A Superior Court jury found each of the defendants guilty of two counts of aggravated rape, two counts of rape of a child through use of force, and two counts of indecent assault and battery on a child under the age of *449fourteen. The complainants, Nancy and Alan,2 are Margaret Brouillard’s biological children. The children were eight and five years old, respectively, at the time of trial; and six and three years old at the time of the alleged abuse. Robert Brouillard is Margaret’s husband, but is neither the biological nor adoptive father of Nancy and Alan. The defendants contend that they suffered a substantial risk of a miscarriage of justice because the complainants’ treating therapist impermissibly vouched for their credibility; because contemporaneous limiting instructions were not given to the jury before the fresh complaint witnesses testified; and because the word “corroborate” in the fresh complaint instruction that was given was not defined.3 We reverse for the reasons discussed below.
1. Vouching for the complainants’ credibility by the treating therapist. Miles Tarter, treating therapist to Nancy and Alan, testified as an expert witness at trial on syndromes associated with sexual abuse, as well as in the capacity of treating therapist and as a fresh complaint witness. The defendants complain that the effect of Tarter’s testimony in these multiple roles was to vouch impermissibly for the Complainants’ credibility. We agree.
Tarter had served as the children’s therapist for more than two years before the trial commenced, and continued to see them professionally on a weekly basis at the time of trial. He gave extensive fresh complaint testimony at trial, recounting in detail what Nancy and Alan had told him about sexual acts allegedly perpetrated on them by the defendants. At various times during his testimony, Tarter linked his observations *450of the complainants and their behaviors to general behaviors or syndromes associated with sexual abuse. In one such exchange, Tarter interspersed his opinion that Nancy was at risk for developing multiple personality disorder with his expert opinion that the disorder is associated with “cases of long term chronic and severe sexual abuse.”4 At another point in his testimony, Tarter informed the jury that the literature regarding true versus false reports from children alleging sexual abuse indicated that children who could not give details of the alleged incidents had likely been coached in their testimony. He then testified that Nancy and Alan had “been giving detail for two years and four months. Great detail, yes.”5 In a third such instance, Tarter testified that he had diagnosed Alan with posttraumatic stress disorder. While he was prevented (due to defense objections sustained by the trial judge) from testifying about the general symptoms and causes of the disorder, Tarter did then testify (over vigorous objection) to a pattern of disclosure among children exposed to “traumatic incidences.”
“Most of the children that I’ve worked with disclose traumatic events piece by piece because it’s so difficult to remember and to go back and experience the events over again by telling them. It’s most typical for disclosures to come out piece by piece. It’s been a very rare experience where the first time a child discloses some traumatic incident like sexual abuse that the first time that they tell the story they’ve told you everything that they’re going to tell you about it. I can’t even remember when that’s ever happened.”
Lastly, on cross-examination, Tarter expressed his concern that Nancy was at risk for developing an eating disorder, then proceeded to offer his (unsolicited) opinion that:
“the link between childhood sexual abuse and the development of eating disorders is fairly well established .... Childhood sexual abuse is the most com*451mon factor in the histories of people who present with eating disorders.”6
It is well established that fresh complaint testimony is not admissible as evidence of the charge, but merely as corroborative of the complaining witness’s in-court allegation. Commonwealth v. Powers, 36 Mass. App. Ct. 65, 70 (1994), citing Commonwealth v. Licata, 412 Mass. 654, 660 (1992). While it is theoretically possible for a fresh complaint witness also to testify about the general characteristics of abused children without thereby impermissibly vouching for the complainants’ credibility, we have suggested that this practice be avoided where possible, given the risk that the juiy will give substantive effect to such a witness’s fresh complaint testimony. See Commonwealth v. Swain, 36 Mass. App. Ct. 433, 444-445 (1994); Commonwealth v. McCaffrey, 36 Mass. App. Ct. 583, 593-594 (1994). In Commonwealth v. Rather, 37 Mass. App. Ct. 140, 148-149 & n.4 (1994), we found error in admitting expert testimony regarding general characteristics of sexually abused children where the expert was not an independent expert, but had evaluated the children for sexual abuse, and, where the sequence of questions posed to the expert juxtaposed questions about the complainant with questions about patterns of disclosure among sexually abused children.7 These Rather factors are present here. Tarter was not an independent expert, having treated the children for over two years, and the testimony cited above juxtaposed discussion of general syndromes with specific descriptions of and opinions about the complainants.
