Palynological analysis of hyrax middens from Southern Africa

Palaeogeography, Palaeoclimatology, Palaeoecology, 76 (1990): 367 379 Elsevier Science Publishers B.V., Amsterdam

367

Printed in The Netherlands

Palynological analysis of hyrax middens from Southern Africa L. SCOTT I and C. B. BOUSMAN 2 t Department of Botany, University of the Orange Free State, Bloemfontein 9300 (South Africa) 2Department of Anthropology, Southern Methodist University, Dallas, TX 75275 (U.S.A,) (Received November 11, 1989)

Abstract Scott, L. and Bousman, C. B., 1990. Palynological analysis of hyrax middens from southern Africa. Palaeogeogr., Palaeoclimatol., Palaeoecol., 76:367 379. Hyrax middens in relatively dry parts of Africa preserve plant material which includes pollen. The bottoms of two stratified hyrax dung middens from Blydefontein Basin in the Karoo shrubland of South Africa, were dated by radiocarbon to ca. 300 yr B.P. and ca. 1130 yr B.P. Comparison of midden pollen spectra with pollen from alluvial sediments near one of the hyrax middens, and in modern surface samples, indicates that middens contain pollen which reflects local vegetation, and that diet probably does not have a major influence on pollen spectra. Changes in pollen composition probably reflect climatic conditions during the early history of the older midden profile. Grass dominated spectra suggest moist conditions existed between 1200 and 300 yr B.P., and an interval of slightly more shrubby karoid vegetation possibly points to a minor dry spell at 1000 yr B,P. A sharp increase of karoid shrub vegetation, and a decrease of grass since 300 yr B.P. is reflected in both middens. The change was apparently triggered by relatively dry conditions, and later, during the last 150 years, it was enhanced by overgrazing from European domestic stock.

Introduction

Palaeobotanical data are required from all major African biomes in order to study longterm environmental change. However, large parts of the continent in both Northern and Southern Hemispheres are arid or semi-arid, and fossil pollen traps are rare in these regions. In contrast to subhumid regions of Africa, where pollen is preserved in lakes and swamps, information about the Quaternary is scarce in arid portions of the continent. Hyrax (Procavia spp.) middens (solidified dung accumulations) are promising sources of palaeoenvironmental data from arid regions. The first fossil pollen data from a hyrax midden were published by Pons and Quezel (1958). They showed that more Mediterranean elements 0031-0182/90/$03.50

occurred at 4700yr B.P. than today in the Hoggar Mountains of the Sahara Desert. The significance of this study was unappreciated until pollen analysis of numerous packrat middens in arid North America demonstrated the potential that animal accumulations have for palaeobotany and palynology (Betancourt et al., 1989; King and Van Devender, 1977; Thompson, 1985). This North American research stimulated a search for similar deposits in Africa (Scott, 1989a). In this search, middens of Procavia spp. and dassie-rats (Petromus typicus) were recognized as potential sources of palaeoenvironmental data. However, the former should be more useful as Procavia species are distributed throughout most of Africa, excluding only the northwestern region (Morris, 1965). As dassie-

(t} 1990 Elsevier Science Publishers B.V.

368

rats are indigenous to the Namib Desert (Smithers, 1983), their middens promise to elucidate the Quaternary history of only this smaller region. Hyraxes consume leaves, grasses, berries, fruit, bark and occasionally flowers from a wide variety of species (Foerie and Perrin, 1989; Hoeck, 1975; Lensing, 1978; Sale, 1965). According to Foerie and Perrin (1989) the diet of hyraxes in Mountain Zebra National Park, about 150 km southeast and botanically similar to our study area, consists of at least 36 species including grasses (22.2% dry mass), woody species (19% Acacia karoo, 10.5% Olea europea africana, 5.3% Grewia occidentalis, 4.3% Cussonia paniculata), Compositae shrubs (8% Felicia filifolia, 3.1% Pentzia spp.), vines (2.4°//0 Clematis brachiata), and various other taxa. Hyraxes live in colonies near rock outcrops or overhangs, and rarely venture far from these outcrops because they are dependent on shelter for protection from predators (Fairall et al., 1986). Hyraxes habitually defecate and urinate in.protected areas, often under rock overhangs, known as latrines. The fecal pellets, which are spherical and have a diameter of about l c m , accumulate in piles. In dry conditions, their concentrated urine, an adaptation to arid climates, solidifies to form hard, amber-like hyracium (Green, 1955; Smithers, 1983). If hyraxes urinate on dung piles and this remains undisturbed until dry, then consolidated middens form. Hair, seeds, dust, and other objects, including pollen, stick to the drying urine and become trapped in the middens. The hyracium preserves these materials extremely well. Hyrax middens do not represent nest accumulations, because they are not deliberate collectors of plant material like dassie-rats in the Namib Desert and packrats in America. Well preserved macrobotanical remains are consequentially rare in hyrax middens, but microscopic pollen grains are very numerous. Pollen in the middens is assumed to represent local pollen rain, and possibly the diet of hyraxes. Orderly vertical stratigraphic ac-

L. SCOTT AND C. B. BOUSMAN

cumulations occur in level places, but parts of hyrax middens seem to flow down rock faces. By comparison, packrat nest extensions may occur in any direction and not necessarily vertically (King, 1976). Thus, unlike packrat nests which are normally considered to represent single points in time, hyrax middens may produce chronological sequences where vertical accumulation takes place over long periods. In Southern Africa the first pollen sequence from a hyrax midden came from a rock overhang in the subhumid region near Clarens (Scott, 1989a and unpublished data). This midden has been radiocarbon dated to AD 1974 (Pta-4493) and AD 1963 (Pta-4232) and is estimated to be only ca. 30 years old (Scott, 1989a,b). In view of the active role moisture plays in the decomposition of hyrax manure, areas with dry climates are expected to preserve ancient middens better than more humid zones. Therefore attention was focused on the semi-arid Karoo biome in Southern Africa. Pollen spectra from two dated hyrax middens discovered in Blydefontein Basin near Noupoort, South Africa (Figs.1 and 2) are described, and the results are interpreted in terms of local habitat differences, and environmental changes during the Late Holocene. In order to assess the relationship between mid-

0 _~,

15°

20 °

25 °

30 °

I

I

I

I

35 ° I

\

20, ',~

,,,o~

~lOornt~

e

":L ~ CLARENS 3o.-

-30.

