In early embryonic mosaics of Drosophila, resulting for example by the
loss of a particular chromosome in early cleavage, the clonal boundaries
form essentially randomly (Hotta and Benzer, 1972, Nature 240:527).
In the blasotderm stage embryo, the ectoderm,
which will give rise to the larval exoskeleton and subsequently of the adult,
arises from only the lateral portions blastoderm, while the ventral region
gives rise to mesoderm and the dorsal region to "extra-embryonic" amnion-
serosa. Thus, any clonal boundary falling in the dorsal or ventral regions,
which happens randomly essentially as a function of the space taken up by
those regions, will give rise to a bilaterally different exoskeleton. Thus,
a much higher frequency of mosaics whose exoskeleton is genetically different
a the midline than any other single place in the exoskeleton.
I am not really sure how this applies to bilaterally assymtric lobster, since
they do not undergo a similar syncytial cleavage pattern. It is probably
relevant, though, that in crustacean embryos, the mesoderm also arises
ventrally and the ectoderm laterally.