Marriage of Pikula v. Pikula

*707 WAHL, Justice. This matter concerns the propriety of the custody award of two minor children in the judgment and decree dissolving the marriage of Kelly Jo Pikula and Dana David Pikula. Both parents sought custody of their daughters, aged 4 and 2. After a two day trial, the trial court awarded custody to Dana, the father. On Kelly’s appeal, the Court of Appeals reversed, 349 N.W.2d 322, concluding that the evidence, considered in light of the statutory factors set forth in Minn.Stat. § 518.17, subd. 1 (1984), was insufficient to support the award of custody. The Court of Appeals remanded the matter with direction to the trial court to enter judgment granting custody to the mother. We granted discretionary review. Kelly and Dana Pikula were married on March 29, 1980, when Kelly was 17 and Dana 20. At the time of their marriage, their older daughter, Tiffany, was 8 months old. Prior to Tiffany’s birth, Kelly and Dana had lived with Kelly’s sister, Denise, in St. Paul. After the baby was born, the family moved to Brainerd, Dana’s hometown, where they had frequent contact with Dana’s parents and sisters. The Pikula family is closely knit, with Dana’s parents at the center of the family. The family members visit each other frequently and spend holidays together. Two of the three adult Pikula children work for their father, and the parents continue to assist the adult children financially. Dana took a job with his father’s trucking company, working a split shift as a driver. Kelly had a second daughter, Tanisha, in 1981, and finished high school while taking care of the children and managing the home. As the Court of Appeals observed, it appears from the evidence that both Kelly and Dana were imperfect parents. Dana and members of his family testified Kelly occasionally had trouble controlling her temper with the two girls, was somewhat ambivalent about her role as mother, and was a poor housekeeper. Kelly did not dispute she was sometimes dissatisfied and frustrated, but by her own account and by the testimony of Dana and his family, she was a good, mother. She testified her dis-satisfactions were rooted in her relationship with Dana and in Dana’s problems with alcohol which at times resulted in physical displays of temper and verbal abuse. These problems persisted throughout the marriage and became particularly severe after Tanisha, their second child, was born. Dana was hospitalized during this period after injuring his hand by putting his fist through a door. He initially agreed to undergo counseling at that time, but soon stopped attending because he “didn’t feel he had a problem with other people.” He did attend AA meetings for a period, but began drinking again after five or six months. According to the report prepared by the custody evaluator, Dana continues to have problems with chemical dependency. Kelly and her sisters also testified Dana’s drinking in part precipitated the couple’s separation. At the time Dana began drinking again, Dana forced Kelly and the ■ children to leave her sister Renee’s home in St. Paul where Kelly had been visiting with the children. Dana appeared at the house at around 9 p.m. and insisted Kelly and the girls leave immediately with him. When Kelly resisted, he took the children, put them in the car, and then dragged Kelly out of the house. In the meantime, Renee’s boyfriend came out of the house and hit Dana on the arm with a baseball bat. Kelly said the children were watching this scene from the car, and once they were underway, Dana drove recklessly, shouted at her, and prevented her from comforting the children. Dana denies he used physical force, had trouble operating the car, or kept Kelly from the children. Kelly’s sisters stated, though, they were sufficiently concerned to report the incident to the police. Kelly did not remain in the home long after their return from St. Paul. She said Dana told her he was going to keep her there and he intended to take the children away so she would know what it was like to be alone. He was angry at her for not taking his side against her sister’s boy *708 friend. Kelly then left the home and moved into the Women’s Center of Mid-Minnesota, a shelter for battered women, where she continued to live until the time of the trial. During this time, the couple agreed to a joint custody arrangement until custody was judicially determined. The arrangement was an uneasy one. For a time, the children remained in the family home while Kelly and Dana alternated living there on a four-day rotation schedule. Kelly began bringing the children to the shelter for her custody period, however, when tensions between Kelly and Dana escalated. The recommendations of three professional social workers were also before the trial court. All three recommended that custody be awarded to Kelly. Social worker Jean Remke met with Kelly and Dana together or separately four times. In her view, both Kelly and Dana are somewhat emotionally immature. In Remke’s opinion Kelly is “decidedly the most functional parent,” because she seemed more capable of “putting herself aside to attend to the physical and emotional needs of others,” while Dana “repeatedly used the children in efforts to control their mother,” and showed “no signs of really understanding this and no signs of altering his behavior.” Social worker Louise Seliski had extensive contact with Kelly at the shelter, both through