iTOWARD DEFINITION:-; :
;OF TIlE NAZC              STYL E E:

A. L KROEBER

UNIVER$ITY OF CAgIIFOENIA PUBLICATIOS$IlI AMERICAN' J

A4C,,:HAOtQLGY, AND  NOU.G..
Volume 43, No.4, pp. 3273-42, plates 3A6, 12 fires in t.xr

tNIV~RSITY-9J? CALIFORNAIA PE$S

BEjkEkLEY AND LQ,S AkNGER
N ~~~~11956

K                N,  !       N-N-00ss;.  ;0000

N                     .-                     N- e  <ri ]'

.^  :  CE K L Y04  L SAN EFe  -: t-N:f '  -  <'- ;N  -  -

TOWARD DEFINITION
OF THE NAZCA STYLE

BY

A. L. KROEBER

UNIVERSITY OF CALIFORNIA PRESS

BERKELEY AND LOS ANGELES

1956

UNIVERSITY OF CALIFORNIA PUTBLICATIONS IN AMERICAN ARCHAEOLOGY AND ETHNOLOGY
EDITORS (BERKEY): T. D. McCowN, E. W. GIrFoRD, D. G. MANDELBAum, R. L. OLSON

Volume 43, No. 4, pp. 327-432, platep 31-46, 12 figures in text

Submitted by editors March 28, 1955

Issued April 6, 1956

Price, $2.00

UNIVERSITY OF CALIFORNIA PRESS

BERKELEY AND LOS ANGELES

CALIFORNIA

-0-

CAMBRRIDGE UNIVERSITY PRESS

LONDON, ENGLAND

PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS

PART I: ERRORS AND OMISSIONS IN RETROSPECT

PAGE

The Gayton-Kroeber Approach of 1927 .........    ....................... 327
General Weaknesses ................................................. 328
Comparison with Ocucaje ............................. 329
Statistical Analysis .................................................. 330
First Doubts ............................. 331
Qualifications on the 1924 Ocucaje Presentation .332
Bearing on the Ocucaje Collection of the 1926 Excavation for Chicago . 334
Resemblances in Design .335
Provincial Relations of Ica and Nazea Valleys .336
Relations of Phases within the Nazea Style .337
Partial Listing of Characteristic B-Phase Designs .338
Summary and Caution .338

PART II: REvISED CLASSIFICATION OF NAZCA STYLE VESSEL SHAPES

Purpose ............................................................ 340

Goblets and Vases .............       .......................... 340

Goblet I ....................................... 341
Goblet L ...........         ............................... 342
Goblet K ....................................... 343
Goblet J and vases M-P .344
Waisted goblet J .344

Tapering vase M, formerly "small" vase  .345
N, Cylindrical vase .346
O and P, Bulbous-convex and bulbous-concave vases .347
Jars .348

Jars R, S, T .348

Lipless jar Q .351
Other Shapes .353

U, Double-spout .353

V, Head-and-spout jar .354
W, Flaring rim vase .355
X, Figure vases .355
Y, Head jars .356

Bowls and Plates .357

Plates: classes Ac, Ar, Aw .358
Bowls .362

Straight-sided bowls, shape B: gambreled bowl .362

riii]

iv                               Contents

Shape E: flat non-flaring bowl .................................... 364
Concavely flaring bowls: Cl, C2, residual from old class C ..... ....... 366
Shape D: flat-bottomed flaring bowl ......... ....................... 366
High flaring bowls of shapes F and H ........  ...................... 367
Summary of concavely flaring bowl classes .......  .................. 370
Bowl shapes G and GG .    .......................................... 370
Residual: unclassified bowls ............ .......................... 373
The Nazea Y Style ................................... 373

True Nazea Y, formerly "Y1" . ................................... 374
Coast Tiahuanacoid ("Y2") .  ................................... 374
Dissolution of Y3 ................................... 376
Summary of Reclassification of Nazea Shapes ............................ 377

Some Nazea-Y shapes .............................................. 381
Grouping of classes by style phases .................................. 381

PART III: DISCUSSION OF VESSELS ILLUSTRATED

Introduction     .................           .................. 382
Plate 31 Ocucaje: Phase A. Graves Al, B4, 5 ............. ............. 382
Plate 32 Ocucaje: Phase A. Graves Cb, F4, 8, 18, 20, 22. B and E

Bowls, A Plate     ...............         ................. 385
Plate 33 Ocucaje: Phase A. Bowls from Graves F4, 6, 14, 18 ..... ......... 387
Plate 34 Ocucaje: Phase A. Various Forms from F2, 3, 6, 10, 13 .......... 388
Plate 35 Ocucaje: Phase A-plus. Graves F3, 9 ........................... 390
Supplementary on Ocucaje. Content of Graves Fl, 4, 7, 12, 17, Bi, 3 ........ 391
Plate 36 Nazea. Interior-Painted Plates of Class Ac ...... ............... 393
Plate 37 Nazea. Low Bowl Types B, D, E ............................... 394
Plate 38 Nazea. Various Bowls, Plates, Goblet, Jar. All B or Late B

except Jar h .................................................... 395
Plate 39 Nazea. Goblets, Shapes J, K, L. Phases B and Late B ....... ...... 397
Plate 40 Nazea. Goblet and Vases. Full and Late B Phases ..... .......... 399
Plate 41 Nazea. F-H High Bowls with Painting Extending over

Bottom ....................................................... 401
Plate 42 Nazea. F-H High Bowls with Cat-Deity Design .....   ............ 403
Plate 43 Nazea. F-H High Bowls with Cat-Deity or Snake Designs ........ 405
Plate 44 Nazea. F-H High Bowls with Bird Designs ............... ...... 406
Plate 45 Nazea. F-H High Bowls with Repetitive Pattern .....  ........... 409
Plate 46 Nazea. F-H High Bowls and Cat Jar .......     ................... 410
Acknowledgments     ............................................ ...... 412
Bibliography .........   .......................................... 413
Plates ........ 417

TOWARD DEFINITION OF THE NAZCA STYLE

BY

A. L. KROEBER

PART I: ERRORS AND OMISSIONS IN RETROSPECT

THE GAYTON-KROEBER APPROACH OF 1927

FOR TWENTY-FIVE YEARS there has been agreement that development took place
within the Nazea (Proto-Nazea) style of ancient Peru. There has also been agree-
ment-except for Tello who inverted the order-that a simpler phase of the style
preceded a more elaborate one, which in turn ended in a slovenly breakdown.

But for years the only effort to specify, illustrate, and validate these phases was
the brief monograph of 1927 by Gayton and myself. This was based on some 650
pottery vessels secured by Uhle in Nazea or Rio Grande Valley for the University
of California. Many of these pots had general locality attributions, such as
hacienda names; scarcely a dozen were ascribed to specific graves. As a whole,
the 650 vessels had been purchased, either from huaqueros or from local collectors,
and were thus without data as to associations.

Gayton and I devised a method that might roughly seriate these hundreds of
vessels; or rather, Gayton devised and I accepted and approved the method. We
divided the collection into 26 classes of shapes, and selected by inspection forty
painted designs-with names largely translated from Seler (1923)-and then
tabulated the co-occurrence of shapes and designs. The co-occurrences showed a
definite trend, in that they lumped or clotted along a diagonal. The direction of
this trend we inferred from variant forms of the same design of which one seemed
"naturally" derivable from the other, but the reverse was not.

In brief, we established 4 classes. We first set apart 94 vessels from Nazea that
were wholly or partly outside the Nazea style. They mostly retained some features
derivative from it, but these were mixed with non-Nazea traits, in some cases with
Tiahuanaco traits. The core of this group we called "Y." Of the remaining 563
vessels which were within the Nazea manner, 27 had to be omitted from the co-
occurrence correlation of shape and design because of their unique or rare shapes.
This left 536 vessels that fell into 25 shape classes, and on which the forty designs
selected appeared 704 times-an average of 17 to 18 times each; or, about one and
a third designs per vessel. In this main group, indubitably of Nazea style, we
recognized three phases, of which we called the first and third, the most distinctive,
"A" and "B," and the second or transitional one, "AB" or "X."

In summary, 5 A, 9 X, 11 B shapes correlated thus with 11 A, 14 X, 15 B de-
signs-or rather, predominantly A, X, B shapes correlated with predominantly
A, X, B designs as follows:

SHAPE AND DESIGN CORRELATIONS, 1927

A designs        X designs        B designs
Shapes                       .

No.    Per cent  No.    Per cent  No.   Per cent

A ....................   76      47       74      45      13        8
X ....................   28      19       83      56      36       25
B .                       7       2      181      46      206      52

[327]

328      University of California Publications in Am. Arch. and Ethn.

This is not too close a correlation, but a trend is definite. A tight correlation
could hardly be expected within what was obviously one style. Either shapes or
designs can, theoretically, persist into periods later than the one in which they
originate; or they can appear before the period in which they become most char-
acteristic.

However, our results have met with only a lukewarm reception among Peruvian-
ists. They have not been subjected to outright attack, but neither have they been
outrightly accepted. The A, B, and Y designations have had some usage; but
probably not so much as a result of our correlation tables as because they obtrude
to the eye, on a little acquaintance with wares from Nazea; and there were no other
names lying around, except Tello's "Pre-Nazea" for B.

Meanwhile, the present is a time of renewed interest in the Nazea style, or styles.
Four or five years ago, Junius Bird began to work over the American Museum col-
lection and came tentatively to a new classification and scheme of probable devel-
opment, still unpublished. In 1952, Duncan Strong made an important expedition
to Nazea with stratigraphic excavations. In the same year, Donald Collier and I
undertook to work over the wholly unpublished collections in Chicago Natural
History Museum which I dug for them in 1926. And meanwhile, in Berkeley, John
Rowe and Lawrence Dawson were once more laying out and arranging the Uhle
Nazea collection to see whether it could not be made to yield more convincing
results than Gayton and I had extracted from it. I was also drawn back to concern
with Nazea by a paper I wrote originally for the 1951 International Congress of
Peruvianists in Lima, and which deals with a problem raised by Gordon Willey
(1951) as to the relations of Chavin and Paracas, but which came to involve Nazea
as well (Kroeber, 1953).

Reexamination of the now more than twenty-five-year-old memoir by Gayton
and myself has accordingly been forced by the times. Our findings then will in-
evitably be corrected soon by new data and new studies, and our deficiencies or
limitations may just as well be pointed out and cleared away now.'

GENERAL WEAKNESSES

I specify, therefore, the principal weaknesses I see, as of 1953, in the Gayton-
Kroeber scheme of 1927.

First, the correlating of shapes and designs was sound in principle, but there
was no need to limit designs to forty, especially as the basis of this sampling was
not defined.

Second, some of our tabulations and pictures correlate rather badly. For
instance, on shape U, the familiar double-spout, which we class as an A shape, we
tabulate 17 occurrences of A designs, 40 of X, and 5 of B. The obvious question is,
why then is the double-spout an A shape? Then, we illustrate 11 such vessels on
plates 1 and 2. Four of these-la, lb, 2d, 2e-are flatter than the rest. In the Lima,

1 The two earlier papers discussed here have plates numbered as follows: Gayton and Kroeber,
1927, ple. 1-21; Kroeber and Strong, 1924, pls. 25-40, of which only pls. 25-29 are re-
ferred to here. In text references to the plates the dates of the papers are omitted since the
plate number alone is sufficient to indicate the source; thus:

1-21 ..           Gayton and Kroeber, 1927
25-29 ..K roeber and Strong, 1924
31-46 ..          the present paper.

Kroeber: Toward Definition of the Nazca Style        329
New York, Chicago, and Berlin collections, lenticular, angled, and cylindrical-
bodied double-spouts often carry conspicuous X and B designs. There are such at
Berkeley too-we happen not to illustrate them. Even on our shown heart-shaped
or rounded double-spouts there are post-A designs-see lc, 2d. It appears that we
should have recognized shapes U1, U2, U3, possibly U4 and U. Further, the cat
demon painted on la is very close to our type cat demon, number 15 of figure 3 in
1927, but it is doubtful whether the cat of lc, le, lf can be offhand equated with
number 15 of figure 3. The idea or personage may be the same; the form differs
at many points. As a matter of fact, this number-15 cat demon is listed as an "X"
design-which was perhaps the only thing to do, since it occurs on 30 examples of
our A shapes, on 20 X, and on 22 B. All in all, while the totals in our tabulations
do carry a certain gross or over-all weight, the illustrations indicate that the piece-
by-piece situation may be much less simple.

We were of course aware that A double-spouts and X-B double-spouts differed.
On page 13 we say, in the description of substyle A: "The typical double-spout jars
have spherical or ovoid bodies; the spouts are short and about parallel (pls. lc, f,
2a, c, e). Those with X or B design have a more lenticular form; the spouts tend
to be longer and divergent (pls. ld, 2d)." The illustrations cited conform only
partly with the ovoid-lenticular distinction, and are based mainly on whether the
spouts run parallel or diverging (though even here lf and 2c diverge, although
attributed to A). At any rate, it is unclear or confusing to label a plate, "Nazea A
Style," and then discuss X or B style designs on some of the vessels shown. I think
we were ridden by the principle of our correlation that A shapes and A designs
tend to go together. The double-spout is an A shape and ought to have mostly A
designs on it (though by count it does not have them), so we felt justified in treat-
ing all double-spouts as of substyle A, even when the design was not A and the
shape was aberrant. As an operational procedure this is certainly vulnerable, even
if we were mostly right in our classification.

COMPARISON WITH OCUCAJE

On pages 11-13 we made a comparison with Uhle's Nazea collection from Ocucaje
in lower Ica Valley,' in order to test the validity of our segregation into the three
substyles A, X, B. We found for Ocucaje:

Vessels                        With Designs

98in5Ashapes ........................  55A  32X   4B
10 in 2 X shapes ........................  6 A  4 X  2 B
3in2Bshapes ........................  OA  2X     2B

From this we concluded that the shape-design correlation was at least as high at
Ocucaje as at Nazea; and second, that the predominance there of one substyle
(which happened to be A), was expectable because Ocucaje was a single locality
and Nazca Valley contained many localities. These inferences still seem probably
valid, except for the fact that any inaccurate or spurious correlations resulting
from limitations of the technique applied to Nazea would presumably be repeated
when applied to Ocucaje.

2 Described in Kroeber and Strong, 1924.

330      University of California Publications in Am. Arch. and Ethn.

We were on the right track in seeing the smaller, more uniform, and mainly
grave-lotted Ocucaje collection as having significance for the interpretation of
the more varied and more widely derived Nazea assemblage. But we could have
done better than subjecting it to a parallel correlation test, by assuming that its
obvious uniformity expressed the fact of its being a "pure style" or "pure style
phase" collection-as was indeed made more likely by its having been excavated
in four small cemeteries within the limited oasis of Ocucaje. Then we could have
used this Ocucaje pure style lot as a lever in analysis of the larger and more varied
Nazea assemblage. We might simply have begun by separating off those pieces
from Nazea that were similar in design and shape to the Ocucaje ones. Another
group would have consisted of vessels partly different from Ocucaje but also
partly similar, and therefore presumably adjacent to them in period. One or more
other groups would have appeared with still more differences from Ocucaje-
differences in new shapes or designs, in elaboration of old elements, or in their
slovening and degeneracy, though still within the general limits of the same style.
And finally there would be the vessels emphasizing an array of features which
everyone would agree could no longer properly be classed as within the Nazea
style. After this stylistic analysis, starting from the fixed base or "zero point" of
the pure Ocucaje phase, had been made, our correlation technique would have
been more fruitfully and convincingly applied, I am now sure. We would have
been driven to realize that the fact of Nazea double-spouts with non-Ocucaje
shape, was significant and had to be taken into account in a stylistic study; and
we would accordingly have distinguished the double-spouts into operational U,,
U2, U3 forms, instead of trying to operate with an undifferentiated U definable
only as an idea or principle but variable in actual form. And we would not have
drifted into awkwardnesses of classification such as designating double-spout
shapes as outright A phase while two-thirds of their designs were X phase.

The general moral seems to be that in attacking a stylistic problem-and since
our Nazea collection in Berkeley was without excavatory data, we had before us
nothing but stylistic materials-the soundest and especially the first procedure is
strictly stylistic analysis of stylistic identities and variations. Such an analysis
is concrete and can be cut as fine as is profitable. But we partly deviated from this
principle, and failed to be as successful as we could have been in that: (1) we
partly substituted concepts of shapes for observable forms of them; (2) in designs
we operated with only a minority of those actually occurring; and (3) we applied
correlation technique before our stylistic observing and drinking-in had been
carried far enough. That our results are, in the gross, largely right, I lay to the
fact that we did bring to bear a reasonable amount of fairly perceptive stylistic
experience and insight-more than we were conscious of, or at least avowed, in
our desire to offer a statistical proof.

STATISTICAL ANALYSIS

I continue to believe that statistical correlations can be used in stylistic analysis,
but that their main function is to convince those who are, by nature or by inex-
perience, stylistically insensitive. Stylistic qualities and their patterns and inter-
relations must be taken in through the senses and digested through the subrational

Kroeber: Toward Definition of the Nazca Style            331
process sometimes called intuition, but for which the term "perception with
aesthetic feeling" is adequate. A mathematical approach, being abstract and
rational, seems best deferred until the pioneering job of analysis has been pretty
thoroughly done by exercise of perception. On the other hand, final proof, in the
sense of formal scientific proof, has apparently to be quantitative. I believe that
most qualitative data, and perhaps even stylistic ones, can also be interpreted
quantitatively in the end; but certainly not so profitably at the beginning of
inquiry.

For the quantitative emphasis of our Nazea treatment I must certainly assume
my full share of responsibility. The period was one in which interest was turning
to the quantitative formulation of culture relationships. In 1925, Strong validated
his distinction of four periods of Ancon pottery by a statistical correlation. In
1926 there appeared in the Anthropologist the Clements-Schenck-Brown paper
applying the Chi-square approach to intra-Polynesian relationships. In 1931
Clements, again in the Anthropologist, analyzed Spier's Sun Dance data with the
coefficients of correlation and association; in 1932 Driver and I published Quan-
titative Expression of Cultural Relationships. In 1935, Klimek initiated the Cul-
ture Element Distribution series in the California series of American Archaeology
and Ethnology, soon transferred to Anthropological Records. In 1937, Chretien
and I issued our Quantitative Classification of Indo-European Languages. In all
these efforts I had at least an indirect hand or association. And of course I have
never repudiated as such the quantitative approach to cultural phenomena-and
do not in the least recant it now. It is only that with experience one learns better
the points at which the method becomes respectively less feasible or more profitable

and secure.                     FIRST DOUBTS

All this was far less clear in anthropology in 1925-1926 when our Nazea paper was
written and submitted for publication, first to a board of editors, and then, in
May, 1926, to the University Editorial Committee.

I began to feel minor dissatisfactions with our results even while the paper was
in press-its appearance was delayed to February, 1927-while revisiting the
Nazea collections in the Museo de Arqueologia Peruana during my June to Novem-
ber trip to Peru in 1926.

One of my notebooks there contains nine pages of sketches of Nazea shapes and
designs or design elements classified into "A" and "B"-some 50 shapes and more
than one hundred designs-and, I think now, somewhat more soundly classified
than in our more limited and rigid published schema.

Through the years, as I saw and resaw Nazea pottery, and discussed it with
Peruvianists, I became more and more aware of the imperfections of our joint
effort of a quarter-century ago. Especially did I come to feel that instead of com-
paring Ocucaje with Nazea in validation, we should have used it as the touchstone
to initiate analysis.3

"Already by 1930 the wish evidently had become occasional father of a misremembering that
we had done so. In "Cultural Relations between North and South America," in the Proceedings
of the 23d International Americanist Congress, I find myself saying: "Calling the homogeneous
lot from (Ocucaje) Ica, Nazea A, Gayton and I were able to segregate the mixed lot from Nazea
Valley into style Nazea A and Nazea B ... Gayton was able to demonstrate the objective validity
of each of the two styles by a statistical correlation." (X is ignored.) This certainly seems to

332      University of California Publications in Am. Arch. and Ethn.

Now, with Donald Collier and myself co-authors in the preparation of a report
on the Nazea collection which I excavated for Chicago in 1926, with the aid of a
splendid series of photographs available in my absence from Chicago, it is ob-
viously incumbent on me to clear away the differences from the quarter-century-
old co-authorship of Gayton and myself. Hence the present setting forth, which
is part recantation and part reconcilement.

QUALIFICATIONS ON THE 1924 OCUCAJE PRESENTATION

There is another venture into Nazea-my first one-on which I have to offer some
present-day qualifications: my and Strong's Uhle Pottery Collections from Ica,
of 1924. The Nazea ceramics described in that monograph are mainly the lot from
Ocucaje in Ica Valley already referred to. Our task in this case was concerned not
primarily with the Nazea style, but with all the seven styles and style phases
recovered by Uhle in various parts of Iea Valley. Our treatment therefore was
discriminatory between styles-such as Nazea, Epigonal, Middle and Late Ica-
not between phases within the Nazea style. See the tables on pages 101 and 103,
which contrast Nazea with the others. It must be remembered that it was Ica-in
fact at the lower Ica river oasis of Ocucaje-in which Uhle first found Nazea
ceramics in situ, in graves, in 1900-1901. Only collectors' trade pieces had pre-
viously come into scientific museums; and Uhle in his first reports still called
Nazea "the new-found style of Ica."

Much the larger part of Uhle's Nazea finds in Ica Valley were made in four small
cemeteries at Ocucaje; and they happened to be nearly uniform in their substyle
or phase. Strong and I did include in our plates and counts a few somewhat aber-
rant Ocucaje pieces-about which more in a moment-though we also did pru-
dently separate off as "Proto-Nazcoid" a lot from Uhle's grave- or cemetery-
Ocucaje A.

But we erred more widely when we included with the nearly homogeneous
Ocucaje assemblage some pieces discovered by Uhle in his site "e" at Santiago
in the main part of Ica Valley upstream from Ocucaje. In this general area he
found chiefly Middle and Late Ica remains-his little Epigonal collection was
again from Ocucaje-but at Santiago site "e" he dug up a series of broken vessels,
in previously disturbed grave soil. Of these, several that had been reconstituted
by 1924 were included by us in our description and pictures, even though they
were manifestly not of quite the same phase of Nazea style as the run of the ware at
Ocucaje. For a preliminary or summary treatment, this slip may be forgiven; but
thirty years later, it needs correction.

The Santiago or site "S" pieces illustrated but definitely requiring to be
separated off from the Ocucaje collection are: Kroeber and Strong, 1924, plate
28a, k, 1 from Se and plate 28d from Sa. A glance at these, especially the first
three, validates their differentiation. Their very shapes are without parallel in the
rest of the plates. As long ago as 1926, in one of my books of memoranda made in
imply, erroneously, that we began with the Ocucaje touchstone and validated by correlation, in-
stead of the reverse.-Style phase A is characterized in this passage as restrained, somber, and
severe; B as bright, flamboyant, proliferating, rapid, flowing; Y as being simple to the point of
degeneracy, but also still containing some flamboyances, extensions, flowing quality-in short,
being an abbreviated, cursive B. (P. 8 of the article.)

Kroeb er: Toward Definition of the Nazca Style        333
Peru, I noted that the 3 Se pieces must be excluded from the Ocucaje Nazea
collection.

Next, there are the few aberrants within Ocucaje. The collection from Ocucaje
was assembled from one grave at site A, 5 at B, 2 at C, 23 at F. The contents of
these thirty-one graves are uniformly homogeneous, with the following exceptions.

1. The vessels from graves F3 and F9, in plate 27b, h, n, x, 28e, o, conform in
part to the other F graves, but also contain pieces done in a different phase of the
style. (See pl. 35 of the present monograph.)

2. The jar of plate 28p, being from "near grave F6," not from within it, cannot
be used to help prove a style, though I believe it belongs in Ocucaje style A.

3. For a time I was inclined to doubt graves Ca and Cb, plates 27e, v, 28i, but
no longer see reason for their disqualification.

4. The small site A which I pointed out as puzzling in my Peruvian Archaeology
in 1942.' Strong and I had shown 5 pieces from it and called them "Proto-Nazcoid."'
Returning from Peru in 1942, and realizing as we had not realized in 1924 that 29f
was negatively painted and very similar to Tello's Paracas Cavernas negative
bowls, I went with great expectations to reexamine the 15 catalogue numbers from
Ocucaje A, only to find them heterogeneous (or indeterminate) in style, and so
regretfully wrote site A off as mixed, owing to burials made at different periods
or perhaps to previous pot-hunting; so that I wrote: "Uhle did not segregate his
[A] findings by graves." This, fortunately, now turns out to be an error; and 8 of
the 15 numbers from A are all from one grave, as my colleagues Rowe and Dawson
discovered, and pointed out to me, by going back to Uhle's manuscript field report
preserved in the Museum. In this, Uhle says: "4489-4494 and 4496 and 4789 [sic,
correct] are all from the grave which contained the mummy 328." This mummy,
catalogued in the skeletal collection as 12-2835, Rowe has opened and will report
on. Uhle's statement, which differentiates the 8 specimens cited as constituting a
grave-lot (Al, it might be called), from the 7 outside the grave, failed to be copied
into the Museum catalogue, which was an error of oversight under our rules of
entering everything that pertained to grave provenience and association. What-
ever is of late period from A, like the 29e bowl, is now shown to be from outside the
grave. The assured grave contents prove all to be Nazca A or earlier. In fact, the
grave evidently dates from the Paracas-Nazea transition. The solidly black bowl
(4492, present fig. 7b) I mentioned in 1944 as "possibly Paracas Cavernas."
The negative bowl 29f (4493) has precise Paracas parallels, and 4494 (present
fig. 5a) is similar and also negative. On the other hand, 4489, now 73864 in Pea-
body Museum at Harvard, (illustrated below as pl. 31a), an E-shape bowl
painted with hummingbirds, is straight Nazea A. The unpainted, incised, rough
bird-jar 29d, 4496, has Nazea analogues from my 1926 excavations. Bowls 29c and
g, 4490 and 4489, have Nazea colors and shapes, but are certainly somewhat un-
usual in pattern, as Strong and I felt them to be in 1924. Finally 4491, shown below
as figure 7a, is a B-shape bowl, somewhat abnormally high and round-bottomed,

4Kroeber, 1944, p. 36; also p. 117, App. A.

P1. 29c-g. Of these 5 pieces, c, d, f, g are from the grave at A; the fifth, 29e, 4495, is a
sort of "cumbrous bowl" of post-Nazca period but is specified as not from the grave. The grave
contained 8 ceramic pieces: the 4 not shown in 29c-d, f-g, and the total assemblage are dis-
cussed in Pt. III in the text section referring to pl. 31 of this paper.

334      University of California Publications in Am. Arch. and Ethn.

with a zigzag or facing triangles design more geometric than Nazea A tends to run,
but within its range. In short the 8 vessels of Ocucaje grave Al are partly regular
Nazea A, partly Nazea A but near-unique in some feature, and partly blackware
and negative ware as it occurs in Paracas Cavernas-but the Chavinoid strain in
Paracas is absent !'

In comparison with later or mixed Nazea sites, Ocucaje Al can therefore be
counted with the Nazea A material from Ocucaje sites B, C, F. But in internal
analysis of Nazea A, it must be kept apart as evidently somewhat earlier. If the
pottery from Ocucaje B, C, F, Cahuachi A, Aja B is reckoned as Nazea A, the 8
pieces from the grave in A look like incipient Nazea A.

After deduction of the foregoing, and of a few other vessels near rather than
actually in graves, there remain about 111 ceramic pieces or fragments in the pure-
style Ocucaje grave-lot collection, from twenty-eight graves. Of these, Strong and
I illustrated 48. This is a number sufficiently large to define a style by, especially
as shapes are definite and nearly all pieces were painted with designs. It is these
100-odd vessels that I now feel Gayton and I might profitably have used as the
initial comparand in analyzing the 500-odd vessels from Nazea.

BEARING ON THE OCUCAJE COLLECTION OF THE 1926 EXCAVATION FOR CHICAGO

As soon as I began recently to see photographs of the ceramics in my own Nazea
excavations for Chicago Natural History Museum, nearly twenty-seven years after
I had dug them up, it became evident that in Nazea also there were pure-style sites,
and that I had struck two or three of them that matched exceedingly closely with
Uhle's Ocucaje collection as just "purified." Not wishing to anticipate the full
report on the 1926 excavations in which Collier is to be co-author, I refrain from
publishing any illustrations here. Moreover, it will be better in the end to have
all our Nazea ceramics in Chicago illustrated and described together in one volume.
However, I shall give here a brief account of the matching lots from Nazea.

These are from Majoro Chico A, graves 18, 20; from Aja B, graves 1-9; anid
from Cahuachi A, thirteen graves (numbered 1-15, less 8 and 14, numbers not
used). The ware in the Aja B graves was badly broken, and textiles were lacking.
The Cahuachi A graves were in soft sand, much of their pottery was intact, and
textiles were numerous: O'Neale (1937), has described them in Field Museum
Memoirs.

I present first a classification by shapes of vessels (table 1). This is essentially
complete for the cemeteries and graves in question, though as sherds are matched
in Chicago, a few more vessels or major fragments may be added to present col-
umns for 1926. The Uhle Ocucaje collection is as counted after its "purging," as
just described. I have kept the Gayton-Kroeber designations of shapes by letters,
though I have reworded some of the phrases descriptive of them, to give them
better fit to the specific Nazea A forms which alone we are concerned with at the
moment.

It is evident that Uhle's vessels from Ocucaje-Ica are closely paralleled by mine
from Nazea. The proportional frequencies of the several shape classes are very

d The painted textile in Kroeber, 1944, pl. 12b, though not from grave Al, is evidently also
Paracas-Nazea transition in time, if not full Paracas, because of its resemblance to Kroeber,
1944, pl. 12b from the Truel-Ocucaje-Paracas collection.

