Final Outcome of Representation Elections in Cases Closed, Transition Quarter (July–September 1976)

p. 261. Footnote markers are present but no footnote text is printed on this crop: a raised 1 on the caption and a raised 2 on the "Total elections" column heading. The table is panel A only ("A. All representation elections"); a part B presumably continues on the following page, so the footnote block is likely printed there. Scan quality: the printed periods in the "Percent won" column drop out intermittently (e.g. the image shows "39 0", "66 7", "100 0", "45 5" while "44. 4", "63.0", "83.3", "79.6", "0.0" keep a visible point). Every value in that column is a one-decimal percentage and each one is consistent with Total won / Total elections, so I transcribed them all with the decimal point. The same loss affects the section line, which I read as "A. All representation elections" (the period after the A is not legible in the scan). Doubtful cells: in row "AFL-CIO v national", column "Employees eligible to vote / In units won by / AFL-CIO unions", the glyph is badly broken and could be read as a stray mark; I transcribed it as 0, which is what the row requires (AFL-CIO won 0 of these elections and all 333 employees in units won went to the other national union). In row "Teamsters v local", the stub-column figure "12" has a damaged second digit; 12 is confirmed by 10/12 = 83.3 percent. Stray ink: a small hand-drawn-looking mark sits above the "Elections won by unions" spanner rule, and another just left of "121,894" in the "1-union elections" row. Both are scan/page artifacts, not characters, and are not transcribed. Arithmetic check: every subtotal and the grand total foot exactly across all fifteen value columns (1-union + 2-union + 3-or-more = Total representation elections, and each component column sums to its "Total won"/"In elections won" column). No misprints detected.

p. 262. Continuation page: the table has 16 columns (stub + 15 data columns), but the individual column headings are printed on the preceding page and do not appear on this crop. The only heading printed here is the panel heading over all 15 data columns, so it is repeated across them; the stub has no heading on this image. Column 2 is a percent column. The 1-bit scan has dropped the decimal points, so the cells read "47 3", "100 0", "0 0", etc. I have restored them as 47.3, 100.0, 0.0: every value equals column 3 divided by column 1 to one decimal (877/1,874 = 46.8; 2/3 = 66.7; 985/2,013 = 48.9; etc.), so the space is a lost decimal point, not a printed gap. The same loss of fine points affects the panel letter and the "v" (versus) in the row labels: on the image they read "B Elections in RC cases" and e.g. "AFL-CIO v Teamsters" with no visible period. I have written the panel heading with the period ("B."), consistent with the restored decimal points, but left "v" bare since no arithmetic confirms it; the original almost certainly prints "v." in each of those labels. Two row labels are hyphen-broken in print and joined here: "AFL-CIO v AFL-CIO v AFL- / CIO" and "AFL-CIO v AFL-CIO v Team- / sters". In the "Other local unions" row, column 11 shows a short em-dash before the usual leader dashes; it is a blank-cell rule and is transcribed as "". All internal arithmetic reconciles exactly, which independently confirms the column each value sits in: within every block, columns 4-7 sum to column 3, columns 3 + 8 sum to column 1, columns 11-14 sum to column 10, columns 10 + 15 sum to column 9, and the three subtotal rows sum to the "Total RC elections" row. No printed value appears to be a misprint.

p. 263. 16 columns; this is the continuation page of Table 13, carrying panels C and D only. Header structure: the horizontal rule under "Employees eligible to vote" and the one under "In units won by" both appear to stop at the vertical rule immediately left of the last "Other local unions" column, which is printed as a full-height heading box. I transcribed the spans as printed (hence "Other local unions" repeated down all three levels rather than nested). By the data this column is clearly the "other local unions" member of the "In units won by" group under "Employees eligible to vote" (e.g. the RM "Other local unions" row puts 78 there), so the short rules look like a typesetting slip rather than a different meaning. Decimal points: this is a coarse 1-bit scan and most periods have dropped out. In the "Percent won" column only two cells show the point clearly ("13.2" for RD Teamsters and "100.0" for RM 2-union elections); the rest print as a gap ("28 6", "13 0", "0 0", "100 0", "22 1", "25 6", "14 6", "13 2", "16 7"). I transcribed all of them as decimals, which every check sum confirms (e.g. 23/90 = 25.6, 29/174 = 16.7). The same loss affects the period after the section letters "C"/"D" and after the "v" in "AFL-CIO v. local" / "v. Teamsters" / "v. national" (the RM "AFL-CIO v. AFL-CIO" row does print its period); I supplied those periods. "Total RD elections", column "In units won by / AFL-CIO unions": printed "3 037" with the comma lost; transcribed 3,037, which equals 1,536 + 1,501 from the rows above. Dash rules ("------") in empty cells transcribed as "". Every row and column total in both panels was checked against its components and all of them add up as printed; no misprints found in the numbers. Footnote 2: the two commas after "was made" and before "for example" are as scanned; the first may be a semicolon whose upper dot did not survive.