Content uploaded by Thomas M. Kaiser
Author content
All content in this area was uploaded by Thomas M. Kaiser
Content may be subject to copyright.
eschweizerbartxxx
Z. dt. Ges. Geowiss. 157/3, p. 411–431, 4 figs., 1 tab., 5 pts., Stuttgart, September 2006
Article
1860-1804/06/0157-0411 $ 9.45
DOI:
10.1127/1860-1804/2006/0157-0411
© 2006 E. Schweizerbart’sche Verlagsbuchhandlung, D-70176 Stuttgart
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene
terrestrial environments in sub-Sahara Africa
Thomas M. Kaiser, Jörg Ansorge, Gloria Arratia, Volker Bullwinkel, Gregg F. Gunnell,
Patrick S. Herendeen, Bonnie Jacobs, Jens Mingram, Charles Msuya,
Andreas Musolff, Rudolf Naumann, Ellen Schulz & Volker Wilde*
Kaiser, T.M., Ansorge, J., Arratia, G., Bullwinkel, V., Gunnell, G.F., Herendeen, P.S., Jacobs, B., Mingram, J.,
Msuya, C., Musolff, A., Naumann, R., Schulz, E. & Wilde, V. (2006): The maar lake of Mahenge (Tanzania) –
unique evidence of Eocene terrestrial environments in sub-Sahara Africa. [Der
Maarsee von Mahenge (Tansania) –
einzigartiger Beleg terrestrischer eozäner Lebensräume in Afrika südlich der Sahara.] – Z. dt. Ges. Geowiss., 157:
411–431, Stuttgart.
Abstract:
Recent excavations at the Eocene maar lake of Mahenge (Tanzania) by a German/Tanzanian team pro-
duced more than 1,900 fossil vertebrate, plant and trace fossils, among which fishes (51 %) make up the largest
part, followed by plant remains (36 %). Surface outcrops of lacustrine sediments suggest that the Eocene crater
lake was about 400 m in diameter. A 3.8 meter long section of lacustrine deposits has been documented, among
which are 1.1 meter of basal lacustrine sediments that never before have been exposed. Despite intense and most
probably early diagenetic dolomitization sedimentary fabrics are well preserved. Sediments from the central part
of the basin are well-bedded and in places show a fine and distinctive lamination. The overall sedimentation rate in
the lake centre is estimated at 0.93 mm/year. Plant fossils are characterized by an abundance of (mostly) unat-
tached Leguminosae leaflets representing at least eight taxa out of the total of 22–23 leaf taxa distinguished so far.
The Eocene vegetation at Mahenge obviously was structurally similar to the modern “miombo” woodlands of Tan-
zania, which are also dominated by caesalpinioid legumes. Combined with the sedimentological data, the fossil
flora indicates an overall dry climate with pronounced seasonality. The Mahenge maar lake has provided five dif-
ferent groups of fishes plus others still undescribed. These comprise the oldest known cichlids which may repre-
sent an early species flock. Ecological requirements of the extant relatives indicate redundant evidence of the ex-
istence of a shallow water area at the lakes margins. Mammals are so far only represented by the type specimen of
Tanzanycteris mannardi
Gunnell et al. (2003).
T. mannardi
differs from all known comparable Eocene microbats
in having a much larger cochlear diameter relative to basicranial width. The extremely enlarged cochlea indicates
that this bat had already developed sophisticated echolocation abilities. It is possible that
Tanzanycteris
represents
a relatively primitive ancestral rhinolophoid but the balance of its other character states aligns it with the middle
Eocene Messel Hassianycteridae. Mahenge and related maar lake deposits represent a rarely sampled temporal
and geographical window, which is expected to produce more significant new information about the evolution of
* Addresses of the authors: PD Dr. Thomas M. Kaiser (thomas.kaiser@uni-hamburg.de), Dipl.-Biol. Ellen Schulz, Biocentre
Grindel and Zoological Museum, University Hamburg, Martin-Luther-King-Platz 3, 20146 Hamburg, Germany; Dr. Jörg Ansor-
ge, Institut für Geologische Wissenschaften, Universität Greifswald, Jahnstraße 17a, 17487 Greifswald, Germany (ansorge@
uni-greifswald.de); Prof. Dr. Gloria Arratia, Naturhistorisches Forschungsinstitut – Museum für Naturkunde, Invalidenstr. 43,
10115 Berlin, Germany, and Biodiversity Research Center, University of Kansas, Lawrence, KS 66045-7561, USA
(garratia@ku.edu); Dr. Volker Bullwinkel, Geowissenschaftliches Zentrum der Universität Göttingen (GZG), Universität Göttin-
gen, Goldschmidtstraße 3, 37077 Göttingen, Germany (vb-public@factworld.de); Dr. Gregg F. Gunnell, Museum of Paleontolo-
gy, University of Michigan, Ann Arbor, MI 48109-1079, USA (ggunnell@umich.edu); Prof. Dr. Patrick S. Herendeen,
Department of Biological Sciences, The George Washington University, Washington, DC, 20052, USA (herenden@gwu.edu);
Prof. Dr. Bonnie Jacobs, Environmental Science Program, Southern Methodist University, P.O. Box 750395, Dallas, TX 75275-
0395, USA (bjacobs@smu.edu); Dr. Jens Mingram, Rudolf Naumann, GeoForschungsZentrum Potsdam, Telegrafenberg, 14473
Potsdam, Germany (ojemi@gfz-potsdam.de); Charles Msuya, Muhimbili University College of Health Sciences Dar-es-Salaam,
P.O. Box 65264, Dar-es-Salaam, Tanzania (cmsuya15@yahoo.com.uk); Dipl.-Geol. Andreas Musolff, Institut für Geographie und
Geologie, Universität Greifswald, Jahnstraße 17a, 17487 Greifswald, Germany (Andreas.Musolff@gmx.de); PD Dr. Volker
Wilde, Forschungsinstitut Senckenberg, Palaeobotanik, Senckenberganlage 25, 60325 Frankfurt am Main, Germany (Volker.
Wilde@senckenberg.de).
411_432_Kaiser.fm Seite 411 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
412
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Paleogene ecosystems in Africa, the origin of the Malagasy biota and the origin of several modern groups of plants
and vertebrates, especially mammals.
Kurzfassung:
Neue Ausgrabungen im eozänen Maarsee von Mahenge (Tansania) durch ein deutsch/tansani-
sches Team haben mehr als 1900 fossile Vertebraten, Pflanzen und Spurenfossilien erbracht, unter denen Fische
(51 %) den größten Teil ausmachen, gefolgt von Pflanzenresten (36 %). Aufschlüsse lakustriner Sedimente bele-
gen, dass der eozäne Kratersee ungefähr einen Durchmesser von 400 m hatte. Ein 3,8 m langes Profil in lakustri-
nen Sedimenten wurde dokumentiert, darunter sind 1,1 m basale lakustrine Sedimente, die noch niemals zuvor
aufgeschlossen waren. Obwohl eine intensive, sehr wahrscheinlich postdiagenetische Dolomitisierung stattge-
funden hat, ist das sedimentäre Gefüge gut erhalten. Sedimente aus dem zentralen Abschnitt des Beckens sind
gut geschichtet und zeigen stellenweise eine feine Laminierung. Die Sedimentationsrate im Seezentrum wird auf
0,93 mm/Jahr geschätzt. Unter den pflanzlichen Fossilien sind die meist isoliert gefundenen Blattfiedern von Le-
guminosen am häufigsten. Sie repräsentieren mindestens acht Taxa aus der Gesamtheit von 22–23 Blatttaxa, die
bislang unterschieden werden konnten. Die eozäne Vegetation von Mahenge war offensichtlich ähnlich struktu-
riert wie die modernen Miombowälder von Tansania, die auch von caesalpinoiden Leguminosen dominiert wer-
den. Kombiniert mit den sedimentologischen Daten, zeigt die fossile Flora ein insgesamt trockenes Klima mit
ausgeprägter Saisonalität an. Der Kratersee von Mahenge hat fünf verschiedene Gruppen von Fischen geliefert,
sowie andere, die noch unbeschrieben sind. Die Fischfossilien umfassen die ältesten bekannten Cichliden, bei
denen es sich um einen frühen Artenschwarm handeln könnte. Die ökologischen Anforderungen der rezenten
Verwandten liefern redundante Hinweise auf die Existenz einer Flachwasserzone an den Seerändern. Säugetiere
sind bislang nur durch das Typusexemplar von
Tanzanycteris mannardi
Gunnell et al. 2003 belegt.
T. mannardi
unterscheidet sich von allen bekannten vergleichbaren eozänen Kleinfledermäusen durch einen im Verhältnis zur
Basicranialbreite viel größeren Durchmesser der Cochlea. Die extrem vergrößerte Cochlea belegt, dass die Fle-
dermaus schon die komplexe Fähigkeit der Echoorientierung entwickelt hatte. Es ist möglich, dass
T. mannardi
eine relativ ursprüngliche Rhinolophide ist. Die Ausgewogenheit ihrer sonstigen Merkmale stellen sie jedoch in
eine Reihe mit den Hassianycteridae aus dem Mitteleozän von Messel. Mahenge und assoziierte Maarseesedi-
mente repräsentieren ein bisher selten belegtes zeitliches und geographisches Fenster, von dem weitere bedeutsa-
me neue Informationen zur Evolution paläogener Ökosysteme in Afrika, dem Ursprung der Malagasi-Biota und
dem Ursprung verschiedener moderner Gruppen von Landpflanzen und Vertebraten, insbesondere der Säugetie-
re, erwartet werden können.