Tarter’s testimony, intermingling his role as treating therapist, expert on behavioral characteristics of sexually abused children, and fresh complaint witness, had the effect of impermissibly vouching for the complainants’ credibility. While an expert may testify regarding general syndromes associated with sexual abuse, Commonwealth v. Dockham, 405 Mass. 618, 628-630 (1989); Commonwealth v. Mamay, 407 *452Mass. 412, 421-422 (1990), an expert who has also treated a complainant may only so testify if he does not explicitly or impliedly connect the complainant to the syndrome. Commonwealth v. Hudson, 417 Mass. 536, 541, 543 (1994). Commonwealth v. O’Brien, 35 Mass. App. Ct. 827, 833 (1994). We have also cautioned that a treating therapist who testifies only to general syndromes might nonetheless be viewed as implicitly endorsing the complainants’ testimony by virtue of the fact that she has accepted them into therapy. Commonwealth v. McCaffrey, 36 Mass. App. Ct. at 592. However, even putting aside this concern about implied endorsement of complainants, Tarter far exceeded permissible testimonial boundaries. He explicitly connected the complainants to general syndromes associated with sexual abuse, thereby impermissibly vouching for the complainants and invading the jury’s province of assessing witness credibility. Ibid. See Commonwealth v. Montanino, 409 Mass. 500, 504 (1991).
The defendants complain that the problem was compounded when Tarter further endorsed the children’s credibility by indicating that he believed the complainants’ allegations. We agree. In addition to connecting the complainants to general syndromes of sexual abuse, Tarter conveyed to the jury his belief in the complainants’ credibility. He offered explanations as to why the children’s allegations changed over time, and why Nancy, the older child, initially lied about her mother’s participation in the alleged abuse.8 In response to a question from the prosecutor regarding Nancy’s “behaviors” associated with the alleged abuse, Tarter responded,
“. . . major sleep disturbance. Same thing for [Alan]. Basically both of these kids, I don’t think, have ever had a restful night of sleep. They’re always terrified at night. Probably because they remember that’s when the abuse happened most frequently. So, there’s very major serious sleep disturbance.”
In response to another question from the prosecutor about *453why Nancy had purportedly lied at first, when she said her mother was not involved in the alleged abuse, Tarter said,
“. . . certainly there is a great deal of shame for both of the kids involved, feeling that it must have somehow been their fault; or you know, why did these things have to happen to them; they do [szc] something wrong to make these things happen to them.”
When asked by the prosecutor how long it had taken Nancy to “tell you what happened completely,” Tarter responded, “To completely disclose? Personally I don’t think she’s done. I think that there is more.”9 The effect of all of these statements was to show that Tarter personally believed the complainants’ allegations to be accurate reports of historical fact. He thereby vouched impermissibly for the children’s credibility.
Tarter’s vouching for the children was particularly egregious because, given his considerable experience10 in dealing with sex abuse, the jury likely perceived him “as especially qualified to judge the credibility of children.” Commonwealth v. Powers, 36 Mass. App. Ct. at 70. Moreover, his cited testimony conveyed to the jury that Tarter believed that the children had in fact been sexually assaulted. While expert opinion testimony will not necessarily be excluded because it reaches the ultimate issue in a case, Commonwealth v. Colin C., 419 Mass. 54, 59 (1994), the court has never “allow[ed] an expert to testify that an alleged victim was in fact sexually assaulted.” Id. at 60. Admission of the foregoing testimony by Tarter was therefore error.
Because counsel did not object11 to all the impermissible *454testimony12 we consider whether the testimony, which we have found impermissible, created a substantial risk of a miscarriage of justice requiring reversal. See Commonwealth v. Freeman, 352 Mass. 556, 563-564 (1967); Commonwealth v. Miranda, 22 Mass. App. Ct. 10, 16-17 (1986).
The evidence was not overwhelming. As is often true in such cases, the Commonwealth’s case rested entirely on the credibility of Nancy’s and Alan’s testimony. There was no physical evidence of sexual abuse nor other eyewitnesses. The children’s credibility was shaky, making Tarter’s bolstering of their credibility especially problematic. Testimony at trial indicated that Nancy could not identify the consequences of lying, and claimed that “God made her lie.” During the pretrial competency hearing, Alan was shown to be quite suggestible, readily agreeing with contradictory propositions posed by opposing counsel. Alan did not tell Tarter that Margaret Brouillard participated in the alleged abuse until he was asked a leading question by Tarter in a therapy session. Nancy did not name Margaret Brouillard until, in another therapy session, Tarter told her about Alan’s allegation. There was ample opportunity for the children to confuse the defendants’ alleged acts with abusive treatment Nancy and Alan apparently received at the hands of other adults. There was testimony that Nancy had been abused by an earlier therapist, who handcuffed Nancy, told her she was bad, and pretended to urinate on the floor. This occurred during therapy sessions conducted at the Brouillard home, and later at Nancy’s foster home. Nancy had also apparently been abused by Margaret’s ex-husband, who is Alan’s biological father. The children testified that Robert Brouillard had tattoos on his arms. However, evidence was introduced at trial that Robert had no tattoos on his arms, but that Alan’s biological father did. There was evidence that the children had been abused earlier by several teenage babysitters. In addition, the adults questioning Nancy made no effort to clarify with her that her allegations about what “Robert” or “Daddy Bob” had done *455to her and her brother concerned Robert Brouillard, and not her foster father, whose first name is also Robert. We therefore find that Tartar's improper bolstering of and vouching for the complainants’ credibility created a substantial risk of a miscarriage of justice requiring reversal.
