dissenting:
This case illustrates both the difficulties created by our ease law and the difficulty of fairly assessing the work capabilities of a Social Security applicant whose major problems are mental and whose disability is due in part to alcoholism and other substance abuse. Under our case law the one factfinder who always has a chance to see the applicant is required to justify his conclusions with more rigor than we demand of juries. The understandable temptation — yielded to here, I believe — is to temper the rigor by letting the ALJ satisfy his obligations with reasoning that superficially accommodates itself to our criteria while actually subverting them. If we are to have a consistent rule of law in Social Security cases, we can’t acquiesce in pro forma compliance by the ALJ. If we are to grant disability benefits to those too mentally impaired to engage in substantial gainful activity, we cannot deny them to a man whose treating psychiatrist and examining psychologist have found him substantially impaired in social interaction and in ability to respond appropriately to supervisors, co-workers and the demands of usual work situations.
The medical evidence in this case essentially consists of reports from one treating psychiatrist and one examining psychologist, and the testimony of a non-examining medical expert. The law regarding the relative weight to be assigned to these reports is established:
As a general rule, more weight should be given to the opinion of a treating source than to the opinion of doctors who do not treat the claimant. Winans v. Bowen, 853 F.2d 643, 647 (9th Cir.1987). At least where the treating doctor’s opinion is not contradicted by another doctor, it may be rejected only for “clear and convincing” reasons. Baxter v. Sullivan, 923 F.2d 1391, 1396 (9th Cir.1991). We have also held that “clear and convincing” reasons are required to reject the treating doctor’s ultimate conclusions. Embrey v. Bowen, 849 F.2d 418, 422 (9th Cir.1988). Even if the treating doctor’s opinion is contradicted by another doctor, the Commissioner may not reject this opinion without providing “specific and legitimate reasons” supported by substantial evidence in the record for so doing. Murray v. Heckler, 722 F.2d 499, 502 (9th Cir.1983).
The opinion of an examining physician is, in turn, entitled to greater weight than the opinion of a nonexamining physician Pitzer v. Sullivan, 908 F.2d 502, 506 (9th Cir.1990); Gallant v. Heckler, 753 F.2d 1450 (9th Cir.1984). As is the case with the opinion of treating physician, the Commissioner must provide “clear and convincing” reasons for rejecting the uncontra-dicted opinion of an examining physician. Pitzer, 908 F.2d at 506. And like the opinion of a treating doctor, the opinion of *604an examining doctor, even if contradicted by another doctor, can only be rejected for specific and legitimate reasons that are supported by substantial evidence in the record. Andrews v. Shalala, 53 F.3d 1035, 1043 (9th Cir.1995).
Lester v. Chater, 81 F.3d 821, 830-31 (9th Cir.1995).
Morgan’s treating psychiatrist and examining psychologist are in agreement that Morgan’s mental state markedly limits him in sustained concentration and persistence and social interaction, two of the four key areas which- are identified by the Social Security Administration’s own regulations as being prerequisite to meeting the mental demands of work. 20 C.F.R. §§ 404.1521, 416.921. They also agree that Morgan lacks one of the three required abilities to perform unskilled work as described in Social Security ruling 85-15. The testifying vocational expert confirmed that an individual with the impairments identified by Drs. Reaves and Grossc-up is precluded from engaging in substantial gainful activity.
In combination, Drs. Reaves and Grosscup diagnosed Morgan with suicidal ideation with a history of suicide in his family; affective instability involving major depression with a history of mental illness in his family; general social dysfunction including tendencies toward catastrophizing, resentment, hostility, intense anger, emptiness, and irritability; and marked limitations in ability to maintain sustained concentration and persistence. Both Drs. Reaves and Grosscup made clear in their reports as well as in their Mental Residual Functional Capacity evaluations that these impairments culminated in marked restrictions upon the two abilities which the regulations emphasize are prerequisites for engaging in work activity: ability to interact appropriately with people in a work environment and the ability to complete a normal workday without interruptions from psychologically based symptoms.
Drs. Reaves and Grosscup do not explain exactly why a tendency toward hostility limits Morgan’s social functioning in a work environment. The explanation is not necessary: the limitation is self-evident. It is enough that they noted particular areas in which Morgan was challenged and then provided a diagnosis and an opinion as to the effect of the impairments on Morgan’s mental residual functional capacity. The ALJ quibbles when he objects that the experts did not articulate the connections. If the doctors’ diagnosis is correct, the limitations are gross.
