A new and innovative approach that uses human-scent evidence to identify bomb
makers and arsonists is presented. The process of identifying and locating a suspect
after an explosion or fire is often complicated by the fact that improvised explosive
and incendiary devices generally employ a time-delay mechanism to allow evasion
well before their functioning. One approach that uses specially trained bloodhound-handler
teams as an investigative tool to identify people who had contact with the devices
has been developed. Pipe bombs containing explosives with varying explosion velocities
were functioned, and metal and plastic gas cans were burned with gasoline. The
purpose of this feasibility study was to demonstrate the survivability of human
scent after being exposed to extreme mechanical and thermal effects from the explosion
and burning of various energetic materials and the potential for use in criminal
The ability of bloodhounds to effectively match collected scent
to the correct person and follow that person through and across
numerous environments to an effective conclusion is accepted in
most courts and has been validated in a recent scientific study
(Harvey and Harvey 2003). The authors identified no published studies
that explore the durability of human scent.
In traditional bloodhound circles, the anecdotal information passed
from trainer to student is that human scent is fragile and easily
destroyed. Many dog handlers in the United States are taught that
identifiable human scent disappears after 24 hours. European studies
using properly trained scent-identification dogs showed acceptable
performance levels with collected scent that was aged two weeks
to six months (Schoon and Haak 2002).
This paper will demonstrate that human scent is durable and will
remain identifiable after being exposed to extreme mechanical and
thermal effects associated with the reactions of various energetic
materials. This feasibility study does not address the durability
of human scent as a function of physical parameters, such as surface
temperature of materials. The effects encountered in the detonation
of improvised explosive devices and the deflagration of improvised
incendiary devices are not consistently reproducible due to the
myriad of variables that can affect energetic material performance
in an improvised device. In addition, conditions encountered by
bloodhound handler teams during crime scene responses are never
identical. Therefore, the test design discussed below reflects conditions
encountered in an urban setting and examines the potential to use
this technique in a criminal investigation.
Generally, there are two methods of bloodhound-handler training
in the United
Statestraditional and specialized. The traditional method teaches
a bloodhound-handler team to have the dog search for matching scent
at the beginning of the trail by casting about. The handler then
determines the presence or absence of matching scent by "reading" the dog's behavior. Some groups have trained their bloodhounds to
return to the handler and provide an alert upon determining that
no matching scent is present in the area.
Scent collection techniques used to acquire human scent vary widely
in the traditional bloodhound community. The most routinely used
methods are direct scenting, swiping, and absorption. In the direct-scenting
method, the bloodhound handler allows the dog to sniff the actual
item of evidence. With the swiping and absorption methods, scent
is transferred onto a gauze pad instead of using the actual item
of evidence. Swiping, a direct transfer of scent onto a gauze pad,
is achieved by wiping the pad across the surface of the evidence.
Absorption, or placing the pad next to the scent article for an
extended period, relies on the gauze pad's ability to gather scent
while being in direct or indirect contact with the evidence. Although
these methods have been used for decades, the potential for negatively
affecting trace evidence is clear.
The specialized bloodhound-handler teams use a different response
system to indicate the presence or absence of matching scent at
the start of a trail. Although traditional handlers must rely on
their ability to "read" the dog's behavior, the specially
trained teams use a simplified yes or no response technique. At
the start of the trail, if matching scent is present at the location
being checked, the bloodhound trails. If no matching scent is present
at the location being checked, the bloodhound refuses to trail.
Once the bloodhound has started trailing, thus indicating the presence
of matching scent, much of the handling techniques used in the traditional
bloodhound community are relied upon.
The specially trained bloodhound-handler teams employ the Scent Transfer Unit
(STU) to collect scent pads. The STU-100 is a portable vacuum collection unit
that uses the flow of air to transport the components of human scent onto 11.25
by 22.86cm sterile surgical pads. (The pads are considered sterile for medical
usage, not to denote an absence of any chemical compounds.) The vacuum's intake
funnel supports the sterile pad to allow the evidence to be placed on or near
the pad. At full charge, the STU-100's 12-volt fan pulls approximately 400 liters
of air per minute across the surface of the evidence and through the pad, thus
trapping the scent-causing materials. Where traditional scent evidence recovery
techniques require direct scenting from the article of evidence or touching the
evidence with a gauze pad, the airflow across the scent pad allows the evidence
recovery personnel to immediately capture scent, thus minimizing the loss of other
forensic evidence. It also provides a consistent type of scent article for presentation
to the bloodhound.
When the investigators develop a suspect, the specially trained
bloodhound-handler team is brought to a location recently visited
by that person to conduct a suspect-location check. Typical locations
for scent checks include the suspect's residence or work because
these locations provide large areas of deposited scent due to the
frequent travels in and out of the buildings. Generally, case law
in the United States requires that the dog-handler team be placed
on a trail where the suspect was known or believed to have passed.
