Hyacinthaceae

Pfosser, M., Wetschnig, W. & Speta, F., 2006, Drimia cryptopoda, a new combination in Hyacinthaceae from Madagascar, Linzer biologische Beiträge 38 (2), pp. 1731-1739 : 1731-1739

Based on plastid DNA sequences a clear separation among the four subfamilies of Hyacinthaceae is supported by bootstrap values ≥ 75% in the phylogenetic analysis ( Fig. 1 ).

Within Hyacinthoideae the classification into the monotypic South African tribe Pseudoprospereae , the north hemispheric, Eurasian and North African tribe Hyaciantheae and the tribe Massonieae with species occurring south of the Sahara, the Arabian peninsula, Madagascar and India, as suggested by MANNING et al. (2004) is also supported. The generitype of Hyacinthus , H. orientalis, forms a highly supported clade (99% bootstrap support) with the north hemispheric genera Puschkinia, Prospero, Hyacinthella, Fessia and Zagrosia and no affinity to any of the members of Massonieae is found. Contrary to MANNING et al. (2004) who placed Hyacinthus cryptopodus based on morphological grounds into the genus Ledebouria of Massonieae , our phylogenetic analysis does not support this decision. Instead, based on plastid DNA sequences, H. cryptopodus clearly falls within subfamily Urgineoideae and appears to be closely related to other Madagascan members of subfamily Urgineoideae like Drimia mascarenensis and D. urgineoides (84% bootstrap support).

In their generic synopsis MANNING et al. (2004) provide a key to the genera of Hyacinthaceae in sub-Saharan Africa and start with the dichotomous split of whether or not the bracts are spurred:

1a. Bracts (at least the lower) usually spurred, sometimes deciduous;.... (subfamily Urgineoideae ) ............................................................................................................ 2

1b. Bracts never spurred.... ............................................................................................... 4

(1a) correctly leads further to the genera Drimia, Igidia and Bowiea , of subfamily Urgineoideae , whereas (1b) comprises all other genera including the genus Ledebouria . In the description of the genus they again emphasize the character state of the bracts as being not spurred. As it is obvious from Fig. 2B , the bracts of H. cryptopodus are clearly spurred and therefore this taxon cannot belong to Ledebouria but has to be placed within subfamily Urgineoideae , a position also corroborated by our molecular data.

Karyologically, H. cryptopodus with its diploid chromosome number of 2n=18 also fits to subfamily Urgineoideae although the most abundant basic haploid chromosome number in this group appears to be x=10 ( SPETA 1998a). The same basic chromosome number of x=9 as in H. cryptopodus is found, for example, in the generitype of Drimia , D. elata but also in the Rhadamanthus-group of Drimia sensu GOLDBLATT & MANNING 2000 . This group also contains the closest relatives of H. cryptopoda, the Madagascan species D. urgineoides and D. mascarenensis ( SPETA 1998a) . In Ledebouria sensu MANNING et al. (2004) a variety of chromosome numbers have been found from 2n=20 up to 2n=68 ( SPETA 1998a). The diploid chromosome number of 2n=10 for Ledebouria humifusa (WETSCHNIG in WETSCHNIG & PFOSSER 2003) appears to be the only case so far where a chromosome number smaller than 2n=20 has been reported for this group.

Based on molecular, morphological and karyological data Hyacinthus cryptopodus should be transferred from subfamily Hyacinthoideae to subfamily Urgineoideae . In the most recent generic synopsis ( MANNING et al. 2004) subfamily Urgineoideae contains the three genera Bowiea , Igidia and Drimia . The genus Bowiea forms a clearly separated monophyletic basal clade in all analyses until now ( Fig. 1 , PFOSSER & SPETA 1999, 2004, WETSCHNIG & PFOSSER 2003, MANNING et al. 2004). Although Igidia was not available for analysis and the genus Drimia itself may prove to be a rather heterogeneous assemblage when analyzed more thoroughly we suggest H. cryptopodus to be placed best within Drimia sensu MANNING et al. (2004) for the time being. This decision is not intended as a final one, but should serve as a starting point for future research in an alliance still suffering from a lack of detailed systematic knowledge on the species level. The same holds true for subfamily Ornithogaloideae which lacks an adequate distribution of species into more uniform genera. To adopt a broad and comprehensive generic concept like that of MANNING et al. (2004) only shifts problems to another level, because the distribution of species into genera mainly defined by molecular data results in genera unrecognizable by morphological synapomorphies. Possiblities for further subdivisions for all subfamilies of Hyacinthaceae based on a combination of morphological, cytological, karyological and chemotaxonomical characters have already been proposed ( SPETA 1998a, b, PFOSSER & SPETA 1999). Recently, Drimiopsis and Resnova have been resurrected as independent genera of their own again ( LEBATHA et al. 2006), after they have been sunk into Ledebouria before ( MANNING et al. 2004). This could serve as an example of how to arrive at a situation with more homogeneous and thus also more practically recognizable genera.

Figures

Fig. 1. Strict consensus phylogenetic tree (maximum parsimony) of family Hyacinthaceae. Bootstrap support values after character weighting are indicated above branches. The position of Hyacinthus cryptopodus within subfamily Urgineoideae is indicated by an asterisk.

Fig. 2. Hyacinthus cryptopodus. The conspicuously spurred bracts are marked in B.