Content uploaded by Gerard Nicholas Labuschagne
Author content
All content in this area was uploaded by Gerard Nicholas Labuschagne on Mar 28, 2024
Content may be subject to copyright.
An Examination of Serial Homicide in South Africa: The
Practice to Research Link
GÉRARD N. LABUSCHAGNE
1,
*and C. GABRIELLE SALFATI
2
1
South African Police Service & Division of Forensic Medicine and Pathology, University of the
Witwatersrand, South Africa
2
John Jay College of Criminal Justice, USA
Abstract
In this paper, South African serial homicide cases will be used to illustrate how practice can
inform research and how research can aid practice by highlighting key questions that need to
be answered and practice-based assumptions that need to be empirically tested. The cases
will also be used to highlight some of the unique features of series in South Africa. Copyright ©
2014 John Wiley & Sons, Ltd.
SERIAL HOMICIDE ACROSS THE WORLD
It is widely accepted that serial homicide occurs throughout the world (Gorby, 2000). Cases
have been documented in academic literature in Australia (Gorby, 2000), Japan (Aki,
2003), Germany (Harbort & Mokros, 2001), Israel (Kallian, Birger, & Witztum, 2004;
Kallian, Bar-El, Durst, & Witztum, 1996), Russia (Myers, Bukhanovskiy, Justen, Morton,
Tilley, Adams, Vandagriff, & Hazelwood, 2008), Canada (Campos & Cusson, 2007;
Keppel & Birnes, 2009), the UK (Jenkins, 1988), the USA (Jenkins, 1989), and Italy
(Campobasso, Colonna, Carabellese, Grattagliano, Candelli, Morton, & Catanesi, 2009).
Egger (2002) comments on cases occurring in Colombia, Ecuador, Peru, Turkey, Thailand,
Nepal, Singapore, France, Nigeria, China, Sweden, and Hungary. Labuschagne (2010) also
comments on cases in South Africa, Namibia, India, and Swaziland.
Serial homicide is defined by the South African Police Service (SAPS) as occurring
when a suspect(s) murder two or more victims on at least two separate occasions and the
motive for the homicides are not for primarily for material gain nor to eliminate a witness
in another matter. This is consistent with the definition conferred at the Federal Bureau of
Investigation’s 2005 Serial Murder Symposium. In South Africa, by 2007, the SAPS had
identified 131 homicide series, of which 74% (97) were solved. The majority of these were
identified since 1990. The earliest noted homicide series occurred in 1936 (Labuschagne,
2007). The Investigative Psychology Section of the SAPS has official court records of a
*Correspondence to: Gérard Labuschagne, South African Police Service & Divisionof Forensic Medicine and Pathology,
Faculty of Health Sciences, University of the Witwatersrand, Private Bag 3, WITS, 2050, South Africa
E-mail: docgnl@saps.gov.za
Journal of Investigative Psychology and Offender Profiling
J. Investig. Psych. Offender Profil. 12:4–17 (2015)
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jip.1415
Copyright © 2014 John Wiley & Sons, Ltd.
series dating back to 1953 where the accused, Elifasi Msomi, killed 15 people in the prov-
ince of KwaZulu-Natal after luring them with a con story of employment, a method still
seen today. The highest number of homicide convictions is attributed to Moses Sithole
who raped and killed 38 Black women from 1994 to 1995.
The predominant understanding of serial homicide in the academic literature is from a
western point of view. In South Africa, a small but growing body of research exists, yet it
is still only a handful of studies (Barkhuizen, 2006; Del Fabbro, 2006a, 2006b; De Wet,
2005; Du Plessis, 1998; Hodgskiss, 2001, 2004; Hook, 2003; Hurst, 2003a, 2003b; Knight,
2006; Labuschagne, 1998, 2001, 2006, 2007,2010; Lemmer, 2003; Pistorius, 1996; Snyman,
1992). Most of these studies to date in South Africa have, however, been part of small-scale
unpublished studies by masters and doctoral students, mainly performed at the University of
Pretoria. Besides Labuschagne (2010), often, the studies lacked any practical applicability and
held little investigative relevance.
Although many South African studies did involve interviewing the actual offenders and
accessing to original police case files, the samples were small, generally two offenders, with
the largest sample being that of Hodgskiss (2001), which included 13 South African serial
offenders. Pistorius, in an unpublished 1996 study, was the first to conduct research on South
African serial homicide and did so from a psychodynamic perspective, based on two investi-
gations she had worked on. Thereafter followed Labuschagne (1998, 2001) who focused on a
systemic interactional approach to understanding serial offenders based on interviews with
offenders and, later (2006), a case study on the use of a linkage analysis as evidence during
the trial of a serial offender based on an investigation and trial he was involved in, and a study
on foreign object insertion that also commented on the occurrence of such insertions in serial
homicide cases (2007).
Du Plessis (1998) explored serial homicide from a grounded theory approach, based on
interviews with offenders. Later, Hodgskiss (2001, 2004) reviewed offender characteristics
and behaviours of 13 incarcerated offenders whom he interviewed. Hook (2003) conducted
a post-structural deconstruction of narratives of a person who committed serial homicide
based on interviews with academics in the fields of criminology and psychology, media
reports, and unpublished research. De Wet (2005) approached the topic from a psychosocial
perspective based on interviews with offenders, and Barkhuizen (2006) used the same method
for his intrapsychic object relations approach. Del Fabbro (2006a, 2006b) explored serial
homicide from a family systems paradigm based on interviews with the offenders, and where
available, their family members. Hurst (2003a, 2003b) conducted a philosophical analysis of
discursive dynamics in the court trial of an individual who committed serial homicide. Her
data source was court transcripts. Lemmer (2003) did a comparative document study between
three South African and four American serial offenders. Knight (2006) published an article
from an object relations point of view on the roots of the behaviour of serial offenders as
narcissists. This paper appears to be based on a brief review of the literature and her own
thoughts on this topic and not based on a research process. Snyman (1992) reviewed interna-
tional literature on the topic of serial homicide. (For a more detailed review of South African
perspectives on serial homicide studies, see Del Fabbro, 2006a, 2006b.)
