The FBI Laboratory: An Investigation into Laboratory Practices and Alleged Misconduct in Explosives-Related and Other Cases (April,1997) Office of the Inspector General, Department of Justice

On April 19, 1995, an explosion destroyed the Alfred P. Murrah Federal Building in Oklahoma City, Oklahoma, resulting in the deaths of 168 people. A massive investigation was undertaken by the FBI, as lead agency, with the assistance of other agencies including the Bureau of Alcohol, Tobacco and Firearms (ATF). Two persons, Timothy McVeigh and Terry Nichols, were indicted for causing the blast, but have not yet been tried.

In 1995 SSA Frederic Whitehurst sent the OIG correspondence criticizing members of the Laboratory for their handling of the Oklahoma City case. On January 9, 1996, Whitehurst sent the OIG a 30-page letter criticizing the September 5, 1995, FBI Laboratory explosives report of SSA David Williams in that case. Whitehurst contends that Williams' report goes beyond Williams' expertise, is biased in favor of the prosecution, and contains unjustified conclusions.

To investigate Whitehurst's allegations, we interviewed Whitehurst, Williams, EU Chief J. Thomas Thurman (who reviewed and approved Williams' report), Steven Burmeister (a CTU explosives residue examiner who worked on the case), James Corby (former MAU Chief), as well as other FBI and ATF personnel, some of whom worked at the scene of the blast. We also submitted written questions to Roger Martz, the CTU Chief who worked briefly on the Oklahoma City case. Additionally, we considered pertinent FBI documents and applicable literature in the field of explosives.

As explained below, we conclude that in his report Williams repeatedly reached conclusions that incriminated the defendants without a scientific basis and that were not explained in the body of the report. We find fault with other aspects of the report as well. We also conclude that Thurman performed an inadequate review of Williams' report by allowing Williams too much discretion and by approving conclusions with which Thurman disagreed and could not support. Further, we conclude that Martz improperly deviated from the explosives residue protocol in his examination of some specimens. Finally, we conclude that Whitehurst's numerous other contentions lack merit.

The following section (Section II) analyzes Williams' report, and Section III addresses Thurman's review of the report. Section IV addresses allegations concerning Martz's examination of evidence. Section V states our conclusions.

As discussed with reference to the Trade Center bombing, Williams is the only EU examiner who has offered opinions of a specific velocity of detonation (VOD) of the main explosive of a bombing based on the damage at the crime scene. See Part Three, Section C, n.24, supra, and accompanying text. Williams has done so in two cases, the World Trade Center case and the Oklahoma City case. Williams' September 5, 1995, Oklahoma City report reads as follows:

In his OIG interview Williams stated that this 13,000 feet per second opinion had a tolerance on either side of 1,000 feet per second.

For the reasons stated in our discussion of the World Trade Center case (see Part Three, Section C, p.118, supra), Williams' specific VOD opinion in the Oklahoma City case lacked an adequate scientific and empirical basis.

Williams also stated in his report that [a] fertilizer base explosive, such as ANFO (ammonium nitrate and fuel oil), among other commercial and improvised explosives, has an approximate VOD of 13,000 fps. Williams thus stated that the approximate VODs of both the main charge and ANFO were each 13,000 feet per second, which supported his theory that the main charge was ANFO.

The statement of the VOD of ANFO, however, is incomplete because ANFO has a broad VOD range. For example, the Dupont Blasters' Handbook (Dupont) shows commercial ANFO products with VODs in the 7,000-15,600 feet-per-second range. When Williams wrote his Oklahoma City report, he was aware of this range:

Additionally, Williams' working hypothesis in the Oklahoma City case was that the ANFO used by the perpetrators was not produced commercially but was rather improvised -- that is, the offenders mixed the ammonium nitrate and diesel fuel themselves. Presumably the quality control of improvisers would be inferior to that of commercial manufacturers. If the ingredients were not combined in the correct ratio, the VOD of the resulting explosive might be reduced. Accordingly, improvised ANFO would have at least as broad a range of VODs as that of commercial ANFO.

Thus, ANFO can detonate at a VOD of 13,000 feet per second, but it can also detonate at lower (7,000 feet per second) and higher (15,600 feet per second) velocities. By only mentioning an ANFO velocity of 13,000 feet per second, Williams suggested too strongly that there was an exact match between the VOD of ANFO and his reported VOD of the main charge.

