Defendant’s assignments of error raise a number of questions relating to the admissibility of evidence that he was insane at the time of the killings. We find that the trial court improperly limited testimony by a psychiatric expert concerning his examination, findings and diagnosis of defendant, and on this ground we *456order that defendant receive a new trial. For guidance of the court on remand, we will also discuss defendant’s assignments of error regarding exclusion of evidence of hereditary insanity and declarations by defendant to third parties as to his state of mind.
The uncontradicted evidence showed that on 20 August 1977 defendant killed his wife and two children and then stabbed himself. Defendant did not deny that he committed the killings but instead tried to show that he was insane. To this end he produced evidence that he had been a good husband and father and a steady worker and that he had had a good reputation in the community. For some two or three weeks prior to the killings, however, he had been depressed and had problems with his job. Lastly, according to two psychiatrists who examined him, defendant’s mental condition at the time of the killings was such that he did not know the difference between right and wrong.
The first psychiatric witness on defendant’s behalf was Dr. Eugene Douglas Maloney, who was employed at the Neuse Mental Health Center and who was duly qualified as an expert in the field of psychiatry. Before the jury Dr. Maloney was permitted to testify that defendant was his patient1 and that he had seen him on three different occasions in 1977: 21 September, 30 September, and 4 November. He also testified about his method of establishing a rapport with and gaining defendant’s confidence and about his method of examination.2 Finally after seven record pages of quibbling among the trial court, the district attorney, and defense counsel over the form of the question and the form of the witness’ response, the doctor was barely permitted to tell the jury that in his opinion at the time of the killings defendant “was incapable on that night of distinguishing between right and wrong.”
*457Most of Dr. Maloney’s testimony took place out of the hearing of the jury in order for defendant to place into the record testimony to which objections were lodged by the state and sustained by the trial judge. In summary this testimony consisted of conversations Dr. Maloney had with defendant, his medical findings, and his medical diagnoses all of which led ultimately to his opinion regarding defendant’s sanity in a legal sense. Dr. Maloney would have testified, had he been permitted, that defendant expressed ideas that on the day of the killings people were after him, that he was being watched, and that they wanted to “hurt me and my family.” Defendant said further that just before the killings one of his children brought in a glass of something to drink and he “thought it was for me to poison myself.” Defendant told Dr. Maloney that when he stabbed his wife, “I lost all senses at that time. I was protecting her, I didn’t want them to hurt her. I was crying, they were going to hurt me and my family. Everybody was watching.” With regard to stabbing himself defendant said, “They were not going to get my family because I had just killed them and they were not going to get me ... for some reason I . . . went to the bedroom with my family and laid down so I could die with my family.” Defendant then made similarly bizarre statements concerning a “pacemaker” which had been installed in his heart so that he could not die and would be able to stand trial, and to the effect that at Dorothea Dix Hospital “they took my glands out to keep me from mating with other people . . . they took my birthright so I wouldn’t have to die, by taking away my glands.”
Dr. Maloney would have further testified, had he been permitted, that on 4 November 1977 he found defendant to be “highly disturbed . . . dangerous to himself or others . . . angry, frightened” and that he “paced the floor, looked behind curtains, looked behind desks, looked behind file cabinets.”
Finally Dr. Maloney would have told the jury, had he been permitted, that he diagnosed defendant as paranoid and *458psychotic. He would have explained “paranoid” as involving “delusions of persecutions” and a “delusion” as being “a false belief that cannot be changed by reason or logic.” He would have explained a psychotic person as one who suffers “from delusions which I have defined, and hallucinations. Hallucinations being seeing things, hearing voices not there, feeling things that are not there.”
[2] We think all of this testimony was erroneously excluded to defendant’s prejudice. The error is not cured by Dr. Maloney’s being permitted to give his ultimate conclusion regarding defendant’s ability to distinguish right from wrong. The defendant was entitled to have the jury know the bases for this conclusion as well as the conclusion itself.
Dr. Maloney’s findings resulted from his personal examination and observation of defendant. His diagnosis resulted from this and listening to and analyzing defendant’s conversation. It is a well-settled rule that an expert may give an opinion based on facts within his personal knowledge without resort to a hypothetical question. State v. DeGregory, 285 N.C. 122, 203 S.E. 2d 794 (1974); 1 Stansbury’s North Carolina Evidence § 136 (Brandis rev. 1973) (hereinafter Stansbury). Clearly Dr. Maloney’s findings are admissible under this rule. Problems arise, however, when a physician’s opinion is derived in whole or in part through information received from another, as Dr. Maloney’s diagnosis was here, because of a second rule articulated in our cases that in general “an expert witness cannot base his opinion on hearsay evidence.” Cogdill v. Highway Commission, 279 N.C. 313, 327, 182 S.E. 2d 373, 381 (1971). Resolution of this conflict has been especially difficult in cases involving the physician-patient relationship, because communication between the two is often an essential, if not the only, way for the physician to form an intelligent opinion.3
*459Cases in this area have dealt with two major issues. The first is the admissibility of the expert opinion when it is based on an out-of-court communication. The second is, assuming the opinion is admissible, to what extent the expert may repeat what was told him out of court in order to show its basis. Here we are faced with both issues. We conclude that Dr. Maloney’s opinion, in the form of his diagnosis, was admissible. We also conclude that on retrial he may recount his out-of-court conversations with defendant in order to explain his diagnosis to the jury.
