PATHOLOGIC CHANGES IN BRAIN CAUSED BY ALPHA PARTICLES 407 



the oranule cells of the cerebellum is not clear to iis. A further point is that 

 with the passage of time the tissue abo\"e the zone of damage in both cerebel- 

 lum and cerebrum underwent profoimd atrophy i Figs. 1 and 2 ) . Measure- 

 ments indicated that reduction in the width of the irradiated part of the 

 cerebellum was relatively somewhat greater than that in the cerebrum 

 'Table II i. This does not necessarily mean that the cerebellum was rela- 

 tively more radio\ulnerable. for the cerebellum underwent greater shrinkage 

 than the cerebrum during processing of the brain. 



Nei*\"e cells and neuroglia of the cerebral cortex appeared equally vulner- 

 able at early stages, as indicated at the 6,000 rad surface dose levels by 

 necrosis of isolated nerve cells and glia (Figs. 7 and 8), and at subsequent 

 stages, as indicated by necrosis of astrocytes and nerve cells (Fig. IIA). In 

 the cerebellum the granule cell seemed the most radiovulnerable. Depending 

 on dosage and time, \arying numbers of these cells in the more intensely 

 irradiated part of the granular layer subsequently underwent necrosis. 

 Purkinje cells situated at comparable levels in the cerebellum also suffered, 

 but the tempo at which the damage occurred seemed slower than that in 

 granule cells. Purkinje cells .sometimes appeared intact at a time when 

 granide cells were pyknotic (Fig. 7A). Since at later stages following irradi- 

 ation the width of the zone of Purkinje cell damage was often not much dif- 

 ferent from that of the pyknotic granule cells (Figs. lA, 2A, 3 A, and 12A), 

 it was concluded that relatively little difference in radiovulnerability of these 

 two cell types existed. Less radio\ulnerable than granule and Purkinje cells 

 were Bergmann cells, nerve cells and glia of the molecular layer, and Golgi 

 cells in the granular layer, in that order of decreasing radiovulnerability. At 

 fairly early stages, damage was also incurred by intrafolial white matter, and 

 the reacti\ ity of its glial cells was about the same as that elsewhere. 



E.xamination of material exposed to the 1,500 rad surface indicated that 

 nerve cells had been destroyed and that blood \essels were not morpho- 

 logically altered (Fig. 4). The same was true at the 3,000 rad level at 20 

 days when a zone of cytologic damage was encountered. At 6,000 rad, nerve 

 cell and glial necrosis occurred as early as 6 hours after irradiation, but it 

 was not until 48 hours that vasodilatation as brought out by the Pickworth- 

 Lepehne method appeared and not until 60 hours that vasodilatation was 

 concentrated in the '"Bragg zone."" Clearly, ner\e cell and glial damage oc- 

 curred before circulatoiy distinbances sufficient to cause vasodilatation were 

 e\iclent. Moreover, only at 48 hours did the blood-brain barrier become per- 

 meable to FLSP, indicating that during the preceding hours any vascular 

 change that might ha\e occurred was not functionally evident. Further, the 

 presence of hemorrhages, taken as e\idence of vascular damage, did not 

 occur until the 3rd day alter irradiation. The conclusion seems inescapable 

 that, in earlier stages at least, nerve cells and glia were primarily damaged 

 by particle radiation. Damage of nerve cells and glia concurrently is most 



