Citation
Dynamic test simulations for determining ground quadrilateral fault impedance measurements

Material Information

Title:
Dynamic test simulations for determining ground quadrilateral fault impedance measurements
Creator:
Smith, Karl M
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
40 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Electric relays -- Testing ( lcsh )
Electric fault location -- Data processing ( lcsh )
Electric fault location -- Mathematical models ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaf 40).
Thesis:
Anthropology
General Note:
Department of Electrical Engineering
Statement of Responsibility:
by Karl M. Smith.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
45210762 ( OCLC )
ocm45210762
Classification:
LD1190.E54 2000m .S33 ( lcc )

Downloads

This item has the following downloads:


Full Text
DYNAMIC TEST SIMULATIONS FOR
DETERMINING GROUND QUADRILATERAL
FAULT IMPEDANCE MEASUREMENTS
by
Karl M. Smith
BSc., Colorado School of Mines, 1986
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Electrical Engineering
Spring 2000


This thesis for the Master of Science
degree by
Karl M. Smith
has been approved
by
4 ~7 -ao
Date


Smith, Karl M. (M.S., Electrical Engineering)
Dynamic Test Simulations for Determining Ground Quadrilateral Fault
Impedance Measurements
Thesis directed by Professor Pankaj K. Sen
ABSTRACT
Complex algorithms in todays microprocessor based relays are capable of
determining ground quadrilateral fault impedance, fault location and direction by
unique measurement calculations from settings and inputs. This complexity,
however, increases the challenge for developing test procedures for these relays.
This is in part due to the limitations of the test equipment and the high volume of
complex calculations required for dynamic testing. Therefore it is necessary to
provide advanced mathematical models and simulations of the power system to
demonstrate how the fundamentals are applied when implementing test procedures
that conform to more rigorous standards than outlined in the test manuals. To
automate the calculation process, a spreadsheet ha been developed to determine
ground quadrilateral measurements by conveniently entering in relay settings and
inputs.
This abstract accurately represents the content of the candidates
thesis. I recommend its publicatit
Signed
Pankaj K(/ Sen
m


ACKNOWLEDGEMENTS
I would like to thank New Century Energies for the vast amount of financial and
technical resources made available to me for the completion of this thesis and my
education. In particular, I would like express my appreciation for the following
individuals;
George Laughlin, a system protection engineer at New Century Energies, for his
guidance and providing engineering support.
Chris Gallegos, coworker, for providing the necessary test data for the extensive
numerical examples throughout this thesis.
Ken Behren, SEL developer, for discussion on reactance element error.
Karl Zimmerman, Brad Heilman and Bill Flemming, SEL application engineers,
for discussion on directional elements.
Edi VonEngeln, classmate, for help in the development of Excel spreadsheets.
Carlene Stroh, coworker, for help in using Microsoft Word commands.
Mickey Pitt, a friend and fellow student who has taken many classes with me, for his
time and commitment in helping me to complete the graduate program. Thanks for
your encouragement.
Dean Nester, my supervisor, for his support and encouragement.
Finally, I would like to thank Dr. P.K. Sen, who has taught the majority of my
courses, for helping me realize my potential as an electrical power systems engineer.


CONTENTS
Figures.........................................................vii
Tables.........................................................viii
Abbreviations and terms..........................................ix
Chapter
1. Introduction.................................................1
2. Theory and fundamentals......................................4
2.1 Reactance measurement........................................4
2.2 Resistance measurement......................................11
2.3 Negative sequence impedance measurement.....................14
3. Modeling a nonhomogeneous system............................19
3.1 Reactance measurement calculation for the...................21
zone 1 fault location with 5 Q fault resistance (T=0)
3.2 Reactance measurement calculation for the...................22
zone 1 fault location with 5 O fault resistance (T=5)
3.3 Resistance measurement calculation for the..................22
zone 1 fault location (true resistance from the fault study = 5 0.)
3.4 Negative sequence impedance measurement threshold...........24
calculations for a bolted fault at the zone 1 fault location
3.5 Negative sequence impedance measurement threshold...........25
calculations for 5 O fault resistance at the zone 1 fault location
v


4. Dynamic test simulations....................................27
5. Results and Conclusion......................................31
Appendices
A. Dynamic test simulation spreadsheet (T=0)...................33
B. Dynamic test simulation spreadsheet (T=5)...................34
C. Residual current compensation factor derivation.............35
D. Zero sequence current compensation factor derivation........38
E. Effect of transmission line conductor changes on Kn and Ko..39
References........................................................40
vi


FIGURES
Figure
1.1 Ground Quadrilateral characteristics.......................1
2.1 Reactance reach dimension..................................4
2.2 Sequence network for SLG fault.............................5
2.3 Measurement error for a two source system..................7
2.4 Voltage phasors for a resistive fault in a nonhomogeneous system. 8
2.5 Zero sequence network for a single line to ground fault....9
2.6 Resistance reach dimensions...............................11
2.7 Negative sequence network for a SLG fault.................16
2.8 Z2 characteristics, MTA = 90..............................17
2.9 Z2 charcteristics, MTA<90.................................18
3.1 Nonhomogeneous system 345 KV line from Rifle to Craig...19
4.1 PROTEST 3 Z plot..........................................27
4.2 Reactance measurement error (T=0).........................29
4.3 Reactance measurement error (T=5).........................30
C. 1 Sequence network for SLG fault............................35
vu


TABLES
Table
3.1 Relay inputs from fault data (Rf = 0 Q)........................20
3.2 Relay inputs from fault data (Rf =5 Cl)........................20
Vlll


ABBREVIATIONS AND TERMS
Zli Positive sequence line impedance at the zone 1 fault location.
Zl2-Negative sequence line impedance at the zone 1 fault location.
ZloZero sequence line impedance at the zone 1 fault location.
Ziang-Positive sequence line impedance angle
MTA Maximum torque angle
Rf Fault resistance
IfTotal fault current
Ir Residual relay current (3Io)
Kn Residual current compensation factor
Ko Zero sequence current compensation factor
T Nonhomogeneous compensation correction factor angle
AZT Reactance element error term
Z2M Negative sequence impedance measurement ( a )
Z2F, Z2RNegative sequence impedance for forward, reverse faults
Z2FT, Z2RT Negative sequence impedance forward and reverse thresholds
SEL Schweitzer Engineering Laboratories
IX


1. Introduction
Ground quadrilateral distance elements provide more fault resistance coverage
than ground mho units when protecting transmission lines from ground faults.
Fault resistance coverage in a ground quadrilateral is controlled by the resistive
reach setting which defines the side boundaries of the quadrilateral in the R-X
plane. The upper boundary of the quadrilateral is controlled by the reactive reach
setting. In recent versions of microprocessor based relays, the reactive reach
setting is defined in terms of the positive sequence line impedance in the R-X
plane instead of the X-axis. The lower boundary defining direction, resides in the
negative sequence impedance plane and is controlled by negative sequence
impedance threshold settings [ 1 ].
jx
Figure 1.1: Ground quadrilateral dimensions.
1


When power systems have unequal source and line impedance angles, the fault
resistance measurement will have a reactive component that is introduced into the
reactance measurement that will cause the relay to overreach or underreach. This
type of system is defined as nonhomogeneous. The difference between the
source and line impedance angles determines the value of the nonhomogeneous
compensation setting required for correcting the reactance measurement error.
To determine the reactance measurement error, a fault study needs to be
conducted to model a non-homogeneous system for increasing increments of
fault resistance at the reactive reach point. The apparent impedances for each value
of fault resistance then need to be converted to current compensated reactance
measurements that would occur in the relay. The difference between the current
compensated reactance measurement and the reactive reach boundary from relay
test results defines the reactance measurement error. Test results for the reactive
reach boundary can be obtained by running a computer generated impedance
search plot. Apparent impedances from the search plot are then converted to
current compensated reactance measurements to confirm that these measurements
are within close tolerances to the reactive reach setting boundary. To simulate
the relays conversion of reactance and resistance measurements, a spread sheet
will be created that expedites the calculation for these measurements from relay
settings, voltages and currents. This method of dynamic testing is necessary since
2


there is no way to determine the reactance measurement error from a computer
generated search plot since the relay is underreaching.
3


2. Theory and Fundamentals
2.1 Reactance Measurement
The ground reactance measurement calculates the positive sequence impedance of
the line at the fault location [1]. The reactive reach dimension of the ground
quadrilateral can be represented by lines sloped at the positive sequence line angle
that originate from the R-axis (figure 2.1). The distance of these lines depends on
the zone of protection which designates the percentage of the line length expressed
terms of positive sequence impedance.
jX
Figure 2.1: Reactance reach dimension
4


If the reactance reach measurement is within the reactive reach defined by the top
portion of the quadrilateral, the reactive reach element will assert. For a trip output
to occur, however, all ground quadrilateral elements must assert.
To derive the expression for the reactance measurement, a voltage drop expression
needs to be written from a single line to ground sequence network (figure 2.2), that
includes a fault resistance term.
RELAY FAULT
LOCATION LOCATION
Figure 2.2: Sequence network for a SLG fault (Phase A)
5


From the A phase relay voltage equation written in terms of the symmetrical
impedances,
Va = Ii(Zli) + I2(Zl2) + Io(Zlo) (2.1)
and,
3Io= Ia+ Ib+ Ic
The voltage drop equation can be expressed in terms of the residual current
compensation factor Kn.
Va= Zli(Ia + Kn) (2.2)
Derivations of residual and zero sequence current compensation factors including
effects of conductor line changes are shown in appendices C-E.
Adding a fault resistance term to the voltage drop equation, the expression now
yields,
VA= ZLi(Ia + KnIr) + IfRf (2.3)
Multiplying both sides of the equation by the complex conjugate of IrZT to make
the voltage across the fault resistance appear real yields,
Va(IrZT )* = ZL1(IA + KnIr)(IrZT )* + IfRf (IrZT )*
where T is the nonhomogeneuos compensation factor correction angle. Taking the
imaginary component of both sides of the equation, the expression now
yields,
hn[VA(lRZT )*] = Im[ZL1(IA + KnIr)(IrZT )*] + hn[IFRF (IrZT )*]
where,
6


Im[IFRF (IrZT )*] = 0
Multiply both sides of the equation by |Zli|,
|ZL1[ = frn[VA(IRZT)*]___________________
Im[ ZL1/|ZL,| (IA + KnIr)(IrZT )*]
where |Zu | is the reactive reach referenced to the positive sequence line angle.
Therefore Zli/|ZLi | = IZZiang
Substituting,
17 ... hnrVAaRZT)*l____________________ (2.4)
1 L1 hn[lZZ1ANG(lA + KnIr)(IrZT )*]
If the T setting is not used for nonhomogeneous systems, an error term will be
introduced into the reactance measurement since the voltage across the fault
resistance now appears reactive (figure 2.3) [3].
Zli(Ia+KnIr)
Figure 2.3: Measurement error for a two source system
7


The voltage drop from equation 2.3,
Va = Zli(Ia + KnIr) + IfR-f
now has an error term in the relay voltage (figure 2.4).
jix
Figure 2.4: Voltage phasors for a resistive fault in a nonhomogeneous system
8


The reactance element error term, AZT, can be determined from zero sequence
currents and impedances shown in the zero sequence network (figure 2.5).
Figure 2.5: Zero sequence network for a SLG fault
There are two methods of calculating AZT. One method requires a theoretical
approach while the other takes a more practical approach.
- [ Zso + Zlq + Zrq]
Zro
AZT = -^ES-
J-so
(2.5)
(2.6)
The reactance measurement error, A |Zli|, can now be expressed in terms of AZT
from the following equation;
9


(2.7)
A|ZU| = Rf C ^(A)8^ T ]_____________________________
[ Ia sin(Z- 0 ZIr) + KnIr sin(ZZiANG +ZKN) ]
The reactance measurement error A|Zli | can easily be obtained from fault studies
and relay settings to determine the compensation required to prevent the relay from
overreaching or underreaching.
10


2.2 Resistance Measurement
The resistance measurement calculates fault resistance for line to ground faults from
relay inputs and settings. The resistance reach dimension is defined by any point on
the reactive reach line that intersects the origin in the R-X plane to the side
boundaries of the quadrilateral. [1]
]X
Figure 2.6: Resistance reach dimensions
11


If the resistance measurement is within the resistive reach boundary of the
quadrilateral, the resistive reach element will assert. Like the reactance measurement
derivation, deriving the resistance measurement can also be accomplished by writing
a voltage drop expression (equation 2.3) from the sequence network found in figure
2.2. [2]
VA= Zl1(Ia + Kn) + IfRf
Multiplying both sides of the equation by the complex conjugate of ZLi(Ia + KnIr)
to make the positive sequence line impedance appear real,
VA (Zl1(Ia + KnIr))* = ZL1(IA + KnIr)(ZLi(Ia + KNIR* + IFRF (ZL1(IA +
.......KnIr))*
Taking the imaginary component of both sides of the equation
Im[VA (Zli(Ia + KnIr))*] = Im[ZLi(lA + KnIr)(Zli(Ia + KnIr))*] +........
.......Im [IFRF(ZL1(IA + KnIr))*]
where,
Im[Zu(lA + KnIr)(Zli(Ia + KnIr))*] = 0
Solving for RF,
R Im[VA (1ZZiang(Ia + KnIr))*]
Kf Im [IF (1ZZ,ang (Ia + KnIr))*]
Since IF includes current contributions from both ends of the line, fault current seen
from the relay, 3Io, is approximated in terms of zero and negative sequence currents
to eliminate the effects of load current, i.e.,
12


3Iq Ii +12 + Io
for a SLG fault,
Ii = I2 = Io
Therefore,
3Io = 3/2(I2 + Io)
Substituting,
RF = hnfVA (1^Ziang(Ia + KnIr))*1______________ (2.8)
hn [3/2(I2+Io) (1ZZ1ANG(IA + KNIR))*]
From the derivation it is evident that Rf will only be half the actual fault resistance if
source impedances are equal for both fault contributions. To calculate the true fault
resistance, the ratio of total zero sequence fault current to relay fault current, 3Io,
must be calculated. [3,4]
The measured fault resistance then needs to be multiplied by this ratio to determine
the true value of fault resistance.
RF(true) = RF (measured) IF/3Io(relay) (2.9)
13


2.3 Negative Sequence Impedance Measurement
The negative sequence impedance measurement Z^, was designed to overcome
several limitations imposed by the negative sequence directional relay [5]. These
limitations occur whenever the negative sequence voltage is minimized by strong
negative sequence sources behind the relay or when the negative sequence current is
minimized due to fault resistance. The negative sequence impedance measurement is
derived by compensating the negative sequence voltage in the traditional negative
sequence torque equation shown below ([3,4]);
T = |V2| |I2| cos( ZV2 ZI2 MTA) (2.10)
where MTA is typically set to the positive sequence
impedance line angle, ZZiang
Expressing equation 2.10 in complex conjugate form,
T = Re[V2(I2ZZ1ANG)*] (2.11)
V2 is compensated by rewriting the expression as V2 aZZiANG^
where a = Z2m, the negative sequence impedance measurement.
Substituting,
T = Re[V2-2^M ZZiangWzZZiang)*] (2.12)
The Z2mZZiangI2 term will now increase V2 for forward faults and decrease I2
for reverse faults. Setting the torque equal to zero to define the balance point,
0 = Re[V2-Z2MZZiANGl2(l2ZZiANG)*]
14


Rearranging,
0 = Re[ V2(I2ZZiang)*] (Z2m)Rc[(I2^Zia>jg) (I2ZZ1ANG)*]
Solving for Z2m,
7 _ Ref VzCIzZZiang)*!____________
2M Re[(I2ZZiANG) (I2ZZ1ANG)*]
From the identity,
(l2ZZlAKG)(hZZlAKG)* = \I2\2
the expression now yields,
Z2M = Ref V9(I?ZZi ano)*1 (2.13)
H212
For the relay to sense fault direction, Z2M must be compared to negative sequence
impedance threshold quantities. These threshold quantities are a function of Z2M and
the negative sequence voltage and currents for faults at the relay point. To determine
threshold quantities, a fault study must be conducted to arrive at Z2 in the forward
(Z2F) and reverse (Z2r) directions at the relay point. The negative sequence network
for ground faults shown in figure 2.7 indicates Is2 for forward faults and Ir2
for reverse faults.
15


NEGATIVE SEQUENCE BUS
Figure 2.7: Negative sequence network for a SLG fault
From V2 and I2 at the relay point,
-Vo
Z2F = = -ZS2
tS2
Z2r = ^Y2_
-Ir2
= Zlo+ Zr2
(2.14)
(2.15)
To obtain Z2F and Z2r from the fault study, several contingencies are required
to obtain the strongest negative sequence source behind the relay. If this method is
not convenient, a simplified approach is to set Z2F equal to half of the positive
sequence line impedance and add 0.1 to arrive at Z2r.
From Z2F and Z2r threshold settings, the negative sequence impedance threshold
quantities can now be calculated [1].
16


THRESHOLD CONDTIONS THRESHOLD QUANTITIES
Z2F < 0; Z2FT = 0.75( Z2F) 0.25 | Z2M 1 (2.16)
Z2F > 0; Z2FT = 1 .25(Z2f) 0.25 | Z2M1 (2.17)
Z2R > 0; Z2RT= 0.75(Z2r) + 0.25 | Z2m 1 (2.18)
Z2R < 0; Z2RT= 1.25(Z2r) + 0.25 | Z2m 1 (2.19)
When Z2M< Z2FT a forward fault is declared and when Z2m> Z2RT a reverse fault
is declared. The characteristics for the threshold quantities for a 90 degree MTA are
shown in the Z2 plane in figure 2.8 .
Z2 PLANE
jX2
REVERSE FAULT
FORWARD FAULT
Figure 2.8: Z2 characteristics, MTA = 90
17


For MTAs other than 90 degrees, the negative sequence impedance threshold
characteristics are perpendicular to the positive sequence line impedance which is
typically set at the MTA. This is required to maintain the same threshold
characteristics since positive sequence line and fault impedances remain unchanged
when mapped in the Z2 plane.
Figure 2.9: Z2 characteristics, MTA <90
18


3. Modeling a Nonhomogeneous System
To verify the accuracy of ground quadrilateral measurements, a fault study was
conducted using ASPEN software on the 345 kV 81.7 mile PSCO Tri State
transmission line from Rifle to Craig substations to model a nonhomogeneous
system.
MODELING A NONHOMOGENEOUS SYSTEM
RIFLE CRAIG
SUBSTATION ^ = -^3I0 ZIF = 5 SUBSTATION
345KVZ8 3Io relay 345 KVZO
PT
Figure 3.1: Nonhomogeneous system 345 KV line from Rifle to Craig
19


Fault data for zone 1 faults (85 % of the line) were obtained for increasing 1.0 ohm
increments of fault resistance to determine how fault resistance effects the reactance
measurement (appendix A). Relay inputs for 0 ohm and 5 ohm fault resistances are
shown in tables 3.1 and 3.2.
( Secondary)___________________(Secondary )
VOLTAGE (V) CURRENT(A)
VA= 37.52Z-3.2 5.53Z-81.6
VB = 63.87Z-117.1 .027Z96.1
Vc = 65.0Z 116.6 .027Z96.1
Table 3.1: Relay inputs from fault data (Rf = 0 Q)
(Secondary)____________________( Secondary)
VOLTAGE (V) CURRENT(A)
VA = 40.8Z-11.1 5.24Z-69.2
VB = 63.7Z-117.7 .026Z108.5
Vc = 65.9Z116.6 .026Z 108.5
Table 3.2: Relay inputs from fault data (Rf = 5 Q)
For bolted faults in a nonhomogeneous system the reactance measurement is
unaffected. However if a fault resistance is introduced, the reactance measurement
will have an error term. From fault data and relay settings the reactance measurement
without compensation can be calculated and compared to the reactive reach of the
relay. The difference from this comparison is the reactance measurement error. When
the compensation factor T, is used in the calculation, the reactance measurement
should be within close tolerances to the reactive reach setting for a zone 1 fault.
20


3.1 Reactance Measurement Calculation for the
Zone 1 Fault Location with 5 Q Fault Resistance (T = 0)
|ZL1| = MVa(IrZT )*1___________
Im[lZZ1ANG (Ia+KnIr)(IrZT )*]
Since IA Ir, then,
|ZL1| = hn[VAaAZ-e(lZT))*3________
hn[lZZIANG(IA(l+ Kn)(IrZT )*]
From the following identity,
Ha|2 = (Ia)(Ia)*
the equation can be simplified to
|ZL1| = Im[VA(lIA|Z-9(lZT ))*]
Im[lZZ1ANG|IA|z(l+ Kn)(1ZT )*]
= ImrVA(lZ-e(lZT))*l
|IA| Im[lZZ1ANG(l+ KnX1ZT )*]
= Imr40.8Z-ll.iriZ-69.2 QZO T)*l
5.24 hn[lZ86(l+ .75Z-19X1Z0 )*]
= 3.92 n
21


3.2 Reactance Measurement Calculation for the
Zone 1 Fault Location with 5 Q Fault Resistance (T=5 )
iZl1|= Imr40.8Z-ll.mZ-69.2 (TZ5 Il*l
5.24 Im[lZ86(l+ .75Z-19)(1Z5 )*
= 3.77 Q
It is also important to model the system since the true fault resistance from the fault
study does not equal the measured fault resistance. The measured fault resistance
depends only on the relays contribution of zero sequence fault current. The true fault
resistance measurement will depend on the ratio of total zero sequence current to the
relay zero sequence fault current. From this ratio, the true fault resistance can be
calculated and checked with the resistance from the fault study.
3.3 Resistance Measurement Calculation For the
Zone 1 Fault Location
(True Resistance From Fault Study = 5 Q)
d hnfVA (1ZZiang(Ia + KnIr))*1_________
F hn [3/2(Ia2 + Iao)(1ZZ1ang (Ia + KNIR))*]
Since,
Ia Ir> I Ia I Ia-^-0
Ia2 Iao W3
then,
22


Rp = Im[VA (|IA| z-eq + Kn)1ZZ,ANq)*1
Im[IA(IA(l+KN)lZZ1ANG)*]
From the following identity,
|Ia|2 = (IaXW*
The equation can be simplified,
pr= ImrVA(lZ-e(l+KN)lZZ,ANn)*1
|IA| Im [((1+Kn) 1ZZ1ang)*]
Rf = Imr40.8Z-ll.l (l Z-69.2 (1 + .75Z-19UZ861*!
5.24 hn [((1+ .75Z-19) 1Z86)*]
= 2.694 n
To calculate the true fault resistance the following equation is used
Rp(true) = Rp (measured) If/3Io(relay)
Rp (true) = 2.694 x 2885.4
1418.3
= 5.48 Q
The fault resistance measurement is now consistent with the resistance from the
fault study when this approximation method is used.
To verify fault direction for the system model, negative sequence impedance
measurements and threshold quantities should be calculated from the fault study at
the zone 1 fault location. Directional quantities should be calculated for both bolted
faults and faults with resistance at the resistive reach setting to understand how the
threshold characteristics are affected by fault resistance in the Z2 plane.
23


3.4 Negative Sequence Impedance Measurement
Threshold Calculations for a Bolted Fault at
the Zone 1 Location
From table 3.1 relay inputs,
V2 = 1/3(Va + a2VB+ aVc)
= 1/3(37.52Z-3.2 + 1Z240(63.87Z-117.1) + 1Z120(65.OZ 116.6))
= 11.04Z-175 V
I2 = 1/3 (IA + a2IB + ale)
= 1/3(5.53Z-81.6 + 1Z240(0.27Z96.1) + 1Z120(0.27Z96.0))
= 1.93Z-81.7 A
3I2 = IA = 5.79 A (exceeds typical 3I2 supervision setting of 0.5A)
z = Ref V2(I2ZZiang)*1
2 lEF
= Ref 11.04Z-175fl.93(Z-81.7('lZ86V)*l
11.9312
= 5.72 Q.
= 1.25(Z2F)-0.25|Z2m|
= 1.25(3.6)-0.25(5.72)
= 3.07 Q
Z2m < Z2ft ; forward fault declared
24


