The objective in having the eight supplementary punched card files is demonstrated by the 

 foregoing example; the files permit an initial and efficient manual selection of cards from the total 

 punched card file, resulting in a group of cards which is of a size practical for handling by machines, 

 if necessary. 



With the preceding as background, it can now be explained how the CBCC used the machines 

 to the advantage they have been described as offering. Though machines are not used by the 

 CBCC to sort through the entire punched card file, the advantage of machines is not lost. It is merely 

 transferred to the secondary step. The cards resulting from manual selection might be regarded as a 

 temporary and specialized file of cards; regarded in this way, it is seen to be cross-indexed by punches 

 according to every one of the information categories coded by the Biology Code. All the advantages 

 described earlier for coding and machine handling in general apply to this special small file which 

 results from the initial manual selection. To carry this idea to the example of cards selected from the 

 punched card file arranged according to organisms , we can regard cards resulting from the selection as 

 a special file, the "special" feature in this example being restriction of information to that on one or 

 more specified organisms . If only the cards for one species of organism were selected, the cards in 

 this specialized file resulting from that initial selection would be arranged according to Chemical 

 Serial Numbers, since that is the way the cards are arranged secondarily under any one species in the 

 punched card file. The remaining information punched on the card (i. e. , other than test organism 

 identity) represents indexes which, however, can be efficiently used as indexes only by applications 

 of machine sorting of all the cards (unless the cards are so few that inspection of the punches of the 

 card or interpretation at the top of the card would be faster than machines). For example, if the initial 

 manual selection results in five hundred cards (all punched with the symbols for the organism in 

 question), the sorting of those cards for only those which are punched with symbols for a specified 

 response or for a specified organ is made practical by use of machines. 



For practical retrieval of information by machines, using only a single file comparable to the 

 CBCC Serial Punched Card File, machines used must have a far greater capacity for rapid selection, 

 assembly, and reproduction from that single file than the IBM equipment used by the CBCC. 



Chemistry IBM Punched Card Files : 



The major file of Chemistry IBM Punched Cards is arranged according to the CBCC Chemical 

 Serial Number and is referred to as the "Chemical Serial File". This single file is not sufficient for 

 the most practical retrieval of information about chemicals of specified structure and therefore a 

 subsidiary file is maintained for these cards. 



This second file is arranged according to structural components instead of the Chemical Serial 

 Number. For this particular file, the coding of each compound is in most cases necessarily punched 

 on more than one card, the number of cards for any compound being equal to the number of types of 

 structural groups which are contained in the molecule and which are given a distinct code symbol. 

 (The Chemistry Code should be examined to understand the specific structural groups assigned unique 

 symbols. ) On these two or more punched cards representing one chemical compound, the punching of 

 the coded structural groups is "rotated". To explain this "rotation" requires first explaining that the 

 punching of code symbols for structural groups of a compound is not fixed to a definite area of the 

 Chemistry Punched Card in the way that a given type of biology information (e. g. , the test organism 

 identity) is fixed to one area of the Biology Punched Card (e. g. , only Columns 18 through 25 of the 

 Biology Punched Card, for test organisms); instead, on the Chemistry Punched Card, the coding for 

 one structural group can be punched in Columns 13 through 16, or punched in Columns 17 through 20, 

 or 21 through 24, etc. (Coding of any structural group requires only four IBM punched card columns, 

 three columns for the code identity of the group and a fourth column to indicate the number of times the 

 structural group occurs in the molecule. ) Therefore, the punched sequence of the coded structural 

 groups can be varied on the cards for any one compound so that each group appears in turn at the 

 beginning (i. e. , at the left of the Chemistry Punched Card, Columns 13 through 16). For example, a 

 compound with three structural components (for which there are three distinct, four-unit code symbols) 

 would occupy Columns 13 through 16, 17 through 20, and 21 through 24 on any punched card (refer to 

 the illustration of the Chemistry Punched Card, Figure 3), but in this special "rotated" file, three cards 

 would be punched, differing only in that the coding for each structural group appears on one of these 

 cards in Columns 13 through 16. This file is referred to as the "Chemistry Rotated File" and it permits 

 manual selection of all compounds containing any given structural group, because the file is arranged 

 according to the coded structural group punched in Columns 13 through 16. 



Two other files of Chemistry IBM Punched Cards, indicated on the card illustrated in Figure 3, 

 are special files maintained more for CBCC internal convenience than for actual retrieval or correlation 



210 - 



