Introduction
Ten years ago Taslitz (1990), in his comprehensive study on dog scent lineups, expressed a view that the uniqueness and stability of human odour had not been established yet and experimental research in the area of dog scent identification still was in its infancy, therefore scent lineups should be excluded from evidence in court. A decade in forensic science is more than a century in law, however, so it seems worthwhile to look at the possibilities of canine human scent identification again.
The new Polish code of Criminal Procedure, which has been in effect since September 1st, 1998 includes the principle of free appraisal of evidence in art. 7: ,,The procedural organs determine their position based on all of the evidence, appraised independently according to the rules of correct reasoning as well as on the basis of knowledge and life experience". According to art. 193 § 1 if the determination of circumstances of significant importance for the solving of a case requires specialised knowledge, the concerned parties consult an expert or experts. However, the Polish judicature, unlike the American, has not elaborated on a general standard for the evaluation of scientific evidence. For this reason, dog scent lineup will be analysed in the light of the standards expressed by the American supreme Court in Daubert v. Merrel Dow Pharmaceuticals (1993): verifiability, publication and peer review, known error rate, standardisation and, last but not least, general acceptance—an echo of Frye.
Is the method verifiable
?It is impossible to find the so called demarcation criterion which would enable us to discern science from nonscience. However, a possibility for empirical verification of hypothesis or statements is vital to any branch of empirical knowledge, hence to forensic science as well. Moreover, the degree of empirical verifiability is considered one of the important development indicators of science (Konieczny, 1984). The American supreme court was therefore right in establishing such a criterion for scientific methods in judicial proceedings. As seen from the European viewpoint, Daubert did not ,,replace a judicial anachronism with a philosophical one" (Allen, 1994. Pp. 1168-1169) and it is not ,,a giant step backwards" (Bohan, Heels, 1995), as some commentators assert. Dog scent lineups undoubtedly meet the verifiability criterion. However, the same cannot be said in respect to the other requirements.
Has the method been tested, peer-reviewed and published?
Although the first (para) scientific experiment on canine olfactory abilities dates from the nineteenth century (Romanes, 1887), comparatively little scientific research in this field has been carried out so far, and even less has been published. It is indisputable that canines are able to detect human scent (King, Becker, Markee, 1964); however, in a scent lineup their task is not to detect but to discriminate. The uniqueness and stability of human odour have not been carried out. Furthermore, it turned out that dogs had difficulties in identifying scent traces of the same person originating from different parts of the body (Brisbin Jr., Austad, 1991; Schoon, De Bruin, 1994).
Our knowledge of factors which could influence dog scent identification is limited as yet. Polish police experts are doing research on the influence of cosmetics, vinegar, sealing wax and other substances on dogs’ indications. A student of mine (Kasprzyk, 1998) examined dogs’ ability to determine whose scent trace had been left on a car seat, as well as the influence of occupational odours (butchers’ and petrol-related) on dog identification. Russian forensic scientists (Gricenko, Obidin, Starowojtow, 1999) performed some preliminary experiments on the time factor in human scent identification. However, a lot of factors like alcohol, drugs, tobacco, trace freshness, etc., which are important in everyday investigative practice, have not been examined yet. We have to come to the conclusion that Taslitz’s (1990) opinion about the incipient stage of the research on dog scent lineups is still valid, even taking into consideration some additional research on their reliability.
Does the method have a known error rate?
The problem of validity and reliability is vital for any forensic science method. Dog scent lineup reliability was examined chiefly by Schoon (1997, 1998). It was measured in diagnostic ratio, i.e. the quotient of ,,hits" and ,,false alarms". Depending on the experiment’s configuration, the raw diagnostic ratio ranged from 4.3 (Schoon, 1997) to 6.9 (Schoon, 1998), whereas the diagnostic ratios of other common forensic identification methods vary from 3 to 160 (Schoon, 1997). It should be noted that the percentage of correct dog scent identifications according to the Dutch research was not impressive: 32% and 58% (Schoon, De Bruin, 1994), 29% (Schoon, 1997) and 22% (Schoon, 1998). Much more promising results were obtained by Settle, Sommerville McCormick, and Broom (1994). The average score of the dogs was 8-% and 85%, but the experimental design was far removed from real forensic circumstances, as the dog was to find a matching scented cloth which had been taken from the same jar as the one initially presented.
