You are on page 1of 29

The Smithsonian Institution

Regents of the University of Michigan

Aja's Chronology: The Problem of Cave Eleven


Author(s): Walter Spink
Source: Ars Orientalis, Vol. 7 (1968), pp. 155-168
Published by: Freer Gallery of Art, The Smithsonian Institution and Department of the History of
Art, University of Michigan
Stable URL: http://www.jstor.org/stable/4629246 .
Accessed: 17/10/2011 13:24
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Smithsonian Institution and Regents of the University of Michigan are collaborating with JSTOR to
digitize, preserve and extend access to Ars Orientalis.

http://www.jstor.org

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE ELEVEN


BY WALTER SPINK
ART HISTORIANS HAVE TRADITIONALLY

as-

signedmany of the Mahayanapaintingsat


Ajantato the seventhcenturyA. D. or even
later; but epigraphistshave seldom concurredwith this view. No epigraphist,to
my knowledge, has consideredany of the
site's painted or incised donative records
to be later than the sixth centuryA. D.1 Instead, they have contended that the donative recordsbelongpartly to the last half
of the fifth centuryand partly to the sixth
century.'

No one can dispute the first part of


this proposition, since the Vakataka king
Harishelia, who was flourishing around
A.D. 475, is mentioned as the reigning

monarch in the dedicatory inscriptions of


Cave I 6 and Cave I 7?3But the secondpart

of the propositionmightbe questioned;for


it seems that the criteria upon which so
many of Ajanta's records have been assigned to the sixth century are subjectto
dispute.
Admittedly my attitudeis conditioned
by the fact that I do not believe artistic
1

Biihler, some hundred years ago, allowed


the possibility of inscriptionscontinuinginto the
early 7th century. But even he seems to prefer
a late sixth century terminus.He states that the
letter forms of the incised donative records in
Cave 26 "probablybelong to the latter half of the
sixth or beginningof the seventhcentury A. D.(and) come very close to the inscriptionsin Cave
i6 and 17-(which) probably (belong) to the end
of the fifth or the beginningof the sixth century
A.D." Biihler'sopinions are here quoted from J.
Burgess,Report on the Buddhist Cave Temples
and 7heir Inscriptions,ArchaeologicalSurvey of
WesternIndia, vol. 4, London, I 883, pp. I128-I3 3.

activity continuedat the site after about


A.D. 500, as I havearguedat lengthelsewhere.4If this hypothesisis true,the recordscouldobviouslynotbelongto thesixth
century,sincetheyaregenerallyassociated
withthegiftsof images.
I amnowencouraged
in my theoriesby
the fact that my friendDr.M.K.Dhavalikar,workingindependently,
hasassigned
an importantnewlydiscoveredinscription

in Cavei i to thisperiod.5
His view has
beensupportedat leasttentativelyby two
eminentauthorities,Dr.V. V. Mirashiand
ProfessorP. R. Srinivasan,both of whom
have beenkind enoughto studya photographof the record.Admittedlynone of
these three scholarshas agreedthat all
AjantaMahayanarecordscanbe assigned
to the fifth century;nor are any of them
dogmaticabouttheirattribution.Butperhapstheirpresentopinionis sufficientto
suggestthat the questionof the developmentof epigraphyat Ajantadeservesreconsideration,because the paintings in

Cavei i withwhichthisnewinscription
is
2 A long incised record which appears betweeen Cave 26 and Cave 27 dates from the
Rastrakiita period (see G. Yazdani, Ajanta, 4
parts, Oxford, I930, 1933, I946, i955; part 4
(text), pp. I2I-I24).
However, there is no evidencethat it is a donative record.
I Harishena'sregnal period is not precisely

known but probably falls at about A. D. 46 5-490.

See W. Spink, Ajanta to Ellora, Bombay, I967,


pp.

I-2.
4 W. Spink, Ajanta and

Ghatotkacha:a PreliminaryAnalysis,ArsOrientalis,vol. 6, pp. I 3 if.


I See Dhavalikar,New Inscriptions,Ars Orientalis,vol. 7, fig. 3.

56

WALTER SPINK

situationin Ajan.ta's
Thepaleographic
Mahayanaphaseis an anomalousone, a
If it is truethatthe Cavei i inscrip- factwhichhasperhapsnotbeensufficienttionis a late fifth-centuryinscription,and ly recognized,for no reallythoroughgoing
if it is also truethat the paintingsassoci- studyof thewholebodyof theseMahayana recordshasyet beenmade.Admittedly,
ated with it belongto Ajanta'slatestmosignificantdifferencescanbe seenbetween
ment of artistic production,then one
wouldhave to concludethat all of Ajan- recordssuchas thatin Cave I 6 andthatin
Cave 26, and it is clearthat the latteris
ta's latest paintingsshoulddate from amore
developed.Thishasbeenrecognized
boutthelate fifth century,andthat donaand is
and statedby variousepigraphists
tive inscriptions
integrallyassociatedwith
be
confirmed
seem
to
themmustalso be assignedto this period. a view whichwould
To negatethisconclusion,onewouldeither by the generallyrecognizeddifferencein
the characterof the excavationsthemhave to rejectthe fifth-centurydatingof
the inscriptionor else reject the theory selves.Buthow canwe gaugethe speedof
thattheCavei i paintings
belongto Ajan- this paleographicdevelopment?Could it
not have occupiedmerelythe courseof a
ta's very latestyearsof patronage.In the
few decadesratherthan the courseof a
presentpaperI shalltry to justifythislate
number
of generations,as has beengenedatingof the paintingsin the interiorof
Cavei i. I mustleaveit to paleographers rally supposed?There are no significant
to decideif the (presumed)sixth-century sixth-centuryrecordsat nearbysites to
whichthelatestAjantainscriptions
canbe
elsewhereat the sitecanbe reinscriptions
Vakataka
are
there
any
compared.6
Nor
assignedto the late fifth century,as I
inscriptionselsewherein India whichbewouldhope.
or to the
Of course,it is possiblethatpaleogra- long to the periodof Harishen.a
phic distinctionsof suchprecisioncannot time of his immediatesuccessors,if such
be made with any real assurancein the
6 Thereis only one donative recordfrom this
particularcase of Ajanta's Mahayana
can be assigned with fair
phasebecauseof the natureof the prob- general region which
assuranceto the sixth centuryA. D. It appearsin
lemsinvolved.If thisis so, it is important a quite damaged condition beneath an image of
that it be stated;for in the end, we may
Lakulisa in Ellora Cave I9, a cave which must
have to rely uponothercriteriato deter- have been undertakenby the Kalachurisat about
A.D. 55o (see W. Spink, Ellora's Earliest Phase,
mine the upperlimits of activity at the
associatedappearto be amongthe very

latest at the site (fig. I 7).

site. We may have to rely on the content

ratherthanthe letterformsof the epigraphicdocuments,


piecingtogethertheprobablehistoryof the sitefroma studyof the
of its art formsandfromour
development
scantyknowledgeof the politicalcircumstancesof the regionduringandjustafter
thelateVakatakaphase.