3.5 Negative Sequence Impedance Measurement
Threshold Calculations for 5 Q Fault Resistance
at the zone 1 Location
From table 3.2 relay inputs,
V2=l/3(VA + a2VB + aVc)
= 1/3(40.8^-11.1 + 1Z240(63.7Z-117.7) + 1Z120(65.9Z116.6))
= 10.53Z-163.4V
I2 = 1/3 (IA + a2IB + ale)
= l/3(5.24Z-69.2 + 1Z240(0.026Z108.5) + 1Z120(0.026Z108.5))
= 1.76Z-69.2 A
3I2 = IA = 5.28 A (exceeds typical 3I2 supervision setting of 0.5A)
Z2m= ^2^ZZiAng)*1
\h\
_ Ref 10.53Z-163.4fl.76Z-69.2QZ86V>*l
|1.71|2
= 5.98 Q
Z2Fr = 1.25(Z2F) 0.25|Z2m|
= 1.25(3.6)-0.25(5.98)
= 3.on
Z2m < Z2Ft ; forward fault declared
25


The margin between Z2m and Z2ft is slightly greater for the 5 ohm fault resistance,
therefore, the fault will appear more in the forward direction when fault resistance is
introduced. Creating a region between the threshold settings will prevent high fault
resistances from causing a fault behind the relay to appear in the forward direction.
For this example, a 5 ohm fault resistance does not significantly affect the directional
capabilities of this model.
26


4. Dynamic Test Simulations
To establish the reactance measurement base line for a dynamic test simulation, an
impedance search plot test was conducted for a zone 1 ground quadrilateral in a
SEL 321 (Schweitzer Engineering Laboratories) relay at Rifle substation using
PROTEST 3 software. The search plot calculates values of apparent ground
fault impedance by setting the equation in the PROTEST 3 Z PLOT binary search
macro to Z=V/I. Apparent impedances are plotted at specified increments of test
current angles (figure 4.1).
Figure 4.1: PROTEST 3 Z PLOT [6]
The voltage and currents from relay test results were then entered into an Excel
spreadsheet to convert the apparent ground fault impedances to current compensated
resistance and reactance measurements that correspond to reach settings of the
27


ground quadrilateral. These results were then checked for accuracy and plotted.
Relay voltage and currents from the fault study were also entered in the spreadsheet.
Fault reactance and true resistance measurements were then plotted against the
results from the search plot to graphically show reactance measurement error when
the T setting was set to zero (figure 4.2). When the T setting was changed to 5, the
reactance measurement was reduced significantly (figure 4.3).
28




ZL1 (OHMS)
FIGURE 4.3: REACTANCE MEASUREMENT ERROR (T=5)
SEL 321 ZONE 1 GROUND QUADRILATERAL FOR 345 KV LINE
TO CRAIG


5. Results and Conclusion
Reactance measurements for ground faults are set in terms of the positive sequence
line impedance to determine the fault location. This is convenient since the positive
sequence line impedance is directly proportional to the length of the line.
However, it is evident from the fault study that the reactance measurement will not
be exactly proportional to the length of the line if the system is nonhomogeneous
and includes fault resistance. A system is nonhomogeneous when the total zero
sequence fault current angle is not equal to relay zero sequence fault current angle.
This angular difference is defined as the nonhomogeneous compensation correction
factor angle setting T. The reactance measurement line from figure 4.2
tilts at this angle when referenced to the reactive reach line generated from the
impedance search plot. If the correction factor T is applied, the
reactance measurement error is now proportional to the length of the line for ground
faults. This method of testing requires a high volume of complex calculations. By
creating a spreadsheet, protection engineers and technicians can conveniently enter
relay settings, impedance search plot points and fault data to augment testing and
verify relay accuracy.
31


Appendices


Appendix A
Dynamic Test Simulation Reactance Measurement Error (T = 0)
SEL 321 Zone 1 Ground Quadrilateral 345 KV line to Craig
IMPEDANCE SEARCH PLOT TEST
PROTEST RESULTS la magnitude Va @ 0 deg la (calc) Region R const X const 1% ERROR R (calc) X (calc)
0 5.98 30 6.000 0.33% 5.017 -0.265
20 5.16 5.197 0.71% 5.036 0.912
40 3.74 3.767 0.72% 5.037 2.789
48.2 3.04 3.038 0.07% 4.996 4.094
60 3.64 3.472 4.84% 2.587 4.092
80 4.3 4.094 5.03% -0.265 4.085
90 4.44 4.222 5.15% 1.452 4.080
105 4.36 4.174 4.93% -3.195 4.088
120 4.12 3.841 7.25% 4.996 4.000
135 5.16 -5.155 0.11% -4.995 1699
165 6.12 -6.129 0.15% -5.008 1.016
SETTINGS
kOM 0.75
kOA -19
Z1ANG 86
T 5
RG 5
XG 4.1
1+K0 t+k01
Magnitude Angle
1.726493 -8.13054
FAULT STUDY
ASPEN RESULTS T- o
la (calc) 1% ERROR R (calc) X (calc)
Region
la magnitude Va angle Va magnitude Fault 2
-81.6 5.53 -3.2 37.52 a 5.206 6.23% -0.064 3.937
-79 5.49 -5.1 37.94 1 5.163 6.33% 0.489 3.934
-76.5 5.44 -3.9 38.5 2 5.111 6.44% 1.041 3.930
-74 5.38 -8.5 39.18 3 5.050 6.54% 1.596 3.926
-71.6 5.31 -9.9 39.95 4 4.982 6.58% 1143 3.924
-69.2 5.24 -11.1 40.8 5 4.906 6.81% 2.694 3.916
33


Appendix B
Dynamic Test Simulation-Reactance Measurement Error (T = 5)
SEL 321 Zone 1 Ground Quadrilateral 345 KV line to Craig
IMPEDANCE SEARCH PLOT TEST
PROTEST RESULTS la (calc) 1% ERROR R (calc) X (calc)
Region
la angle la magnitude Va @ 0 deg R const X const
a 5.98 20 6.000 0.33% 5.017 -0.265
20 5.16 5.197 0.71% 5.036 0.912
40 3.74 3.767 0.72% 5.037 2.789
46.2 3.04 3.038 0.07% 4.996 1.094
60 3.64 3.472 4.84% 2.587 4.092
60 4.3 4.094 5.03% -0.265 4.085
90 4.44 4.222 5.15% 1.452 4.080
105 4.38 4.174 4.93% -3.195 4.088
120 4.12 3.841 7.25% 4.996 4.000
125 5.16 -5.155 0.11% -4.995 2.699
165 6.12 -6.129 0.15% -5.008 1.016
SETTINGS
kOM 0.75 1+K0 1+k01
kOA -19 Magnitude Angie
Z1ANG 66 1.726493 -8.13054
T 5
RG 5
XG 4.1
FAULT STUDY ASPEN RESULTS la angle la magnitude T> Va angle 5 Va magnitude Fault Z la (calc) Region l% SRROR R (calc) X (calc)
-81.6 5.53 -3.2 37.52 0 5.080 8.86% -0.064 3.941
-79 5.49 5.1 37.94 1 5.001 9.78% 0.489 3.908
-76.5 5.44 -6.9 38.5 2 4.914 10.71% 1.041 3.875
-74 5.38 -8.5 39.18 3 4.818 11.67% 1.596 3.842
-71.6 5.31 -9.9 39.95 4 4.718 12.56% 1143 3.811
-69.2 5.24 -11.1 40.8 5 4.610 13.67% 1694 3.774
34


Appendix C
C.l Derivation of Residual Current Compensation
Factor,Kn [7,8]
RELAY FAULT
LOCATION LOCATION
Figure C.l: sequence network for a SLG fault (Z1-Z2, phase A)
V1 p+V2F+Vop 0
35


VlF - (V2F+Vof) (C.l)
Vif=Va-Ii(Zli) (C.2)
V2f = I2(Zu) (C.3)
Vqf = Io(Zlo) (C.4)
Substituting equations C.2, C.3 and C.4 into C.l yields,
VA-Ii(ZLi) = I2(ZL2)+I0(ZLo)
For an a phase to ground fault, Ib = Ic = 0
Therefore,
Ii = I2 = Io = Ia/3
Then,
VA = Ii (Zli +Zl2+Zlo)
Substituting Zli = Zl2, the expression now yields,
Va=Ii(Zli)+I2(Zu)+Io(Zlo)
Rearranging this expression by substitution from the equation,
3Io = Ia+Ib+Ic, yields,
Va= IiZli+I2Zli+(Ia +Ib +Ic)/3Zlo
= Zli[ Ii+I2+(Ia+Ib+Ic)Zlo/3Zli ]
since, IA = I1+I2+I0, or
h = lA-h-h
36


Then,
Va- Zy[ Ia-I2-Io+I2+(Ia+Ib+Ic)Zlo/3 Zl i ]
= Zli[ Ia-Io+(Ia+Ib+Ic)Zlo/3ZLi ]
= Zu[Ia-Io+3Io(ZLo/3Zl1)]
= Zl1[Ia+3I0(-1/3+ZLo/3Zl1)]
= ZL][ Ia+(Ia+Ib+IcXZlo-Zl1)/ 3Zu]
where Zlq-Zli is the residual curr3Zuent compensation factor Kn
3ZLi
Kn =
Zlq-Zli
3ZLi
and Ia+Ib+Ic is the residual current,
Ir Ia+Ib+Ic
Substituting,
Va Zli(Ia+KnIr )
Dividing by the sum of A phase current, Ia and compensated residual current, Ioc,
the expression now yields
Va = Zli(Ia+KnIr) = zL1
Ia+Ioc Ia+KnIr
Kn can also be written, Kp -1
where Kq is the zero sequence compensation factor Zlq/Zli
37


Appendix D
Derivation of Zero Sequence Current
Compensation Factor, Ko [7,8]
Ko is derived in a similar maimer to Kn. From the SLG fault sequence network in
Appendix C,
VA=(Ii+I2)ZL1+I0(ZLo)
Rearranging,
VA=ZL1[I1+I2+(ZL0/ZL1)I0]
where Zlo/Zli is the zero sequence current compensation factor Ko
Ko = Zlo/Zli
and Io is the zero sequence current
substituting
VA = ZL,(Ii+l2+KoIo)
and dividing by the compensated A phase current, IAc, in terms of symmetrical
quantities, the expression now yields,
VA = Zli(Ii+I2+KqIq) =zu
Iac Ii+I2+KoIq
38


Appendix E
Effect of Transmission Line Conductor Changes
on Kiv and Ko [9]
From the following equations,
7 jcopo[l/4+ln(D/R)] po = 4tt xlO'7, co = 2nf
L1 ~2n
7.lo= j 2K
where, D = Geometric mean distance of phase conductors
Dn = Average distance of neutral conductor to phase conductors
R = Geometric mean radius of phase conductors
Rn = Geometric mean radius neutral conductor
And,
Kn= Zl3~f7 ^Zlq/Zu
It is evident that changes in conductor spacing, radius and bundling result in
impedance changes that are not proportional to the length of the line and that vary
from positive sequence to negative sequence. Therefore Kn and Ko do not remain
constant over the length of the transmission line. Settings on the SEL 321 address
this problem by assigning a dedicated residual compensation factor for zone 1.
39


References
[ 1 ] Schweitzer Engineering Laboratories Inc. SEL321 Instruction Manual
Pullman, Washington, March 20, 1998 edition
[ 2 ] S.E. Zocholl. Three Phase Circuit Analysis and the Mysterious K0 factor
Presented before the 22nd annual Western Protective Relay Conference
Spokane, Washington, October 24-26,1995
[ 3 ] J. Mooney, P.E., J. Peer. Application Guidelines for Ground Fault Protection
Presented before the 24th annual Western Protective Relay Conference
Spokane, Washington, Oct 21-23,1997
[ 4 ] E.O. Schweitzer, ID, J. Roberts. Distance relay element design
Presented before the 46th annual conference for protective relay engineers
Texas A&M University, April 12-14,1993
[ 5 ] B. Flemming. Negative Sequence Impedance Directional Element
Presented before the 10th Annual Protest User Group Meeting
Pasadena, California, February 24-26,1998
[ 6 ] C. Gallegos. Provided 321 test results using the Protest 3 relay
testing software for the SEL 321 installation at Rifle substation, 1999
[ 7 ] GE MultilinDistance Relays Fundamentals GER-3966
[ 8 ] GEC Measurements. Protective Relays Application Guide
Published GEC Measurements, Stafford England, 1987
[ 9 ] O. Elgerd. Electric Energy Systems Theoryan Introduction. 2nd edition
McGraw-Hill, Inc. 1982
40


38
predate ceramic technology it is still possible to
compare the ceramics recovered with known types of
Anasazi, Fremont, and Numic origin.
In the Anasazi tradition, the general grayware
types that are noted include Chapin Gray, Moccasin Gray,
Mancos Gray, and Mancos Corrugated. These types are
common throughout the Four Corners region, and are typical
of the Mesa Verde Anasazi variant (Breternitz et al.
1974) .
In the Fremont tradition, the two grayware types
that might occur in this area are Uinta Gray and Emery
Gray (Madsen 1977). Both of these types are similar to
those recovered from Turner-Look, the largest, and
closest, investigated Fremont inhabitation to Luster Cave.
Buckles (1971:531) felt that the ceramics
recovered from Luster Cave were similar to those of the
Fingertip Impressed Type of Uncompahgre Brownware; a type
he feels is typical of the proto-historic cultures that
occupied the Uncompahgre Plateau.
Perishables
Luster and Roth Caves were both protected from the
elements that often afflict more open sites. As a result,
a variety of perishable items that are normally lost
through the processes of decomposition were recovered,
including several fragments of basketry and a quantity of
cordage and quids. The specimens were measured with


39
calipers, and all measurements will be recorded in the
metric system.
Basketry is a class of perishable artifacts that
includes several distinct kinds of items, among which are
rigid and semirigid containers (or baskets proper),
matting, and bags. All forms of basketry are manually
woven without frame or loom. As all basketry is woven, it
is technically a textile class or variety; however, that
term is often restricted to cloth fabrics. Cordage is a
class of elongated fiber constructions that can be
subsumed under the English terms string and rope. The
manufacture of cordage is the oldest fiber-based
technology in the New World archaeological record. The
technological sophistication of these very early cordage
specimens indicates considerable antecedent development.
Cordage manufacture was very likely part and parcel of the
technological repertoire of the earliest migrants to this
hemisphere (Adovasio 1988). j
Basketry specimens were analyzed following
Adovasio (1977) basing the analysis on observances and
measurements of coiling type, foundation type, and stitch
type. Cordage specimens will be allocated to five
structural types. The five structural types were
established on the basis of three interrelated
construction attributes; 1) number and composition of
plies; 2) direction of initial "spin"; and 3) direction of
final twist (Adovasio 1988).


40
The term ply is utilized to mean a strand or bunch
of fibrous material that is almost always twisted. These
strands can be used alone to form single-ply cordage or in
groups to form multiple-ply cordage. Multiple-ply cordage
is produced by twisting two or more "single" plies
together.
An individual ply is simple if it consists of a
single strand or bunch of material with the same twist. A
ply also can be compound. Compound plies are constructed
with multiple strands or bunches of material that are
individually twisted and then twisted with each other in
the opposite direction. Compound plies are therefore
separate pieces of cordage that when twisted with other
such plies form a technically distinct final cordage type.
Spin denotes the initial twist imparted to a fiber
strand or bunch of fibers, and final twist records the
direction imparted to several plies that have been twisted
together. The term spin is used here to designate the
initial twist of a ply because this term is virtually
universal in the archaeological literature on cordage and
because it facilitates cordage description. The direction
of initial spin or final twist can only be S or Z, and
these terms have exactly the same meaning as specified by
Emery (1966:11).
Each specimen was also analyzed for the presence
of splices and knots, as well as length, diameter, number
of twists per centimeter, angle of final twist, and other


41
cordage manipulations, such as rat-tailing, wrapping, etc.
Angle of twist measurements were recorded using procedures
outlined by Emery (1966:11). Cordage formulae follow
Hurley (1979), and the formulae for the five types
identified in the assemblage are shown in Table 1.
TABLE 1
CORDAGE FORMULAE
Type Description Ply Formula
I Single ply, Z twist z
II Single ply, S twist s
III Two ply, Z spun, S twist s z/z
IV Two ply, S spun, Z twist z s/s
V Compound three ply, S and Z S twist spun s z/z s z/z
Adovasio (1975; 1980) states that there is nothing
more homogenous and diagnostic than the basketry
associated with the Fremont culture. Within the Fremont
basketry assemblage are included coiling and twining two
of the three major sub-classes of basket weaves the
third, plaiting, is virtually absent. Coiling is the
numerically dominant subclass of Fremont basketry and is
represented in all Fremont sites where basketry is
preserved. Basketry produced via twining techniques is
relatively uncommon in most Fremont Sites and frequently
is not represented at all.
Trends discernible throughout the culture area
over the 900-year period during which Fremont coiling was


42
produced include a gradual shift from mixed to almost
uniformly R-L work direction, the increased preference of
half rod and bundle foundation to all others, and the
tendency to employ non-interlocking or intentionally split
stitches on the non-work surface to all other types.
Thus, it is reasonable to expect the basketry from Roth
and Luster Caves to reflect these trends, if they are
affiliated with the Fremont Culture.
Bone
Only the worked bone will be discussed within the
context of analysis methodology, as all the unworked
faunal material was placed in the care of another
department at the University of Colorado/Boulder, and has
subsequently been misplaced. Unfortunately even the
curation records only refer to "animal bone" rather than
to specific types precluding even the slightest inference
of information from the records. The worked faunal
material represents a total of 17 specimens from the
collection.
Worked bone has traditionally been categorized in
a variety of classes according to arbitrarily selected
characteristics of the bone from each respective site.
However, in the past ten years the "new archaeology" has
finally reached into the realms of worked bone analysis to
include studies such as the classification of tool types
through usewear analysis (Gooding 1980), experimental


I 43
studies of specific bone implements, such as deer ulnas
(Harrell 1983) and comparative analysis (Dailey 1970a).
Unfortunately, the majority of sites do not produce
samples large enough for major renovations in the analysis
of worked bone, including Roth and Luster Caves. Thus, a
basic analysis determining types of worked bone present -
such as gaming pieces or awls will be conducted, and where
possible, pertinent information as to usewear and
manufacturing techniques will be included.
Macro/Microbotanical Remains
Evidence of cultigens is relatively rare in
archaeological contexts in west-central Colorado. Sites
yielding cultigens are scattered throughout the region in
the lower elevations. Nearly all of the finds are corn,
but two sites, Tabeguache Cave II and 50R243 have yielded
squash remains (Reed 1984). Whether these finds represent
domestic or wild varieties of squash is unknown.
Macro/microbotanical remains provide information
to answer questions about paleoenvironmental data and
subsistence. Several items in the collection could have
been utilized to provide palynological and macrobiotic
information, had certain precautions been taken during the
excavation to prevent contamination (Adams and Gasser
1980; Scott 1983). However, most palynological techniques
were developed 10 20 years after the excavation of Roth
and Luster Caves took place.


44
Macrobotanical remains recovered from the
excavation and curated with the collection will be
examined for charring and other indications that they were
utilized by the population, rather than being by-products
of the local environment (Minnis 1981). Seeds can occur
from several sources, both prehistoric and modern, and in
caves such as Luster and Roth, preservation of both is
typical because of the dry environment, as opposed to open
sites, where generally only charred seeds are preserved at
any depth below the modern surface level (Minnis 1981).
It may often be the simplest, however, to reject all
uncharred undomesticated seeds as modern in origin and to
retain only the charred material as genuine (Keepax
1977:226). Many ethnobotanists use this as a basic rule,
and given the nature of this research, it will be the
basic guideline for determining the status of the
macrobiotic remains.
Of consideration as well is the analysis of the
maize present in the collection. Because of the lack of
funds for re-examination by a paleobotanist, and due to
the relative inexperience in this category by the author,
the analysis of the corn remains by Nickerson provided in
the final publication about Luster Cave will be utilized
(Wormington and Lister 1956) to compare with the current
analysis.


45
Human Bone
The analysis of the human remains will follow
those procedures outlined in Brothwell (1981).
Unfortunately, both the infant burial from Luster Cave
and the child burial from Roth Cave are missing from the
collection. As a result, only the dentition can be
observed, with any conclusions that can be drawn based on
those findings.
Wood
While wood is generally considered only to be
found in the form of charcoal or posts, within the context
of Luster Cave, a variety of worked wood items were
recovered. It may be possible to examine these specimens
for additional information about the types of projectile
points that were being utilized, as well as construction
techniques of the item itself. Artifactual determination
of wooden specimens will be based on any occurrence of
smoothing, cutting, shaping, or binding of the specimens.
Once identified as cultural, the artifacts will be
segregated on the basis of function, although, where use
is not apparent, descriptive categories will be
established.
Charcoal can be utilized for C14 tests, as well
as identified given the appropriate collection with
which to compare it.


46
Minerals
The minerals will be identified according to type.
Any cultural modification that may be present will also be
noted.


CHAPTER 5
LUSTER CAVE ANALYSIS
Lithics
The excavations recovered a lithic assemblage of
211 artifacts. Seventy-nine specimens are classified as
tools while the remaining 132 are debitage. Tools were
classified and analyzed according to basic production
technologies and include flaked or chipped stone and
ground and pecked stone. However, these groupings are
somewhat arbitrary and some overlapping occurs.
Analysis was conducted on both tools and debitage.
Variables which emphasize production technology were the
primary focus of the initial analysis, followed by a
second analysis which focused on specific attributes which
would allow for cross comparison with other lithic
assemblages. Of the 211 specimens present in the
collection from Luster Cave, there were 22 types
delineated within the technological analysis. The
distributions of these types across the site are recorded
in Table 2.
There were two specimens making up the first
type. This type was produced from obsidian; both


TABLE 2
DISTRIBUTION OF TECHNOLOGICAL TYPES IN LUSTER CAVE
48
Area A B C i D E
Level Surf. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 1 12 1
Type 1 1
Type 2 2
Type 3 1
Type 4 1
Type 5 1 5 5 3 1 3 4 4 3 2 3 5 4 2 2 4 2
Type 6 1 2
Type 7 1 3 1 1 3 2 1 1 1
Type 8 1 1
Type 9 3 a 2 5 2 2 2 b 2 3 2 2 1 1 1
Type 10 1
Type 11 1 2
Type 12 1 1 1
Type 13 1
Type 14 1 2 1
Type 15 1
Type 16 1 2 2 1 1 1
Type 17 1 1 1 2 2 2
Type 18 1 1 1
Type 19 1
Type 20 1
Type 21 2
Type 22 1
a = 11, b = 12
Area Level F 13 4 5
Type 1 1
Type 5 1 2
Type 7 2
Type 9 1 d c 3
Type 14 1
Type 16 2
Type 18 2
c
13, d = 17


49
specimens were tertiary flakes. There was one edge which
had been retouched on one specimen; the other specimen had
no visible wear apparent on any edges. Boith of these
specimens are classed as debitage. Because the sample of
obsidian was so small, no further testing on the obsidian
samples was considered.
There were two specimens comprising Type 2. Both
specimens were choppers of basalt, showing mixed cortical
and non-cortical surfaces (Figure 8). On one specimen
there was partial battering along one edge, while the
other chopper exhibited no such wear. Dimensions are 5.38
and 7.4 cm. in length, 6.9 and 6.0 cm. in width, and 3.0
and 3.5 cm. in height.
Figure 8.
Basalt chopper.


