Dampiera

Abstract: Rajput, M .T .M .,1 and Carolin, R .C .2 (xBotany Dept, Sind University, Jamshoro, Pakistan; 2John Ray Herbarium, University o f Sydney, Australia 2006) 1988. The genus Dampiera (Goodeniaceae): systematic arrangement, nomenclatural notes and new taxa. Telopea 3(2): 183-216. — A cladistic analysis of the genus is presented with a discussion of the characters used and the results incorporated in a new subgeneric classification. A new subsection A ngulares is proposed, the previous section Cephalantha is not recognized and the previous section Camptospora is reduced to the rank of series. Species and varieties in the genus Dampiera are provided with lectotypes and neotypes where required. Where there appears to be some doubt as to holotypes, these are discussed. The following new species are described: D. angulata , D. atrip licina, D. decurrens, D. delto id ea , D. fitzgeraldensis, D. fusca, D. galbraith iana, D. heteroptera, D. krau sean a , D. ob liqu a , D. orchard ii, D. pedunculata , D. ram osa, D. rod w ay ana, D. sa lah ae , D. scaevolina , D. sylvestris and D. tephrea; D. dysantha (Benth.) and D. la tea lata (E. Pritzel) are raised to species rank. Introduction The genus Dampiera (Goodeniaceae) is confined to Australia and occurs throughout the continent in a wide variety of habitats. The genus contains 66 species as recognised at present. The last revision is due to Krause (1912). No previous assessment of phytogeny has been made. The new taxa, nom enclatural changes and lectotypifications presented here, relate to the impending treatm ent of this genus in the ‘Flora of A ustralia’. They are also necessary for a discussion of the systematics of the genus. B oth B entham (1868) and Krause (1912) subdivided the genus into sections and although these systems were similar there were some minor differences principally in the level at which the subgeneric taxa were recognised and the emphasis which different characters were given. The results of Carolin (1959) and R ajput and Carolin (1984) suggest that a new assessment is necessary. The m ost satisfactory subgeneric classification should be based upon the putative phylogeny of the genus. Cladistic methods in general are likely to provide a reasonable approximation to phylogeny although it is clear that they have some drawbacks (Carolin 1985; Johnson & Briggs 1985) and that the algorithms may not in fact reproduce the correct phylogeny (Fiala & Sokal 1985). N evertheless, a cladogram obtained by parsimonious m ethods is a satisfactory starting point for considering possible phytogenies (Carolin 1986). The first section of this contribution attempts to develop a phytogeny in this way and provide a systematic framework for the genus. The phytogeny and systematics are considered first because this section necessarily contains a discussion of the characters used in the descriptions of the new taxa. Some of these characters require special terminology, e.g. the hairs, and this is established before they are used. M oreover, the descriptions and nom encla tu ral no tes are arranged in a system atic order and this is also established in the first section. 184 Telopea Vol. 3(2): 1988 Phylogeny and systematic arrangement Characters W hilst the characters which are used in a cladistic analysis are not usually weighted, it is clear that all cladograms are produced after the exclusion of characters which the researcher is prepared to reject (Carolin 1985, 1986). The m ethod adopted here was to generate up to five parsimonious cladograms from characters which are considered to be important in discriminating between the species, using the W A G N ER 78 algorithm of Farris. This was quick and relatively cheap and in the earlier stages of this investigation the m ore sophisticated PA U P package was not available. This preliminary data set included 31 binary characters, some of which represented ordered multistate ch a rac te rs add itive ly coded . H ow ever, a num ber of ch a rac te rs which discriminate effectively between species are exteremely homoplasic. That is they reverse m ore than three times on different branches or they reverse to the prim itive condition and then revert to the advanced condition on the same branch. These appear to be relatively inconstant characters and may well obscure the phylogeny. Homoplasic characters may increase the num ber of most parsim onious trees which can be generated from a given set of data and may even decrease the probability of finding any of the most parsimonious trees. The characters which are herewith rejected as being too homoplasic to reflect phylogeny accurately are: (i) leaves dentate or entire; (ii) leaves lobed or entire; (iii) sepals present or absent; (iv) auriculate wing of the corolla smaller than the others or not. A second data set, w ithout these characters and containing 27 binary characters, some of which also were additive, was then used to generate other cladograms using Swofford’s PA U P package (Swofford 1984). One hundred trees were generated using global branch swapping and a strict consensus tree was generated from them . Some further adjustm ent was then suggested. To produce a cladogram it is necessary to polarize the states of a character and the most efficient way to do this is by outgroup comparison, although philosophically this may involve an infinite regress. Nevertheless, in this case there is a reasonably well authenticated higher level cladogram (Carolin 1977, Fig. 3) which identifies the clade within which Dampiera occurs and thus its sister groups, Anthotium plus Lechenaultia. It also suggests Brunonia as the sister clade to this one. Brunonia, however, shows a number of advanced ch a rac te rs an d , if we accep t C am panu laceae as the o u tg ro u p of the G oodeniaceae (see Carolin 1977), one can arrive at a most parsimonious solution for the primitive states of most characters using the m ethods of M addison et al. (1984). Appendix 1 provides a quick reference to the characters used. A ppendix 2 gives the data matrix used in the cladistic analysis. Below are listed and discussed the characters which are used, together with the states which they may assume. The state coded O ’ is considered to be primitive. W here this is not unequivocally determined by the outgroups some discussion is given. An asterisk (*) indicates that the immediate outgroups of Dampiera indicate the polarity of that character. 1. * Sub-shrub-O; rosette-1 2. * Sub-shrub-O; m ulticaulate-1 M ulticaulate habit is defined as several more or less herbaceous stems arising from a very short stock. New growth starts more or less at ground level. Rajput & Carolin, Dampiera (Goodeniaceae) 185 Sometimes only one stem arises in any one year but over several years several stems arise. Sub-shrubs, on the o ther hand, have one or few basal woody stems which are usually much branched above and more or less long-lived. New growth starts on the branches of old growth. The lateral branches behave like the stems of the multicaulate plants. Indeed the multicaulate habit seems to have arisen as a condensation of the main stems of the sub shrubs into the thick basal stock with very short internodes (Fig. 1). The rosette habit represents a condensation of the lateral branches as well as the main stems.

Abstract: B.J. Lepschi, M.E. Trudgen and S.J. van Leeuwen. Two new species of Dampiera (Goodeniaceae) from the Pilbara region, Western Australia. Nuytsia 15(2): 269-276 (2004). Dampiera anonyma and Dampiera metallorum Lepschi & Trudgen, two geographically restricted taxa from the Pilbara region of Western Australia, are described, illustrated and their distributions mapped.