Bulletin of the American Academy of Benares,


vol. i, November, I967, pp. i i-z2). The inscription was discoveredby SureshVasantJadhav and
impressionstaken, but it has not yet been published. P. R. Srinivasanhas been kind enough to
attempt a readingand states that "all that can be
decipheredtentatively from the impressionsare
some proper names such as Bha[r]gava, Pachayana and Dharan[dha] rana" (in communication
to the author,dated October i, I964).

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

In theabsenceof
indeedexisted.7
successors
howcanonebecertainthat
suchguideposts,
formsin
the developmentof paleographic
this particularregion at this particular
pace?Ajantimeproceededat a "normal"
ta wasa sitewhichdevelopedundera very
receivingdonationsfrom
variedpatronage,
journeyingmerchantsandpilgrimsas well
as courtiersandmonksandlocaldevotees.
Artistsand scribesmusthave beendrawn
in largenumbersfromverydistantregions.
This welter of both patronageand profor thefact
ductionis probablyresponsible
in the
thatthereis littleformalconsistency
treatmentof the letter formsof the different inscriptions,even when they apparentlyareclosein date.A similarobservationcouldbe madeaboutthe treatment
of the paintingsat the sitewhichproveto
be theworkof manydifferenthands.Some
artists were careless, some careful, some
conservativeand some advanced. In such
a period of urgent progressand ready assimilation of ideas and influences, we
might well expect to find a rapid evolu7 Following V. V. Mirashi (Historical Data
in Annals of the
in Dandin's Das'akumaracarita,
BhandarkarOriental Research Institute, vol. 26,
I945,
pp. 20-2I),
I have postulated the reign of
Harishena'sson for a short period prior to the
break-upof Vakatakapower (seeW. Spink,Ajanta and Ghatotkacha, pp. I35-I36). But we have

no inscriptionsfrom his (presumed)reign; nor is


there evidence that Vakataka power was ever
restoredafter aboutA.D. 5oo. The old theory that
the marriage of a Vakataka princess to the
VishnukundinMddhavavarmantook place in the
sixth century is no longer tenable, since recent
studies suggest that Madhavavarmanwas ruling
aroundA. D. 468-5 I 8. See Rama Rao, New Light
on the Vishnukundinsin Proceedingsof the 27th
Indian History Congress, Allahabad, I965; see
also S. Sankaranarayana,New Light on the Genealogy and Chronology of the Vishnukundins,

I57

tion and a varietyof trendsin the treatmentof letterformsjust as we do in the


formsof art.
One otherfactormightbe considered.
Evenwhileassertingthata distinctdifferentiationcan be madebetweenthe earlier
and later Mahayanaletter forms at the
site, paleographers
have generallylimited
thespanof timecoveredto approximately
fifty years.8Now it standsto reasonthat
withina spaceof so few decades,a space
easilybridgedby the life of a singleman,
variousfactorscan complicatethe paleographicpicture.In writingtoday, I use
letter forms which I learned about Ig9,
e. g. TD, 5, aY, 9 . My parents, in writing

today, use letter forms which they learned


a generation earlier, around i 9 I 0, e. g.
R*gS,

Essentially neither their style


nor mine has changed,so that in I968 we
are writingin "I93 5" and "I9IO" styles
<

respectively.Was the situationso different


in the fifth century?Does not the paleographicform (and thus the presumeddate)
of the inscriptiondepend to a certain deJournalof the Oriental Institute,Baroda,vol. i6,
no.4,June,I967.
8 Even Biihler (see footnote i) appears to
have admittedthis possibility, althoughhis statements can also be read as allowing a i 5o year
span. Recently a span of a century or more has
been suggested by W. Begley (see Begley, The
Chronology of MahLiyanaBuddhistArchitecture
and Painting at Ajanta, unpublisheddissertation,
University of Pennsylvania, I966, pp. 48-49).
Begley dates certain inscriptions (e.g. Cave z6
donative record) on paleographicgrounds to the
last half of the sixth century. His late-sixth century dating of certain excavations and images at
Ajanta is based upon this premise. We would
question this premisebut, even in doing so, must
admit that Begley has consideredthe mattercarefully and has consultedwith a numberof eminent
epigraphistsbefore reachinghis conclusions.

WALTER SPINK

1 58

greeupon the age of the writer,a factor


whichcan generallynot be determinedin
ancientrecords.
Of course,such an observationin no
way changesthe fact thatovera givenperiodof timeanyrandomsamplingof letter
formswill tendto showa definitedevelopment.But it doessuggestthat in any particularcomparison
one cannotbe too dogmatic aboutthe relativedatesof the inscriptionscompared,forat anygivenpoint
in timewe shouldexpecta coexistenceof
individual-indeed older and younger-

lateproductionby an olderandconservative scribe;andyearsbeforeeitherhe or a


contemporarycould have inscribedthe
somewhatsimilarrecordsin Cavesi 6 and

in time?

tion is associated. If the inscription can in-

Suchargumentscan of coursebe used


in two ways.The Cave i i recordin question might(by thesametoken)be a rather

deed be proved to belong to the late fifth


century,it is obviously of great interestto
determine the relative position of the

7. I am only proposingthat it is difficult


to arrangeinscribedimagesor caves at the
site in any kind of absolutesequenceon the
basisof paleography.The distinctionsthat
we wish to make are too fine-the time
span underconsiderationis too short-for
paleographyto be anything more than a
suggestive guide. Since we are dealing not
with differencesof centuriesbut with difstyles. Can we really say, with Sircar,that
ferencesof decades,paleographicobservathe shortpaintedinscriptionon the Bud- tions may confuse the problem as much
as clarify it. Art historical factors such
dha'sthronein Cave4 belongsto the first
half of the sixth century,and is thussigas location, state of completion, changes
nificantlylater than the long and pres- in iconographyreflecting changing ritual
tigious rock-cutinscriptionsin Cave I6
needs,changesin decorativemotifs reflecand Cave I 7?9The letter style of the Cave
ting changingpredelictionsof the patrons
4 recordmay be later,but the writermay
-these aresurerbasesuponwhich to judge
havemerelybeena youngercontemporary the site'sinternaldevelopmentand to place
of the otherscribe.Indeed,the scribechoit in relationto the few fixed chronological
sento cut a courtlyrecordsuchas that of
referencepoints in the late fifth century
Cave i 6 would quite possiblybe an estab(e.g. Cave i 6, and Ghatotkacha inscriplishedandprestigious
(andthereforeolder) tions) with which history has provided
us.10
craftsman.Is it not conceivablethat the
two recordsareactuallycontemporary
or
Let me return to the problem of the
(as I wouldpreferto believeon the basis style and date of the Cave i i paintings
of othercriteria)only some20 yearsapart with which the newly discoveredinscrip-

9 D. C. Sircar, Inscription in Cave 4 at Ajan-

ords.Furthermore,
as mentionedabove,the Cave

andCaveI7 records
maywellhavebeendone

pa, in Epigraphia Indica, vol. 33, pt. 6, sec. 49,

I6

pp. 259-262.
10 The

by an oldeirscribe,in a unconsciously
retardatory
style. It is even possiblethat such prestigious
donativerecordsweredonein a consciouslyformal and perhapsslightlyarchaicmode;this is a
commonphenomenoneven today in public or

fact that the earliest productions of


the Mahayana phase at Ajanta provide us with
relatively few inscriptions increases the problems,
for we do not have a random selection of letter
forms to compare with the numerous later rec-

prestigiousinscriptions.

