of specific epithets came almost as an after-thought: as 
late as 1748, Linnaeus had not added them to his manu- 
script draft now in the library of the Linnean Society of 
London. The main purpose of the Species Plantarum was 
to provide diagnostic phrase-names (polynomials), such as 
Cypripedium radicibus fibrosis, foltis ovato-lanceolatis cau- 
linis and Cypripedium bulbo subrotundo, folio subrotundo 
radical, which, taken together, functioned like the con- 
trasts in a modern key and enabled the species to be dis- 
tinguished from their congeners, with relevant synonyms 
cited as a guide to the existing literature about each. 
Hence, the material, either specimens or illustrations, sup- 
plying the information epitomised in the phrase-name, is 
of first importance for the typification of the associated 
specific epithet. Hach phrase-name set out to give as con- 
cisely as possible a statement of the main distinguishing 
features. Thus, in his synopsis of the genus Orc/is in the 
Species Plantarum (1: 989-944), the phrase-names of spe- 
cies 1-10 begin with ‘bulbis indivisis’ (‘with undivided 
root-tubers’) and so distinguish them from species 11-14 
with ‘bulbis palmatis’ (‘with palmately lobed root-tubers’ ) 
as well as from species 15-17 with ‘bulbis fasciculatis’ 
(‘with bunched fibrous roots’). A number of species are 
further distinguished by the length of the spur in relation 
to the ovary, e.g. species 12. (maculata) with ‘nectarii 
cornu germinibus breviore’ (‘with spur shorter than the 
ovary ) being thereby distinguished from species 18 
(conopsea) with ‘nectaril cornu setaceo germinibus lon- 
giore’ (‘with spur setaceous longer than the ovary’) and 
species 14 (flava) with ‘nectarii cornu filiformi longitu- 
dine germinis’ (‘with spur filiform as long as the ovary’). 
A name such as that for species 5 (pyramidalis), Orchis 
bulbis indivisis, nectaru labio trifido antice bidentato, cornu 
longo, petals acuminatis served to distinguish this spe- 
cies from all other orchids then known. 
[ 78 | 
