Paoli & Paoli and Sylvia L. Paoli, Tustin, under appointment by the Court of Appeal, for the Minors.

SILLS, P.J.

Mere weeks before oral argument in this case, the adoptive placement that had appeared to have been working out at the time of the .26 hearing [1] utterly failed. The prospective adoptive parents returned the two boys who are the subject of this appeal, Jayson, now going on 10, and Joseph, going on 7, to Orangewood Children's Home only a few weeks ago.

The case forces us to examine the recurring problem in juvenile dependency law of postjudgment developments during the pendency of an appeal. (E.g., In re Elise K. (1982) 33 Cal.3d 138 , 187 Cal.Rptr. 483 , 654 P.2d 253 [finding it appropriate to accept stipulation reversing order terminating parental custody and control based on postjudgment circumstances].) In this case, the postjudgment developments call into question one of the foundations on which a termination of parental rights must necessarily be basedthat children who have had their ties to their parents legally severed will likely be adopted. (See § 366.26, subd.(c)(1); see In re Elise K, supra, 33 Cal.3d at p. 149 , 187 Cal. Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.) [because termination proceedings required findings of adoptability, evidence should be taken of postjudgment developments].) It is a problem not specifically addressed in the statutes.

There are two basic but irreconcilable models available. The first model is the traditional model of appellate procedure. If there was no error at the .26 hearing, then it makes no difference if the children who are the subject of the proceedings are not really adoptable. The duty of the appellate court ends with examination for error at the trial level. No error, no reversal; the appellate court must be blind to the implications of postjudgment developments for the child.

The second model is based on the underlying reason for the dependency statutes in the first placeto protect children. If the government is going to terminate a child's ties to his or her natural parents, then there must be reasonable certainty that the child is not going to be left a legal orphan. If postjudgment developments cast doubt where the law requires certainty, there is no harm in allowing the trial court to take a second look at a child's adoptability.

Until our Supreme Court or Legislature tells us otherwise, we conclude the second model more closely represents the present intent of the Legislature and is better public policy. A child should not be condemned to legal orphanage merely because possible problems with his or her adoptability were, as in this case, not discovered or glossed over by the trial court. As long as the order terminating parental rights is not yet final, a court should be able to examine whether the child is still likely to be adopted. We shall therefore reverse and remand for an updated review hearing as to whether the two boys really are adoptable. As we explain below, if they are not adoptable, it would be a travesty of the juvenile dependency law to terminate parental rights.

At the updated review hearing the court should ascertain whether the children still are adoptable in light of the new information brought to this court's attention during the pendency of the appeal pursuant to a motion to take additional evidence. [2] If they are adoptable, then parental rights should be summarily terminated. If, however, the court finds that the children are not now adoptable (and minor's counsel has brought to our attention some very serious reasons to believe that they are not), then the court should consider an appropriate order under section 366.26, consonant with not terminating parental rights.

Ordinarily, this appeal would result in an affirmance because of a deferential standard of review. The incident that precipitated the dependency was an incident in August 1999 in which the mother, Melissa, threw a knife at her sister and threatened her with a barbecue fork, and in which Melissa held a knife to Joseph's neck. The evidence at the jurisdictional and dispositional hearing showed that Melissa was suffering from "bipolar" disorder. (Indeed, Melissa's admittance to a residential psychiatric treatment center coincided with the jurisdictional and dispositional hearing in February 2000.) There was expert testimony that Melissa would need to consistently take her medication, lest she experience recurrent attacks of mental illness.

Things hit a low for Melissa during the first six months of reunification services, ending October 2000. She hadn't completed parenting classes, nor was she "compliant" with the requirements of medication or psychological counseling in her reunification plan. [3] Social workers reported that Melissa would exhibit some "bizarre" behavior during her weekly monitored visitation, most notably vacillating between kindness and detachment toward the children.

The evidence on her behalf at the 12-month review hearing held in March 2001 was a little better. She had completed a ten-week parenting class, and she had not missed any visits without good excuse (e.g., the foster parents being on vacation), and she had taken her medication since March 2000 except for periods when she had run out and couldn't get a doctor's appointment for a refill. Even so, the trial court terminated reunification services and set the case for a .26 hearing.

