6 BULLETIN 318, U. S. DEPARTMENT OF AGRICULTURE. 



from Linnaeus 's original illustration (9) considerably, and one may 

 seriously doubt whether it is the same species. 



Two other illustrations of Dolichos lignosus, both colored, re- 

 semble Linnaeus's original much more closely, namely, Smith's (18) 

 and Curtis's (4). Each of these illustrations depicts a plant with 

 red flowers and with small leaflets, as in Linnseus's illustration (9). 

 The plant of Smith shows the leaflets acuminate, not acute, as 

 drawn by Linnaeus and by Curtis. Smith says the mature pods are 

 " an inch long, a little recurved, brownish, smooth." None of these 

 three illustrations agrees with any of the 60 varieties of the bonavist 

 grown at Arlington Farm, and they may well represent another 

 species. Indeed, Smith's plant may be distinct from that of Cur- 

 tis. The latter was considered a new species by Don (5) , who named 

 it Dolichos curtisii. Smith's plant may perhaps be Dolichos jac- 

 quinii DC. 



The opinions of many later botanists who did not haA^e much 

 first-hand knowledge on which to decide whether lignosus was 

 really a different . species from lablab may be disregarded. The 

 opinions of botanists who studied the flora of India are, however, 

 entitled to greater weight, inasmuch as both species are supposed to 

 be from that region. 



Roxburgh (15) differentiates Dolichos lablab from D. lignosus 

 as follows: 



Legumes horizontal, compressed, semilunar, with a straight 



scabrous back ending in a straight, daggered point lablab. 



Legumes linear, oblong, slightly incurved, torulose, both 

 margins turned and rugose, with a subulate, recurved 

 apex lignosus. 



Of the former, Roxburgh cites seven varieties, with three of 

 which he identifies plates 136, 137, and 141 of Rumphius (16). Of 

 the latter, he lists six varieties, with one of which he regards plate 

 25 in Kaempfer as identical. 



Plate 136 of Rumphius (16) has already been referred to. 



Plate 137 gives an oblique view of full-grown pods that apparently 

 are oblong, but probably are broadest toward the apex, as in ordi- 

 nary varieties of the bonavist. 



Plate 141 is clearly a species of Canavalia. 



Plate 25 of Kaempfer (7) has the immature pods falcate, while 

 the mature ones are nearly straight and not broader toward the apex ; 

 in other words, they are linear and not different in form from those 

 on plate 136 of Rumphius (16), except that the tip is incurved. 

 Judging from the plates cited, Roxburgh seems to have given great 

 weight to the form of the tip of the pod, but this character is of 

 little significance. In our opinion, plates 136 and 137 of Rumphius 

 (16), as well as plate 25 of Kaempfer (7), all clearly represent the 



