Palaeogeography, Palaeoclimatology, Palaeoecology, 43 (1983): 129--151
129
Elsevier Science Publishers B.V., Amsterdam -- Printed in The Netherlands
LACUSTRINE PALEOCHEMICAL INTERPRETATIONS
EASTERN AND SOUTHERN AFRICAN OSTRACODES
BASED ON
ANDREW S. COHEN', RICHARD DUSSINGER 2 and JOHNATHAN RICHARDSON 3
'Department of Geology, The Colorado College, Colorado Springs, CO 80903 (U.S.A.)
2R.D. 8 Box 304, Gettysburg, PA 17325 (U.S.A.)
~Department of Biology, Franklin and Marshall College, Lancaster, PA 1 7604 (U.S.A.)
(Received November 23, 1982; revised and accepted April 7, 1983)
ABSTRACT
Cohen, A. S., Dussinger, R. and Richardson, J., 1983. Lacustrine paleochemical interpretations based on Eastern and Southern African ostracodes. Palaeogeogr., Palaeoclimatol.,
Palaeoecol., 43 : 129--151.
Ostracode assemblages in modern African lakes reflect water chemistry variations and
so are potentially useful as paleochemical indicators. The water chemistry ranges of 33
m o d e m ostracode taxa have been evaluated for five parameters: total conductivity (K20),
Na ÷, Ca 2+, C1- and alkalinity (CO~- + HCO~). Strong correlations exist for distribution
patterns of ostracodes across all of these parameters except Ca 2÷. Four, somewhat arbitrary,
Ostracode Range Assemblages can be defined, based upon increasing alkalinity and
salinity, and named for characteristic taxa. These are:
Range
Range
Range
Range
I, the Sternocypris assemblage (K20 < 500 umho)
II, the Mecynocypria assemblage (K~0 = 500--1500 umho)
III, the Gomphocythere assemblage (K20 = 1500--4000 ~mho)
IV, the Limnocythere assemblage (K20 > 4000 umho)
An application of the ostracode typology to the fossil record, using a core from Lake
Nakuru (Kenya) as an example, suggests a fluctuating alkaline lake prior to about 10,000
yr. B.P., followed by fresh-water conditions until about 8000 yr. B.P. and finally a return
to higher alkalinity throughout the remainder of the Holocene. Since the core has also
been studied for paleochemical interpretation based upon diatoms, an independent verification of the technique is possible. The two interpretations agree through 80% of the
core, suggesting a good reliability for the method.
INTRODUCTION
There has been a considerable growth of interest in recent years among
paleolimnologists in developing methodologies for the determination of
lacustrine paleochemistries. These methodologies can be divided into two
classes:
(1) G e o c h e m i c a l - t y p o l o g i c a l approaches, w h e r e s p e c i f i c a u t h i g e n i c m i n e r a l
a s s e m b l a g e s axe c o n s i d e r e d i n d i c a t i v e o f s t a b i l i t y f i e l d s w i t h i n a g i v e n r a n g e
of ionic c o n c e n t r a t i o n s . This t e c h n i q u e has m o s t successfully b e e n applied to
0031-0182/83/$03.00
© 1983 Elsevier Science Publishers B.V.
130
rocks interpreted to have formed from highly concentrated brines (Cerling,
1977). Its utility is considerably reduced in more " n o r m a l " lacustrine environments, that is to say, in situations where metazoans are likely to be c o m m o n
and diverse.
(2) Fossil-typological approaches, where a series of fossil assemblages is
used to indicate a particular overlapping range of ecological tolerances to a
given ionic parameter. This class of techniques is o f greater interest to the
lacustrine or estuarine paleobiologist, w h o is trying to determine the habitat
requirements and niche breadth of fossil organisms. To date, most studies of
this t y p e have centered around diatom floras.
In this paper, we will apply a technique falling in the second of these
classes, lacustrine ostracode t y p o l o g y , to the highly variable inland waters of
Eastern and Southern Africa. After discussing the modern distribution of
ostracode genera and species in African lakes, as they relate to water chemistry, we will illustrate h o w the technique m a y be applied to the fossil record
of Eastern African paleolakes, in an effort to unravel their paleochemical
histories. Whereas the present study is directed towards African paleolimnology, it should apply (as a methodology, rather than in its specific details)
to any region with lacustrine deposits, where modern and fossil ostracode
assemblages are analogous.
The African inland water system, as a whole, is perhaps one of the better
localities in which to c o n d u c t such a study. Temperature-induced assemblage
variations are minimized in the lower~levation (under 2000 m) equatorial to
subequatorial regions of Africa. Water chemistries, on the other hand, do
vary enormously in this region, and the proximity of these lakes and ponds
to ancient ostracode-bearing lacustrine deposits makes the region ideally
situated for this t y p e of investigation. The reader is directed to Delorme
(1971a, b) and Delorme et al. (1977) for other examples of ostracode typology, as applied to Canadian inland waters, and for a more general review of
the paleoecological significance of fresh-water ostracode fossils. More recently,
Carbonel and P e y p o u q u e t (1979) and P e y p o u q u e t and Carbonel (1980) have
published some ostracode-based paleochemical interpretations for the Pliocene--Pleistocene, O m o Basin of southern Ethiopia.
An index map (Fig.l) of Eastern Africa identifies the lakes discussed in
this paper.
METHODS
Water chemistry and ostracode distribution data were collected b o t h in
field research, during the summers o f 1978 and 1979, and from previously
published data. A complete list of ostracode distribution papers used in this
study can be obtained from the senior author on request. Only taxa with
widespread modern distributions were considered for inclusion in this study.
A taxon was only included if it had been recorded from three or more localities with appropriate water chemistry data. The taxa under consideration
131
7
i
F i g . 1 . L o c a t i o n m a p f o r l a k e s d i s c u s s e d in t h e t e x t : 1 = L a k e A b i a t a ; 2 = L a k e T u r k a n a ;
3 = L a k e N a k u r u ; 4 --- L a k e E I m e n t a i t a ; 5 = L a k e N a i v a s h a ; 6 = L a k e V i c t o r i a ; 7 = L a k e
Albert; 8 = Lake Tanganyika; 9 = Lake Chilwa.
were further reduced to include only those with known or suspected fossil
records from the African Neogene.
Ideally, the taxonomic level most appropriate for this type of study is the
species. Any species will have a tolerance range for any given parameter
equivalent to or narrower than its encompassing higher taxon (subgenus or
genus in this case). There is, however, considerable difficulty in identifying
fresh-water ostracode fossils to the specific level, as most neontological
taxonomic schemes for these organisms rely more heavily on soft part, nonpreservable morphology, than on carapace, preservable morphology. Therefore, only highly distinctive forms are discussed in this study at the species
level. This reduces the precision of the technique considerably, b u t improves
the accuracy of the results. An inspection of the data in Figs.2--6 (in particular for the genus Gomphocythere), shows that even highly eurytopic genera
m a y include moderately stenotopic species, which, taken as a group, define
the total, wide range o f the genus. Indeed, within a particular adaptive complex, represented here b y a given genus, some ostracodes are k n o w n to show
considerable environmental plasticity (Delorme, 1969) as individual species.
Thus, the "ranges" shown for genera here should n o t be construed to be
132
equivalent to the range for a fossil species within that genus, but rather to
encompass that individual species' probably more limited range.
Water chemistry data
The sources for water chemistry data are Wayland (1925), Worthington
(1929, 1930, 1932), Omer-Cooper (1930), Phillips (1930), Beadle (1932),
Damas (1937, 1954, 1964), Ricardo (1939), Loffredo and Maldura (1941),
Ricardo-Bertram et al. (1942), Fish (1952, 1953, 1954), Van Meel (1953,
1954), Dubois (1955), Talling (1957, 1963), Verbeke (1957), Van der Ben
(1959), De Kimpe {1960), Harding (1963), Livingstone (1963), Talling and
Talling (1965), Kilham (1971), Visser (1974), and Livingstone and Melack
(1979).