Keywords:
Mahenge, maar lake, Eocene, Paleogene, Tanzania, terrestrial environment, palaeontology, palaeo-
ecology, palaeoclimate
Schlüsselwörter:
Mahenge, Maarsee, Eozän, Paläogen, Tansania, terrestrische Lebensräume, Paläontologie, Pa-
läoökologie, Paläoklima
1. Introduction
The Singida kimberlite field, with more than fifty kimber-
lite pipes and dykes, is located on a broad peneplain, just
to the south of the Iramba Plateau in north-central Tanza-
nia (Mannard 1962). The latter represents a large tilted
fault-block at the south-western extremity of the Gregory
Rift, with a maximum elevation of 1,680 m above sea lev-
el (Teale 1931, 1932, Eades 1936, Williams 1939). Steep
fault-scarps border the Wembere depression and Eyasi
trough to the north and west, and the Mpura-Durumo val-
ley to the east (Fig. 1). To the south, the plateau grades im-
perceptibly into the peneplain. The combination of mature
peneplanation and long-term tectonic stability south of the
Iramba Plateau has resulted in a topography of relatively
uniform relief, with elevations ranging from 1,100 m to
1,500 m, and creating a gently undulating terrain (Fig. 2).
Outcrops of granite are common (Plate 1: Fig. B), forming
small inselbergs or craggy tors that rise above the level of
the peneplain. Most parts of this peneplain can be associ-
ated with the “African Surface” (King 1962) that was
formed at the end of the Cretaceous.
The locality at Mahenge (4º 47' 50.2" S; 34º 15' 54.5"
E) is close to the village of Mwaru, 65 km west of the
town of Singida (Fig. 1) and east of the Wembere Steppe
and may be reached by car via forest trail.
Mahenge and its fossils is the subject of this contribu-
tion. Our main goal is to present the state of the art of the
present knowledge of the locality and its fossiliferous
content, and to compare our findings with those by previ-
ous authors. First, we will analyze the information and
interpretation on the locality, and second, on the fossils
and their importance.
2. Evidence for the age of the lake beds
The relationships of the fish fauna led Greenwood & Pat-
terson (1967) to favour a Paleogene (probably Oligocene)
age for the Mahenge sequence. A kimberlite near Nzega,
120 km to the west of Mahenge, has yielded U-Pb dates
of 52.2 and 53.2 Ma, and fission track dates of 54.3 +
14 Ma and 51.1 + 3.8 Ma (Davis 1977, Naeser & McCal-
lum 1977, Haggerty et al. 1983), indicating an early
411_432_Kaiser.fm Seite 412 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
413
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Eocene (= Ypresian) age. Harrison et al. (2001) report a
middle Eocene (Lutetian) 206Pb/238U age of 45.83 ±
0.17 Ma from a single zircon crystal which was recov-
ered from the stream bed at Mahenge in 1996. There are
several arguments linking the respective crystal to the
Mahenge diatreme. Recent catchment area of the stream
crossing the Mahenge diatreme is small (7.2 km
2
). It cov-
ers Archaean metamorphics and plutonics and no other
known diatreme. The catchment area can be expected not
to have varied much since formation of the peneplain in
the Cretaceous. Therefore, the crystal with a middle
Eocene age found in the stream bed at Mahenge can most
probably be related to the Mahenge eruptive event and
could be only slightly older than the lake at Mahenge. It
is well known from similar maar lakes in Europe and Af-
rica that accumulation of lacustrine sediments starts soon
after the formation of the crater (Lorenz 1973, Smith
1986, Rayner & McKay 1986, Rayner 1987, Giresse et
al. 1991, Cornen et al. 1992). Harrison et al. (2001) there-
fore confidently assume that the fossils recovered from
the maar sediments date to ~ 45–46 Ma.
3. Recent excavations at Mahenge
During the SSPP (Sub-Sahara-Paleogene-Project) field
campaign in July–August 2002 a total of 17 exposures
within the Mahenge crater were studied (Fig. 3) and doc-
umented in detail with respect to stratigraphy and sedi-
mentology. Most of the natural exposures are located
along the cutbank of the Luwala River (Plate 1: Figs. A,
C) crossing the crater from NE to SW (Fig. 3). During the
2002 field work the river bed was dry and largely accessi-
ble. Additional exposures were excavated along a transect
Fig. 1: The diatreme fields in North-Central Tanzania. Note the concentrations near Nzega, Singida, and Shinyanga, which are sep-
arated by the Wembere Steppe. The Mahenge diatreme is located in the far West of the Singida diatreme field.
Abb. 1: Das Diatremfeld in Nord-Zentral-Tansania. Die Konzentrationen in der Umgebung von Nzega, Singida und Shinyanga wer-
den durch die Wembere-Steppe getrennt. Der Mahenge-Diatrem liegt weit im Westen des Singida-Diatremfeldes.
411_432_Kaiser.fm Seite 413 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
414
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
oriented NW–SE roughly perpendicular to the river bank
(Fig. 3). This transect includes pits already exploited by
Mannard (1962), which were subsequently re-activated
and substantially extended and deepened, as well as new
pits dug (e.g. pit 12; cf. Fig. 3, Plate 1: Fig. D).
The main excavation efforts were concentrated in the
putative centre of the maar (pits 5 and 6 of Mannard 1962
and Harrison et al. 2001). At pit 05 (Plate 1: Fig. D), a
section of 3.80 m of mostly well-bedded and fossiliferous
lake sediments has been exposed (Fig. 4) and constituted
the reference profile for the central part of the lake beds.
Part of a similar (shorter) section was described by Harri-
son et al. (2001) and Jacobs & Herendeen (2004), howev-
er correlation was found to be so difficult due to laterally
varying degrees of silicification and cementation, that a
new section had to be documented.
3.1. The “pit 05c” section (P5C)
The base of section P5C is formed by a homogeneous
siltstone bed (0 m to 0.20 m) which is highly solidified by
carbonate and silica. Towards the top this bed grades into
a soft and friable siltstone (0.20 m to 0.60 m). The yel-
lowish grey (5Y 8/1) sediment is homogeneous, contain-
ing a certain fraction of fine sand which is randomly dis-
tributed throughout the entire thickness of the bed.
Occasionally, fine sand is concentrated in thin horizons
and lenses. Above 0.60 m the siltstone is succeeded by a
succession about 1.90 m in thickness of calcareous mud-
stones and mudstones which are in part strongly silici-
fied. Well-consolidated calcareous mudstones are 2 cm to
8 cm thick and contrast sharply with the underlying
facies. The upper surface is also sharp. However, between
1.20 m and 2.50 m the compact mudstones typically be-
come less competent towards the top and grade into fria-
ble or finely fracturing mudstones. Despite their low
Fig. 2: Modelled surface of the Mahenge diatreme area. Altitude data base on GPS-mapping in the field. The arrow points north.
Red dots mark the diatrem’s circumference. Blue line marks the bed of the Luwala River. The diatreme itself forms a shallow de-
pression. It is surrounded by an elevated ring structure, probably consisting of remnants of the tephra-wall, related to the eruptive
event. Vertical scaling is 1:2.3; coordinate system is UTM.
Abb. 2: Oberflächenmodell des Diatremgebietes von Mahenge. Höhenwerte basieren auf GPS-Daten, die im Gelände gewonnen
wurden. Der Pfeil weist nach Norden. Rote Punkte markieren den Umfang des Diatrems. Die blaue Linie markiert das Bett des Lu-
wala Flusses. Der Diatrem selbst ist eine flache Vertiefung. Er ist von einer leicht erhöhten Ringstruktur umgeben, die wahrschein-
lich aus den Resten des ehemaligen Tephrawalles besteht. Vertikaler Maßstab: 1:2,3; Koordinatensystem: UTM.
411_432_Kaiser.fm Seite 414 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
415
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
competence, the friable mudstones are well-bedded
showing a fine and distinctive lamination. Calcitic coat-
ings of cleavages and bedding planes are a characteristic
feature of these finely fracturing mudstone layers.
Strongly silicified mudstones are prominent within
the section due to their hardness and their conspicuous
light olive colour (5Y 7/2). However, they are not suited
for correlation since the degree of their silicification var-
ies in vertical as well as lateral direction.
The P5C section terminates in a hard but vaguely
stratified bed of a strongly lithified mudstone (2.50 m to
2.70 m). About 2 m further west of the section P5C this
top of exposed lake sediments is highly disturbed. Here
Harrison et al. (2001) describe a section of about 1 m in
thickness in which massive blocks of poorly laminated
yellowish grey mudstone occur together with slabs of
consolidated mudstones within a matrix of friable shale.
This section was re-investigated in 2003 and we find our
stratigraphic observations to be in general agreement
with Harrison et al. (2001).
3.2. Abdallah’s hole section (AHS)
The base of section P5C is an extremely hard siltstone
bed of up to 20 cm in thickness. Since section P5C is sit-
uated almost in the centre of the former maar lake, and
thus represents the deepest point that any prior excava-
tion had reached within the lacustrine succession of the
lake, the AHS section was never before exposed. It thus
represents the so far oldest known sediments from Ma-
henge, the top of which Harrison et al. (2001) considered
the base of the lacustrine series. Penetration of the hard
siltstone bed at the base of P5C was laborious, and in
honour of Mr. Abdallah, who exposed the section below
within several days of hard labour, the section is referred
to as “Abdallah’s Hole Section” (AHS).
Fig. 3: Map of the Mahenge diatreme with excavation localities and natural outcrops indicated. Note that excavation so far has only
been carried out along a transect connecting pits 05a–e, 12, and 06a–c. These localities expose well stratified laminated deposits
and all represent the central part of the former lake beds.