The actual .26 hearing did not take place until late September 2001, giving Melissa time, in August, to file a section 388 petition. She averred, in her supporting declaration, that since reunification services had been terminated she had commenced a 52 week child abuse prevention and parenting program and completed 22 weeks of it. She had stayed on her medication since January 2001. She had been gainfully employed since April. She was seeing a psychiatrist and a clinical social worker. And she had maintained constant visitation with the two boys.

However, under In re Kimberly F. (1997) 56 Cal.App.4th 519 , 65 Cal.Rptr.2d 495 , a case conspicuously avoided by Melissa's appellate counsel, Melissa's showing would not have been enough, tested on an abuse of discretion standard, to require reversal of the order denying the modification request. A psychiatrist (ironically called by Melissa herself) testified Melissa suffered from "schizoaffective disorder," which results in alternating depression and agitation. The psychiatrist was also not "convinced" of Melissa's ability to "stay compliant with medication in the future." The children had told a social worker that they wanted to be adopted, even if it meant they could not see their mother and grandmother again. While Jayson testified that he wanted to continue to see his natural mother no matter what, Joseph said he didn't want to live with his mother and grandmother.

In sum, the hearing resulted in evidence that (a) the reason for the dependency was very serious, (b) the relative bonds between the caretakers and the parents were neutral (as far as Melissa's case was concerned), and (c) there was reason to doubt that Melissa's changebasically now taking her medicationcould be trusted for the indefinite future. Under an abuse of discretion standard, reversal would certainly not be compelled. (See In re Kimberly F., supra, 56 Cal.App.4th at pp. 532-533, 65 Cal.Rptr.2d 495 .) [4]

Melissa's appellate argument thus tries to sidestep the Kimberly F. factors by basically asking this court to create an automatic judicial exception to the basic time frames in the dependency statutes for mentally ill natural parents. She argues that section 388 does not provide an adequate "escape mechanism" for the lastminute reform by the mentally ill. (See In re Marilyn H. (1993) 5 Cal.4th 295 , 309, 19 Cal.Rptr.2d 544 , 851 P.2d 826 [importance of availability of modification proceeding as "escape mechanism" to the general constitutionality of state's dependency statutes].)

Melissa's case doesn't quite fit the model of In re Elizabeth R. (1995) 35 Cal.App.4th 1774 , 42 Cal.Rptr.2d 200 , which does indeed stand for the idea that periods of hospitalization for mental illness can furnish a valid reason for exceeding the 18-month maximum for reunification services. True, both cases involved periods of hospitalization for mental illness (though here the total period was less than three months, while in Elizabeth R. the hospitalization took up all but five months of an 18 month reunification period, see id. at p. 1777, 42 Cal.Rptr.2d 200 ). But the two cases present somewhat contrasting ways with which a parent deals with mental illness. In Elizabeth R., the parent recognized the signs of impending psychotic episodes, and in each instance voluntarily sought hospitalization. (See id. at pp. 1779-1780, 42 Cal.Rptr.2d 200 .) Here, by contrast, the hospitalizations were involuntary, because early on Melissa refused to take her "meds." Thus Melissa would not have, as was the case in Elizabeth R., a record showing "exemplary" compliance with the reunification plan. (See id. at pp. 1790-1791, 42 Cal.Rptr.2d 200 .) Given that there was some evidence here that Melissa could not be trusted to always take her medication, Elizabeth R. doesn't help her.

To the degree that Melissa contends this court should extend the rule of Elizabeth R. to all cases of mental illness, we must decline the invitation. The Legislature has made no special provision for mental illness. Mental illness is treated like drug addiction. (See § 300(b) [both mental illness and substance abuse are treated the same].) The same clock begins running on the maximum time for reunification services. (See § 361.5.)

