Document not found! Please try again

Taxonomic and Nomenclatural Notes on ... - David Morrison

4 downloads 0 Views 2MB Size Report
L. subcymosa C . Gardner & A. S. George) are not discussed. Many of the ..... 26248), S. of Badgingarra, George 14197, 5.xi.1975. (PERTH). ...... Druce, Bot.
Brunonia, 1986, 9, 1-28

Taxonomic and Nomenclatural Notes on Lechenaultia R. Br. (Goodeniaceae)

David A . Morrison John Ray Herbarium, Macleay Building (A12), University of Sydney, Sydney, N.S.W. 2006; present address: Department of Applied Biology, New South Wales Institute of Technology, P.O.Box 123, Broadway, N.S.W. 2007.

Abstract Morrison, D. A. Taxonomic and nomenclatural notes on Lechenaultia R. Br. (Goodeniaceae). Brunonia 9:l-28 (1986). There have been a large number of nomenclatural and taxonomic problems within Lechenaultia, and these are resolved here. A total of 24 names are lectotypified, two neotypes are chosen, a series of misapplied names is elucidated, and L. agrostophylla is synonymised under L. Jiliformis for the first time. As well, L . brevifolia, L . lutescens and L. papillafa are described as new species, and the considerable morphological variation within several of the species is reported in detail. The phylogenetic relationships within the genus are also re-appraised, the sectional delimitations are re-defined and section Patentes is described as new.

Introduction The first satisfactory complete treatment of Lechenaultia R. Br. was that of Bentham (1868), and the only full generic treatment since then is that of Krause (1912). Since that time there has been considerable collecting, particularly in areas not previously collected, and also much new work on the morphology and anatomy of the genus (e.g. Carolin 1959, Carolin 1960, Carolin 1966, Carolin 1967, Carolin 1971). Therefore, a full revision has been found necessary. The diagnostic part of this revision will be set out in the forthcoming treatment in the 'Flora of Australia', along with the key to the species. However, there are various nomenclatural and taxonomic problems that need to be considered in detail and a number of new species to be described. Therefore, full consideration is given here to the synonymys, misapplications and typification of those species that warrant it, a full description of each of the new species, and a re-appraisal of the relationships and sectional delimitations within the genus. Those few species which present no taxonomic or nomenclatural difficulties (L. acutiloba Benth., L. divaricata F. Muell., L. macrantha K. Krause, L. pulvinaris C. Gardner and L. subcymosa C . Gardner & A. S. George) are not discussed. Many of the species of Lechenaultia as circumscribed in this revision are very variable, notably L. biloba, L. filiformis, L. floribunda, L. formosa, L. stenosepala and L. tubiyora. The morphological variability is therefore also described fully for each of these, and it may be that each has variants worthy of separate taxonomic recognition. However, it has not been possible to analyse these taxa in enough

2

D. A. Morrison: Lechenaultia

detail to arrive at unequivocal taxonomic conclusions. This would require extensive field analyses at a population level as well as manipulative experimentation, and these studies were not possible for this revision. Therefore, a conservative approach has been adopted, and all infraspecific names have been synonymised under the appropriate species. Lechenaultia is a very coherent group within the Goodeniaceae, and can be recognised by a number of characters: the plants are always subshrubs or small shrubs without radical leaves, the inside of the corolla is hairy, the anthers are connate into a tube around the style when the flowers first open, the ovary is inferior and elongated with numerous ovules in two vertical rows, the indusium is distinctly 2-lipped at maturity with the stigmatic surface surmounting the upper lip, the pollen grains are united in tetrads, and the seeds are surrounded by a woody endocarp derived from the inner layers of the loculus wall (and thus the true fruit) while the capsule (the apparent fruit) is formed only from the outer floral whorls. The genus is restricted to Australia and New Guinea, with all of the 24 species occurring in Australia. Twenty of the species are endemic to south-western Western Australia, with three in arid and semiarid central Australia, and one widespread across tropical northern Australia and extending to the southern coast and off-shore islands of New Guinea. Relationships Within the Genus To analyse the phylogenetic relationships within Lechenaultia, a cladistic analysis was carried out. Cladograms were constructed using the WAGNER78 computer program (see Farris 1970), with several input sequences generating a series of cladograms. The 23 attributes used in the analysis are listed in Table 1. In formulating hypotheses of apomorphy, a priori outgroup comparison was used for attributes 5, 6, 9, 10, 13, 15, and 17-20, using Anthotium, Dampiera and Brunonia as the outgroup (see Carolin 1977). The remaining attributes yielded equivocal evaluations of relative apomorphy, and were analysed using a posteriori functional outgroup analysis (see Watrous and Wheeler 1981). All of the cladograms generated yielded very similar results, with the most parsimonious tree having 48 character changes (Fig. 1). Equally parsimonious trees were produced by minor re-arrangements of some of the branches (e.g. L. juncea and L.filiformis could also be placed next to L. subcymosa), but these differences were resolved by functional outgroup comparison of characters that could not be coded for in the initial analysis. However, all cladograms (including those one or two steps longer) show the same basic tree topology, which implies that these trees are close to the most parsimonious solution. The Fig. 1 cladogram shows a set of relationships that is significantly different from those suggested by Bentham (1868) and Krause (1912). Most notably, these authors subdivided the genus on the presence or absence of a 'beaked' capsule (the result of the distal articles usually not forming), and they consequently united L.filiformis and L. divaricata as a putative monophyletic group, sect. Latouria. The cladogram indicates that the 'beaked' appearance of these capsules is not homologous but has arisen independently; thus these species do not form a monophyletic group on their own. This hypothesis is supported by the data presented by Carolin (1966) on the fruit and seed structure of the genus. The sectional delimitations proposed here are set out in Fig. 1. Sect. Lechenaultia as circumscribed here is a monophyletic group defined by a number

Brunonia, Vol. 9, No. 1 (1986)

3

of apomorphic attribute states: the corolla is red or occasionally green, the corolla tube is longer than the free corolla lobes and forms an erect cylinder that is only hairy in the tube, and the free superior corolla lobes are erect. Sect. Latouria is also monophyletic, and is defined by the shorter lanceolate-falcate and almost wingless superior corolla lobes. Sect. Patentes, however, is a paraphyletic group that is uniquely defined only by plesiomorphic attribute states. While sect. Patentes and sect. Lechenaultia together form a monophyletic group, I have chosen to separate them for taxonomic purTable l. Attributes used in the cladistic analysis Attribute

Plesiomorphous state (0)

Rough bark

Throughout the stems

Leaves Leaf indumentum Flowering habit Inflorescence Bracts Sepal indumentum Corolla tube Corolla tube

Fleshy Glabrous Scapigerous Monochasial or dichasial Foliose Glabrous Open on the superior side Shorter than the free corolla lobes

Corolla tube

Shorter than or equal to the free corolla lobes Throughout the tube and also on the free lobes Throughout the tube

Corolla indumentum Corolla indumentum Inferior side of corolla tube Corolla colour Superior corolla lobes Superior corolla lobes Superior corolla lobes Wings on superior corolla lobes Wings on superior corolla lobes Wings on superior corolla lobes Style Hairs at back of the indusium Capsule

Without gibbous pocket Shades of blue or yellow Spreading Lanceolate Not enclosing the indusium Equal or almost equal to those on the inferior lobes Always present on at least one side of the lobe If present then on both sides of the lobe

Apomorphous state (1) Only at the base or on the lowest branches Not fleshy Hairy Not scapigerous Solitary flowers Reduced Hairy Erect and cylindrical At least as long as the free corolla lobes Longer than the free corolla lobes In the tube only Only in a tuft at the base of the tube With gibbous pocket

Curved Long

Shades of red or green Erect Lanceolate-falcate Enclosing the indusium Not equal to those on the inferior lobes Usually absent on both sides of the lobe Always absent on one side of the lobe and present on the other side Erect Short

All articles formed

Distal articles not formed

poses. Sect. Lechenaultia is united by a relatively large number of apomorphous attribute states (Table 1, attributes 8, 10-12, 14, 15), and is thus distinctive when compared to sect. Patentes. I believe this morphological divergence between the two groups is worthy of sectional status. Uniting the two sections would thus produce a cladistically more appealing but morphologically less descriptive sectional taxonomy. Nevertheless, sect. Patentes is only an artificial grouping for taxonomic convenience, and the existence of the name is an incidental by-product of the cur-

D. A. Morrison: Lechenaultia

4

rent International Code of Botanical Nomenclature which requires the naming of all other subgeneric categories if one is to be named. The section should not be treated as a coherent evolutionary unit. Biogeographically, the cladogram is also informative. The tree indicates that the tropical species (L.filiformis), the two western central Australian species

brevtfolta

-

strtata heterornera lutescens dtvartcata

Fig. 1. Cladogram for Lechenaultia. Cross bars are synapomorphies of the attributes as coded in Table 1, and double bars are reversals.

(L. lutescens and L. striata) and the eastern central Australian species (L. divaricata) all have many derived attribute states and have sister species that occur only in south-western Western Australia, suggesting a relatively recent origin of these species. The cladogram also indicates that the arid and tropical species have separate origins in the temperate south-western Western Australian flora. This

5

Brunonia, Vol. 9, No. 1 (1986)

pattern of recent invasions of the arid zone from temperate areas is repeated in several other genera of the Goodeniaceae (e.g. Dampiera, Goodenia; Carolin 1982) as well as in the Papilionaceae (e.g. Leptosema; Crisp 1982), and may thus be part of a much more general pattern in the evolution of the arid zone flora of Australia.