Abb. 3: Karte des Mahenge-Diatrems mit Grabungsstellen und natürlichen Aufschlüssen. Grabungen wurden bislang nur entlang
eines Transsekts durchgeführt, der die Grabungsstellen pit 05a–e, 12 und 06a–c miteinander verbindet. Diese Lokalitäten schließen
gut geschichtete Ablagerungen auf und repräsentieren den zentralen Teil der ehemaligen Seesedimente.
411_432_Kaiser.fm Seite 415 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
416
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Fig. 4: Stratigraphic section of pit 05c and the
adjacent “Abdallah’s Hole Section”. The vast
majority of fossils so far known from Mahenge
derive from a 20 cm thick section at the base of
the P05c profile at between 65 cm and 85 cm.
This section widely corresponds with the hori-
zons 3–6 of Harrison et al. (2001).
Abb. 4: Profil der Grabungsstelle pit 05c und
der sich anschließenden „Abdallah’s Hole Sec-
tion“. Die überwiegende Mehrzahl der bislang
von Mahenge bekannten Fossilien stammt von
einem etwa 20 cm mächtigen Schichtpaket an
der Basis des pit 05c-Profils zwischen 65 cm
and 85 cm. Dieses Profil entspricht weitgehend
den Horizonten 3–6 nach Harrison et al. (2001).
411_432_Kaiser.fm Seite 416 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
417
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
This 1.10 m thick section (Fig. 4) consists mainly of
calcareous mudstones, silicified mudstones, and sandy
siltstones. The calcareous mudstones as well as the silici-
fied mudstones are lithologically comparable with the
mudstones of section P5C. These generally well-bed-
ded sediments, however, include numerous silt layers
ranging in thickness from <1 mm to 5 mm. Noteworthy is
the frequent occurrence of sandy siltstones of up to 10 cm
in thickness. Between 0.60 m and 0.70 m clasts of lami-
nated mudstones (“intraclasts”) up to 5 cm in width are
irregularly embedded within a homogeneous silt/sand
matrix. Lenticular clasts of reworked lake sediment also
occur at the base of silt/sand layers which are character-
ized by graded bedding of coarse or medium grained
sand at the base to fine sand and silt at the top.
The surface outcrops of lacustrine sediments within
the Mahenge crater area suggest that the lake had a diam-
eter of about 400 m (see also Mannard 1962, Harrison et
al. 2001). The sections of “pit 05c” and “Abdallah’s
Hole” mainly represent the fine-grained sediments which
were deposited near the centre of the lake. It can be in-
ferred that rather steep slopes bordered the crater’s lake
in its early stage. The collapse of crater walls and slump-
ing of debris repeatedly triggered transport of coarse-
grained sediment towards the centre of the lake and led to
a gradual reduction in relief. As the lake margins contin-
ued to be flattened the sediment input became finer in
grain size. This stage of lake development is documented
in the AHS profile. The 10 cm thick homogeneous silt/
sand layers as well as the clasts bearing graded layers
represent the episodic sediment input by mudflows and
turbidity currents respectively.
Compared with AHS there is little evidence of input
by bedload in section P5C. The succession between
0.60 m and 2.50 m is dominated by well-bedded, in part
distinctly laminated carbonaceous mudstones. The re-
markable quality of fossil preservation in combination
with the total lack of bioturbation provides clear evidence
of stable stratification of the water column and anoxic
conditions near the sediment surface (Herendeen &
Jacobs 2000, Harrison et al. 2001, Kaiser et al. 2003).
4. Microfacies analysis of the
lacustrine sequence in the basin
centre
A set of petrographic thin sections, comprising app.
0.5 m from different depths of the recent section has been
studied microscopically. All materials investigated have
been sampled by the SSPP. Light microscopy was en-
gaged as plane light-, polarized light-, dark field-illumi-
nation, and fluorescence in order to obtain information on
the process of sediment deposition and diagenesis. Addi-
tionally, samples were taken for X-ray diffraction analy-
ses to identify the major mineral constituents of the sedi-
ments.
Mineral phases which could be identified by X-ray
diffraction are dolomite, opal-CT, crystalline quartz, and
some palygorskite. Petrographically the sediment is a do-
lostone with some silicified horizons. In all thin sections,
micritic and micro-sparitic xenomorphic dolomite (of
app. 2–15 μm grain size) is the most common mineral
(Plate 2: Figs. A–H). Sometimes calcitic vein fillings and
coarse-grained laminae containing micro-sparitic calcite
could be observed (Plate 2: Fig. C). Despite intense and
most probably early diagenetic dolomitization sedimen-
tary fabrics are well preserved.
Most samples in thin sections exhibit sub-mm scale
lamination, which points to an at least seasonally devel-
oped oxygen deficiency of the lake’s bottom water and a
seasonal climate. Frequently occurring graded event lay-
ers (Plate 2: Figs. A–B), ranging in thickness from sub-
mm to several cm, indicate catchment erosion by heavy
rain showers.
Strong seasonal lamination sometimes is replaced by
sediments with lenticular and flaser bedding and only
faint lamination, which could be caused by climatically-
induced lake level changes and/or changing influx of a
small tributary. Graded event layers often contain re-
worked lake sediments (“intraclasts”), coated lithic
grains (quartz, feldspars, some biotite) with varying stag-
es of replacement by dolomite, dolomitic peloids, some
ooids, and indeterminate ostracode shells (Plate 2: Figs.
D, E, F), with the latter pointing to the existence of a mar-
ginal shallow-water zone, provided they represent ben-
thonic taxa.
Except for the ostracode shells no remains of aquatic
organisms could be observed. But in fact, there are some
laminations which resemble those of diatomaceous lam-
inites from recent and ancient lake sediments (Bullwinkel
2003). But if there were primary diatomaceous layers
they are already totally altered into opal-CT and/or re-
placed by dolomite (Plate 2: Figs. B, C and G, H). This is
in good accordance to the fact that diatoms are known for
fast alteration (Bullwinkel 2003).
As the Mahenge laminated sediments both in fabric
and thickness strongly resemble recent seasonally lami-
nated lake sediments (varves), an estimation of sedimen-
tation rates by varve counting has been undertaken. For a
total thickness of 492.6 mm in thin sections from differ-
ent parts of the profile (including graded event layers),
529 varves could be found, which yields an overall sedi-
mentation rate of 0.93 mm/year which is considerably
higher than previously estimated by Harrison et al.
(2001). Taking into account this rate, the recent 5.5 m
long Mahenge profile represents app. 5,000 years of the
lake’s history.
411_432_Kaiser.fm Seite 417 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
418
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
5. The fossil record of the Mahenge
maar
During the SSPP expedition much attention was paid to
record and collect all fossils that came to our attention
during the excavation. A large bias was accepted in iso-
lated fish scales, isolated fish bones and extremely disar-
ticulated fishes, as well as plant fragments that did not al-
low any closer determination. Excavations in 2002 thus
yielded a total of 1,948 fossil specimens. This number ex-
ceeds the number of all previously known fossils from
Mahenge and is by far the largest number of specimens
ever collected from this locality (Tab. 1). Like in the col-
lections of Harrison et al. (2001) fishes are the most
abundant group (n = 993) and make up 51 % of the SSPP
specimens. Plants are the second abundant group (n =
698) and make up 36 %. However it should be men-
tioned, that we collected about two times as many fishes,
but almost ten times as many plant fossils compared to
Harrison et al. (2001). This remarkable difference is most
likely not due to taphonomy but is expected to reflect the
different sampling strategies of the two teams. In addition
the SSPP collected a fairly large number of coprolites (n
= 51, 2.6 %) and “sand nests” (n = 178, 9.1 %). These pe-
culiar small scale accumulations of fine to medium
grained sand are also common in the deposits of the Mid-
dle Eocene northern hemisphere maar lake of Messel
(Germany) where they have been interpreted as spitting
debris mainly of fishes (Schmitz 1991). At Mahenge as
also at Messel they represent the only sandy inclusions in
the central facies of the basin. Arthropods (<0.4 %) and
frogs (0.2 %; Plate 4: Fig. D) are among the most rare
taxa (see Tab. 1). This rarity has been confirmed by all
parties working at Mahenge so far.
Tab. 1: Representation of fossils recovered in the SSPP exca-
vations at Mahenge in 2002 (n-SSPP and %-SSPP) and fossils
reported by Harrison et al. (2001) in 1994 and 1996 (n-H2001
and %-H2001). Note, that fish remains are the most abundant
fossil group followed by plants. Sand nests and coprolites have
been collected for taphonomic reasons.
Tab. 1: Auflistung der im Rahmen der SSPP-Ausgrabungen in
Mahenge 2002 geborgenen Fossilien (n-SSPP und %-SSPP),
sowie von Fossilien die Harrison et al. (2001) 1994 und 1996
geborgen haben (n-H2001 and %-H2001). Fischreste stellen
die häufigste Fossilgruppe, gefolgt von Pflanzen. Sandnester
und Koprolithen wurden aus taphonomischen Gründen gebor-
gen.
n-SSPP %-SSPP n-H2001 %-H2001
Fish 993 51.0 513 81.40
Plant 698 35.8 85 13.49
Sand nest 178 9.1
Coprolite 51 2.6
17 2.70
Indeterminate
specimens
17 0.9
Invertebrates
(mostly arthropods)
7 0.4 5 0.79
Frogs 4 0.2 4 0.63
Total 1948 624
Plate 1:
Fig. A: The vegetation at Mahenge is miombo woodland, dominated by tree species like Julbernardia, Isoberlinia, and Brachyste-
gia. The Luwala River bed crosses the diatreme from NE to SW (view northward) and potholes (foreground) indicate that it may
have substantial transport energy in the rainy season.
Fig. B: Granite outcrop 39 (cf. Fig. 3) marking the NE boundary of the diatreme, the foreground being inside, the granite pitch be-
ing outside. At this place the Luwala River enters the diatreme as a cascade.
