Appellant challenges an order of the National Transportation Safety Board (NTSB) which denied him first-class airman medical certification in 1979. Jurisdiction is proper pursuant to 49 U.S.C. §§ 1486, 1903. The Air Line Pilots Association, International, has filed an amicus curiae brief in support of appellant’s position. A full discussion of the facts and the administrative hearing is necessary to address the various claims.
I
Clinton McHenry has been a pilot for 39 years and has accumulated over 28,000 hours of flight time. For the last 23 years he has served as a captain for Eastern Airlines. In addition to flying as a commercial pilot, McHenry has participated in aerobatic contests and has established an international reputation as an aerobatic pilot. He has won more aerobatic contests than any other pilot in United States history-
On January 24,1977, McHenry suffered a strange illness which resulted in the revocation of his airman medical certification.1 At 1:00 o’clock in the morning, while landing an Eastern passenger flight in Tampa, he experienced a severe pneumatic headache caused by the rapid decompression that accompanies aircraft descent. McHenry relinquished control of the plane to his co-pilot, Randall Gibson, who completed the landing. Once on the ground, McHenry again took control of the plane, navigated it to the ramp, and parked it. Thereafter, he began to experience a period of amnesia and confusion. Gibson took him to a hospital where the amnesia continued for several hours. A cardiologist and a neurologist evaluated McHenry at the hospital and conducted various tests, including an isotope scan of the brain, an electrocardiogram and an electroencephalogram, all of which produced normal results. McHenry also gave the doctors a verbal medical history which included descriptions of three similar past experiences. The next day McHenry was released from the Tampa hospital and returned home to Miami where he was admitted to another hospital for further examination under the supervision of Dr. Ray Lopez, a neurologist. McHenry gave Dr. Lopez the same medical history and again described the three similar past experiences. After additional extensive testing and evaluation, Dr. Lopez rendered a final diagnosis of Transient Global Amnesia (TGA) and recommended to Eastern that McHenry be returned to full flying status in March of 1977. However, the Federal Aviation Administration (FAA) denied appellant’s application for recertification the following September. McHenry received a final denial of first-class airman medical certification from the FAA in October of 1978. The Federal Aviation Administrator concluded that McHenry had experienced a “disturbance of consciousness without satisfactory explanation of the cause” as contemplated by 14 C.F.R.'§ 67.13(d)(2)(i)(b), and therefore, a denial of medical certification was appropriate.2 McHenry then filed a petition for review with the NTSB. The case was referred to an administrative law judge who commenced evidentiary hearings on the matter in February of 1979.
During the proceedings before the administrative law judge, McHenry argued that he had experienced an isolated episode of TGA which constitutes a satisfactory medical explanation within the meaning of 14 C.F.R. § 67.13(d)(2)(i)(b). The Administrator responded that McHenry had experi*1188enced a total of four disturbances of conciousness, including the three described by McHenry to his treating physicians, and therefore the episode was not an isolated manifestation of TGA. The Administrator argued alternatively that even if McHenry’s illness of January 24, 1977, was an isolated episode of TGA, a diagnosis of TGA does not provide a satisfactory medical explanation within the meaning of the regulations.