Although there may only be a small body of South African serial homicide research, there is
by no means a large body of international research on the topic. International research also
suffers from limiting aspects such as lack of access to actual police records and interviews
with offenders, and small sample sizes. This makes it difficult to determine baselines for
behaviours and nearly impossible to create a comparative sample against which to compare
any further research. In such situations, the individuals involved in the investigation of these
Serial homicide in South Africa: practice to research 5
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
types of crimes become the informal database of experiential knowledge against which cases
are compared. Although this is a valuable source of information, its unstructured nature, lack
of accessibility to a larger audience, and fallibility of the human mind make it far from ideal.
As a result, the need for large-scale studies, using police file information, is essential in creat-
ing an open, scrutinizable body of knowledge. This was a primary motivation for the collab-
orative research reflected in this special edition. Recently, the Investigative Psychology
Section of the SAPS and the Investigative Psychology Research Unit at John Jay College
of Criminal Justice in New York have completed a collaborative study on serial homicide
in South Africa. This is to date the largest empirical study internationally on serial homicide
making use of original police case files and using a thoroughly empirically validated data
collection tool, the Homicide Profiling Index (Salfati, 2005, 2007). The study includes 33
of the total of 53 (62%) identified series between 1953 and 2007, the results from which are
included as part of this special issue of the Journal of Investigative Psychology and Offender
Profiling. It is an attempt not only to develop a more intimate understanding of serial homicide
in South Africa but also to create a body of knowledge against which other similar studies can
be compared to determine similarities and differences.
In this paper, examples of South African serial homicide cases will be discussed, briefly
illustrating some of the unique features of series in this context. The case studies will primarily
be used to illustrate how these practical issues can inform key questions for a research agenda
and how results from practice-led research can ultimately aid practice, providing a stronger
empirically sound basis for evidence-led policing. In doing so, research can help confirm or
refute the assumptions held by investigators gleaned from their experiences and can aid in
providing an understanding of baselines and correlations that may help understand the type
of crime at hand and inform investigations.
Each of the following case studies chosen aims to highlight unique questions and aspects
of serial homicide cases, some of which are peculiar to the South African context.
CASE STUDIES
Case example: the Newcastle Homicide Series (state versus Themba Anton Sukude:
case number cc34/2006 in the High Court of South Africa, Natal Provincial Division)
This case involved four separate incidents. In this series, the offender attacked his victims
in a small geographic area bordering a small town. The offender lived nearby the crime
scenes, a feature consistent with overseas experiences regarding geographic profiling
(Rossmo, 2000). This offender took no precautions in hiding his identity, saving for
committing his crimes in the evenings, nor did he take forensic precautions such as using
a condom during his two rapes to prevent leaving DNA, a fairly common feature of South
African serial homicide offenders
1
and serial rapists as experienced by investigators and
supported by limited research (Woodhams & Labuschagne, 2012). All but one of the
victims in this series, an Indian male, were Black. It is not uncommon for South African serial
offenders to have victims of a different population group to themselves in their series. What
was different about this series in comparison with what is often experienced with South
African serial homicide is that the offender used a sudden violent approach to his victims,
instead of the frequently used con story to lure victims away to a secluded place.
1
Other South African serial murderers who targeted couples include the Thohoyandou Serial Murderer, David
Mbengwa, and the Wemmerpan Serial Murderer, Cedric Maake.
6G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
The first incident
At 1800h on Saturday, 14 February 2004, in the summertime, a young Black couple
walked to a park bordering the town of Newcastle in the KwaZulu-Natal province. After
consuming alcohol, they decided to engage in sexual intercourse. Whilst they were
engaged in intercourse, the offender crept up to them and without warning struck the man
on the head with a large rock. He pushed the male victim off the female victim and slapped
the woman with an open hand whilst she tried to keep the offender at bay. The offender then
hit the male victim with the rock for the second time before dragging the female victim away.
When the female victim screamed, the offenderdisplayed a knife to her, threatened to stab her,
and dragged her to a nearby ditch approximately 10 m from where the initial attack occurred,
where he raped her once. After the rape, he told her to dress and leave. The man died at the
scene at the same location where he was struck. Cause of death was determined to be blunt
force trauma to the head; there were no other wounds inflicted on the homicide victim. The
bloodied rock was found at the scene, approximately 1 m from the head of the deceased,
who was lying on his back. He was clothed, but his pants were unzipped and his belt undone.
Nothing was stolen from either of the victims.
The second incident
On the evening of Wednesday, 27 October 2004, at approximately 2300h, in the summertime,
at the same park as the first incident, a young Black couple decided to engage in sexual
intercourse. The male victim, a mini-bus taxi driver, parked his taxi in an open space.
Whilst engaging in intercourse on a blanket next to the mini-bus taxi, the offender crept
up to them and struck the man on the head with a large rock. The offender told the female
victim not to dress, displayed a knife, and took the female victim to the nearby river,
approximately 20 m from the mini-bus taxi, where he raped her once. Afterwards, he
returned to the deceased’s vehicle with the female victim; he took the keys from the
deceased’s pocket, opened the vehicle, and took two cellphones. The victim asked for
money from her purse for transport home, which the offender agreed to give to her.
Thereafter, the offender set fire to the vehicle and fled the scene without giving the victim
the promised money. The female victim then also fled the scene. The fire quickly burnt
itself out and did not cause any significant damage to the vehicle. The man died at the
scene at the same location where he was struck by the offender. The bloodied rock was
found at the scene also approximately 1 m from the deceased’s head. The cause of death
of the male victim was determined to be blunt force trauma to the head; there were no other
wounds inflicted on the homicide victim. The deceased was found lying on his back, next
to the blanket the victim had placed on the ground. His shirt and shoes were on, and his
pants were around his ankles. His underwear was in place with his penis exposed.