individual counseling and observation. Seliski also found Kelly had been a fit mother to the two girls and believed she would continue to provide a loving and supportive environment for them. She said she observed affection between Kelly and the children, that Kelly never used excessive discipline, and that the children were always clean. Seliski terminated therapy with Kelly because Kelly was “handling her life as well as anyone could expect her to handle it” and had no significant psychological problems or chemical dependency. It was Seliski’s recommendation that custody be given to Kelly. The reports prepared by Remke and Seli-ski were included in the custody evaluation prepared by social worker Nancy Archibald. The evaluation also included reports of interviews with the parties, their families, neighbors and friends, a church premarital evaluation, and letters of recommendation. In Archibald’s opinion, the views expressed by Remke and Seliski were supported by her interviews with Kelly and Dana. She also recommended, based on all the data, that custody be awarded to Kelly with reasonable visitation provided to Dana. Evidence was also introduced at trial concerning the custodial environment each parent would provide the children. Kelly testified she intended to move with the children to her sister’s home in Maplewood until she could find employment and move into her own apartment. Dana objected to this plan, and testified that Kelly’s sister had used marijuana and characterized some of her sister’s friends as “bikers.” Dana testified that he intended to remain in Brain-erd if he were awarded custody of the girls. He continued to work a split shift at the time of trial, and his schedule required him to leave Brainerd at 3:00 a.m. for Wadena, lay over in Wadena from 7:00 a.m. until 3:00 p.m., and return to Brainerd at 7:00 p.m. Occasionally, he would return to Brainerd during his layover, permitting him to spend several hours at home. The child care responsibilities were principally borne by Dana’s mother, however, and the children frequently spent the night with her and were cared for by her during the day. Based on this record, the trial court initially made two key findings of fact in awarding custody to Dana. These findings stated as follows: Finding 11. That there is a strong, stable, religious family group relationship within the Pikula family that has been developed, nurtured and cultivated over the years. It has stood like a bedrock through the depression years and postwar years of plenty and permissiveness. This environment has inbred in the family a unity, respect, loyalty and love that for the most part has been destroyed and lost in most modern American families. *709 It is in the best interest of the children that they be kept in the cultural family environment. Finding 12. That the environment in which petitioner finds herself is almost the exact opposite. It is characterized by self-interest and excessive liberalism. Both findings were amended 4 months later pursuant to Kelly’s motion for amended findings to read: Amended Finding 11. That there is a strong, stable, religious family group relationship within the Pikula family, including respondent and the children, that has been developed, nurtured and cultivated over the years. It has stood like a bedrock through the depression years and post-war years of plenty and permissiveness. This environment has inbred in the family, including respondent, a unity, respect, loyalty and love that for the most part has been destroyed and lost in most modern American families. It is in the best interests and welfare of the children that their custody be awarded to respondent, who shares these attributes and who will assure that these children will be raised in the present cultural, family, religious and community environment of which they have been and are integral parts, which environ- • ment affords them stability, appropriate socializing and family orientation. The children are properly adjusted to their current home situation, broadly defined, and to the greater community within which they have lived virtually their entire lives, the children behave well and have extensive and qualitative contacts with significant persons within this environment, respondent’s personal environment continues to stabilize and improve and is presently satisfactory, as well as gives indications of continuing stability, and it is desirable that the children’s continuity with respondent and significant other persons and institutions here be maintained, respondent offering a permanent, well-established, concerned and involved, as well as supporting home for the children, the overall health of those who likely will here affect the mental, physical, emotional, educational, cultural and religious growth of the children is good, and respondent is inclined to, has and likely will continue to care for the children and raise them in their religion, creed and culture. Amended Finding 12. That the environment in which petitioner finds herself is almost the exact opposite ■ of that in which respondent lives and will raise the children, it would subject the children to considerable uncertainty and instability in home, community, culture, persons and religion, should custody be awarded to petitioner, and further, such an award would disrupt, curtail and likely end the children’s nurturing and constant contacts with the environment, persons and