Systematic Treatment LECHENAULTIA Lechenaultia R. Br., Prodr. 581 (1810); Roemer & Schultes, Linn. Syst. Veg. edn 16. 5:34 (1820); Sprengel, Linn. Syst. Veg. edn 17. 1:719 (1824); Sims, Bot. Mag. 52:t.2600 (1825); Ker Gawler, Bot. Reg. 1l:t.916 (1825); Sweet, F1. Australasica t.26,46 (1827); Don in Loudon, Hort. Brit. 79 (1830); G. Lodd., Bot. Cab. 16:t.1579 (1830); Don, Gen. Hist. 3:727 (1834); Lindley, Intr. Nat. Syst. Bot. edn 2. 242 (1836); Endl., Gen. P1. 508 (1838); Don in Loudon, Encycl. PI. edn 2. 1164 (1841); Paxton, Paxt. Mag. Bot. 8: 151 (1841); Vriese in Lehm., PI. Preiss. 1:414 (1845); Lehm., PI. Preiss. 2:244 (1848); Leenh. F1. Malesiana ser. 1. 5:338 (1957); H. Eichler, Suppl. Black's F1. S. Austral. 294 (1965); Beard, Descr. Cat. W. Austral. PI. edn 2. 125 (1970); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 674 (1975); Carolin in Jessop, F1. Centr. Austral. 360 (1981); J. Green, Census Vascular P1. W. Austral. 100 (1981); Jacobs & Pickard, P1. N.S.W. 133 (1981); Jessop, List Vascular PI. S. Austral. 36 (1983). Lectotype (here designated): Lechenaultia formosa R. Br. Lechenautia A. L. Juss., Ann. Mus. Natl. Hist. Nat. 18:6 (181I), orthographic variant. Leschenaultia Benth. in Endl., Enum. PI. 70 (1837), orthographic variant; Lindley, Sketch Veg. Swan R. 26 (1839); DC., Prodr. 7518 (1839); Hook., Bot. Mag. 72:t.4256,4265 (1846); Lindley, Veg. Kingd. edn 2. 695 (1847); Paxton, Paxt. Mag. Bot. 14:245,246 (1847); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:181 (1854); Don in Loudon, Encycl. PI. edn 3. 1320 (1855); Benth., F1. Austral. 4:37 (1868); Benth. & J. D. Hook, Gen. P1. 2:537 (1876); F. Muell., Syst. Census Austral. PI. 87 (1882); Baillon, Hist. P1. 8:370 (1885); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); Tate, Handb. F1. Extratrop. S. Austral. 245 (1890); C. Moore, Handb. F1. N.S.W. 305 (1893); Bailey, Queensl. F1. 3:891 (1900); E. Pritzel, Bot. Jahrb. Syst. 35:551 (1905); K. Krause, Pflanzenr. IV.54:97 (1912); Maiden & Betche, Census N.S.W. PI. 190 (1916); Ewart & Davies, F1. N. Territory 266 (1917); C. Gardner, Enum. P1. Austral. Occ. 125 (1930); Blackall & Grieve, How to Know W. Austral. Wildflowers edn 1. 409 (1956); J. Black, F1. S. Austral. edn 2. 828 (1957); C. Gardner & A. S. George, J. Roy. Soc. W. Austral. 46: 134 (1963); C. Gardner, J. Roy. Soc. W. Austral. 47:63 (1964); G. Chippendale, Proc. Linn. Soc. N.S.W. 96:262 (1972); G. Cunningham et al., P1. West. N.S.W. 636 (1981). Latouria (Endl.) Lindley, Veg. Kingd. edn 2. 695 (1847), based on Lechenaultia sect. Latouria Endl.; Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:187 (1854). Holotype: Latouria jiliformis (R. Br.) Vriese. Ericopsis C. Gardner, J. Roy. Soc. W. Austral. 9:42 (1923). Holotype: Ericopsis formosus C. Gardner .

6

D. A. Morrison: Lechenaultia

Notes Robert Brown named the genus after the Frenchman Jean-Baptiste LouisClaude-Theodore Leschenault de la Tour (1773-1826), botanist and traveller, who was one of the naturalists with Nicolas Baudin's voyage of exploration (1800-4) which visited Australia in 1801-3, collecting at Cape Naturaliste, Swan River and Shark Bay. The two botanists met at Encounter Bay in South Australia while Brown was the naturalist with Matthew Flinders' circumnavigation of Australia. However, being an educated man, Brown expected (quite naturally) that his friend would spell his surname in the French manner, rather than spelling it with the Germanic 's'. Brown therefore omitted the 's' in the spelling of both the surname and the generic name. The generic spelling thus cannot be considered as an orthographic error, and the original spelling must be maintained (ICBN 1983, Art. 73.1 and 73.3). Botanists consistently used Brown's spelling of the generic name for nearly 30 years after he published it, until Bentham (1837) introduced Leschenault's spelling of his own name (although De Candolle (1839) was the first to unequivocally equate the new spelling with Brown's original spelling). For the next 100 years, this new spelling was taken up by all botanists who had cause to use the name, including those who had in the past used the original (e.g. Don, Lindley, Paxton, de Vriese). However, Brown's original name was re-instated during the late 1950s and early 1960s, and this spelling has come into common (but not universal) usage since then.

Lechenaultia formosa is the species which best agrees with the protologue of the genus, particularly concerning the shape of the articles. It is thus chosen as the lectotype for the genus.

Sect. I. PATENTES Lechenaultia R. Br. sect. Patentes D. A. Morrison, sect. nov. Corolla patens, obliqua, caerulea vel flavida; lobis omni alatis, intus sparse pubescentis vel sparse pilosis. Holotypus: L. biloba Lindley.

I. Lechenaultia biloba Lindley, Sketch Veg. Swan R. 27 (1 Dec. 1839); Paxton, Paxt. Mag. Bot. 8:151 (1841); Vriese in Lehm., PI. Preiss. 1:415 (1845); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:182 (1854); Benth., F1. Austral. 4:42 (1868); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54: 100 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 676 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: Not designated. Lectotype (here designated): Swan River, Drummond, 1839 [Ist Coll.] (CGE). Probable isotype: K (n.v. photo SYD). Possible syntype mounted on the same sheet. Lechenaultia grand~floraLindley, Sketch Veg. Swan R. 26 (1 Dec. 1839); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:181 (1854). Type: Not designated. Lectotype (here designated): Swan River, Capt. Mangles, s. dat. (CGE). Syntype: Vasse River, on the South West coast of New Holland, Mrs Molloy, 1839 (CGE).

Brunonia, Vol. 9, No. 1 (1986)

7

Lechenaultia grandifora DC., Prodr. 7:519 (late Dec. 1839); Vriese in Lehm., P1. Preiss. 1:4l5 (1845). Type: 'In Nova-Hollandia ad Swan-river legit cl. Drummond.' Lectotype (here designated): Swan river, Drummond, 1839 [lst Coll.] (G-DC, n.v. microfiche). Lechenaultia drummondi[i] Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:182 (1854). Type: 'In Herb. Mus. Caes. Vindob. nempe no. 292 Sw. Riv. Drumm. Huc etiam pertinet specimen Preissianum no. 1463.' Lectotype (here designated): In solo limoso-arenoso inter frutices vallis Cataractae ad fl. Cygnorum, L. Preiss 1463, 25.vii.1839 (LD). Isolectotypes: L903.311-303 (n.v.) & 903.31 1-304 (n.v. photo SYD), MEL1522561. Syntype: Swan River, Drummond 292, s. dat. [Ist Coll.] (W, n.v. photo SYD). The names L. biloba and L. grandiflora were published simultaneously by Lindley in his 'A Sketch of the Vegetation of the Swan River Colony', and thus have equal priority when placed in synonymy. Bentham (1868) first synonymised these two names and chose L. biloba as the preferred name. These names only predate that of De Candolle by a few weeks. There is only one specimen of L. biloba in the Lindley herbarium at CGE, and it is chosen as the lectotype. There is also a sheet at K with apparently a mixed collection on it. Two of the specimens are labelled as being collected by Drummond at Swan River in 1839, and are presumably duplicates of the material sent to Lindley. The other specimen bears Lindley's name and is dated the year before. This could be a piece removed from Lindley's own specimen. There is also only one sheet at CGE labelled L. grandifora in Lindley's handwriting. However, two separate collections are mounted on this sheet, the upper collection being by Mrs Molloy at Vasse River and the lower collection by Capt Mangles at Swan River. The reference to L. grandifora is below the Mangles collection, and this is chosen as the lectotype. There are two specimens in this collection, and the left hand one is chosen as the lectotype. There is only one sheet at G-DC labelled L. grandflora DC. However, this sheet contains two specimens collected at Swan River by Drummond, both of which may be part of the same collection. The right-hand specimen is chosen as the lectotype. De Vriese' protologue for L. drummondi[i] is not clear but it appears he intended to provide a substitute name for L. grandifora DC., which is a later homonym of L. grandifora Lindley. De Vriese considered L. grandifora DC. a distinct species, and he had previously (in 'Plantae Preissianae') identified material as belonging to this species. Under this assumption, the type of this nom. nov. would be that of L. grandifora DC. (ICBN 1983, Art. 7.9). However, as de Vriese' intention is not specifically stated, a lectotype is chosen for this name to provide for alternative interpretations. The LD specimen of Preiss 1463 is annotated by de Vriese as both L. grandifora DC. and L. drummondi[i] and is chosen as the lectotype. Notes

L. biloba varies greatly in flower and leaf size as well as in corolla colour, but these characters do not appear to be correlated. Flower size varies throughout the geographical range, varying from flowers with corollas about 12 mm long to those nearly twice this size; flower size is, however, usually consistent within a population. This variability is suggestive of chromosomal duplication in some populations (cf. Peacock 1963).

8

D. A. Morrison: Lechenaultia

Corolla colour also varies greatly, with various collectors suggesting that there are three distinct colour forms: dark blue (the commonest form), pale blue and cream. Gardner (1978) suggests that this differentiation is related to soil type, but all three forms have been reported within a single population. However, the creamflowered forms seem to occur only on the western sandplains. Leaf size varies markedly from west to east, with the longer leaves only occurring on plants of the western sandplains. On the red gravels and laterites of the inland areas the leaves are exclusively of the shorter type (less than 6 mm long), which gives the plants a superficial resemblance to L. brevifolia. 2. Lechenaultia stenosepala E. Pritzel, Bot. Jahrb. Syst. 35552 (1905); K. Krause, Pflanzenr. IV.54:100 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 675 (1975); J. Green, Census Vascular PI. W. Austral. 100 (1981). Type: 'Hab. in distr. Avon septentrionali in arenosis apertis pr. Dandaragan, flor. m. Dec. (D. 5760).' Neotype (here designated): Strathmore Rd Reserve (no. 26248), S. of Badgingarra, George 14197, 5.xi.1975 (PERTH). Typification Pritzel based his description solely on Diels 5760. Diels' herbarium was housed at B; but this specimen has not been found there and was presumably destroyed during World War I1 (cf. Hiepko 1978). No duplicates are known and so a neotype must be designated. Unfortunately, none of the specimens examined (about 50) fits the protologue particularly well: the calyx measurements of the protologue are those of the smaller-flowered specimens (see below), while the corolla colour and venation are those of the larger-flowered specimens, and the corolla measurements are those of the intermediate specimens. The chosen neotype is the best fit to the protologue of those specimens examined. Notes L. stenosepala is very variable in leaf and flower size, varying from plants with small leaves and very small (corolla about 10 mm long) blue corollas to plants with larger leaves and medium-sized (corolla about 25 mm long) creamy mauve corollas with yellow spots at the top of the tube and the style projecting above the lobes. This variation appears to correlate with latitude, the largest leaves and flowers only occurring on plants in the northern part of the distribution, and the smallest leaves and flowers only occurring in the southern part. The variation is suggestive of differentiation into chromosomal races, but this hypothesis remains to be tested.