Kroeb er: Toward Definition of the Nazca Style              335
similar. The main difference is that the higher bowls (shapes F, H) and two classes
of jars (P, Q) are represented at Ocucaje and nearly lacking in the three com-
parable Nazea cemeteries. (If this is confirmed, it might mean that Ocucaje is on
the whole a bit later within phase A than the three Nazea cemeteries, since A in
general is characterized by frequency of low bowls, and the trend to B is toward
higher and higher vessels.)

TABLE 1

COMPARISON OF SHAPE FREQUENCIES

Shape and descriptionUhle        1926        1926      1926

Shape and description    |  Ocucaie   Cahuachi A   Aja B    Matro Chico

Inside-painted bowls

A    Plate bowl ......................  12        11          5          5
Outside-painted bowls

B    Low gambreled bowl ............   13          3          2          2
D, E Low-flaring and flat bowls .......  33       17    -    10          7
F, H Flaring bowls ...................  12         1

Total outside-painted.        [58]      [21]       [12]        [9]
Other shapes

U    Double-spout ...................   9          2          3          1
V    Head-and-spout .................   1

X    Human figure vase .1 1                        2         ..          1
P    Bulbous jar, height 86-103 per cent  6
Q    Lipless jar, sides sloping in 4......

Three-handled: K-S, pI. 28i .1                2          1
T    Two-handled, unslipped, in-

cised, or scored, pl. 28f ........ 4        2          4
Pan's PPe ............................ p.          1
O ther                      .............................  3  2

Total classified .............  100      41         27         16

What is most convincing, however, is that pieces unique at Ocucaje recur at one
or more of the A sites in Nazea. Thus, the human figure vase, class X (pl. 25a) ; a
three-handled jar of quite specific form and painting (pl. 28i); and two-handled
(T) incised pots or jars (28f).

RESEMBLANCES IN DESIGNS

The similarity extends to design. In table 2 the Ocucaje column refers to frequency
of subjects on the B, D, and E shape low-bowl outside designs as illustrated in
Kroeber-Strong plates 25-29 after these have been "purged" of the atypical grave
lots discussed above.

Or again, on red-rimmed, inside-painted plate bowls of class A, compare the
Uhle Ocucaje fish and fruit designs of Kroeber-Strong's plate 28g, h, q, with the
following themes from Cahuachi: fish, single or paired, 2; kinked fish, 3; two birds
with fishes, 1; two birds, 1; bird with flower, 2; flower alone, 1; beans, 1.

336        University of California Publicatiomns in Am. Arch. and Ethn.

TABLE 2

Low BOWLS, OUTSIDE DESIGN FREQUENCIES, SHAPES B, D, F

Design                 Uhle Ocucaje 1926 Cahuachi

Mammals ................................     1
Fishes ...................................

Birds .................................      8          7

Birds, vertical ...........................              1
Toads ..............................         1

Snake (?) heads (?) .....................               3
Double-ended snake ......................    1

Spiders ..................................              1

Trophy heads ............................                1
Fruits .................   ...........       4          2
Flowers (?) ..............................   1

Stars. 1.............................        .
Slings. 1.............................       .
Steps, Step-frets .........................  2          2
D iam onds ...............................   2
Boxes ..................................     1 I
Checkers ............................        1

Ovals ................................... 1

Double spirals ...........................   1
Feet .................................       1

Stripes .................................    1

Other ..................................     1           1

27          22
Paneled .................................    8          9

........................................  .2
Not paneled .............................   17          13

27          22
PROVINcIAL RELATIONS OF ICA AND NAZCA VALLEYS

It seems clear from the foregoing that the type of Nazea ware first discovered by
Uhle at Ocucaje in Ica is not a merely local specialization but that it recurs in
Nazea Valley in essentially identical form. Some minor variations between the
products of adjacent valleys are expectable in principle, in this early, pre-imperial
period of the history of Peru; but the parallels just adduced show that such differ-
entiations as may ultimately be established will be rather small and secondary.
Nazea A, as defined by its pottery, is ruled out as a primarily provincial culture
type characteristic of its valley. It is the culture of a period, shared by two or
possibly more valleys, or at the very least by lower Ica Valley and by the central
or Nazea part of the Rio Grande drainage. When descriptions of the Cahuachi A
and Aja B ceramics and their grave associations shall have been added to those
of Cahuachi textiles by O'Neale and a fuller inventory of Uhle's Ocucaje B, C, F
ware, the A phase of Nazea culture will be reasonably well and reliably defined.

Kroeb er: Toward Definition of the Nazca Style        337
And that in turn means that the process of authentic unraveling of the story of
the development of Nazea culture as a whole will be theoretically possible through
a combination of two procedures: first, the subtraction of A elements and features
from the total mass of Nazea materials; and second, the segregation of this re-
mainder into groups varying from the most similar to A to the most different.
Such a segregation in turn will be subject to check against publication of remains
from various post-A pure-type lots of :graves such as Strong, as well as I, has
excavated in the region, or as others may uncover with record of proveniences; and
above all, to check against the results of Strong's excavations of stratified refuse,
as the most direct evidence of successions. However, while refuse stratification
furnishes the firmest skeleton of relative chronology, the details of vessel shapes and
painted patterns, as well as of textiles, building techniques, and so on, are generally
better provided by systematic excavations, especially of individual graves. And
patient stylistic analysis of the large Nazea collections-there are several thousand
vessels available in public museums, even though they mostly lack data on individ-
ual provenience-is bound to reveal some significant forms and features that hap-
pen not to occur in the necessarily limited materials recovered in stratigraphic
and grave-lot excavating. In short, the ultimately fullest interpretations are the
product of the joint and complementary use of the three methods of stratigraphy,
association, and stylistic definition.

RELATIONS OF PHASES WITHIN THE NAZCA STYLE

The Nazea A ceramic style is vigorous but somber, severe, and basically simple in
its expressions. There is also no doubt that Nazea B, more or less as Gayton and I
have presented it, is brighter, lighter, and more complicated in its treatment, as
well as showing new shapes and new design themes and forms.7 As I have always
seen it, A is the initiating and formative phase of the Nazea style, B its culmination
-the auge, as they say in Peru. X, or AB-it would have been the better label to
retain-is merely the transition from A to B; one might legitimately stretch or
contract or even omit it, if one chose. If X is retained as a phase, it is more easily
defined by a series of characterizing concrete features of detail than in terms of a
namable or expectable stage of stylistic life history. The formation and the cul-
mination of a style are designable aesthetic events. But whether one makes B abut
directly on A, or inserts AB, or a series of AB1, AB2, AB3-these are choices rather
of interest and convenience. Conceptually, A and B are the dominating polarity.

True, with enough finely analyzed data in hand, it will presumably prove pos-
sible to recognize successive stages of A and of B. And Y I would define as still
containing Nazca-style ingredients, though sloppily conceived and executed; but
beginning to be mixed with occasional elements and patterns originating outside
the Nazea tradition. With enough materials and analysis, Y would expectably sub-
divide into two or more stages, with successively less and less Nazea in it. When
the Nazea ingredient, finally, is so diluted or altered that it is no longer apper-
ceived on inspection (though historical analysis may reveal it as still persistent in
transformation)-then we are "outside" or beyond the Nazea style and within
Coast Tiahuanaco, Huari, Ica Epigonal, or Middle Ica.

7 See fn. 3 above.

338       University of California Publications in Am. Arch. and Ethn.

This suggested outline of the history of the Nazea style is not drawn directly
from the joint paper of 1927, but is based rather on subsequent inspection and
reinspection of Nazea wares and published illustrations, on letting the perceptions
sink in, and on gradually converting them into a historic formulation. At the same
time, it arrives at much the same results, in an all-over way, as our 1927 findings;
and where I differ from these in detail it is because I now believe that we framed
our correlational categories prematurely, and sometimes too narrowly or too
rigidly, and consequently arrived at results which, while mainly right, erred in a
fair number of particulars.

For instance, I would now define painted B designs-B in the sense of what is
post-A and pre-Y-by the following themes.

PARTIAL LISTING OF CHARACTERISTIC B-PHASE DESIGNS

Women's yellow faces in series, long eyed, hair lock on cheek (1927, fig. 3, no. 18, but there
attributed to X).

Complementary interlocking fish (or snakes) (1927, fig. 3, no. 26).
Jagged staff deity (1927, fig. 3, no. 33).

("Degenerate") deity head, standing alone (1927, fig. 3, no. 22, attributed to X; Seler figs.
237-241).

Jagged staff scepter or mace, as distinct from jagged staff deity.

Jagged staff (pointed, bearing two rhomboids, or one) used as ray from head.
Ray from head, end bent over on itself.
Ray, end partly curled.

Pointed ray flanked by two rays with bent-over ends ("fleur de lys"; Seler, fig. 212).
Jagged staff ray similarly flanked (Seler, fig. 221).

Cactus-spinelike short black lines, as emphasized by Bird; straight or forked.
"Demon" heads repeated or strung (Seler, fig. 212).
Same, radiating (Seler, fig. 229).

Same, reduced to octopuslike figure, with one eye (Seler, fig. 216).
In series (Seler, figs. 252-254).

Three-dot faces in sliced circle, surrounded by trapezoids (Seler, fig. 214).

Similar faces, hollow design (outlines only, no color fill) (Seler, figs. 178, 208).
Full-length human figure, frontal, with conical cap (Seler, figs. 122-124).
Full-length men with staves, profile (Seler, fig. 131).
Full-length profile hunters, dancers, etc.

Seler, being interested in significance and not in chronology, in his work empha-
sizes meaning-what kinds of spirits, animals, plants were represented. He did not
at all attempt to unravel the sequences of phases within the style. For stylistic
analysis, minute features like the curvature of rays, almost comparable to brush
strokes, seem the surer criteria, and many of the traits just listed here are of this
type. Seler's interpretations seem to me remarkably probable, on the whole. The
collections available to him were large and representative, and his monograph
serves well for convenient reference.

SUMMARY AND CAUTION

To summarize, I am trying to rejustify the main stages of the Nazea style, or its
pivotal polarities, more exactly than Gayton and I defined them in 1927, by widen-
ing the basis of associations dealt with, and especially by more directly utilizing
pure-style lots as an entering wedge or lever. The result is a reaffirmation in

Kroeber: Toward Definition of the Nazca Style        339
fundamentals of the classification and sequence submitted then, with a consid-
erable number of modifications in detail, and perhaps less seeming rigidity of
scheme.

I will only touch on one new difficulty which reexamination and newly available
materials seem to forecast. Truly pure lots of graves are not too common, even in
small cemeteries in Peru. The thirteen graves I opened at Cahuachi A were uni-
form, and so were the nine at Aja B. But at Uhle's Ocucaje F, two graves of
twenty-three were later A or X, instead of the straight A discussed here. At Majoro
Chico A, only grave numbers 18 and 20 were straight A. Soisongo C had two A
graves, three probably B, four of period Y, and four contained insufficient grave
goods for placement. It will be recalled that the 500 or so vessels which Uhle got in
thirty-odd contiguous Mochica tombs by digging into the Huaca de la Luna at
Moche, and which I assumed to be contemporary in describing them, have recently
been found by Rowe to represent three of the five phases into which Mochica is now
subdivided. It is evident that wholly uniform substyle cemeteries are likely to be
the exception rather than the rule, and that no cemetery can be assumed as dating
entirely from one style phase, without fairly full yield and the most careful scru-
tiny. In other words, associations within graves remain and probably will long
remain one of our most important channels to the sharper understanding of
Peruvian prehistory.

PART II: REVISED CLASSIFICATION OF NAZCA

STYLE VESSEL SHAPES

PURPOSE

By now it will be plain that our 1927 effort on the Nazea style has long been due,
or overdue, for revision. In Part I, I have indulged in a sort of Russian style self-
criticism as regards general principle. In the present Part II I shall review in
detail each of the shape classes we then set up and the evidence bearing on it,
suggesting also further classes or subclasses where it seems necessary, or merging
of extant ones. This part of the present monograph will therefore be constructive,
or reconstructive, and I trust will contribute to the widening and firming of the
basic classification of Nazea ceramic shapes, which is one prerequisite to any
precise understanding of the style and its course. The identification, classification,
and validation of painted design features is at least equally important, but done
systematically it will be a long job, and I leave it to other hands.

It occurred to me several times that it might be simpler to start fresh and make
a wholly new classification. The longer I worked over the material, the more
strongly did I come to feel that though the faults of our 1927 grouping were in-
numerable, they were special faults, or that they lay in short-sightedness or shal-
lowness of procedure such as is almost inevitable for novices, as we certainly were,
in dealing with a rich and complex style like the Nazea. In the large, our old classi-
fication holds up right well. I would not really know how to construct a basically
better one; though after nearly three decades of added experience I can of course
correct and improve our old pioneering attempt at a hundred points.

I will therefore now proceed to examine the shape classes one by one, in the
general order in which we arranged and lettered them, with one exception. Bowls
and plates, our old groups A to H, make the most difficulties, the differences be-
tween classes being finer; so I save them for the last.

GOBLETS AND VASES

We recognized 4 "goblet" and 4 "vase" shapes in 1927. Both are normally higher
than wide. Goblets always flare or spread somewhat from bottom up; and 3 of our
4 goblet classes show greater or less pinching-in, usually rather sudden, around
the middle. Vases, as we typed them, were basically cylindrical, to which property
2 of the 4 shapes added a bit of bell-mouth flare above, and 2 a widening toward
the bottom, which we called "bulbous." Our classification of vases was more suc-
cessful than that of the goblets. This is reflected in the fact that we labeled 141
vessels as vases, only 71 as goblets, and that 2 of the latter classes were quite small.

The following is a description of the 4 goblet types as represented in the outline
key of 1927 figure 2:

I, "angled goblet" (30 vessels): lower half cylindrical, upper half a straight cone frustum.
J, "goblet" (24): a wide cylinder gently constricted in the middle.

K, "double-curve goblet" (10): similar to last but narrower and with a second slight narrowing
toward the lip.

L, "conical goblet" (7): a simple cone frustum, higher than broad at the mouth.

[ 340 ]

Kroeber: Toward Definition of the Nazea Style               341

GOBLET I

We assigned the "angled goblet," I, to period B, with A designs numbering 0; X,
16; B, 28. It is an actual type, but in its full form it is late B and Y only. It is well
exemplified by plates 16f and 15d, which we designated as Y and therefore as out-
side our shape classification. Of these two, 16f is the more squat, 15d more like
figure 2I in silhouette, though its lip is quadrilaterally wavy or petaled. Our at-
tributed exemplification of shape I by 9d as within the Nazea style, I would not
today hold to. The base of 9d has four rounded angles or bosses, its surface is

TABLE 3

ANGLED GOBLETS, SHAPE I

Measurements      8724    8725     8726     8731    8553    8410     8409

Height .................  107    130      80      82       87     148     mate
Rim D .................  120     146     105     124      127     161
Waist D ...............  71      80       63      72       80     106
Base D ................  83      83       65      68       78     104
H/rim D per cent .......  89     89       76      66       68      92
Base/rim per cent ......  69     57       62      55       61      65
Waist/base per cent ....  86     96       97     106      103     102
Waist angle ............  1580   1580    1550     1580   curved   1620

rounded

blackish, its painted decoration is in a band around its waist as in 15d, and the
design in this band strikes me as Y rather than A-B Nazca.' Finally, 1927 figure
llb on page 31 is listed by us as an "angled goblet( ?)," but the main justification
here is in the question mark. The specimen is an F or H bowl, with H/D ratio of
75 per cent, a spread in profile of 170 from vertical, and it flares continuously in
its lower part instead of rising vertically or tapering as an I goblet would.

I reassembled most of the 30 I goblets we listed in 1927-a few were misplaced
or had been misnumbered in listing. I found 7 that fairly well fitted the type.
These were all definitely late B-really B-Y transition. They all show a definite
sharp turn or bevel between side and bottom-nearly a right angle. The bottom
seems quite flat, though actually gently rounded; the side at first rises nearly
vertically. Somewhere in the lower half of the total height, there is an outward
bend, resulting in the remainder of the side flaring about 220 from vertical (18?-
25?). Another way of putting it is that the silhouette bends out in an obtuse angle
of around 1580. In 3 of the 7 cases the lower part of the side sloped inward from
the vertical, so that the vessel showed a slightly constricted waist.

The proportions are given in table 3.

As for design, 6 of the 7 were zoned; 8725 carried horizontal stripes. The lower
zone was quartered across the bottom twice, carried women's yellow faces once.
The upper zone in 8724 and 8731 carried rectangular panels alternately filled with
different simple geometric units.

1 The proportioms of 16f, 15d, 9d are: H/D mouth, 85, 98, 147; min. D/D mouth, 58, 48, 70;
base D/D mouth; 60, 52, 95-not a homogeneous type; 9d is particularly deviant.

342       University of Caiifornia Publications in Am. Arch. and Ethn.

In Y graves at Nazea I excavated a few goblets similar to these, but with the
angle at the waist more pronounced and less obtuse around 110? for the arc,
130?-135? for the chord, in one case; 130? and 1500 in another.

The one other sizable group (besides scattering forms, errors of numbering, etc.)
that emerges in our 1927 listing for "I" consists of 9 concave flaring bowls of
generic F-H type. Four of these were late B in design; all of these had quartered
bottoms. The H/D ratios were 71, 73, 76, 60 per cent for 8577, 8702, 8551, 8567.
Five of the 9 bowls were middle B in design: 8560, 8568, 9162 two-zoned; 8638,
8802 one-zoned, with a pair of deity heads or deities. The H/D ratios were 61, 70,
66, 67, 64 per cent; the shape also that of F-H bowls.2 Such bowls of course
resemble angled goblets in that both groups flare with concave profile; and the
height ratios at least overlap. The real difference is that in F-H bowls the flare is
sinuous, the attractiveness of the shape residing in its curvature; whereas the I
goblets basically seek an angled silhouette. A further difference is that F and H
bowls begin in Nazea A and have their main development there, though they carry
on into B; whereas the I goblets begin in late B and reach their climax as a form
in Y.

In the present paper, plate 40a, discussed in Part III, is somewhat hesitantly
assigned to shape I.

GOBLET L

I will temporarily pass over the plain "goblet" J because of its relation to the vase
group, and go on to the "conical goblet" L, which again seems mainly a Y-period
shape, plus at most late B. It is illustrated in plate 39e, f, of this paper. This is a
rarish type: we listed only 7 examples, and we illustrated none. In 1926 I ex-
cavated at least 2 in indubitable Y association, and 1 in equally indubitable B
context; but this last was of less pronounced shape, with the height only slightly
exceeding the maximum diameter: 108 per cent; whereas the 2 Y pieces run around
130 per cent.

I have remeasured and examined the 7 Uhle pieces from Nazea. They run pretty
uniform in height: 123-145 mm. The H/D percentage averages 110, and seriates
as 97, 99, 100, 107, 110, 127, 129.8 Vessel 8395 is somewhat convex in profile; the
others all show slight departures toward the concave somewhere in their cone
shape. The designs are in sloppy late B or in Y style: from the time of transition
from Nazea to post-Nazea, in short; which is also just where my 3 excavated pieces
fall. We were therefore correct in classing L as a B-period shape, though we did
not add that it also ran over into Y which we were not classifying.

2 These 9 vessels have not been included in the description and discussion of F-H bowls below,
nor of Cl and C2. They agree with the F-H class in their H/D ratio of 60 per cent or more; but
F-H designs are phase A and X, not B. Conversely, the present lot agrees with C1-C2 in having
phase-B design, including bottom-quartering; but C1-02 H/D ratio is less than 50 per cent.

8 The lowest and next to the highest of these proportions come out as 96 and 126 per cent on
somewhat ampler measures for specimens 8795 and 8395 as cited by us in 1927 on p. 43 in the
list of smallest and largest vessels of each shape. Or rather, I must say "miscited," since the line
for shape I has been displaced by the printer and put after the lines for J, K, L. When cor-
rected (by the specimen numbers given in the complete listing on p. 44), the shapes on p. 43
run in this order: . . . G, H, J, K, L, I, M, N.... Or, to correct, emend at head of lines as
follows: I > J, J > K, K > L, L > I.

Kroeber: Toward Definition of the Nazea Style                 343

GOBIET K

The "double-curve goblet," K, to which we credited 10 specimens in the Uhle
Nazea collection, was also without photographic illustration of any of its 10 as-
signed exemplars. The type sketch, 1927 figure 2k, is tall and slender (ratio about
170), and the name "double-curve" is appropriate, since both the lower and the
upper half of the generically cylindrical silhouette bulge a bit, with a mild con-

TABLE 4

SHAPE K: DOUBLE-CONVEX GOBLETS

Waist   Max. D in  No.

H     H/D   Roll-in  pinched in  upper 4  zones

Cat. no.  (in mm.) per cent at rim, from max. exceeds lower  of   Design
Cat. o.  (n mm. percent(in mn.)  diam.  ?2 max. by: painted

(in mm.)  (in mm.)  designs

8566 ..........  117   96     1        2        26       2

8908 .........  85     99     0        2        20       1    Fine execution

8748 .........  102    107    0        3        14       3    Bottom quartered
8501 ..........  120   128    2        7         7       3*

8903 .........  93     129    6        8         1       2*t  Fine execution
8664 ..........  140   149    9        3    _    5       3*   Arrows

8662 .........  137    136    1        7      -10        2    Trophy heads, arrows
8957 pl. 39c . .  137  134    2        4      -11        3?   Pastel tints, clear
8473 ........  . 157   139    0       13      -16        1

8474** pl.39a  151     137    3        7      -14        1    Turned over rays, blood

Mean   .   .. ...    125    2.3       5.6     12.4tt
Range from:   151    149

to: .        85    96

* Top and bottom zone alike.

t Middle zone is plain red band. Top and bottom zone alike.

In place of 8958 listed in 1927, probably by error.
fiGeometric design.
Mate of 8473.

tt Irrespective of sign.

striction or concavity between, and with the upper convexity carried on into a
slightly constricted mouth; which last feature is anomalous in the Nazea style.4 As
compared with the drawing, the 2 measured Uhle Nazea specimense are shorter.
127 and 138 per cent in H/D.

A reexamination of the University's 10 K goblets (see also pl. 39c, d, this mono-
graph) yields the results shown in table 4, which confirm the class. There is always
a pinching in at the waist; and there is almost always a rolling in, or contraction
of diameter, at the rim-slight, but conspicuous. On the other hand, as between
the upper and lower bulges, this varies from an excess of 26 mm. (21 per cent) for
the upper to 16 mm. (10 per cent) for the lower diameter. The vessels have been
seriated for this feature. Evidently what is significant for the class is not where
the maximum diameter comes, but that there is a reduced diameter in the middle.

' The lipless jar Q and the bulbous I vase 0 constrict toward the mouth, but in an essentially
straight in-slope, without the in-curl or roll we attributed to type K.

r-Smallest and largest, 1927, p. 43: the line for shape "J" actually gives the K goblets
8903 and 8584.

344       University of California Publications in Am. Arch. and Ethn.

The proportion of height to diameter is also not too important, varying from 96 to
149.

All 10 vessels I would rate as larger B in style. There are no A designs present.
In quality, none are poor, and most are good. We had the period right in 1927:
table 1 lists 0 A designs, 8 X, 8 B.

TABLE 5

GOBLET J AND VASES M-P: MEANS OF PERCENTAGE RATIOS

H/max. D  Con. 1*  Cone. 2 Cone. 3

Shape classes   H/max. MD    bottx.  m n. D/MD minConmaxc.  min. D/MD D bott./MD

bott.  m    . D/MD m    _nr  _

J, goblet (24), 4 illus.....  123  (133)      ..       86        ..       92
M, "small vase" (30)

5 illus .................  (159)  144       ..         t                111
N, cylindrical vase (47),

8 illus .................  203   (217)      83        86:      771       98
0, bulbous I vase (28),

1 illus ................  (186)   154       ..       ..        ..       121
P, bulbous II vase (36),

7 illus ................  (128)   114       96       ..        ..       113

e Concavity means the maximum distance in mm. between the side of the bowl and a ruler laid from the rim to
the side.

t Of, 10h, fig. lid run: 95, 100, 100.
t For 3 vessels only: llb, llc, lid.

GOBLEr J AND VASES M-P

The last of our goblets, J, the unqualified type, is a valid class, and so, on the
whole, are the four vase classes M-P, to which J is related. As a group, these
shapes are rare in Nazca A, at any rate in the Nazea A represented by Uhle's
Ocucaje and my Cahuachi A and Aja B. But they are extremely abundant in
Nazea B, and characteristic of it. Thus, J, "goblet," 24 examples; M, "small vase,"
30; N, "cylindrical vase," 47; 0, "bulbous I vase," 28; P, "bulbous II," 36; total,
165 vessels out of 563 pre-Y, or between a quarter and a third of the collection.
We were thus correct in assigning 4 of the 5 types-J, M, N, 0-to Nazca B, and
the fifth-P-to Nazea X. At any rate some rather low P's occur already in A-
see 25h, 28m, 28n-but apparently none of the four other classes are A.

The name "small vase" for M seems in need of replacement, and the cylindrical
vases N break into two subtypes. Other than that, it remains chiefly to consider
the similarities and differences between the 5 classes.

In table 5, I present first of all the means of ratios of crucial measures for the
5 types.

WAISTED GOBLET J

The "as such" goblet J has a definite waist smaller by measurement than either
top or bottom diameter (75, 85, 89, 91 per cent of mouth). The bottom diameter is
smaller than the mouth (82, 91, 94, 100 per cent). In this feature the J goblet is
like a bowl, especially like the F shape. It is differentiated from F by somewhat
less spread from bottom to top; by the fact that F bowls only rarely (and other
shapes never) are actually constricted in the middle, though they normally show

Kroeber: Toward Definition of the Nazea Style            345
concavity; and especially by greater height for the goblet. In both F and H bowls
the height is at most 81 per cent of the diameter and thence down to 56 per cent.
Our 4 illustrated J goblets (9e, 9g, 9i, 10d) average 123 per cent. This distribution
is bimodal: 105, 109, and 138, 139, so 2 subtypes may be involved. Goblets 9g and 9i,
the squatter pair, are B in design; 9e and 10d seem advanced B. The 2 shorter
also have the relatively smaller base diameters, 82 and 91 per cent; they thus flare
more; the 2 others are more nearly cylindrical. It looks accordingly, so far as the
small numbers can be trusted, as if at least two stages of J goblet were recogniz-
able; the presumably earlier one, Ji, differentiated from the F and H type of bowl
by being taller, though still somewhat suggestive of F-H silhouette; the later stage,
J2, approximating the vase shapes, with a height ratio comparable to that of the
P bulbous II vases.

The bowl 28j, which is from Ocucaje F, grave 2, long troubled me. It is like F
and yet unlike. First of all, it is too high for F-90 per cent. Then, the bottom
rounds quite sharply into the sidewall, much as in the J class. Finally, there is
actual waist constriction as in J.

In indubitable Nazea B association at Ocongalla West site B, I found in 1926
3 vessels, in graves 3, 4, 12,' which resemble 28j. Here are the percentage data
(subject to correction on final measurements):

Grave 3   Grave 4   Grave 12  Plate 28j
H/MD .........       102       90         87        90
Waist/MD .........   82        84         75        82
Bott. D/MD .........  91      107         77        86

The most constant feature is the waist/mouth proportion. It averages 81 per cent
as against 86 for the J goblets illustrated in 1927. The height ratio, 87-102, is also
not too far from the 105-109 of the more squat Ji subtype.' I am tempted to set
up a subtype J3: near-bowl goblet, or near-goblet bowl.

Tentatively, I suggest "waisted goblet" as generic designation for such J types
as may be established. Two further of Uhle's Nazea examples are shown in plate
39a, b of the present paper and discussed in the corresponding text of Part III.

TAPERING VASE M, FORMERLY "tSMALL " VASE

The "small vase," M, was shown schematically in the 1927 key figure as a straight
cylinder rising from a rounded bottom to a height of 175 or 180 per cent of its
diameter. I recall no quite straight unbroken lines of such all-over length in Nazea
pottery. Smallness per se is also scarcely a criterion of shape, except for such
distortion as any extreme miniature is subject to being warped to. But this shape
M is defined by the five illustrations in 1927, plates 9f, 10e, 10g, 10h, figure lld.
It has an over-all likeness to the J waisted goblet just discussed, but with two
inversions. The basal diameter is usually greater than that of the mouth; and
there is often no actual constriction at the waist (except for the one quite nominal
exception of 9f, and this is counterbalanced by the convexity of 10e). M is also
somewhat more slender than J.

8 The Chicago numbers are 170686, 170691, 170764. There may be others in the collection.

7I have also considered in this connection 27b, 27h, but have decided that fundamentally they
are F bowls. Their H ratio is 80 or less; their Base/Mouth ratio, 75 or less.

346      University of California Publications in Am. Arch. and Ethn.

The greater basal than upper diameter of M allies it to the 2 "bulbous" (viz.,
bottom-bulging) vase types 0 and P which are to be considered in a moment. The
relation to these is analogous to the relation of goblet J to bowl F.

M is a valid class as soon as actual examples are examined-see the examples
in present plate 40b-f. But our old name for it and its key sketch failed to bring
out its characteristics.

I examined and measured a random selection of about half the 30 vessels we had
assigned to M. For these 14, the H/D ratio averages 138 per cent, with a range of
116-157 per cent. The 2 with the highest ratio (8959, 8491) are much the tallest
of the lot, with heights of 176 and 170 mm., whereas the dozen others ranged
between 152 and 135, or around 51/2 to 6 inches, whence no doubt our designation
"small."

Twelve of the 14 have their greatest diameter somewhere in their lower half;
2 are exceptional (8408, 8900) in that the maximum (or equal) diameter is at the
rim. Three of the 12 have an over-all convex profile, tapering to the rim. The rest,
plus the 2 with rim maximum, show a concave profile. The concavity, however, is
shallow enough in 5 cases to result in no minimum diameter below the rim; 4 times
there is such, but the minimum is only 1-4 mm. less than the rim; only in the 2
with rim maximum is the constriction as much as 8 and 5 mm.

All 14 pieces bear B design; 6 are two- or three-zoned. We had the class as B in
1927, with 3 A, 20 X, 29 B designs.