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

paintingsin the developmentof the site.


As statedabove,it is my contentionthat
thesepaintingsrepresentworkof the very
latestphaseof Ajanta'spatronageandthus
can be seen as an upperterminusfor activity at the site.Thisopinionis sharedby
Begley in his recent study of Ajanta's
stylisticdevelopment,
as well asby a number of other scholarswith whom I have
discussed
thepaintingsat thesiteitself.1"
Someimmediateobjectionsmaybe anticipated in our attempt to date these
paintingsto the very latestphaseof Ma-

hayanaactivity,for Cavei i hasalways


(and rightly)been consideredamongthe
very first of the cavescut out duringthe
site'sMahayanaphase.12

Cave i i underwentan unusualand


very early development.But we shall attemptto prove that only the excavation
as suchand certainof its paintingsbelong
to the early yearsof Ajanta'sMahayana
phase.The paintingswhichare associated
with the new inscriptionbelongto a much
latermoment,possiblyseparatedfromthe
earlierpaintingsby asmuchastwo orthree
decades.
Therearea few othercavesat Ajanta
whichshowbothvery earlyandvery late
Mahayanaforms,but noneof themshow
quite such a distinct separationof the
phasesas Cavei i. Cavessuchas Cave4
11 See Begley, op. cit.

However Begley dates

these Cave i i paintings to the late sixth century


because of his different view of paleographic
developments at the site (see footnote 8).
12 The view of Fergusson and Burgess (see

J. Fergussonand J. Burgess,Cave Templesof India, London, i8 8o) that Cave i i was actually a
converted Hinayana excavation is of course no
longer accepted; but their opinion does point up
the fact that the excavation as such has many
early and formative features.

159

and the two-storiedCave 6 were begun


very early and progressedso slowly that
workwasstillgoingon whenpatronageat
the site was finallyinterrupted.13
Cave I6
showsthe additionof very late paintings
(all representations
of Buddhagroups)on
therearwall andalongtherearportionof
the left wall of its interior,on surfaces
whichhadbeenpreparedfor paintingearlier but which were never actually adorned."4
Cave I9 showsthe additionof
shrineletsat eithersideof its court,which
mayhavebeenplannedfromthebeginning
butwerecertainlynotcompleteduntilvery
late. But only Cave i i, of all the Mahayana caves,providesa casewherean already completedform or motif was reworkedat a later date. This very fact is
remarkable,for it helps to confirmthe
hypothesisthat the total span of Mahayanapatronageat the sitewas very short;
otherwisewe shouldexpectto find many
such examplesof repairor reworkingin
othercavesas well.15The formin question
13 For a detailed analysisof the development
of the plans of Cave 4 and 6, see Spink, Ajanta
and Ghatotkacha,pp. 1 S-I0 -5
14 The division between the earlier and the
later phaseson the left wall of the interioris clear,
for the narrative scenes occupying the forward
portion of the wall break off abruptly at about
the mid-point, where some of the figures are incomplete.
15 In a few instances we do have situations
where the plan of work was changed while in
progress.In the ceiling of Cave 2 a preliminary
designhas been coveredover. A standingBuddha
panel in Cave 26 was cut over a preliminary
sketch of a seated figure. On the left rear wall of
Cave 2 a painted Buddha was crowded into the
corner of an already completed Sravasti Miracle
scene, while intrusive sculptured panels in the
porch of Cave 4 and in Cave Upper 6 occupy
areas which were presumablyintended to show

6o

WALTER SPINK

is themainBuddhaimage;andas we shall
try to showbelow,its two layersof decoration almost certainlycorrespondwith
the two phases of patronageof which
othermotifsin thecavegiveevidence.

porchdoor (fig. 4) are also early characteristics.16


But perhapsthe most striking
early characteristicof the cave is the
treatmentof its still very undeveloped
shrinewhere,as in othervery earlycaves

Thelocationof Cavei i immediately (CavesLower 6,


suggestsits relativelyearly date, for it is
placedin the very midstof the Hinayana
nucleusaroundwhichthe site'sMahayana
developmentbegan. Its plan (fig.i) has
beensomewhataffectedby its location,for
the normalcomplementof cells has been
omittedalong its right wall, presumably
becauseof the architect'sfear of breaking
into the nearbyHinayanaChaityaHall,
Cave io. But the confusion of its plan,

whichincludesan off-centerplacementof
theshrineaswell as a generallackof symmetry,is probablyalso due to the planners'inexperience
in excavationduringthe
first yearsof Mahayanapatronageat the
site. Suchinexperiencecouldalso explain
the anomalouscharacterof the treatment
of the porchbalustrade,the clumsyform
of the interiorpillarsand the unusualinclusionof rock-carved
seatsalongtheright
wall of the interiorand in parts of the
porch (figs. 2 and 7). Theseare elements
which find no currencyin later excavations.On the otherhandthe porchpillars
and pilastersare comparablewith (and
probablyare the sourceof) similarelementsin other early Mahayanacaves at
the site. The simplecell doorsalso relate
thecaveto otherveryearlyMahayanaexcavations,whilethesimpleformatandthe
floral decoration(here painted) of the
paintings.But all such changeswere made in normal course as work progressed and interests
changed.They are in no sense repairsor reworkings of already completed areas or images, as is
the casewith the Cave i i Buddha.

i6 and I7), one could cir-

cumambulate
the Buddhaimage(fig.

2I).

Furthermore,the Buddha image is enthroned against a clearly defined monolithic stupa,which can be seenas one proceedsaroundthe image (fig. 22). Thisidea
of incorporatinga stupa in the shrinearea
seems to be a very early one in the site's
Mahayana development. The shrines of
Caves Lower 6, I6 and I7, all of which
were probably completed somewhatlater
than that of Cave i i, show no such stupa,
while in still later shrines even the possibility of circumambulationis denied.17
The Buddha image in Cave i i was
completely carved, but the pradakshinapatha was never finished; one can move
throughit, but only with difficulty, for it
has been left very roughat its lower levels
(fig. 2I).

