Industrial Distribution of Cases Received, Fiscal Year 1972

p. 243. Crop: the table is cut off at the bottom after "Crude petroleum and natural gas production" and continues on the following page; no footnote texts are visible in this crop, so `footnotes` is empty even though markers ^1 (caption, omitted per instructions) and ^2 (stub heading "Industrial group") are printed. A rotated running head of the facing page bleeds into the right margin of the scan; it is not part of the table. 16 columns: stub, "All cases", seven unfair-labor-practice columns (All C, CA, CB, CC, CD, CE, CP), four representation columns (All R, RC, RM, RD), then UD, AC, UC. "All cases" and the three right-hand headings span both heading levels, so they are repeated in level 2. Arithmetic check: I summed all 23 industry rows in each of the 15 numeric columns against the printed "Manufacturing" subtotal. Thirteen columns reconcile exactly (18,129; 11,589; 2,750; 330; 95; 13; 78; 6,267; 5,394; 354; 519; 86; 53). Two do not, and both anomalies are in cells I want to flag: 1. "Chemicals and allied products", column CA: I read "295". That row's components (295+99+44+6+0+4 = 448) fall 100 short of its printed All C of 548, and the CA column sums to 8,223 against the printed 8,323 — the same 100. Both discrepancies would be explained by the cell being 395. I cannot distinguish a 3 misread as 2 at this resolution from a genuine typesetting error in the original, so I transcribed the glyphs as I see them ("295") rather than correcting. Treat this cell as doubtful. 2. "Primary metal industries", column UC: the cell appears blank in the scan and is transcribed as "". It is the only empty numeric cell in the body. That row's other values (1,076 + 414 + 8 + 2 = 1,500) fall 4 short of its printed All cases of 1,504, and the UC column sums to 130 against the printed 134 — again the same 4. So a "4" was probably printed here and is lost or dropped in this scan; I did not write it in, per instructions. Treat this cell as missing rather than genuinely zero. Otherwise every row reconciles across (All C = CA+CB+CC+CD+CE+CP, and All cases = All C + All R + UD + AC + UC), including the three post-Manufacturing mining rows, which gives good confidence in the column assignment of each digit. "Miscellaneous Manufacturing" is printed with a capital M on "Manufacturing"; kept as printed. "Manufacturing" is the indented subtotal row set off by a double rule.

p. 244. This crop is a continuation page: no column headings are printed on it, so header_rows is empty. The grid is 16 columns (stub + 15 numeric), counted from the column rules and confirmed on every row. Arithmetic consistency checks all pass as transcribed: column 2 equals the sum of columns 3-8 in every row, and column 9 equals the sum of columns 10-12 in every row; the subtotal rows (Mining, Transportation/communication/other utilities, Services) and the Total, all industrial groups row also equal the sums of their member rows, column by column. Column 1 is NOT the sum of columns 2 and 9 (e.g. Mining 535 vs 407+124=531; Local passenger transportation 417 vs 303+108=411), which looks structural rather than a misprint, so nothing was adjusted. Two stray ink specks, neither treated as a footnote marker or digit: a small dot immediately right of 124 in the Mining row (column 9), and a short tick above 24 in the Motion pictures row (column 5). Both values are confirmed by the column sums. Three row labels wrap with end-of-line hyphens and were rejoined: "miscel-/laneous" (Automobile repairs...), "recrea-/tion" (Motion pictures...), "zoolog-/ical" (Museum, art galleries...). The scan's old-style 5 and 6 are easily confused (e.g. 57, 5,985, 26,852); each such reading was verified against a row or column total.