Individual human scent as a forensic identifier using mantrailing

Abstract

Specially trained dogs have long been used by law enforcement agencies to help in criminal investigations and in searching for missing persons. Still, it is unclear which components of human scent released into the environment contribute to successful searches of individuals. In this study, saliva and axillary sweat samples were taken from a total of 190 people. Additionally, DNA was extracted from whole blood of seven different people and used as an odour sample as well. Overall 675 tests (trails) were performed during a period of 18 months. The ability to track individuals with the odour samples mentioned above was examined with seven dogs, four of which were specially-trained dogs (mantrailer) from the Saxony Police. Results indicated that specially-trained police dogs can track a person with an average success rate of 82% and correctly identify the absence of an odour track with an average success rate of 97% under various conditions. Private rescue dogs were less successful with an average success rate of 65% and 75% respectively. These data suggest that the potential error rate of a well-trained handler team is low and can be a useful tool for law enforcement personnel. Saliva, as a reference odour source, was found to be particularly suitable for the search. The results of the study suggest that the components contained in axillary sweat, saliva and DNA extracted from whole blood are sufficient, serving as a key stimulus for individualized searches.

Keywords

Human scent evidence
Reliability mantrailing
Mantrailer
Forensic science
DNA

1. Introduction

In police investigations, it is often necessary to prove the previous presence of a of a known suspect at the crime scene, or to find evidence for this. Many suspects deny ever having been at the crime scene. This is particularly true in cases of severe criminal offenses, such as theft, gang violence, arson, or homicide. Furthermore, in some cases physical evidence, such as dactyloscopic traces or DNA traces, was not yet found nor was available. In such cases, the police (e.g. in Germany or USA) use special trained dogs — mantrailer. Mantrailing is the search for and the tracking of (individual) human scent by using these specially-trained dogs. The mantrailing canine does not follow mandatorily precise the footstep-to-footstep track of the individual. This distinguishes a trailing dog from a tracking dog [1].

Rather, it is of interest whether the dog actually traces an odour trail of the absent person exclusively, which corresponds to the reference odour presented on a so-called scent article. If there is no matching odour, the mantrailer is expected to refuse to trail. The same is expected if a scent is presented from an individual who is known to have never been there. This aforementioned procedure or use a blank scent article is a so called “negative check” or “negative”. In this “negative” case, the canine will typically try to follow a trail in close vicinity of the starting point. Performing the negative correctly requires the canine not to trail and show his handler a corresponding individual behaviour. Examples for individual behaviours are, that one dog may jump at his handler while another dog moves only a few steps around the starting point, yet another remains in place and some check a larger area and stop or return to the starting point. Finally, only the dog handler can interpret the dog’s behaviour correctly and is able to testify the results in court. Thus, the district courts of Dresden, Chemnitz (both in Saxony) and Lüneburg (in Lower Saxony) have already accepted the evidential value of Mantrailing evidence provided by the mantrailers of Saxony Police Force [2–4].

In principle, a trailing canine can only find and follow a trail from the starting point if a searched person was there and has moved away from this location in a manner which allows an exchange of the released human scent trace with the environment [5–9]. This is in accordance with the Locard’s exchange principle, which states that a perpetrator cannot leave a crime scene without leaving numerous traces [10].

From experience in the field, it is known that various body fluids, such as sweat, saliva or blood are suitable as scent articles for trailing. Previously sterile gauze pads are used to collect perpetrators’ individual scent. The suspect is required to open the sterile package and to wipe himself, e.g. over the forehead or the neck. Subsequently, this gauze pad is sealed by the suspect in an airtight plastic bag, which in turn is stowed in a special jar. After that the dog is confronted with this scent article at the crime scene. Otherwise, the collection sets for DNA samples are especially practical for police work. They include cotton swabs for a swab of the buccal mucosa. Considering this, the suitability of saliva as scent article is also of interest, since making a swab of the buccal mucosa to get a suitable scent article can be carried out with comparatively little effort.

Studies have already indicated the ability of trained dogs (Canis familiaris) to differentiate humans by their individual scent [11–19]. However, most of these studies do not focus on the pursuit of an odour trail. There are only few studies dealing with the question, whether the human individual scent is traceable in real life scenarios by canines [20–23]. For the use in court, it is therefore of great importance whether the results of police operations are sufficiently reliable in terms of evidence according to legal standards. The Daubert standard provides a rule of evidence regarding on the admissibility of expert testimony during federal legal proceedings in the United States [24]. These criteria include, whether the method in question has been scientifically tested, whether it has been subjected to peer review and publication, its known or potential error rates, and finally, whether it is accepted within a relevant scientific community [25,26]. There are no comparable regulations in Germany. The court is free to choose what to consider as evidence. However, the district court of Nuremberg has set up principles for the applicability of the results of the use of mantrailers as sole evidence in a pioneering judgment. Firstly, the dogs must have the relevant certification level of the police. Secondly, the used scent article must be clearly assignable to a concrete person, therefore, only swabs directly from the body of the person concerned should be used. In addition, the acquisition of the scent article must be documented in a protocol. Thirdly, in each case, two dogs must search for the same trail independently of each other and without the involvement of the other dog handler. Fourthly, each assignment has to be filmed completely, in order to allow a posteriori analysis by the court and an expert [27].

However, this alone does not state the suitability of this investigation tool and the reliability of the results in principle. This must be validated in a variety of tests. Therefore, this study was designed to investigate how reliable the results of mantrailing are under real police operating conditions. This has only been shown twice, so far. Once, each with five attempts of eight dogs. The terrain for each individual trail consisted of a combination of grass, asphalt, cement or dirt [20]. The other time with a large number of attempts. However, thereby the traces were solely laid on grass [28]. Moreover, this study tested for the first time whether saliva as well as DNA samples provide sufficient cues to trail an individual.

2. Methods

2.1. Animals and dog handlers

The study was conducted over a period of 18 months, starting in November 2014 and included seven dogs of different breeds (Table 1). Four dogs are mantrailing police dogs from the Saxony Police and trained according to the method described in a book written by Armin Schweda [29]. Three dogs are private rescue dogs, two of which were also trained according to the Schweda method. The dogs are either certified by the police or a rescue organisation. Those canine team certifications include a scent recognition assessment that evaluate the ability of the canine to detect and use a specific person’s scent on a scent article to follow the matching scent trail to this specific person while discriminating from nonmatching scent trails. This trail is aged at least 24 h and approximately 2000–2500 m in length [30,31]. The dogs used in the experiments were trained solely for human scent trailing. All canines, except for the bloodhound/Hanover Hound mongrel, were handled by their handlers and work for the police or were used by rescue organisations. Among the dog handlers were five police officers and two civilian volunteers affiliated with rescue organisations. All handlers have been working with the dogs for at least one year. Everyone had already been deployed to search for missing persons before. Table 1 highlights the variation present among the participating teams which can be divided into eight broad categories, including: breed, sex, institution, canine age, training method, number of years of specific training and certification.