3. Lechenaultia expansa R. Br., Prodr. 1:581 (1810); Don, Gen. Hist. 3:727 (1834); Benth. in Endl., Enum. P1. 70 (1837); DC., Prodr. 7 5 1 9 (1839); Benth., F1. Austral. 4:42 (1868); F. Muell., Syst. Census Austral. PI. 87 (1882); E. Pritzel, Bot. Jahrb. Syst. 35:553 (1905); K. Krause, Pflanzenr. IV.54:98 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 674 (1975); J. Green, Census Vascular PI. W. Austral. 100 (1981). Type: Not designated. Lectotype (here designated): In ericetis apricis ad Portum Regis Georgii 111'" in ora occidento australi Novae Hollandiae, R. Brown, xii. 1801 (Britten sheet No. 2544) (BM, n.v. photo SYD). Isotype: BM (n.v. photo SYD). Lechenaultia pallescens var. congesta Vriese in Lehm., PI. Preiss. 1:4l5 (1845). Type: 'In solos limoso, arenoso, inter frutices ad flumen Canning, distr.

Brunonia, Vol. 9, No. 1 (1986)

9

Sussex, d. 2 m. Nov. 1839. Herb. Preiss. No. 1468.' Lectotype (here designated): In solo limoso-arenoso inter frutices ad fl. 'Canning' et distr. (Sussex), L. Preiss 1468, 2.xi.1839 (LD). Isotypes: K(n.v.), L903.311-308 (n.v. photo SYD) & 903.311-309 (n.v.), MEL1522628. Lechenaultia parvijlora Vriese in Lehm., PI. Preiss. 1:416 (1845). Type: 'In arenosis sylvae ad portum, Princess Royal, Harbour, m. Oct. 1840. Herb. Preiss. No. 1462.' Lectotype (here designated): In arenosis sylvae ad portum 'Princess Royal Harbour', L. Preiss 1462, x.1840 (LD). Lechenaultia tenuifolia Vriese in Lehm., PI. Preiss., 1:415 (1845). Type: 'In Australia occidentali et meridionali, Herb. Preiss. No. 1460 ex parte et in solo glareoso sterili inter frutices prope montem Barrow, Plantagenet, d. 9 m. Nov. 1840. No. 1461.' Lectotype (here designated): In solo glareoso sterili inter frutices prope montem 'Barrow' (Plantagenet), L. Preiss 1461, 9.xi. 1840 (LD). Probable syntype: In Australia occidentali, L. Preiss 1462, s. dat. (LD). Lechenaultia expansa var. congesta (Vriese) Vriese in Lehm., PI. Preiss. 2:402 (1848)' based on L. pallescens var. congesta Vriese. [Lechenaultia laricina auct. non Lindley: Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2 10: 185 (1854), p.p. (as to Preiss 1462 only)]. [Lechenaultiafloribunda auct. non Benth.: Benth, F1. Austral. 4:43 (1868), p.p.; K. Krause, Pflanzenr. IV.54:99 (1912)' p.p. (both as to Preiss 1468 only)]. Typzzcation There are five specimens of Robert Brown's collection mounted on the Britten sheet at BM. The upper left-hand specimen is chosen as the lectotype of L. expansa. De Vriese based L. pallescens var. congesta on Preiss 1468. However, the label of the LD specimen of this collection (the preferred reference collection; cf. Crisp 1983) is not annotated by de Vriese (as are most of the other specimens). It does, rather oddly, bear a separate label in de Vriese's handwriting labelling it as L. glauca Lindley, a name that he never published with regard to this specimen. Furthermore, there are two specimens at MEL labelled as Preiss 1468, both annotated by Bentham. However, only one of these specimens (MEL1522628) appears to be a duplicate of the LD Preiss 1468 specimen. The other specimen (MEL1522630) is quite clearly L. floribunda rather than L. expansa. This may thus be the source of Bentham's (1868) misidentification of this Preiss collection. De Vriese later ('Plantae Preissianae' 2:244) synonymised L. pallescens and L. expansa but did not mention L. pallescens var. congesta. However, in the list of specimens at the end of the volume ('Plantae Preissianae' 2:402) he lists Preiss 1468 as L. expansa var. congesta. His intention is thus clear and the combination is validly made. There are two specimens at LD labelled as Preiss 1462; however, only one of these is annotated as L. parviflora in de Vriese's handwriting (see L. tenuifolia below for a discussion of the other specimen) and this is chosen as the lectotype. No specimens of this Preiss collection number have been located at other herbaria. In 'Plantae Preissianae', de Vriese claimed to base L. tenuifolia on Preiss collections 1460 (in part) and 1461. However, there is only one specimen of Preiss I460 at LD rather than the expected two (the other specimen should be part of the type material of L. pallescens). The Preiss 1460 specimen at LD is not annotated by de Vriese but is mounted with Preiss 1464, which is annotated as L. pallescens. Both Preiss 1460 and Preiss 1464 are L. floribunda rather than L. expansa. The LD

10

D. A. Morrison: Lechenaultia

specimen of Preiss 1461 is annotated by de Vriese as L. tenuifolia and therefore this is chosen as the lectotype. No specimens of Preiss 1461 have been located at other herbaria. There is, however, another specimen mounted with Preiss 1461 at LD and this is also annotated by de Vriese as L. tenuifolia. This latter specimen is labelled as Preiss 1462, not Preiss 1460 as expected from the protologue. The label does, nevertheless, have locality detail as set out in the protologue and does not have a date. Such labels were apparently intended as duplicates of the 'reference' specimens which all bear detailed locality descriptions and dates. There is also another Preiss collection at LD which is labelled as collection no. 1462; this is mounted separately and has full locality details and date. This specimen is annotated by de Vriese as L.parviflora (see above). The only other Preiss collections of Lechenaultia mounted together are Preiss 1460 and Preiss 1464, which de Vriese unites under L. pallescens, and Preiss 1458 and Preiss 1459, which de Vriese unites under L. tubiflora. It is thus very probable that de Vriese based L. tenuifolia on Preiss 1461 and Preiss 1462 (in part) rather than Preiss 1460 (in part) as indicated in 'Plantae Preissianae'. The two specimens annotated by de Vriese as L. tenuifolia appear to be similar, and are distinct from the other Preiss I462 specimen as well as the Preiss 1460 specimen. There is, however, a specimen at L that is labelled as both Preiss 1460 and as L. tenuifolia; I do not know the origin of these labels (see Crisp 1983 for a discussion of the confusion surrounding the distribution of the Preiss specimens). Notes This species is closely related to L. floribunda and is often confused with it, particularly in the area around Perth where the two distributions overlap. However, L. expansa is most obviously different in that the flowers are smaller and have a more densely branched flowering habit, so that they appear as compact convextopped inflorescences; while L. floribunda usually has fewer larger flowers which are more scattered. However, this distinction breaks down in some areas (see L. Jloribunda). In the past, the two species have been diagnostically distinguished by whether the ovary is longer (L, floribunda) or shorter (L. expansa) than the floral leaves, and while this can be very useful it is not consistent enough to be diagnostic. However, L. expansa has larger, often minutely pitted, leaves, a more densely hairy floral tube, smaller and almost equal wings on the petals, a more thickly hairy indusium, smaller fruits with fewer articles, and ovoidal articles without horizontal grooves. 4. Lechenaultiafloribunda Benth. in Endl., Enum. PI. 70 (1837); DC., Prodr. 7:519 (1839); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:186 (1854); Benth., F1. Austral. 4:43 (1868); E. Pritzel, Bot. Jahrb. Syst. 35553 (1905); K. Krause, Pflanzenr. IV.54:99 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 675 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: 'Swan-River. (Hugel.).' Holotype: Swan River, Hugel, s. dat. (W, n.v. photo SYD). Isotype: K (n.v. photo SYD). Lechenaultia glauca Lindley, Sketch Veg. Swan R. 27 (1839); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:184 (1854). Type: Not designated. Lectotype (here designated): Swan River, Drummond, 1839 [lst Coll.] (CGE). Syntype: Swan River, Mr Toward, s. dat. (CGE).

Brunonia, Vol. 9, No. 1 (1986)

11

Lechenaultia pallescens Vriese in Lehm., PI. Preiss. 1:415 (1845) and 2:244 (1848). Type: 'In depressis arenosis et umbrosis sylvae prope urbiculam Perth, d. 22 m. Oct. 1839. Herb. Preiss. No. 1460 et in arenosis sylvae prope Bull'screek, Perth, m. Nov. 1841. Herb. Preiss. No. 1464.' Lectotype (here designated): In arenosis sylvae prope 'Bull's-creek'/Perth/, L. Preiss 1464, xi.1841 (LD). Isolectotypes: MEL1522631, W. Syntypes: In depressis arenosis subumbrosis sylvae prope urbiculam 'Perth', L. Preiss 1460, 22.x.1839 (L903.311-306, n.v. photo SYD, & 903.311-316, n.v. photo SYD, LD, MEL1522629, W, n.v.). Lechenaultia floribunda var. borealis E. Pritzel, Bot. Jahrb. Syst. 35:553 (1905); K. Krause, Pflanzenr. IV.54:99 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 675 (1975). Type: 'Viget in distr. Irwin pr. Nordhampton in apertis arenosis flsr. m. Nov. (D. 5778).' Neotype (here designated): Between Dongara and Northampton, W. E. Blackall 2656, 19.ix.1932 (PERTH). Lechenaultia drummondiana A. Colozza, Nuovo Giorn. Bot. Ital. n. ser. 15:204 (1908). Holotype: Swan River, Drummond, 1839 [lst Coll.] (FI-W, n.v. photo SYD). [Lechenaultia laricina auct. non Lindley: Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:185 (1854), p.p.; K. Krause, Pflanzenr. IV.54:104 (1912), p.p. (both as to Preiss 1460 only)]. Typification There is only one sheet of L. glauca at CGE with Lindley's handwriting. However, two separate collections from the Swan River are mounted on this sheet, the upper collection being by Mr Toward and the lower one by Drummond. The reference in Lindley's handwriting is adjacent to the lower collection and this is chosen as the lectotype. There are two specimens in this collection and the left-hand one is chosen as the lectotype. De Vriese claimed to base L. pallescens on two Preiss collections, numbers 1460 (in part) and 1464. There is, however, only one specimen of Preiss 1460 at LD, and this is not annotated by de Vriese (see L. tenuifolia above for a further discussion of this specimen). Preiss 1464 is mounted on the same sheet as Preiss 1460, and this seems to indicate that de Vriese considered these specimens to be part of the same species (see above). The Preiss 1464 specimen is annotated by de Vriese as L. pallescens, and so this is chosen as the lectotype. Pritzel described L. Jloribunda var. borealis solely from Diels 5778. This specimen has not been found at B, and was presumably a casualty of World War I1 (cf. Hiepko 1978). A neotype must therefore be chosen, as no duplicates are known. Fortunately, the protologue is quite clear in its characterisation of this variety and there is no difficulty in selecting an appropriate neotype. Notes L. floribunda is a very variable species, particularly with respect to leaf length, flower size and flowering habit. There appears to be a morphocline running from north to south, with many of the more northern populations having plants with smaller flowers (i.e. shorter ovaries, sepals, etc.), larger leaves and a more densely branched flowering habit than the more southerly plants. Thus, many of the northerly plants look similar to L. expansa. However, these plants retain the less densely

