By now talk of a serial murderer preying on young backpackers was beginning to have a serious effect on local communities that depended on tourism, adding to the pressure we all felt to make an arrest. Public meetings, rather than reassuring people, were reported by the media in ways that served only to fuel the growing panic about the killer at large.
In the first few weeks our media briefings had tended to focus on the search of the forest, allowing us to deal in facts and avoid vague speculation about the identity of the killer. But our refusal to speculate did not stop the media from speculating. By mid-November I realised I would have to give them what they wanted. Yes, I told them, we were looking for a serial killer. No, we did not have a prime suspect or suspects for the killings, but ‘I believe we are getting a better picture of what happened, a better picture of the direction in which we should be going and to that extent we are making progress.’ I made it clear that the investigation would not be wound down once we left the forest; we had plenty of leads to pursue.
While our search of the forest was over, the process of evaluating what we had found had only just begun. Detailed forensic studies of more than 200 exhibits, including clothing and other personal effects, would have to be carried out by units specialising in crime scene analysis, forensic ballistics, mapping, photogrammetry, fingerprints, document examination and videos. Our own forensic examiners were being helped by outside agencies including the Forensic Services Division and the Firearms and Ballistics Branch of the Australian Federal Police, the South Australian State Forensic Science Laboratory and the School of Materials Science at the University of New South Wales.
By the time we left the forest, several trends or ‘facts’ had begun to emerge. The abductions and killings had all occurred between 30 December 1989 and 18 April 1992. The seven victims were all hitchhikers. Liverpool was a common point in their travels, and the likely area where they had been picked up. All of the bodies had been stabbed multiple times. Two of the bodies had also been shot multiple times; the same weapon, a .22 calibre Ruger rifle manufactured between 1964 and 1982, had been used, and the bullets fired were Winchester Winner. The shooter or shooters appeared to be familiar with and to have a degree of expertise in the use of firearms. There appeared to be an increase in the ritualistic nature of the murders. The abductions and killings appeared to have been well planned. More and more time was being spent at the crime scene: the murders themselves were taking longer to commit. Bindings were found at or near five of the crime scenes, indicating that the victims had been bound and the binding had been removed before ‘burial’. In several cases evidence suggested that the victims had been sexually assaulted, although sex did not seem to dominate the killer or killers’ behaviour. In each case, the victim’s backpack and other personal items they might have been expected to be carrying were missing. Since these items were of little monetary value, robbery was unlikely to be the motive. Were souvenirs being taken and, most importantly, kept? The killer or killers knew the Belanglo State Forest well. Given the rough terrain that had to be travelled in the forest, it was likely a 4WD had been used to pick up the victims and drive them to their fate.
On 11 November 1992 the New South Wales government had approved a reward of $100,000 for information leading to the arrest and conviction of the person or persons responsible for the murders of Caroline Clarke and Joanne Walters. No one had tried to claim it. Twelve months later, following the discovery of five more bodies, the government increased the reward to $500,000 (at the time the largest reward ever offered in Australia) for information leading to the killer’s conviction. A free pardon was also offered to any accomplice not involved in the murders who would give the killer up.
Within 24 hours of phone lines being set up, 5100 calls from around Australia were logged. Dealing with such a volume of information created enormous problems of its own, but the scale of the response convinced me that if we were going to catch the killer, it would be with the public’s help.
6
TOO MUCH INFORMATION
Detective Inspector Rod Lynch had worked on plenty of major task force investigations during his time as a detective, but he was taken aback by what he saw when he arrived at Air. ‘I was tasked with setting up an office in Sydney at the former Technical Services building, known as the Old Hat Factory, at the corner of Smith and Campbell. The building had been occupied by several police units, including the Criminal Investigation Branch until it was disbanded in the mid-1980s, and the building had seen little use since. The conditions were shocking.’
The office area was derelict and the whole building was very run-down. Police Properties was quickly brought in and asked to fix it up. That meant not only cleaning the building from top to bottom, but installing furniture and upgrading electricity and electronics for computer networking and security. While this was being done Rod was busy recruiting investigators, analysts and support staff. He also visited the forest to get a first-hand feel for the needs of the investigation.
It was clear from the outset that one of our biggest challenges would be information management. The case had already generated six different investigations: four after the initial disappearances; a more general inquiry by Kings Cross Patrol into missing persons from that area; and a murder investigation by the South West Region Major Crime Squad after the bodies of Caroline Clarke and Joanne Walters were found in 1992.
Information had been treated differently in each case. Some had been recorded in hard copy and kept locally; some had been entered on police databases that could not be cross-referenced with other databases; the murder investigation used the department’s Task Force Information Management System, which was designed as a stand-alone. The murder investigation itself had logged more than 7500 inquiries and questioned more than 500 people.
After the bodies of James Gibson and Deborah Everist were discovered in early October 1993, the information base exploded. We quickly realised the systems we had were hopelessly inadequate: not only did they lack the capacity to handle such a mass of information, but they were also incompatible with each other, making it impossible to cross-check between systems or to retrieve information as and when we needed it. Quality control was either poor or non-existent.
