Opinion of the Court
EVERETT, Chief Judge:Tried by a general court-martial composed of members, Major Benedict was found guilty of three specifications of conduct unbecoming an officer and a gentleman by taking indecent liberties with a *254female child, in violation of Article 133, Uniform Code of Military Justice, 10 U.S.C. § 933. His sentence was dismissal, confinement for 3 years, and total forfeitures. After approval by the convening authority and affirmance of the findings and of the sentence except for confinement exceeding 18 months by the Court of Military Review, 20 M.J. 939 (1985), we granted review of these four issues:
I
WHETHER THE MILITARY JUDGE ERRED BY PERMITTING, OVER DEFENSE OBJECTION, PHYSICIAN’S TESTIMONY AMOUNTING TO ULTIMATE LEGAL CONCLUSIONS, UNDER THE GUISE OF MEDICAL DIAGNOSIS.
II
WHETHER THE MILITARY JUDGE ERRED BY PERMITTING, OVER DEFENSE OBJECTION, TESTIMONY IN THE FORM OF A MAGAZINE ARTICLE WHICH WAS IRRELEVANT HEARSAY.
Ill
WHETHER THE MILITARY JUDGE ERRED BY PERMITTING IN EVIDENCE, OVER DEFENSE OBJECTION, PROSECUTION EXHIBIT 7, A LETTER CONTAINING THE OPINIONS OF THREE MEDICAL OFFICERS REGARDING THE SANITY OF THE ACCUSED, THEREBY PERMITTING HEARSAY TO BE GIVEN TO THE COURT MEMBERS AND DENYING THE ACCUSED HIS RIGHT TO CROSS-EXAMINE TWO OF THE OFFICERS WHO NEVER TESTIFIED AT TRIAL.
IV
WHETHER APPELLANT WAS ERRONEOUSLY DENIED AN OPPORTUNITY TO PRESENT EVIDENCE OF GOOD MILITARY CHARACTER IN DEFENSE OF CHARGES OF CONDUCT UNBECOMING AN OFFICER.
We resolve these issues against the Government.
I
The Government’s case-in-chief consisted of testimony of the victim, her sister, and another child. After the prosecution had rested, the civilian defense counsel made a brief opening statement referring to an insanity defense. The first defense witness was Dr. David Feazell, a clinical psychologist, who had examined Major Benedict and had concluded that he “suffered from a mental disease or deficiency of the variety known as pedophilia, which is a mental disorder involving an unusual sexual attraction and expression repeatedly with young — in .the case of Major Benedict —young, female children, preadolescents.” Because of this disease, appellant, at the time of the alleged crimes, “lack[ed] the substantial capacity to conform his conduct to the requirements of the law.”
According to Dr. Feazell, pedophilia could be either “homosexual, involving children of the same sex” or “heterosexual, involving children of the opposite sex of the offender”; and in Benedict’s case, it was the latter. It was also the opinion of this expert that Benedict had not “appreciate[d] the criminality of what he was doing.” Feazell had referred appellant to Dr. William Samek.
During the extensive cross-examination, Dr. Feazell was asked if he were “familiar with” the Diagnostic and Statistical Manual of the American Psychiatric Association in its Third Revision (DSM III). He answered in the affirmative and explained that this Manual was “an exhaustive classification of mental diseases and disorders with explanations and criteria for the diagnosis of each.” According to Dr. Feazell, pedophilia was classified in DSM III as a “paraphelia.”
Over defense objection, he explained that a decision to place a disorder in the Manual
*255“is arrived at through studies of large groups of psychiatrists involving what they believe is the state of the art of theories of mental illness and research on the course of illness, the ability to distinguish reliably among specific types, new theories about previously undefined — you mentioned revisions — one major area of revision is better definition of mental illnesses that, at the time of the previous writing research and effective treatment, had not borne out the effective differences. There are, for instance, preceding the issuing of the Third Edition of the Diagnostic and Statistical Manual, there were nation-wide study groups that work — and smaller study groups — there was research conducted about how reliably psychiatrists and psychologists really could make those diagnoses, as they had delineated them in the previous book, er, where the sources of error were in the research, what things were being lumped together in a non-useful way that did not — one of the objections that was raised — whether these classifications, whether they were really showing all that’s necessary for differential treatment, based on one sub-type versus another, as the theory supported. So that’s — does the research support it? That's — it’s a long, complicated process by a large group of people deliberating.”
