dissenting:
I am unable to join in the majority’s opinion for several reasons. First, the Board manifestly applied the wrong standard for review of a claim reopened under 38 U.S.C. § 5108 (formerly § 3008). Second, I do not accept the conclusion that the failure of the Board of Veterans’ Appeals (BVA or Board) to comply with the statutory command, under 38 U.S.C. § 7104(d)(4) (formerly § 4004), that it provide “a written statement of reasons or bases” for its findings and conclusions is harmless error in this case. Third, the Board failed to discuss or apply the applicable regulation regarding the award of service connection when “a chronic condition is shown to exist within a short time following the applicable *547presumptive period”. 38 C.F.R. § 3.307(c) (1991).
I. ANALYSIS
A. Reopened Claim
The majority concludes that “the Board determined that the new evidence evaluated with the old evidence did not present a new factual basis on which to base the claim.” Ante at 545 (emphasis added). I respectfully disagree. In my view, the Board decision clearly failed to evaluate the new evidence “in the context of the other evidence of record and make new factual determinations”, as is required by Jones v. Derwinski, 1 Vet.App. 210, 215 (1990). See also Godwin v. Derwinski, 1 Vet.App. 419, 425 (1991). First, as a prelude to outlining some of the “old” evidence, the Board stated: “Evidence from a prior Board decision will be set forth, in part, for purposes of clarification.” Kunta K. Soyini, BVA 90-02483, at 2 (Jan. 30, 1990) (emphasis added). Then, in its “DISCUSSION AND EVALUATION”, the Board referred to the evidence submitted “[sjince the Board’s earlier decision” and specifically found as “Findings of Fact”:
5. The additional evidence obtained since the Board ... decision in February 1988 does not establish the onset of a chronic acquired psychiatric disorder during service, or manifestations of a psychosis within a year following separation from service.
6. The additional evidence is not sufficient to alter the factual basis upon which the prior decision was predicated.
Id. at 7. (emphasis added). A fair reading of the BVA decision, in light of the above quotations, yields the conclusion that the Board evaluated only the new evidence. That mistake is not reversible error, however, if there was not, as a matter of law, new and material evidence justifying the reopening of the claim. See Thompson v. Derwinski, 1 Vet.App. 251, 253, 253-54 (1991); Godwin, at 425. Although the majority does not discuss this issue, I must do so because of my conclusion that the Board applied the wrong standard in evaluating only the new evidence.
In order to justify reopening the claim, the appellant in 1988 submitted medical records showing treatment at the Connecticut Department of Corrections from May 17, 1974, less than a year after his discharge (on June 19, 1973), until March 14, 1975. It was on the basis of this evidence that the Regional Office (RO) of the Veterans’ Administration (now the Department of Veterans Affairs) (VA) found that “the early manifestations of the veteran’s schizophrenia occurred in 1974.” R. at 173. The VARO found that such manifestations occurred in November 1974 when the veteran was treated for “borderline schizophrenia” based on “complaints of hearing his dead grandmother’s voice and placed on psychotropic medication.” Ibid. The RO thus granted service connection for medical treatment purposes only, pursuant to 38 U.S.C. § 1702 (formerly § 602) which provides that “any veteran of ... the Vietnam era who developed an active psychosis (1) within two years after discharge or release from ... service, and (2) ... before May 8, 1977, ... shall be deemed to have incurred such disability in ... service.”
In contrast, the presumptive period for disability compensation under 38 U.S.C. § 1112(a)(1) (formerly § 312) and 38 C.F.R. §§ 3.307(a)(3), 3.309(a) (1991) for “psychoses” is one year. Hence, unless the 1974-75 medical records contain “new and material” evidence that could trigger the application of that presumption, the claim for compensation would not have been properly reopened and the failure to apply the correct standard for review under Jones and Godwin would be harmless error under Thompson and Godwin. I conclude that such evidence was, indeed, submitted.