50
There was one specimen which made up the third
type. This specimen was a basalt uniface which had been
utilized along one edge, and may be considered a chopper.
It exhibits mixed cortical and non-cortical surfaces.
The following eleven types were all made from
cryptocrystalline silicates, which in all cases from
Luster Cave are chalcedony and chert all of which
appears to have been quarried from local, known quarry
areas (Piontkowski, personal communication).
One specimen makes up this Type 4. It is a
secondary biface of whitish chalcedony. This specimen has
been stained with red ocher in two places along its base.
Fifty-six specimens are included in Type 5, all of
which are tertiary flakes of a variety of chalcedonies and
cherts. Although there did not appear to be visible wear
on the specimens, all of the specimens within this
category are modified bifaces and will be discussed in
detail later in the section. The specimens were recovered
throughout the fill from the cave.
Type 6 is comprised of three specimens of chert.
Each specimen is a tertiary biface, with additional
modification. The three specimens will be discussed later
in the section.
The specimens within Type 7 are all modified
tertiary bifaces of either chert or chalcedony. There
does appear to be visible wear on each specimen.


51
Type 8 is comprised of two unmodified tertiary
unifaces of chalcedony. One specimen appears to have
visible wear along one edge.
Type 9 consists of 84 specimens of various cherts
and chalcedonies. All the specimens are unmodified,
tertiary flakes with wear visible on the surfaces of only
three specimens. All the specimens within this type are
considered debitage, except for the three exhibiting wear,
these specimens are considered utilized flakes. Specimens
of this type were recovered throughout the fill.
Type 10 consists of one specimen. It is a chunk
of chert, without any visible wear or modification.
Type 11 is comprised of three specimens of chert.
The specimens appeared to have mixed cortical and
noncortical surfaces, with little visible wear and no
additional modification. The specimens are considered to
be core fragments.
Three specimens considered to be choppers make up
Type 12. The specimens are of chalcedony, with mixed
cortical and non-cortical surfaces. Some wear was visible
on the edges of the specimens, although there was no
additional modification.
Type 13 is a tertiary biface of siltstone with no
visible wear and no additional modification.
Type 14 is comprised of five specimens of
siltstone. All are secondary flakes with no visible wear


52
or additional modification. All specimens in this type
are considered to be debitage.
One secondary flake of siltstone comprises Type
15. It exhibited no visible wear or additional
modification.
Type 16 is comprised of ten specimens of
sandstone. All specimens are considered cobbles and show
extensive visible wear, although there is no additional
modification. All the specimens in this type are
considered to be groundstone and will be discussed later
in the section.
Type 17 consists of nine specimens of Brushy Basin
quartzite. The specimens are tertiary bifaces, some of
which have additional modification and will be discussed
in further detail later in the section.
Five tertiary flakes of Brushy Basin quartzite
comprise Type 18. There was no visible wear or additional
modification present. All the specimens in this type are
considered debitage.
Type 19 consists of one specimen which is a chunk
of quartzite without any visible wear or additional
modification. The specimen is so clear, it was almost
mistaken for glass.
Type 20 consists of one specimen of quartzite. It
is a cobble with extensive visible wear. This specimen is
considered to be a hammerstone (Figure 9).


53
Figure 9. Quartz hammerstone.
Type 21 consists of two cores with mixed cortical
and non-cortical surfaces of quartzite. There appears to
be limited wear, and no additional modification.
One primary pebble of quartzite comprises Type 22.
There does not appear to be any visible wear or additional
modification. Its possible use is unknown.
Groundstone
There were ten specimens which were determined to
be groundstone all of sandstone. Nine of the specimens
are manos, while one specimen is a piece of sandstone
which has been deeply grooved possibly having been used
to smooth or sharpen bone or wood (Figure 10). The
dimensions of this specimen are 9.27 cm. x 3.14 cm. x 2.15
cm.


54
The manos were of various shapes and sizes (Table
3). Included in the shapes were two rectangular
specimens, one of which had tapered ends (Figure 11).
There were two loaf-shaped specimens, and three round
(Figure 12, 13). There were also two round/asymmetrical
specimens as well. All of the manos were small, and are
considered to be one-handed manos. There were battering
marks on most of the specimens which indicated either
shaping or the use of the mano to pound the grain before
it was ground. There were six two-sided manos and three
one-sided manos. Lengths ranged from 6.51 cm. to 14 cm.;
widths ranged from 4.75 cm. to 10.3 cm.
The mano tool is generally shaped to match the
style of the metate on which it is to be used. Thus, the
small, one-handed mano may be slightly shaped from a river
cobble to fit the basin-shape curvature of the metate.
TABLE 3
DISTRIBUTION OF GROUNDSTONE IN LUSTER CAVE
Type Area Level Material # Handed # Sided
Loaf-Shaped A 5 Sandstone 1 2
Loaf-Shaped B 3 Sandstone 1 2
Rectangular Surface Sandstone 1 1
Rect./Tprd. B 4 Sandstone 1 2
Round B 4 Sandstone 1 2
Round C 2 Sandstone 1 1
Round/Tprd. A 5 Sandstone 1 2
Round/Asym. F 5 Sandstone 1 1
Round/Asym. F 5 Sandstone 1 2
Grvd Sndstne A 1


55
Figure 10. Sandstone arrowshaft smoother.
Figure 11. Rectangular mano


56
Figure 12.
Loaf-shaped mano.
Figure 13. Round mano


57
Bifaces
The bifaces were categorized as follows: 32
points, nine scrapers, eight knives/knife bases, two drill
tips and 24 utilized flakes/bifaces (Figure 14).
Following is a discussion and typological comparison of
the bifaces which are considered to have been modified
into projectile points. Their distribution within Luster
Cave is noted in Table 4.
There were sixteen points with no typological
comparison (Figure 15). The first specimen is corner-
notched with a flaring stem (Figure 15a). It is of
chalcedony. The interesting thing about this point is
that it has a black residue, possibly resin, that covers
parts of the stem.
The second specimen is a side-notched fragment
(Figure 15b). The third specimen is also a side-notched
fragment, although this specimen appears to have been
modified more extensively than the second (Figure 15c).
The fourth through seventh specimens are
notched, spirate triangles (Figure 15d, e).
The eighth and ninth specimens were very rough,
side-notched fragments (Figure 15f).
The tenth and eleventh specimens are small,
well-made points. Both specimens are side-notched,
although one is also notched in the center of the base
(Figure 15g, h).


58
TABLE 4
DISTRIBUTION OF PROJECTILE POINT TYPES IN LUSTER CAVE
Area A B C D E F
Level 12 3 4 1 2 3 4 5 12 3 4 112 15
Untyped 111 1 3 2 3 1 2 1
Buckles' 2 1 1
Buckles' 5 1 1 1
Buckles' 7 1
Buckles' 10 1
Buckles' 18 1
Buckles' 19 2
Elko 1
Pinto 1 1
Gypsum 1 2 1
Desert S-N 1 1
Nawthis S-N 1 1
Uinta S-N 1
Parowan B-N 1 1
Figure 14. Bifaces


59
Figure 15
Untyped projectile points
(actual size).


60
The twelth specimen was a tip fragment, with no
indication of basal modification.
The thirteenth specimen is a non-descript point.
It appears to be a biface, although there is a slight
notching of one side (Figure 15i).
The last two points that were not typed are two
side-notched specimens, although one specimen appears be
somewhat Gypsum-like (Figure 15j).
There were two specimens that compared favorably
with Type 2 described by Buckles (1971). The second of
the two specimens also compares favorably with the Desert
Side-Notched point of Fremont origin (see the Discussion
on Desert Side-Notched points later in section) (Figure
16). Buckles describes this type as small side and
basal-notched points, possibly representing a type that is
more commonly found to the west of the mountains (Buckles
1971:116, Figure 2). This may suggest a Great Basin
origin. Buckles also compares this point type with those
found on the Plains after A.D. 1500 (Buckles 1971:116;
Kehoe 1966).
There are three specimens in the next category
that are similar to Buckles' Type 5 (Figure 15k). Buckles
describes this type as small basal-notched points with
triangular shaped bodies and short expanding bases
(Buckles 1971:119-120, Figure 2). Two of these points
also compare favorably to the Parowan Basal-Notched type
of Fremont origin (Figure 16). Buckles suggests this type


61
Figure 16. Typed projectile points. Elko Side-notched
(a), Pinto Shoulderless (b, c), Gypsum (d,
e), Desert Side-Notched (f, g), Nawthis
Side-Notched (h, i), Uinta Side-Notched
(j) and Parowan Basal-Notched (k, 1). All
points are actual size.


62
resembles large projectile points of Basketmaker and
Pueblo II populations, as well as points found in Fremont
sites (Buckles 1971:146).
The next specimen is comparable to that of
Buckles' Type 7 (Buckles 1971:120, Figure 2) (Figure 16).
It is also comparable to the Gypsum point of Archaic
origin.
The following specimen is comparable to that of
Buckles Type 10 (Buckles 1971:122). Buckles describes
this type as being small triangular unnotched points with
slightly convex to straight sides and slightly convex to
straight bases. Buckles suggests that these points have
their highest frequency relationships with what are
believed to be historic and proto-historic Ute occupations
defined as the Escalante Phase on the Uncompahgre Plateau
(Buckles 1971:122, Figure 2).
The next specimen compares with Buckles Type 18,
as well as Desert Side-Notched (Buckles 1971:130-131,
Figure 3) (Figure 16). Buckles describes this point type
as being large side-notched points with straight or
slightly concave bases. Buckles compares these with
similar specimens from Danger Cave (Jennings 1957:121,
Figures 97a and b), with specimens from the Bitterroot
Phase in Idaho and with early points from Plains sites
(Buckles 1971:130-131).
The next points are similar to Buckles Type 19
(Buckles 1971:131, Figure 3). They are also comparable to


63
Desert Side-notched specimens of late prehistoric origin
(Holmer 1980:60).
There is one specimen which compares favorably
with the Elko Side-Notched type (Holmer 1978) (Figure
16a). The Elko series has traditionally been divided into
three variants: Corner-Notched, Side-notched, and Eared.
The Elko Side-Notched is similar in form to the Elko
Corner-Notched except that the maximum stem width is
approximately equal to the maximum blade width. Tangs are
rarely present; the distal notch angle often approaches
horizontal, causing a shouldered appearance to the blade.
The Elko series projectile points are the most
plentiful but the least temporally diagnostic of the point
types commonly found in the northern Colorado Plateau and
the far eastern Great Basin. About all that can be
positively stated is that they occur after 7600 B.P. They
possibly persisted into historic times; Powell collected a
Paiute hafted knife incorporating and Elko-like point in
1873 (Fowler et al. 1973:41). The great time depth of the
Elko series points refutes the suspicions of several
(Clewlow 1967; Heizer and Baumhoff 1961; O'Connell 1967)
that their temporal occurrence everywhere in the Great
Basin falls between 3500 and 1400 B.P. That estimation is
based mostly bn central and western Great Basin research
with the Danger Cave data being discounted as aberrant
(Heizer and Baumhoff 1961) Hester and Heizer (1973),
however, acknowledge that the Elko series points are


64
probably earlier in the far eastern Great Basin than to
the west, and a diffusion of the Elko series points into
the central and western Basin from the east has been
postulated (Adovasio 1970).
There, are three specimens which compare favorably
with the Pinto Shoulderless (Holmer 1978) (Figure 16b, c).
Holmer (1978) places the age range of the Pinto series at
approximately 8300 to 6200 B.P. at the four Archaic sites
used in his study. Joes Valley Alcove in the northern
Colorado Plateau provides a time span of approximately
8300 to 6300 B.P. for the Joes Valley Tanged points which
are identical to the Pinto series points (Holmer 1978).
These points date more recently at other sites, however,
as they are common between 6500 and 3800 B.P. at O'Malley
Shelter (Fowler et al. 1973) and from approximately 5000
to 3000 B.P. at Swallow Shelter (Dailey 1976). Pinto
points are also reported in western Colorado on the
Uncompahgre Plateau (Wormington and Lister 1956:14)
although no dates are available.
Three points compare favorably with the Gypsum
point (Holmer 1978) (Figure 16d, e). These points have
already been discussed as similar to several types Buckles
distinguished on the Uncompahgre Plateau (Buckles 1971).
The Gypsum projectile point is the most recent
type generally associated with the Archaic stage of the
northern Colorado Plateau and the southern portions of the
Great Basin. The Gypsum point dates from approximately


65
4600 to 1500 B.P. at Sudden Shelter and Cowboy Cave.
O'Malley Shelter (Fowler et al. 1973:42) contained 105
specimens, the most ever found at a single site, and they
date from approximately 5000 to after 1000 B.P. This time
range estimate is supported by radiocarbon dates from
Gypsum Cave (Heizer and Berger 1970:17; Shutler 1967:306)
that indicate a range from approximately 3000 to 2000 B.P.
Gypsum points are unique among Archaic projectile
points and may be of interest in understanding the
development of hafting technologies. It was observed at
Cowboy Cave that most Gypsum points and many later arrow
points have the remnants of pitch on their basal stems
(Jennings 1980). No earlier point types show any trace of
pitch. The discovery of pitch as an adhesive may have
reduced the need for the deep side or corner notches
characteristic of earlier dart points, although deep
notching probably remained a necessity for knife blades
because of the cantilever forces produced during cutting.
The Elko series may represent the type of blades retained
as knives.
Two specimens are similar to the Desert Side-
Notched point, and have already been discussed as being
similar to types Buckles delineated for the Uncompahgre
Plateau (Buckles 1971) (Figure 16f, g). They have been
recovered from excavated sites near the northern and
western periphery of the Fremont area although they have
been reported in uncontrolled situations throughout the


66
Fremont area (Berry and Berry 1976). Of the excavated
Fremont sites they have never been the dominant type,
making up only 12% of the total points recovered (Holmer
and Weder 1980). Of significance is that most of those
sites contain Shoshoni ceramics although they constitute
only 3% of the total ceramic collection (Holmer and Weder
1980). The correlation has been inferred by Fowler et al.
(1973) at O'Malley Shelter, and by Frison (1971) at the
Eden-Farson Site in Wyoming. The conclusion is that the
occurrence of Desert Side-notched points does not result
from Fremont occupations but indicates post-Fremont
Shoshoni use of the area after approximately A.D. 1150
(Holmer and Weder 1980). This conclusion is supported by
the presence of identical point'types in the northern
Plains (Plains Side-notched) after approximately A.D. 1590
(Kehoe 1966).
There were two specimens which compare favorably
with the Nawthis Side-Notched points (Holmer and Weder
1980) (Figure 16h, i). Their distribution is limited to
the southern half of the Fremont region dating from
approximately A.D. 950 to 1250. They are similar, if not
identical, to points associated with Pueblo II occupations
south of the Colorado River. There is no apparent
associations with any single ceramic type although most
sites also produce small quantities of both Kayenta and
Virgin ceramics.


67
One specimen compared favorably with that of the
Uinta Side-Notched points (Holmer and Weder 1980) (Figure
16j). They are widely distributed over the northern half
of the Fremont region dating from approximately A.D. 800
to 1200. Sites where they are the dominant type usually
contain a large percentage of Uinta Gray ceramics. Their
similarity to the Prairie Side-notched points recovered in
the northern Great Plains dating from approximately A.D.
700 to 1300 (Kehoe 1966) supports conjectures of a Great
Plains influence in the northern Fremont areas (Aikens
1966).
There were two specimens which compared favorably
with the Parowan Basal-Notched, both of which have been
discussed previously as being comparable to types
delineated by Buckles for the Uncompahgre Plateau (Buckles
1971) (Figure 16k, 1). The spatial distribution of
Parowan points can be divided into two groups based on
ceramic associations (Homer and Weder 1980). The first
area includes the Virgin River, Santa Clara River, and
Johnson Canyon. The sites in the area have a high
percentage of Virgin ceramics (98%) and very small
percentages of Kayenta, Mesa Verde or Sevier ceramics
(2%). The temporal span clusters between approximately
A.D. 900 and 1200. Parowan points constitute 63% of the
total arrow points from sites in this area.
The second area includes the Parowan Valley, part
of the Sevier River Drainage, and part of Southeastern


68
Nevada. The ceramics from the sites in this area are
predominantly Sevier ceramics. The temporal span clusters
around A.D. 950 to 1150. Parowan points constitute 55% of
the total arrow points recovered.
Parowan points have been noted at other sites at
low frequencies as far east as the Bull Creek drainage.
Not surprisingly, the two pithouses at Bull Creek from
which Parowan Points were recovered were the ones which
had significant quantities of Virgin Kayenta ceramics.
Parowan points are the predominant points in both the
Parowan and Virgin Kayenta cultural regions. They were
used ca. A.D. 950-1150.
Ceramics
There were 32 ceramic fragments present in the
collection; one was a rimsherd and the remaining 31 were
body sherds. Ceramics were only recovered from Luster
Cave. While all the specimens appear to have been
constructed using an obliterated coiling method and all
specimens appear to have some degree of corrugation
present on the outer surface, it appears that there may be
three types of ceramics present, based on core .thickness,
temper, clay, degree of surface manipulation and color.
The first type appears to have been constructed from a
micaceous clay, utilizing small quartz fragments as temper
(Figure 17). Firing was partial oxidation, with the color
a light gray on the surface, and dark gray in the


69
Figure 17. Sangre de Cristo Micaceous Wear.
interior, with a large amount of soot on the outer
surface. There was no slip evident on any of the
specimens. The thickness ranges from .5cm to .6cm. It is
difficult to determine shapes and sizes of vessels without
more rimsherds, but the rimsherd present is of this type,
and the profile indicates that it probably was from a
simple bowl (Figure 18). There appears to have been a
slight corrugation of the exterior surface, but this may
also be due to an only partial obliteration of the coils
on the exterior surface, as well as a heavy layer of soot
which nearly covered the entire outer surface. The
texture of this type was fine-to-medium. There were 25
specimens representing this type, the distributions across
the site are recorded in Table 5.



t
Figure 18. Rimsherd Profile.
TABLE 5
DISTRIBUTION OF CERAMIC TYPES IN LUSTER CAVE
Area B C F
Level 2 3 5 1 4 Surface
Type 1 Type 2 2 5 8 1 3 9
Type 3 1 2
It has been suggested that micaceous ceramics
similar to this first type are variable and have
widespread distributions in context which include
occupations by Utes, diverse Athabascans, Puebloans,
Hispanics and others (Buckles 1988:221; Baugh and Eddy
1987). Baugh and Eddy recommend that such micaceous
ceramics should not be identified with ethnic-specific
classifications, but should be classified as Sangre de
Cristo Micaceous Wear.
The temper of the second and third types was
crushed sandstone. The color was generally a brownish


71
gray to gray, indicating a reducing atmosphere. There was
no slip evident on either the second or third types. The
thickness of the sherds averaged 0.7 cm. to 0.8 cm.,
although there were two specimens with thicknesses of 1.0
cm. The second type indicated a clear corrugation, and
there was little of the sooting on the outer surface that
characterized the first type (Figure 19). Because all of
the sherds from this type were body sherds, it was
difficult to determine the shapes the sherds may have
represented. There were 3 specimens representing this
type.
The third type was very similar to the second
type, except the surface manipulation was comprised of
fingernail impressions, rather than corrugation (Figure
20). One of the specimens had a hole drilled through it,
possibly indicating repair or a way to carry it. The
texture of this type, as well as the second type, was
medium to coarse. There were 3 specimens of this type.
All the types represented are from the same area
of the cave, although separated arbitrarily by Luster's
sections, so it is difficult to separate them temporally,
as well as stratigraphically, from one another. Both
Types 2 and 3 compare favorably with Buckles' Uncompahgre
Brownware (1971:507-522), while Type 1 is considered
similar to the Sangre de Cristo Micaceous Wear.


I o 1 w
Figure 19. Uncompahgre Brownware., corrugated.
Figure 20. Uncompahgre Brownware, fingernail
impressed.


73
Buckles gives a broad range of temporality to
Uncompahgre Brownware, suggesting it was produced
approximately 400-500 years B.P., or with the advent of
the Ute in the area (Buckles 1971:552). The temporality
of the Sangre de Cristo Micaceous Wear can only be guessed
at, since its distribution is across many ethnic and
temporal borders (Baugh and Eddy 1987).
Perishables
There were 171 specimens of perishable materials
present in the collections recovered from Luster Cave.
There is a wide variety of different items considered to
be "perishables". The distribution of the items is
recorded in Table 6.
There were four specimens of basketry present in
the collections, all small fragments (Figure 21). Two of
the specimens were merely a split rod wrapped with a
series of yucca leaves (Figure 21a, b). These specimens
are similar to several examples recovered from Danger Cave
(Jennings 1980:72, 74). The other two specimens are very
small fragments of basketry, one of which had been treated
with some type of preservative (Figure 21c); the other is
merely a segment of a 1/2 rod and bundle held together
with one section of noninterlocking stitch (Figure 21d).
The specimen treated with preservative is a
one-rod-and-bundle foundation held together with


TABLE 6
DISTRIBUTION OF PERISHABLE ITEMS IN LUSTER CAVE
74
Area A B C D E F
Level Sur. 12341234512341112345
Basketry
Quids
Leather:
Modified
Unmod.
Thong
Bundles
Ycca Lvs
Ycca Knts
Cordage:
Yucca
Type 1
Type 2
Type 3
Type 4
Fur
Type 1
Type 2
Type 3
Sinew
Human Hair
Juniper
Unmod:
Hair
Fur
Reed
Ycca Bs.
Atlatl Shft
Arrow Shaft
Gmng Pieces
Shft/Tnon
Wood. Waste
Fbr-wrp. St.
Dcoratd. Rd.
Misc. Wood
5 2
111
2
2
13 2
1 1
1 2
2
1 1
4 1
1 8 2 2 1
1
3
1
1111
4
1
1
1 5
1
1 1
1
1
2 7
2 3 2 5 4
1 1
1
112
1
2
2
1
2
1
1 1
2
2 2
2
2 1
1
4
1
1
1 1
1
2


75

Figure 21. Basketry. Split rod wrapped with yucca
leaves (a, b), basketry treated with
preservative (c) and basketry fragment (d).
non-interlocking stitches. The dimensions of this
specimen are 3.15 cm. x 1.25 cm. x .61 cm.
Adovasio (1971) has traced the differences
between textiles, in particular basketry and cordage, of
the Great Basin and Southwest. Given his descriptions of
the variations between the two areas, the basketry from
Luster Cave is most similar to that of the Great Basin,
where one-rod-and-bundle foundation coiling appears to
have been the standard.
There were 26 yucca quids present in the
collections (Figure 22). There was a note attached to one
of the quids dated May 26, 1956. It said: "Largest quid
Yucca so. One quid evidently has some Aaave fibers in it.
Needs to be checked with A. utahensis. In natural range?
Vorsila L. Bohrer."


76
Figure 22. Yucca quids.
There were 9 specimens of leather present.
Included within this category was one piece of modified
soft hide, six pieces of unmodified soft hide, and two
pieces of thong (Figure 23). The modified soft-hide
category was limited to a specimen of softened skin that
appeared to have several perforations in it for sewing
(Figure 23a). This piece measured 7 cm. by 3 cm, and was
recovered from Area F, Level 3.
The unmodified soft-hide category included scraps
that were probably waste pieces from the manufacture or
repair of soft-hide articles (Figure 23b). This included
six scraps, with sizes ranging from 5.8 cm. to 1.8 cm. in
length, and 5.5 cm. to 1.1 cm. in width.
Two soft-hide thongs were recovered, the length of
the specimens measured 9.3 and 4.8 cm (Figure 23c, d).