Moreover, revisions have been made since the first edition of the Manual; but he did not know whether “the revision process is underway now for a current revision.”
On redirect, Dr. Feazell testified that DSM III is “the primary, authoritative manual in the ... [diagnosis and classification of mental disorders”; and that Benedict had “a mental disease or defect”— namely, pedophilia.
Dr. William Samek, “a clinical psychologist, specializing in the treatment of mentally disordered sex offenders,” testified next for the defense. His qualifications were extensive; he had prepared 4,000 or more written reports on behalf of various courts and had testified “between four and five hundred times” in various courts. He had examined appellant, who was presently his patient in a treatment program. In his opinion Major Benedict had “a mental disease or defect” at the time of the alleged crimes; and “the clinical psychiatric diagnosis of that disease” was pedophilia. Because of this “mental disease or defect,” Benedict “lackfed] the substantial capacity to appreciate the criminality of his conduct” or “to conform his conduct to the requirements of the law.” Samek had discussed his diagnosis with Dr. Feazell, although he arrived at it “independently.”
This expert explained that being able to “appreciate the criminality of” one’s actions “has to do with an emotional understanding and not just your thoughts,” while “[k]nowing right from wrong has to do with a rational, factual understanding.” Asked about the DSM III which had been published by the American Psychiatric Association in 1980, Dr. Samek replied:
It is published by the Psychiatry Society, it’s purely a psychiatric document that is considered in the psychiatric and the psychological community as “the Bible” in terms of diagnosis. There are people who diagnose, using the old, preferring the old DSM. In any field like this, we’ve got a lot of educated people, and you have a lot of opinions, but in terms of a standard reference, it’s unquestionably the reference, it’s used in courts, it’s used in health insurance claims, it’s the accepted authority.
On cross-examination, Dr. Samek acknowledged that pedophilia “is a non-psychotic disorder” and that Major Benedict had shown no evidence of any psychosis. During recross-examination, Dr. Samek conceded that, when he was “referring to mental disease or defect,” he was “using those terms as they are used in the psychiatric or psychology field.”
After the two psychologists had testified, the defense rested. Before calling any rebuttal witnesses, the assistant trial counsel asked the court to take judicial notice of this passage from an article which appeared in 140 American Journal of Psychiatry (June 1983) and was entitled, American *256Psychiatric Association Statement on the Insanity Defense — Insanity Defense Work Group:
The American Psychiatric Association ... suggests that any revision of the insanity defense standards should indicate that mental disorders potentially leading to exculpation must be serious. Such disorders should usually be of the severity (if not always of the quality) of conditions that psychiatrists diagnose as psychoses____“A person charged with a criminal offense should be found not guilty by reason of insanity if it is shown that as a result of mental disease or mental retardation he was unable to appreciate the wrongfulness of his conduct at the time of the offense. As used in this standard, the terms ‘mental disease or mental retardation’ include only those severely abnormal mental conditions that grossly and demonstrably impair a person’s perception or understanding of reality, and that are not attributable primarily to the voluntary ingestion of alcohol or other psychoactive substances.”
Id. at 685. Civilian defense counsel objected strenuously to the prosecution request; but assistant trial counsel responded that “there is no reason in the world why you cannot take judicial notice of this learned treatise”; and the military judge interjected, “I already have.”
Thereafter — and still before prosecution rebuttal was presented — the military judge instructed the court-members on “the legal standards for determining mental responsibility.” His avowed purpose for doing so was “that this is a very complicated area of the law,” and, in light of a question asked by a court member, the judge believed it “appropriate that you be instructed on the basic standards applicable to your decision in this case.” However, he cautioned that the instruction “applies to the testimony given by the defense psychologists,” and “it’s certainly not given at this point to emphasize the prosecution testimony.” His instructions conformed to the guidelines outlined by this Court in United States v. Frederick, 3 M.J. 230 (CMA 1977), and United States v. Cortes-Crespo, 13 M.J. 420 (CMA 1982).
