See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/285022134

Ethnoarcheological contributions to the taphonomy of human sites

Article · January 1980

CITATIONS READS
54 109

1 author:

Diane Gifford-Gonzalez
University of California, Santa Cruz
98 PUBLICATIONS   2,706 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

An Introduction to Zooarchaeology View project

All content following this page was uploaded by Diane Gifford-Gonzalez on 04 December 2016.

The user has requested enhancement of the downloaded file.

1978 Gifford, D. Ethnoarchaeological observations on natural processes affecting cultural materials.
In Explorations in Ethnoarchaeology. R. A. Gould, editor. University of New Mexico Press, Albuquerque.
Pp. 77-101.

Ethnoarchaeological Observations of
Natural Processes
Affecting Cultural Materialsl

DIANE P, GIFFORD
Meñll CoIIege
Uniye¡sity of C¿lifomid, Santd Cruz

INTRODUCTTON
Ethnoarchaeological research has thus far focused mainly upon
gathering data with which to test hypotheses concerning the
relationship between culture process and its direct material
manifestations. While cultural factors a¡e the primary agency
producing the remains with which prehistorians are concemed, there
are in addition noncultu¡al facto¡s which can substantially alter
cultural materials as they become part of the archaeological record.
These processes have never been described in a way generally useful in
evaluating their influence upon archaeological materials. Since these
noncuìtural factors affecting archaeological sites are processes, they

77
 DIANE P. GIFFORD
can-like culturâl processes-be obse¡ved in the contemporary world,
and thei¡ effech clarified by ethnoarchaeological research.
Recently a growing numbe¡ of archaeologists have recognized the
need to define noncultural processes more closely. Ethnoarchaeologi-
cal observe¡s have commented upon the problem of differential pre-
servation of contemporary sites and the skewed representation of cul-
tural materials in preserved contexts (Gould 1974; yellen 1974). Other
researchers have undertaken experimental work over the last fifteen
years in to explicate natural processes of alteration (Isaac
attempts
1967; Jewell and Dimbleby 1966).
Conside¡ation of natural processes of alteration should be of primary
concern to all archaeologists, especially to those working with sites that
lack enduring structures, es do most hunter-gatherer and many pastoral-
ist occurrences. In these cases, behavioral ¡econstructions have been
drawn exclusively from the distribution of materials upon living floors.
Wheat's Olsen-Chubbuck work (1972), for example, and Leroi-Gou¡-
han's Pincevent investigations (Leroi-Gourhan and Brezillon 1966), base
extremely detailed behavioral inferences upon such dist¡ibutions. These
investigations are notable for the integration of microstratigraphic and
sedimentological data with other evidence in drawing such inferences.
In many situations, ho'ivever, direct behavioral reconstruction from
intrasite distributional patterns may be considerably complicated by
both cultural and noncultural facto¡s. Schiffer (l97Za), in his
discussion of primary and secondary refuse, has c¡iticized the
assumption on ihe pårt of many archaeologists thai spatial clusters of
a¡tifacis inva¡iably reflect activity loci. He points out that on rnany
sites the primary factor structtring cultu¡al materials as recovered by
the archaeologist may be refuse-disposal behavior. Noncultu¡al
processes, especially those operant after site abandonment, may
likewise alter spatial patterning in cultural materials. In addition, these
processes may selectively destroy perishable materials, skewing the
originally constituted sample.
This chapter will discuss the more common noncultu¡al processes
affecting archaeological materials prior to burial, and the role of these
in determining "archaeological visibility." Special reference will be
made to the effects of contemporary processes on site preservation in
the East Lake Turkana (formerly Lake Rudolf) region, and to the
implications of this research for evaluation of regional archaeological
sequences.
78
N atutal Processes Affecting Cultural Mateñals
POSTOCCUPATION PROCESSES
Most human ìiving sites that become archaeological sites undergo
some kinds of depositional processes. These may be aqueous or eolian
in nature, and can vary tremendously in their effects upon a site.
Moreover, the span of time intervening between the abandonment of
such a site and its burial can vary considerably; during this period a
range of natural processes may rearrange or destroy va¡íous cultural
materials. The most common of these processes and their effects are
represented schematically in Figure 4.1. It must be emphasized that
the "stages" in Figure 4.I actually form a continuum, and that the
processes listed fo¡ one "stage" may well occur in another. For
example, animal disturbances cân occur during occupation as well as
during later stages. The following cases exemplify specific alterations
caused by such processes.
Plant disturbance can cause major rearrangements in materials ofall
kinds. Treefalls are capable of lifting and reconcentrating artifacts
lying in or on thei¡ root systems in strikingly consistent and patterned
shapes. As the fallen tree decays, it may leave a stain pattern in the soil
liable to furiher misinterpretation. At the Dutch Mesoliihic site of
Bergumermeer, Newell and Musch (personal communication) have
discerned the "ghosh" of treefalls interspersed with those of actuaì
structures. These two features differ in their shape and in the
placement and concentration of artifacts within. It is possible that
some of the hearthless "huts" and "tent circles" in forested areas of
No¡thern Europe, such as those reported by Rust (1943), are actually
the relics ofancient t¡eefalls.
These and other disturbances of the occupation layer can often be
discerned through detailed examination of site stratigraphy.
In his excellent ethnoarchaeological research on the Kalaha¡i
Bushmen, Yellen (in press) noted that nearly all identifiable bone not
trodden into the substrate during occupation was quickly consumed by
wildlife. Any marrow-bearing bone was eaten by camivores, and the
ecology of the Dobe region is such that even bovids consume bones,
apparently to obtain salts and mine¡aìs otherwise lacking in the
envi¡onment. Yellen observed ihat the bones remaining on the surface
tended to be those oflarger animals, especially the diagnostic articular
ends, and that subsurface large-animal elements were predominantly
nonidentifiable fragments. The resulting selective destruction of large
79
 N atural Proces ses Affecting CuIturuI Materí¿ls
STAGE PROCESS MATERIAL MANIFESTATION MODT OF EXPLICATION

