128 PEESIDElfT's ADDKESS. 



soids ; 21 per cent, spheres and discs ; 11'5 per cent, rectangular 

 bodies; 11 per cent, polygonal bodies; 7 tetrahedrons, and 4-5 

 were unclassed. 



The two first, embracing 66 per cent., seemed to indicate that 

 the normal shape must be sought in ellipsoid or sphere. Now, 

 pollen is spherical whilst within the anther cell, and elliptical 

 when it becomes free. Whilst examining the recently-shed 

 pollen of Sarothamnus scoparius one day, to my astonishment I 

 saw in a few seconds some of the spheres change into ellipsoids, 

 and in about two minutes not a sphere remained. On gaining 

 perfect freedom an unequal tension was evidently able to assert 

 itself. I had often suspected this change before, from undoubted 

 spheres having been seen among undoubted ellipsoids. I got 

 fresh pollen and the phenomenon was often repeated, so that all 

 doubt vanished. The pollen of Oxalis acetosella afforded another 

 good example of this change from sphere to ellipsoid, and in this 

 case the ellipsoids continued the change till many became dis- 

 tinctly rectangular, again demonstrating that form is regulated 

 by internal tension. The brown prickly spheres of the pollen of a 

 Cactus, Epiphxjllum truncatum, violaceum, on being re-examined 

 a second day had all changed from spheres to polygons, mostly 

 seven or eight-sided. Doubtless, with these examples in mind, 

 many other cases could be found. The conclusion seems inevit- 

 that the normal primal shape of pollen is spherical, and that not 

 only discs and ovals are modified spheres but also cylinders and 

 rectangles as Avell as polygons. 



Some that seem spheres under a low power are seen under a 

 high power to be many-sided polygons, as Chenopodium Bomis- 

 Henricus, and Phlox Drummondii. I conclude all the varied 

 shapes are modifications or sections of spheres, viz., the ellipses, 

 discs, squares, rhomboids, and polygons, modifications ; the cones 

 and tetrahedrons, sections ; whilst the dumb-bells, trefoils, quatre- 

 foils, etc., are aggregations. 



It must be noticed how certain shapes predominate in certain 

 orders, for example, spheres in CampanulacecB and Cactacea ; 

 discs in Mimosm ; ellipsoids in Liliacece ; polygons in Caryophyl- 

 lacea ; triangles in Onagracew ; rectangles in Leguminosoe i 