Captain Peggy Huddleston testified as a rebuttal witness that she had completed medical school some 4 years before the trial and her residence a few months before. She was serving as a staff psychiatrist at the Eglin Air Force Base Regional Hospital. She testified, over defense objection, that Major Benedict had come there “for purposes of a sanity board evaluation”; and she had served on this board along with two other psychiatrists — “Doctor Henry Dohn, who is the Chief of Inpatient Services at Eglin, and ... Doctor Raja Calnaido, who was the Chief of Mental Health Services at Eglin.” Benedict had been admitted by her as an in-patient; and she had done “an initial in-take interview; formed a working diagnosis as to what the problem seemed to be”; and performed other responsibilities. Over objection, she also testified about the procedures followed by the sanity board.
When assistant trial counsel (ATC) asked her about “the ultimate conclusion of the board,” the defense objected; and, during an Article 39(a), UCMJ, 10 U.S.C. § 839(a), session, the military judge stated his “understanding that you cannot bring to the attention of the ... [court-martial] what the findings of the sanity board were, or the opinion or conclusion of any member other than the witness testifying.” The ATC replied that the Military Rules of Evidence had changed the case law in this regard. Although the defense argued that the Military Rules had not explicitly overruled earlier precedents, the military judge decided “to let the report go to the members.”
The findings of the board — stated in its report of October 28, 1983, which was signed by Captain Huddleston, Major Dohn, and Major Calnaido, and submitted to the court members — were:
a. At the time of the alleged criminal conduct, the accused possibly suffered from a mental disease/defect known as pedophilia. We have inadequate infor*257mation to make a firm diagnosis of the disorder.
b. The provisional diagnostic entity noted in a. above appears to have first manifested itself approximately 4 years ago.
c. The prognosis for heterosexually oriented pedophiliacs is generally guarded. However, Major Benedict’s overall good life adjustment and functioning, coupled with his motivation for help and his genuine remorse for his actions, are factors which improve his prognosis. Therefore, prognosis is considered good.
d. It is unlikely that hospitalization will be required for this patient.
e. The accused, at the time of the alleged criminal conduct, is not believed to have lacked substantial capacity to appreciate the criminality of his conduct.
f. The accused, at the time of the alleged criminal conduct, is not believed to have lacked substantial capacity to conform his conduct to the requirements of law.
g. Mitigating circumstances may have been ext[a]nt at the time of the alleged criminal conduct. It is beyond the scope of this Board to determine this.
h. The accused possesses sufficient mental capacity to understand the nature of the proceedings against him and to cooperate intelligently in his defense. In connection with the report, the judge
later pointed out that, “as a practical matter, had we not given that document to the court members — they already knew three people were on the board — we have one of them present, I think they could surmise, reasonably, that if the other two disagreed with Doctor Huddleston, the defense would have called them, or we would have dropped the case, but I did not intend for” the prosecution to have Dr. Huddleston “testify as to the board’s findings or recommendations, or diagnoses.”
With respect to the questions answered by the sanity board, Dr. Huddleston testified that she “had insufficient information at that time to make a firm diagnosis of the disorder, and my conclusion was that I could render a provisional diagnosis, which means that there is reason for believing that a diagnosis exists, but there is inadequate information to make a firm diagnosis.” At the time of the sanity board, she had been “[unjaware of the legal definition of the term ‘mental disease or defect’ ” and had interpreted it to refer to “[a] mental disorder, without any specifications, along the lines of all of the mental disorders that are encompassed in DSM III.”
According to this expert, no distinction was drawn “in the psychiatric community ... between the terms ‘mental disease,’ ‘mental defect,’ [and] ‘mental disorder’ “the terms would be used interchangeably, synonymously.” In her “opinion,” Major Benedict “did not lack ... substantial capacity to appreciate the criminality of his conduct” or “to conform his conduct” to the requirements of law. She testified that “[p]edophilia is a non-psychotic mental disorder”; and she went on to explain “what a psychotic disorder is.”