Overall, the error rate of this type of human identification seems to be high and, accordingly, the diagnostic value of the method is rather low. As a matter of fact, the diagnostic ratios of some other methods like paint, glass, or fibre analyses are even lower (Schoon, 1998), but it should not be overlooked that those methods do not lead to the identification of an individual. On the other hand, still too few methodologically sound experiments shave been performed to establish the diagnostic value of the method in question.
In a casework the following criteria should be used in the assessment of dog scent lineup reliability:
1. Presence of the perpetrator’s scent races at the crime scene.
2. Total number of traces presented.
3. Similarity of the traces presented: especially the attractiveness of the comparative (i.e. the one collected from the suspect) trace.
4. Previous dog’s knowledge about the traces presented.
5. Number of trials.
6. Independence of trials: number of dogs, number of comparative traces.
7. Configuration of the identification procedure: evidence trace comparative traces and decoys or comparative trace evidence trace and decoys.
8. Presence of a handler or other person.
9. Certification of the dogs.
Is the method standardised?
Dog scent lineups have been used as an investigative tool in many countries, mostly since the sixties or early seventies. In Russia, in the Ministry of the Interior Forensic Science Centre alone, over 200 expert opinions based on scent lineups were provided in 1998 (Primenow, Starowojtow, 1999). In Poland, in March 1993, the so-called ,,osmological sections" were established in the police forensic science laboratories and the specialisation in that area was later also introduced as a field of expertise. In 1996, the peak of the method’s use, 1,963 courts’ and prosecutors’ decisions on scent lineups were realised by the Polish police; in 1997—only 1,170. In 1998, as many as 1,518 such decisions were realised (an increase of 29%, using 98 dogs which performed 46,228 tests altogether (an increase of 36%), in 36 police forensic science laboratories. In 936 test sets (out of 1,518) a scent identity was established (Analiza, 1999). In Cracow, 45 such orders were realised in 1998 and 27 in the first half of 1999.
In Polish criminal trials the dog lineup can have a twofold procedural form: an experiment, performed by a police officer, public
prosecutor or even a judge, or an expert opinion provided by police specialists. The latter form is becoming more and more frequent. The Ministry of Justice Institute of Forensic Research is sometimes asked to assess the value of such opinions in the form of a superopinion (Wojcikiewicz, 1998).
As for dog scent lineups in Poland, a methodology has been established in the police instruction on human scent race examination, issued in June, 1998 (Metodyka, 1998). Although for a year its legal position was rather vague—it has been sanctioned ex post by § 7 of the circular from May, 1999 only (Zarzadzenie, 1999)—it nevertheless has been influencing police practice. According to this instruction, scent traces at the crime scene should be developed for at least 30 minutes with an absorbent and then preserved in a glass jar. Comparative traces from a suspect should be collected, preferably from the palms, for 15 minutes. In particular cases, comparative traces can also be taken from other parts of the body or from blood. Contamination of the traces must be avoided. All scent races (evidence, comparative and decoys) are collected on nappy-like cloths (absorbents) and preserved in glass jars.
Recognition of scent traces may be performed not less than 24 hours after their preservation, using a minimum of two special dogs, in the form of a row (minimum 5 scents altogether) or a circle (minimum 10 scents). Control trials have to be conducted before the actual identifications, in order to check the attractiveness of the scents as well as the dog’s suitability for the work. The examination of scent identity should be repeated at least three times, and the position of the evidence scent should be changed each time. However, a blank lineup, when a suspect’s trace is not presented in a parade, is not mentioned there. A preferable configuration of the identification procedure is from the evidence trace to the comparative one.
A holder cannot of course know where the scent in question is situated.
The instruction unfortunately sanctions the two-phase form of dog scent identification as a valid experiment or expert opinion. On the other hand, it includes a very important clause that any methodological corrections can be introduced only after appropriate research. However, the requirements for a conclusive opinion, assumed in Polish police practice to be matching results obtained by both dogs used in a test, seem to be insufficient when compared for instance to those of haemogenetic examinations.
