The botany of the Cunene-Zambezi Expedition with notes on Hugo Baum ( 1867-1950 )

A record and discussion o f the Cunene-Zambezi Expedition that took place in Angola from 1899 to 1900 is provided. Although the main aim o f the expedition w as to evaluate the economic potential o f southern Angola, it also resulted in sig­ nificant botanical collections made by Hugo Baum (1867—1950), many of w hich serv e as holotypes o f names o f plants from the region, and beyond. The itinerary' is supplemented by a map illustrating the route followed by the expedition, and locality names are clarified and updated. A full list o f the type specimens o f flow ering plants is provided, with their nomenclature updated and an indication o f w here known duplicates are housed.


INTRODUCTION
The collector Hugo Baum (1867-1950) participated in the Cunene-Zambezi Expedition, which took place in Angola from 1899 to 1900.led by the Dutchman.Pieter van der Kellen.In addition to botanical and zoological collections, ethnological observations were also recorded.All ot these were described in German by Warburg (1903) in the now very rare book Kunene-Sambesi Expedition.H. Baum, which can only be found in highly specialized libraries or in the collections of rare book antiquarians.The book includes a folded map (Figure 1) that illustrates the course of the expedition, w hich left Mo^ámedes (Namibe) on 11 August 1899 in an easterly direction, through the actual provinces of ( unene and Cuando-Cubango.reaching the Cuando River in March 1900.After turning around, the expedition finally arrived back at Mo^ámedes on 26 June 1900.more than 10 months later. The expedition resulted in over 1 000 plant collec tions, many o f w hich are types.In Warburg (1903), 281 taxa of flowering plants new to science are described based on these collections, w ith 310 numbers referred to as types.To these must be added an unknown number o f taxa that w ere published later.In the Aluka database o f types o f African plant names (Aluka 2008.accessed in January 2008).there are 1 209 images o f specimens collected by Baum.This figure includes duplicates o f the same collection number.In fact, the Baum collections prov ided a series o f duplicates distributed among Euro pean herbaria.This is fortunate because the holotypes ot the taxa published in Warburg (1903) were deposited in the Herbarium o f Berlin (B) and many o f these speci mens were destroyed in a fire during the Second World War.For that reason the Aluka database presently hosts only 48 images o f Baum specimens from B. Duplicates have been recorded in various herbaria, namely B. BM, BR, COL E. G, K, M, NY.P. S. W and Z (Holmgren, et al. 1990), which include several herbaria that were not listed as depositories o f Baum collections in Lanjouw & Stafleu (1954).However, for some taxa.the holotype destroyed at B w as apparently unique.
In terms o f type material the Baum collection is one o f the most important among those that originated in Angola, and is only surpassed by those o f F. Welwitsch and J. Gossweiler.In addition to their type status, they prov ide information on type locality.This is v ery impor tant in the case o f rare plants, particularly those that were collected only once.Information on the exact collecting localities helps researchers find rare plants in the field, w hich in turn facilitates the ev entual proposal o f specific conserv ation measures.In the case o f the Baum collec tion.pinpointing the collecting localities is often difficult for several reasons: 1, localities recorded on the labels are often not consistent w ith duplicates o f the same col lection.or localities given in the book: 2. the changes of geographical place names that took place in Angola over the last century: and 3. the original spelling o f the locali ties that does not match the more recent spelling used in maps and gazetteers.To facilitate the work o f taxonomists and systematists studying the flora o f southern Africa, the Baum col lections and expedition itinerary were analysed.In the present work we include an historical overview o f the expedition and an analysis o f the flowering plant type material cited in the hook (Warburg 1903).Based on the original map. a new map has been produced, which is complemented by a list o f old locality names, as cited in Warburg (1903), and the corresponding current names and, whenever possible, geographical co-ordinates.

MATERIAL AND METHODS
The information on type collections was obtained from several sources.Initially, the information on types cited in Warburg (1903) was databased.using Figueiredo & Smith (2008) to update the nomenclature.The loca tion of duplicates was determined mostly using internet resources.In this regard, the Aluka database (www.aluka.org) was particularly useful.Several herbaria online, such as K (www.kew.org/herbcat),B (www.bgbm.org)and Z (http://www.zuerich-herbarien.unizh.ch/)also provided much information.Duplicates cited in the literature but not seen (either in herbaria or as images) were included in the database.Information from COI was obtained from a card catalogue.The existence of B types was verified using the resources mentioned above and a list o f fami lies including extant collections at B (Hiepko 1987).It was noted that the localities recorded on the labels are often not consistent with duplicates o f the same collec tion.Furthermore, they often do not match exactly the localities given in the book.As a compromise the collect ing localities given in Warburg (1903) are used.