After commenting that the military judge previously had instructed the court members “that mental disease or defect is legally defined as the result of a deterioration, destruction, malfunction, or non-existence of the mental faculties, rather than the moral faculties,” trial counsel asked Dr. Huddleston “what sorts of disorders are implied by that legal definition?” Her reply was, “Psychotic mental disorders, with the characteristics that I mentioned previously.” Then, after citing a further portion of the judge’s instructions to the court members on mental responsibility, the ATC asked her whether “[a] non-psychotic mental disorder” — such as “pedophilia” — “meet[s] the legal criteria of mental disease or defect, as I read it earlier.” Using a chart to illustrate her answer, the witness responded that “non-psychotic disorders,” including pedophilia, would not be a “mental disease or defect.” Furthermore, in evaluating Benedict, she had noted “[n]o psychosis”; and he had been able “to distinguish fact from fantasy” and “exercise a rational thought process.”
*258On cross-examination, Dr. Huddleston conceded that, although Benedict had been at the Eglin Air Force Base Hospital “in excess of 150 hours,” she “could not make a firm diagnosis” based on “the DSM III criteria.” Moreover, she had diagnosed only one pedophilic before meeting Major Benedict; and she had no “specialized training in the diagnosis, or treatment, for the sexually maladjusted individuals.” According to Dr. Huddleston, the DSM III is “not an authoritative source,” but it is “a useful adjunct.”
Testifying in surrebuttal, Dr. Samek stated that, according to DSM III, pedophilia is a mental defect. In his opinion — as well as “in the opinion of the professional community — psychiatrists and psychologists,” the DSM is “an authoritative source.” Moreover, “[t]here are numerous types of mental diseases and defects; psychosis is only one of them.” During the previous four and a half months, he had spent some 40 to 50 hours in personal observation of Major Benedict. On cross-examination, Dr. Samek conceded that “[t]he words ‘mental disease and defect’ are legal terms, not psychological” and that “[t]he word that DSM III uses is ‘disorder.’ ”
Trial counsel asked Dr. Samek if he was “aware that in a June 1983 article of the American Journal of Psychiatry that the American Psychiatric Association’s Insanity Defense Workshop stated that sanity defense disorders ... should usually be of the severity, if not always the quality, of conditions that psychiatrists diagnose as psychoses.” Samek replied that he was “not aware of” this but “agree[d] with it.” When the prosecutor next asked if Dr. Samek was aware that the article stated that the American Psychiatric Association “would define mental diseases or defects, as we’ve been using that term today, to include only those severely abnormal mental conditions that grossly or demonstrably impair a person’s perception or understanding of reality,” the defense objected. The objection was overruled; and the witness answered that he “was not aware of the official position of the American Psychiatric Association at that meeting.” The military judge then advised the court members that he had “allowed the prosecutor to refer to the American Psychiatric Association’s statement for the main purpose of questioning the witness as to whether or not he had considered that statement in forming his own opinion.”
After a brief recess, during which he had an opportunity to review the article about which he had been questioned on cross-examination, Dr. Samek testified that “the thrust of that article” was to propose “changes to the sanity defense as it presently exists.” On recross, Dr. Samek testified that the article “talked about what the American Psychiatric Association felt should be legally — you know, the definition of legal insanity” but “did not allude to the ... military legal standard of insanity.”
Thereafter, the prosecution called Dr. Huddleston to the stand again to “testify as to the fact that that article does not recommend a change — a substantive change in the legal standard, but rather it is an attempt, by the APA, to, in effect, bridge the gap between the legal terminology and what sorts of disorders, in psychiatric terms, are indicated by that terminology.” Moreover, the witness would testify “as to what exactly the definition in the APA article is as to mental disease” and “that the accused did not fit within that definition of mental disease.” Over defense objection, the military judge allowed Dr. Huddleston to take the stand, but he stated, “I do share counsel’s concern about a psychiatrist testifying about the law.” He added, “I think the article is relevant, I think I’ll let your client [sic] state what the purpose of the article was, as you indicated; discuss her diagnosis in relation to the article.”