I can find no wholly satisfactory descriptive name for M, but suggest "tapering
vase."

N, CYLINDRICAL VASE

N, "cylindrical vase," is indeed such; and it is the tallest and slenderest shape in
the Nazea style. It is illustrated by lOb, c, f, i, lla, b, c, d. Contrary to M, the
mouth tends to be larger than the base: 6 cases out of 8, with a mean ratio of 98.
The mean H/D ratio is 203; or 217 if the basal diameter is compared with the
height.

When the base exceeds the mouth, as in lOb, the shape is nearly indistinguishable
from P, except for being more slender, that is, less bulbous below.

Most of our illustrations, like lOf, lOi, lla, show a long, elegant, straight, and
even silhouette, with a slight bell-like flare at the top.

However in llb, c, d, there is a more ornate form, which in profile shows three
bulges and two concavities, and four zones of painted design. The uppermost
design zone comprises the second constriction plus the (incomplete) flare to the lip,
and is the widest. The three lower zones of design correspond respectively to a
bulge, a concavity, and another bulge. The coordination is skillful, and its com-
plexity lends an air of richness of ornamentation; which was no doubt the aim.
The design zoning occurs also in lla, but without bulges in silhouette. The N vases
of plate 10 are single- or double-zoned only, instead of fourfold.

I did not in 1926 excavate any of the convex-concave profile jars like llb, c, d,
though I puchased a few.

Although the smooth and the wavy outline N cylinder vases are obviously
related, they differ so sharply that it is warranted to distinguish subtypes: N1 and
N2' Presumably N2 is the later; because all other A-phase and B-phase tall vessels

Kroeb er: Toward Definition of the Nazca Style           347
are single- or double-curved in silhouette, like N1; whereas local bulges, bosses,
or swellings are characteristic of the following Y phase, as in 15d, 15f, and
probably 9d. The fact that the bulges of N2 are reenforced by the multiple design
zoning strengthens the significance of the distinction between N1 and N2.8

0 AND P, BULBOUS-CONVEX AND BULBOUS-CONCAVE VASES

Shapes 0 and P, the bulbous vase types I and II, I would rename bulbous-convex
and bulbous-concave. They are valid classes, though the one (lOa) and two (5a, 5b)
examples of them which we illustrated (out of 28 and 36 attributed, respectively)
are few for substantiation.

Both types are largest in diameter somewhere in their lower halves; the mouth
diameter is smaller, though only moderately so. In both, the height normally
exceeds the width, and perceptibly so. The difference between them is that 0 shape
is convex from bottom to top; P is also mainly convex in profile, but before the
mouth is reached there is a slight concave comstriction, above which the lip spreads
again a bit.

Strong and I pictured 5 P vessels, 3 from Ocucaje B and F (25h, 28m, 28n)
which are straight Nazea A as defined in the first part of the present paper, and 2
from Santiago in Ica (28k, 281) whose context is at least mainly post-early-A. The
three first run 93, 92, 102 per cent in H/D ratio; the two latter, 126 and 127 per
cent. This difference suggests that the form P originated in Nazea as a relatively
squat form in phase A and grew taller in late A or X. Our 2 figured pieces from
Nazea Valley fit this inference: 5a, 102 per cent, looks as if it had A-phase painting
on it; 5b, 145 per cent, B-phase.

There is corroboration in 3 vases newly illustrated in the present monograph
and discussed in Part III: plates 31f and 34a from Ocucaje graves B4 and F6 are
Pa; but 35c, from Ocucaje F3 (later than the rest of F) is a tall and banded Pb
form.

Still further corroboration comes from my 1926 excavations for Chicago at site
Ocongalla West B. The data are in preparation for joint publication in full by
Donald Collier and myself, so they will be only summarized here. Graves 8, 9,
10, 12 contained full B-phase material only; among them, 11 characteristic P-shape
vases, whose H/D proportion ranged from 124 to 156 per cent and average 135
per cent.9 This is in marked contrast with the 92 to 102 per cent range and mean of
97 for the four A-phase pieces, 1924 plates 25h, 28m, 28n, 1927 plate 5a. But the
mean height of 135 per cent for the 11 Ocongalla West B phase-B pieces accords
well with the 133 per cent mean of the 3 phase-B or post-A pieces, 1927 5b, 1924
28k, 281.

It is thus clear that P, like N, subdivides into two subclasses: P1 or Pa, about as
high as wide, and centering in Nazea A; P2 or Pb, about a third higher than wide,
centering in Nazea X-B.

"Means, 1917, pls. 3:2 and 4:2, shows views of two sides of a vase that is neither M nor N.
It consists of a bulging bowl, something like the lower half of shape W, surmounted by a
gently spreading cylinder more than twice as high. The total height in his 4:2 is 71 mm., the
maximum width below 45, the minimum above it 40, the mouth 42; H/W thus is 158 per cent,
which is nearer M than N proportion. There are three zones of design: at the bottom a face
(wearing a sling behind), covering the bulge; a wide zone with a cat deity among tadpoles; at
the rim, a narrow band of trophy heads. The style is B.

9 The Chicago museum numbers are: 170708, 712, 723, 730/28, 732-1, 736, 739, 740, 771, 772,
773.

348       University of California Publications in Am. Arch. and Ethn.

The same Ocongalla West B site also gives additional facts on the bulbous I or 0
type. Graves 3, 7, and 10, also full B-phase, contained 5 0 vases ranging from 112
to 142 in H/D and averaging 125 per cent."0 Three of these show a turning of the
convex profile into a slight concavity just below the rim, but without actual con-
striction, that is, diminution of diameter ;"1 2 are convex right up to the rim.

Shape 0 has no known antecedents in pre-B neriod.

While we are with this Ocongalla West B site, I might add that its B-period
graves 3, 4, 7 contained 3 head jars of type Y, with H/D proportions of 125, 108,
130 per cent,' which, so far as their shape is concerned, are P with a nose pinched
out into a projecting little elevation on one side. The painting-on of eyes, mouth,
facepaint, perhaps hair, then completes the "head jar." When I first looked over
their photographs, while concentrating on shape and ignoring painting, I actually
labeled them "P." A designation of shape Pn (the "n" for "nose") would be as
correct as shape Y, besides suggesting their relationship. Indeed, as no shape Y
jars are known in Nazea-A contexts, it seems probable that their antecedents were
the squat Pa vases of A, and that these had developed by peak-B times into two
forms: the tall Pb and the not quite so tall Pn with modeled-on nose and painted-
on facial features.

We had the period-ascription of classes 0 and P roughly but not exactly right
in 1927.

A designs   X designs   B designs

O bulbous I .........  2         28          21       "B"
P bulbous II ........ 0           2           3   "    X"

We had too few P designs to judge; the encircling bands and stripes to which they
are addicted were not in the list of forty designs with which we operated.

JARS

In the class of jars, we distinguished in 1927, apart from the specialized double-
spout and head-and-spout forms, 4 shapes:

Q  lipless jar .......... 12 examples  2 shown in 5e, f
R  wide-mouth jar .... 13 examples  Oshown

S  narrow-mouth jar.. 23 examples  4 shown in 4f, 6d, e, f

T   handle jar ......... 8 examples  6 shown in 4c, d, e, 6a, b, c

JARS R, S, T

While Q is lipless, shapes R, S, T agree in that all of them have a neck and collar,
and are obviously interrelated. The key sketches in 1927 figure 2 illustrate the
relationship well, except that the wide-mouth jar chosen as example is unusually
low in its collar; though it does seem to be true that wide mouths do tend to have
somewhat lower collars. S and T are one shape, except for the addition in T of a
pair of lug handles. These, however, are far enough up on the shoulders so that
they ordinarily do not extend beyond the maxiimum diameter.'"

10 The Chicago museum numbers are: 170685, 687, 688, 702, 725.

"1 Contrast the P-shape vessel 170723 from grave 10, whose rim diameter measures 76 mm. in
its photograph, but 12 mm. below the rim has shrunk to 73.5 mm.

"12Numbers 170682, 690-2, 700.

18 Pl. 4e has an anomalous cylindrical single-rod handle- perhaps rather a spout, since it is
open at both ends. It also has the widest mouth of all jars figured by us in 1927: 54 per cent

Kroeb er: Toward Definition of the Nazea Style               349
When the illustrated specimens of the three classes R, S, T are merged and
seriated, as below, it is evident that the R jars are not only proportionally wider-
mouthed but wider across the neck, and that the height of their collars or lips tends
to the lower range of this feature. In proportional total height or squatness, R
falls into the midst of the S and T series, and these in turn show little distinction
inter se, except for some tendency of the handled jars to average slightly higher,
and both a little wider and higher in the collar or lip.

H/max. D: S R S S T (T, R) (S, T) T T

Mouth D/max. D: S T     S T    (T, S) T   S T   T   R  R
Neck D/max. D: S S T T T T S S T T R R
H coll./H: T  R   (S, R, S, T)    (T, S, T, T) S

TABLE 6

SELER'S JAR SHAPES, SERIATED

Fig. no.           Rim/max. D*    H/max. D      H/coll. H     Phase

246 .21                                         71            14           BB
310 .27                                          84           10            B
117 .29                                         83            16           B
124 ....                           36          (86)           15           B
35 .41                                           94           20           X

199 .41                                         89            19           B
90a .41                                         107           15           X

357 .45                                         101           17           BBt
95 .50                                           92           19           B
50 .52                                          100           19           X

238 .55                                         90             9           BB
180 .58                                         85             8           A
227 .59                                         97            28           BB
33 .60                                           90           15            ?
63 .                               68           94             9           A

8 jars with R/max. D under 49 = shape-class S, mean 35. 7 jars with R/max. D above 50 shape-class R, mean 57.
t Erroneously attributed to Chancay, following misprint in Uhle's Fricakulturen.

The three classes thus are distinguishable, but are closely related. R and S differ
in that in R the neck has a diameter of 50 per cent or more of the body diameter;
in S, less than this ratio. T jars are essentially S jars plus a pair of small lug
handles, without addition to the maximum diameter. "Two-handled small-mouthed
jars" would be a more exact designation for them than merely handled jars. They
are far more closely related to the small-mouthed unhandled than to the three-
handled jars with constricting cambered mouth just referred to in the footnote.

of the height. With R and S arbitrarily distinguished below by Rim/D ratios above and below
50 per cent, this piece, 8578, would be basically R rather than 5, or than T, as we classed it on
account of its handle-spout. But except for the abstract idea of our word "handle," one promi-
nent cylindrical projection is not the same as two cylindrical lugs-either formally or func-
tionally. For the same reason I do not include with the two-handled class T the three-handled
28i, which my 1926 Nazea excavations established as a type instead of an anomaly. The shape
of the body, the tapering cambered mouth, the size of the handles, the standardized texture and
design-all signalize this three-handled type as far removed from class T.

350       University of Caflifornia Publications in Am. Arch. and Ethn.

Seler has 12 drawings of R and S shapes; he shows no handled T's. I have
measured, pereentaged, and seriated his R and S shapes in table 6.

Combining this Seler series with our own for R and S, and adding T, we have
these means, minima, and maxima.

Mouth Ratio        HID Ratio        H coll/H Ratio
9 R..      Mn 58  50-68    Mn 88   79-100     Mn 17    8-28
12 S..    Mn 36   21-51     Mn 89   73-113     Mn 15   10-20
6 T..      Mn 44  38-56    Mn 97   91-103     Mn 14   11-15

It is again evident that we have basically one form    except for the width of
mouth. This mouth width varies from a fifth to two-thirds of body diameter (21-
68 per cent). This range is too great to ignore in classification; but as there is no
marked gap in distribution within it, our arbitrary segregation: R= > 50,
S = < 50 is probably as convenient as any other, especially since the handle T
shapes prove to be all of near-S profile.

I also rated the 15 Seler jars in terms of estimated degree of development of
painted design within the B phase, on the chance that this might coincide with a
development in vessel shape. My scale was: "A"; "X" for early B, X, or possibly
A; "B" for definitely B; "BB" for elaborate or extreme B quality. There were 4
pots with BB design, on the S jars lowest (21 per cent) and highest (45 per cent)
in M/D ratio, and on 2 around the middle range (55, 59 per cent) of the R
jars; but the "X" and "B" ratings run scattering. The outcome was therefore
inconclusive. Only the 2 or 3 A or possibly A vessels fall at the top of the wide-
mouth range-58-68 per cent. With this finding the Ocucaje A-period R jar in
plate 28p agrees:` M/D, 65 per cent; H/D, 79 per cent; H/collar H, 14 per
cent. All the as yet identified A-phase painted pieces are therefore wide-mouth R.
The S shape is mainly later;`5 and presumably T as a painted form developed
later ;16 but the R form was maintained into late B.

Our 1927 shape-design correlation came out as per this tabulation:

A designs   X designs   B designs   Total
Wide R (13) .........  3           8           0        11
Narrow S (23) .......  9          27          8         44
Handled T (8) .......  2           3           1         6

Total .........   14          38           9        61

This hardly left us any choice other than putting all 3 shapes into X, although
the only series of any size are the 44 designs on the 23 S jars. R and T together had

14The piece, 4711, was found near grave F6, not in it, Uhle specifies; but the snake deity and
the stars on it are A-style painting. I first measured the illustration, which yielded 64, 80, 17 per
cent, as against the figures cited above based on measures of the pot itself. In spite of the small
size of the illustration, the percentage proportions come out essentially reliable except for collar
height, which is too minute in the picture.

15 But not always. Fig. 10, discussed in Pt. III, shows an A-period non-handled narrow-
mouth S shape (F12-4748), whose rim D/max.D is as low as 36 per cent, with H/D 93 per cent.
This is a very heavy piece, dark red with a black band, carrying an overpainted perhaps fugi-
tive design.

16 As painted ware. Unpainted two-handled wide-mouth jars already occurred in A: Ocucaje
F4-4702 is a rough buff pot 108 mm. high, 138 body D, 95 rim D, 83 neck D, 21 H of collar;
H/D ratio is 78 per cent, Rim/D 69 per cent, H/coll.H 19. The fairly sizable handles
project beyond the body 11-12 mm. (contrary to those on later painted T jars), spreading the
calipers to 161 mm. This may have been a cookpot.-Similar are F13-4751 and P13-4752, pl.
34e, f of the present paper, described in Pt. III; except that these are incised, as F4-4702 is not.

Kroeber: Toward Definition of the Nazea Style          351
only 17 cases of design (of our selected 40) on 21 jars. As a matter of fact, as
just shown, jars run from early A through late B. My Nazea diggings have led me
to believe that in the early A period most jars remained unpainted, or with a bit
of modeling or incising; and hence their seeming rarity, since plain vessels have
rarely been collected.

Some S jars with B design, like 6e, f, depart from the near-globular shape of
most R, S, and T vessels toward a more flattened or even semilenticular shape.
This holds also for Seler's figures 310, 117, and especially 246, whose H/D ratios
are 84, 83, 71 per cent. Whether it would be desirable to set up a distinct subclass
of S for them remains to be seen. The flattening of the body is similar to that in U
double-spouts painted with full B design; and the same lower-case letter (such as
Sf, Uf for "flattened") might be used for both subclasses. They appear to be far
more common than the rod-handle or spout of 4e.

LIPLESS JAR Q

The Q or lipless jars are allied to R, S, T in general proportion, but their con-
stricting slope merely ends. They leave off the spreading collar, or the neck plus
collar, of R, S, T. As might be expected, they seem to run a bit lower: 67 per cent
for 5e, 82 for 5f. The latter is just higher than the lowest S and R jars with 71, 73,
and 79 per cent. This overlap in height tends to confirm the basic relation to R, S,
T. The key sketch in 1927 figure 2q is somewhat too high, with a proportion of
about 96 per cent. But the ratio of M/body D is correct: figure 2q, ca. 54 per cent;
plate 5e, 54 per cent; plate 5f, 59 per cent.

F4-4701                /          F7-4717

Mates

a                                  b

F 19 4770           Lipless

Jars Q

c

Fig. la-c. Unpainted lipless jars of period A from Ocucaje; a, b, shape Q; c, pseudo-Q.

352        University of Caiifornia Publications in Am. Arch. and Ethn.

TABLE 7

TRUE Q JARS, NASCA AND OCUCAJE

Cat. no.       H      D      M     HID    M/D                Dsg

per cent per cent          Design

Nazea, group 1

8801 ..............  149   191     79     78     41   Largest, two-headed cat-deity

design; centipede body con-
taining 8 paneled fruits. Note
small mouth of vessel

8964 ..............  96    139     70     69     50   P1. 5f. 5 heads, multiple eyes;

allied to "lacework" trophy
heads

8425 ..............  113   164     87     69     53   P1. 5e. Thorns, blood from mouth
8616 ..............  137   170     95     87     56    "Lacework" trophy heads
Nazea, group 2

9102 ..............  109   109     78    100     72

8963 ..............  97    104     76     93     73   Mate of next. Lacework trophy

heads

8962 ..............  97    100     74     97     74    PI. 9m. Mate of last. Lacework

trophy heads

8864 ..............  92    116     87     79     75   2 jagged-staff deities
Ocucaje, unpainted

F4-4701, fig. la...  117   131     96     89     73
F7-4717, fig. lb...  113   130     96     87     74

TABLE 8

PSEUDo-Q BOWL JARS, NASCA AND OCUCAJE

Cat. no.         H      D      M     H/D MMD                Description

per cent per cent

Nazea, painted

8834 ................  85    114     90     75     79   Flat bottom; very heavy
9025 ................  87    139    110     63     79   Flat bottom

9045 ................  73    115     88     63     77   Nearly flat bottom
8826 ........         102    123     86     83     70   Heavy
Ocucaje, unpainted

F1o-4713 ............  75    132    102     57     77

F19-4770, fig. lc ....  107  156    111     69     71   For shape, cf. pl. 46g, 16-7851

The nominal frequency of Q is 12 out of 563 pre-Y Nazea pieces at Berkeley-
between 2 and 3 per cent. This is about the same as the frequency of R, half that
of S, and somewhat greater than T. The 4 classes together number 56, or just 10
per cent of the total collection.

Plate 9m is probably not to be put into Q; we left it in "Miscellaneous." It
constricts toward the top and there is no marking of the lip, but the basic form is
different from Q.

The Nazea occurrence of painted-shape Q centers in period B. But there are in
the Ocucaje collection 4 Q-shape vessels of Nazea phase A: 4701, 4717, 4731, 4770,

Kroeber: Toward Definition of the Nazca Style         353
respectively from graves F4, F7, F10, F19,-the 2 first being mates of a pair! Three
of them are drawn here in outline in figure la, b, c, and discussed in Part III. They
are all without painted design, though they are polished and slipped or washed.
The situation for Q then is similar to that for S and T: the shape occurs in A, but
was still unpainted, and therefore did not get into collectors' assemblages.

However, it is necessary to distinguish between true-Q and pseudo-Q jars.
The former slope nearly straight upward and inward to the mouth, like a cone
frustum; it is this trait that caused us to apply the name lipless to them. The
pseudo-Q jars curve into the mouth, are convex in the whole of their profile, aver-
age lower than the true-Q, and could be described as bowls as properly as jars.

The 8 painted pure-Q vessels from Nazca are seriated in table 7 by their M/
max. D ratio. It appears that they fall into two groups: first, mouth ratio 41-56
per cent, absolute D 190-139; second, mouth ratio 72-75 per cent, absolute D 116-
100. Both our 1927 illustrations are of first type; our figure 2q sketch of the shape
class is of second type. Number 8962 we showed in plate 9m as of miscellaneous
class Z, though it is listed in Q in the summary of numbers of that class.

The designs of both groups are X or B, and mostly conspicuous full-B. This is
automatic in view of Nazea A not painting the shape when it was made. The
stylistically earliest design in the group is the "centipede-body" deity on 8801, the
largest vessel of the shape in the collection. This might be A or later archaizing.

Two of the Ocucaje unpainted jars belong here in shape, and are added in table 7.
The 4 bowl-shaped vessels classified as Q in 1927 are here listed in table 8 in
order of their M/max. D ratio; together with 2 period-A unpainted Ocucaje vessels.
The Nazea specimens are fairly uniform in size and proportions, differing most in
relative height. Two of them are notably heavy ware; 2 are definitely flat-bottomed,
not rocking at all on a table; and a third is nearly flat.

The design in all 4 Nazea pieces is of late type, either geometric or, if Nazea-like,
rough or slovenly.

8834: background black; a zone of small geometric design, two plain bands.
9025: white zone, with alternately red and black double-facing steps.
9045: four full-face trophy heads, laid in line within a zone.

8826: background yellow; a geometric zone, four plain bands.

OTHER SHAPES
U, DOUBLE-SPOUT

The double-spout or U shape has been discussed in the first part of this paper. We
were evidently right in 1927 in classing it as basically an A-period shape, in the
sense that it is definitely characteristic of A, but we erred in omission by not separ-
ating off at least as a distinctive subclass the later lenticular and other forms.

The detailed history of the double-spout within the course of the Nazea style
remains to be worked out. I will only suggest now that the form was relatively
most frequent in A, became less common in B, and had about or wholly gone out
in Y. At any rate, my seven full-B graves at Ocongalla West B contained only
1 example (170777) among 56 restorable vessels; which contrasts with 5 among
68 vessels in about twenty A graves at Cahuachi A and Aja B.

In saying that shape U is relatively less frequent in late than in early Nazca, I

354   w   University of California Publications in Am. Arch. and Ethn.

am excluding altogether those lenticular vessels with long tapering spouts spread-
ing at nearly right angles, like plates 20d, 19d, which bear straight Coast Tiahuan-
aco painting. These may be post-Y; at any rate, I do not know of any having been
reported in a straight Y grave context.

A few notes on variant U-shape notes may be worth while.

A typical high-B-painted lenticular vessel, probably made in two parts, is shown
in Means, 1917, plate 2:5. An extremely flat lenticular U with long and very
spreading spouts is in Muelle and Blas, 1938, plate 4a: the painted design still is
Nazea, but very late, almost with a Tiahuanacoid effect.

Double-spouts with cylindrical body and with stepped rectangular body are
illustrated in our plate la and lb; 2e is double-conical or rather double-frustum."7

The design is occasionally in echelons of small figures. Of such, our 2d seems late
(trophy heads); Muelle and Blas's 5b (birds) may be Nazea A. Their 5a has a
near-cylindrical body with spouts rising from its periphery, much like Kroeber
and Strong's 28e from Ocucaje aberrant grave F3. The spouts have somehow a sug-
gestion of Paracas about them, though I cannot pin down any trait of these spouts
to a parallel with any specific Paracas piece, Cavernas or Necropolis. The design on
the Muelle and Blas 5a is also strange.

I offer this tentative classification of the two dozen double-spouts figured by
Seler:

Seler's Double-Spouts

A. Standard A-period spheroidal shape, resembling that of an ox heart. This shape was re-
tained into X and B, as shown by the painted design; though B also used other shapes.

1. With A-period design: Seler figs. la, 9, 62, 64, 304, 383.
2. With probably A design, or A-X: 25, 70, 283.

3. With X, X-B, or B design: 39, 42, 91, 201 (flattened spheroid), 202.
B. Shapes other than standard A, designs full B-period.

1. Body shows a turn or bevel near or below middle: 40, 46d, 229, 231 (upper slope concave !),

233 (high spouts).

2. Low spheroid, flat bottom: 214.

3. Body high, narrow, conical bottom: 27c.
C. Body beveled, design Nazca A.

1. Spouts suggest Paracas: 339.

D. Body beveled, spouts long and spreading.

1. Design Tiahuanacoid or pseudo-Tiahuanacoid: 427.

Our 1927 illustrations I would now group tentatively as follows.

Possibly very early or formative Nazea A: 2c (flat-bottomed).

Nazea A: Id probably; le, lf, 2a, 2b; 2e (double-frustum) probably formative A.
Nazea X or B: la (low cylinder), lc (low spheroid), ld (almost lenticular).
?: lb (stepped rectangle).

Coast Tiahuanacoid: 19d, 19e (blackware, incised), 20d (in New York).

17 This piece, 8886, has the white outlining of the design perceptibly raised above the colored
figures and black background. The white may have been put on last; I found no sure indications
one way or another; but the white pigment must have been somewhat viscid. I suspected that
it might have been laid into incisions, but could not detect traces of such. The continuous pattern
is one of 6 step-frets complementary with the black background, in the order: Lt R, G, R, Lt R,
Dk Br, V. The lower row follows the same order but begins with R under the Lt R. The
spouts and bridge are painted B, the top of the body R, the bottom is extraordinarily flat. The
quality of work is superb. I consider the piece very early: formative Nazea A, with reminiscences
of Paracas tradition.

Kroeber: Toward Definition of the Nazca Style         355
This Coast Tiahuanacoid form is somewhat anticipated in body form by the
late Nazea lenticular shape, but is not anticipated (or only slightly) in its taper-
ing, spreading, and hump-bridged spouts. The precise shape of Tiahuanacoid body
seems also not to occur in Nazea B-only the principle of an upper and lower half.
I would infer that the Coast Tiahuanacoid double-spout was not a gradual deriva-
tive from the Nazea B lenticulars, but a mutation to a deliberately new form,
probably elsewhere than at Nazea-Ica, though it probably did take over the double-
spout idea from Nazea. As for Muelle's interesting suggestion that the new form
was arrived at through metallurgical technique and retransfer of the effects to
ceramics, the final validation of this of course depends on the fortune of discovery
of appropriate archaeological association; but the hypothesis continues to appeal
to me as probable.

V, HEAD-AND-SPOUT JAR

Shape V. head-and-spout jar, is represented by 8d, 8e, 8f which I take to be Nazea
B, and by 9j, which we classed as B in 1927 but which I now construe as A. This
last is not a true head-bridge-and-spout, but a small spout rising out of the much
larger head of a whole human figure, to which a strap-handle has been added
reaching back down to the body; this handle is flat, but it does not really bridge
anything. A vessel like this is obviously not strictly equatable with the B head-
and-spouts. It might be classed as a related subtype, or even as a variety of X.

Plate 25b, of period A from Ocucaje F10, has the spout rising out of the body
of the vessel, and connected with the head by a bridge. It is therefore a true head-
and-spout by definition or in principle, as 9j is not. But its proportions, profile,
and painting differ markedly from 8d-f, and show definite similarity to 9j. It
might therefore be constituted into another subclass.

Seler illustrates half a dozen V-shape pieces, plus 2 whose drawing is incom-
plete or ambiguous (figs. 212, 348), and 1 (317) which is a drawing of Uhle's 25b
from Ocucaje. The 6 first group as follows:

21, 22, human (man's ) head on bird body.
211, woman's head on woman's body.

222, woman's head on cylinder-plus-cone body.
347, 349, fisherman's head on spherical body.

Of these, 21, 211, 222 are B, 22 uncertain, 347, 349 probably A.

Two American Museum head-and-spouts are shown in our 20a-b and 21f. The
first is a well-modeled, finely painted, and polished figure of a woman; both head
and cloak-covered body, corresponding to 8d-f, are obviously of B phase. The
other has a man's modeled head and simply painted body and arms on the spheri-
cal vessel body, and might be A.

W, FLARING RIM VASE

The flaring rim vase, W, is one of the smaller Nazea style classes, represented by
7 examples in the 563 vessels. The illustration 8c is closely duplicated by 20c from
New York. The outstanding feature is the enormous vase collar, wide, high, and
spreading, and always set on top of a low flattened spherical trophy head, which

356      University of California Publications in Am. Arch. and Ethn.

has a modeled nose, like the Y vessels 7d-f. The collar is nearly as high as the head
in 8c, higher in 20c. It is almost as wide-across the vessel's rim-as the head is
wide. The period of both vessels is obviously B.

A W shape which I excavated at Ocongalla Zero in 1926 (Chicago, 170423) has
its collar 43 per cent of the total height, 80 per cent as wide as the body. The collar
carries the complementary interlocking snake pattern, dating the vessel as B.

X, FIGURE VASES

X, the figure vase, has our dozen examples held together by the abstract idea of
a human body rather than by any particular form of body. For instance, 9a,
b, c are almost certainly A-to judge by 25a, Ocucaje F6-4708, and by vessels I
excavated in 1926-although in 1927 we labeled 9a, b, c as style B. On the other
hand, 8a, 8b, which we also classed as B, are actually such. I would also now class
as B, 15a, a fish-tailed man, which we put into Y. Muelle and Blas's Muestrario
lb seems indubitably A.

In general, phase A represented men, phase B women. This holds for 9a, b, c
as against 8a, b. The A-period 25a and Muestrario la also represent men. The
male figures have the opening of the jar the size of the head; in the female ones
8a, b, the head projects from a great collar or basket about as big across as the
body of the vessel.

Seler shows one male seated figure, 149, which is much like our 9a, b, c and
presumably also Nazea A. He illustrates four seated female figures: 205, 206, 209,
210, which I would adjudge B; for one thing, they have almond eyes. (This is in
addition to a figurine of a naked tattooed woman, fig. 208, and a painting of one
as seen from above, fig. 207, with full modeled head in the center of a shallow
bowl of apparently Aw shape.)

Seler's figures 347-349 show fishermen with nets, done probably in A style. Two
of these are shape V, with bridge and spout; the third, 348, is not clear and may
be V or X.

It will undoubtedly be necessary to distinguish subclasses of shape X, and in
part these will correspond to periods. Xm and Xw for men and women figures may
be indicated, Xa and Xb for phases of style, Xe, Xf for effigies, full figures, and
so on. If necessary, combinations of these could be used, like Xma or Xby.

Y, HEAD JARS

Our shape Y, head jars, manifests some heterogeneity because we have allowed
the concept of a head to overshadow the actual shapes. However, several of these
shapes repeat often enough to constitute valid subclasses. We counted 22 exem-
plars of Y shape, of which we illustrated 6 in 7a-f. I divide this group of 6 into
4 subclasses.