One of the most important conditions for the lineup to have proper diagnostic value is the similarity of all presented scent traces, which also means that the suspect’s scent must not be attractive for dogs. However, routine attractiveness tests which consist of the identification of a control sample in the presence of the evidence scent, sometimes are not sufficient and a blank test, when a dog initially does not smell any odour, is needed to exclude the attractiveness of the suspect’s scent (Gola, Orzechowski, 1998). This is probably caused by the act that a dog prefers to be obedient and correctly recognizes a control scent despite the attractiveness of the suspect’s. Only later, when there are no corresponding scents (because the suspect was not a donor of the scent at the crime scene) does the dog choose only the suspect’s because of its attractiveness. Such dogs’ tendencies, without any counteraction, could doubtless cause miscarriages of justice.
When analysing the methodology of dog scent lineups in different countries, we come to the conclusion that it is far from uniform and standardized. It varies, for instance, in respect to:
Time of trace development: in Russia it cannot be shorter than one hour (Riekomiendacii, 1999),
Nature of comparative traces: usually it is sweat scent, in Russia—dried blood scent (Riekomiendacii, 1999),
Carriers of comparative traces: in the Netherlands, Germany, Belgium, Denmark they use stainless tubes or cloths (De Bruin, 1989; Gricenko, 1997), in Poland—cloth absorbents,
Course of a lineup: in the Netherlands live parades are also possible, the procedure should be repeated only once, generally few methodological lineup variants are used (De Bruin, 1989; Gricenko, 1997), blank lineups are not universally required, and a uniform attractiveness test methodology has not yet been established.
There are also differing approaches towards the evidential use of dog scent lineup results.
Is the method generally accepted?
Dog scent identification has become a true ,,queen of evidence" in Polish criminal trials. There are cases where scent lineup is the only evidence; moreover, such cases even exist where, despite the existence of other traces of high quality, no other appropriate expert had been consulted at all (Wojcikiewicz, 1998). The famous Cracow case of the so-called ,,Collector", an alleged five-time murderer, where almost three hundred dog scent lineups in four police forensic science laboratories had been carried out, ended in acquittal in April 1999, mainly because of many methodological errors made by the police during the lineups and the attractiveness of the defender’s scent.
In May 1999, I surveyed 41 judges ruling in criminal trials in the Cracow district. Dog scent lineup was regarded as ,,scientific evidence". Many respondents had not been misled by the name of the method which in Polish forensic language is called ,,osmology", with a suffix implying that it is a branch of science, not only a method of human identification. The judges were also asked whether they would sentence a defendant on the grounds of a dog scent lineup as ,,the only piece of evidence". Only 9 (22%) respondents would do it whereas, again all of them would sentence a defendant solely on the grounds of haemogenetic evidence. These (preliminary) results show that at least some Polish judges have a rather reasonable attitude towards this kind of evidence, perceiving its advantages and disadvantages.
The Polish judicature as a whole, however, does not take an unequivocal position. In a burglary case, the Supreme Court accepted as evidence the identification made by one dog only and wrongly allowed that it be the same one that had been used as the tracking dog (Wyrok Sadu Najwyzszego, 1996). In another judgement the Court stated that the evidence resulting from a scent experiment is a special piece of evidence and its positive result can be an additional, important link, but that its negative result does not have any value when other evidence of guilt exists (Wyrok Sadu Najwyzszego, 1998). In another decision (Postanowienie Sadu Najwyzszego, 1998b), issued in the above-mentioned ,,Collector’s" case, the Supreme Court rightly stated that osmological opinions, together with facial sketches and a polygraph examination, are sufficient for detention if their results are convergent. In the case of the General Fonkiewicz homicide, the Court expressed an opinion (Postanowienie Sadu Najwyzszego, 1998a) that dog scent identification is circumstantial evidence sufficient for detention, especially when it was not questioned, five dogs were used, and similar results were obtained.
However, it seems that the lower courts are more circumspect than the Supreme Court. The Warsaw Court of Appeals, in the same case, stated that osmological opinions can be treated only as circumstantial evidence to be corroborated by other evidence and are not sufficient for detention (Postanowienie Sadu Apelacyjnego, 1998). The Lublin Court of Appeals, in its judgement from 1998 (Wyrok Sadu Apelacyjnego, 1998), warned against the uncritical treatment of the results of dog scent lineups.
