New Haworthia (1974)

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

Published in CACTUS & SUCCULENT JOURNAL (U.S.), Vol. XLVI, 1974

Fig. 1. Haworthia geraldii Scott.

‘Out of Africa always something new’ is an expression the reader may excusably believe no longer applicable in the case of Haworthia. However, in the last 8 years no fewer than seven new Hawrthia species have been described. Three of these are by Col. C. L. Scott, one by Dave Hardy and Mr. Obermeyer and three by the writer. Many species have been deleted in the past two years and many more are destined to a similar fate—but it is also likely that new species will still emerge.

The real problem in Haworthia has been and still is the question of deciding just what constitutes a species, as this concept is rather difficult to define even for the most general application. A species is obviously not one single plant, but an aggregate of individuals which look alike and comprising a population or series of populations adapted to, and propagating in particular habitats. A species should not be described without duly considering this. In Haworthia one cannot generalize too much about what a species is either, as in the three subgenera (and add Astroloba and Poellnitzia if some authorities exercise their will) one finds apparently quite different systems in operation. Peter Brandham of Kew has recently demonstrated that entire and even widely separated populations of H. reinwardtii may comprise or contain genetically identical plants. This shows this species does propagate vegetatively and a species concept here may be quite unfitted for other species such as H. lockwoodii which do not propagate vegetatively in nature. In many cases the writer has found odd populations which are each quite distinctive and yet which must be outliers of more widely distributed and commoner species. If these are all to have names, taxonomy in the group may become quite unmanageable. It thus seems far wiser to adopt a rather broad species concept and rather include such populations under their apparent points of origin—if it is possible to do this.

H. geraldii Scott was described from the Riversdale area and is found just north and northeast of the town. The writer has only visited the northern locality where the plants do present a remarkable sight. Huge solid clusters of up to 1/2 meter or more in diameter containing 60 or more large rosettes slightly raised from the ground—beautifully colored greenish-yellow, with prominent face lines on the heavy thick leaves. About 200 meters to the west is a smaller population of another plant which appears to have some affinity to H. asperula (synonymous with H. schuldtiana?). This latter plant seems to breed true from field seed, which may seem rather odd when one considers how close it is to H. retusa and H. geraldii. A few kilometers east is the site of H. fouchei, and still further east in the Kafferkuilsriver one finds forms which lead on to H. longebracteata from the lower reaches of this river. The story does not end there, because westward from Stilbaai andsouth of Riversdale one finds populations still very near to H. longebracteata, until at the Duiwenhoks river at Heidelberg one comes to the source of H. nitidula. North of Heidelberg one finds a most classic example of ecotypic transformation in Haworthia. In the sandstone formations are typical H. laetivirens, in theshale a little to the south, the plants are larger and unspined until in the conglomerate at Heidelberg there is a motley complex of plants tending to H. nitidula (and hybrids with H. floribunda!). It is certain that all these populations must be regarded as components ofone species and that names such as H. fouchei, H. longebracteata, H. nitidula and also H.geraldii cannot rationally be maintained. Species such as H. turgida, H. dekenahi and H. caespitosa may also yet be excised. However, this is an extraordinarily difficult complex and good facts are really needed on which to base sound argument.

Fig. 2. Haworthia baylissii Scott.

H. springbokvlaktensis Scott is unquestionably a good species. Known for many years and recognized as H. mutica by collectors such as Major F. R. Long, W. E. Armstrong and other enthusiasts of the pre-war period, Col. Scott has correctly separated this species at Springbokvlakte. Growing together with Lithops terricolor in the typical sandy Karoid area east of Steytlerville, it is a most striking plant. The very rounded leaf tips and thumblike leaves, the purplish, translucent face area and darker face-lines make this a most attractive plant. It does not seem to be very plentiful and is seldom proliferous. More than one locality is known and evidence suggests that it is to some degree intergraded with the H. emelyae complex around Oudtshoorn. At Uniondale one finds plants which also indicate affinity with H. comptonii which in any case may best be regarded as a near ally of H. emelyae. H. mutica is now well-known to occur in the Bredasdorp/Caledon area (H. otzenii is synonymous here)—inclined to a purplish hue with unpointed leaves, it does extend eastward toward the notorious Riversdale area. It is very difficult to convincingly separate H. mutica from H. retusa and even from H. pygmaea at Mossel Bay.