Shape Y1, as already discussed under P (bulbous II vase), is essentially a P
vase with modeled nose and sometimes with modeled conical bosses for ears. Eyes,
eyebrows, mouth, hair, and face decoration are shown by painting. The profile of
the vessel above the head is slightly concave and always carries painted design,
but this is rarely elaborate. While the H/D proportion averages less than in
straight-P vases of the same phase, the difference is modest. In 1927 we classed
all Y shapes as B in style. The 3 Y1 = Pn vessels I excavated at Ocongalla West B
(as mentioned above under P) were in indubitable B associations. There is at

Kroeber: Toward Definition of the Nazca Style          357
least nothing in 7d, e, f to conflict with this attribution; and Uhle's A-period
graves from Ocucaje and mine from Cahuachi, Aja B, and so on, contained no
Y1 head jars. The main type, Y1, is thus clearly B in period.

Seler shows a number of shape-Y1 pieces: figures 138, 143, 144, 145, 146, 147,
148, 151, 153, 154. Of these, 153, 154 clearly represent trophy heads, as Seler says.
The others presumably represent heads of living people.

Shape Y2, in our 7b, 8796, is a modification of wide-mouth jar R by addition
of modeled nose and ear projections and painting appropriate to picturing three
trophy heads of as many colors, treated like medallions. While this appears to be
a type-I was shown a miniature at Nazca-7b is unique in our Uhle collection.
The illustration shows the white trophy face, with rectangular dark red painting
bordered in black below the eyes. A second face is light brown-orange. Here the
paint is white-bordered black and encloses the eye; it is cleft from below, and
falconid. The third face is dark purplish red; the painting is again below the eye
and consists of a black-bordered orange triangle, equilateral and laid on its side,
point toward nose. These three styles of eye paint on the same vessel of course
prove that the three paint styles can scarcely have generic diagnostic value for
time period. The three faces stand out from a black background; the neck collar
or lip is dark red.

Y3 is a modification of lipless jar Q by addition of a modeled nose, and might
be called Qn. We did not show an illustration, but Seler has several, 140, 155, 156,
157, 158, of which at least the last 4 are trophy heads.

Y4 is a head-jar type with bulging turban, or mass of gathered hair, across
which is painted a sling worn as headband: 7a; also Muelle and Blas 2a, and Seler
141, 142. I excavated for Chicago a somewhat similar pottery head placed with a
beheaded corpse in A-style associations; and the 3 figured pieces show nothing
that conflicts with their also being A-period. The sling is a favorite A motif, but
rare afterward.

There are head vessels of other forms of which only 1 piece has been figured, so
that I do not know if they represent further subclasses or are sporadic uniques.

Our 7c is spherical without opening except for a round hole in one side about
where the ear would come. The nose is modeled. This may be related to shape Y3.

Seler's figure 138 is a rather crude piece that is somewhat of a cross between
Rn and Y4.

BOWLS AND PLATES

Open bowls are the most difficult vessels to classify for shape in the Nazea style.
That is why consideration of them has been kept to the last.

In 1927 we recognized 8 shape classes among the 176 Uhle Nazea bowls. These
are, with their letter designations, descriptive names, illustrated exemplifications,
and number of pieces ascribed by us to each class:

A  round-bottom bowl .................   3g, h, 21d  20
B  point-bottom bowl .................   5d          18
C  conical bowl .................        4a, b, 21c  35
D  shallow bowl .................. ........  ......   9
E  angular bowl .................        2f          25
F  cup bowl .................            fig. llc    30
G  straight bowl .................        5c          5
H  flaring bowl ...........              3a-f, fig. 12a 34

358      University of California Publications in Am. Arch. and Ethn.

There is at least one feature characteristic of all or most exemplars of each class.
This is usually the particular trait which is denoted by the name we gave the
class. The total characteristic shape of each class, however, always comprises sev-
eral traits: the proportion of height to width, the shape of the bottom, the slope
and the curve of the sides, the side (in or out) on which design is painted. In the
range of open bowls, 8 classes are a good many to operate with, and when the
going got thick with transitional forms, with hybrids such as a round bottom but
flaring side, we tended to classify vessels by their one name-giving feature and to
disregard all others. Such a procedure cuts through entanglements, indeed, but
at the cost of forsaking a "natural" classification, which has as its first principle
the consideration and balancing of all features, or as many as practicable; at any
rate until it becomes clear which features legitimately override others. A balanced
or natural classification would have given us more than the 8 classes we established,
but would have justified them better; and the greater number of classes would
almost certainly have over-grouped into fewer than 8 basic classes. All this is of
course much easier to recognize a quarter of a century later, and for a first at-
tempt we might have done much worse. We ought, however, to have added de-
scriptive definitions of our types. Our names are brief and one-featured, the 1927
key sketches in figure 2 are quite small, and we showed only fifteen photographic
illustrations for the 8 classes-indeed only three for the 5 classes B, D, E, F, G!

Using the ascriptions of particular bowls listed on our 1927 page 44, I found
some of our bowl classes obviously heterogeneous, and others intergrading. I have
saved such shape classes as could be saved, but have had to break others up and re-
make them. At that I am not wholly satisfied with my new classification: for
instance, F and H still overlap, and their uncompromising separation would either
be arbitrary or would leave an unclassified residuum. Bowls just are the most
difficult shapes in Nazea. And all Nazea ware is given to flowing modulations of
form in comparison with Mochica, where representations are innumerably varied
but these are all contained in 6 or 8 basic vessel shapes.

I would make a primary division between bowls painted inside and those painted
outside. There are very few painted both in and outside; so the dichotomy is
simple as well as conspicuous. Painted design of course is, strictly, not part of
shape; but it predetermines it to some extent. At any rate, it is a feature that
works, pragmatically. I therefore begin by distinguishing "plates," painted in-
side, as a class, from "bowls," painted outside.

PLATES: CLASSES Ac, Ar, Aw

All our 1927 cited illustrations of classes A and C (3g, h, 21d, and 4a, b, 21c) are
plates, by my present definition. I therefore began by keeping A and C as desig-
nations. But only 8 of our 20 A vessels proved to be inside-center-painted; 4 others
we had put into class B because their bottoms were more pointed; and our class C
consisted of 4 inside-rim-painted and 31 not painted inside at all.

I therefore start wholly fresh, with shape A defined as a shallow bowl or "plate"
painted on its inner or upper side.

Kroeber: Toward Definition of the Nazcxt Style                 359
Next, I subdivide into:

Ac, center painted, bottom rounded, with Nazea A design

Ar, rim painted, bottom tending to the conical, B- or Y-style design
Aw, whole inside painted as unit, smallish vessels, B or Y style

Acr can be used if necessary when rim painting is added to center painting, as
in 8466. But when the center of A-phase bowls is painted with a definite design,
and to this there are added merely pairs or triplets of thin lines, usually white,
across the rim, then this addition is so subsidiary in effect that I would disregard
it and call such plates Ac. They occur in my Cahuachi A and Aja B 1926 collec-
tions, but very rarely in Uhle's Nazea material.

Class Ac

We had 7 of these in our old class A; the last 4 of the following list we had put
into shape class B. The signs mean: b, a bevel or angle outside, where center and
rim meet; b+, bevel marked; b-, bevel mild; rd, side rounded. The ratio H/D
averages 38 per cent, with range from 28-45. The style of painted design is in-
dubitably Nazca A.

TABLE 9

CENTER-PAINTED PLATES, CLASS Ac

Cat. no.        Bevel     D        H      H/D       Design in panel  Background

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _   _ _ _ _ _ _   __ _____   ______ per  cent _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

9156 ..................   b       190      85      45    "Heart," etc.         W
9155 ..................   b       151      63      42    High fish             W

9093, p1. 36b ............  rd    155      60      32    Root, etc.            W
8704, pl. 36c .............  b-   169      67      40    Face                  W
8653 .................    b       164      66      40    7-point star          W
8652 ..................   b+      179      75      42    "Fruits"              W

8651, pl. 3h .............  b+    159      55      35    2-headed condor       R
8466* ..................  b+      174      70      40    "Beans"; rim bar8     R

8592, pl. 36e .............  b    178      58      33    High fish             W
8523, pI. 36f .............  b    217      60      28    4 tuna fish           W
8431, pl. 36d ............  b     162      66      41    5 beans               W

* Strictly, 8466 is Acr. The center has a bold vegetal design, which might represent sprouting beans. The rim design
is geometric on R ground. It is crossed by bars alternately Y and B, both W-bordered. The R space between bars is bi-
sected by a B line, and the Y bars themselves are similarly B-bisected. See Pt. III, text for pl. 36.

Eleven Ac plates out of 500-odd Nazea vessels is 2 per cent-a very low pro-
portion compared with what I got in my digging at Nazca, which was around 8
per cent of all mendable recoveries, and toward a quarter of all style-A vessels.
I conclude that plates break easily in the grave and that buyers are not much
interested in them, so that huaqueros mostly do not bring them in. In other words,
there has been an adverse selection in collections assembled at second or third hand
as Uhle's Nazea one was. On the other hand it contains a higher proportion of
double spouts, around 7 or 8 percent, than my excavation yielded.

360       University of California Publications in Am. Arch. and Ethn.

a                                b

d
e

Fig. 2a-f. Gambreled profiles (except b, sphere segment, GG) of plates of class Ac from Nazea-

interior-painted designs shown in plate 36a-f.

B5 - 4649      a                F17-4760

Fig. 3a, b. Gambreled profiles of plates whose interior-painted designs are

shown in plate 31b and fig. 4a; Ocucaje, period A.

It will be seen that while this class of plates is determined by its circular field
of design painted on the concave or upper side, the shape outside is usually bev-
eled; and the exterior profile is essentially that of outside painted class B, gam-
breled bowl. The plates do run somewhat flatter: height 28-45 per cent of diameter,
mean 35 per cent, where the bowls are 37-49 per cent, mean 44 per cent; but this
is a proportional difference only, and there is overlapping between the two classes.

The 6 class-A plates from Nazca, whose design-painted interiors are shown in
the present plate 36a-f, have their exterior profiles drawn in figure 2a-f. It is
at once evident that 5 of the 6 have the bevel or gambrel characteristic of outside-
painted B-shape bowls; but one, figure 2b, is simple hemispherical or shape GG.

Kroeber: Toward Definition of the Nazcaa Style              361
From Ocucaje are B5-4649 shown in plate 31b and in profile in figure 3a of the
present paper; F17-4760, design in figure 4a and profile in figure 3b; and F3-4688,
design in figure 4b.

F7- 4760

f3- 4688

b

Fig. 4. Sprouting bean(s) design on interior of period-A plates from Ocucaje.

Class Ar

Rim-painted plates are perhaps descendants of the center-painted ones, by shift
of the area used as decorative field. At any rate, none of them are clear Nazea A,
and some are late B or even Y. Four of the following we had in 1927 in class C;
I have extracted them and added 9048 from our former Nazea Y period group 2
(Tiahuanacoid), and 8665, to make table 10.

Neither in proportion nor in contour of profile is this a wholly coherent class.
The profiles of two of the group, plate 38e, f, are shown in figure 12a, b, and are
discussed in Part III in the text for plate 38.

TABLE 10

RIM-PAINTED PLATES, CLASS AR

Cat. no.       D        H     HI/D per oent             Style

9008, pl. 4a ......  203      61       30     Profile conical
9032, mate .......  209      63        30     Profile conical
9048, pl. 14e ......  223    64        29     Cf. 9008

8426, pl. 4b .....  191       76       40     Bottom flattish, profile concave, flaring
9023, pl. 38e ......  227     91       40

8665, pl. 38f ......  154     85       55     Shape I in profile

Class Aw

This is a small group of small, shallow plates, of uniform    convex curvature,
painted all over their inner or upper side. The style phase of the design is B or Y.
Three of the 5 are quartered-a design frequent on the outside or bottom in class
Cl. The 2 others carry a plain star or star-rayed face. See table 11.

362      University of California Publications in Am. Arch. and Ethn.

TABLE 11

CENTER-AND-RIM-PAINTED PLATES, CLASS Aw

Cat. no.            D        H    H/D per cent

8600..                        122      38       31    Quartering
8601 .122                              38       31    Quartering
8433 .122                              43       35    Quartering

8432 .138                              47       34    8-pointed star, W on R
9000 .123                              36       29    Face with 24 rays

Mean ....................... .    ..          32

BOWLS

Of our 7 old shapes of bowls painted on the outside, 2 are unquestionably valid
as we grouped them into classes B and E. I have here tried to define these classes
more clearly, and I retain them. Both have sides that are nearly straight in profile.
I treat these two classes first.

However, groups D, F, H, and the residue of C (after the inside-painted plates
we formerly put into it have been subtracted)-all these flare with concavity
somewhere in their profile, and are intergrading and difficult classes. They will
be reviewed after classes B and E.

Straight-Sided Bowls,

Shape B: Gambreled Bowl

The term "gambreled" is by analogy with a two-pitch obtuse-angled roof slope
called gambrel.

The core of this group consists of our former "point-bottom bowls," after trans-
fer from it to Ac of 4 inside-painted plates. The shape is adequately represented
by our old (1927) figure 2b, and by plate 5d, though this is atypical in its bottom
being painted, which is why it was tilted up to be photographed. The form is
characterized by a sharp edge-angle or bevel. Below this, the bottom of the bowl
is a gently sloping cone, more or less rounded. Above the bevel, the side of the
bowl slopes upward much more steeply, sometimes nearly vertically, actually ver-
tically, or even a bit inward. The side may be straight or somewhat convex, never
concave. The diameters of the top of the side (the rim) and of its bottom (the
bevel) are sufficiently close that the side is more or less like a ribbon stood on edge.
This makes it an easy field to paint on. There is little distortion from top to bottom
through shrinkage of the field. Beyond that, the bandlike field is narrow up and
down, but long around the vessel. It lends itself well, therefore, to repetition of
small units. Variety is easily achieved by inverting these, alternating their color
or repeating it in threes, and so on. By "paneling" I designate vertical black or
white lines separating the units of design. Sometimes the units stand free without
such dividers; occasionally they stand continuous, in contact.

I present first, in table 12, a condensed descriptive tabulation of 12 non-Ac bowls
which we classed as B in 1927. I omit, however, 8738, which is deliberately double-
pointed instead of circular, and is obviously Nazea-Y in period.

X~~~~~~~~

m~~~~~~~~~~~~~ I)

>~~~~~~~~4   0 Ca ;4 Pm 4 BBBBB
< e   *aX  00 o b  ~ ~~~~bo  tw ov  o

R~~~~~~~~~~~~~~~ I

k~~~~~~~ bo

11  m  XOX Een $sE  OSlo03

0 0         0~~~~

-,d  m  _1  co  _   _  _  co

ci                       9
ez to00  00t- !M  m t   t-

.S  g O  X 8 0  N O N co co

X       _  _ _l _l _ _l _ _{co  co _l  . t

.0 co   .0  .Z .- .-  .o  .   co . .

t esc  r  te  s U CD  es4 01  co too  *)
1 o   xo  " co  V- ao  u   co  to co Ot-

t   m dsi  C1 x o r  o  co u:

oo oo oo  oo 0o  00 00 oo   0 o  z

364      University of California Publications in Am. Arch. and Ethn.

The first 6 of these dozen bowls are clearly A phase, the next 3 almost certainly
the same. The last 3 however seem post-A, the final piece in the list (8558, pl. 5d)
being in fact late B in design.

The mean percentage ratio of H/rim D of all 12 bowls is 44. The mean absolute
height of the design band is 37 mm.; this is in most cases somewhat more than
half of the total height of the bowl. But since the design band usually slopes some-
what, the lower edge of the design, which practically coincides with the bevel,
comes roughly about halfway the total vertical height.

The last piece in the list shows a drop of design-band height (width) to about
a third of total bowl height. This may be a function of its late period.

From Ocucaje, 26f, g, i, 27c, j belong here. The H/D ratio of these is, in the
same order, 50, 41, 48, 58, 38 per cent: mean 47 per cent.

Also from Ocucaje are B3-4632, figure lla of the present paper, and A1-4491
in figure 7a.

Shape E: Flat Non-Flaring Bowl

Class E is what we called "angular" bowl in 1927 but illustrated only by figure 2e.
Its distinctive quality seems to be that the side slopes up in a nearly straight line
-concavity or bulge is slight. This straightness gives the sides the effect of form-
ing an angle with the base; especially so when the bottom is rather flat, as it tends
to be in this class.

From Ocucaje there is quite a list of illustrated pieces easily assignable to
class E.

CLAss E BOWLS FROM OCUCAJE

1924                HID per cent  1924                 HID per cent
26c .52                           271 ..                   58
e .50                             p........           .. 45
h .51 convex q... 48

j    .41                          u........           .. 40
27a .55                             v........           .. 41
d .57                             w.         ......... 52
e .41 29c ..                                             45
f .47                             g........           .. 46

i .39                            Mean ..                 48 percent

To resolve all doubts, I finally reassembled all 25 E-shape bowls as we had
listed them from Nazea in 1927, and seriated them by H/D ratio in the adjoined
list (table 13). Mr. Dawson kindly measured them while I noted shapes and
painted design.

This proves to be a homogeneous group and a valid class. Not only is it typical
of period A; it seems confined to A. The diameters run from 96 to 207 mm. The
mean H/D ratio is 50 per cent, with range from 37 to 68 (for no. 8925; the next
lower being 3 bowls at 62 per cent). The side is convex five times, straight five,
slightly concave fifteen times. Concave means that a ruler laid along the side
shows a sliver of daylight between. I have entered this eye-judgment of convexity
or concavity in the seriation for height. From H/D 50 per cent up, there are 2
straight-sided, 9 concave.

I also classified the group for relative flatness or roundness of bottom-again
by eye-on a scale: FF, F, (F), (R), R, RR. For this feature the correlation with

Kroeber: Toward Definition of the Nazca Style                  365
height is even stronger. It will be seen from the list that all the flat-bottomed bowls
but one have a H/D percentage ratio from 45 down, all the round-bottomed ones
from 48 up. Also the extreme forms, FF and RR, tend to come toward the ends
of the range.

TABLE 13

SHAPE E BOWLS FROM NAZCA SERIATED BY H/D PER CENT

Cat. no.        H/D per cent     Side        Bottom

8818 .................     37            C            FF
9087 .................     37            (C)           F

8571 ..................    38            0            FF
8572 .................     39            0            FF
8817 .................     43            (C)           F

8647 ..................    43            0            PF
8931 ................      44            V             F

9084 ..................    45            (V)          (F)
8646 .................     45            C             F
8644 .................     45            V             F
8480 .................     48            V             R
8924 .................     49            C             R
8645 .................     50            0             F
9081 .................     50            V             R
8821 .................     50            (C)           R
8934 .................     52            (C)          RR
8640 .................     53            C             R

9002..................     54            CC            R
8912 .................     56            (C)           R
8468 .................     56            C             R
8996 .................     56            0             R
8819 .................     62            C            (R)
9080 .................     62            C             R
8930 .................     62            CC            R
8925 .................     68            CC           RR

Side, scale:

(concave) CC, C, (C), 0, (V), V, VV (convex).
Bottom, scale:

(flat) FF, F (F), (R), R, RR (round).

That E is a shape limited to the A phase of Nazea was pretty well substantiated
by our 1927 correlation table 1. Here the 25 bowls are credited with carrying,
from among our forty designs used, 17 that are A, 3 of X, 0 of B in style. Simi-
larly for the class-A plate, which is also limited to style phase A: A designs 12,
X 1, B 2. By contrast, for shapes that were A phase but also continued later, our
correlation showed much greater period-scatter of design. Thus, double-spout, U
(which ran on into B): A 17, X 40, B 5. For the flaring bowls F and H, which
are mainly A but continue on into "X," our design distributions in the correla-
tion table were F: A 17, X 16, B 1; and H: A 13, X 14, B 5.

366      University of California Publications in Am. Arch. and Ethn.

Concavely Flaring Bowls:

Cl, C2, Residual from Old Class C

The bowls that remain in our old class C after transfer of the 4 rim-painted plates
to shape Ar are not too homogeneous in shape. What nearly all of them share is a
concave, flaring side-which they share with classes F and H-and a H/D ratio
between 40 and 60 per cent-which makes them lower than all but a few extreme H
bowls. Some show a bevel in profile, more do not. Some are fairly conical, some
rounded, some almost flat in their bottom.

As compared with this not too well-defined shape, they seem to belong to 2 or 3
classes in painted design. Only 1 piece, 8739, H/D 45 per cent, appears to have
an A-period design of pods. All the others are B phase, or late B, or decadent B-Y.

One group comprises 10 concave-flaring-sided bowls, mostly late B period, carry-
ing geometric, not representational, designs on their sides, and the outside bottom
painted in quartering. The H/D range is from 40 to 48 per cent, mean 44. They run
from 153 to 187 mm. in rim diameter, with one attaining 214 mm. (nos. 9036, 9095,
8753,18 8674, 8813, 8428, 9012, 8469, 9011, 8591).

With these might be reckoned 8429, H/D 42 per cent, which has a quartered
bottom, but has a bevel and is convex below this, flaring above; also 8723 and 8430,
beveled but not concave, H/D 39 and 46 per cent; 8701, which is concave and
quartered but differs from the group of 10 in being smaller and relatively higher:
131 mm. and 54 per cent. These are all B or late B; 8701 may be Y.

I designate this group provisionally as C1.1'9

Second, there are 10 vessels, all of them concave and flaring and without bevel,
which are painted in two zones (one in three zones), the lower of them containing
women's yellow faces in series. Five of these vessels are in fair quality B style,
mean H/D 47 per cent (8728, 8729, 8965, 8966, 8967). Five are poor in modeling,
texture, and painting, and are very late B or Y; their mean H/D is 59 per cent
(8427, 8588, 8589, 8727, 8730).

I am inclined to unite with these a group of 4 concave-sided bowls intermediate
in H/D ratio: 52 per cent, apparently with B painting (only one of them zoned,
and not in yellow faces), and good in quality of finish (8733, 8734, 8995, 9013).

These last 14 bowls I designate as C2.

Shape D: Flat-Bottomed Flaring Bowl

A second class of bowls which are characteristically concave flaring is shape D.
Our old name for these, "shallow bowls," is too general to serve well: classes E and
A are equally shallow or more so. The D bowls are shallow or low as compared with
classes C2, H, and F, largely because they are flatter-bottomed, though they re-
semble these in their concave profile. "Flat-bottomed flaring" may accordingly
make a suitable name for class D.

18 This piece exceeds our old shape I, "angled goblet" (discussed above), in the degree to
which its inward bend at the middle approaches a right angle.

19 Not counted as Cl-or as C2-are 9 concavely flaring bowls classed in 1927 as I, angled
goblets, but now excluded. Four of these have quartered bottoms, and all are B or late B in de-
sign; but their H/D ratio is too high (60-73, mean 67 per cent) for Cl-C2.

Kroeb er: Toward Definition of the Nazea Style       367
This is not a very frequent type: we counted only 9 examples among 563 vessels
in 1927, and did not illustrate any of these 9. Nor is there a very clear example
pictured in the Ocucaje vessels in the Kroeber-Strong Ica report of 1924.

I have re6xamined the 9 bowls we assigned to shape D, and have added 8486.
They are, together with shape E, as nearly flat-bottomed as Nazca bowls come.
The sides show concavity. They are fairly large, from 168 to 201 mm. across, with a
H/D ratio of 41 per cent, range 35-47 per cent. Six of the 9 seem painted with B
designs, and late rather than early B-interlocking fish, thorns, and the like; the
others are uncertain. We designated the group as of period X in 1927. The present
plate 38c shows 1 of them. Plate 38a and 38b are even flatter bottomed, and the
flare and concavity are similar. We listed these 2 in 1927 as of Nazca-Y2 phase;
but they seem to me to go into shape D if 38c belongs there. Plate 38d seems to
me marginal to class D: in 1927 we included it in class A on account of its bevel,
presumably.

I admit that the demarcation between classes Cl, C2, D is not clean-cut as to
shape; hence my falling back on type of design.

High Flaring Bowls of Shapes F and H

These 2 shapes, "cup" and "flaring" bowl in our 1927 terminology, look very dif-
ferent in the key of figure 2. There the cup bowl, F, is depicted with a conical bot-
tom, sides nearly vertical or parallel, and a sudden roll-out at the rim. The flaring
bowl H has a rather flat-rounded bottom, sloping concave sides (whence the name
"flaring"), and a broad opening. Its H/D ratio, as drawn in 1927 figure 2h, is
about 57 per cent; that of the 2f cup is about 80. Both these figures are within the
range of H/D ratio of the two types, as determined by measurement of all ascribed
specimens. But they are only just within the limits, which are 56-72 per cent for
H, 62-81 for F; and they are at opposite ends of the ranges. Hence the key drawing
for F is over-average tall, that for H over-average broad, so that one would not
suspect from these drawings any similarity or overlap of shape between the types,
such as the measurements cited show to occur on a considerable scale.

In 1927 we showed six photographed illustrations of F (3a-f) but none of H.
We did give then 1 text figure drawing of each shape: of F in figure llc, of H in
figure 12a. The former of these is tilted to show the painted bottom, so the shape
is not clearly recognizable; but a measurement of the specimen, 9043, gives an
H/D percentage of 73, and a slope of the sides of 11 degrees off the vertical. The
H piece of 1927 figure 12a is 8581, whose H/D ratio is 70 per cent and departure
from vertical 21-22 per cent. The significant difference is in the latter feature-
the H specimen "spreads" more than the F; in H/D they are very close. It is also
clear that the concavity of the sides is similarly shallow. We have then a second
indication that the two types intergrade.

A third is the fact that the piece shown in plate 3a was actually included by us
in the roster of those belonging to class F on page 44, although on page 45 it is
given as H, together with 3b-f. This was not an unduly gross error, as is evident
when one compares 3a with adjacent 3b (110, 120 slope; 74, 70 per cent H/D).
But it does show how close some F and some H bowls are.

368        University of California Publications in Am. Arch. and Ethn.

This comes out also in 2 other pairs of bowls to which Mr. Dawson called my
attention, through discovering that they had been misplaced.

D        H      H/D %  Concavity
F 8604 ....                               139       98      71      5

H 9097 ....                               142      104      73      3.5

Here the "H" member of the pair is the less wide and the less concave-the con-
trary of what is implied by "flaring."

D        H      H/D %  Concavity
P 8928 ....                               143       96      67      6.5
H 8917 ....                               143      100      70      6
Again the "H flaring" bowl is the less flaring.

We must accordingly accept the fact that the 2 classes F and H intergrade and
render their distinction difficult.

Nor are we helped by painted design, for on both shapes it is straight or early A
period (like Ocucaje B, F or Cahuachi A) in two-thirds of the cases, late A or
"X" (or earliest B) in one-third.

I give therefore in table 14 means, maxima, and minima of the usual dimension
measures for all2' the 30 and 34 specimens ascribed by us to F and H in 1927; also
the concavity and side slope2' of most of them.

TABLE 14

COMPARISON OF F AND H BowLs, NAZCA

Measurements                  F             H

Diameter in mm.

Maximum ..........................     149           227*
Minimum ..........................     110           130
Height in mm.

Maximum ...........................    117           144
Minimum ..........................      77            83
Height/Diameter per cent

Mean .............................      72            65

Maximum ..........................      81            72
Minimum ..........................      62            56
Concavity in mm.

Mean .............................       4.9           4.9

Maximum ..........................       9             9
Minimum ..........................       3             1
Slope in degrees

Mean .............................      12.40         160

Maximum ..........................      150           220
Minimum ..........................       80           130

* Unique; the next largest run 173, 172, 171, etc.
20 29 F pieces, 31 H were available.

21 I measured 28 F pieces for concavity, 18 for degree of slope of side; 19 H for each. Con-
cavity and slope are measured by holding a ruler edge tangent against the most protruding
parts of the side, wherever they come; concavity is then measured with a narrow ruler across the
daylight showing between the bowl's hollow side and the first ruler. Slope is measured by
manipulating a transparent plastic protractor in front of the first ruler.

Kroeb er: Toward Definition of the Nazca Style         369
It is evident that there is overlap between F and H in every measurement em-
ployed: considerable for four features, and present in the fifth. The H bowls
average a little the larger, around 150 mm. as against 130 in width, 95-100 mm. as
against 110-115 in height. For width, the median of H is at the upper limit of F,
the median of F at the lower limit of H. For height the medians fall somewhat
within the limit of range of the opposite class; but the overlap has a range of 34
mm. as against a total range of 67 mm. for the two groups together.

As regards proportion, the F group is the relatively higher, by 72 per cent as
against 65 per cent; but the overlap is for 10 percentage points out of an extreme
difference of 25 points.

TABLE 15

F AND H BOWLS COMPARED ACCORDING TO PERIOD OF DESIGN

Measurements                  F                     H

Diameter ...........................  21 A    132       21 A        159

8 X        133        9 X         151
H/D per cent .......................  21 A    72        21 A         65

8X          74         9X         65

Concavity ..........................  21 A     4.5      11 A         4.0

8 X          5.8      8 X          6.2
Slope in degrees .....................  14 A  12.5      11 A         15.5

4X          12.1      8X          16.9

For concavity, the F and H means come out exactly alike. This makes the
concavity relatively a bit deeper for F, because of its smaller average size.

For slope, the overlap is not very great-between 130 and 150; and the means
are not very far apart either, 12.40 for F, 160 for H. Again the F mean is near the
H minimum, the H mean near the F maximum. I believe that partial and inter-
grading as its differentiation is, the slope is the feature that most impresses the
eye-has most pattern value. The F "cup bowl" does tend to be somewhat nearer
an ordinary coffee cup in shape; the H "flaring bowl" does flare more, if "flare"
refers to degree of over-all spread, without reference to curve or billowing of the
spread.

Yet when half the specimens of each group fall within the range of the other,
our two classes are obviously not valid, unless there be some unrecognized diag-
nostic criterion that separates them cleanly. I have not been able to find such a
criterion. It might therefore be wise to renounce separation of the classes, and
unite them in "FH." Or F, FH, and H might be distinguished on the basis of the
measurements I have given. The F subclass would then consist of pieces beyond
the H range, H beyond the F range. However, pieces would almost certainly turn
up that were straight F in some features, FH in others; and as new specimens
were described, they would alter the F and H limits, and therewith the FH
range also.

