KASS, J.

Robert B. Haraldstad and Michael J. Purcell, the defendants, were convicted by a jury of the rape of Cathie A. and of assaulting her by means of a dangerous weapon. The victim set the incident during the early hours of July 5, 1980, near Fort Revere in Hull. She charged a third man as joining in the sexual assault.

The defense was that the incident simply did not occur; that the defendants were elsewhere when the victim said they raped her; and that her story was a fabrication. In support of that view of the case the defendants called as a witness Dr. D. Colm Armstrong, a specialist in obstetrics and gynecology, who had examined the victim after she had complained that she had been raped. Dr. Armstrong testified that no sperm were seen in the patient. Did he have an opinion based on medical certainty, the defendant Purcell's counsel asked Dr. Armstrong, whether, assuming that two or more men had ejaculated into the victim's vagina within ten hours of the examination and that she had not douched or bathed in the interim, "sperm would be present in her vagina?" Dr. Armstrong replied, "I don't have an opinion with certainty."

That answer came as a discomfiting surprise to the examining lawyer, whose next question was, "Pardon me?" A noon recess mercifully intervened seven questions later, but by that time counsel had concluded his direct examination. Following brief cross-examination by the Commonwealth, counsel for Purcell undertook to have Dr. Armstrong clarify and correct his testimony. There had been a misunderstanding. Dr. Armstrong had been thinking in terms of motile sperm (i.e., having the power to move). As to the significance of the absence of motile sperm, he was not prepared to venture an opinion with reasonable medical certainty. He was prepared to testify, it developed on an offer of proof, that he would, with reasonable medical certainty, expect to find nonmotile or dead sperm, given the facts posed in the hypothetical question, all of which had a basis in the record.

The Commonwealth objected to putting questions to Dr. Armstrong on redirect to iron out the confusion. The ground for objection was that the defense was taking a second bite at material covered by the direct examination without the prosecution's having inquired into it in cross-examination. Certainly, the evidence in question was significant. The credibility of the victim and the defendants was very much in issue. In addition to Dr. Armstrong, Haraldstad and Purcell took the stand and sixteen witnesses called by the defense gave exculpatory testimony, but those witnesses were in varying ways connected with the defendants by friendship or by blood. Objective medical evidence, therefore, carried more than ordinary weight.

Whether clarification of Dr. Armstrong's testimony was so critical to the defendants that its exclusion, standing by itself, deprived them of their right to present their case fully is a question we need not decide because of later developments in the case related to the excluded evidence.

During the course of closing argument, the prosecutor said:

"A lot is made of the fact that no sperm was found in Cathy [A]'s vagina by the doctor. Well, you heard the doctor testify. . . . [T]he critical thing that Dr. Armstrong told you is that when asked whether or not you would expect to have sperm, he said [he] would not have an opinion based on certainty. And then when asked again — I can't recall exactly the question — but when asked again, he said [he] wouldn't have an opinion."

For some fifteen additional lines of transcript, the prosecutor mined that vein. In so doing he exploited the absence of evidence he had succeeded in excluding, a practice we held to be fundamentally unfair in Commonwealth v. Mosby, 11 Mass. App. Ct. 1, 8-9 (1980). See Commonwealth v. Nordstrom, 364 Mass. 310, 316 (1973); Commonwealth v. Burke, 373 Mass. 569, 574-575 (1977). It is a tactic equally reprehensible when engaged in by a defendant. Commonwealth v. Dias, 14 Mass. App. Ct. 560, 564-565 (1982). Such conduct may be relatively harmless if the excluded evidence is without much significance or if, as in Commonwealth v. Fitzgerald, 376 Mass. 402, 418-419 (1978), the point sought to be made may be inferred from other evidence which has been received. Here, however, the proffered testimony of Dr. Armstrong about the import of the absence of sperm cannot be written off as insignificant. This was not a case in which the government's evidence was overwhelming. Exclusion of the corrected testimony, when coupled with the prosecution's references to the uncorrected testimony of Dr. Armstrong, constitutes reversible error. The point was adequately preserved by counsel for Purcell, who called the problem specifically to the attention of the trial judge and requested that he instruct the jury that they were not to draw any inferences from Dr. Armstrong's unwillingness to express an opinion about the meaning of the absence of sperm. The judge declined so to do.

