Seventy-eight cases of late-onset, non-communicating hydrocephalus were ope
rated upon by third ventriculocisternostomy. The distortion in the anatomy
of the dilated third ventricular floor dictated the selection of the target
area. The optimal site for the perforation was the translucent, bluish and
thinned out part of the floor. This was variable and in 76.9% not in the m
idline with more than one fenestra done in 35.9%. The size of the ventricul
ocisternostomy needed not be around 5 mm. Smaller sized openings in a taut
floor (60.3 %) served the same purpose as bigger ones in a redundant area (
39.7%). The success of the procedure could be predicted from the profuse do
wnward flow of cerebrospinal fluid through the perforation, "Whirl Sign". A
n acceptable assurance of our results was confirmed both clinically and rad
iologically. The outcome in our series had four grades, namely cured in 78.
2%, ameliorated, but still needed diversion, in 16.7%, status quo in 2.5%,
and complicated in 2.5%.