Angela Dawn Gann Bryowsky and Allan Dyoll Gann were divorced in January 1992. The trial court adopted the parties’ divorce agreement as part of its judgment of divorce. That agreement granted to the parties the joint legal custody of their minor child; however, physical custody of the child was vested in the mother.1 Following the divorce, the *1324parties entered into an informal arrangement whereby the minor child actually spent about half his time with his mother and half with his father. Because of the mother’s work schedule, this arrangement benefited the parties and the child. In December 1993, after learning that the mother was planning to remarry and move with the child to Mississippi, the father petitioned to have the divorce judgment modified so as to allow him to have physical custody of the child. After an ore tenus hearing wherein numerous witnesses testified, including the child, both parties, character witnesses, relatives, and professionals who had taught, tested, and worked with the child, the trial court entered a judgment for the mother. The father appealed to the Court of Civil Appeals, which reversed the judgment and remanded the ease with instructions for the trial court to enter a judgment awarding custody to the father. See Gann v. Bryowsky, 676 So.2d 1317 (Ala.Civ.App.1995), for a detailed statement of the facts surrounding this case. We granted the mother’s petition for certiorari review pursuant to Rule 39, Ala.R.App.P. We reverse and remand.
When evidence in a child custody case has been presented ore tenus to the trial court, that court’s findings of fact based on that evidence are presumed to be correct. The trial court is in the best position to make a custody determination — it hears the evidence and observes the witnesses. Appellate courts do not sit in judgment of disputed evidence that was presented ore tenus before the trial court in a custody hearing. See Ex parte Perkins, 646 So.2d 46, 47 (Ala.1994), wherein this Court, quoting Phillips v. Phillips, 622 So.2d 410, 412 (Ala.Civ.App.1993), set out the well-established rule:
“ ‘Our standard of review is very limited in cases where the evidence is presented ore tenus. A custody determination of the trial court entered upon oral testimony is accorded a presumption of correctness on appeal, Payne v. Payne, 550 So.2d 440 (Ala.Civ.App.1989), and Vail v. Vail, 532 So.2d 639 (Ala.Civ.App.1988), and we will not reverse unless the evidence so fails to support the determination that it is plainly and palpably wrong, or unless an abuse of the trial court’s discretion is shown. To substitute our judgment for that of the trial court would be to reweigh the evidence. This Alabama law does not allow. Gamble v. Gamble, 562 So.2d 1343 (Ala.Civ.App.1990); Flowers v. Flowers, 479 So.2d 1257 (Ala.Civ.App.1985).’ ”
It is also well established that in the absence of specific findings of fact, appellate courts will assume that the trial court made those findings necessary to support its judgment, unless such findings would be clearly erroneous. See the eases collected at 3 Ala. Digest 2d Appeal & Error § 846(5) (1993).
There are two different standards for reviewing custody arrangements. If custody has not previously been determined, then the “best interest of the child” standard is appropriate. Ex parte Couch, 521 So.2d 987 (Ala.1988). However, if a judgment has granted custody to one parent, or if one parent has given up legal custody, then custody will be changed only if the change would “materially promote” the child's welfare. Ex parte McLendon, 455 So.2d 863 (Ala.1984). In the present case, the trial court did not state which of these standards it found to be applicable and it made no specific findings that would support its judgment. Emphasizing the fact that the parties had worked out an informal arrangement to share child-rearing responsibilities, and relying on Ex parte Couch, supra, the Court of Civil Appeals held that the “best interest of the child” standard was applicable. The court then reviewed the judgment, using both of the standards, ultimately concluding that the trial court’s refusal to award custody to the father was clearly erroneous.2
The mother contends that the reliance by the Court of Civil Appeals on Ex parte Couch in support of its application of the “best interest of the child” standard was misplaced. She argues that the more stringent McLen-don standard should have been applied because she had previously been granted physi*1325cal custody of the child in the parties’ divorce judgment. She also contends that the record does not support the conclusion reached by the Court of Civil Appeals that the trial court’s judgment is clearly erroneous. We agree.
