t is approximately. the mi en- 



la] :t. 



Or the other hand, . 3 for 



'■ ■■ al Iculated has 

 proached the maximum temperature for enlargement, the coeffi- 

 mt approaches zero. No matter what ranee 



• of maintained t 

 1 elow the ; iximum to some value above I al 



point must be accomplished by a corresponding change ir | 

 rate o r enlargement from a positive value to zero, and the ra - 

 of zero any positive quantity is of course zero. 



From t>>ese considerations it is clear that all graphs 

 of temperature coefficients are vertical and traverse a po " 

 whose ordinate is infinite and whose abscissa is near the 

 value of t^e rri^ : in bemperature for the process considered ; 

 and they are all horizontal and traverse a point whose ordinate 

 is zero and '.Those abscissa is near the value of the maximum 

 temperature for the process considered. These two points are 

 the fixed ends of grapl s, 1 etween them the graph slopes down- 

 ward to the rirh+, having a curvature varying wit* tl 

 tained temperature itself. 3 graphs of fig 4 appear 

 to he all concave upward throughout their upper portions, but 

 they each show a reversal for an ordinate val 'ween 1,0 

 and 2«0, beyond which point each graph is concave downward 

 for some distance, as the tercperatu ^e ma if approached ; 

 i.e., as the c ifficient value approaches zero- 



sr graphs such as those here considered it follows 

 (from the points brought out a'-ove) 4 I I j lope of J 



82 



