For cytoarchitectonic delineation of cortical areas in human brain, the Gal
lyas staining for somata with its sharp contrast between cell bodies and ne
uropil is preferable to the classical Nissl staining, the more so when an i
mage analysis system is used. This Gallyas staining, however does not appea
r to be appropriate for counting neuron numbers in pertinent brain areas, d
ue to the lack of distinct cytological features between small neurons and g
lial cells. For cell counting Nissl is preferable. In an optimal design for
cell counting at least both the Gallyas and the Nissl staining must be app
lied, the former staining for cytoarchitectural delineaton of cortical area
s and the latter for counting the number of neurons in the pertinent cortic
al areas. (C) 1999 Academic Press.