ânimåls'bones skewed the large-species representation in the

recoverable sample by as much as 50 percent.
F atterning in form
pattern.i
human behavior-
behavior- pl¡acenent of
and pl ethnographic model s' At Lake Turkana bones on abandoned campsites are not consumed
cul
ul tural
tural materi al s experimentation by wildlife. But sites exposed to the elements for upwards of five years
I
undergo a loss of identifiable su¡face bones ihrough weâthering
.___L__ I
attrition, with the smalle¡ elements of the surfâce samplé decomposing
aninal
an i¡nal
---.-..\ earlier. Ultimately, if all surface bones were to decay, a skew effect
-'---'\ I
¡- plant -
Dlant tural¡ naterials
cuultturâ similar to that reported by Yellen would occu¡.
'----.-\, \ sel êcti vel ! repatterned
cti vely anaìogy with modern Processes causing differential destruction of culturâl rnaterials may
:^
clinatic/
clinati -'
¿-.'- and/or
and/or destroyed envi ronments
leave little evidence of their operation other than that of the structure

Y'_'
:l"l
aqeuous
-tt'-

___l__
I
I

I
of the sample itself. Ethnoarchaeological research can describe the
distinctive features of assemblages affected by such processes. These
data can then be used in much the same manner as studies of
distinctive structures resulting from obse¡ved geologic processes, at
l-- --__-_. culul tura l nater
nateriais
se ectively
sel ctively repat
s repatterned micros trati qraphic analysis, least indicating, if not achrally proving, the likelihood of operation of
' and/or destro
destroyed
experlnentati such processes on prehistoric samples. For example, Yellen notes that
êolian. on
the preserved subsurface bone samples from his Kalahari sites consist
:'::::- .-____L
.___L I

mainly of whole bones of smaller animals, plus small, largely

I
unidentifiable fragments of larger animals' bones. Recovery of a
¡- soi I pH ----- I
ted destru
ì imi destructlon microstrati graphic analysi s, sâmple in which large animals were represented only by fragments of
- nd repatterni
and repatterning of similar size to small animals' bones may indicate the operation of such
cuì
uì tural material s
mater chenical analysis processes. Àdequate knowledge of regional paleoecology is a basic
erosion / -
I requirement in assessing the likelihood that ce¡tain plocesses were
I
operant upon an assemblage. Weathering and biological damage to
-----l---
---l------ I
perishable materials vary in nature and rate depending upon local
climatological and ecological conditions, and it is imperative to
t'- sanpl i ng
I
-----
--- - culul tural mater
materials establish whether or not observed destruction phenomena could hâve
recovered iin
-..// archaeological
lal -.-":
accidental
haeoloqical context
c
occurred under ancient ecoìogical conditions.
:tion
destructi on
While technically not a postoccupation process, trampling of small
obi€cts in the substrâte by site inhabitants deserves mention he¡e. It
may be a site formation process of as yet unrecognized importance.
FIGURE 4.1. Schemâtic representation of stages of formation of an archaeological site, Yellen's and my own observations of newly vacated sites indicate that
from c¡eation to recovery. Shows processes affecting cultural materials and primary
methods ofdetermining their presence and effects.
normal human activity on a loose subst¡ate will result in subsurface
migration of most smaller elements. At Site 20, mapped on the day of
abandonment and lâter excavated, I recovered about nine times as
many pieces (1,953) from the subsurface layer as had been visible on
the surface (200). These subsurfâce elements averaged three
centimeters in maximum dimension, with some long, narrow pieces
reaching five centimeters. An occupation of four days' duration

