525 F.3d 1168 (2008)

Jonathan L. HAAS, Claimant-Appellee,
v.
James B. PEAKE, M.D., Secretary of Veterans Affairs, Respondent-Appellant.

No. 2007-7037.

United States Court of Appeals, Federal Circuit.

May 8, 2008.

*1171 Barton F. Stichman, National Veterans Legal Services Program, of Washington, DC, argued for claimant-appellee. With him on the brief was Louis J. George.

Todd M. Hughes, Assistant Director, Commercial Litigation Branch, Civil Division, United States Department of Justice, of Washington, DC, argued for respondent-appellant. With him on the brief was Jeanne E. Davidson, Director. Of counsel on the brief were David J. Barrans, Deputy Assistant General Counsel, and Ethan G. Kalett, Staff Attorney, United States Department of Veterans Affairs, of Washington, DC.

Before MICHEL, Chief Judge, BRYSON, Circuit Judge, and FOGEL, District Judge.^[*]

Opinion for the court filed by Circuit Judge BRYSON; Dissenting opinion filed by District Judge FOGEL.

BRYSON, Circuit Judge.

Beginning in 1962, the United States used herbicides such as Agent Orange in Vietnam for the purpose of "defoliation, crop destruction, and on a smaller scale, clearing vegetation around U.S. fire bases and other installations, around landing zones, and along lines of communication." S.Rep. No. 100-439, at 64-65 (1988). Agent Orange consisted of an equal mixture by weight of two chemicals, 2, 4-dichlorophenoxyacetic acid and 2,4,5-trichlorophenoxyacetic acid. It also contained trace amounts of 2,3,7,8-tetrachlorodibenzo-para-dioxin, also known as dioxin. Id. at 64. The use of Agent Orange in Vietnam increased substantially between 1967 and 1969. Agent Orange came under scrutiny after a report from the National Institutes of Health indicated that 2,4,5,-trichlorophenoxyacetic acid was associated with birth defects in animals, although later research indicated that those birth defects were more likely caused by dioxin. Id. at 65; see also David A. Butler, Connections: The Early History of Scientific and Medical Research on "Agent Orange", 13 J.L. & Policy 527, 545-48 (2005); Inst. Of Med., Veterans and Agent Orange: Health Effects of Herbicides Used in Vietnam 30 (1994) ("Veterans and Agent Orange") (discussing later research). The use of Agent Orange was phased out by 1971. Veterans and Agent Orange at 27.

*1172 The impact of Agent Orange on humans has subsequently been the subject of much research and controversy. Congress has enacted several statutes mandating that research be conducted regarding the impact of Agent Orange on human health and providing that veterans be compensated for illnesses resulting from exposure to the chemical. This case concerns the Agent Orange Act of 1991, Pub.L. 102-4, 105 Stat. 11, which provided a special mechanism of disability compensation for veterans exposed to herbicides such as Agent Orange.

To receive disability compensation, a veteran must establish that the disability was service connected, which means that it must have been "incurred or aggravated... in the line of duty in the active military, naval, or air service." 38 U.S.C. § 101(16). The Agent Orange Act provided that for certain veterans and certain diseases, both exposure and service connection are presumed to be established. 38 U.S.C. § 1116(a)(1).

The statutory list of diseases as to which exposure and service connection are presumed includes non-Hodgkin's lymphoma, certain soft-tissue sarcomas, chloracne, Hodgkin's disease, porphyria cutanea tarda, certain respiratory cancers, multiple myeloma, and diabetes mellitus (type 2). See 38 U.S.C. § 1116(a)(2). If a veteran can prove that he or she has one of the listed diseases and "served in the Republic of Vietnam" between January 9, 1962, and May 7, 1975, the disease will ordinarily "be considered to have been incurred in or aggravated by such service." 38 U.S.C. § 1116(a)(1)(A). Consequently, proving service "in the Republic of Vietnam" is important to any veteran who seeks compensation for one of the listed diseases.

This case calls on us to address whether veterans who served on ships off the coast of Vietnam during the Vietnam War served "in the Republic of Vietnam" and thus are entitled to the presumption of service connection if they suffer from one of the listed diseases. The government argues that the phrase "served in the Republic of Vietnam" requires that a servicemember have at some point set foot within the land borders of Vietnam. Mr. Haas contends that the phrase extends to those who served on board ships in the waters off the Vietnamese coast but never went ashore.

By regulation, the Department of Veterans Affairs ("DVA") has interpreted the phrase "served in the Republic of Vietnam" to mean that the veteran's service must have involved "duty or visitation" in the Republic of Vietnam in order for the veteran to be entitled to the statutory presumption of service connection. See 38 C.F.R. § 3.307(a)(6)(iii). That regulation, as interpreted by the DVA, made the statutory presumption of service connection unavailable to veterans such as appellant Jonathan Haas, who served on a naval vessel that traveled in the waters near Vietnam but who never went ashore. The Court of Appeals for Veterans Claims ("the Veterans Court") set aside the DVA's interpretation as unduly restrictive. Haas v. Nicholson, 20 Vet.App. 257 (2006). We hold that the agency's requirement that a claimant have been present within the land borders of Vietnam at some point in the course of his duty constitutes a permissible interpretation of the statute and its implementing regulation, and we therefore reverse the judgment of the Veterans Court.

I

In August 2001, Mr. Haas applied to the Phoenix, Arizona, regional office of the DVA seeking disability compensation for type 2 diabetes, peripheral neuropathy, and loss of eyesight. He claimed that he had been exposed to herbicides while serving *1173 in Vietnam and that based on that exposure he was entitled to a finding of service connection for his conditions.

Mr. Haas served on active duty in the United States Navy from September 1959 to September 1960 and subsequently from May 1963 to June 1970. Service records indicate that from August 1967 to April 1969, Mr. Haas served on the U.S.S. Mount Katmai, which he described as an ammunition supply ship that operated in the West Pacific off the coast of Vietnam. It is undisputed that that Mr. Haas never went ashore, and thus never set foot on the physical landmass of the Republic of Vietnam. Mr. Haas explained that his ship did not visit any ports because it carried highly explosive ammunition and would have posed a threat if docked in a port. Mr. Haas subsequently left active duty and was transferred to the Retired Reserves on July 1, 1982.

Mr. Haas's claim to service connection for his condition is based on his naval service and the presumptive service connection afforded for type 2 diabetes based upon a showing that the veteran "served in the Republic of Vietnam." See 38 U.S.C. §§ 1116(a)(1)(A), (a)(2)(H); 38 C.F.R. § 3.307(a)(6)(iii). In denying his claim, the regional office explained that in order to qualify for a presumption of service connection, Mr. Haas must have "physically served or visited in the Republic of Vietnam." For a sailor serving in the waters offshore, the regional office explained that "the ship must have come to port in the [Republic of Vietnam] and you disembarked." Mr. Haas disagreed with the regional office and contended that "service in the Republic of Vietnam," as defined by 38 C.F.R. § 3.307(a)(6)(iii), should be interpreted to include service in the offshore waters regardless of whether the servicemember's ship came to port and the servicemember disembarked.

On appeal, the Board of Veterans' Appeals affirmed the regional office's decision denying Mr. Haas the presumption of service connection. The Board applied the DVA's regulation, as interpreted by the agency, and ruled that Mr. Haas was not entitled to the statutory presumption for those who served "in the Republic of Vietnam" because he had never "set foot on land in the Republic of Vietnam." As for Mr. Haas's contention that he was actually exposed to herbicides while his ship operated near the coast of Vietnam, the Board rejected his claim on the ground that his allegation was "unsupported by any evidence demonstrating that his ship was located in waters sprayed by herbicides."

Mr. Haas then appealed to the Veterans Court. A three-judge panel of that court reversed the Board's decision. The court first found the phrase "served in the Republic of Vietnam" in 38 U.S.C. § 1116 to be ambiguous. The court explained that "[t]here are many ways in which to interpret the boundaries of a sovereign nation such as the former Republic of Vietnam" and that the "legislative history of the 1991 act ... is silent concerning what constitutes `service in the Republic of Vietnam.'" 20 Vet.App. at 263, 268. Turning to the DVA's interpretation of the statutory language, the court first examined the pertinent regulation, 38 C.F.R. § 3.307(a)(6)(iii). That regulation defines "service in the Republic of Vietnam" as including "service in the waters offshore and service in other locations if the conditions of service involved duty or visitation in the Republic of Vietnam." The court determined that the regulation "do[es] not clearly preclude application of the presumption [of service connection] to a member of the Armed Forces who served aboard a ship in close proximity to the landmass of the Republic of Vietnam." 20 Vet.App. at 259.

*1174 Finding that the regulation "merely has replaced statutory ambiguity with regulatory ambiguity," the Veterans Court then analyzed the DVA's interpretation of the regulation and concluded that the agency's current interpretation of its regulation conflicts with the agency's earlier interpretation of the same regulation. The court noted that the agency's original instructions to its adjudicators in the Adjudication Manual of the Veterans Benefits Administration, M21-1 ("Manual M21-1"), called for awarding presumptive service connection for specified diseases if the veteran had received the Vietnam Service Medal "in the absence of contradictory evidence," and that those provisions were not altered following the issuance of two precedential DVA General Counsel opinions on related topics. See DVA Op. Gen. Counsel Prec. 27-97 (1997) (finding that service on a deepwater vessel off the shore of Vietnam did not constitute service "in the Republic of Vietnam" under 38 U.S.C. § 101(29)(A)); DVA Op. Gen. Counsel Prec. 7-93 (1993) (finding that service in high altitude planes flying over Vietnam without any other contact with Vietnam did not constitute "service in Vietnam" under 38 C.F.R. § 3.313). Consequently, the court found that when the DVA adopted the "foot-on-land" test, it was reversing its previously established course. 20 Vet.App. at 270-72.

The Veterans Court further concluded that the agency's new interpretation was not a reasonable one. In so ruling, the Veterans Court noted that under the DVA's current interpretation of the regulation, the DVA "would afford the presumption of exposure to Agent Orange to a Vietnam-era veteran who served only in the inland waterways of the Republic of Vietnam and never set foot on land; yet, in order for a Vietnam-era veteran serving in the waters surrounding Vietnam to be entitled to the presumption, he or she must have set foot on land, without consideration as to either the length of time spent patrolling in the waters offshore, or the risks of windblown exposure to Agent Orange sprayed along Vietnam's coastline." 20 Vet.App. at 275. The court explained that

given the spraying of Agent Orange along the coastline and the wind borne effects of such spraying, it appears that these veterans serving on vessels in close proximity to land would have the same risk of exposure to the herbicide Agent Orange as veterans serving on adjacent land, or an even greater risk than that borne by those veterans who may have visited and set foot on the land of the Republic of Vietnam only briefly.

Id. at 273. Based on that reasoning, the court concluded that the DVA's interpretation of section 3.307(a)(6)(iii) was "plainly erroneous" and that the regulation "must be read to include at least service of the nature described by the appellant, that is, service in the waters near the shore of Vietnam." Id.

Finally, the Veterans Court ruled that the pertinent provisions of the DVA's Manual M21-1 were "substantive rules" and that the DVA's amendment of those provisions in February 2002 to incorporate the "foot-on-land" requirement was invalid because the DVA had failed to make that change pursuant to the notice-and-comment requirements of 5 U.S.C. § 553. 20 Vet.App. at 277. Alternatively, the court ruled that the February 2002 changes could not be applied retroactively to Mr. Haas's claim, which had been filed in August 2001, because the effect of the rule change was to narrow the scope of Mr. Haas's substantive rights. Id. at 277-78. The court therefore reversed the Board's denial of Mr. Haas's claim to service connection for diabetes and held that in Mr. Haas's case, the Manual M21-1 provision *1175 "allowing for the application of the presumption of exposure to herbicides based on the receipt of the [Vietnam Service Medal] controls." Id. at 279.

