SCHALL, Circuit Judge.

Petitioner David L. Dixon petitions for review of an arbitrator's award denying his grievance that he was improperly removed from his position with the Department of Transportation's ("DOT's") Federal Aviation Administration ("agency") for having altered, adulterated, or substituted a urine sample during a random drug test. Because we conclude that the arbitrator's award is not supported by substantial evidence, we reverse and remand.

BACKGROUND

Following his removal, petitioner filed a timely grievance in accordance with the terms of the labor relations agreement between his union and DOT ("Agreement"). After the grievance was denied, the matter was submitted to an arbitrator. The following statement of facts is based upon the evidence before the arbitrator, including testimony taken at a hearing on February 20, and 21, 1992.

At the time of the events in question, petitioner had been employed by the agency for five years as an electronics technician. Award at 1. In that position, he was responsible for maintaining vital equipment and was classified as a Category I employee by the agency. Id. As such, he was subject to random drug testing as a condition of his federal service. Petitioner had previously passed several random drug tests, the most recent one having been on January 14, 1991. Tab A15, Exhibit A at 8.

At about 9:30 a.m. on April 4, 1991, petitioner's supervisor, Leo J. Tabaka, notified him that he had been scheduled for a random drug test later that morning. Award at 2. Petitioner continued in his work routine until about 11:00 a.m., at which time he went to the lunchroom. There, he waited with two other agency employees who also were scheduled for random testing that day. Id. When these employees were unable to give samples at the time they were scheduled to do so, petitioner volunteered to give his sample. Id. There is no evidence in the record that petitioner left the lunchroom or used the drinking fountain or sinks near the lunchroom prior to giving his sample. Id. at 3; Tab A15, Exhibit A at 1-2.

In the sample collection area, petitioner met Kenneth Zimmerman, a specimen collector for a private health care company contractor. Award at 3. After filling out the first part of a drug test custody and control form ("custody and control form"), Mr. Zimmerman accompanied petitioner to a restroom. Once in the restroom, Mr. Zimmerman taped the faucets in the sink and poured a blue dye into the toilet water. Id. Then he directed petitioner to fill at least half of a collection cup, at which point he left the restroom. Id. After he had filled the cup, petitioner knocked on the door to summon Mr. Zimmerman. Id. Mr. Zimmerman accompanied petitioner back into the restroom. Id. Inside the restroom, Mr. Zimmerman asked petitioner to read a temperature strip attached to the collection cup. Id.

Petitioner testified that the temperature of the sample was 92°F, id., and that the sample was "[t]he color of urine." Tr. at 369. At the hearing, Mr. Zimmerman could not remember the exact temperature of the sample. Award at 3. However, Mr. Zimmerman did mark on the custody and control form that the sample's temperature had been read within four minutes of collection and was in the acceptable range of from 90.5° to 99.8°F. Tr. at 111. When asked at the hearing whether he remembered anything about the color of the sample, Mr. Zimmerman testified as follows: "No. I'll tell you. I — it's really hard to say, you know. It just looked okay to me, you know. I can't — had a good temperature. We looked for that, and I, you know — it looked like it was okay to me." Tr. at 116-17. Mr. Zimmerman did not note on the custody and control form anything unusual about the color of petitioner's specimen. Award at 3: Tab A2 at 5. Patrick Moran, a specialist in DOT's Departmental Drug Office, who managed the contract for the collection of urine samples for DOT's drug testing program, Tr. at 10-11, testified that a clear sample would be unusual. Tr. at 55. He also testified that if a specimen collector observed a clear sample, he or she was expected to make a note of such an observation on the custody and control form. Tr. at 55-56.

After observing that the sample's temperature was within the acceptable range, Mr. Zimmerman poured the sample into a specimen bottle and sealed the bottle with an adhesive seal taken from the custody and control form. Award at 3. At the hearing, Mr. Zimmerman demonstrated the sealing of a specimen bottle. Tr. at 112-114. The resulting seal was smooth and unwrinkled. Tr. at 355. At Mr. Zimmerman's request, petitioner initialed and dated the seal. Award at 3-4. Petitioner testified that the seal he initialed on April 4 looked approximately like the demonstration seal and that he did not notice any wrinkles on it. Tr. at 355-56. Mr. Zimmerman then placed the specimen bottle in a small shipping box, and petitioner and Mr. Zimmerman left the restroom. Award at 4.