The ALJ contends that the expert reports are internally inconsistent in that they include numerous comments which indicate that Morgan is socially functional and able to concentrate and work well and that they are inconsistent with each other in that Dr. Reaves’ opinion of Morgan’s superior cognitive functioning conflicts with Dr. Grosscup’s mental status evaluation. The ALJ relied on these differences not only to discredit the conclusion that Morgan’s social functioning and concentration abilities are markedly limited, but also to reject most of the experts’ opinions and most of Morgan’s subjective complaints as well. Under the Administration’s regulations, however, a finding that a claimant is functional in some areas does not preclude a diagnosis of disability. Disability is defined as the inability to engage in substantial gainful activity. The ultimate question before an ALJ — and before the experts undertaking residual functional capacity assessment — is whether a claimant is impaired to the degree that he/she is incapable of performing work activities in a competitive work environment on a regular and sustained basis. Given that this is the question to be answered, a claimant’s superior cognitive functioning, his ability to maintain some friendships, and his ability to stay focused while caring for a garden do not disprove the finding that he is precluded from substantial gainful work activity. What the ALJ characterizes as inconsistency within the reports is subordinate to the agreement between them that while Morgan can concentrate and can function well socially to some degree and under certain circumstances, the ability to engage in sustained, competitive work activity lies beyond the threshold of his capabilities-
The ALJ engages in further quibbling to find the two-experts at odds with each other. In particular, the ALJ cites Dr. Reaves’ June *6051994 report in which Dr. Reaves noted that Morgan felt he -had experienced “significant improvement,” and Dr. Reaves’ July 1994 report in which he noted that Morgan’s symptoms of depression had improved; the ALJ contrasts these reports with Dr. Grosse-up’s July 1994 report in which Dr. Grosseup recognized only “moderate change.” The ALJ also points out that while Dr. Reaves in his July 1994 and February 1995 reports commented that Morgan’s suicidal ideation was less severe, Dr. Grosseup, during the same time periods, described Morgan as having a strong desire to kill himself. The ALJ finds these differences to be significant, arguing that they rendered the conflicting impressions suspect and adding that they appeared to show that Morgan himself made different reports to the experts thus discrediting his subjective symptoms as well.
A presumption in favor of Morgan’s claim was created when both the treating psychiatrist and examining psychologist arrived at similar conclusions. See Lester, 81 F.3d at 832. The differences isolated by the ALJ are not significant enough to overcome this presumption. The differences are not inconsistencies. Like Dr. Reaves, Dr. Grosseup in her June and July reports did note that Morgan was feeling improvement; Dr. Grosseup went on however, to conclude that Morgan experienced only moderate change because Morgan was nonetheless complaining of significant side effects such as increased anger, hostility and frustration from the medication — side effects which were noted by Dr. Reaves also. Dr. Grosscup’s report and conclusion are not inconsistent with Dr. Reaves’ report and conclusion but represent a more conservative assessment. This slight difference in impression is not significant when both the psychiatrist and psychologist ultimately concluded that Morgan is markedly impaired despite signs of improvement. The same holds true for them divergence of opinion regarding Morgan’s suicidal ideation. Dr. Reaves’ notation that Morgan’s suicidal ideation is less severe is not inconsistent with Dr. Grosscup’s notation that Morgan remains suicidal. These two statements are, in fact, perfectly consistent, especially when one considers that both opinions regarding Morgan’s suicidal ideation represent attempts to characterize Morgan’s fluctuating suicidal tendencies over a long period of time. It is not inconsistent for two experts to arrive at slightly different characterizations of a patient’s up-and-down mental state over a period of a month. We have previously noted that depression is a complex and highly idiosyncratic phenomenon that often waxes and wanes, eluding neat description; it would be naive to expect someone suffering from depression and suicidal ideation to remain consistent in the way in which the ALJ demands. It was unreasonable to discredit the opinions of examining experts who otherwise agree.
In accepting the ALJ’s- reliance on Dr. McConoehie’s opinion, the majority refers to Magallanes v. Bowen, 881 F.2d 747, 751-55 (9th Cir.1989), Andrews v. Shalala, 53 F.3d 1035, 1043 (9th Cir.1995), and Roberts v. Shalala, 66 F.3d 179 (9th Cir.1995), in which we held that an ALJ may reject the opinion of a treating or examining physician, based in part on the testimony of a nontreating, nonexamining medical advisor. In each of these cases, however, there was significant evidence conflicting with the opinion of the treating or examining physicians. In Ma-gallenes, there were laboratory results and contrary reports from other examining physicians. In Andrews, five nonexamining physicians, one of them a specialist in the relevant area, plus the claimant’s own testimony, were used by the ALJ to reject the favorable report from one examining physician. In Roberts, the examining doctor’s conclusion conflicted with his own written report and test results. Here, in contrast, there is no additional persuasive evidence to support the ALJ’s rejection of the opinions of Morgan’s treating psychiatrist and examining psychologist. Moreover, when, as here, the disability claim revolves primarily around allegations of - mental disabilities which are by their nature difficult to diagnose, it is important to give adequate weight to the personal impressions formed by the experts who have interacted with the claimant. The standards set for an ALJ forming a conclusion as to mental capacity for sub*606stantial gainful employment have not been met.