II

This court ordinarily will not hear appeals from the Veterans Court in cases that the Veterans Court remands to the Board of Veterans' Appeals. See Adams v. Principi, 256 F.3d 1318, 1320 (Fed.Cir.2001). Nonetheless, we have held that it is appropriate for us to review such cases in certain circumstances, under the principles set forth in Williams v. Principi, 275 F.3d 1361 (Fed.Cir.2002). This appeal addresses the purely legal question of the proper interpretation of a statute and its implementing regulations, a question that will not be affected by the proceedings on remand. Moreover, postponing review until after completion of the proceedings on remand could deprive the government of its right to review of the legal issue in this case, because the Secretary of Veterans Affairs has no right to seek review of a Board decision in favor of the veteran under 38 U.S.C. § 7252(a). We therefore conclude that this appeal is ripe for review even though the Veterans Court remanded the case for further proceedings before the Board. See Williams, 275 F.3d at 1364.

III

On the merits, the parties disagree about the proper resolution of virtually every issue in this case: whether the phrase "served in the Republic of Vietnam" in the Agent Orange Act of 1991 is ambiguous; whether the DVA's regulation that interprets that phrase is itself ambiguous; whether the agency's interpretation of that regulation is entitled to deference, or instead is unreasonable and inconsistent with the agency's previous, longstanding interpretation of the regulation; and whether the DVA's 2002 modification to Manual M21-1 constituted a substantive regulatory change that could not be given effect without notice-and-comment rulemaking.

A

In order to make sense of the statutory and regulatory arguments made by the parties, it is necessary to review the history of the legislative and regulatory measures directed to the issue of herbicide exposure in Vietnam. That history, both prior to and after the enactment of the Agent Orange Act of 1991, is complex.

Beginning in the late 1970s, Congress responded to widespread expressions of concern by veterans' groups regarding the health effects on Vietnam veterans of exposure to Agent Orange and other herbicides used in the conflict there. In 1979, Congress enacted a provision requiring the Veterans Administration ("VA"), as the agency was then known, to conduct an epidemiological study of persons who, while serving in the armed forces during the war in Vietnam, were exposed to dioxins produced during the manufacture of various herbicides, including Agent Orange, to determine if there might be long-term adverse health effects from such exposure. Pub.L. No. 96-151, § 307, 93 Stat. 1092, 1097-98 (1979). The responsibility for conducting that study was subsequently reassigned to the Centers for Disease Control ("CDC"). See H.R.Rep. No. 98-592, at 5 (1984), as reprinted in 1984 U.S.C.C.A.N. 4449, 4451. Congress directed the VA to publish a description of the actions that it planned to take in response to those reports. Pub.L. No. 97-72, § 401, 95 Stat. 1047, 1061-62 (1981).

In 1984, Congress enacted the Veterans' Dioxin and Radiation Exposure Compensation *1176 Standards Act, Pub.L. No. 98-542, 98 Stat. 2725 (1984). Section 5 of that Act directed the VA to prescribe regulations establishing guidelines and standards for resolving claims for benefits based on exposure during service "in the Republic of Vietnam during the Vietnam era to a herbicide containing dioxin." In particular, the statute called the VA's attention to evidence that three diseases chloracne, porphyria cutanea tarda, and soft tissue sarcoma are associated with exposure to certain levels of dioxin and directed the VA to determine whether service connection should be granted in individual cases involving each of those diseases. Id. §§ 2(5), 5(b)(2)(A)(i), 5(b)(2)(B).

In response, the VA promulgated a regulation that presumed exposure to a herbicide containing dioxin for any veteran who served "in the Republic of Vietnam" during the Vietnam era. The regulation concluded that the development of chloracne manifested within three months of exposure would be presumed to be service-connected, but that porphyria cutanea tarda and soft tissue sarcomas were not sufficiently associated with dioxin exposure to warrant similar treatment. 38 C.F.R. § 3.311a (1986); see 50 Fed.Reg. 34,452 (Aug. 26, 1985). The regulation defined "Service in the Republic of Vietnam" to include "service in the waters offshore and service in other locations, if the conditions of service involved duty or visitation in the Republic of Vietnam." 38 C.F.R. § 3.311a(a)(1) (1986). The VA explained that the regulation was adopting the VA's "longstanding policy of presuming dioxin exposure in the cases of veterans who served in the Republic of Vietnam during the Vietnam era." 50 Fed.Reg. at 34,454-55. That policy was "based on the many uncertainties associated with herbicide spraying during that period which are further confounded by lack of precise data on troop movements at the time." Id. at 34,455. "While it may be possible to approximate areas where herbicides were sprayed," the agency wrote, "it would be extremely difficult to determine with an acceptable degree of precision whether an individual veteran was exposed to dioxin." Id. Accordingly, the agency adhered to its prior policy of presuming exposure for servicemembers who had served in Vietnam. In addition, the agency provided that because some military personnel who were stationed elsewhere "may have been present in the Republic of Vietnam, `service in the Republic of Vietnam' will encompass service elsewhere if the person concerned actually was in the Republic of Vietnam, however briefly." (50 Fed.Reg. 15,848, 15,849) (Apr. 22, 1985) (proposed rule). The VA added that "[i]n view of shifting personnel deployments, absence of on-site measurement of dioxin contamination and other factors the Agency has adhered to a policy of presuming exposure if the veterans served in Vietnam during the relevant period. This section formalizes that existing policy." Id. at 15,849; see also 50 Fed.Reg. 34,452 (Aug. 26, 1985) (adopting proposed rule unamended).

Meanwhile, congressional committees continued to hold hearings to assess the epidemiological studies of Agent Orange that had been mandated in 1979. Those studies were designed to determine whether any component of Agent Orange not just dioxin affected human health, although given its notoriety dioxin often figured prominently in the research and analysis. See Veterans and Agent Orange at x; see also id. at 28-36 (discussing history of research on Agent Orange). The success of those studies depended on determining which veterans had been exposed to Agent Orange and the extent of their exposure, so that health problems among veterans who had been highly exposed could be compared to those of a control *1177 group. See id. at 58. The VA and the CDC ran into a series of problems in attempting to make that determination. Initially, it was believed that exposure could be deduced from studying ground troop movements in conjunction with records of aerial spraying of Agent Orange. See id. That approach proved unworkable, as a representative of the Centers for Disease Control explained in testimony before a subcommittee of the House Committee on Veteran's Affairs:

When CDC got into this, it was assumed there would be records that could determine exactly where an individual was on a given day, and that could be correlated with known [herbicide] use. I think with the finest use of existing records, you cannot separate between exposed and unexposed. You can get some ... approximations, but it would be a disservice to veterans and to everyone to proceed with an expensive study of this nature if you can't clearly differentiate between who's been exposed and who's not exposed. Without that, you have no basis to proceed with doing a study.

Agent Orange Studies: Hearing Before the Subcomm. on Hospitals and Health Care of the H. Comm. on Veterans' Affairs, 99th Cong. 15 (1986) ("1986 House Hearing") (statement of James O. Mason, CDC Director); see also Veterans and Agent Orange at 58.

In light of those difficulties, the CDC attempted to derive an exposure index through other means. Initially, an attempt was made to develop an index by measuring the amount of dioxin present in fat samples from veterans. 1986 House Hearing at 81-83 (statement of James O. Mason, CDC Director). Although the objective was to study Agent Orange, it was expected that determining dioxin levels would indicate the degree of exposure to Agent Orange. See Veterans and Agent Orange at 259-62 (describing use of dioxin as a "biomarker"). That procedure, however, did not bear fruit because of the practical difficulties of obtaining fatty tissue samples. Id. at 82-83. Subsequent research based on blood tests did not reveal any difference in the blood levels of dioxin between a group of veterans stationed in Vietnam and a control group of veterans stationed outside of Vietnam. The CDC ultimately concluded that it had no validated scientific method of identifying a group of veterans who were highly exposed to Agent Orange. Agent Orange Legislation and Oversight: Hearing on S. 1692, the Proposed "Agent Orange Disabilities Benefits Act of 1987"; S. 1787, the proposed "Veterans' Agent Orange Disabilities Act of 1987"; and Agent Orange Oversight Issues Before the S. Comm. on Veterans' Affairs, 100th Cong. 165-66 (1988) (statement of Thomas E. Harvey, Deputy Administrator of the VA). The CDC explained that "the Agent Orange Exposure Study ... cannot be done.... The difficulty is and has always been the inability to discriminate between exposed and unexposed ground troops." Id. at 165 (discussing the inability to derive an exposure index from military records, self-reporting, and direct measurements of dioxin from tissue samples).

Although the CDC was unable to conduct the Agent Orange exposure study as it was originally conceived by Congress in 1979 due to the inability to identify with scientific certainty which Vietnam veterans had been highly exposed to Agent Orange, there remained other sources of scientific information on the health effects of Agent Orange and dioxin in humans. One ongoing study focused on the group of Vietnam veterans who had been involved in the aerial spraying of Agent Orange, known as the "Ranch Hand study" after the name of the mission responsible for *1178 conducting the spraying operation. See Veterans and Agent Orange at 53. Further data has also been available, for example, from populations that were exposed to chemical accidents involving dioxin, workers at factories manufacturing herbicides, and agricultural or forestry workers who were exposed to herbicides similar to Agent Orange or herbicides containing dioxin before their use was largely banned in the United States. See id. at 36-45.

Against the backdrop of the ongoing scientific investigations, the VA declined to change its regulations after 1985 to provide a presumption of in-service exposure for any diseases other than chloracne, on the ground that the scientific evidence did not show a statistically probable association between Agent Orange exposure and any other disease. In litigation initiated by veterans' advocacy groups, however, a federal district court ruled that the agency, by then renamed the Department of Veterans Affairs, had applied too stringent a standard for determining which diseases to include in its regulations promulgated under the 1984 Dioxin Act. See Nehmer v. U.S. Veterans Admin., 712 F.Supp. 1404, 1420 (N.D.Cal.1989). The DVA subsequently amended its regulation, 38 C.F.R. § 3.311a, to include soft tissue sarcomas. See 56 Fed.Reg. 7632 (Feb. 25, 1991) (proposed rule); 56 Fed.Reg. 51,651 (Oct. 15, 1991) (final rule).

In October 1990, the DVA promulgated a separate regulation providing that "Service in Vietnam during the Vietnam Era," together with subsequent development of non-Hodgkin's lymphoma, "is sufficient to establish service connection for that disease." 38 C.F.R. § 3.313. That regulation was based on information in a CDC study that had been released earlier that year. See 55 Fed.Reg. 25,339 (June 21, 1990) (proposed rule). The CDC study found a statistically significantly elevated level of non-Hodgkin's lymphoma among Vietnam veterans by comparing veterans who served in Vietnam and those who served in other locations during the Vietnam era. For purposes of the analysis, the study treated veterans who were stationed off the coast of Vietnam as Vietnam veterans. See The Association of Selected Cancers with Service in the U.S. Military in Vietnam, as reprinted in Centers for Disease Control Selected Cancers Study and Scientific Reviews of the Study: Hearing before the H. Comm. On Veterans' Affairs, 101st Cong.2d Sess. 106 (1990) ("1990 CDC Study"). The study concluded that there was no evidence that the increased risk of non-Hodgkin's lymphoma among Vietnam veterans was related to exposure to Agent Orange in Vietnam. Id. at 81, 125.