Appellant presented numerous experts and other witnesses at the hearing. Dr. Lopez, McHenry’s treating physician in Miami, testified first. Dr. Lopez is a Board-certified neurologist with 15 years experience. He described TGA as a widely recognized but rare syndrome, first documented in 1957, which causes the victim to suffer total amnesia for a period of several hours. After the amnesia has ended, Dr. Lopez explained, the victim is left as before but with no memory of events during the attack. The episode leaves the victim with no damage or impairment that can be detected by modern medical science. Although the exact cause or triggering mechanism of the syndrome is unknown, recurrences are rare according to Dr. Lopez. He has treated between 25 and 30 patients with a TGA diagnosis. It is Dr. Lopez’s opinion that McHenry experienced an isolated episode of TGA on January 24, 1977. His opinion is based on his own experiences with TGA victims, the remarkable similarity between McHenry’s symptoms and classic cases of TGA as described in literature, and the failure of scores of medical tests to produce any other explanation. Dr. Lopez was also questioned about the reference to three previous amnesic incidents in the medical history given to him by McHenry and used as evidence at the hearing. He explained that McHenry was very anxious to aid the doctors with any relevant information. Accordingly, McHenry related three episodes where he had forgotten something. The first occurred in 1964 and involved McHenry forgetting the address of a house while en route to the house through an unfamiliar neighborhood. The incident bothered McHenry and he returned home to his wife and asked her to question him about various matters in order to test his memory. He answered all of her questions perfectly. The second incident occurred just before an aerobatic competition in 1971. After drinking several cups of coffee, which is unusual for McHenry, he suddenly realized that he was unsure of the names of several people around him who were casual acquaintances. He then asked his son to quiz him in the same fashion that he had asked his wife several years earlier. He correctly answered all of the questions, but still decided not to fly in the competition. Finally, the third episode occurred in 1975 and involved McHenry forgetting the maiden name of a recently divorced woman. Dr. Lopez interprets all three of these incidents as medically insignificant normal losses of memory which are common to everyone.
The next witness on behalf of appellant was Dr. Irving Fosberg, a psychologist who at one time specialized in aviation psychology. Dr. Fosberg has repeatedly examined McHenry and administered over a dozen psychological tests. Each of the tests produced normal results. Dr. Fosberg described McHenry as a zealously precise person, a perfectionist who prides himself on his memory. In his opinion, McHenry’s first three episodes of memory loss were just that, a temporarily forgotten name or address, while the later episode on January 24, 1977, was something more, an actual disturbance of consciousness. Dr. Fosberg testified that McHenry thought the three earlier experiences were similar to the later episode only because they all involved a failure of his prized memory.
Dr. Emil Taxay, a Board-certified internist specializing in aviation medicine, also testified on behalf of McHenry. Dr. Taxay estimated that he has examined several thousand pilots during his career. He described McHenry as a perfectionist who overemphasized the significance of the three earlier losses of memory. He also concurred in the diagnosis of TGA for McHenry’s amnesic episode in January of 1977. It is his opinion that McHenry is medically competent to resume flying as a commercial pilot.
*1189A number of other persons, including McHenry’s wife and son, testified as eyewitnesses with respect to the nature and extent of the three earlier episodes of memory loss. Each person described the episodes as mere losses of memory with which everyone is familiar.
The last witness to appear for McHenry was Dr. Charles Fisher, a professor of neurology at Harvard University. Dr. Fisher credits himself with being one of the first doctors to recognize TGA. He has treated over 75 patients with a diagnosis of TGA, and has published numerous articles and lectured internationally on the subject. He examined McHenry several times in 1977 and 1978. It is his opinion that McHenry’s three earlier memory losses were medically insignificant events. He concurs in a diagnosis of TGA for the final episode in January of 1977. The underlying cause of TGA is still unknown according to Dr. Fisher, however he believes that severe pain from a pneumatic headache may have acted as a triggering mechanism in McHenry’s case. He stated that TGA itself is not uncommon, although recurrences of TGA are extremely rare. He has seen one recurrence in the dozens of TGA victims he has treated, and this recurrence took place ten years after the initial episode. He considers McHenry to be in excellent health.
