500 



SCIENCE 



[N. S. Vol. XXXVIII. No. 980 



figured and described by him. The first at- 

 tempt comprehensively to treat Eumpf's 

 plants was by A. W. E. T. Henschel, who pub- 

 lished his " Clavis Rumphiana " in 1833, 

 pages xiv-1-215. In this work he attempted 

 to reduce Eumpf's species, so far as possible, 

 to modern binomial nomenclature. Thirty 

 years later J. K. Hasskarl, a Dutch botanist 

 having an extensive knowledge of the flora of 

 the Malayan region, published his " Neuer 

 Schliissel zu Rumpf's Herbarium Amboi- 

 nense," vi + 247 pages, originally printed in 

 the Ahhandlung der naturforschenden Gesell- 

 schaft, IX. (1866). Both of these works are 

 unsatisfactory for the chief reason that a 

 simple statement that a certain plate repre- 

 sents a certain species is frequently of little 

 or no value, especially when the species is 

 actually based on the plate, as is frequently 

 the case. 



In my work on the Philippine flora during 

 the past ten years I have come very fully to 

 realize that most of the species described by 

 Blanco in his " Flora de Filipinas," none of 

 ■which are represented by type material, can 

 he accurately interpreted only by an intensive 

 knowledge of the Philippine flora, as a whole, 

 and a very special knowledge of the vegeta- 

 tion of those regions from which Blanco se- 

 •cured his botanical material, taking into con- 

 ■sideration also habitats, dates of flowering and 

 fruiting, economic uses and native names, in 

 fact all data given by Blanco regarding each 

 individual species. In many cases one must 

 secure material from the actual localities 

 ■cited by Blanco, and our recent coUeetions 

 must be compared with Blanco's descriptions 

 not only as to the botanical characters given 

 by him, but all other data. Similarly I have 

 come to the conclusion that many of the spe- 

 cies based' on Rumpf's figures can be correctly 

 interpreted and understood only by an inten- 

 sive botanical exploration of the regions in 

 which Rumpf collected his material, and a 

 study of the specimens secured, taking into 

 consideration all the data given by Rumpf 

 and comparing it with data secured with bo- 

 tanical material from Amboina and neighbor- 

 ing islands. 



Many of the species based wholly or in part 

 on Rumpf's figures have been credited with 

 a wide Indo-Malayan range, but in some 

 cases, at least, the " species " are collective 

 ones. Many others are not understood at all 

 and appear in monographs as unrecognizable, 

 doubtful or imperfectly known forms. We 

 have in the Philippines many of the species 

 proposed by the older authors which are typi- 

 fied by Rumpf's figures, and in critical genera, 

 especially in those with numerous species, it 

 is frequently quite impossible definitely to 

 state which of our forms is the species based 

 on Rumpf, and which is a distinct but closely 

 allied one. The same principle holds true for 

 the entire Malayan region. 



In the case of many plants figured by 

 Rumpf, there is absolutely no doubt as to the 

 present status of such as the cocoanut, the 

 papaya, the tamarind, the mango, the beetle- 

 nut palm, and other well-known forms in 

 monotypic or small genera. The difficulties 

 arise in such genera as Calamus, Canarium, 

 Gnetum, Mucuna, Pandanus, etc., where spe- 

 cific differences are frequently not very great. 

 It is frequently quite impossible absolutely to 

 delimit the species from the figures and de- 

 scriptions given by Rumpf, and apparently no 

 serious attempt has ever been made to inter- 

 pret the species from actual Amboina speci- 

 mens. 



To illustrate this matter Mucuna pruriens 

 DC. is based on Dolichos pruriens L. The 

 original publication of Dolichos pruriens L. 

 is in Stickman's " Herbarium Amboinense " 

 (1754), 23, and is based absolutely and only 

 on Cacara pruiens Rumpf Herb. Amboin., 

 v., 393, t. lJf2; the question of specific iden- 

 tity of Mucuna pruriens is not complicated by 

 any additional synonyms m. the original pub- 

 lication of Dolichos pruriens. Most botanists 

 assign to the species a pantropical distribu- 

 tion, as did Linnffius in his later publications; 

 yet a simple examination of the material in 

 any large herbarium will at once show that 

 Mucuna pruriens is a " collective species," 

 and that specimens so named really represent 

 several more or less distinct species. No 

 botanist can definitely state that he actually 



