Growth Increment and Stable Isotope Analysis of Marine

Bivalves: Implications for the Geoarchaeological

Record of El Nino

Harold B. Rollins
Department of Geology and Planetary Science, University of Pittsburgh, Pittsburgh, P A 15260
Daniel H. Sandweiss
Department of Anthropology, Cornell University, Zthaca, NY 14853
Uwe Brand
Department of Geology, Brock University, St. Catharines, Ontario, Canada
Judith C . Rollins
329 Bellwalt Dr., Bridgeville, PA 1501 7

The 1982 - 1983 El Nino event afforded the opportunity to develop criteria for the recognition of ancient El
Ninos using mollusks from archaeological sites along coastal South America. A combination of growth
increment and stable isotope analyses indicated that elevated sea surface temperatures during large scale
El Ninos leave a record decodable from the growth patterns of selected bivalve shells. The intertidal
venerid Chione subrugosa displayed a pronounced break in the valve margin profile following the 1982-
1983 event but provided an inconsistent stable isotope pattern. The subtidal carditid Trachycardium
procerum, on the other hand, preserved a discernible and diagnostic growth interruption as well as an
expected trend in stable isotope indicators of salinity and temperature change.
We conclude that some of the major culturally disruptive El Nino events can be recognized in the
geoarchaeological record by these techniques, especially if ancillary information, such as faunal distribu-
tion patterns, are also considered. Perhaps the most serious constraint upon application of this approach
involves microstratigraphic resolution of shell midden deposits. Stratigraphic sampling of midden mate-
rial should be accompanied, if possible, by sampling of proximal natural strata. The chances of discovery of
major El Nino perturbations in the geoarchaeological record of shell middens is enhanced by the cata-
strophic nature of such events and by the indication that major El Ninos have a high probability of being
closely spaced in time.

INTRODUCTION turbations, utilizing the widespread and

abundant shell midden deposits along coastal
The predominantly maritime-based econo- Peru. This paper discusses the application of
mies of coastal Peru have existed for thou- growth increment and stable isotope analyses
sands of years (Moseley, 1975; Sandweiss et of bivalve shells as a tool for recognition of
al., 1983; Richardson, 1981). For at least the ancient El Nino events.
last 5000 years El Nino must have been, as it
is now, a major influence upon coastal occupa-
tion of that area (Rollins et al., 1986a). The
recognition of ancient large-scale El Nino THE 1982-1983 EL NINO
events would certainly aid our understanding
of the ecological vagaries that must have, in The 1982-1983 El Niiio event was proba-
part, affected the growth and development of bly the most devastating Pacific Ocean cli-
these coastal economies. matic perturbation of this century (Arntz,
The 1982- 1983 El Nino event has afforded 1986). Historical records of El Nino events
us the opportunity to develop criteria for the extend back to the 1700s and, using any avail-
recognition of ancient large-scale El Niiio per- able index, the 1982-1983 El Nifio ranks as

Geoarchaeology: An International Journal, Vol. 2, No. 3, 181-197 (1987)

01987 by John Wiley & Sons, Inc. CCC 0883-6353/87/030181-17$04.00
THE RECORD OF EL NlNO

one of the worst (Cane, 1983; Quinn et al., rapid, ostensibly reflecting a behavorial adap-
1978). The event was atypical in many re- tation triggered by temperature change,
spects. Thermal anomalies (as measured by when there is actually a complex linkage
SST-sea surface temperature) were extraor- among nutrient supply, light intensity, tem-
dinarily high and the effects of the event did perature, productivity, and food (Barber and
not hit coastal South America until late sum- Chavez, 1983). The stresses created by
mer, when local SST would normally be de- reduced food availability are apparently am-
clining to its lowest value. Caviedes (1984) plified up the food web, leading to decreased
presented a detailed chronology of the 1982- growth rates and reproductive failure most
1983 events that affected the coast of South noticeable at higher trophic levels. These cat-
America. In June of 1982, the coast of central astrophic ecological disturbances are often
Chile received high rainfall with repeated unappreciated over short spans of time due to
passage of South Pacific winter depressions. the spatial displacement of many species fol-
By October, 1982, excessive rainfall occurred lowing the onset of El Niiio. Local “catches” of
in coastal Ecuador and southern Colombia some fish and shellfish may actually increase
but did not affect the region south of about due to movement by currents of displaced pop-
latitude 2 degrees south. However, September ulations into shallower, more accessible
and October were times of dramatic rise in coastal regions or due to the temporary intro-
SST along the South American coast. The tor- duction of thermally anomalous “exotic” spe-
rential rainfall moved southward during De- cies. Arntz (1986) noted, for example, that the
cember, following the abnormal southern dis- 1982- 1983 El Niiio caused rapid growth and
placement of the intertropical convergence high densities of scallops and some other in-
zone (ITCZ).Premature and excessive rainfall vertebrate species. On the other hand, inter-
reached northern Peru by January, causing tidal and shallow subtidal mussel banks and
extensive flooding and landsliding. While kelp forests were destroyed by the event. Hu-
there was rainfall in the desert of northwest man inhabitants of these coastal areas who
Peru, drought struck the agricultural regions are dependent upon gathering of nearshore
of southern Peru (Rasmussen and Wallace, invertebrates for food were hardest hit by the
1983). Violent storms continued along the 1982- 1983 catastrophe. This most likely has
coasts of Ecuador and norther Peru until mid- been true in this region for every large-scale
June. Some degree of seasonal atmospheric El Niiio event over thousands of years.
normality was restored by the end of June, The economically important bivalve species
1983 (Caviedes, 1984). Although the ITCZ of coastal Peru are, for the most part, sessile
had passed beyond northern Peru, air temper- benthic and thus incapable of short-term mi-
atures were still higher than usual and the gration with changes of water currents and
SST of coastal waters north of Callao, Peru temperature. Over the long term, however,
(about latitude 12 degrees south; Figure 1) such species exhibit ecological resilience due
was as much as 8°C above normal. The Paita to their high reproductive recruitment poten-
coastal station (about latitude 5 degrees tial. Although scattered anecdotal informa-
south) showed signs of recovery of SST nor- tion is available concerning the effects of El
mality by July, 1983 when the temperature Niiio upon shellfish harvests, little research
decreased to 20°C (Barber and Chavez, 1983). effort has been directed toward understand-
Ecological disruption following the onset of ing physiological and morphological re-
the 1982- 1983 El Niiio was dramatic but not sponses of various shellfish species to the eco-
at all unexpected. Modification of the normal logical stresses of El Niiio. Sessile bivalve spe-
upwelling of cold nutrient-rich water adjacent cies are sensitive indicators of changing
to coastal South America inevitably affects coastal conditions and record, on a daily
populations of phytoplankton, fish, and sea- and/or subdaily basis, a chronology of envi-
birds. The response of larger marine pelagic ronmental perturbations. This record is mani-
organisms to SST anomalies can be very fested in the pattern of shell growth incre-

