This appeal presents questions concerning the means by which a minor claimant to survivor social security benefits may attempt to establish that though he was bom in wedlock he is in fact the illegitimate child of another than this mother’s husband. Wayne M. McMillian, the claimant here, appeals the district court's grant of summary judgment in favor of the Secretary who had denied the claim on the basis that claimant had not established that for purposes of § 202(d) of the Social Security Act (the Act), 42 U.S.C. § 402(d), he was the illegitimate son of the deceased wage earner who he asserted to be his father. We affirm, though on different grounds than those directly relied upon by the Secretary and the district court.
I
Wayne M. McMillian (claimant or Wayne) was bom on March 11, 1976, to Annie M. McMillian (Annie). At that time and continuously since 1966, Annie was married to and living with James McMillian (James). Beginning in 1973, however, Annie had had an extramarital affair with Willie M. Beatty, Jr., that continued until Beatty’s' death on February 11, 1981. On March 10, 1981, Annie applied for survivor’s insurance benefits for Wayne, alleging that Beatty, indisputably an insured wage earner under the Act, was Wayne’s natural father. In the course of the administrative proceedings that then ensued, evidence both supporting and opposing the critical allegation of Beatty’s biological parentage was adduced.
Annie testified to the existence of the affair and asserted, though, ambiguously, that for two or three months preceding Wayne’s conception she had been separated from her husband and during that time had had sexual relations only with Beatty. There was a substantial amount of evidence that Beatty believed that Wayne was his son. He wrote a number of occasion cards and letters to Annie both before and after Wayne’s birth that strongly, if circumstantially, confirmed their ongoing affair and sexual relationship. Some, following Wayne's birth, contained inferable allusions to Wayne as their son, though none identified Wayne by name. Numerous affidavits and live testimony by acquaintances of Beatty’s asserted that Beatty believed and held out Wayne to be his son. An acquaintance of Annie’s corroborated Annie’s version of her separation from her husband and her sexual relationship solely with Beatty at the time of Wayne’s conception. A group insurance application by Beatty on file with the International Longshoreman’s Association identified “Wayne Matthew McMillan” [sic] as an unmarried dependent child of Beatty’s along with *1149Beatty’s one child, Crystal Beatty, by his marriage. There was considerable evidence that up until his death Beatty provided financial support specifically for Wayne. Annie testified that Beatty often gave her as much as $100 weekly for Wayne’s support. Beatty paid dental bills for Wayne and provided weekly day care support and school funds for him. Annie testified that Beatty purchased a mobile home for her so that she and Wayne might have a place to live if James should throw them out. Finally, there was testimony that discord arose between Annie and James because of the latter’s suspicion or knowledge that Wayne was not his son.
From the other side there was no evidence directly refuting Annie’s essential version of her affair with Beatty nor of James’s non-access to her at the time of Wayne’s conception. Nor was there any evidence directly refuting that of Beatty’s subjective belief and holding out that Wayne was his son, nor his provision of significant financial support for Wayne, aside from some dispute as to the reality of the mobile home purchase. There was, however, record evidence suggesting, though circumstantially, that James was Wayne’s father. Wayne’s registered birth certificate identified James as Wayne’s “father” in the block provided for that information on the official form. More critically, a 1978 state court order in a child support proceeding brought by Annie against James identified Wayne as a child born of their marriage for whom support was sought and awarded.
Based essentially upon the evidence of Beatty’s various acknowledgments of parentage and his provision of financial support for Wayne, the claim was first administratively allowed. When Beatty’s wife, however, protested the award (whose effect- was to reduce the survivor benefits payable to her and her daughter by Beatty) the administration reconsidered the claim. Based essentially upon the birth certificate and state court support order which recited James’s parentage, the administration reversed its earlier determination and directed that Annie refund the sum already paid under the award.
Now represented by counsel, Annie requested and was given a hearing before an AU in which Beatty’s widow and their daughter were joined as interested parties. Following receipt of the evidence generally summarized above, the AU denied Wayne’s claim on the basis that claimant had not established that he was Beatty’s “child” for purposes of the Act. The Appeals Council declined to review this hear-' ing decision which thereupon became the final decision of the Secretary. Upon judicial review the district court accepted its magistrate’s recommendation and granted summary judgment upholding the Secretary’s denial of Wayne’s claim and its direction to Annie to make repayment.
This appeal followed.
II
The issue, broadly stated, is whether the Secretary’s decision denying survivor benefits on the basis that claimant had not established that he was “dependent child” of the deceased wage earner, Beatty, is supported by substantial evidence and is not erroneous in law. The answer turns on the proper interpretation and application of important provisions of the Social Security Act that present interpretive problems. We therefore begin by reviewing in summary fashion those provisions most relevant to the issue and by indicating wherein their application to the instant claim bears upon the issue before us.