In most, but not all cases, the water analyses were made at the same time
as the ostracode collections. Such data were used preferentially, but were
insufficient to be used exclusively in some of the large open lakes. Here,
however, variations in water chemistry are small, relative to the tolerance
ranges of their local ostracode faunas. This is demonstrated by comparing
total chemical ranges for a given species with the variability of the same
chemical parameter at the lake in question.
Five chemical parameters are considered in the present study: Na÷, Ca 2÷,
Cl-, CO~- + HCO~ alkalinity, and total conductivity (K20). Na+, Ca2+, and
C1- are expressed in mg/1 concentration, while alkalinity is expressed in meq/1
by convention. Conductivity is expressed in pmho/cm at 20°C, and is
approximately proportional to total ionic salinity in the range of salinities
considered in this paper [the conversion rate being K:0 ~ salinity (p.p.t.) X
103]. Conductivity data are used because they are more readily available
than total salinity data.
The choice of parameters is partially dependent on data availability and
partially on a consideration of the importance of particular ions for crustacean biology (Beadle, 1943). Beadle (1974, ch. 5) has discussed the importance of mineral composition to the ecology of African inland waters at
length, so only a few points will be considered here. Most Eastern African
waters are enriched in Na÷, CO~- + HCO~, and CI-, probably due to weathering of their alkaline volcanic-bearing watersheds and to arid climate fractionation processes in their feeder streams (Eugster, 1980).
Alkalinity, at higher values, regulates the solubility of Ca~+. Above approximately 80 meq/l total alkalinity, Ca 2÷ is available in solution only in extremely
small amounts (usually less than 1 mg/1). While no experimental evidence
exists to determine the minimal concentration of Ca :+ necessary for ostracode skeletal formation, distribution data collected in this study suggest that
it is approximately 1 mg/l. In waters with less calcium, the crustacean fauna
is composed exclusively of forms with less calcified, more chitinous exoskeletons. Beadle (1974) has noted that in the extremely calcium-poor waters
of the Amazon River, conchiolin is substituted for calcite in some bivalves,
133
but this type of p h e n o m e n o n has yet to be observed among ostracodes (or
any highly calcified crustacean to our knowledge) in Africa or elsewhere.
Additional data for major ions (particularly Mg 2÷, K ÷, and SO~- ) would be
extremely valuable for this type of study, b u t are, unfortunately, unavailable
for m a n y of the lakes and rivers considered in this study. The same may be
said of major nutrients (total P, N) and trace elements. For example, La
Barbera and Kilham (1974) have suggested that F- concentrations may be
very significant in structuring the distribution of cladocerans and copepods
in African lakes. U n d o u b t e d l y , the parameters considered in this study are
not the only chemical ones of importance in the distribution of ostracode
species, b u t the consistency of the results presented here indicates that t h e y
are very important determinants of these distributions.
It should be emphasized t h a t the data presented here are not the result of
experimental manipulation b u t rather are empirical distributions of natural
ostracode occurrences. Though it would be of great interest to have such
experimental data in order to assign absolute or lethal limits on tolerance
range for these species, we are uncertain about the significance of such results
for natural distributions. Many of the ions vary in concert in natural waters
and their effects may be synergistic and/or inseparable by conventional
experimental design. Furthermore, we are primarily concerned with the types
of waters which these species inhabit, rather than the ones which they might
theoretically inhabit in the absence of competition and predation. Therefore,
the limits presented here are unlikely to be lethal limits for the individual.
For the paleoecologist, it will ultimately be more interesting to say t h a t the
fauna of a given lacustrine deposit is similar to t h a t of a given modern lake
(which can then be used as a basis for further models of the paleolake) than
to assign a set o f paleochemical values to a given assemblage divorced from
any observable, modern lake model. Assemblage ranges can easily be constructed from these data which have no m o d e r n analogue, and in fact lie outside of the theoretically possible range of water chemistries. For instance, a
high-Ca 2÷ high-alkalinity water chemistry could be construed from a
Gomphocythere--Limnocythere assemblage (a c o m m o n fossil assemblage)
given their overlapping ranges (see Figs.4 and 5), b u t in fact such chemical
conditions are unlikely to occur in natural waters. Thus, the interpretation
of assemblages must be judiciously tempered with knowledge of the types of
waters which are likely to occur and those which are not (see for example
Hardie and Eugster, 1970; and Eugster and Jones, 1979). Most ostracode
assemblages suggest a wide range of possible water chemistries. Their pro bable
ranges are usually much more limited. We will return to this point in the
discussion of our results, with some specific examples.
RESUL TS
Figs.2 through 6 illustrate the k n o w n chemical ranges (for the parameters
previously mentioned) of 33 c o m m o n ostracode species and genera. The
135
RANK
ORDER
Zonocy[~ris costa la
ACOC~
i
h~a
2
3
Slenocypris culvirami
i
2,1
1
Gom~h 9c~'t he re obtusata
Pat astp_noc~pris e ~ m a t
11
29
?
4
a
5
St randesia caudata
6
All Zonocyprrs
7
All Mecynocypria
8
All S t e n9.c~LI~
g
Stenoc~pds ~
10
S t enoc y ~ j ~ plat yba ~is
12
Ca n d ~ s i s
13
afr=cana
Stenoc~tprLs iu~od~
47
?
49
14 5
1101
,
, ,
,,,,io~
Na~(rag/L)
RANK
IJ
11
ORO~R
Cyprinotus congener
145
Dar w + ~ a stevenso~
16
!
69
719
g3
!
(Physocyprla) capensi$ 17
18.5
914
All S cler Ocyprls
18.5
914
All OncocyprFs
20,5
Oncocy r ~
20.5
All G ~ g c
y~here
t
worthingtom
Hempcypris f ossyla tu§
22.5
All Cyprinot us
22.5
102
?
102
107
126
Zono¢~
17,
25
la~
Onc ocj',~js omercooper,
26
All Strandesia
275
Potamoc y ~ s product a
27.5
I
?
182
182
I IIIIJ]1
"11o,
10
Na~(mgJI )
RANK
1o'
102
101
,i
i
All LImnoc•there
29
2,18
Gor~hocythere an~ulata
3O
2~8
All PotamocyDris
31
timnoc~th~a mlchaelseni
32
C ypridel~ tgEgJ~
33
290
?
?
292
28~o
'
i'
Na÷(mg./I.)
'11~2
103
,~,
Fig.3. N a + c o n c e n t r a t i o n ranges (expressed in rag/l) for selected African ostracode taxa,
based o n natural distributions. R a n k order values are based on increasing logarithmic
m e a n s o f m i n i m u m and m a x i m u m c o n c e n t r a t i o n occurrences.
136
101
RA..
10°
ORDER
Limnocythere michaelseni
1
Gomphocythere obtusata
2
i i ,, , I
5.6
? 7.2 ?
A, S c ~
4.5
10.5
All Potarnocypris
4.5
10.5
~
4.5
IO.5
s
producta
Gompho¢//t here angulata
4,5
Hemic ypris fossulatus
7
Zonoc)'pris costata
8
10.5
11.O
I
111.9
Zonocypris ~
10
121"5
All C ~ o t u s
10
1~.5
10
5
121
Onco~
wor t hingtonF
All AIIoc y ~
12.6
12
Paraslenoc~pris perarmata
13
1~0,.,
All Z o no ¢~yjp~FS
16
13.6
1
tDo
.......