Fig. C: Massive beds of reworked pyroclastic material exposed in the river bed upstream of the crater centre. These partially graded
beds dip slightly (8˚ to 16˚) towards the former lake centre and represent the episodic input of coarse-grained sediment from the
crater slopes.
Fig. D: Excavation pits close to the centre of the former lake structure (view ESE). Foreground: pit 12, background right: pit 05d,
background left: pits 05a–c.
Tafel 1:
Fig. A: Die Vegetation von Mahenge wird von Miombowald gebildet. Es dominieren Bäume wie Julbernardia, Isoberlinia und
Brachystegia. Das Bett des Luwala-Flusses durchschneidet den Diatrem von NE nach SW (Blickrichtung Nord). Die Kolke im Vor-
dergrund zeigen an, dass er in der Regenzeit eine erhebliche Transportenergie aufweist.
Fig. B: Der Granitaufschluss 39 (s. Abb. 3) markiert die NE-Grenze des Diatrems. Der Vordergrund ist innerhalb, die Granitstufe
befindet sich außerhalb des Diatrems. Hier tritt der Fluss Luwala in einem Wasserfall in den Diatrem ein.
Fig. C: Massive Schichtenfolgen aufgearbeiteten pyroklastischen Materials sind im Flussbett stromaufwärts des Kraterzentrums
aufgeschlossen. Die teilweise gradierte Schichtenfolge fällt leicht (8° bis 16°) zum ehemaligen Seezentrum hin ein und entspricht
dem episodischen Eintrag grobkörniger Sedimente von den Hängen des Kraterwalls.
Fig. D: Grabungsstellen in der Nähe des Zentrums der ehemaligen Seestruktur (Blickrichtung ESE). Vordergrund: Grabungsstelle
pit 12; Hintergrund rechts: Grabungsstelle pit 05d; Hintergrund links: Grabungsstelle pit 05a–c.
411_432_Kaiser.fm Seite 418 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
419
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 419 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
420
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
5.1. The flora
Plant macrofossils collected from Mahenge now number
over 1,000 specimens and are most common in the lower
stratigraphic levels of pit 05c–d (levels 7–8 of Harrison et
al. 2001 and Jacobs & Herendeen 2004). They are char-
acterized by an abundance of (mostly) unattached Legu-
minosae leaflets (Plate 3: Figs. A, C, E, Plate 4: Fig. C)
representing at least eight taxa out of the total of 22–23
leaf taxa distinguished so far. Included among these are
more complete specimens of
Acacia
mahengensis
(Plate
4: Figs. A, B),
Aphanocalyx
singidaensis
(Plate 3: Fig.
A),
Bauhinia
sp., two taxa superficially similar to
Cy-
nometra
spp. (Plate 3: Fig. D), and an unknown mi-
mosoid (Plate 3: Fig. F; Herendeen & Jacobs 2000). The
vast majority of the isolated leaflets may be assigned to
the subfamily Caesalpinioideae (Plate 3: Figs. C, D), rep-
resenting one or more of cf.
Cynometra
affinity. The re-
maining dicotyledonous taxa are represented primarily
by microphyllous, entire-margined leaves that are distin-
guished by venation patterns and overall shape, but of
still unknown affinity; only two taxa with a toothed mar-
gin have been found. Monocotyledonous remains are
common among the leaves but have yielded only faint in-
dications of epidermal cellular structure, rendering them
unidentifiable. Nevertheless, it is important to note that
plants with “graminoid” leaves were a significant part of
the lake or lake-margin ecosystem. On the basis of what
we know of the composition of the Mahenge palaeoflora,
it was structurally similar to the modern “miombo”
woodlands of Tanzania, which are also dominated by
caesalpinioid legume taxa. Regrettably, no palynomorphs
have been recovered from a number of sediment samples
which were processed.
Plate 2:
Thin sections of lacustrine deposits from Mahenge (pit 06a). All photos made with transmitted light; B and G with parallel polariz-
ers, C, D, E, F, and H with crossed polarizers.
Fig. A: Flat-bed scan, transmitted light with crossed polarizers. Note alternation of seasonal laminae and graded event layers con-
taining reworked material (intraclasts). Sample SSPP MA-2079.
Fig. B: Laminated sediment. Note intercalation of event layers and coarser allochthonous minerals (lithoclasts). Coarse layers in
the lower half of the section contain numerous peloids and ostracode shells. Thickness of one doublet of fine laminae (varve) is
250–300 μm. Sample SSPP MA-2087.
Fig. C: Detail of Fig. B. At the base note a coarse-grained event layer with altered lithoclasts and micro-sparitic calcite matrix. Dark
laminae consist of dense, micritic dolomite (grain size 2–5 μm) and some lithoclasts, whereas lighter laminae are composed of
opal-CT with minor amounts of micritic to micro-sparitic dolomite (grain size 2–10 μm).
Fig. D: Detail of a coarse-grained event layer with coated grains and peloids. Sample SSPP MA-2087.
Fig. E: Quartz and feldspar are partly replaced by micritic secondary dolomite. Same sample as D.
Fig. F: Ostracode shell in a coarse-grained event layer. Sample SSPP MA-2073.
Fig. G: Dark laminae contain more lithoclasts and indicate periodic, most probably seasonal, inwash of allochthonous material.
Sample SSPP MA-2073.
Fig. H: Detail of Fig. G. Seasonal lamination with matrix composed of opal-CT without biogenic remains. The lamination is over-
printed by micritic to micro-sparitic secondary dolomite. Thickness of one doublet of laminae (varve) is 200 μm.
Tafel 2:
Dünnschliffe lakustriner Ablagerungen von Mahenge (pit 06a). Alle Aufnahmen wurden mit Durchlicht, B und G mit parallelen
Polfiltern, C, D, E, F und H mit gekreuzten Polfiltern aufgenommen.
Fig. A: Flachbettscan im Durchlicht, mit gekreuztem Polfilter aufgenommen. Deutlich ist der Wechsel von saisonalen Laminae und
Ereignislagen, die aufgearbeitetes Material (Intraklasten) enthalten. Probe SSPP MA-2079.
Fig. B: Das laminierte Sediment zeigt Einschaltungen von Ereignislagen und gröberen allochthonen Mineralien (Lithoklasten).
Gröbere Lagen in der unteren Hälfte des Profils enthalten zahlreiche Peloide und Ostracodenschalen. Die Dicke eines Doublets
feiner Laminae (Warven) beläuft sich auf 250–300 μm (Probe SSPP MA-2087).
Fig. C: Detailansicht von Fig. B mit einer grobkörnigen Ereignislage mit alterierten Lithoklasten und micro-sparitischer Kalzitma-
trix an der Basis. Dunkle Laminae bestehen aus dichtem, mikritischem Dolomit (Korngröße 2–5 μm) und einigen Lithoklasten,
während hellere Laminae von Opal-CT und geringen Anteilen von mikritischem bis micro-sparitischem Dolomit gebildet werden
(Korngröße 2–10 μm).
Fig. D: Detail einer grobkörnigen Ereignislage mit coated grains und Peloiden (Probe SSPP MA-2087).
Fig. E: Quarz und Feldspat sind teilweise ersetzt durch mikritische Sekundärdolomite (gleiche Probe wie Fig. D).
Fig. F: Ostracodenschalen in einer grobkörnigen Ereignislage (Probe SSPP MA-2073).
Fig. G: Dunkle Laminae enthalten mehr Lithoklasten und zeigen periodische, sehr wahrscheinlich saisonale Einspülungen von al-
lochthonem Material an (Probe SSPP MA-2073).
Fig. H: Das Detail von Fig. G zeigt saisonale Lamination mit einer aus Opal-CT bestehenden Matrix, jedoch ohne biogene Reste.
Die Lamination ist überprägt von mikritischen bis mikro-sparitischen Sekundärdolomiten. Dicke eines Laminen-Doublets (Warve):
200 μm.
411_432_Kaiser.fm Seite 420 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
421
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 421 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
422
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Mean annual and wet months rainfall (all months with
an average of
≥
50 mm) were estimated for Mahenge us-
ing regression formulas derived from modern tropical Af-
rican and South American leaf and climate data (Wilf et
al. 1998, Jacobs 1999, 2002, Gregory-Wodzicki 2000,
Jacobs & Herendeen 2004). Mean annual precipitation
estimates are 643 ± 32 and 776 ± 39 mm/year, or near the
modern annual average of 660 mm. The wet months pre-
cipitation estimates are 630 ± 38 and 661 ± 38 mm, or
somewhat higher than the modern 595 mm, indicating
that, although there was a dry season (or seasons), more
of the annual rainfall may have occurred during the dry
season(s) than today, and therefore the dry interval may
have been slightly less pronounced during Mahenge time.
Such a climate is consistent with the sedimentological
data which also indicate a pronounced dry season and
overall dry climate based upon the dominance of dolo-
mite in the fossil matrix.
5.2. Insects
Insect remains are some of the most rarely recorded fos-
sils at Mahenge. However when preserved, the quality of
preservation is very detailed. Therefore Mahenge is con-
sidered one of the few potential insect lagerstätten in Af-
rica, comparable to the Middle Cretaceous kimberlite di-
atreme of Orapa in Botswana (Rayner et al. 1991). Of
special interest is a rather complete beetle in dorso-ven-
tral compression, with well-preserved legs and one out-
stretched hind wing (Plate 4: Figs. E, F). On the basis of
these characters the fossil is identified as a scarabaeid
beetle [Scarabaeoidea (Lamellicornia)]. Besides an unde-
scribed Scarabaeidae from Orapa (McKay & Rayner
1986, Krell 2004) and some Miocene dung beetles from
Kenya (Paulian 1976), the Mahenge fossil is one of the
few fossil scarabaeid beetles known from Africa.