77
Figure 23. Modified leather (a), unmodified leather
(b) and leather thong (c, d).
There were seven specimens of unmodified fur/hair.
Included within this category were three specimens of
human hair, all of which were brownish-black in color and
fairly coarse in texture. The remaining specimens
represented lagomorph fur from either the white-tailed
jackrabbit (Lepus townsendii) or snowshoe rabbit (Lepus
americanus).
There was one specimen of an unmodified yucca
plant base from which leaves have been cut off.
There were two bundles present in the collection.
The first specimen was entirely of shredded yucca; the
second specimen was of shredded yucca, yucca leaves,
leather and two Cvmopterus umbelliferae roots (Figure 24).


78
Figure 24. Bundle of shredded yucca leaves, leather
and umbelliferae root.
Umbelliferae roots are noted to have been used for food,
seasoning and medicine (Colton 1974:305; French
1971:385-412; Harrington 1967:171-173; Whiting 1939:86).
This bundle may have been used in a healing ritual, or it
may have been used during travel.
There were 66 lengths of cordage made from plant
fibers recovered from Luster Cave. Analysis of the
collection included manufacturing techniques, material
employed and knot-types utilized. There were no artifacts
made of cordage recovered. The majority of cordage
recovered were small scraps which show the effects of hard
and continual use. Some of the specimens are burned, worn
to the point of unraveling, have broken fibers or are even
worn through. Many of the specimens are knotted together,
indicating a frugal effort to save cordage.


79
In a description of cordage, the term olv refers
to a single yarn which is usually plied with another
single yarn to become a two-ply cord or yarn. The
direction of twist is determined as follows: "If the
elements are twisted in one direction so that the slope of
the spirals, when held in a vertical position, conforms to
the central portion of the letter S, the cord is said to
have an S-twist. If the elements are twisted in the
opposite direction, the cord has a Z-twist." (Rohn
1971:114). Two-ply cords are by far the most common
within this particular collection.
A tabulation the cordage specimens from Luster
Cave according to direction of twist shows 57 pieces of
Z-twist cordage, as opposed to eight pieces of S-twist
cordage. Both twist types are present throughout the
fill of the cave.
There are several methods of spinning yarn with a
spindle; one of the most common is to roll it along the
thigh. Ruth Underhill (1944:36) has shown that twist
direction is dependent upon the direction the spindle is
rolled. If it is rolled away from the body, an S-twist
cord results; if the spindle is rolled toward the body,
the cord will be Z-twisted. This would mean that in order
to make a two-ply Z-twist cord, the first yarn would be
rolled away from the body to get the S-twist, then the ply
twist would be achieved by rolling the yarns toward the


80
body. Spinning along the leg can also be achieved without
the aid of a spindle.
Another method of using a spindle is to drop the
spindle and let it spin freely just above the ground.
Here, as before, the twist direction is determined by the
direction the spinner twists the spindle as it is dropped.
It is impossible to say what method was employed at Luster
Cave, although the absence of any spindle whorls may
indicate employment of the first method discussed.
Of the 66 total cordage specimens, 63 are two-
ply, obviously the most popular manufacturing technique.
The remaining three are single ply, which may represent
cordage which has come unplied.
Overhand and square knots appear to have been
utilized for a variety of purposes relating to the
cordage. Overhand knots appearing at the ends of cords
may have prevented the fibers from unraveling. Square
knots, as well as granny knots, were used to join two
pieces of cord together. It has been noted that the
square knot seems to predominate in the Southwest
(Basketmaker through Pueblo), whereas the sheetbend and
overhand knots are more common in the Great Basin
(Lambert and Ambler 1961:57).
Within the collection of cordage, it appears that
there are eleven types, based on material, number of
strands, and direction of ply.


81
There were 41 specimens of yucca cordage which
made up four types.
The first type was a single specimen made up of a
combination of a strand of yucca fibers and a strand of
sinew (Figure 25a). It was a Z-twist, with a length of
5.3 cm., and a thickness of 1.31 cm. There was just one
twist per centimeter, and the angle of the twist was 60
degrees.
The second type is made up of 34 specimens of
two-strand, Z-twist Yucca fibers (Figure 26). There were
four specimens with only two twists per centimeter, with
twist angles of 60 degrees (two specimens), 65 degrees and
75 degrees (one specimen each). There were eleven
specimens with three twists per centimeter, with twist
angles of 45 degrees (3 specimens), 60 degrees (six
specimens), 65 degrees and 75 degrees (one specimen each).
There were eight specimens with four twists per
centimeter, all with twist angles of 60 degrees. There
were five specimens with five twists per centimeter, with
twist angles of 45 degrees (one specimen), 60 degrees
(three specimens) and 75 degrees (one specimen). Lengths
range from 1.61 cm. to 33.5 cm.; thickness ranged from 0.1
to 0.61 cm.
The third type includes five specimens of two-
strand, S-twist Yucca fibers (Figure 26). Lengths range
from 4.4 cm. to 5.7 cm.; widths range from 0.15 to 0.16
cm. Three of the specimens had 4 twists per centimeter,


Figure 25. Yucca and sinew cordage (a) and sinew cordage
(b, c and d).
Figure 26. Yucca cordage. S-twist cordage (a) and Z-
twist cordage (bf c, d, e and f).


83
all with angle twists of 60 degrees. There was one
specimen each which had 3 and 5 twists per centimeter,
with angle twists of 45 and 60 degrees, respectively.
The fourth type is comprised of one specimen of a
single twist of yucca fibers. Its length is 14.4 cm. and
the thickness is 0.1 cm.
There are eleven specimens of fur cordage, of
which there are three types.
The first type is comprised of eight specimens of
two-strands with a Z-twist (Figure 27). There were either
2 or 3 twists per centimeter, .three being the most popular
with six specimens. The angle of the twist was either 30
degrees (three specimens), 45 degrees (four specimens) or
60 degrees (one specimen). The lengths ranged from 2.85
cm. to 22.0 cm.; thicknesses ranged from 0.27 cm. to 0.81
cm.
The second type was comprised of two specimens of
one strand cm.; thicknesses were 0.5 cm. and 0.27 cm.,
respectively.
The third type was comprised of a single specimen.
This specimen was a two-strand, S-twist cord. There were
two twists per centimeter, and a twist angle of 60
degrees. The length was 4.8 cm. and the thickness was 0.8
cm.
There were nine specimens of sinew which made up
two types. Both types were of two strands, eight of which
had a Z-twist and one which had an S-twist (Figure 25b, c


84
Figure 27. Fur cordage.
and d). Eight specimens had three twists per centimeter,
and one specimen (not the S-twist) had four. All
specimens had a twist angle of 60 degrees. Lengths ranged
from 2.85 cm to 16.5 cm; widths ranged from 0.15 cm. to
0.29 cm.

There was one specimen of cordage made of human
hair fiber (Figure 28). It was a loosely woven
two-strand, Z-twist cord, with two twists per centimeter.
The twist angle was 60 degrees. The length of the
specimen was 11.3 cm.; the thickness was 0.3 cm.
A final type of cordage was made of shredded
juniper, all the specimens were two-strands with a Z-
twist (Figure 29). Three of the specimens had one twist
per centimeter, with a twist angle of 60 degrees. A final
specimen had two twists per centimeter, with a twist angle


o i 2 Ben-
Figure 28. Human hair cordage.
Figure 29. Shredded juniper cordage.


86
of 75 degrees. Lengths ranged from 10.5 cm. to 22.5 cm.;
thicknesses ranged from 0.29 cm. to 0.49 cm.
There were 15 specimens of shredded yucca leaves
and 14 yucca leaves consisting of only a knot (Figure 30).
The purpose of these specimens is unknown, although the
shredded leaves may have been in a preparatory state for
cordage. The specimens which consist of only a knot may
be remnants of fiber which was used as cordage. Another
possibility may be that they were "doodles" and have no
significant use. Among the knot-types represented, there
were two Larks-head, seven square knots, three overhand,
one double and one figure-of-eight.


87
There were 26 specimens of wood present in the
collection of artifacts recovered from Luster Cave. The
artifacts were separated on the basis of apparent
function, although where use was not apparent, descriptive
categories were established. Ten types were represented
in the collection.
Two of the specimens from the collection were
determined to be atlatl dart shafts (Figure 31).
Diameters, were 0.72 cm. and 0.74 cm., and 10.6 cm. and
19.6 cm. in length, respectively. Atlatl shafts are
common in dry cave sites throughout the desert Southwest.
Four wooden foreshafts were judged to be parts of
composite arrows (Figure 32b, c; 33a). Two of the
specimens appear to be V-notched proximal sections. There
appears to be resin residue within the notch of both
specimens. Both specimens are 0.62 cm. in diameter; the
lengths are 3.9 and 4.83 cm. The other two specimens are
unidentifiable fragments, although it is most likely they
are sections of arrow shafts. The diameters of these
specimens are 0.62 cm. and 0.6 cm.; the lengths are 2.1
and 2.94 cm.
Wood parts of composite arrows are commonly found
in Southwestern sites where perishable artifacts are
preserved (Janetski 1980).
Two specimens from the collection were classified
as gaming pieces (Figure 34). Both specimens are
undecorated. Cuts on opposite ends of the piece can


Figure 31. Atlatl shafts.
Figure 32. Arrow foreshafts (b, c) and painted reed
fragment (a).


Full Text

PAGE 1

A REANALYSIS OF ARCHAEOLOGICAL MATERIALS FROM ROTH AND LUSTER CAVES by Elizabeth Kae Smith-McDonald B.S., Brigham Young University, 1983 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Department of Anthropology/Sociology 1989 td }

PAGE 2

This thesis for the Master of Arts degree by Elizabeth Kae Smith-McDonald has been approved for the Department of Anthropology/Sociology by .,.

PAGE 3

Smith-McDonald, Elizabeth Kae (M.A., Anthropology/ Sociology) A Reanalysis of Archaeological Materials from Roth and Luster Caves Thesis directed by Assistant Professor Craig R. Janes The research covered in this paper deals with the contribution previously curated materials add to solving current research problems in chronology, sub-sistence, settlement patterns and social organization and relationships. Artifacts recovered from the 1952 excavation of Roth and Luster caves in west-central Colorado were reanalyzed and a radiocarbon sample was tested to provide the data needed to answer the research problems. An Archaic and late Formative occupation are theorized for Luster Cave. Roth Cave provides evidence that it may have been occupied during the Archaic period. The for.m and content of this abstract are approved. I recommend its publication. Signed c iii

PAGE 4

To Emily, may you grow surrounded by love and happiness

PAGE 5

ACKNOWLEDGEMENTS I would like to thank Michael Piontkowski for all the help and advice he provided during the course of my research as well as providing the funds for the radiocarbon test. I would like to thank Diana Leonard and Jeannette Mobley-Tenaka of the Henderson Museum for preparing the collection and providing me with all the available records in the Museum's possession, as well as support during my research. I would like to thank Craig Janes, Lorna Moore, and Jim Grady for their help as members of my thesis committee. I would like to thank Kate Aasen-Rylander for her advice concerning my macrofossil analysis. I would also like to thank Jim Wilde for his comments and criticisms on the rough draft of my thesis. I would also like to thank my family for their support during this project. Last, but not least, I would like to thank my husband Ken for his support, encouragement and love throughout this entire project.

PAGE 6

CONTENTS TABLES. . . . . . . . . . . . . . . . . . . . . . . . X FIGURES ......................................... . . xi CHAPTER 1. INTRODUCTION. . . . . . . . . . . . . . . . . 1 2. THE ENVIRONMENTAL ANED CULTURE-HISTORICAL CONTEXT OF ROTH AND LUSTER CAVES............ 3 Environmental Context. . . . . . . . . . . . 3 Previous Research. in West-Central Colorado. . . . . . . . . . . . . . . . . . 11 Culture-Historical Overview .. 17 The Paleoindian Period.................... 18 The Archaic Period........................ 19 The Formative Period. . . . . . . . . . . 19 The Protohistoric Period.................. 23 3. RESEARCH DESIGN ........................... ... 25 Problem Domain I: Status of the Collection. . . . . . . . . . . . . . . . 2 5 Problem Domain II: Chronology............... 27 Problem Domain III: Settlement Patterns. . . . . . . . . . . . . . . . . 29 Problem Domain IV: Subsistence.............. 30 Problem Domain V: Social Relationships and Organization .... ...................... 32 4. MATERIAL CULTURE ANALYSIS METHODS............. 35 Li thics. . . . . . . . . . . . . . . . . . . 3 6 Ceram.ics.................................... 37

PAGE 7

5. 6. 7. 8. viii Peri shab.le s . . . . . . . . . . . . . . . . 3 8 Bone ................ 42 Macro/Microbotanical Remains................ 43 Human Bone . . . . . . . . . . . . . . . . . 4 5 Wood ..... 45 Minerals ..... .............................. 46 LUSTER CAVE ANALYSIS .... Li thics ............ Groundstone. Bifaces. Ceramics .... Perishables. 47 47 53 57 68 73 Worked Bone. . . . . . . . . . . . . . . . . 94 Macrofossils. . . . . . . . . . . . . . . 99 Human Bone ................................ Minerals ................................. Radiocarbon Date ........................... ROTH CAVE ANALYSIS . . . . . . . . . . . . . . Li thics .................................... .............................. Bifaces .................................. Perishables . . . . . . . . . . . . . . . . Worked Bone Macrofossils ............................... Human Bone DISCUSSION . SUMMARY AND CONCLUSIONS ...................... REFERENCES CITED .................................... 105 106 107 109 109 111 114 117 127 126 128 130 138 140

PAGE 8

ix APPENDIX A. ARTIFACT CATALOG. .............................. 15 0

PAGE 9

TABLES Table 1 Cordage Formulae . . . . . . . . . . . . . . . 41 2. Distribution of Technological Types in Luster Cave.................................. 48 3. Distributionof Groundstone in Luster Cave. . . . . . . . . . . . . . . . . . . . 54 4. Distribution of Projectile Point Types in Luster Cave.................................. 58 5. Distribution of Ceramic Types in Luster Cave ......................................... 70 6. Distribution of Perishable Items in Luster Cave. . . . . . . . . . . . . . . . . . 7 4 7. Distribution of Worked Bone in Luster Cave . . . . . . . . . . . . . 9 4 8. Macrofossils from Luster Cave .................. 100 9. Corn Cob Measurements from Luster cave ......... 102 10. Distribution of Minerals in Luster cave ........ 107 11. Distribution of Technological Types in Roth Cave. . . . . . . . . . . . . . . . . . . . . 110 12. Distribution of Groundstone in Roth cave ....... 112 13. Distribution of Projectile Point Types in Roth Cave ........................ . . . . . . 115 14. Distribution of Perishable Items in Roth Cave. . . . . . . . . . . . . . . . . . . . . 118 15. Macrofossils from Roth Cave .................... 125 16. of Artifacts in Luster cave. . . . . . . . . . . . . . . . . 132 17. Distribution of Diagnostic Artifacts in Roth Cave. . . . . . . . . . . . . . . . . . 132

PAGE 10

FIGURES Figure 1. Luster Cave location map....................... 4 2. Roth Cave location map. . . . . . . . . . . . . 5 3. The Uncompahgre Plateau Region................. 7 4. Excavation plan of Luster Cave................. 10 5. Excavation plan of Roth Cave................... 10 6. Frontal view of Luster Cave.................... 12 7. Location of sites on the Uncompahgre Plateau............... . . . . . . . . . . . 14 8. Basalt chopper. . . . . . . . . . . . . . . . . 49 9. Quartz hammerstone............................. 53 10. Sandstone arrowshaft smoother.................. 55 11. Rectangular mano. . . . . . . . . . . . . . . . 55 12. Loaf-shaped mano............................... 56 13 Round mane . ..... . . . . . . . . . . 56 14. Bifaces. . . . . . . . . . . . . . . . . . . . 58 15. Untyped projectile points...................... 59 16. Typed projectile points ................ ;....... 61 17. Sangre de Cristo Micaceous Wear................ 69 18. Rirnsherd profile............................... 70 19. Uncompahgre Brownware, corrugated.............. 72 20. Uncomphagre Brownware, fingernail imp res sed. . . . . . . . . . . . . . . . . . 7 2 21. Basketry ............. . . . . . . . . . . . . 7 5

PAGE 11

22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. Yucca ..... .............................. Modified leather .............. Bundle of shredded Yucca leaves ............... Yucca and sinew cordage. Yucca cordage. ................................. Fur cordage ... ............................ Human hair cordage. Shredded juniper cordage .... Yucca knots .................. Atlatl shafts .... .......... Arrow foreshafts and painted reed fragment ................... Arrow shaft and modified wood. Wooden gaming pieces. Shafts with tenons ............................ Woodworking waste. Fiber-wrapped twigs ........................... Bone awls ... Bone gaming piece and.bone tubes. Bone shell bead fragment, pendant. bone fishhook and Antler flakers. Macrofossils. Tooth crown fragment. Burial matting. Quartz hamm.erstone ............................ Loaf-shaped mano .............................. Round mano .... Rectangular mano. xii 76 77 78 82 82 84 85 85 86 88 88 89 89 91 92 93 95 97 98 98 100 105 106 111 112 113 113

PAGE 12

49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. Bifaces ... Projectile points. Basketry Basketry fragment. fragment. Worked wood. ................................. Yucca quids. ................................... Unmodified leather. Yucca Yucca Yucca Bone cordage ..... leaf bundle. knots ..... awl fragments. Macrofossils ....................... .......... Human Bone . . . . . . . . . . . . . . . . . xiii 115 116 118 119 120 120 121 122 123 124 125 126 129

PAGE 13

CHAPTER 1 INTRODUCTION In 1951, a joint archaeological excavation was established between the University of Colorado Museum and the Department of Anthropology to pursue research in the Glade Park area of west-central Colorado. Two caves and three arroyo sites were excavated during the summer with the results published in the March 1952 edition of Southwestern Lore, and subsequently in "Archaeological Investigations on the Uncompahgre Plateau in West-Central Colorado" by H. Marie Wormington and Robert H. Lister (1956). The materials from the sites were then curated at the Henderson Museum on the University of Colorado/Boulder campus, and essentially forgotten. This project addresses a research problem dealing with previously curated materials, that is, do these materials have any value in contributing to solutions to research problems archaeologists are addressing at this time? This project is a reanalysis of materials excavated from Roth and Luster Caves in west-central Colorado that have been curated in a museum more than 30 years ago in order to investigate the possibility that they could contribute pertinent information to an area that has been neglected for nearly that same length of time. The area

PAGE 14

2 originally excavated is also of significance as a possible transition zone between two Formative cultures, as well as having been occupied by both an Archaic and Post-Formative group. The two sites in question, Roth cave and Luster Cave, have been placed at various times within the context of several of these cultural units (Buckles 1971; Lister and Dick 1952; Pierson 1980; Wormington and Lister 1956). The main research problem for this project is to attempt to establish more clearly cultural affinity for both of the sites. Projectile point and ceramic typologies will be utilized in tandem with radiocarbon dating to solve this problem. Additional materials recovered from the caves will also be utilized in answering research questions pertaining to subsistence, settlement patterns and social relationships.

PAGE 15

CHAPTER 2 THE ENVIRONMENTAL AND CULTURE-HISTORICAL CONTEXT OF ROTH AND LUSTER CAVES Environmental Context Roth Cave is located in Mesa County, Colorado, while Luster Cave is located in Grand County, Utah, approximately 100-200 yards west of the Colorado/Utah state border. Both caves are in the Glade Park region on the western edge of the Uncompahgre Plateau at elevations of 6400 feet mean sea level (MSL) and 5400 feet MSL, respectively. Luster Cave sits approximately 400 yards west and 100 feet upslope from the Little Dolores River, with Roth cave located at the foot of a sandstone cliff about 500 yards north of the Little Dolores River, and approximately 100 feet above the river (Lister and Dick 1952; Wormington and Lister 1956). The approximate geographic location of Roth.Cave is T12S R103W Section 17 (U.S.G.S. Sieber canyon 1:24,000), with Luster cave at T20S R26E Section 32 (U.S.G.S. Westwater 1:24,000) (Figures 1 and 2) The Uncompahgre Plateau is considered part of the Canyon Lands Section ofthe Colorado Plateau and is

PAGE 16

Figure 1. Luster Cave location map. U.S.G.S. Westwater 1:24,000. Sec. 32, T20S, R26E, Grand County, Colorado. 4

PAGE 17

Figure 2. Roth cave 1:24,000. Colorado. location map. U.S.G.S. Sec. 17, T12S, R103W, .I 4 II u n n II II II 5 Sieber Canyon Mesa County,

PAGE 18

6 located mainly in west-central Colorado, with the northwest end extending about 25 miles into Utah (see Figure 3). The Plateau is about 115 miles long and 25 miles wide and forms the divide between the Uncompahgre-Gunnison and San Miguel-Dolores River systems. Throughout much of its length the Plateau is comparatively level, gently sloping towards the northeast from its highest points on the southwest side. This tableland is broken by numerous steep-sided narrow canyons which cross it from southwest to the northeast. Structurally the Uncompahgre Plateau is a fault block or uplift, and is mostly surfaced by sandstone and mudstone of the Dakota-Morrison formations and the Glen Canyon group. Elevations on the Plateau range from about 5,000 feet MSL on the northwest end to over 14,000 feet MSL to the southeast. Average yearly precipitation varies from about 8 inches at lower elevations to over 23 inches at higher elevations, with themajority of the Plateau receiving about 12 to 16 inches. A large percentage of the precipitation occurs in winter and early spring with January, February and March being the months of heaviest snowfall. Summer temperatures average 55 degrees at higher elevations and 75 degrees at lower elevations. On average, there are 140 days between the last frost of spring and the first frost of autumn at lower elevations, but at higher elevations the average is only about 50 days.

PAGE 19

0 10 20mlles UTAH I I I I I 1 Figure 3. The Uncompahgre Plateau Region. 7

PAGE 20

8 The most current climatic information closest to the sites is derived from the Little Dolores and Little Dolores 5 NE climate stations, located at 6700 and 6380 feet MSL, respectively. Unfortunately, the stations monitored only precipitation information. The temperature information is taken from the Colorado National Monument station, at an elevation of 5780 feet MSL, approximately three miles from Roth Cave. The mean annual temperature is 63.7 degrees Fahrenheit, while the average high and low temperature for January.is 34 degrees Fahrenheit and 13 degrees Fahrenheit. The average high and low temperature for July is 94 degrees Fahrenheit and 64 degrees Fahrenheit. Annual precipitation averages 13 inches. with 1.45 inches and 0.40 inches as the average maximum and minimum amounts occurring in January. The average maximum and minimum amounts occurring in July are 2.91 inches and 0.53 inches. Average annual snowfall is recorded as 78.1 inches from the Little Dolores climate station. This information was not available from the Little Dolores 5 NE climate station. Luster Cave is located in the Plains Lifezone (> 6,000 feet), and is characterized in this area by dry grasslands -mostly sage and yucca, occasional shrubs -typically pinyon, juniper and scrub oak with few trees, except those located along permanent water sources -the Little Dolores River in this case.

