The majority determines that the twins who developed from the ova K.M. donated to E.G. have two mothers rather than one. While I disagree, as I shall explain, with that ultimate conclusion, I agree with the majority’s premise that a child can have two mothers. Our previous holding that “for any child California law recognizes only one natural mother” (Johnson v. Calvert (1993) 5 Cal.4th 84, 92 [19 Cal.Rptr.2d 494, 851 P.2d 776] (Johnson)) must be understood in the context in which it arose—a married couple who intended to become parents and provided their fertilized ova to a gestational surrogate who did not intend to become a parent—and, thus understood, may properly be limited to cases in which to recognize a second mother would inject an unwanted third parent into an *149existing family. When, in contrast to Johnson, no natural1 or adoptive father exists, two women who intend to become mothers of the same child may do so either through adoption (Sharon S. v. Superior Court (2003) 31 Cal.4th 417 [2 Cal.Rptr.3d 699, 73 P.3d 554]) or because both qualify as natural mothers under the Uniform Parentage Act (Earn. Code, § 7600 et seq.) (UPA), one having donated the ovum and the other having given birth (see Johnson, at p. 92).
While scientific advances in reproductive technology now afford individuals previously unimagined opportunities to become parents, the same advances have also created novel, sometimes heartbreaking issues concerning the identification of the resulting children’s legal parents. Declarations of parentage in this context implicate complex and delicate biological, personal, legal and social policy considerations. For these reasons, courts have sought above all to avoid foreseeable disputes over parentage with rules that provide predictability by permitting the various persons who must cooperate to bring children into the world through assisted reproduction to determine in advance who will and will not be parents, based on their expressed and voluntarily chosen intentions. (See, e.g., Johnson, supra, 5 Cal.4th 84, 93-95.)
Precisely because predictability in this area is so important, I cannot agree with the majority that the children in this case do in fact have two mothers. Until today, when one woman has provided the ova and another has given birth, the established rule for determining disputed claims to motherhood was clear: we looked to the intent of the parties. “[I]n a true ‘egg donation’ situation, where a woman gestates and gives birth to a child formed from the egg of another woman with the intent to raise the child as her own, the birth mother is the natural mother under California law.” (Johnson, supra, 5 Cal.4th 84, 93, fn. 10.) Contrary to the majority’s apparent assumption, to limit Johnson’s holding that a child can have only one mother to cases involving existing two-parent families does not require us to abandon Johnson’s intent test as the method for determining disputed claims of motherhood arising from the use of reproductive technology. Indeed, we have no other test sufficient to the task.
Furthermore, to apply Johnson’s intent test to the facts of this case necessarily leads to the conclusion that E.G. is a mother and K.M. is not. That E.G. intended to become the mother—and the only mother—of the children to whom she gave birth is unquestioned. Whether K.M. for her part also intended to become the children’s mother was disputed, but the trial court found on the basis of conflicting evidence that she did not. We must defer to the trial court’s findings on this point because substantial evidence *150supports them. K.M. represented in connection with the ovum donation process, both orally and in writing, that she did not intend to become the children’s mother, and consistently with those representations subsequently held the children out to the world as E.G.’s but not her own. Thus constrained by the facts, the majority can justify its conclusion that K.M. is also the children’s mother only by changing the law. This the majority does by displacing Johnson’s intent test—at least for the purposes of this case—with the following new rule: a woman who has “supplied her ova to impregnate her lesbian partner in order to produce children who would be raised in their joint home” (maj. opn., ante, at p. 138; see also id., at pp. 134, 139, 142, 143, 144) is a mother of the resulting children regardless of any preconception manifestations of intent to the contrary.
