NEELY, JUSTICE:

The Court granted this appeal and subsequent leave to proceed by motion to reverse exclusively for the purpose of reevaluating, in light of rapid changes in society, our well-settled law that "[w]ith reference to the custody of very young children, the law favors the mother if she is a fit person, other things being equal * * *." Syl. Pt. 1 Funkhouser v. Funkhouser, W. Va., 216 S.E.2d 570 (1975); part, Syl. Pt. 2, Settle v. Settle, 117 W. Va. 476, 185 S.E. 859 (1936). In this regard we affirm our prior holdings, and consequently, reverse the judgment of the Circuit Court of Berkeley County in this case.

The record shows that the parties were married in 1968 and that they had one child, a girl born in 1970. In September 1975 the parties separated and their child remained with the appellant wife. The appellant, alleging cruelty, filed a divorce complaint against the appellee husband who answered and counterclaimed for a divorce on the same grounds. The circuit court granted a divorce to the appellee husband and awarded him custody of the child, subject to the appellant's reasonable visitation rights. It was a specific finding of the circuit court that the appellant was not a fit person to have permanent custody of the child.

The trial court relied upon one incident of sexual misconduct on the part of the wife as grounds for awarding custody to the husband. The evidence indicates that very late one evening in December 1974, during a period of trial separation between the parties, the appellant and a male companion parked their car in downtown Martinsburg and entered a bar. Later that same evening, the appellant and her male companion left the bar and returned to their parked car where they were observed by the appellee, the appellee's cousin, and the appellee's minister, all of whom had followed the appellant to the area without her knowledge. Although the appellant denies it, the weight of the evidence is that she committed the act of fellatio with her male companion late that night in the parked car, after leaving the bar.

There was voluminous evidence in this case regarding the conduct of the parties which, except for this one incident of sexual misconduct, demonstrates nothing more than aggravated strife between two adults who quarreled, fought, and even on occasion physically abused one another. The evidence shows that the child received from both parents the type of affection and care which this society expects of competent parents. In fact both parties conceded that the other party was a perfectly fit "baby sitter" for the child, and the trial judge noted in his memorandum of opinion that both parties "took good care of the child when they had the child with them" and "both parties loved the child." The evidence covering the aggravated strife between the adults was offered to establish relative degrees of parental competence; however, it was essentially stipulated that the mother was fit, so the evidence of the parties' treatment of one another was introduced primarily to establish fault on the part of each party to the marriage in the hope that the child would be awarded to the less blameworthy party. It appears to be the contention in these types of cases that where a mother has been at fault in the dissolution of a marriage, that fault casts sufficient aspersions on her character to warrant a finding of unfitness. We reject this argument as it violates our rule that a mother is the natural custodian of children of tender years.

The appellee's primary argument in support of the ruling of the trial court is that our presumption of law that a mother is the natural custodian of children of tender years is unconstitutional, State ex rel. Watts v. Watts, 350 N.Y.S.2d 285 (N.Y.C. Fam. Ct. 1973) or that if the presumption is constitutional, it is certainly unwise. As the proper standard for determining constitutionality is whether the presumption furthers, in a rational way, a legitimate public purpose, we can say that the question of constitutionality and the question of wisdom are inextricably intertwined.

In the first instance, it is incorrect to characterize the presumption as denying the equality of competing parents' rights to have custody of their children, since all parental rights in this respect are subordinate to the interests of the innocent children. The first and overriding principle in child custody cases is the welfare of the children, Funkhouser v. Funkhouser, W. Va., 216 S.E.2d 570 (1975), Lipscomb v. Joplin, 131 W. Va. 302, 47 S.E.2d 221 (1948), and the presumption of maternal preference was initially conceived, and now operates, to serve that principle.