12

D . A. Morrison: Lechenaultia

hairy floral tube, the tomentose rather than hispid indusium and the larger fruits, and are thus easily identifiable as L. floribunda. These northern plants have often been separated as L. floribunda var. borealis but, as they are part of a morphocline rather than a morphologically discrete group, this taxonomic separation does not seem justified. There are also some collections from the Coorow-Green Head area with very short leaves on the lower branches, which makes them look very similar to L. papillata. However, these plants do produce more typical leaves nearer the flowers and are thus easily distinguished.

5. Lechenaultia papillata D. A. Morrison, sp. nov. Fruticuli virgati vel frutices parvi virgati; caulibus floriferis non scapiformibus; foliis confertis, 1.4-4.0 mm longis; sepalis 3 0-5.0 mm longis; corollis caeruleis pallentibus, obliquis, lobis fere aequalibus, lobis superis alis angustatis; foliis, bracteis, bracteolis, sepalis, ovariis et fructibus papillatis. Holotypus: Fault line, northern end of Location 1163, c. 58 km N. of mouth of Oldfield R., Hj. Eichler 20392, 21.x.1968 (AD97018209). Isotypi: CANB (n.v.), PERTH. (Fig. 2). The specific epithet is a reference to the papillate leaves, calyx, ovaries and fruits, which are unique within the genus. Perennial subshrub to small shrub, diffuse, to 45 cm high and 40 cm diam., glabrous. Stems ascending to erect, several to many from a common woody rootstock, moderately branched with the branches ascending to erect, terete, up to 1.5 mm thick, woody, bark pale brown to grey, rough. Leaves densely crowded, somewhat fleshy, linear, 1.5-3.0(-4-O)mm long, 0.3-0.6mm wide, often thickened with a keel below, acuminate, more or less smooth and papillate, glabrous. Flowers in compact monochasial or occasionally dichasial cymes of up to 5 flowers; bracts and bracteoles linear, 3.0-5.5 mm long, acuminate, papillate, glabrous. Calyx lobes not overlapping at the base, linear to linear-lanceolate, 3.0-5-0 mm long, 0.5-0.7 mm wide, acuminate, papillate, glabrous. Corolla pale blue; tube cleft to the base, 5.5-6.5 mm long, outside glabrous, inside with dense short weak simple hairs on the upper three-quarters, with scattered papillae towards the base; lobes almost equal, spreading, lanceolate, 6.0-8-0 mm long, 0.9-1 - 3 mm wide, long acuminate, outside glabrous, inside usually with scattered long weak simple hairs on the lower quarter and also along the margins of the wings; wings on the inferior lobes triangular, 4.0-5.0 mm long, 1.9-2.6 mm wide, margins undulate-entire, transverse veins obscure; wings on the superior lobes rounded, 3.5-5-0 mm long, 0-8-1.6 mm wide, margins entire, transverse veins obscure. Stamens not enclosed in the tube; filaments thin, 3.0-3.5 mm long; anthers linear, 1 ~3-2.0mm long, acute, yellow. Ovary erect, 5.0-8 a0 mm long, papillate, glabrous; style conspicuously curved, moderately robust, not dilated towards the base, 6.0-8.0 mm long, often with scattered very short, erect capitate hairs on the lower half, with dense long, weak simple hairs at the back of the indusium. Fruit 4-valved, 10-16 mm long, not woody, opening annually, slightly constricted between the articles, usually all articles formed, glabrous and papillate. Articles in two rows when mature, about 9-14 per row, cylindrical, 1 mm long, pale greenblack, horizontally grooved with several vertical ridges.

Distribution South Western Australia: Roe District.

Brunonia, Vol. 9 , No. 1 (1986)

Fig. 2 . Holotype of Lechenaultia papillata (Eichler 20392, AD97018209).

13

D. A. Morrison: Lechenaultia

14

Ecology This species has been recorded in yellow or white sand, loamy sand and gravelly loam. It usually occurs in heath, low open scrub or eucalypt scrub mallee, where it may be locally common. Flowering is usually from late October to late November. Notes Flower size is somewhat variable within L. papillata, but the species nevertheless remains distinctive. This species is closely related to L.jloribunda, from which it can readily be distinguished by the shorter more crowded leaves, the papillate leaves, calyx, ovary and fruit, and the much more hairy floral tube. Specimens Examined WESTERN AUSTRALIA: Roe District: Lake King-Norseman rd, between Rabbit Proof Fence & 100 mile Tank, Beard 3784, 22.x.1964 (PERTH); Lacum Cronin, Gardner 15020, 25.xi.1964 (PERTH); 272mls [435 km] from Perth on Lake King rd, Humphreys 190, 28.x.1966 (PERTH); 57.5mls [92.5 km] E. of Lake King crossroads, Humphreys, 14.xi.1965 (PERTH); lower King rd, Humphreys, 28.x.1966 (PERTH); 29 mls [47 km] E. of Cross Roads, Forrestania, Lulgtz 3868, 25.xi.1964 (PERTH); near Mt Madden, Newbey 1835, 31.x.1965 (PERTH); 13 mls [21 km] W. of Lake King, Newbey 2637, 7.ix. 1967 (PERTH); 21 mls [34 km] SW. of 90 Mile Tank, Frank Hann N.P., Newbey 6523, 13.xi. 1979 (PERTH); 15 km W. of Lake Cronin, Newbey 6613, 3.xi.1979 (PERTH); 23 km NE. of Mt Heywood, Newbey 7998,9.xi.1980 (PERTH); 43 km NW. of Mt Ragged, Newbey 8220, 8.xi.1980 (PERTH); 15 mls [24 km] from Lake King towards Newdegate, Wrigley, 6.xi.1968 (CBG031367, NSW125989).

Sect. 11. LATOURIA Lechenaultia R. Br. sect. Latouria Endl., Gen. P1. 508 (1838); Benth., F1. Austral. 4:43 (1868); Benth. & J. D. Hook., Gen. P1. 2:537 (1876); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); Bailey, Queensl. F1. 3:892 (1900); K. Krause, Pflanzenr. IV.54: 108 (1912); C. Gardner, Enum. P1. Austral. Occ. 125 (1930). Holotype: L. Jiliformis R. Br. 6. Lechenaultia heteromera Benth., F1. Austral. 4:43 (1868); K. Krause, Pflanzenr. IV.54: 105 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 678 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: 'Drummond, n. 142, Mylne; E. Mount Barren, Moir's Inlet, Oldfield and Phillips rivers, Maxwell.' Lectotype (here designated): East Mt Barren, Maxwell, s. dat. (MEL1522730). Isolectotypes: BM. Syntypes: Swan River, Mylne, s. dat. (K, n.v. photo SYD); Phillips R, Maxwell, s. dat. (K, n.v. photo SYD, MEL1522727); Moirs Inlet, Phillips river, Maxwell, s. dat. (K, n.v. photo SYD); Oldfield R, Moirs inlet, ?Maxwell, s. dat. (MEL1522728).

As Bentham cites a series of specimens in his protologue, one must be chosen as the lectotype. No specimens of Drummond 142 have been found at any of the herbaria consulted. Duplicates of the remaining collections are mounted together on a number of sheets at K and MEL, and all bear Bentham's monogram. One of the Maxwell collections from East Mt Barren is chosen as lectotype, because this collection has the most complete specimen and is mounted on its own at MEL and is therefore easily identified.

