6 NATIVE BRITISH ORCHIDACE/E 



shown on PI. 39 B, in which a flower of O. morio with three Hps is figured. This 

 conversion of both petals into hps, as also the reversion of the hp to a petal, giving 

 rise to three lips or three petals, is called peloria. A spike of Orchis purpurea, found 

 near Chambery, had sepals which remained closed, and a lip exactly like the petals 

 (PI. A, fig. i), which moved down to a horizontal position forming a platform on 

 which insects could alight and find access to the flower. The flower with spreading 

 sepals was forcibly opened to show its construction. PI. 2, fig. 3, shows a spike of 

 Ophrjs arachnitiformis found near Hyeres, France, in which the lip has reverted to a 

 petal, and, like the sepals and petals, is wliite with a green nerve, a striking example 

 of peloria. PI. 2, fig. 2, shows the typical flower of tliis species (enlarged), which 

 has pink or white sepals and petals, and a dark purple lip with a metaUic shield. 

 These are not accidental malformations, but throw-backs to a remote ancestor. 



The evidence that the orchid flower originally had six anthers is apparently con- 

 clusive. On each side of the base of the column in some genera is a little laiob called 

 a staminode, which is the rudiment of a suppressed anther. ' Usually it is the anthers 

 a^ and a'^ (Text-fig. 2) which are thus represented. PL 2, fig. 8, shows a much-enlarged 

 column of Cephalantheragrandiflora in which these obsolete anthers were fully developed 

 and had pollinia (from Bauer's Illustrations oj Orchidaceous Plants). I found a somewhat 

 similar specimen at Hyeres in 1920. Even in the most higlily specialised genus, 

 Ophrjs, cases occur, though very rarely, of partial reversion to a 6-anthered flower. 

 PI. 2, fig. 5, shows an enlarged column of Ophrjs arachnitiformis from Hyeres with 

 three anthers developed. In another spike of the same species also found there, no 

 less than four anthers were developed, three being the same as in PL 2, fig. 5, and the 

 fourth being the anther a'i {vide Text-fig. 2), face to face with the usual fertile anther 

 A'^, and so close that their beaks overlapped! The development of the anther a^ is 

 of the most extreme rarity, no doubt because it was the first to be got rid of when the 

 hp became a landing-stage, as owing to its position it obstructed access to the 

 rostellum by insects, and hindered cross-poUination. It was so in this case, the 

 stigma being a cup surrounded by the four anthers. A beautiful example of the 

 development of the anther ai is shown in PL 58 C,^ from a spike of Ophrjs apifera sent 

 me from Dorset by Sir Maurice Abbot Anderson in July, 1930. PL 2, fig. 4, taken 

 from the Journal of the Linnean Societj, 1865, PL IX, fig. 18, shows a column (enlarged) 

 of a species of Isochilus common in Trinidad, in wliich no less than five anthers were 

 developed, and the sixth (^3) represented by a filament. Had the latter developed its 

 anther, itwould have been a case of complete reversion to theancient 6-anthered form3. 



' PL K, fig. I, A3 shows a staminode {st) abnormally developed, with a yellowish rudimentary 

 pollinium (^) visible within. 



' p. 237-. 



3 M. Maire {Kevue gen. Bot. p. 516 (1909)) mentions a case of Orchis militaris and also of Ophrjs 

 tenthredinijera found near Algiers with six anthers. 



