AN ACCOUNT OF THE GENUS ARGEMONE. 331 



Papaver alpiiiwn, one particular form of which was named by 

 Linnaeus Argemone pijrenoica. What the phmt was, however, is of 

 little consequence in this enquiry, and as Zinn, working only four 

 years later in Haller's own garden, was unaware what the plant 

 was, we can hardly hope ever to ascertain. We know enough, 

 however, to justify us in refusing to accept the citation here. 



The presence of a variety of this species in Polynesia, if var. 

 glauca be really, as I think, best referred here, is difficult to explain. 

 It is hardly likely that the form is truly indigenous in the Sandwich 

 Islands, and yet it is difficult to see why the plant characteristic of 

 the South-eastern United States should be that which has become 

 established in this particular archipelago. One would rather have 

 expected to find that the species present was characteristic of the 

 Pacific coasts of America ; indeed, Su' William Hooker has identified 

 it with the white-flowered Argemune of Chili. If this be really the 

 case, and it must be admitted that in habit and general appearance 

 it much resembles the Chihan plant — hardly more, however, than 

 it does the plant characteristic of the North American prairies — it 

 is remarkable that it should have assumed a form which, while 

 remaining true to itself in all the specimens I have seen, should 

 have diverged from the other Western American forms in the dis- 

 position of its bracts and in the shape of its sepals, and should in 

 these respects have assumed the features which characterise the A. 

 alba of the Eastern States, It is to be feared that we see in this 

 combination of characters an indication of the necessity for refusing 

 separate specific rank to any of the white-flowered Argemones, and 

 of recognising in the genus only three species — A. fruticosa, A. 

 viexicana, and a very variable A. alba. 



The Sandwich Island variety does not appear to have ever been 

 introduced into European gardens. The date of introduction of 

 the true A. alba is somewhat uncertain. Hornemaun states that it 

 was introduced to Denmaik in 1812, and it has been supposed by 

 some that it was not known till then in Europe, and that conse- 

 quently Lestiboudois's name could not apply here. This is of course 

 an unjustifiable deduction, since we know that the plant was in 

 Europe in the time of Jussieu and Lamarck, and was cultivated in 

 many French and Italian gardens prior to its description by Horne- 

 mann. It has also been generally supposed that it came from 

 Mexico ; Hornemann, however, is careful to indicate that its origin 

 was to him unknown. We know now that its Mexican origin is as 

 mythical as that of A. mexicana itself, and that it must have been 

 introduced from South-eastern America, where alone it occurs in a 

 wild state. The exact date I have not, however, been able to trace. 

 Before leaving A. alba, it is necessary to allude to another point. 

 In the accounts of the genus Argemone it has been usual to speak 

 of the perianth as variable. This statement requires some qualifi- 

 cation. In the very large number of specimens examined by me 

 in the various herbaria mentioned in the introduction I have only 

 seen three specimens with a 4-merous perianth. These have been 

 the specimens marked A. alba Juss. in Mr. Drake's herbarium, the 

 specimen at Ivew which is the type of the Bot. Mag. figure of 