After leaving the restroom, Mr. Zimmerman and petitioner walked down the hall to an office, where the balance of the custody and control form was completed. Id. Mr. Zimmerman testified that it was his general practice next to insert the custody and control form into the shipping box, and to seal the shipping box for delivery to a drug testing company in the presence of the sample donor. Id. Then, Mr. Zimmerman testified, he would place the sealed shipping box in a larger shipping bag for transfer to an airborne shipping company. Tr. at 138. However, petitioner testified that when he left the office, the custody and control form was still on the desk and had not been sealed in the shipping box with the bottle containing his sample. Award at 4; Tr. at 354.

Block VI of the custody and control form has a space where the person who receives the package containing the sample on behalf of the airborne shipping company is to sign. Tab A2 at 5. In petitioner's case, Mr. Zimmerman signed the custody and control form on behalf of Airborne Express. Id.

On the custody and control form, beneath Mr. Zimmerman's April 4, 1991 signatures releasing custody of the sample and indicating receipt of the sample on behalf of Airborne Express, the next person specifically identified as having custody of the sample is Marilyn Snelling, an employee of CompuChem Laboratories, Inc. ("CompuChem"). Id. Ms. Snelling's signature attesting to receipt of the sample from the airborne shipping company appears next to an April 5, 1991 date stamp. Id. However, Dr. Shirley Brinkley, a CompuChem Senior Technical Analyst, who testified at the hearing concerning CompuChem procedures for handling and testing of the sample, explained that Ms. Snelling was "[n]ot necessarily" the person who took the box containing petitioner's specimen out of the Airborne Express package. Tr. at 229.

According to Dr. Brinkley, the shipping box was removed from the larger shipping bag and opened by an unidentified receiving clerk at CompuChem. Tr. at 156, 229-33. Upon removing the specimen bottle from the shipping box, Dr. Brinkley testified, the receiving clerk noticed that the sample was unusually clear in color, and on the basis of that observation sent the sample directly to a "trouble room" rather than through the normal processing routine. Tr. at 161-62. A trouble room clerk attempted to contact Cynthia Schaeffer, CompuChem's customer service representative for DOT, on April 5. Tr. at 162. Ms. Schaeffer was out of town, however, and did not receive the voice mail notification from the trouble room clerk until four days later on April 9. Id.

After speaking to the clerk in the trouble room on April 9, Ms. Schaeffer, on April 10, contacted Ken Edgell, a DOT manager. Id. Mr. Edgell directed CompuChem to take photographs of the bottle containing the sample and to then perform adulteration panel tests on the sample. Tr. at 162, 166. Rather than photographing the sealed specimen bottle, CompuChem photocopied portions of the seal which showed petitioner's initials. Tr. at 162-63. One of the photocopies shows what appears to be a wrinkle near petitioner's initials. Tab A14, photocopy immediately following custody and control form; Tr. at 355-56. A series of adulteration panel tests performed on April 11 and April 18 revealed that the sample in the specimen bottle was not urine. Tr. at 173-74, 177.

On April 12, petitioner was informed that there was a problem with his drug test. Award at 6. William Horstman, the agency's section manager for the State of Missouri, told petitioner that the agency would give him an opportunity to be retested on April 15. Tr. at 302. Mr. Tabaka testified at the hearing that petitioner told him that he wanted to be retested. Tr. at 90; Tab A15, Exhibit A at 5. On April 15, however, Mr. Tabaka informed petitioner that the agency would not retest him because it was agency policy not to retest when there was a substitution of a sample. Tr. at 359-60. When petitioner learned that he was not going to be retested by the agency, he made his own arrangements to be tested on April 15 by a local independent testing facility, Chem-Tox, Inc. ("Chem-Tox"). Tr. at 361-62. The results of the Chem-Tox test were negative. Id.