182 VOL. 2, NO.3

 THE RECORD OF EL NlNO

ments, types of shell microstructure, and cerum were selected for detailed growth incre-
profile of the value anterior margin. ment analysis. Description of the techniques
of bivalve growth increment analysis are
readily available from the geoarchaeological
METHODS
literature and will only be briefly described
The bivalve specimens utilized in this study here (Deith, 1983; Koike, 1980; Rollins et al.,
were harvested alive or purchased from ven- in press). Individual valves were cut with a
dors at several sites along coastal Peru, dur- Felker rock saw along the axis of maximum
ing a series of collecting trips in 1984. Collect- growth (normal to surface growth lines).
ing stations ranged from Tumbes, in extreme Fragile valves were embedded in Epon 815
northwestern Peru, to the Paracas peninsula, epoxy resin prior to cutting (Kennish et al.,
about latitude 15 degrees south (Rollinset al., 1980). Cut valves were ground and polished
1986b; Figure 1). using a Buehler Isomet lapidary machine and
Following preliminary screening of several then etched in 5% HCl for approximately one
harvested bivalve species, 35 specimens of the minute. Etching time varied according to
intertidal Chione subrugosa and 15 speci- species and condition of valves. The etched
mens of the subtitdal Trachycardium pro- valves were carefully washed, air-dried, im-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 183

THE RECORD OF EL NINO

mersed in acetone and quickly placed upon organic matter content of the shells. Accuracy
pieces of sheet acetate (0.1 mm thick). After of chemical values was determined by analyz-
drying, the acetate peel replicas were ing N.B.S. standard rocks (633,634,636) and
mounted between glass microscope slides. Ac- precision of data was calculated using dupli-
etate peel replicas were studied and photo- cate analyses.
graphed using a MPV-2 microscope and auto- The shell samples were also analyzed for
matic camera. Growth increment counts were oxygen and carbon stable isotopes. Powders
made upon assembled photo mosaics. were reacted with 100% phosphoric acid at
Stable isotope analysis of bivalve shells is 50°C for 10 minutes, and evolved gas was
also widely used in geoarchaeological studies, analyzed on a V.G. 903 Micromass spectrome-
mainly as a technique for ascertaining season ter. The isotope ratios are expressed in 8 rela-
of harvesting (Shackleton, 1983; Bailey et al., tive to PDB and corrected for oxygen-17.
1983;Jones et al., 1984;Deith, 1986;reviewed Chemical data were evaluated using the
by Rollins e t al., in press). Killingley and statistical packages (e.g., SPSS) on a Bur-
Berger (1979) were able to detect upwelling roughs B7900 computer, and with the Stat-
events by stable isotope analysis of California view program on a n Apple Macintosh com-
mussel shells. We know of no other study, puter. Graphs were computer-generated.
however, which has attempted to use stable A total of 14 samples from three specimens
isotopes of molluscan shells as indicators of El of Chione subrugosa and 19 samples from four
Niiio sea surface temperature perturbations. specimens of Trachycardium procerum were
Prior to stable isotope analysis all bivalve analyzed for Ca, Mg, Sr, Mn, Na, Fe, Zn, Cu,
shells were initially cleaned of adhering or- and Ni. Also, a representative suite of 15Sam-
ganic and inorganic material. Then the shells ples was analyzed for 8l80 and 613C.
were cut along the axis of maximum growth in
the manner described above and subjected to
further cleaning by immersion in an ultra- RESULTS
sonic bath of de-ionized water (to minimize Growth Increment Analysis
contamination) for 15 minutes. Additional
cleaning consisted of brief immersion in 15% Chione subrugosa and Trachycardium pro-
HC1, rinsing with de-ionized water, and sub- cerum display different patterns of growth.
sequent air-drying (Brand and Veizer, 1980). The former has a reflected mantle margin, as
Shell samples were taken (by drilling and do other venerid bivalves, and easily counted
sawing) at specific intervals along the axis of arcuate growth increments are deposited
maximum growth in order to traverse the (Figure 2). Trachycardium procerum pos-
growth before, during, and after the 1982- sesses a nonreflected mantle margin and
1983 El Nifio event (as indicated by examina- growth increments intersect the shell margin
tion of valve margin profile and growth incre- at high angles (Figure 3). In Chione subru-
ment pattern). These samples were powdered gosa daily growth increments were prominent
and dissolved in 5% HC1, in order to minimize while semi-daily growth increments were
the leaching of elements from material that most distinct in Trachycardium procerum.
was not removed during the cleaning process. Detailed analysis of the daily and semi-daily
All samples were analyzed for Ca, Mg, Sr, growth increments of bivalve shells has
Na, Mn, Fe, Al, Ni, Cu, and Zn using an auto- proven to be a useful indicator of short and
mated and computerized Varian 1475 atomic long term environmental stress, intensity of
absorption spectrophotometer (see Brand and seasonal zonation, and habitat relative to
Veizer, 1980;Brand and Hinsperger, 1986 for onshore/offshore depth gradients (Richardson
complete details of sample preparation and et al., 1980; 1981; Barker, 1964; Jones, 1983;
analysis). Trace element compositions were Rhoads and Lutz, 1980; Hall et al., 1974; Ev-
calculated on a 100% carbonate free basis to ans, 1975; Clark, 1975; Kennish and Olsson,
minimize trends due to fluctuations in 1975; Thompson, 1975; Pannella and Mac-

