50 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN' [VOh. 10 



and it is probable that an abnormality exists in the Diomma specimen and/or 

 it does not exhibit mature characteristics. 



Table 1. Dimensions of tracheary elements. 



Dimnma 









Mean 



Eange 



Fiber-tracheid length 



786 fi 



405-1200 fx 



Pore diameter 



63 /a 



29-104 n 



Vessel element length 



538 m 



300-825 ix 



Angle of vessel element end wall 



54° 



75-30° 



Sohnreyia 





Fiber-tracheid length 



901 fi 



450-1170 ix 



Pore diameter 



97 fi 



40-165 ix 



Vessel element length 



509 fx 



360-780 ix 



Angle of vessel element end wall 



54° 



80-10° 



Table 2. Pore distribution in per cent. 



Solitary Radial multiples Clusters Chains 

 Diomma 36.4 53.0 10.6 few 



Sohnrcyia 34.3 60.0 5.7 few 



DISCUSSION 



Two species of Diomma, D. ulei and D. friiticosa, and Sohnreyia excelsa have 

 been validly described in the literature and represent distinct biological entities. 

 In the course of our researches, we were able to examine more specimens assigned 

 to each of these genera than were available to previous workers. Examination 

 of these specimens brought to light only the single heretofore described species 

 of Sohnreyia. However, besides the two previously described species of Diomma, 

 material of at least two further species was present among our specimens. In 

 this work these have been designated for convenience Diomma species A and 

 Diomma species B. Collections representative of these are listed at the end of 

 the section on methods and materials. 



Relationships of Diomma. 



It is the contention of the authors that the characteristics of Diomma defi- 

 nitely indicate its placement among the Rntaceae. The general structure of 

 flowers and fruits is rutaceous in all respects except for the presence of apo- 

 tropous ovules. This latter feature, however, also occurs in the genera Spathelia 

 and Sohnreyia, both of which have been included in Rutaceae by Engler (1931). 

 IIow r ever, Harms (1931a) considered the presence of apotropous ovules as a 

 characteristic that argued against joining Diomma with any of the families of 

 Geraniales (sensu Engler). 



If we were to extend Harms' thesis on ovule position in Diomma to other 

 groups and exclude certain taxa on the basis of this single characteristic, many 

 unnatural relationships would result. Consider Rhamnales: Rhamnaceae possess 

 epitropous ovules, but Vitaceae and Leeaceae are distinguished by apotropous 

 ovules (Suessenguth 1953). Most Meliaceae are epitropous, but C edreloideae 

 are apotropous; most Simaroubaceae are epitropous, but Alvaradoideae are 

 apotropous. Within other families, ovule position varies among the genera. This 

 is true for example, in Tkeaceae (Melchior 1925) and Guttiferae (Engler 1925). 

 Apparently the taxonomic significance of ovule position, as well as of other 