~ L ~

~f-/

CAPE TOW[' I 15 °

I 20'

25 °

I :30"

AREAABOVE

Ft ~ 2000 m [ ] KARO0 REGION I

35 °

Fig.1. Map of S o u t h e r n Africa s h o w i n g physical features and the location of the K a r o o region. Dry conditions occur west of the 400 mm isohyet.

PALYNOI,OGY OF HYRAX MIDDENS OF SOUTHERN AFRICA

369

,~

li

= HYRAX

MIDDENS

SEDIMENT FARM

I AND

SITES

,/600_

.........

2

:(~:~:_!:oo

~ _ ~

:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:...-.......:..................................... ,..... :...:,:.: :: ===================================== ::: ::::::::: :::?:'.~:

./

:.........:...:.:.

~: :~:...~...~.:.:.: ~-.

:.:.....-...:

.'.-.'.'.'.'.'.'.9.

': - "-:.

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

H O U S E

~:::::!::::::~ :~::!!~)!/.:::-'::i::~:~

N

,

,~': ":" -: ':i:G

":':.~ • ":.'.::.:.':.:::.::::.'.~::.~.:

:'..:::

:• :;;~: .'.'.."

.:,

.::i: :::i:i:::i.!i:~::.i:!.~:.~:~i:i:i:i:i:!:i:!:i:!:i:~,

:::.': ~:.~:-:::~:::~:i~;L~:: " ".'.'.-'.'. ""..'.-'.'.." ....

........

.'-."

:;:-i-:i::::.o':L::::::::::,~::::

::: :::~;;~: '-~i~> " -" "'.'\

,;::: :::i:~i:'~ :i: : :~:!:!:~!: v'C,~.' " ' ".t~." "'

' " ....

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: i!:i~

:i::::::::::::'.:~.~~.:: ::::::::i:: :~-~:

::~L ".

~-~':0~o~ ,. ,...-..:.:.. '..:::...'.....,.....................:.:...:-......:. -..:.:. .....:.CL~.:......:...:.:...:{...:.....:...:.:...:.:.:.:.:.:.:.:.:.:.:.:.:.:.:..`

~ !.:.:.:.:.:.:.:...:...:.:.:.:.'.:. \ :' ..:. ':.:..:.'.'.".:.:.'.:.'.'

".'.'~'," x

\

--. "-

"I

,

I

/

\

I

\

/.':'".'.~.',~0..:": ,',"..

~ f\

~',

.':-.:::...:':.

",~w~',',~,',

,',,,,',

".',-:'~::'

".',+.-,,,',~'."

.:':-:...~'

,~,

/

",,,,,, ..,

•

' ....

/

..~

•

~:.?:-:.:.:..~:.:.:.....:..z:.'..:..:G

.. . . . . . .

,

.:..: .....

'.'. . . . . -

~. v.:~:...'~:::-.

' , " % ~ l ~ f ~ l ' l ~ ' l , , ~ '" ~.

~

:' .:.-.:.::."~

%:."" "-':'5:-

\ ~

'•

.:.,'.'. • "..:.'," ".'.'.. '. "... ~. ".". '. " '. ,~':...:'~lwPPr~l~/,a/.,.~ SHELTER

.'

Jx

.

-• .•' . .

. .

'",

," r~ , I,

'.~" -" ', 2""

~ ~-

~ ~

'

:

'

.

"

. ...

' • . .. .,

..

, ' . . . .. . .. . .. .

.

:. , " .." ..

' . ' . , '.. ' ,.' , . . , '.. ' . ' . ".

, ' ~ t - : ". . . . . >"

.'

•

~To~

/ C/ :

3, "~''".

~.~.

~ ,:".'-':::::::>

"". " ' " . " " ' ". " " ".

*~' .

":-" .: ~.~:~:~:~:~:~:~:~F:~:~:~:~`.`~:~:~:~:~

.

. " ." ' " ' ".

.

.V

~-

C~

/

/

L

~

~

~.

-~

~

~'~

\

\,,-/

~

,'-~ ~

.o. ~

_

/

/. .

/

./

/

--

-,

~'~

~

//

~ ~

---

J ,:.:-'/ N . . . ..", ::' :':.:'.',

-

, ( / :-,-~

.

.

.

~ - - -

~__~

/

/ ~

.

/ ~

"~

./I

::::.::?

.' " " . "' ."" " ' ." "" . "" ". " ~.

~ ".::'-':':

__~_

~ :..'-', \~

:::..::.:~:

-

, i

..

/ ~

USP) / BLYDEFONTE ..,.:....~

•

~

::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::', ,::':

•-:.:...~

:, ",*~ ....

-~

1700~:.~: :: :::..::::'...':V: v ~ .'.'.'.'.' '.'~:~ :~:........~,, "%"::':\ ~ ~ ~

,... . . . . .... ,.....~..,... ,,.., ...-..............,.., ,.,, . ' . , ...,..,

%:7.. ::.:-:.: .... : : : : : : : : : : : : : : : : : : : : : : : : : : : : ~ : : : : ' : : ~ : : : : . : - , / .:.:.:.:.:-:.'....'.:.:...:.: ..... .-.'.:..:::. - .:. ",, k~'," ~ ,,%;. ~-::.:..:.'.'...'.:.:,.'....-,:... ,,............-................~ ~ ~""-':':'.'.'".','.'.':'.':'-':':'":':':':'.':':':':".'.':':':'.%~ "7]

k.:

