Evidence is presented from data covering the period 1940 to 1952 which corroborates the conclusion of previous studies that prior tonsillectomy probably adversely affects the occurrence of brainstem paralysis (bulbar and bulbospinal) in poliomyelitis. Neither this study nor any preceding studies relating to this problem have proved the contention. On the basis that the hypothesis is correct, an attempt is made to find an answer by studying the incidence of the common paralysis of cranial nerves (VII, IX and X and XI) in bulbar and bulbospinal cases of poliomyelitis on the basis of presence or absence of tonsils. Rates of incidence of paralysis of cranial nerves, not adjusted for age, indicate a decidedly higher proportion of paralysis of the facial nerve (VII) among nontonsillectomized patients whereas tonsillectomized persons are preportionately more affected by palatal and pharyngeal paralysis (nerves IX and X). Paralysis of the facial nerve appears from two studies to occur more commonly at earlier ages, particularly in the age group 0 to 4 years. However, age adjustment did not erase, although it did somewhat lower, the TR/TP ratio. This finding lends credence to a real difference but can only be applied to this study, as Paffenbarger in a smaller study found no significant difference in frequencies of paralysis of the facial nerve between groups with tonsils removed and tonsils present, and Southcott, also in a small study, found paralysis of the facial nerve more common among tonsillectomized patients with bulbar (includes bulbospinal) involvement. The differences noted for palatal and pharyngeal paralyses (nerves IX and X) in the unadjusted rates as between tonsillectomized and nontonsillectomized patients remain statistically different and in some instances significant when corrections for age are made. The results of this study are suggestive but give no entirely satisfactory explanation for the differences noted. Various explanations previously offered are cited and briefly discussed. Perhaps more definitive studies in animals along the approach suggested by Southcott will prove more fruitful, namely, labelling virus by some radioactive element in order to trace the route it takes to the central nervous system.