PAGE 21

9 Roth cave is located in the Foothills Lifezone (6,000 8,000 feet) which is characterized by extensive areas of grass and shrubs including pinyon, juniper, scrub oak, sage and yucca, in this area. The method of excavation followed a similar procedure in both caves (Lister and Dick 1952:71, 80; Wormington and Lister 1956:95-96, 106-107). To assist in horizontal control during excavation, and as part of the mapping process, points at five foot intervals were marked and numbered upon the cave walls. These marks were used as reference points for locating all objects encountered during the excavation (Lister and Dick 1952:71, 80; Wormington and Lister 1956:95-96, 106-107). Six areas in Luster and three areas in Roth, all identified.by letters, were excavated (Figures 4 and 5). Material from each area was removed in levels of 12 inches until the cave floor was encountered, allowing for the determination of vertical position of all artifacts found (Lister and Dick 1952:71, 80; Wormington and Lister 1956:95-96, 106-107). An attempt was made to locate all cultural material in place, but screening facilitated the location of smaller items. Approximately three-quarters of Luster Cave and one-third of Roth Cave were excavated (Lister and Dick 1952:71, 80; Wormington and Lister 1956:95-96, 106-107). A map was made of each cave before excavation started, and a photographic record was kept during the operation. Detailed notes were entered into field notebooks

PAGE 22

10 I I I I D ILevel \ I[]) 2Levels I mmJ 4Levels I m SLevels Grooved Sondstone Boulder I 0 5 IOfeet I I Figure 4. Excavation plan of Luster Cave (From Wormington and Lister 1956:95) .... -. ....... --CJ I Level liliill 3 Le ve Is 0 5 10 feet Figure 5. Excavation plan of Roth Cave (From Wormington and Lister 1956:107)

PAGE 23

11 throughout the work, and a field catalog was kept for all specimens recovered with all specimens being numbered so that they could be identified at a later date (Lister and Dick 1952:71). Unfortunately, the photographic record, as well as the field notebooks, could not be found in any files at the Henderson Museum during the course of this project. A recent survey of Luster Cave indicated that there is little left that could be reexcavated (Figure 6). The fill in the cave has been disturbed, not only the excavation in the 1950s, but it also appears as if there has been recent disturbance as well. Previous Research in West-Central Colorado Research on the Uncompahgre Plateau of west-central Colorado over the last fifty years has indicated that a variety of different cultures have occupied the area including the Archaic and Ute. A Formative culture may have also occupied the area, although further research is needed to establish possible affiliation with the Formative cultures bordering the area. The earliest reported work on the Uncompahgre Plateau was a survey of the Paradox Valley and adjacent areas in 1931 by the State Historical Society of Colorado and the Smithsonian Institution (Woodbury and Woodbury 1932). While most of the study was conducted south of the Uncompahgre Plateau in the Paradox Valley,

PAGE 24

Figure 6. Frontal view of Luster Cave, looking into the cave from the east. 12 ( t

PAGE 25

13 some survey work was done on the Plateau, and the Woodburys reported finding a pueblo about four miles west of Norwood along Naturita Creek. Harold Huscher (1939) conducted a survey along the northeastern slope of the Plateau and reported a number of preceramic sites which contained crude grinding stones, thin bifacially-flaked knives and projectile points. Huscher and Huscher (1943) also encountered circular stone structures which they argued were hogans of Athabascan (Navajo) origin. Many of these structures, particularly those located on the southwestern slope of the Uncompahgre Plateau, resemble Tabeguache and Cottonwood Pueblos and were associated with Black-on-White pottery, suggesting that they were not hogans at all but were possibly affiliated with a formative culture. A majority of the sites which have been excavated on the Plateau are rock shelters. C.T. Hurst from 1939 to 1947 worked op several of these sites. The first site, Tabeguache Cave I, was located about ten miles northeast of the town of Nucla and was considered by Hurst to be a Basketmaker II site (Figure 7). This cave contained slab-lined cists with some corn and square-toed sandals. Tabeguache Cave II was located about ten miles downstream from Cave I, and Hurst (1945) believed that it had been occupied at three different time periods. The earliest I occupation he attributed to a pre-Basketmaker people which are represented by the Tabeguache point, bifacial knives

PAGE 26

ROTH cAVE Glade I I I Monticello m,r;;------. ..1 UTAH COLORADO Park GRAND JUNCTION Moore and Casebier Sites \ .Montrose N.F. ,Figure 7. Location of sites on the Uncompahgre Plateau. 14

PAGE 27

15 and milling stones. The second occupation was by nomads who were contemporary with Basketmaker II peoples. Hurst (1945:8) that these nomads had obtained Basketmaker items by trade or conquest. This level contained some corn, corner-notched points, rod and bundle basketry and a slab-lined cist. Hurst attributed the upper level, which contained pottery and various stone artifacts, to the Utes. Hurst (1948) excavated another rock shelter, Cottonwood Cave, about nineteen miles east of Nucla, which he claimed was also a Basketmaker II site. Cottonwood Cave contained a bundle of corn, yucca fiber cordage, metates and square-toed sandals. The westernmost rock shelter excavated by Hurst (1947) was Dolores Cave, located along the Dolores River near the town of Uravan, and was occupied from Basketmaker II times to the early historic period. This cave contained a slab-lined fireplace, a few stone artifacts, and a rope-wrapped bundle. In addition to these rock shelters, Hurst also excavated several small masonry structures on the Uncompahgre Plateau. Tabeguache Pueblo was located 14 miles northwest of Nucla, and Cottonwood Pueblo was 16 miles east of Nucla. On the basis of pottery types, Hurst (1946, 1948) regarded these ruins as belonging to the Pueblo I and II time periods. During 1948 Hurst excavated another structure in the same settlement as

PAGE 28

Cottonwood Pueblo, but he died that winter and no information about this structure was ever published. 16 H.M. Wormington (Wormington and Lister 1956) worked on the northeastern slope of the Plateau where she excavated the Moore, Casebier and Taylor sites. Robert Lister excavated the Alva Site near the Taylor Site. All of these sites were rock shelters which Wormington and Lister felt represented a single complex which they named the Uncompahgre Complex. This precerarnic and prehorticultural complex is believed to have begun sometime during the first or second millennia B.C. Wormington and Lister (1956:81) describe this complex as being characterized by specialized cutting and scraping tools which they called Uncompahgre Scrapers; a great variety of projectile point types; occupation of caves or rock shelters; bifacially-flaked knives; thin, flat milling stones; and various bone implements. Wormington and Lister (1956:81) believed that 'it was probable that this complex was widely distributed on the Uncompahgre Plateau, and attributed the sites reported by Huscher (1939) on the northeastern slope as belonging to this complex. They also placed Hurst's Dolores Cave and the lowest level of Tabeguache cave II in this complex. Buckles (1971) excavated several Uncompahgre Complex rock shelters on the northeastern slope of the Plateau, and he suggested that this Archaic pattern continued on the Plateau until historic times. Buckles (1971:1170) also

PAGE 29

proposed that during part of this time there was a coexistence of the hunters and gatherers of the Uncompahgre Complex with horticultural groups. 17 Lister (Lister and Dick 1952, Wormington and Lister 1956) also worked in the Glade Park area on the northwestern edge of the Uncompahgre Plateau. The sites excavated include Luster Cave, Roth Cave, the Arroyo Sites and the Little Park Caves. The Arroyo sites contained projectile points, knives, milling stones, a clay figurine, corn cobs, and a Fremont-like petroglyph showing human figures with headdresses. Roth Cave had projectile points, metates, basketry but lacked pottery. Luster Cave contained corn, various stone artifacts, basketry, cordage, and Ute-like pottery. Following the work of Wormington and Lister there was little archaeological research completed on the Plateau until Metropolitan State College of Denver held field schools in the area during 1974-77. Unfortunately, there has been little published dealing with these excavations. Crane (1977) completed macrobotanical research on the Weimer Ranch sites, excavated by C.T. Hurst in 1947-48, of which only one had been reported (Hurst 1948). Culture-Historical Overview The cave sites are within the potential geographic range of several aboriginal cultures: the Paleoindian,

PAGE 30

18 Archaic, Anasazi, Fremont and Nurnic (Ute). Researchers have utilized a general culture-temporal framework as a heuristic device for organizing these synchronic cultural manifestations into a diachronic evolutionary framework: the Paleoindian period (ca. 10,000 B.P. to 8,000 B.P.), the Archaic period (ca. 8,000 B.P. to ca. A.D. 500), the Formative period (ca. A.D. 500 to A.D. 1150/1200), and the Protohistoric period (ca. A.D. 1300/1400 to ca. A.D. 1776). The historic period began with the documented travels of the Dominguez-Escalante Expedition. Subsequent decades saw limited use of west-central Colorado by EuroArnericans; utilization became intense only after the gold rush to western Colorado in the 1850s, and the removal of the Utes in early 1880s (Reed 1984). The Paleoindian Period Peoples representative of the Paleoindian period appear to be the first to inhabit the region. Evidence of this big-game hunting adaptation is presently found in the form of projectile points occurring as isolated artifacts or on sites with later cultural materials. Sites in which Paleoindian materials have been recovered include a deep horizon at Christmas Rockshelter, located near Montrose, from which Buckles (1968, 1971) recovered a base of a Paleoindian point (Reed 1984) There has not been an absolute date associated with the point, but since nothing associated with it is younger, Buckles (1968, 1971) feels it is of Paleoindian origin. several Folsom and Plano

PAGE 31

19 points have also been collected as surface finds near Montrose, on the Uncompahgre Plateau, along the Gunnison River along the base of Grand Mesa (Piontkowski, personal communication 1989), in Dinosaur National Park and in Tabeguache Canyon (Cassells 1983). The Archaic Period As te.rminal Pleistocene environmental conditions were supplanted by those mor.e similar to today' s, the big-game hunting adaptation of the Paleoindians was replaced by one emphasizing plant collection and processing and the hunting of a wider variety of smaller fauna. Human populations evidently grew, as sites attributed to the Archaic Stage are quite numerous. These sites outnumber those affiliated with other cultural units in west-central Colorado. The sites that have been investigated include the Taylor, Alva, Casebier and Moore Rockshelters (Wormington and Lister 1956) Levels of Christmas Rockshelter have also been placed within this cultural context (Buckles 1968, 1971). C.T. Hurst also excavated the Tabeguache Cave sites which have been placed within this cultural context (Hurst 1940, 1943, 1944, 19451 1946) o The Formative Period Following A.D. 1, there occurred an important shift in the economic adaptations of prehistoric peoples in the northern Colorado Plateau. Cultigens became an

PAGE 32

20 important source of food, and ceramics and substantial habitation structures appeared. Traditions such as the Anasazi in southwestern Colorado and the Fremont of Utah and northwestern Colorado flourished in many areas. In west-central Colorado, however, there is rather limited evidence of these formative cultures. The degree to which the prehistoric peoples of west-central Colorado conformed to a Formative Stage lifeway is presently not well understood. Present indications are, however, that the transition from an Archaic lifeway to a Formative lifeway may not have been as complete as compared to other contemporary groups of the northern Colorado Plateau. Recent archaeological investigations in westcentral Colorado have produced no firm evidence of Basketmaker III manifestations. Further, this research has not supported the Anasazi cultural affiliation of Formative Stage sites, as once posited for the area (e.g., Schroeder 1964). Whereas the general adaptive strategies represented such sites as Cottonwood, Tabeguache I and II and Dolores Caves may have been similar to the Basketmaker II culture, as no Basketmaker II surface structures such as those found in the Durango area (see Morris and Burgh 1954; Reed and Kainer 1978) have been discovered. The later stone structures lack key architectural features such as kivas and high walls, and have far too few ceramics to represent a typical Anasazi site. Certain artifact types, such as manos, vary

PAGE 33

considerably from contemporaneous Anasazi sites. In short, architectural and artifactual variation seems too great to support the presence of a bona fide Anasazi occupation of the project area. 21 If the Formative Stage sites in west-central Colorado are not considered Anasazi, then it follows that they be compared and contrasted to the other recognized Formative Stage culture in the area, the Fremont. The San Rafael variant is the Fremont variant closest to the project area. The Uinta variant may actually extend southwards to the vicinity of Cisco, Utah, to include the Turner-Look site, which yields a large number of calcite tempered Uinta Gray ceramics (Wormington 1955). Currently, there are several sites that are associated with the Fremont culture on the northwestern Uncompahgre Plateau, including two sites in Sieber Canyon, Roth Cave and Luster Cave (Piontkowski, personal communication 1989). Presently, there is some question as to whether the association is valid for Roth and Luster Caves. If the differences between the Formative Stage sites in west-central Colorado and the Anasazi or Fremont cultures are too great to support notions that they represent regional variants or subcultures of the Fremont or Anasazi cultures, then perhaps the least radical alternative would be to .suggest that these sites represent an in situ development from an Archaic technocomplex (Reed

PAGE 34

22 1984:39). In this scheme, people practicing an Archaic tradition lifestyle adopted a Formative Stage lifestyle as the need to intensify food production arose. Being relatively close to Anasazi and Fremont culture areas, they were able to trade certain items, such as ceramics, and were open to influence for such things as architectural styles. The importance of cultivated foodstuffs relative to collected wild foods. may not have been similar to either the Fremont or the Anasazi; perhaps hunting and gathering techniques were still .able to meet most of the economic needs. It is proposed that these similarities can be explained primarily by two factors. First of all, no culture develops and flourishes in complete isolation. Therefore, diffusion can be said to account, in part, for the ceramic and architectural similarities The other explanation, however, has an ecological basis. For instance, the type of building materials available and the nature of the soil deposit can influence what type of dwelling is built. The type of temper and clay that people use to make their pottery is also influenced by what is available. More important, however, is the extent to which the chosen subsistence technique effects the settlement pattern and sociopolitical organization (Crane 1977). The gross environment of the Uncompahgre Plateau does differ from that of eastern Utah. However, archaeological evidence suggests that the people in both areas had a similar

PAGE 35

23 ecological adaptation in that although they grew some domesticates, they were more heavily dependent upon hunting and gathering (Crane 1977; Reed 1984). It is hypothesized, therefore, that common factors in economy and technology can account for similarities in settlement pattern and the inferred sociopolitical organization. The Protohistoric Period Following the disappearance of the Formative lifeway from west-central Colorado, human adaptation evidently reverted to that similar to the Archaic Stage. Some archaeologists suggest that the term "Post-Formative Archaic Stage" might in fact be a more appropriate term for this stage (Reed 1984). Hunting and gathering was once again the primary mode of subsistence, and a nomadic lifestyle was practiced. The material culture was quite similar to that of the Archaic Stage, although the bow and arrow had become popular. Most of the data obtained so far for this stage concerns the Ute Tradition at the end of this period. The Ute evidently entered the region between A.D. 1200 and 1400, based on linguistic and archaeological evidence (Buckles 1971; Reed 1984). The Ute were expelled from west-central Colorado by 1881 (Reed 1984). The period prior to the apparent immigration of the Ute is poorly understood in west-central Colorado. Only one site definitely dating to this period has been identified in the region. Buckles obtained a radiocarbon

PAGE 36

date calibrated between A.D. 1335 and 1435 from a lithic scatter in the Ridgeway Reservoir project area (Reed 1984:42). 24

PAGE 37

CHAPTER 3 RESEARCH DESIGN This project was undertaken to examine what potentially new information could be gleaned from museum-curated artifacts of excavations that took place over 35 years ago. It addressed a variety of research questions which have been at the forefront of Formative-culture research for the past ten years including chronology, settlement patterns, subsistence and social relationships and organization. An additional research question concerned the status of the collection. Each of the above research questions.will be addressed below, defined as they apply to the caves specifically, with appropriate research topics and.q\lestions following within each section. Problem Domain I: Status of the Collection As archaeologists have become aware of the importance of a variety of scientific methods that are now utilized to complement the investigations of archaeological features, attempts have been made, where appropriate, to leave portions of a site unexcavated in order to "save" it for future generations and improvements in methodology. The problem that prompted this project

PAGE 38

was how much additional information can be gleaned from museum-curated objects before reexcavation is even considered. Research Topic A: Condition of the collection. Question 1: Are the specimens clean or dirty? Data Requirements: Inspection of the collection. Question 2: Have the specimens been treated for long-term preservation? Data Requirements: Inspection of the collection. Question 3: Are the specimens all present? Data Requirements: Inspection of the collection, comparison with acquisition ledger and published literature. Question 4: Were excavation notes and pertinent information curated with the collection, in the same repository, and is it in legible form? Data Requirements: Perusal of records and photos. Research Topic B: What resources were conserved that might be utilized for extraneous tests, and will the tests be accurate? 26

PAGE 39

Question 1: What is needed to provide for accurate C14 tests? Data Requirements: Material appropriate for radiocarbon testing, preferably charcoal. Question 2: What is needed to provide for accurate macro/microbotanical tests, and are these possible with the materials available from the collection? Data Requirements:. Soil samples which have been kept sterile from modern contaminants. Question 3: What is needed to provide accurate information from obsidian testing, and are these tests possible with the material available from the collection? Data Requirements: Quantities of obsidian large enough to test. Problem pomain II: Chronology 27 The time period for which Roth and Luster Cave were occupied was not absolutely established, but fixed within a range in which it was possible that they were occupied, through comparative typological analysis and stratigraphy (Lister and Dick 1956) Luster Cave was suggested to have been occupied within the range of 950 1300 A.D., wh;lle Roth Cave was suggested to have been occupied from about 500 850 A.D. Unfortunately, C14 samples which were taken at the time of excavation were never processed, leavin9 the two caves in a somewhat

PAGE 40

28 ambiguous chronological Over the years, the caves have been claimed to have had occupants of Numic/Ute origin (Pierson 1980), or Formative origin (PII/Anasazi or Fremont) (Lister and Dick 1956; Wormington and Lister 1956; Reed 1984). A well-established chronology would be beneficial because the two caves are located in an area of transition, a boundary of both Fremont and Anasazi occupation. The later Numic/Ute, however, occupied most of the surrounding area. Comparison with current projectile point, ceramic and basketry typologies, as well as reliance on C14 dating will establish a more definite temporal framework for the caves. Research Topic A: The temporal span of Luster and Roth Caves. Question 1: Can Roth and Luster Cave be placed in an absolute-dating time frame? Data Requirements: Utilization of C14 and dendrochronological dating. Question 2: Utilizing known projectile point, ceramic and basketry typologies, can Roth and Luster Caves be placed within a relative-dating time frame? Data Requirements: Examination of collection for comparable diagnostic specimens, comparison of these with known, reliable typologies for Archaic, Formative (Anasazi and Fremont) and Numic cultures.

PAGE 41

29 Problem Domain III: Settlement Patterns Settlement and subsistence patterns are quite complex, being intimately related to social organization, technology and other systems comprising culture. They refer to the manner in which economically important resources are procured, whether through seasonal wanderings from maturing resource to maturing resource, or through sedentary villages, from which procurement forays emanate. The study of settlement and subsistence patterns provides important data that permit the development of cultural ecological models of human adaptation in the region. These models are in turn important in the study of culture process (Reed 1984) Although it is beyond the scope of this project to plan, carry out and analyze an intensive survey of the area around Luster and Roth Caves, it is reasonable to assume that the caves were utilized for a specific purpose, and it is possible to extrapolate, within a reasonable margin of error, what this purpose was. It is also possible to compare these caves within a context of sites that have been located, and possibly investigated, in order to gain at least a peremptory view of the northwestern Uncompahgre Plateau as it was utilized in prehistoric times, as opposed to historic times.

PAGE 42

Research Topic A: Question 1: Seasonal Occupation At what time of year were Roth and Luster Caves occupied? Data Requirements: Seasonality data in the form of plant and animal remains recovered 30 from the caves. Analysis relating to this question must consider the artifact array as representing various activities. The greater the array of discrete activities, or a variety of discrete tasks, the stronger the argument for use of the site over several seasons. Question 2: Assuming compatible locations in the ecozone, were similar activities being carried out at Roth as opposed to Luster? Data Requirements: Items of material culture recovered from controlled contexts which, when considered as clusters, represent various activities that were carried out at or from the sites. Question 3: Are there comparable sites in the area? Data Requirements: Locational analysis to examine site location, use, and temporality. Problem Domain IV: Subsistence Given the fact that maize was recovered from.both caves, as well as additional macrobotanical remains, it seems reasonable to assume that subsistence can be

PAGE 43

31 addressed to a certain extent. The question that remains to be addressed, however, is how dependent were they upon cultivated foods and was the degree of dependence reflected in the length of.time the caves were occupied? Although the caves appear to be in reasonably good locations for year-round occupation, the amount of cultural material recovered and the relatively shallow occupational depth in the caves would suggest that they were occupied sporadically. Macro/microbotanical evidence, as well as faunal material would lend itself to further answering this question. Research Topic A: Use of domesticated :plants. Question 1: Did the inhabitants of the caves plant and harvest the corn recovered during excavation, or did they obtain it through exchange of goods? Data Requirements: Environmental data for compatible growing seasons; examination of the corn itself for comparison with other types located in the area. Question 2: Were the inhabitants of the caves involved in irrigation practices? Data Requirements: Investigation of irrigable flatlands adjacent to the habitation sites. Research Topic B: Wild Resource Exploitation.

PAGE 44

Question 1: What wild plant and animal foods were the inhabitants of the caves collecting? Data Requirements: Analysis of bone scrap, macrofossils and pollen samples. Question 2: What is the geographical distribution of these plants and animals? Data Requirements: Reconstruction and understanding of where particular resources are available, with consideration of seasonal variation. Problem Domain v : Social Relationships and Organization The study of the social organization of the prehistoric cultures which occupied the Uncompahgre 32 Plateau in west-central Colorado is generally limited to analysis of political organization and residential social groupings, due to the simple nature of their social organization (Reed 1984). The collection will provide enough information to identify possible components of the social and community organization present during the occupation of the caves. It is also within the scope of this research domain to ascertain the extent to which the two sites may represent one of the two main Formative cultures known in the region, namely the Anasazi or the Fremont. It may also be possible to ascertain whether the materials recovered from the caves represent a culture which has developed in situ from an Archaic technocomplex.

PAGE 45

By identifying the culture which occupied the caves, it may then be possible to trace any interaction which occurred with an outside group. Research Topic A: Social and Political Organization Question 1: What type of social and political organization was characteristic of the historic and proto-historic inhabitants of the Uncompahgre Plateau? Data Requirements: Perusal of historic and ethnographic literature for insights into the social and political organization of the Utes. Research Topic B: Time and Space Relationships Question 1: Are the artifacts recovered comparable to each other in time and space? Data Requirements: Typological comparison of artifacts and, if possible, an explanation of time and spatial ranges for those artifacts. Question 2: Are there associated artifacts that are dissimilar to the majority of the collection with regards to time and space? If so, can these artifacts be traced to their points of possible origin? Data Requirements: Typological comparison of the artifacts. Question 3: Can inferences be made from the analysis of the artifacts as to cultural 33

PAGE 46

associations? Data Requirements: Comparison of artifacts to find possible points of origin. 34

PAGE 47

CHAPTER 4 MATERIAL CULTURE ANALYSIS METHODS Analysis of the material culture recovered from the caves is a vital component of the research, without which nothing could have been accomplished. Thus, the following subsections will address the analysis methods used for each specific type of artifact mentioned in the original reports, and which are accounted for within the acquisition records of the While it would be desirable to analyze the specimens with respect to the wide range of possible analysis methods available at this point in time, due to the small number of specimens available within a specific material type/database, a large majority of the reanalysis will not be able to address certain statistical analyses. Because there has been a noticeable gap in research on the Uncompahgre Plateau for thirty years, most of the analyses will rely on typologies established for the Great Basin and Southwestern cultures. It will take the collective effort of researchers working on the Uncompahgre Plateau in the future to create typologies for this area.