In the 1990 regulation, the DVA defined "Service in Vietnam" to include "service in the waters offshore, or service in other locations if the conditions of service involved duty or visitation in Vietnam." 38 C.F.R. § 3.313 (1991). That language was similar to the language previously used to define "service in the Republic of Vietnam," but it differed in two subtle, but important respects. First, the 1990 regulation referred to "Service in Vietnam" rather than using the statutory phrase "service in the Republic of Vietnam." Second, the placement of the comma before the word "or" in the definition of "service in Vietnam" in the 1990 regulation, section 3.313, suggested that the requirement of visitation or duty in Vietnam applied to "service in other locations," but not to "service in the waters offshore." Section 3.311 a used the word "and" rather than "or" and did not have a comma separating the reference to "service in the waters offshore" and "service in other locations," which suggested that the requirement of visitation or duty in the Republic of Vietnam applied to both of those forms of extraterritorial service.

*1179 The government does not dispute that the 1990 non-Hodgkin's lymphoma regulation, which is still in effect, applies to veterans who served "offshore" and never visited the landmass of Vietnam, as those veterans were among those found to have an elevated risk of non-Hodgkin's lymphoma in the 1990 CDC study. In fact, in 1993 the DVA issued a General Counsel opinion in which the agency explicitly stated that the non-Hodgkin's lymphoma regulation covers servicemembers who served in the waters off the shore of Vietnam, although the opinion concluded that the regulation does not cover servicemembers whose involvement in the Vietnam theater was limited to high-altitude missions in Vietnamese airspace. DVA Op. Gen. Counsel Prec. 7-93 (Aug. 12, 1993).

By contrast, the government asserts that under the more general 1985 dioxin exposure regulation, section 3.311 a, a veteran who served offshore must have set foot on the landmass of Vietnam in order to satisfy the regulatory definition of having served "in the Republic of Vietnam." The punctuation of the earlier definition in the 1985 regulation, section 3.311a, supports the government's position, as it suggests that the requirement of visitation or duty in the Republic of Vietnam applies to both "service in other locations" and "service in the waters offshore."

In 1991, Congress enacted the Agent Orange Act, Pub.L. No. 102-4, 105 Stat. 11, which established a more comprehensive statutory framework for herbicide-based claims. As enacted, the Agent Orange Act specified three diseases non-Hodgkin's lymphoma, certain soft tissue sarcomas, and chloracne and provided that when one of those diseases became manifest "in a veteran who, during active military, naval, or air service, served in the Republic of Vietnam during the Vietnam era," the disease would be considered to have been incurred in or aggravated by such service.^[1] Pub.L. No. 102-4, § 2(a), 105 Stat. 11, 12 (1991) (now codified, as amended, at 38 U.S.C. § 1116(a)(1)). In addition, the Act directed the DVA to identify other diseases associated with herbicide exposure. The Act provided that any veteran who "served in the Republic of Vietnam during the Vietnam era" and has a disease designated by the Secretary "shall be presumed to have been exposed during such service to an herbicide agent containing dioxin or 2, 4-dichlorophenoxyacetic acid, and may be presumed to have been exposed during such service to any other chemical compound in an herbicide agent, unless there is affirmative evidence to establish that the veteran was not exposed to any such agent during that service." Pub.L. No. 102-4, § 2(a), 105 Stat. at 12 (now codified, as amended, at 38 U.S.C. § 1116(f)).

The legislative history of the Agent Orange Act indicates that Congress sought to strike a balance between waiting for the results of scientific research regarding the effects of Agent Orange and providing benefits for Vietnam veterans with current health problems. The Chairman of the House Committee on Veterans' Affairs stated:

The question of whether compensation should be paid for disabilities allegedly *1180 related to exposure to herbicides has gone on for much too long.... It has received an inordinate amount of attention and energy. It is time to move on and, in doing so, to leave in place a mechanism for continuing scientific scrutiny which, if allowed to work, can assuage the remaining concerns of affected veterans.

137 Cong. Rec. 2348 (1991) (statement of Rep. Montgomery). The Act therefore codified the presumption of service connection for the three diseases already covered by DVA regulations, mandated independent scientific review through the National Academy of Sciences, and instructed the Secretary of the DVA to consider designating additional diseases as service-connected when recommended by the National Academy of Sciences. Importantly for present purposes, the focus of Congress's attention was on the scientific evidence as to what diseases were linked to Agent Orange exposure; there was no indication during the legislative process that Congress focused on the precise scope that should be attached to the statutory phrase "served in the Republic of Vietnam."

When the DVA drafted regulations for the Agent Orange Act, it incorporated the definition of the phrase "service in the Republic of Vietnam" from the 1985 general dioxin exposure regulation, 38 C.F.R. § 3.311a. See 58 Fed.Reg. 50,528, 50,529 (Sept. 28, 1993) (adopting amended section 3.307(a)(6)). Thus, the DVA defined "service in the Republic of Vietnam" to mean "service in the waters offshore and service in other locations if the conditions of service involved duty or visitation in the Republic of Vietnam." 38 C.F.R. § 3.307(a)(6)(iii) (1994). The DVA explained that in light of the enactment of the Agent Orange Act it was no longer necessary to retain the general dioxin exposure regulation, 38 C.F.R. § 3.311a. However, the DVA noted that the definition of the phrase "service in the Republic of Vietnam" in the new regulation would be incorporated directly from the definition in section 3.311a. 58 Fed.Reg. 50,528, 50,529 (Sept. 28, 1993) (proposed rule).

The following year, the DVA issued another set of regulations in which it added Hodgkin's disease and porphyria cutanea tarda to the list of diseases for which the agency would presume exposure and service connection based on presence in Vietnam during the Vietnam era. See 59 Fed. Reg. 5106 (Feb. 3, 1994). The new regulation retained the language from the general dioxin exposure regulation of 1985 and continued to define "service in the Republic of Vietnam" to include "service in the waters offshore and service in other locations if the conditions of service involved duty or visitation in the Republic of Vietnam." 38 C.F.R. § 3.307(a)(6)(iii) (1995).

The question whether the phrase "service in the Republic of Vietnam" included servicemembers whose service was limited to ships that had traveled in waters off the shore of Vietnam came into sharp focus in 1997. First, in a precedential General Counsel opinion issued that year, the DVA construed the phrase "served in the Republic of Vietnam" in 38 U.S.C. § 101(29)(A) not to apply to servicemembers whose service was on ships and who did not serve within the borders of the Republic of Vietnam during a portion of the "Vietnam era." The opinion stated that the definition of the phrase "service in the Republic of Vietnam" in the Agent Orange regulation, 38 C.F.R. § 3.307(a)(6)(iii), "requires that an individual actually have been present within the boundaries of the Republic to be considered to have served there," and that for purposes of both the Agent Orange regulation and section 101(29)(A), service "in the Republic of Vietnam" does not include service *1181 on ships that traversed the waters offshore of Vietnam absent the servicemember's presence at some point on the landmass of Vietnam. DVA Op. Gen. Counsel Prec. 27-97 (1997).

Later that same year, in a proposed regulation addressing incidents of spina bifida among the children of servicemembers who had served in Vietnam, the DVA proposed to use the same regulatory definition for "service in the Republic of Vietnam" that it had used in the 1985 regulation and the Agent Orange regulation. See 62 Fed.Reg. 23,724, 23,725 (May 1, 1997) (proposed rule). A commenter objected to the definitional language and urged that the phrase "if the conditions of service involved duty or visitation in the Republic of Vietnam" be eliminated from the regulation. See 62 Fed. 51,274, 51,274-75 (Sept. 30, 1997) (final rule). The DVA declined to make that change. It explained the reason for not making the suggested change as follows:

Because herbicides were not applied in waters off the shore of Vietnam, limiting the scope of the term service in the Republic of Vietnam to persons whose service involved duty or visitation in the Republic of Vietnam limits the focus of the presumption of exposure to persons who may have been in areas where herbicides could have been encountered.

62 Fed.Reg. at 51,274.

In 2001, the DVA issued a proposed regulation to include type 2 diabetes among the illnesses for which presumptive service connection would be recognized based on herbicide exposure. See 66 Fed. Reg. 2376 (Jan. 22, 2001) (proposed rule). The proposed regulation would presume herbicide exposure based on "service in the Republic of Vietnam," which would continue to be defined to cover service in waters offshore of Vietnam "if the conditions of service involved duty or visitation in the Republic of Vietnam." The DVA subsequently adopted the proposed rule including type 2 diabetes among those diseases as to which presumptive service connection would be recognized. 66 Fed.Reg. 23,166 (May 8, 2001) (final rule).

In the course of the rulemaking proceeding, a comment was made urging the DVA to use that proceeding to make clear that "service in the Republic of Vietnam" includes "service in Vietnam's inland waterways or its territorial waters." The comment was based on the assertion that U.S. military personnel had been exposed to herbicides while serving in those locations. In its final rulemaking order, the DVA responded that it is "commonly recognized" that the statutory term "in the Republic of Vietnam" includes the inland waterways. 66 Fed.Reg. at 23,166. With respect to service in the offshore waters, however, the DVA explained that even before the enactment of the Agent Orange Act, the agency had taken the position that service offshore required some duty or visitation within the Republic of Vietnam to qualify for the presumptions of herbicide exposure and service connection, and that service on a deepwater vessel offshore did not constitute such service. The DVA added that the commenter had cited "no authority for concluding that individuals who served in the waters offshore of the Republic of Vietnam were subject to the same risk of herbicide exposure as those who served within the geographical boundaries of the Republic of Vietnam, or for concluding that offshore service is within the meaning of the statutory phrase `Service in the Republic of Vietnam.'" Id. Accordingly, the agency declined to make the suggested change. Later that year, Congress followed the DVA's lead by adding type 2 diabetes to the list of diseases included in section 1116(a)(2). See Veterans Education and Benefits Expansion Act of *1182 2001, Pub.L. No. 107-103, § 201(b), 115 Stat. 967.

In early 2002, the DVA amended the language of its Adjudication Manual M21-1 to specifically incorporate the agency's "foot-on-land" interpretation of the Agent Orange regulations. Before the amendment, the Manual provided that in determining whether a veteran had "service in Vietnam," it would ordinarily be sufficient that the veteran had received the Vietnam Service Medal, but that it might be necessary in some cases to determine if the veteran's ship had been in the vicinity of Vietnam for some significant period of time. The amended version of Manual M21-1, published in February 2002, stated that, under section 3.307(a)(6) of the regulations, a veteran "must have actually served on land within the Republic of Vietnam (RVN) to qualify for the presumption of exposure to herbicides." M21-1, part III, paragraph 4.24(e)(1) (Feb. 27, 2002). It added that the fact that a veteran has been awarded the Vietnam Service Medal "does not prove that he or she was `in country,'" because servicemembers "who were stationed on ships off shore, or who flew missions over Vietnam, but never set foot in-country, were sometimes awarded the Vietnam Service Medal." Id.