"~

k~

~-

,. ..... ,........ ,....... , ...............-.,...,., ,.,,...... .... ,,.. _~ '.'...'.'.' _ ~ 4..:..~.'. 1.':::'.':::.'::::.':-1":::'.'..:'.':':':" .:.::':':'t':'.:.'.'.'.~.~ ~ ,J ,, :.:G: ,.'~ ~.. ,:..X.: .~eO~t.~,::.: .':'.'.':-.'. :':'?.':'5:':-:.:':.:':':'.-:':': :'.-::.. "..:.:.':'.-., ~ X /-\ ..... :" ~" '):'- • "~" .P:~:. :~': • :':':':':: -&;':':':':" ":" ::::'k',':: :'.".:'.':': :':':.:':. o r ~ ,' ,, h. - -'. ~ I :'.> ..'.'..' '." -.'.' • ::." .':'.. ".':" • ":-:~,~ .:':'..'."." ~. ".- - -', • .' ...... :BSM ~' ~ ~...: ~ • . -'~ '"'" G:':" '~" ":"" ""- """"'"""" ~,'""'"""":,.'"'"'" :'.""'.'-"h'~ IxCH 2~ ~'." / :"." ~ <... ......... . .............. ~.......~ .... ,,.~.~..;~'~ ~ ., _~ ~ ....

..

~ : ,

"~'.~ .......

~'

~

~:': :"OPPERMANSKOP:~'>:'~.':e'~'~"::;':':':':"~..:. "...-...:....:.~..... .... ,r-.'..:.:-:-:.'-:.'.:.:.'.P'-'.'.'...

~:..]

:, .'P, : .'~ : ": I~ : • ~•. ....

""

:~:-:, .:.:-.:.:.'....:.:.:,:.:.~0:.:...-...:.:...-...-~:....:.: .... :..:.:.....:. ::: .:...:. =========================== .:.:.:.:.. :)

~:~o~...:.:.:..t..L:~.>c.....:...:.:.:.7..:.............:....:...:.:....:~-\,,

.

~. '.'.:.'.-'"-J

~ ~' ' \

': .;~,...:2:.;:-..~.,J. . .'.:.:.'L.:..:.'. • - : .::.~.i~...,..> ,.. ,..~..................-.%,. -. -.....• -:-............ ~ .~.'... ,.... : ~--

:..:::~:/:~:.J99~`.~:~.~::~:~.:.....~.9~...~.~.~.:~`~::.~:..~::.:~.~:~:~.~.:~:~:.~:~ .:::::~.:::G?.~::.~:..-~: :-:-:.::.~:4G.%r:..~:.:-" j / :r:~.:.~:.:.:.:.:.:-:.:-:+'-:-:.:.:~, ~..~, , . ...~:.:..:o...~...":.'.::'.':'.'.--:4 ....+. .... :-.->V .-..~ -:::-:.: .... :.:.:.;:~:.~:-:......~",,

\\

~' / ,

: ::: : :i-:::~:;~

L

~I.:'.:'F:::':'").":':':::::::'.',Z,.?:::::::::':.::~.'~:.:~.....>.~ ,~ :::::::::::::::::9:::::::::::: :::'.:~: :::::::::::::::::::::::::.::::::-::7::

~\

:; \\'." - .'.'.'..." ." .'.'..... /::5",~m'~' ~::::.'.:.'.:.:.:.:.:-:::.;. (" ." .'.'.'.'." ..... /:.'.','.'.'.'.'PI'.:

~

~ -,.. , \

. ='27

.-.-,'..:':'.

X

~

:':;"':':':'.'" :'.':~:':':' :':':W:'.' .~,-"':':: ': :':' .':':'- \ '~-: v~-:.:-:.: .: ..'. :[ .:'i '/"' .'.." "" ..... ..... '"" """ "":"2 "' ""'"' """'" """ ':::" " "'"' '"' "'"""'") '""""" ~ : ' ~ %" ""'""~ "" ' .:,..:...~.......:.,:/...p::-.:..: .... :.:......~.~,.~ ,, , ~ /'.'.'.'.'<~9~":" ..~:; ".'.'. ,90U::. ".'." -.'. • .'.: ":.:'.:' ~ ,.:-:.:.:.p'...,v~Z...'. ~.-:.~':~.:.:.~:.:.'.:.:.:.:.:.:.:. , \ ~:~..'..'-~:..:'7."-'.:.:..'.:.'. -.." ::.'.:.'.:.'~.'.'.:.'.'.I _ \

[:::~:: Y.: :~:

~ ' ~ ~ ' - i'

~:::~::i:::::!::i:i::::: :::::: ::::::.::::::::: ~ :{::,"" ......... ::::i: :.!::::: .......::::::::::::::::::::::::::::::::::::::::.

,:.,~.: :":~ ::.:'::~: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

':'i":

I\

/I

i::::~:

([.::;

~: :::::::::::.:-::::::::::::

/

""--

~"

/

.

"-:: : :':'-':'.'~

I

/

/

======="=================================================================================================== =====================================================================~::'::::":::'::":": ==============:~===============,==,=//====//e. =====Oo ============================ :.:.'..: .: : : .:.: ':. ::.:.: .: .:. :.:.: .:.:. :':. :-...'.'. :.:-:.: .I...:.: .:-:..:.:.: : ... :.: -:,:.:.:.:-:.:...:. :-:.:',:-:-:-:.:: .'.'..:% .'. i========================================================================================= i." . . . . . . . . . . . . ......... ...... ... . .........., ~... .. , : ::::""::: ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::"::: ::::::::::::::::::::::::::::::: :::::::::::::::::::::: '~":::::'"" " ::::: : : i . ... :: :-::: '

,

." :: : :::

. ~0.0: . . . . . . . . . .: . . . '

~' ~ ~.

I "N

. . . . . . •. . . . . . . . . . . . . . . . . . ' . . . . . . . . . . " ..... 7. . . . . . . . . . . . . . . x ..... " . . . . . . . . "''" •...... ..... - ........ 0. .................. 200Ore.... ..... ~........ ..... ......:.~...........~..................

","

'-'.

. . . .

.-'.'.'