At the meeting on April 12, in which petitioner was informed that there was a problem with his drug test, he was asked if he would be willing to have his specimen retested. Tr. at 359. At the hearing, petitioner testified: "I didn't think it would help me, because it it wasn't — if the sample wasn't urine, it wasn't the sample that I submitted, and I did not see how it would help, so I didn't — I really didn't agree to have it retested at that time." Tr. at 359-60. Subsequently, on April 19, Robert Lindsey, an agency manager, recommended to petitioner that he have the CompuChem sample retested by another laboratory. Tab A15, Exhibit A at 6. Petitioner testified that he sought to have the CompuChem sample retested at St. Louis University, but that a Dr. Long of that institution was uncooperative with him. Tr. at 363-65. Petitioner testified that in a telephone conversation he had with Dr. Long, Dr. Long stated in response to his question regarding the color of the CompuChem sample that it was "[p]ale yellow, clear what's the difference?" Tab A15, Exhibit A at 8; Tr. at 366. Petitioner did not attempt to have a test done at any other facility. Award at 7.

Mr. Moran testified that DOT policy requires that a supervisor initiate a removal action if a Category I employee is shown to have altered, tampered with, or substituted a urine specimen. Tr. at 21. Later in his testimony, Mr. Moran reiterated that, "[b]ecause of the importance of this program within the Department of Transportation, . . . it was determined that [for] anyone who would subvert the program would — the appropriate penalty should be removal." Tr. at 23.

Acting under that policy, Mr. Horstman issued petitioner a notice of proposed removal on May 6, 1991. Tab A1. The notice charged petitioner with "[a]lteration, adulteration, or substitution of a urine specimen . . ." Id. Mr. Horstman testified that petitioner was a good employee. Tr. at 295-96. He felt, however, that the agency's policy created an overriding consideration, overshadowing all other mitigating factors. Tr. at 297. By letter dated September 23, 1991, Mr. Horstman informed petitioner that, having considered the material submitted by petitioner in opposition to the proposed removal, as well as other material, he had made the decision to remove petitioner. Tab A17. After petitioner's grievance of the removal action was denied, the matter was submitted for arbitration.

At the hearing, petitioner denied altering, adulterating, or substituting the urine sample that was submitted to Mr. Zimmerman. Tr. at 378-79. In addition, petitioner submitted character affidavits from fellow agency employees William Leingang, Wilbert Brewton, Steven Oates, Farnsworth Dye, and James Turner. Each of these individuals stated in his affidavit that he had never known petitioner to use drugs and that he had always known petitioner to be truthful and honest. Tab A15, Exhibits B-F. Most of these character witnesses were available to testify, but agency witnesses themselves testified that petitioner was a good worker and that they had no reason to doubt his honesty. The arbitrator remarked:

Manager Tabaka had a good opinion of the Grievant and had never seen him with any drugs. He was very surprised at the accusation and thought that a mistake had been made. Manager Horstman also stated that the Grievant was a dedicated individual who did a good job. No management official had any reason to doubt the Grievant's honesty and they all felt he was going to be a good Technician.

Award at 7.

Petitioner also submitted the July 5, 1991 affidavit of Arthur S. Hume, M.D., a forensic toxicologist in the Department of Pharmacology and Toxicology at the University of Mississippi Medical Center. Tab A15, Exhibit G. In his affidavit, Dr. Hume stated that he had received the packet which petitioner's attorneys had told him was prepared by CompuChem in connection with the April 4 drug test. Id. at 2. In his affidavit, Dr. Hume declared:

[I]t is virtually impossible for an individual subject to standard drug testing procedures to substitute tap water or any other liquid for his own urine within the precise temperature range (90.5°-99.8°F) attributed to the sample at item V on Copy 1 of the Drug Testing Custody and Control Form. This temperature range reflects the temperature of freshly voided urine. In order to successfully substitute under the standard drug testing procedures another liquid for urine and have its temperature within a 90.5°-99.8°F temperature range, an individual would have to obtain incredibly sophisticated equipment, such as surgical tubing and an enema bag similar to those used to evacuate the lower colon, and then attach the apparatus under his arm. Even then, however, it is extremely difficult to produce a substitute within a 90.5°-99.8°F temperature range.

Id. at 3.

Finally, Lillian Gasway, an electronics technician for the agency, Tr. at 338-39, testified that during a random drug test she had observed a collector removing and replacing a seal without breaking the seal. According to Ms. Gasway, the collector was able to remove the seal from a sealed shipping box in order to place a custody and control form in the shipping box which had inadvertently been left out when the box was originally sealed. Tr. at 339-40.

