Haworthia as a problem genus (1970)


M.B Bayer, National Botanic Gardens of South Africa, Karoo Garden, Worcester.

Although many enthusiasts have recognized the need for a better understanding of haworthias as they grow in the field, South Africa has been very slow in providing the necessary interest.  Mr. G.G. Smith did a vast amount of work only to withdraw from the field before this work could reach fruition.  As a result, even his contribution has done very little towards solving the fundamental difficulties in the group.   Practically all the errors which can contribute to unsatisfactory classification have been made.  Perhaps the most unfortunate, and the most forgiveable, has been the total failure to recognize the importance of locality and variability.  In other words haworthias have never been studied on a population basis.  This has led to unnatural sections in which only superficial morphological characters have been used, and too many superfluous species and varieties which have no basis as far as the distribution of the plants is concerned.   Floral structure has several times been suggested as a possible solution to the problem of identification, but has never led to any further conclusion.  Why, of course, is very obvious.  To try and establish order from false foundations will never succeed.  The basic premise that the species described are true representatives of morphologically and geographically distinct species has never been questioned for the haworthias.  New species and varieties were freely described by Smith, von Poellnitz and Uitewaal (and Resendé) although the real identities of Haworth’, Salm Dyck’ and Baker’s species were in many cases unknown.

We thus have species such as H. retusa, H. reticulata, H. mirabilis and H. herbacea to name but a few, which have not been accurately recorded since they were originally described.  We have varieties in the same category, arid also varieties bearing no relationship in distribution to that of their putative parent species.  One only requires an elementary knowledge of South African geography to realize just how incongruous many of von Poellnitz’ localties are.  Perhaps, as Resendé pointed out, better collecting and recording methods here at home would have obviated many errors.

Description has also left much to be desired, and while it would have been easy to have recognized plants in Haworth’ or Salm Dyck’s possession, it is now almost impossible to establish which population these plants could have been drawn from.  Recent descriptive work has been equally misleading because the new species and varieties have seldom been discussed adequately in terms of their distribution and variability.

The seriousness of the situation only recently hit home here at the Karoo Garden at Worcester.  Seed was collected of a common Haworthia species, and an attempt made to identify the plant so that seed could be distributed to members of the South African Botanical Society.  After several attempts involving several species, the name H. helmiae was first adopted.  At least von Poellnitz had cited Brandwacht (3 miles west of the Karoo Garden) and Worcester itself in addition to Great Brak River, as localities for this species. Further consideration and better fitting description led to the adoption of the name H. aegrota.  This done, it was finally decided to investigate the entire district and see just what could be found.  At the same time Mr. G. G. Smith’s notes and records were obtained.  Also assisting was Mr. F. J. Stayner whose experience of haworthias dates back to his association with Mr. F. R. Long, a well known enthusiast after whom H. longiana was named.  The results of this investigation form the basis of the introductory remarks.

Collecting was done throughout the Robertson Karoo – an area extending from Worcester to Ashton, southward to Drew and McGregor and back to Villiersdorp in the southwest.  An area roughly 50 by 20 miles.  75 collections were made at various localities in the area.  Fortunately it was possible to call on the help of Mr. C. L. Payne who was the source of much of von Poellnitz’ material (via Triebner).  He was able to correct many locality citations, especially of the Muticae, and point out other localities such as the type locality of H. pallida var paynei and H. triebneriana var diversicolor.  Mr. Smith’s field books also yielded much useful information, including locality details and photographs of plants collected and named. The collections broke down into the following:-

H. aegrota (herbacea) – 22 collections.

            H. reticulata -13

H. notabilis (maraisii var notabilis) – 3

            H. schuldtiana (maraisii) – 29

            H. triebneriana var diversicolor (mirabilis)1

            H. margaretifera (pumila)16

H. poellnitziana (minima) – 1

H. marginata – 1

H. setata (var gigas ?) (arachnoidea)- 3

hybrid swarms -3

Analysis of these collections resulted in some interesting conclusions:

Haworthia aegrota:  This species proved to be so common that it was concluded it must have been one of Haworth’s species, and also so variable that it contained the elements of other species described more recently.  The variability was complicated by the discovery of two hybrid swarms, one involving H. aegrota and H. reticulata, and the other involving these two species and also H. notabilis.  Reference to the literature led to the conclusion that the plant complex sampled had provided the plant first illustrated by Boerhaave in 1720, and first named by Miller in 1768, as Aloe herbacea.   The same species that Haworth called H. pumila before adopting the name H. atrovirens.  As both light and dark green forms are found, and because of the appearance of plants in the hybrid swarms, it was also concluded that H. pallida and perhaps H. translucens had their origin in the same complex.  Certainly H. pallida var. paynei is nothing but a small form of the one population.  It was quite easy to establish that the four species H. aegrota,

H. submaculata (of von Poellnitz), H. luteorosea and H. guttata (of Uitewaal) also must have come from this same area.  We have therefore no choice but to regard all these species as synonyms of the single species H. herbacea (Mill.) Stern typified by Boerhaave’s illustration.  No basis for varieties was found at all except for the variety paynei—but even this is negated by the occurrence of one particularly aberrant population.