PAGE 48

36 Lithics The lithic assemblages (which will also include the groundstone) from Roth and Luster Caves are examined separately, following a stylistic and technological analysis. While lately the emphasis in lithic analysis has been on attempts to establish "pure" classification schemes based on statistical manipulations rather than more description (Holmer 1978, 1986; Holmer and Weder 1980; Wilde 1986), the analysis of the materials from the caves will emphasize comparison with existing collections due to the limited number of specimens. The stylistic analysis will compare typologies that have already been established for the Great Basin and Northern Colorado Plateau regions, including Archaic, Formative and Historic groups. Holmer (1978; Holmer and Weder 1980) will be consulted to a large degree for the Archaic and Formative typologies from the eastern Great Basin. Given the absence of a well-dated projectile point sequence for west-central Colorado, literature from other regions must be referred to for comparative purposes. However, caution will be exercised to avoid utilizing the projectile point typologies too casually to determine site age and cultural affiliation. Archaeologists all too often suggest that the dates of point types in other regions are quite similar to the dates of points in west-central Colorado, and sometimes

PAGE 49

37 imply that the cultural groups are the same (Reed 1984). Plains or Great Basin cultures influence or occupation may be posited, instead of realizing that many projectile point types have a very broad geographical distribution, spanning many cultural groups, and were manufactured over long periods of time. The technological analysis will be conducted on both tools and debitage. Variables which emphasize production technology will be the primary focus, including material, type of object, amount of cortex on the specimen, presence of wear on the specimen and other modification (Wilde 1986:59-63). The bifaces and groundstone, although technologically classified within the analysis will be further broken down into more traditionally-named categories such as scrapers, knives, drills, projectile points, manos and metates. Ceramics Ceramics have often been a diagnostic cultural element of archaeological features in the southwest. It is not until the beginning of the formative period that ceramics appear to come onto the scene. Ceramics have also helped to chart external relationships between cultures through the movement of tradewares, and also the movement of ceramic elements such as design and form. Although few ceramics were located at Luster Cave, and none at Roth Cave -possibly suggesting that Roth Cave may

PAGE 50

predate ceramic technology -it is still possible to compare the ceramics recovered with known types of Anasazi, Fremont, and Nurnic origin. 38 In the Anasazi tradition, the general grayware types that are noted include Chapin Gray, Moccasin Gray, Mancos Gray, and Mancos Corrugated. These types are common throughout the Four Corners region, and are typical of the Mesa Verde Anasazi variant (Breternitz et al. 1974). In the Fremont tradition, the two grayware types that might occur in this area are Uinta Gray and Emery Gray (Madsen 1977). Both of these types are similar to those recovered from Turner-Look, the largest, and closest, investigated Fremont inhabitation to Luster Cave. Buckles (1971:531) felt that the ceramics recovered Luster Cave were similar to those of the Fingertip Impressed Type of Uncompahgre Brownware; a type he feels is typical of the proto-historic cultures that occupied the Uncompahgre Plateau. Perishables Luster and Roth Caves were both protected from the elements that often afflict more open sites. As a result, a variety of perishable items that are normally lost through the processes of decomposition were recovered, including several fragments of basketry and a quantity of cordage and quids. The specimens were measured with

PAGE 51

39 calipers, and all measurements will be recorded in the metric system. Basketry is a class of perishable artifacts that includes several distinct kinds of items, among which are rigid and semirigid containers (or baskets proper) matting, and bags. A11 forms of basketry are manually woven without frame or loom. As all basketry is woven, it is technically a textile class or variety; however, that term is often restricted to cloth fabrics. Cordage is a class of elongated fiber constructions that can be subsumed under the English terms string and rope. The manufacture of cordage is the oldest fiber-based technology in the New World archaeological record. The technological sophistication of these very early cordage specimens indicates considerable antecedent development. Cordage manufacture was very likely part and parcel of the technological repertoire of the earliest migrants to this hemisphere (Adovasio 1988) ..:. ; Basketry specimens were analyzed following Adovasio (1977), basing the analysis on observances and measurements of coiling type, foundation type, and stitch type. Cordage specimens will be allocated to five structural types. The five structural types were established on the basis of three interrelated construction attributes: 1) number and composition of plies; 2) direction of initial "spin"; and 3) direction of final twist (Adovasio 1988)

PAGE 52

40 The term ply is utilized to mean a strand or bunch of fibrous material that is almost always twisted. These strands can be used alone to form single-ply cordage or in groups to form multiple-ply cordage. Multiple-ply cordage is produced by twisting two or more "single" plies together. An individual ply is simple if it consists of a single strand or bunch of material with the same twist. A ply also can be compound. Compound plies are constructed with multiple strands or bunches of material that are individually twisted and then twisted with each other in the opposite direction. Compound plies are therefore separate pieces of cordage that when twisted with other such plies form a technically distinct final cordage type. Spin denotes the initial twist imparted to a .fiber strand or bunch of fibers, and final twist records the direction imparted to several plies that have been twisted together. The term spin is used here to designate the initial twist of a ply because this term is virtually universal in the archaeological literature on cordage and because it facilitates cordage description. The direction of initial spin or final twist can only be S or z. and these terms have exactly the same meaning as specified by Emery (1966:11). Each specimen was also analyzed for the presence of splices qnd knots, as well as length, diameter, number of twists per centimeter, angle of final twist, and other

PAGE 53

41 cordage manipulations, such as rat-tailing, wrapping, etc . Angle of twist measurements were recorded using procedures outlined by Emery (1966:11). Cordage formulae follow Hurley (1979), and the formulae for the five types identified in the assemblage are shown in Table 1. Type I II III IV v Description Single ply, Single ply, Two ply, z Two ply, s TABLE 1 CORDAGE FORMULAE z twist s twist spun, s twist spun, z twist Ply Formula z s s z/z z s/s Compound three ply, s and z spun s z/z s twist s z/z Adovasio (1975; 1980) states thatthere is nothing more homogenous and diagnostic than the basketry associated with the Fremont culture. Within the Fremont basketry assemblage are included coiling and twining -two of the three major sub-classes of basket weaves -the third, plaiting, is virtually absent. Coiling is the numerically dominant subclass of Fremont basketry and is represented in all Fremont sites where basketry is preserved. Basketry produced via twining techniques is relatively uncommon in most Fremont Sites and frequently is not represented at all. Trends discernible throughout the culture area over the 900-year period during which Fremont coiling was

PAGE 54

42 produced include a gradual shift from mixed to almost uniformly R-L work direction, the increased preference of half rod and bundle foundation to all others, and the tendency to employ non-interlocking or intentionally split stitches on the non-work surface to all other types. Thus, it is reasonable to expect the basketry from Roth and Luster caves to reflect these trends, if they are affiliated with the Fremont Culture. Only the worked bone will be discussed within the context of analysis methodology, as all the unworked faunal material was placed in the care of another department at the University of Colorado/Boulder, and has subsequently been misplaced. Unfortunately even the curation records only refer to "animal bone" rather than to specific types precluding even the slightest inference of information from the records. The worked faunal material represents a total of 17 specimens from the collection. Worked bone has traditionally been categorized in a variety of classes according to arbitrarily selected characteristics of the bone from each respective site. However, in the past ten years the "new archaeology" has finally reached into the realms of worked bone analysis to include studies such as .the classification of tool types through usewear analysis (Gooding 1980), experimental

PAGE 55

43 studies of specific bone implements, such as deer ulnas (Harrell 1983} and comparative analysis (Dalley 1970a}. Unfortunately, the majority of sites do not produce samples large enough for major renovations in the analysis of worked bone, including Roth and Luster caves. Thus, a basic analysis determining types of worked bone present -such as gaming pieces or awls will be conducted, and where possible, pertinent information as to usewear and manufacturing techniques will be Macro/Microbotanical Remains Evidence of cultigens is relatively rare in archaeological contexts in west-central Colorado. Sites yielding cultigens are scattered throughout the region in the lower elevations. Nearly all of the finds are corn, but two sites, Tabeguache Cave II and 50R243 have yielded squash remains (Reed 1984). Whether these finds represent domestic or wild varieties of squash is unknown. Macro/microbotanical remains provide information to answer questions about paleoenvironmental data and subsistence. several items in the collection could have been utilized to provide palynological and macrobiotic information, had certain precautions been taken during the excavation to prevent contamination (Adams and Gasser 1980; Scott 1983}. However, most palynological techniques were developed 10 -20 years after the excavation of Roth and Luster caves took place.

PAGE 56

44 Macrobotanical remains recovered from the excavation and curated with the collection will be examined for charring and other indications that they were utilized by the population, rather than being by-products of the local environment (Minnis 1981). Seeds can occur from several sources, both prehistoric and modern, and in caves such as Luster and Roth, preservation of both is typical because of the dry environment, as opposed to open sites, where generally only charred seeds are preserved at any depth below the modern surface level (Minnis 1981). It may often be the simplest, however, to reject all uncharred undomesticated seeds as modern in origin and to retain only the charred material as genuine (Keepax 1977:226). Many ethnobotanists use this as a basic rule, and given the nature of this research, it will be the basic guideline for determining the status of the macrobiotic remains . Of consideration as well is the analysis of the maize present in the collection. Becauseof the lack of funds for re-examination by a paleobotanist, and due to the relative in this category by the author, the analysis of the corn remains by Nickerson provided in the final publication about Luster cave will be utilized (Wormington and Lister 1956) to compare with the current analysis.

PAGE 57

45 Human Bone The analysis of the human remains will follow those procedures outlined in Brothwell (1981). Unfortunately, both the infant burial from Luster cave and the child burial from Roth Cave are missing from the collection. As a result, only the dentition can be observed, with any conclusions that can be drawn based on those findings. While wood is generally considered only to be found in the form of charcoal or posts, within the context of Luster Cave, a variety of worked wood items were recovered. It may be possible to examine these specimens for additional information about the types of projectile points that were being utilized, as well as construction techniques of the item itself. Artifactual determination of wooden specimens will be based on any occurrence of smoothing, cutting, shaping, or binding of the specimens. Once identified as cultural, the artifacts will be segregated on the basis of function, although, where use is not apparent, descriptive categories will be established. Charcoal can be utilized for C14 tests, as well as identified -given the appropriate collection with which to compare it.

PAGE 58

46 Minerals The minerals will be identified according to type. Any cultural modification that may be present will also be noted.

PAGE 59

CHAPTER 5 LUSTER CAVE ANALYSIS Lithics The excavations recovered a lithic assemblage of 211 artifacts. Seventy-nine specimens are classified as tools while the remaining 132 are debitage. Tools were classified and analyzed according to basic production technologies and include flaked or chipped stone and ground and pecked stone. However, these groupings are somewhat arbitrary and some overlapping occurs. Analysis was conducted on both tools and debitage. Variables which emphasize production technology were the primary focus of the initiai analysis, followed by a second analysis which focused on specific attributes which would allow for cross comparison with other lithic assemblages. Of the 211 specimens present in the collection from Luster Cave, there were 22 types delineated within the technological analysis. The distributions of these types across the site are recorded in Table 2. There were two specimens making up the first type. This type was produced from obsidian; both

PAGE 60

48 TABLE 2 DISTRIBUTION OF TECHNOLOGICAL TYPES IN LUSTER CAVE Area A B c D E Level Surf. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 1 1 2 1 Type 1 1 Type 2 2 Type 3 1 Type 4 1 Type 5 1 5 5 3 1 3 4 4 3 2 3 5 4 2 2 4 2 Type 6 1 2 Type 7 1 3 1 1 3 2 1 1 1 Type 8 1 1 Type 9 3 a 2 5 2 2 2 b 2 3 2 2 1 1 1 Type 10 1 Type 11 1 2 Type 12 1 1 1 Type 13 1 Type 14 1 2 1 Type 15 1 Type 16 1 2 2 1 1 1 Type 17 1 1 1 2 2 2 Type 18 1 1 1 Type 19 1 Type 20 1 Type 21 2 Type 22 1 a = 11, b = 12 Area F Level 1 3 4 5 Type 1 1 Type 5 1 2 Type 7 2 Type 9 1 d c 3 Type 14 1 Type 16 2 Type 18 2 c = 13, d = 17

PAGE 61

49 specimens were tertiary flakes. There was one edge which had been retouched on one specimen; the other specimen had no visible wear apparent on any edges. Both of these specimens are classed as debitage. Because the sample of obsidian was so small, no further testing on the obsidian samples was considered. There were two specimens comprising Type 2. Both specimens were choppers of basalt, showing mixed cortical and non-cortical surfaces (Figure 8) On one specimen there was partial battering along one edge, while the other chopper exhibited no such wear. Dimensions are 5.38 and 7.4 em. in length, 6.9 and 6.0 em. in width, and 3.0 and 3.5 em. in height. Figure 8. Basalt chopper.

PAGE 62

There was one specimen which made up the third type. This specimen was a basalt uniface which had been utilized along one edge, and may be considered a chopper. It exhibits mixed cortical and non-cortical surfaces. The following eleven types were all made from cryptocrystalline silicates, which in all cases from Luster Cave are chalcedony and chert -all of which appears to have been quarried from local, known quarry areas (Piontkowski, personal communication). 50 One specimen makes up this Type 4. It is a secondary biface of whitish chalcedony. This specimen has been stained with red ocher in two places along its base. Fifty-six specimens are included in Type 5, all of which are tertiary flakes of a variety of chalcedonies and cherts. Although there did not appear to be visible wear on the specimens, all of the specimens within this category are modified bifaces and will be discussed in detail later in the section. The specimens were recovered throughout the fill from the cave. Type 6 is comprised of three specimens of chert. Each specimen is a tertiary biface, with additional modification. The three specimens will be discussed later in the section. The specimens within Type 7 are all moditied tertiary bifaces of either chert or chalcedony. There does appear to be visible wear on each specimen.

PAGE 63

Type 8 is comprised of two unmodified tertiary unifaces of chalcedony. One specimen appears to have visible wear along one edge. 51 Type 9 consists of 84 specimens of various cherts and chalcedonies. All the specimens are unmodified, tertiary flakes with wear visible on the surfaces of only three specimens. All the specimens within this type are considered debitage, except for the three exhibiting wear, these specimens are considered utilized flakes. Specimens of this type were recovered throughout the fill. Type 10 consists of one specimen. It is a chunk of chert, without any visible wear or modification. Type 11 is comprised of three specimens of chert. The specimens appeared to have mixed cortical and noncortical surfaces, with little visible wear and no additional modification. The specimens are considered to be core fragments. Three specimens considered to be choppers make up Type 12. The specimens are of chalcedony, with mixed cortical and non-cortical surfaces. Some wear was visible on the edges of the specimens, although there was no additional modification. Type 13 is a tertiary biface of siltstone with no visible wear and no additional modification. Type 14 is comprised of five specimens of siltstone. All are secondary flakes with no visible wear

PAGE 64

or additional modification. All specimens in type are considered to be debitage. One secondary flake of siltstone comprises Type 15. It exhibited no visible wear or additional modification. 52 Type 16 is comprised of ten specimens of sandstone. All specimens are considered cobbles and show extensive visible wear, although there is no additional modification. All the specimens in this type are considered to be groundstone and will be discussed later in the section. Type 17 consists of nine specimens of Brushy Basin quartzite. The specimens are tertiary bifaces, some of which have additional modification and will be discussed in further detail later in the section. Five tertiary flakes of Brushy Basin quartzite comprise Type 18. There was no visible wear or additional modification present. All the specimensin this type are considered debitage. Type 19 consists of one specimen which is a chunk of quartzite without any visible wear or additional modification. The specimen is so clear, it was almost mistaken for glass. Type 20 consists of.one specimen of quartzite. It is a cobble with extensive visible wear. This specimen is considered to be a (Figure 9).

PAGE 65

53 Figure 9. Quartz hanunerstone. Type 21 consists of two cores with mixed cortical and non-corticai surfaces of quartzite. There appears to be limited wear, and no additional modification. One primary pebble of quartzite comprises Type 22. There does not appear to be any visible wear or additional modification. Its possible use is unknown. Grounds tone There were ten specimens which were determined to be groundstone -all of sandstone. Nine of the specimens are manos, while one specimen is a piece of sandstone which has been deeply grooved -possibly having been used to smooth or sharpen bone or wood (Figure 10). The dimensions of this specimen are 9.27 em. x 3.14 em. x 2.15 em.

PAGE 66

54 The manos were of various shapes and sizes (Table 3). Included in the shapes were two rectangular specimens, one of which had tapered ends (Figure 11). There were two loaf-shaped specimens, and three round (Figure 12, 13). There were also two round/asymmetrical specimens as well. All of the manos were small, and are considered to be one-handed manos. There were battering marks on most of the specimens which indicated either shaping or the use of the mano to pound the grain before it was ground. There were six two-sided manos and three one-sided manos. Lengths ranged from 6.51 em. to 14 em.; widths rangedfrom 4.75 em. to 10.3 em. The mano tool is generally shaped to match the style of the metate on which it is to be used. Thus, the small, one-handed rnano may be slightly shaped from a river cobble to fit the basin-shape curvature of the metate. TABLE 3 DISTRIBUTION OF GROUNDS TONE IN LUSTER CAVE Type Area Level Material # Handed # Sided Loaf-Shaped A 5 Sandstone 1 2 Loaf-Shaped B 3 Sandstone 1 2 Rectangular Surface Sandstone 1 1 Rect./Tprd. B 4 Sandstone 1 2 Round B 4 Sandstone 1 2 Round c 2 Sandstone 1 1 Round/Tprd. A 5 Sandstone 1 2 Round/Asym. F 5 Sandstone 1 1 Round/Asym. F 5 Sandstone 1 2 Grvd Sndstne A 1

PAGE 67

55 Figure 10. Sandstone arrowshaft smoother. Figure 11. Rectangular mana.

PAGE 68

56 Figure 12. Loaf-shaped mano. Figure 13. Round rnano.

PAGE 69

57 Bifaces The bifaces were. categorized as follows: 32 points, nine scrapers, eight knives/knife bases, two drill tips and 24 utilized flakes/bifaces (Figure 14). Following is a discussion and typological comparison of the bifaces which are considered to have been modified into projectile points. Their distribution within Luster Cave is noted in Table 4. There were sixteen points with no typological comparison (Figure 15). The first specimen is cornernotched with a flaring stem (Figure 15a). It is of chalcedony. The interesting thing about this point is that it has a black residue, possibly resin, that covers parts of the stem. The second specimen is a side-notched fragment (Figure 15b) The third specimen is also a side-notched fragment, although this specimen appears to have been modified more extensively than the second 15c). The fourth through seventh specimens are notched, spirate triangles (Figure 15d, e) The eighth and ninth specimens were very side-notched fragments (Figure 15f) The tenth and eleventh specimens are small, well-made points. Both specimens are side-notched, although one is also notched in the center of the base (Figure 15g, h).

PAGE 70

58 TABLE 4 DISTRIBUTION OF PROJECTILE POINT TYPES IN LUSTER CAVE Area A B c D E F Level 1 2 3 4 1 2 3 4 5 1 2 3 4 1 1 2 1 5 Untyped 1 1 1 1 3 2 3 1 2 1 Buckles' 2 1 1 Buckles' 5 1 1 1 Buckles' 7 i Buckles' 10 1 Buckles' 18 1 Buckles' 19 2 Elko 1 Pinto 1 1 Gypsum 1 2 1 Desert S-N 1 1 Nawthis S-N 1 1 Uinta S-N 1 Parowan B-N 1 1 Figure 14. Bifaces.

PAGE 71

e 9 I . . h Figure 15. Untyped projectile points (actual size). 59

PAGE 72

The twelth specimen was a tip fragment, with no indication of basal modification. The thirteenth specimen is a non-descript point. It appears to be a biface, although there is a slight notching of one side (Figure 15i). The last two points that were not typed are two side-notched specimens, although one specimen appears be somewhat Gypsum-like (Figure 15j). 60 There were two specimens that compared favorably with Type 2 described by Buckles (1971). The second of the two specimens also compares favorably with the Desert Side-Notched point of Fremont origin (see the Discussion on Desert Side-Notched points later in section) (Figure 16). Buckles describes this type as small side and basal-notched points, possibly representing a type that is more commonly found to the west of the mountains (Buckles 1971:116, Figure 2). This may suggest a Great Basin origin. Buckles also compares this point type with those found on the Plains after 1500 (Buckles 1971:116; Kehoe 1966). There are three specimens in the next category that are similar to Buckles' Type 5 (Figure 15k). Buckles describes this type as small basal-notched points with triangular shaped bodies and short expanding bases (Buckles 1971:119-120, Figure 2). Two of these points also compare favorably to the Parowan Basal-Notched type of Fremont origin (Figure 16). Buckles suggests this type

PAGE 73

a b c f J Figure 16. Typed projectile points. Elko Side-notched (a), Pinto Shoulderless (b, c), Gypsum (d, e), Desert Side-Notched (f, g), Nawthis Side-Notched (h, i), Uinta Side-Notched (j) and Parowan Basal-Notched (k, 1). All points are actual size. 61

PAGE 74

62 resembles large projectile points of Basketmaker and Pueblo II populations, as well as points found in Fremont sites (Buckles The next specimen is comparable to that of Buckles' Type 7 (Buckles 1971:120, Figure 2) (Figure 16). It is also comparable to the Gypsum point of Archaic origin. The following specimen is comparable to that of Buckles Type 10 (Buckles 1971:122). Buckles describes this type as being small triangular unnotched points with slightly convex to straight sides and slightly convex to straight bases. Buckles suggests that these points have their highest frequency relationships with what are believed to be historic and proto-historic Ute occupations defined as the Escalante Phase on the Uncompahgre Plateau (Buckles 1971:122, Figure 2). The next specimen compares with Buckles Type 18, as well as Desert Side-Notched (Buckles 1971:130-131, Figure 3) (Figure 16). Buckles describes this point type as being large side-notched points with straight or slightly concave bases. Buckles compares these with similar specimens from Danger Cave (Jennings 19.57:121, Figures 97a and b), with specimens from the Bitterroot Phase in Idaho and with early points from Plains sites (Buckles 1971:130-131). The next points are similar to Buckles Type 19 (Buckles 1971:131, Figure 3). They are also comparable to

PAGE 75

Desert Side-notched specimens of late prehistoric origin (Holmer 1980:60). There is one specimen which compares favorably with the Elko Side-Notched type (Holmer 1978) (Figure 63 16a) The Elko series has traditionally been divided into three variants: Corner-Notched, Side-notched, and Eared. The Elko Side-Notched is similar in form to the Elko Corner-Notched except that the maximum stem width is approximately equal to the maximum blade width. Tangs are rarely present; the distal notch angle often approaches horizontal, causing a shouldered appearance to the blade. The Elko series projectile points are the most plentiful but the least temporally diagnostic of the point types commonly found in the northern Colorado Plateau and the far eastern Great Basin. About all that can be positively stated is that they occur after 7600 B.P. They possibly persisted into historic times; Powell collected a Paiute hafted knife incorporating and Elko-like point in 1873 (Fowler et al. 1973:41). The great time depth of the Elko series points refutes the suspicions of several (Clewlow 1967; Heizer and Baumhoff 1961; O'Connell 1967) that their temporal occurrence everywhere in the Great Basin falls between 3500 and 1400 B.P. That estimation is based mostly on central and western Great Basin research with the Danger Cave data being discounted as aberrant (Heizer and Baumhoff 1961). Hester and Heizer (1973), however, acknowledge that the Elko series points are

PAGE 76

probably earlier in the far eastern Great Basin than to the west, and a diffusion of the Elko series points into the central and western Basin from the east has been postulated (Adovasio 1970) 64 There. are three specimens which compare favorably with the Pinto Shoulderless (Holmer 1978) (Figure 16b, c). Holmer (1978) places the age range of the Pinto series at approximately 8300 to 6200 B.P. at the four Archaic sites used in his study. Joes Valley Alcove in the northern Colorado Plateau provides a time span of approximately 8300 to 6300 B.P. for the Joes Valley Tanged points which are identical to the Pinto series points (Holmer 1978). These points date more recently at other sites, however, as they are common between 6500 and 3800 B.P. at O'Malley Shelter (Fowler et al. 1973) and from approximately 5000 to 3000 B.P. at Swallow Shelter (Dalley 1976). Pinto points are also reported in western Colorado on the Uncompahgre Plateau (Wormington and Lister 1956:14) although no dates are available. Three points compare favorably with the Gypsum point (Holmer 1978) (Figure 16d, e). These points have already been discussed as similar to several types Buckles distinguished on the Uncompahgre Plateau (Buckles 1971). The Gypsum projectile point is the most recent type generally associated with the Archaic stage of the northern Colorado Plateau and the southern portions of the Great Basin. The Gypsum point dates from approximately

PAGE 77

65 4600 to 1500 B.P. at Sudden Shelter and Cowboy cave. oMalley Shelter (Fowler et al. 1973:42) contained 105 specimens, the most ever found at a single site, and they date from approximately 5000 to after 1000 B.P. This time range estimate is supported by radiocarbon dates from Gypsum Cave (Heizer and Berger 1970:17; Shutler 1967:306) that indicate a range from approximately 3000 to 2000 B.P. Gypsum points are unique among Archaic projectile points and may be of interest in understanding the development of hafting technologies. It was observed at Cowboy cave that most Gypsum points and many later arrow points have the remnants of pitch on their basal stems (Jennings 1980). No earlier point types show any trace of pitch. The discovery of pitch as an adhesive may have reduced the need for the deep side or corner notches characteristic of earlier dart points, although deep notching probably remained a necessity for knife blades because of the cantilever forces produced during cutting. The Elko series may represent the type of blades retained as knives. Two specimens are similar to the Desert Side Notched point, and have already been discussed as being similar to types Buckles delineated for the Uncompahgre Plateau (Buckles 1971) (Figure 16f, g). They have been recovered from excavated sites near the northern and western periphery of the Fremont area although they have been reported in uncontrolled situations throughout the

PAGE 78

66 Fremont area (Berry and Berry 1976). Of the excavated Fremont sites they have never been the dominant type, making up only 12% of the total points recovered (Holmer and Weder 1980). Of significance is that most of those sites contain Shoshoni ceramics although they constitute only 3% of the total ceramic collection (Holmer and Weder 1980). The correlation has been inferred by Fowler et al. (1973) at O'Malley Shelter, and by Frison (1971) at the Eden-Farson Site in Wyoming. The conclusion is that the occurrence of Desert Side-notched points does not result from Fremont occupations but indicates post-Fremont Shoshoni use of the area after approximately A.D. 1150 (Holmer and Weder 1980). This conclusion is supported by the presence of identical point'types in the northern Plains (Plains Side-notched) after approximately A.D. 1590 (Kehoe 1966) There were two specimens which compare favorably with the Nawthis Side-Notched points (Holmer and Weder 1980) (Figure 16h, i). Their distribution is limited to the southern half of the Fremont region dating from approximately A.D. 950 to 1250. They are similar, if not identical, to points associated with. Pueblo II occupations south of the Colorado River. There is no apparent associations with any single ceramic type although most sites also produce small quantities of both and Virgin ceramics.

