584 



THE AMERICAN NATURALIST [Vol. XL V 



0. grandiflora and 0. Laniard- inn a rather than in 0. 

 biennis, but as already stated, an argument of this nature 

 can not reach finality. Whether Morison's Fig. 7 repre- 

 sents one of the grand iflora-La ma rekiana series or the 

 "European biennis," it seems certain that Barrelier's 

 species (3) came in the large-flowered, open-pollinated 

 series represented by 0. grandiflora or 0. Lamarckiana. 



Unfortunately, the actual specimens in the British 

 Museum, which bear these early names of Tournefort 

 and are supposed to have served as the types for the 

 "Hortus Cliffortianus," are not fully authenticated. 

 The handwriting is said not to be that of Linnaeus, and 

 certain differences between the names employed and 

 those which Linnaeus himself would probably have used, 

 seem to indicate that they were written by an amanuen- 

 sis or that some mistake occurred in the naming. This 

 deprives us of certainty in regard to the names they 

 bear. The characters of the specimens themselves are 

 very well preserved. The first specimen (which I shall 

 refer to as specimen 1) bears on the label the name 

 "Onagra latifolia fore sulphureo"; the second speci- 

 men (2), "Onagra latifolia. T. 302," and beneath it is 

 written " Oenothera octovalvis." But 0. octovalvis was 

 a species of Jussieua. These two specimens are identi- 

 cal in every character. They represent evidently a race 

 of the "European biennis" having larger flowers than 

 American forms of 0. biennis, though not so large as 0. 

 Lamarckiana. The style is short so that the stamens 

 surround the stigma, the buds bear some long hairs, red 

 papillae occur on the stem, and the leaves are rather 

 broadly lanceolate. The stigma lobes are remarkably 

 long in all the flowers on both specimens. These two 

 specimens mav have served as the types of Linnaeus 's 

 species (1) in Hort. Cliff, (see Gates, '11a, p. 102). As 

 far as the characters of these specimens are concerned, 

 they really resemble 0. Lamarckiana more than they do 

 the American races of 0. biennis, but are classed with 

 0. biennis because the flowers are self-pollinating. 



