In Cobbett's Woodlands there is a good specimen of backwards and forwards cross referencing. The author writes:
"Many years ago I wished to know whether I could raise birch trees from the seed.... I then looked into the great book of knowledge, the Encyclopædia Britannica; there I found in the general dictionary:
"'Birch tree—See Betula (Botany Index).'
"I hastened to Betula with great eagerness, and there I found:
"'Betula—See Beech tree.'
"That was all, and this was pretty encouragement."
William Morris used to make merry over the futility of some cross references. He was using a print of an old English manuscript which was full of notes in explanation of self-evident passages, but one difficult expression—viz. "The bung of a thrub chandler"—was left unexplained. In the index under Bung there was a reference to Thrub chandler, and under Thrub chandler another back to Bung. Still the lexicographers are unable to tell us what kind of a barrel a "thrub chandler" really was. I give this story on the authority of my friend, Mr. S. C. Cockerell.
No reference to the contents of a general heading which is without subdivision should be allowed unless of course the page is given.
There are too many vague cross references in the Penny Cyclopædia where you are referred from the known to the unknown. If a general heading be divided into sections, and each of these be clearly defined, they should be cross referenced, but not otherwise. At present you may look for Pesth and be referred to Hungary, where probably there is much about Pesth, but you do not know where to look for it in the long article without some clue. Sometimes cross references are mere expedients, particularly in the case of a cyclopædia published in volumes or parts. Thus a writer agrees to contribute an article early in the alphabet, but it is not ready in time for the publication of the part, so a cross reference is inserted which sends the reader to a synonym later on in the alphabet. In certain cases this has been done two or three times. An instance occurs in the life of the distinguished bibliographer, the late Henry Bradshaw (than whom no one was more capable of producing a masterly article), who undertook to write on "Printing" in the Encyclopædia Britannica. When the time for publication arrived (1885), Bradshaw was not ready, and in place of the article appeared the cross reference, "Printing, Typographic—See Typography." Bradshaw died on February 10, 1886, and the article on "Typography" which was published in Vol. 23 in 1888, was written by Mr. Hessels.
Cross referencing has its curiosities as well as other branches of our subject. Perhaps the most odd collection of cross references is to be found in Serjeant William Hawkins's Pleas of the Crown (1716; 5th ed., 1771; 7th ed., 4 vols., 1795), of which it was said in the Monthly Magazine for June, 1801 (p. 419): "A plain, unlettered man is led to suspect that the writer of the volume and the writer of the index are playing at cross purposes."
The following are some of the most amusing entries:
"Cards see Dice."
"Cattle see Clergy."
"Chastity see Homicide."
"Cheese see Butter."
"Coin see High Treason."
"Convicts see Clergy."
"Death see Appeal."
"Election see Bribery."
"Farthings see Halfpenny."
"Fear see Robbery."
"Footway see Nuisance."
"Honour see Constable."
"Incapacity see Officers."
"King see Treason."
"Knaves see Words."
"London see Outlawry."
"Shop see Burglary."
"Sickness see Bail."
"Threats see Words."
"Westminster Hall see Contempt and Lie."
"Writing see Treason."
This arrangement of some of the cross references is perhaps scarcely fair. They are spread over several elaborate indexes in the original, and in their proper places do not strike one in the same way as when they are set out by themselves. One of the instances given by the critic in the Monthly Magazine is unfairly cited. It is there given as "Assault see Son." The cross reference really is, "Assault see Son Assault."
Hawkins's work is divided into two parts, and the folio editions have two indexes, one to each part; the octavo edition has four indexes, one to each volume.