The underlying entitlement provision, 42 U.S.C. § 402(d) provides survivor benefits to every “child” who survives “a fully or currently insured individual” if the “child” has applied for benefits, is unmarried and within certain age limits, and was dependent on the insured individual at the time of the latter’s death. There is no dispute here about any of these predicates for entitlement except the “child” and “dependency” status of claimant in relation to Beatty.
In definitional provisions of the Act, “child” is defined to include both legitimate and, under special conditions of proof, ille-. *1150gitimate children of insured wage-earners. See 42 U.S.C. §§ 402(d)(1), 402(d)(3), 416(e), and 416(h). Any legitimate “child” is “deemed dependent” per § 402(d)(3). An illegitimate claimant may establish that he is a “child” for eligibility purposes under either of three critical provisions of the Act: §§ 416(h)(2)(A), 416(h)(2)(B), or 416(h)(3)(C). In rough summary these provide, respectively, for establishing “child” status by (1) showing that under the law of the deceased wage earner’s domicile at death the claimant would have been treated as a child of the decedent for purposes of the devolution of intestate property, § 416(h)(2)(A); or by (2) showing that claimant’s legitimacy was prevented only by some “legal impediment” to an otherwise valid marriage between his parents, § 416(h)(2)(B); or by (3) independently establishing one of a number of specific factual predicates from which “child” status is “deemed” to exist, § 416(h)(3)(C). Any illegitimate claimant “deemed” to be a “child” under either § 416(h)(2)(B) (legitimate but for legal impediment) or § 416(h)(3)(C) (independent proof of child status) is, in turn, “deemed ... legitimate,” hence “dependent,” per § 402(d)(3).
From the time that the instant claim first took specific legal form in the hearing before the AU, it has been pressed by claimant only under § 416(h)(3)(C), that is, by attempting to establish “deemed child” status, hence “deemed dependent” status, by one of the specific means of proof provided in that critical subparagraph of § 416. Because the proper interpretation and application of this statutory provision has been critically disputed throughout these proceedings, the relevant text is important to our consideration. That text runs as follows:
(3) An applicant who is the son ... of a[n] ... insured individual, but who is not (and is not deemed to be) the child of such insured individual under [the preceding “state property law” or “legal impediment” provisions], shall nevertheless be deemed to be the child of such insured individual if:
(C) in the case of a deceased individual—
(i) such insured individual—
(I) had acknowledged in writing that the applicant is his or her son or daughter,
(II) had been decreed by a court to be the mother or father of the applicant, or
(III) had been ordered by a court to contribute to the support of the applicant because the applicant was his or her son or daughter,
and such acknowledgment, court decree, or court order was made before the death of such insured individual, or
(ii) such insured individual is shown by evidence satisfactory to the Secretary to have been the mother or father of the applicant, and such insured individual was living with or contributing to the support of the applicant at the time such insured individual died.
On the facts adduced in evidence, there has been no possibility (and no assertion) here that entitlement could have been established under either (i)(II) (court decree of parentage) or (i)(III) (court ordered support). The Secretary, upheld by the district court, so noted, and that is not in dispute at this point. The dispute has been and remains only whether, on a proper interpretation and application of this statutory provision, entitlement might have been, or was, established under either (i)(I) or (ii).
Ill
In denying the claim as pressed under § 416(h)(3)(C), the AU, upheld in turn by the Appeals Council and the district court, adopted and applied two legal positions that are the focus of claimant’s challenge on appeal.
First, adopting the Secretary’s construction, the AU found in the introductory reference in paragraph (3) to “an applicant who is the son ... of a[n] ... insured,” a *1151predicate requirement, applicable to all claims pressed under any provision of that paragraph, that the claimant prove the biological parentage of the insured independently of and in addition to the specific elements of proof spelled out in subparagraphs (C)(i) and (C)(ii). Applying this construction, the AU determined that Beatty’s parentage had not been independently proven, and on that basis alone denied the claim.
Second, the ALJ’s determination that Wayne had not proven Beatty’s biological parentage was actually based upon a conclusion that the evidence proved to the contrary that James McMillian was that parent. This conclusion in turn was based in part upon the recitations (not direct adjudications) of James McMillian’s parentage in the birth certificate and court support orders. But it was based as well upon the North Carolina common law presumption that a child born in wedlock is the child of his mother’s husband (is legitimate), a presumption rebuttable only by proof of non-access or impotence. See, e.g., State v. White, 300 N.C. 494, 268 S.E.2d 481, 490 (1980) (criminal non-support case); Eubanks v. Eubanks, 273 N.C. 189, 159 S.E.2d 562, 568 (1968) (civil non-support).
