In this brief summary, we argue that many widely held beliefs about HLA-G are questionable. Recent research has led to a re-evaluation of many of the characteristics that were thought to make HLA-G unusual among the MHC class I molecules. First, contrary to reports suggesting that the gene encoding HLA-G exhibits marked polymorphism in some human populations, recent data have shown that the HLA-G gene has comparatively little polymorphism - a feature that might allow it to be expressed in the placenta without causing rejection by the maternal immune system. Second, although truncated forms of HLA-G are generated in the placenta, most of them are unlikely to have significant biological effects as they do not reach the cell surface. Third, the hypothesis that a major role of HLA-G is to prevent attack of the placenta by maternal natural killer cells is now the subject of renewed scrutiny. Finally, there is little evidence that the induction of expression of HLA-G is a major mechanism by which tumor cells avoid immune attack. HLA-G has once again become as mysterious as when it was discovered: an MHC class I molecule expressed at a challengingly extraordinary site--the immunologically uneasy interface between mother and fetus.