375 
Adenoclirieae of South Africa. 
androgyna it is arguable that M. androgyna, Roy. (FI. Leyd. Prodr., 203), 
the African plant of Hermann and Boerhaave, may have differed from 
M. androgyna , Linn. (Virid. Cliff*., 98). and that the latter, since Linnaeus 
says so, really was the Surinam Ricinokarpos of Boerhaave. This is to 
contend that we must concede that Linnaeus knew his own species, and 
that therefore we must accept his identification. The argument is plausible, 
and in its favour is the consideration that it influenced an author so careful 
as Muller, who (DC. Prodr., xv. 2,. 798) has quoted M. androgyna , Linn., as 
the equivalent of ‘ Croton Ricinocarpus, Linn, quoad syn. Boerh. et patriam 
Surinamiam ’. But the argument is less weighty than it appears ; it can 
only be employed at all if the period at which Linnaeus’s recollection was 
most vivid be treated as negligible. The word ‘ androgyna ’, as used by 
Linnaeus in 1737, is not a nomen triviale in the sense conveyed by the 
specific epithets first employed in 1753 ; it is only an apt condensation of 
the note appended to the generic definition of 756 Mercurialis (Gen. PI., 
ed. 1, 307) of the same year. The association of Royen with Linnaeus in 
1 737 was so close that even under ordinary circumstances Royen was 
unlikely to have used the name M. androgyna in 1740 for a plant other 
than that so named by Linnaeus only three years before. The circumstances 
in this case were, however, unusual. In 1737 Linnaeus found himself in 
a difficulty through his having failed to observe what Burmann did observe, 
that Ricinokarpos , Boerh., is an ancipital genus. That difficulty centred 
in the statement by Linnaeus that the plant which is Mercurialis androgyna , 
Roy., was also the basis of Ricinokarpos. Later in the year, Linnaeus 
adopted a suggestion made by Royen which overcame the difficulty. The 
acceptance of Muller’s identification involves, then, the assumption that by 
1740 Royen had fallen into error as regards the name applied in 1737 by 
Linnaeus to the plant to whose existence the difficulty that Royen had 
solved was due. The intrinsic improbability of Muller’s hypothesis is 
heightened by the existence of a Cliffortian specimen which certainly is 
M. androgyna , Roy. (1840), and must either be M. androgyna , Linn. (1737), 
or be a plant from Clifford’s garden for which Linnaeus failed to account. 
An examination of the contents of the Mercurialis cover in the Linnean 
herbarium disposes of the suggestion that the recollection of Linnaeus as to 
the identity of his M. androgyna of 1737 was clearer in 1763, when he 
referred it to Croton as C. Ricinocarpos , than it was in 1753, when he 
referred the same plant to Mercurialis as M. procurnbens. 
In that Mercurialis cover five species are represented. Three of these 
have been written up by Linnaeus himself as — 1, perennis ; 2, annua ; 
3, tomentosa. These represent respectively the three species so named in 
*753 (Sp- PL, ed. 1). Another has been written up by Linnaeus, without 
a serial number, as androgyna , this word being afterwards struck through 
by him and the name ambigua substituted. The reason for this change is 