The only witness on behalf of the Administrator was Dr. Harold Stevens, a Board-certified neurologist. Dr. Stevens has never examined McHenry but has' reviewed all of the relevant medical records and was present throughout the hearings. He believes that McHenry has experienced four disturbances of consciousness. He stated that McHenry’s perfectionist character would cause him to relate an accurate medical history to a physician, and thus, McHenry’s description of three similar episodes is very significant and should not be labeled mere losses of memory. The fact that McHenry can specifically recall these episodes of memory loss is also significant to Dr. Stevens. He stated that most people would not remember the circumstances surrounding normal forgetfulness, but they would remember a more serious incident such as an amnesic disturbance of consciousness. Accordingly, Dr. Stevens stated that he would not diagnose McHenry’s fourth episode as TGA because TGA characteristically is an isolated occurrence. He admitted, however, that if it were not for the three previous episodes, McHenry’s illness in 1977 would fit the classic pattern of TGA. He also testified that some experts believe recurrence is common, although he believes recurrence is rare. He carefully followed 11 cases for a substantial period of time, and none of the patients experienced a recurrence. Unlike Dr. Fisher, Dr. Stevens believes that TGA itself is rare. He agrees with Dr. Fisher that the underlying cause of TGA is still unknown.
On June 12, 1979, the administrative law judge issued an initial decision in favor of appellant McHenry. He concluded that McHenry had experienced one “disturbance of consciousness” in January of 1977, and that the three earlier episodes were medically insignificant events. Accordingly, he found that McHenry had suffered an isolated episode of TGA. The law judge then followed NTSB precedent, Petition of Weddle, 1 NTSB 1933 (1972), in holding that a diagnosis of TGA provides a satisfactory explanation for an isolated disturbance of consciousness. Thereafter, the Administrator filed a notice of appeal with the NTSB. On December 12,1979, four NTSB members unanimously reversed the administrative law judge’s initial decision and affirmed the Administrator’s order denying McHenry’s airman medical certification.3 Nine months later, the same four NTSB members evenly divided over several key issues on a petition for reconsideration of the case. A split Board on reconsideration, however, is insufficient to reverse the Board’s original decision, and McHenry now appeals to this court. He primarily claims that the original decision of the NTSB is irrational, arbi*1190trary, and not supported by substantial evidence. He also argues that the Administrator’s notice of appeal to the NTSB was untimely filed, and therefore the initial decision of the administrative law judge is binding. For the following reasons, we agree with appellant on the first claim, but disagree on the latter.
II
In reviewing the decision of the NTSB, we are bound by the narrow standards of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A, E): “The reviewing court shall hold unlawful and set aside agency action, findings, and conclusions found to be arbitrary, capricious .. . [or] unsupported by substantial evidence.. . . ”
Under the “arbitrary and capricious” standard the scope of review is a narrow one. A reviewing court must “consider whether the decision was based on a consideration of the relevant factors and whether there has been a clear error of judgment. . . . Although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not empowered to substitute its judgment for that of the agency.”
Bowman Transportation, Inc. v. Arkansas-Best Freight Systems, Inc., 419 U.S. 281, 285-6, 95 S.Ct. 438, 441-442, 42 L.Ed.2d 447 (1974) (citations omitted). In other words, when reviewing an agency decision under the “arbitrary and capricious” standard, we must defer to the wisdom of the agency provided their decision is reasoned and rational. See Watkins Motor Lines, Inc. v. ICC, 641 F.2d 1183 (5th Cir. 1981). The substantial evidence test is also a narrow standard of review. Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Refrigerated Transport Co., Inc. v. ICC, 616 F.2d 748, 751 (5th Cir. 1980). It is something more than a scintilla of evidence, but something less than the weight of the evidence; the “possibility of drawing two inconsistent conclusions from the evidence does not prevent an administrative agency’s findings from being supported by substantial evidence.” Consolo v. Federal Maritime Commission, 383 U.S. 607, 620, 86 S.Ct. 1018, 1026, 16 L.Ed.2d 131 (1966). Of course, when reviewing the evidence supporting the agency’s findings, we must review the record as a whole, including the administrative law judge’s initial decision, and consider portions of the record which reasonably detract from the supporting evidence. Universal Camera Corp. v. NLRB, 340 U.S. 474, 71 S.Ct. 456, 95 L.Ed. 456 (1950); Nadiak v. CAB, 305 F.2d 588 (5th Cir. 1961), cert. denied, 372 U.S. 913, 83 S.Ct. 729, 9 L.Ed.2d 722 (1963).