Several other issues which the appellants have argued on appeal are likely to recur at a new trial and, therefore, we consider them.

1. Argument and charge on consciousness of guilt. In closing argument, the prosecutor said:

"What about the statements [by Purcell] to Detective Yannizzi [the arresting officer] that [Purcell] denied on the stand but you heard Detective Yannizzi testify to them. And that's in his report. First of all, he denied even being in Hull on July 4th and 5th. What's the significance of that? Well, I would submit to you, ladies and gentlemen, it's what we call consciousness of guilt. Something happened up there that he didn't want to admit. The reason that he lied to Yannizzi about that is because he didn't want to get in to that. I wasn't even in Hull that night. It's the classic lie."

However unintended, this was a distortion of the evidence. As for Yannizzi's report, it had never been admitted in evidence. Yannizzi's testimony was that, upon showing Purcell the arrest warrant, explaining the charges against him and giving him the Miranda warnings, he had asked Purcell "if he was in Hull at the forts that night. . . . [Purcell] told me he didn't know what I was talking about." In context that was not a denial by Purcell that he had been in Hull. Indeed, the prosecutor pressed the point by asking Detective Yannizzi, "And he responded that he was not in Hull that night?" Yannizzi answered, "He said he didn't know what I was talking about." On cross-examination the following exchange took place:

Q. "And he said initially he didn't know — Did you tell him that he was under arrest for rape, by the way?"

A. "I don't remember. He was showed the warrant. I don't remember if I said that or not."

Q. "But the warrant said on it, what?"

A. "Rape."

Q. "Rape. All right. And he said he didn't know what you were talking about, he didn't rape anyone. Is that what he said?"

A. "Yes."

When Purcell later took the witness stand, the government and the defense plowed the same ground. Without further rehearsing the testimony, it added up to Purcell having said, when presented with the arrest warrant, that he didn't know what Yannizzi was talking about.

The doubtful cast which the prosecutor gave to the Purcell-Yannizzi encounter, i.e., a flat denial by Purcell that he was even in Hull on July 4 and 5, was, unfortunately, reinforced by the trial judge's charge. Although the judge left it to the jury to decide whether Purcell had denied being in Hull, he assumed the existence of testimony that Purcell had made such a denial, an assumption the record does not support. It is error to instruct a jury on consciousness of guilt based on facts without reasonable support in the record. Morris v. United States, 326 F.2d 192, 194-195 (9th Cir. 1963). United States v. Myers, 550 F.2d 1036, 1048-1050 (5th Cir. 1977), cert. denied, 439 U.S. 847 (1978). Cf. Commonwealth v. Carita, 356 Mass. 132, 140 (1969) (that defendant not found in his usual haunts was insufficient to support consciousness of guilt demonstrated by flight); Commonwealth v. Ryan, 14 Mass. App. Ct. 901 (1982) (record did not support instruction on joint enterprise).

2. Attempt to establish lack of bias of the witness Hanson. James Hanson, one of the three implicated by the victim, was called by Haraldstad. Only fifteen years old at the time of the incident, Hanson had been tried separately and adjudged not delinquent, i.e., a form of acquittal (see note 2, supra). Without knowledge of that disposition, the jury reasonably would speculate that Hanson would give testimony favorable to the defendants (actually it was favorable only to Haraldstad) to save his own skin. To show that he was free of pressure to give biased testimony, the defendants pressed to have the fact of Hanson's exoneration received in evidence. The request was rightly refused (as the defendant Purcell concedes in his brief) on the principle that evidence of the acquittal of a person jointly charged but separately tried is not probative of the innocence of the others charged. State v. Farmer, 126 Ariz. 569, 572 (1980). Felts v. State, 546 P.2d 265, 266 (Okla. Crim. App. 1976). There are in such cases too many variables in the parallel proceeding. For example, in the first proceeding critical evidence might have been suppressed, leading to acquittal, which would not be suppressed in the proceeding against the other defendants.