H. baylissii Scott is from Oudekraal in the Somerset East district. This species has not been seen by the writer in the field and with two very disparate collections from that locality as well as deviations from the description, doubts as to its validity must exist. A collection at Kirstenbosch apparently received from Col. Bayliss (the collector of the species), is clearly H. angustifolia but with shorter, thicker leaves than normally encountered in that species. The locality is a little far west for H. angustifolia (not to be confused with H. monticola from the southern Karoo!) but the plants do also have the silver flecks as in co-types of H. baylissii, also in cultivation at Kirstenbosch. The main distinction of H. baylissii was said to be the recurving leaves, but this is not a stable character and none of the plants in cultivation either at Kirstenbosch or the Karoo Garden have this character. Why this species should have been placed in the section Muticae really serves only to underscore how abysmally weak the sections actually are. This section was originally created by Berger for H. reticulata, and eventually contained ‘H. hurlingii, ‘H. haageana’, ‘ H . intermedia* (all synonymous now) as well as a few other species including H. caespitosa. H. baylissii has quite patently no close affinity, difficult and obscure as Haworthia taxonomy may seem to be, with any of these species and until better understood should be considered close to H. angustifolia, but not necessarily to all species in the section Loratae.

Fig. 3. Haworthia springbokvlakensis Scott.
Fig. 4. Haworthia koelmanniorum Obermeyer & Hardy

H. koelmanniorum Obermeyer and Hardy was described from Groblersdal in the eastern Transvaal and this is far to the northwest of the nearest species, H. limifolia. The differences by which it was first separated from that species are quite invalid. As far as suckering is concerned, the different forms of H. limifolia do not all sucker freely, if at all; H. limifolia does not have a significantly different base to the rosettes, and if it did, would probably require separation at generic level! The tubercles on the leaves of H. limifolia vary enormously over the distribution range—from colorless, wavy, undulated ridges, to prominent white bars—from small, concolorous tubercles to large, raised pustules. The difference in phyllotaxis is also fallacious, as any admirer of the arrangement of spines on cacti will be quick to grasp. A l l Haworthia have their leaves spirally arranged and H. limifolia together with H. tesselata are rather exceptional in that the angle of leaf divergence is rather high, in the vicinity of 160°. This means that the two primary counter spirals are more readily apparent than the three secondary spirals. When grown together with many other collections of H. limifolia, the writer can see no exceptional reason for separating H. koelmanniorum from that species, other than the color being a deeper, translucent gray-green and the leaves bluntly tapering. The flower in H. limifolia is rather variable and there is no evidence here to suggest the H. koelmanniorum is distinct. Floral characters in the whole subgenus Hexangulares are, however, rather much of a muchness and not very helpful. H. tuberculata, together with H. starkiana, does have unique floral characters and to suggest the H. koelmanniorum has an affinity here is to do justice to neither species and offend the small differences which are of taxonomic value in the genus. Geographically speaking, the species is of note but it is not correct to talk of ‘outliers (of the genus) extending to Zululand, Swaziland and southeastern Transvaal*. It must be obvious from the above that probably only one species, H. limifolia, is involved, but perhaps sufficiently differentiated for local populations to be recognized at subspecies level.

H. pubescens Bayer is, with the next two species, a ‘creation’ of the present writer—and perhaps enough said! There is no doubt that this species has an affinity with the very difficult and variable H. schuldtiana complex. However, it is equally allied to H. herbacea.

Important distinctions include both vegetative and floral characters, flowering times, locality and ecological associations—and the author is not concerned with single plant characters or collections. H. pubescens comes from a very dry area south of Worcester and is hard to grow, requiring strictly winter watering and no direct sunshine. A well-grown plant is a pleasure to behold—the whitish pubescens on the dark green background give the incurved, abruptly pointed leaves a velvety appearance.

Fig. 5. Haworthia pubescens Bayer.

H. pulchella Bayer, nom. nov. in manuscript, can also be put in the gem category. The small, sparkling spined, emerald ball is the real jewel of the Karoo garden collection. First known to the writer through its discovery by the late Mr. Roy Littlewood and the recollected by Mr. F. J. Stayner, this species is from the very dry, desolate hills of the Touws River area. Its affinities are doubtful and the flower is as ordinary as that of an haworthia can be. Associated species do not give any leads either and the nearest species geographically speaking are H. setata and H. tessellata. Much further away are H. wittebergensis, H. lockwoodii and H. margaritifera. If the writer was forced to select a near ally, it would be H. monticola on the basis of floral similarity.