In 2004, the DVA published a proposed rule, as part of a proposed wholesale revision of the DVA's regulations, in which it once again articulated its position with respect to offshore service. Citing the diabetes regulation, the DVA explained that veterans who served on the inland waterways of Vietnam "may have been exposed to herbicides" and that service on the inland waterways "constitutes service in the Republic of Vietnam" within the meaning of 38 U.S.C. § 1116. However, the agency restated that it was

not aware of any valid scientific evidence showing that individuals who served in the waters offshore of the Republic of Vietnam or in other locations were subject to the same risk of herbicide exposure as those who served within the geographic land boundaries of the Republic of Vietnam. Furthermore, we are not aware of any legislative history suggesting that offshore service or service in other locations are within the meaning of the statutory phrase, "Service in the Republic of Vietnam."

69 Fed.Reg. 44,614, 44,620 (July 27, 2004) (proposed rule). Accordingly, the DVA proposed to revise its regulation "to make it clear that veterans who served in waters offshore but did not enter Vietnam, either on its land mass or in its inland waterways cannot benefit from this presumption." Id.

The new benefits regulations, including the proposed rule regarding offshore service, have not yet been finally adopted. However, while this appeal was pending the DVA initiated a rulemaking proceeding that would amend section 3.307(a)(6)(iii) to incorporate the DVA's interpretation of the regulation as part of the regulatory text. The amended version of the regulation would define "service in the Republic of Vietnam" for purposes of section 3.307 to include "only service on land, or on an inland waterway, in the Republic of Vietnam during the period beginning on January 9, 1962, and ending on May 7, 1975." 73 Fed.Reg. 20,566, 20,571 (Apr. 16, 2008). In explaining the reason for the amendment, the agency referred to the litigation in this case and then stated that in its view the statutory reference to service in the Republic of Vietnam "is most reasonably interpreted to refer to service within the land borders of the Republic of Vietnam." Id. at 20,568. The agency explained its position as follows:

It is both intuitively obvious and well established that herbicides were commonly *1183 deployed in foliated land areas and would have been released seldom, if at all, over the open waters off the coast of Vietnam. The legislative and regulatory history indicates that the purpose of the presumption of exposure was to provide a remedy for persons who may have been exposed to herbicides because they were stationed in areas where herbicides were used, but whose exposure could not actually be documented due to inadequate records concerning the movement of ground troops.

Because it is known that herbicides were used extensively on the ground in the Republic of Vietnam, and because there are inadequate records of ground-based troop movements, it is reasonable to presume that any veteran who served within the land borders of Vietnam was potentially exposed to herbicides, unless affirmative evidence establishes otherwise. There is no similar reason to presume that veterans who served solely in the waters offshore incurred a significant risk of herbicide exposure.

Id. Although the DVA conceded that it was "conceivable that some veterans of offshore service incurred exposure under some circumstances due, for example, to airborne drift, groundwater runoff, and the proximity of individual boats to the Vietnam coast," it stated that for purposes of the presumption of exposure, "there is no apparent basis for concluding that any such risk was similar in kind or degree to the risk attending service within the land borders of the Republic of Vietnam." Id. Moreover, observing that offshore service "encompasses a wide range of service remote from land and thus from areas of actual herbicide use," the DVA concluded that "there is no reason to believe that any risk of herbicide exposure would be similarly pervasive among veterans of offshore service as among veterans of service within the land borders of Vietnam." Id.

B

We first address the government's argument that the pertinent language of 38 U.S.C. § 1116 is ambiguous and that the DVA's regulation issued pursuant to that statute, 38 C.F.R. § 3.307(a)(6)(iii), is entitled to deference as a permissible interpretation of the statute. Under the Chevron doctrine, "when an agency invokes its authority to issue regulations, which then interpret ambiguous statutory terms, the courts defer to its reasonable interpretations." Fed. Express Corp. v. Holowecki, ___ U.S. ___, 128 S.Ct. 1147, 1154, 170 L.Ed.2d 10 (2008); Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) (a court will defer to an agency's regulatory interpretation of a statute if the statute is ambiguous or contains a gap that Congress has left for the agency to fill through regulation). "Step one" of the Chevron analysis considers whether "Congress has directly spoken to the precise question at issue," a question that we analyze using the traditional tools of statutory interpretation. Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778; Cathedral Candle Co. v. Int'l Trade Comm'n, 400 F.3d 1352, 1362 (Fed. Cir.2005).

The relevant portion of section 1116(a)(1)(A) provides that for a veteran who suffers from one of several specified diseases, including type 2 diabetes, and who "during active military, naval, or air service, served in the Republic of Vietnam during the period beginning on January 9, 1962, and ending on May 7, 1975," the disease "shall be considered to have been incurred in or aggravated by such service." As applied to veterans who served in waters offshore of Vietnam but not on the landmass of Vietnam, the Veterans Court concluded that the statutory phrase *1184 "served in the Republic of Vietnam" is ambiguous.^[2]

The court first noted that "[t]here are many ways in which to interpret the boundaries of a sovereign nation such as the former Republic of Vietnam." 20 Vet. App. at 263. The court then surveyed different sources that define sovereign nations in different ways, ranging from including only the nation's landmass to including the nation's "exclusive economic zone," which can extend up to 200 miles from the coastline. Id. at 263-64. The government agrees with the Veterans Court that section 1116 is ambiguous in this respect. Mr. Haas, however, argues that the statute has a plain meaning that covers servicemembers in his position.

Addressing the phrase "served in the Republic of Vietnam," Mr. Haas asserts that "[a]ll relevant definitions of the sovereign nation of the Republic of Vietnam include the territorial waters off the landmass of Vietnam." To support that assertion, Mr. Haas cites to two definitions identified by the Veterans Court, Presidential Proclamation 5928 (1989) and the United Nations Convention on the Law of the Sea ("UNCLOS"). Both definitions include the nation's "territorial sea," which is generally defined as extending 12 nautical miles from a nation's coast. Yet Mr. Haas does not explain why other definitions, such as the contrary ones cited by the Veterans Court, are not "relevant." Neither the language of the statute nor its legislative history indicates that Congress intended to designate one of the competing methods of defining the reaches of a sovereign nation. We therefore agree with the Veterans Court that the statutory phrase "served in the Republic of Vietnam" is ambiguous as applied to service in the waters adjoining the landmass of Vietnam.

Based on a textual analysis of section 1116, Mr. Haas asserts that Congress made its intention clear that active duty personnel who served on ships offshore of Vietnam should be considered to have "served in the Republic of Vietnam" within the meaning of 38 U.S.C. § 1116(a)(1)(A). His argument is that if a veteran "served in the Republic of Vietnam" and has one of the diseases listed in section 1116(a)(2), such as diabetes, the veteran does not need to provide evidence that he or she was actually exposed to herbicides. By contrast, under section 1116(a)(1)(B), service connection is presumed only if the veteran "served in the Republic of Vietnam" and "while so serving was exposed to" an herbicide. Because proof of actual exposure is not required under section 1116(a)(1)(A), Mr. Haas argues that there is no reason to require proof of actual presence on the landmass of Vietnam. He contends that the government's asserted justification for the "foot-on-land" approach that herbicides are only sprayed on land is not relevant under section 1116(a)(1)(A), which by its terms does not require direct herbicide exposure.

*1185 Contrary to Mr. Haas's contention, the statutory provision that obviates the need to prove herbicide exposure for certain diseases neither says nor implies anything about the meaning of the phrase "served in the Republic of Vietnam." Congress simply concluded that for those who served in Vietnam, it was too difficult to determine who was exposed and who was not. But in so concluding, Congress did not indicate that service "in" the Republic of Vietnam included service on the waters offshore or in any other location nearby. Nor did Congress suggest that exposure was not important to the determination of service connection. The entire predicate for the Agent Orange Act and its regulations was exposure to herbicides in general and Agent Orange in particular. The fact that Congress presumed exposure for veterans who served in Vietnam does not by any means suggest that exposure was considered unimportant and that veterans in other areas therefore do not have to prove exposure. Thus, there is no force to Mr. Haas's argument based on the difference between section 1116(a)(1)(A) and section 1116(a)(1)(B).

Mr. Haas next contends that the legislative history of the Agent Orange Act demonstrates that Congress intended to give those who served only in offshore waters the benefit of section 1116(a). His argument is based on statements in the legislative history of the Agent Orange Act that Congress intended to codify the DVA's then-existing regulations on diseases meriting a presumption of service connection for Vietnam veterans. See, e.g., 137 Cong. Rec. 2345 (1991) (statement of Rep. Montgomery) ("This compromise would codify administrative decisions of the Secretary of Veterans Affairs in deeming three conditions service-connected for compensation purposes."); id. at 2352 (statement of Rep. Stump) ("H.R. 556 codifies current VA policy regarding agent orange compensation by establishing in statute a presumption of service-connection for non-Hodgkin's lymphoma, soft-tissue sarcoma, and chloracne.").

The problem with that argument is that the references to the regulatory presumptions in the legislative history did not distinguish between the broader definition of "service in Vietnam" provided in the non-Hodgkin's lymphoma regulation (section 3.313) and the narrower definition of "service in the Republic of Vietnam" found in the chloracne/soft tissue sarcoma regulation (section 3.311a). In the absence of any clearer statement in the legislative record, which Mr. Haas has not identified, the remarks about the existing regulations do not support the construction of the statutory phrase "served in the Republic of Vietnam" that he advocates. If anything, the different circumstances that prompted the issuance of the two regulations and the fact that only the chloracne/soft tissue sarcoma regulation used the precise phrase that was later incorporated into the statute "service in the Republic of Vietnam" (section 3.311a) rather than "service in Vietnam" (section 3.313) suggest the contrary conclusion. The chloracne/soft tissue sarcoma regulation was based on scientific evidence linking those diseases to dioxin exposure. The Agent Orange Act was similarly designed to provide compensation for exposure to Agent Orange. The non-Hodgkin's lymphoma regulation, by contrast, was not predicated on exposure, but instead was based on evidence of an association between non-Hodgkin's lymphoma and service in the Vietnam theater, including service aboard ships. Thus, the Agent Orange Act closely tracked the narrower chloracne/soft tissue sarcoma regulation, which defined "service in the Republic of Vietnam" to apply to those who served in the waters offshore *1186 only if their service included "duty or visitation in the Republic of Vietnam."

C

Having concluded that the phrase "served in the Republic of Vietnam" in section 1116 is ambiguous, we next turn to "step two" of the Chevron analysis, which requires a court to defer to an agency's authorized interpretation of the statute in question if "the agency's answer is based on a permissible construction of the statute." Chevron, 467 U.S. at 843, 104 S.Ct. 2778. We therefore address the DVA regulation that defines the phrase "service in the Republic of Vietnam" to mean "service in the waters offshore and service in other locations if the conditions of service involved duty or visitation in the Republic of Vietnam." 38 C.F.R. § 3.307(a)(6)(iii).

First, we note that Congress has given the DVA authority to interpret the statute, both under its general rulemaking authority, 38 U.S.C. § 501, and in the Agent Orange Act itself, 38 U.S.C. § 1116(a)(1)(B). Second, we agree with the Veterans Court that the regulation reflects a reasonable interpretation of the statute in that it requires some presence in Vietnam, even if the veteran's service largely occurred elsewhere.