Thereupon, to show Hanson's freedom from bias, the defendants requested that they be allowed to ask him if he had any charges pending against him. That request was denied, presumably because rehabilitative evidence ordinarily may not be introduced until the witness has been impeached. 4 Wigmore, Evidence § 1104 (Chadbourn rev. 1972). If, however, the opponent has insinuated the lack of truthfulness of the witness through attempted, though unsuccessful, impeachment, the party who put on the witness may adduce evidence to establish his good character and truthfulness. Commonwealth v. Ingraham, 7 Gray 46, 48-49 (1856). See State v. Farmer, 126 Ariz. at 572-573, where evidence of acquittal was allowed after the prosecution asked defense witnesses if they had been indicted along with the accused. In the instant case the prosecutor never attacked Hanson's character or credibility. The whiff of Hanson's bias was inherent in his position as an alleged participant in the crime. If a witness in one case will be a defendant in another case based on the same facts, the presumption that the witness testifies truthfully falls away, and evidence enhancing credibility, if otherwise competent, may be admitted. This is the view taken in Largin v. Commonwealth, 215 Va. 318, 318-319 (1974). It is not unusual in our practice to permit prosecutors in certain circumstances to inquire of a witness whether any deal or inducement has been offered for his testimony. See, e.g., Commonwealth v. Johnson, post 935, 936 (1983). We think that in the case at bar the judge would have acted within his discretion had he allowed the defense to show Hanson's lack of motive to falsify through a stipulation that no charge was pending against him. The resolution of that question, should it arise again, is in the judge's discretion. It is likely that the jury would have suspected bias by Hanson in favor of the defendants in any event because Hanson testified that he was a longtime friend of Haraldstad and had more recently grown friendly with Purcell. Moreover, the impact of his testimony was merely cumulative, i.e., it dealt with Haraldstad's presence on a wall of Fort Revere — and Purcell's absence — at the time the victim said she was pulled into a car by Haraldstad, Hanson and Purcell elsewhere. Thus the marginal value of having the jury know that Hanson need not face charges should be weighed against what damage the Commonwealth asserts that information may inflict on its position.

3. Leading questions to codefendant on cross-examination. Leading questions put by Haraldstad's lawyer to the codefendant Purcell were objected to, and the judge sustained the objections. Haraldstad claims improper limitation on his right of cross-examination. Had Purcell been an antagonistic witness or had he been giving testimony which implicated Haraldstad, leading questions should have been allowed. The leading question is a customary tool in the exercise of the right of confrontation. Using it, the cross-examining lawyer can probe for omissions, inconsistencies in the testimony of the witness, and aspects of his background which may bear on the credibility of the testimony. Hence codefendants who have made inculpatory statements must be made available for cross-examination. Bruton v. United States, 391 U.S. 123, 136-137 (1968). Commonwealth v. Nolin, 373 Mass. 45, 49 (1977). Commonwealth v. Hicks, 377 Mass. 1, 5 (1979).

