Harding. — On Kerns and Serifs. 99 



"serif " does not resemble Kcpaia, yet, through the ambiguous 

 use of our letter c, numerous Greek words beginning with k 

 are by us pronounced with an initial s, and " ceraia " certainly 

 approximates to " ceriph."''' Murray's great dictionary, now 

 in course of publication, gives the only discoverable suggestion 

 as to the etymology of serif. In 1863 a writer in Notes and 

 Queries suggested the Dutch and Flemish schreef, "line, 

 stroke," but the editor is doubtful, and says, " historical evi- 

 dence is wanting." I think my etymology is at least as 

 probable. At the same time, the singular variety of spelling 

 seems to suggest transfer from some such language as the 

 Hebrew, where exact transliteration is not possible. 



I return to " kern." I find only two suggestions as to its 

 origin. The Encyclopaedic Dictionary says, "Perhaps from 

 crena, " a notch." This I think exceedingly unlikely. A 

 kern is not a notch, nor does it resemble one. All types have 

 notches — a most important feature — but these are technically 

 " nicks." Another, in a later dictionary, is attributed to Mr. 

 Thomas MacKellar, a celebrated typefounder of Philadelphia, 

 who thinks that the word may be derived from the Celtic 

 quern, a hand-mill ; but the connection of ideas is not obvious, 

 nor does the word seem to have been one likely to have been 

 adapted to this use by the early printers. Just as the serif is a 

 "horn," so, I suggest, is the " kern." In one case the horn 

 is a feature of the impressed character, in the other of the 

 type which gives the impression. 



The " serif," then, comes from the Greek Kepaia ; the " kern " 

 from the Latin equivalent, cornu ; and, if this be the case, two 

 English words diverging as widely as the two I have been 

 considering — "kern" and "serif" — spring from a common 

 root. The vowel-change need give us no concern, for we find 

 exactly the same in " corn " and "kernel." 



I must add that my friend Sir James Hector has sug- 

 gested another possible derivation of " kern," much more pro- 

 bable than crena, and deserving of consideration. He thinks 

 it may be a form of " cornice," and the overhanging portion 

 of the type might very fairly suggest the comparison. In this 

 case we also go back to the Latin ; but we find the root not 

 in cornu, but in corona, " a crown," and the analogy to the 

 Plebrew would be foimd not in Dr. Clarke's " seraphs," but 

 in his " coronal apices," in which the ancient scribes displayed 

 their skill. 



And once again, to anticipate possible criticism, I repeat 

 that I put forward these etymologies only as suggestions, to 

 my mind more probable than any that have hitherto been 

 made public. 



* We have a good example from the same Greek root in the word 

 *' rhinoceros." 