:'.'.

.','-'

"'.'.

.'.'.'.

"'"

. . . .

.''

")".

""

.

'

''"

.

.

.

.

.

.

....

,t"L'~

".'."'.',~

-

...-.... -..-.........-.-........................................................................;.. eoo......:........................,.:.~ ~....:.:......o.~ ,.. ..

.

-.

...

,,.

•

. . . . . .

.. . . . . .

. .

. . . . . . . .

....,

.

. . . . . . . . .

,

.~

. . . . . .

.,. . . . . . .

.

.....

~.

• .

..:

. . . . . . . . . . . .

~

................................................. •3. ............. ~ ...... ~'I.'" "'" "'" "''~" .... . ' . " . ' . ' . ' . ' . . '.. . .•" .: ..... ..... :. ".'-'.'.'. ..... '.'- :'- :.., - / .4:~.'.-'K. " ".'.'." .... ".'.'-":.:.:...." ' • ,~ .... ................. . ....." ......... . .... . ,........ '.'..-:... ........• • .:." •........ :.:~:...:..: .:.:.:..'.: " ..... .'.:.:.., t a..-..............-.......-....)

Fig.2.

Map

Oppermanskop),

of Blydefontein and

Upper

Basin

near

Section

Noupoort,

Pond

sediment

showing

the

sequence

den pollen and the regional vegetation, spectra are compared with those from modern surface pollen samples including soil scrapings and fresh hyrax dung, as well as nearby Holocene alluvial sediments. The investigation of hyrax middens is integrated with a research project on Later Stone Age archaeology and palaeoenvironments of Blydefontein Basin (Bousman et al., 1988). According to historical accounts the vegeta-

location

of hyrax

middens

~ ,.

N

,-, /

.... "i'i" i ' ' > '::: """:'""':"""": .... ~ ....... :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::

'''.'.'.'.'."

BAS

I \ I , , I \._ I \ "% .... /

'

.-

...... BLYD~-~.,,,,,~-r~

1 (near

/

"

~

"

:~

~I

Meerkat

~,,

Shelter)

_ _ ~ /

and

2 (near

(USP).

tion in the Karoo biome has changed significantly over the past few centuries. Roux and Theron (1987) ascribed these changes to a number of factors including abusive European farming practices, and climatic deterioration. The results of pollen analysis of hyrax middens provide a means of monitoring recent vegetation change in the Karoo, and comparing it with the model of change proposed by Roux and Voster (1983).

370

Setting The two hyrax middens described in this paper occur in small rock overhangs just below 1700m altitude, against the slopes of the Kikvorsberg. This range rises up to ca. 580 m above the surrounding plains, and consists predominantly of sandstones and subordinate mudrocks of the Tarkastad Formation, Beaufort Group. The first midden (1) is situated in the lower reaches of the perched Blydefontein Basin, in the Kikvorsberg, and immediately to the north of Meerkat Shelter, a Later Stone Age archaeological site, on an east-facing slope. The other midden (2) occurs west of Blydefontein Basin on a northeast-facing slope in a deep protected ravine below Oppermanskop peak, and in the Zeekoe Valley drainage (Fig.2). In both cases middens occur in small sandstone overhangs where hyraxes could hide from natural predators such as eagles and lynx (rooikat). The annual precipitation in the region is ca. 366 mm with the Kikvorsberg probably slightly wetter than the lower-lying plains, due to an orographic effect. The vegetation on the Kikvorsberg, classified as Karoid Merxmuellera Mountain Veld, consists of grassland with dwarf composite shrubs, while the surrounding plains consist of shrubby False Upper Karoo Veld (Acocks, 1953). Protected mountain slopes like those containing the hyrax middens are more woody than the surroundings, but these woody communities differ on various slopes. The vegetation at Meerkat Shelter in the floor of Blydefontein Basin, is not very divergent from that of the higher montane surroundings. It is comprised, of woody shrubs like Tarcho-

nanthus camphoratus, Rhus erosa, Euclea crispa, Diospyros austro-africana, and often supporting the ranunculaceous vine Clematis brachiata. Smaller shrubs include Euryops oligoglossus, Eriocephalus spinescens, Elytropappus rhinocerotis, and Polygala hottentotta. Riverine vegetation on alluvial sediments directly below Meerkat Shelter consists of Cyperus marginatus and Phragmites australis in swampy areas, and Artemisia afra on sandy areas.

L. SCOTT AND C. B. BOUSMAN

Woody elements in the ravine containing the second midden near Oppermanskop, are taller and more dense with a poorly developed understory of Compositae, Gramineae, and Melianthus sp. Typical montane woodland elements are: Rhamnus prinoides, Cussonia paniculata, Rhus burcheUi, Maytenus sp., Osyris sp., Diospyros lycioides and Tarchonanthus camphoratus. The increased density of woody species is probably due to the relatively frost free condition of the protected northerly orientated ravine, facing towards the warm low-lying plains. This ravine is apparently drier than Blydefontein Basin proper where Meerkat midden is situated.

Methods The middens were extracted from the overhangs. Vertical 'profiles' were cut by a bandsaw in the lab, and then cleaned. The Meerkat midden was 22 cm thick, and the Oppermanskop midden was 18 cm thick. Samples of ca. 2 g were taken at intervals of roughly 2 cm. Pollen was extracted from each sample by the following steps: boiling in 10~/o KOH; acetolysis treatment; washing in 40~/o HF; and mineral separation using ZnCL 2 solution with a specific gravity of 2. Microscope slides were mounted in glycerine jelly. Pollen concentrations were estimated by means of the exotic pollen method using Alnus pollen. Similar methods were used for the modern soil and alluvial sediment samples, except these samples were boiled in 10~/o HC1 instead of KOH. Three samples were submitted for radiocarbon dating from the midden pollen sequences.