To be “new and material” under section 5108, the evidence must be more than cumulative (“new”), see Colvin v. Derwinski, 1 Vet.App. 171, 173-74 (1991), and “there must be a reasonable possibility that the new evidence, when viewed in the context of all the evidence, both new and old, would change the outcome” (“material”). Colvin, at 174; Smith v. Derwinski, 1 Vet.App. 178, 179-80 (1991); Godwin, at 424-25. Here there is no question that the 1974-75 *548medical records were “new”, since no evidence was previously of record with respect to that 1974-75 period. They were also “material” for two reasons. First, an entry for May 17, 1974, shows: “Plan Hal-dol 1 mg B/D x 3 D”. R. at 149. This apparently refers to 1 milligram of Haldol to be administered daily. Haldol is a drug which the Physicians’ Desk Reference 1282 (44th ed. 1990) identifies as intended to be provided for adults for “use in the management of manifestations of psychotic disorders ... [and] for the control of tics and vocal utterances of Tourette’s syndrome”. The next entry in those records is for September 30, 1974, and shows “RX ... Haldol 2 mg BID X 5 D”. R. at 149. This seems to indicate that Haldol was prescribed on September 30, 1974, at almost 4 times the daily dose identified in the May 17 entry.
The May 17 entry is certainly susceptible of several interpretations. One is that it was only a medication “plan” which was never carried out. However, an equally reasonable inference would be that that level of medication was prescribed and provided at some point prior to September 30 when it was determined that a far heavier dose was necessary. It may also be reasonable under the circumstances, giving the veteran the benefit of the doubt on this issue of material fact under 38 U.S.C. § 5107(b) (formerly § 3107), as Colvin, at 174, requires be done in determining whether new evidence is “material”, to construe that May 17 entry as evidence that the drug was prescribed on or shortly after May 17, since the entry of a specific dosage is more consistent with a prescription than a “plan” for one. That is, if the veteran’s symptoms on May 17 warranted 1 milligram of Haldol 3 times a day, it would seem unusual for a physician to set forth a “plan” for using that dosage at some unspecified point in the future without regard to the veteran’s symptoms at that later point. Indeed, the large increase in the Haldol dosage on September 30 supports this interpretation.
If the fact finder were, then, to conclude that Haldol was prescribed on May 17, 1974, 11 months after the veteran’s discharge, and that the most likely basis for the prescription was symptoms consistent with psychosis, the fact finder could also determine that this was evidence warranting the application of the following provisions of 38 C.F.R. § 3.307(c) (1991):
[It is not required] that the disease be diagnosed in the [one year] presumptive period, but only that there be then [during that period] shown by acceptable medical or lay evidence characteristic manifestations of the disease to the required degree, followed without unreasonable time lapse by definite diagnosis. Symptomatology shown in the prescribed period may have no particular significance when first observed, but in the light of subsequent developments it may gain considerable significance.
To prevail under this regulation, the appellant would also have to show that the psychosis was manifested to a ten-percent degree, but the level of the prescription might itself be sufficient for that purpose. I would thus hold that “there is a reasonable possibility” that the 1974-75 medical records “could change the outcome” under Colvin, Smith, and Godwin and hence were “material” under section 5108.
Alternatively, I would hold that that evidence is also “material” under the very next sentence of the regulation quoted above:
Cases in which a chronic condition is shown to exist within a short time following the applicable presumptive period, but without evidence of manifestations within the period, should be developed to determine whether there was symptoma-tology which in retrospect may be identified and evaluated as manifestation of the chronic disease to the required 10-percent degree.
38 C.F.R. § 3.307(c) (1991). The treatment in November 1974 for assumed “borderline schizophrenia” could be construed by the fact finder as evidence that “a chronic condition is shown to [have existed] within a short time [5 months] following [the close of] ‘the applicable presumptive peri-od_’” Ibid. The Haldol prescription plan entry of May 17, 1974, could then *549reasonably be considered by the fact finder as evidence of symptomatology of the schizophrenia within the presumptive period. The same could be said regarding the attempted suicides in prison on January 9 and 13, 1975. R. at 164-65. Alternatively, the behavior in 1973 during service (drug overdose and antisocial behavior (R. at 52, 54)), might reasonably be construed as that earlier manifestation. Under this analysis, I would hold that there was a “reasonable possibility” that this evidence could “change the outcome”.