5.3. The fish fauna
Fish fossils are the most abundant animal group repre-
sented at Mahenge. The vast majority of individuals are
in articulated preservation, however also isolated scales
are to be found.
The Cenozoic freshwater fossil fishes from Africa are
little known and in many cases, represented only by frag-
ments. An initial review of the fossil record at Mahenge
was provided by Greenwood (1974) and more recently by
Murray (2000a). Among all fossil African localities bear-
ing fishes, Mahenge is probably the most important one,
because hundreds of complete and almost complete spec-
imens have been recovered here. Some of them, as for in-
stance, osteoglossomorphs (Plate 5: Fig. D) and cichlids
(Plate 5: Fig. B), are even represented as ontogenetic se-
ries in the SSPP material.
The Mahenge crater’s lake has provided five different
groups of fishes plus others still undescribed. Up to now,
two osteoglossomorphs (Plate 5: Fig. D), a clupeomorph,
a characiform, a siluriform (Plate 5: Fig. A), and five per-
comorphs (Plate 5: Fig. B) have been formally described.
As with almost all fossils from Mahenge, this sample
also was collected from the central part of the lake (pits
05 and 06) that in all probability corresponds to the deep-
est water column. The extant relatives of the clupeo-
morph, the characiform, and also all larvae and juveniles
however prefer shallow waters. The fossil community
preserved at the lake centre is therefore most likely a
Plate 3:
Fig. A: One leaflet of bifoliolate Aphanocalyx singidaensis (Herendeen & Jacobs 2000), Caesalpinioideae (Leguminosae), SSPP
MA-1956, pit 05d.
Fig. B: Leguminosae fruit (SSPP MA-274, pit 05c).
Fig. C: Single leaflet of unknown Caesalpinioideae (Leguminosae), SSPP MA-1113, pit 05d.
Fig. D: Bifoliolate leaf of “cf. Cynometra” (Herendeen & Jacobs 2000), Caesalpinioideae (Leguminosae), SSPP MA-1139, pit 05d.
Fig. E: Isolated Leguminosae leaflet, SSPP MA-1147, pit 05d.
Fig. F: Rachis with leaflets of Mimosoideae (Leguminosae), SSPP MA-1091, pit 05d.
Fig. G: A single-seeded segment of a Leguminosae fruit with insect damage (SSPP MA-258, pit 05c).
Tafel 3:
Fig. A: Einzelne Fieder eines doppelt gefiederten Blattes von Aphanocalyx singidaensis (Herendeen & Jacobs 2000), Caesalpinioi-
deae (Leguminosae), SSPP MA-1956, pit 05d.
Fig. B: Leguminosenfrucht (SSPP MA-274, pit 05c).
Fig. C: Einzelne Blattfieder eines unbekannten Vertreters der Caesalpinioideae (Leguminosae), SSPP MA-1113, pit 05d.
Fig. D: Doppelt gefiedertes Blatt von „cf. Cynometra“ (Herendeen & Jacobs 2000), Caesalpinioideae (Leguminosae), SSPP MA-
1139, pit 05d.
Fig. E: Isoliertes Leguminosen-Fiederblatt, SSPP MA-1147, pit 05d.
Fig. F: Rhachis mit Fiederblättern eines Vertreters der Mimosoideae (Leguminosae), SSPP MA-1091, pit 05d.
Fig. G: Einfrüchtiges Segment einer Leguminosenfrucht mit Fraßspuren von Insekten (SSPP MA-258, pit 05c).
411_432_Kaiser.fm Seite 422 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
423
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 423 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
424
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
partly allochthonous taphocoenosis and the water body
above does not correspond to the habitat of all these fish-
es. The reduced number of specimens of the clupeo-
morph and of a small unnamed ostariophysan probably
reflects that they were not living in the water column, but
were transported into the central part of the lake during
their taphogenesis. This would also explain why the cich-
lids are most numerous, followed by the osteoglosso-
morphs, because the extant relatives of these taxa also in-
habit the open water column. These observations are
good evidence for the existence of a shallow water area in
the lake basin.
The first described species from Mahenge is the clu-
peomorph
Palaeodenticeps tanganikae
(Greenwood
1960).
Palaeodenticeps
is the fossil relative of extant
Denticeps clupeoides
; both are the only members of the
Denticipitidae, and of the Denticipitoidei. Due to the sim-
ilarities between the fossil and living species,
D. clupe-
oides
is regarded a living fossil. The Denticipitidae are
further interpreted as the sister group of the large subor-
der Clupeoidei. Besides the fossil specimens reported by
Greenwood (1960, 1968), the SSPP recovered a few
more that show differences to those described by Green-
wood and that are presently under description.
The second described species is the osteoglossomorph
Singida jacksonoides
, a sole member of its own family,
Singididae that was referred to the Osteoglossoidei by
Greenwood & Patterson (1967). Although recent phylo-
genetic studies placed
Singida
within the Osteoglossinae,
their sister group is still uncertain.
Singida
appears to be
the sister to the extant genera
Scleropages
and
Osteoglos-
sum
in Li et al.’s (1997) cladogram, but it appears as the
sister of the extant genus
Pantodon
in Hilton’s (2003)
phylogenetic hypothesis.
Murray (2003a, b) described two ostariophysans from
Mahenge. One catfish, identified as a new species of
Chrysichthys
and a characiform, identified as a new ge-
nus and species,
Mahengecharax carrolli
. The assign-
ment to
Chrysichthys
is based, among other characters,
on the presence of long, thick pectoral and dorsal spines.
Numerous catfish specimens are represented in the new
material. Some of them have very long spines. Other
specimens with short spines (Plate 5: Fig. A), which do
not belong within
Chrysichthys
, are under study. In addi-
tion to the catfishes, the new material also comprises un-
described ostariophysans, not assignable to characiforms
or siluriforms.
Apparently, the dominant group populating the waters
at Mahenge were the cichlids (Plate 5: Fig. B). This as-
sumption is based on collections done by previous expe-
ditions (see Murray 2000a) and is confirmed by our ob-
servations based on the newly recovered material. The
Mahenge cichlids were originally identified (Greenwood
1960) as being similar to
Haplochromis bloyeti
(now in
the genus
Astatotilapia
). Later, Greenwood & Patterson
(1967) suggested that these fishes were similar to
Hemi-
haplochromis multicolor
(presently in the genus
Pseudo-
crenilabrus
). Greenwood (1974) identified these fishes as
Haplochromis
only. Recently, Murray (2000b) described
five new cichlid species from Mahenge that were includ-
ed in a new genus,
Mahengechromis
. The diagnostic
characters of the new genus are the distribution of ctenoid
and cycloid scales in certain body regions and the pres-
ence of unicuspid teeth in jaws and pharyngeal bones.
The five species are discriminated by the development
and shape of some cranial bones. Currently, they are in-
terpreted as the oldest known cichlids.
With over 1,300 extant species cichlids are distributed
in Africa, Central and South America, Madagascar, India,
Israel, Syria, Sri Lanka, and Iran (Nelson 1994). The ma-
jority have undergone radiations to form the species
flocks of the great lakes such as Lakes Victoria, Malawi,
and Tanganyika (e.g., Dominey 1984). Most previously
known fossil cichlids are represented by isolated bones or
Plate 4:
Fig. A: Leaf of Acacia mahengensis Herendeen & Jacobs 2000 (SSPP MA-1282, AHS at pit 05c).
Fig. B: Close-up of leaflet.
Fig. C: Undetermined Leguminosae (SSPP MA-2223), related to several isolated leaflets from the SSPP collection.
Fig. D: Undescribed frog of the family Pipidae (SSPP MA-753, pit 05d). The specimen shows pronounced adaptations to aquatic
life.
Figs. E, F: Scarabaeid beetle in dorso-ventral compression (SSPP MA-811, pit 05d). Note the excellent preservation of wing vena-
tion and legs.
Tafel 4:
Fig. A: Blatt von Acacia mahengensis Herendeen & Jacobs 2000 (SSPP MA-1282, AHS, pit 05c).
Fig. B: Nahaufnahme der Blattfieder.
Fig. C: Unbestimmte Leguminose (SSPP MA-2223); ähnlich verschiedenen isolierten Blattfiedern aus dem SSPP-Material.
Fig. D: Unbeschriebener Frosch aus der Familie der Pipidae (SSPP MA-753, pit 05d). Das Tier zeigt ausgeprägte Adaptationen an
das aquatische Leben.
Figuren E, F: Käfer aus der Familie der Scarabaeidae (Blatthornkäfer) in dorso-ventraler Einbettung (SSPP MA-811, pit 05d).
Außergewöhnlich ist die exzellente Erhaltung von Flügelnervationen und Beinen.
411_432_Kaiser.fm Seite 424 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
425
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 425 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
426
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
largely incomplete specimens only (e.g., Greenwood
1957, 1960, 1972, Van Couvering 1977 for Africa, and
Casciotta & Arratia 1993 for South America). Due to the
number of species at Mahenge, Murray (2000b) proposed
that five fossil cichlid species could have been a species
flock. This hypothesis is interesting as it suggests that the
present diversity of cichlids in certain African lakes is not
a recent phenomenon, but also occurred in the oldest so
far known Cichlidae from the Eocene of Mahenge. Since
Mahenge is much smaller (0.2 km
2
proposed surface ar-
ea) than those modern African lakes containing species
flocks, e.g., Lake Victoria (app. 68,000 km
2
and 79 m
deep) and Lake Tanganyika (app. 32,800 km
2
and
1,435 m deep), we can assume that the diversity in differ-
ent habitat types in the Mahenge maar was probably
much smaller than that in the modern Rift Valley lakes.