Brunonia, Vol. 9, No. 1 (1986)

15

7. Lechenaultia lutescens D . A. Morrison & R. C. Carolin, sp. nov. Herbae vel fruticuli diffusa; caulibus floriferis non scapiformibus; foliis non confertis, 9 5-3 1 - 0 mm longis; sepalis 3.5-4.5 mm longis; corollis lutescentibus, obliquis, lobis inaequalibus, lobis superis alis perangustis vel absentibus. Holotypus: Yuendumu Reserve N.T., 22"35', 131°42', N. Henry 334, 13.i.1972 (NT33986). Isotypi: CANB221438, K (n.v. photo SYD), MEL1522789. (Fig. 3). [Lechenaultia helmsii auct. non K. Krause: J. Maconochie & N. Byrnes, Muelleria 2: 135 (1971); G. Chippendale, Proc. Linn. Soc. N.S.W., 96:267 (1972); J. Maconochie, N. Territory Bot. Bull. 3:8 (1980); J. Green, Census Vascular P1. W. Austral. 100 (1981)l. The specific epithet is a reference to the usually pale yellow corolla. Perennial herb or subshrub, virgate, to 30 cm high and 30 cm diam., glabrous. Stems ascending, several to many from a common woody rootstock, moderately branched with.the branches erect, terete, 1 mm thick, woody only towards the base, glaucous, stria&. Leaves not crowded, rigid, linear to oblong, (9.5-)14.5-20.5(-31.0) mm long, 0.7-1.3 mm wide, flat, margins often recurved, central vein prominent below, acuminate, rugose, glabrous. Flowers in loose monochasial or occasionally dichasial cymes of up to 5 flowers or rarely solitary and terminal; bracts and bracteoles linear, 9.5-17.0 mm long, acuminate, rugose, glabrous. Calyx lobes not overlapping at the base, linear-subulate, 3.5-4.5 mm long, 0.5-0-7 mm wide, acute to acuminate, rugose, glabrous. Corolla varies from orange-yellow through (usually) pale yellow to creamy white; tube cleft to the base, 7.0-8.0mm long, outside glabrous, inside with dense long weak simple hairs throughout; lobes not equal, spreading; inferior lobes narrow lanceolate, 9.5-12.5mm long, 0 . 8 - 1 - l m m wide; superior lobes lanceolate-falcate, 6.5-9-0(-10.0) mm long, 0.6-0.8 mm wide; all lobes acuminate, outside glabrous, inside with dense long weak simple hairs on the lower third; wings on the inferior lobes deeply bi-lobed, 5 0-7.5(-9.5) mm long, 2 0-3.0(-4.0) mm wide, margins undulate-serrulate, transverse veins prominent; wings on the superior lobes triangular to deeply bi-lobed, 3.5-4.5(-5.5) mm long, 0.8-1.0(-1 8) mm wide, margins usually entire, transverse veins usually obscure. Stamens not enclosed in the tube; filaments thin, 3.0-4.5 mm long; anthers oblong, 2.2-2.9 mm long, obtuse, yellow. Ovary erect, 7.5-11 a0 mm long, more or less smooth, glabrous; style conspicuously curved, moderately robust, not dilated towards the base, 8.0-9.2 mm long, usually glabrous but sometimes with scattered very short erect capitate hairs on the lower fifth, with dense, short, weak, simple hairs at the back of the indusium. Fruit d-valved, 15-25 mm long, not woody, opening annually, constricted between the articles, usually all articles formed, glabrous. Articles in two rows when mature, about 10-13 per row, ovoidal, 1 mm long, dark purple, with several vertical grooves.

-

Distribution Northern Territory: Central Australia North, and Central Australia South Regions; Western Australia: Mueller, Giles, Carnegie, and Helms Districts. Ecology This species has been recorded in deep red sand dunes, sandy loam plains or around the gravelly edges of lateritic breakaways. It occurs among mallees, desert

16

D. A. Morrison: Lechenaultia

Fig. 3. Holotype of Lechenaultia lutescens (Henry 334, NT33986).

Brunonia, Vol. 9, No. 1 (1986)

17

oak, open Triodia grassland, or Spinifex open scrub, where it is often locally common. Flowering apparently occurs sporadically throughout the year, perhaps only after recent rain. Notes L. lutescens varies somewhat between populations in leaf size, flower size and corolla colour but this does not seem to be significant. This species is closely related to L. heteromera, from which it can readily be distinguished by the yellow corolla and shorter calyx lobes. It has also been confused with L. striata (see below), which has a distinctly scapigerous flowering habit and ridged rather than grooved articles.

Specimens Examined NORTHERN TERRITORY: Central Australia North Region: 41 mls [66 km] from Mt Doreen on The Granites rd, Carolin 7952, 24.viii.1970 (AD97114087, SYD); Ngana outstation, 60 km SW. of Yuendumu, HenshaN 3341, 24.ii.1981 (CBG8110636, NSW, NT65726); 3 km E. of Chilla Well H.S., Latz 8101, 27.ix.1978 (CBG8101108, NT64574, 64575); c. 50 km SSW. of Vaughn Springs, Latz 8668, 30.iv.1981 (CANB317041, CBG8110635, NSW, NT66067); 0 . 5 ml [0.8 km] SE. of Chilla Well, Maconochie 1071,29.vii.1970 (AD97128082, NT27782, PERTH). Central Australia South Region: junction of Sandy Blight Junction Rd and Giles Rd, Carolin 6171, 6.viii.1967 (SYD); 5 mls [8 km] W. of Docker River Settlement, Henry 415, 9.iv.1972 (BRI146601, CANB237342, NSW, NT34911, PERTH); 22 km E. of Docker River, HenshaN 2818, 8.xi.1979 (NSW, NT60459); 8 km W. of Docker River Settlement, Henshall 3431, 11.iii.1981 (CANB317025, CBG8110634, NT66031); Bloods Range, Latz 2417, ll.iv.1972 (CANB234048, NSW, NT34795, PERTH); 72 km NE. of Docker River Settlement, Maconochie 1869, 27.viii.1973 (CANB298603, NT41606). WESTERN AUSTRALIA: Mueller District: 9 mls [14 km] W. of Dovers Hills, George 9020, 27.vii.1967 (PERTH, SYD); Giles District: 44 mls [71 km] W. of Giles Met. Station, George 8254, 3.x.1966 (NSW100422, PERTH); c. 36 mls [58 km] W. of Giles Met. Station, George 8257, 3.x.1966 (PERTH); NE. of Glen Helen, Rawlinson Range, George 8781, 20.vii.1967 (PERTH, SYD); 40 km SW. of Giles Met. Station, Latz 7661,29.v.1978 (BRI410315, NT072838); c. 26 km SE. of Giles Met. Station, Lazarides 8319, 9.v.1977 (CANB264795, NSW, PERTH). Carnegie District: NE. of Miss Gibson Hill, Gibson Desert, George 12004, 18.vii.1974 (PERTH); 53 km SW. of Warburton, George 12178, 27.1%.1974 (CANB291988, PERTH). Helms District: 16 mls [26 km] E. of Cosmo Newbery, George 8105, 28.ix. 1966 (PERTH).

8. Lechenaultia striata F. Muell., Fragm. 8:245 (1874); Tate, F1. Extratrop. S. Austral. 245 (1890); G. Chippendale, Proc. Linn. Soc. N.S.W. 96:262 (1972); Carolin in Jessop, F1. Centr. Austral. 360 (1981); J. Green, Census Vascular P1. W. Austral. 100 (1981); Grieve, How to Know W. Austral. Wildflowers Suppl. 45 (1982); Jessop, List Vascular PI. S. Austral. 36 (1983). Type: 'In Monte Olgae; Giles.' Holotype: Mount Olga, E. Giles, s. dat. (MEL1522797). Lechenaultia helmsii K. Krause, Pflanzenr. IV.54: 105 (1912); C. Gardner, Enum. P1. Austral. Occ. 125 (1930); Beard, Descr. Cat. W. Austral. P1. edn 2. 126 (1970); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 678 (1975). Holotype: Victoria Desert, Camp 44, R. Helms, 7.ix.1891 (K, n.v. photo SYD). Isotypes: AD96620166 & 97603293, MEL1522794 & 1522795, NSW76457-76459. Notes When Krause described L. helmsii, he noted that he had not seen any specimens of L. striata; there are no isotypes known, and the holotype was not available to him. If he had seen an authentic specimen it is doubtful he would have described

18

D. A. Morrison: Lechenaultia

the Helms specimens as a new species, as L. striata is very distinctive and relatively invariant. Krause was apparently confused by Mueller's protologue of L. striata, where the leaves and calyx lobes are described as similar to those of L. superba; this comment is confusing because it appears to be completely untrue. Since none of the specimens available to him fitted this protologue, Krause listed L. striata as a 'species dubia' and described the Helms collection as a new species. This lead was apparently followed uncritically by later botanists until the late 1960s. At this stage, Krause's name was incorrectly transferred to a series of recent collections of an unnamed species (now L. lutescens, see above) from Western Australia. This confusion apparently arose from the fact that the type of L. helmsii was collected from this area, while the L. striata type was from the Northern Territory (these being the only Lechenaultia 'species' known from these areas). In the early 1970s, several collections of this new species from the Northern Territory were also identified with Krause's name. Carolin (1981) only partly resolved this confusion by synonymising L. striata and L. helmsii, as the later misapplications were not explained and the undescribed species was not mentioned. 9. Lechenaultia brevifolia D. A. Morrison, sp. nov. Fruticuli caespitosi; caulibus floriferis scapiformibus; foliis caulibus floriferis ad basem confertis, 2.0-3-0 mm longis; bracteis et bracteolis 3.0-5.5 mm longis; sepalis 3.0-4.5 mm longis; corollis atrocaeruleis, obliquis, lobis inaequalibus, lobis superis alis perangustis vel absentibus. Holotypus: Between Red Kangaroo Hill and Yilgarri, R. Helms, xi.1891 (AD96620151). Isotypi: CANB351296, MEL (n.v.), NSW76545. (Fig. 4). The specific epithet is a reference to the very short leaves, which distinguish this species from the rest of the genus. Perennial subshrub, tufted, to 40 cm high and 50 cm diam., glabrous. Stems ascending, several to many from a common woody rootstock that may sucker, sparsely branched with the branches erect, terete, leafy stems 1 mm thick with the flowering stems narrower, woody only towards the base, glaucous or pale brown, striate. Leaves crowded on short leafy stems and on the lower parts of the flowering stems and becoming scattered distally on the flowering stems, fleshy, linear, 2.0-3.0 mm long, 0.6-0.8 mm wide, triquetrous, acuminate, more or less smooth, glabrous. Flowers in loose monochasial cymes of up to 5 flowers; bracts and bracteoles linear to linear-oblong, 3 SO-5.5 mm long, flat and usually with incurved margins, acuminate to mucronate, more or less smooth, glabrous. Calyx green, lobes scarcely overlapping at the base, linear to linear-subulate, 3.0-4.5 mm long, 0.5-0.7 mm wide, acute to acuminate, more or less smooth, glabrous. Corolla dark blue on the wings and lobes, usually white in the tube; tube cleft to the base, 7.0-8.5 mm long, outside glabrous, inside with dense, short, weak, simple hairs throughout; lobes not equal, spreading; inferior lobes narrow lanceolate, 7.0-10.0(-14.5) mm long, 0.8-1 a0 mm wide; superior lobes lanceolate-falcate, 6.0-9.0(-10.5) mm long, 0.7-0.9 mm wide; all lobes acute to acuminate, outside glabrous, inside with dense, short, weak, simple hairs on the lower third; wings on the inferior lobes triangular, 5 -0-8- 0(-10.0) mm long, 2.5-3.5(-5 0) mm wide, margins usually undulate-entire, transverse veins prominent; wings on the superior lobes rounded, 2.5-5.5 mm long, 0.1-1.0 mm wide, margins entire, transverse veins obscure. Stamens not enclosed in the tube; filaments thin, 3-0-4.0 mm long; anthers linear-oblong, 1.6-2-0 mm long, acute, yellow. Ovary erect, (6.0-)7.5-12.0(-15 -0) mm long, pale brown, more or less smooth, glabrous; style

Brunonia, Vol. 9 , No. 1 (1986)

Fig. 4. Holotype of Lechenaultia brevifolia (Helms s.n., AD96620151).