In other countries, the judicature’s position concerning dog scent lineups is various. The American courts sometimes regard them as admissible evidence, as do the Dutch, Russian, Belorrusian, German, Hungarian, Belgian and Danish courts (Taslitz, 1990, Gricenko, 1997). On the other hand, in Lithuania and Ukraine, dog scent lineups are used by the police only as an investigative method (Pimienow, Starowojtow, 1999).
As concerns the scientists’ views, the situation looks similarly: apart from strong dog scent lineup proponents, like Schoon (1997, 1998), Sommerville, Settle, Darling, Broom (1993) and Bednarek (1998) we can also find their detractors like Taslitz (1990), as well as the cautious and objective critics like Hanausek (1998) and Widacki (1998). In any event, the method undoubtedly has yet to gain general acceptance.
Conclusions
Canine identification of human scent does not yet have a proper scientific foundation. Aslitz’s (1990, p. 134) view that ,,the courts (…) must continue to be alert to the scientific nature of scent lineup evidence" has not become outdated; however, such expertise should not be called ,,junk science" in general. The method has been introduced into trial proceedings too early, by overly hasty police practitioners which have caused miscarriages of justice. Moreover, there have been no field studies. On the other hand, the same could be said ten years ago about DNA analysis, flourishing at present. Criminal trial is not the domain of scientists, but singular cases provide impetus for the development of forensic science (Waltols, 1998); Glazek, Wojcikiewicz, 1998). In this situation the criteria of ,,scientific evidence" like those in Daubert, are of great importance. They are very helpful for judges in evaluating the reliability of dog scent lineups and therefore can be considered as a reasonable, auxiliary standard of the reliability of lineups even in Polish criminal trials, especially in the absence of indigenous criteria. After all, nobody should be convicted solely on the basis of a dog wagging its tail.
References
Allen R.J. (1994). Expertise and the Daubert decision, Journal of Criminal Law & Criminology, 84, 4, 1157-1175.
Analiza zagadnien dotyczacych uzywania I utrzymania psow sluzbowwych Policji w 1998 r. [Analysis of problems connected with the use and keeping of police dogs in 1998] (1`999), Warsawa: Centralne Laboratorium Kryminalistyczne Komendy Glownej Policji.
Bednarek T. (1998). Podstawowe zagadnienia z badan osmologicznych [Basic issues concerning scent examination], Problemy Kryminalisstyki, 222, 32-34.
Bohan T. L., Heels E.J. (1995). The case against Daubert: The new scientific evidence ,,standard" and the standards of the several states, Journal of Forensic Sciences, 40, 6, 1030-1044.
Brisbin I. L. Jr., Austad s. N. (1991). Testing the individual odour theory of canine olfaction, Animal Behaviour, 42, 63-69.
Daubert v. Merrel Dow Pharmaceuticals (1993).
De Bruin J. C. (1989). The detection dog and science, Rotterdam: Dog Training Centre. Frye v. United States (1923).
Glazek A., Wojcikiewicz J. (1998). Judicature as a stimulus to forensic science development. Paper presented at the 9th ENFSI Meeting, Lisbon, Portugal, 23-25 April.
Gola A., Orzechowski J. (1998). Rola i znaczenie badan na atrakcyjnosc w procesie identyfikacji sladow zapachowych [The role and significance of tests for attractiveness in the process of scent race identification], Problemy Kryminalisyki, 222, 40-41.
Gricenko W. W. (1997). Idientifikacija czietowieka po zapachowym sliedam za rubiezom [Human identification on scent traces abroad], unpublished manuscript.
Gricenko W. W., Obidin A. b., Starowojtow W. I. (1999). Wlijanije faktora wriemieni na obrazowanije, sochrannost’ i wozmoznost’ issliedowanija zapachowych sliedow czietowieka [The influence of time on the development, preservation and examination possibilities of human scent traces], Moskwa: Ministerstwo wnutriennych diet Rossijskoj Fiedieracuji, Ekspiertno-kriminalisticzieskij centr.
Hanausek T. (1998). Meandry osmologii [The meanders of osmology], Palestra, 1-2, 41-46.
Kasprzyk J. (1998). Ekspertyza osmologiczna [Osmological expertise], unpublished master’s thesis, Krakow: Uniwersytet Jagiellonski.
King J. E., Becker r. F., Markee J. E. (1964). Studies on olfacotry discrimination in dogs: (3) Ability to detect human odour trace, Animal Behaviour, XII, 2-3-, 311-315.