With these factors in mind, we have reviewed the entire record of this case and the decision of the NTSB. As always, we had hoped to find substantial evidence to support the Board’s factual conclusions and a reasoned application of the FAA regulations to those facts. We found, instead, a decision which is lacking in a sound rationale.4 Portions of the decision are not supported by substantial evidence. The Board’s ultimate finding on a central issue is vague and unclear. The Board has failed to rationally explain its departure from NTSB precedent on another issue. Accordingly, we are compelled to vacate the order of the NTSB and remand for further proceedings. In order to best discuss and explain the errors in the Board’s decision, we will start at the beginning and attempt to logically progress through their faulty reasoning.
The Federal Aviation Administrator denied McHenry airman medical certification because of alleged “disturbance^] of consciousness without satisfactory medical explanation of the cause.”5 McHenry re*1191sponded that he had suffered an isolated episode of TGA which the NTSB previously had found to be a satisfactory medical explanation. There is no question that McHenry at all times carried the burden of proving he was entitled to medical certification under the FAA regulations. Day v. NTSB, 414 F.2d 950 (5th Cir. 1969). Likewise, there seems to be no doubt that McHenry suffered a disturbance of consciousness in January of 1977.
Obviously, one of the first factual issues to be decided is the number of disturbances of consciousness that McHenry has experienced. The administrative law judge found that McHenry experienced an isolated disturbance of consciousness in January of 1977, and that the three earlier episodes were medically insignificant events. Record at 634. The NTSB apparently agreed with this portion of the law judge’s initial decision:6 “[W]e do not have enough information concerning those [three earlier] occurrences to identify them as disturbances of consciousness and we have given them no weight in reaching our decision.” Record at 701, n. 11. Indeed, we find that there is not substantial evidence in the record to reach any other conclusion. All medical testimony in the record describes a disturbance of consciousness as some impairment in a person’s ability to make judgments and be cognizant of his surroundings. There is no evidence that McHenry’s three earlier episodes fit this description. Numerous eyewitnesses described McHenry’s behavior during these experiences as completely normal. McHenry described the episodes as normal forgetfulness, similar to the later episode only in that they all involved a degree of memory loss. The only evidence in support of the Administrator’s position was the testimony of Dr. Stevens who had never examined McHenry and was merely interpreting medical histories taken by other doctors. One of these doctors was Ray Lopez, who testified in support of McHenry’s position and contrary to the Administrator’s position. Therefore, we believe the only rational conclusion is that McHenry experienced an isolated disturbance of consciousness in January of 1977.
We must now discuss whether this isolated disturbance of consciousness was an attack of TGA. The administrative law judge found that TGA was the proper diagnosis. Record at 635. The conclusion of the NTSB on this issue unfortunately is unclear. The Board never expressly held that McHenry suffered from TGA. They did find, however, that an earlier decision, Petition of Weddle, 1 NTSB 1931 (1972), should be overruled in order to reach a decision in McHenry’s case. Record at 700. Weddle had established that TGA provides a satisfactory medical explanation for a disturbance of consciousness. Thus, we believe the logical conclusion is that the Board determined McHenry suffered an attack of TGA in January of 1977; otherwise, there would be no reason to overrule the earlier decision. But the matter is not quite so simple. In the Order on Reconsideration, two Board members specifically state that “the Board did not determine that the disturbance was an episode of TGA.” Record at 782, n. 2. Despite this ambiguity and inconsistency, we find that McHenry’s disturbance of consciousness was the result of *1192TGA. All of McHenry’s medical experts agree with this conclusion. Even the Administrator’s sole witness, Dr. Stevens, stated that if the three earlier episodes were disregarded as medically insignificant events, then the later episode would fit the classical pattern of TGA. Record at 307-8. The only reason Dr. Stevens hesitated to diagnose the later episode as TGA was that TGA characteristically occurs only once, and he believes McHenry suffered three earlier disturbances of consciousness. Record at 297. Therefore, since we have already determined that McHenry suffered only one disturbance of consciousness, we must now conclude that this isolated episode was a manifestation of TGA. There is no evidence to support any other conclusion.