Indeed, since the VARO concluded that an “active psychosis” had been manifested as of November 1974, it would seem difficult for the BVA to conclude otherwise as to that issue. That would seem to be the law of the case as to when active psychosis was present in this veteran.
Accordingly, I conclude that the evidence was new and material for the purpose of reopening the claim under section 5108.1
I stress that my position is not that the outcome should be changed, or that the failure to change it would be a clearly erroneous fact determination by the Board. Rather, I am merely applying the Colvin standard, and its injunction that “the ‘benefit of the doubt’ standard ... necessarily lowers the threshold of whether the new and material evidence is sufficient to change the outcome”, Colvin, at 174, to the facts of this case in order to decide the question of law as to materiality. I am not, as the majority asserts, ante at 545-46, “ ‘bootstrapping]’ the appellant into the one year presumption found in 38 C.F.R. § 3.307.” Nor am I finding the facts. I am attempting to discern what would have been a plausible reading of the evidence for the purposes of determining the materiality of the new evidence.
B. Failure to State “Reasons or Bases”
Having concluded that the claim was properly reopened, I move to the issue of the Board’s compliance with the applicable law and regulations.
As this Court stated in Moore v. Derwinski, 1 Vet.App. 401, 404 (1991):
Beginning with Gilbert v. Derwinski, 1 Vet.App. 49, 56-57 (1990), this Court has consistently held that in order to enable a claimant to understand a decision and the reasons behind it, as well as to assist in judicial review, the BVA is required by 38 U.S.C. § 7104(d)(1) to include in its decision “a written statement of [its] findings and conclusions, and the reasons or bases for those findings and conclusions”. Gilbert, at 56; [citations omitted].... In providing its “reasons or bases”, the Board must include in its decisions “the precise basis for that decision ... [and] the Board’s response to the various arguments advanced by the claimant.” Gilbert, 1 Vet.App. at 56; see also Hatlestad, 1 Vet.App. at 169. This must include “an analysis of the credibility or probative value of the evidence submitted by and on behalf of the veteran in support of [his or her] claim [and] a statement of the reasons or bases for the implicit rejection of this evidence by the Board.” Gilbert, 1 Vet.App. at 58; see also Hatlestad [v. Derwinski ], 1 Vet.App. [164] at 169 [1991].
Here, the Board’s decision, as apparently is conceded by the majority, ante at 541, 546, failed to state with the requisite clarity the reasons or bases for several of its findings or implicit findings of material fact. First, what was the Board’s reason or basis for apparently deciding that the legal conclusion by the RO that the veteran manifested an active psychosis in November 1974, 16 months after service, was not evidence of a “definite diagnosis” of the psychosis “within a short time following the ... presumptive period”, or “without unreasonable time lapse” after manifestation of “characteristic manifestations of the disease”, under the provisions of 38 C.F.R. § 3.307(c) (1991)? As part of the same issue under the regulation, what was the Board’s rea*550son or basis for implicitly concluding that the in-service behavior or Haldol prescription plan were not evidence of “symptoma-tology which in retrospect may be identified and evaluated as manifestation of the chronic disease to the required ten-percent degree”? Ibid.
More fundamentally, if the Board failed to consider this evidence, it erred under 38 U.S.C. § 7104(a) and Gilbert and its progeny because the Board is required to consider “all evidence and material of record”. See also 38 C.F.R. § 3.303(a) (1991) (“Determinations as to service connection will be based on review of the entire evidence of record”). And the Board also erred in not including in its decision its “analysis of the ... probative value” of this evidence and of “the reasons or bases for [its] implicit rejection of [it].” Gilbert, at 58-59.