The government contends that the regulation makes clear that service connection is presumed only for veterans who were at some point present on the landmass of Vietnam. We believe that is probably the most natural reading of the language of the regulation that refers to "duty or visitation in the Republic of Vietnam." That is, we agree with the government that "duty or visitation" in the Republic of Vietnam seems to contemplate actual presence on the landmass of the country. However, the question as to the meaning of the phrase "duty or visitation in the Republic of Vietnam" is not free from doubt, as "duty" or "visitation" could be understood to refer to "duty" or "visitation" within the broader area encompassed, for example, by the territorial waters of the Republic. Thus, both the phrase "duty or visitation in the Republic of Vietnam" and the phrase "waters offshore" are sufficiently ambiguous that the language of the regulation cannot be said to resolve the issue with certainty.

D

For that reason, we must look to the DVA's interpretation of its own regulation and determine whether that interpretation resolves the legal issue before us. Generally, "an agency's interpretation of its own regulations is controlling unless plainly erroneous or inconsistent with the regulations being interpreted." Long Island Care at Home, Ltd. v. Coke, ___ U.S. ___, 127 S.Ct. 2339, 2346, 168 L.Ed.2d 54 (2007) (internal quotations omitted); see also Auer v. Robbins, 519 U.S. 452, 461-63, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997). An agency's interpretation of its regulations is entitled to "substantial deference," requiring a court to defer to the agency's interpretation "unless an alternative reading is compelled by the regulation's plain language or by other indications of the [agency's] intent at the time of the regulation's promulgation." Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512, 114 S.Ct. 2381, 129 L.Ed.2d 405 (1994), quoting Gardebring v. Jenkins, 485 U.S. 415, 430, 108 S.Ct. 1306, 99 L.Ed.2d 515 (1988).

That rule does not apply if a particular regulation merely "parrots" statutory language, because if it did, an agency could bypass meaningful rule-making procedures by simply adopting an informal "interpretation" of regulatory language taken directly from the statute in question. See Gonzales v. Oregon, 546 U.S. 243, 257, 126 S.Ct. 904, 163 L.Ed.2d 748 (2006); Christensen *1187 v. Harris County, 529 U.S. 576, 588, 120 S.Ct. 1655, 146 L.Ed.2d 621 (2000) (an agency cannot "under the guise of interpreting a regulation ... create de facto a new regulation"). In this case, however, we are satisfied that the DVA regulation does more than merely parrot section 1116. The Supreme Court in Gonzales v. Oregon characterized the regulation in that case as a parroting regulation because it "just repeats two statutory phrases and attempts to summarize the others." 546 U.S. at 257, 126 S.Ct. 904. The Court added that the regulation "gives little or no instruction on a central issue." Id. By contrast, the regulation at issue in this case, 38 C.F.R. § 3.307(a)(6)(iii), elaborates on the statutory phrase "served in the Republic of Vietnam" by construing it to include service offshore and service in other locations as long as the service "involved duty or visitation in the Republic of Vietnam." That language qualifies as interpretation rather than reiteration.

The fact that the regulation is itself subject to competing interpretations, depending on whether it is read to require duty or visitation on land, as opposed to duty or visitation within Vietnam's territorial waters, does not mean that the regulation merely parrots the statute. It is not unusual for an interpretive regulation to be itself ambiguous; that happens, in fact, whenever a court is required to look to an agency's interpretation of a regulation that in turn interprets a statute. See, e.g., Auer, 519 U.S. at 461-63, 117 S.Ct. 905; Cathedral Candle Co., 400 F.3d at 1352, 1363-64. In such cases, courts do not disregard the regulation and its interpretation as long as the regulation reflects the agency's exercise of its interpretive authority and does not simply "restate the terms of the statute itself." Gonzales, 546 U.S. at 257, 126 S.Ct. 904; see id. at 256, 126 S.Ct. 904 (deference was accorded to the agency's interpretation in Auer because "the underlying regulations gave specificity to a statutory scheme the [agency] was charged with enforcing and reflected the considerable experience and expertise the [agency] had acquired over time...."). For these reasons, it is appropriate to defer to the DVA's asserted interpretation unless it is plainly erroneous or inconsistent with the regulations.

The Veterans Court concluded that it did not need to grant deference to the DVA's interpretation of section 3.307(a)(6)(iii) for several reasons: because the DVA's interpretation of the regulation has been inconsistent; because the DVA's interpretation was based on what the court considered plainly erroneous statutory analysis in a precedential opinion of the DVA's General Counsel; and because the court regarded the DVA's interpretation as unreasonable in that the agency has interpreted service in Vietnam differently under two different regulations and has failed to point to scientific evidence supporting its interpretation. We address each issue in turn.

1. The Veterans Court first decided that the DVA's current interpretation of section 3.307(a)(6)(iii) conflicts with the agency's prior interpretation of the regulation, and that the agency's current interpretation therefore merits less deference than it might otherwise deserve. We agree with the Veterans Court that there has been some inconsistency in the DVA's application of section 3.307(a)(6)(iii), but we do not agree that the DVA's inconsistency deprives the agency's interpretation of entitlement to deference, particularly in light of the fact that the agency has interpreted its regulation consistently for some years, going back to a time well before Mr. Haas filed the application for benefits that is at issue in this case.

*1188 For several years after the enactment of the Agent Orange Act and the corresponding regulations, the DVA did not formally interpret the regulatory reference to service "in the Republic of Vietnam." During that period the agency did not give any explanation of the meaning of the proviso requiring "duty or visitation in the Republic of Vietnam" in cases involving servicemembers whose principal service was in the waters offshore of Vietnam.

During that period, DVA adjudicators relied on the DVA's Adjudication Manual M21-1, which instructed DVA adjudicators on how to determine whether claimants had served "in the Republic of Vietnam." That 1991 version of Manual M21-1 provided as follows in pertinent part:

(1) It may be necessary to determine if a veteran had "service in Vietnam" in connection with claims for service connection for non-Hodgkin's lymphoma, soft-tissue sarcoma and chloracne.... In the absence of contradictory evidence, "service in Vietnam" will be conceded if the records shows [sic] that the veteran received the Vietnam Service Medal.

(2) If a veteran who did not receive the Vietnam Service Medal claims service connection for non-Hodgkin's lymphoma, soft tissue sarcoma or chloracne and alleges service on a ship in the waters offshore Vietnam, review the record for evidence that the ship was in the vicinity of Vietnam for some significant period of time (i.e., more than just in transit through the area). If the veteran cannot produce evidence that the ship was in the waters offshore Vietnam, contact the Compensation and Pension Service Projects Staff. Be prepared to furnish the name of the ship, the number of the ship, and the dates that it is alleged to have been in the waters offshore Vietnam. M21-1, part III, paragraph 4.08(k). The government contends on appeal, as it did in the Veterans Court, that the "contradictory evidence" mentioned in paragraph (1) has always included evidence that a veteran did not set foot in Vietnam. The Veterans Court concluded, however, that the second paragraph addressing the special case of veterans on board ships, which never mentions a foot-on-land requirement, would not have been necessary if the first paragraph had already implicitly contained a requirement that the veteran set foot on land in order to have "served in the Republic of Vietnam." 20 Vet.App. at 276.

We agree with the Veterans Court's analysis of the Manual M21-1 provision. The government's argument that the Manual provision incorporates the requirements of section 3.307(a)(6)(iii) simply reads too much into the "contradictory evidence" provision of Manual M21-1. In particular, the government's contention that M21-1 has always contained a "foot-on-land" requirement is unconvincing given that the Vietnam Service Medal was awarded to a broader class of service members than those who served on the landmass of Vietnam. See Exec. Order No. 11231 (July 8, 1965) (establishing award of the Vietnam Service Medal "to members of the armed forces who serve[d] in Vietnam or contiguous waters or air space").

Moreover, paragraph (2) of the Manual M21-1 provision, which refers to the possible need to review evidence that a veteran's ship was in the vicinity of Vietnam for some period of time, suggests that the Adjudication Manual did not exclude the possibility of benefits being granted to a veteran who never set foot in Vietnam. We therefore reject the government's suggestion that the DVA's current interpretation of the "service in the Republic of Vietnam" language in section *1189 3.307(a)(6)(iii) could be discerned from the outset in Manual M21-1.

Even though the 1991 version of the Manual and later versions issued on several occasions during the 1990s do not reflect the DVA's present interpretation of section 3.307(a)(6)(iii), the Veterans Court was nonetheless mistaken to conclude that the inconsistency between the early versions of the Manual and the agency's current interpretation of the regulation deprives the DVA's current interpretation of the right to judicial deference. As noted above, the DVA adopted its current interpretation of section 3.307(a)(6)(iii) in 1997. Since that time, it has reiterated its interpretation on numerous occasions, including by amending Manual 21-1 in 2002 to expressly incorporate the "foot-on-land" interpretation of the Agent Orange regulations and then formally rescinding the Manual provision in 2008. See 73 Fed.Reg. 20,363 (Apr. 15, 2008). Thus, any lack of clarity or inconsistency in the DVA's interpretation of the Agent Orange regulations has long since been resolved, and the "foot-on-land" policy is now firmly in place.

The DVA made its interpretation clear first in DVA General Counsel Opinion 27-97, the 1997 General Counsel opinion that ruled that sailors on deepwater vessels who did not set foot on land in Vietnam were not "in the Republic of Vietnam" within the meaning of 38 U.S.C. § 101(29)(A). In the course of analyzing section 101(29)(A), the opinion noted that the regulatory definition in 38 C.F.R. § 3.307(a)(6)(iii) "requires that an individual actually have been present within the boundaries of the Republic to be considered to have served there." The opinion concluded that the definition of "service in the Republic of Vietnam" in the regulation was consistent with the definition of the same phrase in section 101(29)(A), which the General Counsel interpreted to require physical presence on the landmass of Vietnam.

During the same year, the DVA set forth its interpretation of the regulatory language again in its response to comments on the spina bifida regulation. See 62 Fed.Reg. 51,274 (Sept. 30, 1997). The DVA explained that "[b]ecause herbicides were not applied in waters off the shore of Vietnam, limiting the scope of the term service in the Republic of Vietnam to persons whose service involved duty or visitation in the Republic of Vietnam limits the focus of the presumption of exposure to persons who may have been in areas where herbicides could have been encountered." More significantly for purposes of this case, in the very regulation that made type 2 diabetes the subject of presumed service connection (and thus provided the basis for Mr. Haas's claim), the DVA noted that service offshore does not constitute "service in the Republic of Vietnam." 66 Fed.Reg. 23,166, 23,166 (May 8, 2001).

To be sure, during the 1990s the DVA was not entirely consistent in its adjudications of claims arising under the Agent Orange Act. Mr. Haas cites four Board of Veterans' Appeals decisions that he contends support his position that a servicemember is entitled to presumptions of exposure to herbicides and service connection based on service offshore of Vietnam. The two earliest Board decisions support his argument, but the other two are at best unclear as to their interpretation of section 3.307(a)(6)(iii). For its part, the government cites a number of other decisions in which the Board applied the regulation as urged by the government, i.e., requiring proof of some duty or visitation onshore in Vietnam. The dates of the decisions cited by the government range from 1998 to 2005; both of the Board decisions that support Mr. Haas's position are from 1997.