184 VOL. 2, NO. 3

 THE RECORD OF EL NINO

Figure 2. A-C. Chione subrugosa (photomicrographsof acetate peels). A. Specimen

# Ptl-10, collected 4-16-84, Tumbes, Peru. Note distinct daily lines and subtle,
discontinuous sub-daily lines. Crossed-lamellar microstructure (XI is in typical un-
stressed pre El-Nino orientation. Growth margin of the shell is to the left. B. Same
specimen. Area immediately below El Nino break. Note transgressive crossed-
lamellar microstructure (X) invading upper composite prismatic layer. C. Specimen #
FPT-1, collected from shell midden, Pampas las Salinas, Peru. Narrow break in valve
profile is typical of a storm break, not a major El Nino disturbance. Recovery was rapid
following the storm break. Growth margin of the shell is to the right. D-E.
Trachycardium procerum. Arrows indicate onset of El Nino break. D. Specimen #
TP1-6, collected 4-18-84, Chiclayo, Peru. E. Specimen # 2TP2-10, collected 12-1-84,
Trujillo, Peru.

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 185

THE RECORD OF EL NlNO

Figure 3. Truchycurdium procerum (Speciman # TP3-11, collected 4-20-84,

Chimbote, Peru, A-B. Acetate peel, negative print (scale = 1 mm). Note broad
shallow El Nino break in valve profile and increase in crossed-lamellar microstruc-
ture at the valve margin. Anterior tip is to the right in all of the photographs.
Semi-daily growth increments are shallowly inclined below the El Nino break. Ar-
rows indicate 307 days (614 increments). c. Photomicrograph (scale = 0.1 mm).
Enlargement of area C in Figure 2A. Note massive crossed-lamellar microstructure
and nearly horizontal growth increments. D. Photomicrograph (scale = 0.1 mm) of
area D in Figure 2A. Cross-lamellar microstructure still present, but growth incre-
ments have resumed normal inclination to the valve margin.

Clintock, 1968; Ekaratne and Crisp, 1982). colored bands, the result of calcium carbonate
There is general agreement that bivalve deposition when valves are open and the for-
shells are gauges of a hierarchy of lunar and mation of a thin dark line, presumably due to
solar rhythms, as well as extrinsic environ- anaerobic dissolution of shell material, when
mental events. An individual growth incre- valves are closed (Lutz and Rhoads, 1977).
ment consists of a couplet of light and dark Intertidal bivalves (e.g., Chione subrugosa)

186 VOL. 2, NO.3

 THE RECORD OF EL NlNO

therefore have incremental patterns that re- stresses may also result in pronounced growth
flect both lunar tidal (exposure and immer- breaks in bivalve shells. Such stresses may
sion) and solar day rhythms where valve clo- accompany storms, temperature shocks (hot
sure accompanies exposure and increased or cold), or spawning. Breaks due to hot and
input of heat (light). Some controversy exists cold shock may be quite similar, exhibiting
over the temporal significance of single sudden bunching of growth increments at the
growth increments in bivalve shells. Richard- onset, with slow recovery. Storm breaks, how-
son et al. (1980,1981)have made a strong case ever, typically display rapid onset and recov-
for tidal exposure and temperature control- ery, and may result in the depositional incor-
ling the expression of growth increments. poration of sediment grains at the base of a
Prominent growth lines are the result of tidal narrow notch in the shell margin. Excellent
exposures coinciding with elevated tempera- discussions of the types of growth patterns
tures, as when low tide occurs during daylight and breaks are provided by Kennish and
hours. Alternating strong and weak lines oc- Olsson (1975) and Cunliffe (1974). Kennish
cur when one low tide is in early morning and and Olsson (1975) determined that the major
the other is in mid-afternoon (Richardson et controlling factors affecting the growth of the
al., 1981).We suspect that such factors lead to venerid clam Mercenaria mercenaria were
a pronounced difference in strength of growth temperature, tides, substrate type, water
lines so that bivalves which are totally ex- depth, and age ofthe individual. Environmen-
posed a t low tide will appear to form one line tal stress also may be indicated by irregulari-
per day whenever one low tide coincides with ties in the shell microstructure and in the
daylight hours. Counting of growth lines is profile of the valve margin. Kennish and
difficult and the tendency is to observe and Olsson (1975) noted that thermal shock
record only the more prominent lines. Pre- breaks in venerids may cause a transgression
sumably, growth line counts on high inter- of massive crossed lamellar microstructure
tidal bivalves will tend to emphasize daily into the outer prismatic or composite pris-
periodicities whereas counts on lower inter- matic microstructure. Wide notches may oc-
tidal and subtidal bivalves stress semi-daily cur on the valve margins as a result of the
intervals. This may explain why daily lines temporary withdrawal of the mantle during
have been commonly recognized in venerids prolonged stress. They are quite distinct from
such as Mercenaria mercenaria, Meretrix the narrow storm-induced notches mentioned
Zusoria, and Chione spp., which inhabit inter- above and may result from long-term climatic
tidal mud flats or lagoons (Koike, 1973; perturbations such as El Nifio events. Under
Pannella and MacClintock, 1968). The sub- such conditions the ventral margin of the
tidal carditids, in contrast, tend to display valve may be blunted with a precipitous slope
nearly equal development of semi-daily lines, (Rollins et al., 198613).
corresponding to each low tide. Our analyses Conspicuous El Niiio-induced growth
of Trachycardium procerum and Chione sub- breaks were present in all the specimens of
rugosa support this interpretation. Chione subrugosa collected in 1984 from
As noted by many workers (e.g., Thompson, Tumbes (Figure 4). The growth breaks dis-
19751, shell increments may also be grouped played rapid onset and slower, often erratic,
into alternating close and wide-spaced bun- recovery. Most specimens did not achieve to-
dles that correlate with semi-monthly neap tal recovery and the ventral shell margins are
and spring tides. The tidal patterns are gener- strongly blunted. In contrast, the ventral
ally less obvious in subtidal bivalves (e.g., margins of unstressed specimens of Chione
Trachycardium procerum). In more temper- subrugosa are much thinner, as indicated by
ate regions annual patterns of increment de- comparison of these specimens with a collec-
position reflecting seasonality may also be tion of 194 Chione subrugosa valves from an
discernible. archaeological site at Valdivia, Ecuador.
Specific environmental or biological Sixty-one percent of the Ecuadorian shells