1o1
Ca+ '~mg./l.)
RANK
100
ORDER
All Gomphocythere
C~(Physoc~pria)
I
101
i
, i
capensls 16
CandonopslS afrlcana
16
i
I
13.6
j
16
13.6
13.6
All Mecynocy~ria
16
Darwinula stevensoni
19
13,6
All Onco~ypris
20
14,1
All Strandesia
21
14.5
13.8
3tenocypris plalybasis
22.5
15.0
$tenocypriscurvirami
22,5
15.0
A~I Stenoc~p~$
24
16,9
StenocyprJs~
26
1~.3
Stenocywisstagnalis
26
1~.3
Cyprinotus congener
27
17.3
Strandesia caudata
23
1,~
1
RANK 1~
ORDER I
Oncocypri_s omercooperi
Ca-*(mg./I.)
........
10'
I
20.2
i
22.4
29
Acocvwis ha ~ a
30
avocvvr~s
31
All Limno¢t h ~
32
Cvorideis toross
33
lO~
,0~
. . . . . .
I
I
I
I
l
l~iJil
r
~
i iI1~1
j
4~.7
45.0
I
45.3
I#
, , , ,~,,,11
10
p , , ,~,~,11~
I
,03
Ca, ,(mgJl.)
F i g . 4 . C a ~+ c o n c e n t r a t i o n
ranges (expressed in rag/l) for selected African ostracode taxa,
b a s e d o n n a t u r a l d i s t r i b u t i o n s . R a n k o r d e r v a l u e s are b a s e d o n i n c r e a s i n g l o g a r i t h m i c
means of minimum and maximum concentration values.
137
RANK
ORDER
Z onocy~tls costata
1
Pa r a st e n oc~vJ)jts perarma!a
2
C a ndgrmps~ ~f [ic ana
3
S~randesia caudata
4
All Sten~cyprl $
5.5
~tenoc p ~
55
S~e n o ~ r
~ta~la!~
IS platybasl~
100
lOl
i
114
12.0
i
~130
13.1
t3.1
13.4
7
Gomphoctj~he[e abtu§at§
,o2
lO
r38
i
8
~5 5
Stenocypr!s curv)rarnl
10
S t e n o c / p n s junod,
11
Cypr,netus conaener
12.5
Zon£E~rls 9~abbr,~
12.5
All ZOflOCypn~
17.9
21.9
222
22.2
224
14
i
......
io~
, ,
, io2
,
=
cl (mg./i)
RANK
ORDER
,j
100
I
ii
,o2
25.4 ?
=CyJ~rm ( ~ p r i a )
c a~}ensts
AC OCy~£t8 h y l a ~
19
All Cyptinotus
19
Hemicypri8 IoossuLatua
19
All Oncocypris
19
One oc ypris wgrthinqt oni
19
All Gomphocythere
22.5
All Sclerocy[~rL§
22.5
All Po Lam 9¢~y~r LS
24
OncocyprLs omercoopefl
25
All StrandesLa
26
Pg~m ~£~s
=oroducta
Da~winula st evensonl
324
16
?
38.0
38.o
36.0
38.0
L
38.0
I
43.0
45.2
1
63.0
68.6
69.0
27
821
28
t=,=
t
i
ii
101
102
4
,02
ct (reg./L)
RANK
ORDER
Gomphocytherean~ta
29
Limnocythere michaelseni
3G
o
1~
!;
32
Cyprideistorosa
33
17
92i.0
108
31
AIIL~mnocythere
18
2~5
2slao
,oo
lo~
,°2
Io3
,o"
Cr(mg./I.)
Fig.5. C]- concentration ranges (expressed in mg/l) for selected African ostracode taxa,
based on natural distributions. R a n k order values are based on increasing logarithmic
means of m i n i m u m and m a x i m u m concentration values.
138
RANK
lO 0
I
..... 11111
costata
~zI~J~C3LOL~ curvirami
caudata
~ o h o c v t h e r e obtusala
Ilvocvoris g J J ~
101
i
i
i
r
I Jill
2.00
2
2.31
2.55
3
I
3.13
I
4
55 ?
3.)50
5.5
3.50
I
3.60
7
~LLP~no c v o r is olat vbasis
All Stenocvoris
9
~
~o
t~z0~J
All ,7onocvoris
I
3.66
,.,p-
s
,>
3.67
I
3.93
r
4.~)4
11
4.05
Parastenocvoris oerarmata 12
All ~clerocvBrls
13
Slen~cvorls ~
14
4.24
1
5.00
.....
1!0
r , I, 'r1~11
ALKAL~NITY(meq/L)
RANK
ORDER
Cyprln0tus c ~ r
15
Candonopsis africana
|6
All MecynocyprLa
17
All StranOesla
18
All Gomphocyt here
19
All C~'prlnolu$
20
Hemlcvorrs fossulatus
21
O;~rwinula stevensoni
22
o
Io
11
,
l
5F46
5.69
5 .t9 4
6.1oo
9
6.38
?
6.j39
p
?
r.loo
7.123
(PnvsocvDrla) Dapensis 23
All AIIOC~,prla
25
potam~cxwis •roducl•
26
All L pmn,~cyther e
27
Qnocypr=s
28
omercooperl
7,40
24
AJ~ Polamocyptis
Jl,I
, 1,1=I
5.45
,--f,-7,72
I
7 .j73
?
9.16
°o
1~'112
.......
I
10
I
ALKALINITY(meq
/I )
2
ORDER
I
,
,
,
I:,,1
All Oncocypri$
26
9 177
Zonocyp rl$ ~
30
11 ~ O
GomphocytherQ angulala
31
Oncoc¥ r~J~ wor thingtoni
32
LlmnO~yfher @ mK]haelseni
33
12.52
1272
13,j93
poo
.......
Po~
'
......iP02
ALK ALINITY(meq./L)
Fig.6. Alkalinity (CO~- + H C O ~ ) concentration ranges (expressed in meq/l)for selected
African ostracode taxa, based on natural distributions. R a n k order values are based on
increasing logarithmic means of m i n i m u m and m a x i m u m concentration occurrences.
139
k n o w n distribution of each species or genus m a y b e obtained from the senior
author on request. For each Figure, the ostracode taxa are arranged in order
of increasing logarithmic means (between maxima and minima) in their ranges
for the chemical species in question. These means should n o t be interpreted
as either the optimal or median portion of the ostracode's range. Rather, their
importance lies in allowing a comparison of the ordering of species between
chemical parameters. Using this information, it is possible to determine those
parameters, if any, which can be isolated from the others as having unique
effects on the types of taxa present as ionic concentration is increased, and
those which are n o t empirically separable using natural distribution data alone.
All species and genera are assigned a rank number in order of increasing
ionic concentration of their means, for each parameter (see Figs.2--6). A
Spearman Rank Order Correlation was then performed for each pair of parameters. These are illustrated in matrix form in Fig.7. Na ÷, C1- and K20 are all
very highly and significantly correlated (p < 0.01), indicating that: (a) ostracode assemblages vary uniformly along concentration gradients for each of
these three parameters; and (b) using distribution data alone, the effects of
any one of these parameters cannot be factored out from the other two.
Thus, for instance, a high K20, low Na ÷ ostracode fossil assemblage could not
be detected from these data. In fact, such noncoordinate variation in these
three water chemistry parameters does not presently exist in Eastern Africa.
Alkalinity is also positively correlated with these three parameters, b u t at
a less significant level. The species primarily responsible for this lowered
correlation ( Cyprideis torosa and Ilyocypris gibba) are euryhaline-estuarine
species, which are evidently tolerant of a wide range of alkalinities, from the
relatively low CO~- + HCO~ present in most estuaries, to the highly alkaline
soda lakes.