One of the earliest significant cases on the principles involved here was State v. Alexander, 179 N.C. 759, 103 S.E. 383 (1920). Defendant in Alexander offered into evidence the answers he had given an expert (a physician) during an examination from which the expert had formed an opinion as to defendant’s sanity. The trial court allowed the expert to give his opinion but prevented him from recounting certain declarations by defendant. This Court, through Chief Justice Clark, affirmed, giving its view of the case as follows, id. at 765, 103 S.E. at 386:
“His Honor in these rulings was drawing a distinction between facts drawn out in Doctor Hall’s conversation with the defendant, which tended to show the state of defendant’s mind and those which did not. Conversation with one alleged to be insane is, of course, one of the best evidences of the present state of his mind. If, however, there is incorporated in the conversation self-serving declarations which in themselves do not throw any light upon the present condition or the past condition of the man’s mind, then these declarations are not admissible.”
Because of the trial court’s ruling, the court in Alexander was not squarely presented with the issue of the admissibility of an expert opinion based in part on out-of-court conversations. Insofar as admitting the conversations themselves was concerned, Alexander imposed a limit only to the extend that they “do not throw any light upon the present condition or the past condition of the [defendant’s] mind.”
In Penland v. Coal Co., 246 N.C. 26, 97 S.E. 2d 432 (1957), a physician testifying at a workmen’s compensation proceeding gave an opinion as to the plaintiff’s physical condition. On cross-examination, Jiowever, he stated that he had discovered no objec*460tive symptoms and that his opinion was based on subjective statements by the plaintiff.4 On appeal defendants contended that the opinion should have been disregarded and an award to plaintiff disallowed as not based on competent evidence. This Court disagreed, stating, id. at 31, 97 S.E. 2d at 436:
“As to this contention, the rule is that ordinarily the opinion of a physician is not rendered inadmissible by the fact that it is based wholly or in part on statements made to him by the patient, if those statements are made, as in the present case, in the course of professional treatment and with a view of effecting a cure, or during an examination made for the purpose of treatment and cure. ‘In such cases statements of an injured or diseased person, while not admissible as evidence of the facts stated, may be testified to by the physician to show the basis of his opinion.’ ” Quoting, 20 Am. Jur., Evidence § 86 at 729.
Penland thus (1) allowed in evidence an opinion when it was reasonable to assume that the information upon which it was based was reliable and (2) held that testimony about that information was proper to show the basis for the opinion.
Subsequently, Seawell v. Brame, 258 N.C. 666, 129 S.E. 2d 283 (1963), made the reliability element mentioned in Penland a prerequisite for admissibility of the opinion. In Seawell a physician testified that in his opinion the cause of plaintiff’s asthmatic attacks and ulcer was his being struck by a machine operated by defendant. He based his opinion on conversations with plaintiff’s wife, other members of his family and his former employee, all of whom said that plaintiff had not had such problems before the accident. This Court held his opinion inadmissible. Seawell is distinguishable from Penland as involving communications outside the physician-patient relationship. When a patient seeks treatment, “it is reasonable to assume that the information which [he] gives the doctor will be the truth, for self-interest requires it.” State v. Bock, 288 N.C. 145, 163, 217 S.E. 2d 513, 524 (1975). The *461law makes no such assumption with regard to the communications of others to physicians.5
In State v. DeGregory, supra, 285 N.C. 122, 203 S.E. 2d 794, this Court was called on to determine whether a psychiatrist’s opinion that defendant was sane was improperly admitted. The psychiatrist testified that he had seen defendant for about three hours on three different occasions and that he had formed his opinion on the basis of this examination and information supplied him by members of his staff. Defendant argued that the opinion should have been excluded because it was based in part “on information obtained by someone else, which information was inadmissible in evidence.” Id. at 132, 203 S.E. 2d at 801. In making this argument, defendant relied on the following quotation from State v. David, 222 N.C. 242, 254, 22 S.E. 2d 633, 640 (1942): “There are two avenues through which expert opinion evidence may be presented to the jury: (a) Through testimony of the witness based on his personal knowledge or observation; and (b) through testimony of the witness based on a hypothetical question . . . .” The Court, thorugh Justice Huskins, rejected defendant’s argument, finding his reliance on the “personal knowledge or observation” rule misplaced:
“Defendant’s interpretation of the quotation from State v. David, supra, is too limited. The quotation states that an expert may base his testimony on facts within his personal knowledge or observation, or may base his opinion on facts presented in a hypothetical question, but it does not purport to limit facts and information within the personal knowledge of an expert to knowledge derived solely from matters personally observed. As demonstrated in opinions of this Court since State v. David, supra, an expert witness has wide latitude in gathering information and may base his opinion on evidence not otherwise admissible.” 285 N.C. at 132, 203 S.E. 2d at 801. (Emphasis original.)
