VARIOUS AUTHORS 285

LEGUMINOSAE TYPIFICATION OF SC HOT!A CAPITATA Bolle, in Peters Reise Mossamb. Bot. I: 18 relied upon. Only two Schotia species are known (1861), based his description of Scholia capitata on from Mozambique, namely, S. hrachypetala Sond. a flowering specimen collected by Peters at Inham- and the species interpreted as S. capitata. Collections bane in Mozambique. Unfortunately the holotype from the type locality have failed to reveal a third in the Berlin Herbarium was destroyed during the species. Bolle provided a detailed floral description war, and all efforts to trace an isotype have proved o f S. capitata and fortunately the reference to five unsuccessful. In the British Museum (Natural History) petals eliminates S. hrachypetala from contention. there is, however, a sketch prepared in 1926 of a Although the application of the name S. capitata portion of the holotype in the Berlin Herbarium. is no longer in doubt, in the absence of any type The sketch consists of one leaflet (2,6x1,6 cm), material it seems desirable to select a neotype. I now one flower bud, and carries the comment “ material select the flowering specimen Balsinhas 204 in the very poor”. Unfortunately this sketch does not Kew Herbarium from Mozambique, Lourengo enable S. capitata to be positively identified. Marques Prov., Goba, proximidades do rio Maivavo, As mentioned by Codd in Bothalia 6 (3): 522 as the neotype of 5. capitata. (1956), in establishing the identity of S. capitata, circumstantial evidence had, to some extent, to be J. H. Ross

POACEAE A NOTE ON A POSSIBLE BIGENERIC HYBRID BETWEEN DIG IT A RIA A N D In 1932 Goossens described Anthephora aequiglumis, distinguishing it from the closely-related Nees by the distinctly-elongated branches of the rhachis (at least in the lower part), which are about 7 mm long, and the subequal glumes which are shorter than the lemmas which do not form a false involucre. Later in that same year. Stent (1932) recombined the species into a new , Tarigidia, which she described in order to accommodate this . She maintained that . . the structure of the spikelet is very similar to that of D igitaria, but the elongated lower glume and the type of inflorescence place Tarigidia beyond the limits of that genus”. More recently, Launert (1957) described this grass as Digitaria otaviensis, completely overlooking Tarigidia aequiglumis. Subsequently in 1970 he sank his species under T. aequiglumis. On examination of the type of T. aequiglumis and F ig.— 6.—Distribution of Tarigidia aequiglumis within the all available material, it became evident that this region of overlap of Digitaria sect. Erianthae and Anthe­ grass shares characters of Anthephora and D igitaria. phora pubescens. representing a possible bigeneric hybrid between these two genera, the putative parents in this case Attempts to hybridize Anthephora pubescens and being Anthephora pubescens and one of the species of one of the species of Digitaria sect. Erianthae artifi­ Digitaria sect. Erianthae. cially should be undertaken to show that the progeny resemble material of the presently known Tarigidia Tarigidia resembles Anthephora in having the same aequiglumis. type of spicate inflorescence, spikelets in clusters, the production of the lower glumes, which are almost as Material examined long as the lower lemmas, and by the occasional S.W.A.—1917 (Tsumeb): Otavi (-CB), Dinter 5589. This fusion of the lower glumes. Resemblance to Digitaria specimen was designated as the type of Digitaria otaviensis is evident in the nervation of the lower lemma, and by Launert (1957). the characteristic flattened upper lemma with broad, O.F.S.—2726 (Odcndaalsrus): Hoopstad district, Odendaals- thin margins. One specimen in particular, Dinter rus, on an empty plot (-DC), Schultz s.n. sub. PRE 8344 (PRE, 5589 (collected at Otavi in South-West Africa), has holo.); 2727 (Kroonstad); Groenebloeme Station (-CA), the branches of the panicle produced on the lower Potts 2674 (PRE, para.). half with the lower glumes of many of the spikelets C a p e .— 2624 (Vryburg): 35,2 km N.W. of Vryburg on Genesa reduced as in Digitaria. Here the similarity to that Road (-DC), Acocks 12692. genus is closer than it is to Anthephora. REFERENCES It is generally accepted today that Anthephora belongs to the tribe , (c.f. Chippindall 1955; B r o w n , W. V. & S m i t h , B . N., 1972. Grass evolution, the Simon 1971 and Clayton 1972) however, Tatoeka Kranz syndrome ,*C/'*C ratios and continental drift. Nature 239: 345-346. (1957) and Brown & Smith (1972) follow Pilger (1954) in keeping Anthephora in a separate tribe, C h i p p i n d a l l , L. K. A., 1955. Paniceae in Meredith et al. Anthephorae. The grasses and pastures o f South Africa. 317-453. Johannesburg: C.N.A. Up to now T. aequiglumis has only infrequently Ci ayton, W. D., 1972. Gramineae in Ft. W. Trop. Afr. 3 , 2 : been collected within the region of overlap of Digitaria 354-356. sect. Erianthae and Anthephora pubescens (Fig. 6). From the number of collections of T. aequiglumis G o o ssf .n s , A. P., 1932. The genus Anthephora Schreb. Trans. R. and from its disjunct distribution, it is evident that Soc. S. Afr. 20: 189-200. hybridization, under natural conditions, if occurring, L a u n e r t , E ., 1957. Neue arten und erstfunde aus Sudwest- is a rare phenomenon. afrika. Mitt. Bot. Miinchen 2: 307. 286 NOTES ON AFRICAN

L a u n e r t , E., 1970. Gramineae in Prodr. FI. S.W. Afr. 160: S t e n t , S . M. 1932. Notes on African grasses, XII. Kew. Bull. 202. 1932: 151.