Even if we were to concede, however, that the parents' rights to custody of their children should be subjected to an equal protection analysis, there is no doubt that the presumption would withstand judicial scrutiny. So far, the United States Supreme Court has not decided that gender is a suspect category for equal protection purposes or that classifications based on sex must be subjected to strict scrutiny. It does appear, however, that the United States Supreme Court is examining gender-based distinctions under more stringent standards than the "rational basis" standard of review ordinarily applied to non-suspect categories. The emerging middle level standard is that ". . . classifications by gender must serve important governmental objectives and must be substantially related to the achievement of those objectives." Craig v. Boren, 429 U.S. 190 at 197 (1976). Under the West Virginia Constitution, the presumption's validity would depend upon whether it bears a rational relationship to a legitimate state purpose. See State ex rel. Piccirillo v. City of Follansbee, W. Va., 233 S.E.2d 419 (1977); State ex rel. Harris v. Calendine, W. Va., 233 S.E.2d 318 (1977). Should the presumption of maternal preference be measured by either the applicable state or federal standard, or even a strict scrutiny standard, should one be adopted in the future, it would in our view be constitutional. The sociological, biological and evidentiary reasons which are discussed in the following sections provide sufficient grounds for sustaining the presumption, in the absence of compelling and preponderant evidence of a superior alternative.

The appellee's argument of both the unconstitutionality and the unsoundness of our presumption is predicated upon recent, dramatic, social changes, which appear to indicate that the presumption is outmoded in today's world. In support of this position he cites In re Marriage of Bowen, 219 N.W.2d 683 (Iowa, 1974) where Justice McCormick says:

We do not think either parent should have a greater burden than the other in attempting to obtain custody in a dissolution proceeding. It is neither necessary nor useful to infer in advance that the best interests of young children will be better served if their custody is awarded to their mothers instead of their fathers. We previously emphasized the weakness of the inference; we now abandon it. [ 219 N.W.2d at 688]

Certainly there is respectable authority for the position that the abandonment of gender as a rational determinant of social roles in countless thousands of commercial occupations implies a similar abnegation of the current preference for the mother in the award of custody. See, e.g., Roth, The Tender Years Presumption in Child Custody Disputes, 15 J. Fam. L. 423 (1976-77) and the authorities cited therein. While this Court has carefully evaluated the arguments in favor of determining the relative competence of each parent for child custody on the evidence which each parent produces at a hearing, and awarding custody on that basis, the Court unanimously rejects any rule which makes the award of custody dependent upon relative degrees of parental competence rather than the simple issue of whether the mother is unfit.

Even though lifestyles are rapidly changing, and men are now performing tasks which, as recently as ten years ago, were almost exclusively performed by women, we nonetheless are persuaded that the primary responsibility for the maintenance of the house and the care and upbringing of minor children in this society still rests with the woman, even though simultaneously, as in this case, the woman may be earning an outside income. The socialization patterns which prevailed during the formative years of the current generation of parents with young children encouraged women to develop certain attitudes such as surpassing patience and a high tolerance for a close, grating, aesthetically unpleasant, and frequently oppressive, yet nonetheless absolutely indispensable physical relationship with children. We are not being normative in our reliance upon the socialization process; we merely avail ourselves of it for the benefit of young children in the same way that a physicist relies upon the law of gravity or a doctor relies upon osmosis. When the socialization pattern changes to the extent that the traditional roles of mother and father are reversed with such frequency that the presumption no longer bears any relation to reality, then the law, perforce of changed circumstances, will inevitably change.

Certainly the generalities which we infer from the socialization pattern alone, absent other considerations, would be slender reeds upon which to base our rule. However, there are other factors which militate in favor of the presumption and against the available alternatives. From a strictly biological perspective, children of the suckling age are necessarily accustomed to close, physical ties with their mothers, and young children, technically weaned, are accustomed to the warmth, softness, and physical affection of the female parent. The welfare of the child seems to require that if at all possible we avoid subjecting children to the trauma of being wrenched away from their mothers, upon whom they have naturally both an emotional and physical dependency. While a child is usually emotionally dependent upon his father, he seldom has the same physical dependency which he has upon his mother.

