It is a frequent observation that twinning runs in the families [5] and we have recently reported that such is also the case with Gregor Mendel's pedigree [4]. The problem we are concerned with, now, is the identification of the “gemellogenetic” factor postulated by Gedda already in 1951 [2].

We thought we should start by going back to experimental embryology. Clearly, as shown by the experimentally induced MZ twinning in Triton or the spontaneous multiple MZ twinning in Armadillo, there must exist a temporal threshold below which individual parts of a single conceptus can still show totipotence.

Interestingly, recent studies of molecular histology have stressed the role of cell membranes and of membrane receptors in the transmission of specific messages related to the position and function of the cell. This has led Edelman [1], in particular, to identify a specific kind of molecule that appears early in-embryonic development and is responsible for cell adhesion, thus contributing to organogenesis and the shaping of the embryo by regulating cell movements and the topologic development of tissues.

These so-called cell-adhesion molecules (CAM) are proteins that can differentiate both in function and in expression. In terms of function, there are molecules responsible for the adhesion of nerve cells (N-CAM), of liver cells (L-CAM), etc. In terms of expression, there are embryonic forms (E-CAM) and adult forms (A-CAM), the former being characterized by larger quantities of sialic acid. This chemical difference accounts for a different aggregation effect on the cells, that is, for a different cell adhesion. Moreover, the E-CAM and A-CAM can be interconverted, at different times and with variable speed.