GEOARCHAEOLOGY:AN INTERNATIONAL JOURNAL 187

THE RECORD OF EL NINO

Figure 4. Chione subrugosa. Hachure marks indicate position of El Nino breaks.

A-B. Specimen # PT1-8, Tumbes, Peru, collected 4-16-84. C-D. Specimen #
Vald-2, Valdivia midden, Cut J, coastal Ecuador. E-F. Specimen # PT1-17,
Tumbes, Peru, collected 4-16-84. G-H. Specimen # PT1-19, Tumbes, Peru, col-
lected 4-16-84. I-J. Specimen # PT1-10, Tumbes, Peru, collected 4-16-84. K-L.
Specimen # 2PT1-6, Tumbes, Peru, collected 11-25-84. M-N. Specimen #
2PT1-11, Tumbes, Peru, collected 11-25-84.

had smooth profiles, with no stress breaks growth interruption in the valve profile. The
(Figure 4C, D). When breaks were observed El Nifio-stressed specimens did not exhibit
they were not the major interruptions de- such recovery.
scribed above (Rollins et al., 1986b). Growth increment counts taken on the
Moreover, the Ecuadorian shells that had photomosaics of acetate peel replicas of
stress breaks near the ventral margins of the Chione subrugosa extended from the ventral
shells showed rapid recovery and a thin ven- tips to the beginning of the major growth in-
tral margin was reestablished after the terruption. Increment counts ranged from

188 VOL. 2, NO. 3

 THE RECORD OF EL NINO

300-375 (average of 328) for specimens col- growth directed inward as there was horizon-
lected alive on April 16, 1984. The increment tally along the radial growth gradient. This is
counts on another specimen of Chione sub- indicated by the scatter and overlap of the
rugosa killed on November 25, 1984 totaled measurements plotted in Figure 6. The er-
531 (Table I). Counting backward from the ratic pattern of growth recovery in this spe-
ventral margin and assuming one increment cies substantiates the interpretation of a
addition per day places the time of onset of stress-induced break in contrast to some type
major stress on Chione subrugosa in late May, of ontogenetically programmed slow-down of
1983. This is coincident with the maximum shell growth.
SST anomalies measured at the Peruvian Increment analysis of Trachycardium pro-
ports of Paita and Chicama when water tem- cerum corroborates the results of the Chione
peratures reached 10°C above normal. Such subrugosa counts. Comparable counts from
increment counts are probably underesti- the ventral tip to the initiation of the growth
mates because of the likely anaerobic dissolu- break were about double those in Chione
tion or nondeposition of increments during subrugosa (Table I), and denoted major stress
intervals of maximum stress. Our results commencing on April 30, 1983 and June 20,
closely agree with the growth increment 1983 for the two specimens examined. These
counts of Pallant (in press), who used growth dates are well within the interval of elevated
breaks and the absence of sub-daily lines to SST.
infer stressful habitats of Chione subrugosa El Nino-induced stress profiles are less ac-
in a study of environmental changes in Costa centuated on Trachycardium procerum than
Rica. Our data support Pallant’s interpreta- on Chione subrugosa and may be subtle
tion as we noted that sub-daily lines were bunching of growth lines or broad shallow
deposited only prior to the onset of the El Nino depressions (Figure 3). Recovery of Trachy-
SST anomalies. Stress resulted in closely cardium procerum following El Nino stress
spaced lines of about equal strength. involved resumption of relatively normal
Specimens of Chione subrugosa exhibited growth and the addition of several centime-
slow and erratic recovery from the El Nino- ters of shell material. Only a few specimens
induced stress (Figure 5). The growth inter- displayed pronounced inward growth but in
ruptions recorded by notches in the valve these cases the slopes of the valve profiles are
profiles, unlike storm breaks, sometimes more gentle, presumably due to the nonre-
spanned several months and presumably in- flected mantle margin in this species. During
dicate extended periods of partial mantle re- major stress the angle of inclination of the
traction. The trauma of the El Nino-induced growth increments to the valve surface is only
stress is further indicated by a lack of corre- about 10 degrees. Normally (i.e., under
spondence, among individuals, in the amount nonstressful conditions) the inclination angle
of post-break shell deposition. In fact, many is about 35 degrees.
specimens displayed as much post-stress Trachycardium procerum showed trans-