2.2. Subjects

The participation was voluntary, and the subjects were informed about the aims of the study. The appropriate institutional boards, ethics boards and research boards have approved this study (German Institutional Research Board: The Ethics Committee of the University of Leipzig — Faculty of Medicine, reference number 402-14-15122014). Samples were collected from 190 individuals, ages 7–86. Among them were 134 male and 56 female participants (Table 3). Since the study was supposed to imitate real operating conditions, the subjects did not need to use any hygiene articles for the scent collection beforehand. However, the participants were interviewed whether they had consumed body odour-influencing foods (e.g. garlic or onion) before the experiments started.

2.3. Scent collection

Previous studies as well as own experience have shown that various items are suitable for scent collection [17,32–35]. In this study, two different materials were used — customary medical gauze swabs and specially made foam blanks. These were either prepared with axillary sweat, saliva, DNA extracted from whole blood, or enzymatically decomposed DNA (as described below) and then used as scent articles.

2.3.1. Armpit sampling

A MaiMed 10 cm × 10 cm gauze swab, 8-ply-sterile, made of surgical gauze type was used to collect scent. The subjects were required to wear a new pair of latex gloves, then open the sterile package and clamp the contained gauze pads for 10 min under their left or right armpit. Subsequently, this gauze pad was sealed by the subject in an airtight 250 × 350 mm — 50 μ ziplock bag and stowed in another airtight 250 × 350 mm — 50 μ ziplock bag. The ziplock bag was provided by the experimenter and marked with a number in lieu of names in order to carry out a double-blind study, i.e., that neither the dog handler nor the accompanying persons (escort, camera operator) nor the runners were aware of the target scent article (A, B or a negative) presented to the dog. The experiments were performed the same day about 30 min on average after the sampling. Overall the experiments on one day takes an average time of about 4 h and a maximum of max. 7.5 h, depending on the number of dogs participating and the location of the trial.

Additionally, specially made foam blanks, PANAPOR filter foam PPI 60T/PUR-ether foam, RG ca.25 kg/m³, 11 mm diameter × 40 mm, PANA Foamtec GmbH, were used to collect scent. The experimenter handed the subjects airtight ziplock bags — 50 μ, 250 × 350 mm with foam blanks. The subjects were required to wear a new pair of latex gloves, then open the ziplock bag and clamp the contained foam blanks for 10 min under the left and/or right armpit. Then, the foam blanks were sealed by the subject. The stowing procedure thereafter that was the same as the one with the gauze swabs. The experiments were performed the same day after the sampling.

2.3.2. Saliva sampling

The same PANAPOR filter foam blanks used for the armpit sampling were used for saliva scent sampling. The experimenter handed the subjects an airtight ziplock bag — 50 μ, 250 × 350 mm with a number of foam blanks and ziplock bags. The subjects were required to wear a new pair of latex gloves, then open the ziplock bag and take the foam blanks in their mouth and moisten them with saliva afterwards for about five seconds. Thereafter, each moistened foam blank was sealed by the subject in an airtight 250 × 350 mm — 50 μ ziplock bag and then put in another airtight 250 × 350 mm — 50 μ ziplock bag. This ziplock bag was provided by the experimenter with a number in lieu of names in order to carry out a double-blind study. The experiments were performed the same day on average about 30 min after the sampling. Overall the experiments on one day takes an average time of about 4 h and a maximum of max. 5 h, depending on the number of dogs participating and the location of the trial.

The number of gauze pads or foam blanks that were provided varied, pending on the number of dogs participating in the experiment on each of the days.

2.3.3. DNA sampling

From seven subjects blood samples of 100 ml were taken by a physician through venepuncture. The DNA was isolated by means of isopropanol precipitation [36] in the Institute for Legal Medicine of the University of Leipzig, Department of Forensic Molecular Genetics. Thus, 1000 μl of buffer solutions were prepared for each sample with a DNA content between 0.7245 ng/μl and 1.386 ng/μl determined using real-time polymerase chain reaction PCR with the PowerQuant System (Promega Corporation, Madison, WI) [37]. Nevertheless, due to the method, it is not excluded that some proteins, polysaccharides and RNA precipitate simultaneously with the DNA and were therefore still present. In each case 100 μl of the DNA sample was dripped onto foam blanks under sterile conditions and dried afterwards. In order to check the essential influence of the DNA as a key stimulus for the dogs, corresponding negative samples were prepared. As a negative sample, DNA was extracted from whole-blood samples as described above. Then, the DNA precipitate was finally removed from the buffer solution and the absence of DNA in the samples was detected by means of real-time PCR with the PowerQuant System. The prepared foam blanks were stored in air-tight sealed ziplock bags at 5 °C. The experiments were performed up to a maximum of six weeks after sampling.

2.4. Experimental design

Tests were conducted in twelve different villages, towns and cities throughout Saxony under operational conditions. Overall, there were 49 urban test locations. Each test was performed at a T-junction, i.e. the urban development had been selected so that the dog could only move into three directions: to the left or to the right or backwards (Fig. 1). At the respective T-junction, at least two to five runs were carried out in succession with different trail layers (runners) for each dog. No dog carried out a search at the same location one of the participating dogs had already used on the same day. One location was used only once for one dog a day. Runner A and Runner B walked a split trail for the experiments absolutely simultaneously to avoid minute differences in “freshness” of the trail, that could be a cue for dogs and were sent along the pavement either 100 m to the left or 100 m to the right, waiting there behind a wall or in a doorway. This was to avoid a visual incentive for the dog. The test area was not closed off so pedestrians walking by and traffic occurred.



[list=links-for-figure]
[*]Download high-res image (654KB)
[*]Download full-size image
[/list]

Fig. 1. Experimental design of a T-junction. Model overview of a T-junction with dog’s starting area, decision-making area and position of the target persons. During each test trail layer A and B was asked to walk along the pavement either 100 m to the left or 100 m to the right and to hide. The dog was presented with a randomly chosen scent of one of the two persons in the pair or of a negative trail and asked to differentiate. Illustration using material from ClkerFreeVectorImages, provided on Pixabay https://pixabay.com/, licensed under Creative-Commons-Lizenz CC0 1.0 Universal https://creativecommons.org/publicdomain/zero/1.0/legalcode.