institutions now significantly and positively affecting their lives, petitioner’s behavior and practices of child rearing as well as her interest in her children are at least subject to serious question and doubt, and it would not be in the children’s best interest to award their custody to petitioner. The Court of Appeals, in reversing, held . that the trial court had abused its discretion in awarding custody of the children to Dana on the facts of this case. The Court of Appeals noted little reference was made by the trial court to the statutory factors set forth in Minn.Stat. § 518.17, subd. 1 (1984) in fashioning its findings, that the trial court had instead emphasized the desirability of Dana’s extended family, and had dismissed the expert opinions of three social workers. Moreover, the court observed it appeared the trial court had penalized Kelly for the divorce and remarriage of her parents, for actions and attitudes of other close relatives, and for her involvement with persons concerned with women’s issues, without finding that any of these factors affected the relationship between children and mother. The Court of Appeals viewed the matter as controlled by our decision in Weatherly v. Weatherly, 330 N.W.2d 890 (Minn.1983), where we reversed the trial court’s award of custody to the father because the court had disregard *710 ed all evidence reflecting negatively on the father, disregarded substantial evidence in favor of the mother, and disregarded statutory criteria which would have weighed in favor of the mother. On appeal to this court, Dana claims the Court of Appeals erred in reversing the trial court’s award of custody. He contends the Court of Appeals exceeded its scope of review for custody awards, that the trial court’s findings are sufficiently supported by the evidence, and that the award is in conformity with law. 1. The petitioner asserts the Court of Appeals exceeded the scope of its power of review in determining the issue before it. Appellate review of custody determinations is limited to whether the trial court abused its discretion by making findings unsupported by the evidence or by improperly applying the law. Weatherly v. Weatherly, 330 N.W.2d 890 (Minn.1983); Berndt v. Berndt, 292 N.W.2d 1 (Minn.1980). The Court of Appeals saw the issue to be “whether the trial court properly granted custody of the parties’ two children to respondent,” suggesting that de novo review of the entire record was appropriate. Such a review is inappropriate. That the Court of Appeals may have misconstrued the scope of its review in reaching its decision, however, does not automatically mandate reversal. This court must review the sufficiency of the evidence in the underlying record to determine the propriety of the trial court’s decision and may or may not reach the same result as that reached by the Court of Appeals. 2. The trial court's findings must be sustained unless clearly erroneous. Weatherly, supra; Berndt, supra. The trial court’s first finding describes the custodial environment provided by Dana and his family. The court found that the extended Pikula family surrounded the girls with a social milieu imbued with the “traditional” values, shared and fostered by Dana. The court further found that the girls were significantly attached to that environment, so as to be afforded “stability, appropriate socializing and family orientation,” and that Dana would continue to raise them within “their religion, creed and culture.” Given the close knit and interdependent character of the Pikula family, these findings appear reasonably supported by the evidence, although they do emphasize the desirability of the environment that would be provided by the paternal grandparents and the extended family rather than that of the proposed custodian, the father. The trial court’s findings regarding Kelly’s fitness as a custodial parent are troubling in light of the whole record. The court found Kelly’s environment would “subject the children to considerable uncertainty and instability in home, community, culture, persons and religion,” that granting Kelly custody would sever the children’s relationship with the Pikula family, and that Kelly’s “behavior and practices of child rearing as well as her interest in her children are at least subject to serious question.” Each of these findings was contradicted by evidence submitted by Kelly, and inconsistent with testimony from Dana and his family that Kelly was a good mother. In fashioning these findings, the trial court also discredited the custody evaluator’s report and the recommendations of two other professional counsellors. The trial court is not, however, bound to adhere to such expert testimony if it believes it is outweighed by other evidence. While the grounds for the trial court’s failure to consider this evidence are not apparent, given our limited scope of review we cannot conclude there was not sufficient evidence on the record to outweigh it. We therefore hold that the evidence was adequate to support the findings which the trial court did make. 3. We conclude, however, that the trial court erred in determining that custody of the children should be awarded to Dana on the basis of the facts that were found. Minn.Stat. § 518.17, subd. 3 (1984) provides that “[i]n determining custody, the court shall consider the best interests of the child * * The statute further *711 defines ‘ best interests of the child” as “all relevant factors” to be considered and evaluated by the court, including: (a) The wishes of the child’s parent or parents as to his custody; (b) The reasonable preference of the child, if the court deems the child to be of sufficient age to express preference; (c) The interaction and interrelationship of the child with his parent or parents, his siblings, and any other person who may significantly affect the child’s best interests; (d) The child’s adjustment to his home, school, and community; (e) The length of time the child has lived in a stable, satisfactory environment and the desirability of maintaining a continuity; (f) The permanence, as a family unit, of the existing or proposed custodial home; (g) The mental and physical health of all individuals involved; (h) The capacity and disposition of the parties to give the child love, affection, and guidance, and to continue educating and raising the child in his culture and religion or creed, if any; and (i) The child’s cultural background. The court shall not consider conduct of a proposed custodian that does not affect his relationship to the child. Minn.Stat. § 518.17, subd. 1 (1984). In Berndt v. Berndt, 292 N.W.2d 1 (Minn.1980), we held the enumerated statutory criteria, even absent consideration of other relevant factors, mandate that, when the evidence indicates that both parents would be suitable custodians, the intimacy of the relationship between the primary parent and the child should not be disrupted “without strong reasons which relate specifically to the [primary] parent’s capacity to provide and care for the child.” Berndt, 292 N.W.2d at 2. Awarding custody to the non-primary parent without such strong reasons, when the primary parent has given the child good care, may constitute reversible error. Id. The guiding principle in all custody cases is the best interest of the child. Berndt, 292 N.W.2d at 2, citing Minn.Stat. § 518.17, subd. 1 (1976); Rosenfeld v. Rosenfeld, 311 Minn. 76, 249 N.W.2d 168 (1976); LaBelle v. LaBelle, 296 Minn. 173, 207 N.W.2d 291 (1973). The importance of emotional and psychological stability to the child’s sense of security, happiness, and adaptation that we deemed dispositive in Bemdt is a postulate embedded in the statutory factors 1 and about which there is little disagreement within the profession of child psychology. See J. Goldstein, A. Freud and A. Solnit, Before the Best Interests of the Child 31-35 (1979); Leonard & Provence, The Development of Parent-Child Relationships and the Psychological Parent, 53 Conn.B.J. 320, 326 (1979); Okpaku, Psychology: Impediment or Aid in Child Custody Cases? 29 Rutgers L.Rev. 1117, 1121-22 (1976) cited in Klaff, The Tender Years Doctrine: A Defense, 70 Cal.L.Rev. 335, 348 (1982). For younger children in particular, that stability is most often provided by and through the child’s relationship to his or her primary caretaker — the person who provides the child with daily nurturence, care and support. As we further noted in Berndt, a court order separating a child from the primary parent could thus rarely be deemed in the child’s best interests. Courts in three other states have reached similar conclusions in construing their custody statutes and rules. 2 Garska v. McCoy, 278 S.E.2d 357 (W.Va.1981); In re Maxwell, 8 Ohio App.3d 302, *712 456 N.E.2d 1218 (1982); VanDyke v. VanDyke, 48 Or.App. 965, 618 P.2d 465 (1980); see also Commonwealth ex rel. Jordan, 302 Pa.Super. 421, 448 A.2d 1113 (1982). We follow the reasoning of those states in adopting the rule that when both parents seek custody of a child too young to express a preference, and one parent has been the primary caretaker of the child, custody should be awarded to the primary caretaker absent a showing that that patent is unfit to be the custodian. Continuity of care with the primary caretaker is not only central and crucial to the best interest of the child, but is perhaps the single predicator of a child’s well-being about which there is agreement, and which can be competently evaluated by judges. The other indicia of a child’s best interests set forth in section 518.17, while plainly relevant to a child's wellbeing and security, are, by contrast,, both inherently resistant of evaluation and difficult to apply in any particular case. Subdivision 1(g) and (h) require judges to assess the proposed custodians’ “mental and physical health,” and “capacity and disposition” to give the child “love, affection, and guidance.” A trial court is further required to consider all other “relevant factors” in reaching its decision. We are mindful that trial courts, seeking to apply these factors to reach an intelligent determination of relative degrees of fitness, must aspire to a “precision of measurement which is not possible given the tools available to the judges.” Garska, 278 S.E.2d at 361. Moreover, as one author has observed, “[ejmpirical findings directly or indirectly relevant to questions for which judges deciding difficult [custody] cases need answers are virtually nonexistent.” Okpaku, supra at 1140. The legislature, in enacting Minn.Stat. § 518.