Fig. 6. Haworthia pulchella Bayer, Touws River.


H. serrata Bayer is no less contentious than H. pubescens. Occurring at several localities within a relatively small area, this species is one of the affiliates of the Heidelberg complex. In some forms the leaves are incurved, suggesting a robust plainly marked specimen of H. herbacea, while in others the leaves recurve as in H. heidelbergensis. The flower is characteristically rather bright green in the veins and does not have the closely plicate tips of the upper, outer tepals as does its expected relatives. That many leaves, among other considerations, remove it from H. retusa, and it must have close ties with H. rossouwii at Bredasdorp. H. serrata demands recognition, but how best to express its distinctiveness in among several almost unique populations in that area still eludes.

In concluding it may not be necessary to say that taxonomy in Haworthia has passed beyond cursory examination of the simple, external structure of the plants. What is really needed is some good morphological study aimed at analyzing characters which might be of better value in separating and circumscribing the species. Nevertheless, some kind of structure is necessary even to base such a study on, and any contribution must be gratefully welcomed.

Fig. 7. Haworthia serrata Bayer.

Haworthia Revisited – 28. Haworthia pubescens

28. Haworthia pubescens Bayer, JS.Afr.Bot. 38:129(1973).  Bayer :147(1976).  Bayer :50(1982).  Type: CAPE‑3319 (Worcester): Sandberg Hills (‑CB), Bayer 163 (NBG).

pubescens: with many short hairs.

Rosette stemless, seldom proliferous, to 4cm φ.  Leaves 20-35, short incurved, opaque grey-green, covered with minute spines.  Inflorescence simple, to 20cm.  Flowers 10-15, with upper lobes flared, white with pinkish venation.

1982 – H. pubescens has the same growth form as H. herbacea but it is smaller and seldom exceeds 30mm in diameter.  The incurved leaves are dark grey‑green and finely pubescent.  The flowers are quite unlike those of H. magnifica to which it may have been related, as the buds are long and slender.  Also the upper perianth lobes are widespread as in H. herbacea.  It grows in very close association with H. herbacea and flowers in November/December ‑ after H. herbacea and before H. magnifica.  The distribution is extremely limited and only occurs on two low quartzitic hills east of the locality for H. maculata.  It appears to be represented again about 15 km south by a form which is less pubescent and with more turgid leaves.  This latter form at Lemoenpoort is in a quartzitic ridge again near H. herbacea, and it resembles H. maculata in the relative proportions of its leaves.  H. maculata occurs in a more recognisable form about 5 km to the west at Moddergat, at a far southern locality for the species.

1999 – The geographic scale of species and their distribution ranges changes from the summer rainfall areas of the country to the southwestern winter rainfall area. H. pubescens probably has to be seen in that context. There seems to be little doubt that it is filling the space of H. maraisii, but that could also be mooted for H. maculata. The former species is the most probable relative as there are growth forms with the same leaf texture, and rosette shape near to Robertson. However, H. maraisii occurs in its typical form very close to Lemoenpoort at Trappieskraalkloof, just to the east.  G.J. Payne did inform me that he had collected a small dark species just above the Brandvlei Prison and this would probably fall into this context if it is re-collected.

a. var. pubescens.

Distribution: 3319 (Worcester): Sandberg Hills (‑CB), Bayer 163 (NBG)

b. var. livida var.nov. 
Type: CAPE-3319(Worcester): S. Lemoenpoort (-CD), Bayer 1128 (NBG, Holo.).

livida: bluish-grey.

Differs from the typical variety in being less pubescent, with slightly broader and fewer leaves, and partly with pellucid spots on the leaves.  (A var. pubescens foliis latioribus minus pubescentibus cum maculis pellucidis differt).

Distribution: 3319 (Worcester): S. Lemoenpoort (-CD), Bayer 1128 (NBG).

[ed.] Bayer now considers var. livida as H. maculata var. livida

H. maculata var. livida (M. B. Bayer) M. B. Bayer (Haworthia Nomenclator, 10, 2012). Type: RSA, Western Cape (Bayer 1128 [NBG]). — Distr: RSA (Western Cape: Robertson Karoo: S of Worcester).
Haworthia pubescens var. livida M. B. Bayer (1999) ≡ Haworthia maraisii var. livida (M. B. Bayer) M. Hayashi (2000) ≡ Haworthia intermedia var. livida (M. B. Bayer) J. Esterhuizen (2003); incl. Haworthia livida Breuer (2011) (nom. inval., ICN Art. 38.1a, 41.5).
Differs from var. maculata: Ros smaller; L less spotted.

