A CRITICAL-EVALUATION OF THE RELATIONSHIPS BETWEEN AREALOMETER INSTRUMENT VALUES AND COTTON FIBER PERIMETER .1. FUNDAMENTALS AND COMPUTER-SIMULATION

Arealometer instrument values are computer simulated to study the erro r in the calculated fiber perimeter P due to instrumental errors. (Thi s airflow device senses the fiber's specific surface at two different specimen compressions; P and wall thickness t are computed from the re adings.) Systematic changes over time in instrument values from design ated true values are classified into thirteen nontrivial combinations. Each combination results in different P values and, consequently, a u nique drift or error in P. Simulations of 50,000 Arealometer observati ons using actual standard deviation data from a control cotton are als o run. These simulations show that random variations in instrument val ues over time are, in effect, an unordered sequence of the nontrivial combinations. Also, a mean taken from only two to six Arealometer obse rvations-the accepted practice-does not adequately characterize a cott on sample. Fundamental relationships are probed with partial and total derivatives of Arealometer functions. There is more variability in th e calculated P than in the t value. The total differential dP accurate ly measures the error in P, given the error in the instrument values.