2. Fresh complaint instruction. The defendants complain that they suffered a substantial risk of a miscarriage of justice because there were no contemporaneous fresh complaint instructions during any of the four fresh complaint witnesses’ direct testimony. The only contemporaneous instruction came during the cross-examination of the third fresh complaint witness, Miles Tarter. In addition, the defendants complain that the word “corroborate” was not defined for the jury, thus rendering the fresh complaint instruction ineffective. We do not agree that the judge failed to make clear the meaning of “corroborate.”13 However, we do find that the lack of contemporaneous fresh complaint instructions also independently created a substantial risk of a miscarriage of justice.
The first two fresh complaint witnesses were Lorraine Le-Page, a child abuse investigator, and Nancy’s foster mother, Terry Haley. LePage testified that Nancy told her that, “Daddy touches my pee-pee,” and that when he did that, her mother was “out.” Haley gave graphic testimony describing what Nancy had told her about the sexual acts “Bob” and her mother had committed against her, and testified that Nancy said they laughed while they did these things. On redirect, Haley repeated the direct testimony, and expanded on it, adding that Nancy had said that, “when they were finished with her, they did [Alan]; and then they made the two children perform the same acts on each other thát they *456had done.” No fresh complaint instruction was given during LePage’s or Haley’s testimony.
Similarly, no instruction was given during the testimony of Alan’s foster mother, Donna Rogers, who testified that Nancy had told her that “Daddy Bob” had touched her. She then gave graphic testimony describing what Alan had told her regarding the sexual acts “Daddy Bob” had committed against him, and that Margaret Brouillard had held Alan down while Robert Brouillard committed these acts. Mrs. Rogers’ testimony was highly emotional, interjecting comments concerning her personal feelings about what Alan told her:
“I just couldn’t see how a mother could do it, could hold him down. He was just a little boy. And he was telling me that she held him down, and that Bob was there doing this to him. And I just couldn’t understand how she could do that to that baby.”
Lastly, no instruction was given during therapist Tarter’s direct testimony, when the bulk of the impermissible vouching noted earlier occurred. It was only during his cross-examination that a fresh complaint instruction was given.14
Here, as in Commonwealth v. Trowbridge, 419 Mass. 750, 760-762 (1995), where four fresh complaint witnesses testified, one of them as both an expert and a fresh complaint witness, it is plain that the greater the number of fresh complaint witnesses, the greater the risk that fresh complaint testimony will be taken by the jury as substantive evidence that the al*457leged crimes were actually committed. While there is no per se rule about the number of permissible fresh complaint witnesses,15 it is well to remember that the purpose of the contemporaneous fresh complaint instruction recommended in Commonwealth v. Licata, 412 Mass. 654, 660 (1992), is to assure that the jury will not use fresh complaint testimony as substantive evidence. Commonwealth v. Trowbridge, 419 Mass, at 761. Here, the intermingled vouching and fresh complaint testimony by Tarter during his direct examination was permitted without the benefit of a contemporaneous fresh complaint instruction, giving rise to the great risk that the jury would take his fresh complaint testimony as substantive evidence. Rogers’ testimony was also damaging because of its emotional nature, and Haley’s bécause of its graphic detail. We therefore find that there is a substantial risk of a miscarriage of justice, even though Licata had not been decided at the time of trial. See Commonwealth v. Trowbridge at 761. This is not a case like Commonwealth v. Rather, 37 Mass. App. Ct. at 149-150, where there were other eyewitnesses to the alleged abuse, or Commonwealth v. Houghton, 39 Mass. App. Ct. 94, 98-99 (1995), where there was substantial physical and circumstantial evidence linking the defendant to the crime. This case depended entirely on the credibility of the complainants, which was improperly bolstered by impermissible expert testimony, and by the fresh complaint testimony, given without contemporaneous limiting instructions.