The types are listed in alphabetical order by family, genus and species, in their currently accepted taxonomy (Figueiredo & Smith 2008).An index to all the scien tific names as well as a sequential index o f type numbers referring to their name in this list, is also presented.
A list ot several geographical names, cited in Warburg (1903) as collecting localities, as well as those shown on the map.was compiled.The location o f the names was determined using maps and gazetteers.An annotated copy o f the Warburg (1903) map.produced at LI SC. was examined.A new map (Figure 2) showing the itinerary and updated locality names was produced based on the information compiled.

HISTORICAL OVERVIEW AND ITINERARY OF THE EXPEDITION
Hugo Baum was bom on 17 January 1867 in Forst Niederlehmsitz.the son o f a railwayman.He was raised in Guben on the river Neisse (then in Brandenburg.Ger many; now Polish and called Gubin) in a forested area where he developed an interest in plants and insects.After concluding his school education in Guben, he started his professional training as a gardener in Nettkow (Silesia: then German now Polish) and later attended a two-year study course in pomology in Proskau (Silesia).Following his compulsory military service in Magde burg, he worked for ten years at the old Botanical Gar den in Berlin.In 1899, thanks to his extensive botanical knowledge, he was invited to participate in the C unene-Zambezi Expedition.& Pax o f which only a few plants were found.At the same location he found for the first time a plant that was referred to as Gardenia thunbergia Thunb.[of which the collected specimen is Gardenia volkensii K.Schum.subsp.spathulifolia (Stapf & Hutch.)Verde.], which was called stumpdorn, the German version o f the Afri kaans stompdoring, a name recorded by Smith (1966), and mulavi by the local Africans.This tree appeared fre quently along the Kunene, Chitanda and Kubango Riv ers, but also along the Longa and Kuito Rivers and grew predominantly in the low grounds o f the rivers or on the river islands.
The expedition arrived at Ediva on 3 September and there they met the leader Van der Kellen and two Boers and two Africans with horses, who had been hired by the Dutchman at Humpata as hunters to provide game meat for the expedition.Van der Kellen's brother Emil man aged a large farm in Ediva.It was owned by the Compagnie de Mo^ámedes, and cultivated various food and cash crops, and carried between 800 and 1 000 head o f cattle.
After Ediva, the first place reached by the expedition on their way to Humbe was called Otjikutsie.Humbe used to have a trading post for cattle.However, it had lost importance since the outbreak o f the cattle plague.The local population in Humbe cultivated sorghum, pennisetum, groundnuts, groundpeas and to a lesser extent beans, locally called macundi.From Humbe the expedition trav elled on the same day to the Kunene River.On 12 Septem ber they crossed the Kunene w ith the support of the people o f Humbe, o f whom some 40 men had joined the expedi tion as porters.On the left bank of the Kunene the expedi tion headed upstream to the mouth of the Chitanda River.
At the Chitanda, the houtbosch, Berlinia baumii Harms [= Julbernardia paniculata (Benth.)Troupin], w hich they considered the most common tree in southern Angola, was noticed for the first time.The local population used the bark of this tree for the production o f boats.In Goudkopje, on the banks o f the Chitanda, a place also locally called Ompopo, the expedition spent several days searching for alluvial gold in the river; they did find a small quantity.In early October the expedition reached Kassinga.also on the banks of the Chitanda.From there, they w ent on to the Nambali stream which flows into the Kubango River.
On 8 October the explorers reached the Kubango at Linghonung, the first village they came across since Kassinga.Following the Kubango, on 15 October the expedition came to the village o f Chirumbo, located on a river island to protect the population against assaults by the Kuinghama tribe (Kwanhama?)from Ovamboland.On 16 October they crossed the Kubango for the first time beween the Kutsi River and Massaca.a fortified village inhabited by some 500 people.From Massaca the explorers travelled the Kubango downstream to a village called Kanjundu, where they left the Kubango to proceed to the Kuebe River, which they followed down stream to its confluence with the Kubango.Then they continued along the Kubango to the mouth o f the Quatiri River.Dow nstream from the mouth o f the Quatiri, the expedition passed through the village o f Kavanga on the Kubango that was inhabited by the Kangella tribe.From there, the explorers followed the Kubango downstream along the border with German Southwest Africa (pres ently Namibia) to Imbala, the principal village o f the Kuangaris.On 27 November, the expedition left Imbala accompanied by a dozen Kuangari guides follow ing the Kubango upstream to the confluence with the Habungu and then along this stream further upstream.
