I'd like to add my own bias about this issue of nicotine smell in mother's
milk.  I don't have the reference, but I am quite confident that I've read
that mothers who smoke have very low breastfeeding rates.  I would imagine
that very few of the children of smoking mothers ever breastfed.  I
bet that if they tried to do a prospective, or even a retrospective,
study to see if the breastfed children of smokers had any different
smoking rates than the bottle-fed ones, they couldn't find a high
enough number of breastfed children to give their finding statistical
power.

We do know (sorry no reference immediately handy on this either) that
there is a definite increased risk of taking up smoking in children whose
parents smoke.  These kids aren't just exposed to parental behaviors which
may influence them.  If we want to postulate the role of passive intake of
nicotine and many other constituents of cigarette smoke, there's all that
second-hand smoke (the passive smoker has elevated urinary cotinine
levels, as though he smokes too), which they are exposed to for years and
years.  All these other exposures are such confounding factors that it
would be hard to separate (unless, of course, that minuscule subset of
smoking mothers who breastfeed are also those who take special care with
avoiding second-hand exposure for their children and provide a more
positive parenting and role modeling for their children - I wouldn't be
surprised to actually find a lower rate of smoking in breastfed children
of smoking mothers).

Mother's milk, IMHO, is the *least* likely cause of smoking in
these kids, because so few of them are breastfed and, if breastfed at all,
for a very short period as compared with their other passive exposures.
This is an example of research which may show a very definite finding but
whose results are open to ridiculous interpretations.  Sometimes I wish
people would refrain from doing research of this sort, although I
understand the scientific curiosity behind it.  Sigh, Alicia Dermer, MD,
IBCLC.