Table I. Counts of Growth Increments-Ventral Margin to Onset

of El Nino Break.
Chione subrugosa
Date of # of # of Onset of El Nino
Specimen # Locality Death Increments Days Break-Date
2PT1-11 Tumbes 11/25/84 53 1 53 1 6/16/83
PT1-19 Tumbes 4/16/84 325 325 5/28/83
PT1-10 Tumbes 4/16/84 300 300 6/22/83
PT1-17 Tumbes 4/16/84 375 375 4/8/83
Trachycardium procerum
TP1-3 Chiclayo 4/18/84 715 358 4/27/83
TP3-1 Chimbote 4120184 614 307 6/19/83

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 189

THE RECORD OF EL NINO

Figure 5. Chione subrugosa. Specimen # FT1-19, collected 4-16-84, Tumbes, Peru.

A. Acetate peel, negative print (scale = 1.0 mm). Arrow indicates onset of El Nino
break at 325 days (325 increments) prior to death. Note extreme bunching of growth
increments from the initial break in valve profile to the anterior tip of the valve. B -D.
Photomicrographs (scale = 0.1 mm). Enlargements of the stress profile shown in
Figure 4A.

gressive development of crossed lamellar support to the recent studies of Cerastoderma

shell microstructure as far as the shell margin edule, another carditid clam (Richardson et
in association with the El Nifio break. In al., 1980,1981).In that species, tidal level was
Chione subrugosa the inner crossed lamellar determined to be a major factor influencing
layer also extended part way into the outer the rate of growth and more rapid growth
composite prismatic layer concomitant with correlated with longer periods of immersion.
the El Nino break, but never came close to the
shell margin (Figure 2). Stable Isotope Analysis
The nearly equal strength of semi-daily
growth increments in Trachycardium pro- For the most part, the elemental analysis of
cerum follows the pattern expected for-low the shells of Chione subrugosa and Trczchy-
intertidal and subtidal bivalves and lends cardium procerum resulted in anomalous or

190 VOL. 2. NO.3

 THE RECORD OF EL NlNO

15' X

14.

13.

12 '
X

11,

10.

-E 9.

-
E
0 .
X

X 0
0

I
> 7' 0
n TNOV 25,'84

30 31 32 33 34 35 36 37 30 39 40 41 42 43
TH (mm)
Figure 6. Two collections of Chione subrugosa.Graph shows lack of correlation between amount of post-El
Nino shell growth (as measured by DVM-horizontal distance from El Nino break to growing margin of the
shell) and total shell height (measured from umbo to growing margin of the shell). The El Nino-induced
shock resulted in long-term growth interruption. DVM does not record the amount of inward-directed, or
vertical, component of growth, however.

insignificant chemical differences between 613C in the carbonate shell material precipi-
the pre- and post-El Nino events. However, tated during the incursion (Killingley and
Sr/Na and stable isotope trends in the shell Berger, 1979). At the same time, the warmer
composition of Trachycardiumprocerum were waters should result in lighter 6l80 values in
indicative of changing oceanographic condi- molluscan shell carbonate.
tions during the El Nino event. In contrast, The anomalous values of both 613C and
Chione subrugosa exhibited a statistically in- 6lSO exhibited by specimens of Chione
significant Sr/Na and stable isotope pattern subrugosa (Figures 8 , 9) are consistent with
(compare Figures 7 - 12 and Table 11). the habitat of this species. Lagoonal mud flats
Upwelling currents are generally enriched are subject to frequent episodic localized tem-
in the light carbon isotope because the total perature spikes and we anticipate that the
dissolved inorganic carbon contains more shell record of such an intertidal species
C-12 than C-13. Therefore, the incursion of would exhibit a virtually indecipherable pat-
warm surface waters of the equatorial counter tern of temperature sensitive stable isotope
current into the colder waters of the Peruvian ratios.
Province should result in heavier values of In contrast, the elemental and isotopic com-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 191

THE RECORD OF EL NlNO
+ 0.5

0.0

-0.5

CI -1.0
0
a.
0
0 0 PTI-17 ,\"

t AA PTI-8 "
v
-1.5 I
~~
EN

Relatlve G r o w t h
0. PTI-19

Increment
Figure 7. SrINa-growth increment distribution dia-
n
u)

-2.0
iI
gram for three specimens of Chione subrugosa. Solid
symbols represent pre-El Nifio shell growth. EN is the
I
onset of the El Nino-induced shock, as indicated by
growth abnormality along the valve margin profile. -2.5
I
Open symbols represent shell carbonate deposition dur-
ing and after the El Nino shock.
EN
- 3.0 I I I

Table 11. Unpaired student t-Test of Sr/Na (wt) in aelative G r o w t h Increment

