The typical habitat of Haworthia koelmaniorum variety koelmaniorum in the western part of its distribution range. This photograph was taken on 28 August 2008 and indicates the fragmented nature of grass fires. The burnt hillside in the background is home to a large colony of Haworthias which grow amongst fixed rocks.

SUMMARY
Haworthia koelmaniorum is a very interesting cryptic species that frequents both a savannah and a grassland biome in western Mpumalanga. The plants have been studied by Charles Craib for about 20 years and this account discusses the findings drawn from this research. Some of the work on Haworthia koelmaniorum variety mcmurtryi has been published and that on Haworthia koelmaniorum variety koelmaniorum is presently in preparation for publication. This latter research is discussed with particular reference to the unique ways in which a fire driven ecology has produced variation in these plants. A recommendation is also made that H. koelmaniorum is treated as a variable plant without any rank below species level.

NEWSLETTER
THE HISTORY, DISTRIBUTION AND HABITAT OF HAWORTHIA KOEL-MANIORUM
H. koelmaniorum was described in 1967 by Mrs. Mauve (Amelia Obermeyer). The plant was brought to the attention of the then Botanical Research Institute by Mrs. T.H.C. Koelman from the Groblersdal area of the central Transvaal (now western Mpumalanga). At the time of her description of the new species Amelia Obermeyer-Mauve remarked that it was with considerable surprise, that this Haworthia was found in the Groblersdal district. It was moreover of great scientific interest that these plants were discovered so far away from the predominantly Cape distribution of the Genus (1967, 1502).

H. koelmaniorum continued after the type description to be found at various places in the Groblersdal area and a map of its currently known distribution is provided by Pieter Bosch (2004: 57). In this map the plants are shown as occurring mostly in a large area south south east and south west of Groblersdal with a narrow tongue of distribution in the hills and mountains in the extreme south west.

Haworthia mcmurtryi now Haworthia koelmaniorum variety mcmurtryi was described by Charles Scott in 1984, (1985: 140 – 141). The plants were originally found by Doughlas Mcmurtryi and the late Mary Bleck in typical highveld hilly grassland with Faurea saligna and Protea caffra two types of trees often found close to the transition between highveld and savannah. H. mcmurtryi was reduced to varietal status under Haworthia koelmaniorum by Bayer in 1999 (1999: 180 - 181). A sophisticated account, upholding the varietal status of H. koelmaniorum variety koelmaniorum and H. koelmaniorum variety mcmurtryi was published by Pieter Bosch in 2004 (2004: 56 – 60). Bosch was the first researcher to look at how the plants live in their environment. He related geological, pedalogical and climatic influence to variety generation with respect to the two varieties of Haworthia koelmaniorum.

Charles Craib questioned the varietal status of H. koelmaniorum variety mcmurtryi in a paper entitled “The Autecology and Status of Haworthia koelmaniorum variety mcmurtryi (Craib 2007: 22 – 27). He argued that the retuse leaved “mcmurtryi” were simply adaptations to a fire driven ecology under circumstances where longer leaved haworthias, the typical “koelmaniorum”, would not be able to survive frequent dry season grass fires.

In a current detailed research paper in preparation (Craib, September 2008) Charles evaluates the status of the two varieties of Haworthia koelmaniorum. Some of the findings are discussed below with specific reference to H. koelmaniorum var. koelmaniorum in the western part of its distribution range and H. koelmaniorum var. mcmurtryi in its typical hilly highveld grassland habitat.

THE CONCEPTS OF SPECIES, SUBSPECIES AND VARIETY IN THE ORGANISATION AND PRESENTATION OF BOTANICAL KNOWLEDGE
In botany species, subspecies, varieties and forms are concepts for organising and presenting knowledge. These concepts cannot be demonstrated empirically. In recent decades attention has been paid to matters such as chemotaxonomy and DNA sequencing which are empirical scientific techniques. Scientific techniques have been used randomly alongside concepts with an absence of necessary theorisation not only at an appropriate philosophical level but also at the level of practical application. At best the techniques of empirical science can provide some level of authenticity for the concepts currently in use.

