The following statement is quoted from a letter written by Dr. L. C. Dunn, managing editor of Genetics, to Dr. L. J. Stadler, a member of the board of editors:-

"The chief difficulty from the standpoint of publisher and printer comes from the frequent employment of subscripts which as you know have to be set in by hand and sometimes require special characters to be cast. This represents extra cost to the journals. If it is absolutely essential it must be done, but I’m not convinced that it is essential. In the present paper A1 would serve as well as A, etc. except that the habit of subscripts has crept in through use. Jones had a rule against them but I notice that he didn’t enforce it in Emerson’s papers and I haven’t either. There’s no avoiding superscripts for multiple allelic series, but subscripts aren’t generally essential and when both are required, e.g. Ab1, the system approaches physical limits for the compositor and looks rather absurd. I don’t propose any sudden revolution. I do suggest it might be discussed by the maize group, keeping in mind that a system needn’t necessarily be frozen by the first ten years of use and that economies in publication, if done without harm to clarity and preciseness, give our journals greater stability and security for the future."

Dr. Dunn’s example illustrates the confusion which might often result from following his suggestion. Arabic figure “1” in typed manuscript cannot be distinguished from l.c. letter “l”. The symbol “al” might be read “a-one” or “albescent”. If the literal part of the symbol were always italicized and the numerical part not italicized, there need be no confusion. Or, if the numeral is joined to the letter by a hyphen, there should be no trouble. Again, if the numeral could be set in smaller type than the literal part of the symbol, the printer’s problem might be solved, but certainly not the typist’s. It seems likely, however, that two sizes of type might be as bad as subscripts for the compositor. In a recent personal conference with Dr. Dunn, he suggested omitting the numeral “1” in all cases. No numeral would then indicate either that there is only one gene with that literal symbol or that it is the first one reported. Thus, we would have a (= a1) a2, a3, etc. In order that you may see how you like it, the latter plan is followed throughout this News Letter. Let me know what you think of it. The principal difficulty noted in its use here appears first in Anderson’s Table (p. 3) where gl3 = glossy 3 not golden 13. In the inventory of seed stocks l7 is not seventeen but luteus 7. Perhaps a period would help, thus: gl.3 and 1.7.

R. A. Emerson