It occurred to me that the intergrading might be partly accounted for by
change in the proportions of F or H or both as the Nazca style moved from its
early A phase to its late A to X phase already referred to. I made the comparison,

370       University of California Publications in Am. Arch. and Ethn.

but nothing much eventuated. Concavity did seem to grow with time, but in both
F and H. In table 15 are the figures for what they are worth.

Summary of Concavely Flaring Bowl Classes

We have now encountered 4 or 5 groups of bowls that "flare" with concavely
spreading sides22 but vary in H/D ratio:

D   H/D: 41 per cent flat-flaring ........................ full B

Cl H/D: 44 per cent (bottoms painted quartering) ...... B-late B
C2 H/D: 47 per cent (design in zones) .................. late B-Y
H   H/D: 65 per cent high flaring ....................... A, X
F   H/D: 72 per cent high flaring ....................... A, X

It is obvious that the 3 shallowest of these 5 groups are late in Nazea-style his-
tory; the 2 deepest, early. I do not however believe that D, Cl, C2 grew up out of
H, F. The designs suggest that, within later Nazea, flaring bowls grew somewhat
higher (41 to 44 to 47 per cent), but that this trend had nothing to do with the
much higher (65, 72 per cent) earlier bowls.

It would rather look as if the idea of concave flaring bowl sides persisted all
through the Nazea style, but that the feature was played with rather variously
in different periods-as indeed it was in goblets and vases also.

Bowl Shapes G and GG

Our last class of bowls, G, "straight-bowl," again makes difficulties. In the 1927
type drawing (fig. 2g) its height runs about 90 per cent of the diameter. We at-
tributed only 5 vessels to G. The one illustration, plate 5c, has sides that are neither
quite vertical nor straight, though nearly enough so to give the effect of a short

TABLE 16

BOWLS OF CLAss G, NAZCA

Cat. no.   H     Rim D    Max. D  H/Rim D H/Max. D  Profile        Design

(in mm.) (in mm.) (in mm.)  per cent  per cent

8434 .....  95      116     123       82       77   convex    3-prongs

8670       102      124      124      82      82    convex    Vertical stripes
8668        124     127      135      98      102   concave   Vertical stripes

8690        124     101      110     123      113   concave   Trophy heads, arrows
8831        95 .....  114    117      83      81    concave   Parallel angles

cylinder: the H/D ratio is 83 per cent. This piece seems to be a low or short mem-
ber of the cylindrical vase series; it is probably unrelated, as a shape concept or
mode, with the bowl classes. I presume we put it in the bowl group largely because
it seemed too low for a "vase."

In the Ocucaje plates there are 2 pieces that approach the G "straight bowl"
type. One is 25h, though this is somewhat more concave in the side and more
rounded in its bottom than the exemplar 5c. It is also somewhat higher, propor-

22Nine flaring bowls which I exclude from their 1927 attribution as angled I goblets (see n.
2, 19) have not been included in any of these five classes. They resemble the first three in being
of late period, the last two in their H/D ratio.

Kroeber: Toward Definition of the Nazca Style             371
tion 91 per cent. It can perhaps best be construed as a low example of class P,
"bulbous II vase."' A second possible G is 28j. This, however, is pinched in a good
deal at the waist for a cylindrical shape.

The measurements of our 5 G-shape bowls of 1927 are listed in table 16.

It is hard to see common features, except that these 5 differ from all other bowls
in having a diameter somewhere in their lower half equal to or greater than the
rim diameter. Three have a H/D ratio (at rim) of 82-83 per cent; but the others,
of 98 and 123 per cent. These last could be construed as P-shape vases, which turn
concave above where their convexity attains its maximum diameter toward the
base. However, the third specimen with concave profile (8831) agrees in propor-
tions closely with the two whose total profile is convex.

Al-4494       a                 4762         b

F3-4688                         F

d

Shape

4802                               GG

e

Fig. 5a-e. Hemispherical bowls of shape GG, period A, from Ocucaje; a, b,

negatively painted; b from outside of grave F17.

All in all, it would seem that our old G is both too rare and too variable to be
a satisfactorily usable shape class. More examples may reveal a significant diag-
nostic feature; until then, the class is best held in abeyance.

GG is a bowl shape which we failed to recognize in 1927, or at least confused
with interior-painted plates. In vertical section it is simply the geometric seg-
ment of a circle-its profile the arc of a circle. This is exemplified by the negatively
painted bowl from Ocucaje Al, 4493, shown in plate 29f; and again by F12-4746

23This is not forcing the point, since 25h is early Nazea and the tall P forms are middle or
late Nazea, in line with the obvious tendency of most vessels to become higher-or lower-as
the style developed. For low P vases in Nazea A at Ocucaje, see pl. 28 m, n, H/D 92, 102 per
cent.

372      University of California Publications in Am. Arch. and Ethn.

in 27r. Here the dominant stylistic feature is the almost perfect rounding of the
bottom. Both height ratios are 47 per cent. Both specimens are early Nazea A; the
negative one seems especially early because in the south other negatively painted
vessels have so far been described only from Paracas. It seems profitable to set up
for these round-bottomed bowls a provisional class GG; the letter is doubled in
order to replace without confusion of terminology our old and dubious class G.
My present figure 5 shows 5 phase-A vessels of this shape from Ocucaje.

In this class GG there also belong 3 bowls from Nazea, outlined in present
figure 6. Of these, 8998 and 8999, different in size, but mates in proportion (H/D
38, 39 per cent), shape, and design are figure 6a, b. This design is exterior and

8812

C

Fig. 6a-c. Hemispherical bowls of shape GG, period A, from Nazea.

consists of alternately red and black "pods" separated into panels by red lines, on
a white background-a typical phase-A design. There is also the bowl 8812, figure
6c, H/D 41 per cent, segment-spherical in profile, with a white, dark red, and
brown-orange repeated amphibian, chiefly eyes and gills, painted on a black band
outside; I am uncertain whether this design is only A or occurs in B as well. All 3
of these bowls we classed as "shape" A in 1927, in spite of their exterior painting,
on account of their rounded bottoms.

There seems in these pieces-and there may be others not yet noted-warrant
for a distinctive shape class of A period.

The "plate" subclass Aw, inside-painted with B-Y period design, also is evenly
rounded, but is flatter than GG, that is, it represents a thinner slice of a larger
sphere.

Again, in Nazea Y, we have a class of bowls which I designate as shape DBY
(deep bowl of Y period), which round continuously. They are represented by
1927 plates 15e, 16b, c. These are of heavy ware and higher than the GG class; the
H/D ratio runs from 60 to 75 per cent. The designs are also of quite different
content and disposition. Toward the rim, their curvature sometimes flattens out
somewhat (never wholly so), sometimes continues into an in-curve. I would assume
that these DBY represent an independent development, not a persistence from the
GG bowls of A times.

Kroeber: Toward Definition of the Nazca Style          373
However, there are 4 bowls wholly painted with outside design-bottom as well
as sides-to which Mr. Dawson called my attention. They lack any shoulder, bevel,
or concavity, have a thoroughly rounded bottom, and then rise to the rim with con-
tinuing but somewhat diminished convexity. They might thus be put into shape
GG; but what is even more striking than the shape is the similarity of the subject
of design, a masked deity, coupled with its all-over coverage of the vessel.

H       D  HID per,cent             Design

8505 .86                  142*    60     Panache, snake    striped body,

} thumb claw,
8920 .84                  143     60     Falcon eye, wings?  spread legs,

trophy heads held
8921 .83                  141     58     Head at both ends,mask whiskers pointing

up and out, 2 snakes as tails along side,
8922t .83                 142     58     no trophy heads

* Actually 127-157, oval; in fact, rim is pinched in to be bean-shaped.
t Mate of 8921.

It seems clear that we have here a distinct tradition-perhaps within a single
workshop-of a highly specific, elaborate, rather difficult unit of form and orna-
ment. If more examples turn up, we shall have to recognize a class characterized
by design as well as shape. Until that happens, it will be most economical to con-
sider the group a variant of GG-perhaps GGa, the "a" standing for all-over paint-
ing. In 1927, we put them all into "miscellaneous" class Z.

I would estimate the design to be late A or early B-more or less what we meant
by X in 1927.

Residual: Unclassified Bowls

The 9 outside-painted pieces that we included in shape A in 1927 are low, aver-
aging 38 per cent in H/D, with range from 32 to 46. They average poor in quality
and are heterogeneous. Two or 3 of them look like A style, 3 are B style, 2 Y or
later, 1 or 2 too nondescript to classify. The seeming A-style ones are without bevel
in the side-they just round on up. The B-style bowls round sharply instead of
being angled. Of the 2 Y or late bowls, 9014, shown in present plate 38d as mar-
ginal to class D, has in profile a flat-coniical bottom (the only one like that), a
sharp bevel, and rather deeply concave and sharply spreading sides. No. 8710 is
virtually a "cumbrous bowl" as Kelly long ago, in 1930, described them, except
that the rim painting-which comes in segments-is outside instead of inside.

These 9 seem a collocation, not a group. Some of them are like shape-class B,
except that the painted sides slope or spread more.

I leave these pieces to be placed later. My present purpose is to determine valid
natural classes, not to give at all costs a non-residual pigeonhole assignment to
every vessel that Uhle collected in Nazea.

THE NAZCA Y STYIE

We did not in 1927 classify the Nazea-Y part of the Uhle collection into shapes.
We did subdivide the 50-odd pieces we diagnosed as Y into three stylistic groups,
Y1, Y2, Y3.

Y1 we characterized as a hasty, careless, decadent Nazea, generally on dull yel-
low or buff or muddy background.

374      University of California Publications in Am. Arch. and Ethn.

Y2 we described as a hard ware, polished, usually with a red background, its
painted design tending toward the geometric.

Y3 included 7 freely modeled birds, animals, and men, and a few others.
I would now describe these 3 groups as follows:

"Y1" is simply Nazea Y-it might also be called "Nazea C"-subsequent to Nazea
B, and the terminal phase within the Nazca style. It is fairly abundant in the Rio
Grande drainage, but not being a fine or even well-made ware, it has got into collec-
tions proportionally much less numerously than Nazea A and B.

"Y2" is the local phase of Coast Tiahuanacoid or Huari. At least this is true
basically, though there probably were some interinfluences with Y1.

"Y3" I now consider to be an artificial mixture-mostly Nazea A, plus some
B and some Y.

Our old listing of "Y1" covered 23 jars (1927, p. 27); 2 goblets or jars (p. 28);
and 7 bowls and a goblet which we considered transitional between Y1 and Y2
though nearer to Y1 (p. 29). The total was 33 vessels.

Y2 included 2 plates painted inside, 2 bowls, 5 jars (p. 27), and 3 oblong bowls
(p. 28); total, 12.

Y3 (p. 29) consisted of 7 vessels shown in 15a, 17a-f, a pair including 9k,2' and
a white bowl, 9007, and a bird jar, 8423-11 in all.

This makes 56 vessels; which with 563 A or X or B, and presumably 38 of Ica
or Inca or foreign type (18a-1, 19a-g) adds up to the 657 mentioned on page 5 of
our 1927 monograph.

TRUE NAZCA Y, FORMERLY tY 1"

These are the principal true Y shapes recognizable in the Uhle Nazea collection.

a. Strap-handled face pitchers, like 12a-f. These I subsequently found diag-
nostic in excavating. There are 12 in the Uhle collection; also I would include the
strap-handled bird pitcher 17e. Except for the almond-eyed face, I know no Nazea
A-X-B shape from which they might be derived. Symbol, FPY-face pitchers (in)
Y (style).

b. Cylindrical-collared jars, like 13c, d, f. They may have three handles, a single
lug, or neither. Symbol, CJY.

c. Transitional forms like 13e (strap-handle,- no face, cambered neck, body five-
cornered).

d. Bowls, with continuous rounding and rather deep lip sometimes incurved,
sometimes not. They come both circular and oval in horizontal plan: 15e, 16b, c.
They have already been mentioned in discussion of bowl shape GG. These deep
bowls are also diagnostic. Symbol DBY.

e. Oblong bowls, like 15c and probably 16a. Symbol, OBY.

f. Goblets, the lower cylindrical part definitely smaller in diameter than the
upper flaring portion: 16f, probably also 15d, and allied, 9d. These seem examples
of angled-goblet shape I which we attributed to B in 1927 but which attained its
full development in Y. Symbol AGY-angled goblet (in) Y.

The following Uhle pieces I am not sure about, but incline to consider Y:

Two-chambered jar, facing figures, 13a-Y on account of noses and handle.

24 We say: "91," but the description shows that 9k is meant.

Kroeber: Toward Definition of the Nazca Style          375
Two-chambered monkey-bridge-and-spout, 17d-on account of color and painted
design.

Wide-mouthed jars 13b,n 15f, seemingly derivative from Nazea A-B shape R.
Symbol, WJY.

Incurved or lipless jar, 16d (H/D ca. 100 per cent). The white "eyes" or "stars"
suggest Tiahuanacoid style, but the larger figure is similar to that of 12d, which
is surely Y.

The inside-rim-painted plates of shape Ar, shown in 4a, b, do not quite seem Y,
as 4e also does not; but they also do not fit too happily in the B-period niche where
I have put them. There are design resemblances between 4b and 12e.

Nazea-Y painted designs are very sloppily executed, but nevertheless char-
acteristic.

1. Detached "demon" (deity) heads, very deliquescent or degenerated; even
amoebalike: 16b, 12a.

2. Armed or rayed forms suggesting suns, flowers, and so on, perhaps derived
from preceding. The ends of the 5-12 rays are usually bent over, and there is
mostly a round center or eye. The rays or arms may be long, narrow, and pointed,
or short and clustered: 12c, d, e, 15c, e, 16c, d.

3. Simple crosses: 13f, 12b, 17d. These may be related to the last; cf. the series:
12e, 12d, 16d, 12b, 13f.

4. Circular panels containing designs: 12b, 13d, 15e, 16c, 17d. Some bowls that
I excavated in Y graves in 1926 have the panels white or plain (reserved), with-
out any contained pattern. The circles seem allied to the sunlike and flowerlike
raying forms such as 12c, e; or they may contain such: 15e, 16c.

5. Small white dots in: (a), spatter, 12b, 13c; (b), rows, 12e, 13f, 15f, 16d in
dark stripes, and cf. 4b; (c), somewhat large white disks or circles, 13d, 17d.

6. Multicolored stripes grouped into a rectangular panel, sometimes with black
dashes or spatter: 13b, 17e (12e?).

7. Juxtaposition of rectangular panels containing contrasting patterns of simple
geometric elements, such as checker with zigzags, stripes with dots: 16a, f. This
manner has some precedent in late B.

8. Rows of eyed diamonds: 16a, 9d.
9. Step-frets: 16c.

This is an impoverished array of design compared with preceding Nazea A-X-B,
but some of it is patently derived from that, and some of the rest might be de-
rived by shrinkage selection.

COAST TAHUANACOID ("Y2")

I merely list the illustrated vessels in which a highland influence is stronger than
the Nazea remnants or persistences.

Jars, spherical, with or without lugs: 14a, b, 16e.

Jars with longish tapering spouts, with human faces and strap handles or with-
out: 19a, c, 9k, perhaps 91.

2513b is hard to place, though there is design similarity to 17e.

2 The shape is Tiahuanacoid, the painted design element resembles that of 17c, which I now
class as Nazea A. See below.

376      University of California Publications in Am. Arch. and Ethn.

Lenticular-bodied tapering double-spouts, spread at least 600: 19d, e, 20d.
Bowls painted outside: 14d, f.

Plates, rim-painted inside: 14c, perhaps 14e.

Probably the side-mouthed jar 15f; probably not 13b.

The designs vary greatly in the distinctiveness and purity of their Tiahuana-
coidness.

DISSOLUTION OF Y3

Of the several supposedly Y3 pieces on the 1927 plate 17, one, 17d, has already
been transferred to true Nazea-Y style, the old "Y1," on account of its painted
design resembling that of 13d, f,' as well as because of its brownish color. Double-
chambered vessels are of course characteristic of northern Peru.28 The bird jar
17e has also been mentioned as akin to the strap-handled face pitchers so typical
of Y.

The 4 other vessels of plate 17 I now consider, in the light of further exposure
and experience, not only Nazea in our old narrower sense of A-X-B, but specific
Nazea A. My colleague Dawson and I sounded each other out recently on the at-
tribution of the mammal and bird figures 17b and f and agreed that they probably
are from quite early in the Nazea development-A or even proto-A. I was helped
to this idea by having found one or two analogues in my Nazea excavations; he
deserves credit for coming to the conclusion through unaided stylistic sensitivity.

Granted 17b as Nazea A, 2 unillustrated pieces, a bird jar (8423) and a white
spiral whorl-modeled bowl (9007),29 plate 37f, of the present paper, fall in line as
also Nazea A; and so does, probably, the double-bowl of 17c. Already in 1927 we
commented on the "Y3" tendency toward a white background for fanciful pieces;
this tendency I would now transfer to early A.

The quail-bodied head-and-spout 8423, just referred to as not Y but probably A,
has its body black, the breast speckled with gray, the underside white. The beak
is white, the head black, the eye has elongated concentric stripes of gray, orange
brown, and white surrounding a circular white pupil with black center. The wings
have a gray rectangular body, and gray feathers. The right wing is barred red,
white, gray; the left white, red, gray; these bars being outlined in black and
crossed by four black lines. On the black upper side of the tail are five transverse
wiggly zigzag lines, respectively white (proximal), orange-brown, gray, deep red,
white (distal). The spout and bridge are deep red, almost purple. The texture is
good, the painting firm, the colors saturated, and Y thinness and sloppiness of
pigment are altogether lacking.

How did we come to regard this group of A vessels as belonging in a phase of Y?
I think it was because we drifted into the assumption that A was simple, severe,
and limited to few shapes. Therefore all free figure modeling was non-A-in fact
seemed at the opposite end of the scale, B or even Y. Thus we credited the 3 figure
jars and the spout-and-bridge figure of plate 17 to Y3-as well as the head jar 7a

` Commented on in 1927, p. 29.

2 By Nazea-Y times Mochica influence may well have been occasionally reaching the Rio
Grande Valley. Compare the stirrup-mouth 19b, which is polished R, Y, B, and W, well painted.

`9 The bowl is chalk white, with a black line (both on the inner and the outer side) in the
groove between each of the convexly raised whorling spirals. The bottom is completely flat (ex-
cept for the swelling of the whorls); the sides rise at about 700; D is 103 mm., H 60, percentage
37-basic shape E.

Kroeber: Toward Definition of the Nazca Style          377
and the man jars 9a, b, c, j to period B; all of which I now see as A. That modeling
occurred in A as well as in X, B, and Y has long been plain from the Uhle
Ocucaje vessels 26a, b. And Paracas is certainly rich in successful attempts at
freely modeling individual effigy forms; so that impulses in the same direction in
Nazea A should not be surprising.

There remains the remarkable fish-tailed man or god cylindrical jar 15a-b. This
is altogether too fine in modeling, texture, and painted design to fit into Y. Nor
is it A. The curled-over rays and jagged staves of the headdress definitely place
this piece in period B.

SUMMARY OF RECLASSIFICATION oF NAZCA SHAPES

A, inside-painted plate, formerly included in A, "angular bowl" and in C,
"conical bowl." There are at least 3 subclasses.

Ac, circular center panel painted (rim may or may not have simple radiating
lines). The style of painting is Nazea A. The height ranges from 28 to 45 per cent
of the diameter, mean 38 per cent. There is usually a bevel between center panel
and rim (shape B in bowls) but some are sphere-segments (shape GG).

Ar, rim painted. A less well-defined and less homogeneous group than Ac, pos-
sibly derived from it. It is at any rate from periods B and probably Y.

Aw, whole upper (inside) surface painted as unit; no bevel, profile uniformly
convex, plate shallow, H/D = 32 per cent. Period, B or Y.

B, gambreled bowl, formerly "point-bottom bowl." Painted outside, like all
other bowl types. There is an angle in profile, sharper than that in shape A, so
that the slope of the bowl side approaches verticality and makes a convenient and
effective though narrow, ribbonlike design field, especially for repetitive elements.
The H/D ratio is 44 per cent. Period, mostly A, but carrying over into X-B.

E, flat non-flaring bowl, formerly "angular bowl." Bottom relatively flat, side
sloping up essentially straight in profile, with not more than nominal concavity
or convexity. This is a well-defined class, consistently A in style of painting. The
H/D ratio is around 48-50 per cent, with range from 37 to 62 per cent.

D, flat-bottomed flaring bowl, formerly "shallow bowl." Flat-bottomed much
like E, but the sides show a concave profile, and they are lower than E, around 41
per cent in the mean. These tend to be large for bowls. This is an infrequent shape.
Those of the designs which are distinctive are fully developed or late B.

(C, residual from old class C, "conical bowl," after subtraction of subclass Ar
plates. Concave and flaring-sided, bottom roundish without bevel.)

Cl, non-representational geometric design on outside; underside of bottom also
painted, usually in quartering. H/D mean, 44 per cent. Period B, mostly late B.

C2, bottom unpainted, sides zoned, the lower zone often with yellow faces in
series. H/D means, 47 to 52. B to late B to Y.

F-H, high flaring bowls, formerly F, "cup bowl," and H, "flaring bowl," form a
numerous class of rather wide range of shape. The bowls we ascribed to type H
average somewhat larger than F, somewhat lower and broader in H/D ratio (65
per cent as against 72 per cent), and, as a function of this last feature, the depar-
ture from vertical of their side slope (or spread in profile) is greater, as expressible
in degrees of arc. The absolute depth of concavity of the side is about the same. In

378      University of California Publications in Am. Arch. and Ethn.

H/W ratio, about half of all these bowls fall in the overlap area of their distribu-
tion; about a quarter are "pure F," that is, above the maximum of any bowl origi-
nally put into H; and a quarter are "pure H," below the minimum for F. Still,
for half of the group, I have found no tangible or definable criterion on which to
base ascription to either F or H; and it proves that at least twice we assigned one of
a pair of essentially identical bowls to F and the other to H. Until someone succeeds
in finding a diagnostic differential, it will be most correct to deal with the group as
undifferentiated F-H. They are flaring bowls higher than D, Cl, C2, and higher
also than the non-flaring B and E shapes.

Most F-H bowls are from period A, but they also run over into "X," that is,
late A and early B.

G, formerly "straight bowl," was based on only 5 examples, which prove too
heterogeneous to form a valid or at any rate useful class.

GG, sphere-segment bowl, new class, has an even rounding of surface upward.
It occurs somewhat scatteringly, beginning with 2 negatively painted bowls of
Ocucaje grave Al, and contains some positively painted vessels (on inside or out-
side), most or all of which are also A in design phase. GG differs from the plate Aw
in that the latter, painted on the inner side, is flatter than GG.

I, angled goblet. Lower half or third cylindrical, sometimes even contracting
slightly; upper portion in a nearly straight flare of 20?-25?, sometimes more. The
height is less than the rim diameter. This form seems to be limited to the transition
from late B to Y. Some Y forms have the bend accentuated toward a right angle;
see AGY below.

J, waisted goblet, former "goblet." A slightly flaring cylinder, somewhat con-
stricted in the middle, with a long concave curve, without angle. The rim diameter
is largest; the height greater than it. Period B. Suggested subclasses:

Ji. H/D 90-110. Differentiated from F-H bowls by greater height and less flare.
Early or middle B ("X").

J2. H/D 130-140, or in the range of vases. Differentiated from bulbous-concave
vase P by flatter bottom and by the waist being near the middle instead of in upper
part. Full or late B.

?J3, near-bowl goblet. Definitely waisted, 75-85 per cent of mouth; low, H/D
90-100 per cent.

K, double-convex goblet, former "double-curve" goblet. In profile, a moderately
but conspicuously constricted middle separates a convex lower and a convex upper
half; the upper continues into a slight turn-in (0-9 mm.) at the rim. Either the
upper or the lower bulge may have greater diameter; if it is the upper, the vessel
is relatively squat (H/D 96-129 per cent); if the lower, it is tall (H/D 134-149
per cent). The design comes more often than not in two or three zones. Full B
period.

L, conical goblet. An inverted cone frustum, with relatively flat bottom. The
sides are reasonably straight, but tend more often to the concave than convex.
H/D ratio 97-130 per cent. Late B and Y.

M, tapering vase, former "small vase." Leans both toward J, waisted goblet, and
O and P, the 2 bulbous vases. It resembles the 2 latter in over-all proportion and
in the maximum diameter coming low down. It resembles J in usually appearing

Kroeber: Toward Definition of the Nazca Style        379
to be constricted, though this is an illusory effect of merely concave profile in the
upper half. Painted design may be vertical, staggered, or zoned. Period B, as
ascribed in 1927.

N, cylindrical vase. The tallest and slenderest Nazea shape, with H/D per-
centage more often above 200. The mouth is usually slightly larger than the diam-
eter below. When the reverse holds, N resembles P except for being slenderer.
Period B. Suggested subclasses:

Ni, as described, essentially cylindrical. Painting one- or two-zoned.

N2, three bulges and two concavities in profile, with design in four zones, fitted to
these curves. Probably begins later in B than Ni, and derivative from it.

0, bulbous-convex vase, former "bulbuous vase I." The diameter is greatest in
the lower half, whence the "bulbous" effect. The general profile is convex. This
may turn slightly concave just below the rim, but without actual diminution of
diameter below that of the rim. The H/D ratio is around 125 per cent with the
amount of variation usual in the Nazea style. The painted design may be zoned, but
tends to be vertical. Period B.

P, bulbous-concave vase, former "bulbous vase II." Differs from 0 in that
toward the top there is a slight but conspicuous constriction of diameter, above
which the lip spreads a little. There are 2 subclasses, differing in time.

Pa, in period A, H/D 90-105 per cent.

Pb, in period B, H/D 125-150 per cent.

The time spread of these 2 subclasses may be what led us in 1927 to ascribe
shape P to phase X.

It would be possible to recognize a third subclass:

Pn = Pb with modeled nose (and sometimes ears) = Y1 as below.

Q, liples jar. Somewhat like the collared jars R, S, T minus their collar; though
the upper half tends to slope with rather minimal curvature. The mouth diameter
is from 50 to 60 per cent of the body diameter. The period for painted examples is
certainly B prevalently, though we set it as X in 1927. A very similar shape occurs
in A but unpainted.

Qp, pseudo-Q, really a convexly incurved bowl.

R, wide-mouthed jar, and S, narrow-mouthed jar form a continuous series and I
suggest they be distinguished, when convenient, by the arbitrary criterion whether
the neck diameter is more or less than half the body diameter.

R has a wide time-range, from A to late B. It occurs in A both painted with
design and unpainted.

S is rare in A. The extreme forms-Rim D/body D < 40 per cent-seem to be B.
T, two-handled small-mouthed jar, former "handled jar." This is shape S with
two small handles or lugs set in vertical plane far enough up on the vessel's
shoulders not to extend the body diameter. Like S, typically B period. (Unpainted
two-handled but wide mouth jars go back to A: either incised or as plain cookpots.)

TTT, three-handled jar, with constricting cambered mouth, and a characteristic
design painted on its roughish surface (Ocucaje 28i), is A period.

U, double-spout jar. Most frequent in A, progressively less common through B,
not reported from Y. There are 2 main subclasses, plus probably subvarieties of
these and occasionals.

380      University of California Publications in Am. Arch. and Ethn.

U1,0' ox-heart-shaped, the body usually a little less high than wide, but higher
if the spouts are included. The effect at any rate is slightly elongated vertically,
rather subglobular. The typical A form, continued, however, into X and at least
early B.

U2, lenticular, with varying degrees of flatness, edge, or bevel, and unrounding
of bottom. I raise the question whether these vessels may not have been made in an
upper and a lower half joined together. The spouts remain surprisingly like those
of Ul in the gross. U2 is characteristic of full or elaborate B; the time of its first
appearance remains to be worked out.

Cylindrical, hourglass, double-cone, rectangular, stepped, and other forms such
as double-spouts on effigies appear sporadically, but unless they recur as types they
need not be elevated into subclasses. These aberrant or experimental types seem
early in period as well as late.

The familiar Coast Tiahuanacoid double-spout has not been found occurring in
Nazea-style association, so far as I know, and seems to represent an invention in
another style, perhaps after impingement on this of Nazea B.

V, head-and-spout jar. There are several subtypes.

Vi, body more or less globular, man's head, short spout, short bridge. Period A.
(Vla, with spout rising out of head, and the bridge turned into a curved strap
handle, is so far represented only by 9j. Period A.)

V2, woman's (?) head, spout-and-bridge longish, body of vessel beveled, not
globular. Period B.

V3, body of vessel modeled into effigy of bird, musician, or the like. Probably
period A, continuing into B.

(V4, double-chambered, with bridge connecting effigy and spout, so far repre-
sented only by 17d, probably period Y.)

W, flaring rim jar. Woman's globular head or trophy head, mainly painted,
surmounted by flaring collar roughly as high and wide as the head. Period B.

W has affinity in concept with shape Y, and also with R.

X (human) figure vase. There are almost inevitable ties with V, which also
represents human bodies; but shape X dispenses with spout-and-bridge and has a
simple, large opening on top; which fact, in turn, makes a link with W." Except
in the effigy subclass, modeling is fairly carefully executed for the head, but
sketchy for the seated body (sometimes knees only), details being done in painting.
Subclasses are:

Xi, seated man.'2 Usually A period.

X2, seated woman, carrying receptacle. B period.

(X3, whole figure more or less modeled. Plate 15a, b is standing, has a fish tail,
and "carries" a wide collar; but other freely modeled or imaginative effigies can be
expected to turn up.)

Y, head vase or jar, formerly "head jar." I alter to "vase" because of the close
relationship in shape between the most common form of head vessel to the bulbous-

30 I use Ul, U2, not Ua, Ub, because the Ul form runs on from period A into period B.
81 Cf. the wide collars of 8a, shape X, and 8c, shape W.

Xm, Xw, Xe might also be used for Xl, 2, 3, (man, woman, effigy).