Claimant attacks the legal premises underlying both of these positions. He contends first that the different means of establishing “deemed” dependent child status provided in subparagraphs (C)(i) and (C)(ii) are alternative, free-standing means of proving that dispositive status, and that their invocation does not therefore require any additional independent proof of biological parentage.
Beyond that, he contends that the state common law presumption of parentage could not properly be applied in assessing claims of dependent child status under any of the provisions of § 416(h)(3)(C).1 On this basis, he contends that his essentially unrebutted evidence of Beatty’s acknowledgment of parentage established entitlement under subparagraph (i)(I), or that, alternatively, his independent evidence of Beatty’s parentage coupled with his evidence of significant support by Beatty established entitlement under subparagraph (ii), if assessed free of the state law presumption.
IV
We admit, but need not in this case authoritatively accept, the force of claimant’s contention that in logic § 416(h)(3)(C) should not be construed to impose on claimants invoking either subparagraph (C)(i) or (G)(ii) any further independent requirement of proving biological parentage; i.e., that the requirements for establishing deemed dependent child status are fully stated within the two subparagraphs as freestanding, independent alternatives.2 And *1152we agree with claimant’s contention (though for different reasons) that a state presumption of in-wedlock parentage is not applicable of its own force (or by congressional incorporation) to claims made under § (h)(3)(C).
Nevertheless, for reasons that follow, we conclude that any error committed by the Secretary and the court below in respect of these two matters was harmless, and that the judgment upholding the Secretary’s denial of claims may therefore be affirmed.
The basic difficulty in applying the relevant statutory provisions to the survivor benefit claim in this case traces to the relatively rare circumstance that the claim here is that of an illegitimate child born in wedlock. The relevant statutory provisions for claims by illegitimates were obviously drawn with the more common circumstance of out-of-wedlock births primarily in mind. But the statutes make no distinction, as surely they should not, between illegitimates born in and those born out of wedlock. There is thus no question but that a claimant born in wedlock may in appropriate circumstances establish his entitlement to survivor benefits as the illegitimate child of another than his mother’s husband under one or the other of the provisions of § 416(h). See, e.g., Torres v. Gardner, 270 F.Supp. 1 (D.P.R.1967) (entitlement established under § (h)(2)(A) by proving inheritance rights under state intestate succession law). No one really challenges that general proposition here, but the special circumstance has undoubtedly led to considerable confusion and difficulty in dealing with this claim.
Until 1965, § 416(h)(2) provided the sole means by which illegitimates could establish entitlement to benefits as dependent children, with § (h)(2)(A) the primary vehicle. Under that provision, an illegitimate claimant could establish entitlement to benefits by proving his entitlement to inherit from the insured wage earner as a “child” under the intestate succession law of the state of the insured’s domicile. In 1965, § (h)(3)(C) was added specifically to provide other means by which entitlement might be established. See Mendoza v. Secretary of Health & Human Services, 655 F.2d 10, 13 (1st Cir.1981); Severance v. Weinberger, 362 F.Supp. 1348, 1350-51 (D.D.C.1973) (3-judge court).
Under the state inheritance law provisions of § 416(h)(2)(A), a claimant born in wedlock who sought to establish entitlement to benefits as the illegitimate child of another than his mother’s husband was and is perforce confronted with any presumption of in-wedlock parentage integral to the state’s inheritance law. See Torres, supra. That element of state law is necessarily included in the general congressional incorporation of state inheritance law into § 416(h)(2)(A).
But when § 416(h)(3)(C) was added in 1965 to provide supplementary and alternative means of proving entitlement by illegitimates, it did not expressly or by implication incorporate any element of state law respecting parentage, inheritance rights or any other matter of possible relevance to dependent child status. So far as the text of § 416(h)(3)(C) is concerned, all questions of its interpretation and application are referable solely to federal law.
Accordingly, in this case, the North Carolina common law presumption of in-wedlock parentage had no application of its own force or by congressional incorporation to Wayne’s claim under § 416(h)(3)(C)(i)(I) and (ii). The Secretary and the district court therefore erred in applying that presumption as a matter of state law in denying the claim.3
But, as indicated, we think this error was harmless under a proper construction and *1153application of those provisions of § 416(h)(3)(C) under which the claim here in issue was made. This flows essentially from our perception — which we now hold— that the relevant state law presumption of legitimacy of children born in wedlock should apply by judicial adoption as an essential element of federal common law respecting claims by persons born in wedlock seeking to establish entitlement under either § 416(h)(3)(C)(i)(I) (parental acknowledgment) or § 416(h)(3)(C)(ii) (independent proof of parentage and support) as children of insured wage earners other than their mothers’ husbands.4 Our reasons are as follows.
