&NA;This study evaluated the changes in pathways of cerebrospinal fluid (CSF) outflow that accompanied acute and compensated hydrocephalus in the rabbit. Intraventricularly injected99mTc antimony sulfide was used as a tracer of outflow pathways, and specified structures were counted 12 to 24 hours after injection. Fifteen rabbits were divided into three groups: 1) an acutely hydrocephalic group in which 3 cisternal injections of blood were followed by a study of CSF pressure, ventricular size, and CSF outflow pathways 1 week after the last injection; 2) a control group treated according to the same protocol, except that sterile saline was injected instead of blood; and 3) a chronic group also treated according to the same protocol but in which the animals were maintained an average of 4 weeks after the last blood injection. Ventricular size was measured by computed digitation and expressed as an area ratio of ventricle to brain (VBR). In control animals, 11.8% of the injected colloid dosage was found in cranial perineural lymphatic channels, and 4.8% appeared in the spinal cord. The mean CSF pressure was 149 ± 20.2 mm H2O (mean ± SE) and the mean VBR was 0.040 ± 0.003. In animals evaluated 1 week after subarachnoid injection, accessory cranial perineural lymphatic outflow decreased significantly to 3.4%, and spinal cord activity increased to 9.8% (P< 0.05, two‐tailedt‐test). These animals were hydrocephalic and had CSF pressure of 247 ± 25.1 mm H2O (mean ± SE) and VBR of 0.083 ± 0.009. By 4 weeks, the percentage of activity in both the cranial perineural lymphatic channels (7.4%) and spinal cord (5.9%) had returned toward normal, as had the CSF pressure (154 ± 19.3 mm H2O) and VBR (0.045 ± 0.004). These data suggest that so‐called “accessory” spinal cord pathways are important in clearance of particulate matter and CSF after subarachnoid hemorrhage. In acute hydrocephalus, more CSF is shunted into the spinal cord and less escapes via the cranial lymphatic channels; with compensation of hydrocephalus, these both return toward control values. (Neurosurgery24:187‐192, 1989)