Cytologists have seldom made a quantitative study of the location of t
he nucleus within the cell. We have mathematically defined three param
eters in two dimensions: the number of distinct cell/nucleus boundary
contacts (Ncon), the length of such contacts (Lcon), and the degree of
nucleus location eccentricity (Ecc), expressed as a function of the g
reatest distance between cell and nucleus boundaries, and of the nucle
us/cytoplasm ratio (N/C). Based on the analysis of 68 circular or elon
gated cell and nucleus models, we identify 21 topographies, for a rang
e of N/C values between 20% and 70% and for a range of cell surface ar
ea values between 2900 and 6700 pixels on a hexagonal raster (256 x 25
6). Beyond 70%, a maximum of two topographies may be distinguished; be
low 20% - a figure nearer a nucleolus-nucleus ratio - the relevance an
d reliability of the approach are subject to some doubt. We conclude t
hat the problem of describing the location of one structure within ano
ther remains dependent on N/C. The method has been applied to 59 lymph
oid cells on smear photographs recognized internationally by hematolog
ists as characteristic of lymphoid diseases. The results show the robu
stness of the combination of four criteria in relation to the overall
elongation of nucleus or cell profiles, and to significant contour irr
egularities. The method has demonstrated its ability to describe and q
uantify,the eccentric position of the nucleus within plasmocytic lymph
ocytes, and provide a further degree of discrimination.