1960] 



MORPHOLOGY AND RELATIONSHIPS OF DIOMMA 



5.3 



Regarding aestivation of sepals, we can see no insurmountable obstacle to our 

 scheme, for both types have been recorded in Rutaceae. Engler (1931) mentions 

 aestivation of sepals in Rutaceae as mostly imbricate. Eichler (1878), in his 

 analysis of rutaceous flowers, mentions that open, valvate, and imbricate aestiva- 

 tion occur in almost all subfamilies of Rutaceae. For example, sepals are valvate 

 in Calodendron and Melanococca according to Bentham and Hooker (1862) ; 

 valvate in Bauerella and mostly valvate in Boronia sect. Yalvatae Benth. (Engler 

 1931). It appears then that in Rutaceae, aestivation of sepals is at most a char- 

 acteristic of generic significance and that related genera may differ in this 

 regard. 



A few original observations on aestivation in sepals of Spathelia indicate 

 that, contrary to reports in the literature, aestivation in this genus may be 

 valvate | e.g. S. glabrescens Planch, and 8. subintegra Marie-Vict.) or imbricate 

 (e.g. S. cubensis P. Wils. and S. simplex L.). Other authors have used the term 

 "subvalvate" in referring to aestivation of sepals in Spathelia. Such a term 

 is likely to be confusing a nd certain to be subjective. 



In Engler 's general statement of the characteristics of Rutaceae (1931), he 

 notes: "Somen . . . mit oder ohm Nahrgewebe." A survey of Engler's data 

 indicates that within subfamilies of Butaceat differences in presence or absence 

 of endosperm are found. A few examples may be pertinent: (1) In the sub- 

 family Rutoideae, most members of the tribe Xanthoxyleae possess exendo- 

 spermous seeds, whereas in the tribes Ruteac and Boronieae members possess 

 endospermous seeds; (2) in the subfamily Toddalioideae, members of the tribe 

 Toddalieae may or may not have endosperm. Thus, the presence or absence of 

 endosperm in the seeds of Rutaceae does not seem to be a constant characteristic 

 of a subfamily and is probably merely of generic importance. 



From the above, the authors conclude that neither the differences in sepalar 

 aestivation nor the presence or absence of endosperm in the three genera under 

 consideration represents a valid criterion against the erection of the subfamily 

 Spathelioideae to contain Diomma, Sohnreyia, and Spathelia. In regard to our 

 proposal of Spathelioideae sensu novo, Erdtman (in correspondence) states. 



. . pollen morphology does not speak against your view of referring Sohn- 

 reyia, Diomma. and Spathelia' to the Rutaceae, as a special subfamily." Spath- 

 elioideae, as constituted herein, differs from all other subfamilies of Rutaceae 

 in the presence of apotropous ovules, much as the Cedfeloideae differ from other 

 subfamilies of Meliaceae. 



It appears that within this subfamily, Spathelia is more primitive than either 

 Diomma or Sohnreyia. Several conditions in Diomma and Sohnreyia seem to 

 represent reductions or specializations from that which obtains in Spathelia. 

 For example, although Spathelia regularly has tricarpellate ovaries, both Diom- 

 ma and Sohnreyia regularly have bicarpellate ovaries. Tricarpellate ovaries do 

 appear sporadically in the latter genera, which may indicate a reduction from 

 that state. Seeds in Spathelia are endospermous, whereas endosperm is absent 

 in the seeds of Diomma and Sohnreyia. Wings on the fruits of Spathelia are 

 relatively undeveloped, which may indicate a primitive condition. Fruit wings 

 are broad in Diomma and Sohnreyia. There also seems to be a lack of stability 

 in certain features of Spathelia, which again may be interpreted as a state of 

 primitiveness. This is found in the inconstant type of sepalar aestivation, and 

 in the wide variety of form in staminal processes. It may also be true that the 

 well-developed secretory glands, such as are found in carpels of Spathelia, are 



