The only mistake of any real importance refers to the large numbers of the Israelites, who are said to have left Egypt,--some 600,000 men, besides children, or probably over two million altogether. For on two subsequent occasions, when the census of the tribes is given, it totals up to about the same number.[102] This is no doubt a serious difficulty; as anyone can see, who will take the trouble to calculate the space they would require on the march, or in camp. If we assume, for instance, that they crossed the arm of the Red Sea in, say, _forty_ parallel columns, these would still have to be of enormous length to contain 50,000 persons each, with their flocks and herds. [Footnote 102: Exod. 12. 37. Num. 1. 26.] Perhaps the best explanation is that suggested by Professor Flinders Petrie, that the word translated _thousands_ should be _families_,[103] so that the tribe of Reuben, for instance,[104] instead of having forty-six _thousand_ five hundred men, would have forty-six _families_, (making about) five hundred men. The chief arguments in favour of this are, first, that the same word is used in Judges 6. 15, where it so obviously means family and not thousand, that it is so translated in both the Authorised and Revised Versions. [Footnote 103: Egypt and Israel, 1911, p. 43.] [Footnote 104: Num. 1. 21.] And secondly, it would account for the remarkable fact that though there were twelve tribes, and they were each counted twice, yet the number of the hundreds is never 0, 1, 8 or 9; but always one of the other six digits. It is extremely unlikely (practically incredible)[105] that this would occur in an ordinary census, but the proposed theory explains it at once. For the hundreds could scarcely be 0, or 1, as this would mean too few men in a family; or 8 or 9, which would mean too many; while the other digits always work out to what (allowing for servants) is a reasonable proportion, from 5 to 17. On this theory the number of men would be reduced to 5,600, which is much more intelligible. But some other passages scarcely seem capable of this interpretation, so it must be admitted that the number forms a difficulty, whatever view we adopt. [Footnote 105: The chance of its occurring would be only (6/10)^24 or less than 1 in 200,000.] (3.) _Modern discoveries._ Lastly, as to the effect of modern discoveries on the accuracy of the Old Testament. In the case of the Pentateuch, as we have seen, there is very little _direct_ evidence either way; but it is different in regard to some of the later books.