On methods of measuring nerve fibres

Various methods of measuring axonal circumference and cross sectional area were examined. In measuring circumference, the Reichert MOP semi-automatic image analysis system varied more than a map meter. Several methods of area measurement were compared. Though all were relatively accurate in the long run, variability of individual measurements differed from one method to another, being greatest with a rolling disc planimeter and least with the method of counting millimetre squares within the profile to be measured. The map meter and the MOP were compared in measuring the circumferences of a large sample of axons: the latter method was significantly the more variable. Also, the MOP gave values significantly greater than the map meter. Comparison was also made of area measurements on the same large sample of axons, using various methods. Apart from the MOP and the rolling disc planimeter, each method gave area estimates significantly different from all others. The differences were in some cases as great as 4%. There was no significant correlation between the error involved and the size of the axon in measuring axonal circumference or area. Of all the methods studied, the MOP was the most efficient in measuring both circumference and area in terms of time taken and versatility. So long as a particular study is confined to comparing sets of measurements made by the same method, the systematic errors reported above are relatively unimportant. However, in comparing sets of measurements made with different methods, systematic errors may be important and should be borne in mind.