Notes on the taxonomy of Rubus in southern Africa

The taxonomy of Rubus in South Africa is beset with problems. These include the introduction of extra-African species as ornamentals and crops, the apparent segregation of new forms and finally hybridization with indigenous species. These problems are compounded by poor and incomplete collecting of Rubus in South Africa, and by the difficulty of relating introduced taxa to the many and varied species, varieties and ecotypes occurring in other countries.


RESUME NOTES SUR LA TAXONOM IE D U RUBUS EN AFR1QUE AUSTRALE La taxonomie du Rubus en Afrique du Sud est assiegee de problemes. Ceux-ci incluent I 'introduction d 'especes extra africaines comme plantes ornementales et de cultures, I 'apparente segregation de nouvelles formes et finatement /'hybridation avec des especes indigenes. Ces problemes se compliquent par des collections incompletes et pauvres de Rubus en Afrique du Sud et parla difficulte d ' apparenter les taxa introduits aux especes nombreuses et variees, aux varietes et aux ecotypes des autres pays.
Prior to the cytogenetic studies of Gustafsson (1942Gustafsson ( , 1943) ) there had been only a few studies such as those o f Sudre (1908( -1913( ), Focke (1911( -1914) ) and Bailey (1941), that had attempted to produce an overall taxonomy of Rubus.By 1913 there were already some 3 350 Latin names in existence, many of which were applied to primary hybrids or very localized varieties (Gustafsson, 1943). As Newton (1975) has suggested, this may have been due to the vague species concepts prevailing at that time or it may have resulted from parochial attitudes adopted towards plants o f widespread distribution.Never theless, from the work done by Gustafsson and other geneticists, we now know that much o f the m or phological diversity was a result o f natural hybridiza tion, polyploidy and apomixis.
Pseudogamy, a process whereby heterozygous segregates and hybrid derivatives can be maintained in nature, has played a particularly important role in the development o f many agamic complexes in Rubus (Grant, 1971).A batologist not only has to contend with these basic genetic difficulties, but also has to deal with considerable phenotypic plasticity (Beijerinck, 1953;Heslop-Harrison, 1963).it is little wonder then that this remarkable genus had defied taxonomists for over three centuries and that there is still no consensus on supraspecific categories.
The problems o f Rubus taxonomy in South Africa are aggravated by the introduction and naturaliza tion o f exotic species, the apparent segregation of new forms in areas surrounding cultivated black berries, the role o f hybridization among local, as well as between local and exotic species, and finally by inadequate herbarium material.
Harvey (1862) recognized five species o f Rubus in South Africa: Rubus fruticosus, R. ludwigii, R. pinnatus, R. rigidus and R. rosaefolius.The last overall revision was by C. E. Gustafsson (1933) who added a further six species: R. adolfi-friederici, R. chrysocarpus, R. ecklonii, R. immixtus, R. intercurrens 1975) could find no trace o f hybridization having occurred in Victoria, Australia, although they did not preclude this as a future possibility once the introduced species had increased their ranges.
These preliminary observations suggest that a tax onomic revision of Rubus in South Africa is highly desirable, but such a study would obviously be a long-term project.Although many modern revisions o f Rubus are based largely on cytological investiga tions, it is nevertheless still necessary, in many parts of the world, to rely on herbarium material.If this material is inadequate or incomplete then the task becomes very difficult.The importance o f complete herbarium collections has been stressed by a number o f workers (Am or & Miles, 1974;Beijerinck, 1953;Edees, 1959;Watson, 1958).My impression, after having seen much of the available herbarium mate rial in South Africa, is that few collectors have been aware o f what would constitute adequate pressed material for the naming of a specimen of Rubus. For this reason, and because so little material is available for study and because some areas are undercollected, I have outlined the four components which make up a useful herbarium voucher (see also Am or & Miles, 1974;Beijerinck, 1953) -this in the hope that it will encourage collectors to collect more material.
It is important to collect: 1.One 10 cm section, with leaves, selected from the middle of a first-year cane o f vegetative growth (primocane).Rubus usually, but not always, flowers in the second season.The first season or primocanes are easily recognized by their lush and robust growth.
2. One 10 cm section, with leaves, from the middle o f a flowering cane (floricane).This is necessary as in most species of Rubus in South Africa the floricanes differ markedly in shape, size, leaf-shape and number, and presence or absence o f a white bloom.
3. A complete inflorescence with flowers, and fruits if these are available.4. A few petals dried separately.The petals o f some species, if not collected separately, become lost dur ing drying as they abscise rapidly after collection.
This dried material should be accompanied by full descriptive notes.O f great taxonomic value are the colours of petals, young and old fruits, primocanes and floricanes.The relative length and colour of stamens and styles, as well, as the relative lengths of calyx lobes and petals which are often very diagnos tic.It should also be noted whether canes are erect, arching or looping.Looping canes may tip-root dur ing the Autum n equinox.O dd forms or unusual plants should also be collected with a note to that ef fect.It is important to accurately record the localities o f oddities, particularly if they are collected near blackberry orchards or in areas known to be heavily treated with herbicides, as these chemicals are known to cause chimeras and unusual phenotypes.

I
am indebted to Professor K. D. Gordon-Gray, Drs R. L. Am or, O . A. Leistner and Mr M .J. Wells for their valuable comments and criticism.
and R. transvaalensis.Also mentioned by Gustafsson were two species o f introduced brambles: R. affinis and R. argutus, of which only the former was noted as a naturalized weed.Not included in any o f these are the now naturalized weeds R. niveus Thunb.(Java bramble), R. cuneifolius Pursh (American bramble) and R. phoenicolasius Maxim, (wine berry).It is often difficult to decide whether certain species are indigenous or introduced.A case in point is R. immixtus C. E. Gust.The type locality of this species is Hogsback in the eastern Cape, an area which has an extremely variable Rubus flora and well noted for its large number of naturalized European plants such as gorse (Ulex europaeus), roses (Rosa spp.) and hawthorns (Crataegus).It seems quite pro bable, therefore, that R. immixtus may be either a European species or a hybrid with R. rigidus Sm. in .plicatus x pinnatus and R. pinnatus x rigidus.The tremendous variability of R. rigidus and R. pin natus in South Africa may well be explained by their apparent ability to hybridize with other species.It is interesting to note that Amor & Miles ( * Botanical Research Institute, Department of Agriculture and Fisheries, Private Bag X101, Pretoria, 0001.studies R