It seemsevident that the first

phase of work on the cave was brokenoff


at just this point. If it appearscuriousthat
the pradakshinapathawas not finished
duringthe later phaseof work in the cave,
the explanationis probably that modes of
worship changed,making the renewal of
work on the passage pointless; as mentioned above, none of the latest viharasat
the site have pradakshiniapathas
in their
shrines.
For a brief survey of the developmentof
doorway forms at Ajan.ta,see Spink, Ajanta and
Ellora, pp. 38-43.
17 The changingarrangements
of shrineplans
in the Mahayanacaves at Ajanta are discussedin
detail in Spink, Ajantai and Ghatotkacha, pp.
16

I38

ff.

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

i6i

were calculatedto impress.To give some


of the splendidorganizationof
a considerable impression
Cavei i was interrupted,
herethe
amountof paintinghadalreadybeendone. theCavei i ceilingwe reproduce
At the timewhenexcavationworkon

This is not surprising;a review of painting

central section (fig. 3), with its alternating

procedures
at the sitewouldshowthatthe
paintersalwaysfollowedcloseon theheels

floral and zoomorphic forms, ranged respectively upon a greenand red "checkerboard"pattern,the latter now so darkened
that its colorscan hardlybe seen.
The Cave i i porchceiling,like all ceilings of the Mahayanaphase at Ajant1a,reflects structuralprototypes. But it is interestingto note, as a kind of confirmation
of the early date of the ceiling painting,
that the artist is here concernedto "support" the painted transverse"beams"by
means of atlantes painted at the points
where the "beams"meet the porch's rear
wall (fig. 5). Slightly later, in the porch of
Cave I7, theseatlantesare moved into the
ceiling area itself, for they must have been
consideredas obtrusiveelementswi ich interruptedthe painted panels planlied for
the wall surfaces.Still later, with a similar
logic, they are omitted entirely-see for
instancethe paintedceilingof the porchof

of the excavators.For instance,they gen-

erally paintedthe porcheswhile the cuttingoutof theinteriorswasstillunderway;


and they often beganthe paintingof the
mainhall while the shrineand lessercells
were still being excavated.This is easily
discerniblefrom evidence provided by
many of the partly finishedcaves at the
site.

Certainmajorpaintedelementsin the
porchareof a characteristically
earlytype
and,whencomparedwith thoseof the interiorhall,serveto confirmthedivisionof
the paintingof the cave into two distinct
phases.The ceiling(fig. 2), muchdamaged
by smokedepositsandby someone'swellmeaningattemptto cleanit with a broom,
is for all of its discoloration
stilla splendid
exampleof the boldly executedbut carefully organized early Mahayana-phase Cave2.
Thebeautifulbodhisattvas(fig. 4) (now
ceilingforms.18
It canbe readilycompared
muchobliterated)which appearon the rear
to ceilingssuch as those in the porch of

CaveI7 or in theambulatory
of CaveI9.
Laterexamplessuchas thosein the unfinishedinteriorof Cave2I, or eventhosein
the interior of Cave 2, appear fussy and

lackingin exuberanceof designby comparison, for all of the fact that later ceilings incorporatenew characteristics(pan-

els with tiny humanor yaksha figures,


lavishuseof expensivepigment,etc.) which
18 The depositswere probablyfrom the cooking fires of hermits who in some later century
must have inhabited the cave. There are various
painted symbols (includingtridents)in the porch,
which recall their presence.

wall of Cave i i's porch, on either side of


the door, must also belong to this early
phase of work. They were probably the
first of these conventional pairs which
flank so many of the major doorways at
Ajanta. These great guardians,like those
which remainin the porchesof Caves I7
and 2 and in the interiorsof Caves i and
2, stood against a now nearly obliterated
backgroundof rocks and vegetation and
clouds. Fortunatelyjust enough traces of
the clouds remain at the very top of the
wall to show that these motifs sometimes
run ever so slightly up onto the ceiling,

WALTER SPINK

I62

overlappingthe edge of the ceiling design.19


Thuswe canbe surethatthe ceiling
wasunderwayandprobablycompletedby
the time the guardianpanelswere begun.
Thisis as we shouldexpect,for clearevidencein numerousothercavesalso shows
that it was customaryto completeceiling
areasbeforeadjacentwall surfaceswere
decorated.Thisis not to say thatsuchceilings (which for practical reasonswere
painted first) belong to a separateearlier

phaseof work.Suchan assumption


would
hardlybe logical.Indeedin the porchof
Cave I i we have clearevidencethat the
greatguardianswereplannedalongwith
the ceiling,for a borderof paintedpearlmotifswhich "depends"from the ceiling
(tracesstill being visible along the left
lateralwall andalongpartof therearwall
of theporch)is interrupted
justabovethese
two panels.It wouldhavebeencontinued
at thesepointshadnot theguardianpanels
been alreadyconceived,perhapsalready
begun,when it was added.By this same
token we can say that the large flaming
Buddhaat theleft endof therearwall was
conceivedafter the ceilingwas finished,
for parts of its mandorlaare actually
painted over this pearl-motif(fig. 6).2?
Here, again,the situationis the expected
one;for evidenceof overlappingsin other
caves generallyshowsthat motifs at the
lateralendsof the rearwall of the porch
werenot painteduntil afterthoselocated
closer to the cave's axis. This is not to say
19

Observationsregardingthe tracesof clouds


above the bodhisattvaswere made at the site but
are not adequatelyvisible in the photographs.
20 Tracesof the ruinous"pearl-motif"can be
seen in fig. 6 at the very top of the wall. On the
spot observationsmake it clear that the borderof
the mandorlais paintedover this motif.

that the flamingBuddhadoesnot belong


to the samegeneralphaseof paintingas
the greatguardiansand the porchceiling.
Stylisticfactorssuggestthat most of the
paintingsin the porchbelongto thissame
early periodof activity;but someof the
motifsareso obscuredby smoke,by abrasion, or by laterSaivitesymbolsthat one
cannotbe too dogmaticabout this. Admittedly,theunfinishedstateof therestof
thecavemakesit quiteconceivable
thatthe
porchdecorationwasnot fully achievedin
thefirstphaseof work(theobscuredseries
of seatedBuddhasalongtheupperedgeof
the right rearwall and of the right side
wall couldwell be additionsof the second
phaseof work). Sufficeit to say that in
the porchwe have a considerable
bodyof
painting(includingthe fragmentarymotifs which definethe door, window,and
pillar designs)which belongto the first
few yearsof Mahayanaactivityat thesite,
and which were completedwhile the excavation of the interior of the cave was

stillunderway.21
Movingintotheinteriorwemeeta very
differentsituation.Here we find that the
rearand lateralwalls are decorated,but
that the ceiling(althoughpreparedwith
mud plaster)is not (fig. 7). Nor is there
any evidencethat the interiorpillars(as
opposedto the porch pillars)were ever
A ruinouspainted Litany scene at the left
end of the porch, which we assumealso belongs
to the first phase of work on the porch, has a
short (unpublished)inscriptionbeneath.It would
be of interest to comparethe letter forms of this
inscriptionwith those (also unpublished)beneath
a now ruinousstandingBuddha(see below) on the
rearwall of the interior.The latter image,like the
seated Buddha associated with the newly discoveredinscription(figs. i6 and 17), mustbelong
to the secondphaseof work on the cave.
21