8t
 DIANE P. GIFFORD
was sufficient to effect this size-dependent sorting, with a fìne sand
substrate.
Trampling may have several important effects upon cultural
materials, which can be invesligated through both experimental and
ethnoarchaeological research. It shouÌd be possible to use results of
such resea¡ch to disce¡n trampling effects in archaeological
occuffences,
As in Yellen's Kalahari example, trampling of smaller elements into
the substrate, if combined with destruction of larger surface elements,
can substantiaìly bias the presewed sample. Surface samples of imper-
ishable as well as perishable materials can be skewed, should the sur-
face be modified by geologic processes. Experimental and ethnoar-
chaeoìogical observations should be directed at ascertaining whether
subsurface zones consistently develop with permeable substrates and
occupation activities, and whethe¡ these zones can be distinguisl.red.
Hypotheses which should be tested include the following:
(l) Given a permeable substrate, normal occupation activities on a
site ("trampling") will generate a subsurface zone of like-sized
items, definable within a given range of maximum dimensions,
and a surface zone composed of all items of larger maximum
dimension than those in the subsurface zone.
(2) The amount of subsurface migration of pieces due to trampling
depends directìy upon the permeabiìity of the substrate and the
âmount of trempling activity.
(3) The greater the median grain size of the substrate, the larger the
median maximum dimension of elements migrating subsurface.
There may be a highly significant relationship between trampling
phenomena and the preservation of primary refuse. Schiffer (1972a,
1975d, Chap. 9, below) has pointed out the important distinction
between primary and secondary ¡efuse. The former consists of
materials discarded at thejr use location, the latte¡ are discarded away
from the location of use. On the basis of students' obse¡vations of
Naci¡ema discard behavio¡, Schiffer proposes that size is a major
factor in determining whether an item becomes primary or secondary
refuse. Elements less than one inch (2.7 cm) in maximum dimension
appear highly likely to become primary refuse (Schiffer: Chap. 9,
below). This is close to the median size of subsurface pieces tecovered
from Site 20. A second set of hypotheses concerning this relationship
may be proposed for testing:
82
N dtural Processes Affecting Cultural Mateúals
(4) Elements below 3 cm meximum dimension are highly likcly to
become primary refuse, while those above this size are less likely
to be primary refuse.
(5) Given the conditions desc¡ibed in (l), elements.likely to be
primary refuse are also likely to migrate subsu¡face due to
trampling.
(6) A subsurface trampling zone will contain a relatively higher
proportion of primary refuse than will the surface zone ofa site.
I would agree with Schiffels (I972a) proposition that increasing
intensity of occupation (a function of population size and duration of
occupation) will be paralleled by decreasing primary refuse and
increasing secondary refuse. However, when considering the actual
material consequences of the interaction of these variables, I would
add the additional variable of substrate permeability. Substrate
permeability may be a major factor in primary refuse preservation,
acting in a feedback system with cultural fo¡mation processes during
occupation. A nonpermeable substrate (stone, beaten earth, and the
like) will not absorb culh:ral mate¡ials to any degree, increasing the
probability on intensively occupied sites that all refuse will become
secondary. A permeable substrate, on the other hand, may absorb
smaller primary refuse elements, as well as lost items, allowing them to
be preserved through an out-of-sight-out-of-mind process.
I am proposing here that trampling by site inhabitants may preserve
a ce¡tain class of cultural materials from subsequent cultural and
natural disturbances. Further, I am proposing that deposits of
trampled materials may have diagnostic features. Ethnoarchaeological
and experimental ¡esearch can test these propositions.
With regard to the effects of geologic processes upon cultural
matèrials, archaeologists can profit from a rapidly expanding literâture
on contemporary processes and thei¡ effects in the fields of
sedimentology, geomorphology, and paleontology (J. R. L. Allen 1963;
Bigarella 1972; Selley 1970; Visher 1974. Of special interest to
archaeologists concerned with postoccupation processes is the recent
¡esearch of vertebrate taphonomists,' which has included considerable
observation of natural processes and experimentation (Behrensmeyer
1975; A. P. Hill 1975; Voorhies 1969a, 1969b). Since the major concern
of taphonomists is closer definition of the processes of alteration
intervening between the death of an organism and the deposition of its
¡emains in geological context, these investigators share many of the
83
 DIANE P. GIFFORD

same interests as ârchaeologists studying postoccupation processes.

The major conclusion de¡ived from these studies, which should be
considered carefully by archaeologists, is that attempts at paleoecolog-
ical and behavioral reconstructions must be based upon a thorough
knowledge of the processes intervening between life situation and
geologic deposition (4. P. Hill 1975; Voorhies 1969b).
One a¡ea of common concern is that of biological alteration proc-
esses, and numerous data on bone damage and weathering are availa-
ble to interested prehistorians (Behrensmeyer 1975 A.P. Hill 1975).
Wíth regard to geological processes, taphonomists have drawn upon
and augmented sedimentological process studies. A primary principle
is that bones are sedimentary particles which, in the course of
deposition, behave according to definabìe, predictive rules. Since most
fossil-bearing deposits are aqueous in origin, most research has
focused on defining the behavior ofbones in flowing water, at different
velocities, with various types of bed load (Behrensmeyer i975; Voor-
hies 1969a). The ¡elevance of this research to many archaeological
situations is great, as ihe following discussion of my ou'n ¡esearch will
indicate.

DASSANETCH SITES

Formation P¿tterns
I worked in northern Kenya arnong the Dassanetch people who live
on the easte¡n shore of Lake Turkana, immediately south of the
Ethiopian border (Fig. 4.2). The Dassanetch homeland extends from
the Omo River Delta in Ethiopia down the eastern shore o[the Iake, as
far as fifty kilometers south of the Kenyan border. The Dassanetch
have been studied for the first time during the past seven yeârs
(Almagor l97l; Carr 1977). They view themselves âs pastoralists,
keeping cattle, donkeys, sheep, and goats; dogs are rare. They do,
however, cultivate millet ând maize in the few areas suitable for crops
in their arid land. Poorer members of the tribe supplement their diei
with animals caught in the lake. ,{t Ileret, where I conducted most of
my work, the majority of people belong to the Inkoria territorial
section. Raiding by traditionally hosiile tribes to the south and east still
occurs. Due to these hostilities, the Dassanetch in the region live in FIGURE
4.2. East Lake Turkana study âreå. Area 100 contains 15 manyatta and
defensive settlements, with the house area entirely surrounded by stock-camp sites. Area000contaìns l9gal diøs hunting-fishing sites.
thorn-brush animal pens (Fig. 4.3).

