ORCHIDACE.E 



to rise above the stigma as a naked mass. In flowers that are ex- 

 amined at the time of anthesis it is possible to detect a very thin 

 membrane which is continuous with the body of the gynoste- 

 mium. To this membrane the polhnia are in part agglutinated. 

 That this membrane is deeply bifid, as is the case in related spe- 

 cies, is by no means clear. In fact the use of this membrane in 

 generic classification would, I am sure, be regarded as reprehen- 

 sible by cautious investigators. Unless one were pretty well ac- 

 quainted with the more delicate characters of the flower in Spi- 

 ranthes, it would be natural to suppose that development of the 

 rostellum in Spiranthes novaezelandiae had been suppressed. 

 However interpreted, the gynostemium does not bring the spe- 

 cies within the proper genus if Schlechter's monograph of the 

 Spiranthinae is depended on for identification. Indeed, the con- 

 dition of the rostellum removes S. novaezelandiae from the group 

 in which Schlechter placed it. Gattungsreihe I of Schlechter's sys- 

 tem, in which S. novaezelandiae is included under Spiranthes, is 

 defined as follows : " Gattungen, bei denen die Klebscheibe der 

 PoUinien zwischen den Fortsatzen des verlangerten, zweispalti- 

 gen Rostellum festgehalten, resp. eingeklemmt ist." Under Spi- 

 ranthes, the first genus in the group, the column is described as 

 follows: *'Columna brevis, pede brevi, apice incurvulo; rostello 

 alte bifido, cruribus erectis, subulatis." And yet in Schlechter's 

 notes under Spiranthes novaezelandiae he tells us that the column 

 in the flowers he has examined had not developed a rostellum. 

 By the preceding remarks it is not my purpose to demonstrate 

 that Schlechter's system is open to general condemnation. I have 

 found it a very helpful effort toward clarity in a highly technical 

 group. One would regard it more favorably, however, if the new 

 genera proposed and some of the genera reinstated had been kept 



as sections or as tentative groups, on which to build up a more 



[ 129 ] 