I find the majority’s reasons for not applying the Johnson intent test unpersuasive. The majority criticizes the test as basing “the determination of parentage upon a later judicial determination of intent made years after the birth of the child.” (Maj. opn., ante, at p. 144.) But the task of determining the intent of persons who have undertaken assisted reproduction is not fundamentally different than the task of determining intent in the context of disputes involving contract, tort or criminal law, something courts have done satisfactorily for centuries. The expectation that courts will in most cases accurately decide factual issues such as intent is one of the fundamental premises of our judicial system. Indeed, the majority itself expresses willingness to continue applying the Johnson intent test to determine whether gestational surrogacy agreements are enforceable. This position leaves no plausible basis for refusing to apply the same test to determine whether ovum donation agreements are enforceable. Ovum donation and gestational surrogacy agreements are two sides of the same coin; each involves an ovum provider, a gestator, and an agreement about who will become the parent or parents of any resulting offspring. Indeed, when two women divide in this way the genetic and gestational components of motherhood, only an examination of their intent permits us to determine whether we are dealing with an ovum donation agreement, a gestational surrogacy agreement, or neither. If courts can perform one of these tasks acceptably, they can also perform the other.2
No more persuasive is the majority’s suggestion that to respect the formally expressed intent of the parties to an ovum donation agreement is prohibited by the rule that parental obligations may not be waived by contract. (Maj. opn., ante, at p. 144.) We expressly rejected a similar *151argument directed against a gestational surrogacy agreement in Johnson, supra, 5 Cal.4th 84, 95-97. Certainly parental obligations may not be waived by contract. (Earn. Code, § 7632.) But Johnson's intent test does not enforce ovum donation and gestational surrogacy agreements; it merely directs courts to consider such documents, along with all other relevant evidence, in determining preconception intent.
As a final reason for rejecting the intent test, the majority suggests that to apply the test outside the context of Johnson might shield from the obligations of fatherhood, contrary to existing law, a man who, lacking the intent to become a father, “engages in sexual intercourse with a woman who assures him, falsely, that she is incapable of conceiving children ...” (Maj. opn., ante, at p. 143.) But no one, to my knowledge, proposes to apply the intent test to determine the parentage of children conceived through ordinary sexual reproduction. This court adopted the intent test to resolve cases of assisted reproduction in which disputes over motherhood arise because one woman has provided the ova and another has gestated them. Both Johnson, supra, 5 Cal.4th 84, and the case before us belong to that category. Although the majority may be correct in asserting that “[u]sually, whether there is evidence of a parent and child relationship under the UPA does not depend upon the intent of the parent” (maj. opn., ante, at p. 143), we adopted the intent test precisely because the UPA does not expressly resolve conflicting claims to motherhood arising from ovum transplants. The majority’s speculation about men who engage in sexual activity despite mental reservations about fatherhood is irrelevant.
The new rule the majority substitutes for the intent test entails serious problems. First, the rule inappropriately confers rights and imposes disabilities on persons because of their sexual orientation. In a standard ovum donation agreement, such as the agreement between K.M. and E.G., the donor confirms her intention to assist another woman to become a parent without the donor becoming a parent herself. The majority’s rule vitiates such agreements when its conditions are satisfied—conditions that include the fact the parties to the agreement are lesbian. Although the majority denies that its rule depends on sexual orientation (maj. opn., ante, at p. 138, fn. 3), the opinion speaks for itself. The majority has chosen to use the term “lesbian” no less than six times in articulating its holding. (Id., at pp. 134, 138, fn. 3, 139, 143, 144.) Moreover, the majority prevents future courts from applying its holding automatically to persons other than lesbians by stating that it “decide[s] only the case before us, which involves a lesbian couple who registered as domestic partners.” (Id., at p. 138, fn. 3.) I see no rational basis—and the majority articulates none—for permitting the enforceability of an ovum donation agreement to depend on the sexual orientation of the parties. Indeed, lacking a rational basis, the rule may well violate equal protection. (See Romer v. Evans (1996) 517 U.S. 620, 631-636 [134 L.Ed.2d *152855, 116 S.Ct. 1620]; Gay Law Students Assn. v. Pacific Tel. & Tel. Co. (1979) 24 Cal.3d 458, 467-475 [156 Cal.Rptr. 14, 595 P.2d 592].) Why should a lesbian not have the same right as other women to donate ova without becoming a mother, or to accept a donation of ova without accepting the donor as a coparent, even if the donor and recipient live together and both plan to help raise the child?3
Having created a new rule to decide certain cases of disputed parentage arising from assisted reproduction, the majority seeks through various means to limit the number of cases to which the new rule will apply. Through this effort, the majority creates more problems than it solves.