We have unsuccessfully been called upon to reconcile the welfare of children with the vindication of an ideal of non-gender-based societal roles. However, society is sharply divided about the desirability of such an ideal, and its implementation even among those strongly committed to the ideal is yet highly experimental. In order to reinforce the non-discriminatory nature of our presumption, it is important to remember that we are talking about a rule which operates in a situation where both parents are fit. In order to be fit, it is obvious that a mother must be willing to offer the type of closeness and physical contact which we assume on the part of mothers. Where a mother is emotionally unsupportive, fails to provide routine cleanliness, fails to prepare nourishing food, or otherwise demonstrates her unfitness, the presumption, by its own terms, will not apply.

All exceptions to the contrary notwithstanding, families in which the traditional roles of mother and father are reversed are extraordinary, at least in West Virginia. In the case before us we are not asked to inaugurate a new rule of law based upon preponderant, reliable social and biological evidence proving that our prior rule is unserviceable; rather, we are asked to inaugurate a new rule, the rationale for which is unsupported by preponderant creditable evidence. In this type of situation tradition and precedent, representing the collective wisdom of this society, provide an appropriate guide for our decision.

In this case we are undoubtedly concerned with a child of tender years, but in other cases the question will inevitably arise whether the child is of such an age that the tender years presumption is applicable. Accordingly we must attempt to establish the age perimiters within which the presumption is intended to operate.

The concept of "tender years" is somewhat elastic; obviously an infant in the suckling stage is of tender years, while an adolescent fourteen years of age or older is not, as he has an absolute right under W. Va. Code, 44-10-4 to nominate his own guardian. Between the two extremes are children who are more or less capable of expressing a preference concerning their custody. Where a child is under the age which entitles him to nominate his own guardian, but is, nonetheless, sufficiently mature that he can intelligently express a voluntary preference for one parent, then the trial judge is entitled to give that preference such weight as circumstances warrant, and where such child demonstrates a preference for the father, the trial judge is entitled to conclude that the presumption in favor of the mother is rebutted.

The dilemma of trying to reconcile the need for general rules with the need for just results in individual cases is a daunting presence in the law of domestic relations. We can justify our reliance upon a presumption which we perfectly well recognize will not operate perfectly in every case only by observing that the science of law bears in one respect a close analogy to the science of mechanics. The mathematician can easily demonstrate that a certain system of pulleys will suffice to raise a certain weight. But his demonstration proceeds on the supposition that the machinery is such as no load will bend or break. If the engineer, who has to lift a great mass of real granite by the instrumentality of real timber and real hemp, should absolutely rely on the propositions which he finds in treatises on dynamics, and should make no allowance for the imperfection of his materials, his whole apparatus of beams, wheels, and ropes would soon come down in ruin, and, with all his geometrical skill, he would be found a far inferior builder to those painted barbarians who, though they never heard of the parallelogram of forces, managed to pile up Stonehenge.

Recognizing the imperfections of our own materials we can justify our presumption only on the grounds that the presumption will achieve greater justice over a wider spectrum of cases than the alternative of endless hearings about issues which cannot, in any meaningful sense, be satisfactorily resolved in the adversary system. The presumption of favor of the mother, while obviously subject to challenge, is no more subject to challenge than expert testimony, the demeanor of the parties, or the competence of counsel.

In any hearing concerning child custody, does not the party who emotes most easily in public, the party who is most articulate, and the party who makes the most attractive physical appearance have the greater chance of persuading the trial court in a close case? Yet none of these qualities is particularly germane to the issue of which parent will make a better custodian. Furthermore, does not a hearing raise the specter of the richer party retaining more numerous and better qualified experts to offer evidence of the richer parent's own superior parental competence? The fact finding process of courts is imperfect at best; it is unacceptable when the decision to be rendered has both subjective and objective elements, the mix of which is almost impossible to preordain by general rules, and the outcome almost uncontrollable by appellate review. If nothing else, the presumption provides a definite standard and a predictable result which is not conceivably related to the trial court's knowledge of the families involved, a serious problem whenever we are dealing with men, no matter how honest, in a rural setting where courts and litigants are frequently well known to one another.