The remaining question in this analysis is whether TGA constitutes a satisfactory medical explanation for the cause of an isolated disturbance of consciousness. In 1972, the NTSB determined that it was a satisfactory medical explanation. Weddle, 1 NTSB 1931. In its decision of McHenry’s case in 1979, the Board concluded that Weddle was incorrectly decided and that TGA is not a satisfactory medical explanation for the cause of a disturbance of consciousness. We have several problems with the Board’s action on this issue.
An administrative agency may reverse a prior decision during an adjudicative proceeding. Ka Fund Chan v. INS, 634 F.2d 248 (5th Cir. 1981). “An administrative agency concerned with furtherance of the public interest is not bound to rigid adherence to precedent.” New Castle County Airport Commission v. CAB, 371 F.2d 733 (D.C.Cir.1966), cert. denied sub. nom. Board of Transportation v. CAB, 387 U.S. 930, 87 S.Ct. 2052, 18 L.Ed.2d 991 (1967). This does not mean, however, that an agency may abandon its own precedent without reason or explanation. “An agency must either conform to its prior norms and decisions or explain the reason for its departure from such precedent.” Mississippi Valley Gas Co. v. FERC, 659 F.2d 488, 506 (5th Cir. 1981). Agencies commonly use adjudicative proceedings to formulate standards and policies. These standards and policies are in turn used to predict future agency action in much the same way that the doctrine of stare decisis is used in the courts. A settled course of agency action also insures that the agency is pursuing the policies committed to it by Congress. Atchison, Topeka, & Sante Fe Ry. Co. v. Board of Trade, 412 U.S. 800, 807, 93 S.Ct. 2367, 2374, 37 L.Ed.2d 350 (1973).
There is, then, at least a presumption that those [congressional] policies will be carried out best if the settled rule is adhered to. From this presumption flows the agency’s duty to explain its departure from prior norms. The agency may flatly repudiate those norms, deciding, for example, that changed circumstances mean that they are no longer required in order to effectuate congressional policy. Or it may narrow the zone in which some rule will be applied because it appears that a more discriminating invocation of the rule will best serve congressional policy. .. . Whatever the ground for the departure from prior norms, however, it must be clearly set forth so that the reviewing court may understand the basis of the agency’s action. . . .
Id. at 808, 93 S.Ct. at 2375 (citations omitted). This principle is a corollary to the requirement that an agency sufficiently explain the basis for its decision and articulate the standards and rationale used in its decision:
The fundamental principle of reasoned explanation embodied in . . . [agency] decisions serves at least three interrelated purposes: enabling the court to give proper review to the administrative determination; helping to keep the administrative agency within proper authority and discretion, as well as helping to avoid and prevent arbitrary, discriminatory, and irrational action by the agency; and informing the aggrieved person of the grounds of the administrative action so that he can plan his course of action (including the seeking of judicial review).
Matlovich v. Secy, of the Air Force, 591 F.2d 852, 857 (D.C.Cir.1978). In other *1193words, the requirements that an agency explain its departure from precedent, and adequately explain the rationale of its decision, are prerequisites to a judicial finding that an agency’s action is not arbitrary and capricious. We find that the NTSB has not met either requirement in this case.