True, as Melissa suggests, a mentally ill parent may not be, at the commencement of a dependency, fully appreciative of the need to demonstrate consistent "compliance" with the need to take medication, and a proper reunification plan should be designed with that in mind. (Cf. In re Elizabeth R., supra, 35 Cal.App.4th at p. 1790 , 42 Cal.Rptr.2d 200 [need for reunification plans to accommodate mental illness].) Any deficiency in the reunification plan was, however, a matter for her trial counsel at the dispositional hearing, when any deficiencies in the plan could have been remedied. It is enough to say that Melissa's appellate counsel has not brought any proceeding to assert Melissa was the victim of ineffective assistance of trial counsel in that regard. [5]

But things happened after the .26 hearing that force us to rethink what the result should be. The children's appellate counsel has brought to light some disturbing facts casting serious doubt on their adoptability. Counsel first filed a minor's brief, sans declaration, to the effect that she favored reversal because the current adoptive parents were considering sending the boys back to Orange County Social Services "because of problems with the boys that" the adoptive parents are "afraid might adversely affect" their two natural children. Apparently, they believe that Joseph has "reactive attachment disorder, in a rather severe way." Counsel was worried that it was a clear signal the prospective adoptive parents were about to give up.

A couple of weeks then passed. The social services agency made a motion to strike minor's appellate brief based on its lack of evidentiary foundation. Appellate counsel (who, as she explained at oral argument, had had a death in the family and had not been able to prepare a formal motion to take evidence to accompany her minor's brief when it was filed), then made a motion to take additional evidence. [6] The additional evidence went beyond the mere possibility that the placement might fail. By the time it was made, it had failed. Counsel's declaration averred that she had just learned that the two boys had already been returned to Orangewood Children's Home. (The return took place sometime between January 19, 2002 and February 7, 2002.)

Counsel also averred that the prospective adoptive mother had now concluded that even Jayson was "possibly affected" with reactive attachment disorder. The declaration also stated that social workers were apparently considering the two boys (and a younger sister who is not the subject of this appeal) to be a "sibling set," so that the prospective adoptive parents were not allowed to keep just the sister, or the sister and Joseph. Moreover, Jayson told his counsel "repeatedly" that he had been "moved around too much and [the two boys] did not want to move anymore." And to add to their sense of dislocation, the boys had "been told" they could not stay at Orangewood, a location they liked.

"Reactive attachment disorder" is a psychological condition that means an inability to form loving attachments. It entails difficulty bonding, poor impulse control, and insensitivity to the needs of others. (See Department of Social Services v. Superior Court (1997) 58 Cal.App.4th 721 , 68 Cal. Rptr.2d 239 ; Comment, Transnational Adoption: The Creation and III Effects of An International Black Market Baby Trade (2000) 21 J. Juv. L. 25, 35.)

It is a relatively serious allegation: The child is at least at some risk of turning out to be a psychopath. Ironically enough, one factor that can lead to reactive attachment disorder is "multiple placements." (See Shink, Law Day 2001: Promoting Justice for Children (May 2001) 30 Colo. Law. 27 , 28 [citing the Diagnostic and Statistical Manual of Mental Disorders (4th ed. 2000) p. 128 (DSM-IV-TR) ]; e.g., In re Krystle D. (1994) 30 Cal.App.4th 1778, 1792 , 37 Cal.Rptr.2d 132 ["This disorder, while related to having inadequate care and multiple caregivers prior to placement in foster care, is further exacerbated by the lack of permanent placement in a stable, nurturing and supportive home."].) [7]

As of this writing, as minor's counsel mentioned at oral argument, these two boys are on their fifth placement, and that is only a 30 day emergency foster home. No matter what we do, then, a sixth placement is inevitable. Given the evident lack of stability that these boys have experienced since being removed from their mother by the county, their appellate counsel urges we not affirm, but return the case for "further actions" so that the boys can find "a stable situation." [8]

For its part, the social services agency stresses that the minors' trial counsel filed no appeal, and that nothing as yet regarding actual unadoptability has been found, as a matter of fact, by a trial court.

We should add here that the issue of adoptability was not actually litigated at the trial levelit was a nonissue. Given the circumstances at the time, minors' trial counsel had no reason to raise iteverything was going well with a then-viable adoptive placement. The two boys (and their sister) expected to be there forever. Thus Melissa would have found that raising the issue would be futile: One can readily imagine what she would anticipate the trial judge saying, "What do you mean your two boys are not adoptable? They have already been placed with a couple who want to adopt them."