PAGE 79

67 One specimen compared favorably with that of the Uinta Side-Notched points {Holmer and Weder 1980) {Figure 16j). They are widely distributed over the northern half of the Fremont region dating from approximately A.D. BOO to 1200. Sites where they are the dominant type usually contain a large percentage of Uinta Gray ceramics. Their similarity to the Prairie Side-notched points recovered in the northern Great Plains dating from approximately A.D. 700 to 1300 (Kehoe 1966) supports conjectures of a Great Plains influence in the northern Fremont areas (Aikens 1966) There were two specimens which compared favorably with the Parowan Basal-Notched, both of which have been discussed previously as being comparable to.types delineated by Buckles for the Uncompahgre Plateau {Buckles 1971) {Figure 16k, 1). The spatial distribution of Parowan points can be divided into two groups based on ceramic associations (Homer and Weder 1980). The first area includes the Virgin River, Santa Clara River, and Johnson Canyon. The sites in the area have a high percentage of Virgin ceramics (98%) and very small percentages of Kayenta, Mesa Verde or sevier ceramics (2%). The temporal span clusters between approximately A.D. 900 and 1200. Parowan points constitute 63% of the total arrow points from sites in this area. The second area.includes the Parowan Valley, part of the Sevier River Drainage, and part of Southeastern

PAGE 80

68 Nevada. The ceramics from the sites in this area are predominantly Sevier ceramics. The temporal span clusters around A.D. 950 to 1150. Parowan points constitute 55% of the total arrow points recovered. Parowan points have been noted at other sites at low frequencies as far east as the Bull Creek drainage. Not surprisingly, the two pithouses at Bull Creek from which Parowan Points were recovered were the ones which had significant quantities of Virgin Kayenta ceramics. Parowan points are the predominant points in both the Parowan and Virgin Kayenta cultural regions. They were used ca. A.D. 950 .-1150. Ceramics There were 32 ceramic fragments present in the collection; one was a rimsherd and the remaining 31 were body sherds. Ceramics were only recovered from Luster Cave. While all the specimens appear to have been constructed using an obliterated coiling method and all specimens appear to have some degree of corrugation present on the outer surface, it appears that there may be three types of ceramics present, based on core.thickness, temper, clay, degree of surface manipulation and color. The first type appears to have been constructed from a micaceous clay, utilizing small quartz fragments as temper (Figure 17). Firing was partial oxidation, with the color a light gray on the surface, and dark gray in the

PAGE 81

69 Figure 17. Sangre de Cristo Micaceous Wear. interior, with a large amount of soot on the outer surface. There was no slip evident on any of the specimens. The thickness ranges from .5cm to .6cm. It is difficult to determine shapes and sizes of vessels without more rimsherds, but the rimsherd present is of this type, and the profile indicates that it probably was from a simple bowl (Figure 18). There appears to have been a slight corrugation of the exterior surface, but this may also be due to an only partial obliteration of the coils on the exterior surface, as well as a heavY layer of soot which nearly covered the entire outer surface. The texture of this type was fine-to-medium. There were 25 specimens representing this type, the distributions across the site are recorded in Table 5.

PAGE 82

Figure 18. Rirnsherd Profile. TABLE 5 DISTRIBUTION OF CERAMIC TYPES IN LUSTER CAVE Area Level Type 1 Type 2 Type 3 2 2 B 3 5 5 8 c 1 1 F 4 3 1 surface 9 2 It has been suggested that micaceous ceramics similar to this first type are variable and have widespread d -istributions in context which include occupations by Utes, diverse'Athabascans, Puebloans, Hispanics and others (Buckles 1988:221; Baugh and Eddy 1987). Baugh and Eddy recommend that such micaceous ceramics should not be identified with ethnic-specific classifications, but should be classified as Sangre de Cristo Micaceous Wear. The temper of the second and third types was crushed sandstone. The color was generally a brownish 70

PAGE 83

71 gray to gray, indicating a reducing atmosphere. There was no slip evident on either the second or third types. The thickness of the sherds averaged 0.7 em. to 0.8 em., although there were two specimens with thicknesses of 1.0 em. The second type indicated a clear corrugation, and there was little of the sooting on the outer surface that characterized the first type (Figure 19). Because all of the sherds from this type were body sherds, it was difficult to determine the shapes the sherds may have represented. There were 3 specimens representing this type. The third type was very similar to the second type, except the surface manipulation was comprised of fingernail impressions, rather than corrugation (Figure 20). One of the specimens had a hole drilled through it, possibly indicating repair or a way to carry it. The texture of this type, as well as the second type, was medium to coarse. There were 3 specimens of this type. All the types represented are from the same area of the cave, although separated arbitrarily by Luster's sections, so it is difficult to separate them temporally, as well as stratigraphically, from one Both Types 2 and 3 compare favorably with Buckles' Uncompahgre Brownware (1971:507-522), while Type 1 is considered similar to the Sangre de Cristo Micaceous Wear.

PAGE 84

Figure 19. Uncompahgre Brownware corrugated. Figure 20. Uncompahgre Brownware, fingernail impressed. 72

PAGE 85

73 Buckles gives a broad range of temporality to Uncompahgre Brownware, suggesting it was produced approximately 400-500 years B.P., or with the advent of the Ute in the area (Buckles 1971:552). The temporality of the Sangre de Cristo Micaceous Wear can only be guessed at, since its distribution is across many ethnic and temporal borders (Baugh and Eddy 1987). Perishables There were 171 specimens of perishable materials present in the collections recovered from Luster Cave. There is a wide variety of different items considered to be "perishables". The distribution of the items is recorded in Table 6. There were four specimens of basketry present in the collections, all small fragments (Figure 21). Two of the specimens were merely a split rod wrapped with a series of yucca leaves (Figure 21a, b). These specimens are similar to several examples recovered from Danger Cave (Jennings 1980:72, 74). The other two specimens are very small fragments of basketry, one of which had been treated with some type of preservative (Figure 21c); the other is merely a segment of a 1/2 rod and bundle held together with one section of noninterlocking stitch (Figure 21d). The specimen treated with preservative is a one-rod-and-bundle foundation held together with

PAGE 86

74 TABLE 6 DISTRIBUTION OF PERISHABLE ITEMS IN LUSTER CAVE Area A B c D E F Level Sur. 1 2 3 4 1 2 3 4 5 1 2 3 4 1 1 1 2 3 4 5 Basketry 1 2 1 Quids 5 2 1 1 1 2 1 8 2 2 1 Leather: Modified 1 Urunod. 1 1 3 1 Thong 1 1 Bundles 2 Ycca Lvs 1 3 2 1 2 1 1 1 1 1 1 Ycca Knts 1 1 1 2 4 1 4 Cordage: Yucca Type 1 1 Type 2 1 5 2 1 1 2 7 2 3 2 5 4 Type 3 1 1 1 1 1 Type 4 1 Fur Type 1 4 4 Type 2 1 1 Type 3 1 Sinew 2 2 1 1 3 Human Hair 1 Juniper 2 1 1 Urunod: Hair 1 Fur 1 2 1 Reed 1 1 2 1 1 1 Ycca Bs. 1 Atlatl Shft 2 Arrow Shaft 2 2 Gmng Pieces 1 1 Shft/Tnon 2 1 1 Wood. Waste 1 2 1 Fbr-wrp. St. 1 2 Dcoratd. Rd. 1 Misc. Wood 2

PAGE 87

Figure 21. Basketry. Split rod wrapped with yucca leaves (a, b), basketry treated with preservative (c) and basketry fragment (d) stitches. The dimensions of this specimen are 3.15 em. x 1.25 em. x .61 em. Adovasio (1971) has traced the differences between textiles, in particular basketry and cordage, of 75 the Great Basin and Southwest. Given his descriptions of the variations between the two areas, the basketry from Luster Cave is most similar to that of the Great Basin, where one-rod-and-bundle foundation coiling appears to have been the standard. There were 26 yucca quids present in the collections (Figure 22) There was a note attached to one of the quids dated May 26, 1956. It said: ''Largest quid Yucca sp. One quid evidently has some Agave fibers in it. Needs to be checked with A. utahensis. In natural range? Vorsila L. Bohrer."

PAGE 88

Figure 22. Yucca quids. There were 9 specimens of leather present. Included within this category was one piece of modified soft hide, six pieces of unmodified soft hide, and two pieces of thong {Figure 23). The modified soft-hide category was limited to a specimen of softened skin that appeared to have several perforations in it for sewing {Figure 23a). This piece measured 7 em. by 3 em, and was recovered from Area F, Level 3. 76 The unmodified soft-hide category included scraps that were probably waste pieces from the manufacture or repair of soft-hide articles {Figure 23b). This included six scraps, with sizes ranging from 5.8 em. to 1.8 em. in length, and 5.5 em. to 1.1 em. in width. Two soft-hide thongs were recovered, the length of the specimens measured 9.3 and 4.8 em {Figure 23c, d).

PAGE 89

Figure 23. Modified leather (a), unmodified leather (b) and leather thong (c, d). 77 There were seven specimens of unmodified fur/hair. Included within this category were three specimens of human hair, all of which were brownish-black in color and fairly coarse in texture. The remaining specimens represented lagomorph fur from either the white-tailed jackrabbit (Lepus townsendii) or snowshoe rabbit (Lepus americanus). There was one specimen of an yucca plant base from which leaves have been cut off. There were two bundles present in the collection. The first specimen was entirely of yucca; the second specimen was of shredded yucca, yucca leaves, leather and two Cymopterus umbelliferae roots (Figure 24).

PAGE 90

Figure 24. Bundle of shredded yucca leaves, leather and umbelliferae root. Umbelliferae roots are noted to have been used for food, seasoning and medicine (Colton 1974:305; French 1971:385-412; Harrington 1967:171-173; Whiting 1939:86). 78 This bundle may have been used in a healing ritual, or it may have been used during travel. There were 66 lengths of cordage made from plant fibers recovered from Luster Cave. Analysis of the collection included manufacturing techniques, material employed and knot-types utilized. There were no artitacts made of cordage recovered. The majority of cordage recovered were small scraps which show the effects of hard and continual use. Some of the specimens are burned, worn to the point of have broken fibers or are even worn through. Many of the specimens are knotted together, indicating a frugal effort to save cordage.

PAGE 91

79 In a description of cordage, the term l2.lY refers to a single yarn which is usually plied with another singleyarn to become a two-ply cord or yarn. The direction of twist is determined as follows: "If the elements are twisted in one direction so that the slope of the spirals, when held in a vertical position, conforms to the central portion of the letter S, the cord is said to have an S-twist. If the elements are twisted in the opposite direction, the cord has a z-twist." (Rohn 1971:114). Two-ply cords are by far the most common within this particular collection. A tabulation the cordage specimens from Luster Cave according to direction of twist shows 57 pieces of z-twist cordage, as opposed to eight pieces of s-twist cordage. Both twist types are present throughout the fill of the cave. There are several methods of spinning yarn with a spindle; one of the most common is to roll it along the thigh. Ruth Underhill (1944:36) has shown that twist direction is dependent upon the direction the spindle is rolled. If it is rolled away from the body, an s-twist cord results; if the spindle is rolled toward the body, the cord will. be Z-twisted. This would mean that in order to make a two-ply z-twist cord, the first yarn would be rolled away from the body to get the s-twist, then the ply twist would be achieved by rolling the yarns toward the

PAGE 92

80 body. Spinning along the leg can also be achieved without the aid of a spindle. Another method of using a spindle is to drop the spindle and let it spin freely just above the ground. Here, as before, the twist direction is determined by the direction the spinner twists the spindle as it is dropped. It is impossible to say what method was employed at Luster Cave, although the absence of any spindle whorls may indicate employment of the first method discussed. Of the 66 total cordage specimens, 63 are twoply, obviously the most popular manufacturing technique. The remaining three are single ply, which may represent cordage which has come unplied. Overhand and square knots appear to have been utilized for a variety of purposes relating to the cordage. Overhand knots appearing at the ends of cords may have prevented the fibers from unraveling. Square knots, as well as granny knots, were used to join two pieces of cord together. It has been noted that the square knot seems to predominate in the Southwest (Basketmaker through Pueblo), whereas the sheetbend and overhand knots are more common in the Great Basin (Lambert and Ambler 1961:57). Within the collection of cordage, it appears that there are eleven types, based on material, number of strands, and direction of ply.

PAGE 93

There were 41 specimens of yucca cordage which made up four types. 81 The first type was a single specimen made up of a combination of a strand of yucca fibers and a strand of sinew (Figure 25a). It was a Z-twist, with a length of 5.3 em., and a thickness of 1.31 em. There was just one twist per centimeter, and the angle of the twist was 60 degrees. The second type is made up of 34 specimens of two-strand, Z-twist Yucca fibers (Figure 26). There were four specimens with only two twists per centimeter, with twist angles of 60 degrees (two specimens), 65 degrees and 75 degrees (one specimen each). There were eleven specimens with three twists per centimeter, with twist angles of 45 degrees (3 specimens), 60 degrees (six specimens) 65 degrees and 75 degrees (one specimen each) There .were eight specimens with four twists per centimeter, all with twist angles of 60 degrees. There were five specimens with five twists per centimeter, with twist.angles of 45 degrees (one specimen), 60 degrees (three specimens) and 75 degrees (one specimen). Lengths range from 1.61 em. to 33.5 em.; thickness ranged from 0.1 to 0.61 em. The third type includes five specimens of twostrand, s-twist Yucca fibers (Figure 26). Lengths range from 4.4 em. to 5.7 em.; widths range from 0.15 to 0.16 em. Three of the specimens had 4 twists per centimeter,

PAGE 94

0 2 3 4 Scm Figure 25. Yucca and sinew cordage (a) and sinew cordage (b, c and d). Figure 26. Yucca cordage. s-twist cordage (a) and ztwist cordage (b, c, d, e and f). 82

PAGE 95

all with angle twists of 60 degrees. There was one specimen each which had 3 and 5 twists per centimeter, with angle twists of 45 and 60 degrees, respectively. 83 The fourth type is comprised of one specimen of a single twist of yucca fibers. Its length is 14.4 em. and the thickness is 0.1 em. There are eleven specimens of fur cordage, of which there are three types. The first type is comprised of eight specimens of two-strands with a z-twist (Figure 27). There were either 2 or 3 twists per centimeter, three being the most popular with six specimens. The angle of the twist was either 30 degrees (three specimens), 45 degrees (four specimens) or 60 degrees (one specimen). The lengths ranged from 2.85 em. to 22.0 em.; thicknesses ranged from 0.27 em. to 0.81 em. The second type was comprised of two specimens of one strand em.; thicknesses were 0.5 em. and 0.27 em., respectively. The third type was comprised of a single specimen. This specimen was a two-strand, S-twist cord. There were two twists per centimeter, and a twist angle of 60 degrees. The length was 4.8 em. and the thickness was 0.8 em. There were nine specimens of sinew which made up two types. Both types were of two strands, eight of which had a z-twist and one which had an s-twist (Figure 25b, c

PAGE 96

84 Figure 27. Fur cordage. and d). Eight specimens had three twists per centimeter, and one specimen (not the s-twist) had four. All specimens had a twist angle of 60 degrees. Lengths ranged from 2.85 em to 16.5 em; widths ranged from 0.15 em. to 0.29 em. There was one specimen of cordage made of human hair fiber (Figure 28}. It was a loosely woven two-strand, Z-twist cord, with two twists per centimeter. The twist angle was 60 degrees. The length of the specimen was 11.3 em.; the thickness was 0.3 em. A final type of cordage was made of shredded juniper, all the specimens were two-strands with a ztwist (Figure 29). Three of the specimens had one twist per centimeter, with a twist angle of 60 degrees. A final specimen had two twists per centimeter, with a twist angle

PAGE 97

85 0 2 3crr ,_ ........ __ ___. Figure 28. Human hair cordage. Figure 29. Shredded juniper cordage.

PAGE 98

of 7 5 degrees. Lengths ranged from io. 5 em. to 22. 5 em. ; thicknesses ranged from 0.29 em. to 0.49 em. 86 There were 15 specimens of shredded yucca leaves and 14 yucca leaves consisting of only a knot (Figure 30). The purpose of these specimens is unknown, although the shredded leaves may have been in a preparatory state for cordage. The specimens which consist of only a knot may be remnants of fiber which was used as cordage. Another possibility may be that they were "doodles" and have no significant use. Among the knot-types represented, there were two Larks-head, seven square knots, three overhand, one double and one figure-of-eight. Figure 30. Yucca knots (actual size).

PAGE 99

87 There were 26 specimens of wood present in the collection of artifacts recovered from Luster cave. The artifacts were separated on the basis of apparent function, although where use was not apparent, descriptive categories were established. Ten types were represented in the collection. Two of the specimens from the collection were determined to be atlatl dart shafts (Figure 31). 0.72 em. and 0.74 em., and 10.6 em. and 19.6 em. in length, respectively. Atlatl shafts are common in dry cave sites throughout the desert Southwest. Four wooden foreshafts were judged to be parts of composite arrows (Figure 32b, c; 33a). Two of the specimens appear to be v-notched proximal sections. There appears to be resin residue within the notch of both specimens. Both specimens are 0.62 em. in diameter; the lengths are 3.9 and 4.83 em. The other two specimens are unidentifiable fragments, although it is most likely they are sections of arrow shafts. The diameters of these specimens are 0.62 em. and 0.6 em.; the lengths are 2.1 and 2.94 em. Wood parts of composite arrows commonly found in southwestern sites where perishable artifacts are preserved (Janetski 1980). Two specimens from the collection were classified as gaming pieces (Figure 34). Both specimens are undecorated. Cuts on opposite ends of the piece can

PAGE 100

Figure 31. Figure 32. Atlatl shafts. Arrow foreshafts {b, c) and painted reed fragment (a) 88

PAGE 101

89 2 3 4 Scm Figure 33. Arrow shaft (a) and modified wood (b, c, d and e) Figure 34. Wooden gaming pieces.

PAGE 102

90 clearly be seen on the first of the two specimens, while the other two sides have been smoothed. Deep transverse striations cover approximately 75 % of one side, while the other side is smooth. The specimen has burn marks along one of the smoothed edges. The second specimen appears to be unfinished. The dimensions of the specimens are 3.61 em. and 4.8 em. in length; 2.19 em. and 2.0 em. in width; and 0.5 em. and 0.3 em. in height, respectively. Three of the four shafts with tenons are nearly identical, varying only in length, and with slight fluctuations in diameter; the fourth is different only in the fact that the end opposite the tenon is fragmented (Figure 35). The four specimens have carefully smoothed shafts which taper to slightly blunted points, except for the fragmented one, with the other ends cut off by a method described by Cosgrove (1947, Figure 71). The process consists of cutting opposite notches on the stick about a third of the way through, leaving the center intact. Then, after scoring the stick at the desired breaking point and prying on the notch to weaken the fibers, the tab is broken out by simply bending the tenoned portion of the stick. All four of these shafts have the tabs still intact. Lengths are 6.9, 6.2, 3.3 and 8.0 em.; diameters are 0.6, 0.52, 0.6 and 0.65 em. Similar items with tenons have been recovered in various sites in Utah (Dalley 1976, 1970b:167; Janetski 1980:83-84, Figure 35), as well as in Arizona (Cosgrove

PAGE 103

Figure 35. Shafts with tenons. 1947), and they are generally considered to be waste produced in notching arrow or dart foreshafts to receive projectile points. It seems unlikely that the three carefully smoothed and tapered foreshafts would be waste material, though it is possible that they were wooden arrow foreshafts which had been cut off to allow for conversion from wooden to flint points. This might explain the slightly blunted tips on these shafts. A comparable artifact, called an arrow foreshaft, is described by Dalley (1970b:167). 91 Four specimens can be identified as waste from woodworking activities (Figure 36). All are round sticks of varying diameters and lengths, but all are less than 1.8 em. in diameter and 10 em. in length. All are smoothed for a distance at one end and are cut off, most likely by notching, either around or on opposite ends. The uncut ends are rough and frayed. Lengths measure 9.7, 8.1

PAGE 104

92 Figure 36. Woodworking waste. and 3.9 em; diameters vary from 0.85, 1.31, 0.89 and 1.39 em. The third specimen exhibited charring on both ends. Woodworking residue has been recovered from Juke Box Cave {Jennings 1957:193), Cowboy Cave {Janetski 1980:85, Figure 34), Hogup Cave {Dalley 1970b:181), various caves in New Mexico reported by Cosgrove {1947, Figure 137), and Sand Dune Cave {Lindsay et al. 1968:71, Figure 45). Four twigs were wound with a fine vegetal fiber. Two of the specimens were small twigs which have been wrapped together with the fiber, and appeared to be charred on one end (Figure 37). The fourth specimen is larger, although there are no additional distinguishing characteristics present. Neither specimen is formally distinctive nor do they show what could be considered special attention or care in manufacture. It may be

PAGE 105

93 0 2 3 4 Scm Figure 37. Fiber-wrapped twigs. possible these are merely examples of "doodling" and have no special significance. Lengths are 11.1, 8.39 and 3.72 ern. One fragmented section of reed was found which had been split lengthwise and was painted in the middle third of the specimen (Figure 32a). Length is 6.1 em. and width is 0.7 ern. There were seven specimens of unmodified reed. Two specimens of wood which do not appear to have anything distinctive about them, except for the fact that they have been decorticated, and blunted and smoothed on the ends. Lengths are 6.32 and 2.48 ern.; diameters are 1. 1 and 0 7 em. The presence and variety of wood and reed artifacts suggests that they were an important component

PAGE 106

94 of the prehistoric tool kit on the Colorado Plateau. Activities reflected in the remains include: limited hunting, as eyidenced by the few arrow parts, and atlatl shafts; and social or leisure activities, suggested by the occurrence of wood gaming pieces and bound twigs. Worked Bone There were 14 specimens of worked bone and three specimens of worked antler recovered from Luster cave representing awls, flakers, beads, and gaming pieces. The distribution of the worked bone artifacts in Luster Cave is noted in Table 7. TABLE 7 DISTRIBUTION OF WORKED BONE IN LUSTER CAVE Area Level Awls: Type A Type B Type c Gaming Pieces Bead Fragment Bone Tubes Antler "Fishhook" Unid. Fragment Shell Pendant A B C D E F Surf. 1 3 1 2 3 4 1 1 5 1 1 1 1 1 2 1 ],. 1 1 1 1 1 There were five specimens of worked bone representing three types of awls. Three specimens comprise Type A; which appear to have been utilized splinters of (Figure 38a, b and c). All three specimens exhibit polish on the tip, while

PAGE 107

Figure 38. Bone awls. Type A (a, b, and c), Type B (d) and Type c (e) one specimen exhibits polish on the edge as well. 95 Diagonal striations appear on the exterior surface of one of the specimens, while transverse striations appear on the tips of the other two. It is most likely that the diagonal striations were.formed during the utilization of the specimen, possibly as a result of punching and twisting the material on which the awl was being used. It is possible the transverse striations were formed as a result of manufacture, as it appears that the tips of the specimens were shaped. Lengths are 6.16 em., 7.2 em. and 4.1 em. One specimen comprises Type B; which has been extremely modified on the articular end (Figure 38d) This specimen appears to have been shaped -with the articular end removed, and the tip tapered (not to an extreme point, but somewhat blunted). This specimen has

PAGE 108

96 been burned along the majority of its length, but appears to have been utilized after its subjection to fire, because the polish extends over the burned area. There are transverse striations along the edge and tip of the specimen, possibly due to manufacture, given the fact that the specimen is so extremely modified. Length is 6.83 em. One specimen comprises Type c (Figure 38e) The articular end was incorporated into the design of this awl, as the entire end remains, as if to make grasping it easier. The tip was noticeably shaped, although not to an extreme point. There are transverse striations on the tip of the specimen, as well as diagonal striations along the edges. It is most likely the transverse striations were created during the manufacture of the specimen, while the diagonal striations were created during the utilization of the specimen. This specimen was heavily encrusted with caliche. Length is 7.26 em. There is one specimen which is presumed to be gaming piece. The specimen is a flattened piece of unidentifiable bone, with incisions present (Figure 39e). The length is 2.15 em.; the width is 1.08 em. There are four specimens which appear to be tubular beads (Figure 39a, b, c and d). The specimens are cut at both ends and exhibit incisions on two of the specimens. The specimens are polished as well. Lengths are 1.61, 2.0, 2.0 and 2.2 em.