In order to fulfill these requirements, the handler is placed on
this fresh trail location and asked to introduce the previously
collected scent pad to the hound. The handler knows that he has
been placed on a known trail but is not told details of potential
outcomes, thus keeping him blind. In addition, the handler does
not know if the scent pad for presentation is a negative control
or a scent pad collected from an article of the crime.
Because these specially trained bloodhounds provide a yes or no
response, a positive response indicates to the investigator that
additional investigative efforts should be exerted to determine
the reason that the dogs matched scent from the evidence to the
location. This type of positive-scent match is most often associated
to a resident or frequent visitor to that location. Assuming that
the scent article being used contains a viable amount of scent,
a negative response during a location check provides strong evidence
to eliminate the suspect from the investigation.
Bombers and arsonists typically employ some time-delay method in
improvised devices to remove themselves safely from the scene of
the crime, and many of these events go unsolved. Using scent collected
from the devices, qualified bloodhound-handler teams can use scent
pads to conduct suspect-location checks, thus providing a new tool
to assist in the identification and capture of these individuals.
Four pipe bombs and two gas containers were used for scent articles.
Four 2.7 x 20.3cm schedule-40 steel pipes and eight end caps were
purchased wrapped in plastic. The gas containers, one metal and
one plastic, were purchased new and immediately placed into large
Twelve test subjects were selected from a local search-and-rescue
organization that had not been used in any previous training or
testing. Most target and decoy pairs chosen were the same sex and
age. In order to deposit scent, the targets handled their respective
items for approximately one to two minutes, placed the items into
resealable bags, closed, and labeled the bags. During the explosion
and collection process, the bomb technicians and scent-pad collectors
were monitored by the test planners to minimize any scent cross
contamination. To accomplish this, the technicians and collectors
were required to wear a new pair of latex gloves each time a new
device was handled. Because assembly and collection personnel could
have contributed scent even while wearing gloves, they were not
permitted to be present during the testing.
Pipe Bomb Preparation
Four pipe bombs were constructed using two low-explosive powders
and two high-explosive materials. Goex (Doyline, Louisiana) black
powder, a 6:1.2:08 mixture of potassium nitrate, charcoal, and sulfur,
and Bullseye (Alliant Powder, Radford, Virginia) double-base smokeless
powder, a combination of nitrocellulose and nitroglycerin, were
chosen for the two low-explosive filled pipe bombs because of their
availability and common use in domestic bombing incidents. In its
legitimate form, smokeless powder is used for reloading ammunition,
and black powder is typically used for a type of sport shooting.
Kinepak, (Slurry Explosive Corporation, Oklahoma City, Oklahoma)
a binary explosive consisting of a mixture of ammonium nitrate and
nitromethane, and Composition C4, a military's cyclotrimethylenetrinitramine
(RDX)-based explosive were the selected high-explosive fillers.
Figure 1. Photograph of a Black Powder Explosive
Figure 2. Photograph of a Smokeless Powder Explosive
Figure 3. Photograph of a Binary Explosive
Figure 4. Photograph of a C4 Explosive
To ensure the safe initiation of each buried pipe bomb, a detonating
cord booster was placed into the energetic material. Holes were
drilled in one of the two end caps to allow for the insertion of
a length of Dupont (Dupont-ETI, North Bay, Ontario, Canada) 70-grain
per-foot detonating cord with a pentaerythritol tetranitrate (PETN)
core. After the pipes were half filled with the explosive material,
approximately 10cm of a 61cm length of detonating cord was inserted
into each pipe with the remaining length protruding from the containment
vessel. A U.S. military nonelectric blasting cap and black powder
core time fuse were used as a time-delay system to initiate the
detonating cord boosters.
Each assembled device was placed inside a 20-liter plastic bucket
that was packed with dirt. The bucket was then suspended inside
a 189-liter steel drum and detonated. This technique was used to
recover as much postblast fragmentation as possible. Some devices
required the steel drum to be partially buried to further restrict
scattering of fragments. Fragment recovery was completed for each
device immediately after detonation, and before the next device
was exploded. A screen sifter and magnet were used to locate and
collect the smaller pieces. The recovered fragments were placed
inside polyethylene resealable bags.
The maximum approximate reaction product temperature of black powder and double-base
smokeless powders are 2380K and 2200 to 3800K, respectively (Picatinny Arsenal
1962). The reaction product temperature in detonating explosives can exceed 5000K
(Persson et al. 1993). The ignition source temperature of gasoline is 1083K (Henderson
and Lightsey 1984), with a much higher flame temperature that is dependent on
oxygen content. These temperatures are not the surface temperatures of the containers
but are provided to demonstrate the range of temperatures that can occur in the
reaction zone of various energetic materials.