Results The bottom section of the Meerkat midden, between 18 and 22 cm depth, was dated by radiocarbon to 300__ 35 yr B.P. (Pta-4403). Two samples the Oppermanskop midden at 4.37.3 cm and 15-18 cm depths were dated to 460+45 yr B.P. (Pta-4571) and 1130+80 yr B.P. (Beta-14658), respectively. The slightly thicker

PALYNOLOGY OF HYRAX MII)DENS OF SOUTHERN AFRICA

Meerkat midden seems to represent most of the last 300 years, whilst the thinner Oppermanskop midden may represent up to the last 1130 years. The Clarens midden from the Orange Free State apparently represents an accumulation of 15cm in approximately 30 years (Scott, 1989a and unpublished data). Causes for differential accumulation rates are not well understood, but factors such as hyrax population density, colony size, frequency and length of latrine use, and physical features which may involve secondary accumulation due to flowing urine, could be involved. Pollen diagrams of hyrax midden, alluvium, and surface pollen samples (Figs.5.3-6) suggest that Gramineae and Compositae pollen are the two most abundant types. Modern botanical surveys in the region demonstrate that both taxa represent numerous species (Acocks, 1953; Roux and Blom, 1979 and unpublished). Also important, especially in the Oppermanskop midden, is Scrophulariaceae-type pollen (Fig.4). Due to overlapping morphologies this group of' pollen cannot be assigned to Scrophulariaceae exclusively. However, as the family comprises a relatively prominent group of herbs and small shrubs in the Karoo, it is likely that most of the pollen grains recorded actually represent Scrophulariaceae species. Other prominent non-arboreal pollen types (NAP) in the diagrams are Chenopodiaceae/Amaranthaceae, Artemisia, Anthospermum-type, Ranunculaceae, and Cyperaceae. Arboreal pollen (AP) is only prominent in midden and slope spectra, and represents forms like Tarchonanthus, Ebenaceae (Diospyros and Euclea), Rhus and Lycium. The pollen concentrations (Figs.3 and 4) in the hyrax middens from the Karoo are comparable to those found in the midden from Clarens, (Scott, 1989a), in being generally below 100,000 grains per gram. The section of the Meerkat pollen diagram representing the last 300 years (Fig.3) suggests a gradual increase in Compositae and Chenopodiaceae/Amaranthaceae pollen, at the expense of Gramineae, Cyperaceae, and Ranunculaceae pollen. The Oppermanskop midden (Fig.4) shows a high peak of grass

371

pollen (76%) at the base of the 18 cm sequence estimated at ca. 1200yr B.P. This declines strongly between ca. 17 and 14 cm (ca. 1000 yr B.P.), when more Compositae. Scrophulariaceae, and Cyperaceae pollen types are present. Between 13 and 5 cm Gramineae pollen values recover partly, and Compositae and Scrophulariaceae-type values decline. A reduction in grass pollen is again shown in the upper 5 cm (younger than 460yr B.P.) of the midden, coinciding with increases in Rhus, Euclea, Compositae, and Chenopodiaceae/Amaranthaceae pollen. This change is comparable to that found in the Meerkat midden and likewise seems to span the last three hundred years. The relatively high Compositae and Chenopodiceae/Amaranthaceae pollen values, and low Gramineae values in the youngest sections of' both middens correspond with the surface pollen counts (Fig.5). Arboreal pollen (AP), especially Rhus, Euclea, Diospyros, and Tarchonanthus, is generally higher in the Oppermanskop midden (Fig.4) than in the Meerkat midden (Fig.3), probably because of denser woody vegetation around the former. Surprisingly, this is not reflected by the surface pollen samples near Meerkat midden (Fig.5), which show anomalously high Diospyros and Lycium pollen values. Four alluvial sediment sequences (BFS, BSM, CH2, and USP) from Blydefontein Basin are described by Bousman et al. (1988). The pollen from the alluvial deposits of the Upper Section Pond (USP), opposite the stream at Meerkat Shelter and midden, are used here for comparison to the Meerkat midden complex. USP shows a gradual succession from pool conditions with relatively high grass pollen values, ca. 2000 yr B.P., to a local marshy situation in more recent times, trapping mainly Cyperaceae, Chenopodiaceae/Amaranthaceae, and Compositae pollen (Fig.6). However, pollen spectra in the younger levels compare well with surface spectra from the area, and with the midden spectra. The differences in composition between the alluvial pollen spectra and those of the middens, i.e.

r

i

l

l

,

,

.

.

.

.

.

.

.

.

.

.

.

.

i

,

l

J

14C DATE (YR BP) ,

I> --

~

SAMPLE NO.

o~

X

~0OOCARPUS ~YRICA :~HUS

ITI Z

~

rARCHONANTHEAE

r11

~UCLEA

.YCIUM U) 1"

GRAM~EAE

m :~

¢D v

YPERACEAE AN~SPERMUM RANUNCULACEAE GUNNERA OXYGONUM ALOE-TYPE ASPARAGUS LILIACEAE (OTHER)

SCROPHULARIACEAE TYPE CARYOPHYLLACEAE CRASSULA GERANIACEAE LINACEAE SCABIOSA CAMPANULACEAE POLYGALACEAE GALIUM

v q

TRILETE SPORES ~IONOLETE SPORES ~ICCIA JMBELLIFERAE

TYPE

-ABIATAE ~YCTAGiNACEAE

3OMPOSITAE

~'RTEMISIA 3TOEBE TYPE ~6.SSERINA RESTIONACEAE

CHENO/AMARANTHACEAE AIZOACEAE

TYPE

RUSCHIA

Ir

TRIBLILUS TyPE A OTHER

POLLEN/GRAM

NVNSfl08 '~ "D (~NV &&03S "rI

g g C~

[YR BP) DEPTH (CM)

I

×

SAMPLE NO. ~DOCARPUS

~

;USSONfA ~YRICA THUS )LEA WAYTENUS ~HAMNACE~ EUCLEA

~RCHONANTHEAE v

Ec P

© CYPERACEAE ANTHOSPERMUM RANUNCULACEAE i

©

PODOSTEMACEAE ~LOE T Y P E ~SPARAGUS TYPE

v r

-ILJACEAE (OTHER) ~U~RBIA

TYPE ;OMMELINACEAE :RASSULA ]ERANIACEAE : AMPANULACEAE ~OLYGALACEAE 3ALIUM rRILETE SPORES ~ONOLETE SPORES ~ELIANTHUS T Y P E )MBE LLIFE RAE TYPE .ABIATAE

