BOTANICAL HISTORY AND NOMENCLATURE OF SORGHUM. 43 



scribed b} T Linne. All the species are excellently illustrated on folio 

 plates, and the descriptions are more exact and complete than those 

 of any previous author. Arduino's paper had been presented in 1780 

 and his experiments had been in progress since 1775, and perhaps for 

 a longer period. He was probably the first author to grow and care- 

 fully study in the field all the forms he could obtain. His three new 

 species are Holcus cafer, H. cernuus, and II. niger. 



His Holcus cafer is the most interesting of all, because it was a 

 recent arrival in Italy from " Cafreria " (Natal), in South Africa, 

 which Arduino describes as " an exceedingly vast province of Africa." 

 How it was obtained is not stated. The description and figure show 

 a form with an umbellate panicle, the long, heavily seeded branches 

 drooping in the form of an umbrella (fig. 13). In appearance it is 

 identical with the drooping strain of Planter sorgo now grown in this 

 country, and which is of known Natal origin. We thus have our 

 Planter sorgo antedated by seventy-five years. Arduino's variety 

 was described as 8 feet or more in height, with stalks as large as 

 American corn, 12 or 13 leaves, and an umbrella-like panicle with 

 drooping branches 6 inches long. The heavy stalks, filled with sweet 

 juice, weighed three or four times as much as those of the ordinary 

 forms, and the stems and foliage remained green until frost, even 

 when the heads were harvested. The reddish seeds were considerably 

 exserted from the small hairy glumes. It is the Sorghum arduini of 

 Jacquin (Eclog. Gram., pi. 18, 1791) : the variety cafer of Koernicke 

 (1885, vol. 1, p. 307) ; and is doubtfully admitted by Hackel (1889, 

 p. 519). 



Arduino believed his Holcus niger to be the black sorghum of Pliny 

 and Tournefort. He illustrates (fig. 14) a loose, ovate-pyramidal 

 panicle not nearly so lax and spreading as that of Holcus sacchara- 

 tus. It is much like some of our more compacted Amber forms and 

 the Oomseeana of Wray. 



Arduino's Holcus cernuus (fig. 15) is our white durra, based on the 

 white-seeded variety of Arabia, etc., discussed by so many of the 

 older authors. Arduino is naturally puzzled by the action of Linne 

 in uniting this form with his Holcus sorghum, as he did in his 

 "Mantissa" (1771). 



In Arduino's interpretation of Linne's Holcus sorghum (fig. 12), 

 it is described as from 6 to 8 feet in height and an inch in diameter, 

 with 8 to 10 leaves and an erect, oval, compact panicle full of seeds 

 of various shades of red and yellow, some included and some partly 

 exserted, the lemmas being awned or awnless. Linne's Holcus sac- 

 charatus (fig. 12) Arduino describes as a tall, slender cane with long, 

 slender leaves and a sparse, lax panicle about a foot long, with droop- 

 ing branches, awned or awnless spikelets, smooth or hairy glumes, and 



175 