The NTSB below merely stated: “The Board now believes that its acceptance of TGA as a satisfactory medical explanation in that case [Weddle] was in error. Moreover, we now believe that our acceptance of the representation that a recurrence is so unlikely as to present no future risk was not a correct conclusion.” Record at 700. The Board’s reasoning for reaching this conclusion is twofold: (1) McHenry failed to establish the eause of his disturbance of consciousness; and (2) he failed to establish that it will not recur. Record at 700. The Board determined that “TGA” is only a descriptive term and that the actual underlying cause of the syndrome is unknown. All of the medical testimony supports this conclusion. In the past, however, the Board has held that proof of the underlying cause of a disturbance of consciousness is not necessary if the disturbance can be diagnosed as a medically accepted phenomenon, and accordingly, a reliable analysis of futare risk can be made. In’ Petition of Mosely, 2 NTSB 1824 (1975), the petitioner suffered a disturbance of consciousness because of a subarachnoid hemorrhage. The underlying cause of the subarachnoid hemorrhage was unknown but a “possibly disqualifying defect.” The Board held, despite the underlying unknown cause, that an accepted medical diagnosis would allow consideration of future risk under a more directly applicable regulation.7 Therefore, we believe the Board previously has interpreted “satisfactory medical explanation of the cause” to mean a medically accepted diagnosis which will allow a reliable analysis of future risk. We are not certain if the Board will continue this approach, or in the future require proof of direct biological and medical causation.8 If the Board has ’abandoned the approach in Mosely and adopted a new standard, they should say as much and explain what motivated the shift.
The Board also based its decision on a finding that McHenry failed to prove TGA has “little or no likelihood of recurrence.” We believe that this conclusion is not supported by substantial evidence.9 Every medical expert at the hearing, including the Administrator’s witness, Dr. Stevens, testified that he believes TGA rarely recurs.10 *1194The Board acknowledged that the medical experts agreed recurrences of TGA are rare, but continued “we believe it is logical to conclude that, since diagnoses are rare, documented histories are rare.” Record at 700. This “logical conclusion” is also not supported by the evidence and cannot serve as a basis for the Board’s decision. The amount and depth of research into TGA was not a debated issue at the hearing. Moreover, three doctors who testified at the hearing have treated an aggregate of over 100 TGA victims. Dr. Stevens testified that as many as 200 cases are reported in literature. Record at 815. Therefore, the small amount of evidence in the record is contrary to the Board’s conclusion.
We believe the primary reason for the lack of supporting evidence on this entire question is that the issue of whether TGA constitutes a satisfactory medical explanation for an isolated disturbance of consciousness was never properly argued by the parties. The Administrator did not submit substantial evidence to support the Board’s reasoning. We are not willing to state, however, that there is substantial and compelling evidence in support of McHenry’s position. As discussed above, we are uncertain of what standard the Board has used and we cannot consider McHenry’s evidence until a standard is clearly established. For these reasons, and in the interest of public safety, we remand this issue for further proceedings before the NTSB. The opportunity for both parties to introduce additional evidence will facilitate the resolution of the issue.
Ill
Finally, we must address McHenry’s claim that the Administrator’s notice of appeal to the NTSB was untimely filed. The NTSB rules of practice provide that an appeal from a law judge’s decision is initiated by the filing of a simple notice of appeal within 10 days after a written initial decision is served:
A party may appeal from a law judge’s order or from the initial decision by filing with the Board and serving upon the other parties (pursuant to § 821.8) a notice of appeal within 10 days after an oral initial decision has been rendered or a written decision or an order has been served.
49 C.F.R. § 821.47. Filing can be accomplished by mail or personal delivery, and the method for establishing the date of filing is dependent on the method of filing:
Documents to be filed with the Board shall be filed with the Office of Administrative Law Judges, National Transportation Safety Board, Washington, D. C. 20594, by personal delivery or by mail and shall be deemed to be filed on the date of actual personal delivery or on the date as shown on the postmark, as the case may be.
49 C.F.R. § 821.7(a). The failure to file a timely notice of appeal results in the initial decision becoming final. 49 C.F.R. § 821.-43.
In this case, the administrative law judge’s initial decision in favor of McHenry was served on June 12, 1979. Therefore, the Administrator had until June 22 to personally deliver or place in the mail a notice of appeal to the NTSB. The notice of appeal filed by the Administrator was dated June 22 and was accompanied by a certificate of service in which the signator represented that the document was placed in United States mail on June 22. The Administrator, however, used a franked government envelope which received no postmark, and the notice was received by the NTSB on June 27. Thus, there is no way to objectively guarantee that the notice was actually mailed on June 22.