Haworthia reticula: This also proved to be a very common species and like the preceding species must have been found during the early exploration of the Cape Colony.  The first two collections were obtained from the eastern and western limits of the population and named at the Karoo Garden as H. hurlingii and H. haageana respectively.  Subsequent collections in the area between showed that a single species complex was involved and that this was Haworth’s missing H. reticulata.  ‘Missing’ because the locality has not been precisely recorded prior to this.  Mr. Payne’s observations on the locality of H. intermedia, H. hurlingii var. ambigua, H. haageana var. subreticulata and H. reticulata var. acuminata, plus investigation at these sites, show beyond doubt that these are all forms in a single species.


(Fig. 1. (Above) Haworthia herbacea (Miller) Steam, two common forms. Fig. 2. (Below) H. herbacea from a somewhat anomalous population in which the plants are smaller and densely proliferous.)

Haworthia notabilis:  Although three collections are recorded, they were all made in the same general locality.  The first collection produced a varied assortment of plants taken to be H. aegrota, while the third collection followed the discovery of a hybrid swarm on a slope a short distance away.  While it is possible that H. notabilis is a direct product of this hybrid swarm involving H. aegrota and H. reticulata, it does seem to contain another species element, possibly a form of H. schuldtiana.  Further investigation is needed to clarify whether H. notabilis is a good species or not.

Haworthia schuldtiana:  The very first collection of this species was so variable as to discourage any attempt at attaching varietal names to later collections.  The species is widespread, but difficult to find on account of its coloration and choice of locality.  Also it does not proliferate as do many other haworthias.  Despite the number of collections, it resists most attempts to resolve into distinct species components.  There is some indication that we are dealing with from 4 to 6 such components together with intermediates.  It is also suspected that the species H. parksiana, H. floribunda, H. magnifica, H. maraisii and H. sublimpidula have a direct affinity here.

Haworthia triebneriana var. diversicolor:  This variety appears to be an intermediate between H. schuldtiana var. whitesloaniana and H. triebneriana, which is a large species complex on its own, south of the Robertson Karoo.


(Fig. 3. (Above) Haworthia reticulata (Haw.) Haworth, a ubiquitous form. Fig. 4. (Below) H. reticulata, glabrous forms; the right hand plant is very pale and chiorotic. In the hot, dry summer the plants develop very beautiful shades of red—in contrast to the bright green. of the turgid winter plants.)

Haworthia margaretifera:  16 collections were made of this species although it is the most common and adaptable of all the species encountered here.  It was found growing at both high and low altitude, in rocky situations, on sandy plains, in dense scrub and on eroded bare screes.  Variability is not excessive but it is clear that this species has suffered most from now outdated concepts.

Haworthia poellnitziana: The type locality at Drew is restricted to an area of about 100 sq. yards, and from what can be gleaned from the literature and from what is still apparent in the field, this is seemingly a product of hybridization between H. margaretifera and H. marginata.  The third hybrid swarm found was between these two species at Ashton and some of the products were identical to H. poellnitziana. However, there is still a suggestion that it is a western form of a variety of H. margaretifera (or species of the Margaretiferae) still to be investigated at Swellendam and westward.  It is worth recording Mr. Payne’s observation that H. marginata var. virescens had come from the same locality at Drew.

Haworthia marginata: A very good and beautiful species threatened by extinction.  This species is closely related to H. margaretifera and with a natural composition of the sections may eventually prove to find its position with this species.  Fortunately its distribution extends eastward to Heidelberg and it may perhaps still be found in the south west near Bredasdorp, where it has been reported in the past.  Hybridization with H. margaretifera has resulted in the naming of many varieties which cannot be upheld.

Haworthia setata (var. gigas?):  The identity of this plant is doubtful.  We already find so much difficulty in attaching the names H. setata and H. arachnoidea to natural populations, that the naming of varieties is not readily attempted.  Some of the specimens found were enormous (up to 6 inches in diameter), while others were barely 1 inch across.  At one locality H. aegrota intrudes and intermediate forms apparently occur—a phenomenon still being considered.

Much work has still to be done before the Robertson Karoo complex is finally understood. This area also does not form a closed environment and has many ecological associations elsewhere.  This investigation has shown that classification will not be really possible on the basis of a potted collection.   Despite the doubts and difficulties encountered it is felt that the genus is not as involved as its written taxonomy may suggest.  The dissolution of many well-known species under the one name H. herbacea may give rise to concern, as may also the suggestion that H. schuldtiana and related species be invalidated. There is consolation in the fact that such changes will necessitate definition of as yet unnamed or incorrectly recorded species.

Clonal and vegetative propagation will always exacerbate nomenclatural problems in the Haworthia, as it tends to camouflage inherent variability.  Two plants from one population may be seen to be conspecific at collection,  but ten offspring from each plant, grown under different conditions in different parts of the world may lead to quite another conclusion.  The presence of natural hybrid swarms should also be noted, because unless a plant is considered in proper perspective in a population, which may only be a small part of a still larger entity, justice will not be done.  There are many examples which can be quoted to confirm that this is what the genus now suffers from.  There is merit in naming individual plants if they are already widely represented in collections or have sufficient distinction to be worthy of note.  The term variety should only be used if we recognize that we are speaking of forms, and if the term ‘sub-species’ is utilized as the botanic concept of a population based variant.

It is hoped that the studies commenced at the Karoo Garden will eventually be extended to other areas. The detailed and intensive coverage of area required before a sample can be regarded as reliably representative may preclude this.  It will be a tragedy, as time and advancement of agriculture with concomitant exploitation of the environment are steadily eroding what remains of our floral heritage. Haworthias, like many other species do not regenerate and establish themselves once disturbed.