Arson Device Preparation
Two gasoline containers, one plastic and one metal, were placed
on the ground and covered with one half liter of gasoline. The gasoline
was ignited and allowed to burn for two minutes. The fire was then
extinguished with water. After cooling, the remains were placed
in separate paper bags.
Twenty professional and novice bloodhound-handler teams were used
for this study including 16 handlers and 20 dogs. The bloodhound-handler
teams using two dogs in this test were designated with the same
numeric identifier but with a different alpha identifier (i.e.,
4 and 4a).
Thirteen of the handlers were specially trained. The remaining three
(10, 11, 14) were traditionally trained. Five handlers were full-time
law enforcement dog handlers, three were reserve officers, and eight
were civilians. The handlers' experience levels ranged from 700
cases worked to no field-case experience. The dogs ranged in age
from under one year to seven years old. Twelve teams had previously
trained on arson debris. Three teams had previously trained on bomb
Follow this link to
Scent pads were collected from the pipe-bomb debris by placing
the fragments onto the STU-100 and running the machine for approximately
30 seconds. Scent pads were collected from the arson debris by placing
the STU-100 intake funnel inside each bag and running the machine
for approximately 30 to 90 seconds. The decision to place the evidence
article directly on the scent pad or to place the intake funnel
inside the evidence container was made specifically for ease of
processing. No empirical data has been gathered to determine if
either collection method is superior to the other.
In order to eliminate cross contamination, the STU-100 intake was
cleaned according to manufacturer's recommendation by using isopropanol
swabs. The STU-100 was allowed to dry prior to collecting scent
pads from each of the six devices. To minimize preparation time,
most of the scent pads were split evenly with scissors into two
sections after scent collection, thus creating two pads from one.
Before cutting the pads collected from a new device, the scissors
were cleaned with isopropanol swabs and allowed to dry. Half of
the pads were collected with the STU-100 the same day the devices
were functioned. The remaining halves were collected two weeks later.
All of the pads were packaged in polyethylene, resealable bags and
maintained at room temperature. The scent pads were randomly aged
two days and 16 days before being presented to the dogs.
The trails were run in an urban public park frequented by joggers
and people walking pets. Because it was previously demonstrated
that bloodhounds are capable of identifying human scent that was
vacuumed onto a scent pad (Harvey and Harvey 2003) and capable of
matching that scent to a suspect on aged and contaminated trails
through an urban environment (Harvey and Harvey 2003), no attempts
were made to age or contaminate the test trails. For the test trails,
12 people walked a split trail; six scent targets and six decoys.
On each trail, two people (target and decoy) were started at the
same point and walked the first section of the trail side by side.
After approximately 14 meters, they split at a 45-degree angle and
continued another 18-27 meters to their respective hiding locations.
The bloodhound-handler teams were not permitted to see the trails
being laid, and the target and decoy were hidden at the end of their
Six stations were set throughout the park, one for each device.
At each station, the target and decoy laid a new trail in a different
location for each bloodhound-handler team so that no team ran the
same trail. Each bloodhound-handler team completed one trail at
each station. For each starting location, the teams were placed
directly on the target and decoy trail. After placing the dogs in
harnesses, the handlers were given an arbitrarily selected scent
pad for presentation to their dog. The parameters recorded were
- Did the dogs begin to trail?
- Did the dogs identify the target person?
The following outcomes were recorded. For beginning to trail, a
YES was recorded if the dog indicated the presence of matching scent
at the start of the trail and began to follow. A NO was recorded
if the dog gave no response to the presence of matching scent at
the start of the trail. (Table 2)
For the identification of the target person, a YES was recorded
if the dog trailed to and alerted on the person at the end of the
trail. If the dog trailed to the decoy person and gave a positive
identification, a false positive (FALSE) was recorded. If the dog
trailed but did not identify either the target person or the decoy,
a negative identification was noted. Because the teams were given
two minutes to complete their test, the positive or negative identification
results only reflect the immediate response given by the dog at
the end of the trail. The trails were monitored and recorded by
people without knowledge of the correct outcome.
The test design replicated common crime scene practice; therefore,
no negative-control pads were introduced. In actual casework, the
bloodhound-handler team is placed on a known trail and given scent
collected from an instrument of the crime. If an identifiable amount
of scent is present on the scent pad and the bloodhound finds matching
scent at the start of the trail, the bloodhound-handler team follows
the trail to its logical conclusion.
Follow this link to
The overall percentage for a positive beginning to trail indication
was 78.3 percent. Beginning to trail indications were calculated
by using the scores of the bloodhound-handler teams that indicated
positive at the start of the trail. A no-response indication at
the start of the trail did not necessarily signify that there was
no matching scent present on the pad or at the trail beginning.