19

20

D. A. Morrison: Lechenaultia

conspicuously curved, moderately robust, not dilated towards the base, 6.5-8-0 mm long, with scattered very short erect capitate hairs on the lower half, with dense long weak simple hairs at the back of the indusium. Fruit 4-valved, 22-29 mm long, not woody, opening annually, not constricted between the articles, usually all articles formed, glabrous. Articles in two rows when mature, about 15-21 per row, cylindrical, 1 mm long, pale brown, with several vertical grooves.

Distribution South Western Australia: Coolgardie and Roe Districts. Ecology This species has been recorded from deep yellow or shallow red sand, and is usually found in low scrub or heath. Plants are usually scattered, but they are sometimes common in patches. Flowering is from late October to early December. Notes There is some variability in flower and ovary size between populations of L. brevifolia, but this does not seem to be significant. This species is closely related to L. heteromera, from which it can readily be distinguished by the distinctly scapigerous flowering stems and much darker corollas, and to L. juncea and L. striata which have much longer leaves and paler corollas. Specimens Examined WESTERN AUSTRALIA: Coolgardie District: 10 rnls [16 km] E. of Southern Cross, Beard 5177, 23.x.1967 (PERTH); 242 mile post on the Great Eastern Hwy, Demarz 4868, 20.xi.1973 (PERTH); near Gnarlbine, Helms, 12.xi.1891 (AD97603432, MEL1522739); near Moorine Rock, Lullfitz s.n., 14.iii.1980 (PERTH, SYD); Mt Moore, MerraN, 1889 (MEL1522747); 12 rnls [19 km] E. of Southern Cross, Wittwer 1351, 10.x. 1974 (PERTH); Victoria Spring, Anon., 3O.i~.1875 (MEL1522731). Roe District: Lake King-Norseman rd, between Rabbit Proof Fence and 100 Mile Tank, Beard 3785, 22.x.1964 (PERTH); Scaddan, Cranjield 1059, 6.xi.1978 (CANB323248, MEL597764, PERTH); 7 km W. of 90 Mile Tanks, Demarz 7987, 16.xii.1979 (PERTH); 10 rnls [16 km] E. of Newdegate, George 2270, 14.xii.1960 (PERTH); c. 85 rnls [I35 km] ENE. from Esperance Bay, Gwynne, 2.xi.1891 (AD97603433, MEL1522739, NSW76544); 46.2 rnls [74 km] E. of Lake King crossroads, Humphreys, 14.xi.1965 (PERTH); 42 rnls [67 km] E. of Lake King crossroads, Humphreys, 16.xi.1965 (PERTH); 29 rnls [46 km] E. of crossroads, Forrestania, Lullfitz 3861, 25.xi.1964 (PERTH); between RPF and crossroads, Forrestania, LuNJitz 3963, 8.xii.1964 (PERTH); 82.9 km E. of Lake King crossroads, Frank Hann N.P., Morrison 225, 29.xi.1984 (SYD); towards the Tone R, Muir, 1880 (MEL1522740); c. 35 rnls [56 km] E. of Hyden, Newbey 3322, 4.xii.1970 (PERTH, SYD); 6 km SW, of 90 Mile Tank, Newbey 6470, 12.xi.1979 (PERTH); 21 km SW. of 90 Mile Tank, Newbey 6522, 13.xi.1979 (PERTH); 5 km ENE. of Lake Ace, Newbey 8047, 16.xi.1980 (PERTH); W. of Salmon Gums, Frank Hann National Park, Royce 10229, 10.xii.1971 (PERTH); 7 rnls [I1 km] from Lake King towards Newdegate, Wrigley, 6.xi.1968 (CBG037596, PERTH); 72 km from Esperance towards Norseman, Wrigley 5321, 2.xi.1968 (CBG057830). CULTIVATED: Annangrove, N.S.W., Wrigley, 29.ii.1968 (CBG022981).

10. Lechenaultia juncea E . Pritzel, Bot. Jahrb. Syst. 35353 (1905); K. Krause, Pflanzenr. IV.54:105 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 678 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: 'Hab. in distr. Irwin in fruticetis arenosis pr. Watheroo Dec. flor. (E. Pritzel PI. Austr. occ. 982); et in locis similibus pr. Northampton flor. m. Nov. D.5636)'. Lectotype (here designated): Inter flumen Moore et

Brunonia, Vol. 9, No. 1 (1986)

21

Murchison, E. Pritzel 982, xi.1901 (AD97603266). Isolectotypes: BM (n.v.), K (n.v. photo SYD), L903.311-315 (n.v.), NSW76468, W (n.v.). Typification Pritzel based L. juncea on Diels 5636 and on his own collection 982. The Diels specimen has not been found and was presumably destroyed at B during World War I1 (cf. Hiepko 1978). The B specimen of Pritzel 982 apparently also has been destroyed, and one of the duplicates is selected as the lectotype. 11. Lechenaultia filiformis R. Br., Prodr. 1581 (1810); Don, Gen. Hist. 3:727 (1834); DC., Prodr. 7519 (1839); F. Muell., Fragm. 6:9 (1867); Benth., F1. Austral. 4:44 (1868); Bailey, Queensl. F1. 3:892 (1900); K. Krause, Pflanzenr. IV.54:108 (1912); Ewart & Davies, F1. N. Territory 267 (1917); Leenh., F1. Malesiana ser. 1. 5:338 (1957); G. Chippendale, Proc. Linn. Soc. N.S.W. 96:262 (1972). Type: Not designated. Lectotype (here designated): Carpentaria Island h, R. Brown, 20.xii.1802 (Britten sheet No. 2543) (BM, n.v. photo SYD). Isotypes: BM (n.v. photo SYD), K (n.v.), MEL1522620. Latouria filiformis (R. Br.)Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10: 187 (1854), based on L. filiformis R. Br. Lechenaultia agrostophylla F. Muell., Fragm. 6:8 (1867); Benth., F1. Austral. 4:44 (1868); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); Bailey, Queensl. F1. 3:892 (1900); K. Krause, Planzenr. IV.54:109 (1912); Ewart & Davies, F1. N. Territory 267 (1917); G. Chippendale, Proc. Linn. Soc. N.S.W. 96:262 (1972); J. Green, Census Vascular P1. W. Austral. 100 (1981), syn. nov. Type: 'Ad rivulos rupestres plagae elevatae (Sandstone-Tableland) juxta fluminis Victoriae partes australes passim proveniens.' Lectotype (here designated): Upper Victoria River, F. von Mueller, s. dat. (K, n.v. photo SYD). Syntypes: Depot Creek, Arnhem's Land, Ferd. Mueller, iii.1856 (MEL1522474); Arnhem's Land (MEL1522475).

Typification There are 11 elements mounted on the Britten sheet of Robert Brown's material at BM. The specimen second in from the upper left-hand corner is chosen as the lectotype of L. filiformis. No specimens of L. agrostophylla at MEL bear labels with locality data exactly as cited by Mueller in his protologue. There are two specimens labelled 'Arnhem's Land', one of which is also labelled 'Depot Creek' and dated 1856, which were presumably collected by Mueller on his north-west expedition in 1855-57. As he only travelled along the Victoria and Roper Rivers during this trip (see Willis 1949), these are probably syntypes. However, the K specimen is chosen as the lectotype of L. agrostophylla, as it is the only specimen which bears the exact locality data. Notes This species is very variable in flower size and colour, varying from small flowers (corolla about 11 mm long) with cream corollas to medium-sized flowers (corolla about 23 mm long) with dark blue corollas. This variation appears to correlate with longitude, as plants with the larger flowers predominate in the western part of the distribution and plants with the smallest ones predominate in Queensland and New

D. A. Morrison: Lechenaultia

22

Guinea. The number of articles per fruit (and hence fruit size) varies greatly within any one plant, as they also do in L. divaricata. In the past, this taxon has been divided into two species based on leaf width and fruit 'pedicel' length; however, this separation cannot be maintained. There is an infinite gradation between linear and filiform leaves, and the fruits are actually all sessile. The capsules sometimes appear to be pedicellate simply because the proximal seeds have not developed; this is often particularly true of the larger fruits. Sect. 111. LECHENAULTIA Lechenaultia R. Br. sect. Lechenaultia; Endl., Gen. P1. 508 (1838). Lectotype: L. formosa R. Br. Lechenaultia R. Br. sect. Euleschenaultia Benth., F1. Austral. 4:40 (1868), orthographic variant; Benth. & J. D. Hook., Gen. P1. 2537 (1876); Schonl. Nat. Pflanzenfam. IV.5:75 (1889); K.Krause, Pflanzenr. IV.54:97 (1912); C. Gardner, Enum. P1. Austral. Occ. 125 (1930). 12. Lechenaultia laricina Lindley, Sketch Veg. Swan R. 27 (1839); Paxton, Paxt. Mag. Bot. 14:246 (1847); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:185 (1854); Benth., F1. Austral. 4:41 (1868); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54:104 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 677 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: Not designated. Lectotype (here designated): Swan River, Drummond, 1839 [lst Coll.] (CGE). Possible isotypes: K (n.v. photo SYD), MEL 1522749. Lechenaultia splendens Hook., Bot. Mag. 72:t.4256 (1846); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:186 (1854). Type: Not designated (cult.). Lectotype (here designated): the figure, t.4256. Lechenaultia splendens var. stricta Hook., Bot. Mag. 72:with t.4256 (1846); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:187 (1854). Type: Not designated (cult.). [Lechenaultia tubifiora auct. non. R. Br.: Vriese in Lehm., P1. Preiss. 1:415 (1845), p.p.; Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:183 (1854), p.p. (both as to Preiss 1458 and Preiss 1459 only)]. [Lechenaultia formosa auct. non R. Br.: DC., Prodr. 7519 (1839); Benth., F1. Austral. 4:40 (1868), p.p. (as to Preiss 1458 only); K. Krause, Pflanzenr. IV.54:107 (1912), p.p. (as to Preiss 1458 only)]. Typification The specimen chosen as the lectotype of L. laricina is the only one in the Lindley herbarium at CGE. The specimens regarded as possible isotypes are the only other Drummond collections of this species that agree with the protologue and are not labelled as part of his collection 414 (not regarded as part of the type material because the lectotype is not numbered). L. splendens was described by Hooker from plants grown by Lucombe, Pince, and Co., who obtained the seed from James Drummond. There is only one specimen at K identified as L. splendens, but it is marked 'ex horto 1850' and thus cannot be regarded as a type specimen. Therefore, the illustration accompanying the protologue is designated as the lectotype.

