208 JOURNAL OP FORESTRY 



ingly long under these conditions ; so we arbitrarily assigned numbers 

 from I to 9 to these States — for instance, M4 for Massachusetts — and 

 then gave the serials published by Massachusetts numbers from M4 to 

 M49. But even this did not simplify matters sufficiently in N, where 

 the States beginning with "New" have been very prolific publishers. 

 It was seen that four or even five figure numbers were inevitable. Con- 

 sequently we tried the plan of treating "New" and "No" as if they 

 were two others letters following N, and gave New Hampshire, for 

 instance, the numbers from New3 to New39. Now here is the interest- 

 ing fact : "New37" contains 5 symbols and so does N4958, for example ; 

 but the former is remembered more easily because it is remembered 

 as two units, namely, "New" and "thirty-seven." The N followed by 

 four decimal figures, in connection with a Dewey number of three or 

 four places, is almost impossible to carry in the head while looking for 

 a pamphlet ; the New followed by two figures is quite possible. The 

 advantage to memory by this change offsets any disadvantage of the 

 possible disarrangement of alphabetic order in case of the admission 

 of a new State whose name should begin with Nex, Ney, Nez, Ni, or 

 Ny. Which is preliminary to saying that if this is true of the author 

 numbers, that a combination of letters and figures may be easier to 

 remember than a row of decimals of equal length, why may it not be 

 true of class numbers (with all due respect and apologies to Mr. 

 Dewey) ? In other words, why necessarily make the forestry classifica- 

 tion a Dewey Decimal one ? 



The writer's suggestion would be this : By all means let the Society 

 standardize the classification, after due discussion. But why not pub- 

 lish the classification adopted as an outline, without Dewey numbers 

 (they are easy enough to put in if one wishes to use them) ? Again, 

 why confine the number of main headings to 10? It seems to me that 

 this would be a serious mistake in a young and growing science like 

 forestry, where, almost overnight, a new minor section may spring up 

 to the importance of a main division. Why not leave it flexible — use 

 as many main divisions as seem important now — the number would 

 probably be between 10 and 20 — and leave it possible to elevate some 

 of the lesser divisions or put in new ones later if it seems advisable? 

 This need not mean that a library using the Dewey system could not 

 use the scheme, for such could use two successive numbers at the left 

 of the decimal point, e. g., the numbers of the main divisions might run 

 from 5.0 to 6.9 instead of from 5.0 to 5.9. Of course a general library 

 could not so easily do this, but it would be possible without any longer 

 numbers than the other scheme ; and then most general libraries would 

 not need very elaborate divisions anyway and could easily unite some 