McHenry moved to strike the Administrator’s notice of appeal as untimely filed. He argued that the plain language of 49 C.F.R. § 821.7(a) dictates two methods of establishing a timely notice of appeal: (1) personally delivering the notice within 10 days after the initial decision is served; or (2) placing a notice of appeal in the United States mail and having it postmarked within 10 days after the initial decision is served. McHenry claimed that the Administrator failed to establish a timely *1195filing since he did not use either method provided in the regulation. The NTSB rejected this argument and denied McHenry’s motion:
[I]t would be inequitable to penalize the Administrator by restricting the proof of the date of mailing to a postmark, which is not applied to government mail using a franked envelope. Rather, the more reasonable approach, and the one most in keeping with the spirit of the Board’s Rules, is to accept, as the date of filing, or the date of mailing, the date given on the document filed by the Administrator and on the certificate of service attached to that document.
Record at 672. McHenry now argues that the Board improperly promulgated a new regulation in reaching its decision on this issue. He contends that 49 C.F.R. § 821.-7(a) has been amended to allow the establishment of a timely notice of appeal through the representations of the appealing party, instead of by an objective means such as a postmark or date of receipt. We cannot agree. We do not interpret the Board’s decision as establishing a new rule. Instead, we believe that the Board was simply interpreting the regulations and applying them to the facts of this case. A reviewing court must defer to an agency’s decision if the agency has made a reasonable interpretation of its own regulations. Homan & Crimen, Inc. v. Harris, 626 F.2d 1201 (5th Cir. 1980). Here, the NTSB interpreted 49 C.F.R. §§ 821.7(a) and 821.47 as clearly requiring personal delivery or mailing of a notice of appeal within a ten day period. A postmark, however, is not an absolute requirement for proof of mailing within the ten day period. Postmarks are often illegible and government franked mail is frequently not postmarked. Therefore, in these cases, the Board can look to other factors to determine whether the notice was timely filed. The Board considered the relevant factors in McHenry’s ease, including the certifications by the Administrator’s staff, and concluded that the notice was timely filed. We believe this is a reasonable interpretation of the regulations and affirm the Board’s decision.11
McHenry alternatively claims that the Administrator has acted in bad faith. He primarily relies on two pieces of circumstantial evidence to prove that the notice was not mailed on June 22. First, in 1979 the FA A Chief Counsel’s office and the NTSB Office of Administrative Law Judges were located in the same building, only one floor from each other. McHenry claims that it is highly unlikely that it would take the United States mail five days to deliver a document from one floor to another floor of the same building. Unfortunately, we cannot say that five days is an unreasonable or unbelievable amount of time for the United States mail to complete this process. Second, McHenry claims that the FAA attorney who signed the notice of appeal should have known this problem might develop since the document was prepared on the tenth and final permissible day, and accordingly he should have hand delivered the notice to the NTSB office. We recognize that this would have been an admirable effort, but it is not required or expected. It is not unreasonable or negligent for an attorney to mail a notice of appeal on the tenth day, as the regulations allow, without concern over a postmark. Absent other evidence, we cannot find that the Administrator or his staff has falsified documents and acted in bad faith.
IV
For the reasons discussed in sections II and III of this opinion, the order of the NTSB is affirmed in part, and reversed and remanded in part for further proceedings consistent with our findings.
AFFIRMED IN PART, REVERSED AND REMANDED IN PART.
. For an extensive discussion of the airman medical certification process, see Delta Airlines, Inc. v. United States, 490 F.Supp. 907 (N.D.Ga.1980).
. 14 C.F.R. § 67.13 provides in part:
(a) To be eligible for a first-class medical certificate, an applicant must meet the requirements of paragraphs (b) through (f) of this section.
