76 
Endlicher Endlicher (Genera Plantarum 843) followed closely 
1839. De Candolle’s classification of the family, recognising 
a number of small genera, and remarking on the 
artificiality of the tribe Paeonieae, which he rather increased by 
the inclusion of T'rautvetteria. 
Bentham & Hooker Bentham & Hooker (Genera Plantarum 
i. 1) adhered closely to the arrangement of 
De Candolle. They placed the so-called 
spurious genera Actaea and Xanthorrhiza in the subtribe Cimici- 
fugeae of the tribe Helleboreae and wisely limited Paeonieae to 
the genus Paeonia. 
Baillon Baillon (Histoire des plantes i. 1.) classified Ranuncu- 
1867.  laceae in four tribes I. Aquilegieae, Il. Ranunculeae, 
Ill. Clematideae, IV. Paeonieae, beginning with the 
genus Aquilegia and ending with Crossosoma, the latter now being 
regarded as a distinct family, the Crossosomataceae, more closely 
related to the Dilleniaceae and Rosaceae. 
Prantl Prantl (Engler & Prantl, Natiirl. Pflanzenf. iii. 2: 54) 
188 ivided the family into three main tribes, I. Paeonieae, 
I. Helleboreae, 111. Anemoneae. 
By over-emphasizing the importance of the relative lengths 
of the outer and inner seed coats, he associated in the tribe 
Paeonieae such divergent genera as Glaucidium, Hydrastis and 
Paeonia, thus rendering the tribe a most unnatural one. Another 
character used by Prantl for further subdividing his tribes into 
groups, namely, the presence of one or two integuments in the 
ovule, led to further artificiality, such genera as Thalictrum and 
Adonis being associated by this character alone. Further, 
Prantl’s reductions of several distinct and well-defined genera 
such as Knowltonia to Anemone, Naravelia to Clematis, and 
Cimicifuga to Actaea, have not been generally accepted. 
From our point of view there is little evidence of a phylo- 
genetic arrangement in any of the above classifications, De 
Candolle’s system beginning with the highly evolved and reduced 
genus Clematis, and Baillon’s with the highly specialized genus 
Aquilegia. There appear to be two distinct phylogenetic groups 
in the family which for convenience may well be regarded as 
subfamilies. Of these the structure of the flowers of the Helle- 
boroideae, and especially in regard to its gynaecium, is probably 
the more ancient, the often green leaf-like carpels corresponding 
to our conception of a primitive type. In this subfamily there 
is a remarkable range in floral structure from an apparently 
primitive, genus such as Helleborus to a so highly evolved and 
specialized zygomorphic type of flower as is found in Aconitum 
and Delphinium. The best position for the Paeonieae remains 
a problem. I consider it to be an isolated type and have placed 
it first, as it shows no clear connection between the two other 
tribes. The second subfamily, Ranunculoideae with its achenial 
