In a letter dated October 13, 1996, Whitehurst wrote to the OIG expressing concerns about testimony by CTU examiner Roger Martz in Florida v. George Trepal. In 1991, Trepal was found guilty of one count of murder and six counts of attempted murder as a result of his adding the poison thallium nitrate to bottled Coca-Cola. Trepal is now challenging his conviction, for which he was sentenced to death. At the trial, Martz testified that [b]ased on [a diphenylamine] test I concluded that thallium nitrate was added to the Coca-Cola. Martz also testified that a white powder, which had been found in a bottle in Trepal's garage, was thallium nitrate.


Whitehurst complained in his letter that Martz had misstated the significance of certain analytical tests or otherwise testified inaccurately. Some of Whitehurst's allegations, such as suggesting that Martz should have volunteered in testifying that his undergraduate major was in biology, do not merit further discussion here. As a result of Whitehurst's letter, however, we did identify several concerns about Martz's work in this case that are similar to ones we noted in certain other matters discussed earlier in this Report.


To investigate this matter, we obtained the Laboratory's case file, reviewed transcripts of testimony by Martz in a pretrial deposition and at the trial, and questioned Martz in a sworn interview. After the FBI provided written comments on a draft of this portion of the report, we also interviewed the two Laboratory examiners who had provided scientific information for the FBI's comments: chemist Thomas Jourdan, who recently became the unit chief for the newly created Materials and Devices Unit and Steven Burmeister, who now is an examiner in the CTU.


Given the tests that Martz actually performed, he could have properly stated in his dictation and testimony that two samples of Coca-Cola, identified as Q1 and Q2, were consistent with thallium nitrate having been added to them. Alternatively, he correctly could have observed that Q1 and Q2 had elevated levels of thallium and nitrate ions as compared to unadulterated Coca-Cola. Martz, however, did not limit his conclusions in this way, and his work on the case was deficient in several respects: (1) his dictation stated that the nitrate ion was identified in samples Q1 through Q3 and those samples were consistent with thallium nitrate having been added to them; this was incorrect insofar as he had not performed tests necessary to reach these conclusions with regard to Q3; (2) Martz did not acknowledge certain data obtained from the tests he performed; (3) he failed to perform additional tests that were appropriate under the circumstances; (4) in testifying, Martz improperly offered a stronger opinion about the identification of thallium nitrate than he had expressed in the dictation reviewed by his supervisor and included in the Laboratory report; (5) Martz did not adequately document his work, his case notes were incomplete, undated and inaccurate, and the charts were not accurately or clearly labeled; (6) he lacked a sufficient analytical basis to opine that a bottle containing thallium nitrate found in Trepal's garage, identified as Q206, contained no other drug residues ; (7) he also gave an unsupported opinion about the purity of the thallium nitrate in Q206; and (8) Martz in his deposition and trial testimony made various inaccurate, incomplete, or unsupported statements.


In December 1988, several unopened Coca-Cola bottles were sent to the FBI Laboratory for examination to detect tampering. The bottles had been found in the home of a women who, along with her two sons, had become suddenly ill. The woman, Peggy Carr, ultimately died from thallium poisoning; the children survived. After the bottles were received in the Laboratory, examiner Don Havekost of the Elemental Analysis Unit determined through the use of inductively coupled plasma atomic emission spectroscopy (ICP) that the contents of several bottles contained thallium. The ICP results showed that samples Q1 through Q3 contained from 403 to 915 milligrams of thallium, while none was identified in unadulterated Coca-Cola.

After Havekost identified thallium in Q1 through Q3, Martz was asked to examine these samples further to possibly determine the form in which the thallium was present. Thallium can appear as a soluble salt in different compounds, including thallium chloride, thallium sulphate, and thallium nitrate. Martz conducted tests that included a diphenylamine (DPA) test and ion chromatography (IC). Based on the DPA and IC tests, Martz concluded that nitrate ions were present in Q1 through Q3.


Martz prepared dictation that was approved by his unit chief Steve Allen in February 1989 and later incorporated verbatim into a Laboratory report dated July 10, 1990. The report was disclosed to the defense attorneys and Martz was questioned about it, albeit perfunctorily, in a pretrial deposition and at trial. In his dictation, Martz stated:


The contents of Q1 through Q3 . . . were analyzed for an anion associated with thallium. The nitrate ion was identified in Q1 through Q3. The Q1 through Q3 cokes are consistent with thallium nitrate having been added to them.


(Emphasis added).


