Haworthia Revisited – 54. Haworthia scabra

54. Haworthia scabra Haw., Suppl.Pl.Succ. :58(1819).  Bayer :152(1976).  Scott, Cact.Succ.J(U.S.) 52:274(1980).  Bayer :73(1982).  Scott :4(1985).  Type: Cape, not preserved.  Lectotype: icon (K):  H. tuberculata V.Poelln., Feddes Repert.Spec.Nov. 29:219(1931).  Scott loc.cit. :275 (1980).  Type: Oudtshoorn Distr., Mrs van der Bijl.  Not preserved:  H. tuberculata var. acuminata V.Poelln. loc.cit. 44:218(1938).  Type: Keurbooms River Heights, Woolley in Long 394.  Not preserved:  H. tuberculata var. sublaevis ibid. :219.  Type: Oudemuragie, near DeRust, Mrs Blackburn in Triebn.1062.  Not preserved:  H. tuberculata var. subexpansa V.Poelln., Cact.J 7:56(1938).  Type: Ladismith, Joubert in Triebn.813.  Not preserved:  H. tuberculata var. angustata V.Poelln., Feddes Repert.Spec.Nov. 49:27(1940).  Type: Oudtshoorn Distr., Vergelegen, STE7793.  Not preserved.

scabra: rough.

Rosette stemless, slowly proliferous, to 160mm tall.  Leaves to 160 X 22mm, attenuate, rotate, incurved, nearly as thick as wide, surfaces scabrid or smooth, with or without distinct raised non-confluent concolourous tubercles.  Inflorescence sparsely branched, lax.  Flowers tepals fused, tube straight, inner lower tepals revolute.

1982 – H. scabra is a very dark species with scabrid, tubercled leaf surfaces similar to H. nigra.  It is widespread from Baviaanskloof to Ladismith.  Size is variable and the plants may be as small as 30mm or as tall as 150mm.  The form of the plants varies from having the leaves distinctly tiered, in either 3 or 5 vertical tiers; and either incurved or suberect and spreading.  In some forms there are few leaves and the plants are acaulescent, but in others with many leaves, the plants do have a short leafy stem.  The plants may be highly proliferous or solitary.  The habitat requirements are quite interesting and reflect the tendency in the genus for populations to be highly localised.  Thus H. scabra may, in the mountains north of Oudtshoorn, be found crowded in white quartz outcrops, in the same way that H. venosa ssp. tessellata occur in such formations in the Richtersveld.  The var. morrisiae is a much smoother element which is known only from the Schoemanspoort area north of Oudtshoorn, in the same local area as H. starkiana.  Von Poellnitz actually reported that it was collected growing next to H. starkiana and H. scabra, suggesting that it was a hybrid.

1999 -It is curious that Von Poellnitz, followed by Scott, placed H. scabra in a different section to H. tuberculata. It is evident from Von Poellnitz’s citations and Scotts distribution maps that the distributions of the two elements overlap. The difference between the two (and therefore at sectional level also) was the argument that the former had minute warts and the latter distinct tubercles. It has already been pointed out that tuberculate and glabrous leaf surfaces may occur in one and the same species. Von Poellnitz’s comment that var. morrisiae occurred with H. starkiana and with H. scabra (he used the name tuberculata) is now fairly confidently thought to bear also on the problematic identity of H. smitii. Plants of that species were given to Mrs Lategan by Jacob Smit and still survive as more than one clone in JD. Venter’s collection. There has been a problem in relating them to field populations and it is only recently that a collection by Mary Parisi was considered to be this species. However, that collection and an extensive population examined in the closer Schoemanspoort area contain elements which can positively be related to both var. morrisiae and to H. smitii. The only detracting fact is that the surviving cultivated plants of the latter are much larger and robust than those seen in the field . This is presumed to be from its long history of cultivation.

Based on these observations, it has been decided to simply include H. starkiana under H. scabra and widen the circumscription of the var. morrisiae to include H. smitii. The var. lateganiae is probably as much a variant of H. scabra as is H. starkiana and it is more sensible for the arrangement to reflect that fact. The var. starkiana occurs from the Raubenheimer Dam south-westwards to the mouth of Schoemanspoort, while var. lateganiae occurs in the area to the east around Oudemuragie. There is clearly a great deal of interaction between these two varieties and the typical scabra which grows in the same area. They do not seem to co-occur in the way that one would expect discrete elements would, and the var. morrisiae seems to be the intermediate. Over the rest of the range of the species, only the identifiable var. scabra is found in a range of guises. Near Prince Alfred’s Pass, the leaves may be long and slender, more numerous and less tubercled. Whether or not this leads directly on to H. sordida is an open question but probably unlikely.


a. var. scabra
The var. scabra is the widespread, very variable element. The plants may be quite small and blackish-green, lightly or heavily tubercled and the leaves may be manifestly trifariously arranged, or quinquefariously. The leaves may be tightly incurved and rotate, or erect-spreading.