In the course of its examination of relevant authorities, the Court quoted with approval from Birdsell v. United States, 346 F. 2d 775, 779-780 (5th Cir.), cert. denied, 382 U.S. 963 (1965): “With the increased division of labor in modern medicine, the physician mak*462ing a diagnosis must necessarily rely on many observations and tests performed by others and recorded by them; records sufficient for diagnosis in the hospital ought to be enough for opinion testimony in the courtroom.” Using this analysis, the DeGregory court concluded that it was proper for the expert to base his opinion “upon both his own personal examination and other information contained in the patient’s official hospital record.” 285 N.C. at 134, 203 S.E. 2d at 802.
The most recent of our cases in this area is State v. Bock, supra, 288 N.C. 145, 217 S.E. 2d 513. Defendant in Bock complained of the exclusion of testimony by an expert witness that in his opinion defendant had amnesia concerning the events surrounding the killing with which he was charged. The expert’s opinion had been elicited by way of a hypothetical question that did not refer to defendant’s history of excessive drinking followed by blackout spells or periods of amnesia because there was no evidence before the jury of these events. The expert was aware of this history through conversations with defendant, his family and friends. Defendant had not gone to the expert, a psychiatrist, for diagnosis and treatment. Instead the psychiatrist had only seen him for about two hours two days prior to trial for the purpose only of preparing himself to testify. On these facts, the court held that the expert did not have the requisite personal knowledge of defendant’s history to give his opinion without the use of a proper hypothetical question.
[1] Although none of these cases articulates any sort of universally applicable rule, the pattern of their holdings supports the following propositions: (1) A physician, as an expert witness, may give his opinion, including a diagnosis, based either on personal knowledge or observation or on information supplied him by others, including the patient, if such information is inherently reliable even though it is not independently admissible into evidence. The opinion, of course, may be based on information gained in both ways. (2) If his opinion is admissible the expert may testify to the information he relied on in forming it for the purpose of showing the basis of the opinion. Penland v. Coal Co., supra, 246 N.C. 26, 97 S.E. 2d 432.
[2] Applying the first of these propositions to the present case, we find two grounds supporting the admission of Dr. Maloney’s *463diagnosis. There is uncontradicted evidence that defendant was sent to Dr. Maloney as a patient for treatment. Under Penland v. Coal Co., supra, this lends reliability to the statements made by defendant to the doctor. Secondly, we find a sufficient indication of the reliability of these statements in the nature of Dr. Maloney’s entire examination. The examination, as described in note 2, supra, was a thorough, carefully designed attempt to gain an understanding of defendant’s state of mind. Dr. Maloney did not rely for his conclusions on any one statement by defendant or on any particular fact he disclosed. Instead he took into account the entirety of what defendant said together with his own interpretation and analysis of it and the objective manifestations that accompanied it. The assertion of State v. Alexander, supra, 179 N.C. at 765, 103 S.E. 2d at 386, that “[conversation with one alleged to be insane is, of course, one of the best evidences of his present state of mind” is still true. Conversation, and its interpretation and analysis by a trained professional, is undoubtedly superior to any other method the courts have for gaining access to an allegedly insane defendant’s mind. When it is conducted with the professional safeguards present here, it provides a sufficient basis for the introduction of an expert diagnosis into evidence.
For the reasons stated, Dr. Maloney’s findings and diagnosis as to defendant’s mental state should have been admitted into evidence. It follows also from what we have said that on retrial Dr. Maloney should be allowed to testify as to the content of his conversations with defendant in order to show the basis for his diagnosis. The reason for allowing such testimony and the safeguards that should accompany it were well stated by the Arizona Supreme Court in State v. Griffin, 99 Ariz. 43, 49, 406 P. 2d 397, 401 (1965):
“In the same vein to allow a psychiatrist as an expert witness to answer without any explanation . . . would impart a meaningless conclusion to the jury. The jury must be given an opportunity to evaluate the expert’s conclusion by his testimony as to what matters he took into consideration to reach it. Therefore the psychiatrist should be allowed to relate what matters he necessarily considered as a ‘case history’ not as to indicate the ultimate truth thereof, but as one of the bases for reaching his conclusion, according to ac*464cepted medical practice. The court should therefore exercise care in the manner in which such testimony is elicited, so that the jury may understand that the case history does not constitute factual evidence, unless corroborated by other competent evidence.”