Kroeber: Toward Definition of the NazCa Style       381
concave vase P. Other Y forms relate to Q and R. Mainly, it is the applied paint-
ing that makes a vessel Y.

Y1 is a P vase with modeled nose and sometimes ears added. Synonym, Pb plus
n, Pn. Period B. The lower Pa of period A seems to have had no shape-Y counter-
part.

Y2, a somewhat corresponding modification of wide-mouth jar R, but with three
medallion faces (Y2 = Rn). The period has not been determined.

Y3, modification of lipless jar Q (by addition of modeled nose): Y3 = Qn. The
presumable period is B.

Y4, turbaned head, approaching life size. Period A. There is no such form
relationship to P, Q, R as in the 3 preceding, but rather a suggestion of an enlarge-
ment of part of shape X, human figure vase.

Some Nazea-Y Shapes

The following are non-Tiahuanacoid Nazea-Y period shapes that have been de-
scribed above. I suggest for them capital letter designations ending in "Y," in
which the letters preceding Y have no relation to the arbitrary system of nomen-
clature so far used for the shape classes, but are the initials of terms descriptive of
them; this procedure to prevent confusion between the non-Y classifications de-
veloped in 1927 and the partial classification for Y developed above.

FPY, face pitchers, of Y period ("a" above, under "True Nazea Y, form-
erly Y1").

CJY, cylindrical-collared jars, with or without lugs or handles ("b" above).
WJY, wide-mouthed jars, perhaps developed from A-X-B shape R.
DBY, deep bowls ("d" above), with circular and oval subvarieties.
OBY, oblong bowls, rectangular ("e" above).

AGY, angled goblets of Y period ("f" above), an exaggeration of angled goblet I.
Other shapes of period Y remain to be isolated and defined.

Grouping of Classes by Style Phases
Limited to Nazea A:

Ac, Pa, TTT, Vi, V3, X1, Y4.

Nazea A and continuing into "X" or B:

B, F-H, GG, (P), R, Ul. (Q, S, without design in A.)
Nazea B:

C1, D, Ji, J2, K, M, N, N2, 0, Pb, Q, S, T, U2, W, X2, X3, Y1, Y3.
Late B to Y:

Ar, Aw, C2, I, L, V4.
Nazea Y only:

FPY, CJY, WJY, DBY, OBY, AGY.

PART III: DISCUSSION OF VESSELS ILLUSTRATED

INTRODUCTION

This section contains photographic reproductions of 105 ceramic objects in the
Nazea style, together with a piece-by-piece commentary and interpretation.

The first four plates show vessels from Ocucaje of the same "straight-A" phase
of the Nazea style as Strong and I (mainly) described in our 1924 monograph
on Uhle's pottery finds from Ica Valley. The vessels figured here now in plates
31-34 might have been included in the 1924 monograph but were not. They further
define and affirm the A phase. Three of these vessels (pls. 31a, d; 32a) are included
by courtesy of the Peabody Museum of Harvard University. Although collected by
Uhle, they were exchanged about 1905, when F. W. Putnam was Director of both
Museums.

The fifth plate (pl. 35) shows vessels from two stylistically aberrant graves,
numbers 3 and 9, at site F at Ocucaje. Although the rest of the graves at site F are
straight A, I interpret these two as subsequent, and provisionally denote them as
"late A."

The last eleven plates, 36-46, are all from the collection made for the University
by Uhle in Nazea or Rio Grande Valley, which was the basis of the Gayton-Kroeber
1927 analysis. With the exception of two or three graves dug by himself, Uhle
secured this material from either huaqueros or hacendados-pot-hunters or plan-
tation owners, in English. He did associate considerable parts of the collection
with localities, mainly haciendas, which means he had reason to believe, or perhaps
felt certitude, that such vessels were excavated at the localities specified to him.
However, without actual grave attributions, associations of objects are mostly
lacking, and their bearing on stylistic problems is therefore more over-all, indirect,
or conjectural than that of the grave-lotted collection from Ocucaje.

The vessels in these eleven plates might have been included in the Gayton and
Kroeber monograph of 1927 on Nazea. They now supplement that earlier publica-
tion, adding 73 vessels to the 143 photographically represented there.1 The addi-
tions have been selected primarily with a view to illustrating shape classes under-
represented in the earlier work; or new subclasses established now. Beyond that,
the pictures have been chosen for their bearing on matters of painted design and
phase of style development. They range from Nazea A to Nazea Y.

Colors are abbreviated as follows: B, black; Br, brown; G, gray; 0, orange;
R, red; V, violet; W, white; Y, yellow.

Plate 31

Ocucaje: Phase A
Graves Al, B4, 5

31a: Peabody Museum 73864, by exchange, original Uhle 4489 from grave Al.
Shape E. Diameter 144 mm. Design: hummingbirds. The feet are very short and
almost turned upward. Each bird's beak nearly touches the wing and overlaps the
tail of the one before it.

l Nazea style, A-Y, 112; style of Ica, 12, foreign style from Nazea, 7; total Uhle from Nazea
region, 131; American Museum, 12; total, 143.

[ 382]

Kroeber: Toward Definition of the Nazca Style          383
This specimen completes the illustration or description of the vessels attribut-
able to grave Al as it is defined above (Pt. I, "Qualifications on 1924") as a true
burial distinct from non-grave vessels and objects at site A. The Al grave lot com-
prises:

4489-73864, the present bowl.
4490, plate 29c.

4491, B bowl, zigzag band, alternate up- and down-pointing triangles Dark R
and Light R; profile and outline of design shown in figure 7a below.

4492, B bowl, made by reduced firing (B all through paste), shape E, figure 7b.

Al- 4491

a

Al- 4492

b

Fig. 7a, b. Two bowls from Ocucaje grave Al: a, 4491, gambreled B bowl

with design; b, 4492, E bowl, blackware all through.

4493, negative-painted; plate 29f.

4494, negative-painted bowl, similar to last, also shape GG, vertical stripes out-
side, blurred on bottom: figure 5a.

[4495, specified as not from this grave.]

4496, owl jar, rough ware, light buff, unpainted, incised; plate 29d.

4789 [sic, correct, from grave Al], painted with varicolored short stripes,
plate 29g.

In summary, the grave contained 7 bowls and 1 modeled jar. Of the 7 bowls, 1
is B, 2 are negatively painted, 4 positively painted-2 in more or less regular
Nazea-A style, 2 somewhat aberrant; the jar is unslipped, unpainted, incised.
Compared with Nazca generally, the genre is Nazea A. But compared with the
rest of Nazea A, the contents of the grave are only Nazcoid, and must evidently be
construed as earlier than straight A-"proto-A" or still transitional from Paracas.

384      University of California Publications in Am. Arch. and Ethn.

31b: B5-4649; profile shown in figure 3a. Shape A, center-painted plate-bowl.
D 175 mm., H 60, H/D 34 per cent. B and Br on W, rim R. The design appears to
represent a vegetable, and occurs chiefly on plates. See figure 4a, b for similar
designs in duplicate instead of quadruplicate, also on Ocucaje class-A plates.
Martin-Vegue 44-A-4 is related. Also from grave B5 are the vessels c, d on the
present plate; also plates 25h, 25c, 26b, 26f (B5-4644-45-46-48); and an unde-
scribed bowl in Peabody Museum, original Uhle number 4647.

31c: B5-4643. Shape E, flat non-flaring bowl, unusually large. D 237, H 129,
H/D 54 per cent. The cat design is simple in paws, tongue, whiskers, lack of mouth
mask, and "centipede" body. Similar painting occurs below on F-H bowls 42a, b.
Maskless cats appear in Seler figures 20, 24, 32, the centipede body in 58-69.

B4-4638

Fig. 8. Design on E-shape bowl, period A, Ocucaje.

31d: Peabody Museum 73857. Originally Uhle B5-4647. Shape B or E. D 133.
The bird is unusually simple and "archaic." It may be resting on the water, like
Seler figure 310.

31e: B5-4646. The 4 pieces 31b-e, together with plates 25c, h, 26f, now make
all 7 vessels from grave B5 available in illustration. Shape D or F bowl. D 116 mm.,
H 66, H/D 57 per cent. The eight figures (colors Br, R, Y, B, G, R, Br, G-on B)
presumably represent lizards-as in 27t, F17-4759.2 Some have the hind-foot toes
turned backward, some forward as if they were frogs. The tails, too, curve some
backward, some forward. The vessel is unusual in being painted over its whole
outer surface-as in 33a, 37b, 38a-f, although the painting on the bottom is pretty
well confined to tails. The freedom with which these are interwoven contrasts with
the schematic stiffness of the bodies.

This piece was previously shown on an inadequate scale in 1924, plate 26b.

31f: B4-4640. Shape Pa, bulbous-concave vase. D 173, H 179, H/D 103 per cent.
This is the Nazea-A shape out of which the later Pb developed. Cf. plate 34a below,
and 25h. There are five units, alternately Y and G, on Dark R background, in
the horizontal trophy headband of design.

The pottery of grave B4 is now completely described. Besides the present 31f,
text figure 8 shows bowl 4638. The other numbers from 4635 to 4642 are illustrated
in Kroeber and Strong's original Ica report of 1924: in sequence, 25f, 27k, 26i,
26d, 29b, 29a.

2 27t has six figures, five of which I construed as frogs, one as a salamander. But salamanders
are rare in South America, and I doubt if any occur on the arid Peruvian coast.

Kroeber: Toward Definition of the Nazca Style        385
Of the present 6 phase-A vessels, the only one with light background for the
design is the plate-bowl b (and this has its rim R) ; c and f have their designs out-
lined in B; the other 4 are unoutlined, except for some W inner borders, toes, and
eye rims of the lizards in e.

Plate 32

Ocucaje: Phase A

Graves Cb, F4, 8, 18, 20, 22
B and E Bowls, A Plate

32a: Peabody Museum 73865, original Uhle 4654 from grave "b" of Ocucaje
site C. Shape E, flat non-flaring bowl. D 165. For the design compare 1924, plate
26c from grave F2; also plate 33b below.

32b: Cb-4651. Shape E. D 164, H 70, H/D 43 per cent. Four colors: R, Br, B
on W. Water-bird design; compare 1924 plate 27o, 1927 plate 3b, and Seler
figure 295.

Grave Cb contained 7 vessels, 4651-4657. These are now all illustrated: 32a, b
here; a class-A plate as drawn in figure 9a, with defaced design of a crustacean
and fishes and 27e, i, v, 28i.

32c: F22-4778. Shape B, gambreled bowl. D 155, H 85, H/D 55 per cent. There
are ten units in the design band, each bordered (or divided and paneled, as one
will) in W, the upper right half B, the lower left half with this color sequence:
R G O-R G O-R O-R 0 (R can be construed as Dark R, 0 as Light R).

The "oblique" steps of the design show that these oblique and "penetrating"
forms of the step and step-fret are not necessarily late derivatives from a "natural"
restrained and rectangular step; this vessel is straight-A phase like all others in
plates 31-34.

Grave F22 contained 1 other bowl, 27c.

32d: F8-4718. Shape B, gambreled bowl. D 195, H 100, H/D 51 per cent. The
design here is of eight fishes, Dark R and Dark Y, with B and W on B background.
This form of fish recurs in Nazea A; cf. Seler figure 336. The B longitudinal line,
mouth, and diagonal stripes, together with W anterior underside and tail edge,
give the fishes a curiously cut-up effect.

32e: F20-4772. Shape B, but approaching GG, hemispherical bowl; for which
see figures 2, 5, 6 in Part II. D 135, H 72, H/D 53 per cent. There are twenty-two
repeats of the design unit, including the crowded ones at the left of the photo-
graph; apparently eleven of these are Dark R, some defaced. The heavy B contin-
uous zigzag line can be seen as dominating the pattern, with the varicolored
unbordered triangular figures serving as filler. Or again, on account of their area,
these figures may be primary, and the zigzag serves chiefly to accentuate their
alternation. I have no idea what the figure represents.

32f: F18-4764. Shape B, gambreled bowl, but the bevel is not pronounced. D
200, H 95, H/D 48 per cent. There are eleven repeats of the design unit outlined
in W on B background, the contained arcs R and Y in alternate units, in the
eleventh partly R, partly Y; the panel dividers are Y-Br. The denotation of the
unit is unknown.

8 Another crustacean, suggestive rather of a river-mouth prawn (ca'nar6n) than of a lobster,
is shown in fig. 9b. It is on the side of a bowl, F1-4676, H 108, D 195, H/D 55 per cent.

386      University of California Publications in Am. Arch. and Ethn.

a

Fl - 4676

b

Fig. 9a, b. Crustacean designs on interior and exterior of period-A bowls, Ocucaje.

Grave F18 contained 4 bowls: this 32f, 33c, 27o, and 4766 whose design band
contained seven units and one condensed unit of a fruit or seed of type of Seler
figures 396, 397 (palm fruit?), alternately R and Y-Br, bordered in W on B
background, with W panel dividers.

32g: F4-4693. Shape A, center-painted plate. D 159, H 49, H/D 31 per cent. The
pairs of short W lines across the Dark R rim border do not seem to recur in Uhle's
phase-A material from either Ocucaje or Nazea, but I collected plates with these
rim lines repeatedly in my Nazea excavations of A graves in 1926 for Chicago.
The center star is B, bordered with W, on Dark R. A ten-pointed star is not easy
to draw freehand, and this potter was not too successful as regards symmetry. Also
her ten points and fifteen pairs of rim lines do not match too well. However, the
idea of ten rays seems to have taken hold at Nazea.

Kroeber: Toward Definition of the Nazca Style        387
Of the 6 bowls in plate 32, 3 have a light background, 3 R or B one. Borders or
outlines of figures tend to be B or W in contrast with the background; but in 2
vessels, b and d, there is no outlining. Vertical lines serving to divide or panel
design units come B in a, W in d and f, can be construed as W in c, and are lacking
in b and e.

Plate 33

Ocucaje: Phase A

Bowls from Graves F4, 6, 14, 18

33a: F4-4695. Bowl painted over bottom as well as sides. It is shown inverted,
bottom up; the whitish streak across it is a mend, not the near rim. Shape E. H 116,
D 45, H/D 39 per cent. Background B; the design of W, R, G, Light R (and Br
or 0?) annular disks is approximated again in 35a and f.

33b: F14-4754. Shape F, concave and round-bottomed. D 134, H 87, H/D 65
per cent. Design, eight fruits (?) somewhat similar to plate 32a, successively G-V,
Y, B, W, B, Br, B, W, bordered and paneled (divided) by R vertical lines.

33c: F18-4765, another side shown in 27w. Shape E, side slightly concave,
bottom flattish. D 147, H 76, H/D 52 per cent. The design is of hummingbirds,
some with beak in front, some with head reversed; the present photograph shows
two birds on their backs ( !), one with beak forward, one with head turned to rear.
All the birds have a tail with lance-shaped widening at the end of a long stalk.

33d: F14-4753. Shape F. D 193, H 120, H/D 62 per cent. The design, confined
to the upper half, consists of nine parallel zigzag lines, in five colors, namely, down-
ward, W, 0, W, R, Light R, W, 0, G, 0 on B background. There are fifteen double
zigzags in the circumference. The band of "color exercise" is 63 mm. wide-a bit
,znore than half the total height. The varicolored zigzags are not far from 29g,
A-4789, except that there the encircling zigzag lines are broken up into discon-
tinous, slightly rounded chevrons in stacks of eight. In both cases the simplicity
of pattern and variety of colors used suggest the effect of a painter trying out
his palette of colors.

Somewhat similar is F5-4703, an E-shape bowl, D 127, H 57, H/D 45 per cent,
with W, R, W, Y, W encircling zigzags on B background.

33b and 33d were the only vessels found in grave F14.

33e: F6-4710. Shape E, flat-bottomed bowl. D 131, H 59, H/D 45 per cent. The
upper part of the outside is W, the lower part and bottom are B, with thirty-two B
lines rising to the rim to divide the area into as many W finger-shaped panels. This
is effective decoration; it can hardly have "represented" anything even sym-
bolically. The B seems painted over the W. Compare the following figure; and, for
all W, plate 37f.

Grave F6 contained also a P vase, 4709, plate 34a; a double-spout, 4706, plate
34c; the figure jar 4708 of 25a; bowl 4707, 26h; and a bowl 4705.

33f: F4-4699. Shape F (?), near E, side concavity very slight. D 147, H 77,
H/D 52 per cent. W outside, B base, thirteen thin R vertical lines dividing the W
into empty panels. In contrast with the preceding, these are square-cut, not
rounded at the bottom.

Grave F4 contained plates 32g, 33a, 33f; the present figure la, above (4701,

388      University of California Publications in Am. Arch. and Ethn.

lipless jar of shape Q, without design, which is a mate of F7-4717, figure lb) ; 25e,
27u, 28g; and nos. 4692, 4696 (plate, mate of 28g), 4697, 4700, and 4702.

This last, 4702, is an unincised, two-handled cookpot of D 140, H 111, MD 96,
Neck D 85, D inel. handles 162. The proportions are: H/D 79, M/D 69, Neck/D 61,
Handles/D 116 per cent.

Plate 34

Ocucaje: Phase A

Various Forms from F2, 3, 6, 10, 13

34a: F6-4709. Shape Pa, bulbous-concave vase, similar to plate 31f. D 156,
H 139, H/D 89 per cent. The design is much simpler than 31f. A B band encircles
the middle; above this is a zone of thin vertical lines, thickened in the middle,
in sets of six colors separated by one B. Below the B "belt," the color is R slip. This,
with 31f, illustrates the low-vase shape Pa out of which the post-phase-A tall vases,
Pb and perhaps others, evolved.

34b, 1, 2, 3: three pottery whistles from grave F10: 4733, 4734, 4824 (sic, cor-
rect). All have a flange which is perforated for suspension, since the whistles are
both small-two inches or less in length- and light for handling. All three still
blow.

34bl, 4733 (top, left), represents a person stretched out, with head half up and a
conical cap. It is reddish Br. The end of the mouthpiece has been broken off; 47 mm.
remain.

34b2, 4734 (bottom, left), represents an animal, probably a mammal, perhaps
a dog or sea mammal, probably not a bird. It is reddish and polished, 55 mm. long.

34b3, 4824 (right), but also from grave Fl10, has a large human face occupying
most of the surface that bears the vent; the perforated flange is thus on the back
of this head instead of ventral as in the two preceding. The length is 45 mm.;
maximum breadth at ears now 19; before the breaking off of one of these, about 20;
at level of jaw, 17. The modeling is rubbery, as if done on the outer surface with
the fingers only, where bl and b2 have been finished in spots with a tool. Also, this
piece differs in being partly painted: dull off-W or light muddy Y for the centers
of eyes and mouth; a dark brownish R around the W of eyes and mouth, also to
indicate tiny arms and fingers on the two sides, from about the level of the jaw
down; and a Light Br for the remainder of the face. The nose is pinched up, the
ears project upward; otherwise there is little modeling. The face suggests mon-
strosity in its viscid formlessness; there is no trace of cat-deity symbolism;
although the dark border of the eyes includes a vertical bar above and below,
which is an early Nazea style of eye-paint.

Somewhat similar is the whistle g on this plate, F3-4689, though grave F3 ap-
pears to be somewhat later that the other F graves. This piece is also crude in the
softness of its plastic effect, flattish, with modeled nose, painted with W in part;
but there is a headdress instead of ears, and the length is about equally divided
between legs (where it was blown), body, and head.

We have so little on Nazea-style whistles that I might mention another, though
it is without data, having been given to me in 1926 as a personal souvenir or gift
token in the manner customary in Peru. It is now number 16-7852 in the Univer-
sity Museum. It is really a flageolet, having two parallel blowholes and two finger-

Kroeber: Toward Definition of the Nazca Style        389
stop vents, producing two (not three) tones. It is modeled into a bird (possibly
a dove), head projecting upward, body rounded flattish, tail flat and blown into.
There are three clear perforations from top to bottom, one in front of the throat
for suspension, two in longitudinal line down the tail, probably for suspension
also, at any rate not affecting the tone when stopped. The plan, modeling, finish,
and W and dark painting are a bit more sophisticated than in the preceding
Ocucaje pieces, but this one might well be of the same A period.

34c: F6-4706. Smallish double-spout, spheroidal except for being flat-bottomed.
It could not be designated as shaped like an ox heart. D 133, H of body 117, H to
top of spout 156, body H/D 88. The B-bordered, crested head, neck, and sketchy
body of the painted birds are of striking W. The wing is pointed, the tail a B
oblate circle, legs are lacking.

34d: F2-4678. Miniature head-and-spout, only 59 mm. high to top of spout;
D 48, body H 44, H/D 92 per cent. The decoration of the spherical body is B-
centered W disks on the R background. Mouth and eyes in the face are similar
disks. The only real modeling is the nose.

The two other vessels in grave F2 are shown in 26c, 28j.

34e, 34f: F13-4751, F13-4752. Two unpainted, incised, two-handled cookpots.
See Part II, footnote 16.

34e, D 188, H 112, H/D 60 per cent; D of mouth 144, or 77 per cent of body;
D of neck 142.5; H of collar 15. The cylindrical handles drop from the rim to well
above the body maximum. Between the handles on each side, just below the neck,
are five elevated rectanguloid bosses, each crossed by nine to eleven thin vertical
incisions. Between each two bosses, there are about five (sometimes only three)
longer incisions. The vertical inner rim is incised with a zigzag line, the triangles
below which contain a number of short vertical scratches or jabs with a point-
giving the effect of facing, alternatively rough and smooth triangles.

34f, D 250, H 221, H/D 89 per cent; rim D 165-170, neck D 127-130, Rim/body
D + 76 per cent, Neck/body D 58 per cent; width of top of sloping collar, ca. 40
mm. The two strap-handles, 20-22 mm. wide, are set down the shoulder, a little
above the maximum body diameter. The rim lip is concave, sloping, and about
40 mm. wide on top. The incisions on the shoulder and side, and the slashes within
their area, were made with a sharp edge and are narrow. Their pattern is like
that of the inside of the collar of 34e, but on a much larger scale and more hap-
hazard.

These 2 vessels are of interest as being, with 28f from F10 (and compare also
the bird jar of 29d from Al), the unpainted, incised, formative prototypes of the
later Nazea painted and standardized jar shapes R, S, T. We have seen that the
wide-mouth R shape (M/D > 50 per cent) tends to be earlier, the narrow-mouth
S to be associated with full- or late-Nazea painted designs. This agrees with the
2 present pieces 34e, f, whose mouths are respectively 75 and 76 per cent of the
body-a true cooking pot proportion. The handles of course are functional in an
actual cookpot. As the jar became painted and decorative, the handles might be
retained (shape T) or more frequently dispensed with (R, S).

Besides 34e, f, grave F13 contained only 2 bowls: 27a, F13-4750, and its mate
4750A.

390      University of California Publications in Am. Arch. and Ethn.

34g, the whistle F3-4689, has been discussed above in connection with 34bl, 2,
3. It is well to recall that grave F3, with F9, is probably of later A phase than the
remainder of site F; see the discussion of plate 35.

Plate 35

Ocucaje: Phase A-plus

Graves F3, 9

The two graves F3 and F9 differ from the remaining ones at Ocucaje F in appar-
ently dating from a somewhat later phase of Nazea-style development, which I
provisionally designate as "A plus." It is certainly not B, perhaps not yet "X,"
but it includes shapes and designs somewhat different from those occurring at sites
Al, B, Cb, and most of F. The distinctness especially of grave F3 was repeatedly
pointed out to me by Lawrence Dawson as he helped me in finding, measuring,
and drawing specimens.

The first two figures of this plate, 35a and b, have been previously illustrated,
but on an inadequately small scale.

35a: F3-4681 is plate 27h. This F-H bowl has been touched on in Part I, under
"Qualifications." Its dimensions are: base D 87, waist constriction 85, rim D 135,
H 98, H/rim D 73 per cent. The design is more complex than straight A-phase
cat deities, and the long row of attached fruits, the spotted tongue, and the object
held in the left hand all seem post-A.

35b: F3-4683 is plate 28e. D 94, body H 86, H/D 92 per cent. The cylindrical
body is unparalleled among straight-A double-spouts. The annular disks or flowers
( ) on maroon background are also unique, on account of their size and "petaled"
effect; although disks without petaling occur in f of the present plate, and-
smaller in size-in 33a and 34d. In the present double-spout they are mostly
divided into seven "petals"-which would be difficult to execute geometrically,
but less so freehand; although eight- and nine-segment disks also occur, and a
small quartered one. Compare also 27r, F12-4746, with five-petaled disks.

The division of a circle into sevenths recurs in 9003, plate 37e (and in Martin-
Vegue, pl. 44-A-1, a seven-pointed star).

35c: F3-4684. This is a bulbous-concave vase, of subtype Pb. D (body and rim
both) 135, of waist 123, H 162, H/D 120 per cent. Compare this proportion with
the straight-A Pa proportions of 103 and 89 per cent in 31f and 34a. Also, there
is no straight-A vase with three zones of figured design as this has them. On the
other hand, the B background of the zones in this piece appears to be a reminis-
cence of earlier A. The sinuous dolphins shown have more flexibility than most
straight-A portrayals, but lack the basic rectangularity and elaboration of design
of characteristic later pieces. There are six to eight animals in each of the three
zones, in the order Light R, R, G. The B background has rows of minute W dots.

35d: F3-4682. A high bowl, perhaps assignable to shape F. D 151, H 116, H/D
77 per cent. The idea of zoning is incipiently present. The only figure design is a
series of twenty simple eight-pointed stars (four crossing lines), in the order:
O-R-B, the three together forming a repetitive unit. The first horizontal stripe
is R, the lower one 0 (or Light R).

Kroeber: Toward Definition of the Nazca Style       391
35e: F3-4686. Lower than a and d, but otherwise similar in shape. D 161, H
94, H/D 58 per cent. This is one of the few repetitive over-all patterns in straight-
A or A-plus Nazea-though see 27b, e, v; and, probably of later period, 28a. The
colors are R, Light R, B, W. The W disks with B center we have already encoun-
tered in 33a and shall see again in 35f. The overlapping double-arcs, with their
fish-scale effect, are new, and apparently unique for the Nazea style.

35f: F9-4727. Shape near E. D 103, H 60, H/D 58 per cent. The simple design of
large alternately W and R disks with B centers, and in a B zone, is striking. They

Fl12 -474 8

Fig. 10. Heavy S-shape jar, period A, from Ocucaje.

are not "petaled" (divided or radiated), but are connected by a W line. We have
already mentioned annular disks in 33a, 34d, and 35b; and might add 27r, F12-
4746, where the centers are W instead of dark and the sectoring is into fives.

35f is the only vessel from grave F9 shown here, although the 2 others in F9
are represented in 27n and 27x. Grave F9 seems less definitely divergent from
straight A than is F3.

The contents of grave F3 comprise 35a-e; the bowl 27b; the center- and rim-
painted plate 28o; another inside-painted bowl 4688; a pottery whistle, 4689, de-
scribed under 34g; and some beads of "turquoise and other materials," 4690. Not
one of these ceramic pieces, except the whistle, has a close parallel in any straight-
A-phase specimen; the closest to straight A is the design painted on 35c, though
the shape of this is post-A.

Supplementary on Ocucaje

Contents of Graves Fl, 4, 7, 12, 17, Bl, 3

Grave F12 contained 3 vessels: 4746-4748. The 2 first have been illustrated in
27r, 27q. The third, 4748, is a narrow-mouthed jar of shape S, which (see Pt. II,
"Jars R, S, T") I have been inclined to consider as more typical of late than of
early Nazea. The design has become faint and would be indistinguishable in a
photograph. Dawson was good enough to sketch it, and from this the formal draw-
ing in figure 10 has been made. The colors appear to be B on Dark R plus perhaps

392      University of California Publications in Am. Arch. and Ethn.

fugitive or post-fired painting. The "hooks" are reminiscent of those in plate 46c,
which in turn has a design almost identical with that of figure 8 except that in 8
the "hooks" are closed into loops.4

This jar F12-4748 is unusually heavy, weighing twenty-eight ounces. It suggests
highland ware in a certain stoniness. The dimensions and proportions are: H 142,
max. D 152, MD 54, neck 40 mm., H/D 93 per cent, Neck/MD 74 per cent, and the
crucial M/max. D only 36 per cent. It has a wholly flat base.

Grave Fl contained bowl 4676 with a prawn or crayfish design shown in figure
9b, which has been discussed in a footnote to the description of Cb-4651 in present
plate 32b.

a

B 1- 4624

Unpainted
b

Fig. lla, b. Period-A bowls from Ocucaje.

Figure 1 in Part II shows 3 lipless jars of shape Q, but without design paint-
ing, from graves F4, F7, and F19 respectively.

F4-4701 and F7-4717 (fig. la, b) are evident mates though they had been
placed in separate graves. They differ slightly in height. Both have been washed
with a purplish-R slip. The context of figure la in grave F4 has been cited in the
discussion of plate 33f.

The context of figure lb is that F7 held 3 other vessels, bowls 4714-4716, shown
in 27p, 27j, 26e.

Figure lc shows what might be classed with the 2 preceding as a lipless jar,
especially as it is coated somewhat like them, with a rusty R wash over its burnt
O or brownish paste. But it is curved inward where these slope inward like a cone
frustum, and in that respect it is rather pseudo-Q shape and reminiscent of the

4 B4-4638 of fig. 8 has five units of design on what I consider an E-shape bowl in spite of a
slight concavity of its side. The D is 124, H 72, H/D 58 per cent.

Kroeber: Toward Definition of the Nazca Style          393
modeled and painted cat-head bowl-jar of 46g, though this rises and curves in
more than figure lc.

The grave context of F19, besides 4770 of figure lc, was 2 bowls: 4768 and 4769,
shown in 26j and 26g.

Measurements have been given in Part IT, under "lipless jar Q," tables 7, 8.
Figure lla, b shows 2 bowls from B graves.

B3-4632, figure lla, is a shape-B gambreled bowl, H 81, D 196, H/D 41 per cent.
It is painted with twelve repetitions of a three-part horizontal zigzag, alternately
B and R on W, paneled. This design is similar to that shown in plate 37a (8692,
Nazca).