Martz had sufficient analytical data to support this conclusion with regard to samples Q1 and Q2, but not Q3. His IC results clearly showed increased concentrations of nitrate ions in samples Q1 and Q2 as compared to known, unadulterated Coca-Cola. The case notes, however, do not reflect any IC analyses for Q3. The DPA test referred to by Martz is a color spot test for a range of oxidizing compounds, which include but are not limited to nitrates. Given the tests he performed, Martz should not have stated in his dictation that the nitrate ion was identified in Q3.


At trial, when asked if he had done any test other than the DPA test to determine if there was thallium nitrate in the Coca-Cola, Martz stated that he did one other test called ion chromatography. In his deposition, after describing the DPA and IC test results, Martz said he had cover[ed] his entire investigation of this case. Martz failed to acknowledge data he had obtained from other analytical tests. In addition to the DPA and IC analyses, Martz's notes indicate he also performed silver nitrate and barium chloride spot tests on Q1 through Q3 and a sample of known Coca-Cola. These tests, respectively, will indicate the presence of chloride or sulphate. The results of these tests are ambiguously stated in Martz's notes and, given his testimony, he evidently did not rely upon them in reaching his conclusions.


Martz also analyzed sample Q1 by mass spectrometry (MS), scanning electron microscopy/energy dispersive x-ray analysis (SEM/EDX), and x-ray powder diffraction (XRPD). The XRPD results, obtained after the sample was treated by burning, indicated the presence of thallium chloride, but not thallium nitrate. The MS data, also obtained after burning or other treatment of the sample, indicated the possible presence of thallium chloride, phosphate, sulphate, and possibly oxide, but not thallium nitrate. The SEM/EDX results showed the presence of thallium, sodium, potassium, calcium, chlorine, and phosphorous.

The fact that the MS and XRPD results for Q1 did not show the presence of thallium nitrate may be explained by the sample preparation process. Martz, however, did not acknowledge this analytical data in his trial or deposition testimony. In his interview with the OIG, he said he did not rely on the MS, SEM/EDX, or XRPD data in reaching his conclusion that thallium nitrate had been added to the Coca-Cola. This conclusion, Martz said, was based on the DPA and IC tests and the information about thallium he received from Havekost.


Martz also failed to perform additional tests that were appropriate under the circumstances. First, he did not perform the simple validation experiment of adding thallium nitrate to known unadulterated Coca-Cola and analyzing the mixture in the same manner as Q1. This could have usefully indicated whether the results that Martz observed from his tests other than DPA or IC were due to the sample preparation process. Moreover, Martz could have quantified the nitrate he identified in the questioned samples. Havekost had identified the quantity of thallium in Q1 through Q3. If Martz had quantified the nitrate, he could have determined if the relative amounts of the thallium and nitrate were consistent with their having been present in the Coca-Cola in the form of the compound thallium nitrate.


By testifying that in his opinion thallium nitrate had been added to the Coca-Cola, Martz overstated the significance of the analytical results in a manner similar to what he did in the World Trade Center case. As discussed in Part Three, Section C, there Martz as the chief of the CTU approved Lynn Lasswell's conclusion that mass spectrometry had identified urea nitrate on certain evidence, when the results in fact merely established the presence of urea and nitrate ions. In Trepal, the test results that Martz said he relied upon showed the presence of nitrate in samples Q1 and Q2. The ICP results from the tests earlier performed by examiner Havekost identified thallium in samples Q1, Q2, and Q3. Taken together, these results showed that nitrate and thallium ions were present in Q1 and Q2, but they do not necessarily demonstrate that the compound thallium nitrate had been added to the Coca-Cola.


When Martz opined that thallium nitrate had been added to the Coca-Cola, he also went beyond his written report, which said only that the samples were consistent with thallium nitrate having been added to them. We do not accept Martz's view that an examiner may properly offer an opinion about the identification of a questioned substance that is stronger than the conclusions described in the Laboratory report where, as here, the opinion rests on the same data and analyses as the report. In our interview, Martz's explanation for his testimony appeared to be that he was convinced he was right. The process of supervisory review -- which is intended to assure that conclusions are reasonably supported -- is seriously undermined if examiners feel free to offer opinions that are stronger than those their unit chiefs have reviewed and approved for inclusion in reports. In this case, Martz should have confined his conclusions to those contained in his dictation.


