lum and Clematepistephium, 

 i Neuwiedia Griffithii and N. 



ifolia through the reduction in the numb 



integument 



some 



steps in an ancient ontogeny which are pointing to that direc- 

 tion, and which have remained unaltered in the cul-de-sac of 

 saprophytism. Such possible steps may be changing from ovoid 

 seeds (Vanilla) to lenticular ones with a cellular edge (Cyrtosia 

 javanica) to several-layered keels (Galeola septentrionalis), to 

 truly alate seeds (Eriaxis, Epistephium, Clematepistephium). 

 The step (from here ?) to the one-layered, tunicate testa is, how- 

 ever, a major one which remains yet to be demonstrated. 



As a matter of fact, the recently published new information 

 about certain orchids with sclerotic seeds stimulated me to 



fam 



time 



m 



vided by George and Cook in 1981. Although differently sculp- 

 tured, the seeds are those of the Vanilla type, and so are also the 

 fleshy, indehiscent fruits with parietal placentation; these fruits 

 also have a scent like Vanilla (Anonymous 1982). Both Veyret 

 (1981) and Dressier (1983) have published excellent photographs 

 of the cross sections of ovaries of various Palmorchis species 

 revealing an axile placentation similar to that of Apostasia, 

 while the sclerotic seeds are also of the Vanilla type. These bits 

 of new information shed more light on their actual phylogenetic 

 affinities than the various speculations already offered in print. 

 These papers also prompted me to survey all the genera and 

 species which have been attributed in the broadest sense to the 

 relationship of Vanilla, Palmorchis and Rhizanthella in the past. 

 It must be emphasized that in any reliable systematic, evolu- 



examination 



m 



important information. All of the so-called systems, including 

 the recently published cladistic surveys, are based only on what- 



225 