Murray’s (2000b) hypothesis is based on the comparably
small number of described species and in the observation
that sympatric speciation also took place in small crater
lakes with limited diversity of habitat types (Schliewen et
al. 1994).
In the SSPP material there are also specimens with
ctenoid scales that differ from the typical cichlid mor-
phology. One of these has two dorsal fins separated by a
space; the anterior fin is spiny, while the posterior one is
preceded by a short spine. Another SSPP specimen has
ctenoid scales and differs from all other percomorphs col-
lected at Mahenge in its small size, moon-shaped body,
and very large scales.
In general the fishes from Mahenge represent a unique
phylogenetic puzzle bearing curious combinations of ple-
siomorphic and apomorphic characters. Traditionally,
this fauna would have been interpreted as endemic of
Mahenge. Based on the available information this inter-
pretation may be correct, but because the Eocene African
fish fauna is so poorly known yet, this hypothesis remains
to be tested, when hopefully more material from Ma-
henge and related fossil localities will be known in the
future.
5.4. The oldest known placental mammal of
Sub-Saharan Africa – a bat from Mahenge
The first Eocene Sub-Saharan mammal was discovered
July 24, 2000 at Mahenge pit 05, level 5 (stratigraphy af-
ter Harrison et al. 2001) by Mr. Charles Msuya. It is rep-
resented by the anterior half of a microbat skeleton (
Tan-
zanycteris mannardi
Gunnell et al. 2003) including skull,
lower jaws, vertebral column, both shoulder girdles and
humeri, and right ulna. The hind limbs and most of the
pelvis are missing (Plate 5: Fig. E).
T. mannardi
is a relatively small bat about the size of
Palaeochiropteryx tupaiodon
from the Middle Eocene of
Europe.
T. mannardi
retains a primitive axial skeleton
with unfused vertebrae throughout the column, has an ili-
ac blade that extends dorsally beyond the level of the ili-
osacral articulation, has a scapula with a double-faceted
infraspinous fossa, has a ventrolaterally curving coracoid
with a blunt tip, and has ribs with posterior laminae.
T.
mannardi
differs from all known comparable Eocene mi-
crobats (
Icaronycteris
,
Palaeochiropteryx
,
Archaeonyc-
teris
,
Tachypteron
, and
Hassianycteris
) in having a much
larger cochlear diameter relative to basicranial width. In
most other comparable features
T. mannardi
differs from
archaic microbats and is most similar to Old World rhino-
potamoids and rhinolophoids.
Features shared in common between rhinopotamoids
and
Tanzanycteris
include an enlarged cochlear fenestra,
a clavicle that articulates with the coracoid, and a trochit-
er that extends proximally well beyond the humeral head
(the latter two character states are shared with craseonyc-
terids only, not rhinopotamids). None of these shared
character states is exclusive to
Tanzanycteris
and rhino-
potamids suggesting that they may represent convergenc-
es or retention of plesiomorphic features. Character states
held in common between
Tanzanycteris
and rhinolo-
phoids include an extremely enlarged cochlea, having the
first rib broader than all other ribs, having a clavicle that
articulates with the coracoid, having a trochiter that ex-
Plate 5:
Fig. A: Partially disarticulated siluriform fish (SSPP MA-2222, pit 05).
Fig. B: Lateral view of a cichlid, Mahengechromis (SSPP MA-419, pit 05c).
Fig. C: A non-cichlid perciform (SSPP MA-1288, pit 05c).
Fig. D: Lateral view of an osteoglossomorph (SSPP MA-128, pit 05c).
Fig. E: The type specimen of the bat Tanzanycteris mannardi (Gunnell et al. 2003), pit 05a.
Tafel 5:
Fig. A: Teilweise disartikulierter siluriformer Fisch (SSPP MA-2222, pit 05).
Fig. B: Seitenansicht eines Cichliden, Mahengechromis (SSPP MA-419, pit 05c).
Fig. C: Ein nicht zu den Cichliden gehörender perciformer Fisch (SSPP MA-1288, pit 05c).
Fig. D: Seitenansicht eines Osteoglossomorphen (SSPP MA-128, pit 05c).
Fig. E: Das Typusexemplar der Fledermaus Tanzanycteris mannardi (Gunnell et al. 2003), pit 05a.
411_432_Kaiser.fm Seite 426 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa
427
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 427 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
428
Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
tends proximally well beyond the humeral head, and in
having an iliac blade that flares dorsally resulting in a
well developed iliac fossa. Of these character states, the
presence of a broad first rib is uniquely shared between
Tanzanycteris
and rhinolophoids suggesting that
Tanzan-
ycteris
may be more closely related to this group than to
any other. Among the other characters shared in common
between
Tanzanycteris
and rhinolophoids, all but the flar-
ing iliac blade are shared with a variety of other extant
microbat groups (and several other Eocene taxa). The
flaring iliac blade with a well developed iliac fossa is
likely plesiomorphic for bats. It is possible that
Tanzan-
ycteris
represents a relatively primitive ancestral rhinolo-
phoid but the balance of its other character states aligns it
with the middle Eocene Messel Hassianycteridae.
The extremely enlarged cochlea indicates that Tan-
zanycteris had developed sophisticated echolocation abil-
ities. No other known Eocene microbat has as enlarged a
basal turn of the cochlea as Tanzanycteris. Among extant
forms only rhinolophids and one mormoopid (Pteronotus
parnellii) have comparable cochlear enlargement. These
extant bats are characterized by “High Duty Cycle Con-
stant Frequency Echolocation” in which the pulse and
echo are separated in frequency rather than time. The ma-
jority of extant microbats use low duty cycle echoloca-
tion, a system in which emitted pulses and returning ech-
oes are separated in time. Bats that use the high duty
cycle system can forage for fluttering insects in dense
forest close to vegetation or the ground. This behaviour
is beyond the limits of most low duty cycle echolocating
bats.
While the discovery of Tanzanycteris is interesting
and important for bat palaeobiogeography, the ramifica-
tions of the first mammal found in the Eocene of Sub-Sa-
haran Africa go far beyond the origin and diversification
of Chiroptera. Except for a few places, the early Cenozo-
ic mammalian fossil record of Gondwanan (southern)
continents is poor. The early Tertiary record of mammals
is almost completely unknown from the African conti-
nent except for isolated sites in Egypt, Morocco, Tunisia,
Libya, Senegal, Ethiopia, and Algeria (Arambourg &
Magnier 1961, Savage 1969, 1971, Sudre 1979, Coiffait
et al. 1984, Mahboubi et al. 1986, Sigé et al. 1990, Si-
mons 1992, 1995, Godinot & Mahboubi 1992, 1994,
Gheerbrant et al. 1993, 1998, Godinot 1994, Gheerbrant
1995, Hartenberger et al. 2001, Kappelman et al. 2003).
Other than the Fayum in Egypt, Bassin d’Quarzazate in
Morocco and recently described material from Chilga in
Ethiopia (Kappelman et al. 2003, Sanders et al. 2004),
mammal specimens from these various localities are rep-
resented only by isolated teeth and dental fragments.
The earliest known records of African Paleogene
mammals are from the late Paleocene of Morocco and
represent an interesting mix of African endemic species
with some immigrant taxa that share commonalities with
European groups. There are scarce Early Eocene faunal
samples from Chambi in Tunisia and Morocco that are
dominated by African endemic taxa. Poorly described
faunas from Algeria and Egypt represent all that is known
of Middle Eocene African faunas, while good records of
Late Eocene and Oligocene mammals are known from
several areas in northern Africa and from Chilga in Ethio-
pia. Nearly all African localities of Early Paleogene age
are in the North and were situated along the southern
shoreline of the ancient Tethys seaway. Only Mahenge
and Rukwa Valley in Tanzania and Chilga in Ethiopia of-
fer opportunities to examine what interior continental
vertebrate faunas may have been like. Much of the vast
African continent remains unsampled during the crucial
interval when archaic mammalian faunas were giving
way to more modern mammals.
The phylogenetic and geographic origins of most
modern mammalian groups remain a mystery. While Asia
and Indo-Pakistan have been suggested as possible places
of origin for many, or most, modern mammalian orders
(Krause & Maas 1990, Beard 1998), Africa remains a vi-
able alternative for the geographic origin of at least some
of these groups. In fact, it has been suggested that mod-
ern bats and anthropoid primates may both have originat-
ed in Africa. However, with the exception of the samples
from the late Paleocene in Morocco, the African mamma-
lian record is dominated by taxa already too derived to
have been ancestral to or stem taxa of most modern
groups of mammals. The lack of Paleocene and Early to
Middle Eocene samples from nearly all of Africa prohib-
its analysis of Africa’s possible role in origination of
modern mammals. The importance of discovering new
localities in Africa that represent this important interval
in mammalian evolution cannot be over-stated.
6. Temporal and biogeographic
significance of the Singida Diatremes
The maar lakes of the Singida Region (Fig. 1) offer the
first opportunity to collect an Eocene fauna and flora
from the interior of Africa. When more remains of mam-
mals are found at Mahenge and other Singida crater
lakes, they will provide evidence that may impact a
number of long-standing issues including: the origin and
diversification of several modern orders of mammals, the
extent and development of African mammalian ende-
mism, the implications of Afrotheria (Murphy et al.
2001), and the origin of anthropoid primates and modern
bat families.
A marginal aspect of discovering new Paleocene and
Eocene localities in interior Africa involves the origin of
the unique flora and fauna present on Madagascar today.
The island of Madagascar was in contact with the African
mainland until, at least, the Early Cretaceous (Krause &
411_432_Kaiser.fm Seite 428 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa 429
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Maas 1990) and did not reach its present position until
the Late Cretaceous. During the time that Madagascar
was contiguous with the African continent it was an ex-
tension of what is now Tanzania. Therefore it is almost
certain that the mammals and cichlid fishes present on
Madagascar today shared common ancestry with groups
that would have lived in what is now Tanzania in the Late
Cretaceous. If evidence of common ancestry can be
found in the Eocene of Tanzania and surrounding areas,
the origin of some Malagasy vertebrate groups may be-
come better known.