P il g e r , R., 1940. Unterfamilie . Gramineae III in T a teo ic a , T ., 1957. Proposition of a ne.v phylosere ic systen Pflanzenfam. Band 14 de. of . J. Jap. Bot. 32, 9: 275-287. S im o n, B. K., 1971. Rhodesian and Zambian grass lists. Kirkia 8, 1: 3-83. A. E. Loxton

NON-KRANZ (C3) LEAF ANATOMY IN THE PANICOID GRASSES: ADDITIONAL RECORDS FROM SOUTH AFRICA Anatomically the Kranz syndrome is characterized Listed below, together with specimen citations, by the presence of a parenchymatous bundle sheath are further non-Kranz species of the Paniceae as composed of large, thick-walled cells, possessing judged by leaf anatomical structure: specialized chloroplasts and often surrounded by Acroceras macrum Stapf (F ig. 7.) radially-arranged mesophyll cells. Smith and Brown (1973) and Brown and Smith (1972) point out that T r a n s v a a l .—2330 (Tzaneen): Lakeside Holiday Resort (-CA ), Ellis 1595. these anatomical specializations are related to physio­ logical characters associated with the photosynthetic C a p e .—3228 (Butterworth): Mazeppa Bay (-BC), Ellis 266. fixation of carbon. At least two different photosyn­ Alloteropsis semialata (/?. Br.) Hitchc. thetic pathways are present in the Poaceae (=Grami- T r a n s v a a l .—2531 (Komatipoort): Ida Doyer Nature neae), each being consistently associated with a Reserve (-CC), Ellis 378. particular form of leaf anatomy (Carolin et al. 1973). N a t a l .—3030 (Port Shepstone): Eureka Farm, Izotsha Thus the highly-specialized parenchyma sheath is ( - C D ) , Ellis 733. restricted to those plants exhibiting the C4 photo­ Panicum aequinerve Nees synthetic pathway and non-Kranz, C3 plants T r a n s v a a l .—2531 (Komatipoort): 15 km from Barberton have a thin-walled parenchyma sheath with a on the road to Shya la Ngubo dam (-CA), Ellis 1541. few plastids, smaller than, but of the same type as N a t a l .—2829 (Harrismith): Cathederal Peak, stream banks those of the mesophyll. Anatomical evidence is (-CC), Ellis 1453; in Rainbow forest, Ellis 1457. 2929 (Under- therefore reliable for predicting the Kranz and non- berg): 16 km from Underberg on the road to Mondi (-DD), Kranz syndromes. Ellis 214. 3030 (Port Shepstone): Eureka Farm, Izotsha (-CD), Ellis 1470. Smith and Brown (1973) give 13C/12C ratios for C a p e .—3227 (Stutterheim): Fort Warden, Komga (-DB), 198 grass species and correlate this with anatomical Ellis 283; Hogsback forest (-CA), Ellis 1476. and physiological data from the literature. Their Panicum ecklonii Nees (F ig. 8.) results show that all the species in the subfamilies N a t a l .—2729 (Volksrust): Majuba pass (-BD), Ellis 30. Arundinoideae, Oryzoideae and Festucoideae are non- 2829 (Harrismith): Cathedral Peak (-CC), Ellis 1421; Ellis 1449. 2929 (Underberg): Kamberg (-BD), Ellis 53. Kranz and all species of the subfamilies Eragrostoi- deae and Aristidoideae are Kranz. Only in the sub­ Panicum hymeniochilum Nees family Panicoideae do both Kranz and non-Kranz T r a n s v a a l .—2530 (Lydenburg): Badplaas (-DC), Ellis 465; 19 km from Nelspruit on the road to Alkmaar (-BD), types occur. The tribes Andropogoneae, Anthepho- Ellis 1560. 2630 (Carolina): 16 km from Amsterdam on the reae, Arundinelleae, Maydeae and Melinideae are road to Piet Retief (-DA), Ellis 1523. Kranz, and the tribe Isachneae is non-Kranz. Only Panicum laticomum Nees

in the tribe Paniceae itself are both Kranz and non- C a p e .—3227 (Stutterheim): Fort Warden, Komga (-DB), Kranz genera found and the genus Panicum is reported Ellis 1101. as being the only genus with both Kranz and non- Panicum natalense Hochst.

Kranz species. Non-Kranz, panicoid genera repre­ T r a n s v a a l .—2528 (Pretoria): National Botanical Gardens, sented in South Africa and recorded by Smith and Brummeria (-CA), Ellis 417. 2530 (Lydenburg): Long Tom pass (-BA), Ellis 453; Sabie, Ellis 460. 2531 (Komatipoort): Brown, are Sacciolepis and Oplismenus. None of the Ida Doyer Nature Reserve (-CC) Ellis 376.

non-Kranz species of Panicum mentioned by the above N a t a l .—2829 (Harrismith): Cathedral Peak (-CC), Ellis authors are found in South Africa. 1442. 3029 (Kokstad): Franklin (-AD), Ellis 223.

F ig . 7.—Acroceras macrum. Transverse section of leaf blade showing paren­ chyma sheath cells devoid of chloroplasts and dis­ tinctly radiate chloren­ chyma. Ellis 266. > 40.