A myth corrected to – Haworthia maculata var. livida (Bayer) Bayer – and flowers ignored

(Haworthia maculata var. livida (Bayer) Bayer, comb.nov.  H. pubescens var. livida Bayer in Haworthia Revisited, p.134, 1999, Umdaus)  Type: Cape-3319 (Worcester): S Lemoenpoort (-CD), Bayer 1128 (NBG, Holo.).

I described Haworthia pubescens var. livida in Haworthia Revisited (Umdaus, 1999), in the full knowledge that it was in a twilight zone of inadequate information. It is a good example of how Latin names give plants a false reality. The system forces decision making without any slack being cut for doubt. This is thus a good opportunity to demonstrate what inexperience and ignorance add to the process of classification. In the small area along the Breede River north of the Brandvlei Dam near Worcester, the species H. herbacea, H. maculata and H. pubescens grow in close proximity. H. herbacea is ubiquitous throughout the Worcester/Robertson Karoo, while H. maculata has a curious distribution in that area.  It occurs at widely separated localities on the western fringe of H. herbacea and I have wondered about its relationship to that species because of the similar flowers and flowering time. H. pubescens is only known from a small set of low ridges east of the Brandvlei Dam where it grows in close proximity to H. herbacea.  It also has similar flowers but it flowers a little later in late spring as opposed to early spring.

When I first discovered plants at Lemoenpoort (Hammansberg) about 20km further south I found very few plants and the only thing I could do with them was regard them as in the same context of the Breede River species and hence their description as a variant of that H. pubescens. H. herbacea is very abundant in this area too but almost exclusively in Bokkeveld Shales and Dwyka Tillite. The doubtful plants at Lemoenpoort occur in Witteberg Sandstones. A population of plants was also known from the top of the Moddergat Mountain (Ouhoekberg) to the west that created some problems for me as it seemed to relate to the absence of H. mirabilis from both the Hammansberg and Ouhoekberg. Not only that but it also looked very similar to H. maculataH. mirabilis is known at its closest east of these two mountains at several localities along the eastern ridges of the Ouhangsberg and Droogerivierberg mountains that are in turn east of Hammansberg. H. maculata is very abundant along the long ridge of low hills north of the Brandvlei Dam. Very similar plants occur high in the Audensberg peaks north of Worcester and also lower down in the Kanetvlei (Sandhills) area west of Heatlie Peak. These hills are very difficult to explore and I speculate a connection to a population at Buitenstekloof to the east. There is a population here that obfuscates the interpretation of the name H. intermedia VPoelln. I lumped this population under H. mirabilis var. notabilis. I have to rely on memory here because having set flower morphology and flowering time aside (without completely ignoring them), I did not have the methodology in place to record detail. The flower was rather similar to H. herbacea and H. maculata and it also flowered late spring. I was not, and am not, sure that there is not a connection between the Buitenstekloof plants and the Kanetvlei plants.

It is unfortunate that Gerhard Marx has expressed his opinion that I have ignored flowers in my classification of Haworthia. While there is no doubt that this is partially true, there are very good reasons why I did this. The main reason is the historic one. It was understood and accepted that the flowers in the species as they were then recognized, were very similar. Even the subgeneric differences were not seen to be significant. My own early attempts at classification also indicated that the flowers were not going to be of assistance at the level at which differences between populations and species were being recognized. In fact I had observed similarities of the flower in populations that were clearly different taxonomic entities. Conversely I observed big floral differences in populations that I imputed to be the same “species”. Because of the added seasonality of flowers, I thus put them aside to focus on exploration and explanation on the basis of the vegetative structures and geographic distribution. Marx’ observation therefore comes at a time when that work is substantial and flowers can now be used to clarify and verify relationships. My conviction is that those relationships are best understood in relation to distribution and we are only now in a position to assess what value the flowers may add to those interpretations.

This article thus reports on the exploration of the Ouhangsberg, Hammansberg and Moddergat (Ouhoekberg) mountains as well as some work done in the Brandvlei Dam area. Again I put flower and flowering time aside – for the moment. This is in the same way I am inclined to the considered view that geographic information and field observation are the essential ingredients whereby ANY method has to provide explanation that satisfies physical experience. The alternative is that what we see and experience physically and mentally, is just the illusion that metaphysicists ascribe to creation.