The third incident
In the early hours of Friday, 26 November 2004, nearby the scene mentioned in the second
incident, the body of an adult Black man was found along a footpath by a passerby; he was
wearing only his underwear and a shirt. A shoe was found approximately half a metre from
his body. The pants and the other shoe were not found at the scene. At the scene, a large
bloodied large rock was found approximately 1 m from the deceased’s head. The cause of
death was blunt force trauma to the head. No female victim came forward to report any rape
or assault.
Serial homicide in South Africa: practice to research 7
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
The fourth incident
On the evening of Friday, 7 January 2005, at about 2100h, the deceased, an adult Indian man,
left his relative’s residence with his sports bag containing his personal belongings. He then
proceeded to the park mentioned in the first and second incidents. In the early hours of
Saturday, 8 January 2005, he was found unconscious with severe blunt force trauma to the
head and a large bloodied rock nearby, approximately 8 m away, down a small embankment
next to where the victim was discovered. He was found wearing a yellow t-shirt and Black
jeans but no shoes. One shoe was found nearby the body. This scene was nearby the second
incident and in between the locations of the first and second incidents. A large roll of money,
belonging to the victim, was found next to his body. He was taken to hospital by ambulance
and died 2days later without ever regaining consciousness. His sports bag was missing from
the scene. No female victim came forward to report any rape or assault.
The state’s evidence
The state’s evidence was strongest on the first two incidents. Both the rape victims had pointed
the offender out at either an identity parade (line-up) or a photographic identity parade. Unfor-
tunately, there was no DNA evidence although both victims went for a medicolegal examina-
tion and a sexual assault evidence collection kit was taken for each rape victim. During
interrogation, the offender admitted to killing the men in the first two incidents but refused
to do a formal pointing-out of the crime scenes or a formal confession to an independent
officer not previously involved in the investigation or a magistrate. Unfortunately, the inves-
tigating officer in this case was a non-commissioned officer, and in South Africa, a confession
to a non-commissioned officer is not admissible as evidence. Confessions are only admissible
in court if they are made to a commissioned officer (i.e. from the rank of captain and above)
who was not involved in the investigation or to a magistrate. A confession must also be an
‘unequivocal acknowledgement of guilt, the equivalent of a plea of guilty before a court of
law’as indicated in R v. Becker (Schwikkard, 2009). In both instances, the confession must
be in writing, after the necessary legal warnings have been given to the offender. Therefore,
if an offender spontaneously confesses, even after having been warned of his rights in
accordance with South African law (the equivalent of the Miranda Rights), this will not be
admissible. The purpose of this is to build in safeguards to avoid false and coerced confes-
sions. However, the offender had also admitted to a member of the public that he had killed
people in the park. In South Africa, a confession to a member of the public is admissible as
evidence in court, the weight thereof which will be determined by the court.
In the last two incidents, however, there were no eyewitnesses, surviving victims, nor
physical evidence linking the offender to the crimes. It was here that the decision was made,
in consultation with the Director of Public Prosecutions, to also charge the accused on the third
andfourthcountsofmurder,relyingonsimilar fact evidence. A linkage analysis was then
compiled in support of similar fact evidence, which was accepted by the court. The offender
was convicted on all counts. This was the first time that linkage analysis evidence had been
presented in the trial of a serial homicide offender in South Africa (Labuschagne, 2006).
From questions in the field to research agenda
The first issue in this series was the targeting of two victims in at least two of the instances by
one offender, in particular with regard to how common it is that offenders target couples.
Although investigators have come across series where couples were targeted, there is no
8G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
research to inform how often this happens in South African serial homicides. As such, this high-
lights a key practice-led research question. A related issue that has clear practical utility for a
possible linkage analysis relates to how often offenders only kill one of the two victims, if they
start targeting couples. And if they start targeting couples, do they remain consistent across their
series, or if they change, what is their pattern? Because of questions such as these, a recent study,
aiming to provide baselines relating to the aetiology of serial homicide in South Africa, has spe-
cifically looked at detailing the nature of the victims targeted as part of these series and has
shown that 48.5% of offenders target male and female victims during the course of their series
and 40.6% had a mixture of single and multiple victims (Salfati, Labuschagne, Horning,
Sorochinski & De Wet, 2015b). Salfati, Horning, Sorochinski, and Labuschagne (2015a) have
additionally provided figures relating to consistency patterns and have illustrated that couples
and live victims are not consistently targeted in South African series as a whole with couples
being targeted in only 13.4% of incidents and a victim being left alive in 19.1% of incidents.
Targeting of male victims is also not a common feature of South African serial homicide with
men being victims in 36.4% of all incidents. In terms of consistency, their study illustrates that
across the first three incidents, 42.3% (11/26)
2
consistently targeted a particular type of victim.
Eleven per cent (3/26) of offenders selected men across the first three offences. Only 3% (1/30)
of offenders targeted couples consistently across the first two offences, and none continued to
target couples after the first two incidents. Of offenders, 3.8% (1/26) had live victims across
the first three incidents, but this pattern did not continue past the third victim in the series if
the series continued. Across the first three offences, 30.8% (8/26) of offenders consistently
targeted vulnerable victims. Research, as stated earlier, could have been used to provide statis-
tics to support the experienced-based assumptions in the linkage analysis presented in the court.
This offender took no precautions in hiding his identity, saving for committing his crimes in
the evenings. He let the two female victims live, he did not make any efforts to destroy DNA
evidence, he did not take fingerprint precautions, nor did he wear a mask to hide his face, a
feature also seen in South African serial rapists (Woodhams & Labuschagne, 2012). Questions
remain regarding whether these types of behaviours change over the series, so that offenders
become more forensically aware, and thus only targeting one victim, leaving less evidence,
and so on. Following up on questions like these, Sorochinski, Salfati, and Labuschagne
(2015) found that South African serial homicide offenders rarely engaged in what are typically
regarded in developed countries as post-offence planning behaviours, those being behaviours
to cover up the crime and avoid detection.
