Range data as commonly encountered, including estimates of the 95% range, represent a 

 mixture of the variability existing between individiials and the variability existing within 

 individuals. 



(a) By the method of greatest accuracy, the 95% range is obtained by fitting a recog- 

 nized type of frequency curve to a group of measured values and excluding the extreme 

 2.5% of area under the curve at each end. (See sketch on preceding page). Estimate is 

 made by this procedure only when the group of values is relatively large. 



(b) By a less accurate method, the 95% range is estimated by a simple statistical 

 calculation, assuming a normal distribution and using the standard deviation. This esti- 

 mate is used when the group of values is too small for curve fitting, as is usually the case. 



(c) A third and still less accurate procedure for estimate of the 95% range is sim- 

 ply to take as range limits the highest value and lowest value of the reported sample group 

 of measurements. It underestimates the 95% range for small samples (3 or 4 values) and 

 overestimates for larger sample sizes, but may be used in preference to the preceding 

 method when the sample shows convincing evidence that the variable is asymmetrical in 

 distribution. 



(d) The upper and lower limits of the ordinary range of variation, as estimated by 

 an investigator experienced in measuring the quantity in question and based solely on gen- 

 eral experience, constitute still another estimate of the 95% range. The trustworthiness 

 of limits so placed should not be underestimated. 



In many instances range data have not been available. In other instances an esti- 

 mate of the 95% range is given, but information on the manner of estimate is lacking. Ef- 

 fort to assemble both types of missing information is continuing. 



The data in each table are in the judgment of the contributors and reviewers es- 

 tablished fact and free of questionable material, and represent the consensus of expert 

 judgment and experience in the special field from which the table was drawn. It is recog- 

 nized, however, that all data are subject to continuing revision as investigators improve 

 techniques and make more measurements. The reader is warned against attributing sig- 

 nificance to small differences from species to species. He is invited to submit any values 

 or ranges he feels should be given consideration, and is particularly invited to add to the 

 coverage of animal and plant forms. 



