(dissenting in part):
This is the rare state prisoner habeas-appeal where there is reasonable doubt whether petitioner committed the crime. At the time of the alleged rape the victim did not see her assailant’s face; she caught only a glimpse of his clothing and noticed that he was wearing a green upper garment. In direct testimony she claimed to have seen Cannon standing across the street from her while she was walking home and to have seen him walking through a gas station across from her shortly before she was attacked, but this was substantially weakened on cross-examination. While I agree that, at minimum, there must be a determination whether the lineup was impermissibly suggestive, I would go further and direct that even if this investigation should be concluded favorably to the State, the writ should nevertheless issue on the usual conditions because the prosecution was improperly permitted to introduce evidence of the contents of the typewritten statement.
*269The prosecutor’s use of the typewritten statement to rehabilitate Detective Funk was highly questionable from the first. In the preliminary hearing, Funk testified unequivocally that Cannon was first questioned about the Rippel incident at about 9:30 a. m. This would be logical since the first written statements, each preceded by oral questioning, related to the Stevens and Sprague attacks. At trial, however, Funk changed his story and said that Cannon had been questioned orally about the Rippel incident at about 8:35 or 8:40 a. m. in between the oral and written statements on the Stevens attack. The difference, of course, was critical — and highly suspect — since the trial judge had found that Cannon requested an attorney in the course of the stenographic statement taken at 8:45 a. m. and had therefore properly excluded any of Cannon’s admissions made thereafter. On cross-examination, defense counsel read at length from Funk’s testimony at the preliminary hearing. Funk insisted that this testimony was accurate, in spite of the clear conflict between it and his testimony on direct examination at the trial. Defense counsel concluded his cross-examination with the following exchange:
Q. Well, at that time, fifteen days after the questioning, was that a true statement you made that at about 9:30 you questioned him about this incident?
A. Yes, sir. That was the first statement we took in regards to this incident.
Q. At 9:30?
A. At 9:30, yes sir.
Immediately thereafter, on redirect examination, the prosecutor seized upon the word “statement” in Funk’s penultimate response and utilized the latent ambiguity in the word as a wedge:
Q. What do you mean by “statement,” Detective?
A. Statement would be something which was transcribed onto paper. In other words, a stenographic statement or a statement which we would type ourselves in response to our questions.
Q. So, you’re referring to these later questions, or statements as you call them, after nine o’clock to what?
A. To the statement which we took from Mr. Cannon on the typewriter.
While it hard to avoid believing that the detective and the prosecutor had prepared this “rehabilitation” by arranging for the detective to use the word “statement” on cross-examination, we need not go on that ground. Once the ambiguity about the word “statement” was established and Detective Funk asserted he had been referring to the later typewritten statement, the rehabilitative purpose of the redirect examination was achieved. Yet the prosecutor did not rest at that point but proceeded to introduce as much evidence about the contents of the statement as he could.
The prosecutor first asked Detective Funk to describe how the statement was taken; he then elicited the testimony that Cannon had signed the statement; and finally, and most important, he asked, “And in sum or substance does that document relate what the defendant previously told you orally?” Funk answered, “Yes, sir, it is.”
This testimony brought out critical information about the contents of the typewritten statement, which went well beyond any legitimate rehabilitative purpose. Coming in the wake of detailed accounts of Cannon’s alleged oral admissions, Funk’s testimony amounted to far more than an “oblique” reference to the contents of the typewritten statement, as the majority characterizes it. In effect, it constituted direct evidence that Cannon had signed a written confession of guilt, which the jury could use to convict him or to buttress testimony about his alleged earlier oral admissions.
The majority concludes that the judge’s instructions to the jury were sufficient to cure any error in admitting evidence *270of the contents of the statement. I disagree. After having deliberated for about two hours, the jury asked the court, “Can we consider the typewritten statement of Cannon as evidence?” The court responded as follows:
My response to that ... is that a typewritten statement of Cannon was alluded to during the trial and was marked for identification. Reference was made to one on one or more occasions, that one was typed, where it was typed and some of the circumstances concerning it.
Now, these references are testimony and may be considered by you. However, the statement itself was never introduced into evidence, and therefore you may not consider it as evidence. All right?
The jury’s question makes it clear that they were seeking advice whether they* could consider the testimony about the contents of the typewritten statement as evidence of Cannon’s guilt. The natural interpretation of the court’s answer would be that they could consider all the testimony about the statement for any purpose, but they could not consider “the statement itself,” since the piece of paper containing the statement had never been introduced into evidence. Under this direction, it seems entirely likely that the jury understood that it could consider Funk’s testimony that the typewritten statement and the oral statement were identical in “sum'and substance,” and that it could rely on Funk’s characterization of the contents of the typewritten statement in determining guilt or innocence.
Partly on the basis of its erroneous analysis of the court’s instruction, the majority proceeds to conclude that Harris v. New York, 401 U.S. 222, 91 S.Ct. 643, 28 L.Ed.2d 1 (1971), rescues the conviction. The argument is that since the judge could have allowed the prosecutor to introduce the typewritten statement to impeach Cannon, his error in admitting evidence of the contents of the statement “would have had no independent significance had the trial judge allowed the prosecutor proper latitude in cross-examination.” However, use of the contents under Harris would have had to be strictly limited to impeachment. Assuming, as the Court did in Harris, that a properly instructed jury can distinguish between use for impeachment and use in the case in chief, the introduction of the typewritten statement solely for impeachment would have had no such destructive effect as it did. Although the majority appears to doubt a jury’s power to make such fine distinctions, the belief in this ability lies at the heart of Harris. Moreover, the jury in this case seemed exceptionally receptive to limiting instructions. The question to the trial judge invited an instruction that they must ignore the contents of the typewritten statement. Unhappily, while that instruction was required, it was not given.