Friday, May 9, 2014

Bruce Rogers on Drop Caps

The following notes on large (or decorative) initial capital letters are taken from Bruce Rogers, Paragraphs on Printing, 1943, New York. The copyright was not (as far as I can tell) renewed, placing the work out of copyright.

If you are short of time skip to the end; figures g, h and i represent, for Bruce Rogers, “how it should be done”. The opinion in the article on whether to align with small caps on the first line may be disputed in the case that the entire first line is set in small caps. The opinion about complex shapes not being worth the trouble to do right comes from the difficulty of hand-set metal type and no longer applies to computer type. Notes in square brackets are mine.

*

Usually it is desirable to distinguish the opening of a book, or a chapter, or sometimes even of a paragraph, by inserting an initial letter larger than the type. When individual types [i.e. metal pieces of type - Liam] were more precious than they are now a wise printer would have several founts of what was called ‘titling’ type. These were often the larger letters of one of the standard faces in used in his [printing] office, case without shoulder at the bottom so that they could be easily justified with any body type. A printer, lacking this titling type, would just insert one of the large capital letters of the fount [font], but usually without cutting off th shoulder, which would damage the type for further use with its own lower case.  This naturally resulted in an undue amount of white space below the letter, and to balance this he would leave a corresponding space between the side of [p. 113] the letter and the text type in all the lines abutting on the initial, with the exception of the first line which, with the first word usually in capitals, was brought over close against it (a and c, p. 114).





This became such a convention in composition that when rectangular or other shapes of ornamental engraved initials were inserted, the white spacing between the block and the type was still exaggerated, even though the bottom of the letter aligned perfectly with one of the text lines (e).
This was known as the ‘river and bridge’ practice, when it came in for criticism at the hands pf writers on typography; but to this day many compositors do not seem to have ever encountered any condemnation of the custom, and in the work of otherwise careful printers there is an astounding indifference to the manner of fitting in an initial letter.



One of the most common errors occurs when the initial is the beginning of a word which is completed in small capitals instead of capitals.  Almost invariable the top of the [page 114: figures, inserted at references] [page 115] letter will be found aligned with the top of the small capitals, instead of being lifted slightly above them to align with the capitals of the fount. (b).
If your initial will not range [that is, align] at the bottom with any of the lines of type [because pieces of metal type are physical objects of fixed sizes] it is preferable to raise it slightly above the first line, rather than to sink it (d, f).





Of course a safe way out of these perplexities is to use an initial only a little larger than the text ans set it so that it aligns at the bottom with the first line of type, extending upward into the blank space left for the chapter heading (j).
This is a ‘modern’ rather than an ‘old-style’ fashion, and is generally more appropriate when the composition is in modern type, though in fact it does occur very early in the history of printing and was common all through the eighteenth century. Bodoni followed this fashion continually. In many of his books the word, after the capital, is continued in lower-case type.


If calligraphic or other initials of free and irregular form are employed it is impossible to give any explicit directions for their setting. They should be dealt with according to their various shapes, and your eye alone must be the final arbiter.  It is usually better to let them stand above the type page (in the Bodoni manner) rather than to leave a blank rectangle in the type for them to fill. (Illustration 63).

Because of their shapes the letters A and L are especially difficult to fit as initials. If the A is an article [that is, a stand-alone single-letter word “a”], it can have a blank rectangle to itself (g), but if it is the beginning of a word it is usually better to cut into the top of the body to permit fairly close setting of the remainder of the word (h).
It is also an advantage to letter-space the word slightly to help equalize the space between it and the initial.

The letter L is still more difficult to handle, but it it be large enough to occupy more than two lines of type it should be cut in for the first line (i), as with A (h).
If the opening sentence is a quotation it is always a question how to indicate it, especially if the initial be a rectangular decorative one. The custom in some offices is to indent the whole initial to permit the quotes to be set outside it, but this always looks a misfit. If the quotes can be put in the margin they are better so; but otherwise it is allowable to ignore the opening quotes entirely. Their absence very seldom causes any confusion to the reader.

*

Figure 63 mentioned in the text is a reproduction of a full page, with the dotted lines showing the boundary of the original page:



















5 comments:

Unknown said...

Where do you think the term 'river and bridge' came from?

Possibly the ws under the cap and the 'bridge' of text between the cap and the text body?

You do manage to find some awkward ones Liam!

Liam Quin said...

Yes, I think the bridge is the line of text at the top and the water is the bottom.

Unknown said...

Next question then. If you were to support one for say CSS adoption, which of the above styles would you support / prefer?

Quite a variation.

Unknown said...

Next question then. If you were to support one for say CSS adoption, which of the above styles would you support / prefer?

Quite a variation.

Liam Quin said...

Figures g and h (and i, which is just the same as g/h but with L instead of A) show the two variant that I feel should be supported (and that I had proposed for XSL-FO 2).

These are also supposed to be supported in the CSS inline draft, but the examples there are badly wrong.