Almost 30 years after the initial study by Richard W. Smithells and coworkers, it is still unknown how maternal periconceptional folic acid supplementation prevents human neural tube defects (NTDs). In this article, questions about human NTD prevention are considered in relation to three groups of mouse models: NTD mutants that respond to folate, NTD mutants and strains that do not respond to folate, and mutants involving folate-pathway genes. Of the 200 mouse NTD mutants, only a few have been tested with folate; half respond and half do not. Among responsive mutants, folic acid supplementation reduces exencephaly and/or spina bifida aperta frequency in the Sp(2H), Sp, Cd, Cited2, Cart1, and Gcn5 mutants. Prevention ranges from 35 to 85%. The responsive Sp(2H) (Pax3) mutant has abnormal folate metabolism, but the responsive Cited2 mutant does not. Neither folic nor folinic acid reduces NTD frequency in Axd, Grhl3, Fkbp8, Map3k4, or Nog mutants or in the curly tail or SELH/Bc strains. Spina bifida frequency is reduced in Axd by methionine and in curly tail by inositol. Exencephaly frequency is reduced in SELH/Bc by an alternative commercial ration. Mutations in folate-pathway genes do not cause NTDs, except for 30% exencephaly in folate-treated Folr1. Among folate-pathway mutants, neural tube closure is normal in Cbs, Folr2, Mthfd1, Mthfd2, Mthfr, and Shmt1 mutants. Embryos die by midgestation in Folr1, Mtr, Mtrr, and RFC1 mutants. The mouse models point to genetic heterogeneity in the ability to respond to folic acid and also to heterogeneity in genetic cause of NTDs that can be prevented by folic acid.