It is at this point that one of the major conundrums of our dependency system comes into sharp focus. What happens when events happen during the -pendency of an appeal from an order terminating parental rights which cast doubt on the adoptability finding on which the order terminating parental rights was necessarily premised? The question of what an appellate court should do when it receives information from a child's appellate counsel indicating that an order terminating parental rights should be reversed, even though there was, strictly speaking, no "error" by the juvenile court, is now before the California Supreme Court, in the context of the role of minor's counsel. It will be the Supreme Court's first encounter with the problem since Elise K.

Actually, the facts in In re Elise K, supra, 33 Cal.3d 138 , 187 Cal.Rptr. 483 , 654 P.2d 253 are directly on point with this case: There, at the time of the termination of rights (under old Civil Code section 232), the "outlook for adoption was exceedingly favorable." (See id. at p. 148, 187 Cal.Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.).) Then the adoptive placement failed, and the child's age rendered her unlikely to be adopted. ( Ibid. ) She was in the position of "having neither a parent nor a prospect of gaining one through adoption." ( Ibid. ) The Court of Appeal, however, "declined to give weight" to the new postjudgment evidence. ( Id. at p. 139, 187 Cal.Rptr. 483 , 654 P.2d 253 (maj.opn.).) Our Supreme Court took the case and, by contrast, found it "appropriate" to accept a stipulation allowing the reversal of the termination order. ( Ibid. ) The Chief Justice, joined by Justice Reynoso, wrote separately to demonstrate the wisdom of having appellate courts take cognizance of the changed circumstances. ( Id. at pp. 139, 151, 187 Cal.Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C. J.).)

While the entirety of our high court certainly impliedly rejected the "see no error" approach of the Court of Appeal in Elise K., the stipulation obviated the need to explore the issue in depth. That gauntlet was taken up by Chief Justice Bird, but anything she said in her separate concurring opinion could only be, at most, persuasive authority for future cases. (For what it is worth we are persuaded by it.)

There are two possible models that might be adopted by the courts (or by the Legislature if it were to directly confront the issue) concerning developments during the pendency of an appeal that cast doubt on a child's adoptability. The first model says that regardless of what happens to the child, the only thing the appellate court should look at is whether there was any prejudicial error at the .26 hearing.

It should be remembered, at this point, that adoptability under section 366.26 is relatively easy for social service agencies to establish. (The exact language from the statute is whether "it is likely the child will be adopted.") True, the determination must be made on "clear and convincing evidence." (§ 366.26, subd. (c)(1).) However, it is only common sense that when there is a prospective adoptive home in which the child is already living, and the only indications are that, if matters continue, the child will be adopted into that home, adoptability is established. In such a case, the literal language of the statute is satisfied, because "it is likely" that that particular child will be adopted.

Case law has also established that adoptability may be satisfied by a general adoptability, independent of whether there is a prospective adoptive family (as the term of art goes) "waiting in the wings." ( In re Sarah M. (1994) 22 Cal.App.4th 1642, 1649 , 28 Cal.Rptr.2d 82 ; see In re Jeremy S. (2001) 89 Cal.App.4th 514, 525 , 107 Cal.Rptr.2d 280 [evidence that there is a prospective adoptive parent willing to adopt a child gives rise to the inference that other individuals might be willing to adopt the same child].) That is, the child needn't either be already placed with prospective adoptive parents, nor is it necessary even to have identified prospective parents, as long as there is clear and convincing evidence the child will be adopted in a reasonable time. (See In re Jennilee T. (1992) 3 Cal.App.4th 212, 223-225 , 4 Cal.Rptr.2d 101 .)

The upshot, in terms of this model, is that if a child has been found to be adoptable at the .26 hearing, social workers should, in theory, be able to find another prospective adoptive home. One might describe this model as, "Social workers will provide."

