182 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN [Vol. 9, No. 2 



basis of correlated morphology, relationships, and distribution as discussed 

 below. The 20 species recognized here include 9 subspecies and 12 varieties for 

 a total of 36 taxa. 



A subspecies is differentiated from a variety in this work on the basis of fixed 

 variations and correlated distributions. If the population concerned occupies a 

 range essentially distinct from the balance of the species, it has been treated as 

 a subspecies. On the other hand if it is morphologically but not geographically 

 distinct, it has been treated as a variety. 



MORPHOLOGICAL CONSIDERATIONS 



Taxonomic investigation of a common, frequently collected genus such as 

 Sagittaria offers ample opportunity for study of variations in populations. This 

 study illustrates the point that the number of proposed taxa of variable organisms 

 may vary inversely with the frequency of collections. That is, there is a tendency 

 for infrequently collected plants found in widely scattered areas to appear distinct 

 until such times as collected intermediates show them to be no more than the 

 extremes of a variable population. 



In order to avoid the possibility of consciously or unconsciously setting up 

 preconceived values for the various characters observed, each specimen was 

 examined and analyzed in detail without attempting to force it into any taxon. 

 These analyses were recorded on individual work-sheets and only after more than 

 2000 specimens were examined was any attempt made to separate them into groups. 

 In this manner it soon became apparent that certain characteristics were com- 

 pletely useless in themselves to segregate the various taxa. Even those that were 

 useful could not in themselves, except in rare cases, safely identify a plant. 

 Only after the author had correlated several different morphological characters 

 with both distribution and habitat did the chaos of variability gain a semblance of 

 order. 



With the framework thus laid down, each new specimen .that was examined and 

 analyzed helped to complete the generic and specific structure by filling in miss- 

 ing gaps or extending the over-all picture. 



The following survey reviews the morphological variation and attempts to 

 assess its value in the taxonomy of the genus. 



Habit. The entire genus is herbaceous. Most species are perennial; the few 

 individuals that seem annual probably represent ecological responses of warm- 

 temperate species to a colder climate and a shorter growing season. Perennation 

 is by rhizomes, runners, and corms. The rhizome is mostly short and thick, usually 

 producing a single crown of leaves each growing season; very rarely however, 

 leaves are produced along its entire length,, The runners or stolons are extensive 

 noded branches that generally terminate in a leafy bud. These buds can success- 

 fully winter over, or, more frequently and depending on climatic conditions, imme- 

 diately form another plant. In temperate areas many species develop winter storage 

 organs instead of buds on the stolon. These storage organs develop from the last 

 two internodes of a runner (Arber 1920, p. 15). They do not exactly fit the accepted 

 definitions of any of the standard names for storage organs and have been called 

 bulbs, tubers, and corms by different authors. Following Mason (1952), the author 

 has considered them to be corms. 



Leaves. These organs are in every respect the most variable structures of the 

 genus. All have sheathing, cellular petioles. Depending on edaphy these leaves 

 may grow submerged, floating, or emersed. The type of leaf is a response to the 