Williams testified at his OIG interview that determining that the main charge had a specific VOD of 13,000 feet per second, with a tolerance on either side of 1,000 feet per second, did not limit the main charge to a specific explosive. Williams acknowledged that there are a lot of different explosives in the range of 12-14,000 feet per second. Williams also acknowledged that although ammonium nitrate crystals were found at the post-blast scene, there are many explosives in the range of 12-14,000 feet per second that have ammonium nitrate in them. Nevertheless, Williams concluded in his report ( it is also the opinion of this examiner ) that the main explosive used at Oklahoma City was ANFO. He acknowledged that he reached this conclusion, in part, because Terry Nichols, one of the defendants in the case, purchased ammonium nitrate and diesel oil prior to the bombing. Without the evidence of these purchases, Williams admitted he would have been unable to conclude that ANFO was used. Indeed, Williams stated that based on the post-blast scene alone [i]t could have been dynamite; I'm suggesting that there could have been other things.

We conclude that it was inappropriate for Williams to render a categorical opinion in his report that the main charge was ANFO. As discussed with reference to the World Trade Center case, it is inappropriate for a forensic Laboratory examiner to identify the main charge based in whole or in part on prior knowledge of the explosive components purchased by a defendant. Such an identification is not based on scientific or technical grounds and appears to tailor the opinion to evidence associated with the defendants.

Moreover, Williams' report does not mention that the defendant's purchases were the basis of the ANFO opinion. The report is presented as an FBI Laboratory report. It begins with the phrase, Results of examination: . The reader is left with the impression that the opinions presented are based on the scientific analyses of the FBI Laboratory. Accordingly, Williams' opinion that the main charge was ANFO appears to be based solely on his technical expertise as an explosives examiner and thus appears to be very incriminating to someone (like defendant Nichols) who allegedly purchased ANFO components before the Oklahoma City explosion. The opinion is thus misleading and presents the case in a way most incriminating to the defendants. Had Williams explicitly stated in his report that the ANFO opinion was based on the defendant's purchases, the opinion could have been appropriately discounted as a non-expert conclusion that seeks to match the characteristics of the explosion with evidence associated with the defendants.

As indicated, Williams told us that the crime scene was consistent with the use of an ammonium nitrate dynamite, which could have had a VOD in the range Williams estimated. The major components of ammonium nitrate dynamite (ammonium nitrate and nitroglycerin) were found at the crime scene. A dynamite wrapper was also found. Williams' report, however, fails to address the possibility that the main charge consisted of dynamite, which an objective report would explicitly have discussed.

We conclude that Williams' categorical conclusion that the main charge was ANFO was not scientifically justified and was based on improper grounds.

We question the basis for Williams' conclusion that the weight of the main charge was approximately 4000 pounds of ANFO.

As discussed in the section on the World Trade Center (see Part Three, Section C, pp. 132-134, supra), Williams' method of determining weight is impressionistic and depends on his VOD estimate, which is itself speculative. As noted in that section, other members of the EU do not routinely estimate the quantity of the explosive from a damage assessment because the placement and confinement of the explosive have a significant effect on the amount of the damage. Nevertheless, we concluded in the Trade Center case that Williams' size estimate of 1,000-1,500 pounds, which he characterized as a ballpark figure, was not, as such, an unreasonable opinion because he offered such a broad range.

With respect to the weight of the explosive, Williams' Oklahoma City report differs from his trial testimony in the Trade Center case in two respects. First, the Oklahoma City report does not offer a broad range but limits the estimate to approximately 4000 pounds of ANFO.

Second, it appears that Williams' opinion was based in part on the recovery of receipts showing that defendant Nichols purchased 4,000 pounds of ammonium nitrate. Williams testified at the OIG interview:

Saying that his 4,000 pound estimate is not solely based on the receipts implies that the opinion was based on the receipts in part. To this extent, the opinion was flawed for the same reason Williams' ANFO opinion was flawed because it was based on the receipts. Moreover, if Williams' opinion was based, in part, on the receipts, his report should have said so.

We conclude that Williams' weight estimate was flawed because it was more specific than warranted by the application of the forensic science and because it was based in part on collateral sources unrelated to laboratory or crime scene observations.

Several other conclusions in Williams' report were overstated and conformed to evidence associated with the defendants.

