330
mediated by the transfer of an unknown molecule. More broadly, these experiments,
along with others, showed that germ layer identity is not fixed before gastrulation,
and they confirmed Pander’s belief that interactions between the germ layers are
essential for normal embryonic development.
The first indication that ectodermal cells could be induced to form mesodermal
and endodermal tissue, came from experiments in which differentiated tissues from
adult guinea pigs were implanted into newt gastrulae, or cultured in sandwiches
between two ectodermal explants (Toivonen 1953 ; Takata and Yamada 1960 ; Saxen
et al. 1964 ). Two important conclusions were drawn from these results. First, they
showed that in the proper environment, presumptive ectoderm could be converted to
the other germ layers. Secondly, they showed that some tissues produced instructive
signals that could change the fate of ectoderm. This raised the possibility that such
interactions occur during normal development to form the germ layers. Pieter
Nieuwkoop (1917–1996) devised an experiment to identify the source of the endog-
enous mesoderm inducing signals. In a variation on Holtfreter’s explant experi-
ments, he explanted animal caps and vegetal pieces from salamander embryos and
grew them either separately or in combination (Nieuwkoop 1969 ). He found that
neither the animal caps nor the vegetal pieces were able to form mesoderm when
grown separately, but mesoderm was present when the vegetal pieces were grown in
combination with the animal caps (Fig. 7.6a). In salamanders, mesodermal tissue
was derived exclusively from cells in the pigmented animal cap ectoderm
(Nieuwkoop 1969 ). Similar results were obtained in Xenopus laevis (Nakamura
et al. 1970 ). These experiments identified the vegetal pole as the likely source of the
endogenous mesoderm inducing signals, although Nakamura cautioned that “‘meso-
dermal induction’, meaning the conversion of presumptive ectoderm into meso-
derm, is an abnormal phenomenon” which is only suggestive of the process that
occurs in normal development (p. 318) (Nakamura et al. 1970 ).
7.5.1.3 The Nieuwkoop Center
A series of more spatially and temporally refined recombination experiments showed
that the vegetal cells contained at least two types of inductive signals, and that these
signals act before gastrulation. Dorsally located vegetal cells were able to induce
ectodermal cells to become notochord, a dorsal mesodermal structure, with high fre-
quency, but could only rarely induce ectoderm to form blood, which is derived from
the ventral mesoderm. Ventral vegetal cells were able to induce ventral and interme-
diate mesodermal cell types easily, but could only rarely induce notochord (Dale
et al. 1985 ; Dale and Slack 1987b). A series of heterochronic recombinants showed
that vegetal pieces from young embryos could not act on gastrula-stage explants, but
that gastrula stage vegetal pieces could act on younger ectoderm (Dale et al. 1985 ).
This showed that the vegetal pieces continue to produce the mesoderm- inducing sig-
nal after gastrulation, but that the ectoderm loses the competence to respond to these
signals. The dorsal mesoderm inducing activity was further localized to the dorsal-
most vegetal cells of a 64-cell stage embryo, and its descendants (Gimlich and
Gerhart 1984 ). When transplanted to the ventral side of a host, these cells could
W. Tseng et al.