From the Habungu the journey continued to the Bondo, a small stream that contained little water even during the rainy season, which they followed for only a small distance.From there the expedition went to the lake-like pan called Ungombekike from where some of the Kuangari guides returned home.After passing through two other pans the group reached the Kuito River.Mov ing upstream along the Kuito.after a few days the group reached the mouth o f the stream Jonkoa and from there.
the nearby mouth o f the Longa.On 26 December, the explorers arrived at the village Napalanka.situated in a valley between the Longa and inaccessible swamps.After following the Longa upstream for four days, the expedition arrived at a place called Minnesera.which is situated in the sw ampy area o f the mouth o f the Quiriri.
On 16 January 1900, the group crossed the Longa, o f which the banks w ere so swampy that 80 oxen were needed to pull the wagons out o f the swamp.Down stream from the mouth o f the Quiriri the expedition erected a camp where they w ere visited by the chief o f Minnesera and 30 o f his people.A few kilometres dow n stream from Minnesera.the Longa was crossed again with a float.On 31 January the expedition reached the same location at the Longa w here they had been on the 12th.Due to a lack o f food supplies for the expedition, following the local population's refusal to barter for food supplies, sweet potatoes were simply taken in sacks from nearby fields to feed the carriers and the horses.The shortage o f maize rations had weakened the horses, which forced the expedition to reduce the hunting par ties.w hich, in turn, resulted in a lack o f game meat.
Subsequently the expedition reached a ford at the Longa, which rubber dealers used as passage-way.At that place, several residing dealers exchanged root rubber.Carpodinus chylorrhiza K.Schum.[= Landolphia thollonii Dewevre] for commodities.Two-wheeled carriages were also sent to collect the rubber in the surrounding region.The dealers kept considerable rubber stocks tied up in bun dles of 40 finger-strong rubber strips (see below), which as an exchange unit were called manga.The weight of one manga was about one kilogram.The expedition man aged to purchase elev en sacks of maize from a local Portu guese man.That became their only foodstuff and had to be rationed.Due to the food shortage, 17 of the porters were sent back to Humbe and Ediva.
One of the ox-wagons left the Longa for the Quiriri.while the two others remained at the Longa to trade.The expedition avoided contact w ith the Portuguese traders at the Quiriri.because '...the Portuguese in the colonies were often deported persons...' (Warburg 1903;93).However, they associated with the Swedish trader Swanstrom who sent them freshly baked bread.Later the two ox-wag ons that were left at the Longa rejoined the expedition at the Quiriri.where in a few days.300 mangos would be exchanged for cotton cloth.From the Quiriri dow nstream the inhabitants of the many villages were busy w ith the preparation of rubber.In a village at the Quiriri called Sakkemecho. the expedition exchanged consumer goods and cattle for 297 mangos on credit that would be collected after a few months during the return journey.
In the rubber-producing area, rubber completely replaced money as local currency.Root rubber w as pre dominantly found on large grass fields on sandy soils, called chanas (sannas).which were surrounded by for ests.The plant did not grow near the streams and rivers, but on the waterless chanas between the streams.For merly, the root rubber plant had occurred commonly at the Kutsi and the Kubango Rivers at Massaca.where it had now become completely extinct.How ev er, the plant was said to exist in the territory o f the Kuinghama tribe and occurred abov e the mouth o f the Luassingua.reach ing the Longa, Quiriri.Kampuluve and Kuito Rivers.Its range went beyond the latter and extended even as far as the Cuando.After having been dug out with short hoes, the roots, tied up into stacks o f ± 2 m tall, were watered regularly to soften the bark so that it later peeled off easily.Thereafter, the bales were again dried in the sun and the single roots were cut into 30 to 40 cm long pieces that were beaten on a board to remove the bark from the wood.The pieces o f bark were then worked with a wooden hammer until they took the form o f a cake.This cake was boiled and again beaten with clubs.Finally the cake was cut into square pieces, which were placed in boiling water and.while still hot, were shaped into cylindrical, finger-to thumb-thick strips, about the length o f a span o f the hand (± 22 cm) called matali, of which 40 pieces constituted one manga.There were a total o f eleven traders in the entire rubber area, o f whom nine were Portuguese, one was an Englishman and one a Swede.
The expedition continued the journey from Sakkemecho, crossing a waterless and hilly area between the Quiriri stream and the Kampuluve, which ran into the Kuito River.The expedition reached the Kampuluve at its source in a swampy valley that extended for 25 to 30 km up to the mouth o f the stream.They arrived there on 27 February 1900 and reached the Kuito River for the second time.At that point the Kuito was at most 150 m wide, but with very deep and fast-flowing water.On 14 March, the expedition crossed the Kuito with canoes and on the same day reached a stream called Chinpulu on foot.Due to the lack o f grass for the oxen at the Kuito the wagons and oxen had been sent back with two people to the Onschingue River where the others were expected to arrive at the end o f March.From the Chin pulu River the men reached a small stream called Miané that they followed up to its source.Then the march con tinued through waterless hills to the Kuisi stream in the Zambezi area.The expedition followed the Kuisi down stream.then crossed it and after several hours reached the village Konjongo at the Konjumbu.a small stream.
By the time the expedition reached Konjongo, Baum was weakened by fever and stomach pains and he decided to return with two porters to the Kuito River, which they crossed to the right bank.From there the group went to the Onschingue where on 1 April they met part o f the expedition that had remained at the Cuando.On 4 April, the expedition initiated the return journey along the previous route, upstream along the Kuito and Kampuluve Rivers up to the Sakkemecho village, where they delivered to the local inhabitants part o f the rubber they owed them from the previous transaction.
From the Sakkemecho village, the men followed the Quiriri upstream and then crossed the sandy hills to the Longa.From the Longa the expedition travelled to the Lazingua.and then westwards passing the Quatiri to the Bundju stream, where they remained for some time to allow the oxen some rest.The journey continued to the Kuebe and the Kulei, a tributary o f the Kutsi.The pas sage across the Kutsi.which was rather deep, took three days.From there the journey went on to the Kubango.The crossing o f the Kubango with more than a hundred head o f cattle and the wagons that had to be dissembled to be ferried across the river, took another five days.The group then followed the Cubango upstream and turned westward behind Kohi.
On 19 May, the expedition arrived at the Chitanga River and shortly afterwards at nearby Kassinga.After a half-day stay in Kassinga, the expedition followed the Chitanga downstream to Goudkopje.The Chitanga area was rich in game.On 29 May the expedition reached the Kunene at Kiteve.The local population provided eight large boats to help the expedition cross the Kunene within two days.From Kiteve the expedition followed the Kunene downstream.On 5 June the men arrived at Humbe where the porters from that place left the expe dition after having been paid for their services.For the remaining journey to the coast the expedition recruited other local porters.On their way to Ediva at the Kaculovar River that was reached after two and a halfdays, they met the Dutchman Emil van der Kellen who had come by horse from Ediva when he had heard about the expedition's return.
At a place called Kahama. the expedition crossed the Kaculovar a second time and then followed it upstream.In the clay soils along the Kaculovar, Baum again fre quently found Acacia kirkii Oliv., which was said to produce the best gum.Within two days the expedition crossed the passage through the Chela Mountains.After leaving the mountains, the men went along a stream called Tschikutt passing consecutively through Pietfontain [Pietfontein in modem Afrikaans], the dry Olefantsrivier [Olifantsrivier], the Bumbo and the Kitibe.At Monino v illage they had a longer rest before they walked the long and dry way to the Giraul River.From there the expedition travelled the last part to Mov'ámedes, where they arrived on 26 June 1900.
During the ten-month journey, members o f the expedi tion killed 175 antelopes, 3 hippos, 2 boars, and 3 croco diles.Baum collected a total o f 1 016 plant collection numbers, o f which more than 300 were later described as new taxa.More than 60 o f the new species were named after Baum.Besides this haul.Baum caught more than 300 butterflies, prepared them and sent them to Europe.With regard to the economic importance o f southern Angola.Baum recommended the areas close to the coast at the small rivers Bero, Giraul, Monino, and Coroca for the establishment o f plantations, particularly for the cultivation o f cotton.In his opinion, the high plateau, predominantly constituted o f heavy loamy or clay soils was suitable for the cultivation o f maize, wheat, rye, oats, barley, and European vegetables, subtropical fruits such as citrus, and grapev ines.According to Baum, there was much fertile land for colonization at the Cuando.The part o f this area east o f the Cuebe and north o f the Lomba was characterized by a wealth o f root rub ber plants.Fie considered the region particularly rich in trees providing tanning agents, such as the mopane, Colophospermum mopane (J.Kirk ex Benth.)J.Leonard, and the houthosch.Julbemardia paniculata (Benth.) Troupin.The distribution area o f the acacia that provided the best gum was located in the region o f the C'aculuvar and Kunene.Thanks to its large forests, southern Angola was considered very rich in timber; however, the cost o f transport to the coast was very high.Although gold was panned in the Chitanda River at Goudkopje beneath Cassinga, the main value o f southern Angola was that it was very suitable for livestock.In Baum's opinion the transport o f cattle to the coast was easy, given the availability o f water and pastures along the whole way.Finally, he emphasized the abundance o f fish resources at Mo^ámedes (Namibe), Tiger Bay, and particularly Porto Alexandre (Tombua), where salterns that produced salt for the preservation o f fish were located.

LIST OF BAUM TYPE COLLECTIONS GIVEN IN WARBURG (1903)
All the holotypes (or syntypes) o f the names published in Warburg (1903) were deposited at B. Information is available in the literature and on the internet concern ing the families which survived the fire that consumed the herbarium during the Second World War (see www. bgbm.org).At the time o f the bombardment, the staff of the herbarium was undertaking the task o f transferring the type specimens to a safer location.Types belong ing to certain families are therefore still extant at Berlin.Likewise, material on loan to other herbaria obviously also survived.The remaining material has been reported as lost.In the case o f the Baum collections, out o f the 311 type collection numbers o f flowering plants cited in Warburg (1903), only 40 holotypes and four syntypes are confirmed to exist.There are 181 holotypes reported as destroyed, and 34 belonging to families for which type material apparently survived, could not be located in any o f the databases cited above.In rare cases, types that were reported as destroyed in the literature were found to be extant as they are available in the Aluka database.This is the case for the holotype o f Lobelia minutidentata Engl.& Gilg.that had been reported as destroyed (Thulin 1984).This was probably due to the misplace ment o f the specimen as the locality data prov ided to Aluka wrongly assigns it to Namibia.It was noticed that several Baum collections have been attributed to the wrong country, and depending on the arrangement of herbaria, this may be a reason for the material not having been found in the past.For instance.Baum 894 (Cephalaria retroseta Engl.& Gilg), is represented in Aluka (accessed in February 2008) by three duplicates: the K duplicate is attributed to South Africa, the BM duplicate to the Democratic Republic o f Congo and only the dupli cate in E is correctly assigned to Angola.In other cases it was found that types belonging to families that were reported as destroyed still exist.For example, in the Rubiaceae.at least four Baum types are extant.

FIGURE
FIGURE I.-Original map o f expe dition itinerary divided into three sections (Warburg 1903).

FIGURE 2 .
FIGURE 2.-Actual map o f expedition itinerary showing route taken by ox-wagons (continuous line), routes travelled on foot (dotted lines), anddates when some localities were reached (in italics for journey out, underlined for return journey).
of the interest o f German Southwest Africa (now Namibia) in the planned train connection between Porto Alexandre (presently Tombua) in Angola and the then independent Zuid-Afrikaanse Republiek.later the Transvaal province o f South Africa.At the time, there was a lack o f knowl edge o f w ild-growing cash crops and local African cul tures in the area.In his expedition report Hugo Baum did not only describe the flora, but also provided informa tion on the geography, geology, climate, the local people encountered on the expedition, and some o f their cus toms.The leader o f the expedition was the Dutchman, Pieter van der Kellen.who was the representative o f the Compagnie de Mogámedes in Ediva.On 11 August 1899, the expedition left Mo^ámedes (presently Namibe) with three ox-wagons for the rivers Kunene, Kubango, Kuito, up to the Cuando.and on 26 June 1900 it returned to the place from which it departed (Figures 1. 2; refer to Appendix 1 for actual name and co-ordinates o f localities, which are mentioned in this text with the original spelling).At the time Mo^ámedes was a town o f 1 500 inhabitants with a local economy dominated by the trade in dried fish.The most common acacia species found between Ediva and Humbe was Acacia verrucifera Harms [= Acacia kirkii Oliv.subsp.kirkii var.kirkii] that provided the best transparent gum (gummi arabica) o f great bond ing strength.The distribution area o f this acacia extended from the Chela Mountains to the Kunene River, and from the Chitanda and the Kubango Rivers downwards to the territory o f the Kuangari tribe.Near Chihinde, Baum collected Strophanthus ambaensis (Schinz) Engl.