A4
 )< N aturdl Ptocesses Affecting Cultural Materials
O
9tn Because water is scarce, settlements lie relatively close to the lake.
Most ere located along the lower reaches of major ephemeral rivers,
,fr=fr where wells fo¡ subsurface water can be dug. Lake water is used mainly
/=
q to water stock; it is potable, but somewhat alkaline, and water from
riverbed wells is prefer¡ed for human consumption. Aerial photographs
(I:24,000) of the region clearly show the arrangement of larger settle-
ments along major watercourses. Two types of pastoralist sites exist:
large semipermanent ones inhabited by men, women, and children, re-
ferred to here by the Swahili term manyatta, anià mobile stock camps

o staffed by boys and young men. Manyattas are usually occupied for a
span ofyears, but settlements of a few months' duration may be rnade in
co E areas where waterand forage are temporarily good. Milking
stockfor the
,C families is kept at the manyattas, but most ofthe he¡ds are sent to stock
camps. Hundreds ofhead ofstock may be kept in these camps, which

o Ê shift location to take advantage of new vegetation that springs

up after
scattered showers. In times of scarcity, cattle, sheep, and goats are all
o
o flo herded in sepârate areas according to their differing forage require-
Ê ments. During the
time I was at lleret, the land had been very dry for

to ouoËcoeË o about seven years, and nearly all the caitle had been sent north to better
grazing lands near the Omo Delta, leaving only donkeys and small stock
in the south. Stock camps a¡e often located inland or along the lake, in
areas lacking sufficient watér for larger human populations. In the
=
O upper reaches of the ll-Erriet River Valley herdboys were expected to
o l¡l live on meat, milk, and whatever water they could obtain from the

t%?3 F--
(/)
ut
o riverbed.
The manyattas are characterized by a relativelyìarge core living area,
with houses a¡ranged in roughly linear order according to the kin rela-
f z. tions of the households. Manyâttås may contain between I0 and 80
houses (Fig. 4.3). The houses, built by each married woman with off-
t- t-
I
U. spring, are entirely portable one-room huts of skins and hides, lashed
over a saplingframe. They are low (1.5 m) ovals, measuring about 3 by 4
=
:tr
ut
Iz meters. Each contains a hearth, always placed to the immediate left of
Ld the entrance. Depending upon local availability of stone, hearthstones
U) o may or may notbe left in place upon abandonment of the site. The su¡-
f (9
o
I -Ul É
o
rounding animal pens are shared by male kin and friends. Small stock
and calves are kept in the inner pens, larger stock in the outer pens. In-
formants agreed that although some areas are habituaìly reoccupied, it
og@ was forbidden to establish a new manyatta directly on top of the visible
remains ofan old one, aìthough fences and other components could be
reused.

e'7
 DIANE P. GIFFORD

During and after occupation, the house area is distinguishable from

the pens by its relatively ìighter soil color, higher proportion of
artifacts, and ash lenses with hea¡thstones. The manure-stained pen
regions show up clearly, even on aerial photographs. T\e outermost
large stock pens are marked by piles of burned manure, up to fifty
centimeters high and three mete¡s across, swept up and fired every day
by the women. Inner pens are usually marked by deposits of food SHADE HUT
refuse and ash, formed as women habitually dump household debris
into or over the inner pen fences. Household food refuse consists
primarily of bones and other inedible parts of caprines; cattle are not LIVESTOCK
normally slaughtered for household consumption, being eaten away
from the settlement by older males. SLEE P ING
PENS
Stock camps during their occupation lack formal structures, other AREA ---
than pens and sometimes an openwork shade hut for newbom kids and
lambs. Herdboys do not normally construct shelters for themselves;
their actual living space is usually a sleeping area with a hearth,
centered under a shade tree, with the pens around it (Fig. 4.4).
Herdboys are given few items of any real wo¡th from family stores, and
their cultural debris is minimal, consisting mainly of animal bones.
Since the câmps are inhabited for shorter periods than âre the
manyattas, rnanure staining ofthe pen area is imperceptible. Once the 10m
fences of a stock camp have decayed, the site is hard to discern.
A third type o[ siie in Dassanetchland is the hunting-fishing camp.
Persons who have, through past misfortunes, lost thei¡ he¡ds are FIGURE 4.4, Plan ofa stock camp.
known as gal dies alternative rneanings of "poor people" and
-wilh
"fishermen"-and form a distinct class among the Dassanetch. If the
loss was suffe¡ed several generations back they tend to live in their own
settlements,Iocated in areas suitable for cultivation. Younger men and supplies. Thegal di¿s a¡e sometimes joined by younger sons of poorer
boys conduct fishing and hunting expeditions, traveling as fa¡ as 120 herding families and by men who have lost thei¡ he¡ds in the course of
kilometers down the lake in their dugout canoes. They lack firearms their lifetimes, but still reside at he¡ding settlements.
and are so poor in trade goods that even iron for spear tips is precious; Gal diøs camps lie along the lakeshore, varying in size from small,
until recently, they had no fishnets. Nonetheless, they manage to single-occupation "lunch stops" to large, repeatedly occupied sites
procure a variety of fish-including the huge Nile perch, terrapins, near good fishing grounds. Normal occupation deb¡is consists of bone
soft-shelled turtles, and crocodiles-with their spears and detachable- associated with hearths, plus occasional stone tools; low windbreaks of
head harpoons. They occasionally take larger land mammals and hip- stone or thatch sunshades a¡e sometimes constructed. A survey of a
popotamuses in cooperative hunts. If their catch is large enough, they ì9-by-1.5-km strip ofshore south ofKoobi Fora (see Fig .4.2) yielded ZZ
transport d¡ied meat back to their settlements; if not, the men are at such sites which, iudging from correlation with recent beach levels,
least able to live off the land without drawing upon the family grain span about fifteen years.