At each run, the dog handler randomly chose one out of three identically looking ziplock bags, containing either the scent article of runner A or runner B, or the scent article of an absent person (negative). The plastic bags were marked arbitrarily with a coloured dot (red, green, blue), so they could be distinguished by the experimenter. In order to mimic real-live police scenarios, there was no prior performance test for any of the dogs. This was different than, for example, what was conducted by Schoon, in relation to the interest of the dog for the odour of the target person [38] or what Harvey et al. described [20,28], in relation to the correct identification of a negative. Each experiment was taken into account. Five minutes after the trail was laid, the search was performed. The start was carried out in a starting area in front of or in the T-junction, in a right angle to the runner’s running direction (Fig. 1). To begin the experiment, the handler and the dog were taken to the starting corridor. The handler presented the scent article, which was randomly chosen, to the dog, that was then given the command to trail. During the testing, the handlers were asked to judge whether their dog presented them with a negative or positive trail. The result of a trail was noted after the dog handler had stated that their dog displays a negative or after the dog had left the decision area on its way to the runner. The decision-making area extended from the T-junction 50 m to the left and 50 m to the right. After leaving the decision area, the dog handler was informed via walkie-talkie whether the decision of their dog was correct or wrong. Indicating the target person was not an evaluation criterion, so the dogs were not expected to complete the trail to the runner. Nevertheless, if the decision with a target present was stated as correct all dogs made a correct tag of the target person in 100% of the cases. Since the police dogs were still used in criminal proceedings during the time period when the experiments took place, negative effects of field operations should be avoided. This means that regardless of whether the test was considered wrong, the handler was allowed to finish the correct trail, so the dog could find and tagging the target person for motivation and training purposes. If the dog was successful in finding and tagging the target person, it received a reward, for example, it received a treat like cheese spread or liver sausage or was being pet.

A time limit used to make the decision was not set because the tests should simulate real operational scenarios. After a short break, another round followed with two other runners: C and D, E and F, etc.
Test series 1 was performed with axillary odour on gauze pads as scent articles, test series 2 with axillary odour on foam blanks as scent articles, test series 3 with saliva on foam blanks as scent articles and test series 4 with DNA on foam blanks as scent articles or saliva on foam blanks as scent articles. The dog handlers were not informed about the DNA samples; they have still assumed the use of saliva samples. All tests were performed blindly, meaning that neither the dog handler nor the accompanying people (escort, camera operator) nor the runners were aware of the target scent article (A, B or a negative). The dogs performing the trails were videotaped and tracked by Global Positioning System (GPS). For this purpose, a Garmin Astro 320 handheld GPS device and a DC 50 dog tracking collar (Garmin International Inc., Olathe, KS) were used.

2.4.1. Weather

The weather data, including the wind direction and speed, were recorded on site with the weather station Davis Vantage Pro 2 6152EU that combines rain collector, temperature and humidity sensors and anemometer (Davis Instruments, Vernon Hills, IL). Since the tests were supposed to simulate real operations, a special alignment of the T-junction depending on the daily wind direction was not implemented. The weather conditions varied throughout the 57 test days. The temperature readings were between −4,9 °C and 36.8 °C. On average, the wind speed was between 1.6 km/h and 16 km/h, with some winds gusting up to 33.8 km/h. The humidity reached values between 16% and 90% and the air pressure was between 933 to 1033 hPa. The effect of weather on the overall performance of the dogs is subject for further examination.

2.5. Statistical analysis

The following constellations were used for analysis.

Test series 1–4

“CORRECT” — The dog leaves the decision area in direction of the target person from whom the presented scent article originates — a true positive.

“CORRECT” — The dog shows a negative indication within the decision area in case of a presented scent article from an absent person — a true negative. “CORRECT” is also stated, if the dog shows a negative indication within the decision area in case of a presented scent article from blood extracts without DNA.

“WRONG” — The dog leaves the decision area in the direction of the person from whom the presented scent article does not originate — a false positive.

“WRONG” — The dog does not show a negative indication within the decision area in case of a presented scent article from an absent person — a false positive.

“WRONG” — The dog shows a negative indication within the decision area in case of a presented scent article from a present person — a false negative. “WRONG” is also stated, if the dog does not show a negative indication within the decision area in case of a presented scent article from blood extracts without DNA — a false negative.

A binomial test was used to demonstrate the abilities of the canines to use a specific person’s scent on a scent article to detect and to follow the matching scent trail of this specific person while discriminating from nonmatching scent trails and to demonstrate the absence of a matching scent trail better than chance. Percentages for “CORRECT” and “WRONG” findings were calculated. A goodness of fit test was used to test for differences between experimental groups. A test of independence was used to test for differences between the used scent articles, to test the influence of the position (left or right) in the T-junction of the target person/runner and to test the influence of the sex of the subjects on the success rate. A probability of p < 0.05 was assumed to be statistically significant, a value of p < 0.01 as highly statistically significant.

3. Results

3.1. Test series 1

During test series 1, the gauze swab scent article with a sample from the armpit was used. Runner A and B walked along the pavement on a road with normal traffic and hid behind an object or in a doorway (Fig. 1). The dogs were required to differentiate between the scent of the correct person and a negative among the scents of all present people, like for example, camera operator, attendants and pedestrians.

All dogs performed significantly better than expected. The overall performance of the group was 76%. The police dogs scored “CORRECT” with a percentage of 86%, the private rescue dogs with a percentage of 66%. Both, the results for the groups as a whole as well as the results for the two subgroups (police dogs and private rescue dogs), were highly significant (p < 0.01; p_All = 1.71 × 10⁻¹⁶, p_Pol = 9.14 × 10⁻¹³, p_Res = 3.86 × 10⁻⁶). The individual dogs successfully identified the odour of the target person from 64% up to 92%. These results are also highly significant. The only dog with only significant results was dog H4. There are no results for dog H6, as it was not able to participate in test series 1 (Fig. 2).




[list=links-for-figure]
[*]Download high-res image (336KB)
[*]Download full-size image
[/list]

Fig. 2. Comparison of trails during test series 1. Dogs’ individual performance and p-values with the gauze swab scent article with a sample from the armpit. None of the dogs indicated a false negative.

3.2. Test series 2

During test series 2, the foam blank scent article with a sample from the armpit was used. The experimental setup corresponds to that of test series 1 (Fig. 1). Due to the pregnancy and maternity protection of the dog handler, dog H4 could not participate in the trials. All dogs performed better than expected. The overall performance of the group in finding the “CORRECT” runner was only 66%. The police dogs identified the target person successfully 71% of the times, the private rescue dogs have a success rate of 62%. Both, the results for the groups as a whole as well as the results for the two subgroups (police dogs and private rescue dogs), were highly significant (p < 0.01; p_All = 2.88 × 10⁻²⁰, p_Pol = 3.37 × 10⁻¹³, p_Res = 6.05 × 10⁻⁹). The individual dogs successfully identified the odour of the target person from 58% up to 85%. These results are also highly significant. The only dog with just significant results was dog H6 (Fig. 3).




[list=links-for-figure]
[*]Download high-res image (324KB)
[*]Download full-size image
[/list]


Fig. 3. Comparison of trails during test series 2. Dogs’ individual performance and p-values with the foam blank scent article with a sample from the armpit. Dog H5 indicated one and dog H7 two false negatives.

3.3. Test series 3

During the 3rd test series, the foam blank scent article with a sample from saliva was used. The experimental setup corresponds to that of test series 1 (Fig. 1). All dogs performed better than chance. The overall group performance in finding the “CORRECT” runner was 78%. The police dogs identified the target person with a success rate of 85%, the private rescue dogs with a success rate of 65%. As in test series 1 and 2 the results for test series 3 group results were highly significant (p < 0.01; p_All = 3.78 × 10⁻⁶⁷, p_Pol = 2.91 × 10⁻⁵⁹, p_Res = 4.15 × 10⁻¹³). The individual dogs successfully identified the odour of the target person from 64% up to 89%. These results are highly significant (Fig. 4).