Populations recorded

The following populations (see map) were found and observed:-

1. KG669/69 N Brandvlei Dam – set 1
2. MBB164 N Brandvlei Dam – set 1
3. MBB1119 (and 6815) Audensberg – set 2
4. MBB1120 (and 7994) Sandhills, Kanetvlei – set 2
5. MBB1128 (and 7066) W Lemoenpoort (type locality var. livida) – set 3
6. MBB1145 Ouhoekberg, Moddergat – set 5
7. MBB2591 NE Brandvlei Dam – set 1
8. MBB4461 (and 6514) Buitenstekloof – set 6
9. MBB7270 Ouhoekberg W – set 5
10. MBB7266 E Lemoenpoort – set 4
11. MBB7271 Cilmor Winery, Dewetsberg – set 7
12. MBB7526 Die Nekkies W. Brandvlei Dam – set 1
13. MBB7991a Ouhoekberg E – set 5
14. MBB7991b west Ouhoekberg E – set 5
15. MBB7992 Hammansberg midpoint – set 8
16. MBB7993 Hammansberg W – set 8
17. MBB7997 Sandberg H. pubescens­, type locality – set 9 are for MBB7997 cf H. pubescens N of point 11
18. MBB7995 and 7996 Brickfield H. herbacea ‘submaculata’ – set 10
19. MBB7984, 7985 and 7988 Droogerivierberg, H. mirabilis – set 11
20. Unknowns – set 12

There are 4 records for the Brandvlei Dam but in Update 6, I report on H. maculata from the pump station east of the Resort, very abundant to the eastern point of Die Nekkies.  Three of the records are for west of the Resort while I have not confirmed Etwin Aslander’s observation that he had seen it at the extreme western end of Die Nekkies. G.J. Payne in 1970 communicated to me that he had seen a Haworthia south of the Brandvlei Dam above and west of the Brandvlei Hot Spring. I was unable to confirm that in a single brief visit there ca. 1976. I have not been able to relocate plants at Sandhills but have no doubt that it will be found there if more thorough and extensive fieldwork is done there. In fact a very important consideration is that there is still enormous scope for exploration.

On each visit to the Ouhoekberg I have found plants in different local populations and none of the last four records is the same as seen in my first visit. At the original Lemoenpoort I always struggled to find plants at all, and the same applies to the locality to the southeast. At my last visits to these two places there have been many plants but relatively difficult to find.

H. herbacea is abundant in the Hammansberg and Ouhangsberg area as well as east of Cilmore.  It has not been observed near the Ouhoekberg. It occurs between Cilmor and Die Nekkies in a larger form that acquired the name H. submaculata V.Poelln.; and I did suspect that there was a continuity with H. maculata that does not now seem to be the case. I exclude H. pubescens from this report. It occurs with H. herbacea east of Cilmor, but it also occurs on the northeastern slope of Dewetsberg north of the H. maculata recorded habitat. There are/were plants of H. mirabilis vegetatively very similar to H. pubescens at the Breede River Bridge southwest of Roberson. Their flowers and flowering times (March) correspond with H. mirabilis in that area. H. pubescens flowers in November (later than H. maculata at Die Nekkies or of H. herbacea – Sept/Oct) and the flower is similar to that of H. herbacea. It seems fairly obvious that H. pubescens fills the gap for H. mirabilis and the flower and flowering time are problematic.

The plants illustrated:

Set 1.  Map points 1, 2, 7, 12. Die Nekkies
Figs. 1.1 to 1.5 7526 H. maculata. Die Nekkies.
Figs 1.6 to 1.10, sn* H. maculata, Die Nekkies – East.

A chapter in Update 6 deals with H.maculata – Die Nekkies Biomes, and Haworthia maculata These few images give scant insight into the variability in respect of rosette and leaf form, colour and marking, and proliferation.

Set 2.  Map points 3, 4. Audensberg and Sandhills
Figs. 2.1 to 2.5 MBB6185 H. maculata. Audensberg.
Figs. 2.6 to 2.36 H. maculata, Vreesniet, Kanetvlei.

