This suggests that the node produces additional neural inducers that are not BMP antagonists. .
Similarly, mis-expression of BMP-4 or BMP-7 within the neural plate does not induce epidermis (Streit et al 1999). Consistently with these findings, mouse knockouts of noggin, chordin, follistatin, BMP-2, BMP-4 or BMP-7 show no obvious defects in early ectodermal patterning. In addition, recent experiment in avian embryos suggest that neural induction requires signals originating in the underlying mesoderm in addition to those from the organiser (Pera et al 1999). Thus the data from amniotes seem to indicated that the neural induction is more complex than the neural default model initially suggested. Other signals appear to be required in addition to BMP inhibition, and these signals may also affect the formation of border cell types such as the neural crest.
When studying a development process, some degree of variability between organisms is to be expected. Some variation may be the result of the obvious differences in embryonic morphogenesis between amphibian which at the onset of gastrulation consists of a hollow ball of cells and avian which develop from a flat epiblast.
Unlike amniotes, early Xenopus development relies on localized maternal stores of mRNA and protein, since zygote transcription does not commence until mid-blastula stages. A consequence of the early localization of cell fate determinants is that some cells may initially be exposed to signals that actually interfere with the subsequent fate decisions, and mechanisms would then be necessary to counteract these signals. In addition, cell proliferation is significantly greater in amniotes, and increased cell numbers may allow for a greater separation of patterning events, both spatially and with respect to cell generation. Despite all these differences, however, it would seem rather unlikely that amniotes and anamniotes have evolved fundamentally different mechanisms for specifying neural fate at the molecular level.