All but one of the victims in this series, an Indian man, were Black. Questions regarding
consistency in type of victim targeted remain uncalculated to date, and although investigators
will state that it is not uncommon for South African serial homicide offenders to have victims
of a different population group to themselves in their series, this needs to be backed up by
more empirical derived numbers in order to hold as linking evidence. Salfati, Labuschagne
et al. (2015b) found in their study that 31.2% of offenders had multiple racial groups as vic-
tims. This has obvious relevance for case linkage when using victim similarity as a criterion,
illustrating the need for further research in this area. This is not only relevant when testifying
about victimology during a trial as part of a linkage analysis, but this is also relevant during the
investigation phase when deciding which cases to include as part of a serial investigation
when other evidence, such as DNA, is being processed or is not available.
2
One offender consistently targeted young boys. The breakdown of victim consistency by type includes his con-
sistency in both targeting vulnerable victims and men across the first two, three, and four offences. However, in
the total percentages for victim consistency, this offender is only counted once.
Serial homicide in South Africa: practice to research 9
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
In this instance, the offender attacked his victims in a small geographic area bordering a
small town. The offender living nearby the crime scenes is a pattern of behaviour that has
been noted in international research on serial offenders (Rossmo, 2000), but as yet, there
have been no studies to look at whether this is comparable with patterns in South Africa,
although this has been observed by investigators.
As stated earlier, what differed in this series from other South African serial homicides is
that the offender used a sudden violent approach to his victims, instead of a con story to lure
victims away to a secluded place. However, the number of cases that fall into each of these
categories is still only alluded to, and no research-based figures exist, nor do figures regarding
whether an offender’s pattern changes over time in relation to approach style. In this series, the
suspect made exclusive use of what is often referred to as a blitz-style approach, something, as
later research has shown, only seen in 25.9% of South African incidents (Horning, Salfati, &
Labuschagne, 2015), which is comparable with figures seen in serial rape of 23% (Woodhams
& Labuschagne, 2012). Furthermore, an offender, first watching a couple engaging in sexual
activity, is only present in only 2% of serial homicide incidents (Horning et al., 2015).
Case example: the Quarry Homicide Series (state versus Richard Jabulani
Nyauza: case number cc97/2007 in the High Court of South Africa, Transvaal
Provincial Division)
This series involved 16 separate homicide victims, in two separate time frames and place
locations. From an investigator’s point of view, this offender is perhaps a textbook example
of South African serial homicide offender. He lured unsuspecting women, often with the
promise of employment, to a secluded place where he raped and killed them, leaving their
bodies at the same scene where they were killed. He also clustered his scenes around two
designated geographical areas approximately 2.8 km from each other. In between his two
series, he was in jail awaiting trial for an unrelated attempted rape of a child. Once he was
found not guilty of the attempted rape and was released from custody, he continued his homi-
cides within a few weeks of being released. In this series, the offender, Richard Nyauza, had
three different methods of killing his victims; these were by means of stabbing, strangulation,
and blunt force trauma. From an investigative experience perspective, it has been suggested
that it is a fairly regular feature of South African serial homicide offenders to make use of
more than one method of killing their victims, an important aspect to be aware of when
assessing cases for linkage purposes (Labuschagne, 2010). As with the previous case
example, the offender took no precautions to hide his identity; he committed his crimes during
the day, was seen by witnesses with some of his victims, took no forensic precautions to pre-
vent leaving DNA, and took traceable items from some of his victims such as cellphones—
which played a role in leading the police to him.
The first series
From January to September 2002, the bodies of five unidentified adult Black women
were found along a small river near to a highway just outside of Pretoria near a township
known as Olievenhoutbos. These five cases were initially investigated as part of another
series of bodies found along highways, known as the ‘Highwayman’series, but were
later treated as a separate serial investigation once the Highwayman, Elias Chauke,
was linked to certain homicides but not those near Olievenhoutbos, which occurred
whilst he was in prison for another matter. As a result, after all investigations were
10 G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
exhausted for the 2002 Olievenhoutbos homicides, these cases were closed off as
inquests before a magistrate.
The second series
Then, in early January 2006, two more bodies of adult Black women were found approxi-
mately 2.8km from the nearest 2002 body. This sparked concerns that the same offender
was again committing homicides after a 4-year hiatus. The 2002 task team was reformed with
the first author of this paper as the head of the task team. This series was labelled as ‘Quarry’
homicide series because of the bodies being found in close proximity to a nearby open-pit
mine known as a ‘quarry’. From January to September 2006, a total of 11adult Black women
were found killed, and a 12th case, that of attempted homicide, also took place. To date, only
six of the total 16 homicide victims have ever been identified, a common problem experienced
in South Africa.
The state’s evidence
In eight of the homicide cases, there was evidence linking that offender to the crimes. This
evidence was in the form of either cellphone evidence in two cases (the offender was in the
possession of two victims’cellphones) or DNA in four cases (from vaginal swabs) or a formal
pointing-out of the crime scene in seven cases. A ‘pointing-out’is a formal police procedure
where the offender agrees to point out any locations that he wants to, to an independent
member of the police who is a commissioned officer and who was not involved in the inves-
tigation; all locations indicated by the offender are photographed, and what is reported during
this procedure is also written down and regarded as a confession. The legal warnings and
requirements are similar to those described earlier in relation to a confession, except that a
pointing-out is performed only to a member of the police and not with a magistrate. Only
one case had all three of these evidential aspects.