Until recently, there was almost no strong evidence
that would indicate the potential place of origin for the
Malagasy biota, although Africa (Krause & Maas 1990)
or the Indian subcontinent (Marivaux et al. 2001) were
viewed as possible sources. Seiffert et al. (2003) reported
the discovery of the earliest known strepsirhines primates
(in this case lorises and galagos) from Bartonian (Late
Middle Eocene) sediments in Egypt. Strepsirhines still
live on the African mainland while some groups of strep-
sirhine primates only live on Madagascar today. The dis-
covery of strepsirhines from the Late Middle Eocene of
Egypt suggests that additional evidence from other local-
ities of similar age and older sediments may help to clari-
fy the ultimate origin of this group and also to elucidate
the patterns of faunal interchange that produced the con-
figuration of strepsirhine distribution now present in Afri-
ca and Madagascar.
The role that Africa played in the origin and diversifi-
cation of modern mammalian groups has been given little
serious attention, partly because only little evidence has
been available predating the Late Eocene and from areas
other than those that bordered the ancient Tethys seaway.
The expansion of the African fossil record both temporal-
ly and geographically is the only way to gather additional
evidence that will impact on origination theories. The
same significance applies to the fish fauna and the flora of
Sub-Saharan Africa. Hypothesised divergence times and
biogeographic histories based upon molecular phyloge-
netic studies can only be tested by fossils. The African
maar lakes of the Singida Region offer the potential to
provide much data on Eocene African vertebrates and
plants from interior regions of Sub-Saharan Africa. It is
probable that Africa was a source area for some modern
mammalian groups and plant taxa, but this will not be
known until the African record is much better sampled.
7. Acknowledgements
We gratefully acknowledge the community of Mwaru vil-
lage (Tanzania) in particular all those who worked with us
on the SSPP excavation at Mahenge. We also wish to thank
the local authorities of Singida Region, the Tanzanian
Commission of Science and Technology (COSTECH, Dar-
es-Salaam), the Director of Antiquities (Dar-es-Salaam),
and Mr. Ferdinand Mizambwa (Antiquities Department,
Dar-es-Salaam). We gratefully thank Gerd Alberti and Pe-
ter Michalik (Greifswald) for sharing 8 phantastic weeks
of field work with ES, CM, VB and TMK. Herbert Lutz
(Mainz) is gratefully acknowledged for his helpful com-
ments on an earlier version of this manuscript, and Chris-
toph Zahn (Hamburg) for editorial assistance. We thank
the Deutsch Forschungsgemeinschaft for funding this re-
search (KA 1525 3/1, KA 1525 3/2, AR 275/10-1) and in
particular we wish to thank our sponsors, the companies
Heraeus Kulzer (Dormagen), Carl Zeiss (Wetzlar) und
ZSP Geodätische Systeme (Jena).
8. References
Arambourg, C. & Magnier, P. (1961): Gisements de vertebres
dans le bassin tertiaire de Syrte (Libye). – C. R. Acad. Sci.,
252: 1181–1183, Paris.
Beard, K.C. (1998): East of Eden: Asia as an important center
of taxonomic origination in mammalian evolution. – Bull.
Carnegie Mus. Natur. Hist., 34: 5–39.
Bullwirkel, V. (2003): Organische Petrologie und Mikrofazies
der mitteleozänen Seesedimente des Eckfelder Maares
(Südwesteifel): 123 pp. http://webdoc.sub.gwdg.de/diss/
2003/bullwinkel/index.html.
Casciotta, J. & Arratia, G. (1993): Tertiary cichlids from Ar-
gentina and reassessment of the phylogeny of the New
World cichlids (Perciformes: Labroidei). – Kaupia, 2: 195–
240, Darmstadt.
Coiffait, P.E., Coiffait, B., Jaeger, J.J. & Mahboubi, M. (1984):
Un nouveau gisement a mammifères fossiles d’age Éocène
superieur sur le versant sud des Nementcha (Algerie orien-
tale): decouverte des plus anciens rongeurs d’Afrique. – C.
R. Academie Sci., 299: 893–898, Paris.
Cornen, G., Bandet, Y., Giresse, P. & Maley, J. (1992): The na-
ture and chronostratigraphy of Quaternary pyroclastic ac-
cumulations from Lake Barombi Mbo (West-Cameroon). –
J. Volcanol. Geotherm. Res., 51: 357–374.
Davis, G.L. (1977): The ages and uranium contents of zircon
from kimberlites and associated rocks. – Carnegie Inst.
Washington, Yearbk., 76: 631–635.
Dominey, W.J. (1984): Effects of sexual selection and life his-
tory on speciation: Species flocks in African cichlids and
Hawaiian Drosophila. – In: Echelle, A.A. & Kornfield, I.
(eds.): Evolution of fish species flocks: 231–249, Universi-
ty of Maine (Orono Press).
Eades, N.W. (1936): The geology of the Iramba Plateau. –
Short paper; Geol. Surv. Tanganyika, 15: 1–33, Dar es Sa-
laam.
Gheerbrant, E. (1995): The Paleocene mammals from the
Ouarzazate Basin (Morocco) 3, Adapisoriculidae and other
mammals (Carnivora, ?Creodonta, Condylarthra, ?Ungula-
ta et incertae sedis). – Palaeontographica, Abt. A, Palaeo-
zool.-Strat., 237: 39–132.
Gheerbrant, E., Cappetta, H., Feist, M., Jaeger, J.-J., Sudre, J.,
Vianey-Liaud, M. & Sigé, B. (1993): La succession des
411_432_Kaiser.fm Seite 429 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
430 Thomas M. Kaiser et al.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
faunes de vertebres d’age Paleocène superieur et Éocène
inferieur dans la basin d’Ouarzazate, Maroc. Context
géologique, portée biostratigraphique et paleogéo-
graphique. – Newsl. on Stratigraphy, 28: 33–58.
Gheerbrant, E., Sudre, J., Sen, S., Abrial, C., Marandat, B.,
Sigé, B. & Vianey-Liaud, M. (1998): New data on Thane-
tian and Ypreseian mammals of the Ouarzazate Basin (Mo-
rocco) and their stratigraphic context. – Palaeovertebrata,
27: 155–202.
Giresse, P., Maley, J. & Kelts, K. (1991): Sedimentation and
palaeoenvironment in crater lake Barombi Mbo, Cam-
eroon, during the last 25,000 years. – Sediment. Geol., 71:
151–175.
Godinot, M. (1994): Early North African primates and their
significance for the origin of Simiiformes (=Anthropoi-
dea). – In: Kay J.G. & Fleagle, R.F. (eds.): Anthropoid ori-
gins: 235–295, New York (Plenum).
Godinot, M. & Mahboubi, M. (1992): Earliest known simian
primate found in Algeria. – Nature, 357: 324–326, Lon-
don.
Godinot, M. & Mahboubi, M. (1994): Les petits primates
simiiformes de Glib Zegdou (Éocene inférieur du moyen
d’Algérie). – C. R. Academie Sci., 319 (II): 357–364, Pa-
ris.
Greenwood, P.H. (1957): Fish remains from the Mumba cave,
Lake Eyasi. – Mitt. Geol. Staatsinst. Hamburg, 26: 125–
130, Hamburg.
Greenwood, P.H. (1960): Fossil denticipitid fishes from East
Africa. – Bull. Brit. Mus. Natur. Hist., Geol., 5: 1–11.
Greenwood, P.H. (1968): The osteology and relationships of
the Denticipitidae, a family of clupeomorph fishes. – Bull.
Brit. Mus. Natur. Hist., Zool., 16 (6): 215–273, London.
Greenwood, P.H. (1972): Fish fossils from the late Miocene of
Tunisia. Travaux de Géologie Tunisienne no. 6. – Extrait
des Notes du Service Géologique, 37: 41–72, Tunis.
Greenwood, P.H. (1974): Review of Cenozoic freshwater fish
faunas in Africa. – Annals geol. Surv. Egypt, 4: 211–232.
Greenwood, P.H. & Patterson, C. (1967): A fossil osteoglos-
soid fish from Tanzania (E. Africa). – Zool. J. Linn. Soc.,
47: 211–223, London.
Gregory-Wodzicki, K.M. (2000): Relationships between leaf
morphology and climate, Bolivia: implications for estimat-
ing paleoclimate from fossil floras. – Paleobiology, 26:
668–688.
Gunnell, G.F., Jacobs, B.F., Herendeen, P.S., Head, J.J., Ko-
walski, E., Msuya, C.P., Mizambwa, F.A., Harrison, T.,
Habersetzer, J. & Storch, G. (2003): Oldest placental
mammal from sub-Saharan Africa: Eocene microbat from
Tanzania – evidence for early evolution of sophisticated
echolocation. – http://palaeo-electronica.org/2002_2/afri-
ca/main.htm.
Haggerty, S.E., Raber, E. & Naeser, C.W. (1983): Fission track
dating of kimberlitic zircons. – Earth and Planetary Sci.
Letters, 63: 41–50.
Harrison, T., Msuya, P., Murray, A.M., Jacobs, B.F., Báez,
A.M., Mundil, R. & Ludwig, K.R. (2001): Paleontological
investigations at the Eocene locality of Mahenge in North-
Central Tanzania, East Africa. – In: Gunnel, G.F. (ed.): Un-
usual occurrences and rarely sampled habitats: 39–74,
New York (Kluwer).