Brunonia, Vol. 9, No. 1 (1986)

23

L. splendens var. stricta was described by Hooker from plants from the same source as the type variety. However, there are no specimens at K which fit the description and could be considered as a type, and there is nothing on the figure to illustrate this variety. This is thus a typeless name. The characteristics Hooker used to distinguish between the two varieties are no more than the usual morphological variability among plants of a single population. While Bentham and Krause merely misidentified the Preiss collection no. 1458, de CandoIle seems to have had a genuine misconception of L. formosa and L. laricina. He apparently considered Brown's L. formosa to be conspecific with Lindley's L. laricina, and therefore used Loddiges' name L. multijYora (a synonym of L. oblata anyway, see below) for what is treated here as L. formosa. Why he thought this is not clear, although the generally large number of specimens incorrectly identified as L. laricina in the Australian herbaria checked for this revision suggests that it is not easy to identify these species when they are pressed.

13. Lechenaultia superba F. Muell., Fragm. 6: 10 (1867); Benth., F1. Austral. 4:41 (1868); K. Krause, Pflanzenr. IV.54:104 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 677 (1975); J. Green, Census Vascular PI. W. Austral. 100 (1981). Type: 'Ad flumen Phillip's River, in collibus quarzosis. Mx.' Lectotype (here designated): On shattered quartz ranges, Phillips R., Maxwell, 26.vii.1863 (MEL1522806). Probable isotypes: K (n.v. photo SYD), MEL1522805. Typification There are two Maxwell specimens at MEL but only one bears locality data exactly as set out in Mueller's protologue, and this specimen is chosen as the lectotype. The other MEL collection has a label in Mueller's handwriting and presumably is a duplicate separated by Mueller at the time. 14. Lechenaultia hirsuta F. Muell., Fragm. 6:9 (1867); Benth., F1. Austral. 4:42 (1868); K. Krause, Pflanzenr. IV.54: 105 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 678 (1975); J. Green, Census Vascular PI. W. Austral. 100 (1981). Type: 'In Australia occidentali. Drummond (145).' Lectotype: SW. Australia, Drummond 145, s. dat. [6th Coll.] (MEL1522742). Isotypes: BM (n.v.), K (n.v. photo SYD), LD, MEL1522743, NSW76461.

Typification There are two sheets of Drummond 145 at MEL, both with labels in Mueller's handwriting. The better of these sheets has been selected as the lectotype. 15. Lechenaultia longiloba F. Muell., Fragm. 6:10 (1867); Benth., F1. Austral. 4:42 (1868); K. Krause, Pflanzenr. IV.54: 100 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 677 (1975); J. Green, Census Vascular PI. W. Austral. 100 (1981). Type: 'In Australia occidentali. Drummond.' Lectotype (here designated): W.A., J. Drummond, s. dat. [6th Coll.] (MEL1522783). Probable syntypes: K (n.v. photo SYD), MEL15227741522779, 1522781, 1522782, 1522784 & 1522787.

24

D. A. Morrison: Lechenaultia

Typification At MEL there are 12 sheets of unnumbered Drummond material, as well as a specimen of Drummond 179. This latter collection (also at BM) is not regarded as part of the type material as no collection number was cited by Mueller in the protologue. The specimen chosen as lectotype is the most complete specimen of the choices at MEL which best agrees with Mueller's protologue. All the other material may be isotypes but there is no evidence for this. 16. Lechenaultia tubijlora R. Br., Prodr. 1581 (1810); Don, Gen. Hist. 3:727

(1834); DC., Prodr. 7:519 (1839); Vriese in Lehm., P1. Preiss 1:415 (1845); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:183 (1854); Benth., F1. Austral. 4:41 (1868); E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54:103 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 676 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981). Type: Not designated. Lectotype (here designated): In ericetis ad Portum Regis Georgii 111~"in ora occidento australi Novae Hollandiae, R. Brown, xii.1801 (Britten sheet No. 2545) (BM, n.v. photo SYD). Isotypes: BM (n.v. photo SYD), K (n.v.), MEL1522812. Lechenaultia pinastroides Lehm., P1. Preiss 2:244 (1848); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:183 (1854). Type: 'In glareosis arenosisve inter montem Manypeak, Tjilberup, et Cape Riche, d. 23. Nov. 1840. Herb. Preiss. No. 430.' Lectotype (here designated): L. Preiss 430 (S). Isotypes: L903.311-321 (n.v. photo SYD); MEL1522813; W (n.v.). Ericopsis formosus C. Gardner, J. Roy. Soc. W. Austral. 9:42 (1923). Holotype: Hotham River, Popanyinning, C . A . Gardner 1880, 5.xii.1922 (PERTH). Isotypes: PERTH. Typification There are six specimens of Robert Brown's collection of L. tubiflora mounted on the Britten sheet at BM. The left-hand specimen is chosen as the lectotype. When choosing lectotypes of Preiss specimens the preferred reference collection is that at LD (cf. Crisp 1983). However, there is no specimen of Preiss 430 at LD. Therefore, the S collection has been chosen as the most suitable lectotype of L.pinastroides, as this is part of Lehmann's own herbarium. The right-hand specimen is selected as the lectotype. Notes Apart from corolla colour and leaf size, which can vary considerably within a single population, L. tubiflora varies markedly in stature. The larger shrubby plants with erect stems have greenish cream corollas, greener, less scarious leaves, and are confined to the coastal strip between Albany and Cape Le Grande, usually occurring within 15 km of the coast. The prostrate plants usually have corollas that are red or tinged with red, greyer more scarious leaves, and are widespread in both the coastal and inland districts.

17. Lechenaultia linarioides DC., Prodr. 7519 (1839); Benth., F1. Austral. 4:40 (1868); E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54:102 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 676 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981).

Brunonia, Vol. 9, No. 1 (1986)

25

Type: 'In Nova-Hollandia ad Swan-river legit cl. Drummond.' Holotype: Swan River, Mr Drummond, 1839 [lst Coll.] (G-DC, n.v. microfiche). Probable isotype: (K, n.v. photo SYD). Scaevola grandiflora Benth. in Endl., Enum. PI. 70 (1837); DC., Prodr. 7:512 (1839). Type: 'Swan-River. (Hugel.).' Holotype: Swan River, Hugel, s. dat. (W., n.v. photo SYD). Lechenaultia arcuata Vriese in Lehm., PI. Preiss 1:4l6 (1845); Hook., Bot. Mag. 72:t.4265 (1846); Paxton, Paxt. Mag. Bot. 14:245 (1847); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10:186 (1854). Type: 'In arenoso-calcareis prope urbiculam Freemantle, d. 17 m. Dec. 1838. Herb. Preiss. No. 1465.' Lectotype (here designated): In arenoso-calcareis prope urbiculam 'Freemantle', L. Preiss 1465, 17.xii. 1838 (LD). Isotypes: L903.311-319 (n.v. photo SYD) & 903.31 1-318 (n.v.), MEL1522750. Lechenaultia grandzxora (Benth.)Druce, Bot. Exch. Club Brit. Isles Rep. 4:632 (1917), based on S. grandiflora Benth.

18. Lechenaultia chlorantha F. Muell., Fragm. 2:20 (1860); Benth., F1. Austral. 4:40 (1868); J. Green, Census Vascular P1. W. Austral. 100 (1981); Grieve, How to Know W. Austral. Wildflowers Suppl. 45 (1982). Type: 'In vallibus rupestribus ad flumen Murchison.' Holotype: Rocky gully, Murchison R., ?Oldfield, s. dat. (MEL1522479). Isotype: K (n.v. photo SYD). Lechenaultia formosa var. chlorantha (F. Muell) K. Krause, Pflanzenr. IV.54: 108 (1912), based on L. chlorantha F. Muell.; Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 679 (1975).

There is only one specimen at MEL which bears locality data as cited by Mueller in the protologue; however, this specimen has no collector's name. The specimen at K, which is presumably a duplicate, bears Oldfield's name. Notes Bentham (1868) expressed doubts about the taxonomic status of this species but only because he used the shape of the corolla lobe tips to distinguish between it and L. formosa. He thus included in L. chlorantha some specimens from the Kalgan River area which are included here in L. formosa. The corolla lobes are in fact fairly variable within L. forrnosa and L. chlorantha, and thus are not useful for distinguishing between the two taxa. Krause (1912) followed Bentham's lead to its logical conclusion and reduced L. chlorantha to a variety of L. forrnosa. However, while L. forrnosa is very variable, L. chlorantha s. str. is distinctive in a number of ways: the leaves are generally longer and thinner with a more punctate-rugose surface, the bracts and bracteoles are much longer and thinner, the flowers are generally much larger (i.e. longer and wider sepals, corolla lobes and corolla wings), the corolla is green rather than orange or red, the stamen filaments are thinner and longer, the style is apparently glabrous, and the articles are larger. This suggests that, despite the obvious similarities between the two taxa, L. chlorantha is worthy of specific status. 19. Lechenaultia fsrmosa R. Br., Prodr. 1:581 (1810); Sims, Bot. Mag. 52:t.2600 (1825); Ker Gawler, Bot. Reg. 1l:t.916 (1825); Sweet, F1. Australasica t.26

26

D.A. Morrison: Lechenaultia

(1827); Don, Gen. Hist. 3:727 (1834); Vriese in Lehm., PI. Preiss 1:414 (1845); Vriese, Natuurk. Verh. Holl. Maatsh. Wetensch. Haarlem ser. 2. 10: 185 (1854); Benth., F1. Austral. 4:40 (1868); Schonl., Nat. Pflanzenfam. IV.5:75 (1889); E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54: 107 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 679 (1975); J. Green, Census Vascular P1. W. Austral. 100 (1981); Jessop, List Vascular PI. S. Austral. 36 (1983). Type: Not designated. Lectotype (here designated): Bay I, South Coast, R. Brown, i.1802 (Britten sheet No. 2546) (BM, n.v. photo SYD). Isotypes: BM (n.v. photo SYD), K (n.v.), MEL1522652. Lechenaultia oblata Sweet, F1. Australasica t.46 (1828); Don, Gen. Hist. 3:727 (1834); F. Muell, Fragm. 6:226 (1868). Type: Not designated (cult.). Lectotype (here designated): the figure, t.46. Lechenaultia baxteri Don in Loudon, Hort. Brit. 79 (1830), nom. superfl., syn. L. formosa R. Br. Lechenaultia multi'ora G. Lodd., Bot. Cab. l6:t. 1579 (1830), nom. nud. Lechenaultia multz@ora G. Lodd. ex DC., Prodr. 7:519 (1839), nom. supfl., syn. L. oblata Sweet. Lechenaultia formosa var. oblata (Sweet) E. Pritzel, Bot. Jahrb. Syst. 35552 (1905), based on L. oblata Sweet; K. Krause, Pflanzenr. IV.54:107 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 679 (1975). [Lechenaultia chlorantha auct. non F. Muell.: Benth., F1. Austral. 4:40 (1868), p.p. (as to 'specimens from the Kalgan river, Oldfield' only)].