On May 14, 1992, the arbitrator issued his award denying the grievance and sustaining the agency's decision to remove petitioner. This appeal followed.

DISCUSSION I [25] THE STANDARD OF REVIEW

On appeal, petitioner contends that the arbitrator's award is not supported by substantial evidence and that the agency abused its discretion in imposing the penalty of removal.

Under the provisions of Article 6, section 2, of the Agreement, the agency was required to prove before the arbitrator "by a preponderance of the evidence" that petitioner had committed the offense with which he was charged. Award at 12.

The arbitrator's award sustaining petitioner's removal is reviewed by this court under 5 U.S.C. § 7121(f) (1988). That statute provides that "section 7703 of this title pertaining to judicial review shall apply to the award of an arbitrator in the same manner and under the same conditions as if the matter had been decided by the [Merit Systems Protection] Board." That means that the arbitrator's award must be affirmed unless it is found to be:

(1) arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with the law;

(2) obtained without procedure required by the law, rule, or regulation having been followed; or

(3) unsupported by substantial evidence.

5 U.S.C. § 7703(c) (1988); Grigsby v. Department of Commerce, 729 F.2d 772, 774 (Fed. Cir. 1984).

The critical issue in this case is whether the arbitrator's conclusion that the agency proved the charge against petitioner by a preponderance of the evidence is supported by substantial evidence. See Bradley v. Veterans Admin., 900 F.2d 233, 234 (Fed. Cir. 1990); Jackson v. Veterans Admin., 768 F.2d 1325, 1330-32 (Fed. Cir. 1985); see also Envirotech Corp. v. Al George, Inc., 730 F.2d 753, 758, 221 USPQ 473, 477 (Fed. Cir. 1984). The Supreme Court has defined the "substantial evidence" standard as follows:

"[s]ubstantial evidence is more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Consolidated Edison Co. v. Labor Board, 305 U.S. 197, 229[, 59 S.Ct. 206, 216, 83 L.Ed. 126]. Accordingly, it "must do more than create a suspicion of the existence of the fact to be established. . . . it must be enough to justify, if the trial were to a jury, a refusal to direct a verdict when the conclusion sought to be drawn from it is one of fact for the jury." Labor Board v. Columbian Enameling Stamping Co., 306 U.S. 292, 300[, 59 S.Ct. 501, 505, 83 L.Ed. 660].

* * * * * *

Whether or not it was ever permissible for courts to determine the substantiality of evidence supporting a Labor Board decision merely on the basis of evidence which in and of itself justified it, without taking into account contradictory evidence or evidence from which conflicting inferences could be drawn, the new legislation definitely precludes such a theory of review and bars its practice. The substantiality of evidence must take into account whatever in the record fairly detracts from its weight.

Universal Camera Corp. v. NLRB, 340 U.S. 474, 477, 487-88, 71 S.Ct. 456, 459, 463-64, 95 L.Ed. 456 (1951).

Because "[t]he substantiality of evidence must take into account whatever in the record fairly detracts from its weight," we must canvass the entire record. Spurlock v. Department of Justice, 894 F.2d 1328, 1330 (Fed. Cir. 1990). In so doing, we are to bear in mind that "[e]xaggeration, inherent improbability, self-contradiction, omissions in a purportedly complete account, imprecision, and errors detract from the weight to be accorded the evidence upon which an administrative board bases its decision." Id.

It also is important to recognize what our substantial evidence standard of review does not involve. Specifically, we do not make a de novo decision on the weight of the evidence. De Cicco v. United States, 677 F.2d 66, 70 (Ct.Cl. 1982). Rather, we look at the evidence which supports the arbitrator's award and then review the entire record to determine whether factors such as "[e]xaggeration, inherent improbability, self-contradiction, omissions in a purportedly complete account, imprecision, and errors" detract from the weight of that particular evidence. Spurlock, 894 F.2d at 1330. Only if such factors so detract from the weight of the evidence which supports the arbitrator's award, or the agency's evidence is so sparse, that a reasonable fact-finder would not find the charge proved by a preponderance of the evidence may we reverse. See Jackson, 768 F.2d at 1330.