Initially, the state prosecutor only intended to indict the offender on the eight incidents
where there was one of the three previously mentioned evidential aspects and on the attempted
homicide where there was eyewitness evidence and circumstantial evidence to link the
offender to the offence. Consultations were held between the investigation task team and
the prosecutor regarding indicting the offender on 16 murders and the attempted murder.
Eventually, the prosecutor agreed to do so. The accused was subsequently found guilty on
all counts, eight of the incidents’conviction relying solely on linkage evidence.
From questions in the field to research agenda
The offender in this series displayed behaviours that are often associated with South African
murder series by investigators, notably the luring of unsuspecting women, often with the
promise of employment, to a secluded place where he raped and murdered them, leaving their
bodies at the same scene where they were murdered. Although this was the experience of
investigators, there was no empirical research to support the experiences of the investigators
of what per cent of offenders display this behaviour and how many are consistent in this
behaviour throughout their series. Horning, Salfati and Labuschagne (2015) in a study follow-
ing up on the issues relating to consistency of victim type showed that the offender, using a
con story to lure his victims, occurred in 47% of incidents in South Africa. As many cannot
afford a vehicle to use in the commission of their crimes, they have to lure the victim with
Serial homicide in South Africa: practice to research 11
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
them by foot and public transport to the final crime scene and to the location of where the body
was ultimately discovered.
The offender clustered his scenes for each series around a designated geographical area.
The geographic pattern of the crime scenes was an aspect noted locally and internationally,
thus indicating a feature that can possibly be regarded as a base-rate feature of serial murder
behaviour throughout the world. The 2002 murders were grouped close by to each other,
whereas the 2006 murders were grouped close by to each other approximately 2.8km away.
During the 4-year hiatus, housing construction had started to take place in the area of the 2002
murders, which most likely contributed to the offender relocating 2.8 km away when he began
murdering again in 2006.
In between his two series, the offender was in custody awaiting trial for an unrelated
attempted rape of a minor. Once he was found not guilty of the attempted rape, he contin-
ued his murders within a few weeks of being released. Questions regarding recidivism, as
well as possible change in behaviour due to intervening time or experience, are key to
understand, especially in relation to linking. This is also relevant in relation to the issue
of recidivism. This period of incarceration did not deter nor modify the offender’sbehav-
iour upon release. It is often believed by practitioners that serial murderers cannot be
rehabilitated. Salfati (2008) discusses the issue of what specificfactorsmayinfluence
consistency, especially with regard to issues relating to other events in the offender’slife,
including other crimes, andtime in prison. However, this issue as yet has not been investigated
in South Africa.
The offender had three different methods of murdering his victims: by means of
stabbing, strangulation, and blunt force trauma. It is highlighted by investigators that it
is a fairly regular feature of South African serial murderers to make use of more than
one method of killing their victims in a series, an important aspect to be aware of when
assessing cases for linkage purposes. However, no figures existed on this at the time of
the series, nor of its relationship to differentiating offenders, or in relation to determining
consistency patterns across an offender’s series, an issue relevant for case linkage and
signature analysis.
In this example, only five of the 16 murder victims were identified; Salfati, Labuschagne,
Horning, Sorochinski, and De Wet (2015b), in their later study, showed that in South Africa,
18.2% of serial murder victims are never identified.
All victims were murdered outdoors and bodies recoveredwhere the murders took place—
a feature that Horning et al. (2015) later showed occurs in 78.1% of South African serial
murder incidents.
As with the preceding case study, one of the keyissues highlighted was the lack offorensic
or identification precautions taken by the offender, despite having been previously arrested
and put on trial for the attempted rape of a young girl.
3
The offender took no precautions to
hide his identity; he committed his crimes during the day, was seen by witnesses with some
of his victims, took no forensic precautions to prevent leaving DNA, and took traceable items
from some of his victims such as cellphones—which played a role in leading police to him.
The role of forensic awareness or an offender’s care to manipulate this aspect of the crime
scene remains key for investigations and is thus essential to look at in further detail.
This was also the second time that a linkage analysis had been used as evidence during a
trial for its probative value (Labuschagne, 2010).
3
However, the attempted rape case involved his then girlfriend’s daughter, which can possibly be one factor that
later led to him choosing stranger victims for his murder series.
12 G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
Case example: the Concordia Homicide Series (state versus Robin Cloete: case number
ks35/2005 in the High Court of South Africa, Namakwaland Provincial Division)
In contrast to most homicides, serial offenders have been recorded as tending to target
strangers as their victims (Hickey, 2006; Egger, 2002; Proulx, Beauregard, Cusson, &
Nicole, 2007). However, the definition of serial homicide does not specify what the nature
of the relationship must be between victim and offender. In this case example, the offender,
Robin Cloete, targeted three victims, all known to him. His first homicide victim was his ex-
fiancé; the second and third homicide victims were his ex-girlfriend and her mother. Most
people who commit pre-planned intimate partner homicides receive lengthy sentences in
South Africa, except in this case where the first homicide occurred in 1993 and the offender
received a relatively short sentence and was out on parole in under 10 years. As described later
in the text, this atypical victimology raised the question during the trial as to whether the
offender could be classified as a serial homicide offender. What was also different in compar-
ison with the two case examples discussed earlier was the large distance and time frame of the
homicides. His first homicide was committed in Cape Town, whereas the second and third
homicides (occurring at the same time) were committed in the small town of Concordia,
which is in a different province and approximately 500km away.
The first incident
In 1993, the victim, living in Cape Town in the Western Cape province of South Africa, had
terminated the engagement with the offender after learning that he had fathered a child with
another woman. In the ensuing 3weeks, the offender tried to make contact with the victim
but was refused. On the morning of the incident, the offender consumed a small amount of
alcohol, took his brother’s police issue firearm, borrowed a friend’s vehicle, and drove
50 km to the victim’s place of work, a university in Cape Town. He then entered her building,
walked into her office, and proceeded to argue with the victim. Co-workers of the victim came
to investigate and were fired upon without injury. The offender then proceeded to shoot the
kneeling victim eight times. He was arrested on the scene. After he shot the victim, he told
a bystander the following: ‘I have shot her, I am satisfied’. The offender was found guilty
of murder, one count of attempted murder, one charge of theft of ammunition, and illegal pos-
session of a firearm. He was found fit to stand trial and found to have acted with premeditation.
