(concurring in part and dissenting in part):
Although the evidentiary experiment performed by one of the court-martial members was improper, I agree that appellant was not prejudiced thereby. (Issue II.)
However, I dissent from the majority opinion’s approval of the military judge’s instructions. (Issue I.) Hargrove was found guilty of murdering two other soldiers by firing a round from a gun of one tank into another tank parked in line about ten feet away, this being “an act which is inherently dangerous to others and evinces a wanton disregard of human life.” See Art. 118(3), Uniform Code of Military Justice, 10 U.S.C. § 918(3). The military judge instructed that, as one of four conditions “[f]or an act to be inherently dangerous to others,” the act must “be such that its probable results, if known to the accused, would be death or great bodily harm.” (Emphasis added.) After deliberating the remainder of the day and much of the following day, the members returned to ask that certain instructions be repeated. In fact, one member specifically requested that the explanation of “wanton disregard” be repeated. Twice more, the military judge gave the quoted instruction.
At this point, defense counsel objected that the word “if” had been incorrectly inserted; and he asked for an instruction without that word. This request simply was for an instruction in conformance with the model instruction which is now on page 3-172 of the Military Judges’ Benchbook that the lethal act must “be such that its probable results, known to the accused, would be death or great bodily harm.” (Emphasis added.) However, the military judge repeated a fourth time the instruction which included the word “if.”
Appellant now asserts that the instruction with this word inserted implied that no actual knowledge of the probable results was required of the accused. Thus, the difference between murder and involuntary manslaughter was obscured. I am not enough of a grammarian to know what use of “if” would have suggested to the court members. I cannot understand why the judge insisted on retaining the word in the face of defense counsel’s objection and in light of the court members’ obvious desire to obtain a precise understanding of the meaning of the term “wanton disregard.” On balance, I cannot say that the instruc*74tion as given is so confusing as to merit reversal; but I believe that the military judge should have instructed as requested by the defense.
The military judge proposed to give preliminary instructions to the members that “lack of substantial capacity exists when there is a substantial or great impairment of that capacity, but a complete impairment is not required.” (Emphasis added.) This instruction conformed to the model now suggested on page 6-4 of the Military Judges’ Benchbook (May 1982). Defense counsel objected to the instruction. He explained that, assuming for purposes of argument that “substantial” means 75%, then “lack of substantial capacity” would mean that the accused would have 74% or less capacity. On the other hand, if again it is assumed that “substantial” means 75%, then “substantial or great impairment of that capacity" would require that the impairment be 75% or greater. Counsel argued that the latter actually is the inverse of the former, correct standard. Thus, under the correct standard, an accused with 74% capacity would not have the requisite capacity, while under the incorrect one he would. In response, the military judge proposed language that “[t]he legal criteria is a lack of substantial capacity. A complete impairment is not required.” Defense counsel agreed to this, and the military judge preliminarily instructed in this manner.
When it came time for final instructions, the military judge reversed himself. He offered two reasons for doing so: “One, I think the legal ambiguity pointed to by the defense counsel is, as I said, well recognized in the law. I don’t think it’s unique to the instructions which I propose.” Second, he pointed out that defense counsel’s questioning of his expert witnesses had been concerned with substantial impairment — which counsel had urged was the incorrect standard — and not with “lack of substantial capacity” — which counsel had urged was the correct test. After expressing concern over defense counsel’s tactics in arguing that the standard used in the model instruction was incorrect and then “rather meticulously ... questioning” his witnesses along the lines of that standard, the military judge concluded, “[I]f there is any undue emphasis on this standard, it originated not with the court, but with the defense, which ... made this subject the focus of rather pointed questioning.” Thereafter, the military judge instructed as he had originally proposed to do.
Under one interpretation the two standards are different. Moreover, I agree that the standard requested by the defense — rather than that found in the Benchbood — conforms to the American Law Institute’s criteria for mental responsibility which were adopted by this Court in United States v. Frederick, 3 M.J. 230, 234 (C.M.A.1977). Therefore, the focus must be on the reasons for the judge’s change of position.
As to the judge’s first justification, I cannot understand why his recognition of “the legal ambiguity” justifies his denial of the defense’s requested clarification. The judge’s primary reason for changing his instruction was probably his concern that counsel had trapped him by successfully arguing that the Benchbook’s model instruction was incorrect and then examining his witnesses in terms of the allegedly incorrect standard embodied in that instruction. The remedy, however, was for trial counsel or the judge to seek clarification from the witnesses while they were being questioned, rather than for the judge later to compound confusion in instructing the members.
Whatever his reason for initially agreeing to change the instruction, the fact is that the error in the model instruction was timely brought to this judge’s attention and was well-illustrated by a hypothetical. The judge had the sole responsibility for rendering correct instructions, but he failed to do so.
Usually, I would be reluctant to reverse for an instructional error like this. However, in light of the evidence in this case and the court members’ obvious concern about Hargrove’s mental state at the time of the incident, I conclude that appellant was prejudiced and that a rehearing should be granted so that his mental responsibility can be determined under proper instructions.