Systematics of Stenosiphonium Nees (Acanthaceae): the Legacies of Hooker and Thompson

Mark A. Carine

The publication, in 1855, of Joseph Hooker and Thomas Thomson's Flora Indica was planned as the first volume of a detailed and exhaustive account of the Indian flora. The aim was to "present a systematic account of the vegetable products of British India arranged according to natural principles [and] besides the descriptions.....all matters of importance connected with anatomical, structural, morphological and physiological points, wherever it is practicable, [were to] be treated and in other cases pointed out as subjects worthy of future attention".

The need for a flora of British India was widely recognised, and large collections of Indian plants, including The East India Company Herbarium ('The Wallich Collection'), and the herbarium of Robert Wight from southern India, were available for use in its preparation. In addition, Hooker and Thomson had gathered their own herbarium (labelled 'Herb. Ind. Or. Hook. fil. T. Thomson'), comprising their collections from northern India, supplemented by the herbarium of southern Indian plants made by Thomas Thomson's brother, Gideon Thomson, and by the collections of several other botanists. Upon their return to Kew from India in 1851, Hooker and Thomson set about sorting and distributing their herbarium, and preparing Flora Indica (see Burkill, 1965 for details)

The descriptions in the first volume were very detailed, and completion of the work would clearly have been both time consuming and very costly. Indeed, it was the problem of money, and more specifically the refusal of the East India Company to sponsor Flora Indica that led to its abandonment after the first volume (Desmond, 1992).

However, whilst work on Flora Indica was halted at an early stage, Hooker and Thomson's herbarium was to prove a valuable source of material for people working on the Indian flora. Thomas Anderson, Superintendent of the Calcutta Botanic Garden from 1861-1869, was amongst those, and whilst working at Kew, he made use of the herbarium in producing his Enumeration of the Indian species of Acanthaceae published in 1867. This work included an account of Stenosiphonium Nees, a genus of the Acanthaceae currently under study at Oxford.

In his Enumeration, Anderson described two new species of Stenosiphonium, namely Stenosiphonium setosum and Stenosiphonium parviflorum. S. setosum was based on a single specimen from the Hooker-Thomson herbarium, collected by Gideon Thomson, and originally labelled 'Strobilanthes 67'. Three specimens were cited in the account of S. parviflorum, including another Hooker-Thomson herbarium specimen, again collected by Gideon Thomson, but labelled 'Strobilanthes 68'.

Hooker and Thomson had clearly thought the two plants similar, placing them closely together (as Strobilanthes 67 & 68) and Anderson's only criteria for distinguishing the two species were the size of flower parts, as the epithet parviflorum ('small flower') would suggest . The two other specimens cited in the protologue account of S. parviflorum (one in the Wallich Collection, and one in Cleghorn's herbarium) upon close inspection are not conspecific with the Thomson specimen. Both have small inflorescences, but are immature specimens of S. cordifolium (Vahl) Alston, a widely distributed and relatively common species. This was recognised by Clarke (1885) and Gamble (1935). Size of floral parts however, was clearly a significant factor in Anderson's delimitation of species in this group.

A number of specimens currently on loan to Oxford had been determined as S. parviflorum. However, all proved to be species of Strobilanthes, the large, widely distributed and morphologically variable genus with which Stenosiphonium has long been associated. Thus, the only material of Anderson's two new species available were duplicates of the Gideon Thomson specimens, 'Strobilanthes 67 & 68', from Hooker and Thomson's herbarium.

Little is known about the collections of Gideon Thomson. The Hooker-Thomson herbarium labels lack detail, with the locality of both the S. parviflorum specimen, and that of S. setosum given as 'Maisor [Mysore] and Carnatic' covering a large area of southern peninsular India. We do know from Flora Indica that the herbarium was gathered, chiefly by means of collectors, from the plain of the Carnatic (chiefly in the neighbourhood of Madras), the Nilgiri and Curg [Coorg] mountains, and in the Courtallum Hills. However, no more precise locality details were available. Our knowledge of these plants, based on material in European and North American herbaria was extremely limited.