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

I63

demandspedecorated.Theseobservations
cial consideration,
for the weightof evidenceat Ajantasuggeststhat the ceilings
and pillarswere normallycompletedbefore the walls.22Furthermore,
as pointed
out above,the Buddhain the shrineis remarkablein that it shows two distinct
layersof paint.
After making such observations,we

beenpainted(or was hurriedlypaintedin

can reconstructthe over-all development

only to be taken up again some years


later. Perhapsinterestin continuingwork

of workin thecaveasfollows.
The excavation, started very early in
the Mahayanaphase,proceededin the normal fashion, the porch being completed
first. Paintingof the porchprobablybegan
while the interior hall was still being cut
out. When the excavation of the interior
hall was complete, and excavation of the
shrinewas in progress,the interiorhall was
readiedfor painting by the applicationof
mudplasterto its ceilingand its four walls.
It is possible that the Buddha image (and
its halo) was painted at this samemoment
(or even beforethe walls and ceilingof the
main hall were plastered) for we have
proof in other instances that this crucial
element was sometimes completed relatively early, presumablyto make the cave
useful for worshipeven when its total decorative schemehad not been realized.23In
any case,it musthave beenjust at this moment that the work in progresswas halted,
when the pradakshinapathawas still incomplete,when the Buddhaimagehad just
22

This is evident from a study of such major

caves as Caves i6, I, 2, and 2I. Fully completed


caves such as Cave I 7, Lower 6, or i 9 cannot of
course provide such evidence; nor can caves such
as 3, 4, 5, Upper 6,

14,

15,

22,

23,

24,

25,

27,

etc.,

where a consistent program of painting was never


undertaken due (in most instances) to their being
underway at such a late date.

anticipation of work being interrupted)

and when the interiorhall had beenprepared (i.e., mud plastered)to receiveits
decoration.Justwhy workwasbroughtto
a haltwe shallprobablyneverknow,anymorethanwe shallknow why work was
interruptedrelatively early on certain
otherearlyMahayanacavessuchas I 6 and
20,

on the decorationsflaggedonce the caves


were actually put into use, just as present

day buildings sometimesnever are quite


finishedor are continuedonly sporadically
once an initial surgeof activity has run its
course.
When work was taken up again in
Cave ri, a very characteristicchange in
attitude was evidenced. The walls of the
previouslyunfinishedinteriorhall (except
for the unfinishedand less important-because less visible-front wall) were now
literally coveredwith Buddhasand bodhisattvas of all different shapes and sizes.
Their placementis quite random,suggesting a reflection of the interestsof various
separate donors rather than of one controlling patron working with a masterplanner.On the right wall (figs. 8 and 9),
dreadfully obscured by grimy deposits,
arerankupon rankof tiny seatedBuddhas,
23 See Caves 20, Upper 6, 4, and even Caves
and i at Ajantii, as well as the Ghatotkacha
vihara. The objectionthat the Buddhawould not
be paintedbefore the pradakshin.apatha
was completely cut cannotbe sustained,for thereare many
instances at Ajanta where painted areas are directly contiguousto areasstill being workedon by
the excavators. For instance, the ceiling of the
rear aisle of Cave 21 was painted while the cutting of the contiguouspillars of the shrine antechamberwas underway.

i6

WALTER SPINK

i64

in themeandstyleof the
very reminiscent
in Cave
multipleBuddharepresentations
2's shrinevestibule,an area which must
have beendecoratedduringthe very last
yearsof thesite'sactivity.24
TheseCave i i
figures,whichmay be intendedto repre-

sent the "IOOO Buddhas"(the groupin


Cave2 is so inscribed),surrounda larger
standingimageflankedby two attendants
holdingmusicalinstruments(fig. io).
On theopposite(left)wall (figs.i i and
I2),

numerous
largerBuddhas
The
appear.

upperlevelof thewall showsa barelyvisible seriesof standing(perhapswalking)


Buddhasat the left, and a seriesof seated
Buddhasat the right(thelatterseriescontinuesaroundon theleft rearwall).Justto
the rightof the secondcell doora hieratic
standingimage(Avalokitesvara?)
can still
be seen.Towardthe frontend of the wall
betweenthe first two cell doorways(fig.
I i), thereweretwo largeandsplendidenthronedimages,donein a flashybuthighly
skilledstyle with muchuse of a brilliant
orangepigmentwhichis oftenfoundin the
intrusivelate panelsaddedto othercaves
at the site (e. g. Caves4, Upper6, 2I).
Noteworthy are the long-neckedbirds'
headsspringingfromthethronebacks
(figs.
I3 and I4). In sculpturedshrine groups,
which are easierto arrangechronologically
becauseof their integral connectionswith
the excavations proper, such motifs only
appearin the latest phasesof work at the
site.2 A study of painted groups,or of intrusive panels, suggests the same conclusion, for such birds' heads only appear in

distinctlylate contexts:one never finds


them, for instance,amongthe numerous
paintedBuddhapanelsin the earlyChai-

tya Cave I9, but they are commonin


ChaityaCave26.
Near the left endof the rearwall (fig.
15) another
largeBuddhaappeared;
it is
now practicallyobliterated,but the head
of oneof the two attendantshasbeenpreservedby theDepartmentof Archaeology,
and shows clear relationshipswith the
headsof the attendantsof the groupson
the left wall. Thereis certainlyeveryreason to assumethat this group,and indeed
all of the figureson the backwall, belong
to the samephaseof late activityas those
on the side walls. This assumptionis furthersubstantiated
by the curiousfact that
this groupand someotherfigureson the
backwall werepaintedovercloth,as were
someof the figureson the left wall. Nei-

therin theporchof Cavei i norindeedin


any portionof any othercave at the site
do we finda counterpart
to this.Wereturn
to thisproblembelow.
Justto theleft of theshrinedooron the
rear wall (fig. I5), there are a numberof
tiers of repetitive seated Buddhas, raised
upon stemmed lotus pedestals. This was
clearly a hieratic formulation of an epiphany scene, probably representingthe
SravastiMiracle,a compositionwhich becameincreasinglypopularat Ajanta as the
site developedin its Mahayanaphase.
The right rear wall of Cave i i's interior displays,at its right end, an elaborate
The long necked birds' heads (which replace earlier lotus tendrils) become particularly
common as a throne motif in sixth-centuryBuddhist compositions(see Spink, Ajanta to Ellora,
p. 3'). Still later, they are replaced by complete
birds,which often hold flowers or beads.
25

Cave 2 was one of the last fully excavated


caves at this site (see Spink, Ajanta and Ghatotkacha, pp. I43-I46),
and its shrine vestibule, being at the rear of the cave, must have been one of
the last partsof the cave to have been decorated.
24