Grave B3 contained 7 vessels, 4626-4633, of which 3 are shattered and only 1
other has been illustrated, the inside-painted plate of 28h, 4630, H/D 27 per cent.
Specimen 4629, also inside-painted, is notable for its size, which when whole must
have exceeded 400 mm. diameter, and for what may have been a unique, bold de-
sign, B and R on W.

B1-4624, figure llb, the only vessel in its grave, is an unpainted R bowl notable
for its size: H 145, D 244, H/D 59 per cent. There is a trace of bevel or gambrel,
but it is not definite, and the piece does not fit too well into any bowl-shape class,
since it also scarcely flares with any continuity.

Uhle Ocucaje Grave Lots Completely Enumerated

and Described or Illustrated

(See text of explanation of plates)

Passage in                      Passage in
Grave        which described   Grave        which described
Al           31a               F7 .........  Ocuc. Suppl.
B .Ocuc. Suppl.                F9 .........  35f

B4 ........  31f               F12 .........  Ocuc. Suppl.
B5 .31b, e                     F13 .........  34e, f
Cb .32b                        F14 .........  33d
F2 .34d                        F18i S ...   32f

F3 .35a-e                      F19 .........  Ocuc. Suppl.
F4 .33f                        F22 .........  32c
F6 .33e

Plate 36
Nazea

Interior-Painted Plates of Class Ac

Herewith begins a series of illustrations of vessels from Nazea without recorded
grave associations, which will be analyzed for style of shape and design.

The present plate shows 6 inside-center-painted plates of "shape"-class Ac,
which are characteristic of style-phase A (at least in the larger sense). All of them
have a plain R border. The design is restricted to the circular central panel which
occupies about two-thirds to three-fourths of the total diameter of the vessel as
it appears photographed from above. The background of this panel in all 6 of
the pieces here shown is W or whitish. Four of them have the painted figures out-
lined in B; 2 are without outline. The W radiating lines crossing the margin,
usually in pairs, which I encountered frequently on shape-A plates in my 1926

394      University of California Publications in Am. Arch. and Ethn.

digging of A-phase graves for Chicago, are unrepresented in Uhle's material,
whether from Nazea or Ocucaje; except for the one Ocucaje case of plate 32g,
F4-4693, with the ten-pointed star, and here the central field is R and the bor-
dering is W; and Nazea 8466, discussed in Part II, table 9, footnote.

The sizes, proportions, and degree of bevel of the Nazea plates of this class are
given in table 9 of Part II. Figure 2a-f gives horizontal side views of the same 6
vessels in the same order as in plate 36a-f, except that 2b has not the shape-B bevel
or gambrel of the others, but is hemispherical, of shape GG.

36a was given me personally at Nazea as a memento or token in 1927. It is now
no. 54-41-10/34461 in the Peabody Museum. The design suggests an ear of maize
in husk with two snakes or worms emerging; I do not know what it actually rep-
resented.

36b is Uhle's 9093. It may represent a sprouting or growing tuber. I recall no
counterpart in Seler.

36c, 8704, is a full-face visage with an unusual amount of hair. One would ex-
pect it to be a trophy head, though neither eyes nor mouth suggest this. The
narrow mouth slit and the threadlike tongue recur in other figures.

36d, 8431, shows five simple pallar beans.

36e, 8592, shows better than 3g a high fish, perhaps of tuna type. This is a
common type, especially on plates: compare 28h (B3-4630); Seler figure 343, 344.

36f, 8523, shows four fishes, each generically similar to that of 36e but sharper
nosed and larger tailed, headed to the right, which is the less usual direction in
the Nazea style.

Plate 37
Nazea

Low Bowl Types B, D, E

37a: 8692. Gambreled (formerly "point-bottom") bowl of type B. See table 12
in Part II. The design element, B background, panel divisions all point to style-
phase A, during which this shape B was chiefly in favor. The odd half-unit of
pattern is typical. About ten elements were planned, but there proved to be room
only for nine and a half.

37b: 8558. Also shape B, gambreled, but the vessel is of style-phase B, as shown
by the bottom painted in quartering. See table 12 just cited. A poorer view of
this piece is shown in plate 5d and it is there classified as of shape B, though not
so listed on page 44 of the 1927 paper.

Other than plate 5d, we illustrated no shape-B bowls in 1927. But the 1924
Ocucaje report shows 26f, g, i, 27c, j, all straight-A-phase B shape.

37c: 8533. Shape D, flat-bottomed flaring bowl (originally "shallow bowl"),
whose sides flare considerably and are somewhat concave in profile, in distinction
from shape E; whereas the difference from F-H is that D is both lower and flatter
bottomed. We failed to illustrate the type in 1927-it is not very frequent though
definite enough; and there are no examples shown in the 1924 report on Ocu-
caje, since that deals with Nazea A, whereas the D shape is usually or always
Nazea B. The present piece shows the unitary (non-complementary) interlocking
fish or snake as design unit.

Kroeber: Toward Definition of the Nazcxa Style       395
37d: 8577. Also D shape. The painting is vigorous but roughly done, and I would
assign it to decadent B approaching Y. The birds shown are discussed below, in
comparison with Seler's birds, under plate 44c, d.

37e: 9003. Another D bowl, with considerable spread. The painting is somewhat
sloppy and hasty, of B to Y period, I would estimate. The larger stars have seven
and eight points with about equal frequency. Cf. plate 35b.

37f: 9007. This is basically of shape E, but its spiral fluting is unusual enough
so that we put it in the miscellaneous class "Z" in 1927. It is a striking piece with
its bone W relieved only by B accentuation of the grooves in its twist. This is the
piece referred to above (under description of pl. 33e) as one which I would now
withdraw from late Nazea and construe as early A phase.

37g: 9084. Herewith we come to straight shape-E, flat non-flaring bowl, which
we called simply "angular bowl" in 1927. The bottom is as flat as bottoms get in
the Nazea style, the side then slopes upward at an angle but nearly straight, with
little bulge or flare. See table 13 in Part II. E-shape bowls are characteristic of
Nazca A. The bird represented is what Seler calls "maize-bird" or dove (his figs.
286, 287).

37h: 8818. E-shape bowl, the bird's webbed (?) feet being carried onto the
underside of the vessel, which is quite flat. For other instances of paint on the
bottom, see plates 31e, 33a, 41a-f.

Plate 38
Nazea

Various Bowls, Plates, Goblet, Jar

All B or Late B except Jar h

Plate 38 is a miscellany, in shape and in design. Only a common period holds the
vessels together; they are from late in the development of the Nazca style except
for the jar 38h. Most of them in 1927 were put into shape classes other than those
I assign to them now.

The first 4 vessels, 38a-d, I was tempted to set together as a group "C3." They
are certainly allied to classes Cl and C2 as discussed in Part II. But they differ
from these in not having been put into our 1927 class C, as present Cl and C2
were; and also in being neither quartered on the bottom nor zoned. They are all
fairly flat-bottomed, shallow, definitely flaring, and concave in profile. They differ
from class B in not being gambreled, from E in not being straight-sided, from F
and H in being lower. But they do resemble the small class of D bowls in their
relatively flat bottoms, shallowness, considerable flare, and concavity of side; also
in seeming to be of late Nazea period. Accordingly, I leave them provisionally in D.

38a, 38b: 8741, 8740. D shape. In 1927 they were classed as Y2, that is, Nazca Y
with Tiahuanaco influence. They are figured in plate 14f, d; but inverted there,
set with bottom up, which makes them look like inside-rim-painted plates of class
Ar. Their finish is too hard and their coloring too brilliant to be Y; and their
designs seem to me now more in the Nazca than in the Tiahuanaco manner of Y.
There is parallel within Nazea of the hanging rods or bars of the design of 38a
in 40d (cf. Seler, fig. 172). The lower tier of figures in 38b may be much reduced
trophy heads; cf. Seler figures 201-204. Otherwise I have noted no parallels in
Seler to the designs of 38a, 38b. Seler, with his interest in symbolic-representative

396      University of California Publications in Am. Arch. and Ethn.

meaning, has slighted mere geometric designs, which are far more numerous in
Nazea ceramics than the illustrations in his basic monograph would lead one to
infer.

38a dimensions: D 169, H 75, H/D 44 per cent.

38c: 8943. Shape D; so listed in 1927. H 183, D 76, H/D 42 per cent. The design
seems to me in poorish late-B manner, verging toward Y. Colors: Y, R, Dark R
(purplish), B, on W.

38d: 9014. Marginal to shape D, with somewhat rounded bottom, definite bevel,
low and heavily concave sidewall. It was classed as shape A in 1927. D 195, H 62,

a~

b

Fig. 12a, b. Profiles of interior-rim-painted "plates" from Nazea shown in plate 38e, f.

H/D 32 per cent. The design painting seems late, with groups of alternately W
and R vertical stripes on B background, broken by four W-bordered squarish
panels, alternately B and W-stippled B. It is an aberrant piece. B backgrounds
and W outlining are usually early.

38e, 38f: 9023, 8665. Two rim-patterned, inside-painted "plates" of type Ar,
whose side profiles are shown in figure 12a, b. In 38e, the W-stippled meander and
the B-stippled "worm" suggest 38d; as do the W-bordered rectangles. 38f of course
is late and deliquescent B: multiple disintegrated heads, bent-over rays, pronged
"thorns," all very hastily cursive.

38e we classed as shape "C" in 1927; 38f is unlisted, apparently through a mis-
print of its number; it approaches F-H bowls in outline although it flares more-
the ratio of base to top diameter is 2:5. It might be classed as shape I, angled
goblet. It is Ar, of course, only in virtue of being rim-painted on the inside,
rather than of any formal shape as such.

38g: 8732. Shape I, angled goblet, as we classed it in 1927, except that the bot-
tom is rounded-conical, instead of essentially flat; from the bevel to the base is

Kroeber: Toward Definition of the Nazca Style        397
nearly 30 per cent of the total height. It is the only one of our angled goblets with
round bottom. For this reason it is also much the highest in proportion. In terms
of mouth diameter, the total height is 102 per cent, the waist 82 per cent, the bevel
diameter 84 per cent. The angle between below-waist and above-waist profile is
about 1600, or 20? bend.

The design of 38g consists of six pin-feathered, wingless, running birds-much
like Seler's figure 301 (plus the dots or pin-feathers of figure 300)-in the upper
zone, quite hastily painted; and in the lower, a row of paneled full-faces, abbre-
viated to four horizontal strokes each-a far remnant of the characteristically
phase-B women's Y faces in rows-as shown also in the lower zone of Seler's figure
301 (except that here there are five strokes to a face). Seler's piece is also an
angled goblet-so the 2 specimens agree almost point for point. The period of both
obviously is the transition from terminal Nazea B to Y; if not indeed full Y.

38h: 8520, shape R, wide-mouth jar, as listed in 1927, though we did not then
show any outright R jar in photograph. The mouth diameter is 68 per cent of the
maximum diameter; the neck 64 per cent, the height 92 per cent. The collar height
is 11 per cent of the total height. The proportion of 68 per cent is greater than for
any R, S, or T jar figured and measured in the Uhle collection' or figured by Seler.
It approaches the 75 and 76 per cent of the cookpots of present plate 34e, f from
Ocucaje F13. We may therefore infer that this jar is early, in or toward the Nazea-
A range. With this the W dividers are in accord, which set off each trophy head
in its own panel. The heads are also in the unusual position of being prone re-
cumbent-face down. Only the one right of center is set upright because the space
remaining for it was too narrow to lay it like the others.

Plate 39
Nazea

Goblets, Shapes J, K, L
Phases B and Late B

Nazea goblets are tall vessels, spreading from the bottom up. Sometimes the up-
ward spread is considerable, either beginning halfway up by means of an angled
profile (shape I), or by means of an inverted cone shape beginning at the bottom
(shape L). These two forms are characteristic of late B and continue into phase Y.

Characteristic of developed phase B, but not continuing into Y, are 2 other
shapes with moderately constricted but sinuous, unangled middle: J, waisted gob-
let (in 1927, merely "goblet"), and K, double-convex goblet (in 1927, "double-
curve goblet"). These 2 differ in that J flares somewhat upward, at times ap-
proaching the high F-H bowls; whereas K may be larger at bottom than at top,
and then approaches the concave cylindrical vase P. These 2 phase-B shapes,
favoring gently flowing profiles, thereby approach non-goblet shapes such as the
higher bowls and lower vases, sometimes differing from them chiefly in propor-
tions. As the Y-phase decadence was approached, the feeling for curvatures of
tactile movement became lost, and the silhouette turned abrupt, harsh, and geo-
metric in I and L.

6 The nearest is pl. 7b, a head jar, then attributed to shape-class Y, where the ratio is 60 per
cent.

398      University of California Pulblications in Am. Arch. and Ethn.

Examples of shape I, though somewhat aberrant ones, have already been shown
in plate 38f, g, and figure 12b; and another, aberrant in a different way, appears in
40a. The present plate 39 shows two illustrations each of shapes J, K, L.

39a, 39b: 8501, 8408, both shape J, waisted goblet. They were construed in 1927
as "K, double-curved goblet," and "M, small vase," respectively-which indicates
the tendency of the earlier goblets to intergrade. They may in fact have originated
in phase B through modifications of flowing surfaced bowls and vases; at any rate,
there seem to be no goblets or direct prototypes of goblets in A.

39a is largest at the top, but genuinely waisted. Its measures are: D at top 94
mm., waist 80, bottom maximum 87, H 120, H/D 128 per cent. I confess I feel no
great assurance in assigning this piece to class J rather than K. The top and bot-
tom zones of painted design are the same, reversed; the legged figures are perhaps
insects or spiders, corresponding somewhat to Seler's "water beetles," figures 364,
365. The middle zone is geometric, but the steps in the squares are rounded off, as
are some of the points in top and bottom. The execution is still precise, but tending
toward rapidity. I would estimate the design to be late B.

39b is nearly cylindrical, which is no doubt why we put it in "M, small vase," in
1927. The diameters, downward, are about 118, 110, 113 mm. (corresponding to
67, 62, 63 in the photographic negative), the height 151, H/D 128 per cent-this
last the same as in 39a. The waist constriction is slight, but attracts the eye.

The painted pattern of 39b is unusual, in vertical panels instead of zoned; and I
can recall no analogues for the striking "H" figures (except possibly Seler figure
365). I can only guess the period of the painting: B-Y transition, with the simpli-
fication of Y already present but its sloppy execution not yet accepted, in this
piece.

Other J goblets are shown in plate 9e, g, i.

39c, 39d: 8957, 8474, double-convex goblets, K. These both taper toward the top,
show a definite waist, are very similar in H/D ratio, and evince the same stage of
design development, namely full B, or late B with retention of quality. They even
contain the same crossed-rectangle design unit in background and subsidiary
zoning respectively, also the rays with bent-over squarish ends.

39c measures 91, 87, 102 mm. in its three diameters, in downward order, H 137,
H/D 134 per cent.6 The central animal design is strange-Nazea wing and paw,
but body truncated and suggestive of Huari Tiahuanacoid. The quartered bottom
is full or late B, but more commonly found on bowls of type C.

Could Huari have begun early enough to influence the Nazea B style in Nazea
Valley?

39d: 8474, has diameters of 96, 86, 110 mm., and is therefore unusually slim-
waisted; H 151, H/D 137 per cent. The head reappears on the Q-shape lipless jar
of plate 5e, even to the blood in or under the mouth; and again in plate 4e. Seler
shows a similar headdress in figures 220 following, and the same "open" or jawless
and blood-dripping mouth in figures 234-237, and especially 238, 240. Our 5e and
the Seler figures agree in bearing a "3-battlement" headdress instead of curled-
over rays. The small figures floating in the background of 5e evidently are halves

o 8957 does not occur in our 1927 classification list, but there are two 8958's, under K and under
0. I assume there was a misprint, and that 8957 was put in K then as now.

Kroeber: Toward Definition of the Nazca Style         399
of the rectangles that accompany the head in the present piece. A certain clarity
in the pastellike colors is also shared by 5e and the present 39c and 39d.

39e, 39f: 8395, 8516, are L, conical goblets, and were so classed in 1927.

39e: 8395, measures D 134, H 171, H/D 128 per cent. The elegance of form of
39c, d is gone. The painting is mannered and hasty-probably later and certainly
less skilled than the very similar painting on 40a. I would say it is incipient Y.
The diamond eyes in the rim zone are Y or incipient Y: cf. plate 9d.

39f: 8516, measures D 143, H 158, H/D 110 per cent. There is slight concavity
of profile, whereas 39e is convex. The design is quite unusual. The three-pronged
element recurs elsewhere, as in 43d, but the flag- or hatchet-shaped units are inex-
plicable. As a Nazea design, I would be at a loss to place this one in a period, though
the vessel as a whole suggests Y.

Plate 40
Nazea

Goblet and Vases

Full B and Late B Phases

The first figure, 40a, is a goblet rather indeterminate as to shape class. The other
figures are all M, tapering vases, as I have renamed our 1927 "small vase" shape.

40a: 8811, we put in 1927 into class I, angled goblet. I leave it there, though it is
not angular. It curves gracefully, in a long concave line. This fits my definition of
J, waisted goblet, especially of subclass Ji, "like high bowls F-H, but higher, up
to H/D 110 per cent," (which is the proportion of 40a). However, this piece has
no "waist" by measurements; the appearance of one is illusory. One could make
out a case for K, double-convex goblet, because of the rounding at top and bottom;
nevertheless what dominates the profile is the long concave sweep. This in turn
argues against its being L, a conical goblet. So I return to our 1927 start, as shape I,
an angled goblet, with the proviso that it represents an early form, before the
typical curve of period B had turned into the angle of period Y-as exemplified
by plate 38g.

This finding fits with the painting, which is less near the incipient-Y manner of
39e, and far from the fully degenerate-Y manner of 38g. Note, by the way, that
the present piece still shows full B-phase bent-over rays, which do not appear in
39e. Also, we have here three zones of trophy heads; but in 39e, geometric "eyes"
replace the small heads of the rim zone.

It is not clear what the vertical bars in the bottom zone represent. We could see
them as limb bones; but I do not recall that Nazca art ever concerned itself with
loose human long bones, in spite of its preoccupation with heads.

The middle figure in the middle zone is of course merely the headdress of the
heads on both sides of it, inserted because lack of planning brought it about that
there was no space left for a full head.

Herewith we leave the goblets with their tendency to intergrade, and turn to the
vases.

All the examples here shown seem to fall into the class M, tapering vase, which
presents the only difficulties in the vase group. There is no single feature which
defines the group sharply: that is presumably why we fell back on the irrelevant

400       University of California Publications in Am. Arch. and Ethn.

"small vase" in 1927. Of the 5 examples here shown, 40f tapers upward consider-
ably, 3 others modestly, but in 40c the rim is 1 mm. larger than any other di-
ameter.7 Four of the 5 vases show a slight constriction, enough to give elegance to
the shape; but 40b is on the convex side of being a cylinder.

In short, class M is basically cylindrical, with a tendency to a gentle taper, or
a mild constriction about two-thirds way up, or both. It differs from N, cylindrical
vase, in its H/D ratio being mostly less than 150 per cent whereas N is around 200;
and from 0 and P, the 2 bulbous-vase classes, in lacking their bulbous bulge at
the bottom.

The measures for shape-class M are given in table 17.

TABLE 17

VASE M MEASURES

Plate        Cat. no.  Max. D  H/D per cent

40b .............   8463       103       133    Cf. p1. 10e
40c .............   8900       100       135    Cf. P1. 9f
40d .............   8405       97        145
40e .............   8518       99        146
40f .............   8954       105       139
log .............   8614       111       135
10h .............   8546       97        147

It is evident that in general proportions as well as size, there is fairly close
standardization. It is in the nuances of more or less taper, constriction, or con-
vexity that there is enough variation to make it difficult to find a precise water-
tight definition for the shape class.

The shape is also close to J, waisted goblet, except that this tends to flare, shape
M to taper.

As for painted design on the group, it is evident that this is full B, and oc-
casionally perhaps more than full: 40e for instance, with its B on Y zone on Br,
is at least approaching the Y phase; compare 9d, 15d; and, above, 39b, 39e.

Plate 40b, which is a clearer rendition of 10e, resembles Seler's figures 131, 132,
but may represent what he construes his figures 133, 134, 186 to refer to, namely
a dance of the ghosts of those whose heads were taken as trophies.

Paintings of large-scale erect human figures, full-face or profile, without animal
attributes, were of course hardly possible on Nazea pottery until taller forms were
evolved than the Nazea A-shape Pa-in other words, became possible when stand-
ard goblet and vase shapes had been evolved; though there are occasional attempts
on double-spouts. It is rather remarkable how rare on the whole the erect natural
human figure is in Nazea art. I list below the readily available illustrations.

40c emphasizes the bird body as plate 9f features the human head of a composite
being repeated three times on the same jar 8900. The nearest Seler analogues are
figures 101-103, 105 for the bird feet (and wing, and tail), but figure 90 for the
non-mouth-masked face. The design is accordingly an unusual combination.

7 So under the calipers; in the photograph the maximum is in the lower half.

Kroeber: Toward Definition of the Nazca Style            401
40d has triple zoning, and all the elements conform to this indication of B pe-
riod: points and bent-over rays; face-arrows (?); and bars alternately hanging
and standing, like cog gears; cf. 38a, and Seler figure 172.

40e, 8518, has already been mentioned, just above, as showing or foreshadowing
period Y color and design.

40f, 8954, illustrates how a good bulge below can give the illusion of a waist
constriction, even though actually there is a continuous taper up to the rim, merely
the rate of the taper varying. This is one of the higher M shapes, and has bor-
rowed its zoning, design subjects, and color clarity from the high B tall cylindrical
vases. The maximum diameter is near the base and is 105; the mouth 85, or 81 per
cent; the height 146, or 139 per cent of the diameter.

I must confess that I find it hard to distinguish this "M" piece from shape P,
bulbous-concave vase.

Erect Human Figure Designs

G-K 1927:                                    Seler:

Illus.                 Shape          Illus.             Shape
Ic ................    U             124 ...............  S
5b ...................   Pb          129 .................  M
9g ...................   J           130 .................  M
9i ...................   J           131 .................  M

lOa ................... 0            132 .................  M
lOd ...................  J           133 .................  N

21b (?) .................. R         134 .................  M

122, 123, 137, 186 ...... shape not shown
Plate 41
Nazca

F-H High Bowls

With Painting Extending over Bottom

The remaining illustrations are, with one or two exceptions, figures of the high,
flaring, concave-profiled bowls of the overlapping classes F and H. On account of
their intergradation as discussed in Part II, these bowls will be treated here as a
unit, consideration being given to individual peculiarities of shape and designs.
The period is A, extending on into early B ("X").

The present plate shows six views of 5 bowls in which the painted design ex-
tends over the bottom as well as the side. Three other such bottom-painted bowls
have already been described here: plates 31e, 33a, 37h. The vessels of 41a-f were
uniformly posed for photographing so that the bottom and side sloped evenly away
from the camera lens at an angle of 450, so far as the curvatures of both bottom
and side allowed.

(All these bowls were originally classed by us in 1927 as F, except that 41e was
considered H.)

The measures are given in table 18.

41a: 9043, shown in 1927 drawn in outline as text figure lle. The design shows
one long snake on W background. I am unsure whether to rate this design as of
style-phase A or X, but incline to the former, for reasons which would take too
long to detail.

41b: 8929. The design is like Seler figures 101-110 and plate 44a, b below. It

402       University of California Publications in Am. Arch. and Ethn.

shows two figures of a type of bird which has a crop, trapezoidal head, rhomboidal
eye, blunt and hairy beak holding a trophy head. The feet are human, the thighs
marked with abbreviated trophy-head designs. The wing ends in seven feathers,
each tipped with a point and a three-dot face: the ends show in the photograph
just in front of the head. Above the feathers is a cogged or spiked edge frequent
in cat-deity representations. The trophy head, both feet, and one feather are below
the turn of the side into the bottom.

This design is later than the A of Ocucaje, where it does not occur. I do not
know whether it is late A, "X," or early B. It is discussed further after plate
44c below.

TABLE 18

BOTTOM-PAINTED F-H BOWLS IN PLATE 41

Fig.      Cat. no.   Type       D        H       Conc.* H/D per cent Spreadt

a.........     9043       F        116        85       3.5     73        110
b.........     8929       F        120        77       4       64        120
C, d ........  9101       F        120        92       6       77        120

e.........     8828       H        152        90       9       59        170
f.8923                    F        119        90       5       76        120

Concavity means the maximum distance in mm. between the side of the bowl and a ruler laid from the rim to the
side. This holds also for tables 19 to 23.

t Spread is the departure from vertical, in degrees read through a transparent protractor, of a similar ruler (touching
rim and side), the rim edge being held level. This holds also for tables 19 to 23.

41c, d: two views of 9101, c showing the left side of face and anterior body of the
cat-deity figure, d its tail end, right hand, and right edge of face. The hand holds
something like a truncheon or atlatl; the snake tail ends in a frog or salamander.
The "pleated" body has a "navel eye." Along the back lies a series of trophy heads,
face down, with slings to hang them from. The head wears a whiskered mouth
mask, with four W disks on a band falling on each side. It is a complex picture,
which crowds the B background and took some dexterity and experience to impose
on the combination of circular curve and convex-concave vertical curvature. The
design is post-early-A, probably "X."

41e: 8828, shape H, also shown, but without bottom, in plate 3c. This vessel is of
average height for shape-class F, but has conspicuously greater diameter, spread
or flare, and concavity, which is why we attributed it to class H in 1927. I construe
the figures as naked, possibly female dancers. There are no genitals, but a "navel
eye." They are "tangled" in a web which in 1927 we called motive "no. 13, over-
lapping lines." Another pair of figures is on the opposite side. Between, on the
sides, is a repetitive pattern of rounded rectangles, somewhat staggered as an
incident of the increase of diameter upward. They suggest a fish-scale pattern, but
I have no idea what they represent. Compare plate 21a, American Museum no.
41.0-1020, shape T, in which the paired fi!gures may be either wrestling or dancing,
the same "overlapping lines" are present, but the "scale" pattern is lacking. There
are strands of falling hair between the two heads. This suggests that what looks
like W lines across B background between the heads in the present 41e should also
be construed as locks of B hair, between which show streaks of W background.

Kroeber: Toward Definition of the Nazca Style          403
There is an interesting act or ritual represented here as to which we can only
guess.

The whole scene is unusually lively in action. There is nothing as animated in
the entire repertory of phase-A painting. The vessels must be at least "X" in pe-
riod, and really are akin, in their movement, to plates 5b, lOd, f, llb, c, and Seler
figures 131, 132, 133, 137, 186-most or all of which are indubitable high or
climax B.

41f: 8923. The bottom is covered with the same "overlapping lines" web as 41e
bears. Above are eleven columns each of four ripe maize ears, according to Seler's
interpretation-see his figures 382-386.

I incline to an "X" attribution; the field is rather too crowded to be of A phase.

Plate 42
Nazea

F-H High Bowls with Cat-Deity Design

Six more F-H bowls are shown in plate 42, with the cat deity painted on their
sides. Their dimensions are listed in table 19.

TABLE 19

CAT-DEITY DESIGN F-H BowLs IN PLATE 42

Fig.     Cat. no.   Type      D        H       Conc. H/D per cent  Spread

a.            8913       H        167      120      4       72       120
b .8773                  F        147      108      6       73       110
c .9100                  F        142      105      5       74       150
d .8909                  H        157      102      7       65       160
e .8917                  H        143       99      6       69       110
f .9075                  F        126       92      9       73       110

42a: 8913. The figure fates the less usual way, to the right, and encircles the
bowl with its "centipede" body. There is no mouth mask, only a point of whiskers
at each side of the face, and two more downward, like a forked beard. The rounded
eyes are cut off by a straight line above, with the pupils against this-a reminis-
cence of the Chavin eccentric eye? The mouth is triangular, with a small simple
tongue projecting. The paws have five square fingers plus a short thumb with long
pointed nail.

I construe this as an early-Nazea piece.

42b: 8773, is generally similar to the last, but it heads to the left, the more usual
way. The eyes are similar; the mouth too is small. There are side whiskers, but
none down from the jaws. The tongue, though simple, is long, reaching as far as
the nails of the paws. These are again square, but only four, and without any
thumb. The centipede body is similar to the last. The background is B, as against
W in 42a. Both pieces seem about equally old, but the execution of 42a is better.

In both the foregoing the face is stretched out in the axis of the body. In the
next 4 pieces, the face is in front view, the body in profile.

404      University of California Publications in Am. Arch. and Ethn.

42c: 9100. The head is similar to that of 42a, except for lacking chin whiskers
and showing what appear to be erect ears. There are: side whiskers in a point, eyes
cut off above, small mouth, short tongue (though with a capsicum hanging from it),
a three-part headdress, square-cut fingers (four) plus a thumb with long claw. The
profile body is arched and mottled, like Seler's figures 1-3, 8-10, 13-16; the (hind)
feet show just to the right of the head. The body ends in a swung-up tail. Below
the mottled part of the body is a longit-udinal stripe filled with a pattern of "over-
lapping lines" (1927 fig. 3, no. 13) ; also as in plate 41e, f. The background is dark.

42d: 8909. In general, the design is like 42c, with the following differentiations:

Background W

Eyes rhomboidal, not cut off on top

Mouth mask present, with short whisker points downward and long ones curving up past
headdress

No ears

Mouth rectangular
Tongue plain, long

Headdress with four uprights instead of three
Feet rounded, paddle-shaped, no thumbs

Body with two rows of "bird tracks" or "arrowheads" instead of mottles; but it does have a
similar stripe of "overlapping lines"

Triangular spines along the body back and the side of head may represent fruits

The figure is set in a panel, not encircling the bowl.

42e: 8917. For the first time in this series, the head fills the whole height of the
front of the bowl; there are no hands. Also for the first time, there is a mouth
mask with only horizontal, spreading side-whiskers. Also there are two W disks
at each side of the head. Eyes are rhomboidal, mouth semicircular, no protruding
tongue. Body: three rows of vertical arrows, each with two eyes = snake (?).

