It is interesting to note some of these comparisons. Goddard, for example, considers it conservative to estimate that 2% of the school population is “feeble-minded” (_112_, p. 6). In the same publication he says: “There are between 300,000 and 400,000 feeble-minded persons in the United States” (p. 582). Since the elementary school enrollment is about 20,000,000 (_208_), the feeble-minded school children alone on his first estimate would account for 400,000 feeble-minded in the United States without allowing for any feeble-minded outside of the ages in the elementary school. The report of the British Royal Commission, published in 1908, forms the starting point for many of the estimates made today. The commission added together the number of school children which were thought to require special classes with the number of defectives found in institutions, prisons and almshouses, or reported by its medical investigators. The total gave 0.46% of the general population as “mentally defective persons,” not including certified lunatics. From this amount should be deducted .06% who were insane but had not been certified as such, leaving 0.4% mentally deficient. This was not regarded by the Commission as an estimate, but was the number actually “enumerated by the medical investigators” in sixteen typical districts studied in England and Wales with a total population of 2,362,222 (_83_, VIII, p. 192). Turning to the school children we find that in the areas investigated there were 436,833 school children of whom 0.79% were found defective. Since this was an enumeration and not an estimate, the commission paid no attention to the discrepancy between 0.79% of the school children and 0.31% of the rest of the population. Tredgold, moreover, based his estimates of the frequency of the mental deficiency in England and Wales on the data of the Royal Commission without attempting to harmonize this discrepancy. This oversight has apparently been one source of the not uncommon difference between the estimates for school children and for the general population. One suspects that the fact that the elementary school population is about a fifth of the general population, has also mistakenly contributed to this error. The discrepancy of three to five times as large a frequency of deficiency among school children as in the general population certainly needs clearing up.