THE MANUSCRIPT 



229 



40r- 



30 



o 



•o 20 



10 



X 



J_ 



10 20 30 



Time in minutes 



40 



Fig. 17-1. Left, a properly prepared graph of the results of a typical experi- 

 ment. The first few minutes were the most interesting, so readings were 

 made at shorter intervals. Right, the same data : upper, axes interchanged; lower, 

 time scale not linear. 



classifies the data, and its organization should be self-evident. Because 

 the numbers in a table usually could be arranged in several different 

 ways, it is desirable to try several forms and select the one that will dem- 

 onstrate the results to the reader most effectively. Table 17-1 presents 

 the same figures as Table 16-3, but in a different arrangement. If the 

 two are compared, Table 16-3 is preferred because additions of vertical 

 columns are easier than addition across a page. 



Table 17-1. Data on Analysis of Variance from Table 16-3 



Source of Variation 

 Degrees of Freedom 

 Sum of Squares 

 Mean Square 



Typing the manuscript 



The manuscript should be typed on good white paper, preferably 16- 

 pound or 20-pound bond. If the surface is too rough, erasures are impos- 

 sible; if too smooth or glazed, ink will smear. Use a typewriter with sharp, 

 clean type faces and a ribbon in good condition. 