There are six elements on the Britten sheet of Robert Brown's collection at BM. The lower left-hand specimen is chosen as the lectotype of L. formosa. Sweet described L. oblata from a plant cultivated by Robert Barclay, of Buryhill, from seed collected by William Baxter. As no specimen has been found that could be considered to be a type, the figure accompanying the protologue is designated as the lectotype. In the protologue of L. baxteri, Don refers to both the 'Botanical Register' and the 'Botanical Magazine' of 1825. Both of these references are to descriptions and illustrations of L. formosa. The name, then, is valid but it is nevertheless a superfluous synonym of L. formosa. Both illustrations referred to appear to have been also derived from material collected by William Baxter, and this is presumably why Don refers to them. In the 'Botanical Cabinet' of 1830, Loddiges illustrates but does not describe a new species named L. rnultzjlora; thus this is a nom. nud. The illustration is a representation of L. formosa, and Loddiges notes that it was based on cultivated material obtained from Robert Barclay, the same source from which Sweet obtained his material for L. oblata several years earlier. De Candolle validated the name L. multi'ora in 1839 by providing a description; and he referred directly to Loddiges as the source of this name. However, he incorrectly gave the date of Loddiges' publication as 1826 (although dated 1829, it was in fact issued in June 1830; cf. Stafleu and Cowan 198I), and thus considered the name to be a prior synonym of L. oblata. De Candolle also considered that the 'Botanical Magazine' and 'Botanical Register' illustrations of L. formosa (and consequently the name L. baxterq are in fact referable to L. oblata, presumably following Don (1834).

Brunonia, Vol. 9, No. 1 (1986)

27

Notes L. formosa varies greatly in flower and leaf size as well as in corolla colour and habit. Large-flowered prostrate plants occur throughout the geographical range but the smaller-flowered more erect forms are confined to the coastal strip between Albany and Cape Le Grande, usually occurring within 25 km of the coast. The variation in flower size (from corollas about 14 mm long to those nearly twice this size) is suggestive of chromosomal duplication in some populations (cf. Peacock 1963). In the very northern part of the distribution the plants have leaves that are much longer and much paler green than those of plants throughout the rest of the range, and this gives these plants a superficial resemblance to L. chlorantha. However, the corollas remain an orange-red colour rather than pale green. In the Eucla and eastern Roe Districts, this species is confined to the small patches of siliceous sand topsoils. These plants are erect, have much thicker leaves than the plants elsewhere, erect ovaries, and the cohering margins of the superior corolla lobes always have small wings. This prompted Parsons (1970) and Nelson (1974) to identify these plants as L. tubiflora or L. aff. tubiflora. However, the plants are still recognisably L. formosa because of the reflexed inferior corolla lobes. On the southern slopes of Frenchman's Peak are a number of plants that are erect, have thickened leaves with acuminate tips, erect ovaries, wings on the cohering margins of the superior corolla lobes, and more erect inferior corolla lobes than usual. Wrigley and Fagg (1979) postulated a hybrid origin of these plants, from a cross between L. formosa and L. tubiflora. Species Incertae Sedis Lechenaultia tubiflora var. purpurea E. Pritzel, Bot. Jahrb. Syst. 35:552 (1905); K. Krause, Pflanzenr. IV.54: 104 (1912); Grieve & Blackall, How to Know W. Austral. Wildflowers edn 2. 676 (1975). Type: 'Hab. in distr. Eyre pr. Gibsons Soak in arenosis, flor. m. Nov. (D. 5973)', n.v. Typification Diels 5973 has not been found at B and was presumably destroyed (cf. Hiepko 1978). No other specimens (out of 152 examined) have been found that fit the protologue; in particular, 'floribus . . . saturate purpureis' is unmatched. Species Exclusa Lechenaultia humilis (R. Br.) Sprengel, Linn. Syst. Veg. edn 17. 1:72O (1824)=Anthotium humile R. Br., Prodr. 1582 (1810). Acknowledgments Most of those type specimens cited in this paper that have not been seen by me have been checked by Roger Carolin, who kindly made all of his notes available to me in preparing this revision; Chas Chapman, Neville Marchant, Ken Newbey, Sue Patrick and Rae Paynter were very helpful and showed me much kindness during my limited time in Western Australia; Kerri Gallagher spent a lot of her spare time helping with the more mundane parts of the work, as well as keeping me company in W.A.; Nick Lander was very helpful as ABLO at Kew; Alex George was kind enough to answer all my queries; Geoff Burrows spent much of

D. A. Morrison: Lechenaultia

28

his time improving the usability of the revision; Malcolm Ricketts took the photos; Surrey Jacobs helped with the Latin; Roger Carolin, Jim Armstrong and Peter Weston commented on an earlier version of the manuscript; the Directors of AD, BRI, CGE, LD, NT, PERTH and S kindly loaned me specimens under their care; the Directors of CANB, CBG, MEL, NSW and SYD allowed me to see the specimens under their care; the Directors of FI-W and W sent me photographs of type specimens; the Western Australian Fisheries and Wildlife Department and the National Parks Authority of Western Australia gave me permission to collect plants under their protection; and the research was funded by the Australian Biological Resources Study, the Linnean Society of N.S.W. (through the Joyce W. Vickery Scientific Research Fund) and the University of Sydney.

References

.

Bentham, G. (1837). In 'Enumeratio Plantarum . . . Collegit . . Baro de Hugel.' (Ed. S. L. Endlicher.) p. 70 (Beck: Vienna.) Bentham, G. (1868). 'Flora Australiensis.' Vol. 4. (Reeve: London.) Carolin, R. C. (1959). Floral structure and anatomy in the family Goodeniaceae Dumort. Proc. Linn. SOC.N.S. W. 84, 243-55. Carolin, R. C. (1960). The structures involved in the presentation of pollen to visiting insects in the order Campanulales. Proc. Linn. Soc. N.S. W. 85, 197-207. Carolin, R. C. (1966). Seeds and fruit of the Goodeniaceae. Proc. Linn. Soc. N.S. W. 91, 58-83. Carolin, R. C. (1967). The concept of the inflorescence in the order Campanulales. Proc. Linn. Soc. N.S. W. 92, 7-26. Carolin, R. C. (1971). The trichomes of the Goodeniaceae. Proc. Linn. Soc. N.S. W. 96, 8-22. Carolin, R. C. (1977). The systematic relationships of Brunonia. Brunonia 1, 9-29. Carolin, R. C. (1981). In 'Flora of Central Australia'. (Ed. J. P . Jessop.) p. 360. (A. H. & A. W. Reed: Sydney.) Carolin, R. C. (1982). A review and critique of studies on the phytogeography of arid Australia. In 'Evolution of the Flora and Fauna of Arid Australia'. (Eds W. R. Barker and P. J . M. Greenslade.) pp. 119-23. (Peacock Publications: Adelaide.) Crisp, M. D. (1982). Evolution and biogeography of Leptosema (Leguminosae: Papilionoidae). In 'Evolution of the Flora and Fauna of Arid Australia'. (Eds W. R. Barker and P. J . M. Greenslade.) pp. 3 17-22. (Peacock Publications: Adelaide.) Crisp, M. D. (1983). Plantae Preissianae types at Lund. Austral. Syst. Bot. Soc. Newsletter No. 36, 4-6. De Candolle, A. P. (1839). 'Prodromus Systematis Naturalis Regni Vegetabilis.' Vol. 7. (Treuttel & Wurtz: Paris.) Don, G. (1834). 'A General History of the Dichlamydeous Plants.' Vol. 3. (Rivington: London.) Farris, J. S. (1970). Methods for computing Wagner trees. Syst. Zool. 19, 83-92. Gardner, C. (1978). 'Wildflowers of Western Australia.' 13th edn. (West Australian Newspapers Ltd: Perth.) Hiepko, P . (1978). Die erhaltenen teile der sammlungen des Botanischen Museums Berlin-Dahlem (B) aus der zeit vor 1943. Willdenowia 8, 389-400. ICBN (1983). 'International Code of Botanical Nomenclature.' (Eds E. G. Voss et al.) Regnum Veg. 111. Krause, K. (1912). Goodeniaceae und Brunoniaceae. PBanzenreich (IV) 54, 97-109. Nelson, E. C . (1974). Disjunct plant distributions on the south-western Nullarbor Plain, Western Australia. J. R. Soc. W. Aust. 57, 105-17. Parsons, R. F. (1970). Mallee vegetation of the southern Nullarbor and Roe Plains, Australia. Trans. R. SOC.S. Aust. 94, 227-42. Peacock, W. J. (1963). Chromosome numbers and cytoevolution in the Goodeniaceae. Proc. Linn. Soc. N.S. W. 88, 8-27. Stafleu, F. A., and Cowan, R. S. (1981). 'Taxonomic Literature.' Vol. 111, 2nd edn. Regnum Veg. 105. Watrous, L. E., and Wheeler, Q. D. (1981). The outgroup comparison method of character analysis. Syst. Zool. 30, 1-1 1. Willis, M. (1949). 'By Their Fruits: a Life of Ferdinand von Mueller, Botanist and Explorer.' (Angus & Robertson: Sydney.) Wrigley, J. W., and Fagg, M. (1979). 'Australian Native Plants.' (Collins: Sydney.)