Of a genus thought to contain six species (Mabberley, 1987), two were known only from a single collection, made before 1855, from an unknown locality. It was therefore with great interest, but low expectations of finding material of these species that I embarked on a visit of Indian herbaria during March and April of 1996 to study their collections of Stenosiphonium and Strobilanthes. Many herbaria containing important historical as well as recent material were consulted. These included the Central National Herbarium in Calcutta (CAL), BSI Industrial Section Herbarium, Calculla (BSIS), BSI Southern Circle Herbarium, Coimbatore (MH), Rapinat Herbarium, Tiruchirapalli (RHT), French Institute, Pondicherry (HIFP), Tropical Botanical Garden and Research Institute, Trivandrum (TBGT), Presidency College Herbarium, Madras (PCM), and the Herbarium of the University of Kerala.

Two of the herbaria, the Central National Herbarium, and the Tropical Botanic Garden and Research Institute (TBGRI) contained Stenosiphonium specimens collected in the early part of this century, in 1912 and 1914 by M. Rama Rao and K. Vencoba Rao respectively, for the 'Flora of Travancore'. Rama Rao's specimen had been identified as S. parviflorum (it has small flowers), and Vencoba Rao's as S. russellianum Nees (a synonym of S. cordifolium). Further, researchers at TBGRI had collected living material of Stenosiphonium from the evergreen forests of the Western Ghats in Kerala during 1993 (Thomas et al., 1995). This material, in the living collection at TBGRI, along with the two 'Flora of Travancore' specimens shared the characters that distinguish Anderson's two species from other species of Stenosiphonium i.e. lanceolate bracts and the presence of two stamens and two staminodes. However, placing the specimens into either of Anderson's two species was difficult. Whilst the two Thomson specimens, very different in size, were easy to separate, the increased sample made distinguishing between the two species impossible. Detailed study of the specimens showed that Anderson's two species are conspecific.

Clearly, there is a need for adequate sampling in resolving problems of species delimitation, and problems of synonymy will result from the description of species using few specimens. This was recognised by Hooker and Thomson in their Introductory Essay to Flora Indica. They rather boldly estimated that "one half of the recorded species of Indian plants are spurious, and that in many natural orders the undescribed species hardly equal in number those which require to be cancelled". 'Hair-splitting naturalists' who " separate accidental forms by trifling characters" were blamed for this, and they went on to urge botanists to "endeavour, by selecting good suites of specimens, produced under all variations of circumstances, to determine how few not how many species are comprised in the flora of [a] district".

'Good suites of specimens' however are not always available. In all of the herbaria consulted, I have found just three further specimens conspecific with 'Strobilanthes 67 & 68' of the Hooker-Thomson herbarium. Similarly, in another species of Stenosiphonium, Stenosiphonium wightii Brem. only seven collections are recorded. Whilst a lack of specimens probably reflects the rarity and limited distribution of these taxa, it does make species delimitation difficult, as Anderson's attempts prove.

In Strobilanthes, a genus closely related to Stenosiphonium, forty six species are recognised in the former Madras Presidency, including a large number of endemics (Gamble, 1935). In this genus we find a similar problem, with many of the endemic species known only from type material, or from limited collections, often decades old.

It is over one hundred and forty years since Hooker and Thomson set out their aims for Flora Indica, and identified the requirements for a rigorous investigation of taxonomic problems in the Indian Flora. Today, at least in Stenosiphonium and Strobilanthes, we are still far from meeting those requirements. Under-collection still poses a significant barrier to our understanding of these groups.


Anderson T. 1867. An Enumeration of the Indian Species of Acanthaceae. Journal of the Linnean Society 9: 425-526.

Burkill IH. 1965. Chapters on the History of Botany in India. Delhi: Goverment of India Press.

Clarke CB. 1885. Acanthaceae. In: Hooker JD, ed. Flora of British India. Vol. IV. London: Reeve & Co.

Desmond R. 1992. The European Discovery of the Indian Flora. Oxford: Oxford University Press.

Gamble JD. 1935. Flora of the Presidency of Madras. Vol III. London: Adlard & Son Ltd.

Hooker JD and Thomson T. 1855. Flora Indica. Vol. I. London: W Pamplin.

Mabberley DJ. 1987. The Plant-Book. Cambridge: Cambridge University Press.

Manilal KS. 1988. Flora of Silent Valley. Calicut: Mathrubhumi Press.

Thomas J, Santhosh Kumar ESS, Jabbar A. 1995. Stenosiphonium wightii Bremek. (Acanthaceae): A new record from Kerala forests. Journal of Economic and Taxonomic Botany. 19 (2): 483-485.