There is a certain rough justice to this model. For example, one can point (as courts often do when affirming terminations in juvenile dependency cases) to the deficiencies on the part of the parent that led to the denial of reunification services. A court can usually say in such cases, in substance, that if a parent really, really cared, he or she (usually she) would have gotten her act together from the very inception of the dependency. (Cf. Suzanne J. v. Superior Court (1996) 46 Cal. App.4th 785, 788 , 54 Cal.Rptr.2d 25 ["The dependency scheme is designed to aid those parents who seriously want to maintain a healthy parental relationship with their children."].) [9] As it is there is usually a too-little-too-late flavor to these cases.

The second model stresses the law's abhorrence of legal orphanage, and the high importance that any adoptability finding be correct. (See In re Amelia S. (1991) 229 Cal.App.3d 1060, 1065 , 280 Cal.Rptr. 503 [using language such as "high probability," "no substantial doubt," and "sufficiently strong to command the unhesitating assent of every reasonable mind" to describe nature of the evidence required to sustain an adoptability finding]; see also In re Elise K, supra, 33 Cal.3d at p. 147 , 187 Cal.Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.) [stressing "essential role of adoptability"].) The trial court cannot terminate parental rights unless the child is adoptable, because if the child isn't adoptable, the proceeding will have done nothing to benefit the child other than to permanently separate the child from the child's natural parents, and condemn the child to permanent foster-care. The second model doesn't quite have as much faith in social workers to come up with another adoptive home after one has already failed, and is more willing to trust the trial court to decide the issue of adoptability in light of subsequent facts, rather than trust social workers to necessarily find a new adoptive home regardless of intervening circumstances.

The present case forces a decision upon us. So, until the Supreme Court or the Legislature tells us otherwise, we conclude that, as between the two models, the second is by far the superior.

First, as underscored by recent scandals regarding abuse in foster care placements in Southern California (fortunately, as far as we are aware, none regarding the particular county social service agency before the court here), the Legislature has wisely authorized minors' counsel to monitor possible problems in placement. After all, social workers were the ones who made the actual placement in the first place, and they may have reasons to continue it that may not be completely in sync with the minors' best interests (including psychological vesting, and internal bureaucratic pressures for high adoption statistics). It is the minor's counsel who has the most incentive to bring any problems with a placement to a court's attention.

Thus, section 317, subdivision (e), imposes on the child's counsel the statutory duty to look out for his or her interests, including when those interests are opposed to what social workers want. The statute says that counsel must "investigate the interests of the minor beyond the scope of the juvenile proceeding and report to the court other interests of the minor that may need to be protected by the institution of other administrative or judicial proceedings." (Ibid., emphasis added.) To that end, minors' counsel "shall make ... further investigations ... necessary to ascertain the facts, including the interviewing of witnesses ...."(§ 317, subd. (e), emphasis added.) Moreover, ifas in the present casethe child (that is, the client) is four years old or older, counsel must interview the child "to determine the child's wishes and to assess the minor's well-being, and shall advise the court of the minor's wishes." ( Ibid.; see also Cal. Rules of Court, rule 1438(c)(4) ["The attorney for the child must have sufficient contact with the child to establish and maintain an adequate and professional attorney-client relationship."].)

The text of section 317 makes no differentiation between minors' trial and appellate counsel. Again, the wisdom is selfevident. In many cases, like this one, new evidence will come to light during the pendency of an appeal that casts doubt on an adoptability finding which everyone took for granted at the trial level. It is also telling that this court would have been kept completely ignorant about the possibility of these children becoming legal orphans if an independent counsel had not brought it to our attention.

Unless the Legislature was just wasting its time by specifying that minor's counsel was to report to "the court" the minor's "other interests ... that may need to be protected" by "judicial proceedings," the logical inference is that appellate courts necessarily will have the power to act to protect the child's interests on appeal. (See also Code Civ. Proc., § 187 [if a court has jurisdiction over a matter, it has implied power to use "all the means necessary" to carry that jurisdiction into effect].) [10]

The converse quickly devolves into absurdity. Try writing this sentence without flinching: "Regardless of whether a child will wind up being a legal orphan, the appellate court is powerless to take any action unless there was some error apparent in the trial record." No one on the Supreme Court wanted to write that in Elise K, and we won't endorse it now. In the present case, for example, we are presented with a virtual mockery of the "permanence" that is usually cited to justify termination. If we summarily affirm, these two boys may never have any "permanence" or a natural parent.