88 89
 DIANE P. GIFFORD N atural Processes Affecting Culturdl Mdteridls
It should be noted that the Dassanetch intensively curate nearly all (Frg. 4.2, Area 300). The oldest site was inhabired in Ì957_5g,
items of their mate¡ial culture, due to the relative difficulty of the
newest in-1973. Gal diøs camps were surveyed along a l9_km strip of
replacement, \¡/ith a resulting ìow rate of discard. This leaves ìittle shoreline between Koobi Fo¡a ancì,{llia Bay (Fig. 4.2, Area 000). Dates
artifactual material for potential presewation in the archaeological of occupation for the 19 sites in the zone ranged from 1960 ihrough
record. Government restriction of Dassanetch contact with other 1973. In addition to the intensive studies in these two areas, brief
tribes has severely limited trade, and even wealthy persons cannot observations were made on site preseúation in two other areas (Fig.
readily obtain or replace trade goods. Gourds, pottery and meial 4.7, Areas I00, 200).
containers, metal implements, beads and other omaments, and cloth If a site remains exposed to the elements for l5 years, conside¡able
all must be acquired through a tenuous set of trade connections into alteration of perishable elements takes place. On manyattas and stock
Ethiopia, at vastly inflated prices. Such items a¡e used and mended camps, thorn-brush fences become relatively fragile after 5 years, and
until completely beyond repair. Among the few locally manufactured have disappeared completely after 15. Manyatta pen areas continue
household items are wooden milk containers, made by women; even to
be set off by manure-stained soil, and piles of burned dung in outer
these are intensively curated, since trees of sufficient diamete¡ for use pens, on which plants cannot grow,
are easily distinguishabie afte¡ 15
are rare in this arid region. Animal skins are recycled through various years. The inner pens continue
to be set off by small scatiers of bone
functions as shelter components, shoes, or clothing. Observations of and ash refuse.
abandoned manyattas indicate a relatively low proportion of artifact Bone specimens f¡om the surface zone of the oldest site (1957_5g)
discards to food refuse and a relatively high proportion ofprobable loss were deeply invaded by weathering
lines and we¡e extremely fragile but
items (isolated beads, fishhooks, a tin crucifix, arid so forth) to actual still identifiable. Subsurface
bones were less heavily .eaihered,
artifact discards. These sites contrast sharply with Maasai pastoralist although leached of organic content and fragile.
It appears from these
settlements about seventy kilometers from Nairobi, where proportions observations that bones of medium-sized
mammals, in this case
of artifactual refuse are about equal to those of food refuse on caprines, can ¡emain on the surface in the East Lake Turkana region
abandoned sites. for at least 20 years before toially disintegrating.
It is probable that a ma jor shift in refuse pattems occurs when stone In the course of 20 years, manyattâs, stock camps, and gal dies sites
implement technology with a high overaìì disca¡d rate is replaced by lose nearly all perishable materials if exposed to
the elements. Once
metal implement technology with a very low discard rate. While the weathering has destroyed the bone component on
such sites, very little
timing of this shift in the East Lake Turkana area is not yet known, remains, since the low artifact-discard rates produce little in the
way of
there are in the region large surface scatters of lithic mate¡ials with imperishable materials. I saw a number of manyattas representing
this
some preserved bone which may represent pastoralist sites of this decay stage, on which all that remained were soil stains, hearthstones,
former phase. and scatte¡ed ash lenses, with ra¡e artifacts. Burial relatively soon after
deposition will protect cultu¡al materials from further decomposition
and, to some extent, disarrangement. But if the geological processes
P ostoc cu þation P ro cøss involved in burying a site a¡e sufficiently intense, they can alter the
distribution of materials or even completely destroy the occur¡ence. In
Through informant inte¡views I was able to arrange abandoned sites order for a site to be preserved as originally constituted, a fine balance
in two age-ranked series, establishing a site decay sequence for all three must be struck between geological and other postoccupation
types of occupation. The series, which spanned l5 years, permitted me processes.
to isolate various postoccupation processes affecting cultural The range and intensiþ of geological processes to which a site may
mate¡ials. One area intensively studied was a 0.5-by-0.3-km area in the be subject depend upon its geomorphic setting. The Dassanetch are
ll-Erriet Valley, wherein lay I5 abandoned manyattas and stock camps well aware of the ¡elationship between setting and geological process

