Indexes are not confined to proper names, and therefore it is necessary to add some calculations as to the proportions of the several letters in indexes of subjects. The following table is formed from three large indexes, each different in character. I. represents Gough's Index to the Publications of the Parker Society, which may be taken as a very good standard index. The subjects are very varied, and there are no specially long headings; it also contains proper names as well as subjects. II. represents an index of subjects in Civil Engineering which contains a good number of large headings. III. represents the index to the Minutes of a public board, and also contains a considerable proportion of large headings. It will be seen that the numbers vary so considerably as to be of very little practical value. The percentages are, I think, interesting, but they show conclusively that indexes will vary so considerably that in order to obtain a satisfactory percentage a separate calculation will have to be made in each case. Large headings will vitiate any average; in fact, I have lately had to do with an index in which R was the largest letter, on account of such extensive headings as Railways and Roads.
One striking point in the averages is that B is found to be displaced from the pre-eminent position it occupies in the percentages of proper names.
I. II. III. A 10·67 2·63 5·58 B 6·94 5·07 6·28 C 15·63 8·26 8·84 D 2·48 4·50 4·65 E 3·23 6·94 11·39 F 2·85 3·38 1·63 G 4·34 3·56 1·86
H 4·34 3·19 2·09 I 1·74 2·72 1·39 J 3·97 0·14 0·46 K 0·74 0·05 0·23 L 5·58 4·97 15·12 M 5·71 5·82 7·67 N 1·37 0·19 0·93 O 1·74 1·31 1·63 P 9·31 6·75 7·67 Q 0·12 0·94 0·47 R 2·48 12·38 8·14 S 8·44 13·32 8·14 T 3·60 5·72 1·40 U 0·50 0·05 0·47 V 0·99 0·61 2·33 W 2·61 7·41 1·51 X 0·03 0·00 0·00 Y 0·22 0·00 0·00 Z 0·37 0·09 0·06 ——— ——— ——— 100·00 100·00 100·00
| I. | II. | III. | |
| A | 10·67 | 2·63 | 5·58 |
| B | 6·94 | 5·07 | 6·28 |
| C | 15·63 | 8·26 | 8·84 |
| D | 2·48 | 4·50 | 4·65 |
| E | 3·23 | 6·94 | 11·39 |
| F | 2·85 | 3·38 | 1·63 |
| G | 4·34 | 3·56 | 1·86 |
| H | 4·34 | 3·19 | 2·09 |
| I | 1·74 | 2·72 | 1·39 |
| J | 3·97 | 0·14 | 0·46 |
| K | 0·74 | 0·05 | 0·23 |
| L | 5·58 | 4·97 | 15·12 |
| M | 5·71 | 5·82 | 7·67 |
| N | 1·37 | 0·19 | 0·93 |
| O | 1·74 | 1·31 | 1·63 |
| P | 9·31 | 6·75 | 7·67 |
| Q | 0·12 | 0·94 | 0·47 |
| R | 2·48 | 12·38 | 8·14 |
| S | 8·44 | 13·32 | 8·14 |
| T | 3·60 | 5·72 | 1·40 |
| U | 0·50 | 0·05 | 0·47 |
| V | 0·99 | 0·61 | 2·33 |
| W | 2·61 | 7·41 | 1·51 |
| X | 0·03 | 0·00 | 0·00 |
| Y | 0·22 | 0·00 | 0·00 |
| Z | 0·37 | 0·09 | 0·06 |
| ——— | ——— | ——— | |
| 100·00 | 100·00 | 100·00 |
When the whole index is pasted down it is not yet ready for the printer, as it will require to be marked for the instruction of the compositor. The printer will have general instructions as to the kind of type to be used and the plan to be adopted, but it will be necessary to mark out those words that are not to be repeated and to insert lines indicating repetition. There are also sure to be little alterations in wording, necessitated by the coming together of the slips, which could not be foreseen when the slips were first written out.
In a large work it is probable that your employers are importunate for "copy," and you will be urged to send this to the printer as you have it ready. If possible, it should be kept to the end, so that you may look over it as a whole, and so see that the same subjects are not in more places than one. You will probably have to make modifications in your plan as you go along, and this may cause difficulties which you will now be able to set right.
Much of the value of an index depends upon the mode in which it is printed, and every endeavour should be made to set it out with clearness. It was not the practice in old indexes to bring the indexed word to the front, but to leave it in its place in the sentence, so that the alphabetical order was not made perceptible to the eye.
There is a great deal to arrange in preparing for the press. Lines of repetition are often a source of blundering, specimens of which have already been given.
The dash should not be too long, and very often space is saved and greater clearness is obtained by putting the general heading on a line by itself, and slightly indenting the following entries.