Ca 2+ ranges for African ostracodes are negatively and insignificantly related
t o the other four factors. Many nonalkaline, Ca2*-poor waters of Eastern
Alkalinity
.38
CI-
.03
.49
Conductivity
(K20)
.11
.74
.82
Na +
.10
.66
.82
.96
Ca- +
AIk.
CI-
K20
Fig.7. R a n k o r d e r c o r r e l a t i o n c o e f f i c i e n t m a t r i x . T h i s m a t r i x illustrates the r e l a t i o n s h i p
b e t w e e n t h e five ionic p a r a m e t e r s c o n s i d e r e d in this s t u d y w i t h r e s p e c t t o t h e i r c o o r d i n a t e
e f f e c t s o n o s t r a c o d e assemblage variations. See t e x t for full discussion. K20, C l - , alkalinity
a n d Na ÷ are all positively a n d s i g n i f i c a n t l y c o r r e l a t e d w i t h e a c h o t h e r (p < 0.01), indicating t h a t t h e i r individual effects o n t h e o s t r a c o d e assemblage c a n n o t be f a c t o r e d o u t f r o m
n a t u r a l d i s t r i b u t i o n s alone.
140
Africa (for example, Lake Victoria) are similarly poor in other ionic constituents, while, as previously mentioned, more highly alkaline waters, like Lake
Turkana, are also low in Ca 2÷, due to low saturation values. It would appear
that few of these widespread species can live in very low Ca 2+ waters (below
5 mg/1). However, in a situation analogous to adaptation for high alkalinity
in Lake Turkana, there are numerous endemic, low-Ca 2÷ species in Lake
Victoria (Lindroth, 1953; A. C. Cohen, pets. obs.). Extracting Ca from water
across such a tremendous concentration gradient is a highly energy-intensive
activity for shelly invertebrates (Beadle, 1974). The upper "limits" o f Ca 2÷
values illustrated for these ostracode taxa almost certainly are artifacts of the
typically low Ca 2+ concentrations of African inland waters. Again, wideranging values for I. gibba and C. torosa result from their dual presence in
high Ca 2+ estuarine environments and low Ca 2÷ sodic environments. Thus, the
Ca 2+ concentration does n o t appear to be the principal limiting factor in the
distribution of these t w o species.
Ostracode assemblages in relation to alkalinity
Alkalinity is perhaps the chemical parameter most widely discussed in
reference to African lacustrine faunal and floral distribution (Jenkin, 1936;
Hustedt, 1949; Talling and Talling, 1965; McLachlan and McLachlan, 1969;
H e c k y and Kilham, 1973; Beadle, 1974). Four broad ranges of alkalinity
values appear to correlate with some aspects of ostracode distribution. The
absolute boundaries of these ranges are, of course, somewhat arbitrary,
reflecting a combination of sampling bias and reality. The assemblage ranges
(Fig.8) are named for the most distinctive or diagnostic member taxa.
Range I: less than 0.5(?)--5 meq/l (the Stenocypris Assemblage). Nearly
all species of the cosmopolitan (both Paleotropical- Ethiopian and Neotropical in distribution) genus Stenocypris occur in the relatively fresh waters of
this range. Preliminary evidence suggests that the closely related genera
Chrissia and Parastenocypris m a y also be typical of this range. Particularly
notable as indicators of this range are the c o m m o n species S. platybasis and
Zonocypris costata, the latter of which has already been described from the
Eastern African fossil record (Carbonel and P e y p o u q u e t , 1979). The absence
of ostracode assemblages representative of c a r b o n a t e - b i c a r b o n a t e - p o o r
waters (<0.5 meq/1) merely reflects the fact that such lakes are relatively
rare and u n d o c u m e n t e d in this region of Africa.
Range H: 5--15 meq/l (the Mecynocypria I Assemblage). Mecynocypria is
an extremely useful indicator of this water chemistry range, occurring only
rarely in Range I waters. Additionally, Allocypria and Physocypria (members
o f the same clade) occur in Range II waters, though Physocypria may be
found in Range I waters more c o m m o n l y than Mecynocypria. The genus
IThis Assemblage was referred to as the Paracypria Assemblage in C o h e n (1982). Paracypria is no longer considered a valid genus for these African species, R o m e (1962) having
placed it in either Allocypria or Mecynocypria (K. G. McKenzie, pers. commun., 1981).
141
RANGE#
I
II
NAME
Stenocypris
Assemblage
Mecynocypria
Assemblage
Na +
<75
CI-
<20
Alkalinity
-<5.0
K20
<500
Comparison
with Tailing
and T a i l i n g
classification
Class
1
(<600)
75-150
20-50
5-15
500-1500
Class 2
(600-6000)
III
Gomphocythere
Assemblage
IV
L imnoc.yt here
Assemblage
150900
50-500
>900
>500
15-30
>30(-80?)
15004000
>-4000
Class 3
(>6000)
Fig.8. Ostracode assemblage ranges. The range number is followed by the most distinctive
or commonplace genus of that range. Only for Range II is the genus absolutely restricted
to that range, however. See discussion in text. Na÷ and C1- values are in rag/1 alkalinity
(CO~- + HCO~)in meq/1 and K20 (conductivity)in umho/cm. Talling and Talling's (1965)
classification scheme (values are K20) is included for comparison.
Allocypria t o d a y occurs only in lakes Tanganyika and Albert (and possibly
Europe), b u t was apparently m o r e widespread in Eastern Africa during the
Late Pleistocene (fossils of Allocypria have been recovered f r o m the Nakuru,
Elmentaita, and T ur kana basins). Its e x t r e m e s t e n o t o p y , however, for alkalinity (6--9 meq/1), as well as f or ot her ionic parameters, suggests t hat it m ay
be a very valuable paleochemical indicator. Additional genera, in particular
Eucypris and Pseudocypris, are currently under study as potential Range I
and II indicators.
Cypris, a very i m p o r t a n t Range I and II ostracode genus, has a worldwide
distribution. U n f o r t u n a t e l y , most African species of this genus seem to have
very limited geographical ranges, with the possible exceptions of C. neumanni
and C. latissima ( L i ndr ot h, 1953). T he identification of particular species of
this genus, or of t e n the genus itself, f r o m fossils is usually difficult, thus
reducing th e utility o f an otherwise potentially useful group for paleochemical analysis.
Range III: 15--30 meq/l (the Gomphocythere Assemblage). This range of
alkalinities is occupied largely by e u r y t o p i c species, which also live in less
alkaline waters. No ostracode genus f r o m Africa (including Gomphocythere)
is k n o w n t o o ccur exclusively in waters m o r e alkaline than 7 meq/l. Endemic
species o f Gomphocy there, Cyprinotus, Potamocypris, Hemicypris, Cypridopsis, and Schlerocypris occur in Ethiopian lakes of this t ype, as well as in Lake
T u r k an a (alkalinity: 19 meq/l). Additionally, these same genera, as well as
Oncocypris, Ilyocypris and Cyprideis (and a n u m b e r o f less i m p o r t a n t genera),
are all represented by cosmopolitan species whose alkalinity tolerances
e x t e n d into this range.
142
Range IV: >30 meq/l (the Limnocythere Assemblage). Few species of
ostracodes are k n o w n from highly alkaline waters for reasons already discussed, and none to our knowledge are absolutely diagnostic of them. Two
species, Limnocythere michaelseni and Gomphocythere angulata, are the
only ones currently k n o w n from such African waters, the latter only barely
extending into the range.
The species Cyprideis torosa and Ilyocypris gibba, b o t h of worldwide as
well as African distribution, can tolerate high (Range IV equivalent) salinities,
b u t have not been reported from lakes more alkaline than 25 meq/1. This
suggests that these species are truly estuarine adapted, as o p p o s e d to being
generalists for all types of ionically concentrated waters.
