Industrial distribution of collective bargaining elections, winner, eligible voters, and valid votes cast, during the fiscal year 1949

p. 190. CROP: the table is cut off at the bottom after "Miscellaneous manufacturing"; the non-manufacturing industry groups, any total lines below them, and the text of footnote 1 (whose marker appears on the stub heading "Industrial group 1") are not on this image, so `footnotes` is empty. DECIMAL POINTS: every Percent column is printed to one decimal place, but in this 1-bit scan most decimal points have dropped out, so the cells read "100 0", "37 9", "66 9" etc. A number of points do survive ("5.1", "1.1", "3.5", "2.2", "1.0", "34.9", ".5", ".1", ".0"), confirming the format. I have written the decimal point in every Percent cell. No digit was changed, added or removed, and nothing was computed or rounded. Likewise, percents below 1 are set with a leading point and no leading zero; where that point did not survive I restored it, in these cells: Rubber products / Elections Percent (.8); Aircraft and parts / Elections Percent (.6); Professional, scientific, and controlling instruments / Elections Percent (.9) and Eligible voters Percent (.7); Printing, publishing, and allied industries / Eligible voters Percent (.9); Tobacco manufacturers / Eligible voters Percent (.5), Valid votes cast Percent (.5), Elections Percent (.1). CROSS-CHECK: in all 25 rows the four Winner "Number" cells sum exactly to that row's Elections "Number" cell (e.g. Total 2,092+858+939+1,625 = 5,514; Machinery 69+77+187+120 = 453). The Number columns are therefore well supported. DOUBTFUL CELLS: - "Apparel and other finished products made from fabrics and similar materials", Elections Percent: reads 1.8, but 96/5,514 = 1.74, which would print as 1.7; every other row in that column rounds correctly. Either a misprint or a 7 misread as 8. Transcribed as it appears (1.8). - "Tobacco manufacturers", Eligible voters Percent: reads .5, but 2,623/588,761 = 0.45, which would print as .4. The glyph reads as a 5; transcribed as it appears. - "Paper and allied products", Valid votes cast Percent (3.7 vs computed 3.75) and "Machinery (except electrical)", Valid votes cast Percent (12.2 vs computed 12.26) differ from a half-up rounding by one in the last place; transcribed as printed. HEADINGS: the C.I.O. heading prints as "C I. O" — the periods after the C and after the final O appear to have been lost in the scan; transcribed as it reads. A small stray ink mark sits in the header band between "Winner" and "Eligible voters"; it is a scan artifact, not a heading. The stub heading "Industrial group 1" carries a superscript footnote marker and spans all three heading levels, as do the Number/Percent pairs under Elections, Eligible voters and Valid votes cast (those headings sit at the second level and span down to the third). STRUCTURE: rows set off by double rules in the original are "Transportation equipment" (a subtotal whose components are the three indented rows that follow: Aircraft and parts; Ship and boat building and repairing; Automotive and other transportation equipment) and the return to industry-group level at "Professional, scientific, and controlling instruments". All rows carry values; there are no label-only section lines. The vertical text at the right edge is the running head (Appendix A: Statistical Tables Covering the Fiscal Year 1949) and was excluded.

p. 191. Crop is a continuation page: no column headings are printed, so header_rows is empty. The grid has a stub plus 14 numeric columns (7 count/percent pairs), counted from the column rules. Decimal points: in this 1-bit scan the period in the one-decimal percent columns is frequently broken away and reads as a space ("43 4", "1 5", "12 0", "100 0"). I restored the decimal point in those cells. The column assignment and these readings are corroborated throughout: in every row the four count columns (cols 3, 5, 7, 9) sum exactly to the case total in col 1, each adjacent percent equals count/total, the seven transportation detail rows sum to the "Transportation, communication, and other public utilities" line in all count and employee columns, and the four mining detail rows sum to the "Mining" line. Cells transcribed as "(2)" (Coal mining, cols 2, 12 and 14) contain only the raised parenthesized footnote reference 2 ("Less than 0.1 percent"); there is no value printed in them. Doubtful cell: "Finance, insurance, and real estate", col 4 — I read the second digit as 4 (43.4), but 13/30 = 43.3, so this is either my misreading of a degraded digit or a misprint. Every other percent in the table agrees with its count to the rounded tenth. Stray marks treated as scan speckle / leader-dot remnants rather than decimal points, confirmed by arithmetic that the cells are integers: a dot precedes "0" in "Crude petroleum and natural gas production" col 7, "9" in "Nonmetallic mining and quarrying" col 7 (9/46 = 19.6, matching col 8), and "24,768" in "Communication" col 11 (the seven transportation employee figures sum to 58,286 with this value). Rows above "Mining" belong to the earlier part of the table on the preceding page and are outside this crop. Double horizontal rules separate the section groups (after the mining detail, after "Finance, insurance, and real estate", after the transportation detail, and above "Services"); these are rules, not rows, so they are not transcribed. Two or three "Wholesale trade"/"Retail trade" percent cells also show the faded-period pattern (col 2 "8 0", "9 1"); read as 8.0 and 9.1.