90 9I
 DIANE P, GIFFORD N atural Processes Affecting CuttutuI Mdteridls

and accordingly locaie their semipermanent settlements in areas of
minimal disturbance. Stock camps and manyattas are deliberately sited
away from areas liâble to flooding. Owing to this understandable
human preference, abandoned sites are subject to little if any
deposiiion. The main geological processes acting upon them are
eolian, with wind deflation ouþacing deposition. Deflation exposes
formerly subsurface items to surface weathering processes, thereby
accelerating site destruction processes.
Nearly all the manyattas and stock camps I observed were located in
areas of minimal geologic activity. One short-term mânyâttâ \Ã'âs sited
in an environment potentially favorable to aqueous deposition, a large,
low island in the main r¡/atercourse of the Il-Erriet. lt was, however,
completely destroyed during severe flood inlApril 1974. During one
flood, a c¡est of wate¡ about four meters high swept over the island,
carrying off nearly all cultural materials. Test scrapes of areas formerly
rich in bone refuse yielded next to nothing. The two extremes in
depositional process observed on Dassanetch sites in the Il-Er¡iet Val-
ley would have much the same outcomes in terms of archaeological
preservation. Whethe¡ all the organic materials so valuable in beha-
vioral and economic ¡econstructions disappear over a 30-year period or
within 24 hours, the resultant site impoverishment is the same.
Although ìying in a different so¡t ofgeomorphic context,lhe gal díes
hunting-fishing camps I observed were likewise not in locations
favorable to speedy deposition. In the case of lake-margin sites, the
situation was not so much the ¡esult of human intention to locate sites
away from sedimentary processes âs of the continuing regression of the
lake (Butzer 1971). The important point to note here is that had the
lake instead been in a transgressive pËase, many of these sites would be
covered with lacustrine sediments. It is important to recognize the role
that duration of occupation may play in determining the placement
and, in terms of geologic process, the sedimentary environment of a
site. Hunting-fishing camps are occupied for only a few days at a
stretch, are easily occupied, and are iust as easily vacated. There is no
premium in this case on selective avoidance of certain sedimentary
environments, as in the case of pastoralist settlement patterns. Since a
number of gal dies sites are located in environments with good
poteniial for sedimentation, these short-term Dassanetch occupations
have a relatively higher probability ofpreservation than do longer-term
ones. In addition, sedimentary processes in the lake-margin zone,
whether lacustrine or deltaic, are considerabìy gentler
than those in
the major river cou¡ses âlong whlch the pasto¡alist siro. l;- r,r
two sites subiect t","y ki"d d;;;;jù;;ilä;.jä::ffi ;ff .:iï
of obse¡v¿tion were short-term "?
g¿l dí¿s sites, on. lo"rì"ã'"on'ir,*"
margin, the other in a delta. r"t.
Site 02 a hippo_butchery camp created
-Because of-is_the problems of managing , hippo in September ä"'ll_
1972.
situated close to the kill site, jrrt *.ti't"r-töil"ïii: ffj
"itUãand
""rc"r.
shore. The posterior axial skeleton limbs of rh" ;;;;i *"r.
transported to a spot about sixty meters inland from the
fr-r-ofi, n."¿
and ce¡vical ve¡tebrae. Two rock windbreaks ,"a ,-
riJí""_år"f.a
hearth were const¡ucted in the inlandzone._The lake 1.".i"",
.iri"g
the time of the kill, and within a week the head and "t
Uonï, *"r"
standing in shallow water, while the rest of the animalt"""t
U"àv irv å" Ory
ground. The lake level continued to ¡ise in OctoUe,
lSä u'r,iit tÌ,"
outer scattet was totally covered by gently lapping water.
The lake level
fell markedly in December 1974, exposing tt e h."d ,n¿
.,""î-1o.,",
again. About three ðentimete¡s of fine silt had built up in
the bones, which were in relatively better condilor, "nã "ro,_r.ra
ìiì"-',fr"i.
subaerial counterparts. Had the lake level stayed high, those
bones
would probably have been preserved long aft..ìhe ruú"e.i"fon",
¡",1
disintegrated. The remaining site would thus probably consist
àf the
head and neck of a hippo, plus two stone semicircles, and
a rock_ringed
heartl area about fifty meters northeast of the bone. Th" ,,r*iui.rg
bones mightprovide a clue to the actual riatu¡e ofthe site.
Site 20 is another campsite, created over four days in November
1973 by eight men and boys. They made their camp in the shallow,
braided dist¡ibutary system ofthe delta of a small ephãmeral river. The
camp was situated about th¡ee hundred meters f¡om the shore, in the
conjunction of the main stream channel and h.vo side channels. The
sandy substrate of the channel was much mo¡e suitable for occupation
than was the su¡rounding landscape, which is largely covered with the
spikeleaved local grass. The camp consisted of a hearth, with three
stones to support a ke¡osine tin for cooking; a scatter of ash dug out of
the hearth; a woodpile; and a sleeping area with a lakeweed ,,mattress."
During their stay, the men procured and ate 40 small terrapins, 4
small-to-medium-size crocodiles, about l4 catfish, and a few small Nile
perch and other fish. In addition, they were able to scavenge meat
from two lion-killed zebras. The meat of one they cut off the carcass

9Z
9)
 DIANE P. GIFFORD

and caûied back to camp in their tin; from the other they took ân
entire hind leg. The zebra meat was laid out to dry about twelve meters
up the larger channel from the camp, and ihe bones of the leg were left
there. I mapped the resulting scatter a day after they left the camp
(Fig. a.5).
In the spring of 1974, unusually heavy rains fell in the East Tu¡kana
area. The region usually receives about 300 mm per year, but in two
weeks that spring ii got well over 70 mm (Butzer l97l; East Rudolf
Fisheries Research Project: personal communication). I was in the -!¿u¡
area at the time and observed the stages of the burial of Site 20. The
rç¡::
ãë--å
fi¡st rains falling in the watershed were light. The runoff dispersing äËË¡Éi
across the delta was gentle and confined largely to the main channel, .4. ;¡ ¡ E à
with slight late¡al invasion of the overbank area. Bones in the rnain
channel, including the zebra leg, were swept downstream as much as
25 m. Materials only a few centimeters upslope from the channel in 2
the overbank region were not displaced and were covered in place by
a
2-5 cm of sand and silt. The silts and sands of the first rain dried to a
E
hard cap over and around these mate¡ials prior to a second, heavier .2
¡ain fou¡ days later. Massive runoff from this episode flowed over the
entire site, adding another 5-10-cm deposit to the overbank region,
o
completely covering all materials.
In the summer oÍ 1974 I excavated the core a¡ea of the site to
ascertain the degree to which depositional processes had affected the o
original distributions. The effecis oF flooding differed in the main
channel zone and the overbank zone. Bones in the main channel had
I
been swept away; a trench eight meters farther down the channel f¡l
course yielded only a few shell fragments. Distributions in the ð Þ
overbank zone, on the other hand, were preserved nearly intact; even 9 o
the ash layer remained in place (Fig. 4.6).
Dr. A. K. Behrensmeyer (University of California, Santa Cruz),
whose geologic and taphonomic work at East Lake Turkana has
received much attention, was very interested in the site and undertook
to study its microstratigraphy. She was able to discern in the site's
section the two depositional events I witnessed in the spring of 1974.
We have subsequently analyzed sediment samples from the
depositional phases and facies and also conducted flume and settling
velocity experiments with bones ¡ecovered from the site. The results
have allowed us to estimate the rates of wate¡ flow that selectively
preserved and destroyed portions of the site.