Chione subrugosa and Trachycardium procerum Figure 8. 613C-gro~th increment distribution dia-
erowth increment samdes. gram of Chione subrugosa. Explanations and symbols as
S r N a (wt) X in Figure 7.
Allochem N EL t-value P
C. subrugosa 0.234 0.248 0.756 s 0.375 Water salinities and temperatures can be
T. procerum 0.413 0.271 -3.796 =S 0.005
calculated with:
N = pre-El Nido growth increment samples; EL = post-
and El Nifio increment shell samples. Salinity (k1.0,ppt) = -5.037 In A 28.627, +
positions of Trachycardium procerum appear where A is the Sr/Na (wt.) of individual arago-
to reflect the physiochemical conditions of the nite samples (Brand, 1984) and with:
ambient seawater. The Sr/Na ratios in the
shell increment samples decrease after the El
Nino event (Figure 10). The pre- and post-El
Nifio values are significantly different at the
99.5% confidence level (Table 11). Similar where the tiA is the 1sO/'80 ratio of the mol-
trends are also depicted by the 613C composi- luscan aragonite and 6, is the lsO/''O ratio
tions of Trachycardium procerum, and the of the ambient seawater (Grossman and Ku,
heavier 613C values correlate with the ele- 1981). Application of these equations to the
vated SST along coastal Peru (Figure 11). SrINa values of Trachycardium procerum
Concurrently, the S1'0 values are lighter in (Figure 10) resulted in calculation of an aver-
the post-El Nino precipitated shell carbonate age pre-El Nifio water salinity of about 33 ppt
(Figure 121, presumably an indication of the and a post-El Nifio average value of 35 ppt.
warmer counter current water. Similarly, the calculated water temperature

192 VOL. 2, NO. 3

 THE RECORD OF EL NINO

- 1.4 will never be attained by these techniques,

but we are optimistic that some of the major
4 culturally disruptive El Nino events can be
-I 6 I recognized in the geoarchaeological record.
As such, we urge that attention be paid to
I the following assumptions and constraints:

-I 8
I 1. Clearly, detailed growth increment and
stable isotope analyses of bivalve shells are
I
h

a
0 limited by time and cost constraints. The type
n

-
0

,\" -2.0
l of growth increment analysis employed in
this study utilizing numerous photomosaics of
0 acetate peel replicas is labor intensive and
00 any study must necessarily be restricted to a
Lo
relatively small number of bivalve shells. By
-2 2 the same token, geochemical analysis of bi-
valve shells is constrained primarily by cost
and the large number of samples required per

t
-2.4 shell.
Techniques for more rapid processing of
EN molluscan shell material will be essential.
The results of our study suggest that the pro-
-2.6 1 II I I
file of the valve margin will provide a quick
Relatlve Growth Increment and relatively easy method for recognition of
Figure 9. 6"O-growth increment distribution dia- El Nino-induced stress among large collec-
gram ofChoine subrugosa.Explanations and symbols as
in Figure 1. tions of bivalve shells. The El Nino "signa-
tures" recognized in Chione subrugosa and
Trachycardiurn procerum involved unique
increased from a pre-El Niiio average of 17°C patterns of growth disruption that could be
to 22°C after the incursion of the warm equa- clearly separated from shorter-term environ-
torial counter current. mental stress, such as isolated storm events.
The use of the valve profile technique is en-
GEOARCHAEOLOGICAL IMPLICATIONS hanced by shell growth in tropical and sub-
tropical latitudes where strong seasonal
We believe that the results of this study growth disruptions are absent.
have the potential for establishing a baseline We suspect that the role of detailed growth
for recognition of ancient El Nino events, uti- increment and stable isotope analyses will be
lizing the abundant and widespread shell restricted to occasional checks of randomly
midden material along coastal South Amer- selected shells from midden material (and as-
ica. For the most part, species that are har- sociated natural deposits) and confirmation of
vested today along coastal South America are suspected El Nino-induced patterns.
available from middens and associated natu- 2. Our study demonstrated a strong habi-
ral deposits that reach back through thou- tat and/or biological constraint upon the use
sands of years of maritime culture, although of geochemical analysis of bivalve shells for
not without occasional widespread changes in recognition of El Nino-induced stress. We can-
species distribution patterns (Rollins et al, not expect that intertidal species will respond,
1986a). in terms of stable isotope ratios, similarly to
Recognition of large scale El Nino events subtidal species. The results of our study sug-
from analysis of shell midden material will gest that subtidal species will be more reliable
have to proceed with caution, however. We geochemical prospects, but even intertidal
suspect that a detailed chronology of El Nino species can be used if detailed growth incre-

GEOARCHAEOLOGY:AN INTERNATIONAL JOURNAL 193

THE RECORD OF EL NINO

0.E

0.5

0.4

+
v
3

z0 0.2
2
v)

I
00 TP3 - I
0
. 2TP2-10
0.2 0
. TPI-6

EN
I I I I 1 I

Relative Growth Increment

Figure 10. Sr/Na-growth increment distribution diagram
for four specimens of Trachycardium procerum. Explanation
and symbols as in Figure 7.
tli
ment analysis is performed. We emphasize
the desirability of establishing a careful set of
El Nino indicators for every species used. This
clearly cannot always be tied into an opportu-
t 1.0 nity for direct observation (as was the case
with the 1982- 1983 event) but should always
involve careful and detailed integration of
growth increment and stable isotope tech-
t 0.8 niques.
3. Ancillary information regarding sudden
and dramatic changes in taxonomic distribu-
tion should be considered in any attempt to
t 0.6
EN
t recognize ancient El Niiio events. Our obser-
vations during the 1982-1983 El Nifio sug-
gest that such changes will most likely be
selective. Large clams (e.g., Mesodesma don-
Relative Growth Increment
aciurn) and rock-dwelling gastropods (e.g.,
Figure 11. 613C-growth increment distribution dia-
gram of Trachycardium procerum. Explanations and Fissurella spp.) might be preferentially af-
symbols as in Figure 7. fected, and an abrupt decline in abundance of