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

pavilion composition, now very ruinous


(fig. i6). Possibly it representsa Buddha
paradise motif and as such would tie in
convincingly with the apparently salvation-orientediconographyof the surrounding walls. It is conceivablethat the pavilion revealed a central Buddha image; but
the damagedconditionof the frescomakes
this a mereconjecture.
Between the pavilion scene and the
shrine door, on this same right rear wall,
an impressivegroup of large seated Buddhas was painted (fig. i6). This group,
now badly damaged, forms a kind of
counterpartto the groupof smallerimages
just to the left of the shrinedoor; however
the artistshave made no attempt to relate
the two compositionsin formal terms, as
one might expect them to have done in a
cave with a more consistentoverall scheme
of decoration.
This group of seated Buddhas to the
right of the shrinedoor is in itself a rather
disorganized conception, again suggesting

that we are dealing with a patchwork of


separate donations rather than with a
carefully pre-planned composition. The
upper third of the wall area, at least part
of which was surfacedwith fabric, shows
two elaborately enthroned Buddhas in
(theso-calledEuropean
pralambapadasana
pose) (fig. I7). The newly discovereddonative inscription of Mitradharma,which
Dr. Dhavalikarhas dated to the latterpart
of the fifth century,appearson the throne
pedestal of the image at the left.26The
other image appears to have had no inscription.
26
See fig. 3 in the preceding article by Dr.
Dhavalikarin this sameissueof Ars Orientalis.

These two Buddhaimagesalong the


upperportionof the areato the rightof
the shrinedoor are barelyvisibledue to
surfaceloss as well as obscuringgrime
and needto be photographed
with special
(infra-red?)techniques.Until this is done
one can hardly discussthem properly.
However,it can be saidthat they appear
to have beenskillfullyexecutedin a style
somewhatlike that usedfor the similarly
conceivedlateBuddhason therearwall of
theinteriorof Cave I 6, andfor the elaborately enthronedand inscribedlate Buddhaon theright"triforium"
of Cave9.
The bestpreservedof thesetwo Buddhas is the inscribedone at the left (fig.
I6). Fainttracesof its facearevisible,as is
the face and muchof the body of the elegantlyposedbodhisattvaat its properleft.
Two kneelingdevoteescanalsobe seenon
thissameside(we wouldprobablybe correctin assumingthatsimilarattendantfigures appearedat the properright).This
Buddha'sthroneis supportedby tensely
posedcrouchinglions of a type often encounteredon the thronesof both sculpturedandpaintedBuddhasduringthelater
phasesof workat thesite.
TheadjacentBuddhaimage,whichalso had a pairof bodhisattvaattendants,is
seated upon a somewhatmore conventional four-leggedthrone(fig.I 7). His feet
rest upon an expandedlotus, which serves

ashis "footstool."Suchlotusfootstoolsare
not in themselvesa surehallmarkof late
date,butit is perhapssignificantthatthey
areusedexclusivelyfor suchpralambapadasanaimagesin the laterphase,whereas
a few decadesearlierrectangular
pedestals
or plinthsare often found. In this same
regard,it mightbe notedthat thepralambapadasana
posebecameincreasingly
pop-

WALTER SPINK

I66

ular as iconographydevelopedin the last

The assumptionthat the interiorpaint-

third of the fifth century at Ajaiuta;at


other sites, in the sixth century, its popularity increasedstill further.

ings were done during this late period


when, as many other instancesthrough-

Directly beneaththe Buddhaimages


justdiscussed,
three(orperhapsfour)other
largeBuddhasonce appeared,alongwith
a numberof tinyseatedones,fourof which

uponimageryratherthanuponornamenting or even upon completingexcavations


is furthersuggestedby the fact that the
ceilingis not decorated.This is in direct
contrastto thesituationin theporchwhere,
as we have noted,the ceilingwas undertakenfirst,and with greatattentionto its

can still be seen just to the left of the simple cell doorway(fig.i6). Only oneof the

largerimagesremains;but the otherswere


probablyof the sametype, renderedin a
conventionalcross-legged
pose.Theywere
all shelteredby still evident three-tiered
umbrellas
similarto thoseovertheseriesof
seatedBuddhaswhichin the very late period of work at the site were added to the

left wall of theinteriorof Cave i 6. In the

latter cave, just as in Cave i i, iconographicemphasishad shiftedto the representationof multipleBuddhaimagesduringthesecondphaseof work.
Althoughthe lower portion of this sec-

tionof therightrearwallin Cavei i has


lost all of its painted surface, it is likely

that it too was filled with seatedBuddha


images.

Thus we see that the interiorhall of


Cave i i is literallyfilledwith a varietyof
Buddhaimageswhich,by virtueof certain
of styleandiconographyas
characteristics
well as by virtueof theveryideaof filling
an interiorwith sucha randomassortment
of purelyiconicforms,shouldbe assigned
to a very lateperiodof activityat the site.
The fact, too, that the hall still has available surfaces(suchas the front wall) which
are unadornedwith images also suggests
that it was being decoratedat a very late
period;that is, it was never fully completed, becausetime ran out.

out the site attests,interestswere centered

lavish effect. It is also in direct contrastto

the situationin all otherearlyMahayana


excavations, at least in those where work

progressedwithoutinterruption.It seems
clearthattheoriginalplannersof the cave
hadeveryintentionof decoratingthe ceiling beforetheirworkwas interruptedbecausetheceiling,likethewalls,hasa coating of mud plaster(fig.7). It is difficult
to believe that this mud plasteringwas
done in the second,later phaseof work,
becauseif later artisanshad taken the
troubleto applyit, theywouldalmostcertainlyhave decoratedit. (Thesamecould
be said for the plasteredbut unpainted
front wall of the interior.)Indeed,various

examplesin Cave4, Upper6 and22 make


it quiteclearthatduringthe lastfew years

of workat thesite,whenartistswerecalled
upon to add imagesto wall surfacesnever

previouslyplastered,they applied mud


plasteronly to the immediateareaof wall
with which they were concerned,sometimes adding a small amountto the ceiling

areadirectlyabove,whichthey decorated
as a kind of canopyfor the icon below.27
The idea of plasteringa whole wall or a
27

The well-known sculptured Litany scene


in the porch of Cave 4 is a case in point. Numerous other (and equally late) examples appear in
the interiorsof Cave Upper 6 and Cave 22.

PLATE 1

SPINK

FIG. 1.-AJANTA. CAVE 11. Plan. (Published in J. Burgess, Report on the BuddhistCave Temples,

Archaeological Survey of Western India, vol. 4, London, pl. XXVIII, no. 2.)

PLATE 2

SPINK

z-!v

FIG

2.-A

T_

igan>

PLATE

SPINK

M-

ft

w C i i -?!(s7'

_l

Ak.~~~~~~~~A'

. ...
..N

K,[e

..

-|

>g

4~~~~~~~~~~~~~~~~~.