Although the majority at one point seems to suggest that its holding applies only to persons who have registered as domestic partners (maj. opn., ante, p. 138, fn. 3), it elsewhere articulates its holding without reference to domestic partnership. {Id., at pp. 134, 138, 139, 143, 144.) The resulting ambiguity flags á serious problem. K.M. and E.G. registered as domestic partners in October 1994 and terminated their domestic partnership in March 2001. Not until January 1, 2003, however, did California law give domestic partners rights and responsibilities with respect to each others’ children. (Earn. Code, § 297.5, subd. (d), added by Stats. 2003, ch. 421, eflf. Jan. 1, 2005.) The new law expressly does not apply to persons like K.M. and E.G., who terminated their domestic partnerships before January 1, 2005. {Id., § 299.3, subd. (a).) For the majority to base its holding, even in part, on K.M. and E.G.’s domestic partnership is to hold, contrary to statute and apparent legislative intent, that domestic partnership under prior laws did affect parental rights and obligations.
Other problems arise from the majority’s attempt to limit its holding to cases in which the ovum donor and birth mother intend to raise the children together. Except in the context of the majority’s new rule, a person’s pre-conception intent to participate in raising a child has no relevance to the determination of natural parentage. The duty to raise children (by personal care or through payment of child support) is imposed by law regardless of the parents’ intent or wishes. Many persons who become parents do not intend to raise children (e.g., casual inseminators and parents who abandon their babies) and, conversely, many people intend to raise children without becoming parents (e.g., nannies and some stepparents and grandparents). To make the determination of natural parentage rest in part on the intent to raise a child injects into that determination a best interests factor—something we *153have previously refused to do. (Johnson, supra, 5 Cal.4th 84, 93, fn. 10.) I realize the court in Johnson wrote that “she who intended to procreate the child—that is, she who intended to bring about the birth of a child that she intended to raise as her own—is the natural mother under California law.” {Id., at p. 93.) But the phrase “raise as her own” {ibid., italics added) in that context did not refer simply to providing childcare; instead, it meant that the woman in question intended to be a parent—to raise a child of her own. In no sense did the Johnson court base its decision of parentage on the question of who would provide childcare. By analogy, Family Code section 7611, subdivision (d), which creates the presumption that a man who “receives [a] child into his home and openly holds out the child as his natural child” is not satisfied simply because a man receives the child into his home; to become a presumed father he must also hold out the child as his natural child.
Perhaps the most serious problem with the majority’s new rule is that it threatens to destabilize ovum donation and gestational surrogacy agreements. One important function of Johnson’s intent test was to permit persons who made use of reproductive technology to create, before conception, settled and enforceable expectations about who would and would not become parents. Johnson, supra, 5 Cal.4th 84, thus gave E.G. a right at the time she conceived to expect that she alone would be the parent of her children—a right the majority now retrospectively abrogates. E.G.’s expectation has a constitutional dimension. (See Troxel v. Granville (2000) 530 U.S. 57, 65 [147 L.Ed.2d 49, 120 S.Ct. 2054] [due process clause protects a parent’s fundamental right to make decisions concerning the care, custody and control of her children].) We cannot recognize K.M. as a parent without diminishing E.G.’s existing parental rights. In light of the majority’s abrogation of Johnson and apparent willingness to ignore preconception manifestations of intent, at least in some cases, women who wish to donate ova without becoming mothers, serve as gestational surrogates without becoming mothers, or accept ovum donations without also accepting the donor as a coparent would be well advised to proceed with the most extreme caution. While the majority purports to limit its holding to cohabiting lesbians, and possibly only to those cohabiting lesbians who are also domestic partners, these limitations, as I have explained, rest on questionable legal grounds and may well not stand the test of time.