The judgments are reversed, and the verdicts are set aside.
So ordered.
Pseudonyms are used for the children.
The defendants also complain that there was a “piling on” of fresh complaint testimony, that the convictions on the charges of indecent assault and battery of a child and rape of a child was duplicative, that the motion judge erred in finding the children competent to testify, and that their counsel provided ineffective assistance. In light of our disposition, we need not reach the defendants’ claims of ineffective assistance; nor need we reach the competency issue as that issue must be addressed anew at any retrial. As to the defendants’ duplicative conviction argument, we agree with the Commonwealth that, although indecent assault and battery of a child is ordinarily a lesser included offense of rape of a child, the facts in this case, including numerous allegations of sexual abuse “separate from and unnecessary to the acts of penetration,” support the separate convictions. Commonwealth v. Fitzpatrick, 14 Mass. App. Ct. 1001, 1002-1003 (1982). For a discussion of the issue of “piling on” of fresh complaint testimony, see note 15, infra, and accompanying text.
Counsel for both defendants objected to this line of questioning.
Defense counsel did not object to this testimony.
Defense counsel did not move to strike this testimony.
We note that additional factors contributed to the error in Rather, which are not present here. In particular, hypothetical questions matching the complainant’s assertions and behavior were used in examining the therapist in Rather, while hypothetical questions were not employed here. See also Commonwealth v. Perkins, 39 Mass. App. Ct. 577, 583-585 (1995).
The prosecution proceeded against Margaret Brouillard on the theory that Nancy had lied when she initially said that Margaret was not present during the abuse committed by Robert; and that Nancy then told the truth when she later said that Margaret had not only been present, but had been an active participant in the alleged abuse.
Defense counsel did not object to any of these three statements by Tarter.
Tarter testified that, at the time of the trial, he was a clinical therapist on the sexual abuse treatment team of the New Bedford Child and Family Service office in Fall River; he had a masters degree in psychology and was presently enrolled in a doctoral program in clinical psychology in which he was specializing in child sexual abuse; and he had counseled more than 200 children in approximately six and one-half years of working with children.
We note that trial was in December, 1991. While Commonwealth v. Montanino, 409 Mass. 500, was decided prior to trial, many of the major cases involving vouching by treating therapists and experts were published *454in 1994. See, for example, Commonwealth v. Powers, 36 Mass. App. Ct. at 70; Commonwealth v. Swain, 36 Mass. App. Ct. at 444-445; Commonwealth v. McCaffrey, 36 Mass. App. Ct. at 592-594; Commonwealth v. Rather, 37 Mass. App. Ct. at 148-149.
Counsel objected to two of the seven impermissible statements discussed above.
In his instruction during the cross-examination of Miles Tarter, the judge described the purpose of fresh complaint testimony as bearing on the “credibility of the witness” [i.e., the complainants]. In his charge to the jury, the judge explained that, “. . . fresh complaint is only taken not as positive evidence of the occurrence of the events mentioned, but on the question of corroboration of the statement made by the witness here in the courtroom. You may take such statements into account when you evaluate the credibility of the witness’ testimony here on the stand. You may not consider such statements as positive evidence that the rape or sexual offense did, in fact occur.” This explanation was sufficient to convey to the jurors that they could not use the fresh complaint testimony as substantive evidence. See Commonwealth v. Scanlon, 412 Mass. 664, 674 (1992); Commonwealth v. Gardner, 30 Mass. App. Ct. 515, 523 (1991).
We note that at no prior time had counsel requested a fresh complaint instruction and that the law at the time was not clear that contemporaneous instructions were required. Moreover, the fresh complaint instruction that the trial judge did give, during Tarter’s cross-examination, was given sua sponte. Since counsel on both sides had not objected to a plethora of irrelevant hearsay throughout the trial, the judge’s task of conducting a fair trial was made exceedingly difficult, and troubled him greatly. Indeed, the judge indicated his concern to counsel on a number of occasions. During the trial he commented, “So much of this has not only been hearsay without objection, but so much of it seems to be quite extraneous to the indictments.” After closing arguments the judge said, “I truthfully — and I will say for the record, that if it wasn’t for the experienced defense ... I really would have felt I had to interrupt [because of the excessive hearsay]. . . . If [defense counsel] were both novices ... I think I would have had to [interject] myself in the trial [to] stop[ ] that hearsay coming in.”
The defendants argue that the sheer number of fresh complaint witnesses was prejudicial and constituted impermissible “piling on” of fresh complaint testimony. The use of four witnesses in a case involving two victims and two defendants would not, in itself, be impermissible, if the other infirmities recited above were not present here.