The upper limit of alkalinity tolerance for ostracodes as a group is unknown. An occurrence of Limnocy there michaelseni in Lake Abiata, Ethiopia,
when the lake had an alkalinity of 80 meq/1 is well d o c u m e n t e d (OmerCooper, 1930; Lowndes, 1932). A characteristic feature of highly alkaline
lakes, however, is their enormous variability in water chemistry. Such lakes
tend to be shallow pans, subject to seasonal or more infrequent fluctuations
o f water level and chemistry due to local weather variations (Livingstone and
Melack, 1979).
A study by Wiederholm {1980) of Lake Lenore {Wash., U.S.A.), an alkaline
lake being progressively diluted b y inflowing irrigation projects, showed that
ostracodes colonized the lake when the alkalinity had dropped to 15--20
meq/1. McLachlan and McLachlan {1969) reported no ostracodes in a study
of Lake Chilwa (Malawi) during a drying phase in 1967 when conductivities
rose over 10,000 ~mho. From correlations developed b y Tailing and Talling
{1965) between conductivity and alkalinity in these types of lakes, it is poss-.
ible to estimate the alkalinity during this drying phase at over 100 meq/1.
McKenzie (1970), in a biogeographical review of non-marine ostracodes from
the Southern Hemisphere, noted that a number o f Australian ostracodes
have been recorded in waters with total ionic concentrations in excess of
30 p.p.t., the genus Diacypris being recorded in waters as concentrated as
131.4 p.p.t. (Bayly and Williams, 1966; McKenzie, 1970, after Bayly). It is
clear from the research of Rawson and Moore {1944), Hammer et al. (1975)
and Wiederholm (1980) that ionic concentration tolerances are quite variable
for lacustrine invertebrates, dependant principally on the ionic species present,
with considerably higher tolerances being noted for sodium chloride and
sodium sulphatochloride waters than for sodium carbonate--bicarbonate
waters {the principal t y p e of Eastern Africa).
No long-term studies of ostracode populations in the alkaline lakes of
Africa have y e t been attempted, so we can only speculate a b o u t their occurrences in waters more alkaline than 80 meq/1. In t w o cases, however, highly
alkaline lakes which do n o t suffer regular drying are k n o w n to be devoid of
ostracodes. Both of these lakes, Sonachi Crater Lake (near Lake Naivasha)
and Central Island Crater Lake A (at Lake Turkana) are believed, from altimetric studies for this report, to be in subterranean communication with
their larger, neighboring bodies of water. Local volcanic runoff in b o t h cases
143
serves to maintain alkalinities higher than the parent lake, which nevertheless
keeps the water volumes and levels in these craters much more constant than
would otherwise be possible. Since both of these crater lakes lie in the range
of 50--100 meq/1 alkalinity, and neither appears to have any unusual chemical characteristics for lakes of this alkalinity, the occurrence of ostracodes at
80 meq/1 would appear to be near the natural limit.
The identification of ostracode assemblages as being diagnostic of Range
III or IV paleochemistries obviously relies on the negative evidence of absence
of lower~alkalinity-range ostracode taxa. Since no widespread species can be
considered diagnostic of these ranges, this is an unavoidable limitation on the
m e t h o d o l o g y presented here. Field studies, however, suggest that this may
be a relatively minor problem. In the course of over 300 dredge hauls from
Kenyan lakes for this report, and numerous accounts of other workers (e.g.,
Klie, 1933; Lowndes, 1932; Lindroth, 1953; Rome, 1962), no Range III or
IV assemblages -- t h a t is to say assemblages which were totally devoid of
fresher-water species -- were ever recovered from lakes whose water chemistries did not correspond with those ranges respectively. For example, while
L i m n o c y t h e r e michaelseni, a Range IV assemblage species, has been reported
from Lake Tanganyika (a Range II lake) and Lake Turkana (a Range III lake),
no monospecific assemblages of L. michaelseni have ever been found in either
Tanganyika or Turkana. When found in either of these lakes, it always seems
to occur with fresher-water species as well. Therefore, paleoecological inferences based upon negative evidence (in this case the absence of key taxa),
although always less desirable t h a n those based upon a taxon's presence, may
be justified, as in this case where other supporting evidence is simply unavailable, and where sufficient empirical evidence exists to support its application.
PALEOECOLOGICAL APPLICATIONS
Of the 33 taxa discussed here, 21 have been found as fossils in Miocene-Holocene Eastern African lake sediments. Since the study of fossil fresh-water
ostracodes from Eastern Africa is only in its infancy, it seems likely that
even more of the listed taxa will eventually be recovered as fossils. Fossil
occurrences thus far d o c u m e n t e d are listed separately (available from the
senior author on request).
Ostracode occurrence data from several cores taken in Lakes Nakuru,
Elmentaita and Naivasha of the Kenyan (Gregory) Rift Valley were gathered
by two of the authors {R. D. and J. R.). The original data appeared in an
unpublished manuscript (Dussinger, 1973). One of these sets of core data
(from Lake Nakuru) was chosen to illustrate the application of the technique
discussed in this paper to lacustrine paleochemical analyses. This core has
also been analyzed by Richardson (1979) for its diatom content (again in
relation to water chemistry), which will serve as an independent check on
the validity of our interpretation.
144
The Lake Nakuru core
Lake Nakuru, a small alkaline lake, lies in the Gregory Rift Valley, approximately 120 km northeast of Nalrobi. At the present time, it occupies a
closed basin at approximately 1750 m above mean sea level. Its modern
water chemistry (Table I) is quite variable due to its shallow and greatly
fluctuating depth (Zmax = 2.5 m).
Richardson extracted a 19.5-m core from Lake Nakuru during his 1969--70
field season. The core was sampled every 20 cm for its ostracode content.
The ostracode stratigraphy is illustrated in Fig.9A. A detailed account of the
analysis of this core will appear shortly (Richardson, in prep.).
Many of the genera (and some of the species) described by Dussinger from
the Nakuru core are among those discussed in the first part of this paper, and
are therefore of use in this application. For example, the presence of Mecynocypria (cf. M. reniformis) at several levels of the core suggests t h a t the alkalinity at those levels was between 5 and 15 meq/1. For other levels, the overlapping ranges of two coexisting taxa yield the probable water chemistry at
t h a t level. For example, at 370 cm the co-occurrence of Limnocythere
michaelseni (Na ÷ range 20--3200 mg/1) with Gomphocythere obtusata (10-115 mg/1) suggests a Na ÷ range of 20--115 mg/1, for Lake Nakuru's water at
the time of deposition (and for the part of the lake where the core was taken).
More alkaline conditions (Range IV) must have been present in Lake Nakuru
at the 850-cm level time, when a monospecific Limnocythere assemblage
formed. The absence of ostracodes for 180 cm above and 60 cm below this
point in the core may suggest that alkalinity concentrations during these
time intervals exceeded 80 meq/1.
Fig.9B compares composite interpretations of the Nakuru core paleochemistries based upon both ostracode and diatom typology. The agreement
of the two techniques appears to be quite good, and it must be emphasized
t h a t the two techniques were derived entirely independently of each other.
This should be encouraging news to diatomists studying lacustrine paleochemistries in the absence o f corroborative evidence. Major variances in interpretation between the ostracode and diatom methodologies occur at two
intervals: 1175--950 cm and 360--310 cm. At the m o m e n t , no satisfactory
explanation can be offered to explain these discrepancies, although t h e y are
under further study in an effort to resolve this problem. For 80% of t h a t
portion of the core which contains both ostracodes and diatoms, the trends
of increasing and decreasing ionic concentrations follow each other. Fluctuating alkaline to moderately fresh water occurred between 800--300 cm time
at the core site, with fresh-water conditions following from 300--30 cm, and
finally a return to alkaline conditions near the top of the core.
Up to this point, the fossil ostracode occurrences have been evaluated
entirely on presence--absence data. Whereas it is tempting to consider valve
ratios as being of importance in evaluating probable paleochemistries, this
approach must be considered carefully (and indeed was rejected for this preliminary study) for two compelling reasons.
13 0 0 0
10 340 +- 150
10 340 -+ 1 5 0 to
12 8 5 0 -+ 190
3755
present
Age ( a p p r o x . )
(yr. B.P.)
20--900
10--900
25--100
20--115
20--6500
3300--38 000
Na ÷
(mg/1)
5--22
5--40
ca. 9
5.5--9.5
1--31
0--10?
Ca 2÷
(mg/1)
4--500
4--500
18--37
4--50
3--1500
1020--13 000
C1(mg/1)
2.4--25
0.9--25
6--9
5.5--9.5
2.4--80
120--1440
Alkalinity
(meq/1)
210--3300
90--3300
600--750
210--1100
210--10 700
10 0 0 0 - - 1 6 2 500
K20
(umho/cm)
III
III
II
II
IV
(no ostracodes)
IV
Assemblage
range
P a l e o c h e m i s t r i e s o f s e l e c t e d levels f r o m a 19.5-m core f r o m Lake N a k u r u , K e n y a . S t r a t i g r a p h i c h o r i z o n refers to d e p t h b e l o w t h e m u d /
w a t e r interface. T h e m o d e r n lake w a t e r c h e m i s t r i e s f r o m b o t h high- a n d l o w - w a t e r t i m e s are i n c l u d e d for c o m p a r i s o n , d a t a are f r o m
K i l h a m ( 1 9 7 1 ) , a n d Tailing a n d Talling ( 1 9 6 5 ) , a f t e r B e a d l e ( 1 9 3 2 ) for m o d e r n w a t e r chemistries. See Fig.9 for e x p l a n a t i o n o f u n i t s
used.
470
580
0 (surface
mud)
210
370
--
(cm)
Stratigraphic
horizon
Lake N a k u r u m o d e r n a n d p a l e o c h e m i s t r i e s c o m p a r e d
TABLE I
Ol
/>
o
o
T
i i i r r ~ i
o
I
t
I
oo
o
o
[
o
)
o
i
I
o
I
r
8
t
r
o
I
0
o8
r
o
i
~
mm
m m.
v
o
mm
I
l
o
)#
§
#
°>"
~
m
Sclerocypris
A
o
mum
Mecynocypria
m
<.~
I
m
~
.....
-
.
.
.
cf. M. reniformi,s
C)'pretta
.
.
.
.
•
--
~
'
c f . C. m u r a t i
.
Zonocypris
..........
]
.
.C y p r i a
Oncocypris
........
r~ I
mm
C~idops!))
ms
ml i m l
C y_pridopsis
Cypridopsis
mr
--
.
mm
mm
mmq
mm
•
sp.
sp. D
sp. B
sp. A
-
T
i
qI "
m .m
•
sp. C
-
-
Gomphocyther.e
........
m
m.-m~mmm~r=
. . .mm
.
sp.
calcarata
Cypridopsi_s
...........
-
.........
r
ram.
m
F
-
venusta
l
mm
8?°
i
iNN
~
............
"11,
•-
m -
m
........
mmm
mm
ms
Limnocythere
. . . . . . .
t
•~
0btusata
michaelseni
Limnocythere
sp. C
Limnocythere
sp. B
sp. A
Limnocythefe
O
"10o3
Oo
~> m
~
r-~
:: ::.:.:.::::::::..~ m ~.
~::. ~)~O~i~)::::i:i:i~i::!:i:):i:
))::
))):| "0 0 0
0 0
rn.-Jr'
0
jr.m~.
.......i:i:i~.
: : .....
I, m~
iii
=¢
J
.............................
-~ _¢ =
: ~ ::i~*::
---5 o >
~GO
m
m~=
$>
m
i
I
I
r
I
0
0
i
l
0
I
I
0
(
r
0
J
I
0
d
I
0
~
I
0
I
P
0
1
]o3
-r
0
I
"n
~" 111
I
147
First, the initial distribution ratios of ostracodes during life are controlled
b y a variety of environmental and biotic factors aside from water chemistry.
Temperature variability is probably insignificant in lowland {under 2000 m)
equatorial Africa, b u t local, microenvironmental variations (for example,
algal v s macrophytic substrates) will often lead to several ostracode associations within a given lake. This has been confirmed in studies of the ostracode
fauna of Lake Turkana, where two distinctive ostracode associations may be
recognized (Cohen, 1982) associated with water depth variations, for reasons
which are not yet entirely clear. Thus, an abundance of a given species at the
expense of another m a y not reflect a more advantageous water chemistry for
the first, b u t rather could reflect an advantage along a n y environmental
gradient of significance to ostracodes.
The second problem, potentially even more intractable than the first, concerns the selective post-mortem transport and preservation of ostracode valves.
Ostracode valves have been observed in the course of this study to float quite
readily after death and be transported, often in large numbers, o u t of their
life habitats. This again makes the counting of valve ratios of dubious value,
until such time as these taphonomic problems can be sorted out. Surprisingly
however, this admixture of valves actually works to the advantage of a paleochemical analysis by assuring that a greater percentage of the lake's ostracode
fauna (by species) will be represented in a single sample. This in turn allows
for a greater probability that a narrow range of paleochemical interpretations
m a y be inferred. Returning to the Lake Turkana example, a typical sublittoral, silty b o t t o m sample contains 3--5 species of live ostracodes, whereas
the sediment from the same locality will often contain up to 8--9 species,
represented b y valves only, an admixture from deeper and shallower water.
CONCLUSIONS
Typologies have proved useful in characterizing lake water chemistries b y
their inhabitant floras (Richardson, 1968) and faunas (LaBarbera and
Kilham, 1974). When the group of organisms subjected to a typological study
leaves behind a significant, taxonomically identifiable fossil record in lacustrine deposits, the t y p o l o g y can be inverted to yield paleochemical data about
the ancient lakes in question. Eastern and Southern African lakes provide an
ideal setting for such studies for two reasons: (1) a very wide range of water
chemistries is represented there; and (2) many of these lakes have geologic
records extending well back into the Pleistocene or earlier.
Fig.9. O s t r a c o d e s t r a t i g r a p h y a n d p a l e o c h e m i c a l i n t e r p r e t a t i o n s f r o m L a k e N a k u r u , for
t h e u p p e r 12 m o f a 19.5-m c ore. A. O s t r a c o d e s t r a t i g r a p h y a f t e r Dussinger ( 1 9 7 3 , p. 24 b)
s h o w i n g valve a b u n d a n c e s o f all t a x a p r e s e n t p e r cm 3 w e t s e d i m e n t . B. P a l e o c h e m i c a l
i n t e r p r e t a t i o n s of t h e L a k e N a k u r u core b a s e d u p o n o s t r a c o d e f a u n a a n d d i a t o m flora.
D i a t o m d a t a are f r o m D u s s i n g e r ( 1 9 7 3 , p. 24c) a n d R i c h a r d s o n ( 1 9 7 9 ) . T h e i n d e p e n d e n t
interpretations of paleochemistry based on ostracodes and diatoms show good agreement
for all intervals e x c e p t 1 1 7 5 - - 9 5 0 c m a n d 3 6 0 - - 3 1 0 cm.
