PER CURIAM.
Appellant was convicted on three counts of violation of narcotics laws; three concurrent sentences were imposed. At trial the principal defense was that appellant was not guilty by reason of insanity. Various errors in the conduct of the trial are asserted and we find that they are without merit save as to one contention, i. e., the claim of undue intervention in the trial by the judge in a manner prejudicial to the defendant.
The appellant’s brief of necessity takes utterances and questions of the trial judge out of context and there is no way to evaluate his claims of undue and prejudicial intervention except by an examination of the entire transcript of the trial, which we have done, in order to be sure that we “guard against the magnification on appeal of instances which were of little importance in their setting.” Even a close examination of a transcript cannot, as everyone experienced in litigation knows, truly reflect the trial itself. Sometimes a trial judge intercedes because of seeming inadequacy of examination or cross-examination of witnesses by counsel; sometimes to draw more information from reluctant witnesses or experts who are either inarticulate, less than candid or not adequately interrogated. This is permissible, of course.
At best it is difficult on appellate review to appraise the impact of intervention by the presiding judge and determine whether his participation exceeded permissible bounds. However this transcript reveals what seem to us an inordinate number of instances of extensive ■examination and cross-examination of witnesses and comments by the court. Fairly read, no single comment or question, or line of questioning, can be regarded as prejudicial, but the cumulative impact of all the trial judges activist participation could well have been prejudicial at the very least and could have led jurors to give undue weight to points treated by the judge. In this case the responses elicited by the judge were largely adverse to appellant. In itself this does not render the judicial intervention impermissible but in it were the ■seeds of tilting the balance against the .accused and casting the judge, in the eyes of some jurors, on the side of the prosecution. This risk is always present when a presiding judge undertakes ■to interrogate witnesses at length. If a trial judge has definite ideas as to what lines of inquiry ought to be pursued, he is free to call both counsel to the bench, •or in chambers and suggest what he ■wants done. That the judge may be able to examine witnesses more skillfully or •develop a point in less time than counsel requires does not ordinarily justify such participation. That is not his function.
There are and can be no hard and fast rules as to how much questioning a judge may or should engage in because what would be appropriate in one setting would be otherwise in another. •One obvious general rule is that, since the judge is something more than a moderator, but always a neutral umpire, the interrogation of witnesses is ordinarily best left to counsel, who presumably have an intimate familiarity with the case. A presiding judge can control the trial without participating actively in examination of witnesses. In a non-jury case, as in an appellate court, needless or active interrogation by judges, although not always helpful, is rarely prejudicial. But in a jury case, a trial judge should exercise restraint and caution because of the possible prejudicial .consequences of the presider’s intervention. Cf. United States v. Paroutian, 299 F.2d 486 (2d Cir. 1962).
On the whole record we cannot say, with that degree of assurance required in a criminal case, that the activities of the trial judge may not have prejudiced the defendant, notwithstanding the strong evidence presented against him. Accordingly there must be a new trial.
Reversed and remanded for a new trial.
. Glasser v. United States, 315 U.S. 60, 83, 62 S.Ct. 457, 471, 86 L.Ed. 680 (1942).