':.^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~w
.J

_!
1_* _

:_A

I,

-i .....

P1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~J

..,.;'D

FIG. 3.-AJANTA.

CAVE

11. Porch. Ceiling, detail of central section (AAA, 20, 216).

tv)~~~~~~~~~~~~~~~~~~~~~~~~~~J
ku
o r +
k ~~~~~~~~~~~~~~~~~~~~~~N
-;~~~~~~~~~~~~~~~~~~~~~~~~~~F
"TOf-J
IT"NA-

> *

Si *::-

...._~~~~~~~~~~~~~~~~~~~~s.

? l . _% _ . I } j.
1E

10

..I

F....f.

1'~~~~~~~~~~~~~~~~~V

-~~~~~~~~~~~~~~~~~~~~~iz

S_k.
8rs1lI~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
21
l:
....

=s'L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
tStr-we==wS;ePtv
-~~~~~~~~~~~~~~~~~~~~~~~~A
A4 ! !,-i*W
,~~~~~~~~~~~~~~~~~~~~~~~~~~P
t ___>i
W

i-~~~~~~~~~~~v
;:' '

';

w }

y~~~~~~~~~~~~~~~

^ Tj_

-ijv<<w,i:de,
'P.

rA

FIG.4.-AJANTA. CAVE1 1. Porchdoor, with tracesof flankingguardiansat left and right (AAA, 15,

PLATE5

SPINK

g4,-Siw1
M_s
d. wo

s,aa*-n
zF;2

-'i

ffS)

^4

.J,

i'^'-gf

-w"-.t

-1
-_

__

_t __
,_

<i

_w

*-

=-_

_F>

w J.,

.......

v;*9

X/

*Ll

'h
i

jliS!

^4'...

'

zw

'aFi

-_

8ssm

-t$

eF

,.2

X F

*_s

_2e;#
w

'>

'

9.. .*
<.

__-

i3i

!_!
_

painteddetailsaboveleft
A.CAVE11. Porch,
FIG.S. AJANT.

paintedceiling
paintedfiguresupporting
window,showing
beam(AAA, 20, 214).
Z
Z

_7_afi

fi =

B= t
F

_r_-

...

_i_
sifef

wN

''
ss

ii

-lliw

is

S,,s

.#t,.

is

IX^

S. ,,prS.g.8+.8.

__

1 _

.A,.
F

tw

_lli
,,

S_

...t

<

R |-u
.0

oSiiS.

...

.:_.

]_j

By>?

_jij;_

sj

w]

x,.

!_!ro

_L

{'B
7Z05-d;

_F
R
v

?
'Ss

.s

'

.
wes

fi

.......
i

~
~~~~~~~

_is114s+*t<_

FIG

AJANTA CAVE

11 Interior View toward shrine (AAA 15 115)

tj~~~~~~~~~~~~~~~~~~~~~~~~~~~~a

--

FIG. 7.-

AJANTA.

CAVE11. Interior.Viewtowardshrine(AAA,15, 115).

PLATE

SPINK

CAVE 1 1. Interior. Right half of right wall, showing


traces of a "Thousand Buddhas" (photography courtesy of the
Government of India, Department of Archaeology, no. 55-1-307
[UNESCO Series]).

FIG. 8.-AJANTA.

tE E

____E__

. 1-l _

8,1AA.20 fg.

1_
4i

~~~~~~~~~~~~~~~~~~I-AJ
FIG.~8.AANA~~

11 nInterior. CRighthal
CAVE
~~~~~~~~'

showing small seatedBuddhas (AAA 20 225)

of

Tntght
wall showing

224)

SPINK

PLATE

.'a..

-~~~~~~~~~~~~~~~~~~~~~~~~~2
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..
...

r:,,,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.......
........

:]

l~~~~~~~~~!

PLATE 9

SPINK

FIG.~

AJATA

neirLftwl

CAE1

ewe

is

13.

atedatan

FIG.' 14AAT

hon

etis

AE1.Itro.Lf

tte

lef (A,

al

20 234)

ewe

or
ealo
Bdh
ru,soigbr'
adF seodcl
hea motif on thrnbc
(AAA 20 54)

is

PLATE 1 0

SPINK

_-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

FIG. 15^.-AAT.
sadn
Budh

CAVE
gru

. Inero.

at lef

an

Rea

wal,

.Sravasti

>

;I

aS

t-

i.

at righ
,ue"S

_-~~~~~~=______

(AA

E - | - l l _

ruinouIs
1

121_

_~~~~~~~~~~~~~~T

W
e !! s j!
F_~~~~~~~~~~~~~~~~~~Ste

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~_
4O~;

_5

_Ei

shoin

leto.hrnor
Miracl

4;1i-W)

_:
~~~~'io

PLATE 11

SPINK

CAVE 1 1. Interior. Rear wall, showing Buddhas at right of shrine door


see upper left portion of fig. 16 [AAA, 549-66]).
location,
(for

FIG. 1 7.-AJANTA.

w~~~~~~~~~~~Fig

Figure 17

Figure 18
See fig. 3,
Dhavalikar,
this issue

location,see
f
orlefiores
Diagram

18 a

196within figr

1).

SPINK

PLATE12

_*;ss-u

_"

__1W-z_XieRi

F _
t

ws

_s
_

xt

v=|b1|
I bes''4K}\M

Fw

[9

v4

;s1z_

m1.N?

i.

S-<<8188i7s

w. $ ..XQi,Z}t | F saUW.

>

S .
+...aY
*

ENt's.s

sX

[;

M__|_

*t
i [4E9'w EW
s
;._

.<.*'S-

^te;

..iSos

_gS

4.-4^

EN; l

2 :5DS ID_
;

l_lEi.,_

_5

11

VW

;;si lb| _m_

<''"8Yt,,rt.,*

i9e

3nl

:''',

,@t ',
Mi ,',i'',i,ts

t"

8t

65wxkK<l

/-

u___
p>

t(0_

;_

FIG.18. AJANTA.
CAVE11. Interior.Rearwall.Detailof

figure17,showingappearance
of thefabricuponwhich
paintingsin thisareaaredone(AAA, 20, 220 [detail]).

_*

_l

'1_
_s

_
_

_.

,_

..,_
*s_

. ___

__
___

F _u

_^.

l__;

Iw

:_

_r

..w

\
.s.

.d
.'

^'
".

s
v

iS

. .S.]
^
t_

__

=..]

_il

tS "

:';

t,

41bX

':

__
--__sS

,;

_,

___

:S

*_.''

*t[.

'{

'

s
]

St

_X_

|_.e

w:
......................

'_
.

,<.

2-

:_

.'4"
*

i''

,s

::U#:^
_

4-

':.
s,

'.

S!

:'

_s--g
;

,-Si.,=t*s1

s4

_*
:.g_

iSS E_
@8._
.:2..

s__;

5_.S.

.>^S^w_

.......