We emphasize again that such testimony is not substantive evidence. Thus there is no conflict between its introduction and the rule that a criminally accused’s declarations to third parties generally to show his state of mind are not admissible as an exception to the hearsay rule. See Stansbury, supra, § 161 at 541, discussed below.
The questions presented by two of defendant’s assignments of error are likely to recur at trial on remand. We therefore discuss them briefly here.
[3] Defendant attempted through several witnesses to show evidence of mental illness in his maternal ancestors. The trial court properly excluded this testimony. While it is true that evidence of hereditary insanity has been held admissible, State v. Christmas, 51 N.C. 471 (1859), there was not an adequate foundation for its admission here.. In order for insanity among a person’s ancestors or relatives to be relevant, it must first be shown that (1) there is independent evidence of insanity on the part of the person, (2) the same type of mental disorder is involved, and (3) the mental disorder is hereditary in character. In re Will of Kemp, 236 N.C. 680, 73 S.E. 2d 906 (1953); State v. Cunningham, 72 N.C. 469 (1875). Defendant offered no evidence at trial on the second and third points. Without such a showing, any evidence of hereditary insanity was irrelevant.
Finally, several witnesses on defendant’s behalf were not allowed to testify about declarations by defendant that would have tended to show his state of mind. We cannot tell if these exclusions were prejudicial, because defendant had only a few of the answers put in the record and those are inconclusive. The first of these exclusions occurred in the following exchange during the testimony of defendant’s grandfather, James Iredell Wade:
“A.: ... he [the defendant] was down and out and I asked him what was — seemed to be his trouble and he asked me, he said —
*465MR. Barker: Objection.
The COURT: Sustained, I’m sorry you can’t tell what your son said to you.”
Also while Wade was on the stand, an objection was sustained to the question, “Did Robbie indicate to you at any time he was troubled with his work because of having to go to Jacksonville?” During the testimony of Timothy Edwards Penny, statements made by defendant in a conversation the day before the killings, in which he apparently indicated he was having problems, were excluded. Lastly, the court sustained objections to a number of questions asked of Clifton Edwards, defendant’s former employer.6 The substance of these questions and the answers Edwards would have given were later placed in the record as follows:
“(Q. Mr. Edwards, was there ever an occasion that this defendant followed you home to complain about treatment received at his work?)
(A. Yes.)
(Q. Would you describe that occasion?)
(A. This was the time when I had given him a directive to come to work the following work day in a clean uniform, cleanly shaven; he objected to this directive and was very angry. After talking with him the anger subsided and he came to work on the following work day with a completely different attitude, everything was fine.)
(Q. Now, Mr. Edwards, was there ever an occasion that this defendant told you and I quote, ‘You are not treating me right?’)
(A. This was some time during the last week of his employment and I did think that rather strange because I hadn’t said anything to him to cause this answer. And he had a wild or strange look in his eye which I have never seen before. And I did think something of it even before this happened.)”
*466 [4] In each instance the declarations by defendant and the other testimony that was excluded would have tended to show defendant’s state of mind. Declarations of a person as to his state of mind are in general admissible as an exception to the hearsay rule. See McRae v. Malloy, 93 N.C. 154 (1885); Stansbury, supra, § 161. The trial judge, as we understand his comments set forth in note 6, supra, apparently thought that this exception does not apply to defendants in criminal cases. The state argues he was correct, citing State v. Scott, 8 N.C. 24 (1820), and Stansbury, supra, § 161 at 541, which states: “Even statements of an existing emotion or other mental state, when uttered by an accused person and offered on his behalf, are generally excluded, on the ground that to receive them would permit the accused to make evidence for himself.” We think this statement has been interpreted too broadly by the state. Examination of the cases from which it was derived shows in each instance that the statements excluded were made after the commission of the crime. State v. Moore, 104 N.C. 743, 10 S.E. 183 (1889); State v. Hildreth, 31 N.C. 440 (1849); State v. Scott, supra. But cf. VI Wigmore on Evidence § 1732 at 161-62 (Chadbourn rev. 1973) (arguing strongly that statements of existing state of mind should be admitted even when made by accused after commission of crime). All the declarations defendant sought to introduce here were made before the date and not in contemplation of the killings. They should have been admitted.
Defendant’s other assignment of error is not likely to present a problem on remand. For the reasons stated, we order that he receive a
New trial.
Justices BRITT and BROCK did not participate in the consideration or decision of this case.