1. The report concludes that [t]he explosive main charge was contained in 50 gallon size white plastic barrels and white plastic barrels with blue plastic trim. Recovered at the blast site were white, blue, and black plastic fragments. Williams testified at the OIG interview that these fragments showed very unique explosive damage. Assuming the pieces were from a plastic barrel, [y]ou could tell the inside as compared to the outside of the barrel [fragments]. And in many cases, you could see that the explosive force came from inside to outside of the barrels. According to the AE dictation, the markings on one of these plastic fragments are similar to markings on 50 gallon size white plastic barrels and white plastic barrels with blue plastic trim recovered at defendant Nichols' residence.

We think it is unwarranted to render a categorical conclusion that the main charge was contained in plastic barrels of the same description as those found at Nichols' residence. First, Williams assumed that the main charge was ANFO, which would need containers for transport or storage. As explained above, the conclusion that the main charge was definitely ANFO was unwarranted. Second, since the Laboratory apparently has not made measurements such as the radius of curvature of the fragments (assuming they came from containers), it is virtually impossible to know that the containers definitely were 50 gallon barrels that were white or white with blue trim.

We conclude that Williams lacked a proper basis to state categorically that the main charge was contained in 50 gallon size plastic barrels of the description of those found at Nichols' house.

2. Williams' report states that [t]he initiator for the booster(s) was either a detonator from a Primadet Delay system or sensitized detonating cord. Primadet systems were found at defendant Nichols' house and an accomplice's house. Detonating cord normally contains PETN, which laboratory examinations associated with defendant McVeigh. No evidence of a Primadet system or sensitized detonating cord was found at the crime scene. As Williams told us at his OIG interview, the device used in the bombing is consistent with a Primadet system or detonating cord. I can't say yes and I can't say no. EU Chief Thurman told us in his interview that the appropriate conclusion was that the Primadet system or sensitized detonating cord could have been used. We conclude that it was improper for Williams to render a categorical conclusion identifying the initiator for the booster.

3. Williams' report also states that [t]he initiator for the primadet or the detonating cord was a non electric detonator; [n]on electric, burning type fuse of either hobby fuse or a commercial safety fuse was used as a safe separation and time delay system ; and [t]he time delay for the burning fuse was approximately 2 minutes and 15 seconds. Evidence linked the defendants to a burglary in which non-electric detonators were taken, and the named fuses were found at locations associated with the defendants. No evidence of a non-electric detonator or the named fuses, however, were found at the crime scene. Williams' conclusions were based in part on a videotape showing a Ryder truck appear near the Murrah Federal Building 2 minutes, 15 seconds before the explosion. Based on the tape, Williams posited that a 3 foot burning fuse was used, which he said would correlate with 2 minutes, 15 seconds.

We find that Williams' conclusions are overstated. The scenario he posits is one of many possibilities. For example, as acknowledged by Thurman, the initiator could have been electric, and the fuses named in the report were possibilities but not the only ones. Further, there could have been a longer time delay that was initiated before the truck appeared in the video.

Williams also stated in the OIG interview that his conclusion that the bomber used a 3 foot fuse was based on his assumption that the perpetrator had a military background. (Both defendants have military backgrounds.) It was improper for Williams to make that assumption unless he could do so based on the scientific evidence, and there is nothing in his report suggesting that the evidence indicates that the blast was perpetrated by someone with a military background.

We conclude that the categorical conclusions discussed above were inappropriate.

Williams' report is twenty-eight pages long and treats many subjects. The last two pages consist of conclusions, many of which are discussed above and most of which are categorical statements. The bases for these conclusions, however, are absent from the report. A reader of the report, for example, does not know why Williams concluded that the main charge was approximately 4,000 pounds of Ammonium Nitrate and Fuel Oil (ANFO) mixture or why the initiator for the primadet or the detonating cord was a non electric detonator. As acknowledged by Thurman, the conclusions in a report should be based on, and flow from, the contents of the report. Williams' report failed to meet that standard. As we discuss in Part Six, infra, we recommend that opinions in reports should be reasonably supported by the analysis and data, which should be described in the report.

Although many examiners from different units in the FBI Laboratory may work on a given case, the Laboratory requires one of the examiners (the principal examiner or PE ) to issue the official Laboratory reports in the case. The other examiners (the auxiliary examiners or AEs ) submit their reports ( dictation ) to the PE for incorporation in the official report. In 1992 Whitehurst complained that Thurman did not incorporate some of his AE dictation verbatim, and the matter was reviewed by the management of the Laboratory. See Section H10, infra. On September 1, 1994, the FBI Laboratory issued a policy memorandum explicitly requiring verbatim inclusion of AE reports. The memorandum purported to be a restate[ment of] long-standing policies. Thus, at the time of Williams' Oklahoma City report (September 5, 1995), he should have been well aware of the policy of verbatim inclusion of AE dictation.