_2_

PLATE13

SPINK

'^

~~~~~':

CAE1 1. Ineir

FIG 20-JNA

:.;] ~

(AA

.,...S.-..

..

w:S

n
.....
AO

MF

%,

=4

b
'ib~~~Tj

1|

Shin do__or_

11, 210).

E_rWLN

SPINK

PLATE14

~~t

0
i

_t

'

t$St;00 0'*i
t .

*
.~~~~~4

;*

>.;stP4

;;

W i~~~~~~~~~~~~

4r...-'

<

.Usw,,;

' M

-s

?E :
.

e.

2~~~~~~

3~~~~~~~~~~~~~~M.

FIG.22. AJANTA.CAVE11. Interior.Shrine,showingstupa behind


the Buddha'sthroneand evidenceof repaintingon halo and underarm
(AAA, 20, 239).

W_
.-,A_.

te.

0zg

.ww

AJANTA'S CHRONOLOGY: THE PROBLEM OF CAVE II

whole room first, before startingto paint,

wasa thingof thepast.


Another factor may also support the

assertionthattheplasteringof the interior


was doneconsiderably
priorto the painting. As mentioned above, a considerable
portion of the plaster on the interiorwall
has (or had) a surfacing of cloth, over
which the paintings were applied. I hesitate to explain this curiosity in any dogmaticway, but it would seemthat the cloth
was applied in order to improve or repair
certainareasof the mud plasteredsurface
prior to their being decorated.If this is so,
it would lead us to infer that some years
had passed since the plaster was applied,
duringwhich time the surfaceof the plaster had becomeabraded.
The rearwall of the interiorhad a considerableamount of fabric on it, much of
which is still intact, althoughin extremely
fragile condition, and peeling off (figs. i8
and I9). On the left wall between the second and third cell doorwaysa whole piece
of this fabricwhich eitherhad beenpressed
into the wetted surfaceof the mud plaster
or (morelikely) had beenglued againstthe
dry mud plasterhas becomedetachedleaving a distinct impressionof its shape as
well as its weave on the remainingmud
plasterbeneath(fig. I2). The fact that this
appears to have been the only piece of
cloth applied to the whole left wall and
the fact that its shape was so irregular
stronglysuggeststhat it was usedas a patch
(i.e., for repair) and applied rather carelessly. It is hardto believethat it (or similar
areas on the rear wall) could have been a
painted banner;not only is it irregularly
applied but its edges appear to have been
very frayed-as if it had been ripped to
the requisitesize before application. Fur-

I67

thermore,
images(nowbarelyvisible)seem
to have been paintedon the wall at this
point,butwithno referenceto thepieceof
cloth in question;their upperportionsap-

pear to have projectedabove the cloth's


top edge.
It seemsclearthat suchclothsmerely
formedpart of the baseupon which the
paintingsweredone.Apparently,oncethe
clothswereglued(?)ontothemudplaster,
the thin groundof pigmentupon which
the paintingswere renderedwas applied
uniformlyto the wall surfaces,and the
paintingswere then done without reference to the position of the cloth "patches."

For this reasonit is extremelydifficultto


see,when we turnto the rearwall of the
cave, just wherethe cloth areasend, for
both cloth-basedareas and mud-based
areashave a similarcoatingof ground,
color, and (unfortunately)thick obscuring
grime.In fact it is quitepossibleto observe
the cloth-basedsurfacesintensely,without
realizing the paintings are actually done
on this kind of background.28
It is only
when one realizes that large areas of the
paintings have actually become detached
or raisedup from the mud plasterand yet
retainenoughstrengthto keep from crumbling away that one is led to look closer
for the explanation,namely that they are
sustainedby the still preserved(although
very dry and fragile)fabric.
Another suggestionthat the cloths representsecond-phaserepairsto the plaster
rather than first-phase applications is to
As far as I know,no noticehas everbeen
recordedof this phenomenonin Cave i i, althoughpartsof the rearwall surfacehave been
restored.The restorersmusthave beenawareof
the situation,but perhapsdid not realizehow
interestingandunusualit is.
28

WALTER SPINK

i68

be seenin the fact thatno clothsappearto


have beenappliedto the frontwall or to
theceilingof theinterior,neitherof which
areaswerepaintedin thesecondphase.Also no clothsappearunderthe first-phase
paintingsof theporch.
The decorationof the shrinedoorway
(fig. 20)
is probablycontemporarywith
the other paintingsof the interior.The
factthatit doesnot havethefiguralmotifs
whichone mightexpectat thistimecould
be explainedby its beinga ratherhastily
conceivedand painted conception.The
painterappearsto havemimickedthe decorative formsof the earlierporch door,
whichprovideda readypattern,although
his use of rosetted"dividers"may reflect
the influenceof somewhatlaterdoorway
designs.29

Theanomalouscharacter
of Cave i i is
of coursecompounded(and reaffirmed)
by thetreatmentof thesculptured
Buddha
in the shrine(fig.21). It is clearthat the
image was satisfactorilycompletedand
paintedat one point in time, and then at
anotherpointin timewas "restored."
The
halodistinctlyshowstwo phasesof painting-for anearlierandmorecarefuldesign
is revealedby the flakingoff of portions
of the second layer of decoration(fig.

Whetherthe thronebasewas simi-

22).3

It is conceivable that the shrine doorway


was decorated during the first phase when the
main Buddha image, upon which it enters, was
painted.A comparativestudyof the pigmentsused
mighthelp to resolvethis question.
30 This repaintingis very obvious in on-thespot examinationbut is hardto seein photographs.
29

larly restoredis problematicbecauseno


such overlay now remainson the nicely
executeddesignswhichdecoratethat portion(fig.23). However,theBuddhaimage
itself was surprisinglytreatedin this second phaseof work. Its proportionswere
actuallyalteredby the applicationof a
thick layerof mud plasterto portionsof
the figure (notably the waist) (fig. 22).

Upon this newly built up surfacea new


layerof paintwasapplied.
Thereasonfor thischangein theactual
shapeof theBuddhafigureis problematic.
Perhapsthe changeonly represents
a personalpreferenceon the part of the artist
or patrons.On the other hand, perhaps
canonshadchangedto thisdegreeoverthe
courseof a decadeor two and the image
wasbrought"upto date"whentheinterior
was painted.If the latter shouldbe the
case,suchcanonicaldistinctionsare elusive
ones to the modern devotee's eye, at least
at this point in our investigations.What is
more pertinentto suggestat presentis that
the two phases of work on the Buddha
image would seem to correlate with the
early and late phasesof work on the cave
as a whole and thus with the earliestand
latest phases of Mahayana patronage at
the site itself.We have suggestedpreviously
that patronageactivity at the site started
about A.D. 465 and ended about A.D.
5oo. The tentative late fifth-century dating

of the newly discoveredCave i i inscription would support this hypothesis since


the inscriptionis directly associatedwith
paintingsof a characteristicallylate type.

You might also like