If this Court felt that rational child custody decisions could be made in close cases by having each spouse attempt to prove himself or herself a relatively more competent custodian than the other, we might be inclined to question more critically our reliance on the presumption of maternal preference. As matters stand now, however, there is no reliable way to determine a custodian's degree of relative competence with respect to individual children except, possibly, by the use of behavioral science methods. Unfortunately, behavioral science has not advanced to the point where it can really help courts confronted by difficult child custody decisions. In fact, behavioral science is yet so inexact that we are clearly justified in resolving certain custody questions on the basis of prevailing cultural attitudes which give preference to the mother as custodian of young children.

The basic problem with behavioral science assessment of parental fitness and predictions of custody arrangements best suited to promote children's welfare is that "empirical findings directly or indirectly relevant to questions for which judges deciding difficult [custody] cases need answers are virtually nonexistent." Okpaku, Psychology: Impediment or Aid in Child Custody Cases?, 29 Rutgers L.R. 1117 at 1140 (1976). Existing theoretical research in this area is seriously deficient, and research opportunities for those hoping to improve on past efforts are inherently limited by the confidentiality of custody proceedings.

Because of the lack of relevant empirical data, expert witnesses in the behavioral sciences can contribute very little to the resolution of difficult custody problems, despite the inclination of some courts to rely on expert testimony in this area. For an expert to form an opinion concerning the most desirable custody alternative, he must first assess the personalities of the adults involved and the emotional adjustment of the children. Normally the assessment is done by the use of in-depth interviewing or psychological testing. The expert must then integrate his assessments of the individuals with general personality theory as a basis for predicting future behavior patterns and the consequences those behavior patterns have upon the children's emotional well-being. At this point, however, the trial court confronts the insurmountable obstacle that there are no reliable, empirical studies that can be used to predict the consequences of an adult's assumed future behavior upon a child. The end result is that ". . . without empirical data specifying how adult behavior affects children, the behavioral science expert is without a scientific basis for an opinion on any issue in a difficult case." id at 1143.

In the case before us we have one isolated instance of sexual misconduct which in the context of this case is a wrong against the husband, but totally unrelated to the mother's relationship with her child. Except for this one incident, the record is devoid of any evidence that the appellant is an unfit parent.

We are thus confronted with the age-old problem of whether one party's sexual misconduct vis-a-vis the other party to a marriage, standing alone, is a sufficient ground to rebut our presumption in favor of awarding the custody of children of tender years to the mother. We are not confronted in this case by a mother whose sexual peccadilloes interfere with, or have a bearing upon, her relationship with her child. We can certainly envisage a situation in which a woman's desire to consort with numerous men causes her to absent herself from the home and casts aspersions upon her fitness as a parent, and we can similarly envisage circumstances in which a woman would engage in such outrageous conduct, given contemporary moral standards, that her conduct per se would call into question her fitness as a parent. This, however, is not the situation in the case before us. While the appellant's conduct in this case might be outrageous to some, reasonable men would differ about whether it were sufficiently outrageous, per se, to lead us to conclude she is an unfit custodian, given the lack of consensus about these matters in contemporary society. In this type of situation, where we are drawing the inference that immoral behavior, per se, makes a person unfit as a guardian of a child of tender years, the only workable standard for the rebutting of the presumption is that the conduct must be so outrageous that reasonable men cannot differ about its deleterious affect upon the child.

No court can condone the actions of the wife in this case because, as the parties were married at the time of her sexual indiscretion, she committed a serious marital wrong against the husband. The award of child custody, however, should not be an exercise in punishment of an offending spouse. In punishing the offending spouse one may also punish the innocent child, and our law will not tolerate that result. To the extent that Rohrbaugh v. Rohrbaugh, 136 W. Va. 708, 68 S.E.2d 361 (1951) or any other prior cases in this jurisdiction are contrary to this holding, they are overruled. Therefore, unless the mother's sexual indiscretions are so outrageous as to cast aspersions on her capacity to care for and raise children, isolated acts of sexual misconduct are not to be considered by a trial judge as relevant evidence on the issue of whether a mother is fit.

Accordingly, for reasons set forth above the judgment of the Circuit Court of Berkeley County is reversed and the case is remanded with directions to award custody of the minor child to the appellant.

Reversed with directions.