644 
Telopea Vol. 6(4): 1996 
only Dasypogon and Kingia, it should be recircumscribed to include Calectasia and 
Baxteria, but in two informal groups (Dasypiogon/Calectasia and Kiiigia/Baxteria). From 
rbcL analysis (Chase etnl. 1995a) Dasypogonaceae are among the first-branching families 
of the commelinoid clade (Fig. 6), close to palms. In the combined analysis of Chase et 
al. (1995b) they are together with the palms and Hanguana. However, relationships 
within the commelinoid clade are tentative at this stage, as there are many missing 
data. Habit and method of growth would support a relationship with palms, but ovule 
structure and silica position and morphology indicate an affinity with Rapateaceae. 
Xanthorrhoeaceae Dumort. (1829) 
Sensu Dahlgren et al. (1985). Genera included: Xanthorriwea Sm. only 
Xanthorrhoea is taxonomically isolated and correctly placed in a monogeneric family 
Xanthorrhoeaceae (Dahlgren et al. 1985), with obscure relationships. It differs from 
the other genera of Xanthorrhoeaceae sensii lato in several respects, such as number 
of ovules per ovary locule (Table 3) and and presence of a hypostase (Table 4), 
although in both of these respects it has the (probable) plesiomorphic asparagoid 
conditions. On the basis of inflorescence morphology, Waterhouse (1967) considered 
it close to Agavaceae, which it also resembles in vascular bundle structure (Fahn 
1954), and presence of an STM. However, these characters are probably all 
homoplasious in Asparagales, at least to some extent. Molecular data (Chase et al. 
1995a) put Xanthorrhoea in a phormioid-asphodeloid clade (Fig. 6). It is usually 
sister to Asphodelaceae, which includes genera with an STM such as Aloe, Asphodelus 
and Bulbine (Fig. 5), or sometimes sister to the phormioids, a (mainly) Phormiaceae- 
Johnsonieae group, including Caesia, Dianella, Dri/mophila, Geitonaplesiiim, Hemerocallis, 
Hcnsmannia, fohnsonia, Phormium, Pasithea, Stawellia and a few other (largely 
Australian) genera, lacking an STM. Xanthorrhoea differs from most of the other 
phormioid-asphodeloids in having successive microsporogensis (Rudall, 
unpublished), in which respect it resembles 'higher' asparagoids. 
Lomandraceae Lotsy (1911) 
To be recircumscribed. Genera included: Acanthocarpiis Lehm., Chamaexeros Benth., 
Lomandra Labill., Roinnalda P.F. Stevens, Xerolirion A.S. George and several other 
genera (Chase, Rudall and Conran in prep.). 
The exact circumscription and relationships of the Lomandra-group require further 
analysis, since some genera are unknown for some of the critical characters. On the 
basis of similar leaf anatomy (Fahn 1954, this paper) and sulculate pollen (Chanda 
& Ghosh 1976) Acanthocarpiis and Chamaexeros seem to be closely allied. Acanthocarpiis, 
Chamaexeros, Lomandra and Romnalda share similarities in leaf anatomy, such as 
enlarged outer bundle sheath cells and sclerenchyma girders from the inner bundle 
sheath (Table 5); however, these characters occur elsewhere and may well be 
homoplasious. On present evidence, the family Lomandraceae should therefore 
include at least Acanthocarpiis, Chamaexeros, Lomandra and Romnalda (not Baxteria). 
Other characters, such as nucellus structure (Rudall 1994 and unpublished), presence 
of an STM (in Lomandra and Thysanotns: this paper and Rudall 1995) and rbcL (Chase 
et al. 1995a) further link Lomandra with the arthropodioids (Fig. 6): Arthropodium, 
Cordyline, Chamaescilla, Eiistrephus, Sowerbaea, Thysanotns and Trichopetalum, although 
there are still gaps in the data sets. Several of these genera (e.g. Eiistrephus, Sowerbaea, 
Thysanotns and Trichopetalum) have markedly tuberous roots. Eiistrephus, Thysanotiis 
and Trichopetalum all have fimbriate inner tepals. The family Lomandraceae should 
be recircumscribed to include all of these genera (Chase, Rudall & Conran in prep.). 
The relationships of Xerolirion, a monotypic genus from southern Western Australia, 
have always been problematical. It has sessile, reduced leaves and female flowers are 
