• 8 J. Pfrkins: 



^enus to the Tlujmelaeaceae , subdivision Gyrinopeae. Baillox, in 

 Hist, des Plantes VI, p. 103 (1877) has classed it with the Aiini- 

 lariees. Both Miquel and Baillon state that tlie genus Go7iysti/his 

 is distinguished by a very short, campanulate receptacle with 5 al- 

 most valvate sepals, petals Avanting, very numerous (35) iiliform 

 squamae inserted on the edge of the receptacle, 10 stamens in- 

 curved in the bud, peltate, basifixed anthers, a 4 — 5-carpelled ovary 

 with one suspended ovule in each carpel, a very long curved style 

 and a very thinly fleshy drupe (the size of an orange) containing 

 4 — 5 large, exalbuminous seeds. In Genera Plantariim, III, p. 205, 

 Bentham says, on the contrary, that tlio stamens are indefinite 

 ("stamina plurima") and that the same are "in fundo calycis circa 

 ovarium seriata". In order to make a statement so contrary to any- 

 thing before recorded of the genus, it must be taken for granted 

 that Bentham must have made careful investigation, especially as Miquel 

 states "stamina ni fallor 10" and in his plate (fig. 3) gives a trans- 

 verse section through the flower in which 9 — 10 anthers are easily 

 recognizable. Moreover, Hooker states that Gonystylus Maingayi 

 Hook, f has numerous stamens. This, Gilo, after a careful exami- 

 nation of very abundant material, was able to confirm. Van Tieouam 

 in Ann. Sc. Nat. VII. ser. vol. XVII, p. 240 (1893), has studied the 

 anatomy of Gonysiyhis and upon anatomical differences has sepa- 

 rated the two genera Asclmim and Amyxa. 



Ga.G, in Engl. Bot. Jahrb. XVIIl, p. 517 (1894) was the first 

 to exclude Gonystylus from the Thymelaeaceae. He says that, in the 

 Thymelcieaceae, even in the whole tribe Thymehicalcs, not a single 

 case of indefinite flowering organs occurs, and that in the perfect 

 flower we have, at most, two rows of stamens. Also that a typical 

 receptacle is wanting, "stamina in fundo calycis circa ovarium 

 seriata". Furthermore, according to Solkhkdkr, Syst. Wert der Holz- 

 structur p. 232, no intrahadntmatic Icptom (hicollateral vascular bundles) 

 occurs, this ciiaracter distinguishing the TJiyniclaeales from man}- 

 other families, and in the leaf no secreting cells are known. These 

 morphological and anatomical differences make it apparent that Gony- 

 stylus must be excluded from the Thynielucaccae. Gn.G writes that 

 he is not certain to what family Gonystylus is related, but tiiat on 

 account of some of the above mentioned characters, he believes it 

 may perhaps be ascribed to the Tilt'urrur. Later, in Nachtr. Kngl. 

 und i'nintl. Pflaiucniiii., p. 231 (1897), he bases upon the genus 

 a new family (Gonystylaceae). At present Gii.o places the (iony- 



