ECCENTRIC GROWTH. 283 



on the upper side of the shoot, and their anisophylly might tend 

 to disappear.* 



While, again, a marked degree of eccentricity is rarely met with 

 in the case of asymmetrically constructed foKage-shoots, such a 

 phenomenon is frequent in the case of spirally constructed flowers ; 

 and in such types the very general attempt to convert the floral 

 diagram into a symmetrical circular expression has not only led 

 to confusing results, but in many instances has served to conceal 

 the essential asymmetry of the floral structure. Thus, in a typical 

 Angiosperm flower the assumption of symmetry in the sporophyll 

 region is so remarkable and so definite, that the circular plan of 

 the floral diagram becomes the common convention, to which a 

 quincuncial calyx or perianth as an outer investment appears 

 almost as an accident of the construction; and when the asym- 

 metrical region is thus limited to the members of a single contact- 



* The case of Salvinia is of special interest : tlie embryo, as is well known, 

 commences a normal asymmetrical development which, is continued for three 

 leaves constituting a single cycle. At the fourth node, symmetrical alternating 

 whorls of three commence ; that is to say, a (2 + 3) primary system adds after 

 one complete cycle around the axis a new path of distribution, the system thus 

 immediately becoming symmetrical of the type (3 + 3) : less frequently (2 + 2) 

 symmetry is first attained at the third node (Pringsheim). Normal eccentricity 

 is, however, superimposed on the construction, thus giving three rows of leaves 

 on either flank. It is interesting to compare this architectural scheme for the 

 position and even approximate primary shape of such leaf-members and the 

 apparent order of development of these successive whorls with the ontogenetic 

 or building account given by Pringsheim (GesammeUe Abhandlungen, vol. ii. 

 p. 354), although so little of any scheme can be definitely checked at the 

 actual algal-like apex.' Thus, according to Pringsheim, the water-leaf arises 

 first, and is followed by the foliage-leaf farthest removed from it, the other 

 leaf on the same side as the submerged leaf last. The whorls then alternate 

 in the same fashion. This agrees with the geometrical construction (fig. 97), 

 but Pringsheim's theoretical schemes are based on a preconception of the 

 importance of quadrant division in the segmenting cells behind the apex. 

 Once it is admitted that the apical cell is not the ruler of the space-form, but 

 an accident due to the special type of segmentation into relatively bulky cells, 

 and that lateral members are not localised accurately to special segments, 

 beyond the fact that regularity in production of cells and members must 

 necessarily involve a certain coincidence between the two, the value of these 

 segmental schemes is miich diminished, and the drawings on which they are 

 based do not afford any convincing evidence of their theoretical importance. 



