24 A MILL SCALE STUDY OF WESTERN YELLOW PINE. 



PART II. GRAPHIC ANALYSIS. 



Part II of this publication is a detailed graphic analysis of all the 

 woods and mill data pertaining to the volume, value and growth of 

 immature and mature western yellow pine; resulting from field work 

 described in Part I. 



The methods of graphic analysis afford, without doubt, the most 

 authentic means of determining the underlying principles and average 

 values from a large amount of data. Average values obtained by such 

 methods depend not only upon the arithmetical averages of the parts 

 which make up the whole, but each part in turn depends upon every 

 other part. Values and underlying principles determined by graphics 

 from a large number of observations are dependent upon the majority 

 of observations made, and are not influenced to any great extent by 

 erratic values, which are usually the result of some error in observation. 

 Such errors as a rule are compensative, and the principal tendency they 

 have is to enlarge the field of variation. They do not affect the curve of 

 averages such as would be located by one of skill and experience. 



Compensative errors may occur in large numbers without any 

 influence upon the final results. They may, however, make the work 

 more difficult and obscure, but so long as they are well within the 

 minority their influence should never reach the skilfully located locus 

 of averages. 



Accumulative errors have an entirely different effect. Their influ- 

 ence can not be eliminated from the final determinations unless a proper 

 correction based upon the known errors be applied. Their effect will 

 be to either raise -or lower the curve of averages with an accompanying 

 change of its slope. Such errors are introduced usually by using 

 incorrect measuring instruments, such as a scale with divisions marked 

 inches and tenths of inches which are a little greater or a little less than 

 inches or tenths of inches, or by using incorrectly adjusted calipers, etc. 

 Such errors are easily prevented by standardizing all instruments and 

 scales before putting them into use, and then keeping them in good con- 

 dition. 



Still another form of error may creep in which is neither compensa- 

 tive nor accumulative, but which holds constant for each individual 

 observation. This form also can not be eradicated by methods of 

 graphics, unless known and the proper correction applied. It may be 

 caused by measuring from a point on a tape which is not a true "zero 

 point, or by using a tape which has been spliced, such that beyond the 

 splice it is longer or shorter than it should be. The effect upon the 

 curve of averages is to raise or lower it by a constant amount without 

 changing the slope. It is an error which should never enter into care- 

 fully performed work. 



Notwithstanding the great advantages of graphics, it is a very poor 

 method for an inexperienced person to use, since many inconsistencies 

 may result. For instance, it is very possible to establish a set of values 

 which absolutely contradict another set, and both sets be based upon the 

 same original data. This may result from changing from one variable 

 as the basis of comparison to another. 



