After establishing a displacement distance, an additional correlation could 
be shown between this distance and the animal’s weight.^ This correlation 
suggested that there is a tendency for brain structures to be posterior to the 
average in animals which have a body weight less than the mean, whereas the 
opposite tendency is present in animals in which the body weight is greater than 
the mean. 
Differences among animals in the position and size of the external auditory 
meati are apparently a major cause of the stereotaxic variabilities that have been 
described. This conclusion is based on measurements in which other cranial 
loci were used to obtain reference planes. Table II lists useful landmarks which 
can readily be determined by X-ray examination. 
Table II— AVERAGE STEREOTAXIC COORDINATES OF CALVARIAL 
STRUCTURES 
Structure 
Anterior- 
posterior 
Horizontal 
Millimeters 
29±2 
20 
20 
9±2 
8. 5±0. 5 
5±2 
—22 ±2 
Millimeters 
17±1 
14. 5±0. 5 
10. 5±0. 5 
0. 5±0. 5 
5.5±1.0 
31±2 
15±2 
Anterior edge of posterior clinoid process 
Posterior limit of cranial vault 
A good approximation to the horizontal zero plane is provided by a line pass- 
ing through the orbitale and just below the floor of the temporal fossa. The 
rostral boundary of the posterior clinoid processes falls at 8.5 ±0.5 mm. anterior 
to the ear canal (AP 0) of the average animal, while at 20 mm. anterior, the floor 
of the olfactory groove and orbital roof should lie at 10.5±0.5 mm. and 14.5±0.5 
mm. above the horizontal zero. Figure 6 is a scatter diagram for 8 animals in 
which these skull landmarks were used to establish a stereotaxic frame of refer- 
ence independent of the ear canals. Animals selected for this drawing included 
those with a wide deviation from the mean body weight. For comparison, the 
coordinates of SMI 2 are indicated by the heavier lines and the cross hatching. 
Connecting lines have been drawn through the centers of the structures of each 
animal to show angular relationships and the relative positions of these structures 
“ For the 29 animals with complete measurements on the position of all three structures, the 
regression line determined by the equation 
“displacement” (mm) =2.75 — 3.41 Xweight(kg) 
is significant (0.01 confidence level, T=3.47 with 27 degrees of freedom). Weight variations, 
however, in any one animal in our laboratory may be as much as 200 gms.; it should also be 
noted that after assuming this weight correction, displacements greater than 0.5 mm. would 
still be needed in 12 (41 % ) animals to obtain the dotted curves in figures 3—5. 
for any one animal. The range of .scatter in this figure for the Corpus mam- 
millare (M) and Commissura anterior (CoA) is approximately 1.5 mm. and is 
thus less than one half that shown in figure 3—i. The fact that coordinates for 
the Nucleus habenularis (Ha) show a wider scatter than the other structures 
suggests that cerebral distortion attributable to embryological development of the 
mesencephalic flexure contributes to a greater variation in the position of dorso- 
caudal thalamic structures than of those lying more rostral and ventral. 
Fisure 6 — -Scatter diagram of the parasagittal location of the Commissura anterior (CoA), Corpus 
mammillare (M), and Nucleus habenularis (Ha) in 8 animals in which the stereotaxic reference 
planes were determined by lateral skull X-ray without reference to the position of the ear 
canals (see table II and text). The cross-hatched areas show in addition the position of these 
structures in animal SM 12 used for the atlas. Lines are drawn between the approximate 
centers of these structures for each animal. 
In summary, the findings indicate that in carrying out stereotaxic exploration 
under routine experimental conditions one would have a 50% chance of coming 
within 0.5 mm. of a point described by the coordinates of the present atlas. 
There would be a 90% chance of falling within 1.5 mm. of this target. In 
experiments in which a higher degree of accuracy is required it would be desirable 
to obtain appropriate X-ray data and to avoid the use of unusually large or 
small animals. In our experience the degree of accuracy that can be achieved in 
stereotaxic work in the squirrel monkey compares to that in the cat, and is 
significantly better than in the macaque. 
5 
