I was casting some type this week when I noticed an oddity on the font schemes listed in the back of the The Monotype Casting Machine Manual, which is the basis for my usual casting font schemes. The font scheme gives the count of how many of each sort are cast to make up a font, and these were based on English text from around the early twentieth century.
Although some foundries would sell type with uppercase, lowercase, points (punctuation), and figures each being packaged separately, allowing the customer to make their own upper-/lowercase balance, the Monotype schemes are for a complete set including both upper- and lowercase as well as figures and points. There are two basic schemes: The ‘body’ scheme represents the letter frequencies encountered when composing running text in full sentences and paragraphs. The ‘jobber’ scheme represents the frequencies encountered when composing titles, headings, point-form lists, and other such items which generally require many more uppercase letters.
If you compare the two classes of schemes, it is generally the case that the counts of lowercase are the same, but in the Jobber schemes, the uppercase count if increased by about 2 less than 0.28 times the count of the corresponding lowercase. So, for instance with the letter ‘e’, either scheme has 118 lowercase, the body scheme has 12 uppercase, and the jobber scheme has 44 uppercase, which is close to 12+0.28×118-2 (equaling about 43).
That 0.28 figure is very vaguely in the ballpark of the reciprocal of the average word length (5 letters), and so what you would expect for titles, where pretty much every word is capitalized.
For the rarer letters, this formula gets fudged a bit, with uppercase ‘Z’ having 4 sorts in either case, and ‘F’ having 6 in body and 10 in jobber, but with 24 lowercase, the formula yields 13.
Some other letters are one or two extra uppercase away from the formula as well.
The real outlier, though, is M, where the jobber scheme actually has 4 fewer than the body scheme, 14 in jobber, 18 in body. It seems that the inconsistency is in the count in the body scheme, where, for instance M far outnumbers L, C, or U (6, 8, and 6 respectively) which have similar lowercase counts (14 or 16).
It is not clear if this is a typo in the table itself, showing 18 instead of 8, or whether they did not, in fact use general English text as their sample, but instead their own books and literature, which would be peppered with the capitalized “Monotype” trademark throughout!
As an aside, my earlier remark about using “early twentieth century” text as a sample is reflected in the figure frequencies, where 1 and 9 are heavily represented, as one would expect when usually referring to years in the 20th century. I should probably modify the schemes I use to add more 2’s and 0’s to better reflect the years one might now be mentioning.
Leave a Reply