Second, and even more importantly, the second model is child-centered rather than parent-centered. It is concerned with the child's best interests, not punishing the parent. The first model takes an isolated view of the various stages of dependency proceedings in a vacuum, as if they had no overall purpose of protecting the child's welfare. But there are cases, quite independent of what the natural parents "deserve," where the best interests of the child are enhanced by the preservation, of parental rights. The whole purpose of section 366.26, with its convoluted structure and requirements of multiple factfinding, is to assure that children do not fall through cracks in the system. If parental rights are terminated, children should go to a better situation, a permanent adoptive home, not a worse one (permanent foster care and no parent). If the children are not adoptable, a series of foster-care placements is the least preferable option, not the de facto second option because social workers want a natural parent completely out of the picture. (See § 366.26, subd. (b).)

Finally, adoptability findings are fraught with a unique danger at the trial level. As the present case and Elise K. illustrate, adoptability can easily go from being an uncontested nonissue at trial because a child is doing well in an adoptive placement to dominating appellate proceedings in the wake of the unanticipated failure of that placement.

Consider this trap that a trial court may easily fall into: If a child is doing well in an adoptive placement at the time of the .26 hearing, it is too easy to avoid confronting the child's general adoptability. (Cf. In re Jeremy S., supra, 89 Cal.App.4th at p. 525 , 107 Cal.Rptr.2d 280 [differentiating between adoptability "based solely " on existence of adoptive home and general adoptability based on child's traits].) The child fits the statute because there is an adoptive home "waiting in the wings," so the court never really confronts the question of what will happen if the specific adoptive placement fails. Indeed, the trial court may not have enough information to determine, in advance, whether that child will be generally adoptable after an adoptive placement has failed. (Much obviously would depend on the particular reasons for the failure.)

The second model is, however, from the point of view of legal logic, inconsistent with the traditional structures of appellate procedure. That structure places a high premium, as it must, on predictability. No legal system is worth anything unless it provides for stable, final judgments, which can only be attacked in limited, definable ways. Build too much play into the system and, legally speaking, the fabric of time and space will collapse.

But there are countervailing concerns. Each dependency case involves an innocent child, whose life is tossed about by a whole industry of well-meaning adults, of whom all levels of court are a part. And courts must not lose sight of the purpose of the dependency system in the first place. "[T]he paramount concern of the appellate court in all dependency proceedings is for the protection and welfare of the child." ( In re Jeremy S., supra, 89 Cal.App.4th at p. 527 , 107 Cal. Rptr.2d 280 .) The children who are taken from their parents and put into the juvenile dependency system deserve nothing less than that the courts' primary concern should be their interests, not the traditional strictures of appellate procedure. The world will not end if appellate courts possess the power to overturn judgments terminating parental rights when there is evidence that the children may not be adoptable and the children's interests will be served by the preservation of parental rights. (Accord, In re Elise K, supra, 33 Cal.3d at pp. 149-151, 187 Cal.Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.) [appellate courts only inflict "gross injustice" when they blindly refuse to take postjudgment events in child's life into account].)

We must next deal with a relatively disturbing point made by county counsel at oral argument, namely, that if the failure of an adoptive placement post-appeal were to be grounds for reversal of a termination of parental rights, there would be many more reversals than there are now, because plenty of adoptive placements do not work out.

There are at least three answers to this point. First, we are only dealing with cases where adoptive placements fail during the pendency of an appeal, i.e., when there is no final judgment terminating parental rights. The fact is that the Legislature has provided for the appeal of termination orders (see § 395), which necessarily contemplates the possibility of a reversal of such orders. Nowhere has the Legislature said that the Court of Appeal (or the Supreme Court for that matter) are potted plants whose job it is to rubberstamp every order terminating parental rights in some kind of pantomime of due process. It was the manifest intent of the Legislature, as exemplified in section 366.26, to assure that children be adopted if parental rights are terminated. The Legislature didn't care a fig for the affirmance rate as such. [11]

Second, it is not necessary to go so far as to say the failure qua failure of an adoptive placement, even during the pendency of an appeal, is automatically grounds for reversal. There must also be some basis to conclude that doubt is cast on the adoptability finding made by the trial judge. Children who are clearly adoptable even though an adoptive placement has not worked out would not be affected by taking cognizance of postappellate developments. In marginal cases, however, considering postjudgment evidence will make a beneficial difference to the child. County counsel forgets that reversal of a .26 order because of doubt of a child's adoptability is to prevent legal orphanage, not reward the parent.