148
Ostracodes are well suited for this t y p e of study (Delorme, 1969). Thirty
three taxa (genera and species) were investigated and their water chemistry
ranges plotted for five chemical parameters (Na ÷, Ca 2÷, Cl-, CO~- + HCO~
alkalinity, and total conductivity K20). Ostracodes, being highly calcified
crustaceans, are excluded from highly alkaline waters with low (less than
1 mg l) Ca ÷÷ solubilities.
The ordering of the ostracode tolerance ranges for four of the parameters
(Na ÷, COl- + HCO] alkalinity, CI- and K20) are highly coordinated, suggesting that the effects of these parameters cannot be individually factored o u t
in distribution studies of ostracodes in natural waters from this region.
Ostracode assemblages typical of particular levels of ionic concentration
can be characterized by four broad assemblage ranges. Range I (the Stenocypris assemblage) and IV (the Limnocythere assemblage) roughly correspond to Talling and Talling's (1965) Class I (K:0 < 600 u m h o ) and lower
Class III (K20 > 6000 p m h o ) respectively, whereas Ranges II (the Mecynocypria assemblage) and III (the Gomphocythere assemblage) would in toto
be approximately equivalent to their Class II waters (K20 = 6 0 0 - - 6 0 0 0 pmho).
An example of the application of ostracode t y p o l o g y to fossils in a 19.5-m
core from Lake Nakuru illustrates the technique. Nakuru, which t o d a y is a
highly concentrated, alkaline brine, was much more dilute in the recent geological past. Fluctuations between Range II and IV ostracode faunas can be
d o c u m e n t e d in this core through much of the latest Pleistocene and Early
Holocene. These results are corroborated b y studies of diatom t y p o l o g y
(Richardson, in prep.) of the same core. The broad similarity of results
b e t w e e n paleochemistries determined using these two entirely independent
techniques is very encouraging.
As a final note of caution, we have considered (and rejected for this first
study) one easily suggested refinement of our technique. Tabulating ratios of
species present in a fossil assemblage rather than merely their presence or
absence (as we have done here) is a normal tool of the trade among paleoecologists. Experience with living populations of ostracodes in Lake Turkana
has led us to believe that such ratios may be misleading when taphonomic
and ecologic complications (derived from significant post-mortem sorting
and microhabitat variations) arise.
This technique is still in its infancy, and many refinements are certain to
be made to it, b o t h in an effort to "fine t u n e " it and make it more comprehensive. Nonetheless, it has already proven its w o r t h as a valuable tool in the
interpretation of the complex yet fascinating history of the Eastern African
Rift Valley lake's.
ACKNOWLEDGEMENTS
We would like to thank a number of people who have made this work
possible. Messrs. P. Kongere, P. Ojuok, and B. Ogilio of the Kenya Department of Fisheries and Wildlife were very supportive of our field research, as
149
was Mr. E. K. Ruchiami of the Office of the President, and the staff of the
Lake Turkana Fisheries Station at Kaliokwell--Ferguson's Gulf, Kenya. To
Cohen's assistants Karen Higgins and Nancy Dickinson we owe an enormous
debt for their time spent in the field collecting data. Drs. Leo Laporte and A.
Kay Behrensmeyer have generously supported this work on NSF Grant
#EAR-77-2349 (University of California at Santa Cruz), and with many
hours of discussion. Additional funding was provided by grants from the
University of California at Davis and Amoco Inc. The manuscript was read
by D. Livingstone, R. Bate, F. Swain, D. Delorme, K. G. McKenzie, R. Cowen
and L. Laporte. For all of their comments and criticisms we are grateful.
REFERENCES
Bayly, I. A. E. and Williams, W. D., 1966. Chemical and biological studies on some saline
lakes of Southeast Australia. Aust. J. Mar. Freshwater Res., 17 : 177--228.
Beadle, L. C., 1932. Scientific results of the Cambridge Expedition to the East African
Lakes, 1 9 3 0 - 3 1 . 4 . The waters of some E. African Lakes in relation to their fauna and
flora. J. Linn. Soc. (Zool.), 38: 157--211.
Beadle, L. C., 1943. An ecological survey of some inland waters of Algeria. J. Linn. Soc.
(Zool.), 41: 2 1 8 - 2 4 2 .
Beadle, L. C., 1974. The Inland Waters of Tropical Africa. Longman, London, New York,
NY, 365 pp.
Carbonel, P. and Peypouquet, J., 1979. Les Ostracodes des S~ries du Bassin de l'Omo.
Bull. Inst. G4ol. Aquitaine, Bordeaux, No. 25: 167--199.
Cerling, T. E., 1977. Paleochemistry of Plio-Pleistocene Lake Turkana and Diagenesis of
Its Sediments. Dissertation, University of California at Davis, CA, 180 pp.
Cohen, A. C., 1982. Ecological and Paleoecological Aspects of the Rift Valley Lakes of
East Africa. Dissertation, University of California at Berkeley, CA, 314 pp.
Damas, H., 1937. Recherches hydrobiologiques dans les lacs Kivu, Edouard et Ndagala
(Congo Belge). Verh. Int. Vet. Limnol, 8: 51--68.
Damas, H., 1954. Etude limnologique de quelques lacs ruandais. III. Le plancton. Mem.
Acad. Sci. Colon. Belge, Classe Sci, N.S., 3: 1--66.
Damas, H., 1964. Le Plancton de quelques lacs d'Afrique Centrale. Verh. Int. Ver. Limnol.,
15: 128--138.
De Kimpe, P., 1960. Observations hydrobiologiques au Laupula--Moero. Hydrobiologie et
p~che en eaux douces. In: Colloque sur les probl~mes des Grand Lacs, Lusaka, 1960.
CCTA/CSA LACS, 3 9 : 1 7 pp.
Delorme, L. D., 1969. Ostracodes as Quaternary paleoecological indicators. Can. J. Earth
Sci., 6: 1471--1476.
Delorme, L. D., 1971a. Paleoecological determinations using Pleistocene freshwater ostracodes. In: H. C. Oertli (Editor), Pal~o~cologie d'Ostracodes. Bull. Centre Rech. SNPA,
Pau, pp. 341--347.
Delorme, L. D., 1971b. Paleoecology of Holocene sediments from Manitoba using freshwater ostracodes. Geol. Assoc. Can. Syrup., Spec. Pap., No. 9, pp. 301--304.
Delorme, L. D., Zoltai, S. C. and Kalas, L. L., 1977. Freshwater shelled invertebrate indicators of paleoclimate in Northwestern Canada during the late glacial. Can. J. Earth Sci.,
14: 2029- 2046.
Dubois, J. T., 1955. Etude hydrobiologique d ' u n lac africain: Le Lungwe. Etude physique
et ch~mique. Bull. Centre Belge Etude Doc. Eaux, 53 : 79--82.
Dussinger, R., 1973. The Environmental Reconstruction of African Rift Lakes Using
Ostracodes as Paleoecological Indicators. Thesis, Franklin and Marshall College,
Lancaster, PA, 37 pp. (unpublished).
150
Eugster, H. P., 1980. Geochemistry of evaporitic lacustrine deposits. Annu. Rev. Earth
Planet. Sci., 8: 35--83.
Eugster, H. P. and Jones, B. F., 1979. Behavior of major solutes during closed basin brine
evolution. Am. J. Sci., 279: 609---631.
Fish, G. R., 1952. Appendix A. Annual Report for 1951. E. Aft. Fish Res. Org. Jinja,
Uganda. E. Aft. High Comm., 49 pp.
Fish, G. R., 1953. Appendix A. Chemical Analyses. Annual Report for 1952. E. Afr. Fish.
Res. Org., Jinja, Uganda. E. Aft. High Comm., 42 pp.
