korstian's classification of forestry literature 307 



tralian (including New Zealand), British Isles proper, and others. 

 Furthermore, we have found it advisable to make a separate division 

 altogether for periodicals, co-ordinate with forestry itself, as there are 

 so many of them in forestry and closely related subjects. 



But now as to the length of the call numbers. It is more necessary 

 in a library to have these as short as possible than it is in a file of cards 

 or clippings, for various reasons, and it is still more necessary in an 

 open-shelf library, where people will be using the material who are 

 not accustomed to carrying long numbers in their heads while diligently 

 searching for small pamphlets. The librarian can remember longer 

 numbers than the average user of the library, due to practice and to an 

 immediate visualizing of the general location of the book wanted; but 

 there is a limit, and it is annoying to all concerned to have to consult 

 a slip of paper constantly during the search, especially among thin 

 books and pamphlets bound in cardboard covers. Besides, the longer 

 the numbers, the greater the chance of misplacement on the shelves and 

 consequent unavailability. 



Call numbers consist of three parts — class, author, and book num- 

 bers. The book number, however, seldom has to be remembered with 

 the rest of the call number, because it is usually either the serial num- 

 ber of the publication one wants, as Bulletin No. 177, or it is the year 

 number of a report, as 915 for 191 5, or it is i, 2, or 3, in the case of 

 several books by the same author. These things one remembers sepa- 

 rately and easily ; therefore I leave the book number out of account. 

 The class number (Dewey number when the Dewey system is used) is 

 longest, obviously, when the classification is most detailed. The author 

 or Cutter number is longest, contrariwise, when the classification is 

 least detailed. But both class and author numbers are at their most 

 efficient length (every bit of their length serving a useful j^urpose. none 

 being wasted) and the combination is shortest, at the same time : that 

 is, when the average minor subdivision contains a reasonable number 

 of books, not so many as to require extremely long Cutter numbers, 

 not so few as not to need any. 



This most efficient length is as long in number of symbols, whatever 

 the system used ; but just here the writer wishes to be heretical and 

 state that the approved systems do not always have the simplest num- 

 bers for memorizing. To give an instance : In our section containing 

 serial publications of v^tate forest officers and commissions, we found 

 exceedingly long Cutter numbers necessary in tlie letters M and N. 

 (How many of the readers of the Journal have ever stopj)ed to realize 

 that one-third of the States of the Union have names bcgiiniiiig with 

 M or N ?) Of course, numbers from the Cutter tables would \k- exceed- 