-167 (1984), which permits trial courts to order professional custody evaluations, has recognized the special needs of judges in that regard. That custody evaluations may not adequately provide a judge with such needed insight in particular cases, however, is embedded in the rule that such evaluations may be disregarded when outweighed by other evidence. This inherent lack of objective standards aside from primary parent status in custody determinations has several related effects which are not in the best interests of children. Imprecision in the application of the law may result in “wrong” results, and in unpredictability of outcome. Parents already estranged may be tempted to use a threatened custody contest strategically when neither parent can predict with any certainty which parent will ultimately be awarded custody. The availability of such strategies cannot in any sense be viewed as in the best interests of the children involved. See generally Mnookin & Korn-hauser, Bargaining in the Shadow of the Law: The Case of Divorce, 88 Yale L.J. 950 (1979). This situation is exacerbated by the fact that the two parents may be unequally situated with respect to other matters at issue in the negotiation process. A parent who has remained at home throughout a marriage to raise the children will often have sacrificed economic and educational opportunities in order to perform that role, and he or she will likely be in greater need of economic support upon dissolution of a marriage. A spouse in that position has *713 only one issue available to “concede” in the division of marital assets: custody of the children. At the same time, as the Garska court observed, “uncertainty of outcome is very destructive of the position of the primary caretaker parent because he or she will be willing to sacrifice everything in order to avoid the terrible prospect of losing the child in the unpredictable process of litigation.” Garska, 278 S.E.2d 356 at 360 (W.Va.1981). Moreover, in'practical fact, many primary caretakers may simply be unable to afford the expense of litigation at all, further weakening their bargaining position when the uncertainty in the outcome of a trial is necessarily high. The rule we fashion today should largely remove the issue of custody from the arena of dispute over such matters, and prevent the custody determination from being used in an abusive way to affect the level of support payments and the outcome of other issues in the proceeding. The inherent imprecision heretofore present in our custody law has, in turn, diminished meaningful appellate review. We have repeatedly stressed the need for effective appellate review of family court decisions in our cases, and have required specificity in written findings based on the statutory factors. Rosenfeld v. Rosenfeld, 311 Minn. 76, 249 N.W.2d 168, 171 (1976); Peterson v. Peterson, 308 Minn. 297, 242 N.W.2d 88, 94 (Minn.1976); Wallin v. Wallin, 290 Minn. 261, 267, 187 N.W.2d 627, 631 (1971). We are no less concerned that the legal conclusion reached on the basis of those findings be subject to effective review. We recognize the inherent difficulty of principled decisionmaking in this area of the law. Legal rules governing custody awards have generally incorporated evaluations of parental fitness replete with ad hoc judgments on the beliefs, lifestyles, and perceived credibility of the proposed custodian. See, e.g., Jarrett v. Jarrett, 78 Ill.2d 337, 36 Ill.Dec. 1, 400 N.E.2d 421 (1979), cert. den. 449 U.S. 927, 101 S.Ct. 329, 66 L.Ed.2d 155 (1980) (mother’s cohabitation contrary to the moral standards of the state). It is in these circumstances that the need for effective appellate review is most necessary to ensure fairness to the parties and to maintain the legitimacy of judicial decisionmaking. For these reasons — the recognized need for stability in children’s lives, the uncertainty of other indicia of a child’s best interests in custody decisions, and the pressing need for coherent decisionmaking on the trial court level and for effective appellate review — we hold the factors set forth in section 518.17, subd. 1, require that when both parents seek custody of a child too young to express a preference for a particular parent and one parent has been the primary caretaker, custody be awarded to the primary parent absent a showing that that parent is unfit to be the custodian. We adopt the indicia of primary parenthood set forth in ■ Garska to aid trial courts in determining which, if either, parent is the primary caretaker: While it is difficult to enumerate all of the factors which will contribute to a conclusion that one or the other parent was the primary caretaker parent, nonetheless, there are certain obvious criteria to which a court must initially look. In establishing which natural or adoptive parent is the primary caretaker, the trial court shall determine which parent has taken primary responsibility for, inter alia, the performance of the following caring and nurturing duties of a parent: (1) preparing and .planning of meals; (2) bathing, grooming and dressing; (3) purchasing, cleaning, and care of clothes; (4) medical care, including nursing and trips to physicians; (5) arranging for social interaction among peers after school, i.e. transporting to friends’ houses or, for example, to girl or boy scout meetings; (6) arranging alternative care, i.e. babysitting, day-care, etc.; (7) putting child to bed at night, attending to child in the middle of the night, waking child in the morning; (8) disciplining, i.e. teaching general manners and toilet training; (9) educating, i.e., religious, cultural, social, etc.; and, (10) teaching elementary skills, i.e., reading, writing and arithmetic. Garska, 278 S.E.2d at 363. When the