All that is said here contains a high level of speculation! Hence, the optimal H & W genotypes in terms of longevity may differ among genotypes so that in the estimation of H, W and M, the relationship may have great dispersion due to a heterogeneous society, and could even form clusters per genotype. Obviously, we can get BMI on the ordinate and M on the abscissa to satisfy this equation per population--if we try ! Note that both males and females are accomodated by BMI. Try H versus M.