*1190 While it is true that "[a]s a general matter ... the case for judicial deference is less compelling with respect to agency positions that are inconsistent with previously held views," Pauley v. BethEnergy Mines, Inc., 501 U.S. 680, 698, 111 S.Ct. 2524, 115 L.Ed.2d 604 (1991) (citing Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 212-13, 109 S.Ct. 468, 102 L.Ed.2d 493 (1988)), the DVA never formally adopted the position urged by Mr. Haas either in General Counsel opinions or in the rulemaking process. And even though the agency's current interpretation of its regulations differs from the position it took in some previous adjudications and seemed to take in its Adjudication Manual, that inconsistency does not mean that its current interpretation does not deserve deference. The Supreme Court made that point clear in its recent decision in Long Island Care at Home, 127 S.Ct. at 2349:

[W]e concede that the Department may have interpreted these regulations differently at different times in their history.... But as long as interpretive changes create no unfair surprise and the Department's recourse to notice-and-comment rulemaking in an attempt to codify its new interpretation ... makes any such surprise unlikely here the change in interpretation alone presents no separate ground for disregarding the Department's present interpretation.

See also Smiley v. Citibank (South Dakota), N.A., 517 U.S. 735, 742, 116 S.Ct. 1730, 135 L.Ed.2d 25 (1996) (change under the Chevron doctrine is "not invalidating, since the whole point of Chevron is to leave the discretion provided by the ambiguities of a statute with the implementing agency").

In this instance, the agency's position has been consistent for more than a decade, and there is "no reason to suspect that the interpretation does not reflect the agency's fair and considered judgment on the matter in question." Auer, 519 U.S. at 462, 117 S.Ct. 905. Moreover, because the agency adopted its current interpretation long before Mr. Haas filed his claim, and long before the statute and regulations were amended to include type 2 diabetes among the diseases entitled to special consideration, there is no issue of "unfair surprise" here. Accordingly, we conclude that the DVA's interpretation of section 3.307(a)(6)(iii) merits deference unless that interpretation is plainly erroneous or inconsistent with the language of the regulation.

2. The Veterans Court concluded that the DVA's interpretation of section 3.307(a)(6)(iii) is "plainly erroneous" in part because it is based on what the court regarded as flawed legal analysis in DVA General Counsel Opinion 27-97. As noted, that General Counsel opinion construes 38 U.S.C. § 101(29), a related statute that defines the term "Vietnam era" for purposes of title 38 and in the course of the discussion sets forth the DVA's interpretation of section 3.307(a)(6)(iii). We find nothing in the opinion's analysis that renders the DVA's interpretation plainly erroneous.

The General Counsel opinion examines the question whether veterans who served on deepwater Navy vessels in the vicinity of Vietnam between 1961 and 1975 are considered to have served "during the Vietnam era," as that phrase is used in 38 U.S.C. § 101(29). That question arose because the Veterans' Benefits Improvements Act of 1996 enlarged the statutory period of the "Vietnam era" to the period beginning on February 28, 1961, to May 7, 1975, "in the case of a veteran who served in the Republic of Vietnam during that period." Pub.L. No. 104-275, § 505, 110 Stat. 3322, 3342 (1996). The General Counsel opinion addresses whether service *1191 on an aircraft carrier would constitute service in the Vietnam era for purposes of section 101(29) during the period between February 28, 1961, and August 5, 1964, the period for which service "in the Republic of Vietnam" was required. DVA Op. Gen. Counsel Prec. 27-97 (1997). Focusing on legislative history that emphasized Congress's concern with ground troops who had been present on the landmass of Vietnam before August 1964, the General Counsel determined that service offshore was not included within the meaning of service "in the Republic of Vietnam."

Although the General Counsel opinion does not directly support the DVA's interpretation of section 3.307(a)(6)(iii), it makes clear that the agency viewed the regulatory definition of "service in the Republic of Vietnam" in section 3.307(a)(6)(iii) as closely parallel to the definition of that term in 38 U.S.C. § 101(29)(A). Having interpreted section 101(29)(A) as requiring actual service "within the borders of the Republic of Vietnam" during the pertinent period, i.e., on the landmass of Vietnam, the opinion noted that section 3.307(a)(6)(iii) also requires that individuals "not actually stationed within the borders of the Republic of Vietnam" have been "present within the boundaries of the Republic to be considered to have served there."

We do not agree with the Veterans Court that the General Counsel opinion was legally flawed. While it is true that the amendment to section 101(29)(A) was meant to encompass veterans who may have been at risk for exposure to herbicides prior to 1964, as the Veterans Court stated, the General Counsel opinion merely pointed out that in addressing soldiers who may have been exposed to herbicides during that time period, Congress's express focus was on ground troops. The opinion correctly noted that there was no indication in the legislative history that Congress intended for the definition of section 101(29)(A) to include service on a deep-water vessel off the shores of Vietnam within the scope of the phrase "served in the Republic of Vietnam."

What is particularly important about the General Counsel opinion is that it made clear at least as early as 1997 that the agency interpreted section 3.307(a)(6)(iii) to require presence on the landmass of Vietnam. We see nothing in the General Counsel opinion that renders that interpretation of section 3.307(a)(6)(iii) plainly erroneous.

3. The Veterans Court then found the DVA's interpretation of "service in the Republic of Vietnam" in 38 C.F.R. § 3.307(a)(6)(iii) to be unreasonable because it was not the product of "valid or thorough reasoning." 20 Vet.App. at 273.

First, the court criticized the DVA's interpretation of the phrase "service in the Republic of Vietnam" in section 3.307(a)(6)(iii) because it differs from the DVA's interpretation of the phrase "service in Vietnam" in the non-Hodgkin's lymphoma regulation, 38 C.F.R. § 3.313. 20 Vet.App. at 274. The court's criticism of that inconsistency, however, fails to account for the differences in language, scientific basis, and legal authorization between the two regulations. Section 3.307 (formerly section 3.311 a) was the regulatory predecessor of the Agent Orange Act; it was based on the Veterans' Dioxin and Radiation Exposure Compensation Standards Act, and it included diseases that had been found to be linked to herbicide exposure. Section 3.313, however, was based on the agency's more general authority to adopt regulations "with respect to the nature and extent of proofs and evidence ... in order to establish the right to benefits." 38 U.S.C. § 210(c) (1982). It was not based on herbicide exposure, but *1192 on a CDC study of the occurrence of non-Hodgkin's lymphoma in different groups of veterans, which was specifically found not to be related to herbicide exposure. See 55 Fed.Reg. 25,339 (June 21, 1990) (proposing section 3.313); 1990 CDC Study at 81, 125. Because the CDC study included veterans who served exclusively aboard ships that traveled off the coast of Vietnam among the tested group of Vietnam veterans, it made sense for section 3.313 to include those veterans as beneficiaries of the regulation. Under these circumstances, it was not unreasonable for the agency to interpret the two regulations differently.^[3]

Second, the Veterans Court also found the DVA's interpretation of section 3.307(a)(6)(iii) unreasonable based on the agency's failure to offer scientific evidence in support of the line it drew at the Vietnamese coast and the seeming arbitrariness of some results produced by that line. 20 Vet.App. at 274-75.

Due in part to problems of testing for herbicide exposure and in part to the difficulties in tracking troop movements, it has proved difficult to determine which groups of veterans were exposed to herbicides and to what extent. Congress and the DVA have therefore resorted to a line-drawing process that concededly does not closely track levels of actual exposure. Thus, Congress has determined that for certain diseases, all veterans who served for any period of time in Vietnam will be presumed to have established service connection, even if there is no showing that they were exposed to herbicides or were in areas of herbicide use. The DVA, required to draw a line where Congress's intention was unclear, has construed the statute not to extend presumed service connection to those who were in the Vietnam theater but who served only offshore or in other locations. The DVA has explained the rationale for its line-drawing, which is that Agent Orange was sprayed only on land, and therefore the best proxy for exposure is whether a veteran was present within the land borders of the Republic of Vietnam. In a statement accompanying its recent proposed amendment to section 3.307(a)(6)(iii), the DVA explained:

As a factual matter, our legislative interpretation accords with what is known about the use of herbicides during Vietnam. Although exposure data is largely absent, review of military records demonstrate[s] that virtually all herbicide spraying in Vietnam, which was for the purpose of eliminating plant cover for the enemy, took place overland.... Regarding inland waterways, Navy riverine patrols reported to have routinely used herbicides for clearance of inland waterways.... Blue water Navy service members and other personnel who operated off shore were away from herbicide spray flight paths, and therefore were not likely to have incurred a risk of exposure to herbicide agents comparable to those who served in foliated areas where herbicides were applied.

73 Fed.Reg. at 20,568. In light of that explanation, which accords with the position taken by the DVA for the past decade, *1193 and in the absence of evidence that the line drawn by the DVA is irrational, we are not prepared to substitute our judgment for that of the agency and impose a different line.

The Veterans Court pointed out that service on land could be fleeting and could occur far from the area where herbicides were used, while service on the water could include extended service in coastal waters close to areas where herbicides were used. Under the DVA's interpretation of its regulation, a servicemember in the first category would be entitled to a presumption of service connection for one of the designated diseases, while a servicemember in the second category would not, even though the second servicemember would seem intuitively more likely to have been exposed to herbicides than the first. 20 Vet.App. at 273.

There are no doubt some instances in which the "foot-on-land" rule will produce anomalous results. That is not surprising. Line-drawing in general often produces instances in which a particular line may be overinclusive in some applications and underinclusive in others. As the Supreme Court has explained, "any line must produce some harsh and apparently arbitrary consequences." Mathews v. Diaz, 426 U.S. 67, 83, 96 S.Ct. 1883, 48 L.Ed.2d 478 (1976). But just because some instances of overinclusion or underinclusion may arise does not mean that the lines drawn are irrational. See Vance v. Bradley, 440 U.S. 93, 108, 99 S.Ct. 939, 59 L.Ed.2d 171 (1979) (line-drawing is upheld even if the classification "is to some extent both underinclusive and overinclusive, and hence the line drawn by Congress is imperfect"); Mass. Bd. of Ret. v. Murgia, 427 U.S. 307, 314, 96 S.Ct. 2562, 49 L.Ed.2d 520 (1976) ("Perfection in making the necessary classifications is neither possible nor necessary.").

The asserted arbitrariness of the line-drawing done by the agency in this case is in part the result of Congress's decision to extend the presumption of service connection to all persons who served for any period and in any area within the Republic of Vietnam. Because that blanket rule provides a presumption of service connection to some persons who were unlikely to be exposed, it makes virtually any line-drawing effort appear unreasonable as applied to those who were outside of Vietnam but near enough to have had some chance of exposure.

In our view, it was not arbitrary for the agency to limit the presumptions of exposure and service connection to servicemembers who had served, for some period at least, on land. Drawing a line between service on land, where herbicides were used, and service at sea, where they were not, is prima facie reasonable. Moreover, the line drawn by the agency does not cut off all rights of sea-going veterans to relief based on claims of herbicide exposure, in that even servicemembers who are not entitled to the presumption of exposure are nonetheless entitled to show that they were actually exposed to herbicides, as Mr. Haas has endeavored to do in this case. See 38 C.F.R. § 3.309(e). The DVA's interpretation of section 3.307(a)(6)(iii) as excluding servicemembers who never set foot within the land borders of Vietnam thus was not unreasonable, and it certainly did not rise to the level of being "plainly erroneous or inconsistent with the regulation." Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 414, 65 S.Ct. 1215, 89 L.Ed. 1700 (1945); see Smith v. Nicholson, 451 F.3d 1344, 1349-51 (Fed.Cir.2006).