Two passages in Williams' report (concerning specimens Q18 and Q171) raise questions about whether Williams attempted to circumvent the verbatim-inclusion rule with respect to two AE dictations of Steven Burmeister. We are unable to conclude that Williams intentionally violated the rule because Williams told us that Burmeister orally agreed to the modifications and Burmeister cannot recall whether he did so. However, one of the modifications makes no sense, and should at least have been rewritten. Both modifications exemplify the need for strict adherence to the verbatim-inclusion rule and the problems that can arise with oral agreements to modify reports.

The pertinent part of Burmeister's dictation regarding specimen Q18, the knife seized from defendant McVeigh at the time of his arrest, is as follows:

Williams reproduced this passage verbatim in a report issued before his September 5, 1995, explosives report. But in the September 5th report Williams stated as follows: Traces of PETN were located on specimen Q18, however could not be confirmed. This statement comes in the section of Williams' report dealing with the booster used in the explosion, and the report states that a booster can take the form of several different high explosives including PETN.

The statement [t]races of PETN were located on specimen Q18, however could not be confirmed is internally inconsistent and nonsensical. Confirmation is a prerequisite for a determination that a substance is located on an exhibit. Without confirmation of PETN, the exact identity of the traces on Q18 is unknown. Because there was no confirmation here, the report should not have said that traces of PETN were located on specimen Q18.

Williams insisted at his OIG interview, despite vigorous questioning on the matter, that the statement was not internally inconsistent. Of the many reasons for the verbatim-inclusion rule, we assume one is that the PE normally will lack expertise in the AE's area, and will therefore lack the competence to review or change the AE's conclusions. That reason may apply here. Because Williams apparently does not understand the role of a confirmation test in determining whether a substance is located on a specimen, he would have been unqualified, on his own, to rephrase Burmeister's AE dictation, and should have reproduced it verbatim in his explosives report.

Williams, however, stated in his OIG interview that Burmeister approved Williams' language [t]wice. Burmeister, however, told us that he did not recall discussing the statement with Williams. In response to a hypothetical question, Burmeister stated:

Given the conflict between Williams and Burmeister, we are unable to say whether Burmeister approved Williams' language. If Burmeister did approve the language, he would have erred for the reasons given in his answer to the hypothetical question.

Williams told us that the reason he wrote the statement as he did was to draw attention to the statement so it could be more fully addressed in court. That purported justification, however, is an insufficient basis for including a self-contradictory statement in a report.

We conclude that either Williams (for writing it) or Burmeister (for possibly approving it), or both, erred by including the statement in the report. The better course would have been for Williams simply to have reproduced the AE dictation verbatim.

The pertinent portion of Burmeister's dictation regarding specimen Q171, an item recovered at a location associated with one of the defendants, is as follows:

Williams reproduced this passage verbatim in a report issued May 24, 1995. But in the section of the September 5, 1995, explosives report concerning ANFO, Williams included only the first paragraph of the passage and replaced the second paragraph with: You should refer to the Laboratory report dated May 24, 1995 for additional information and examination techniques concerning specimen Q171.

Williams gave the following reasons for replacing the second paragraph of Burmeister's dictation:

These reasons do not justify replacing the second paragraph of the dictation. As for the first reason, if Burmeister thought the concentration of the ions had significance, the proper way of calling attention to this finding would have been to add it to the dictation. The second reason implies that one purpose of replacing the second paragraph was to, in effect, hide it so that the first paragraph was not brush[ed] away. This would be improper. If the chemist's conclusion embraces the second paragraph, it is beyond Williams' expertise, and outside his discretion as PE, to discount it.

Williams stated in his OIG interview, however, that Burmeister approved the replacement of the second paragraph. Burmeister did not recall whether or not he approved the replacement. Accordingly, we are unable to confirm Williams' assertion that Burmeister sanctioned the replacement.

During the course of Williams' discussion of Q171, he expressed the following opinion:

These are troubling sentiments. A forensic scientist should make his/her best effort to reach accurate conclusions, regardless of their impact on the prosecution's or defense case. It is up to the prosecution or defense to establish, through expert testimony, the relevance or weight that such data should be accorded.