He was subsequently sentenced, and after 9 years and 10 months of his 14-year sentence, he
was granted parole in September 2004.
The second incident
After parole from the 1993 case, the accused returned to his home of origin in the Northern
Cape province, to a town called Concordia. He met the next victim a few months after his
parole was granted, and he dated her for 2years and 10 months prior to the homicide, which
occurred in 2007, and fathered a child with the victim. Similar to the first incident, approxi-
mately 3weeks prior to the incident, the victim terminated their relationship after the offender
assaulted her one morning. There ensued a 3-week period where he attempted to have contact
with the victim, which was unsuccessful. On the morning of the incident, he was under the
influence of alcohol and attempted to contact her at her parental home twice. The third attempt
was when the homicide occurred. The offender took a knife from his home, went to the
victims’residence during the daytime, and, in the presence of numerous witnesses who were
visiting, forced his way inside the house. The two victims then fled to a neighbouring house.
Serial homicide in South Africa: practice to research 13
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
He kicked the door of the neighbour’s house in and entered. He proceeded to stab both
victims. The girlfriend was stabbed three times in the torso, and the mother was also stabbed
three times in the torso. The offender then began to damage the interior of the house and
remained on the scene until the police arrived and arrested him.
The offender was found guilty of assault with the intention to do grievous bodily harm
(for the assault on the girlfriend 3 weeks prior to the homicide), assault common (on one of
the witnesses at the deceased’s home at the time of the homicide), two counts of murder,
and two counts of malicious injury to property.
The state’s evidence
In both incidents, there were eyewitnesses to the attacks, and the offender was arrested at the
scene. In the first incident, the offender claimed to have been suffering from a dissociative
disorder at the time of the actual homicide, and in the second incident, the offender claimed
to be acting in self-defence. In both trials, the version of the accused was rejected. During
the second trial, held in 2009, the first author testified that the accused now qualified as a serial
homicide offender because of the second and third victims. This was hotly contested by the
defence who stated that the accused’s behaviour was not consistent with the ‘mainstream’
serial homicide offender in South Africa because of his victimology and modus operandi.
The state argued that although the victimology and modus operandi of the accused were not
similar to the typical South African serial offender, he did fulfil the criteria as he had killed
at least two people and on at least two separate occasions and that the homicides were not
primarily for financial gain nor to eliminate a witness in another case.
From questions in the field to research agenda
The definition of what constitutes serial homicide has been hotly debated and has been
influenced by both clinical, investigative, legal, and statistical criteria. How cases are
defined and included will have a large impact on the results of studies and needs to be taken
into account in interpretation of results.
Like the Newcastle example earlier, the issue of multiple victims per offence was relevant.
The offender had one homicide victim and two attempted homicide victims with his first
incident and two homicide victims in his second incident, challenging the belief that serial
homicide is always a one-on-one event. Finally, as was raised during the sentencing proceed-
ings of the second trial, serial homicide classification issues were disputed, more specifically
the required criteria versus the popular image of serial homicide as informed by the media.
Linked to this was the issue of serial offenders targeting intimate partners and other victims
with whom they have a close relationship. In South Africa, serial offenders target strangers
as victims in 73.3% of incidents (Salfati, Labuschagne, Horning, Sorochinski, & De Wet,
2015b). This research therefore supports the practice-based assumption that serial homicide
victims are known to the offender in many instances.
Unlike most serial homicides in South Africa (65.7% according to Salfati et al., 2015b),
there were no overt sexual themes observable on these crime scenes. This is similar toa South
African study by Abrahams, Martin, Jewkes, Mathews, Vetten and Lombard (2008) who
found that intimate partner femicides rarely have a sexual component to them. This hints
towards the assumption that when serial homicide is committed in relation to intimates, there
is less likelihood that it will have a sexual theme unless staged as such. This also contradicts
the common misperception that serial homicide is always sexually motivated, as has even
been mentioned in certain earlier definitions of serial homicide.
14 G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
In contrast to most single homicide offenders, serial homicide offenders have often been
described as more likely to target stranger victims (Hickey, 2006; Egger, 2002; Pakhomou,
2004), but it is not unheard of that South African and indeed overseas serial homicide of-
fenders do have victims that are known to them. South African serial offender Stewart
Wilken killed one young stranger runaway aged 12 years, four sex-workers, an ex-girlfriend’s
son aged 12 years with whom he was acquainted, and his own daughter aged 10years
(Hickey, 2006). The so-called Moffat Park serial offender Gcinumzi Richman Makwenkwe
in Johannesburg killed five victims: the first victim was his ex-girlfriend, the second was the
boyfriend of his ex-girlfriend, and the final three were strangers to him. Little is, however,
known about the exact role of type of victim in different types of series and whether victim type
is a consistent element as the series evolves, or whether it changes, and if so, in what way.
What was also different in comparison with the two case studies discussed earlier was the
large distance and time frame of the homicides. The first homicide was committed in Cape
Town, whereas the second and third homicides (occurring at the same time) were committed
in the small town of Concordia, which is in a different province and approximately 500 km
away. However, in both cases, as victims were known to the offender and in the offender’s
social circle, this added proximity. Proximity and geography must thus be looked at as a
feature of the offender’s own lifestyle. Although the offender was initially incarcerated for a
number of years after the first incident, his second and third homicides occurred more than
2 years after being released on parole. Like with the Quarry case, a period of incarceration
did not prevent the offender from committing subsequent murders after release.