[list=links-for-figure]
[*]Download high-res image (328KB)
[*]Download full-size image
[/list]


Fig. 4. Comparison of trails during test series 3. Dogs’ individual performance and p-values with the foam blank scent article with a sample from saliva. Dog H1 indicated three and dog H4 two false negatives.

3.4. Trails with negative-samples

As stated above, the dog handler randomly chose one of three scent articles. Hence every selection of a negative was possible. If only the tests with a chosen negative are considered, individually, all dogs performed better than chance, with the police dogs having highly significant results. One of them, H6, had a success rate of 100% and for the three others (H1, H2, H3), only one wrong decision per dog was recorded in these tests. This corresponds to a hit ratio of 100%, 96%, 95% and 96%, respectively. One of the private dogs (H5) committed an error three times. This still corresponds to a hit ratio of 87%. It is important to keep in mind that the scent article, which was a negative, actually contained an odour and was sampled in the same way as the scent articles of the present trail layers. In about a fourth of all trials a negative was chosen (
 = 23.40%, SD = 3.12). In these tests group results were also highly significant (p < 0.01; p_All = 4.15 × 10⁻⁴⁹, p_Pol = 3.32 × 10⁻⁴³, p_Res = 6.61 × 10⁻¹¹, Table 2).

3.5. Test series 4

Test series 1 to 3 were evaluated to determine whether scent articles with sweat or saliva as a source of individual odour are suitable for identification. Furthermore, a negative versus a positive trail had to be identified. (Practical) experience has shown that dogs are able to track an individual person using scent articles with a variety of substances (including blood). An individualised assignment of traces is essential in criminal proceedings. In police preliminary proceedings, the DNA analysis of found traces can provide valuable indications for the identification of the perpetrator. In addition to test series 1–3, a “proof-of-concept” approach, with only a small number of subjects, tested whether DNA extracted from whole blood is sufficient as an individual odour source. The individual dogs successfully identified the target person from 60% up to 89%. With the exception of dogs H5 and H7, the results were highly significant (p < 0.001). The success rate corresponds to that of the previous scent articles. The overall performance of the group identifying the trail of the target person was 76%. This corresponds approximately to the dog’s success rate with sweat or saliva. The uneven number of attempts resulted from the random choice and from the fact that not all dogs were available on each day. Dog H4 could not participate due to personal circumstances of the dog handler. Out of the 61 test rounds using DNA samples as odour source, 17 were performed using blood extracts without DNA. These were, except in one case, indicated by the dogs as “negative”. In addition, two other rounds with DNA were tested as “negative”, since the corresponding people were not on site. Both cases were reported as “negative” (Fig. 5).



[list=links-for-figure]
[*]Download high-res image (321KB)
[*]Download full-size image
[/list]


Fig. 5. Comparison of trails during test series 4. Dogs’ individual performance and p-values with the foam blank scent article with a sample from DNA. Dogs H2 and H5 indicated one false negative.


Significant differences in the success rate with respect to the scent articles used were observed. The dogs were significantly more successful with saliva than with axillary sweat (X² = 6.6412, p = 0.009965). It was observed that three dogs, especially with the foam blank armpit sampling material, had lower success rates than with the other scent articles. However, statistical significance was only found in two dogs (H3, ² = 9.71 p = 0.007789, H6, X² = 9.71 p = 0.02353). Across all attempts, it was found that the position of the target person in the T-junction (left, right) had no effect. When the runner was placed at the left side of the T-junction, the success rate was 72%, when the runner was placed at the right side of the T-junction a success rate of 70% was observed (_left = 70,98%, σ²_left = 0,12%, SD_left = 3,45,

_right = 69,51%, σ²_right = 0,93%, SD_right = 9,66). Altogether, the success rate of the police dogs was significantly much better than that of the private rescue dogs (X² = 27,2269, p = 0,00001).

3.6. Subjects

A total of 190 subjects were involved in the trials. See Table 3 for the distribution of age and sex. The subjects’ sex had no statistically significant effect on the success ratio of the dogs (X² = 0.2242, p = 0.635881). Of the subjects involved, 58 subjects stated that they had consumed foods that could have an intense effect on the body odour (e.g. garlic) within the last 48 h before the test. However, a statistically detectable effect on the success or failure ratio could not be observed (X² = 0.0129, p = 0.909658).

4. Discussion

Our work shows success rates (correctly identified trails) between 50% and 92%. The success rates for studies with a comparable experimental design, but with other scent articles (STU-100 gauze pads), are between 70% [21], or 54.5% and 91.7% [22].

In our study, the police dogs performed more successfully with a success rate of 82%, compared with the private rescue dogs, whose average success rate was 65%. There may be some various reasons why the police dogs performed better. On the one hand, the police dogs were trained in the same way over a period of two and a half years, according to the training method of Armin Schweda, starting when they were puppies [29]. On the other hand, these dogs already had some years of experience and had to undertake a regular, weekly training, alongside a variety of other tasks. Furthermore, the dog handlers have been experienced police dog handlers for many years. Three out of four of the participating police dogs were those of the Bloodhound breed. These dogs also had the highest success rates in the overall performance. The exceptional scenting ability of the bloodhound had been shown in other studies [20,21,28]. Due to the small number of participating dogs, it is not possible to say whether the breed-specific suitability of bloodhounds influenced the results of this study. In general, bloodhounds have a comparatively large olfactory epithelium with the corresponding number of olfactory receptor cells [39,40]. Overall, the private rescue dogs involved in the study were not as extensively trained and do not have as much experience in field operations as the police dogs. The most successful private rescue dog, with a total success rate of 72%, was trained, as the police dogs, by the method of Armin Schweda. The least successful dog usually works with another handler than the one who participated in the tests. Therefore, this team did not have extensive training and operational experience together. This dog scored a total success rate of 57%. These results are consistent with findings of other studies, which state that experienced dogs or dogs undergoing intensive training are more successful [19,20].