We conclude that Williams should not have replaced the second paragraph of Burmeister's dictation and that Burmeister erred if he agreed to the replacement. If the concentration of the ions was significant, Burmeister should have amended his dictation to include this finding.

In his January 9, 1996, letter, Whitehurst makes numerous additional allegations, most of which lack merit.

1. Whitehurst claims that on several occasions in Williams' report Williams exceeds his expertise (points 1, 3, 6-12, 15, 19). On each occasion, however, Williams relied on other experts, including AEs, whose dictation he included verbatim.

2. Whitehurst questions Williams' conclusion that none of the structural damage evident within the Murrah building was caused by secondary explosive devices or explosions. Whitehurst asserts that Williams cannot test his hypothesis unless he rebuilds the Murrah building and explodes it again with secondary charges to see whether the damage is different. We reject Whitehurst's assertion that the Murrah building must be rebuilt and destroyed again in order to render a valid opinion about secondary explosive devices. We find reasonable support for Williams' opinion. This included the failure to find damage consistent with another seat of an explosion and the opinions of seismologists who concluded that there was a single blast with ancillary or auxiliary blast effects traveling surface and subsurface.

3. Whitehurst questions Williams' conclusion that tire fragments displayed high explosive damage. Williams told us that there is really unique kind of explosive damage, something that in the courtroom someone could see that this tire wasn't a blowout, it wasn't overinflated. We have not viewed the tire fragments. For Williams' opinion to be valid, it should be supported by empirical studies of high explosive damage to tires or substantial experience with such damage.

4. An AE submitted the following dictation, which was reproduced verbatim in Williams' report:

Whitehurst criticizes this opinion on the ground that the FBI's paint protocol has not been validated.

Although we have not viewed the data or charts underlying the AE's opinion, we have no problem with the opinion's form. In the above passage, the AE stated what he did ( examined [the film] microscopically, microchemically and instrumentally ) and stated that his opinion of a match was [b]ased upon the comparison examinations conducted, which was accurate. Assuming the AE accurately reported his results, his conclusion ( that Q507 could have originated from the box portion of a Ryder truck like that represented by K169 ) does not appear to be overstated.

The FBI's paint protocol should be validated. This could be done through the FBI's own documented tests or by determining that the protocol uses peer-reviewed procedures commonly accepted in the literature or in industry. We were told by James Corby, the former MAU Chief, that at the time of his retirement from the FBI (December 1995) the FBI was in the process of validating the paint protocol for the purpose of ASCLD/LAB accreditation.

Whitehurst asserts:

Corby denied Whitehurst's assertion. Corby thought the FBI could analyze paint pursuant to the protocol and render valid opinions, even though the protocol was not yet validated. In any case, the AE did not state categorically that the two compositions were the same, as Whitehurst contends, but merely that based on specified examinations/comparisons, one could have originated from the other.

5. Williams observed in his report:

Whitehurst contends that Supervisory Special Agent Richard Hahn's testimony about pitting and cratering in the Avianca case (see Section E, supra) constitutes exculpatory evidence relative to the Oklahoma City case. Whitehurst is incorrect. Hahn testified that the pitting and cratering in Avianca indicated a high explosive with a VOD in the area of 20,000 feet per second. Hahn contends that in this testimony he was not asserting that pitting can only occur at that velocity, but only that the unique pitting in Avianca indicated that VOD. It is well-known that pitting and cratering can result from the detonation of a high explosive with a VOD of 10,000 feet per second or more. Hahn's testimony about pitting and cratering in Avianca is irrelevant to the Oklahoma City case.

6. Whitehurst faults Williams for failing to mention a contamination study Whitehurst performed in June 1995 that found PETN contamination in a location in the EU and in an evidence storage area. Certain items of evidence in the Oklahoma City case were examined in the Laboratory in about April or May 1995 and findings of PETN were made. Any connection between the findings and the contamination is at this point speculative. The findings of PETN were made by the AE (Burmeister). We think the AE rather than Williams was the appropriate examiner to determine whether the AE's conclusions concerning PETN should be modified because of the contamination study.

7. Whitehurst criticizes the following passage on page 19 of Williams' report:

Whitehurst contends that [a]s an objective scientist, Williams might also have reported that the ammonium nitrate could be used as a fertilizer and the aluminum powder as an additive to paint.