After hearing argument, the military judge ruled:
I realize that this witness did not rely upon this particular article, at the time she made her diagnosis. However, I’m willing to let her testify that the standard — or whatever you want to call it— *259set out in that article, is consistent with military law and military practice.
Thereafter, Dr. Huddleston testified that she was “familiar with the June 1983 article in the American Journal of Psychiatry, by the APA Work Group, on the insanity defense”; and that it does not “represent a recommendation by the APA to change the substantive legal standard on insanity.” After reading to the court members the definition of “mental disease” in the article, she then testified that this did not differ from “the legal definition of mental disease or defect that” she had heard when she was on the stand previously. Moreover, the disorder of pedophilia did not “fit within that definition,” because it “is not a ... mental condition that grossly and demonstrably impair[s] a person’s perception or understanding of reality.”
No other witnesses were called; and after counsel had presented their arguments, the military judge instructed the court members at length. Subsequently, they returned findings of guilty as charged.
II
An expert witness may testify about an opinion if this would be helpful to the factfinders in determining matters in issue. Mil.R.Evid. 702, Manual for Courts-Martial, United States, 1969 (Revised edition). Opinion testimony otherwise inadmissible is not objectionable because it embraces an ultimate issue. Mil.R.Evid. 704. However, as noted in the Analysis to Mil.R. Evid. 704, “The rule does not permit the witness to testify as to” an “opinion” concerning “the guilt or innocence of the accused or to state legal opinions.” Manual, supra at A18-94.
The purpose of allowing expert testimony is to provide the factfinders the benefit of “specialized knowledge” which “will assist” them in “understandpng] the evidence” and “determining]” the facts. Mil. R.Evid. 702; United States v. Snipes, 18 M.J. 172, 179 (CMA 1984). Captain Huddleston — whatever her qualifications in the medical field — was not an expert on the legal definition of mental responsibility. However, in several instances the military judge let her expound on matters of law, rather than psychiatry. For example, her comparison of the definition of insanity in the article with military law and practice moves from the medical into the legal field. From our reading of the record, it appears that to some extent the military judge permitted Captain Huddleston to usurp his role in providing legal guidance to the court members.
Military law has never recognized an absolute rule that an accused must suffer from a psychosis in order to merit acquittal by reason of insanity. However, in view of the leeway granted to Dr. Huddleston in discussing legal principles, the court members may very well have received the impression that appellant was not entitled to acquittal unless they believed he suffered from a psychosis; and none of the experts testified that he was psychotic.
Ill
The harm was aggravated by the prosecution’s attempt to bring to the court members’ attention the contents of the article in the American Journal of Psychiatry. Prior to adoption of the Military Rules of Evidence, learned treatises relied upon by experts could be used only for purposes of cross-examination or for impeachment. See para. 138(e), Manual, supra; cf. Brown v. United States, 419 F.2d 337 (8th Cir. 1969). However, Mil.R.Evid. 803(18), which was in effect at the time of this court-martial, provided:
To the extent called to the attention of an expert witness upon cross-examination or relied upon by the expert in direct examination, statements contained in published treatises, periodicals, or pamphlets on a subject of history, medicine or other science or art, established as a reliable authority by the testimony or admission of the witness or by other expert testimony or by judicial notice [are not hearsay]. If admitted, the statements may be read into evidence but may not be received as exhibits.
*260We doubt that the passage from this article, which was referred to at various times in the examination of witnesses, is “on a subject of history, medicine or other science or art.” Instead, its “subject” seems to fall much more in the field of law. Moreover, the article seemed to be concerned with a proposed standard, rather than with a description of existing law. In our reading of the record, the references to the article have been confusing. We suspect that these references — including a reference by trial counsel in his final argument — were also unenlightening for the court members; and in any event, they resulted in consideration by the members of a medical expert’s inadmissible hearsay opinion on a legal issue.