Some studies showed that dogs (C. familiaris) are able to match items touched by a person to this individual person or to another item touched by individual in the so called line-up [12–19,38,41,42]. Brisbin and Austad described that dogs were not able to distinguish their handler’s elbow scent from scent samples taken of the hand of other persons [42]. Subsequent studies show that dogs are able to distinguish odours from different body parts [13,14,18]. In addition to the methodical characteristic that the dogs that Brisbin and Austad used were (only) trained to recognize the hand scent of the dog handler, Brisbin and Austad themselves point out that scent discrimination tests and tracking experiments are two types of scent-matching tasks [43]. Also, there is scientific evidence that dogs can distinguish individual humans by odour [11,44–48]. Possibly, this includes the ability to generalize odours from different body parts to the human odour trail. A few studies showed the ability of dogs to track the odour trail of a person, even under real conditions [11,20–23,28]. In these studies, handkerchiefs [11], gauze pads contaminated using a special vacuum device — the scent transfer unit STU-100 [20–22,28] and, sterile gauze compresses wiped at various body sites [23] were used as scent articles. In addition, it is important to note that it is still unclear which components of the human scent are used by the dogs for individual trailing. A terminology of three components for factors of human scent has been described. First, the “primary odour”, which contains constituents that are stable over time regardless of diet or environmental factors. The “secondary odour” contains constituents which are present due to diet and environmental factors and the “tertiary odour” contains constituents which are present because of the influence of outside sources. These factors are described as either stable over time (genetic factors) or vary with environmental or internal conditions [33,49,50]. Independently, it is unclear whether the dog’s olfactory perception of individual human scent traces in tracking or trailing tasks is really based on the classification of primary, secondary and tertiary odours. According to our understanding, the primary odour can be considered as a key stimulus for individual scent tracing. It is known from personal experience, that several items, like clothing, saliva, cigarette butts, breathing air, etc. can be used as scent articles for well trained dogs. From these experiences and considering studies carried out so far, the question arises as to whether and which similarities scent articles have and which components allow an individual discrimination. For the individual trailing of scent traces, it is very probable that there must be a common feature in the scent articles but in particular amongst the scent articles and the traces left behind. This can possibly be DNA.

Test series 1 has shown that using a gauze pad with a sample from the armpit as a scent article dogs can follow individual human scent trails and determine the direction of this trail statistically significant (p < 0.05) or highly statistically significant (p < 0.01) within a range of 50m. This also applies to test series 2, where a foam blank with the sample from an armpit was used. Previous studies considered that besides axillary sweat, saliva is also suitable for the distinction between individuals for the chemical analytical examination [51–54]. To our knowledge, there are no studies using saliva samples as scent articles to determine an individual human odour trail. Therefore, in test series 3 a foam blank contaminated saliva was used as a scent article. All dogs performed statistically highly significant (p < 0.001), using the saliva sample. A statistical analysis of the success rate as a function of scent articles shows, that for three dogs the success rate for the saliva sample on foam as scent articles was slightly above and for two dogs slightly below the success rate of the armpit sweat sample on gauze pad as scent articles. However, dog H1 had a better success rate for the sampling from the armpit on gauze pad as scent articles (success rate 92%) in contrast to the sample from saliva on a foam blank (success rate 76%). Nevertheless, these results were not statistically significant for the individual dogs. All dogs, except dog H1 as mentioned above, were more successful with the saliva sample on a foam blank as scent article than with the axillary sweat sample on a foam blank. For two dogs, the difference in performance between the two scent articles was statistically significant. For one of the dogs the success rate increased by 31% from 58% to 89%, when using the saliva sample. The success rate for the second dog was 64% for the axillary sweat sample and 88% for the saliva sample (p_H3 = 0.002366, p_H6 = 0.02353, p < 0.05). Taking all results into account, the dogs achieved better results using the saliva scent articles than using the axillary sweat scent articles. Saliva consists mainly of water and, in addition to proteins and trace substances, also contains epithelial cells, as well as erythrocytes and leukocytes [55–58]. Likewise, saliva contains sufficient DNA for a genome analysis, allowing the identification of a specific person [59–61]. For a review comparing saliva-derived genomic DNA to blood-derived DNA with regards to quality, quantity and convenience, see Sun and Reichenberger [62]. These findings underline the approach in police investigations, in which buccal swabs are used for comparison specimen [63].

Significant results were observed in test series 1–3. In test series 4, scent articles with samples of DNA, extracted from whole blood for a feasibility trial were used. Also in these tests, the results of the police dogs were statistically highly significant. For the private rescue dogs no significant results could be achieved with these scent articles. Due to the low number of samples, the results are to be considered critically. However, the following effects were observed. Using wet samples (buffer solution), the overall success rate was 81% (26 out of 32) and thus statistically significant (p < 0.001, p = 2.78 × 10⁻⁸). Taking the results of the individual dogs, only the results of the four police dogs were statistically significant (p < 0.05, p_H1,H6 = 0.012, p_H2,H3 = 0.008, p_H5 = 0.2548, p_H7 = 0.33). After letting the samples dry for 24 h, a dramatic drop in performance was observed with only 5 out of 12 correct identifications. The other seven samples were reported as a “negative”. This also occurred in 16 out of 17 cases using the blood extracts without DNA as scent article.

The reliability of the dogs used in searching for missing persons or in the reconstruction of criminal offences is essential. Above all, when this is offered as evidence in court, the question arises as to the validity of the method. Schoon described a method for determining a diagnostic ratio for positive and negative identification (ID). This is expressed as the quotient of the percentage of correct ID in suspect = perpetrator cases divided by percentage of false ID in suspect ≠ perpetrator cases or percentage correct non-ID in suspect ≠ perpetrator cases divided by misses in suspect = perpetrator cases [17]. In the present study, however suspect = perpetrator cases were used in the positive experiments and suspect ≠ perpetrator cases in the negative experiments. Therefore, another way, test sensitivity and specificity, for evaluating the results is applied. The sensitivity results from the quotient of true positive results and the sum of the true positive and false negative results. The specificity results from the quotient of true negative results and the sum of the true negative and false positive results. The sensitivity in the police dogs was 0.98 (median 0.99), the specificity 0.97 (median 0.96). In the private rescue dogs, the sensitivity was 0.95 (median 0.95) and the specificity was 0.75 (median 0.68). For the identification of a perpetrator in forensic investigations a high specificity is particularly important. Therefore, the practical meaning of the results observed can be seen especially from the error rate at the “negatives”, max. 5% in the police dogs, max. 31% in the private rescue dogs. In the police dogs were de facto three misses in 101 attempts, i.e. a success rate of 97%. This is also underlined by the fact that among all the trials, the dogs performed only ten times a “false negative”, four times in the police dogs (<1%) and six times (<2.3%) in the private rescue dogs. In criminal proceedings, especially the correct response of a “negative” is particularly important. The faulty indication may lead to false suspicions. In each and every case, the dog handlers have to decide whether their dog presented them with a negative or positive trail and whether this is reliable. In this study, especially the police dogs presented a negative trail with high reliability. One dog always performed accurately and the other three dogs were observed to have made only one mistake. This may result from a misinterpretation of the dog handler, external influences, or the temperament of the dog [64]. According to the German legal system, the public prosecutor determines which means and methods are applied in criminal proceedings. Therefore, private dog handlers are also used in investigations in Germany. Appropriate certification or verification prior utilizing is not required. In this study, the police mantrailer dogs tested under operating conditions had a success rate of up to 92% (median: 83%). Furthermore, they differentiate highly significantly between the presence and the absence of an odour trace using a scent article containing human scent with a success rate up to 100% (median 96%). In contrast, private rescue dogs had an error rate between 13% and 35% for this particular task. Therefore, only well-trained and appropriately tested or certified dogs should be used to support law enforcement authorities. Indeed it should always be taken into account that both, dog and handler, are living beings, and therefore cannot be expected to perform like a machine, both dog and man can make mistakes. In criminal proceedings, this must be compensated by the use of at least two dogs and video recordings for a posteriori analysis [27].

