Knoll et al. (1989b) have shown that the cell line DON13 described in the restriction fragment length polymorphism (RFLP) portion of my study (Donlon 1988) is actually not derived from the patient with Angelman syndrome that was described by Kaplan et al. (1987). It is unfortunate that this was not detected prior to publication, particularly in light of recent developments regarding the apparent consistency of maternally derived de novo deletions in Angelman syndrome patients (Cooke et al. 1988; Williams et al. 1988, 1989; Knoll et al. 1989a) and requires an explanation as to its possible origin. The dosage blot hybridizations on DON13 (Donlon 1988) were performed on DNA that was extracted from an early passage of this cell while the RFLP analysis (Donlon 1988) and those results reported by Knoll et al. (1988b) were performed on DNA from a later passage. Since the dosage blot results from my study of DON13 are consistent with results obtained on WJK48 reported by Knoll et al. (1989b), it is presumed that cell line DON13 either became contaminated and overgrown by another cell line, or was misidentified, at some time after the initial dosage analysis. While this is very unfortunate, a modification of the parental origin results does not change the general thesis of my paper (Donlon 1988), which I believe deserves restating. This manuscript showed that the Prader-Willi and Angelman syndromes can have similar deletions, with regard to the nine DNA segments that were used for analysis, so that the deletion, per se, is not solely responsible for the particular phenotype that is expressed, thus invoking mechanisms other than variation in breakpoint to account for the manifestation of these widely disparate syndromes. With these new parental origin results (Cooke et al. 1988; Williams et al. 1988, 1989; Knoll et al. 1989a) there are presently no exceptions to the association of Angelman syndrome with a maternally derived de novo deletion and only a few reported cases that the Prader-Willi syndrome is associated with a paternal deletion. This is an important point as it blatantly points toward a sex-related mechanism to account for the phenotypic differences in these two disorders, whether it be gamete selection or sex-related DNA modification that occurs during gametogenesis. The basic hypothesis (Donlon 1988) in my manuscript that was proposed to account for phenotypic difference still holds but now points conspicuously towards a sex-related modification of gene combinations rather than the inheritance of recessive allele combinations, as originally outlined. For example, as shown in Fig. 4 from Donlon (1988) a deletion will eliminate a set of loci on one chromosome and allow the expression of hemizygous alleles on the cytogenetically normal homolog. If the solid black rectangles represent loci that have been inactivated through sex-specific modification, then it can be seen that the resultant individual will be functionally "nullizygous" for them. Modification of complementary sequences in a parent of the opposite sex will result in a functional "nullizygosity" of different but complementary genes. Modification of the mother's loci shown by solid black rectangles in "a" and inheritance of a de novo deletion from the father will result in Prader-Willi syndrome, whereas modification of the father's loci in "b" and inheritance of a deletion from the mother will lead to Angelman syndrome. This hypothesis is supported by studies in the mouse that have shown that nuclear contributions from both parents are necessary for normal development of the zygote (Surani et al. 1984; McGrath and Solter 1984). While my study (Donlon 1988) has the misfortune of being associated with a contaminated or misidentified cell line leading to an erroneous result as to the parental origin of one case of de novo deletion, the basic tenet of the paper is accurate.