The study of Haworthia calls into question the usefulness of the concepts themselves. Species in this genus rarely fit neatly into conventional boxes of classification. Students of Haworthia accordingly vary between those who classify virtually each different population of one species as a separate species to those who reduce the same plants to one species with the variations described under species rank. Neither of these situations are particularly satisfactory as both are limited by the prescriptions and limitations of existing concepts. The reductive approach is certainly to be preferred should one’s interest be in trying to understand the development of these plants in space and over time.

The concept of variety with respect to Haworthia koelmaniorum must essentially be based on differences between “koelmaniorum” and “mcmurtryi”. During some 20 years of detailed field research it was found that none of the characters used to distinguish the two varieties of Haworthia koelmaniorum were accurate. This matter is discussed in full detail in the paper currently in preparation. Only one factor that could usually be consistently used to separate the varieties was found. This was the character that H. koelmaniorum var. koelmaniorum often grows more than one flower stem in a given flowering season usually 2 – 3. These are frequently produced over a period of a few weeks. H. koelmaniorum var. mcmurtryi nearly always produces a single flowering stem. These characters however hardly justify any distinction at varietal or any other level.

H. koelmaniorum is best regarded as one variable species, the variation related to the different kinds of niches it frequents in a predominantly bushveld biome and a predominantly grassveld biome. This approach however begs the question as to what we do with the classificatory labels. The logical approach seems to be to abandon them entirely and to simply use the names of places near the colonies to describe the variation. On the other hand if the autecology of the species across its distribution range is adequately understood it is unlikely that detailed descriptions of the variations are either desirable or necessary.

The discussions that follow have been elicited from about 2 decades of field observations. Variations in the plants are entirely related to the kinds of rocky niches they can colonise and grass fire cycles particularly the load of flammable material in different habitats that gets burnt in dry season veld fires.

THE HABITATS OCCUPIED BY HAWORTHIA KOELMANIORUM AND THE ROLE PLAYED BY GRASS FIRES.
H. koelmaniorum var. koelmaniorum occurs in arid savannah at an altitude of about 900 – 1200 metres. The plants are always associated with fixed rocks, usually quartzites tightly wedged into the ground. There is invariably mixed mainly deciduous woodland present at the localities frequented by variety koelmaniorum. The presence of deciduous trees and the leaves that fall from them in the autumn have a notable and hitherto unrecorded role to play in preserving many of the plants from the worst effects of fire. This is discussed below.

The cavities in quartzites tend to be deeper and sometimes also broader than equivalent places in the variety mcmurtryi habitat. In addition the variety mcmurtryi habitat carries much heavier loads of dry grass during the fire season, mostly from May – September. The rocks in the variety mcmurtryi habitat are rhyolite. This rock type forms fissures and cavities that are quite often only slightly elevated above the ground.

The retuse habit of variety mcmurtryi ensures that most of the Haworthias do not get their leaves badly burnt in the grass fires that are such a regular occurrence in their habitat. Dry grass is highly flammable and burns right down to ground level in fires, unless these are driven by high winds. In such instances the burns are more superficial. Occasional variety mcmurtryi plants, with longer leaves, are found in places with deeper cavities amongst large rhyolite rocks.

The smallest variety mcmurtryi plants yet discovered were found on 28 August 2008. They were growing in an area cobbled with rhyolite pebbles and small blocks of these rocks deeply fixed in the ground and protruding just above it. The full rosettes of highly retuse leaves were mostly 2 x 2 cms in size, some even less. These Haworthias appear to represent an extreme adaptation to regular fires. Their small size and very retuse habit render them almost immune to the effects of grass fires. This cobbled habitat is fairly common where variety mcmurtryi occurs but rarely as extensive as at the locality recently found.

The variety koelmaniorum habitat often has thinner loads of dry grass during the winter months and many dry leaves shed by the surrounding trees and shrubs. The deeper cavities between rocks which the Haworthias occupy and scatterings of dry leaves in them, which often only burn partially in fires, permits the establishment of plants with long leaves. In most variety koelmaniorum colonies typical long leaved plants as well as some shorter leaved examples are found. The latter are often indistinguishable in appearance from typical variety mcmurtryi. This data implies that it is neither an advantage nor a disadvantage to have long or retuse leaves in a rocky environment with grass and trees. The considerable variation in variety koelmaniorum colonies, with respect to the size of the plants and length of their leaves, seems to be directly related to the kinds of rocky niches found in quartzite and the composition of the flammable load which consists of grass and leaves.