V

)LANTAGO

.ACTUCOqDEAE ;TOE BE TYPE ;LIFFORTIA [RICACEAE =ASSERINA

;HENO/AMARANTHACEAE UZOACEAE TyPE

v m

~USCHIA &CANTHACEAE

)THER

~L

I

l

I

•

I

I

I

•

,~ ~

I

l l

i i

I

-

AP

I

.~

I

•

l

I

I

I

I

I

I

~'

II

/

l I

•

I

1

,~

•

I

•

~

I

I

•

I

I

I

•

I

SCALE

20%

I

i

I

=< '~ .~

Fig.5. Surface pollen percentages from Blydefontein Basin, Noupoort.

SLOPE l

I

I

II

SLOPE 2 110414

I

t

•

I

=

l,o,,,s

I

~413] I

E

~"~"1

OPPERMANSKOP:

~ x

B

DUNG 10416

ALLUVIAL MUD ~

~

SHELTER

SORCE'ENSPECTRA ! l

I

I

I

• l

•

1

~

iii I

1

~

I

I

I

I

I

I

I

1

I

I

NAP

I

I

I

II

|

~

I

I

I

II

I~

I

II

H i

I

II

I

I

l

= ~ ~oi

z

I

i

I

i

I

I

~

I

It

I

3E

I

I

i

I

!

I

375

PALYNOLOGY OF I-fYRAX MIDDENS OF S O U T H E R N AFRICA

more swampy elements in the alluvial deposits and more woody elements in the slope samples, are assumed to reflect local vegetation patterns rather than dietary preferences of hyraxes. The lack of local Artemisia pollen in alluvial deposits (Fig.6), and the high proportion of it in the midden adjacent to Meerkat Shelter (Fig.3), may arguably be due to hyrax dietary preferences. The bottom two Meerkat midden pollen spectra show low percentages of local Artemisia pollen in the area surrounding the midden at that time. The uppermost USP samples may be roughly coeval with these two Meerkat midden samples. Presently Artemisia pollen production is continuing as shown by the surface spectra (Fig.5). At Clarens (Fig.l), palynological studies of Holocene alluvial material and hyrax middens from two different sites, show a contrasting picture (Scott, 1986, 1989a). Near both sites Artemisia afra can be found, but its pollen is much more prominent in alluvial sediments than in the hyrax midden. In this case selective concentration of Artemisia pollen in the midden did not occur. The high proportion of Stoebe-type pollen in the recent slope and alluvial samples are surprising as they are comparable to late Pleistocene or Early Holocene spectra from Blydefontein Basin (Boysman et al., 1988). Similar Stoebe-type values are not recorded in the 300 year old Meerkat midden (Fig.3) nor in the alluvial sediment sequences of the area (Fig.6, and Bousman et al., 1988). This suggests that an increase in production of this pollen is associated with recent droughts of the early to mid 1980s. Dense stands of Elytropappus rhinocerotis were observed at higher elevations in the Basin, and their pollen is probably incorporated in the alluvium by water transport. Very recent botanical changes in the Basin are not represented in the hyrax midden, and this suggests that the middens apparently have not been used in the last one or two decades. It is possible that: intensified hunting pressure during the last 20 years has caused a disruption of hyrax latrine use in the Basin (Dou Lessing, pers. comm.), and this may account for the

discrepancy between the surface samples and the uppermost midden samples. Discussion

Modern botanical survey data for the Karoo region (P. W. Roux and C. D. Blom, 1979, unpublished) and pollen data from Blydefontein are being analyzed and compared statistically, and this promises objectively to calibrate past vegetative changes with modern vegetation patterns in the region (Bousman and Scott, unpublished data). However, a preliminary intuitive interpretation of the fossil pollen data follows. In general, pollen from Oppermanskop midden suggests relatively moist grassy conditions between ca. 1200 and 300yr B.P., with the exception of a short interval around 1000 years ago when the vegetation was more shrubby. As inferred from pond deposits and grassy pollen spectra from Upper Section Pond, relatively wet conditions must have prevailed since at least 2000 yr B.P. After 300 yr B.P. dry shrub vegetation increased, and grassy elements declined, as reflected by pollen in both hyrax middens. It is not clear if the drying trend in the younger samples from USP, which were probably deposited at or before 300 yr B.P., is due to local sediment filling the pond, or if it marks the beginning of veld deterioration as suggested by the midden sequences. The changes indicated in the vegetation of the Blydefontein area during the last one or two millennia can be ascribed to either climatic change or human influence. The most recent and severe deterioration of the area's veld, during the last 100 to 200 years, in part, is the result of European stock overgrazing (Acocks, 1953; Roux and Theron, 1987; Roux and Voster, 1983). However, a major drought in the late 19th and early 20th centuries coincides with the most intensive overgrazing in the region (Roux and Theron, 1987; Vogel, 1988; and Bousman, unpublished data), and indicates that the changes observed in fossil pollen frequencies at this time are the result of a combination of factors.

o

DIATOMACEOUS POOL DEPOSITS

~U'":

' , , ' 2 0O' '~ , = '',, ~, ,

~

0

~

0

' ." ,. . . •. . . . . " .

"~ ~ ~ 0

0

0

~~

-~ ~ ~ 0 ~D

0

~ 0v0M . L .L

•

.

"

~ 0

0

co m ~

~

~

0

0

0

~

f '~~?~~'~~1 .':

'.,.'.:..:.,.2:-,..