The 1969 Manual for Courts-Martial — which was in effect when appellant was tried — and its 1951 predecessor provided for reports by sanity boards and prescribed in paragraph 122c (original version):
So much of the report of a board of medical officers or any other medical record as pertains to entries of facts or events which are properly admissible under the official records or business entry exceptions to the hearsay rule ... may be received in evidence. The opinions as to the mental condition of the accused contained in such a report are not within these exceptions to the hearsay rule.
(Emphasis added.)
Consistent with this language, it was held that opinions contained in a sanity-board report should not be brought to the attention of the court members — either directly or indirectly. United States v. Smith, 47 CMR 952 (ACMR 1973); United States v. Rausch, 43 CMR 912, 917 n. 3 (AFCMR 1970); United States v. Parmes, 42 CMR 1010 (AFCMR 1970). Thus, in Smith the Court of Military Review set aside the findings of guilty because the government psychiatric witness had been allowed to testify over defense objection to the “unanimous” and collective opinion of the members of the sanity board. The Court of Military Review commented that “we can’t pass lightly over the fact that one side got in effect the weight of three witnesses for the price of only one cross-examination in a field so fertile for such.” 47 CMR at 953.
In appellant’s case, trial counsel and the military judge apparently took the position that these precedents — and the express language of paragraph 122(c) — had been superseded by adoption of the Military Rules of Evidence in 1980. In this connection there was special reliance on Mil.R.Evid. 803(6), which allows introduction into evidence as a hearsay exception of
[a] memorandum, report, record, or data compilation, in any form, of acts, events, conditions, opinions, or diagnoses, made at or near the time by, or from information transmitted by, a person with knowledge, if kept in the course of a regularly conducted business activity, and if it was the regular practice of that business activity to make the memorandum, report, record, or data compilation, all as shown by the testimony of the custodian or other qualified witness, unless the source of information or the method or circumstances of preparation indicate lack of trustworthiness. The term “business” as used in this paragraph includes the armed forces, a business, institution, association, profession, occupation, and calling of every kind, whether or not conducted for profit. Among those memoranda, reports, records, or data compilations normally admissible pursuant to this paragraph are enlistment papers, physical examination papers, outline-figure and fingerprint cards, forensic laboratory reports, chain of custody documents, morning reports and other personnel accountability documents, service records, officer and enlisted qualification records, logs, unit personnel diaries, individual equipment records, daily strength records of prisoners, and rosters of prisoners.
(Emphasis added.)
We disagree with this conclusion. In the first place, we find nothing in the Analysis *261of Mil.R.Evid. 803(6) to indicate that this Rule was intended to change the express language of the 1969 Manual and the well-established case law. 1969 Manual, supra at A18-104 to A18-105. Second, we conclude that a report from a sanity board established pursuant to paragraph 122 of the 1969 Manual for Courts-Martial — or R. C.M. 706, Manual for Courts-Martial, United States, 1984 — is not a report of “opinions or diagnoses ... kept in the course of a regularly conducted business activity" or that it "was the regular practice of that business activity to make the ... report.” See Mil.R Evid. 803(6). While opinions and diagnoses recorded in a hospital as part of a patient's treatment will usually fall within this hearsay exception, a report of a sanity board, convened in connection with possible criminal prosecution, does not.
Each sanity board is created ad hoc to evaluate a specific individual and is not “regularly conducted.” There is no “regular practice” of preparing such reports; and there is no “business” of preparing sanity-board reports. Moreover, if we gave Mil.R.Evid. 803(6) the broad interpretation for which the Government contends, there would be significant problems under the Confrontation Clause of the Sixth Amendment.
“Opinions and diagnoses of mental or psychiatric conditions have been generally considered too complex and speculative to be admitted without cross-examination.” Nauni v. State, 670 P.2d 126, 131 (Okla.Cr. 1983), citing Fry v. State, 529 P.2d 521 (Okla.Cr.1974), and Bennett v. State, 448 P.2d 253 (Okla.Cr.1968). Assuredly, the opinions contained in the sanity-board report deprived Benedict of important safeguards which would tend to assure the reliability of the evidence being received.