(d) Mental and neurologic—
(2) Neurologic
(i) No established medical history or clinical diagnosis of either of the following:
(b) A disturbance of conciousness without satisfactory medical explanation of the cause.
. There are five members of the NTSB but only four participated in the original decision and reconsideration of this case.
. We do not intend to ridicule the Board’s decision and think it important to note that we have not forgotten that one of the Board’s primary functions is to insure safety in the air. Day v. NTSB, 414 F.2d 950 (5th Cir. 1969). On the contrary, we are deeply concerned that an agency vested with such an important and essential duty can act so erroneously.
. Although McHenry only seeks first-class *1191medical certification, he was denied first, second, and third-class certification for the same reason. The disqualifying provisions for the three classes are identical and will not be distinguished in this opinion. The relevant portion of the first-class medical regulations is 14 C.F.R. § 67.13(d)(2)(i)(fo). See note 2 supra. The second-class medical regulation is 14 C.F.R. § 67.15(d)(2)(i)(b), and the third-class medical regulation is 14 C.F.R. § 67.17(d)(2)(i)(b ).
. We state that “the NTSB apparently agreed” because the Board also stated that McHenry failed to prove his illness was an isolated episode. Record at 701. Moreover, in the Order on Reconsideration, two Board members repeat this argument. Record at 782. We are uncertain of the meaning of these statements. The only other episodes at issue were resolved in favor of McHenry. We are unable to think of anything more McHenry could do to prove that the later episode was his only disturbance of consciousness. The Board members possibly were referring to McHenry’s failure to prove that his illness will not recur, but this is a separate issue and is discussed at length later in the opinion.
. In Mosely, the alternative regulation was 14 C.F.R. § 67.17(d)(2)(ii). In McHenry’s case, the analogous regulation would be 14 C.F.R. § 67.13(d)(2)(ii), which provides in part: “[The applicant has] no other convulsive disorder, disturbance of consciousness, or neurological condition that the Federal Air Surgeon finds (a) makes the applicant unable to safely perform the duties ... of the airman certificate that he holds ... or (b) may reasonably be expected within 2 years ... to make him unable to perform those duties....” The standards and burden of proof under this section are different than the standards and burden of proof under 14 C.F.R. § 67.13(d)(2)(i)(fo).
. We make this point because there are many illnesses that can cause a disturbance of consciousness, such as cancer and hypoglycemia, which are widely recognized and have familiar and predictable characteristics, but for which an underlying cause is unknown.
. The only evidence to the contrary was the statement by Dr. Stevens that some doctors have reported frequent recurrences. Record at 308. However, there was no other substantiation of this claim. Dr. Fisher testified that all of these “other” reports can be clearly discounted. Dr. Stevens apparently agrees since he also believes recurrences are rare.
Moreover, we are puzzled as to why the Administrator chose to present so little evidence in support of his case. He used one witness in support of the following two inconsistent propositions: First, McHenry could not have suffered TGA because TGA occurs only once and McHenry has suffered four disturbances of consciousness. Record at 297. Second, TGA is not a satisfactory medical explanation because it is likely to recur. Record at 307-8.
. In its decision, the Board stated that Dr. Stevens testified that there is a strong likelihood of recurrence. Record at 699. This simply is not true. In 1974, Dr. Stevens wrote in reference to TGA victims: “Prognosis is excellent, recurrences rare.” Stevens and Aldeman, Transient Global Amnesia, 43 Medical Annals of the District of Columbia 593, 596 (1974). He still supports that statement. Record at 309.
. The Board’s decision on this issue is also consistent with its long standing policy of preferring to determine a case on the merits rather than on a procedural technicality. See, e.g., Petition of'Mosely, 2 NTSB 1833 (1975); Petition of Mixon, 2 NTSB 1442 (1975). We believe this policy is extremely important, if not essential, to an agency charged with protecting the public from physical danger. See NLRB v. Monsanto Chemical Co., 205 F.2d 763 (8th Cir. 1953).