Konieczny J. (1984). Metodologiczna charackterystyka kryminalistyki [The methodological character of criminalistics], Katowice: Uniwersytet Slaski.
Metodyka pobierania, zabezpieczania i rozpoznawania zabezpieczonych slladow zapachowych ludzi (1998) [Methodology of collection, preservation and identification of preserved human scent traces], Legionowo-Warzawa: CSP-CLK KG Policji.
Pimienow M. G., Starwojtow W. I. (1999). Kriminalisticzieskoje issliedowanije zapachowych sliedow czietowieka [Criminalistic examinations of human scent traces], unpublished manuscript.
Postanowienie Sadu Apelacyjnego w Warszawie [Decision of the Warsaw Court of Appeals] (1998). II Aka 259/98, from 10 July (unpublished).
Postanowienie Sadu Najwyzszego [Decision of the Supreme Court] (1998a). II KZ 93/98, II KZ 98/98, from 12 August (unpublished).
Postanowienie Sadu Najwyzszego [Decision of the Supreme Court] (1998b). IV KO 101/98, from 21 December (unpublished).
Riekomendacja po rabotie s zapachowymi sliedami czielowieka pri prowiedienii sliedstwiennych diejstwij i opieratiwno-rozysknych mieroprijatij [Guidelines for the treatment of human scent traces in trial and investigative practice] (1999), Moskwa: Ekspiertno-kriminalisticzieski centr MWD Rossii.
Romaes g. J. (1887), Experiments on the sense of smell in dogs, Nature, 21 July, 273-274.
Schoon G.A.A. (1997). Scent identifications by dogs (canis familiaris): A new experimental design, Behavior, 134, 531-550.
Schoon G.A.A. (1998). A first assessment of the reliability of an improved scent identification line-up, Journal of Forensic Sciences, 43, 1, 70-75.
Schoon G. A. A., De Bruin J. c. (1994). The ability of dogs to recognize and cross-match human odours. Forensic Science International, 69, 111-118.
Settle R. H., Sommerville, B. A., McCormick J., Broom, D.M. (1994). Human scent matching using specially trained dogs, Animal Behaviour, 48, 1443-1448.
Sommerville B. A., Settle R. H., Darling F.M.C., Broom D.M. (1993). The use of trained dogs to discriminate human scent, Animal Behaviour, 46, 189-190.
Taslitz A. e. (1990). Does the cold nose know? The unscientific myth of the dog scent lineup, The Hastings Law Journal, 42, 1, 14-134.
Ustawa z dnia 6 czerwca 1997 r., kodeks postepowania karnego [Act from 6 June, 1997, Code of Criminal Procedure], Dz. U. Nr 89, poz. 555.
Waltos S. (1998). Proces karny. Zarys systemu [Criminal procedure. An outline of the system], Warszawa: Wydawnictwa Prawnicze PWN.
Widacki J. (1998). Kilka uwag o identyfikacji zapachow ludzkich przez psa na uzytek procesu karnego [Some remarks on the identification of human scent by a dog for use in crimnal proceedings], Palestra, 11-12, 102-108.
Wojcikiewicz J. (1998). Metaekspertyza osmologiczna [,,Osmological" metaopinion], Z Zagadnien Nauk Sadowych-Problems of Forensic Sciences, XXXVII, 158-164.
Wyrok Sadu Apelacyjnego w Lublinie [Judgement of the Lublin Court of Appeals] (1998). II Aka 142/97, from 29 September (unpublished).
Wyrok Sadu Apelacyjnego [Judgement of the Supreme Court] (1996). IV KKN 46/96, from 22 July (unpublished).
Wyrok Sadu Apelacyjnego [Judgement of the Supreme Court] (1998). V KKN 44/97, from 30 January, Orzecznictwo Prokuratury i Prawa, 1998, 6, 12.
Zarzadzenie Nr 10/99 Komendanta Gtownego Policji (1999), z dnia 28 maja, w sprawie metod i form wykonywania zadan zwiazanych z organizacja, sposobami uzywania i utrzymania psow stuzbowych Policji [Circular of the Chief Commander of the Police from 28 May, on the methods and forms of performing tasks connected with the organisation, anners of use, and keeping of police dogs].