A couple of lines before the above-quoted passage, however, Williams refers to paint grade aluminum powder. Further, when Williams' report discussed ammonium nitrate in connection with Q507, it included part of Burmeister's dictation, which stated, Crystalline particles removed from specimen Q507 identified the presence of ammonium nitrate. Burmeister's dictation also included the sentence, Ammonium nitrate is considered an oxidizer and is used in some fertilizer and explosive compositions. This sentence was omitted from Williams' September 5th report, but it was included verbatim in an earlier report.

Thus, taking all the reports together, information was provided that aluminum powder could be used in paint and ammonium nitrate in fertilizer. Moreover, the alternative uses for aluminum powder and ammonium nitrate are well-known. Accordingly, although we think generally the better practice is to include, where feasible, common, relevant uses that might reasonably bear on the applicable investigation, and also include the entire AE dictation, in the explosive examiner's report, any error in this aspect of Williams' report was harmless and insignificant.

Along the same lines, Whitehurst criticizes Williams' discussion of binary explosives because Williams fails to mention that nitromethane (recovered in one of the searches) can be used in model airplane engines. Again, although the better practice is to include pertinent alternative uses in the explosives examiner's report, Whitehurst does not suggest, and we are unaware of, any recovered evidence of model airplane engines. Accordingly, any error in this aspect of Williams' report appears to be harmless and insignificant.

8. Although not mentioned by Whitehurst, in the paragraph containing the above-quoted passage concerning aluminum powder, Williams states that [g]round ammonium nitrate was identified in certain specimens. The applicable AE report did not contain the word ground and described the substance as powder. Williams told us that he looked at the specimens under a microscope and could tell that they consisted of broken prills, although he did not document this examination in his casenotes. Williams stated that it was significant that the prills were ground to rebut defendant Nichols' claim that he possessed the ammonium nitrate to sell as plant food.

Assuming Williams in fact made the microscopic examination, we nevertheless think a word more neutral than ground should have been used -- for example, broken prills. Ground implies that someone deliberately converted the prills to a use other than fertilizer, which would be incriminating and would apparently rebut a defense. But all Williams would have been able to see under the microscope was that the prills were broken or were in powder form. He would not have been able to tell whether they were broken or crushed intentionally or accidentally. Further, Williams should have maintained casenotes of his microscopic examination.

9. Whitehurst criticizes the following passage in Williams' report:

Whitehurst asserts that Williams should have noted that the explosives analysis protocol is very limited and may have missed residues from such explosives as PYX, Dinitronapthalene, TATB, DATB, TATP and on and on and on.

The AE dictation, which was reproduced verbatim in an earlier report, states:

The explosives named by Whitehurst are unusual, and we do not think Williams needed to specifically comment on them. However, we think it would have been preferable for Williams to quote the AE dictation verbatim in his explosives report (even though he had already done so in an earlier report) because the dictation states the analytical results more precisely.

At the time of the Oklahoma City bombing, Supervisory Special Agent J. Thomas Thurman was the Unit Chief of the EU. Under FBI Laboratory policies, Thurman as Unit Chief was required to conduct a complete supervisory review of Williams' Oklahoma City report. This review should have included, among other things, [a] review of all PE work notes, graphs, charts, and photographs, and other materials to determine if the examiners conclusions can be supported and have been properly documented.

We conclude that Thurman failed to properly review Williams' report. Thurman's principal failing was to approve a report with unsupported conclusions, with which he disagreed and could not justify. Thurman's approach to his supervisory review was to approve the report if the examiner was comfortable with it. Thurman's OIG interview: as long as he portrays to me that he is in fact comfortable with that, I'm not going to change it. This was an abdication of supervisory responsibilities.

With respect to Williams' specific VOD opinion of 13,000 feet per second, Thurman acknowledged in his OIG interviews (as noted above and in the section on the World Trade Center) that in his long experience in the unit he never rendered such a specific VOD opinion; rendering such opinions is not normally done in the unit; and it is not part of EU training to give a specific VOD opinion. Thurman stated that he would not have included a specific VOD opinion in the Oklahoma City case if he were the examiner. Thurman should not have approved the VOD opinion without data justifying it, which Williams did not have. For example, Williams told us that his opinion was based in part on the size of the recovered pieces of the Ryder truck: if the VOD had been 16,000 feet per second, according to Williams, the pieces would have been half the size. But Thurman told us that such a specific relationship between debris size and VOD was not a commonly held view, and he knew of no data in the EU supporting it. Accordingly, Thurman should not have approved Williams' VOD opinion.