Hartenberger, J.-L., Crochet, J.-Y., Martinez, C., Marandat, B.
& Sigé, B. (2001): The Eocene mammal fauna of Chambi
(Tunisia) in its geological context. – Topics in Geobiology,
18: 237–250.
Herendeen, P.S. & Jacobs, B.F. (2000): Fossil legumes from
the middle eocene (46,0 Ma) Mahenge flora of Singida,
Tanzania. – Amer. J. Botany, 87 (9): 1358–1366.
Hilton, E.J. (2003): Comparative osteology and phylogenetic
systematics of fossil and living bony-tongue fishes (Actin-
opterygii, Teleostei, Osteoglossomorpha). – Zool. J. Lin.
Soci., 137: 1–100.
Jacobs, B.F. (1999): Estimation of rainfall variables from leaf
characters in tropical Africa. – Palaeogeogr., Palaeoclima-
tol., Palaeoecol., 145: 231–250.
Jacobs, B.F. (2002): Estimation of low latitude paleoclimates
using fossil angiosperm leaves: examples from the Mio-
cene Tugen Hills, Kenya. – Paleobiology, 28: 399–421.
Jacobs, B.F. & Herendeen, P.S. (2004): Eocene dry climate
and woodland vegetation in tropical Africa reconstructed
from fossil leaves from northern Tanzania. – Palaeogeogra-
phy, Palaeoclimatology, Palaeoecology, 213: 115–123.
Kaiser, T.M., Alberti, G., Bullwinkel, V., Michalik, P., Msuya,
C. & Schulz, E. (2003): Mahenge – Ein Fenster zum Mit-
teleozän Afrikas. – Naturwiss. Rdsch., 664: 540–546.
Kappelman, J., Rasmussen, D.T., Sanders, W.J., Feseha, M.,
Bown, T.C.P., Crabaugh, J., Fleagle, J., Glantz, M., Gor-
don, A., Jacobs, B., Maga, M., Muldoon, K., Pan, A., Pyne,
L., Richmond, B., Ryan, T., Seiffert, E.R., Sen, S. & Todd,
L.E. (2003): Oligocene mammals from Ethopia and faunal
exchange between Afro-Arabia and Eurasia. – Nature, 226:
549–552.
King, L. (1962): The morphology of the Earth: 577 pp., Edin-
burgh (Oliver and Boyd).
Krause, D.W. & Maas, M.C. (1990): The biogeographic ori-
gins of late Paleocene-early Eocene mammalian immi-
grants to the Western Interior of North America. – In:
Bown, T.M. & Rose, K.D. (eds.): Dawn of the Age of
Mammals in the northern part of the Rocky Mountain Inte-
rior, North America, Geol. Soc. America, Spec. Pap., 243:
71–105.
Krell, F.T. (2004): Catalogue of fossil Scarabaeoidea (Coleop-
tera: Polyphaga) of the Mesozoic and Tertiary. – http://
www-museum.unl.edu/research/entomology/fossilrecord-
web.pdf.
Li, G.-Q., Wilson, M. & Grande, L. (1997): Review of Eohio-
don (Teleostei: Osteoglossidae) from the Eocene of North
America and their phylogenetic relationships. – J. Vert.
Paleontol., 71: 1109–1124.
Lorenz, V. (1973): On the formation of maars. – Bull. Volcan-
ol., 37: 183–204.
Mahboubi, M., Ameur, R., Crochet, J.-Y. & Jaeger, J.-J.
(1986): El Kohol (Saharan Atlas, Algeria): a new Eocene
mammal locality in northwestern Africa. Stratigraphical,
phylogenetic, and paleobiogeographical data. – Palaeonto-
graphica, Abt. A, Pa-laeozool.-Stratigr., 192: 14–49.
Mannard, G.W. (1962): The geology of the Singida kimberlite
pipes, Tanganyika. – Ph. D. thesis, McGill University: 348
pp., Montreal.
Marivaux, L., Welcomme, J.-L., Antoine, P.-O., Metais, G.,
Baloch, I.M., Benammi, M., Chaimaneii, Y., Ducrocq, S.
411_432_Kaiser.fm Seite 430 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
The maar lake of Mahenge (Tanzania) – unique evidence of Eocene terrestrial environments in sub-Sahara Africa 431
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
& Jaeger, J.J. (2001): A fossil lemur from the Oligocene of
Pakistan. – Science, 294: 587–591.
McKay, I.J. & Rayner, R.J. (1986): Cretaceous insects from
Orapa, Botswana. – South African J. Entomol., 49: 7–17.
Murphy, W.J., Eizirik, E., Johnson, W.E., Zhang, Y.-P., Ryder,
O.A. & O’Brien, S.J. (2001): Molecular phylogenetics and
the origins of placental mammals. – Nature, 409: 614–618.
Murray, A.M. (2000a): The Paleozoic and Early Cenozoic
fishes of Africa. – Fish and Fisheries, 1: 111–145.
Murray, A.M. (2000b): Eocene cichlid fishes from Tanzania,
East Africa. – J. Vert. Paleontol., 20 (4): 651–664.
Murray, A.M. (2003a): A new species of catfish (Claroteidae,
Chrysichthys) from an Eocene crater lake in East Africa. –
Canad. J. Earth Sci., 40 (7): 983–993.
Murray, A.M. (2003b): A new characiform fish (Teleostei: Os-
tariophysi) from the Eocene of Tanzania. – Canad. J. Earth
Sci., 40 (4): 473–481.
Naeser, C.W. & McCallum, M.E. (1977): Fission-track dating
of kimberlitic zircons. – Ext. Abstr., 2
nd
Int. Kimberlite
Conf., American Geophys. Union, Washington, D.C.
Nelson, J.S. (1994): Fishes of the World: 600 pp., New York
(Wiley).
Paulian, R. (1976): Three fossil dung beetles (Coleoptera,
Scarabeaidae) from the Kenya Miocene. – J. East Africa
Natur. Hist. Soc., Nat. Mus., 31 (158): 1–4.
Rayner, R.J. (1987): March flies from an African Cretaceous
springtime. – Lethaia, 20: 123–127.
Rayner, R.J. & McKay, I.J. (1986): The treasure chest of the
Orapa diamond mine, Botswana. – Notes and Records, 18:
55–61.
Rayner, R.J., Waters, S.B., McKay, I.J., Dobbs, P.N. & Shaw,
A.L. (1991): The mid-Cretaceous palaeoenvironment of
central Southern Africa (Orapa, Botswana). – Palaeogeo-
gr., Palaeoclimatol., Palaeoecol., 88: 147–156.
Sanders, W.J., Kappelman, J. & Rasmussen, D.T. (2004): New
large-bodied mammals from the late Oligocene site of
Chilga, Ethiopia. – Acta Palaeontol. Polonica, 49 (3): 365–
392.
Savage, R.J.G. (1969): Early Tertiary mammal locality in
southern Libya. – Proc. Geol. Soc., London, 1657: 167–
171, London.
Savage, R.J.G. (1971): Review of the fossil mammals of Lib-
ya. – In: Gray, C. (ed.): Symposium on the geology of Lib-
ya: 215–226, Tripols (University of Libya).
Schliewen, U.K., Tautz, D. & Pääbo, S. (1994): Sympatric
speciation suggested by monophyly of crater lake cichlids.
– Nature, 368: 629–632.
Schmitz, M. (1991): Die Koprolithen mitteleozäner Vertebra-
ten aus der Grube Messel bei Darmstadt. – CFS Courier
Forschungsinst. Senckenberg, 137: 1–199.
Seiffert, E.R., Simons, E.L. & Attia, Y. (2003): Fossil evidence
for an ancient divergence of lorises and galagos. – Nature,
422: 421–424.
Sigé, B., Jaeger, J.J., Sudre, J. & Vianey-Liand, M. (1990): Al-
tiatlasius koulchii n. gen. et sp., primate omomyidé du
Paléocène supérieur du Maroc, et les origines des Eupri-
mates. – Palaeontographica, A 214: 31–56.
Simons, E.L. (1992): Diversity in the early Tertiary anthropoi-
dean radiation in Africa. – Proc. Nat. Acad. Sci., USA, 89:
10743–10747.
Simons, E.L. (1995): Egyptian Oligocene primates: a review. –
Yearbook of Physical Anthropology, 38: 199–238.
Smith, R.M.H. (1986): Sedimentation and palaeoenvironments
of Late Cretaceous crater-lake deposits in Bushmanland,
South Africa. – Sedimentology, 33: 369–386.
Sudre, J. (1979): Nouveaux mammifères éocènes du Sahara
occidental. – Palaeovertebrata, 9: 83–115.
Teale, E.O. (1931): Shinyanga diamond fields. – Short paper;
Geol. Surv. Tanganyika, 9: 5–6.
Teale, E.O. (1932): The kimberlite and associated occurrences
of the Iramba Plateau. – Short paper; Geol. Surv. Tangan-
yika, 10: 1–10.
Van Couvering, J.A.H. (1977): Fossil cichlid fish of Africa. –
Spec. Pap. Palaeontol.: 103 pp., London.
Wilf, P., Wing, S., Greenwood, D. & Greenwood, C. (1998):
Using fossil leaves as paleoprecipitation indicators: an
Eocene example. – Geology, 26: 203–206.
Williams, G.J. (1939): The kimberlite province and associated
diamond deposits of Tanganyika Territory. – Bull. Geol.
Surv. Tanganyika, 12: 1–38, Dar es Salaam.
Manuscript received: 08.07.2005
Accepted for printing: 23.05.2006
411_432_Kaiser.fm Seite 431 Freitag, 15. Dezember 2006 9:32 09
eschweizerbartxxx
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
411_432_Kaiser.fm Seite 432 Freitag, 15. Dezember 2006 9:32 09