Third, to the degree that failed adoptions and legal orphanages are all too common in our juvenile dependency system, perhaps more cases ought to be reversed if reversal would serve the children's interests.

Finally, county counsel's argument that postjudgment events cannot be taken into account because a child did not file a notice of appeal at the trial level is particularly ill taken. As Justice O'Leary recently wrote for this court, "juvenile cases are hardly ordinary," the proper focus "is always on what is best for the child," and children are not the "typical prevailing party"in "most dependency cases the minor simply gains the right to hope to be adopted by a family who will love, care, and nurture him." ( In re Jeremy S., supra, 89 Cal. App.4th at p. 526 , 107 Cal.Rptr.2d 280 .) Thus in Jeremy S. this court considered the merits of a minor's argument that he was not adoptable, even though it was raised for the first time in a respondent's brief on appeal. ( Id. at pp. 527-529, 107 Cal.Rptr.2d 280 .)

The plain fact is that there will be times, as Elise K. and this case both illustrate, where a child's interest will change after the order terminating parental rights. The child's trial counsel are usually delighted that there is a prospective adoptive home and are not likely to take a position that would derail an adoption (opposing termination of parental rights) just to be "extra sure" that the adoption will come to fruition. Somebody, however, has to be able to protect the child's interest if the adoption doesn't come to fruition, and to forbid the child's own appellate counsel from doing so elevates the technicalities of appellate procedure over the whole reason for the dependency system in the first place. (See In re Jeremy S., supra, 89 Cal.App.4th at p. 527 , 107 Cal.Rptr.2d 280 .)

So back to where we began. In this case, minor's appellate counsel has brought to our attention strong evidence that these two boys, who even social workers are treating as a "sibling group," may be unadoptable. Clearly, if the trial judge knew what we know now, he might have made a different decision. He might have selected among the nontermination options in subdivision (c) of section 366.26, including the "let's see if social workers really can find a permanent home" option of continuing the case for 180 days under subdivision (e)(1)(D).

Given the awful consequences of being wrong, adoptability findings must be relatively certain. The statute requires clear and convincing evidence. Thus, while adoptability may be relatively easy to establish at a .26 hearing when a child is already in an adoptive home, it is still so critical to the process that any genuine doubt should merit a second look. Here, for example, surely any evidence that the adopting parents considered one of the two boys to be suffering from reactive attachment disorder throws the adoptability finding into serious doubt.

We therefore reverse the judgment terminating parental rights and send the case back for an "updated review hearing." (See In re Eileen A. (2000) 84 Cal.App.4th 1248, 1259 , 101 Cal.Rptr.2d 548 [analogous procedure in cases of ineffective assistance of counsel]; In re Arturo A (1992) 8 Cal. App.4th 229, 245 , 10 Cal.Rptr.2d 131 [same].)

We have spelled out in the introduction to this opinion the rules the trial court should apply at that hearing.

WE CONCUR: BEDSWORTH and MOORE, JJ.

[1] All statutory references in this opinion are to the Welfare and Institutions Code unless otherwise specified. All references to .26 hearings will be to the hearing concerning the possible termination of parental rights held pursuant to section 366.26 of that Code.

[2] We hereby grant the motion of minor's counsel to take additional evidence on appeal. (See Code Civ. Proc., § 909; Cal. Rules of Court, rule 23; see also In re Elise K. (1982) 33 Cal.3d 138 , 187 Cal.Rptr. 483 , 654 P.2d 253 .) That declaration, accompanying the motion, is made part of the record on appeal.