Fish, G. R., 1954. Appendix A1--A6. Annual Report for 1953. E. Afr. Fish. Res. Org.,
Jinja, Uganda. E. Afr. High Comm., 44 pp.
Hammer, U. T., Haynes, R. C., Heselfine, J. M. and Swanson, S. M., 1975. The saline lakes
of Saskatchewan. Verh. Int. Ver. Limnol., 19 : 589--598.
Hardie, L. A. and Eugster, H. P., 1970. The evolution of closed basin brines. Mineral. Soc.
Am. Spec. Publ., 3 : 273--290.
Harding, D., 1963. Studies of Lake Nyasa and associated rivers. In: Jackson et al. (Editor),
Report on the Survey of Northern Lake Nyasa. Joint Fish. Res. Org. 1954--55. Governm e n t Printing Office, Zomba, Malawi, 171 pp.
Hecky, R. E. and Kilham, P., 1973. Diatoms in alkaline saline lakes; Ecology and geochemical implications. Limnol. Oceanogr., 18: 53--71.
Hustedt, F., 1949. Sfisswasser-Diatomeen. Exploration du Parc National Albert. Mission
H. Damas (1935--36), No. 8. Inst. Parcs Natl. du Congo Belge, Brussels, 199 pp.
Jenkin, P. M., 1936. Report on the Percy Slade Expedition to some Rift Valley Lakes in
Kenya in 1929. VII. Summary of the ecological results with special reference to the
alkaline lakes. Annu. Mag. Nat. Hist., Ser. 101, 18: 133--181.
Kilham, P., 1971. Biogeochemistry of African Lakes and Rivers. Dissertation, Duke
University, Durham, NC, 191 pp.
Klie, W., 1933. Report on the Percy Slade Expedition to some Rift Valley Lakes in
Kenya in 1929. Die Ostracoden der Rift-Tal-Seen in Kenia. Int. Rev. Gesamten Hydrobiol. Hydrogr., 29: 1--14.
LaBarbera, M. C. and Kilham, P., 1974. The chemical ecology of copepod distribution in
the lakes of East and Central Africa. Limnol. Oceanogr., 9: 459--465.
Lindroth, S., 1953. Taxonomic and zoogeographic studies of the ostracode fauna in the
inland waters of East Africa. Zool. Bidr. Upps. Univ., 30: 43--156.
Livingstone, D., 1963. Chemical composition of rivers and lakes. U.S. Geol. Surv. Prof.
Pap., 4 0 0 ~ : 64 pp.
Livingstone, D. and Melack, J., 1979. The Lakes of Subsaharan Africa. SIL-UNEP Workshop on African Limnology, 46 pp.
Loffredo, S. and Maldura, C. M., 1941. Risultati generali delle richerche di chimica limnologica sulle acque dei laghi dell'Africa orientale italiana esplorati dalla Missione
ittiologica. In: A. Piccioli (Editor), Esplorazione dei laghi della Fossa Galla. Coll. Sci.
Documen. Africa Ital., III, 1: 181--200.
Lowndes, A. G., 1932. Report on the ostracoda. Mr. Omer-Cooper's investigation of the
Abyssinian fresh waters. Proc. Zool. Soc. Lond., Part 3: 667--708.
McKenzie, K. G., 1970. Paleozoogeography of freshwater ostracoda. In: H. J. Oertli
(Editor), Pal6o$cologie d'Ostracodes. Bull. Centre Rech. SNPA, Pau, pp. 207--237.
McLachlan, A. J. and McLachlan, S. M., 1969. The b o t t o m fauna and sediments in a drying phase of a saline African lake (Lake Chilwa, Malawi). Hydrobiologia, 34: 401--413.
Omer-Cooper, J., 1930. Dr. Hugh Scott's Expedition to Abyssinia. A preliminary investigation of the freshwater fauna of Abyssinia. Proc. Zool. Soc. Lond., Parts 1--2: 195-207.
Peypouquet, J. and Carbonel, P., 1980. Les ostracodes indicateurs des pal~oenvironnements lacustres des rifts de l'Afrique orientale au Plio-Pl~istoc~ne. C.R. Acad. Sci. Paris,
290: 563--566.
Phillips, T., 1930. Crocodiles in Africa. London Times, 11 April, 1930, p. 12.
151
Rawson, D. S. and Moore, J. E., 1944. The saline lakes of Saskatchewan. Can. J. Res., 22:
141--201.
Ricardo, C. K., 1939. Report on the Fish and Fisheries of Lake Rukwa in Tanganyika
Territory and the Bangweulu region in Northern Rhodesia. Crown Agents for the
Colonies, London, 78 pp.
Ricardo-Bertram, C. K., Borley, H. J. and Trewavas, E., 1942. Report on the Fish and
Fisheries of Lake Nyasa. Crown Agents for the Colonies, London, 181 pp.
Richardson, J., 1968. Diatoms and lake typology in East and Central Africa. Int. Rev.
Gesamten Hydrobiol., 59: 299--338.
Richardson, J., 1979. Paleolimnology of Lake Nakuru, Kenya (abstract). SIL--U.N.
Environ. Progr. Workshop on African Limnology, Abstr. of Poster Papers, p. 41.
Rome, D. R., 1962s Exploration hydrobiologique du Lac Tanganyika (1946--1947):
Ostracodes. Inst. R. Sci. Nat. Belg., 3(8): 305 pp.
Tailing, J. F., 1957. The longitudinal succession of water characteristics in the White Nile
Hydrobiologia, 11 : 73--89.
Talling, J. F., 1963. Origin of stratification in an African Rift Lake. Limnol. Oceanogr.,
8: 68--78.
Talling, J. F. and Talling, I. B., 1965. The chemical composition of African lake waters.
Int. Rev. Gesamten Hydrobiol., 50: 421--463.
Van der Ben, D., 1959. La v~g6tation des rives des lacs Kivu, Edouard et Albert. Exploration Hydrobiologique des lacs Kivu, Edouard et Albert (1952--54), 3. Inst. R. Sci. Nat.
Belg., 4(1): 191 pp.
Van Meel, L., 1953. Contribution ~ l'6tude du lac Upemba. A. Le milieu physicochimique.
Exploration du Parc National de l'Upemba, 9. Inst. Parcs Natl. Congo Belge, 190 pp.
Van Meel, L., 1954. Le phytoplancton. Exploration hydrobiologique de lac Tanganyika
(1946--47) 4. Inst. R. Sci. Nat. Belg., 681 pp.
Verbeke, J., 1957. Recherches ~cologiques sur la faune des grands lacs de l'Est do Congo
Belge. Exploration hydrobiologique des lacs Kivu, Edouard et Albert (1952--54) 3.
Inst. R. Sci. Nat. Belg., 177 pp.
Visser, S. A., 1974. Composition of waters of hakes and rivers in East and West Africa. Afr.
J. Trop. Hydrobiol. Fish., 3: 43--60.
Wayland, E. J., 1925. Petroleum in Uganda. Geol. Surv. Uganda Mem., No. 1. Crown
Agents for the Colonies, London.
Wiederholm, T., 1980. Effects of dilution on the benthos of an alkaline lake. Hydrobiologia, 68: 199--208.
Worthington, E. G., 1929. The life of Lake Albert and Lake Kioga. Geogr. J., 74: 109-132.
Worthington, E. B., 1930. Observations on the temperature, hydrogen ion concentrations
and other physical conditions of the Victoria and Albert Nyanzas. Int. Rev. Gesamten
Hydrobiol., 24: 328--357.
Worthington, E. B., 1932. A Report on the Fisheries of Uganda Investigated by the
Cambridge Expedition to the East African Lakes 1932--1933. Crown Agents for the
Colonies, London, 88 pp.