Wells and Hepper state that the trails being run from left to right could be identified more easily by the dogs than those from right to left [65]. These preferences could not be confirmed in this study (X² = 0.3352 p = 0.562643). Nevertheless, a dog successfully performed in 92% of the cases where the target person walked to the right side of the T-junction (X² = 5.6064, p = 0.017895), whereas this was only the case in 71% of the cases in which the runner walked to the left. For another dog, there was observed an error rate of 63% in the cases when the target person walked to the right side of the T-junction as opposed to the error rate in the cases of left-handed trails which have a score of 28% (X² = 9.4839, p = 0.002073). The findings here coincide with the findings of Mackenzie and Schultz [66], who state that there is no side preference. It is noteworthy that in almost all experiments none of the dogs choose the third possibility, to go back to the starting area, instead of turning to the right or to the left. If at all, they just walked back a few metres. Dogs have the ability to determine the direction of an odour trail left by a human, in other words, dogs can determine the time-dependent course of a trail — five steps are sufficient [67]. This allows the dog to differentiate the direction. Another aspect is the scent attractiveness of the target person to dogs. Jezierski et al. found that 1.1% of the subjects had a high attractiveness (50–75% false alarms) and 76,3% of the subjects had a low attractiveness (0–25% false alarms) to the dogs [68]. In this context, the influence of the sex on the scent attractiveness to dogs may be seen. According to Jezierski et al. dogs identified women significantly better than men in lineups [69]. Neither the sex nor the age of the target person had a significant effect on the search results. This is in line with the findings of Rogowski [70].

To our knowledge, this is the first study demonstrating that saliva and DNA samples provides enough information for the dog to search for an individual. In a proof of concept approach, the search was also performed with scent articles from samples made of DNA, which has been extracted from whole blood. It turned out that all odour samples, containing no DNA at all, indicated as “negative” except for one, thus no trailing of the corresponding person was possible. It requires a larger number of experiments to verify this approach. For further research, trailing dogs ability to follow aged trails under real-life scenarios is of interest, especially for the use in criminal proceedings. Harvey showed trailing dogs ability to follow aged trails for 48 h old trails [20], Wolf describes aged trails after 24 h, one week and one month [23]. However, in these cases, the person, who had laid the trail, returned to the location. There are only a few reports about aged trails where the person, who had laid the trail did not return after the track had been laid [1,48,71]. Here, further investigation is needed to determine the dogs’ ability to follow older trails left by people who either were never found or never returned after they left a trail. This would also correspond to the scenarios described in the introduction of this work. Such findings could be an additional basis for the courts to assess and determine the reliability of the use of mantrailers.

5. Conclusion

Test series are often carried out under laboratory conditions. In this study mantrailer dogs were tested under real police operating conditions. With a scent article from a person, they could follow the individual human scent trail of that person statistically highly significantly. Furthermore, they differentiate highly significantly between the presence and the absence of an odour trace using a scent article containing human scent. The best results were achieved with scent articles using saliva and sweat as an odour source. Additionally, highly significant results were obtained with scent articles from wet DNA samples. These findings may provide valuable information for the practical work of dog handlers for selecting suitable scent articles. The study shows that the success rate was higher depending on training method and operational experience level of the teams and for this reason only well-trained and certificated teams should be used in criminal proceedings. It has been shown that well-trained mantrailing-dogs can be in principle a reliable and useful tool for law enforcement authorities.

Acknowledgements

The authors wish to thank the participating dog handlers and their dogs, especially the private rescue dog handlers for driving on various occasions and in all kinds of weather without compensation, to participate in this study. Furthermore, special thanks go to Gerald Spindler and Achim Grüner of the Institute for Tropospheric Research Leipzig for supplying a weather station and the staff of the Department of Forensic Molecular Genetics of the Institute for Legal Medicine of the University of Leipzig for the processing of blood samples. We would like to thank Kai-Uwe Goss of the Department Analytical Environmental Chemistry, Helmholtz-Centre for Environmental Research — UFZ, Dirk Labudde and Hanna Siewerts, Forensic Science Investigation Lab, University of applied sciences Mittweida as well as Jason Smith, Schneeberg Police Academy who reviewed earlier versions of this paper.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

[1]
L. WoidtkeMantrailing – Fakten und Fiktionen

Eigenverlag der Hochschule der Sächsischen Polizei (FH), Rothenburg/Oberlausitz (2016)
[2]LG Dresden, Urteil vom 10.04.2013, Az.: 3 Kls 396 Js 201/11, unpublished, LG Dresden, 2013.[3]LG Lüneburg, Urteil vom 7. Mai 2014, Az.: 31 KLs/6001 Js 23887/13 (19/13), unpublished, LG Lüneburg, 2014.[4]LG Chemnitz, Urteil vom 08.10.2014, Az.: 250 Js 32750/13 unpublished, LG Chemnitz, 2014.[5]
B.d.L. Costello, A. Amann, H. Al-Kateb, C. Flynn, W. Filipiak, T. Khalid, D. Osborne, N.M. RatcliffeA review of the volatiles from the healthy human body

J. Breath Res., 8 (1) (2014), p. 29
[6]
G.S. SettlesSniffers: fluid-dynamic sampling for olfactory trace detection in nature and Homeland Security—the 2004 Freeman Scholar Lecture

J. Fluids Eng., 27 (2) (2005), pp. 189-218

CrossRefView Record in Scopus
[7]
B.A. Craven, G.S. SettlesA computational and experimental investigation of the human thermal plume

J. Fluids Eng., 28 (6) (2006), pp. 1251-1258

CrossRefView Record in Scopus
[8]
D. LicinaHuman Convective Boundary Layer and Its Impact on Personal Exposure

Department of Building National University of Singapore, Department of Civil Engineering Technical University of Denmark, National University of Singapore, Technical University of Denmark (2015)

p. 197
[9]
W.G. SyrotuckScent and the Scenting Dog, 8. Auflage 2013 ed

Barkleigh Productions, Inc., Mechanicsburg, Pennsylvania (1972)
[10]
E. LocardDie Kriminaluntersuchung und ihre wissenschaftlichen Methoden, 2. Auflage ed

Kameradschaft Verlagsges, Berlin (1930)
[11]
H. KalmusThe discrimination by the nose of the dog of individual human odours and in particular of the odours of twins

Br. J. Anim. Behav., III (1955), pp. 25-31

ArticleDownload PDFView Record in Scopus
[12]
J.E. King, R.F. Becker, J.E. MarkeeStudies on olfactory discrimination in dogs: (3) ability to detect human odour trace

