TRELEASE — OBSERVAT's SUGGESTED BY PRECEd'g PAPER. 289 



across the middle of the diagram, while the third (3) of this set 

 is displaced so much as to interlace with the lower set. The fact 

 is remarkable. 



In 1881, I was struck by the absence of the mid-styled form of 

 our common O. violacea, which, from the structure of individual 

 flowers, should be trimorphic. For purposes of comparison, I 

 have rearranged the measurements made at that time,* according 

 to the longest organ of each flower, and have plotted them on the 

 scale used by Mr. Eliot — Fig. 3. 



Fig. 3. — OXAI.IS VIOLACEA. 



Diagram compiled from tables published in the American Naturalist (iSSj), pp. 14, iS> 

 showing the relative lengths of stamens and pistils in Si flowers : — i. Pistils of long-styled 

 flowers; 2 and 3, stamens of same. 4 and j, stamens of short-styled flowers; 6, pistils of 

 same. — The bottom of diagram represents a height of i mm., and the top a height of 6 mm. 

 above the bottom of the calyx. 



The percentage relations of the different forms in lOo plants 

 of 0. Suksdorfii, and 8i of (9. violacea, are as follows : 



O. Suksdorjii. O. violacea. 



Long-styled 25 63 



Mid-styled 54 o 



Short-styled 21 37 



These proportions would doubtless vary toward uniformity in 

 a much greater number of measurements, but I believe that the 

 indications they give are fairly reliable. In both species the short- 

 styled form is less abundant than the long-styled ; but the mid- 



* From a paper in the American Naturalist for 1S82, p. 13. 

 V. — I — 19 