In this series, the offender was an extreme example of not taking precautions to prevent
identification. He targeted victims well known to him, during the day, either at their work-
place or home, in the presence of multiple victims. Also, in each instance, the offender did
not flee the scene and was arrested on the crime scene.
As with the previous case studies, the offender used multiple methods to kill his victims:
first by a firearm and second by means of stabbing. This is one of the few South African
serial offenders who used a firearm in his series, with this only generally occurring in
30.2% of overall incidents (Horning et al., 2015; Salfati et al., 2015b).
CONCLUSION
Serial homicide is a phenomenon documented throughout the whole world (Aki, 2003;
Campobasso et al., 2009; Campos & Cusson, 2007; Egger, 2002; Gorby, 2000; Harbort &
Mokros, 2001; Jenkins, 1988; Jenkins, 1989; Kallian, Birger, & Witztum, 2004; Kallian,
Bar-El, Durst, & Witztum, 1996; Keppel & Birnes, 2009; Myers et al., 2008), yet to date, it
has been predominantly defined from a western point of view. Definitions have been formulated
to include the common features of the western presentation of serial homicide. These cases
illustrate a sample of series from across the spectrum of serial homicide as it is presented in South
Africa and illustrate how some practice-based assumptions made by investigators can, and must,
be validated and improved by practice-focused research. Although some features are similar to
serial homicide cases occurring outside of South Africa, other features are unique to South Africa.
Other behaviours take on a different meaning in the South African context than they may have
in Western contexts, highlighting the need for caution about generalising statements about
serial homicide, this being especially relevant for offender profiling and linkage analysis.
Information relating to the issues discussed earlier is scant in local and international
research; therefore, the frequency of these issues to date has been speculative in nature. It is
these investigator or practitioner assumptions that have drawn the criticism of academics
Serial homicide in South Africa: practice to research 15
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
and researcher. In an attempt to deal with these empirical shortcomings, the studies reflected in
this special issue of the Journal of Investigative Psychology and Offender Profiling (Salfati,
Labuschagne, et al., 2015b; Salfati, Horning, et al., 2015a; Horning et al., 2015; Sorochinski
et al., 2015) and quoted earlier were performed as a response to practice-led questions and, in
turn, have provided the first generation of studies that have aimed to answer some of these
questions and support or refute these assumptions, and provide a basis for a more evidence-
led investigation. In turn, this has also provided a comparison between experience-informed
and empirically informed literature regarding the aetiology of serial homicide in South Africa.
It is through these initiatives that a growing, practice-relevant, body of knowledge gleaned
from practice and research is being developed, contributing to the global view of serial homicide.
REFERENCES
Abrahams, N., Martin, L. J., Jewkes, R., Mathews, S., Vetten, L., & Lombard, C. (2008). The epidemiology and
the pathology of suspected rape homicide in South Africa. Forensic Science International,178, 132–138.
Aki, K. (2003). Serial killers: A cross-cultural study between Japan and the United States. Unpublished
master’s thesis, California State University, Fresno, California.
Barkhuizen, J. (2006). An exploration of the intrapsychic development and personality structure of
serial killers through the use of psychometric testing. Unpublished master’s thesis, University of
Pretoria, Pretoria, South Africa.
Campobasso, C. P., Colonna, M. F., Carabellese, F., Grattagliano, I., Candelli, C., Morton, R. J., &
Cantanesi, R. (2009). A serial killer of elderly women: Analysis of a multi-victim homicide investiga-
tion. Forensic Science International,185(1–3), 7–11.
Campos, E., & Cusson, M. (2007). Serial killers and sexual murderers. In J. Proulx, E. Beauregard,
M. Cusson, & A. Nicole (Eds), Sexual Murderers. A Comparative Analysis and New Perspectives
(pp. 99–105). Chichester: John Wiley & Sons, Ltd.
De Wet, J. A. (2005). A psychological perspective on the personality development of a serial
murderer. Unpublished master’s thesis, University of Pretoria, Pretoria, South Africa.
Del Fabbro, G. A. (2006a). A family systems understanding of serial murder. Unpublished PhD
thesis, University of Pretoria, Pretoria, South Africa.
Del Fabbro, G. A. (2006b). A review of South African perspectives on serial murder. Unpublished
master’s thesis, University of Pretoria, Pretoria, South Africa.
Du Plessis, J. J. (1998). Towards a psychological understanding of serial murder. Unpublished
master’s thesis, University of Pretoria, Pretoria, South Africa.
Egger, S. A. (2002). The Killers Among Us. An Examination of Serial Murder and Its Investigation.
2
nd
Ed. New Jersey: Prentice Hall.
Gorby, B. (2000). Serial murder: A cross-national descriptive study. Unpublished master’s thesis,
California State University, Fresno, California.
Harbort, S., & Mokros, A. (2001). Serial murderers in Germany from 1945 to 1995: A descriptive
study. Homicide Studies,5(4), 311–334.
Hickey, E. W. (2006). Serial Murderers and Their Victims 4
th
Ed. Belmont, CA: Thomson Wadsworth.
Hodgskiss, B. (2001). A multivariate model of the offence behaviours of South African serial
murderers. Unpublished master’s thesis, Rhodes University, Grahamstown, South Africa.
Hodgskiss, B. (2004). Lessons from serial murder in South Africa. Journal of Investigative Psychology
and Offender Profiling,1,67–94.
Hook, D. (2003). Reading Geldenhuys: Constructing and deconstructing the Norwood killer. South
African Journal of Psychology,33(1), 1–10.
Horning, A., Salfati, C. G., & Labuschagne, G. N. (2015). South African serial homicide: A victim-
focused behavioural typology. Journal of Investigative Psychology and Offender Profiling. DOI:
10.1002/jip.1426.
Hurst, A. (2003a). Killer in our midst: Part one. An analysis of court transcripts pertaining to the defence of
Stewart Wilken in Die Staat Teen Stewart Wilken.South African Journal of Philosophy,22(4), 289–305.