Anim. Behav., 12 (2–3) (1964), pp. 311-315

ArticleDownload PDFView Record in Scopus
[13]
R.H. Settle, B.A. Sommerville, J. McCormick, D.M. BroomHuman scent matching using specially trained dogs

Anim. Behav., 8 (6) (1994), pp. 1443-1448

ArticleDownload PDFView Record in Scopus
[14]
G.A. Schoon, J.C. De BruinThe ability of dogs to recognize and cross-match human odours

Forensic Sci. Int., 9 (2) (1994), pp. 111-118

ArticleDownload PDFView Record in Scopus
[15]
G.A.A. SchoonScent identification lineups by dogs (Canis familiaris) experimental design and forensic application

Appl. Anim. Behav. Sci., 9 (3) (1996), pp. 257-267

ArticleDownload PDFView Record in Scopus
[16]
G.A.A. SchoonScent identifications by dogs (Canis familiaris): a new experimental design

Behaviour, 34 (7) (1997), pp. 531-550

CrossRefView Record in Scopus
[17]
G.A. SchoonA first assessment of the reliability of an improved scent identification line-up

J. Forensic Sci., 3 (1) (1998), pp. 70-75

View Record in Scopus
[18]
V.V. Vyplelová Petra, Pinc Ludvík, Pacáková Zuzana, Bartos Ludek, Santariova Milena, Capková ZuzanaIndividual human odor fallout as detected by trained canines

Forensic Sci. Int., 234 (2014), pp. 13-15
[19]
S. Marchal, O. Bregeras, D. Puaux, R. Gervais, B. FerryRigorous training of dogs leads to high accuracy in human scent matching-to-sample performance

PLoS One, 11 (2) (2016), p. e0146963

CrossRef
[20]
L.M. Harvey, J.W. HarveyReliability of bloodhounds in criminal investigations

J. Forensic Sci., 48 (4) (2003), p. 6

View Record in Scopus
[21]
R.A. StockhamSurvivability of human scent

Forensic Sci. Commun., 6 (4) (2004), pp. 1-9

View Record in Scopus
[22]
A.M. Curran, P.A. Prada, K.G. FurtonCanine human scent identifications with post-blast debris collected from improvised explosive devices

Forensic Sci. Int., 199 (1–3) (2010), pp. 103-108

ArticleDownload PDFView Record in Scopus
[23]
A. WolfUntersuchung des Einflusses der Alterung menschlicher Geruchsspuren auf die Ausarbeitung der Fährten durch Personensuchhunde

Tierärztliche Hochschule Hannover, Verlag, Gesellschaft Service GmbH, Gießen (2016)

p. 229
[24]
N.R. CouncilReference Manual on Scientific Evidence

(third edition), The National Academies Press, Washington, DC (2011)
[25]SCUS, Daubert v. Merrell Dow Pharmaceuticals, Inc. 509 U.S. 579, Supreme Court of the United States, 1993.[26]
A.M. Christensen, C.M. Crowder, S.D. Ousley, M.M. HouckError and its meaning in forensic science

J. Forensic Sci., 59 (1) (2014), pp. 123-126

CrossRefView Record in Scopus
[27]LG Nürnberg-Fürth Urteil vom 13.12.2012, Az.: 13 KLs 372 Js 9454/12, juris LG Nürnberg-Fürth, 2012.[28]
L.M. Harvey, S.J. Harvey, M. Hom, A. Perna, J. SalibThe use of bloodhounds in determining the impact of genetics and the environment on the expression of human odortype

J. Forensic Sci., 1 (5) (2006), pp. 1109-1114

CrossRefView Record in Scopus
[29]
A. Schweda, T. Schweda, A. NestlerVon der Basis zum erfolgreichen Mantrailing, Finden statt suchen

Müller Rüschlikon Verlag (2012)
[30]SMI., Fachkonzept für die Aus- und Fortbildung sowie Prüfung von Personensuchhunden (Fachkonzept Personensuchhunde), Anlage 15 (zu Abschnitt 3 Ziff. I Nr. 3) VWV DHW Pol, nicht veröffentlicht, 2016, p. 7.[31]DRK, Prüfungs- und Prüferordnung für Rettungshundeteams Mantrailing, Deutsches Rotes Kreuz e. V., 2012, p. 35.[32]
B.A. Eckenrode, S.A. Ramsey, R.A. Stockham, G.J. Van Berkel, K.G. Asano, D.A. WolfPerformance evaluation of the Scent Transfer Unit (STU-100) for organic compound collection and release

J. Forensic Sci., 1 (4) (2006), pp. 780-789

CrossRefView Record in Scopus
[33]
A.M. Curran, C.F. Ramirez, A.A. Schoon, K.G. FurtonThe frequency of occurrence and discriminatory power of compounds found in human scent across a population determined by SPME-GCMS

J. Chromatogr. B Anal. Technol. Biomed. Life Sci., 846 (1–2) (2007), pp. 86-97

ArticleDownload PDFView Record in Scopus
[34]
L.E. DeGreeff, A.M. Curran, K.G. FurtonEvaluation of selected sorbent materials for the collection of volatile organic compounds related to human scent using non-contact sampling mode

Forensic Sci. Int., 209 (1–3) (2011), pp. 133-142

ArticleDownload PDFView Record in Scopus
[35]
A. Schoon, R. Fjellanger, M. Kjeldsen, K.-U. GossUsing dogs to detect hidden corrosion

Appl. Anim. Behav. Sci., 153 (2014), pp. 43-52

ArticleDownload PDFView Record in Scopus
[36]
J.A. Zeugin, J.L. HartleyEthanol precipitation of DNA

Focus, 7 (4) (1985), pp. 1-2

View Record in Scopus
[37]
PowerQuantTMTechnical Manual

(2015)
[38]
G.A.A. SchoonThe effect of the ageing of crime scene objects on the results of scent identification line-ups using trained dogs

Forensic Sci. Int., 47 (1) (2005), pp. 43-47

ArticleDownload PDFView Record in Scopus
[39]D. Shier, J. Butler, R. Lewis, Hole’s Human Anatomy and Physiology, 11th ed., Michelle Watnick 2007.[40]
S.M. StejskalDeath, Decomposition, and Detector Dogs: From Science to Scene

(1st ed.), Taylor & Francis (2012)
[41]
P.G. HepperThe discrimination of human odour by the dog

Perception, 7 (4) (1988), pp. 549-554

CrossRefView Record in Scopus
[42]
I.L. Brisbin Jr., S.N. AustadTesting the individual odour theory of canine olfaction

Anim. Behav., 2 (1) (1991), pp. 63-69

ArticleDownload PDFView Record in Scopus
[43]
I.L. Brisbin Jr., S.N. AustadThe use of trained dogs to discriminate human scent: a reply

Anim. Behav., 46 (1993), pp. 191-192

ArticleDownload PDFView Record in Scopus
[44]
G.J. RomanesExperiments on the sense of smell in dogs