The Court of Civil Appeals correctly noted that the McLendon standard does not always apply in joint custody situations. However, Ex ‘parte Couch involved joint legal custody and shared physical custody of the children where no judicial determination had been made preferring either parent. Therefore, when the father in Ex parte Couch sought custody, we applied the “best interest of the child” standard because under the particular facts of that case both parents were on equal ground in attempting to gain custody of the children. In the present case, the parties had joint legal custody, but a previous judicial determination had granted physical custody to the mother. This distinguishes this case from Ex parte Couch. See Blackmon v. Scott, 622 So.2d 393 (Ala.Civ.App.1993), for a case substantially similar to this one, where the Court of Civil Appeals applied the McLendon standard. We further note that it makes no difference that the parties here had worked out an informal arrangement to care for the child. Although the nature of that arrangement and its impact on the child would be factors to be considered in determining whether a transfer of custody from the mother to the father would materially promote the child’s welfare, the arrangement itself did not constitute a waiver on the mother’s part of her right to preference under the divorce judgment. Ex parte Couch. Therefore, under the facts of this case, the McLendon standard is applicable, and we assume that the trial court found that awarding custody to the father would not materially promote the child’s welfare, unless such a finding would be clearly erroneous.3
As previously noted, the trial court’s judgment followed extensive testimony, much of which was disputed, concerning the respective lifestyles of the mother and the father; their respective abilities to provide for the child’s emotional, social, moral, material, and educational needs; and the family support that was available to the parents in caring for the child. The Court of Civil Appeals noted that “[t]he record reveals that both parents love [the child] very much and that both parents can provide for his material needs.” 676 So.2d at 1319. However, it is apparent from a comparison of the opinion of the Court of Civil Appeals to the record that that court, in support of its decision to reverse, proceeded to set out only the evidence that reflected most favorably on the father. The Court of Civil Appeals did not discuss the evidence indicating that the father smoked marijuana; that he had been known to abuse alcohol and that alcohol abuse was one of the factors that had led to the parties’ divorce; that the father had refused to cooperate in the mother’s efforts to treat the child’s attention deficit disorder; that the father on occasion had acted irresponsibly while the child was in his care (e.g., there was evidence that the father had on many occasions arrived to pick the child up from the grandparents’ house smelling of alcohol); that the child had expressed concern that his father preferred to party with friends rather than spend time with him; that the father’s apparent hatred for the mother was affecting the child (e.g., there was evidence that the child had expressed feelings of guilt concerning his father’s hatred for his mother); and that the father had in the child’s presence cursed the child’s stepfather. There was also evidence that the mother was a good parent; that the child’s new house in Mississippi was *1326“nice” and was within approximately a two-hour drive from Hamilton, Alabama, where his grandparents and the father lived; that the child made new friends easily; that the child’s new school in Mississippi had personnel trained to help him with his attention deficit disorder; that the work schedules of the mother and the stepfather were such that the child would not be left alone; and that (according to the testimony of the mother’s father, who had visited the child in Mississippi) the child was adjusting well in Mississippi and was happier than he had been in years.
Neither the Court of Civil Appeals nor this Court is allowed to reweigh the evidence in this ease. This case, like all disputed custody cases, turns on the trial court’s perception of the evidence. The trial court is in the better position to evaluate the credibility of the witnesses (at one point in its opinion the Court of Civil Appeals stated that witness Susan Brown was the only “impartial witness” with respect to the question whether the child’s behavior and ability to concentrate had improved on medication; 676 So.2d at 1320-1321) and the trial court is in the better position to consider all of the evidence, as well as the many inferences that may be drawn from that evidence, and to decide the issue of custody. As previously noted, we assume that the trial court found that a change of custody would not materially promote the child’s welfare. Based on our review of the evidence, we must conclude that that finding was not clearly erroneous.
For the foregoing reasons, the judgment of the Court of Civil Appeals is reversed and the case is remanded for an order or proceedings consistent with this opinion.
REVERSED AND REMANDED.
HOOPER, C.J., and MADDOX, SHORES, KENNEDY, INGRAM, and BUTTS, JJ., concur.
COOK, J., concurs in the result.