42. Haworthia attenuata Haw., Syn.Pl.Succ. :92(1812). Haw., Rev.Pl.Succ. :54(1821). Berger, Das Planzen. :93(1908). Bayer :100(1976). Bayer :63(1982). Scott :15(1985). Aloe attenuata Haw., Trans.Linn.Soc. 7:11(1804). Salm-Dyck, Monogr. 6:t12(1834). Aloe radula sensu Ker-G., Curtis’ Bot. Mag. t.1345(1811). Type: Not preserved. Neotype (designated here): icon t1345, Curtis’ Bot. Mag.: H. clariperla Haw., Phil.Mag. :186(1928). Aloe attenuata var. clariperla (Haw) Salm-Dyck, Monogr. 6:12b (1834). H. attenuata var. clariperla (Haw.) Baker, J Linn.Soc. :204(1880). Farden, Cactus J 8:34(1939): H. attenuata fa clariperla (Haw.) Bayer :107(1976). Bayer :63(1982): Type: Not preserved. Neotype (designated here): icon, 6:12b, Salm-Dyck, Monogr.: H. fasciata var. caespitosa Berger, Das Pflanzen. 33:92(1908). Type: Not preserved: H. brittenianae V.Poelln., Feddes Repert.Spec.Nov. 41:196(1937). H. attenuata var. britteniana V.Poelln., Kakt.u.and.Sukk. 10:166(1937). H. attenuata fa britteniana (V.Poelln.) Bayer :104(1976). Bayer :63(1982). Type: Cape, Oudtshoorn, Mrs van der Bijl in Triebn. 807. Not preserved. Neotype (B&M)): icon (B). Epitype: Plutosvale, Smith 1118 (NBG). H. attenuata var. odonoghueana Farden, Cactus J 8:34(1939), et vars. linearis, uitewaaliana, deltoidea, minissima, inusitata. Types: not preserved.
attenuata: tapering leaves.
Rosette stemless, proliferous, to 130mm tall. Leaves to 130 X 15mm, attenuate, spreading, lanceolate-deltoid, surfaces scabrid with distinct raised non-confluent tubercles. Inflorescence sparsely branched, lax. Flowers tepals fused, tube straight, lower inner tepals revolute.
1982 -This species is widespread in the Eastern Cape from beyond the Fish River in the east, to the Humansdorp area in the west. Old records suggest that it was collected near Umtata in the Transkei and also in the Kei River valley. It may appear to be closely related to H. fasciata but it is distinguished by having longer, thinner and generally more spreading leaves. The leaf faces are also always tubercled. H. attenuata occurs in the drier Valley Bushveld vegetation. In the Hankey area, H. attenuata and H. radula grow together. There do not appear to be any direct and obvious intermediates between very similar species but breeding experiments may prove them to be the same. H. attenuata fa clariperla is very common in gardens and collections throughout South Africa and is recognised by the large white confluent tubercles in bands across the back of the leaves. The original origin of this form is not known but it has been collected at Fort Brown. H. attenuata fa britteniae was collected in Plutosvale and is recognised by its many large white discrete tubercles. Farden recognised a host of varieties based on minor differences. Smith (1950) pointed out some of the difficulties which arise when one goes to this extreme, saying that it would be possible, on the basis of leaf characters, to derive several varieties from the same plant.
1999 – The recognition of two separate forms is abandoned, and it is suggested that variants are referred to by locality of origin. There is as yet no further confirmation of a distribution extending to beyond the Fish River in the east. Breuer and Metzing neotypify the species name by a specimen which is cited under the var. radula in this work. There really is no need to depart from early illustrations for typification purposes for which there are several precedents. The species is well known, and Breuer and Metzing have introduced the first confusion of identity in the history of the species.
a.var. attenuata.
This variety is recognised from necessity arising from the inclusion of H. radula within this species, and in preference to recognising the many other varieties already in the history of the species.