Whatever the intended effect of the concededly ambiguous introductory reference in paragraph (3) of § 416(h) to “an applicant who is the son,”5 it is clear that Congress intended that a claimant invoking either of the alternative means of establishing entitlement provided by § 416(h)(3)(C) should be put to proving as one ultimate fact that the insured was his natural parent. The only question created by the statutory ambiguity is whether the elements of proof laid out in subparagraphs (i) and (ii) were intended wholly to suffice as the predicates from which “child” status was to be “deemed” for survivor benefit purposes. For our purposes, as earlier noted, we need not attempt to resolve that thorny issue, it being sufficient to note that, by whatever exact means, parentage of the insured is an ultimate fact to be proven under either subparagraph (C)(i) or (C)(ii).
That being so, the question inevitably arises whether in applying this federal statutory requirement to the claim of one born in wedlock, the common law presumption of in-wedlock parentage should operate as an interstitial element of federal law. As indicated, it could only do so by judicial adoption; the statute neither provides such a presumption as a matter of express federal law, nor does it incorporate such a presumption by express reference to state law sources as does § 416(h)(2)(A).
As further indicated, we are persuaded that such a presumption should be so adopted. The presumption is one of the most venerable, persistent, and continuously pervasive in the common law. See generally, Lawson, The Law of Presumptive Evidence, pp. 140-152 (2d ed. 1899). In variant forms, now frequently codified, it is currently applied in most, if not all, states. See Krause, [] — A Proposed Uniform Act on Legitimacy, 44 Tex.L.Rev. 829, 844-45 & nn. 29-35 (1966). The reason is plain. Though evidentiary in form, it is substantive in its intended effect as a conservator of generally recognized fundamental social values related to the institutions of marriage and family and to the stability and predictability of property interests. Id. at 844. It would be anomalous if federal statutory law concerned with or dependent in any way upon determinations of the parentage of children born in wedlock did not include as an interstitial common law element a presumption so universally applied by the states, the fundamental repositories of “domestic relations” law. It is therefore an altogether fit subject for judicial adoption as an element of federal common law. See Board of County Commissioners v. United States, 308 U.S. 343, 349-52, 60 S.Ct. 285, 287-89, 84 L.Ed. 313 (1939) (state law respecting remedy for illegal exaction of taxes “absorbed” as “governing federal rule”).
There then arises the question of the proper source and form of the presumption to be adopted. Because there is in general “no federal law of domestic relations,” DeSylva v. Ballentine, 351 U.S. 570, 580, 76 S.Ct. 974, 979-80, 100 L.Ed. 1415 (1956), we can only look to general common law sources or to the law of particular states for the most appropriate source. As between these two, we are satisfied that the *1154more appropriate source is the same source directly incorporated by Congress in § (h)(2)(A) — the law of the insured’s domicile at the date of his death.
The general sources — the original common law presumption and its many variant forms as now applied — do not readily yield a particular form so obviously adapted to the underlying purposes of the federal statutory scheme that it should be adopted for general application. In its common law evolution it took many variant forms, See Lawson, supra, and, as indicated, it exists today in many variant state law forms. For that reason and others more critical we think the proper form for adoption into federal law is the law of the relevant state as defined in § (h)(2)(A). By this means, consistency in particular applications of the presumption under § (h)(2)(A) and § (h)(3)(C) is insured, as well as consistency with general state law applications of the presumption in determining questions of parentage. See DeSylva v. Ballentine, 351 U.S. 570, 76 S.Ct. 974, 100 L.Ed. 1415 (1956) (whether reference to “children” in copyright statute includes illegitimates answered by reference to state law rule).
V
Reviewing the judgment on the basis that the state law presumption of parentage might properly have been applied as an element of federal law,6 we cannot declare the finding that Beatty was not Wayne’s father clearly erroneous. Aside from the force of the presumption, there was other evidence supporting the finding — most notably the entry on the birth certificate identifying James McMillian as the father and the specific claim in the non-support proceeding that claimant was a child of McMillian’s. The information upon which these assertions were made was traceable, at least by permissible inference, to Wayne’s mother. As such, it of course conflicted with her litigation testimony identifying Beatty as claimant’s father. The resulting credibility issue was for the trier of fact. Its resolution against the mother’s litigation testimony is not made manifestly implausible by any other evidence in the case.
