Distribution of Unfair Labor Practice Situations Received, by Number of Employees in Establishments, Fiscal Year 1986

p. 278. CROP: the table is cut off at the bottom — the last printed row visible is "3,000–3,999" (cumulative percent 97.2, so roughly 2.8 percent of situations remain in size classes below the crop). No footnote texts are visible on this image, so `footnotes` is empty even though the caption carries a raised 1 and the Total row's "30,326" carries a raised 2 (transcribed as 30,326^2). STRUCTURE: 22 columns. "Total number of situations" is a tall standalone stub-side column; the spanner "Total" covers only the two columns "Percent of all situations" and "Cumulative percent of all situations"; "Type of situations" covers the remaining 18 (nine two-column type groups). In the Total row the cumulative-percent cell holds a dash rule, transcribed as "". DECIMAL POINTS: the scan renders the decimal point in the percentage columns as a faint gap ("27 7"), but the cumulative column verifies them as decimals (27.7 + 8.7 = 36.4, + 6.7 = 43.1, and so on through 97.2), so all percentages are written with a decimal point. Likewise the thousands commas in the stub labels "2,000–2,999" and "3,000–3,999" and in "1,072" are faint but present. VERIFICATION: every data row's nine type-of-situation counts sum exactly to its "Total number of situations", and every "Percent of all situations" value matches count/30,326, so the integer counts are solid. PRINTED PERCENTAGES THAT DO NOT MATCH THE COUNTS (transcribed as printed, not corrected). Since the counts are confirmed by row sums, these are either misprints in the source or faint-digit misreads; the affected cell is the percent, not the count: - "Under 10", CC percent by size class: printed 42.2; 414/985 = 42.0. - "Under 10", CD percent by size class: printed 40.4; 93/231 = 40.3. - "Under 10", CA–CB percent by size class: printed 24.1; 317/1,311 = 24.2. - "10–19", CD percent by size class: printed 10.5; 24/231 = 10.4. - "100–109", CC percent by size class: printed 5.5; 53/985 = 5.4. - "200–299", CA–CB percent by size class: printed 6.5; 87/1,311 = 6.6. - "700–799", CG percent by size class: printed 28.1; 6/21 = 28.6. This is the largest gap and the least certain reading — the trailing digit could be a 6 rather than a 1. Two further patterns look like the source's own rounding convention rather than errors, since they recur consistently: a CE count of 1 is always printed as 1.8 (1/58 = 1.7), and an Other-C count of 3 is always printed as 2.2 (3/140 = 2.1).

p. 279. This image is the tail of Table 18, continued from the previous page: only the last three size classes are printed, so the earlier size-class rows and any total row are not on this crop. The column headings are reprinted in full on this page, so header_rows is filled. Decimal points: the 1-bit scan has dropped nearly all small dots (note the dotless i's throughout "situations"), and the percent values therefore read as e.g. "0 7", "97 9", "100 0" with a gap where the decimal point is. I have transcribed them with the decimal point restored (0.7, 97.9, 100.0), since every value in those columns is a one-decimal percentage. No digits were altered. Footnote marker 2 is defined under the table but its reference does not appear anywhere on this crop; it is attached to a cell or heading on the first page of the table. Footnote 1 is referenced in the caption ("Fiscal Year 1986^1"), which is excluded per instructions. Both footnote strings end without a visible period, again likely a casualty of the dot loss described above. Column-position check: in the "5,000–9,999" row the lone 1's sit under CD "Number of situations" (paired with 0.4), CG "Number of situations" (paired with 4.8), and "Other C combinations" "Number of situations" (paired with 0.7); CE and CP are 0/0.0 for all three rows. The dash in "CA-CB combinations" is printed as a short rule; I have written it as a plain hyphen. The row-label ranges use an en dash as printed.