Distribution: 3226 (Fort Beaufort): Middeldrift (-DD), Smith 3511 (NBG). 3324 (Steytlerville); E. Hankey (-DD), Smith 2910 (NBG); Kleinriver (-DD), Smith 7380 (NBG). 3325(Port Elizabeth): Enon (-BC), Bruyns in KG40/77 (NBG); 21km Port Elizabeth to Uitenhage (-CD), Smith 5820 (NBG); Addo (-DA), Barker 5090 (NBG), Smith 3549 (NBG); (-DC), Britten in PRE 34850; Swartkops (-DC), (BOL). 3326 (Grahamstown): Plutosvale (-BA), Marloth 5962 (PRE), Harcourt Wood in BOL20729, Britten in PRE 34891, NBG338/34, Smith 914, 1118 (NBG); Fort Brown (-BA), Smith 3512, 3825 (NBG); Glenelg (-BB), Smith 5406 (NBG); Ballinafad (-BB), Smith 5400, 5407 (NBG); E. Fish River (-BB), Smith 5409, 5410 (NBG).
Inadequately located: Near Grahamstown, Luyt in NBG304/45 (NBG); Albany dist, Britten in PRE 34892; ex hort, Simpson in NBG233/42, Peers in NBG34/36, Howard in NBG82/25, Meterlerkamp in NBG1450/29 (NBG), van der Merwe in NBG1696/29 (NBG), Marloth 4210, 5963 (PRE); Kingwilliamstown, Carter (BOL); Sandile Koppie, Grant & Blenkins in NBG2257/26 (BOL); Smith 617, 2822, 3225, Bayer 169 (NBG), Henderson 1399 (NBG).
Haworthia attenuata var. attenuata JDV90/42 Addo. The tubercles are conspicuously white on both leaf surfaces. The leaves have no fibres although viscid strands can be seen on breaking the leaves. 
Haworthia attenuata var. attenuata MBB6831 Hankey. Not surprisingly the tubercles can be smaller and evenly spaced.
b. var. radula (Jacq.) Bayer comb.nov. Haworthia radula (Jacq.) Haw., Syn.Pl.Succ. :8(1812). Berger, Das Pflanzen. 38:94(1908). Bayer :148(1976). Bayer :71(1982). Scott :19(1985). Aloe radula Jacq., Hort.Schoenb. 4:11.t422(1804). Salm-Dyck, Monogr. 6:t8(1834). Type: icon, t422, Jacq., Hort.Schoenb. Epitype (B&M): CAPE-3324 (Steytlerville): 1.5km E. Hankey (-DD), Smith 3190 (NBG).
radula: like a rasp.
1982 – H. radula is only known from the immediate vicinity of Hankey in the Humansdorp area where it grows together with H. attenuata. There does not appear to be a random association, however, and the two species appear to be discrete in the small localised population as far as is known. H. radula always has many minute crowded white tubercles in contrast to the larger fewer tubercles on the leaves of H. attenuata. Despite the variability of the latter species there do not seem to be obvious intermediates between the two. It is a very attractive plant in cultivation as the many tubercles give the plant a silvery appearance.
1999 – On reflection, some of the above comments are not strictly true. There is some difficulty in deciding if a plant properly qualifies as ‘radula’ or not. There is no evidence at all of range and distribution for the element outside of H. attenuata; perhaps even varietal rank is unnecessary.
Distribution: 3324 (Steytlerville): Sandland, 20 km E. of Patensie (‑DA), P.L. Perry 660 (NBG); Humansdorp (-DD), Marloth 5961 (PRE); Hankey (‑DD), Smith 3190, 7382 (NBG), Holland 4141 (PRE); 1.5km E. Hankey (-DD), Smith 2933, 3190 (NBG); 1.5km N. Hankey (-DD), Smith 3692 (NBG); 1.5km SE. Patensie (-DD), Smith 3182 (NBG).
Inadequately located: Cape, Marloth 10086 (PRE); ex hort, Buckland in NBG428/31 (BOL, NBG), Malherbe in NBG486/42 (NBG); Hansmeykraal, Fourcade 3550 (BOL).
Haworthia attenuata var. radula MBB6803. The tubercles are small and plentiful and the leaves thus appear scabrid rather than tuberculate.