194 VOL. 2, NO. 3

 THE RECORD OF EL NlNO

oc involves the nature of the midden record. Cer-

tainly, the search for ancient El Nifio events
will be most successful under conditions con-
-0.2
ducive to microstratigraphic sampling of rela-
tively undisturbed midden material. Shell
middens, however, are often culturally dis-
rupted (although the midden material associ-
-0 4 ated with seasonally occupied base camps or
shellfish harvesting sites along coastal Peru
may be less culturally disturbed than shell
-0.6
refuse heaps associated with more continu-
ously occupied sites). Stratigraphic sampling
of midden material should be supplemented,
wherever possible, with sampling of proximal
- 0.8 natural strata. Many of the sites along coastal
South America provide this opportunity (see
Rollins et al., 1986a1, but even in these cases
-1.0 care must be taken to avoid seriously time-
averaged faunal assemblages. Ideal condi-
tions for detection of ancient El Nino events
involved catastrophic burial; fortunately this
-1.2 may commonly accompany an El Nino pertur-
bation (e.g., beach ridge deposits associated
Relative Growth Increment with El Nino flooding events as discussed by
Figure 12. 6180-growth increment distribution dia- Sandweiss, 1986). The temporal “window”
gram of Trachycardium procerum. Explanations and available to be recorded is admittedly small,
symbols as in Figure 7. dependent upon the duration of the El Nino
event and the longevities of the bivalve spe-
these species in a particular stratum might be cies. Adequate conditions do exist, however,
significant. Sudden increases in the percent- in the fossil record, as documented by Clark
age of small mussels (especially Semimytilus (1987) during growth line analysis of a
algosus and Brachiodontes purpuratus) and Pliocene scallop assemblage. The chances of
the large gastropod Thais chocolata may sig- recovering a large scale El Nino event in the
nal the presence of a large scale El Nino event. ancient record are enhanced by the fact that
These species appeared to be minimally af- major El Ninos have a high probability of
fected during the 1982-1983 event. The being closely spaced in time. Quinn et al.
small beach clam, Donax obesulus, recovered (1978) assessed the probability of strong El
quickly following the 1982- 1983 El Nino and Nino events recurring within 7 - 8 years at 82
its sudden abundance in a stratigraphic se- percent.
quence might also be indicative. Dramatic in- We gratefully acknowledge the research support of a
creases in certain motile species (e.g., the scal- National Geographic Society grant and a Senior Faculty
lop Argopecten purpuratus) should also be grant from the Latin American Studies program a t the
noted. University of Pittsburgh. We profited from discussions
with Mr. Thomas DeVries of Oregon State University
This approach must be carefully applied, as and we are thankful for the drafting assistance of Mr.
many taxic changes in shell middens may Frank Benacquista.
indicate changes in resource exploitation un-
related to El Nino events (e.g., shifts in har- REFERENCES
vesting technologies or cultural preferences,
alterations in coastal landscapes, etc.). Arntz, W. E. (1986). The two faces of El Nifio
1982-1983. Meeresforsch. 31, 1-46.
4. Perhaps the most serious potential con- Bailey, G. N., Deith, M. R., and Shackleton, N. J. (1983).
straint upon application of these techniques Oxygen isotope analysis and seasonality determina-