. : •: : . S O L ' . ' : : ' . . ' . .

o') ql ~ ~ 0

~

0

~ ~,

0 rO 0 0

~

~

C

L

oIDEPTH(cm} ]

~D 03

s~PLE ~o. PODOCARPUS

Z O

CLUTIA TYPE

c~

B C w~ T~ e-F o o

THUS )LEA

!!;!!!!,, I'

&Q

•

"

"I

•

--

m

EUCLEA )IOSPYROS ~RCHONANTHEAE

;RAMINEAE

i I

•

;YPERACEAE ANTHOSPERMUM $UNNERA

I--

RANUNCULACEAE TYPHA 3XYGONUM JLIACEAE cb

•

-

SCROPHULARIACEAE T Y P E

m

3ARYOPHYLLACEAE

~T~

3RASSULA SCABIOSA

i

POLYGALACEAE TRtLETE SPORES ~ONOLETE SPORES 3PHIOGLOSSUM RICCIA

i

ANTHOCEROS ~YDROCOTYLE

i

JMBELLIFERAETYPE

II1,111

nI n

OOMPOSITAE ARTEMtSIA LACTUCOIDEAE STOEBETYPE THYMELAEACEAE

I

I'll

•

•

u

•

"

=m

i

•

n i

n i

CHENO/AMARANTHACEAE AIZOACEAE TYPE

i

RUSCHIA TRIBULUS i ~

i

i

m

-

3THER

-

POLLEN/GRAM

NVI~ISgiO~t "~t "3 (]NV J,,l, OOS "'I

9L~

P A I , Y N O L O G Y O F H Y R A X M I D D E N S OF S O U T H E R N A F R I C A

The possible influence of prehistoric San hunter-gatherers and Khoi pastoralists should be considered in interpretations of vegetation change for the Karoo and Blydefontein Basin. By associating archaeological residues of San camps with classic botanical indicators of veld disturbance Sampson (1986) suggests that San groups could have damaged only roughly 2% of the plant cover in the upper Zeekoe Valley west of Noupoort before the arrival of the European Trekboers in the middle 1700s. This estimate of prehistoric vegetation damage is only a small fraction of that estimated for modern European stock (Roux and Theron, 1987). Additionally, only limited evidence of Khoi herders was recovered at Blydefontein Rockshelter (Bousman, unpublished data). This herder occupation is dated between 1300 yr B.P. and 1200 yr B.P. when the most grassy conditions existed in the Basin, but not during the period of veld deterioration at 1000 yr B.P. Thus it seems unlikely t h a t overgrazing by prehistoric Khoi domestic stock was a significant factor in botanical changes documented at this time, and, therefore, palaeoclimatic fluctuations, rather than human interference, most likely caused the changes in veld conditions before 200 yr B.P. Events reflecting veld deterioration at approximately 1000 yr B.P. and 3000 yr B.P. are probably due to drier climatic conditions. Evidence in support of a brief dry interval at Blydefontein at 1000 yr B.P. is collaborated by a major dip in a continuous 13C/12C ratio curve on ostrich eggshell excavated from Blydefontein Rockshelter and spanning the interval between ca. 4300 yr B.P. and 500 B.P. (Bousman et al., 1989). The dip in carbon isotope values indicates that a significant increase in C3 plants occurs at the same time as the Oppermanskop midden pollen spectra demonstrate a reduction of grasses (C 3 and C 4 species) and an increase in composites (only C 3 species). Also, a pollen sequence from spring deposits at Deelpan in the Orange Free State approximately 200 km north of Noupoort registers a similar reduction of grass coeval with the Oppermanskop grass reduction (Scott,

377

1988). This suggests that the 1000 yr B.P. grass fluctuation was not a local event. In general, there seems to be a remarkable similarity between Roux and Voster's (1983) model of veld deterioration, based on historical accounts and modern veld observations, and the pollen data in the hyrax middens. According to the model deterioration goes through five stages: (1) primary degradation loss of grasses, (2) denudation, (3) revegetation increase in composites, (4) secondary degradation, and (5) desertification. The resolution of pollen analysis is, however, not high enough to identify the five stages through which veld deterioration is thought to pass. The correspondence between the hyrax pollen data and this model, and that of Karoo spreading proposed by Acocks (1953), demonstrates the usefulness of hyrax midden analysis in palaeoenvironmental reconstructions of arid and semi-arid regions. In view of hyrax abundance in Africa, a search for more middens promises to be fruitful. A series of hyrax middens covering relatively long time spans in rocky and mountainous areas, can potentially provide Late Quaternary proxy palaeoenvironmental data for large parts of arid and semiarid Africa. Numerous studies have demonstrated the great variability in pollen spectra associated with packrat nests in north America (King and Van Devender, 1977; Thompson, 1985), and this possibility with all its attendant problems should be considered for hyrax middens in Africa. However, at present the general agreement between the Meerkat and Oppermanskop pollen sequences, their similarity to the Upper Section Pond alluvial pollen sequence and the Deelpan pollen sequence, and the strong similarity of signal from the ostrich eggshell stable carbon isotope curve and the Oppermanskop pollen changes, all suggest that significant variability does not occur, and that hyrax middens are fairly unbiased fossil pollen traps. Variability of A r t e m i s i a pollen discussed above, cannot conclusively be attributed to hyrax diet, and is possibly related to local vegation patterns. The processes of pollen

378

taphonomy are at issue, but modern studies on pollen survival in hyrax digestive systems, correlation between hyrax diet and pollen composition in hyrax dung, and other questions concerning the nature of pollen transportation and entrapment should be addressed by future research. Stable carbon isotope analysis or other chemical analyses of hyrax middens along with pollen studies could provide important new dimensions for palaeoenvironmental reconstructions (Erasmus et al., 1978). While macrofossils are not numerous in hyrax middens, examples from Jordan have produced a variety of macrobotanical rests (Lindquist and Fall, 1987), demonstrating even greater potential for hyrax midden studies. Analysis of both micro- and macrofossils from hyrax middens could provide more accurate palaeoenvironmental reconstructions. An example of a midden, other than hyrax, and probably belonging to the dassie-rat (Petromus typicus), has been reported from the southern Namib Desert region (Scott, 1989a). This suggests that in some areas hyrax middens studies can potentially be supplemented and tested by the investigation of other kinds of middens. More research is necessary in order to identify additional midden-producing species in Africa.