In an effort to demonstrate that the DVA's interpretation was not only unsupported by science but was contrary to scientific studies, Mr. Haas argues that servicemembers serving offshore could have *1194 been exposed to Agent Orange through several mechanisms, such as "runoff" carrying toxic chemicals into the sea, "spray drift" transporting toxins via the wind, and the shipboard consumption of drinking water produced by evaporative distillation. As support for the last of those contentions, he cites to a study conducted for the Australian Department of Veterans Affairs suggesting that Vietnam veterans of the Royal Australian Navy may have been exposed to herbicide compounds by drinking water distilled on board their vessels. Nat'l Research Ctr. for Envtl. Toxicology, Queensland Health Scientific Servs., Examination of the Potential Exposure of Royal Australian Navy (RAN) Personnel to Polychlorinated Dibenzodioxins and Polychlorinated Dibenzofurans via Drinking Water (Dec. 12, 2002).

The Australian study and the other cited sources were not part of the record below and were not considered either by the Veterans Court or by the DVA in its prior rulemaking proceedings. Judgments as to the validity of such evidence and its application to the particular problem of exposure to herbicides in Vietnam are properly left to Congress and the DVA in the first instance; this court is not the proper forum for an initial analysis of such evidence and its implications for the DVA's policies. We note, however, that in its most recent rulemaking proceeding the DVA made the following observations with respect to the Australian study:

VA scientists and experts have noted many problems with the study that caution against reliance on the study to change our long-held position regarding veterans who served off shore. First, as the authors of the Australian study themselves noted, there was substantial uncertainty in their assumptions regarding the concentration of dioxin that may have been present in estuarine waters during the Vietnam War.... Second, even with the concentrating effect found in the Australian study, the levels of exposure estimated in this study are not at all comparable to the exposures experienced by veterans who served on land where herbicides were applied.... Third, it is not clear that U.S. ships used distilled drinking water drawn from or near estuarine sources or, if they did, whether the distillation process was similar to that used by the Australian Navy.

73 Fed.Reg. 20,566, 20,568 (Apr. 16, 2008). Based on that analysis, the DVA stated that "we do not intend to revise our long-held interpretation of `service in Vietnam.'" Id. As to other cited studies, the DVA stated in connection with the publication of the rescission of the Manual M21-1 provision at issue in this case that none of those studies "bears significantly on the specific question whether herbicides used, and as administered, by the U.S. military during the Vietnam Era could have been blown by the wind into the ocean, or into inland waters that then carried the chemical into the ocean, to reach a boat offshore and result in any significant risk of herbicide exposure." 73 Fed.Reg. 20,363, 20,364 (Apr. 15, 2008).

Without reference to evidence, the Veterans Court stated that "it appears that these veterans serving on vessels in close proximity to land would have the same risk of exposure to the herbicide Agent Orange as veterans serving on adjacent land." 20 Vet.App. at 273. The dissenting judge in this court likewise concludes, also without reference to supporting evidence, that veterans such as Mr. Haas "have asserted a reasonable claim that they may have been exposed to herbicides." But focusing on the facts of Mr. Haas's claim, including his assertion that his ship was within 100 feet of the coast of Vietnam, does little to help answer the question of how the statutory phrase "served in the Republic of Vietnam" *1195 should be interpreted. The Veterans Court, for example, did not suggest what would constitute the proper interpretation of the statute, but merely concluded that the DVA's regulation "must be read to include at least service of the nature described by the appellant, that is, service in the waters near the shore of Vietnam." A standard such as "near the shore" is unmanageably vague, not to mention its lack of mooring in the statutory or regulatory language. By contrast, the DVA's interpretation is a plausible construction of the statutory language and it is based on a simple but undisputed factthat spraying was done on land, not over the water. Applying the substantial deference that is due to an agency's interpretation of its own regulations, we uphold the DVA's interpretation of section 3.307(a)(6)(iii).

E

Finally, the Veterans Court concluded that the pertinent provision of the DVA's Manual M21-1, although styled as an interpretation of the law, was actually a substantive rule that could not be changed without compliance with formal notice-and-comment rulemaking procedures. Accordingly, the Veterans Court concluded that the 2002 change in Manual M21-1, in which the DVA made clear that "service in the Republic of Vietnam" would not apply to servicemembers who had not visited the landmass of Vietnam, was not valid because the change was not effected through notice-and-comment rulemaking. 20 Vet. App. at 277. On appeal, the government contends that the Manual M21-1 provisions are properly viewed as interpretive rules, and thus could be changed by the agency without formal rule-making procedures.^[4]

Sections 4 of the Administrative Procedure Act ("APA"), 5 U.S.C. § 553, requires agencies to publish proposed rules in the Federal Register for notice and comment. Although that requirement does not apply by its terms to matters "relating to ... benefits," 5 U.S.C. § 553(a)(2), the "benefits" exception does not apply to rules and regulations promulgated by the DVA, 38 U.S.C. § 501(d). The DVA's rules relating to benefits are therefore subject to the notice and comment requirements of the APA. Importantly, however, those requirements do not apply to "interpretative rules, general statements of policy, or rules of agency organization, procedure or practice." 5 U.S.C. § 553(b)(3)(A). Because interpretive rules are not substantive rules having the force and effect of law, they are not subject to the statutory notice-and-comment requirements. See Shalala v. Guernsey Mem'l Hosp., 514 U.S. 87, 99, 115 S.Ct. 1232, 131 L.Ed.2d 106 (1995); Chrysler Corp. v. Brown, 441 U.S. 281, 301-02 & n. 31, 99 S.Ct. 1705, 60 L.Ed.2d 208 (1979).

While substantive rules are those that effect a change in existing law or policy or that affect individual rights and obligations, interpretive rules "clarify or explain existing law or regulation and are exempt from notice and comment under section 553(b)(A)." Paralyzed Veterans of Am. v. West, 138 F.3d 1434, 1436 (Fed. Cir.1998); see also Animal Legal Def. Fund v. Quigg, 932 F.2d 920, 927 (Fed.Cir.1991); Am. Hosp. Ass'n v. Bowen, 834 F.2d 1037, 1045 (D.C.Cir.1987). An interpretive rule "merely `represents *1196 the agency's reading of statutes and rules rather than an attempt to make new law or modify existing law.'" Nat'l Org. of Veterans' Advocates, Inc. v. Sec'y of Veterans Affairs, 260 F.3d 1365, 1375 (Fed.Cir. 2001), quoting Splane v. West, 216 F.3d 1058, 1063 (Fed.Cir.2000).

We conclude that the pertinent provision of Manual M21-1 is an interpretive statement, not a substantive rule. As the DVA has explained, Manual M21-1 "is an internal manual used to convey guidance to VA adjudicators. It is not intended to establish substantive rules beyond those contained in statute and regulation." 72 Fed.Reg. 66,218, 66,219 (Nov. 27, 2007). The provision at issue in this case did not set forth a firm legal test for "service in the Republic of Vietnam," but simply provided guidance as to how an adjudicator should go about gathering information necessary to determine whether the regulatory test had been satisfied. As such, the Manual provided reasonably easily applied guidance for adjudicators in an effort to obtain consistency of outcome; it did not define the boundaries of the DVA's legal responsibility with precision.

The 1991 version of Manual M21-1 noted that ordinarily the statutory and regulatory test would be satisfied by proof of receipt of the Vietnam Service Medal. The reference to the Vietnam Service Medal did not displace the legal test for service "in the Republic of Vietnam," but merely directed adjudicators to perform a simple initial analysis, which was sufficient to determine compliance with that test in the great majority of cases. For that reason, we conclude that the pre-2002 version of Manual M21-1 was not a substantive rule that could be amended only by notice-and-comment rulemaking.^[5] Indeed, to treat receipt of the Vietnam Service Medal as a "test" of eligibility for the statutory presumption would be clearly contrary to the Agent Orange Act, because it is undisputed that some servicemembers who received the Vietnam Service Medal were never either in Vietnam or in its territorial waters; accordingly, those servicemembers could not properly be regarded as having served "in the Republic of Vietnam" under any definition of that phrase.

Importantly, it was through notice-and-comment rulemaking that the DVA set forth its position with regard to offshore service in connection with the very regulation that is at issue in this case. In May 2001, the DVA issued the regulation in which it made type 2 diabetes a disease subject to the regulatory presumption of service connection. In so doing, the agency clearly set forth its view as to the status of servicemembers who had served in the waters off Vietnam and had not set foot on shore. Those servicemembers, the agency explained, were not within the scope of the regulatory presumption. See 66 Fed.Reg. 23,166 (May 8, 2001).

That regulation became effective in July 2001, a month before Mr. Haas filed his claim for service connection for diabetes. The agency had thus formally taken a position by then that excluded Mr. Haas from the scope of the regulation. The fact that the DVA did not also subject the amended version of Adjudication Manual M21-1, which followed the position taken in the 2001 rulemaking proceeding, to notice-and-comment rulemaking did not make the agency's actions with regard to *1197 Mr. Haas's claim unlawful. In sum, the agency's formal position with respect to the requirement of visitation or duty on land was established well before Mr. Haas's application for benefits and was reiterated in the diabetes rulemaking proceeding in May 2001. Contrary to the suggestion of the Veterans Court, it was not necessary for the agency to conduct a parallel rulemaking proceeding before incorporating the same rule into its more informal Adjudication Manual.

Because the DVA properly followed its established interpretation of statutory section 1116 and regulatory section 3.307(a)(6)(iii) when it rejected Mr. Haas's claim, we also disagree with the Veterans Court's ruling that the DVA's decision in Mr. Haas's case represents an impermissible retroactive application of the 2002 amendment to Manual 21-1. The agency's interpretation of the statute and regulation were clear by 2001, before Mr. Haas filed his claim. The fact that the agency waited until early 2002 to amend its internal Adjudication Manual to correspond with that interpretation did not prejudice Mr. Haas and does not confer any rights on him.

IV

For the foregoing reasons, we reverse the Veterans Court's ruling rejecting the DVA's interpretation of section 3.307(a)(6)(iii) of the agency's regulations as requiring the servicemember's presence at some point on the landmass or the inland waters of Vietnam. We remand to the Veterans Court for further proceedings consistent with this opinion. Before the Veterans Court on remand, Mr. Haas is free to pursue his claim that he was actually exposed to herbicides while on board his ship as it traveled near the Vietnamese coast. However, he is not entitled to the benefit of the presumptions set forth in 38 U.S.C. § 1116 and the corresponding DVA regulations, which are limited to those who "served in the Republic of Vietnam."

Each party shall bear its own costs for this appeal.

REVERSED and REMANDED.

FOGEL, District Judge, dissenting.

Although I agree with much of the majority's thorough analysis of the relevant legislative and regulatory history, I respectfully disagree with its ultimate holding. Because I conclude that the VA's refusal to apply the presumption of 38 U.S.C. § 1116(a) to Haas and others similarly situated is inconsistent with the intent of the statute and thus is based upon an unreasonable interpretation of the subject regulation, I would affirm the judgment of the Veterans Court. See Haas v. Nicholson, 20 Vet.App. 257 (2006).

While judicial deference to the experience and expertise of administrative agencies is an important principle of our jurisprudence, the historical context in which both courts and agencies act also is important. The present case is the latest skirmish in a decades-long dispute between Vietnam-era veterans and the VA over the health effects of Agent Orange. In 1984, Congress enacted the Veterans' Dioxin and Radiation Exposure Compensation Standards Act, Pub.L. No. 98-542, 98 Stat. 2725 (1984) ("Dioxin Act"), the purpose of which was "to ensure that Veterans' Administration disability compensation [was] provided to veterans who were exposed during service in the Armed Forces in the Republic of Vietnam to a herbicide containing dioxin...." Id. Following its enactment, a group of Vietnam-era veterans and surviving spouses brought suit against the VA for its alleged failure to comply with the Act's provisions. Nehmer v. U.S. Veterans *1198 Admin., 712 F.Supp. 1404 (N.D.Cal. 1989).