The evidence of Beatty’s various acknowledgments of his parentage is of course direct evidence only of his subjective belief. Though obviously of great relevance to the issue, indeed perhaps sometimes sufficient standing alone to support an inferred finding of parentage, see 42 U.S.C. § 416(h)(3)(C)(i)(I), it does not compel such a finding here in view of the opposing force of the presumption of legitimacy and the inherently suspect quality of all acknowledgments of parentage made in circumstances of conceded promiscuity such as those reflected in the evidence here. See Montgomery v. Schweiker, 523 F.Supp. 1128, 1131-32 (D.Md.1981) (parentage by virtue of acknowledgment under § (h)(3)(C)(i)(I) should be found only on basis of clear and unequivocal evidence).
*1155While by the very terms of the presumption the mother’s testimony as to her husband’s non-access was entitled to consideration, it clearly did not suffice to compel a determination that the presumption’s force had been negated. The credibility of this testimony like the parallel testimony respecting Beatty’s parentage was drawn seriously in issue by other evidence and was properly subject to rejection by the trier of fact. Fed.R.Civ.P. 52(a).
Because the dispositive factual finding that the insured wage earner, Beatty, was not claimant’s natural parent was not clearly erroneous we affirm the district court’s judgment.
AFFIRMED.
. The actual basis for claimant’s contention as to the non-applicability of the state presumption was, confusedly, that the state’s inheritance law, within which the presumption was applied, was unconstitutional in its denial of equal protection to illegitimate children — a challenge comparable to that considered in such cases as Latli v. Lalli, 439 U.S. 259, 99 S.Ct. 518, 58 L.Ed.2d 503 (1978).
Though the state, by intervention, formally defended its inheritance law against this challenge, and though the district court formally rejected it, the challenge was wholly misdirected. Such a challenge, even if successful, would only have invalidated those portions of the inheritance law that treated natural illegitimate children differently than natural legitimate children. It would not touch a presumption affecting the question whether a person were in the first instance a natural child of the person from whom right of intestate succession as a child was sought to be shown.
. In summary, claimant points out the redundancy and internal inconsistency of a reading that requires an applicant to prove biological parentage as a threshold predicate and then in addition to prove parental acknowledgment (subpar. (C)(i)(I)), or decrees of other courts establishing parentage (supars. (C)(i)(II) and (III)); and the incomprehensibility of a reading that requires an applicant to prove biological parentage as a threshold predicate and then in addition to "show[ ] by evidence satisfactory to the Secretary [that the insured was] the ... father of the applicant, ...” (subpar. (C)(ii)).
The Secretary, however, defends these readings as necessary to give any meaning to the introductory reference in par. (3) to “[a]n applicant who is the son ..." and supports this by reference to a regulation that arguably so interprets the statute, see 20 C.F.R. § 404.355, and *1152claims for the regulation the judicial deference due such agency interpretation. Perhaps significantly, the Secretary advances a fail-back position if this interpretation should not be accepted under which the case would be remanded for consideration only under subpar. (C)(ii).
As indicated, because of our basis for decision, we need not resolve this problem of statutory construction.
. Though not on the constitutional ground urged by claimant. See note 1, supra.
. If entitlement is sought under either (C)(i)(II) or (C)(i)(III) (prior court decrees of parentage or support based on parentage) proof of either type of decree establishing natural parentage of the other-than-husband insured will indicate that the presumption has already been judicially overcome.
. See note 2, supra.
. Though we must always be wary about deciding critical issues on appeal on grounds other than those relied upon by lower tribunals, the wariness traces most powerfully — though not entirely — to concerns about fairness to the disfavored party. Here, we are satisfied that our basis for decision does not unfairly prejudice the claimant. Though the possibility that the common law presumption might apply as adopted federal common law (though not as state law by its own force or by incorporation) was not addressed either by the parties or the lower tribunals, we raised the possibility sua sponte on oral argument, and invited supplemental briefing by post-argument memoranda. Understandably, given the fact now established by our own research that the question is apparently one of first impression, the parties were unable to cite any judicial authority either adopting or rejecting adoption of the presumption into federal law. The Secretary cited only to McKenzie v. Harris, 679 F.2d 8 (3d Cir.1982), but that case (and there are others) only involved application of a comparable state law presumption as an element of congressionally incorporated state law under § (h)(2)(A).
Under the circumstances, we are persuaded that either allowing rehearing for full adversary development of the question in this court, or remanding to the district court for first instance application of the presumption in its same form but now as federal law would obviously be wasteful exercises not required in the interest of justice. Cf. United States v. The Barge Shamrock, 635 F.2d 1108, 1111 (4th Cir.1980).