Acknowledgements Dr. J. C. Vogel of the Council for Scientific and Industrial Research, Pretoria is thanked for supplying two radiocarbon dates, and Professor P. S. Martin, University of Arizona, funded the third age determination. Also, we wish to thank Julio Betancourt for information on packrat middens. Fieldwork for this research was funded by grants from the National Science Foundation, the Wenner-Gren Foundation, the L. S. B. Leakey Foundation and the Institute for the Study of Earth and Man, S.M.U. Mr. and Mrs. Dou Lessing, the owners of Blydefontein, generously allowed our research activities to occur on their farm; we thank them. Danny Boy Meyer assisted in

L. SCOTTAND C. B. BOUSMAN

the location and extraction of the Oppermanskop hyrax midden.

References Acocks, J. P. H., 1953. Veld types of South Africa. Mem. Bot. Surv. S. Afr., 28. 1-192. Betancourt, J. L., Van Devander, T. R. and Martin, P. S., 1989. Fossil packrat middens: The last 40 000 years of biotic change. Univ. Arizona Press, Tucson, in press. Bousman, C. B., Patridge, T. C., Scott, L., Metcalfe, S. E., Vogel, J. C., Seaman, M., and Brink, J. S., 1988. Palaeoenvironmental implications of Late Pleistocene and Holocene valley fills in Blydefontein basin, Noupoort, C.P., South Africa. Paleoecol. Afr., 19: 43-67. Bousman, C. B., Scott, L., Shackleton, N. and Vogel, J. C., 1989. Pollen, stable isotopes and paleoenvironmental reconstructions in Africa. Paper presented at the Society of American Archaeology, Atlanta. Erasmus, T., Penzhorn, B. L., and Fairall, N., 1978. Chemical composition of faeces as an index of veld quality. S. Afr. J. Wildl. Res., 8:19 24. Foerie, L. J., and Perrin, M. R., 1989. Quantitative and qualitative aspects of the diet of the rock hyrax (Procavia capensis Pallas, 1766) in the M o u n t a i n Zebra National Park. Afr. J. Zool., in press. Hoeck, H. N., 1975. Differential behaviour of the sympatric hyrax Procavia johnstone and Heterohyrax brucei. Oecologia (Berl.), 22: 15-47. Green, L. G., 1955. Karoo. Howard Timmins, Cape Town, 248 pp. King, J. E. and Van Devander, T. R., 1977. Pollen analysis of fossil packrat middens from the Sonoran Desert. Quat. Res., 8: 191-204. King, T., Jr., 1976. Late Pleistocene-Early Holocene history of conifer woodlands in the Luzerne Valley region, Mohave Desert, California. Great Basin Nat., 36(2): 227-238. Lensing, J. E., 1978. The feeding ecology of the rock hyrax, Procavia capensis Pallas 1766, in southern South West Africa. Final Rep., Dep. Natl. Conserv. Tourism, South West Afr. Admin., 253 pp. Lindquist, C. A. and Fall, P. L., 1987. Fossil hyrax middens from the Middle E a s t - a new source of palaeobotanical evidence. In: Int. Union Quat. Res. 12th Int. Congr., Ottawa, July 1987, progr, with abstracts. Natl. Res. Counc., Canada, p. 212. Morris, D., 1965. The mammals. A guide to the living species. Hodder and Stoughton, London. Pons, A. and Quezel, P., 1958. Premieres remarques sur l'~tude palynologique d'un guano fossile du Hoggar. C. R. Acad. Sci., 246: 2290-2292. Roux, P. W. and Blom, C. D., 1979. Vegetation surveys in the Karoo region. Grootfontein Agric. College, Middelberg, South Africa (unpublished). Roux, P. W. and Theron, G. K., 1987. Vegetation change in the Karoo biome. In: R. M. Cowling and P. W. Roux (Editors), The Karoo Biome: a preliminary synthesis. P a r t 2 - - vegetation and history. South Afr. Nat. Sci.

PALYNOLOGYOF HYRAXMIDDENSOF SOUTHERNAFRICA Progr. Rep., 142. Found. Res. Dev. Coune. Ind. Sci. Res., Pretoria, pp. 50 69. Roux, P. W. and Voster, M., 1983. Vegetation change in the Karoo. Proc. Grassland Soc. S. Afr., 18:25 29. Sale, J. B., 1965. Some aspects of the behaviour and ecology of the rock hyrax (Genera Procavia and Heterohyrax). Thesis. Univ. College, Nairobi, Kenya, 217 pp. Sampson, C. G., 1986. Veld damage in the Karoo caused by its pre-Trekboer inhabitants: preliminary observations in the Seacow Valley. Naturalist, 30: 37-42. Scott, L., 1986. Pollen analysis and palaeoenvironmental interpretation of Late Quaternary sediment exposures in the Eastern (}range Free State, South Africa. Palaeoecol. Afr., 17:113 122. Scott, L., 1988. Holocene environmental change of western

379 Orange Free State pans, South Africa, inferred from pollen analysis. Palaeoecol. Afr., 19:109 118. Scott, L., 1989a. Hyrax (Procaviidea) and dassie-rat (Petromuridae) middens in paleoenvironmental studies in Africa. In: P. S. Martin et al. (Editors), Fossil Packrat middens: the Last 40 000 Years of Biotic Change. Univ. Arizona Press, Tucson, in press. Scott, L., 1989b. Pollen analysis and palaeoenvironmental significance of Quaternary faecal deposits in Africa. Paper presented at the 4th Int. Conf. Environmental Quality and Ecosystem Stability, Jerusalem, 1989. Smithers, R. H. N., 1983. The mammals of the southern African subregion. Univ. Pretoria, 736 pp. Vogel, C. H., 1988. 160 years of rainfall of the Cape has there been a change. S. Afr, J. Sci., 84:724 725.