!”

The right-hand screen showed five soup cans that were not present in the earlier image.

“Five?” Boldt asked in a panic.

“Something wrong?” the young man queried. “She’s not a lifter at all,” he repeated.

“Is it possible?” Boldt muttered.

“Five cans of soup?” the young man asked, misunderstanding the question. “You should see some of the clothing that’s been used, the amount of stuff they can hide. We offer a seminar on clothing used in shoplifting-you wouldn’t believe some of the stuff!”

Boldt could only account for two, possibly three, cans: the Chin girl and Slater Lowry. So where were the other three cans?

“Listen up,” Boldt announced to his squad: LaMoia, Gaynes, Danielson, and Frank Herbert.

Herbert, fifty, stood five feet five with a pot belly that made him equally wide. His balding head was spot- shined. Lieutenant Shoswitz stood by the door.

Homicide’s situation room contained a half-dozen Formica desks, a retractable projection screen, and a large Wipe-It board that at the moment contained several profanities and a graphic cartoon.

Boldt briefed them on the case, taking them through his visit to Shop-Alert and the discovery of several unaccounted-for soup cans. He had been on the phone the entire afternoon; his voice was hoarse. Or maybe that was nerves.

“We’re missing two to three cans: Lori Chin’s mother doesn’t have any on her shelf, so I’m guessing there are still three at large. The surveillance video has given us a window of time during which the suspect was inside the store, making the drop,” Boldt continued. “Thanks to a computerized cash register system, we can identify any bank check or credit card purchases and then trace them back to whoever made that purchase. We’ve identified thirty- four people who we know were shopping inside this Foodland supermarket at the exact same time as our suspect. We have also identified eleven customers who purchased the Adler chicken soup within