54 



UNITED STATES NATIONAL MUSEUM BULLETIN 285 



Each point on the scatter diagram is a two-digit speci- 

 men number, with the lower left-hand corner of each two- 

 digit cell representing the actual point. These specimen 

 numbers are not the specimen numbers read in with the 

 data, but rather are numbers, from 1 to n, which are as- 

 signed to n specimens in the data deck in the order in which 

 they appear in the deck. The arbitrary set of minimum 

 values is also assigned a specimen number, which is n + 1, 

 and so also is the set of maximum values, the specimen 

 number of which is n + 2. Where two or more points co- 

 incide, a double letter is printed, with the specimen numbers 

 of the coincident points printed after the same double 

 letter on the right side of the diagram. 



In the sample output, specimens No. 13 and No. 14 repre- 

 sent, respectively, the plots of the specified minimum and 

 maximum for the plotted variables. 



The end points of the reduced major axis for the first 

 diagram (DF versus AM) are plotted as follows: The 

 y-intercept (the y-coordinate of the point on the reduced 

 major axis where .v = 0), given as BASE 1 in the output 

 of DASAN, is — 1.884. Because this is in the scale units of 

 the diagram, it is only necessary to count down — 1.9 dot's 

 along an imaginary vertical line through the lower left 

 corner of the two-digit cell occupied by specimen No. 13, 

 which represents the origin. Similarly, the point at which 

 the reduced major axis crosses the right-hand margin of 

 the diagram is indicated in the DASAN output by TOP 1, 

 which is 2.659. This is plotted by counting 2.7 dots down- 

 ward along a vertical line through the lower left corner 

 of the two-digit cell occupied by specimen No. 14, the 

 maximal point. 



The values DX and DY, printed in the lower right corner 

 of each output page, are, respectively, the values between 

 each dot on the y-axis and between every second dot on 



the x-axis. It will be seen that there are 49 such scale-units 

 on the y-axis and 35 on the x-axis between the minimal 

 and maximal points. 



The regression of y on x is plotted in the same manner, 

 but the end points of the regression of x on y are given in 

 terms of units on the x-axis rather than on the y-axis. 

 Thus, TOP 3 in the output of DASAN (on the last page 

 of the output listed below) refers to the number of x-axis 

 units to the left ( — ) or right ( + ) of point No. 14, and 

 BASE 3 refers to the number of x-axis units to the left ( — ) 

 or right ( + ) of point No. 13. 



The second diagram shown is a double log plot (LOG 

 LO versus LOG AM) in which the coordinates of the speci- 

 fied minimal point are ( — 1.0, —1.0). Here the basal end 

 points of the reduced major axis and the regression of 

 y on x have been measured along a vertical line through 

 the dot (drawn for the purpose of illustration) which rep- 

 resents the coordinates (0,0). The basal end point of the 

 regression of x on y is measured along a horizontal line 

 through the same dot with measurement in units of the 

 x-axis. The upper end points of all three lines are plotted 

 as in the preceding example. 



It will be found that artificially constructed scales match- 

 ing the spacing of units on the axes will increase the speed 

 and accuracy of line-plotting. When only a small number 

 of points are plotted, there may be an apparent discordance 

 between the plotted points (specimens) and the hand-drawn 

 line of best fit. This is because the two-digit plots are only 

 approximations, the accuracy of which is determined by 

 the print-spacing of the printer, whereas the regression lines 

 and reduced major axis, which are plotted by hand, are 

 limited only by the accuracy of the hand-plotting operation 

 and by the data themselves. 



Error Messages in VPLOT 



Only one error message has been written into VPLOT, 

 which reads NEGATIVE VALUE IN INPUT. It is printed 

 out, and the run is terminated, whenever a negative value 



appears where not permitted on a control card or in the 

 data where this value is to be transformed to a logarithm. 



Availability of Program Decks 



Duplicates of the FORTRAN II source decks may be ob- 

 tained through the Smithsonian Institution by writing to 



the author care of the Department of Paleobiology, Smith- 

 sonian Institution, Washington, D.C. 20560. 