Martz also failed to prepare notes and charts that adequately described the analyses performed and the results obtained. We have commented on this same issue in our discussion of Martz's work in the VANPAC case. In Trepal, Martz's handwritten notes indicate that he performed spot tests with AgCl (sic) and BaNO3 (sic) with the results same for all. In his OIG interview, Martz acknowledged that his notes were inaccurate and said he meant that he did a silver nitrate (AgNO3) test for chloride and a barium chloride (BCl) test for sulphate. His notes do not indicate whether same for all means the results were positive or negative, and Martz in his interview said he thinks the results for chloride were positive and he cannot remember what the results for sulphate were. This again illustrates the importance of accurate notes, both to permit thorough supervisory review and to allow an examiner later to refresh his or her memory. We also are concerned that Martz's notes do not describe the use of appropriate blanks or traceable standards.


Samples were not clearly identified on certain MS charts, and Martz acknowledged that other MS charts were actually labeled incorrectly. The IC charts, which Martz thinks reflect work performed by someone in the MAU, do not identify who ran the tests, are not dated, do not specify the instrumental conditions, and contain unidentified handwritten notes. The deficient documentation for the IC tests is particularly troublesome because Martz principally relied upon the IC results in concluding that thallium nitrate had been added to the Coca-Cola.


Martz lacked a sufficient analytical basis to opine, as he did in his pretrial deposition, that a brown bottle containing thallium nitrate contained no other drug residues. This bottle, which had been found in Trepal's garage in December 1989, was identified as Q206. Martz testified that, based on XRPD and FTIR analyses, no drug residues were present in the thallium nitrate found in the bottle. With regard to the FTIR, Martz relied on spectra for the questioned sample and a sample of known thallium nitrate of unspecified origin that were each of unacceptably poor quality. Spectra of this quality would not permit the identification of the questioned sample as thallium nitrate, much less rule out the presence of other possible drugs. Furthermore, XRPD, which identifies only crystalline compounds, is not the method of choice for identifying drug residues.


During his OIG interview, Martz acknowledged that based on the XRPD and FTIR results, his conclusion was debatable. Quite inexplicably, in testifying about drug residues, Martz did not refer to MS results that were probably his best data, although at the trial he later mentioned the MS analysis in testifying on another issue. To determine if drug residues were present, Martz should have made an extraction of the sample and followed a protocol of the type routinely used in forensic laboratories for drug identification.


Martz also made unsupported statements in testifying about the purity of thallium nitrate identified in specimen Q206, the powder found in a bottle in Trepal's garage. In his deposition, Martz said that Q206 was easily over 90 percent thallium nitrate. And possibly as high as 95 to 99 percent. At trial, he stated, In my opinion it's greater than 95 percent pure, because I did not find any other impurities. Martz lacked sufficient data to support these statements, given that he said they were based only on the FTIR and XRPD analyses and, as he acknowledged in his testimony, he did not attempt to quantitate the powder.


Finally, Martz in his deposition and trial testimony made various inaccurate, incomplete or unsupported statements. In his deposition, he stated that to his knowledge, thallium nitrate would not be absorbed into the hand. During his OIG interview, he acknowledged that this statement was not based on specific literature, but instead reflected his general belief that such chemicals are not absorbed through the skin. We readily identified, however, an article published in 1988 which states that severe systemic poisoning has occurred from skin contact, and other basic reference sources note that thallium nitrate may be absorbed through the skin. On this topic, Martz should have said he did not know whether it would be absorbed through the skin since he was not able to provide a properly informed answer.


At trial, Martz misspoke in stating initially that he concluded based on the DPA test that thallium nitrate had been added to the Coca-Cola. He acknowledged in an OIG interview that his conclusion was based on both the DPA test and the IC results, which he did later discuss in his testimony. In describing the IC results at trial, however, Martz testified that all samples were tested while his charts indicate that only two of three samples were. Martz told the OIG that sometimes when he had three samples that were suspected to be the same, he would not test all of them. This might be acceptable if his reports and testimony properly reflected the actual work performed; they do not. Martz also testified that no nitrates were present in the known Coca-Cola, without noting that the IC results indicated at least trace amounts of nitrates. During his OIG interview, Martz described these various misstatements as oversights or technicalities. By so characterizing his conduct, Martz seemed not to appreciate the importance of accurate testimony.


In his work in Trepal, Martz appeared to have a lower threshold of scientific proof than is generally accepted in forensic science and to lack appropriate scientific rigor in his approach to examinations. Martz did not conduct additional tests that were appropriate under the circumstances, and in reaching his conclusions, he did not consider analytical results of his tests other than the DPA and IC tests. His case notes and charts were inaccurate and incomplete, and he testified inaccurately on several points. He committed a serious error when, based on the same data, he rendered an opinion at trial that was stronger than the opinion in his dictation. Such conduct by an examiner is of concern whenever it is displayed in casework, and it is particularly disturbing in a matter such as Trepal where the death penalty is a potential result.