42f: 9075. The head is much as in 42e; but the mask does not meet above the
mouth, and there are four W disks instead of two in the band or hair tie falling
past the face. The headdress is similar to 42e, with a face symbol protruding down
over the middle of the forehead. There seem to be ears. The body is "pleated," with
a "navel eye" in the middle, like 41c, d; along the back are spines, with prone
trophy heads between.

These 6 pieces form a series as arranged from 42a to 42f. I am less certain that
the series is also a chronological sequence.

Seler's cat deities usually have a mouth mask. It is lacking in his figures 20, 24,
32 (mask becomes headdress, has bent-over rays), 59, 60, 68, 69.

The mask does not meet above mouth (as in 42f) in Seler figures 1-6, 8-16, 19,
26-28, 30-31, and occasionally beyond.

Raised whiskers (pointing upward), as in 42d, occur in Seler figures 1-6, 8-19,
23, 36, and others. The co-occurrence of this and the last previous trait applies only
in Seler's 1-6, 8-16.

Ears, as in 42c, perhaps f, show in Seler figures 1-5, 8-11, 16, 23, 24, 28-29 (?),
37 ( ?), 45 (?), 46 ('?). In general, they tend to disappear when mouth mask or
headdress are fully developed.

Chavinoid eyes cut off above (42a-c) are relatively rare in Seler: figures 6, 14
(eccentric pupil), 23, 24, 59, 69, 90, 97, 99.

Kroeber: Toward Definition of the Nazea Style          405
Thumbs with claw (42a, 42c) appear in Seler's figures 5, 23, 26, 27, 29-32, 35-38,
and so on.

Disks along the side of the face are lacking in all Seler's figures of the "mottled
cat," namely, 1-24; are normally present on the cat deity, 25-89; are lacking in
figures 32, 35, 36, 42, 44, 46d, as in 58-69 (jagged snakes), and in 74, 76, 78-82, 84,
87, 90-97.

The frequency of number of disks on each side of the face in Seler is:

1 disk ...               2 occurrences (figs. 50, 89)

2 disks ..           .   8 occurrences
3 disks ...             24 occurrences
4 disks ...             10 occurrences

This makes more occurrences of three disks than of all others together: usually
the third of four is displaced by the mouth mask. In Uhle's Nazea collection in
Berkeley four disks seem most frequent.

Plate 43
Nazea

F-H High Bowls with Cat-Deity or Snake Designs

Plate 43 adds 2 more to the cat-god designs on high bowls, and follows with 4
showing snakes.

The measurements are given in table 20.

TABLE 20

F-H BOWLS, CAT-DEITY OR SNAKE DESIGN IN PLATE 43

Fig.     Cat. no.   Type      D         H      Cone. H/D per cent  Spread

a.........    9072       F        143      113       7      79       120
b.........    9073       F        144      109       7      76        9O
c.........    8804       F        145       93       3      64       130
d.........    8599       F        125       84       4      67       120
e.........    8803       H        158      101       4      64       15?
f.......       9079      H        171      121       7       71      15"

(1927: 3e)

There is the usual intergrading of shapes classed as F and H in 1927, but it is
true that the 2 H bowls are larger and spread a few degrees more than the 4 F
ones.

The body of 43a is unusually cylindrical.

43a: 9072. Cat deity. Rhomboidal eyes, mouth mask not meeting above rectan-
gular mouth, short tongue; whiskers turned upward but spreading, not pointed;
four disks and trophy head at side of face. Body of varicolored, horizontal, spotted
stripes; on back, row of trophy heads lying supine.

43b: 9073. Cat deity. Eyes rhomboidal; mouth mask meeting along upper lip,
mouth rectangular, tongue short; whiskers rounded at end; ears (?); band of

406      University of California Publications in Am. Arch. and Ethn.

three disks and trophy head; headdress like 43a in basic plan but different in pro-
portions. Body, a short area horizontally pleated and with navel eye; above and
beyond this, one stripe containing dots; along back, prone-lying trophy heads be-
tween spines.

43c: 8804. The first of 4 bowls painted "with snakes." In this case, there are side
hairs, gills, or milliped legs. The head is relatively small, eyes circular, mouth
small and rectangular.

43d: 8599. A series of double-ended snakes, each an S, interlocking with another
at each end. Eyes are small round dots, mouth a short bar. A dark stripe runs the
whole length of body between the two heads. In the triangular area between each
interlocked pair is a three-pronged figure, similar to 39f and 46a. None of the
elements is bordered.

43e: 8803. The design looks more right if viewed upside down, with the bowl
rim below. This is confirmed by the fact that when the bowl is reversed, the snake
is headed to the left, but when set on its bottom, the head is to the less usual right.
The snake head is large, rectangular, with rhomboidal eyes set vertically; tongue
longish, curved, pointed. There is a bunch of pointed whiskers projecting out
from each cheek; and beyond these, a pair of W hands (or whiskers?). The wind-
ing body contains a longitudinal W stripe, in which are round dots connected by a
line. The dark background is dotted with whitish three-pointed figures which may
represent the same as the side fringes of 43c or the three-prong elements of 43d.

43f: 9079. Two intertwined snakes, both headed left. Heads indented, eyes
longitudinal rhomboids, tongue curved and pointed. Each body contains a central
W stripe carrying a row of (unconnected) dots; one snake has semicircles "bitten
out" of the W stripe, the other not. Background B.

All 4 of these snake-design bowls seem to be Nazea A in the larger sense. Com-
pare plate 27q, Ocucaje F12-4747, and plate 28p, Ocucaje 4711, near F6. Seler
shows double-headed snakes in his figures 313-316. Three of these have curved
pointed tongues; two, three-pronged or four-pronged elements; two, stripes con-
tain line-connected small disks; two are "bitten into." Martin-Vegue has plate
45-B-5 much like 43d; also 45-B-2 with apparently a series of parallel single-
headed snakes. All these are either phase A or not far from it.

By contrast, the snakes on the rim of the later plate-bowl 4a are segmented or
noded, have circular heads, and hold salamanders (?) in their mouths.

Plate 44
Nazea

F-H High Bowls with Bird Designs

The proportions of the 6 bird-design F and H bowls in plate 44 are given in table
21.

Seler devotes his third section to "The Cat Demon [really Deity or Spirit]8 as
Bird." Here he analyzes figures which combine a feline head and mouth mask with
wings, and usually human legs also, these last either stretched out in aerial flight
or vertical and supporting the head and body; and sometimes the figure is

8 The German word Diimon preserves more of the original sense of Greek daim6n than does
English "demon," which implies evil.

Kroeber: Toward Definition of the Nazca Style          407
feathered with arrows (Seler, figs. 72-97). Plate 40c, above, might be construed
as such a bird deity; so might our plates 4f, 9f, perhaps 21e.

Seler's fourth section is devoted to "Another Bird Spirit," illustrated in his
figures 98-120. This usually has wings, human legs or less often bird feet, and a
curious bird head which is always in profile even when the rest of the figure is
shown full-face. Most characteristic is the beak, always directed vertically, with
the mandibles not coming to a point but parallel, or even everted, that is, with
their facing edges curved away from each other. The tips of the mandibles are W,

TABLE 21

F-H BOWLS WITH BIRD DESIGNS IN PLATE 44

Fig.     Cat. no.   Type      D         H      Conc. H/D per cent  Spread

a ........    8928       F        140      105      5.5     75       12.50
b ........    9074       F        120       77      5       78       130

c ........    8521       H        162      102      5       62       220
d ........    8994       H        152      100      8       66       160
e ........    9083       F        148       92      5       62

f ........    8932       F        113       83      4       73       100

and resemble the thumbnails of some cat-deity hands. The mandibles always clasp
a trophy head. The bird's head may be rectangular, trapezoidal, ovate, or irregular,
and may or may not be provided with side whiskers coming to a point. Some
figures may be wholly birdlike, like Seler's 101, but more are partly anthropomor-
phized, clothed, or elaborated.

I have already shown one design of this type, the class F bowl 41b. In the present
plate I show two more, 44a, b, both also on F bowls.

44a: 8928, is the more naturalistic, that is, birdlike, of the two-though that is
not necessarily the earliest form of the representation. There are bird feet, and
no clothing or artifacts, except for the trophy head in the beak. The bird is
repeated in full two times in the circumference of the bowl, plus a reduced minia-
ture (on the left of the illustration), set upright to squeeze into leftover space, and
carrying what looks as much like a capsicum or other fruit as like a trophy head,
though it also shows eye and mouth dots. Six colors.

44b: 9074, has human feet, outcurved mandibles, side whiskers, and above these
bent-over rays. There seems to be a crop on the throat.

Plate 41b, already discussed, has features of both 44a and b. The bird head is
more as in 44a, though the trophy head is larger and hangs down on to the bottom
of the bowl; and the beak is feathered or bristly. There is a crop on the throat,
human feet, and a wing of seven bars (feathers), each ending in a snake head
and tongue-all as in 44b.

Martin-Vegue, 1949, shows two examples of this same bird spirit, both on F-H
type bowls. His plate 43-A-1 is much like the University's 44b; the mandibles are
feathered. Plate 44-B-5 has a similar head, again with feathered mandibles and a
crop. The body is "pleated," the feet are not clear, the back carries trophy heads.

408      University of California Publications in Am. Arch. and Ethn.

In all these examples, as in Seler's, the single eye is in a line with the beak and is
rhomboidal or bipointed-oval.

44c, d: 8521, 8994. In both these the design consists of files of wingless birds, per-
haps downy, immature, or with pinfeathers, like Seler's figures 300, 301. In 44c,
there are two tiers of birds, and they are filing to the right; their legs are W. They
are also W-speckled, whereas in 44d the spots are dark on W.

Our plate 3f (8554) is very similar to the present 44c, even to W speckling, two
tiers, and filing right.

The birds painted in the upper zone of the angled goblet plate 38g are also
similar to the foregoing and to Seler's figures 300-301.

Martin-Vegue's plate 45-A-7 is very similar to Seler's figure 300 in design as
well as in the shape of the shallow bowl. In addition, his plate 45-A-7 shows quar-
tering on the bottom, which indicates phase B as the approximate period of this
style of bird representation.

The present plate 37d, already described above, shows another bird similar in
being executed with crudity, in silhouette, heavy footed, but with a certain free-
dom of movement. It also has a peculiar blunt rectangular beak. This beak recurs
unmistakably in Seler's figure 302. The feet there are simply crossed, whereas in
37d they are two-toed; both, at any rate, are aberrant: Nazea bird feet are normally
three-toed. The Seler birds are humped (like 44c), with small triangular wings,
whereas in 37d the wings are three-feathered, with the feathers snake-headed: they
are square-ended, as against more normal rounding; this seems in conformity with
the square-ended beak. Seler's birds are in three tiers around a shape-J goblet
higher than wide, and are unquestionably B rather than A in phase, with which
attribution the wings of 37d are in accord-if indeed their sloppy execution does
not rather suggest B-Y. (In line with this lateness are the somewhat different
though also sketchily silhouetted birds in Seler's figure 301, which are associated
with yellow-women's-faces-in-rows so reduced in execution as to stand stylistically
at least as near full Y as full B.)

A third (or fourth) species of silhouetted bird appears in 1927 plate 10g, an
M-shape tapering vase, with three tiers of figures. This bird has a turned-over
parrotlike beak. (We called it a parrot in 1927, fig. 3, no. 37.) It shows an erect,
longish neck, two narrow erect wings, a long forked tail, three-toed feet directed
forward and painted cursively. All these last four features recur in Seler's figure
299b; though the beak of this is different: long, downcurved, open.

What all these birds beginning with 44c, d have in common is that (1) they
occur in files instead of as substantially single units; (2) they are largely or wholly
in sketchy silhouette, unbordered, whereas most Nazea birds are taken to pieces
and their parts presented in contrasting colors, and bordered; (3) the feet in this
group are large, "plantigrade," and drawn with a wide brush; (4) wings are often
lacking, never elaborate; instead there may be a humped back; and (5) in about
half the cases the body or its contour is dotted, as if to indicate pinfeathers; which
accords with the winglessness. As several different species are obviously repre-
sented among the examples of this class, it may be that the common feature repre-
sented is immaturity of the bird. Part of the examples manifestly date from
fairly late in the development of the Nazea style-B or B-Y-and none are clear A.

Kroeber: Toward Definition of the Nazca Style        409
44e: 9083. The design is one of the typical representations of birds as they were
developed in Nazea A, perhaps to continue part way into B. The crop on the throat
is conspicuous. Note the S-line over the eye, the feet turned both fore and aft, and
the narrow ray-feathers bordering (and overlapping) the tail. The several repeti-
tions are crowded: the beak of the whole figure shown in the photograph is pushed
into the wing of the preceding bird-in fact it has crowded out the middle feather.
The remainder of the forward bird is also condensed: note the shortened tail. It is
clear that the bird in the center of the photograph 44e was the first one painted.
When the potter had nearly completed the circuit of the bowl, there was insufficient
room left, and the final bird (the one on the left of the photograph) had to be
telescoped.

44f: 8932. I do not rid myself of a half-conviction that the designs just below the
rim represent birds-with neck, head or eye, a decoratively doubled beak still
remaining. But why the "millipede" effect? Perhaps a chimera or combination
animal? The banding suggests relative lateness of the piece.

Plate 45
Nazea

F-il High Bowls with Repetitive Pattern

The measurements of the 8 simple-design F-H bowls shown in this plate are given
in table 22.

45a: 8936. Slings on a B background; the miniature is a space filler. Compare
Seler figures 423-426.

45b: 9089. Also on a B background; eleven loops of six festoons; in successive
order from the top down: W, R, Y, G, Y, W. For this sort of simple "color exercise"
compare plate 29g, A-4789; also 33a, F4-4695.

45c: 9097. The unit of design resembles Seler figures 393-395, which he inter-
prets as a capsicum pepperpod. The middle of each unit is W with a row of dots,
bordered along each side by a streak of color. The sequence of these colors is:
Y 0 Y B-Y 0 Br-Y B-Y 0 B-Y 0 B-Y 0 B. There is some wavering from
a basic Y 0 B grouping to Y 0 Br to Y 0 or Y Br or Y B. The background is
Dark R.

45d: 8604. A very similar bowl, the colors, however, being a regularly alternating
brownish 0 and a B, around a W middle. This piece which we classed in 1927 as
shape F, with H/D ratio of 71 per cent, and the preceding 45c, H/D 73 per cent,
again evidence the intergrading of these two "shape classes."

45e: 8446. Design, bean pods; compare Seler figures 375-378.

45f: 8438. Design, trophy heads. The half-circle eye is on a level with the nose;
the upper lip is abnormally long; a bar of paint crosses the cheek; there is beard
on the jaw; and a sort of turban is worn. Compare plate 3a, 8598, shape H, and
1927 figure 12a, 8581, shape H, two rows of heads; also Seler, figures 187, 188.

45g, 8535, F, and 45h, 9086, H, differ about as much in design as in shape. Both
are W-bordered, but 45h also has W divisions into panels (like 45c, d, f ), of which
there are only half lengths in 45g. In both, half the "step" is oblique-angled; but
45g combines it with a "fret," square at the corners but round at the end. In both,

410       University of California Publications in Am. Arch. and Ethn.

the design is complementary, the formal design unit consisting of two interlocking,
reversed halves.

The same designs, but consistently rectangular, undivided, and also W-outlined,
appear in 2e, f. Compare also Martin-Vegue plates 45-B-7, 46-A-1, 46-B-5.

TABLE 22

REPETITIVELY PATTERNED F-H BOWLS IN PLATE 45

Fig.      Cat. no.   Type      D         H       Cone. H/D per cent  Spread

a........      8936       F        143        96      5        77       120
b........      9089       F        132       86       4        65       120
c.........     9097       H        142       104      3.5      73       13"
d.........     8604       F        140       97       4        69       13"
e.........     8446       F        122       85        6       70       15"
f.........     8438       H        153       107       7       70       17"
g.........     8535       F        147       108      4        73       12"
h.........     9086       H        170       95       2        56       17"

Plate 46
Nazea

F-H High Bowls and Cat Jar

Plate 46 continues with examples of F-H bowls (a-f) and concludes with an
unusual cat-head jar (g).

The measurements are presented in table 23.

TABLE 23

F-H BOWLS AND MODELED CAT IN PLATE 46

Fig.      Cat. no.   Type      D         H       Conc. H/D per cent  Spread

a........      8470       F        143       117      4        81       80
b.8648. 864               H        133       83       3        62       12"
c........      8579       H        166       117      4        70       12"
d........      8997       H        135       91       7        67       15"
e........      8709       F        136        95      4        70       15"
f........      8484       H        142        95       7       67       16"
g.......      16-7851    ps-Q      111*       76     ..        68       ...

Mouth diam., 49 mm., is 44 per cent of max. diam.

46a, 8470, F, is unusually tall. It shows five repetitions of a design, with double
paneling dividers, on a B background. It is not clear what the three-pronged design
(cf. pls. 39f, 43d) stands for, nor the W dots that border it. The latter recur in
plates 4b, llb, which are Nazea B, and in 12e, 13f, which are Nazea Y, whereas
the present bowl, in spite of its height, is almost certainly Nazea A. A nearer
parallel is to be found in the next illustration, 46b, and perhaps in Martin-Vegue
plate 45-B-2.

46b, 8648, H, carries four and one half units of a double spiral on a B ground
studded with W dots. The spiral recurs in Martin-Vegue plate 45-B-4; similar dots,
though fewer, in 46a.

Kroeber: Toward Definition of the Nazca Style         411
46c, 8579, H, has no close parallel known to me for its main design of concentric
loops or U's except Ocucaje B4-4638, shown drawn in figure 8. The surrounding W
hooks are closely paralleled in plate 29c, an E-shape bowl from Ocucaje, A-4490.
As this is early A period or pre-A, a similar or not much later dating is indicated
for the present specimen. The leftover space proved too narrow for a compressed
version of the design unit, so a wholly different one was used for filler.

46d, 8997, H, shows an over-all, highly conventionalized trophy-head pattern.
The nearest analogue in Seler is figure 173, and that is not too close; also 1927
figure 3, no. 30, "curvilinear head series."

46e, 8709, F, is similar to the last in its columning, and in carrying four tiers in
each column. The slight curvature of the contained four strokes is reversed in each
tier; but the main organization of the patterning is owing to the varying slope of
the three pairs of lines dividing the tiers. These have something of the effect of the
"overlapping lines" element of 1927 figure 3, no. 13, of 1927 plate 21a, and of
present plates 41e, f, 42c, d. But there might also be connection with the capsicum
pods of Seler figures 392, 393.

46f, 8484, H, is a better rendition of the color, values, and crucial right and left
ends of the design of plate 3d. This seems a later piece than any other in the
present plate; or perhaps a late imitation of an early design. Some of the outlining
is quite unsteady. Note the stepped mouth area without mask or whiskers; trident
tongue ending in snake ( I) heads; W snake-headed locks (?) or fillet ends (?)
beside the face; long bent arm holding an atlatl ( ?); feet just behind head; spined
body with swifts ('?) in the interstices; tail ending in a head with side whiskers
and hands having thumb nails.

Plate 46g, no. 16-7851, given to me personally as I left Nazea in 1926, is the only
vessel in the last six plates which is not a high F-H bowl. It is a recurved convex
lipless jar of shape pseudo-Q, but worked into a cat head with modeled projecting
nose and ears. It is 111 mm. in diameter, 49 across the rim, 76 high. The back-
ground is blackish, the painting W, B, R, Br. There are conspicuous feline whisk-
ers, but no mouth mask, and conspicuous eyebrows. Over the middle of the lower
lip (and "central tooth") is painted an oval yellowish Br spot.

I paid no particular attention to this piece, assuming it to be Tiahuanacoid. It is
certainly outside the main Nazea tradition as we know it so abundantly. But when
I showed it to Dawson, he immediately diagnosed it as early, containing Chavin
influences; and with this judgment I now agree. It is somewhat outside the Nazca
stream because it is literally proto-Nazca. It is not straight Chavin nor straight
Paracas Cavernas as represented at Ocucaje in the Truel collection; but it seems
to lean in their direction.

The nearest analogue I can point to is Martin-Vegue, plate 45-A-3, a small W
or whitish bowl or jar painted, in outline only, with a cat head on its front. This
vessel curves in toward the rim much less than 46g; but it turns in more than any
determined bowl shape in Nazea; and the author has labeled it "miscellaneous" as
style phase-or perhaps with reference to our 1927 shape class "Z, miscellaneous."
The painting shows these features in common with 46g: pointed whiskers at the
sides of the mouth; nostrils (modeled in 46g); eyebrows large and conspicuous,

412      University of California Publications in Am. Arch. and Ethn.

of several parallel lines or stripes; mouth roughly oval, with vertical lines sepa-
rating the teeth.

Our 46g, though related to shape-Class Q, "lipless jar," nevertheless is actually
nearer to the pseudo-Q jars or bowls discussed in Part II. It continues the convex
curve of its lower half on into the upper half, where Q jars tend to straighten out
toward the rim, whether they are plain A-period or design-painted B-period pieces.
Coupled with this difference is the fact that the ratio of mouth dia-meter to maxi-
mum body diameter is visibly lower for 46g: 44 per cent, as against 50-75 per cent
for all but one of the Q jars in table 7, and 70-79 per cent for the pseudo-Q jars
of table 8 in Part II. The Martin-Vegue piece is 76 per cent, and is more decidedly
bowl shape than any other.

ACKNOWLEGMENTS

I appreciate help rendered me by many: Victor Duran for superb photographing;
Jane Bendix for putting the text figures into final form; Evelyn Lilge for careful
typing; and Mary Anne Whipple for editing the whole.

Lawrence Dawson found mislaid specimens, measured many vessels, sketched
their contours and sometimes design, and discussed point for point with me, out
of his varied knowledge of Peruvian ceramics and particular study of the Uni-
versity's Nazea-style collections. He is engaged in an intensive analysis of the
development of that style, in comparison with the prospective constructive results
of which my present ones are mainly corrective only.

Dawson and John Rowe, Donald Collier, William Duncan Strong, Anna Gayton,
Junius Bird, Rose Lillien, and Robert Stigler have aided by reading portions or all
of the text of Parts I and II.

A general donation from the Ford Foundation, Division of Behavioral Sciences,
has aided materially in the prosecution and preparation of this study.

BIBLIOGRAPHY

BENNETr, WENDELL C.

1953. Excavations at Wari, Ayacucho, Peru. Yale Univ. Publ. Anthr., No. 49.
GAYTON, A. H., and A. L. KLoBBER

1927. The Uhle Pottery Collections from Nazea. Univ. Calif. Publ. Am. Arch. and Ethn., 24:1-

46, pls. 1-21.
KELLY, ISABEIL T.

1930. Peruvian Cumbrous Bowls. Univ. Calif. Publ. Am. Arch. and Ethn., 24:325-342.
KRonBna, A. L.

1930. Cultural Relations between North and South America. Intern. Congr. Americanists,

26:5-22.

1944. Peruvian Archaeology in 1942. Viking Fund Publications in Anthropology No. 4. Pp.

1-142, pls. 1-48.

1953. Paracas Cavernas and Chavin, Univ. Calif. Publ. Am. Arch. and Ethn., 40:313-348, plS.

26-32.

KROEBEE, A. L., and W. DUNCAN STRONG

1924. The Uhle Pottery Collections from Ica. Univ. Calif. Publ. Am. Arch. and Ethn., 21 (no.

3):95-133, pls. 25-40.
MAPTIN-VEGUE, GEORGE B.

1949. Nazea Pottery at Florida State University. Am. Jour. Arch., 53:345-354, pls. 40-B-46-B.
MEANS, PHILIP AINSWOETH

1917. A Survey of Ancient Peruvian Art. Trans. Conn. Acad. Arts and Sciences (New Haven),

21:315-442.

MUELLE, JoRGE C., and CAMILO BLAS

1938. Muestrario de Arte Peruano Precolombino. Revista del Museo Nacional, Lima, 7:163-

280.

O'NE.AB, LILA M.

1937. Archaeological Explorations in Peru, Part III: Textiles of the Early Nazea Period.

Field Museum of Natural History, Anthropology Memoirs, 2:110-218 (no. 3), pls. 32-68.
SEEdE, E.

1923. Die Buntbemalten Gefiisse von Nasca, in: Gesammelte Abhandlungen 4: 169-338, 430

figs.

WILLEY, GORDoN R.

1951. The Chavin Problem: A Review and Critique. SW Jour. Anthr. 7:103-144.

[ 413 ]

PLATES

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                              [KROEBER] PLATE 31

Plate 31. Ocucaje graves Al, B4, 5.

b-.----...........

a

b

C                                                d

e                                                f

Ocucaje graves Al, B4, 5.

[417]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                         [KROEBER] PLATE 32

Plate 32. Ocucaje graves Cb, F4, 8, 18, 20, 22.

b
a

c

e                                       f

9

Ocucaje graves Cb, F4, 8, 18, 20, 22.

[ 418 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                         [KROEBER] PLATE 33

Plate 33. Ocuicaje graves F4, 6, 14, 18.

:~~~~~~~~1

a

C

X~~~~ f:

Ocucaje graves F4, 6, 14, 18.

[ 419 ]

UNIV. CALI F. PUBL. AM. ARCH. AND ETHN. VOL. 43                     [KROEBER] PLATE 34

Plate 34. Ocucaje graves F2, 3, 6, 10, 13.

_ i iS _ i S ~~~~~~~
_  l.   _  l .      _a

, . l l . _~g
.C      _....

~~~~~~~~~~~~~~~~~~~~~~~~ff

-F~~~~~~~~~~[2                   1

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                                                                                              [KROEBER] PLATE 35

Plate 35. Ocucaje graves F3, 9 (A-plus phase).

:::4:::   Te                                             :V:i:7-:            .;   .S.  _       J             _  .................................  E :-i:5 -:::.;:i::. .  . .   . . . . .. . .

0      _           '              D                .                        fi;",'.''-'',;   1              I ff't~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. .... .

:                               ff; t'g 'f' g'' 'g' '''  .'-E''.''X'''-; 0't;'-'"'.,'-"sS-sa .,,. f40.~~~~~~~~~~~~~~.. .  .   ..   .. ..  .

0                                                                             .i0 7 500   i;:   i  : ' ''  ' _  -   !' 4; 4 "i; ''0 ':0 0 S,' ':"':' : .............. >: jS :. i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... ... ..
tEE :i,:tli0','":.S '..'';S .S.':0: 0,o''' .......'

:                                                  :   th   r;; r ;  . .   )  .   __ , .  . . .......... , <, s. - -:- T: :;.-E tXT: [.o S!! s.;_ 0t~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.  ..  ..  . .

fS iAfiE:gE~~~~~~~~~~~~~~~~7TRW,:i-i- e;;,i:00 D0 -S , s_''

:::: ::: t:i-:S:t:.XP.:0~~~~~~~~~~~e--                                                                              .. s _ c--: .ff::1: .....0_L_-t
:7 ~ ~      ~       ~      ~     Oc     c   j      gr ve        F3   9,,  (A-lu         phase),,Ed,i-.E .f:   1                     F

E-~~~~~~~~~~ 42                                              ]--6.  f _t-ii ,0: _F-

UNIV. CALIF. PUBL. AM ARCH. AND ETHN. VOL. 43                         [KROEBER] PLATE 36

Plate 36. Nazea interior-painited plates Ac.

0                                                b_
ab

c                                           d

e                                            f

Nazca interior-painted plates Ac.

[ 422 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                                 [KROEBER] PLATE 37

Plate 37. Nazea low-bowl shapes B, D, E.

---- .i  ;SE.: .......  ---iLSX. ....       ..   . ..........0gSLLXS

b
a

d
C

f

g                                                       Xh

Nazca low-bowl shapes, B, D, E.

[ 423 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                                     [KROEBER1 PLATE 38

Plate 38. Nazea bowls, plates, goblet (late), jar (early).

E i E T-  Cy  _                                            ....      f.~~~~~~~~~~~~~~~~~~~~~~~~~~......

a

d
C

f

--i - ---------

e

g                                                    h

Nazca bowls, plates, goblet (all late), jar (early)

[ 424 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                   [KROEBER] PLATE 39

Plate 39. Nazea goblet shapes J, K, IJ.

:~~~~~~~~ -' -----

5~~~             ~~~~~~~ 4I25 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                           [KROEBER] PLATE 40

Plate 40. Nazea cylindrical jars, shapes I, M.

a                                         b

C

e                                          f

Nazca cylindrical jars, shapes I, M.

[ 426

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                       [KROEBER] PLATE 41

Plate 41. Nazea high-flaring F-H bowls, bottom-painited.

ab
cd

e                                           f

Nazca high-flaring F-H bowls, bottom-painted.

[ 427 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                         [KROEBER] PLATE 42

Plate 42. Nazea F-H bowls, cat-deity designs.

ab

d
C

f
e

Nazca F-H bowls, cat-deity designs.

[ 428]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                           [KROEBER] PLATE 43

Plate 43. Nazea F-H bowls, cat-deity and snake designs.

Co~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:

a                                               b

e                                             f

Nazca F-H bowls, cat-deity and snake designs.

[ 429 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                          [KROEBER] PLATE 44

Plate 44. Nazea F-H bowls, bird designs.

a                                             b

c                               ~~~~~~~~~~~d

e                                             f

Nazca F-H bowls, bird designs.

[ 430 ]

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                         [KROEBER] PLATE 45

Plate 45. Nazea F-H bowls, r epetitive (lesignLs.

w~~~

c~~~~~~~~~~~

e

Nazca F-H bowls, repetitive designs.

[ 43' _

UNIV. CALIF. PUBL. AM. ARCH. AND ETHN. VOL. 43                              [KROEBER] PLATE 46

l']ate 46. Nazc a P'-H bowls all(l mo(leledcl at jar.

7 _         b                        __~~~~~~~~~~~~

e~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ f
_ ~   ~~~~~~     ~~~~~      4:g 11