I find the majority’s extensive discussion of Family Code section 7613, subdivision (b), irrelevant and illogical. The statute provides that “[t]he donor of semen provided to a licensed physician and surgeon for use in artificial insemination of a woman other than the donor’s wife is treated in law as if he were not the natural father of a child thereby conceived.” The majority concludes the statute does not apply to the case before us. I agree. Although provisions of the UPA that determine the father and child relationship apply “[i]nsofar as practicable” (Fam. Code, § 7650) to determine the mother and *154child relationship, the act’s drafters did not contemplate that all provisions concerning fatherhood would be construed as affecting motherhood. Indeed, the drafters considered it “obvious that certain provisions [of the act] would not apply in an action to establish the mother and child relationship” and expressly left to courts the decision of which provisions should apply. (9B West’s U. Laws Ann. (2001) U. Parentage Act (1973) com. to § 21, p. 494.) That the statute governing sperm donations (Fam. Code, § 7613, subd. (b)) was not intended to govern ovum donations is easy to conclude since the act was drafted in 1973, long before assisted reproduction and gestational surrogacy became commonplace. (9B West’s U. Laws Ann., supra, U. Parentage Act (2000) Prefatory Note, p. 297.) The drafters’ response to these scientific developments has not been to endorse a counter-textual reading of the provision governing sperm donation, but instead to withdraw the 1973 act entirely and replace it with a new act expressly addressing some of the issues that have arisen from the use of reproductive technology. (9B West’s U. Laws Ann., supra, U. Parentage Act (2000) p. 303 et seq.) In short, Family Code section 7613, subdivision (b), has nothing to do with this case.
The majority seems to believe that, having concluded the sperm donation statute (Fam. Code, § 7613, subd. (b)) does not apply, one must necessarily conclude that K.M. is the mother of the children who developed from the ova she donated to E.G. This reasoning entails a non séquito. The statute, when it applies, merely excludes someone as a possible parent; it does not establish parentage. In order to reach the further conclusion that K.M. is a parent, the majority must entertain a string of questionable assumptions: first, that we would refuse to apply the sperm donation statute (Fam. Code, § 7613, subd. (b), quoted ante, at p. 139), despite its plain language, to cut off the parental rights and responsibilities of a man who donates his sperm through a physician to a woman who is not his wife but with whom he lives (maj. opn., ante, at pp. 140-142), and, second, that two women who live together and divide between themselves the genetic and gestational aspects of pregnancy must be treated in exactly the same way as the man and woman just posited (id., at p. 141). The latter assumption, in turn, embodies additional, unstated assumptions about the effect of the equal protection clause. But ovum donation, which requires substantial medical and scientific assistance, is not sufficiently like sperm donation, which can easily be accomplished by unassisted laypersons, to require equal treatment under the law for all purposes. Accordingly, to recognize the sperm donation statute’s inapplicability does not dispose of this case; it merely leaves us with the same question with which we began, namely, whether K.M. is a second mother of E.G.’s children. Until today, the Johnson intent test would have required us to answer the question in the negative. In my view, it still should.
*155Perhaps the best way to understand today’s decision is that we appear to be moving in cases of assisted reproduction from a categorical determination of parentage based on formal, preconception manifestations of intent to a case-by-case approach implicitly motivated at least in part by our intuitions about the children’s best interests. We expressly eschewed a best interests approach in Johnson, supra, 5 Cal.4th 84, explaining that it “raises the repugnant specter of governmental interference in matters implicating our most fundamental notions of privacy, and confuses concepts of parentage and custody.” (Id., at p. 93, fn. 10.) This case, in which the majority compels E.G. to accept K.M. as an unintended parent to E.G.’s children, in part because of E.G.’s and K.M.’s sexual orientation and the character of their private relationship, shows that Johnson’s warning was prescient. Only legislation defining parentage in the context of assisted reproduction is likely to restore predictability and prevent further lapses into the disorder of ad hoc adjudication.
Respondent’s petition for a rehearing was denied October 19, 2005. Kennard, J., did not participate therein. Werdegar, J., was of the opinion that the petition should be granted.
As when an unmarried woman becomes pregnant through physician-assisted artificial insemination pursuant to Family Code section 7613, subdivision (b).
Having alternately embraced and criticized intent tests, the majority ironically concludes by making intent part of its own new test for parentage. That new test appears to require courts to determine whether the donor and recipient intended any offspring to be raised in their joint home. (See maj. opn., ante, at pp. 134, 137-138, 139, 143, 144.)
That the majority’s rule expressly applies only to lesbians creates the additional problem of requiring a formal, legal definition of “lesbian.” Unless we are willing to adopt such a definition and to authorize courts to inquire into the private facts necessary to apply it, the rule is likely unworkable.