Unfortunately I found non-digital photography to be costly, cumbersome and unmanageable. So I have no field images for either population on the high Audensberg or at my original Sandhills site. However, I revisited the area and found the plants quite abundant a little further north at Vreesniet that is also Kanetvlei/Sandhills. The plants were almost confined to narrow rock cracks and their presence suggests to me that there is a very probable connection now eastwards to the Unknowns and even further on to the Osplaas H. arachnoidea, Etwin Aslander’s Hex Pass plants, and then still further to the H. marumana ‘dimorpha’ questionable.

Figs. 2.1 to 2.5 MBB6815 H. maculata. Audensberg.

Figs. 2.6 to 2.36 H. maculata, Vreesniet, Kanetvlei.


A myth corrected – part 4

Set 7.  Map point 11.  Cilmor.
Figs. 7.1 to 7.3 MBB7271

An interesting point here that H. herbacea (map point 18) occurs between this point and all the Die Nekkies populations. I am not sure that the area between can be fully explored although I have been into it.

Set 8.  Map points 15 and 16.  Hammansberg.
Figs. 8.1 to 8.21 MBB7992 H. maculata, Hammansberg.

The first locality 7992 midway between the eastern and western ends of the mountain was very exposed and very sparsely vegetated. The sandstone was very mineralised and fractured. The second locality to the west was well vegetated. In the view northward, the low hill in the near middle is the Draaivlei mountain that is Dwyka Tillite and hosts an abundance of H. herbacea as does the Dwyka strata on the northern side of the Hammansberg. The south side is Witteberg Sandstone with Bokkeveld shale on the lower south-facing slope. The view from the third locality southeastwards, shows massive valleys running into the Table Mountain Sandstone of the Riviersonderend mountains on the right. If one takes in mind that this is a short way east of the Wolfkloof Valley where H. herbacea ‘lupula’ occurs, then this area begs exploration. I do have a recall of H. herbacea in low altitude sandstone either near there or at a similar site at MacGregor to the east. The interesting view is the one looking eastwards that really gives the scale of exploration needed. The furthest mountains are the Langeberg Table Mountain Sandstsones and Buitenstekloof is nested in foothills there and in intrusive igneous rock. The coming closer is the Rooiberg, then Gemsbokberg, then Gannaberg, then Ouhangsberg and Hammansberg itself. Rooiberg is fairly well explored although the significance of H. herabcea ‘flaccida’ is not explained. I doubt if a anything new will crop up on the Hammansberg. But the intervening three definitely hold key populations marginal to these covered in this posting. The Langeberg? I only have the set 12 unknowns and cannot see myself resolving those.

Set 9.  Map point 17. H. pubescens. Sandberg.
Figs. 9.1 to 9.20 MBB7997 H. pubescens, Dewetsberg.

I long ago observed H. pubescens southwest of Sandberg and south of the Breede River on the Dewetsberg. This is not much further than 500m north of a population of H. maculata at Cilmor Cellar. Visiting the site again I observe that the plants are slightly different to the Sandberg plants. While they are equally small, cryptic and dark coloured, they have less surface spinuliferation that gives rise to the species name. Also some plants are distinctly spotted in the way that H. maculata are and these spots may give way to translucens when the plants grow in less light. See the images showing leaf surface detail of H. pubescens compared to the two collections of the Brickfields plants (images 21-27) of H. herbacea in Set 10, map point 18.

Haworthia maculata ↔ Haworthia pubescens, MBB8002 Cilmor

In Haworthia Update Vol 9 there is a report of a population MBB7997 identified as Haworthia pubescens from north of the Cilmor wine cellary. This is approximately 2-3km southwest of the type locality for the species. I noted that the plants have less spinuliferous leaf surfaces and there is a degree of surface translucens and maculation (spotting). I also presented 3 pictures of MBB7271 of what I identified as H. maculata from south of the Cilmor cellar. When I first visited this locality I had no problem identifying the few plants I saw as H. maculata on account of their marked spotting. However, on a recent visit we struggled to find plants at all and the few plants we found were too embedded in rock cracks to make any worthwhile identification. So we revisited the site to explore more extensively and located a large number of plants higher up and slightly west of our first sightings. These plants are illustrated here. They incline more to H. maculata than the plants at MB7997 and I have accessioned the population as MBB8002. There is the usual expected large variation in respect of superficial and observable characters. The plants can be proliferous and cluster, more so than at MBB7997. Similarly the leaves can have more translucens and even less spinuliferousness of the surfaces. Some plants have few and quite thick swollen leaves while others may have more and very slender pointed leaves. I have not observed the flowers and really do not expect them to make any difference to the problematic classification of populations that again are neither here nor there in a narrow concept of species. H. herbacea occurs at all four geographic positions at a radius of about 2km. At the brickfield to the northwest as well as just northeast of the Brandvlei Dam wall it is evident to me that there is a transition between H. maculata and H. herbacea. I did report the known distribution of H. maculata in Update 9. While there is no suitable habitat between Die Nekkies hills at the Brandvlei Dam and the Audensburg or Kanetvlei, there is unexplored suitable habitat southwards to Moddergat and Hammansberg. There is no evidence of H. maculata eastwards to where H. reticulata is known about 15km east on Ribbokkop. Westwards no Haworthia is known although G.J. Payne did inform me that he had observed plants in the hills immediately southwest of the Dam at the now submerged hot spring in the Brandvlei prison area.