The Task Force Information Management System had been developed by the police in 1990 to manage information during the investigation of major crime. While the system had been widely used since then, it was not designed to handle an investigation of the size and scope of Task Force Air. Furthermore, as a stand-alone system, it couldn’t be used to search records or retrieve information provided by government departments, interstate police agencies and the Australian Federal Police.
The only solution was a more sophisticated computerised information management system, devised and implemented by the police themselves. Working closely with intelligence analysts from the State Intelligence Group as well as police data-management and computer experts, we outlined the kind of system we would need and how we would use it. By combining elements from about a dozen different software packages, some developed internally and others bought off the shelf, they were able to put together a package that allowed us to record, mine and cross-reference data from a variety of sources. Crucially, the task force office in Sydney was networked with Bowral and with the Criminal Research Bureau of the State Intelligence Group, enabling us to update information constantly, and ensure nothing was lost between systems.
Essential as it was, this re-engineering came at a cost. Since the old systems were no longer usable, all the data they contained had to be re-entered into the new system. Inevitably, this took time and caused frustration among task force members who were temporarily unable to access information. But the prize, once it was finished, was a fully integrated system in which every item of information could be logged and retrieved, and the chances of missing a vital clue would be drastically reduced. During the period the system was being developed and put into operation, the investigation depended on the ability o
f individual detectives to maintain a strong and detailed picture of what was occurring.
Our problems with information management were one issue I decided not to share with the media. I could imagine the sort of headlines we could expect if the newspapers found out about the mess we had inherited. However, when news of the new information system broke in March 1994, the tone was positive. Under the headline ‘High tech tool speeds up backpackers’ murder probe’, The Australian Financial Review reported: ‘Criminal investigators working on the backpackers murder case are sieving through more than a million items of information with the help of a powerful Australian high-tech tool . . . The technology, known as NetMap, is helping them probe far deeper than with traditional methods and could save hundreds of man hours as it digs for clues . . . Names, addresses, vehicles, times of day, locations all make up a mountain of information from which the system “mines” similarities and connections. It massages the information, transforming it into a multidimensional, graphical representation.’
The report also mentioned the use of a tool called Farm Image, a remote-sensing package for interpreting satellite images of the Belanglo State Forest to identify changes in the light intensity of vegetation on the dates the backpackers were last seen, and changes between dates, and to attempt to locate possible burial sites by identifying areas of new growth in the forest.
While the task force was getting to grips with NetMap, Sergeant Dutton at the Forensic Ballistics Unit was painstakingly examining the fired cartridges and bullets. Analysis of the 47 Winner cartridge cases found 165 metres from Gabor Neugebauer’s remains confirmed that several had been fired from the same Ruger 10/22 model .22 calibre rifle that had fired four of the six bullets found in two nearby trees as well as the cartridge cases found at the Clarke murder scene. The 46 fired .22 calibre Eley cartridge cases found in the same area and two of the spent bullets recovered from the tree trunks were found to have been fired by an Anschutz .22 calibre bolt-action rifle. Dutton was certain the bullets had come from the Winner and Eley cartridge boxes found at the scene.
Detective Sergeant Royce Gorman went to the Winchester factory in Geelong, Victoria. The manufacturer was able to identify the Winner cartridge packet stamped with the partial batch number ‘. . . CD . . . CF2’ as batch ‘ACD1CF2’. The batch had comprised 320,000 bullets. Working through its records, the company was able to calculate the total stock of .22 calibre bullets on hand at the time the batch was manufactured. It concluded that the batch had been sold to 55 outlets, mostly gun shops, across Australia between 2 June 1988 and 30 November 1988, a year before the first known disappearance. Twenty-seven of those outlets were in New South Wales.
The manufacturer of the Eley bullets clarified the partly unreadable batch number on the Eley cartridge box as ‘J23CGA’, a batch manufactured on 23 March 1979 and distributed to retailers well before the first known disappearance. Unfortunately, the company had no record of which outlets had bought the bullets.
Dutton identified the weapon used in the killings as a .22 calibre Ruger 10/22 model self-loading rifle with a ten-round rotary magazine. It suggested that the killer or killers had fired a full magazine into Clarke. Tens of thousands of such weapons had been imported into Australia. Dutton, however, was able to narrow the search. The fired cartridge cases bore the impression of a firing pin that left an upward indentation. This was a manufacturer’s fault in Rugers made between 1964 and 1982. The fired cartridges had other common features that could help identify the rifle used, while the bullets themselves all had gouge marks along their length caused by a silencer.
The Australian Federal Police (AFP) had built a significant database on Ruger rifles after the 1989 murder of Colin Winchester, an assistant commissioner in the AFP responsible for the Australian Capital Territory. Winchester had been shot twice in the head with a Ruger 10/22 .22 calibre semi-automatic rifle fitted with a silencer as he parked his police car in the driveway next door to his house in Deakin. (David Eastman was eventually arrested and convicted of the murder.) In the course of the investigation the AFP test-fired more than 500 Ruger rifles, the results of which were given to the task force. We also benefited from the advice of Detective Superintendent Ian Prior, head of the AFP’s Firearms and Ballistics Branch, who had been in charge of the forensic ballistics analysis in the Winchester investigation.