The arbitrator's award states that "[t]he key to this case is the question of whether the sample that was given by the Grievant was the sample that reached CompuChem and whether, when the adulteration panel was performed, it was [performed] on the same sample that the Grievant gave." Award at 13. We agree with the arbitrator's characterization of the issue as whether petitioner altered his sample or placed a substitute liquid in the specimen bottle. However, having reviewed the record, we hold that the arbitrator's conclusion that the agency proved by a preponderance of the evidence that petitioner "altered or substituted the sample." Award at 17, is not supported by substantial evidence.

II [35] ANALYSIS

A. Evidence Supporting the Arbitrator's Award

The evidence supporting the arbitrator's award consists of the chain of custody presented by respondent to establish as fact that the liquid tested by CompuChem could have come from no source other than petitioner. The chain of custody begins with Mr. Zimmerman's collection of petitioner's sample and ends with the CompuChem report indicating that the sample attributed to petitioner was not urine.

Mr. Zimmerman testified that he received a liquid sample from petitioner on April 4, 1991. He also testified as to how he and petitioner completed the custody and control form. In addition, Mr. Zimmerman testified that it was his general practice to insert the custody and control form into a shipping box along with the bottle containing the donor's sample and to then seal the shipping box (those acts taking place in the donor's presence). Mr. Zimmerman added that later he would place the sealed shipping box in a larger shipping bag for transfer to an airborne shipping company.

Mr. Zimmerman's testimony is supported in part by the custody and control form which bears his signature and that of petitioner. It also is supported in part by petitioner's testimony that he gave Mr. Zimmerman a liquid specimen in a bottle and that he and Mr. Zimmerman completed the custody and control form. The body of evidence just described supports the proposition that petitioner submitted a liquid specimen which was packaged for shipment to CompuChem.

Respondent also relies upon the evidence consisting of the records of CompuChem and the testimony of Dr. Brinkley. Based upon the custody and control form and various CompuChem documents, Dr. Brinkley testified as to CompuChem's chain of custody procedures in general and as to CompuChem's receipt of a bottle with a label on it bearing petitioner's initials and the date April 4, 1991. She also testified that an unidentified receiving clerk at CompuChem removed the specimen bottle from the shipping box, noticed that the liquid in the bottle was unusually clear in color, and, on the basis of that observation, sent the sample directly to a "trouble room" rather than through the normal processing routine. Dr. Brinkley further testified as to discussions between Cynthia Schaeffer. CompuChem's customer service representative for DOT, and Ken Edgell as to what to do with the sample. Finally, Dr. Brinkley testified that adulteration panel tests were run on the specimen and that those tests revealed that the liquid in the bottle was not urine. The body of evidence just described tends to establish that a bottle was received at CompuChem with a label on it bearing petitioner's initials and the date April 4, 1991, that the fluid in the bottle was tested, and that the test revealed that the fluid in the bottle was not urine.

Standing alone, the above evidence may fairly be said to support the conclusion that petitioner committed the offense with which he was charged. To be sure, there is no direct evidence that petitioner altered his urine sample or substituted another fluid for that sample; respondent's case is circumstantial. However, an agency may carry its burden by presenting a case based upon circumstantial evidence. Naekel v. Department of Transp., 782 F.2d 975, 978 (Fed. Cir. 1986). Were the above evidence the only evidence in the record, we would have no difficulty affirming the arbitrator's award. It is not the only evidence, however, and as explained above, in determining whether the arbitrator's award is supported by substantial evidence, we may not view the evidence supporting the award in isolation, but also must consider whatever in the record detracts from that evidence. It is to the balance of the record that we now turn.

B. Other Evidence

Testimony of petitioner and Mr. Tabaka establishes that petitioner did not know until 9:30 a.m. on April 4, 1991, that he was to receive a random drug test. The only evidence in the record concerning petitioner's movements and whereabouts during the period between 9:30 and approximately 11:00, when petitioner volunteered to give his sample, is petitioner's unchallenged testimony. Petitioner testified that, after being told about the random drug test, he continued at work until approximately 11:00, at which time he went to the lunchroom, and from there to the testing area. There is, in the record, no counter version concerning petitioner's movements and whereabouts between 9:30 and 11:00 and no evidence that petitioner left the lunchroom or used the drinking fountain or sink near the lunchroom prior to entering the testing area.