Hurst, A. (2003b). Killer in our midst: Part two. An analysis of court transcripts pertaining to the defence of
Stewart Wilken in Die Staat Teen Stewart Wilken.South African Journal of Philosophy,22(4), 306–326.
16 G. N. Labuschagne and C. Gabrielle Salfati
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
Jenkins, P. (1988). Serial murder in England 1940–1985. Journal of Criminal Justice,16,1–15.
Jenkins, P. (1989). Serial murder in the United States 1900–1940: A historical perspective. Journal of
Criminal Justice,17, 377–392.
Kallian, M., Birger, M., & Witztum, E. (2004). Reassessing “Jacob’s case”: A serial killer re-examined.
Medicine and Law,23(1), 59–71.
Kallian, M., Bar-El, Y. C., Durst, R., & Witztum, E. (1996). Jacob - The case of a serial killer. Israel
Journal of Psychiatry and Related Sciences,33(4), 221–227.
Keppel, R. D., & Birnes, W. J. (2009). Serial Violence. Analysis of Modus Operandi and Signature
Characteristics of Killers. Boca Raton, FL: CRC Press.
Knight, Z. G. (2006). Some thoughts on the psychological roots of the behaviour of serial killers as
narcissists: An object relations perspective. Social Behavior and Personality,34(10), 1189–1206.
Labuschagne, G. N. (1998). Serial murder: An interactional analysis. Unpublished master’s thesis,
University of Pretoria, Pretoria, South Africa.
Labuschagne, G. N. (2001). Serial murder revisited: A psychological exploration of two South African
cases. Unpublished PhD thesis, University of Pretoria, Pretoria, South Africa.
Labuschagne, G. N. (2006). The use of a linkage analysis as evidence in the conviction of the Newcastle
serial murderer, South Africa. Journal of Investigative Psychology and Offender Profiling,3,183–191.
Labuschagne, G. N. (2007). Foreign object insertion in sexual homicide cases: An exploratory study.
Unpublished master’s thesis, University of Pretoria, Pretoria, South Africa.
Labuschagne, G. N. (2010). The use of linkage analysis as an investigative tool and evidential
material in serial offences. In K. Borgeson, & K. Kuehnle (Eds.), Serial Offenders in Theory
and Practice pp. 187–215. Sudbury, MA: Jones & Bartlett Press.
Lemmer, C. (2003). A comparative study between South African serial killers and their American
counterparts. Unpublished master’s thesis, Rand Afrikaans University, Johannesburg, South Africa.
Myers, W. C., Justen, E., Morton, R. J., Tilley, J., Adams, K., Vandagriff, V. L., & Hazelwood, R. R.
(2008). The relationship between serial sexual murder and autoerotic asphyxiation. Forensic
Science International,176(2–3), 187–195.
Pakhomou, S. M. (2004). Serialkillers: Offender’s relationship to the victim and selected demographics.
International Journal of Police Science and Management,6(4), 219–233.
Pistorius, M. (1996). A psychoanalytic approach to serial killers. Unpublished DPhil thesis, University of
Pretoria, Pretoria, South Africa.
Proulx, J., Beauregard, É., Cusson, M., & Nicole, A. (2007).Sexual Murderers. A Comparative Analysis
and New Perspectives. Chichester: John Wiley & Sons.
Rossmo, D. K. (2000). Geographic Profiling. Boca Raton, FL: CRC Press.
Salfati, C. G. (Ed.). (2005). The Homicide Profiling Index (HPI): A tool for measurements of crime
scene behaviors, victim characteristics, and offender characteristics. In C. G. Salfati (Ed.),
Homicide research: Past, present and future: Proceedings of the 2005 meeting of the Homicide
Research Working Group (pp. 152–159). Chicago, IL: Homicide Research Working Group.
Salfati, C. G. (2007). Homicide Profiling Index (HPI): Crime scene and background variables coding
dictionary version 4. Unpublished manuscript.
Salfati, C. G. (2008). Offender profiling: Psychological and methodological issues of testing for be-
havioural consistency. Issues in Forensic Psychology,8,68–81.
Salfati, C. G., Horning, A., Sorochinski, M., & Labuschagne, G. N. (2015a). South African serial ho-
micide: Consistency in victim types and crime scene actions across series. Journal of Investigative
Psychology and Offender Profiling. DOI: 10.1002/jip.1428.
Salfati, C. G., Labuschagne, G. N., Horning, A., Sorochinski, M., & De Wet, J. (2015b). South African
serial homicide: Offender and victim demographics and crime scene actions. Journal of Investigative
Psychology and Offender Profiling. DOI: 10.1002/jip.1425.
Schwikkard, P. J. (2009). Confessions in criminal trials. In P. J. Schwikkard, S. E. Van Der Merwe,
D. W. Collier, W. L. De Vos, & E. Van Der Berg (Eds.), Principles of Evidence 3
rd
Ed (pp. 333–358).
Wetton, South Africa: JUTA & Co.
Snyman, H. F. (1992). Serial murder. Acta Criminologica,5(2), 35–41.
Sorochinski, M., Salfati, C. G., & Labuschagne, G. N. (2015). Classification ofplanning and violent be-
haviours in serial homicide: A cross-national comparison between South Africa and the US. Journal
of Investigative Psychology and Offender Profiling. DOI: 10.1002/jip.1427.
Woodhams, J., & Labuschagne, G. N. (2012). Serial rapists: The offenders, their victims and their
offences. Sexual Abuse: A Journal of Research and Treatment. DOI: 10.1177/1079063212438921.
Serial homicide in South Africa: practice to research 17
Copyright © 2014 John Wiley & Sons, Ltd. J. Investig. Psych. Offender Profil. 12:4–17 (2015)
DOI: 10.1002/jip
