TAXONOMIC PRINCIPLES 27 



similarity to convergent evolution (discussed below). Hutchinson con- 

 tends that there is a considerable and fundamental phylogenetic gap 

 between a buttercup and a magnoha tree and that, although the 

 herbaceous habit has developed independently in several woody 

 famihes, the preponderance of morphological evidence supports his 

 arrangement. 



Figures 2-8 and 2-9 do not show the arrangement of all the 

 famihes within the orders recognized by Bessey and Hutchinson, but 

 this information is included in their original presentations. It is im- 

 portant to remember that these systems, while agreeing in parts, are 

 contending hypotheses. The authors recognized this, for, as Hutchin- 

 son (1959) stated in the preface of his latest work. 



Botanical systems can never remain static for long, because new facts 

 and methods of approach are liable at any time to modify them. Like 

 other things in this changing world, that which seems to be a prob- 

 ability or even a certainty one day may quite weU prove to be a fallacy 

 the next. 



Diagrams of the type mentioned above enable the interested 

 worker to tell at a glance the presumed phyletic relationships within 

 the groups concerned; however, it cannot be overemphasized that 

 these are, at the most, hypothetical in nature and only in the rarest 

 instances are they free of gross oversimplifications. For the experienced 

 taxonomists such schemes may prove more irksome than instructive, 

 but to the systematically inclined organic chemist (possibly even for 

 speciahsts such as palynologists, embryologists, floristic cataloguers, 

 and so forth) they might provide some insight not apparent from the 

 more formalized monographic treatments. 



Parallelism as a factor in classification 



Grant (1959) attributes to two principal factors the main 

 responsibility for the differing generic treatments accorded the phlox 

 family (Polemoniaceae) by several workers on the basis of facts 

 available. These are: reticulate relationships following ancestral 

 hybridization and parallelism in evolution. As indicated in Fig. 2-10, 

 the two phenomena are often concomitant. Several workers have 

 felt that convergence and parallelism per se make it diflicult, if not 

 impossible, to erect meaningful phylogenetic classificatory schemes, 

 and some discussion of these phenomena will be included here. 



Parallelism may occur as a result of hybridization and sub- 

 sequent backcrossing (Fig. 2-lOb). This type of parallelism, whether 