94
 N atur¿l Processes Affecting Culturul Materials

As the lárgest sedimentary particles on the site, bones

in
were useful
esiimating flow velocities. Two experimental approaches to bone
transport were utilized. First, bones were sublectJ to calib¡ated
flow
conditions in a water flume belonging to the Department
of Hydraulic
Engineering, University of California, BerkeÈy. With a subst¡ate
approximating that of the site, we dete¡mined that flow velocities
of
about 40 centimeters per second were necessâry to put bones
of all
sizes into continuous movement, as was the case in the channel.
Ove¡bank flow conditions probably did not exceed l0 cm/sec, the point
at which all small elements are in motion, for any appreciable amount
of time. Second, the same bones and othe¡s eithe¡ tàã large or small to
be placed in the flume were observed settling in standin-g water. The
setding velocity of a bone is verj' heavily influenced by its shape, as

.ffi well as by its weight, and two bones of disparate weijhts may have
similar settling velocities. From an element,s settling velocity one can

,ru\
calculaie ihe diameter of a round quartz grain with equivalent settling
a velocity. This nominal quartz diameter equivalence is useful because it
g allows one to consult the sedimentological lite¡ature on transport of
! regularly shaped materials (1. R. L. Allen 1963; Compton 1962).
Nominal quartz diameter equivalences (dq) from fossil and nonfossil
o bones may be calculaled in the followìng ways (Behrensmeyer
éi 1975:492-9)):
Nonfossil Elements:
a Fossil Elements:
ü , 2. (vJ'
(1-=-
f¡ do=(P5- 1) .db/1.65
Ì ú
Þ
" 413' s. (pq- I)
d¡ : Nominal diameter of the
sirnplified to:
Lr.t
q, dq:.000928.v,, bone:\y'ï3TJlVõiìñ
pb = Bone density
o vs = bone settling velocity
l!
t- g : 980 cm/sec2
pq : quartz densi{ : 2.65

tr,l Table 4.I shows examples of dO and flume data from Site Z0 speci-
mens. A detailed report on sedimentary processes and effects at Site Z0
is currently being prepared by Behrensmeyer and myself.
Generalizing from these observations, I would conclude that Das-
sanetch site location preferences, based upon knowledge of local
geological processes, determine that most pastoralist sites will lie in
zones unfavorable to good preservation of cultu¡al materials. These
sites reflect the settlement patterns and economic base of about 85

o?
 DIANE P. GIFFORD N atutal Processes Affecting Culturdl Mdteridls
percent ofthe Dassanetch people. Sites most liable to be presewed are TABLE 4.1
short-te¡m gal dies camps, located in zones avoided by the more S]TE 20. EXPERIMENTAL BONE D.{TA
sedentary pastoralish; the Iikelihood of aqueous inundation of the sort F lume Exþ¿ñment

that would preserye â site is the very reâson given for this avoidance. Element weisht(Ð v sþnñhec) dq(m'¡.) TtdnsÞott V elocíties (cmls¿c).
Inìtidl Contin¿ous
Hunting-fishing camps reflect the protein procurement activities of a
small segment of the total regional population and are not typical of Terrapin
câ€pece '14.7 161.0 Z.4t 25
the total range ofactivities ofeven thât group.
Terrapin
Relative degree of sedentism may be extremely influential in caIapace
determining the likelihood of site preservation. A hypothesis fragment 14.3 177.9 2.94 l0
concerniug these relationships may be framed as follows: Given a Terrapin
range ofgeomorphic zones in an inhabited region, increasing duration plastron