We turn next to Mr. Zimmerman's collection of petitioner's sample. In view of the testimony of petitioner and Mr. Zimmerman, as well as the custody and control form, it is undisputed that (1) Mr. Zimmerman accompanied petitioner into the restroom, where he taped the water faucets in the sink and poured blue dye into the toilet; (2) the temperature strip attached to the sample collection cup revealed a temperature for petitioner's sample within the acceptable range of 90.5°-99.8°F; and (3) Mr. Zimmerman, who was expected to put a notation on the custody and control form if he observed a colorless sample, made no such notation and testified that the sample "looked okay" to him.

As seen in section II A above, the starting point for respondent's case is the evidence of Mr. Zimmerman's collection of petitioner's sample. In that regard, it goes without saying that a necessary premise of any conclusion that the arbitrator's award is supported by substantial evidence is the proposition that what Mr. Zimmerman received from petitioner was an altered or substituted specimen. Put another way, the sine qua non for the government's chain of evidence is the starting point of an altered or substituted specimen. If that starting point is not established, there is no case, let alone a case supported by substantial evidence. The evidence described above in this section is such that, in view of the record in this case, it renders inherently improbable the existence of the essential starting point for any determination that the arbitrator's award is supported by substantial evidence — receipt by Mr. Zimmerman of an altered or substituted specimen.

The evidence does so for two reasons. First, at the critical point, when petitioner provided his sample, there is no evidence of tampering, adulteration, or substitution. Indeed, the record establishes that Mr. Zimmerman undertook steps to make that impossible (pouring dye in the toilet, taping the faucets). Second, and most importantly, the fluid which petitioner presented to Mr. Zimmerman not only appeared "okay" to Mr. Zimmerman, but also tested within the proper temperature range. This second point, which the arbitrator did not discuss in his decision, is of particular significance.

In his affidavit, Dr. Hume stated that it would be "virtually impossible" for someone substituting water or other fluid to present it within the precise temperature range to enable it to pass for urine, because one would need to come to the test equipped with an "incredibly sophisticated" apparatus attached under his arm. Tab A15, Exhibit G at 3. There is no evidence that petitioner entered the testing area equipped with any kind of apparatus. At the same time, between 9:30 and 11:00, petitioner did not have the opportunity to obtain the kind of apparatus described by Dr. Hume.

In sum, the evidence of record concerning the events of April 4, 1991, is to the effect that (1) the sample petitioner gave Mr. Zimmerman was urine (color "okay" and a temperature reading within the range of 90.5° and 99.8°F) and (2) petitioner did not undertake the measures necessary to provide an altered or substituted specimen (no evidence that those measures were undertaken and evidence of lack of opportunity to undertake them, since petitioner had no chance between 9:30 and 11:00 to obtain and set up the kind of apparatus described by Dr. Hume and since Mr. Zimmerman taped the faucets of the sink and poured blue dye into the toilet).

The proposition that the arbitrator's award is supported by substantial evidence also depends upon respondent's chain of custody evidence. That evidence, in turn, rests largely on two premises: (1) that the seal on the specimen bottle was tamper proof; and (2) that the handling procedures used by CompuChem in transporting the specimen bottle were sufficient to impute to petitioner sole responsibility for the contents of the bottle. We do not find either of these premises to be supported by substantial evidence. Regarding the first premise, in the arbitrator's view, "the key point is that the collection bottle was sealed with the Grievant's signature thereon. These are tamper-proof seals." Award at 16.

Our review of the record has revealed two instances of testimony discussing the quality of the seal in question.

Q. Mr. Zimmerman, I'd like to go back just a minute to when the bottle is sealed, and the sealing strip put across it. What happens if someone tries to remove that strip? What does it look like? What does it do?

A. Well, I really don't know. I haven't seen any removed, but it — it's supposed to be a tamper-proof seal, and being familiar with the police department, when you remove a tamper-proof seal, it just about destroys the whole seal. We've had them there quite similar, but it is a tamper-proof seal, and, you know, certainly if someone opens it or something, there surely should be some damage there to the seal.

Tr. at 123 (emphasis added).

Q. [To Dr. Brinkley] Does the seal on that bottle appear other than when it was cut, appear to be tampered with in any way?