facts demonstrate that responsibility for and per- *714 formanee of child care was shared by both parents in an entirely equal way, then no preference arises and the court must limit its inquiry to other indicia of parental fitness. Once the preference does arise, however, the primary parent should be given custody unless it is shown that the child's physical or emotional health is likely to be endangered or impaired by being placed in the primary parent’s custody. The indicia of primary parenthood set forth above make plain that a parent who has performed the traditional role of homemaker will ordinarily be able to establish primary parent status in a custody proceeding involving young children. That this is so reflects no judgment by this court on the competence or fitness of parents who choose or are compelled to fashion less traditional divisions of labor within a family. Our decision today merely encompasses our understanding of the traumatic impact on children of separation from the primary caretaker parent. Nor do we mean to suggest that a parent who works outside the home may not be deemed the primary parent. We would expect that, as between any two parents, one will be the primary parent even if neither conforms to the more traditional pattern of one parent working outside the home and one within it. See Wagoner v. Wagoner, 310 S.E.2d 204 (W.Va.1983) (both parents employed, mother found primary caretaker). Turning to the facts of this case, we conclude that the matter must be remanded for a determination of which, if either, parent was the primary caretaker of the children at the time the dissolution proceeding was commenced. 3 Any disruption in the relationships between the children and their parents occasioned by the events leading to the divorce is irrelevant to that determination. If either parent was the primary caretaker, custody should be awarded to that parent absent a strong showing of unfitness. The Court of Appeals is affirmed in part in reversing the trial court's award of custody to the father, reversed in part in awarding custody to the mother, and the matter is remanded for proceedings consistent with the rule set out in this opinion. KELLEY, J., took no part in the consideration of this case. Footnotes [1] Four of the nine statutory criteria rest on the centrality of continuity of care and environment to the best interest of the child. See Minn.Stat. § 518.17, subd. 1(c), (d), (e), (f) (1984). [2] Historically, similar but not identical considerations supported the "tender years” doctrine, the legal rule that when a child is of a young age, custody should ordinarily be awarded to the mother. That doctrine, unlike the rule we announce today, was premised on a presumed natural capacity of women to selflessly and instinctively raise children, often articulated in terms such as: [N]othing can be an adequate substitute for mother love — for that constant ministration required during the period of nurture that only a mother can give because in her alone is duty swallowed up in desire; in her alone is service expressed in terms of love. *712 Jenkins v. Jenkins, 173 Wis. 592, 593, 181 N.W. 826, 827 (1921). While at one time the tender years doctrine was universally adopted in the state courts, most jurisdictions have repudiated the doctrine as sex discriminatory. A recent survey of state laws lists 38 states which have rejected the presumption, four which retain "tiebreaker" versions, and eight states with doubtful or unique laws. Freed & Foster, Divorce in Fifty States: An Overview as of August 1, 1981, 7 Fam.L.Rep. 4049, 4063 (1981). The primary parent preference, while in accord with the tender years doctrine insofar as the two rules recognize the importance of the bond formed between a primary parent and a child, differs from the tender years doctrine in significant respects. Most importantly, the primary parent rule is gender neutral. Either parent may be the primary parent; the rule does not incorporate notions of biological gender determinism or sex stereotyping. In addition, the rule we fashion today we believe will encourage co-parenting in a marriage unlike the tender years doctrine which, for fathers, meant that whatever function they assumed in the rearing of their children would be deemed irrelevant in a custody contest. [3] The phrase “at the time the dissolution proceeding was commenced” is used to indicate the point in time at which the family relationships were physically disrupted by events leading to the dissolution of the marriage, e.g., at the time of the parties’ separation or the interruption of the functioning full family unit. In a hearing on remand to determine which, if either, parent was the primary caretaker at that time, and whether that parent was a fit custodian, the trial court would, of course, use the record developed at trial and available to the court when the initial custody determination was made. While it would be inappropriate to permit factual consequences caused by an improper initial decision to determine on remand which parent is found to have been the primary caretaker, the trial court must look at present circumstances to determine the limited issue of the proposed custodian’s fitness. After an award of custody on remand, subsequent changes in circumstances of the children or custodian would be addressed by the noncustodial parent in a motion for modification of custody under Minn.Stat. sec. 518.18.