Tuesday, December 25, 2012

Problems in Australian Taxonomy

I have never been happy with Robert Geesink's treatment of Portulaca in Blumea 17(2): 299, 1969.

Geesink unfortunately lumped all of the terete-leafed Indo-Australian species in with the North American species Portulaca pilosa L. He also wrongly assumed that P. pilosa was a species of tropical Indo-Pacific origin rather than American. This caused much confusion. It soon became apparent that Geesink's Portulaca pilosa subsp. decipiens (V. Poelln.) could not be easily separated from two apparently endemic species, namely P. filifolia F. Muell. and P. australis Endl. Moreover the degree of variation was simply too vast to be covered by Geesink's Portulaca pilosa subsp. pilosa.

Hellmut Toelken ponders this dilemma when he authored the section on Portulaca in J. Jessop's Flora of Central Australia (1981) in the late 1970s. On page 40 Toelken writes: "a wider knowledge of Australian populations will show whether P. filifolia, an annual with fibrous roots and a seed coat with all cells nippled, and P. remota a perennial with tuberous roots and a seed coat with cells along the margin nippled, should be included in the wide range of variation of the subsp. pilosa."

An interrogation of the SA Plant Census database shows no entries for P. pilosa or P. remota. In fact all South Australian occurrences of P. pilosa were re-determined in 1981 as P. filifolia F. Muell. The revision unfortunately did not make it into the Flora of Central Australia, which had been published earlier that same year. There were evidently no records of the naturalized American species P. pilosa in South Australia at that time, although it was already common and widespread in New South Wales and Queensland and in the tropical half of Western Australia. However it appears to have unfortunately been confused with at least one yellow-flowered native species in Western Australia, and it is still mistakenly regarded as a native species there. P. pilosa has been recorded from the far north of South Australia in more recent times.

The other species mentioned by Toelken, Portulaca remota Poelln. is now regarded as an unresolved name and is therefore no longer recognized, at least not in Australia. The name may have been applied at a time when botanists did not realize that P. filifolia generally has a small tuber in the early stages of its growth, which it frequently loses upon reaching maturity. At this time the roots become fibrous (from my personal observations of cultivated specimens). However I cannot explain the differences that Toelken noted in the seed coat.

It is difficult to understand how Geesink reconciled the yellow-flowered, mostly upright plants of the native species P. filifolia Muell. with the mostly prostrate, bright pink to purple-flowered plants that we now know to be the tropical American species P. pilosa L. Yet he appeared to be content to suggest that both are mere manifestations of a species that he presumed to be highly variable. One possible reason for this assumption is that Geesink had, at that time, narrowed the focus of his research to the Indo-Pacific region. Considering that P. pilosa already had a widespread presence in this region by 1969, the species must have appeared to Geesink to be representative of a species complex that was endemic to this region.

Geesink's misplaced application of the name P. pilosa in Australia was unfortunate, but he at least understood the wide degree of variation that was evident in Portulaca filifolia Muell. The extremes of this species are illustrated well on page 40 of Flora of Central Australia.

I have long noticed that some plants that key to P. filifolia from the northern tropics frequently have a larger corolla (10-16 mm long), a more upright growth habit and very sparse foliage. These plants are a very good match for the Toelken description of Geesink's Portulaca pilosa subsp. decipiens (V. Poelln.) shown in Fig. 62 above. The seeds of these plants need to be scrutinized carefully against the illustrations and descriptions in Toelken, which show a colliculate seed coat (i.e. one that is covered with lots of small mounds). Toelken describes the seed coat further as having flat oblong elevations that are arranged in dense concentric rings. If any specimens match this description, then these plants may represent a taxon that is indeed distinct from P. filifolia. In this case the Portulaca decipiens of Karl von Poellnitz (1933) should probably be reinstated, given that Geesink gives the distribution of the taxon as "Australia, north of the Tropic of Capricorn".

The type of P. decipiens Poelln. was collected at "Fanny Bay" (presumably Fannie Bay near Darwin) in December 1869 (Coll. No. Fr. Schultz 705). In his description of P. decipiens on page 159 of Repertorium specierum novarum regni vegetabilis XXXIII (1933/1934), von Poellnitz states that the flowers were not seen and this may be the reason why he seems reluctant to place P. filifolia Muell. in synonymy. He writes (German translation):

It is certainly close to P. filifolia F. Muell. and P. sclerocarpa A. Gray*. These two growing in Australia species have also alternate, very narrow leaves and clearly visible axillary hair. P. filifolia has even narrower, threadlike leaves, longer and more numerous axillary hair, and a capsule that opens below the centre. P. sclerocarpa* has also numerous and long axillary hair, a very hard, at the bottom opening capsule, and seeds with a stellate pattern.

Although the "Fanny Bay" specimen was lacking flowers and the description provided by von Poellnitz was rather sketchy, Geesink nevertheless became convinced that P. decipiens Poelln. belonged to a taxon distinct from P. filifolia Muell. He clearly had access to more recent specimens that could be split on the basis of consistent characters.

Hellmut Toelken (1981) clarified the distinction further by separating the two mainly on the basis of flower size, leaf length, and seed texture. But he also suggested that only a greater understanding of Australian populations can resolve the problems that he noted with abberant specimens.

The names of the two taxa could possibly be reworked now as P. filifolia Muell. subsp. filifolia and P. filifolia Muell. subsp. decipiens (Geesink ex Poelln.). This would of course be on the proviso that a sufficient number of consistent specimens be found that are not merely representing extreme forms within the total population. The full range of variation must first be established.

Typical specimens of P. filifolia have an erect to decumbent habit, corolla 4-10 mm long, and seed coat stellate and tuberculate with all or at least the marginal cells prominently nippled (Fig 61). This is true for most of the specimens found throughout inland Australia.

Specimens of P. filifolia differ in all cases from P. australis by the elongated stigma which always places the stigmatic lobes above the stamens.

Bentham and Mueller's key has stood the test of time. Even the somewhat obscure P. napiformis of Mueller is still recognized today. It occurs in far NW Western Australia apparently as a distinct taxon, yet clearly has affinities with P. australis and P. filifolia. I have never seen this species so have no first hand knowledge of it. Mueller described it as a taxon similar to P. australis but with 4 stigmatic lobes, a lack of axillary hairs, an oblong-shaped tuber, very short branches, and black seeds similar to but smaller than P. oleracea. A photo of P. napiformis appears on the Western Australia FloraBase site. I have seen remarkably similar plants here in Queensland, however the resemblance may prove to be superficial.

Footnote: * Portulaca sclerocarpa A. Gray is now known to be native to the islands of Hawaii where it is endemic. The species does not occur in Australia.

No comments: