Encyclopédie internationale
des histoires de l’anthropologie

Adding Colour to Und Afrika Sprach: Carl Arriens’ Image and Leo Frobenius’ Text

Richard Fardon

SOAS University of London

Richard Kuba

Frobenius Institute for Research in Cultural Anthropology, Frankfurt (Frobenius-Institut für Kulturanthropologische Forschung an der Goethe-Universität Frankfurt)

Mise en couleurs par Agnès Boulmer

Référence complète

Fardon, Richard & Richard Kuba, 2021. “Adding Colour to Und Afrika Sprach: Carl Arriens’ Image and Leo Frobenius’ Text” (Mise en couleurs par Agnès Boulmer), in Bérose - Encyclopédie internationale des histoires de l’anthropologie, Paris.

Télécharger en pdf

Publié dans le cadre du thème de recherche « Histoire de l’anthropologie et des ethnologies allemandes et autrichiennes », dirigé par Laurent Dedryvère (EILA, Université de Paris, site Paris-Diderot), Jean-Louis Georget (Sorbonne Nouvelle, Paris), Hélène Ivanoff (Institut Frobenius, recherches en anthropologie culturelle, Francfort-sur-le-Main), Isabelle Kalinowski (CNRS,Laboratoire Pays germaniques UMR 8547, Ecole Normale Supérieure, Paris) Richard Kuba (Institut Frobenius, recherches en anthropologie culturelle, Francfort-sur-le-Main) et Céline Trautmann-Waller (Université Sorbonne nouvelle-Paris 3/IUF).
Dossier “Leo Frobenius” coordonné par Hélène Ivanoff et Richard Kuba

‘The Holy Buffalo Mask and Lera Dance of the Daka’

Der heilige Büffelmasken und Laerentanz der Dakka’ by the artist Carl Arriens seems to record a performance witnessed by members of the expedition led by Leo Frobenius that traversed Nigeria and North Kamerun between 1910-12. This essay first shows why this is not the case : in short because the event depicted is an impossibility ; and then why the image features in a work of ethnology despite illustrating an event that never occurred. We ask more generally what the discrepancies around an image tell us about the patchwork composition of the extensive published record of Frobenius’ Africa expeditions.

Figure 1
Der heilige Büffelmasken- und Laerentanz der Dakka
From Und Afrika Sprach, Volume III, facing 256

The depiction around which our account circles appears in the third, so far untranslated, volume of Frobenius’ Und Afrika Sprach (1913a) in its scholarly or scientific edition (wissenschaftliche Ausgabe). The work also had a single-volume popular edition (volkstümliche Ausgabe) in German (1912a) which was quickly translated into English in two volumes as The Voice of Africa (1913b). With one instructive exception, the main substance of the third volume of the scholarly edition is not included in the popular edition. That exception concerns the last chapter of the popular edition which is identical to the first chapter of the third volume of the scholarly edition. This has significance for our later analysis, but it does not affect our main image which appears only in the third volume of Und Afrika Sprach.

The published plate of what we shall call for short ‘The holy buffalo mask’ is a halftone (Stulik and Kaplan 2013) ; it was produced by a photomechanical process which was popular at the time, or more precisely by two mechanical processes and one human : the first stage transforms an original image into series of dots, the second prints the resultant image ; in the third step, the human eye sees the dots as smooth continuous tones. Just how smooth depends on the resolution of the print : the graininess of magnified images from old newsprint is familiar, an effect exploited in colour and by intention, for instance, in Roy Lichenstein’s graphic, cartoon-strip images. Most halftone images were produced from photographs, usually monochrome. Arriens’ original is currently lost. [1] Not knowing its original medium, we believed it had been in colour, painted in oils, watercolour or gouache. Late in the day, consulting an old register of the Frobenius Institute’s ethnographic picture archive, it transpired that the image had been recorded as ‘Tu’, a contraction of Tusche, to identify it as a drawing in Indian ink. So, the lost original had been monochrome all along, and our efforts to discover its original colours had been an inviting red herring. Once our initial consternation had passed, we decided that retaining this latest invention might be an appropriate addition to what turns out to have been a sequence of them. The colourization concludes this essay. Other archived works by Arriens help us to show the variously transformed relations to his lived experience which his images underwent in different mediums. As expedition artist, Arriens highlights the imaginative endeavour of the entire output from Frobenius’ expeditions, but he does not have a monopoly on it. The combinatorial strategy behind Arriens’ image of ‘The holy buffalo mask’ turns out also to be an image of the combinatorial collage characteristic of Frobenius’ narrative. To discuss this further we shall need some context.

What Leo Frobenius later considered to have been his fourth Africa ‘expedition’ had traversed Nigeria to enter Kamerun late in the wet season of 1911. [2] Its European party, made up, in addition to Frobenius himself, of our artist Carl Arriens (Pfisterer 2019) and the young engineer-surveyor Albrecht Martius (Kuba 2019), had arrived in Lagos in mid-October 1910. Fourteen months later, in December 1911, shortly after the events that concern us here, and as the party became increasingly unwell, Leo Frobenius ended his own involvement in Nigeria to continue his research in Kordofan in central Sudan from where he returned to Europe in May 1912 (Zwernemann 1987 : 120). [3] Unlike Frobenius’ previous extended African expeditions ̶ to Central Africa (the Congo) and through Sahelian West Africa (contemporary Mali, Guinea, Burkina Faso and Togo) – no chronological travelogue of the fourth Africa expedition was published (Kuba 2010). We are, however, able to augment very brief published accounts of the later stages of the fourth expedition, offered by Frobenius (1912a/1913a/1913b) and fifteen years later by Arriens (1928), with reference to the diaries, letters and dated sketches that survive in the archive of the Frobenius Institute, themselves only a fraction of the material that the expedition must originally have produced.

Despite allegations of underhand dealings and theft from Yoruba in Ibadan and Ife in 1910 that became so notorious as later to be cited by Wole Soyinka on accepting the Nobel Prize for Literature (Ita 1972 : 681-6 ; Penny 2002 : 116-122 ; Platte 2010), subsequent relations on the ground with the British were cordial. As they headed eastwards, Frobenius’ party, at times as many as a hundred strong under the direction of Frobenius’ assistant, known to them as Bida, [4] also the name of the Nupe capital, relied on logistical support from the British, who in their turn could not have controlled what since 1900 had been the Northern Nigeria Protectorate (not to be consolidated into Nigeria until 1914) other than through the assistance of its erstwhile ‘Hausa-Fulani’ rulers, who governed the constituent emirates of the Sokoto Caliphate. At the eastern edge of the Protectorate, where they had hardly mingled with Hausa, on the bank of the River Benue, the Fulani had established the last capital of the Adamawa Emirate at Yola, which was stormed by troops under British command in 1901. When Frobenius crossed eastward into German Kamerun in 1911, he became dependent on the Fulani of Garoua (‘Garua’ in German texts), who had been nominal subjects of Yola until these places were separated by the colonial border. The expedition subsequently spent over two months in Fulani-ruled lamidates (a European coinage for the subordinate chiefdoms of an emirate ruled by a lamido) which had been installed south of the River Benue on the western bank of the Faro-Deo river valley, between those rivers and the Alantika Mountains. Rather than the Fulani, [5] Frobenius’ interest was attracted by the people over whom they ruled on behalf of the European powers, peoples he termed (in English translation) ‘disruptive’, an idea akin to what later anthropologists would call segmentary, acephalous or at the least decentralized. Some of them were organized into small chieftaincies, others were uncentralized. From north to south on the river system, these peoples included the Bata, Verre, Chamba (Daka to the west and Leko to the east), and the Koma. The variable degree of their subordination to the Fulani before the twentieth century is too complex a topic to broach here ; suffice to say that early colonialism reinforced the proxy colonialism of the bearers of Fulani ethnic identity, and this in turn both positioned the Fulani to act as intermediaries between the expedition and the local populations and, in doing so, at the same time made their mediation suspect. Frobenius grasped this circumstance practically and, to encourage local peoples to visit his camp as informants, he sent ahead the misinformation that his expedition intended to address Fulani abuses of power. Collection of objects was particularly intense among the Verre who, as specialist metal workers, would probably also have made some of the pieces collected from the Chamba, the subjects of Arriens’ image and our concern here. The Chamba provided Frobenius sufficient materials for two full chapters of his ethnographic writings ; while the Koma and their dwellings were the subjects of several of Arriens’ larger surviving water colours. Triangulating the published materials with the sketches and diaries in the archive of the Frobenius Institute, we can fix the expedition’s presence in particular places with some degree of precision. Interpreting these movements through their traces is key to understanding the local experience that went into the halftone reproduction of Arriens’ image as ‘Der heilige Büffelmasken und Laerentanz der Dakka’.

Figure 2
Map of the Benue, Faro and Deo River confluences

The Daka (‘Dakka’) of the image’s title are western Chamba, nowadays usually distinguished as Chamba Daka from the Chamba Leko living to their east. These Chamba neighbours share an ethnic identity in larger contexts despite speaking distinct and mutually incomprehensible languages. One of the contrasts between Daka and Leko is linguistic, but the same terms are used to describe geography by contrasting west and east (mostly bounded by the Shebshi and Alantika Mountains), and history (some Daka are recalled to be Leko by origin and vice versa), and culture (some Daka speakers strike their western neighbours as Leko on cultural grounds despite sharing a Daka language ; and some Leko speakers are described as Daka by more easterly Leko). The same terms are serviceable in a number of ways, and the parties to a conversation need to share an understanding of context to pick up quite what is being conveyed when they are used. Frobenius grasps a substantial part of the significance of all this. Although he devotes distinct chapters in his works (1913a, 1925 to the Chamba Leko (‘Tschamba’) and the Chamba Daka (‘Dakka’)), he recognizes that they are closely related. His chapter on the Daka is informed particularly by visits paid to him by the chief of Kiri (Kirri), living to the west by the southern Shebshi Mountains, and the chief Yeli or Dayela (‘Jelu’ or ‘Dajella’) living in the Alantika Mountains (Frobenius 1925 : 164-69 ; Frobenius 1987 : 10, translated by Mohammadou). Although Leko-speaking, Yeli is considered the place of origin for many of the Chamba Daka chiefdoms to its west. And the people of these chiefdoms prefer to identify as Nnakenyare rather than Daka. Frobenius records this term (‘Nagajare’), although he apparently does not realize that the label derives simply from a conventional greeting on the lines of the English, ‘How are you doing ?’. Ethnic labelling is highly contextual in ways we cannot pursue further here (Fardon 1988), and we remark it only because Arriens’ image, although labelled ‘Dakka’, has elements that we shall see clearly derive from observations made among Chamba Leko. At the least, he has combined observations from Chamba speakers of entirely different languages.

The expedition in Adamawa

An undated sketch by Carl Arriens shows the expedition making a river crossing somewhere between Garoua and Tchamba (KBA 11821) [6] which may well be the occasion Frobenius describes in the late afternoon of 28 August 1911 when they encountered the swollen River Faro at the height of the rains. Receiving no help from the Galadima who governed Tchamba, Frobenius himself crossed unassisted, occupied the official German compound, and let it be known that he would summon a nearby rival to assume the Galadima’s office. At this threat, the Galadima roused himself and the following day passage of the main party was achieved (Frobenius 1913b : 667-68). That same day, the industrious Arriens found time to paint two landscapes of the Alantika Mountains seen from a distance, one of them dated 29 August 1911 (EBA-B 02715/6). According to Martius’ diaries, [7] a sortie from Chamba to visit the Koma atop the Alantika Mountains at Ndera, described in detail by Frobenius, took place on 1 September (Notizbuch IV - LF 250 ; Frobenius 1913b : 670 et seq). When, after a few nights, Frobenius and Martius descended from the mountain, Arriens was left at this higher altitude to recover from repeated bouts of malaria (1913b : 676), hence the comparatively large number of images depicting the mountains and their inhabitants, including the arcadian scene remarked by Kuba (2020 : 6, EBA-B 00486 ; also Hambolu 2010 : 41), which is the source of another halftone plate in Und Afrika Sprach (see Figure 27b). [8] The expedition remained in Tchamba for over a month where, repenting his initial reluctance, they were assisted by the Fulani Galadima. (Galadima is usually the title for an administrator working for a Lamido, but here, because the Lamido had moved south to Tibati in the mid-nineteenth century, he had effectively become the governor himself.) Only when he was prepared to continue southwards did Frobenius send for ‘Arriens to come down from his sanatorium’. Consistent with this, Arriens’ pencil drawing of the expedition’s camp in Tchamba is dated 28 September 1911(EBA-B 02702). An undated oil painting by Arriens depicting the Galadima’s compound was probably painted around the same time (EBA-B 00487 ; see Figure 3).

Figure 3
Carl Arriens, ‘Compound of the Galadima’ (oil, 53 x 34 cm, EBA-B 00487)

A mock military enactment, durbar-style, took place in front of an identical baobab tree and entrance hut (Figure 3 1912a : facing page 518, 1913b : facing page 526, with caption translation below, based on a water colour that is currently lost), [9] and this event may have occurred around the occasion of the ‘gorgeous feast’ Frobenius reports the Galadima put on before their departure from Tchamba (1913b : 677).

Figure 4
Carl Arriens, ‘Horse games before the Galadima’s compound in Tshamba’
From Voice of Africa Volume II, facing 526

On 2 October 1911, the expedition departed two short distances southwards to spend 3-4 October in Yelba (‘Jelba’), and 5-6 October in Laro, reaching the larger lamidate of Koncha (‘Konscha’) on 7 October (Martius Notizbuch V - LF 251, ‘Start of September – 7 October, Tchamba – Konscha’). Logistics suggest the main expedition could stay for longer periods only in the larger settlements, like Tchamba and Koncha, that were able to provision them. This sequence, as well as his own later account, make it more than likely that Arriens incorrectly dated his sketch of a cult performance in Laro to 6 September (Figure 5, KBA 10931) when it should have been 6 October.

Figure 5
Carl Arriens, ‘Cult performance in Laro’
Pencil and watercolour, KBA 10931

There is a gap in Martius’s surviving diaries (between 7-25 October), but the next extant diary (Martius Notizbuch VILF 252 ‘25 October – 12 December, Konscha – Lokodja’) suggests that the party remained in Koncha throughout October and into early November. Frobenius gives a glowing report of a few weeks spent there (1913b : 677). Arriens dated a painting of a zebu cattle head towards the start of the month as 8 October (EBA-B 02714) and the portrait of a Bamum man as the 5 November 1911 (EBA-B 00547), noting Koncha as location in both cases.

Supposing Arriens’ sketch of a gourd horn player, drummer and two ritual practitioners in Laro was correctly to be dated to 6 October, then it would be separated by a couple of days from his sketch of two flute players and a drummer in Yelba, supposing that was executed when the expedition stayed there between 3-4 October (Figure 6, KBA 10932).

Figure 6
Carl Arriens, Schamba in Jelba
Pencil and watercolour, KBA 10932

Even if we cannot be confident of the exact dates, we do know that Arriens’ images of these two Chamba performative genres were made in different places. In a later memoir (1928 : 91), Arriens describes the cult performance as a ‘hocus pocus’ he had seen briefly in Laro : a number of men dancing in a circle, one with an iron rattle and others playing metre-long trumpets made from sections of calabash or gourd. Outside the circle danced a masquerade with a gruesome (greulich), heavy wooden head resembling that of the head of a cow or buffalo, and a formless body of rushes. The surviving sketch inspired by witnessing this scene does not show the masquerade (unless the pencilled doodle below the sickle at the top right represents it), and this may go some way to explain the inaccuracy of the masquerade he portrayed in the work we are analysing that was reproduced in halftone in Und Afrika Sprach.


Examining the entirety of Arriens’ work in Frobenius’ service during the 1910-12 expedition, the variety of media and genres in which he worked is striking. While we do not know in every case who took the field photographs surviving from the expedition, and those published are credited most often to Frobenius himself, or occasionally to Martius, fifty-seven photographs still in the archive are currently attributed to Arriens. So far as we know, Frobenius sent all exposed photographic plates (13 x 18cm) back to Berlin. [10] In return he received prints even in the Nigerian hinterland which allowed him to check that his camera was working properly. [11] We cannot say when, or even if, Arriens had access to photographic prints during the expedition. Given Frobenius’ views, he may not have seen them in the field at all ; in one of his letters, Frobenius instructed that ‘my painters should not see the photos, so that their works remain as original and free of the lens as possible.’ [12]

As is already evident from his notes on the horn and flute bands, Arriens’ surviving sketchbook pages contain rapid indications of shape, colour, shadowing and dynamics, that vary in their degree of schematism. His scrawled annotations are difficult to decipher, and we have succeeded in doing so only in part. [13]

Figures 7a and 7b
Figure 7 a and 7b : Arriens’ sketchbook annotations

The freedom of these impressionistic sketches contrasts with the meticulous penmanship of the detailed monochrome sketches documenting individual objects which find their way onto accession records and may well have been composed after return to Germany by Arriens. These are illustrations of the objects destined either for the major German ethnological museums which had sponsored the expedition (notably Hamburg, Leipzig and Berlin) or else for sale via commercial dealerships, which in many cases was simply an alternative route to the German museums. The illustrator’s attention to detail draws out features easily missed even in contemporary digital photographs as, for instance, in those objects that also appear in our halftone, or are related to those that do so. Several are attributed to neighbouring Verre rather than Chamba, though this does not preclude their being in use by Chamba.

Figures 8a, 8b, and 8c
Types of ritual ‘sickle’, ‘lama’
KBA 10935, KBA 10933, KBA 10934
Figures 9 and 10
Carl Arriens, Leopard’s paw as a pouch for circumcision knives, Kontcha, 20.10.1911 ; Horizontal drum.
Fig. 9 : Watercolour, 31x47 cm, EBA-B 00677 ; Fig. 10 : KBA 09920
Figures 11 and 12
Verre iron cult rattle ; Shaker rattle.
Fig. 11 : KBA 09486 ; Fig. 12 : KBA 09435
Figures 13a, 13b, 13c, and 13d
Gourd horn instruments
KBA 09173, KBA 09174, KBA 09175, KBA 09177

Some of Arriens’ colour paintings are similarly meticulous. A Chamba triptych of masquerade, flautist and drummer is an excellent example of this documentary style.

Figure 14
Carl Arriens, ‘The procession of the Nassa mask’
Watercolour, 47x31 cm, EBA-B 00676

Like the accession records that were drawn in the presence of collected objects, it seems that these documentary paintings were composed from photographs, a practice demonstrated in other cases from the expedition records (for instance, the musician playing the kakaki, a 3-4-metre-long metal trumpet, Pfisterer 2019 : 126). In this painted triptych, both the masquerade and the two Chamba musicians are depicted in postures identical to photographs that survive as contemporary prints : the flautist is on the viewer’s left in the photograph (EBA-B 00679a/FoA 04-5874) ; the Chamba masquerade, seen side on, is based on the same photograph ; the drummer appears to have been painted from another photograph (EBA-B 00680b/FoA 04-5872). Note that none of these figures was used to compose our halftone. While we do not know with certainty whether they were painted in the field or after return to Germany, the fact of copying from a photograph, allied to Frobenius’ antipathy towards expedition artists seeing photographs, would argue for the later composition.

Figures 15a a to 15i
Five field photographs (four of these photographs – archived with the prefix EBA – survive as contemporary prints which might have been available to Arriens ; they are slightly cropped compared to recent prints from the same negatives)
Figure : 15a : FoA 04-5874 ; Figure15b : EBA-B 00679a (vintage print, 17x23 cm) ; Figure15c : FoA 04-5875 ; Figure15d : EBA-B 00679a (vintage print, 17x23cm) ; Figure 15e : FoA 04-5876 ; Figure 15f : FoA 04-5873 ; Figure 15g : EBA-B 00680a (vintage print, 17x23 cm) ; Figure 15h : FoA 04-5872 ; Figure 15i : EBA-B 00680b (vintage print, 17x23 cm).

Looking at Arriens’ larger format works, we are struck by their varying degrees of verisimilitude. At one pole, many of the easel paintings from the field, often in watercolour or gouache, are taken directly from life, as a surviving photograph of the artist at work in a Tiv village would suggest (Figure 16).

Figure 16
‘Arriens painting in Salatu’

Another instance would be the watercolour we have seen already (Figure 3, EBA-B-00487, above) which inspired the background to our halftone subject. It shows the compound belonging to the Galadima of Tchamba, shaded by the massive baobab tree that Arriens recalled years after the event (1928 : 89). Baobab trees are typical of villages in this region, and Arriens noted how numerous they were in Koncha (1928 : 91), where they are all that remains to indicate the former extent of what is now a border place of minor importance (compare Arriens’ sketch EBA-B 02701).

The free expression of our halftone occupies an opposite pole to such restrained verisimilitude ; here Arriens allows himself a high degree of creative licence. Other images can be placed somewhere between the two polar genres. Why does Arriens adopt such different genre conventions in works that are contemporaneous ? Or, given that we cannot explore his motives directly, to put the question more practically, when does he do so ? And does answering this question give us any clues as to why ?

‘The holy buffalo mask and lera dance of the Dakka’ is not the only example of this genre in the three volumes of Und Afrika Sprach, but it is the freest of them. A number of plates in the earlier two volumes of Und Afrika Sprach are based on Arriens’ studies from the Yoruba and the Nupe. [14] While atmospheric, compared with the image which is our focus, these are straightforwardly descriptive scenes, particularly in the case of the Nupe series concerned with artisanal production and the market. In all but one instance, the medium is specified as an oil sketch, a study or a painting : the single exception being a drawing. Treatments similar to that given our subject characteristically concern other, in Frobenius’ terms, ‘disruptive’ peoples. Although not published in Und Afrika Sprach (or elsewhere that we have yet traced) a halftone of Tiv musicians and dancers (Figure 17) is the most similarly animated scene of performance saturated with exoticism.

Figure 17
‘Tiv musicians and dancers’
KBA 12006

In a different, and more pastoral vein, Arriens’ scene (Figure 18) of Koma women grinding corn by hand on the granite outcrops of the Alantika Mountains shares in the exotic ambiance (1912a, facing 658 ; 1913b, facing 672, original medium not specified ; reproduced again in Volume III, 1913a, facing 224).

Figure 18
‘Komai women grinding the corn for the evening porridge in the Alantika Mountains’
From The Voice of Africa II, facing page 672

We shall return to this relationship between genre and subject in our conclusion. To what extent these Koma and Tiv scenes depict events Arriens ever witnessed we cannot say ; further research might tell us, but we can be definite about the image which is our main subject.


The traces of ‘Der heilige Büffelmasken und Laerentanz der Dakka’ that we can find in the archive provide sufficient evidence to be certain that Arriens never witnessed the scene he portrayed. What he did instead was to combine in a single image several elements that he saw and recorded over a few days in Tchamba, Yelba and Laro (to which he added some exotic zest). We can put this more strongly. The scene portrayed not only never happened, unless it was forcibly staged it could not have done and that, in short, is because it is a cultural impossibility.

To begin with the caption, ‘holy buffalo mask’ is a mistranslation in several senses. ‘Holy’ is an odd adjective to apply to what Arriens described as a piece of ‘hocus pocus’ designed to gain money, although he also wondered whether the mask might formerly have been used in the secluded bush ceremonies of a secret society. The usual terms for the masquerade in the two Chamba languages reference either its unruliness as a creature of the wild (nam-gbalang, lang-gbadna), or else the resemblance of its central element to a human skull (vara). While it may have the horns of a bushcow, they are just one part of a fused assemblage of human, dead and wild animal characteristics of what is, in short, a theranthrope. [15] With only slight local variations in its design, the masquerade is shared by Chamba Leko and Chamba Daka. The wooden heads of the most westerly Chamba Daka masquerades tend to be larger and flatter than those of the eastern Chamba Leko, and this difference goes along with a greater emphasis on the mask’s theriomorphic characteristics in the west and on its anthropomorphic characteristics in the east. [16] A version of the mask head had already been collected by the expedition from the emigrant Chamba conquest chiefdom of Donga, situated in the plain south of the River Benue, [17] which the party had visited before it reached Adamawa (sketched for accession as KBA 10858, Figure 19 below) ; it is possible that Arriens recalled this example.

Figure 19
Chamba mask from Donga
KBA 10858, now in the Linden-Museum, Stuttgart, 83.091

Even considering that Arriens’ masquerade is captioned as Daka (‘Dakka’) rather than Chamba (‘Tschamba’), its mask head is misshapen, particularly in its pointed horns, which may have been half-remembered from the Donga variant. More damagingly, Arriens has given the Chamba masquerade a body belonging to altogether different masquerades from the Benue region, such as those of the Ankwe or Jukun which we know from his own illustrations (see also Frobenius 1923 : 72-73). The bodies of Chamba masquerades in this region are a single mass of dried fibres, from which it would be improper for an arm to protrude let alone hold a staff.

Figure 20a
Carl Arriens, ‘Masks of the Jukum’, watercolour, 47x31 cm
EBA-B 00662
Figure 20b
Carl Arriens, ‘Ankwe masks’, watercolour, 45x32 cm
EBA-B 00668

This substantial inaccuracy is striking, not least because it differs from Arriens’ watercolour of a Chamba (that is Chamba Leko) masquerade (see Figure 14) which is entirely accurate thanks, as noted already, to being demonstrably based on a photograph and probably painted in Germany. This evidence suggests that the Arriens’ Chamba Daka masquerade known to us from the published halftone predates his watercolour copy of a photographed Chamba Leko masquerade. He may have done the Daka masquerade from memory rather than from a rough sketch (unless in the doodle noted above, no sketch of a Daka masquerade survives in the archive, although this absence of evidence now does not in itself clinch evidence of an earlier absence).

Figure 21
The elements combined in Arriens’ drawing of Der heilige Büffelmasken- und Laerentanz der Dakka, annotated by Agnès Boulmer for colourization
From Und Afrika Sprach Volume III, facing 256

According to his later recollection, Arriens witnessed the masquerade perform together with a band of horn players and a drummer. Three of the figures and the drummer (6, 7, 8, 9) depicted in the published halftone correspond to those in his pencil and watercolour sketch. As in his sketch, the drummer of a horizontal drum (9) in profile faces three figures respectively carrying a dancing sickle and staff (8), an iron rattle, staff and skin bag, apparently containing the wooden lama (7), and a wicker shaker rattle and gourd horn (6). Two of these figures (7, 8) have leopard’s paw pouches hung around their necks, and sport both goatee beards and topknots of hair, the second of these features being absent from the sketch. The horn band has been augmented : the figure to the viewer’s right of the masquerade holds another wicker rattle (10), while more horn players are suggested (for instance the figure at the rear (between 7 and 8).

The first three figures of the flute band are similarly based on the much more schematic pencil and wash sketch Arriens made in Yelba. The postures of the drummer (5) and two flute players are similar (4), as is the detail that the first two of them wear smocks while the third has a bare upper torso. The sketch notes five flute players, although the halftone appears to show seven. Arriens may have interpreted his notes on his own sketch as implying five players additional to the two he drew. Flute ensembles vary regionally between five and six in Chamba communities, but never seven. Like the horn players, the first two flute players are represented with central tufts of hair.

Taken together, our evidence shows how Arriens drew at least upon the two surviving sketches of particular events and combined them when composing the foreground of the image that became our halftone. The backdrop to the whole scene (A & b), as already seen, is the compound belonging to the Galadima of Tchamba, although neither of the events Arriens witnessed occurred there. And he has added other witnesses : a man in Fulani-style robes and turban holding a long sword (1) ; behind him a figure wearing a pith helmet of the kind that Europeans in the expedition, particularly Arriens himself (as in the photograph in a Tiv village, above, Figure 16) and Martius, often wear in photographs (2) ; a third robed and seated figure is suggested behind him. The seated figure in the foreground may well be based on a sketch like that which Arriens made of the mallam of Tchamba (Tschamba) (also noted in 1928 : 89).

Figure 22
‘Mallam of Tchamba’
KBA 11823

Against the compound matting fence, there stands a figure with a wrapper worn over one shoulder and wearing a cap, a kind of dress that Arriens elsewhere depicts as typical of Hausa (EBA-B 00538), another figure perhaps of a child is apparent at his feet. One wonders if this might be Bida.

As an ethnographic portrayal, the single event Arriens depicts not only did not take place but could not have taken place. Flutes cannot perform together with calabash horns, it is not only that they have distinct repertoires and would be cacophonous, but performances with these instruments constitute different types of event. Calabash horns perform in the context of some of the men’s cults, vooma in Chamba Leko or jubi in Chamba Daka ; they should never be seen by uninitiated boys or by women who at the very least have to play up to the supposition that they are ignorant of how these rasping tones are made. Flutes, lera in both Chamba languages, perform publicly at all manner of celebrations, particularly those that accompany funerary rites ; their songs are known generally, particularly by women who are the most important vocal performers. By combining so many incompatible elements, Arriens has unwittingly conveyed insights into the accumulative nature of the expedition itself. Here are the expedition members apparently witnessing performance of the very accoutrements that they are collecting for museums. It seems too good to be true, and it is. The calabash horns, sack rattles of basketry and gourd, iron rattles, the priest’s dancing sickle insigne and the wooden version poking from their leather bags (lama), [18] the leopard’s paw pouch containing the circumcision knives, and so forth were all recorded in the more controlled documentary forms of monochrome ink sketches or, in the case of the leopard’s paw pouch containing circumcision knives, as a watercolour that was also reproduced as a monochrome plate in Und Afrika Sprach. Arriens has combined all these different and incompatible elements in an impossible collage not unlike the staged tribal photographs of his time combining the most exotic dresses and paraphernalia to portray a seemingly ‘typical’ people.

Image and Text

With the exception of the sketches intended only as aide-memoires, examples of each of the different types of illustrations Arriens composed saw publication. Opening the richly illustrated third volume of Und Afrika Sprach (1913a), the volume which documents the ethnographic findings from the late phase of the 1910-12 expedition that has concerned us, we discover as frontispiece a coloured illustration of five masquerades from the Benue region (see Figures 14 & 20 a-c). They start with the ‘buffalo’ masquerade, based on the photograph noted above, now put into a series with masquerades of the Ankwe, Jukun (including the capital Wukari), Hausa and ‘hill dwellers from south of Bauchi’. The ‘buffalo’ masquerade is here labelled Chamba (‘Tschamba’) rather than Daka (‘Dakka’) as it is in the halftone. The opening chapter begins by introducing to the reader people whom Frobenius calls the ‘Ethiopians’, or in accordance with this volume’s sub-title (Unter den unsträflichen Aethiopen), among the ‘blameless Ethiopians’. He uses this Homerian name to designate non-Islamized and largely uncentralized African farmers, those he also calls the disruptive peoples.

The relationship between the popular and scholarly editions of the text requires a detailed explanation here. The opening chapter of Und Afrika Sprach Volume III in the scholarly or scientific edition (1913a : 10-37) is identical to the last chapter in the popular editions (1912a : 649-67 in the German popular edition ; 1913b II : 650-79 in its English translation). This final chapter in the translated popular edition was called ‘Journey into the country of problems’, or in German ‘Die Fahrt in das Land der Fragen’, the journey to the land of questions ; in the scholarly edition it becomes ‘Ethiopian Studies’. What was it about this material that made it worth printing twice ? And how was this possible ? To take the second question first : the first two volumes of the German scientific edition and their single volume popular counterpart were both published in 1912. Although it did not appear until the following year, the third volume of the scientific edition, or at least its opening chapter, must also have been written by then for it to be possible to duplicate it in the popular edition. If that is why it was feasible to do so, the question remains, why bother ? The short answer is that both narratively and structurally this chapter holds Frobenius’ baggy text together : the movement from the second to third volumes involves a journey that the reader is invited to witness, and that journey takes us to the land of the ‘Ethiopians’, metonymically represented by images of the montagnard Koma, whom Arriens considered hardworking, sturdy, long-lived and innocent.

Four images of the expedition on the move form a transitional sequence : 1) traversing North Kamerun in ‘autumn’ 1911, a file of carriers stretches into the disappearing point of the image led by the three Europeans and their African guide on horseback, with at least two other Africans, one of them turbaned, also on horseback later in the column (Figure 23) ; [19] 2) the three Europeans organize traversal of a river in North Kamerun during the ‘rainy season’ (Figure 24) ; 3) canoes (‘indigenous boats’) in convoy transport the expedition along the Taraba-Benue River (Figure 25a, and 25b) ; and 4) the African members of the expedition are busily occupied setting up an encampment to spend a night on the sandbanks of the Benue, while the three Europeans apparently converse at a table behind which we see their beds and mosquito nets awaiting them for the night (Figure 26a and 26b, see also Pfisterer 2019 : 120 described as pencil and gouache ; possibly painted in Germany from sketches and photographs 2019 : 123 ; Kuba and Hambolu 2010 : 52 ). In only two of the four cases is the coloured original in the Frobenius archive, but it seems unlikely all four halftone images were not based on colour originals.

Figure 23
‘March of the expedition through North Cameroon’
From Und Afrika Sprach III, facing page 8
Figure 24
‘Crossing a stream in northern Cameroon during the rainy season 1911’
From Und Afrika Sprach III, facing page 11
Figures 25a and 25b
Figure 25a : ‘Return of the expedition in indigenous boats on the Taraba-Benue’
Figure 25b : Carl Arriens, ‘Downstream trip on the Benue’
Figure 25a : from Und Afrika Sprach III, facing page 32 ; figure 25b : watercolour, 53x37 cm, EBA-B 02705
Figures 26a and 26b
Figure 26a : ‘Expedition night camp on the sandbanks of the Benue’
Figure 26b : Carl Arriens, ‘Camp on a sandbank in the Benue’
Figure 26a : from Und Afrika Sprach III, facing page 40 ; figure 26b : watercolour, 53x37 cm, EBA-B 02704

The further choice of introductory illustrations provides a response to questions about the goal of this travelling. A solitary Koma man surveys the valley of the River Faro from a mountainous outcrop at the settlement of Ndera (Figure 27a, an image that also survives in the archive of the Frobenius Institute as an original watercolour, Figure 27b, see also Kuba and Hambolu 2010 : 41).

Figures 27a and 27b
‘Settlement of the Komai in Ndera (Alantika Mountains) with view of the Faro Plain’
Carl Arriens, ‘View from Upper Ndeera into the Faro Valley’
Figure 27a : from Und Afrika Sprach III, facing page 24 ; figure 27b : watercolour, 45x30 cm, EBA-B 00486

This is followed in the popular editions, by an image of Koma women sociably grinding corn for the evening meal while caring for their children (Figure 18, 1912a : facing 658 in the German edition, and 1913b : facing 672 in the English editions). The original Indian ink image by Arriens has survived in the archive of the Frobenius Institute (Figure 28).

Figure 28
Carl Arriens, ‘Koma women in the Alantika Mountains stamp and grind corn’
Indian ink, 51x34 cm, EBA-B 00611

This image is repositioned in the scholarly edition (see below), but a similar purpose can be attributed to Frobenius’ own photograph of four smiling Tiv youths, arms draped around one another’s shoulders, coiffed, with chipped teeth and adorned with beads. Captioned ‘Tiv lads, who find themselves decried as particularly malicious’ (Junge Muntschiburschen, die als ganz besonders bösartig verschrien sind), it is placed prominently in the preface of this work described as a new construction ‘Neubau’ (Figure 29a).

Figures 29a and 29b
‘Tiv lads, who find themselves decried as particularly malicious’
From Und Afrika Sprach III, facing page xxiv ; figure 29b : FoA 04-5783

The text that follows is divided into three sections dealing respectively with the western, central and eastern branches of the Ethiopians as Frobenius distinguished them. As the fourth Africa expedition was mostly concerned with Central Ethiopians, other than in the introductory chapter, it is in that section of Volume III where we find Arriens’ artworks. The ‘accession’ sketches for instance of calabash or more accurately gourd horns, bullroarers, rattles, wooden figures and so forth appear, most often arranged in sequences. In Frobenius’ archive they were part of a pictorial register containing some 20,000 small-format pictures divided into 96 subject areas, each of them geographically arranged. Then there are photographs, as well as the sketches explicitly based on photographs, and the images which open chapters, presented in the style of woodblock prints. Halftone plates, if we restrict our attention to those, predominantly deal with human and material subjects of study. The four plates of the expedition itself, which are an unusual subject matter in Frobenius’ voluminous publications, having appeared both in the conclusion to the popular editions, and in the introduction to Volume III of the scholarly edition, find no echo in the section on Central Ethiopians ; their function was performed in transitioning us to this country. We find a half dozen images stylistically to be set alongside that of the watchful solitary Koma man of Frobenius’ introduction.

  • (Facing page 136) An image depicting the ‘skull worship’ of three peoples (Verre ‘Vere’, Jen ‘Djenn’, Dowayo ‘Namdji’, ‘Bokko’) ; it is based on a ‘watercolour’ original that survives in the archive (EBA-B 00659). This static image is similar in style to the accession records and photographs known from the archive, and the Verre example specifically seems to be related to the skulls photographed in situ, and collected, by Frobenius (facing page 128).
  • (Facing page 192) An image based on ‘watercolour’ originals of a Koma woman (with a baby carried on her back) and a Koma man (the original is EBA-B 00541).
  • (Facing page 224) A plate based on an original drawn in Indian ink (EBA-B 00611, see above Figure 28). This is the sole image moved from its position in the concluding text of the German and English popular editions (see above). Koma women with their children and babies are grinding corn by hand for the evening porridge (staple food) ; long use of grinding stones has worn indentations into the flat, granite outcrops that characterize the summits of the Alantika Mountains.
  • (Facing page 240) is a drawing (Zeichnung) of the holy ‘Bo’ enclosure of the Koma at Ndera. No original survives of this work.
  • (Facing page 256) ‘The holy buffalo mask and lera dance of the Daka’. The main subject of this essay which we now know to have been drawn in Indian ink, although no original survives in the archive.
  • (Facing page 272) Based on a surviving ‘watercolour’ original is a depiction of the circumcision knives of the Chamba and the leopard’s paw pouch (EBA-B 00677, see Figure 9 above).

If we discount the records of objects (1 and 6) or places (4), we find few depictions of humans. Of these, the solitary Koma individuals on the hillside at Ndera, and in the portraits (2) are static. Only the Koma women grinding corn, and the participants in the masquerade and lera dance, both executed in Indian ink, are shown while involved in lively human activities. And of these two images, only the second includes observers, hence connecting it back to the images of the expedition’s column making their way by land and water through obstacles to these ‘Ethiopian lands’. Recall that Arriens had been left without European companions in the Alantika Mountains, so it is understandable that none of them appears as an observer of the Koma women at work, a presence that would in any case have been disruptive of the gendered scene. So, our image transpires to be unique in the literal sense among the copious illustrative materials found over the three volumes of Und Afrika Sprach.

Conclusion : Atypical and Exemplary ?

Can an image be at once atypical, because unique, and yet exemplary ? We would argue the portrayal we have been examining at least lends itself to being interpreted as such. Arriens depicted a scene which we are confident on the basis of the expedition documents did not take place, and certain from cultural familiarity could not have taken place. The masquerade performing with gourd horns was witnessed at a different time and place from the lera flutes and masquerade. Arriens’ rough sketches of both events survive with their place and date ; the first performance appears in his later memoir ; and we additionally have a photographic record of the second performance. There would be no room for doubt that these are distinct events, even if it was possible for Chamba horn and flute bands to perform simultaneously. What is more, neither of the performances Arriens did witness took place against the backdrop of the compound depicted. From his memoir, we learn that the horn performance was seen more or less accidentally, while the photographs of the flute performance give the impression of a staged performance. Arriens did indeed witness a performance in the public space in front of the compound of the Galadima of Tchamba, but that was a durbar to mark the departure of the expedition. The relief of Frobenius’ hosts from accommodating so substantial a column may have given them ample cause for celebration ; whatever the case, Arriens painted that scene as well.

The detail of Arriens’ image is no more convincing than the overall scene as an accurate record of a singular event. Arriens has added beards and hair knots to the expressively elongated bodies of the performers, presumably because the expedition learned that men in the Benue region, and particularly in the Jukun region which included emigrant Chamba, used to sport them. The top knots are not evident in the surviving photographs of the flute players, so the addition is anachronistic at best, and may even be transposed from another ethnic setting. The material culture on display includes just about everything the expedition had collected from the Chamba that might conceivably be present at such an imagined celebration : circumcision knives in leopard skin paws, the priestly sickles with iron lightning points and their wooden counterparts, basket rattles and iron rattles, personal ornaments and protective amulets, staffs, two different kinds of drum … All this is witnessed by men in European, Fulani and (probably) Hausa or Nupe dress. The image is totalizing and synchronic : an imagined ‘everything’ at an imagined moment of ‘all at once’. Everything all at once : an image of the cultural whole as witnessed by the expedition’s members and now presented in all its exotic trappings to a German and then English language readership, and to the museums and dealers whose interest made the venture feasible.

That the image is fanciful, in the sense that it never happened, is a matter of factual record. But the formative fancy that went into its composition is more a matter of interpretation. What makes the image exemplary in our estimation is its complete coherence with Frobenius’ text, and particularly with the section of that text concerned with the Central Ethiopians. A reader of the text consecutively would reach this point of the narrative by accompanying the expedition on its arduous course towards a little-known land introduced in terms of perplexity and recall the chapter heading that avers to ‘questions’ or ‘problems’ (Fragen). The diligent reader would make this journey whether as a reader of the popular or the scholarly edition. In the case of the popular edition, readers would not reach their intended destination, which probably indicates an intention for there to be a further volume, in the event never written, of the popular as well as the scholarly edition. Once arrived in this exotic land, the reader of the scholarly edition is invited to experience the world of the ‘blameless (unstraflichen) Ethiopians’, as they are introduced by the title of the third volume of Und Afrika Sprach. But this is a culture even more than usually assembled in the imagination of the investigator, not by the ethnographer’s method of taking part in everyday activities but through questions and answers, the dictation of myths and historical recollections, accounts of rituals never witnessed, and the collection and classification of material objects, many of them sourced via local assistants. [20] Seen in this way, Arriens’ image is exemplary of an ethnological ideal that Frobenius’ Ethiopian reconstruction sought to realize in text and illustration.

Given general acceptance that, insofar as anyone is ever able to do so, investigators should try reflexively to situate their own efforts, we return to the initial confession about our own imagination. Convinced that the halftone image had been created from an oil, gouache, or watercolour original by Arriens, we asked the graphic designer Agnès Boulmer to colourize it and scoured the archival sources discussed here to try to work out what the original colours might have been and how they might have looked given other examples of the palette Arriens employed for similar subjects. The outcome more than met our expectations ; many details that seemed obscure in the halftone now stood out clearly. Belatedly, as confessed earlier, we discovered in a ledger of images in the Frobenius Institute from the 1930s, that the then still surviving original had in fact been executed in Indian ink and so was monochrome from the outset. We offer it in conclusion to remind us of the pitfalls of imagination, our own most particularly, while recognizing the clarity and vivacity a new ‘original’ can bring to the inventions that a work like Und Afrika Sprach continues to invite.

Figure 30
Colourization of Arriens‘ Der heilige Büffelmasken- und Laerentanz der Dakka by Agnès Boulmer


Arriens, Carl 1928 Am Herdfeuer der Schwarzen. Erlebtes aus Westafrika, Weimar : Verlag für Urgeschichte und Menschenforschung.

Becker, Carl Heinrich 1913 ’Frobenius und die Brille des Islam“. Der Islam 4(3) : 303-312.

Berns, Marla C., Richard Fardon and Sidney Littlefield Kasfir 2011 Central Nigeria Unmasked : arts of the Benue River Valley, Los Angeles : Fowler Museum at UCLA.

Chappel, Tim, Richard Fardon and Klaus Piepel 2021 Survining Works : context in Verre Arts, Special Issue Vestiges : Traces of Record 7(1), ISSN : 2058-1963

Fardon, Richard 1988 Raiders and Refugees : trends in Chamba political development 1750‑1950, Washington : Smithsonian Series in Ethnographic Inquiry, Smithsonian Institution Press.

Fardon, Richard 1991 Between God, the Dead and the Wild : Chamba interpretations of religion and ritual, Edinburgh : Edinburgh University Press for the International African Institute ; Washington : Smithsonian Institution Press.

Fardon, Richard 2006 Lela in Bali : history through ceremony in Cameroon, Oxford : Berghahn, Cameroon Studies, Volume 7.

Fardon, Richard 2007 Fusions : masquerades and thought-style east of the Niger-Benue confluence, West Africa, London : Saffron Press, Afriscopes.

Frobenius, Leo 1912a Und Afrika Sprach. Bericht über den Verlauf der dritten Reise-Periode der D.J.U.F.G. in den Jahren 1910 bis 1912, Volkstümlick Ausgabe, Berlin : Vita Deutsches Verlagshaus.

Frobenius, Leo 1912b Und Afrika Sprach. Erster Band : Auf den Trümmern des klassischen Atlantis, Wissenschaftlich Ausgabe, Berlin : Vita Deutsches Verlagshaus.

Frobenius, Leo 1912c Und Afrika Sprach. Zweiter Band : Auf der Schwelle des verehrungswürdigen Byzanz, Wissenschaftlich Ausgabe, Berlin : Vita Deutsches Verlagshaus.

Frobenius, Leo 1913a Und Afrika Sprach. Dritter Band : Unter den unsträflichen Aethiopien, Wissenschaftlich Ausgabe, Berlin : Vita Deutsches Verlagshaus.

Frobenius, Leo 1913b The Voice of Africa. Being an account of the travels of the German Inner African Exploration Expedition in the years 1910-12, translated by Rudolf Blind, 2 volumes, London : Hutchinson and Co. (English version of 1912a).

Frobenius, Leo 1923 Das sterbende Afrika, Munich : O.C. Recht.

Frobenius, Leo 1925 Dichten und Denken im Sudan, Atlantis – Volksmarchungen und Volksdichtungen Afrikas Volume 5, Jena : Eugen Diederichs.

Frobenius, Leo 1987 Peuples et sociétés traditionelles du Nord-Cameroun, chapters from Frobenius 1925, translated and with an introduction (pp. 7-13) by Eldridge Mohammadou, Wiesbaden : Franz Steiner Verlag.

Hambolu, Musa O. 2010 ‘Nigeria 1910-1912 : the context of Leo Frobenius’ expedition’, in Richard Kuba and Musa Hambolu (eds) 2010 Nigeria 100 Years Ago. Through the eyes of Leo Frobenius and his expedition team, National Commission for Museums and Monuments, Nigeria, and Frobenius Institute, Frankfurt, pp. 33-42.

Ita, J.M. 1972 ‘Frobenius in West African History’, Journal of African History XIII (4) : 673-88. (pp. 681-6 for Ife).

Ivanoff, Hélène, 2020 ‘Les “compagnons obscurs” des expéditions de Leo Frobenius’, Bérose - International Encyclopaedia of the Histories of Anthropology, Paris.

Kuba, Richard and Musa Hambolu (eds) 2010 Nigeria 100 Years Ago. Through the eyes of Leo Frobenius and his expedition team, National Commission for Museums and Monuments, Nigeria, and Frobenius Institute, Frankfurt.

Kuba, Richard 2010 ‘Travelling Frobenius’, in Richard Kuba and Musa Hambolu (eds) 2010 Nigeria 100 Years Ago. Through the eyes of Leo Frobenius and his expedition team, National Commission for Museums and Monuments, Nigeria, and Frobenius Institute, Frankfurt, pp. 45-54 ; republished online in 2020, ‘The Expeditions of Leo Frobenius between Science and Politics : Nigeria 1910-1912’, Bérose - International Encyclopaedia of the Histories of Anthropology, Paris.

Kuba, Richard 2019 ‘Albrecht Georg Martius (Osnabrück 1884 – 1970 Toronto/Kanada)’, in Frobenius Institut Frobenius. Die Kunst des Forschens, Petersberg : Michael Imhof Verlag, pp. 216-19.

Kuba, Richard 2018 ’Portraits of Distant Worlds. Frobenius’ Pictorial Archive and its Legacy’, in : Helff, Sissy & Stefanie Michels (eds) Global Photographies. Memory – History – Archives. Bielefeld : Transcript, pp. 109-131.

Penny, H. Glenn 2002 Objects of Culture : ethnology and ethnographic museums in imperial Germany, Chapel Hill.

Pfisterer, Ulrich 2019 ‘Carl Arriens (Heide/Holstein 1969 – 1952 Berlin’, in Frobenius Institut Frobenius. Die Kunst des Forschens, Petersberg : Michael Imhof Verlag, pp 120-131.

Platte, Editha ‘Found and lost in Ife : Leo Frobenius and the Olokun bronze head, in Richard Kuba and Musa Hambolu (eds) 2010 Nigeria 100 Years Ago. Through the eyes of Leo Frobenius and his expedition team, National Commission for Museums and Monuments, Nigeria, and Frobenius Institute, Frankfurt, pp. 73-80.

Stulik, Dusan C. and Art Kaplan 2013 Halftone. The atlas of analytic signature of photographic processes, Los Angeles : J. Paul Getty Trust.

Zwernemann, Jürgen 1987 ‘Leo Frobenius und das Hamburgische Museum für Völkerkunde. Eine Dokumentation nach der Korrespondenz‘, Mitteilungen aus dem Museum für Völkerkunde Hamburg, Neue Folge 17 : 111-27.