The US manufacturer was able to tell us that about 55,000 Rugers of the suspect type had been imported into Australia between 1964 and 1985. Starting with the manufacturer’s and AFP’s records, we set out to build a database of those 55,000 Rugers.
The first task was to rule out Rugers that had come into police possession by way of seizure or other means, and were either still in police possession or had been destroyed before September 1992, the time of the first killing known to have involved a Ruger. Once that was done we could start checking gun shop records (both sales and repairs), along with the records of licensed firearm dealers and holders, to identify individual owners.
During December 1993 task force investigators visited 27 firearm dealers across New South Wales, checking records for both Winner cartridges and Ruger rifles. In fact, we checked the records of several different types of .22 calibre rifle, not wanting to draw attention to Rugers in the initial stage of our investigation, as we were concerned that the killer or killers would get rid of the weapon if alerted that the task force was looking for a particular type and model.
As in other aspects of the investigation, we adopted two strategies in our search for the Ruger rifle, one macro and the other micro. We wanted to trace every Ruger rifle manufactured between 1964 and 1985 that had been imported into Australia. At the same time we would check the records of gun shops, gun owners and gun clubs, and test-fire individual weapons where necessary, in order to eliminate as many Rugers as possible from the investigation.
Working from the top down and from the bottom up would, we hoped, enable us to identify both a number of unaccounted-for Rugers and a number of people who could not account for their Ruger. The two-pronged strategy ought to expose omissions and highlight mismatches, while allowing us to narrow the focus to specific geographic areas and to individual gun-owners who could not be eliminated.
It was laborious and time-consuming, but we were not in a race.
7
HOTLINE
On Saturday, 16 October 1993, Detective Sergeant Kevin Hammond and a colleague visited the Belanglo Pistol Club during a competition shoot to ask whether anybody had seen anything suspicious or unusual, or had any information that might be useful to the investigation. Hammond had been seconded to the task force from his job as head of the Bowral detectives’ office. Bill Ayres, from nearby Buxton, suggested to Hammond that he should speak to another member, Alex Milat, who might have some information that would interest him. After speaking to Alex at the club, Hammond arranged for him to be interviewed at Bowral Police Station.
Two days later, 52-year-old Alexander ‘Alex’ Milat told Hammond that around 4 p.m. on Sunday, 26 April 1992, he and Bill Ayres had left the pistol club after a day’s shoot. They were in Alex’s Holden Rodeo 4WD, with Ayres driving and Alex in the passenger’s seat. They were travelling down the unsealed Belanglo Road towards the Hume Highway, said Alex, driving slowly to avoid hitting any of the wombats or kangaroos that often crossed the road in the late afternoon. As they approached an intersection with a track that led into Belanglo State Forest, Alex saw two vehicles coming towards them, the leading vehicle signalling its intention to turn right into the forest. After passing the intersection Ayres moved to the right-hand side of the road, causing the oncoming cars to pass on the passenger’s side. According to Alex, the first vehicle was a chocolate-brown 1980-model Ford Falcon sedan and the other a 4WD dual-cab utility, either a Holden Rodeo or a Nissan Navara, beige on the bottom and brown on the top.
Alex said that as the cars passed each other, he noticed that the driver of the Falcon had what appeared to be a tattoo on the fingers of his le
ft hand. ‘[H]e appeared to be a tall person . . . a Caucasian of medium complexion . . . thin build . . . about 100 kilograms in weight, aged in his mid-twenties . . . [with] . . . a prominent nose and Adam’s apple . . . a flat top hair style and . . . “mutton chop” style side levers. His hair was coloured orange, red colour . . . [and] . . . he had acne spots or marks on the cheeks of his face.’
Alex went on to describe the physical characteristics of the Falcon’s male passenger in similar detail, before adding that he was ‘holding what appeared to me to be a shotgun 410 model. He had hold of its barrel which was pointing upwards towards the roof of the vehicle and the stock appeared to be on the floor beside him.’ In the rear passenger seat, Alex claimed to have seen a woman aged in her twenties:
with shoulder length or slightly longer mousy coloured hair, a Caucasian or fair complexion and she appeared not to be a heavy build. I saw what appeared to me to be a ‘gag’ which consisted of a length of honey coloured material which was wrapped around her head across her mouth. Outside the mouth the material appeared to be knotted a number of times . . . I saw that this female was looking at me and she appeared to sit up in the back seat with her arms by her side as if she was trying to attract my attention.
According to Alex, the woman was sitting between two men, although he could not describe either of them.
As the second car passed, he saw two men in the front and another man in the back, who was sitting next to a woman. The woman ‘had a “gag” consisting of a piece of honey coloured material’, though it was a different colour to the material he had seen wrapped around the woman in the Falcon. It was, Alex said, ‘wrapped around her head across her mouth . . . and as we were driving past she also sat up in the seat and looked at me with her eyes wide open as if she was frightened . . . this female was aged in her twenties, a Caucasian of fair complexion . . . a heavier build than the female I had seen in the Falcon [and had] dark brown hair’.
Milat Page 4