This negative alert may also indicate that the dog was not able
to detect such low levels of material. (Table 2)
The overall combined score for positive identifications was 70
percent. The score for dogs that indicated matching scent by beginning
to trail and correctly identifying the target person was 88.6 percent
with no false-positive indications. (Table 2)
Eight dog-handler teams with a casework experience level under
five conducted 48 trails with 34 positive begin-to-trail indications
(70.8 percent) and 31 positive identifications (64.5 percent). (Table
Five dog-handler teams with a casework experience level more than
five and fewer than 100 conducted 30 trails with 24 positive begin-to-trail
indications (80 percent) and 19 positive identifications (63.3 percent).
Seven dog-handler teams with a casework experience level over 100
conducted 42 trails with 36 positive begin-to-trail indications
(85.7 percent) and 34 positive identifications (80.9 percent).
Follow this link to
Several aspects of this study must be considered when factoring
the significance of the findings. Overall, the dogs correctly identified
the target person in 53 of the 80 bomb-debris experiments and 31
of the 40 arson-debris experiments with no false-positive identifications.
The combined results and the absence of false-positive identifications
supports the general reliability of this procedure and indicates
that dogs can detect and identify human scent on bomb and arson
Some positive identification could have occurred because a dog
alerted on the first visual cue that it received. In the training
of various teams, trails are set up so there is only one choice
at the end of the trailto identify the scent target. Teams that
have trained with multiple decoys on the scent trail typically have
dogs conditioned to check each person for a scent match. The specially
trained teams in this study use multiple decoys in training. It
is unknown whether the three traditionally trained teams use similar
The number of target identifications may have increased if there
had been no time limit for the completion of each trail. The two-minute
limit did not provide enough time for some of the bloodhounds to
make a choice. This time limit may have also had a beneficial effect
because the handlers did not necessarily have the time to entice
their dogs to choose one target over another in order to complete
the trail with a find, thus causing false-positive identifications.
No readily identifiable differences were observed that indicated
the scent pads collected on the day of device functioning produced
better results than the pads collected 14 days after the event.
The following tables specify the scent-pad collection information
and the test results for each pad.
Follow this link to
Follow this link to
Follow this link to
Follow this link to
Follow this link to
Follow this link to
Follow this link to
The explosion and burning of the test materials in this series
was conducted to maximize the chances of recovery. There are too
many uncontrollable environmental parameters associated with the
explosion and burning of materials to reliably replicate events
associated with an actual crime scene. In each of these events the
materials would have been handled differently and subjected to scattering,
weather, and the influence of the actions of emergency personnel.
Likewise, it is impossible in an experimental test scenario to control
all of the environmental variables to accurately replicate trailing
conditions experienced in casework. Consequently, the results derived
from this type of feasibility test series only demonstrate the survivability
of identifiable human scent and the potential to use it in an investigation.
It does not indicate the ability of a particular breed, nor will
it provide sufficient data to predict a scent dog's reliability
in casework or testing.
Caution must be applied when dealing with human-scent evidence.
Because scent is easily transferred, a positive trail or identification
resulting from any scent article only shows a scent relationship
to the scent article and must be verified and corroborated through
other investigative means (Stockham et al. 2004). This scent relationship
generally establishes a direct or indirect link between a person
and an article of the crime; it does not prove complicity.
This feasibility study demonstrated the ability of human scent
to survive the extreme mechanical and thermal affects associated
with the explosion and burning of various energetic materials. Furthermore,
the ability of specially trained bloodhound-handler teams to match
the collected scent to the correct person after these violent energetic
events was demonstrated. By conducting suspect elimination checks
with scent pads collected by the STU-100, a portable vacuum collection
unit, this specialized approach has shown that it can assist in
providing valuable lead information for investigators, focus valuable
and often limited resources, and aid in the solution of crimes.
Harvey, L. and Harvey, J. Reliability of bloodhounds in criminal
investigations, Journal of Forensic Sciences (2003) 48(4):811-816.
Henderson, R. W. and Lightsey, G. W. Effective flame temperatures of flammable
liquids, Fire and Arson Investigator (1984) 35(12):8.
Persson, P., Holmberg R., and Lee, J. Equation of state of the
explosion products. In: Rock Blasting and Explosives Engineering,
CRC, Boca Raton, Florida, 1993, p. 102.
Picatinny Arsenal, Encyclopedia of Explosives and Related Items,
PATR 2700, Volume 2, Dover, New Jersey, 1962, B170, C34-35.
Schoon, A. and Haak, R. Stability of the odor left on an object. In: K9
Suspect Discrimination, Training and Practicing Scent Identification Line-ups,
Detselig Enterprices, Calgary, Alberta, Canada, 2002, pp. 47-48.
Stockham, R. A., Slavin, D. L., and Kift, W. Specialized use of
human scent in criminal investigations, Forensic Science Communications
[Online]. (July 2004). Available: www.fbi.gov/hq/lab/fsc/backissu/july2004/research/2004_03_research03.htm.