A. No, the seal does not appear to be tampered with.

Q. What would you notice if the seal had been tampered with?

A. I believe this particular seal sort of comes apart, and you can't really put it back the way it was before. I'm not as familiar with this particular one from the D.O.T. as I am from the ones that we prepare.

Tr. at 173 (emphasis added). Thus, neither Mr. Zimmerman nor Dr. Brinkley could testify from personal knowledge concerning the seal. We find no other testimony or demonstration in the record to substantiate the quality of the seal. We are left with the impression that the arbitrator accepted as "fact" the agency's general assertion that the seal was tamper-proof.

The record, however, contains evidence which detracts from the testimony of Mr. Zimmerman and Dr. Brinkley supporting the arbitrator's conclusion regarding the tamper-proof nature of the seal. As already seen, at the hearing, a demonstration by Mr. Zimmerman produced a specimen bottle having a smooth, rather than a wrinkled seal, and petitioner testified that the seal he initialed on April 4 was not wrinkled. Yet, one of the photocopies of the specimen bottle shows a wrinkle on the seal next to petitioner's initials. In addition, Ms. Gasway testified that she had observed similar seals used to secure the shipping box being opened and replaced by a collector without breaking or destroying the seal.

As far as the second premise underlying respondent's evidence is concerned, the arbitrator acknowledged in his award that the chain of custody in the case was "not absolutely pristine. . . ." Award at 15. He concluded, however, that the chain of custody had been proven "in all aspects by a preponderance of the evidence." Award at 16.

The arbitrator's award notes petitioner's testimony that Mr. Zimmerman failed to seal the shipping box in his presence and evidence that the previously smooth seal subsequently appeared wrinkled in the photocopies. The award also notes the fact that no one from Airborne Express signed for the receipt or transfer of the shipping bag containing petitioner's specimen bottle.

In addition, neither the testimony of Dr. Brinkley nor the custody and control form identifies the receiving clerk who made the critical observation regarding the color of the sample which was later tested and who made the decision to send the sample to the trouble room. As noted above, according to Mr. Moran of DOT, a clear sample should have been considered "unusual" and the fact of such a sample noted on the custody and control form. Mr. Zimmerman, a trained collector who observed petitioner's sample on April 4, made no such notation on the form and testified that the sample looked "okay" to him. Yet, the sample attributed to petitioner was immediately recognized as a problem by a receiving clerk at CompuChem the following day. Plainly, there is a discrepancy between Mr. Zimmerman's observation of the sample and that of the receiving clerk, whose skill and training the record does not disclose.

As far as the color of the sample is concerned, there are two possible explanations for the discrepancy between Mr. Zimmerman's observation and the apparent observation of the CompuChem receiving clerk. One explanation favors respondent; the other explanation favors petitioner. The explanation that favors respondent is that Mr. Zimmerman's observation of the sample's color in the restroom was not as careful as that of the receiving clerk at CompuChem and was inaccurate. The explanation that favors petitioner is that both Mr. Zimmerman and the receiving clerk made careful and accurate observations but that they were observing different samples. The arbitrator concluded that the discrepancy between the two observations should be resolved in favor of respondent. However, this conclusion is unsupported by substantial evidence, since Mr. Zimmerman testified that he used proper collection procedures which specifically required an observation of the sample's color and the unidentified receiving clerk did not testify at all.

CONCLUSION

A petitioner who challenges the substantiality of the evidence supporting an arbitrator's award faces a heavy burden. In the unusual circumstances of this case, however, that burden has been carried. We "cannot consciously find that the evidence supporting the [arbitrator's award] is substantial, when viewed in the light that the record in its entirety furnishes, including the body of evidence opposed to the [arbitrator's] view." Universal Camera, 340 U.S. at 488, 71 S.Ct. at 464. As a result of the contradictory evidence in the record before us, the evidence supporting the arbitrator's award thins to the point that it must be considered insubstantial. Accordingly, the arbitrator's award is reversed and the case is remanded with instructions that petitioner's removal be cancelled; that petitioner be reinstated to his former position retroactive to the date of his removal; and that petitioner be awarded backpay and other relief to which he may be entitled as a result of this decision.

Costs to petitioner.

REVERSED AND REMANDED.