The majority here finds, “based upon the evidence [supporting the BVA decision] in this case”, that such “reasons or bases” deficiencies are “not prejudicial”. Ante at 541. I disagree. First, I do not agree that the evidence is “overwhelming ... in support of the result [reached by the BVA]”_ Ante at 546. For the reasons outlined above, I believe that a reasonable interpretation of the evidence could support the opposite conclusion. Moreover, this Court has not, until this case, held that a failure to comply with the statutory “reasons or bases” requirement can be “harmless” error under 38 U.S.C. § 7261(b) (formerly § 4061). I believe our precedents suggest the opposite conclusion. In Gilbert, at 59, after outlining the extensive scope of the “reasons or bases” requirements, the Court stated that “strict adherence by the Board to the requirements of 38 U.S.C. § [7104](d)(l) is required.” And in Sammarco v. Derwinski, 1 Vet.App. 111, 113-14 (1991), we stressed that:
Whether the BVA’s ultimate conclusions are correct or not, we hold that the incomplete nature of the decision below does not permit proper review by this Court.
Indeed, in Gilbert itself, after remand, the Board’s decision adverse to the appellant was affirmed by the Court, stating “the Board’s Supplemental Decision of April 5, 1991, provided ample reasons and bases for the denial of service connec-tion_”. Gilbert v. Derwinski, 1 Vet.App. 49, 62 (per curiam) (1991). Against this background, I would be reluctant to find that a “reasons or bases” deficiency is not prejudicial. The requirement for a reasoned, written BVA decision is a statutory right fundamental to the essence of the BVA fact-finding process. In Gilbert, the Court repeated at least four times that the “reasons or bases” statutory requirement is based on the need “to explain both to the veteran and to this Court”, to enable the veteran “to understand the reason for the denial of [his or her] claim”, to be “helpful to the veteran”, and to “ ‘enable a claimant to understand, not only the Board’s decision but also the precise basis for that decision ... [and] the Board’s response to the various arguments advanced by the claimant.’ ” Gilbert, at 59, 57, 56 (emphasis added).
In any event, I am unable to agree with a determination that the Gilbert violation was not prejudicial error on the facts of this case.
C. Failure to Apply Applicable Regulation
I also dissent because the facts asserted by the appellant require consideration of 38 C.F.R. § 3.307(c) and the Board did not consider that regulation. I have discussed above in parts A and B how that regulation might be applicable under various reasonable interpretations of the facts. “Even if not raised by the appellant, the Court has consistently ruled that the BVA is not free to ignore its own regulations.” Karnas v. Derwinski, 1 Vet.App. 308, 313 (1991). See also Peyton v. Derwinski, 1 Vet.App. 282, 286-87 (1991) (semble); Payne v. Derwinski, 1 Vet.App. 85, 87 (1990); cases cited in Karnas, at 313. That our cases have consistently imposed this obligation on the Board regarding VA regulations should be no surprise. The law expressly requires the BVA to base its decisions “upon consideration of all ... applicable provisions of ... regulation”, which are binding on the Board as construed by the precedent opin*551ions of the VA General Counsel. 38 U.S.C. § 7104(a), (c) (formerly § 4004). Moreover, “[t]here is no requirement in the law that a veteran must specify with precision the statutory provisions or the corresponding regulations under which [the veteran] is seeking benefits.” Akles v. Derwinski, 1 Vet.App. 118, 121 (1991).
II. CONCLUSION
For the above three reasons, I do not agree with the Court’s decision to affirm the BVA decision in this case. Rather, the BVA decision should be vacated and the matter remanded for readjudication.
. Under the majority’s analysis of the record as including "no evidence that [appellant’s] in-service diagnosis of personality disorder was a misdiagnosis or that he was suffering from a compensable psychosis within the requisite period", ante at 546, no “material” evidence was submitted and, therefore, the BVA decision should be affirmed because the claim was not properly reopened.