As to Williams' categorical identification of the main charge as approximately 4,000 pounds of ANFO, Thurman thought he could rule out all commercial explosives because in this country with a blast of this size and quantity of explosives we've never seen anything wherein a commercial high explosive was used, and this much of explosive. They've been homemade-type mixtures. We think this is an inadequate ground to eliminate the commercial explosives in total. Unless the commercial explosives could have been eliminated scientifically, they remained a possibility. Moreover, Thurman acknowledged that 2,000 pounds of ANFO and 500 pounds of commercial dynamite could have been used in the blast. Therefore, Thurman should not have approved Williams' opinion as to the weight and identity of the main charge.

As for Williams' opinion that the initiator for the booster was either a detonator from a Primadet Delay system or sensitized detonating cord, Thurman acknowledged in his OIG interview that the proper conclusion was that the named initiators could have been used. As noted, Williams reached a categorical conclusion because Primadet systems were recovered from defendant Nichols' and an accomplice's house. Regarding such a basis for an opinion, Thurman said at the OIG interview:

Accordingly, we conclude that Thurman should not have approved Williams' opinion as to the initiator for the booster.

Regarding other conclusions in Williams' report, Williams said the initiator for the primadet or detonating cord was a non-electric detonator. Thurman acknowledged in his OIG interview that this is a possibility, but said that [i]t may not be the absolute one possibility. And Thurman did not see any reason why it could not have been an electric initiator. Williams also concluded that a non-electric burning-type fuse of either hobby fuse or a commercial safety fuse was used. Thurman acknowledged at his OIG interview that this was one of the possibilities [b]ut certainly not the only one. Further, Williams' conclusion that the time delay was 2 minutes, 15 seconds was just one of many possible scenarios. Accordingly, Thurman should not have approved any of these conclusions.

Additionally, we believe Thurman should have recognized the internal inconsistency in the statement Traces of PETN were located on specimen Q18, however could not be confirmed. He should have required that the sentence be rewritten so that it made sense.

Thurman's method of reviewing Williams' report was deficient. For example, regarding Williams' VOD opinion, Thurman told us in his OIG interview that before approving the report,

And so the specific VOD opinion remained in the report.

This is an improper way for a supervisor to review an examiner's report. The issue for the supervisor should not be whether the examiner is comfortable with the conclusion but whether the supervisor is. It should not be up to him as an examiner to do that, but up to the supervisor to determine whether the conclusions in the report are valid. By leaving the matter to Williams, Thurman failed to perform an appropriate review.

Throughout his OIG interview, Thurman attempted to justify his approval of the report by saying he based it on the conversation he had with Williams about the report, but then pleaded a lack of memory of the details of the conversation. Thus, in the interview he could not justify some of the conclusions in the report, but he implied that Williams provided justification in the conversation although Thurman could not remember what it was. We find this unsatisfactory. Most of the conclusions criticized above came at the end of the report without explanation, and the preceding pages do not support them. For example, at one point Williams stated that ANFO, among other explosives, has the VOD he estimated for the main charge. Then, suddenly, without explanation, he concludes that the main charge was ANFO. If, in fact, Williams provided to Thurman additional reasons justifying his conclusions, Thurman should have required Williams to include those reasons in the report. Thurman acknowledged in his OIG interview that the conclusions in a report should be based on, and flow from, the contents of the report. Thurman, however, failed to make sure that this was done in this case.

Shortly after the Oklahoma City bombing, Martz examined some evidence in the case for explosives residue. The evidence consisted of clothing and a knife seized from one of the defendants. Martz did this as an extra pair of hands while Burmeister was busy outside the Laboratory. Martz's function was to perform instrumental examinations to assist Burmeister who was in charge of the residue analysis. Some issues have arisen concerning Martz's examinations.

A. Supervisory Special Agent Burmeister told us in a 1995 interview that Martz erred in some examinations in the Oklahoma City case. In a 1996 OIG interview, however, Burmeister stated, I don't think he erred in any of these exams. . . . I think he did an acceptable job there. In the later interview Burmeister said that his first interview was based only on Martz's sparse notes and that, between the first and second interview, Martz told Burmeister he did more than what is reflected in the notes. Burmeister told us in 1996, [W]hen I had an opportunity to talk with him [Martz], I didn't find that there was anything wrong with what he did. Burmeister acknowledged in 1996, however, that Martz did not, but should have, followed the provision in the explosives residue protocol that directs examiners to view specimens microscopically before any residue is extracted.