Distribution: 3321 (Ladismith): (‑CA), Dekenah 91 (NBG), Oddie in NBG1971/32 (BOL); E. Ladismith (-CB), Smith 5511 (NBG).  3322(Oudtshoorn): Cango (-AC), Smith 5224, 5225, 5225a, 5226 (NBG); Oudemuragie (-AD), (NBG); Vanwykskraal (-CB), Fourcade 217 (NBG); Kleinpoort (-CB), Taylor 20409 (BOL); Mannetjiesberg (-DB), Esterhuysen 18369 (BOL).  3323(Willowmore): Haarlem (-CB), Fourcade 3610 (BOL), Smith 3668 (NBG); DeHoop (-CB), Wisura in KG507/71 (NBG); Saptou (-CB), Bruyns 7078 (BOL); Keurboomsrivier (-CC), Bouwer (NBG); Luiskraal (-DA), Forrester 101 (NBG); Braamriver (-DB), Bruyns 7083 (BOL).

Inadequately located: Avontuur, James (BOL); Uniondale, Coates (BOL); Calitzdorp, ex hort Whitehill (NBG).


b.var. lateganiae (V.Poelln.) Bayer comb.nov.  H starkiana var. lateganiae (V.Poelln.) Bayer :128(1976).  Bayer :75(1982).  H. lateganiae V.Poelln., Des.Pl.Life 9:103(1937),  ibid., Feddes Repert.Spec.Nov. :99(1938).  Type: Cape, Vanwykskraal, Mrs L. Lategan in Triebn.1023.  Not preserved.  Lectotype (B&M): icon (B):

lateganiae: in honour of Mrs L. Lategan.

This variety has rather long, slender leaves which are smooth and shiny, but darker green than in the var. starkiana. It was originally considered to be with a short stem but this cannot be considered to be diagnostic in any way. It is quite probable that this may just manifest itself because the leaves tend to be thinner and thus have a greater ratio of size to stem than does var. starkiana.

Distribution: 3322(Oudtshoorn): Groot Kruis (‑AD), Otzen in NBG69248 (BOL), Smith 3859, 3860 (NBG).


c. var. morrisiae (V.Poelln.) Bayer :137(1976).  Bayer:74(1982).  H. morrisiae V.Poelln., Kakteenkunde 9:l32(1937).  ibid., Feddes Repert.Spec.Nov. 43:100(1937).  Type: Cape, Cango, Mrs F. Morris in Triebn.916.  Not preserved.  Lectotype (B&M): icon (B).

morrisiae: for Mrs G. Morris.

As indicated in the discussion of the species, the concept of this variety has been widened to cover the intermediates between the varieties scabra, starkiana and lateganiae. This interaction is known only in the Rust-en-Vrede area of Cango but it should be expected to extend to the Grootkruis area to the south-east. The tubercles are generally very small and confluent so that the surfaces are really minutely and uniformly scabrid. However, a widened concept needs to allow for a lighter coloured variant with occasional, more prominent tubercles.

Distribution: 3322(Oudtshoorn): Schoemanspoort (-AC), Venter 6 (NBG); Schoemanspoort (-AD), Bayer in KG119/72 (NBG).

Inadequately located: Oudtshoorn, Van Ryneveld in NBG 69245, Peers 29 (BOL).


d. var. starkiana (V.Poelln.) Bayer comb.nov.  H. starkiana V.Poelln., Feddes Repert.Spec.Nov. 33:73(1933).  ibid. 43:101(1938).  Type: Oudtshoorn, Taylor.  Not preserved:  Lectotype (B&M): icon (B):  H. taylori Barker in ms. (NBG).

starkiana: in honour of Prof. Peter Stark.

1982 – This species is particularly remarkable in that it has a very limited distribution and also in its relation to H. scabra.  It occurs only as far as is known in Schoemanspoort, north of Oudtshoorn, and in the valley extending eastwards to the Raubenheimer and Melville dams.  It occurs on steep, hot north facing slopes, forming large yellow clumps which are conspicuous against the brown and greys of the rocky terrain.  The plants show exactly the same range of variability as do those of H. scabra; i.e. the leaves may be short, incurved and twisted spirally to one side, or they may be longer, suberect, and less twisted.  The var. lateganiae is a darker, dull green in contrast to the bright yellow‑green of the var. starkiana.  The leaves are also less turgid and swollen so that the plants are correspondingly smaller for the same number of leaves.  The var. lateganiae grows in an adjoining valley at a slightly lower altitude and is so far known only at the eastern end of the valley.  Its distribution and variability are thus still really unknown.  The relationship of H. starkiana and H. scabra is very obvious from the similar branched, slender inflorescences, and from the remarkable similarity of the various growth forms.  Some collections of H. scabra var. morrisiae (a relatively smooth variety) are practically as smooth as H. starkiana itself, and have included forms practically identical to that species.  The localities are, however, very close to one another and it is not possible to say that these intermediates are hybrids or not.  H. smitii was possibly a genuine hybrid of H. starkiana and H. scabra as it has the colour and texture of the former species, and the tubercles of the latter.

1999 – Speculation surrounded the identity of H. smitii until the very last moment and it was not expected that a population of such a plant would be found outside of the area of the var. starkiana.  Initially it was included here with var. starkiana as a tuberculate variant.  However, further collecting during the writing of this book confirmed the direct integradation between typical H. scabra and var. starkiana. What is curious is that while the evidence of this intergradation is irrefutable in two populations, there is also the situation where the two elements virtually co-occur in the absence of intermediates.  It is the intermediates which are recognized in a wider concept of the var. morrisiae.

Distribution: 3322(Oudtshoorn): Schoemanspoort (-AD)., Taylor 6511 (PRE), Taylor in NBG696/29, 20408 (BOL), Mavis in NBG1682/36, Lategan (NBG).

Inadequately located: Oudtshoorn, Morris in NBG 68240.