The submitted pictures include two views. View 1 is looking north of east across the Breede River to the Sandberg where H. pubescens occurs. Its full occurrence on those low hills is not known and this I will explore soon. View 2 is looking eastwards looking at a Dwyka Tillite hill across the river in the upper right. We found no Haworthia on that hill although both H. pumila and H. herbacea are present on the smaller rise to the right and behind it – also Dwyka. H.herbacea is very abundant on a Dwyka tillite hill about 10km to the south. The corresponding hill on the left is Ribbokkop where H. herbaea, H. reticulata and hybrids are present, and H. arachnoidea also occurs. The limits of H. mirabilis are the higher hills in the background viz. Rooiberg, Gemsbokberg and those are Witteberg sandstones. ♦

Addendum Haworthia pubescens MBB8011, SW Sandberg

I need to point out that there is a still earlier article which covers Haworthia maculata (Haworthia maculata <–> Haworthia pubescens) that lays the basis for this discussion. In that article I note the position of the Sandberg to Cilmor and DeWetsberg and intended to include the Sandberg H. pubescens in that article. We could not get landowner contact and so that fell away. However, this problem was overcome and we first explored a Dwyka Tillite outcrop southeast of Sandberg. There is a vast accumulation of windblown sand on the first hill and we saw no Haworthia. There is a smaller hill further to the southeast that is also Dwyka and erosion exceeds wind deposition so smaller non-geophytes do quite well. We found both H. herbacea (see fig.1 MBB8014) and H. pumila there. From there we went to the southernmost point of the Sandberg. A misjudgment landed our vehicle in mud and the drama to get out limited the time we had to explore. We found a lone H. herbacea (fig. 2 MBB8012). Returning a week later we approached the Sandberg from the southwest, and almost immediately on reaching the top we found H. pubescent. Fig. 3 is a view towards Cilmor and DeWetsberg where the plants appear to be intermediate H. pubescens↔H. maculata. The picture is useful to get some idea of the role of geographic and geological considerations. The high mountains in the background are Table Mountain Sandstone and no Haworthia is known there. I am not certain that this is true and G J Payne did tell me that he had seen plants on the extreme lower right and south of the Brandvlei Dam. But also on the absolute distant and absolute left, is the Riviersonderend Mt. That is also TMS. The deep Wolfkloof Valley behind that is the locality for the much unexpected H. herbacea ‘lupula’. (These inverted single commas are not entirely necessary but I use them to underscore my informal use of names that have less reality. The var. lupula is real). The mountains ahead of that last line are Hammansberg on the left and the Moddergat to the right. Between there and DeWetsberg has not so far turned up Haworthia, but this is an exploration problem. Behind the DeWetsberg is also underexplored. H. herbacea does occur between DeWetsberg and behind the mountains on the low right just in the picture and also east of the brickfield out further right. H. maculata is only known in this area along the Nekkies north (further to the right) of the Brandvlei Dam just visible in the picture.

Prior to this exploration H. pubescens was to me only known from the northern part of the Sandberg that lies south of a road going eastwards to Eilandia between Worcester and Robertson. Here H. herbacea does occur on the lower northwest warm slopes.  H. pubescens seems to occur only on the upper two ridges and H. herbacea is not known to intermingle with it. This is all Witteberg Sandstone that as a formation overlies Bokkeveld Shale and underlies TMS.