Specifically, the veterans challenged the VA's final rule, 38 U.S.C. § 3.311a(d), which stated that "`sound scientific and medical evidence does not establish a cause and effect relationship between dioxin exposure' and any other disease but chloracne." Nehmer, 712 F.Supp. at 1408. The district court held that the "cause and effect test" employed by VA in 38 C.F.R. § 3.311a(d) to determine the relationship between dioxin exposure and various diseases was inconsistent both with the VA's prior practice and with the purpose of the Act. Nehmer, 712 F.Supp. at 1418. In reaching this conclusion, the court relied on the statement of one of the Act's principal supporters, Senator Alan Simpson, that the "[Dioxin] Act was intended to ensure that veterans `have their exposure claims adjudicated under uniform and consistent regulations that incorporate rational scientific judgments', as opposed to the prior system, in which the claims are `committed to the sound judgment of the VA's adjudication officers' who decide them on `a case-by-case basis.'" Id. at 1422.

The statute at issue in this case, the Agent Orange Act, Pub.L. No. 102-04, 105 Stat. 11 (1991), was adopted subsequent to and informed by the issues raised in Nehmer. The Agent Orange Act required that the National Academy of Sciences conduct a comprehensive review of "all the available and future evidence on the long-term health effects of exposure" to herbicides. Haas, 20 Vet.App. at 268. It codified, in similar form, the 1984 note to 38 U.S.C. § 354, which the Dioxin Act amended, at 38 U.S.C. § 316(a)(3), which provided:

For the purposes of this subsection, a veteran who, during active military, naval, or air service, served in the Republic of Vietnam during the Vietnam era and has a disease referred to in paragraph (1)(B) of this subsection shall be presumed to have been exposed during such service to an herbicide agent containing dioxin or 2,4-dichlorophenoxyacetic acid, and may be presumed to have been exposed during such service to any other chemical compound in an herbicide agent, unless there is affirmative evidence to establish that the veteran was not exposed to any such agent during that service.

See Haas, 20 Vet.App. at 268.

As the majority points out, the legislative history of the Agent Orange Act is silent as to what constitutes "service in the Republic of Vietnam." However, both the legislative history and the language of the statute itself indicate the intent of Congress that a fair and independent system be established to determine the relationship between herbicide exposure and the manifestation of certain diseases. Congress was seeking to make it easier, not more difficult, for Vietnam veterans to assert claims arising from exposure to Agent Orange. In this context, it is reasonable to expect that an administrative interpretation limiting the benefits of the presumption at issue here would be based on at least some scientific evidence.

I agree with the majority that in the present case the VA's interpretation of its own regulation is entitled to controlling weight unless that interpretation is plainly erroneous or inconsistent with the regulation. Majority Opinion, slip op. at ___ However, an interpretation is reasonable only if it "`sensibly conforms to the purpose and wording of the regulations.'" Martin v. Occupational Safety & Health Review Comm'n, 499 U.S. 144, 150-51, 111 S.Ct. 1171, 113 L.Ed.2d 117 (1991) (quoting N. Indiana Pub. Serv. Co. v. Porter County Chapter of Izaak Walton League of Am., Inc., 423 U.S. 12, 15, 96 S.Ct. 172, 46 L.Ed.2d 156 (1975) (emphasis added)). *1199 I agree with the Veterans Court that in the absence of any scientific evidence in the record that supports a "foot on land" requirement, the VA's position is unreasonable.

Congress created the presumption at issue both because exposure to Agent Orange could not be determined by tracking troop movements and because the VA could not pinpoint which veterans were deployed at or near locations where Agent Orange was sprayed, facts which as a practical matter made it very difficult for veterans to prove their claims. Although the plain purpose of the statute is to ensure that all veterans who risked exposure have their claims adjudicated in accordance with uniform, scientifically-based standards, the "foot on land" requirement arbitrarily excludes from the benefits of the statutory presumption an identifiable group of veterans who the available evidence suggests risked exposure.

For example, the VA's interpretation grants the presumption to a veteran who served on a vessel that traveled on inland waterways but not to a veteran who served on a vessel in the waters immediately off the coast of Vietnam, even at no greater distance from land. A veteran whose only contact with Vietnam was a one-hour stop at an airfield would have the benefit of the presumption, while a veteran who spent months on a coastal patrol boat would not. Citing to the administrative record, the Veterans Court noted that "[u]sing VA's risk-of-exposure test outlined in its June 2001 notice of final rulemaking, given the spraying of Agent Orange along the coastline and the wind borne effects of such spraying, it appears that these veterans serving on vessels in close proximity to land would have the same risk of exposure to the herbicide Agent Orange as veterans serving on adjacent land, or an even greater risk than that borne by those veterans who may have visited and set foot on land of the Republic of Vietnam only briefly." Haas, 20 Vet. App. at 273. The Veterans Court concluded that "[t]he Secretary has provided no rational distinction between these types of service and the Court can divine none." Id. Appropriately, the Veterans Court held that:

Absent any discussion regarding the scientific studies mandated by Congress on this subject or any other evidence that contributed to VA's decision to limit the definition, the Court can only conclude that VA's asserted interpretation of this regulation is not the product of agency expertise.

Id. at 275.

Perhaps anticipating that this Court might equally be concerned with the absence of relevant scientific evidence, the VA submitted to the Court during the pendency of this appeal proposed amendments to the regulation that expressly adopt the "foot on land" test and explain the agency's rationale for the amendments. The VA acknowledges the possibility that some veterans who were deployed immediately offshore may have been exposed to herbicides but at the same time asserts there is no evidence that the risk of such exposure was comparable to that faced by veterans who were deployed on land. The VA reaches this conclusion not on the basis of any affirmative data but by discounting the findings of the Australian study upon which Haas and others similarly situated rely. Like the VA's most recent interpretation of the regulation, the proposed amendments appear to be based on uncertainty rather than the careful scientific assessment required by the statute. Thus, despite the clarifying language, I remain convinced that the VA's interpretation is not entitled to deference.

The majority concludes that the "foot on land" rule is rational because there appears to be no clear scientific evidence defining the extent to which different *1200 groups of veterans were exposed, leaving the task of line-drawing to Congress and the VA. Majority Opinion at 1192. Indeed, an interpretation that excludes veterans whose only contact with the Republic of Vietnam was a high-altitude flyover or service in deep offshore waters would be perfectly sensible, as such individuals would not have had a potential risk of exposure. See DVA Op. Gen. Counsel Prec. 27-97 (1997) (finding that service in a deepwater vessel off the shore of Vietnam did not constitute "service in the Republic of Vietnam" under 38 U.S.C. § 101(29)(A)); DVA Op. Gen. Counsel Prec. 7-93 (1993) (finding that service in high altitude planes flying over Vietnam without any further contact with Vietnam did not constitute "service in the Republic of Vietnam" under 38 C.F.R. § 3.313). However, veterans like Haas who have asserted a reasonable claim that they may have been exposed to herbicides deserve to have such claims "adjudicated under uniform and consistent regulations that incorporate rational scientific judgments." See Nehmer, 712 F.Supp. at 1422.^[1] It is the VA's burden, not the veterans', to show that the VA's line-drawing was both informed by scientific evidence and consistent with the remedial purposes of the statute. Because I agree with the Veterans Court that the VA has not met that burden, I respectfully dissent.

NOTES

[*] Honorable Jeremy Fogel, District Judge, United States District Court for the Northern District of California, sitting by designation.

[1] Congress included non-Hodgkin's lymphoma on the list of diseases specifically identified in the Agent Orange Act based on evidence that, contrary to the conclusion of the 1990 CDC study, non-Hodgkin's lymphoma was in fact associated with exposure to Agent Orange. See Report to the Secretary of Veterans Affairs on the Association Between Adverse Health Effects and Exposure to Agent Orange, reprinted in Links Between Agent Orange, Herbicides, and Rare Diseases: Hearing before the Human Resources and Intergovernmental Relations Subcomm. of the Comm. on Gov't Relations, 101st Cong., 2d Sess. 22, 41 (1990).

[2] In its brief, the government mistakenly refers to section 1116(f) as the provision at issue in this case. Because Mr. Haas's disease is one of those listed in section 1116(a)(2), it is section 1116(a)(1), not section 1116(f), that governs his claim. Section 1116(f) was originally enacted as subsection (a)(3) of the first section of the Agent Orange Act, and it applied to diseases referred to in subsection (a)(1)(B). When the Act was amended in 2001, subsection (a)(3) became section 1116(f), and it was modified to apply to diseases other than those referred to in subsections (a)(1) or (a)(2). The legislative history of the 2001 amendment makes it quite plain that the new section 1116(f) was designed to make the Act applicable to new diseases, not to affect the preexisting scope of subsection (a)(1). S.Rep. No. 107-86, at 10-12 (2001). The erroneous reference makes no difference to the analysis in this case, however, as the pertinent phrase "served in the Republic of Vietnam" appears in both sections 1116(a)(1) and 1116(f).

[3] Mr. Haas argues that the non-Hodgkin's lymphoma regulation, section 3.313, not the general dioxin exposure regulation, section 3.311 a, was the true predecessor to section 3.307(a)(6)(iii). That contention is plainly wrong. When proposing section 3.307(a)(6)(iii), the Secretary of Veterans Affairs specifically stated that the definition of "service in the Republic of Vietnam" was taken from section 3.311 a, see 58 Fed.Reg. 50,528, 50,529 (Sept. 28, 1993), and the text of the two regulations is virtually identical (and significantly different from the text of section 3.313).

[4] As we have noted, while not changing its legal position the DVA has recently acted to obviate this issue for the future by publishing a formal notice in the Federal Register rescinding the pertinent provision of Manual M21-1. See 73 Fed.Reg. 20,363 (Apr. 15, 2008).

[5] Mr. Haas argues that Fugere v. Derwinski, 972 F.2d 331 (Fed.Cir. 1992), supports his argument regarding the amendment of Manual M21-1. In that case, however, the only issue before this court was whether a provision of the Manual conflicted with a statute. This court did not address whether the Manual provision in question constituted a substantive rule that could be amended only through notice-and-comment rulemaking.

[1] The majority notes that the Veterans Court did not cite any specific record evidence in support of Haas's position and opines that any interpretation other than the "foot on land" test would be "unmanageably vague." Majority Opinion at 1195. Haas received the Vietnam Service Medal for his service in the Republic of Vietnam. As the Veterans Court pointed out and as the majority acknowledges, id. at 1187-88, the VA itself previously applied the presumption in cases in which a veteran received the Vietnam Service Medal or the veteran's "ship was in the vicinity of Vietnam for some significant period of time." See Haas,, 20 Vet.App. at 271-272 (citing M21-1, part III, paragraph 4.08(k)(1)-(2)). I have no reason to doubt that the VA could develop a manageable and consistent standard that would include veterans such as Haas.

Jonathan L. HAAS, Claimant-Appellee, v. James B. PEAKE, M.D., Secretary of Veterans Affairs, Respondent-Appellant.

I

II

III

A

B

C

D

E

IV

NOTES

Jonathan L. HAAS, Claimant-Appellee,
v.
James B. PEAKE, M.D., Secretary of Veterans Affairs, Respondent-Appellant.