Because the most an appellate court should do is to return such a case for further factfinding, such motions are not nearly as disruptive in the juvenile dependency context as they might be in ordinary civil cases, where appellate courts are loath to grant them. (See In re Elise K, supra, 33 Cal.3d at pp. 149-150, 187 Cal.Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.) (pointing out that taking evidence on appeal in juvenile dependency cases does not pose the usual problems as in other cases).)

[3] The word "compliance" in the context of a mentally ill person who needs to take certain medication to alleviate the effects of a mental illness has a rather ominous sound. A more humane phrase would be on the order of simply, "taking her medication." We use "compliance" because it is the shorthand that appears most in the record.

[4] The case might easily have gone the other way too. If the trial judge had granted Melissa's section 388 request, it too would have been upheld on an abuse of discretion standard. There was considerable evidence in the record that Melissa might prove yet to be a very good parent. One of the social workers observed that in the wake of the 12-month review she was "demonstrating `good role modeling,'" she demonstrated excellent progress in her group therapy, and in fact was said, by the group facilitator, to have "excellent parenting skills."

[5] We do not explore in this opinion the complexities inherent in reunification plans that might arise in the event that a parent in Melissa's appellate position were to raise an ineffective assistance of counsel argument in an appeal from an order made at a .26 hearing based on a neglect to do anything to assure that the reunification plan was meaningfully tailored to the parent's needs. The adequacy of reunification services is typically a matter for the 12- and 18-month review hearings when services are about to be terminated.

[6] Because we grant minor's counsel's motion to take additional evidence, we deny as moot county counsel's motion to strike the brief. The county counsel's motion was made before minor's counsel's brief was supported with evidence taken pursuant to section 909 and rule 23 of the California Rules of Court.

[7] We express no opinion as to whether either boy really has reactive attachment disorder that is properly a matter for the trial court on remand. It is enough to say that, for whatever reason, the prospective adoptive parents thought that Joseph has the disorder and returned all three children. (Cf. In re Elise K., supra, 33 Cal.3d at p. 148, fn. 19 , 187 Cal. Rptr. 483 , 654 P.2d 253 (cone. opn. of Bird, C.J.) (noting that the record did not disclose the reason for the postjudgment failure of the adoptive placement).)

It is also worth mentioning that there is some concern among lawyers who practice in the area that children will "act out" for ulterior motiveseither the natural parent put them up to it to sabotage a placement, or because the children genuinely want to be with their natural parents, and the children intuitively grasp that sabotaging a placement will bring them one step further toward reunification. To the degree that this concern might have any application to this case, it is properly a matter for the trial court in a contested proceeding.

[8] It is a position opposite from minor's trial counsel, whose position was all in favor of termination. This about-face should not carry any necessary implications concerning the work of either minor's trial or appellate counsel. At the .26 hearing, for example, the two boys and their sister were seemingly doing well in a prospective adoptive home. Understandably, their trial counsel would thus favor termination, and take no appeal.

[9] Usually that means kicking a drug habit immediately, though (ironically) in the case before us it meant immediately beginning to take drugs, albeit prescription drugs, to alleviate the effects of mental illness. Usually, only by immediately ceasing to take illegal drugs is a parent able to demonstrate sufficient reform to assure the court that it would be safe to return the child to the parent.

[10] It would be anomalous to read "judicial proceedings" to exclude appellate proceedings in the very case that will determine whether the child might end up being an orphan, i.e., the judicial proceedings that are likely to be the most meaningful to the child. Of course, the phrase would certainly include the (far less likely) situations where the child's own legal interests, independent of dependency, were affected, e.g., as plaintiff or defendant in a personal injury action, or as claimant in a probate matter.

[11] Allowing postjudgment developments to influence the outcome of a dependency appeal has the effect, a little bit like Heisenberg's uncertainty principle (the act of shining a light on an electron makes its location impossible to determine), of making the very existence of an appeal play a substantive part of the process. To be plain, a parent might file a frivolous appeal in the hope that a placement might not work out during the pendency of the appeal, and the parent would "win" back the child merely because he or she had slowed down the judicial machinery. The answer to the point is that the courts should deal with frivolous or bad faith tactics in their own context, and not allow them to affect what would be the best outcome for the child.