Coming back to the southwest corner of the Sandberg where we found H. pubescens. The plants seem very similar to the species as it occurs to the north. They were very cryptic and often in shady rock retreats where they were really hard to see. It was mid- to late-morning when we were there and the plants were not going to be better exposed as the sun moved further west. Although there was very suitable well-drained habitat lower down on the shaded east slopes, there were no plants and I speculate that this may be because the plants may need the cooling effect of wind movement up on the ridge. The pictures tell the story of variability in respect of a whole range of leaf and rosette characters.

It is worth noting fig. 43 of the dead remains of a plant under a clump of restioid. It seems that seedling survival is closely coupled to early protection giving rise to the concept of nurse-plants. Plants are often very difficult to find because they are so hidden beneath accompanying vegetation. But they do need light and the dynamics of vegetation growth and densification must have quite a big impact on the ageing and survival of plants. It raises again the question of how long do the plants live? For plants like Aloe ferox and A. dichotoma I do have a real experience of a lower limit of about 35 years and a top limit in the several hundred. In  the field, the plants seem comfortably ageless.

The really interesting part is this. While I was busy tediously cleaning a plant to photograph it, Daphne called to me to come and see a lighter green plant she had seen. Moving in that direction I saw a plant that registered as H. herbacea but with some hesitation and doubt (see fig.4). I then went to see what Daphne had observed. They seemed less obviously H. herbacea but that seemed to be a logical and conservative opinion (see figs. 5-11). That was until Daphne found two adjacent rosettes at the foot of a restioid clump that left me in no doubt that they were hybrid H. pubescens/maculate (see figs 9-12). These were in bud whereas H. pubescens plants showed no sign of impending flowering. Note the buds are less well developed than MBB8014 further east, even if possibly insignificant. Going back to the other plants we confirmed my doubts. They were a lighter colour and apparently softer texture that we would have expected in H. herbacea. These were the only plants we saw in a space bridging the occurrences of plants of H. pubescens.

12-46 MBB8013 H. pubescens, SW Sandberg.

We did not explore the western slopes where habitat would have been more suitable for H. herbacea and I expect it does occur there. What puzzles me is that so frequently have I found very distinctive hybrids between species in close proximity and very seldom where the species are some distance from one another. I cannot say I have ever found a hybrid in the clear absence of both parents. The example of Astroworthia bicarinata at Lemoenkloof, east of Barrydale, may be an exception where only Astroloba corrugata (syn A. muricata, A. aspera) is present but H. pumila apparently not. Hybridization is thought to be an important element in the “evolution” of new species. I doubt this as it is quite evident that separation into two species is a pre-requirement. If new species have evolved in Haworthia by hybridization, how did they evolve as such in the first place? The answer to me lies in the continuities between populations. I observe, and have experienced of expected continuity between populations. While the Cilmor populations are thought to be H. pubescensmaculata it cannot be said anymore that they are hybrid, or populations where the morph or drift to discrete elements has not reached a conclusion. The latter is more likely. As there is already apparent geographic continuity of H. maculata and H. herbacea, I was expecting some evidence of a similar relationship between H. herbacea and H. herbacea.  So here it is. Hybridization as a factor in speciation in Haworthia does not seem to very likely. It confirms for me that there is a fractal “chaotic” order to species in Haworthia and the reality is that a view of many truly discrete species is a fabrication and a very ill-considered view.

Acknowledgement
We are always greeted with such kindness and helpfulness that we might have expected this from the Sandberg landowners too. It came in no small measure. Driaan Griesel was most enthusiastic and interested and also helped us with extracting our vehicle from the mud on the one occasion, and then jump-starting it after a flat battery on the second. Our imposition did not so much as touch his view of the day.

[Ed.] Bruce made another visit to Southwest Sandberg on 9 December 2012 and includes the following flower pictures. He makes this comment; personal correspondence 27 December 2012.

I am actually not sure at all about flowering time now. I used to be quite sure of being able to collect seed of pubescens mid-Nov. But I observed at Humansdorp that gordoniana peak flowering could be out by 6 weeks. In  any case the plants can produce successive spikes so one can get delayed flowering and added to that energy in the first or the second flower set. I know mirabilis at MacGregor can flower from Nov. thru to March while at Montagu mirabilis can flower as late as April/May. Retusa and geraldii are quite happy to produce flowers in either Spring or Summer and Kobus observed that splendens did that too. Maculata can flower from Sept. thru to late Dec. And each population does its own thing.

.♦