The genera of Sapindales in the southeastern United States

Recent reseach in the fields of floral morphology and anatomy, emi • • ■ j i mill':'...,. [■■ i. •:-., in.jiiiii'. .In. t In in! i ry, etc., has made possible more accurate determination of the systematic positions of many families and their arrangemenl into more natural affinity-groups or orders than was attainable at the time when the classical Englerian system was proposed.Since orders usually form the skeleton of every proposed scheme for a "natural system" n\ the angiosperms.the achievement of the greatest possible homogeneity seems to be of primary importance.With regard to this, the splitting of at least some very large and heterogeneous orders of the Englerian system (e.g., Geraniales, Sapindales), followed in some instances by rearrangement of the tannin imon-m i h< segregated orders, appears to be expedient.Such changes have been carried out more or less successfully in several recent systems (e.g., Hutchinson, 1926Hutchinson, & 1959;; Gundersen, 1950; Cronqnist.1057: Takhtajan, 1959).Although the sequence of a number of families of the southeastern United States which have been surveyed by the author in the Journal of the Arnold Arboretum (vols.43 & 44, 1962 & 1963) generally corresponds with that of Engler's system (1936), Takhtajan has been followed primarily in the delimitation of the orders.Thus, the families Rutaceae, Simaroubaceae, Burseraceae, and Anacardiaceae (as well as Meliaceae, unpublished) are regarded as members of the order Rutales; the families Sapindaceae, Aceraceae, and Hippocastanaceae (as well as several others occurring beyond he Arnold Arboretum and the Gray Herbarium of Harvard University made possible hrough the support of George R. Cooley and the National Science Foundation and nder the direction of Carroll E. Wood, Jr., and Reed C. Rollins.This treatment fol-)ws the pattern established in the first paper in the series (Jour.Arnold Arb.39: 96-346.1958) and continued through those in volumes 10-1 I ( 1') .The area overcd is bounded by and includes North Carolina, Tennessee, Arkansas, and Louisiana.The descriptions are based primarily on the plants of this area, with anysupplementary malm I in .1 ReliTem .inch the author has not seen are suggestions; to James W. Hardin, for his kiixlncv hi leadm*'.ilu-manuscript of Hip-Thomas, respectively, for fresh or preserved materials o! Sapindus and Dodonaea helpinthepreparat110.11 In. huhum 1 ipl ilu Jin lutnm