UNITED STATES NATIONAL MUSEUM BULLETIN 285 



25 



15. (0-2) 



16. (1): 



17. (i; 



first variable specified in Instruction 13, the 

 second variable to the second variable, etc. 

 Punch in fields of 3 in columns 1-72. 



The card is punched 008 for the problem in 

 Figure 1. 



Column numbers of variables in the univar- 

 iate output array which are to be transformed 

 to logarithms to the base 10 and which are to 

 be entered as new columns in the univariate 

 output array. (See Instruction 7f.) Punch in 

 fields of 3 in columns 1-72. 



The card is punched 002003 for the prob- 

 lem in Figure 1. 



(a) Columns 2-3: The number of univariate 

 variables which are to be fitted on an output 

 page. (This includes a column of specimen 

 identification numbers on each page.) 



(b) Columns 5-6: The number of segments 

 into which the univariate output array must 

 be broken in order to fit on standard output 

 paper. Each output segment of the array be- 

 gins with a column occupied by the specimen 

 identification numbers, so that rows in the 

 segment may be readily identified. Calculate 

 the number of segments by means of the fol- 

 lowing formula: 



N= 



T-C 



+ 1 



C-l 



where A r , the number of univariate output 

 segments, is rounded to the next highest whole 

 number; T is the number of variables in the 

 univariate output array (Instruction 3) ; and 

 C is the number of columns which can be 

 fitted on an output page (Instruction 16a). 

 For the sample output (see listing), T is 

 6, and C is 14, so that N is 3, after rounding 

 2.7 to the next highest whole number. 



(a) Columns 2-3: The number of bivariate 

 variables which are to be fitted on an output 

 page. (This includes a column of specimen 

 identification numbers.) 



(b) Columns 5-6: The number of segments 

 into which the bivariate array must be broken 

 in order to fit on standard output paper. Cal- 

 culate by means of the formula in Instruction 

 16b, where T is the total number of variables 

 in the bivariate array (Instruction 4) and C 

 is the number of bivariate columns which are 

 to be fitted on an output page (Instruction 

 17a). 



For the sample output (see listing) , T is 42 

 and C is 12, so that N is 4 after rounding 

 3.8 to the next highest whole number. 



18. (0-10) : Indexing of column numbers of variables to 



be used in computation of ratios, y/x, begin- 

 ning with the first ratio to appear on the 

 first bivariate output page and proceeding 

 from left to right through consecutive pages. 

 Each ratio is represented by a group of 1 2 

 card columns, punched as follows: 



(a) Columns 2-3: Column number of y in the 

 univariate output array. 



(b) Columns 5-6: Column number of x in the 

 univariate output array. 



(c) Columns 8-9: Column number which the 

 ratio will have on its respective output page. 

 (The column of specimen identification num- 

 bers is column 1 on each page.) 



(d) Column 12: Signal for the log (base 10) 

 transformation of both y and x in order to 

 produce bivariate statistics of the ratio both 

 before and after log transformation. Punch 

 zero if no transformation; punch one for 

 transformation. 



Indexing for the next ratio begins in the 

 next field of 3 columns, and so on, through 

 the first 72 columns of the card (6 ratios in- 

 dexed per card). Indexing then begins again 

 on a second card. Because the maximum num- 

 ber of ratios permissible is 55, the maximum 

 number of cards allowed here is 10. 



If no ratios are to be computed (i.e., if the 

 control number described in Instruction 4 is 

 zero) , no cards should be present for this in- 

 struction. 



For the schematic problem in Figure 1 , two 

 cards would be required which would be 

 punched as follows: first card: 02 03 02 01 

 03 04 03 01 06 03 04 00 06 05 05 01 09 04 06 

 01 10 04 07 00; second card: 02 06 08 01 12 

 13 09 00. 



19. (2n) : Names of variables to be placed as column 



headings on each output page. The column 

 headings of each page occupy two cards (col- 

 umns 1-72 of the first card and columns 1-48 

 of the second card i , with the spacing begin- 

 ning with card column 1 , corresponding to 

 print positions on the output page (120 print 

 positions per line") . The names of variables on 

 consecutive pages occupy consecutive pairs of 

 cards. The number of cards required here is 