J. Linn. Soc. Zool., XX (117) (1887), pp. 61-70
[45]
L. LöhnerÜber menschliche individual- und regionalgerüche

Pflüger’s Archiv für die gesamte Physiologie des Menschen und der Tiere, 202 (1) (1924), pp. 25-45

CrossRef
[46]
R. Menzel, R. MenzelDie Verwertung der Riechfähigkeit des Hundes im Dienste der Menschheit, Kameradschaft

Verlagsgesellschaft m. b. H., Berlin (1930)
[47]
B.A. Sommerville, R.H. Settle, F.M.C. Darling, D.M. BroomThe use of trained dogs to discriminate human scent

Anim. Behav., 46 (1993), pp. 189-190

ArticleDownload PDFView Record in Scopus
[48]
R.A. Stockham, D.L. Slavin, W. KiftSpecialized use of human scent in criminal investigations

Forensic Sci. Commun. Res. Technol., 6 (3) (2004), p. 13
[49]
A.M. Curran, S.I. Rabin, K.G. FurtonAnalysis of the uniqueness and persistence of human scent

Forensic Sci. Commun. (2005), p. 16

View Record in Scopus
[50]
A.A. Schoon, A.M. Curran, K.G. FurtonOdor biometrics

S.Z. Li, A. Jain (Eds.), Encyclopedia of Biometrics, Springer, US, Boston, MA (2009), pp. 1009-1014

CrossRef
[51]
D.J. Penn, E. Oberzaucher, K. Grammer, G. Fischer, H.A. Soini, D. Wiesler, M.V. Novotny, S.J. Dixon, Y. Xu, R.G. BreretonIndividual and gender fingerprints in human body odour

J. R. Soc. Interface, 4 (13) (2007), pp. 331-340

CrossRefView Record in Scopus
[52]
H.A. Soini, I. Klouckova, D. Wiesler, E. Oberzaucher, K. Grammer, S.J. Dixon, Y. Xu, R.G. Brereton, D.J. Penn, M.V. NovotnyAnalysis of volatile organic compounds in human saliva by a static sorptive extraction method and gas chromatography-mass spectrometry

J. Chem. Ecol., 6 (9) (2010), pp. 1035-1042

CrossRefView Record in Scopus
[53]
M. Kusano, E. Mendez, K.G. FurtonComparison of the volatile organic compounds from different biological specimens for profiling potential

J. Forensic Sci., 8 (1) (2013), pp. 29-39

CrossRefView Record in Scopus
[54]
J.S. Brown, P.A. Prada, A.M. Curran, K.G. FurtonApplicability of emanating volatile organic compounds from various forensic specimens for individual differentiation

Forensic Sci. Int., 226 (1–3) (2013), pp. 173-182

ArticleDownload PDFView Record in Scopus
[55]
J.K.M. Aps, K. Van den Maagdenberg, J.R. Delanghe, L.C. MartensFlow cytometry as a new method to quantify the cellular content of human saliva and its relation to gingivitis

Clin. Chim. Acta, 321 (1–2) (2002), pp. 35-41

ArticleDownload PDFView Record in Scopus
[56]
J.K.M. Aps, L.C. MartensReview: the physiology of saliva and transfer of drugs into saliva

Forensic Sci. Int., 150 (2–3) (2005), pp. 119-131

ArticleDownload PDFView Record in Scopus
[57]
R.G. Schipper, E. Silletti, M.H. VingerhoedsSaliva as research material: biochemical, physicochemical and practical aspects

Arch. Oral Biol., 52 (12) (2007), pp. 1114-1135

ArticleDownload PDFView Record in Scopus
[58]
W. BuchallaSpeichel Zusammensetzung und Funktion eines oft unterschätzten Helfers

Zahnärztliche Mitteilungen, 2013 (08) (2013), p. 130

View Record in Scopus
[59]
T. Rylander-Rudqvist, N. Håkansson, G. Tybring, A. WolkQuality and quantity of saliva DNA obtained from the self-administrated oragene method—a pilot study on the cohort of swedish men

Cancer Epidemiol. Biomark. Prev., 15 (9) (2006), pp. 1742-1745

CrossRefView Record in Scopus
[60]
D. Quinque, R. Kittler, M. Kayser, M. Stoneking, I. NasidzeEvaluation of saliva as a source of human DNA for population and association studies

Anal. Biochem., 53 (2) (2006), pp. 272-277

ArticleDownload PDFView Record in Scopus
[61]
Y. Hu, E.A. Ehli, K. Nelson, K. Bohlen, C. Lynch, P. Huizenga, J. Kittlelsrud, T.J. Soundy, G.E. DaviesGenotyping performance between saliva and blood-derived genomic DNAs on the DMET array: a comparison

PLoS One, 7 (3) (2012), p. e33968

CrossRef
[62]
F. Sun, E.J. ReichenbergerSaliva as a source of genomic DNA for genetic studies: review of current methods and applications

Oral Health Dent. Manage., 3 (2) (2014), pp. 217-222

View Record in Scopus
[63]Landeskriminalamt, Standards der Kriminaltechnik, Landeskriminalamt Sachsen Projektgruppe “Standards der Kriminaltechnik”, nicht veröffentlicht, VS-nfD, 2016, p. 231.[64]
O. Leitch, A. Anderson, K. Paul Kirkbride, C. LennardBiological organisms as volatile compound detectors: a review

Forensic Sci. Int., 32 (1) (2013), pp. 92-103

ArticleDownload PDFView Record in Scopus
[65]
D.L. Wells, P.G. HepperDirectional tracking in the domestic dog, Canis familiaris

Appl. Anim. Behav. Sci., 84 (4) (2003), pp. 297-305

ArticleDownload PDFView Record in Scopus
[66]
S.A. Mackenzie, J.A. SchultzFrequency of back-tracking in the tracking dog

Appl. Anim. Behav. Sci., 17 (3–4) (1987), pp. 353-359

ArticleDownload PDFView Record in Scopus
[67]
P.G. Hepper, D.L. WellsHow many footsteps do dogs need to determine the direction of an odour trail?

Chem. Senses, 30 (2005), pp. 291-298

CrossRefView Record in Scopus
[68]
T. Jezierski, J. Ensminger, L.E. PapetCanine Olfaction Science and Law: Advances in Forensic Science, Medicine, Conservation, and Environmental Remediation

CRC Press, Boca Raton (2016)
[69]
T. Jezierski, M. Sobczynska, M. Walczak, A. Gorecka-Bruzda, J. EnsmingerDo trained dogs discriminate individual body odors of women better than those of men?

J. Forensic Sci., 7 (3) (2012), pp. 647-653

CrossRefView Record in Scopus
[70]
M. RogowskiAn attempt to determine the influenco of a female scent placed in a male selective sequence on the indications of dog used for the identification of human olfactory traces

Prob. Forensic Sci., LIII (2003), pp. 74-90

View Record in Scopus
[71]
L. WoidtkeMantrailing at the Police of Saxony

Kwartalnik Policyjny, 3 (2016)

2017 Elsevier B.V. All rights reserved.