As Purcell's testimony developed, however, it was friendly to, and exculpatory of, Haraldstad, notably in that it corroborated that Purcell had driven off alone to look for Cathie A. at a time when she said the rape occurred. Nothing Purcell said on the stand implicated Haraldstad. In the circumstances, the examination of Purcell by Haraldstad's lawyer was cross-examination in name only and the trial judge, in the exercise of his discretion, could require that the friendly witness not be led. United States v. Hansen, 583 F.2d 325, 328-331 (7th Cir.), cert. denied, 439 U.S. 912 (1978). See Fed. R. Evid. 611(c) (" Ordinarily leading questions should be permitted on cross-examination" [emphasis supplied]). The Advisory Committee's note accompanying rule 611(c) says that "[t]he purpose of the qualification `ordinarily' is to furnish a basis for denying the use of leading questions when the cross-examination is cross-examination in form only." Saltzburg Redden, Federal Rules of Evidence Manual 410 (3d ed. 1982). Compare Parker v. United States, 404 F.2d 1193, 1196-1197 (9th Cir. 1968), cert. denied, 394 U.S. 1004 (1969); People v. Brown, 27 Ill. App.3d 569, 576 (1975); State v. Mayberry, 52 N.J. 413, 422 (1968); Commonwealth v. Dreibelbis, 217 Pa. Super. 257, 262 (1970); State v. Boggs, 16 Wn. App. 682, 691 (1977).

4. Admissibility of hospital records. Although the records of the victim's examination at South Shore Hospital were substantially illegible, that deficiency was remedied by Dr. Armstrong, who deciphered his writing and read his notes into the record. The requirement of exclusion in Commonwealth v. Brattman, 10 Mass. App. Ct. 579, 586 (1980), of illegible portions of a hospital record does not, therefore, apply. Other objections to admission of the hospital records under G.L.c. 233, § 79, have been dealt with adversely to the defendant in Commonwealth v. Concepcion, 362 Mass. 653, 655-656 (1972), Commonwealth v. Lannon, 364 Mass. 480, 483-484 (1974), and Commonwealth v. DiSanto, 8 Mass. App. Ct. 694, 704 (1979), cert. denied, 449 U.S. 855 (1980). Compare Commonwealth v. Brattman, 10 Mass. App. Ct. at 586.

5. Improper closing argument. In addition to the improper aspects of the closing argument previously condemned by us, the defense protests that other parts of the prosecutor's argument were unfair. Certain of these complaints are overblown. The prosecutor's rhetorical questions and suggestions of inferences were based fairly on the evidence. His amateur psychologizing was no worse than that which the court in Commonwealth v. Johnson, 372 Mass. 185, 197 (1977), regarded with jaundiced eye, but tolerated. The prosecutor did not suggest he had knowledge of facts not in evidence. Contrast Commonwealth v. Kater, 388 Mass. 519, 533 (1983). The prosecutor is entitled to marshal the evidence, suggest inferences to the jury, and call upon the jury's common knowledge and experience. See Commonwealth v. Drayton, 386 Mass. 39, 52 (1982); Commonwealth v. Silva, 388 Mass. 495, 508 (1983).

The prosecutor, however, stepped over the foul line when he said, "Second of all, what about the way in which [the] testimony [of defense witnesses] was prepared or at least gone over, what I would call rehearsed, I would submit to you rehearsed before they came into this courtroom." Of course witnesses are prepared by competent lawyers, and to imply otherwise is shabby. It is qualitatively different from casting doubt on testimony by calling attention to extraordinary parallels between what a group of witnesses who could talk to each other have said on the stand. The prosecutor's argument was the more offensive because the use of the word "rehearse" in a similar context had been objected to earlier in the trial and ordered struck from a question.

The prosecutor would also be well advised to recognize that postrape behaviour by victims is an area of specialized knowledge, Terrio v. McDonough, ante 163, 175-176 (1983), and on retrial it would be better that the prosecutor avoid asking the jury to speculate on the subject. See generally Commonwealth v. Ryan, 8 Mass. App. Ct. 941, 942 (1979).

Other issues raised by the defense are not likely to recur. We are quite sure the prosecutor will not ask Haraldstad again, "You didn't testify at that [probable cause] hearing?" This was a flat violation of G.L.c. 278, § 23, which the prosecutor conceded and in response to which the judge gave prompt and effective curative instructions.

Judgments reversed. Verdicts set aside.