282 TRANS. ST. LOUIS ACAD. SCIENCE. 



Length of pistil in C 7.0S mm. 



" " siiort stamens in B 7-77 " 



" " long " " A 5.09 " 



Average length of middle fart 6.65 " 



From A, B, and C, rearranging- according to length of part, 

 we have 



Ratio \ tfe 7'.?? q.ls \ . /.""T/"' . . 



4.4 : 6.7 : 9.5 



Of the one hundred specimens measured, as has aheady been 

 stated, the numerical portion of the diflerent forms was 



21 : 25 : 54 = 4.1 : 5.0 : 10.8 



From Table IV. it will be seen that the pistils in either the long- 

 styled or short-styled forms are intermediate in length between 

 the two sets of stamens of the other two forms ; while the pistils 

 of the mid-styled form are intermediate between the longer parts 

 of the two other forms. In the long-styled and short-styled forms 

 the distance between the stigmas and nearest set of anthers is 

 greater than the distance between the anthers, in a ratio nearly 

 constant. 



In the Diagram here inserted, a graphic representation of the 

 measurements is given. In plotting the curves the flowers were 

 arranged accortling to the length of the lon<^est part. This causes 

 the curves 1,4 and 7 to descend uniformly, without interfering 

 necessarily with their relations. They interlace, thus showing 

 that in the three forms the length of the long part is nearly the 

 same. The curve -^ does not interlace so well with 1 and 7 as 

 it should, because the greater number of points plotted extends it 

 away to the right. In the curve 7 the stamens of one of the flow- 

 ers were abnormally small, and I call attention also to the two 

 very long sets of stamens (12.7 mm. and 13.3 mm.) of the first 

 two flowers, whose points in the curve lay outside the limits of 

 our coordinate paper. It will be seen by referring to Table IV. 

 that the average length for the long stamens of the long-styled 

 flowers is unexpectedly below the average for the middle part of 

 the other two forms. This is seen very nicely in the Diagram, 