Archival resources Frobenius Institute :

LF 250 Notizbuch IV, Albrecht Martius, November 1911

LF 251 Notizbuch V, Albrecht Martius, September – October 1911

LF 252 Notizbuch VI, Albrecht Martius, October – December 1911

LF 1581-3 Letter Leo Frobenius to his father Herrmann, dated 13 August 1911

LF 1584-8 Letter Leo Frobenius to Pagenstecher & Co, dated 13 August 1911

LF 1591-4 Letter Leo Frobenius to Witt and Büsch, dated 4 March 1911

LF 1592-4 Letter Leo Frobenius to Elfriede Zimmermann, Berlin, dated 28 November 1911

[1A number of images once in the Frobenius Institute in Frankfurt are today missing from the archive, which currently holds 396 images from the 1910-12 expedition that are attributed to Arriens : mostly watercolours (139) and Indian ink (119) but also pencil (41), oil (27) as well as (57) photographs. According to the register book from 1937, an image with the number 678 was then part of the ethnographic image archive, under the title ‘Büffel- u. Lärentanz der Dakka’. Whether it was lost during the bombing of Frankfurt in March 1944 when the Institute was severely hit or disappeared in the decades after the war, remains an open question.

[2Frobenius’s published account is sub-titled ‘Verlauf der dritten Reise-Periode ; however, in the official count of his expeditions apparently introduced after the First World War, Frobenius considered a brief visit to the Maghreb in 1910 as his third journey, pushing the subsequent expeditions back by one.

[3The popular edition of Und Afrika Sprach advertises a four-volume scholarly edition. The fourth volume, which would have covered the Kordofan expedition, was never published.

[4Frobenius first met Bida in 1909 in Togo where the Nupe prince served as a sergeant in the German colonial army. He spoke a German not wholly comprehensible to Frobenius, in addition to Pidgin and several other African languages, including Ewe, Yoruba, Hausa and of course Nupe (Ivanoff 2020 : 6). Carl Arriens later wrote of him, ‘The successful completion of a scientific expedition would hardly be conceivable without the mediation of intelligent natives such as this man. He is chairman of the black staff, interpreter, travel marshal and negotiator in one person’ (Arriens 1928 : 22). When Frobenius left his companions towards the end of the fourth expedition, he quickly met up with Bida to pursue research in Kordofan that should have furnished Und Afrika Sprach with its fourth volume.

[5In sub-Saharan Africa, Frobenius considered Islam as foreign disruptive element, destroying the ‘old warm-blooded African culture’, he hoped to document in a state of original purity, see Chapter 17 of Frobenius 1912a and 1913b volume 2, ‘The spectacles of Islam’, cf. also Becker 1913. His attitude towards the Fulani was similarly critical, accusing them of oppressing the hill peoples and using slave labour (1913b : 666, 669).

[6This and the following reference numbers refer to the Frobenius Institute’s image archive, available online at

[7In relation to these records, Richard Fardon acknowledges research undertaken with Klaus Piepel on a project in parallel with this one, see Chappel, Fardon and Piepel 2021.

[8A pull-out map at the back of Und Afrika Sprach Volume 3 reveals a partial survey of the hills, presumably by Martius, with peaks named after himself and Arriens (though neither name has stuck).

[9Over a dozen photographs in the Frobenius archive also document this event (FoA 04-5882 to FoA 04-5898).

[10Letters dated 4 March 1911, (LF 1591-4), 13 August 1911 (LF 1584-8) and 13 August 1911 (LF 1581-3).

[11Letter to Elfriede Zimmenann, Berlin, dated 28 November 1911 (LF 1592-4).

[12Letter to Elfriede Zimmenann, Berlin, dated 28 November 1911 (LF 1592-4). Concerning Frobenius general preference of painting over photography as means of documentation, see Kuba 2018 : 115-18.

[13Figure 7a Left : top (from left to right) : 3 wind instruments, red wood, leopard skin ; bottom : iron bells ; yellow fur bag with pieces inside (?), ragged tobe [tunic].
Figure 7b Right : top : circle of crouching figures, drummers in front then 5 flutes, the others dancing in circle, dances with mask (?) ; bottom : moon shine, all silhouettes against bright air but all in large surfaces (?) , …quite…runner (?), Schamba in Jelba

[14From the Yoruba (the plates are inserted without pagination and so referenced by facing page ; references are from the two volume English edition, 1913b) ; Volume I : 126 a partly-lit Ibadan interior from an ‘oil painting’ (Oelbild) ; 212 a holy lake at Ife from an ‘oil study’ (Oelstudie) ; 272 the salutation of the Oni of Ife from a drawing (Zeichnung), the shading lines of which are evident ; from the Nupe ; Volume II : 426 the night market of Bida from an ‘oil sketch’ (Oelstudie – note the variant English translation) ; 434 glass workers in Bida from a ‘study in oils’ (Oelstudie) ; 442 bronze-chasers in Bida from a ‘study in oils’ (Oelstudie) ; 452 bead-cutters in Bida from a ‘study in oils’ (Oelstudie) ; 476 iron foundry in Nupeland, medium not specified in the English or German editions. Whether paintings, studies and sketches in oil have real differences is not self-evident. Several of Arriens’ halftone prints from Bida are also reproduced in his later memoir, which includes a tall, vertical masquerade with some generic resemblance to our image that intriguingly was not published in Und Afrika Sprach (see Arriens 1928, facing page 96) where it would have had to share space with the more mundane images enumerated here.

[15This is explained in previous works (Fardon 1990 and 2007) ; in short, while the horns belong to a bushcow, the central dome with its ears, nose and eyes is human or dead, while the gaping mouth is taken from another creature, sometimes specified as crocodile or hippopotamus, both aquatic.

[16Chamba masquerades of this part of the Benue Valley have departed in some respects from the conventions of those in the ‘homelands’ (see Fardon 2007 : 73-77). Those from Donga are particularly upright.

[17Several parties of Chamba raiders had left their homelands in the early nineteenth century at the beginning of the Fulani jihad and went on to found chiefdoms among the peoples they conquered. Donga was the most significant of these in the Benue Valley.

[18Lama are found in several forms and are typically skeuomorphs of the everyday sickle, thus associating them with cutting and harvesting, and contrasting with hoes associated with women and with planting. The dancer priest in the halftone image carries a sickle with a lightning point (like KBA 10934), similar to those of the priests Fardon knew in Mapeo on the Nigerian side of the Alantika Mountains. The wooden lama with a forked ending is of the type used by priests in Yeli (KBA 10933). While Fardon did not see the simpler curved wooden form (KBA 10935), Kaberry and Chilver documented them in the emigrant Chamba chiefdoms in the Bamenda Grassfields where they were associated with harvest, and this may be consistent with Arriens having seen them at a performance during October, which is the time of harvest festivals in the region.

[19The African on horseback with a floppy straw hat resembles Ali Garua, a Kanuri, in a reproduction of an oil sketch published by Arriens in 1928 (facing page 16). That he was able to reproduce images, some of which also appear in Und Afrika Sprach, a decade and a half after that work appeared, suggests Arriens may have retained some of his original artwork.

[20It would take another essay to explore adequately, but Frobenius’ textual production seems to have involved typed versions being made of his and his colleagues’ fieldnotes which were then collaged together with connective passages. In short, there is a close formal resemblance between this way of making a representation of ‘a culture’ in text and the composition of the image we have been examining.