Although reliability need not be considered with respect to evidence admissible under “a firmly rooted hearsay exception,” see Ohio v. Roberts, 448 U.S. 56, 66, 100 S.Ct. 2531, 2539, 65 L.Ed.2d 597 (1980), we have already made clear that admission of this evidence would go far beyond the business-record exception, as previously recognized in military law. Likewise, we are not aware of civilian precedents for admitting such a report.
Mil.R.Evid. 803(6) should be construed to avoid confrontation issues; but if it receives the broad construction given at trial by the military judge, appellant’s right of confrontation was violated. “In either event, the result is the same: The challenged report was inadmissible.” See United States v. Broadnax, 23 M.J. 389, 393 (CMA 1987) (footnote omitted).
Indeed, Broadnax is very much on point. There, the Government contended that a laboratory report on handwriting exemplars was admissible under Mil.R.Evid. 803(8), which authorizes a hearsay exception for “public records and reports.” Judge Sullivan, writing for the Court, pointed out that
[w]e cannot equate what substantially amounts to an opinion of guilt with the factual type of result concerning the identity of an unknown substance which is produced by chemical analysis. See United States v. Evans, 21 U.S.C.M.A. 579, 581-82, 45 C.M.R. 353, 355-56 (1972). Moreover, although scientific or professional methods may be employed to some degree in reaching such an opinion, the element of subjectivity involved is greater than that in chemical analysis.
23 M.J. at 393. While handwriting analysis involves many subjective elements, they pale into insignificance in comparison with the subjective elements involved in evaluation of an accused’s sanity. Thus, Broadnax — which relies alternatively on an interpretation of Mil.R.Evid. 803(8) and the constitutional right to confrontation — mandates that the report in this case should have been excluded.
Furthermore, we have no doubt that, whatever the effect of the other evidentiary errors, this error was prejudicial. The defense had presented testimony by two experts with great forensic experience — especially in dealing with mental disorders that resulted in sex offenses. The only government expert who appeared had *262finished her medical residency only a few months before trial, and she had relatively limited experience with sex offenders. It was important for the Government to be able to strengthen her testimony by placing before the court members a written report signed by her two fellow board members— each of whom apparently had much more experience than she. By the same token, Benedict was prejudiced by the sub rosa admission of the opinions of two experts who were neither observed by the factfinders nor subjected to cross-examination.
IV
During his case-in-chief, the defense attempted to introduce Benedict’s awards, decorations, and officer effectiveness reports as evidence of the specific character trait of “good conduct as an officer.” We conclude that, under Mil.R.Evid. 404(a)(1), this evidence was incorrectly excluded by the military judge. United States v. Court, 24 M.J. 11 (CMA 1987); United States v. Belz, 20 M.J. 33 (CMA 1985).
According to the Court of Military Review, appellant was not prejudiced by this error because his defense was predicated on the lack of mental responsibility, and “[ejvidence of his good military character would have had no impact on a diagnosis of pedophilia.” 20 M.J. at 944. We disagree with this analysis. When an accused defends on grounds of insanity and offers evidence of his good character, he is contending that his acts must have resulted from insanity because — as demonstrated by his past good character — he would never have committed a crime had he been in his right mind. How convincing this contention may be will vary with the facts of the case; but, unlike the court below, we see no reason why the evidence of good character is per se inadmissible.
Here, the judge’s error in excluding the evidence would normally be subjected to the analysis for prejudice prescribed by United States v. Weeks, 20 M.J. 22 (CMA 1985). Certainly, in this hotly contested case, appellant would have a strong argument that he had been prejudiced by this error, if no other evidentiary errors had occurred. However, we need not consider this question, because we have already ruled that admission of the sanity-board report was prejudicial error.
V
The decision of the United States Air Force Court of Military Review is reversed. The findings of guilty and the sentence are set aside. The record of trial is returned to the Judge Advocate General of the Air Force. A rehearing may be ordered.
Judge SULLIVAN concurs.