of site occupation will result in placement of long-term sites in frågment 3.8 154.' 2.2r t0 z0 J5
Terrapin
geomorphic zones of minimal depositional activity. ¡adius 1.0 282.5 7-4t 8 z5 35
C¡ocodile
half-
ARCHAEOLOGICAL VISIBILITY mandible t62.] 500.0 27.13 15 15 50
In discussing the effects of postoccupation processes upon sites, the C¡ocodile
concept of archaeological visibility is potentially useful, but it hâs not
hume¡us 44.6 378.0 t3.25 Z5 35 50
Crocodile
thus far been adequately defined. Here I use the term as an abst¡act radius 9.1 272.3 6-87 no no 40
concept that reìates ethnographic reality on one time level to its move-
probable material evidences on some futu¡e time level. Drawing in
part upon Cowgill's (1970) discussion of archaeological sampling * Water depth: 15.2 cm (0.5ft);
subshate well-sorted coarse sand (median dq = L 3mm).
problems, I propose that a spatial focus of some human activity must
meet three conditions to be archaeologicaìly visible: (1) the human chaeological visibility have no necessâry connection with a particular
activities must hâve material consequences; (2) the material conse- type of subsistence pattern, but rather with ihe interplay of cultural
quences must be potentially presewable; and (3) natural processes and natu¡al factors. Certainly a group of people who habitually erect
must act upon these material consequences in such a manner as to du¡able structures on their occupation sites may be more archaeologi-
presewe them. If an occur¡ence fulfills all three conditions, it will be cally visible than those who do not, and celtainly this type of cultural
archaeologically visible. behavio¡ is rare among hunter-gatherers. But it is likewise rare among
Preservation of cultural materials in much their original condition shifting agriculturalists and pastoralists. For the reasons given earlier,
and disposition (fulfillment of the third condition) is often the result of some transhumant or seminomadic pastoralists may be even more
a complex interaction of cultural and natural p¡ocesses. As indicated "invisible" than hunter-gathere¡s. In such groups low discard ¡ates
in the discussion of trampling, the substrate may inte¡act with cultural (first coirdition) combine with poor pleservation conditions (third con-
processes in determining whether or not primary refuse is presewed on dition) to limit archaeologicál vkibility. Although the research situa- i
Ì
a site. Cultural factors may likewise determine whether or not geologi- tion is changing (see Hole: Chap. 6, below), pastoralists' archaeologicaì ì
cal processes capable of presewing a site can act upon it. invisibility may well account for the lack of serious attention given the
Deetz (1968b) discusses the notion of a "th¡eshold" of archaeological role of that subsistence pattern in the origins of food production in the
¡
visibility, which, he has proposed, most hunter-gatherers do not sur- Near East. !
I

pass. According to the definition given here, the prerequisites of ar- Regional surveys of archaeological sites have been the basis of re-

98 99
 DIANE P. GIFFORD Natural Prccesses Affecting Cutturdl Mdteúdls
cently proposed models of culture chânge in both the Old and
New to_ assess how truly representative these samples a¡e of the total rarrge
worlds. These studies all assume that the sites observed are
the result of prehistoric phenomena.
of regularities raiher than vagaries in human cultural b.h"uio.. If archaeologists wish to derive detailed inferences regarding
Another assumption implicit in such studies is that the obse¡ved prehistoric cultu¡al behavior from patterns in the ar.chaeãlogical
sites
are a random sample of the original range of cultural
occu¡re¡rces. evidence, they cannot afford to ignore the påtterning effects of
But, as stressed earlier, sites of certain types may not be randomly noncultural processes upon such materials. Ethnoarchaeological
distributed as regards geomorphic zones in a regior,, and in
fact certain resea¡ch should encompass all processes, cultural and natural,
geomorphic zones of high preservation potential may
be intentionally affecting the material t¡aces ofhuman behavior.
avoided in location of ìong-term occupation sites. The interaction
of
cultu¡al and natural processes will in such cases lead to a skewing in
the preserved archaeological sample. In the Dassanetch situation, NOTES
for
example, pastoralist sites will generally be poorly preserved
surface 1. EquipmeDt and labor expenses for my work in Kenya were provided by
scatters, which, unless the culturàl and natural processes involved in ^ National
Scie¡ce Foundation Dissertation Resea¡ch Grant CS-40206. Living expenses were paid
site preservation are understood, may be avoided by researchers by the fìnal year.of.an N.D.E.À. (Title IV) Graduate f"ffo"rlifl il.-ú"ìr,ã._C..r,
look_
ing for "good" sites. Foundation for AnthropologìcaÌ Research, in remuneration for my
ãi f"un"l
Only an understanding of the ancient geomorphic processes operat_ "rr"ly.is
mâterial from the site ofLongs Drìft, Kenya, during my time in Naiiobi, páid my ai.fare
to Kenye.
ing in a region combined with ethnoa¡chaeological settlement_pattern I wish to tha¡k the Ofñce of the president of the Republic ofKenya for permission
to
data can indicate potential skewing ofa regionai site sample. conducl ¡eseelch and the Trustees of the National Museums åf Kenya for their
Construc_
tion of a comprehensive regional model of ancient geomorphic envi- sponsorship and aid. Pe¡sonal thank ti.. Glyrrn L.
are given to my acailemic advisor,
kaac, University of Califomia, Berkeley, and to lri¡. R. E. Le"k"y, Ad-inistrrtive
ronments and processes seems a basic requirement in any regional Di¡ector ofthe National Museums, for their invaluable advice and material aid.
archaeological survey. My fielcl
assistanb, MÌ. fack Kilonzo and M¡. Andrea Kilonzo, receive my thanks for
their
patience and hard work.
My deepest gratitude, ofcourse, goes to the people oflleret.
CONCLUSIONS 2. Olson (1962) defines taphonÃy as the;tuåy of the låws of burial, including all
âspects of the trensference of organic remains from the biosphere to the Ìitiosphere.
In assessing the effects ofnoncultu¡al processes upon archaeological
sites, ethnoarchaeological research can do more than simply provide
cautionary notes to archaeoìogists. This paper has presented several
testable-propositions concerning the operation of such processes upon
cultural materials. It has further stipulated probable diagnostic
features of one type of affected sample. Further ."r""rãh *ill
undoubtedly yield many more such propositions.
Shrdies of natural alteration processes in the present, of course,
cannot conclusively prove the operation of such factors in any given
prehistoric sample. They can, however, be used predictively to
improve interpretation of the archaeological record. Às Schiffer and
Rathje (1973) have stated in their discussion of ..n-transforms,,' a
systematic knowledge of natural alteration ptocesses can be helpful in
locating abundant and weìl-preserved archaeological materiali. The
same knowledge that allows one to locate such samples can perírit one

100

View publication stats

l0l