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL 195

THE RECORD OF EL NlNO

tions: Limits and potentials of a new technique. paleoclimatic and life history records in mollusc
American Antiquity 48, 390-398. shells. Geobios Memoire special 8, 333-339.
Barber, R. T., and Chavez, F. P. (1983).Biological conse- Kennish, M. J., Lutz, R. A., and Rhoads, D. C. (1980).
quences of El Nifio. Science 222, 1203-1210. Preparation of Acetate Peels and Fractured Sections
Barker, R. M. (1964). Microtextural variation in for Observation of Growth Patterns within the Bi-
pelecypod shells. Malacologia 2(1),69-86. valve Shell. In D. C. Rhoads and R. A. Lutz, Eds.,
Brand, U. (1984).A salinity equation: Chemical evalua- Skeletal Growth of Aquatic Organisms, Appendix 1,
tion of molluscan aragonite. Society of Economic pp. 597-601. New York, Plenum.
Paleontologists and Mineralogists, Abstracts 1, 16. Kennish, M. J., and Olsson, R. K. (1975). Effects of
Brand, U., and Hinsperger, M. S. (1986). Geochemical thermal discharges on the microstructural growth of
analysis of carbonate rocks: Insoluble-residue free Mercenaria mercenaria. Environmental Geology 1,
method. Brock University,Research Report27,l- 13. 41-64.
Brand, U., and Viezer, J. (1980).Chemical diagenesis of Killingley, J. S., and Berger, W. H. (1979). Stable iso-
a multicomponent carbonate system I. Trace ele- topes in a mollusk shell: Detection of upwelling
ments. Journal of Sedimentary Petrology 50,1219- events. Science 205, 186-188.
1236. Koike, H. (1973). Daily growth lines of the clam, Mere-
Cane, M. A. (1983). Oceanographic events during El trix lusoria. Journal Anthropological Society of
Nino. Science 222, 1189-1195. Nippon 81, 122-138.
Caviedes, C. N. (1984).El Nifio. 1982-3. TheGeograph- Koike, H. (1980).Seasonal dating by growth-line count-
ical Review 74(3), 267-290. ing of the clam, Meretrix lusoria. Towards a recon-
Clark, G. R., I1 (1975). Periodic Growth and Biological struction of prehistoric shell-collecting activities in
Rhythms in Experimentally Grown Bivalves, in G. Japan. The University Museum, The University of
D. Rosenberg and S. K. Runcorn, Eds., Growth Tokyo, Bulletin 18, 1-120.
Rhythms and the History of the Earth’s Rotation, pp. Lutz, R. A., and Rhoads, D. C. (1977).Anaerobiosis and a
103-117. New York: John Wiley and Sons. theory of growth line formation. Science 198,
Clark, G. R., I1(1987).Growth line analysis ofa Pliocene 1222-1227.
scallop assemblage. Geological Society America, Moseley, M. E. (1975). The Maritime Foundations of
South Central Section, Abstracts with Programs, 147. Andean Ciuilization, Menlo Park, California: Cum-
Cunliffe, J. E. (1974). Description, interpretation, and mings, 131 pp.
preservation of growth increment patterns in shells Pallant, E. (in press). Applications of Molluscan Micro-
of Cenozoic bivalves. PhD Thesis, Rutgers Univer- Growth Analysis to Geoarchaeology: A Case Study
sity, New Brunswick, New Jersey., 169 pp. from Costa Rica, In J. Donahue and N. Lasca,
Deith, M. R. (1983). Molluscan calendars: The use of Archaeological Geology. Boulder: Geological Society
growth-line analysis to establish seasonality of shell- of America.
fish collection a t the Mesolithic site of Morton, Fife. Pannella, G., and MacClintock C. (1968).Biological and
Journal of Archaeological Science 10, 423-440. environmental rhythms reflected in molluscan shell
Deith, M. R. (1986). Subsistence strategies at a growth. Journal of Paleontology, Paleontological
Mesolithic camp site: Evidence from stable isotope Society Memoir 2, 64-79.
analysis of shells. Journal of Archaeological Science Quinn, W. H., Zopf, D. O., Short, K. S., and Kuo Yang,
13,61-78. R. T. W. (1978). Historical trends and statistics of
Ekaratne, S. U. K., and Crisp, D. J. (1982).Tidal micro- the southern oscillation, El Nino, and Indonesion
growth bands in intertidal gastropod shells, with an droughts. Fishery Bulletin 76, 663-678.
evaluation of band-dating techniques. Proceedings Rasmussen, E. M., and Wallace, J. M. (1983). Meteoro-
Royal Society London, B 214, 305-323. logical aspects of the El Niiio/Southern Oscillation.
Evans, J. W. (1975). Growth and Micromorphology of Science 222, 1195-1202.
Two Bivalves Exhibiting Non-Daily Growth Lines. Rhoads, D. C., and Lutz, R. A. (1980).Skeletal Growth of
In G. D. Rosenberg and S. K. Runcorn, Eds., Growth Aquatic Organisms. New York, Plenum, 750 pp.
Rhythms and the History of the Earth$ Rotation, pp. Richardson, C. A., Crisp, D. J., Runham, N. W., and
119-134. New York: John Wiley and Sons. GruiTydd, L1. D. (1980).The use oftidal growth bands
Grossman, E. T., and Ku, T. L. (1981).Aragonite-water in the shell of Cerastoderma edule to measure sea-
isotopic paleotemperature scale based on the benthic sonal growth rates under cool temperate and sub-
foraminifera Hoeglundia elegans. Geological Society arctic conditons. Journal of the Marine Biological
of America, Abstracts with Programs, 464. Association United Kingdom 60,977 -989.
Hall, C. A. Jr., Dollase, W. A., and Corbato, C. E. (1974). Richardson, C. A,, Crisp, D. J., and Runham, N. W.
Shell growth in Tivela stultorum (Mawe, 1823) and (1981). Factors influencing shell deposition during a
Callista chione (Linnaeus, 1758) (Bivalvia): Annual tidal cycle in the intertidal bivalve Cerastoderma
periodicity, latitudinal differences, and diminuation edule. Journal of the Marine Biological Association
with age. Palaeogeography, Palaeoclimatology, Pal- United Kingdom 61, 465-476.
aeoecology 15, 33-61. Richardson 111, J. B. (1981). Modeling the development
Jones, D. S. (1983). Sclerochronology: Reading the re- of sedentary maritime economies on the coast of
cord of the molluscan shell. American Scientist 71, Peru. Annals of the Carnegie Museum 50,139-150.
384-391. Rollins, H. B., Richardson 111, J. B., and Sandweiss, D.
Jones, D. S., Williams, D. F., Arthur, M. A., and Krantz, H. (1986a). The birth of El Nino: Geoarchaeological
D. E. (1984). Interpreting the paleoenvironmental, evidence and implications. Geoarchaeology 1, 3 - 15.

196 VOL. 2,NO. 3

 THE RECORD OF EL NINO

Rollins, H. B., Sandweiss, D. H., and Rollins, J. C. B. (1983). Landscape alteration and prehistoric hu-
(1986b).Effect of the 1982-1983 El Nino on bivalve man occupation on the north coast of Peru. Annals of
mollusks. National Geographic Research 2,106- 112. the Carnegie Museum 52, 211-298.
Rollins, H. B., Sandweiss, D. H., and Rollins, J. C. (in Shackleton, J. J. (1973). Oxygen isotope analysis as a
press). Mollusks and Coastal Archaeology: A Review. means of determining season of occupation of prehis-
In J. Donahue and N. Lasca, Eds., Archaeology Geol- toric midden sites. Archaeometry 15, 133-141.
ogy, Boulder: Geological Society of America. Thompson, I. (1975).Biological Clocks and Shell Growth
Sandweiss, D. H. (1986). The beach ridges at Santa, in Bivalves. I n G. D. Rosenberg and S. K. Runcorn,
Peru: El Nino, uplift, and prehistory. Geoarchaeology Eds., Growth Rhythms and the History of the Earth’s
1, 17-28. Rotation, pp. 149-161. New York: John Wiley and
Sandweiss, D. H., Rollins, H. B., and Richardson 111, J. Sons.

GEOARCHAEOLOGY:AN INTERNATIONAL JOURNAL 197