Haworthia koelmaniorum variety koelmaniorum habitat close to a rural village at the western extremity of the distribution range. This photograph was taken on 29 April 2008 at a time when the habitat was drying out in late autumn after the summer rains.

FIRES IN THE MODERN HAWORTHIA KOELMANIORUM ENVIRONMENT
H. koelmaniorum exists and thrives in a fire driven ecological system. Prior to the advent of widespread human settlement in the habitats fire would have swept across vast areas, missing out some places by virtue of natural barriers and changes in directions of the wind during specific fires. The current situation is very different with a number of significant implications for the future health of the plants. The land use patterns in the habitats are responsible both for the incidence and control of fires. Game farms and the Loskop Dam Nature Reserve are primarily concerned with the conservation of rare, valuable and often introduced fauna, which may in former times have occurred in the area. The tendency with this type of land use is for fewer artificially managed fires that burn much heavier loads of flammable material. This often results in losses of Haworthias and preliminary indications are that it shrinks the sizes of populations confining them largely to areas protected by the biggest rocks.

Stock farms, which usually have herds of cattle, are fewer these days than formerly. Fires are also irregular here and controlled by fire breaks as they are on game farms and in the Loskop Dam Nature Reserve. There is, as yet, not enough data to draw long term conclusions about the effects of limited fires on game farms, in nature reserves and on cattle ranches. There is however a great deal of data on the effects of fires in the western part of the variety koelmaniorum distribution range and also the variety mcmurtryi in the general vicinity of the type locality. Regular fires have little effect on variety koelmaniorum and improve the conditions for the germination of seeds in summers directly after a winter grass fire.

Variety koelmaniorum has the advantage of usually producing a few flower stems which emerge and flower over several weeks. This habit and the relatively superficial burns ensure that the plants are capable of flowering directly after a grass fire. The same does not hold true for variety mcmurtryi. Fires in the second half of winter and at flowering time usually destroy most of the buds and inflorescences. The burning of heavy loads of tinder dry grass frequently damages the plants, usually destroying the tissue at the end of the leaves. Only the very large robust plants seem capable of flowering under these conditions.

No single fire is like any other. Fires may burn thoroughly for hours in still conditions or be driven by high winds, particularly in August which is a very windy time of the year in both habitats. The variation in fires that occur in the same place in successive seasons ensures the long term health of H. koelmaniorum habitats.

HABITAT DEGRADATION
H. koelmaniorum habitats are degraded via the exclusion of fires, prevalence of certain kinds of livestock and expansion of rural settlements. These factors are discussed in some detail in the paper by Charles Craib (2001: 22 – 27) with respect to variety mcmurtryi.

H. koelmaniorum colonies are found in suitable rocky habitats in rural villages, on the outskirts of these settlements and in more remote areas away from human habitation. Factors that cause colonies to decline are the loosening of fixed rocks by livestock, almost entirely donkeys, pollution of habitats with broken glass in and near villages and felling of trees and shrubs for kindling and firewood by the local people.

It is surprising how healthy some of the H. koelmaniorum habitats are close to large rural settlements. There is little doubt that in the case of variety koelmaniorum the plants benefit from regular fires ignited by the local people for promoting spring grazing at the beginning of the rainy season.

The depletion of shrubs and trees in some of the variety koelmaniorum habitat will change the composition of the flammable load burnt in grass fires. The lack of protective dry leaves that are superficially burnt in grass fires are likely to result in hotter burns particularly as annual weeds grow amongst grass in disturbed places increasing the likelihood of more thorough fires.

The two varieties of H. koelmaniorum have not yet been recorded amongst the medicinal plants gathered in habitat by local people. Rural villagers are well aware of the plants in some areas but, at present, they are only collected very occasionally by children as “play objects”. It seems that plants are also only gathered from within the confines of villages where they are most readily available.

It may ironically be that the future of H. koelmaniorum is least secure in protected places such as in nature reserves and on game farms owing to the artificially managed burning regimes in these places. This is certainly an area that merits attention from researchers with the necessary background to and understanding of the autecology of these plants.

A very large robust form of Haworthia koelmaniorum variety koelmaniorum in cultivation. The plants are grown close together to enhance seed production when they flower. They are cultivated out in the open in Johannesburg and exposed to afternoon sunlight. The leaves may be damaged from time to time in typical highveld hail storms.