We cut a loaf into slices of 12 mm and counted the holes in each. The holes are not distributed evenly. There are fewer of them near the top of the loaf and the ones that are there are larger, a result we did not expect and have now reproduced eleven times. This paper reports the count, the areas, and the two occasions on which the keepers disagreed about whether a hole was one hole or two.
The loaf was cooled for two hours, which is longer than anyone wanted to wait. Slices were cut with a serrated knife using a guide, photographed against graph paper at 1:1, and the holes traced by hand with a pencil. Tracing by hand was slower than the automatic method and produced a different answer, so the automatic method was abandoned.
A mean of 52 holes per slice, with a mean area of 0.36 cm². The largest single hole measured 2.4 cm² and was located 14 mm from the top crust. Holes larger than 1 cm² account for 9 % of the count and 34 % of the void area.
Two keepers disagreed on 17 of 1,240 holes. Both were correct; a hole that has been cut through twice is genuinely two questions.