As the ova are formed at as early a period as the rest of the maternal
structure, Galton notices that it seems improbable that they would be
correspondingly affected by subsequent modifications of parental
structure. Of course it is not certain that this is a valid argument. We
know that the paternal half of the reproductive elements does not enter
the ovum till a comparatively late stage in its history, and it is quite
possible that maternal elements or gemmules may also enter the ovum from
without. If reproductive elements were confined to one special part or
organ, we should be unable to explain the reproduction of lost limbs in
salamanders, and the persistent effect of intercrossing on subsequent
issue by the same mother, and the propagation of plants from shoots, or
of the begonia from minute fragments of leaves, or the development of
small pieces of water-worms into complete animals.