The explosives residue protocol requires the examiner to perform a microscopic examination of the specimen before any residue is extracted from it. Martz failed to do this. He only made a visual examination of the clothing and knife before he vacuumed the former and swabbed the latter. When asked about his failure to perform the microscopic examination, Martz initially told us that the explosives residue protocol does not require microscopic examinations. When we requested a copy of that protocol, Martz stated that [n]o protocol in the Chemistry/Toxicology Unit (CTU) requires any examiner to perform a certain type of analysis, and he failed to send us a copy of a protocol. In a response to the preliminary draft of this Report, Martz stated: My interpretation of visual and microscopic analysis, which was part of the protocol at that time, was that if something was observed by visual examination, that microscopic analysis would be performed and that is what I did in this case. Martz Reply (Feb. 4, 1997) at 3. By this Martz apparently means that because he did not observe anything on the visual examination, he did not do a microscopic examination. Because this is Martz's third explanation for his failure to perform a microscopic examination and is an explanation that is inconsistent with the other two, it lacks credibility. Moreover, Martz's present interpretation of the protocol is unpersuasive, because one of the purposes of the microscopic examination is to detect objects not observed in an unaided visual examination.

As noted, Martz vacuumed the clothing. In his 1995 interview Burmeister, described this as an unqualified technique. In his 1996 interview, however, Burmeister said that vacuuming is a qualified . . . credible technique. As noted by Burmeister, if an examiner finds material through visual or microscopic examination, the examiner should first remove the material with forceps or a scalpel, before vacuuming, to preserve the integrity of the sample and to avoid commingling two different residues. Although there is no reference to extraction techniques in the explosives residue protocol, the techniques are addressed in the training of explosives residue analysts. Martz had not undergone the one-year training in explosives residue analysis.

Regarding the knife, Burmeister said in 1995 that Martz should have rinsed it rather than used a moistened swab. In 1996 Burmeister said that both swabbing and rinsing were permissible, but added that the decision on which technique to use should be based on a microscopic examination of the knife. The protocol does not address the issue of which technique to use.

We conclude that Martz erred by (1) deviating from the protocol requirement of a microscopic examination without first consulting with Burmeister, (2) initially telling us that such an examination is not part of the protocol, and (3) not including all of his examinations in his notes.

B. In a letter to the OIG dated June 18, 1995, Whitehurst stated as follows:

Whitehurst's doubly hedged opinion ( consistent with . . . could have been ) is speculative. Burmeister told us that he was not surprise[d] that he did not find PETN on the knife after Martz had swabbed it. Although Burmeister found nitroglycerine on the clothing after Martz failed to do so, Burmeister told us he performed different examinations. Finally, if there was contamination, it was not necessarily by Martz.

Williams' report contains several serious flaws. His opinion as to the VOD of the main charge was unjustifiable; his statement of the VOD of ANFO was incomplete; his categorical identification of the main charge as ANFO was inappropriate; his estimate of the weight of the main charge was too specific and based in part on improper grounds; his conclusion as to the containers for the main charge was unjustifiably categorical; his categorical identification of the initiator for the booster was improper; his conclusions concerning a non-electric detonator, the fuse, and the time delay were scientifically insupportable; his conclusions were not supported by the contents of the report; and he included some AE dictation in a selective or confusing way. These errors were all tilted in such a way as to incriminate the defendants. We are troubled that the opinions in Williams' report may have been tailored to conform to the evidence associated with the defendants. We conclude that Williams failed to present an objective, unbiased, competent report.

We also conclude that Thurman did not properly review Williams' report. Thurman left too much discretion to Williams to include certain opinions, and Thurman allowed several categorical conclusions to stand, although he told us he does not agree with them, he could not justify them, and the conclusions are not supported in the body of the report. Thurman did not perform the complete supervisory review, as required by the policy of the FBI Laboratory, to determine if the examiners conclusions can be supported and have been properly documented. We are deeply troubled that in a case of this importance and magnitude the EU Chief did not take greater care in making his supervisory review.

As to Martz's examinations, we conclude that he erred by failing to perform the microscopic examination required by the applicable protocol without Burmeister's approval, by initially telling us that such an examination is not part of the explosives residue protocol, and by not including all of his examinations in his notes.

#####

The FBI Laboratory: An Investigation into Laboratory Practices and Alleged Misconduct in Explosives-Related and Other Cases (April,1997) Office of the Inspector General, Department of Justice