PAGE 109

97 Figure 39. Bone gaming piece (e) and bone tubes (a-d). There is one specimen which is interpreted to be a bead, although 1/2 of the specimen is missing (Figure 40a). is 1 em. There is one unidentifiable fragment. There are four antler specimens which are possibly flakers. All the specimens have been cut and shaped on one end to a blunted, flat end (Figure 41). All the specimens exhibited diagonal striations, either on the edges of the specimens or on the tip and interior surface. Lengths are 3.85 em., 5 em., 5.45 em. and 3.94 em. There was one specimen present in the collection made of bone that resembles the bottom part of a fishhook (Figure 40b). It was recovered from Area B, Level 1 and may possibly be an example of an item copied from the settlers coming into the area.

PAGE 110

Figure 40. Bone bead fragment (a), bone "fishhook" (b) and shell pendant (c). Figure 41. Antler flakers. 98

PAGE 111

There was also one specimen of modified shell present in the collections. It was a small rectangular pendant with a hole drilled through one end (Figure 40c). The dimensions of the specimen are 1.51 em. x 0.75 em. x 0.15 em. Macrofossils 99 There were 68 seeds and three corn cobs present in the collection from Luster cave. Of these, there were 52 corn kernels; four juniper seeds -three of which included both the inner seed and fleshy covering, one yucca seed, four acorn hulls -one of which contained a seed and one wild onion bulb (Table 8; Figure 42). There were also two roots of Cymopterus umbelliferae present in a bundle of leather and fiber (see the discussion in the perishables section). Because the large majority of the undomesticated seeds recovered during the excavation of Luster cave were unburned, there is a strong possibility that they were not placed there through human action. The domesticated seeds, all of which are corn, are assumed to be the result of human action. There were no soil samples curated from which additional pollen and macrofloral evidence could be gleaned. It is an accepted practice in archaeological studies to reference ethnological plant uses as indicators of possible plant uses in prehistoric times. Ethnographic sources do document that with some plants

PAGE 112

Sample 8583 8584 8585 8586 8587 8588 8589 8590 8591 8592 8593 8594 8595 8596 8597 8598 8598 8599 8600 8601 8641 8690 8702 8750 8776 8814 No. r l.,. !: TABLE 8 MACROFOSSILS FROM LUSTER CAVE Location Burned Area Level Type Whl Frg B 2 Corn Kernels A 3 Corn Kernels A 4 Corn Kernels B 5 Corn Kernels c 2 Corn Kernels c 1 Corn Kernels c 2 Corn Kernels E 1 Corn Kernels F 1 Corn Kernels F 2 Corn Kernels F 3 Corn Kernels F 4 Corn Kernels F 5 Corn Kernels A 1 Corn Cob 1 B 2 Juniper Seed (only) A 1 Juniper Seeds, aril A 1 Unknown Seeds B 4 Juniper Seed, aril 1 F 3 Unknown F 5 Yucca Seed F 5 Acorn Hull A 3 Acorn Hull 1 B 3 Wild Onion Bulb 1 c 2 Corn Cob c 4 Acorn Hull 1 F 2 Corn Cob I) ; . .. ;;-Figure 42. Macrofossils 100 Unburned Whl Frg 3 6 2 3 6 1 2 9 3 5 7 3 2 1 2 2 1 1 1 1 1 1

PAGE 113

101 the historic use was developed and likely continued from the past. The ethnobotanic literature serves only as a guide indicating that the potential forutilization existed in prehistoric times, not as conclusive evidence that the resources were used. There were three corn cob fragments and 52 kernels presently in the collection as compared to nine cobs and an unknown number of kernels (six packets) that were recovered from the excavation. Norton H. Nickerson and Ding Hou analyzed the corn kernels and cobs for Lister during the 1950s (Wormington and Lister 1956:105-106), Table 9 shows his measurements. Because the cobs are fragmented it is difficult to surmise the shape of the cobs, although Nickerson commented that there were equal numbers of tapered, cigar-shaped and straight specimens (Wormington and Lister 1956:105). Nickerson also concluded that the maize from Luster Cave was a thick-kerneled somewhat mixed race, with some Mexican influence based on tapering 12-14-rowed cobs with small shank diameters (Wormington and Lister 1956:105). This is similar to the maize recovered from Turner-Look {Wormington 1955). However, given the presence of 8-rowed straight cobs with high rachis-flaps and a cupule depth of zero, Nickerson also felt there was also Eastern influence (Wormington and Lister 1956:105). Also present were cigar-shapedcobs and some isodiamteric kernels, possibly

PAGE 114

indicating a mixture with, or derivative from, Anasazi maize from the south (Wormington and Lister 1956:105). TABLE 9 CORN COB MEASUREMENTS FROM LUSTER CAVE Row Number: Cupule width: Cupule depth: % of cobs 8-rowed: Height of rachis-flaps: Kernel thickness: % of cobs tapered: Lower glurne width! 3 8 10 12 2 3 Reprinted from Wormington and Lister (1956:105) 1 14 7.5 0.0 33.0 1.6 4.5 33.0 5.2 Zea was an important cultivated food for many prehistoric cultures, as well as for many historic and 102 present cultures. The ethnographic literature indicates a number of ways maize was prepared. The kernels were parched, soaked in water with juniper ash, and boiled to make hominy; dried were ground into a meal, which is used as a staple. Whole ears were husked immediately upon harvesting, then boiled and eaten. Clean husks were saved for smoking and other uses, such as wrapping food. (Cushing 1920; Robbins et 1916; Stevenson 1915; Whiting 1939). There were four juniper berries present in the collection, although only one of these indicated any sign of charring. Juniper berries are an abundant crop and are available throughout the year. The ethnographic literature suggests many uses for both juniper berries and wood in food, medicine, as fuel and as construction

PAGE 115

103 material. The berries were dried for storage; they could then be ground into a meal and used to make mush or cakes. Beverages and medicines were made by steeping the fruits or leaves in hot water. The berries may also have been used to flavor meat (Angell 1981; Harrington 1967). Juniper berries have also been used to make a decoction for use as a medicine (Densmore 1974; Gilmore 1977). Juniper bark was used to line pits where dried fruits were stored (Chamberlin 1964). Whiting (1939:21) maintains that the Hopi ate the berries only during times of famine. In times of acute food shortage some groups would peel off the inner bark of the juniper and chew it (Harrington 1967) 0 Gallagher (1977) notes that juniper berries were an important food for the Apache. The berries were eaten fresh, boiled, pounded to form a kind of bread, or soaked and pounded to make a liquid drink. Goodwin (1935) also notes that juniper berries were a staple for the Apache. Smith (1974) reports that the Northern Utes rubbed juniper berries with a mano to separate the seeds from the pulp. The pulp was then either eaten fresh or dried and ground on a metate. Harrington (1967) notes that Rocky Mountain juniper berries were collected in the late summer or fall and may have been eaten raw or cooked (boiled or roasted) The presence of four acorn hulls -one of which still contained the seed inside -indicate the possibility

PAGE 116

104 that acorns were utilized at Luster Cave. Two of the hulls indicated signs of charring. Acorns are noted to contribute to subsistence wherever Quercus grows around the world. In many areas, it may provide the main source of nutrition. Like pinyon nuts, acorns are unpredictable. In some years the local crop is abundant, and in other years it is almost non-existent (Harrington 1972). Some Shoshonean groups such as the Luiseno ground the acorns into meal. The tannic bitterness .of the acorns was leached out by placing the meal in a loosely woven basket and soaking it in hot water. The acorn meal was used to make mush, bread, pancakes or to thicken soup; sometimes it was mixed with cornmeal. Ground up acorns or possibly only acorn shells are noted to have been used to make a beverage that was used as a substitute for coffee. One unburned yucca seed was present in the collection. The etl'mographic literature indicates that yucca flower stalks and fruits were eaten and stored for winter use. The yucca roots were used for soap and leaf fiber for weaving cordage, sandals, and other items (Woodbury and Zubrow 1979). One wild onion bulb was present in the collection -the stalk appeared to have been burned, while the remainder of the bulb present was charred. Wild onion (Allium) is noted to have been frequently exploited by many Native American groups. Wild onion was utilized as flavoring for stews or meats, boiled and eaten as a

PAGE 117

vegetable, or the juice was used medicinally. The bulbs were also dried and stored for future use (Gilmore 1977; Harrington 1967; Hellson and Gadd 1974; Smith 1974; and Yanovsky 1936). 105 There was one coprolite recovered from Luster Cave. It was not analyzed because it would not have produced a comparable database. It also did not appear to be human. The coprolite was recovered from Area F, Level 2. Human Bone There was one tooth crown present in the collection from Luster Cave. The tooth crown present is unidentifiable beyond being a molar or premolar, as only the enamel of the crown remains (Figure 43). The infant burial, unfortunately, had either been lost or had disintegrated, as the matting associated with the burial was present, but there was no sign of bone within the matting (Figure 44). Both specimens were recovered from Area C, in Levels 2 and 3. 0 1 2 3 4 Scm Figure 43. Tooth crown fragment.

PAGE 118

106 Figure 44. Burial matting. Minerals There were nine specimens of minerals present in the collections from Luster Cave -three lumps of resin, three lumps of limonite, two lumps of yellow ochre, and one lump of red ocher. Resin was often used as a compound which "glued" points and other materials into their respective places. The resin would be heated, then as it hardened, it would lock solid. Ochres have been utilized as paint both historically, and prehistorically, on a variety of including human bodies, ceramics, basketry, and leather. Limonite may have also served as a coloring medium, but may also be present due to the

PAGE 119

107 geologic formations of the area. One specimen, however, does appear to have been smoothed on three sides, as if it had been rubbed against something. Table 10 shows the distribution of the minerals throughout the site. TABLE 10 DISTRIBUTION OF MINERALS IN LUSTER CAVE Area Level Limonite Resin Yellow Ochre Red Ochre A 1 3 2 1 B 1 2 1 1 Radiocarbon Date F 1 2 4 1 1 1 1 One charcoal sample was tested for a radiocarbon date, and produced a date of 3410 130 years B.P. (1850 B.C.+ 210 years, with MASCA correction). The charcoal had been taken from a cis4 that was located in Area B, Level 3. Associated artifacts included two projectile points -a Parowan Basal-Notch point and a Gypsum point -five Type 1 ceramic sherds and two manes -one loaf-shaped fragment and one round, two-sided/ one-handed. The association of ceramics and a Parowan Basal-Notched point in the same level and area as the tested charcoal suggests that either the charcoal that was tested had somehow been contaminated, or later occupants of the cave disturbed the fill, the latter of the two ideas being most probable. This may suggest that there was actually a slight

PAGE 120

108 difference in stratigraphy that went unnoticed during the excavation of the cave. <

PAGE 121

CHAPTER 6 ROTH CAVE ANALYSIS Lithics There were 31 specimens present in the collection. Within the technological analysis, there were 11 types. Table 11 shows the distribution of these types in Roth Cave. The first type is a chert core fragment with cortical and non-cortical surfaces present. There was no visible wear or additional modification present. The second type consisted of chert and chalcedony tertiary bifaces. There were 21 specimens, all of which showed additional modification, and will be further discussed later in the section. There were 21 specimens comprising Type 3. The specimens were tertiary flakes with no visible wear or additional modification. All the specimens in this type are considered debitage. Type 4 consists of two chert choppers. There is slight wear visible, but no additional modification. There were five comprising Type 5. All are sandstone cobbles with extensive visible wear present. There does not appear to be any additional

PAGE 122

110 TABLE 11 DISTRIBUTION OF TECHNOLOGIC TYPES IN ROTH CAVE Area A B Level Surf. 1 2 3 1 2 3 Type 1 1 Type 2 4 5 8 2 1 1 Type 3 9 5 5 2 Type 4 2 Type 5 4 1 Type 6 1 1 Type 7 1 Type 8 1 Type 9 2 Type 10 1 Type 11 1 The specimens within this type are considered to be groundstone and will be discussed later in the section. There were two tertiary flakes of siltstone comprising Type 6. These specimens are considered to be debitage. There was one quartzite chunk comprising Type 7. There were no cortical surfaces There was one tertiary uniface of chert comprising Type 8. It had possible wear along one edge, with no additional modification to the body of the specimen. There were two quartzite cobbles comprising Type 9. Both specimens exhibited cortical surfaces over the entire specimen, as well as extensive-visible wear, much of which appears to be battering of the ends. Both specimens are considered to be hammerstones (Figure 45).

PAGE 123

111 Figure 45. Quartz hammerstone. One quartzite cobble exhibiting extensive wear on its cortical surfaces comprises Type 10. This specimen is considered to be groundstone and will be discussed later in the section. One basalt cobble exhibiting extensive wear comprises Type 11. This specimen is considered to be groundstone and will be discussed later in the section. Grounds tone There were seven specimens of groundstone recovered from Roth Cave. All the groundstone specimens are manes, with a variety of different types present (Table 12). There was one each of loaf-shaped, rectangular with tapered ends and round (Figures 46 and

PAGE 124

112 TABLE 12 DISTRIBUTION OF GROUNDSTONE IN ROTH CAVE Type Area Level Mat. #-Handed #-Sided Loaf-Shaped B 1 1* 1 1 Rectangular Surface 1 1 1 Rectangular Surface 1 1 2 Rect/Tapered B 1 2* 1 2 Round/Tapered Surface 1 1 2 Round/Tapered B 1 3* 1 1 Indet. Fragment Surface 1 ? ? 1 = Sandstone 2 = Basalt 3 = Quartz Figure 46. Loaf-shaped mano.

PAGE 125

113 47). There were two rectangular, as well as one fragment which was difficult to determine shape (Figure 48) All of the manos appear to be one-handed, with two specimens that were utilized on one surface only, and four specimens that were utilized on both sides. All the specimens appeared to have been exposed to charred material, as all the specimens.were blackened on their use surfaces. 0 2 3 4 Scm Figure 47. Round mano. 0 1 2 3 4 Scm Figure 48. Rectangular mano.

PAGE 126

114 Bifaces In addition to the eight projectile points present in the collection, there were three scrapers, one knife base fragment, one drill tip and eight bifaces (Figure 49). The distribution of the projectile points in Roth Cave is noted in Table 13. There was only one point which could not be typed with available typologies. This specimen was side-notched as well as corner-notched (Figure 50a) Both the base and tip have been fragmented. There were five points which compared favorably to the Elko Corner-Notch (Holmer 1978) (Figure SOb, c). As noted previously, the Elko series projectile points are the most plentiful but the least temporally diagnostic of the point types commonly found in the northern Colorado Plateau and the far eastern Great Basin. Points of this type are known to occur after 7600 B.P. {Holmer 1978), possibly persisting into historic times (Fowler et al. 1973:41). There was one specimen present in the collections which compared favorably to the Gypsum point of Archaic origin (Holmer 1978) {Figure 50d) As noted previously, the Gypsum projectile point is the most recent type generally associated with the Archaic stage of the northern Colorado Plateau, dating from approximately 4600 to 1500 B.P. {Holmer 1978).

PAGE 127

115 Figure 49. Bifaces. TABLE 13 DISTRIBUTION OF PROJECTILE POINT TYPES IN ROTH CAVE Area A B Level surf. 1 2 3 1 2 3 Untyped 1 Elko 1 2 1 1 Gypsum 1 Rocker S-N 1

PAGE 128

Figure 50. Projectile points. Untyped point (a), Elko Corner-Notched points (b, c) Gypsum (d) and Rocker Side-Notched point (e). All points actual size. 116 There was one specimen which appeared similar to the Rocker Side-Notched point of Archaic origin (Holmer 1978) (Figure 50e). The material from which this point is made may have been quarried in Unaweep Canyon which is not far from this site (Piontkowski, personal communication 1988). The Rocker Side-notched point type is the most tentative of the several types named during the analysis of Sudden Shelter material (Jennings et al. 1980). It bears some resemblance to a few Elko Side-notched points described by Heizer and Clewlow (1968:78, esp. Fig. 4, o and p), but the tight stratigraphic distribution of these points at Sudden Shelter suggests their importance as time markers. They date from 6800 to 5300 B.P.

PAGE 129

117 Perishables There were 33 specimens present from Roth cave that fall under a perishable category that includes several artifact types and materials not commonly recovered in open sites, but are typical of items recovered in dry caves. Included within this category is basketry, cordage, fur /hair, leather, yucca quids and one specimen of wood. The distribution of these specimens in Roth Cave is noted in Table 14. Two basketry fragments were present in the collection. The first basketry specimen was a wall fragment, recovered from the surface of Roth Cave (Figure 51). It had a 1/2 rod with lateral bundle close-coil foundation and non-interlocking stitches. The dimensions of the specimen are 10.88 em. x 2.83 em. x .51 em. There appears to be a slight residue on the specimen, possibly pitch. The second specimen was a center fragment (Figure 52). It had a 1/2 rod and bundle close-coil foundation. The stitch type was non-interlocking, and accidentally-split stitches were visible. The inner surface appears to have.been worn, most likely from utilization The dimensions of this specimen are 8. 2 em. x .7 em. Both basketry specimens are similar to those that developed in the Great Basin (Adovasio 1971).

PAGE 130

118 TABLE 14 DISTRIBUTION OF PERISHABLE ITEMS IN ROTH CAVE Area A B c Level Surf. 1 2 3 1 2 3 1 2 3 Basketry 2 Worked Wood 1 Unmodified Fur 1 Unmodified Leather 3 Quids 2 2 Yucca Leaves 6 1 1 2 Cordage: Yucca Type 1 1 1 1 1 Yucca Type 2 1 2 1 2 Fur Type 1 1 Figure 51. Basketry fragment.

PAGE 131

Figure 52. Basketry fragment. A rectangular piece of wood was present in the collection, which had been recovered from the surface. This particular specimen had been purposefully cut and smoothed (Figure 53). The dimensions are 9.2 em. x 5.7 ern. x 41 em. 119 Four quids of yucca were present in the collection, collected from Areas A, Level 2 and c, Level 1 (Figure 54) Three unmodified pieces of leather were present in the collection, as well as one specimen of unmodified fur/hair (Figure 55). The fur is white and is possibly lagornorph fur from either the white-tailed jackrabbit (Lepus townsendii) or snowshoe rabbit (Lepus americanus).

PAGE 132

120 Figure 53. Worked wood. Figure 54. Yucca quids.

PAGE 133

121 Figure 55. Unmodified leather. There were eleven specimens of cordage present in the collection. There are three types of cordage that were delineated within the collection. The first type includes four specimens of twostrand, z-twist yucca cordage (Figure 56). Two of the specimens had four twists per centimeter, while one each had three and eight .. All twist angles were 60 degrees. Lengths were 16.7, 13.5, 3.1 and 7.4 em. Thicknesses were 0.2, 0.26, 0.21 and 0.18 em., respectively. The second type included six specimens of twostrand, s-twist yucca cordage. Twists per centimeter included 1, 3, 5 and 6, with two Specimens having a twist angle of 45 degrees and the remainder having a twist angle

PAGE 134

Figure 56. Yucca cordage .. z-twist (a) and s-twist (b -d) of 60 degrees. Lengths ranged from 4.0 em. to 33.1 em.; thicknesses ranged from 0.16 em. to 0.33 em. The final type was comprised of one specimen of two-strand, z-twist fur cordage. There were two twists per centimeter and a twist angle of 60 degrees. The length was 20.5 em.; the thickness was 0.8 em. 122 There were 10 specimens of yucca leaves present in the collection. Six of these specimens are knotted, while the other specimen represents a bundle of yucca leaves (Figures 57 and 58) One in particular may represent a comb of some sort (Jennings 1980). Four of the six specimens which are knotted have been knotted with overhand knots, while the other two are square knots.

PAGE 135

123 Figure 57. Yucca leaf bundle. Worked Bone There were two specimens of worked bone recovered from Roth Cave, both were recovered in Area A -one in Level 1 and the other in Level 3. Both specimens are tips from fragmented awls (Figure 59). Both specimens also exhibit impact fractures of the tip and transverse striations on the exterior surface of the bone. One of the awls only exhibits polish along the tip and edge and was burned as well. The other awl exhibits diagonal striations along the exterior surface. The specimens are 3.8 em. and 4.5 em. in length, respectively.

PAGE 136

124 Figure 58. Yucca knots (actual size).

PAGE 137

125 Figure 59. Bone awl fragments. Macrofossils There were eight seeds and two corn cobs recovered from Roth Cave (Table 15). Five of the eight seeds were corn kernels, while two were yucca and one was a pinyon nut hull (Figure 60). Only the pinyon nut hull appeared burned. There were two complete cobs and five kernels present in the collection. The cob with the stick inserted appears to be tapered on the end opposite the stick; the other cob is fragmented at the end, so it is difficult to judge its shape. It would be premature to conclude the history of the corn present in Roth Cave based on two cobs and five corn kernels, except to note

PAGE 138

Sample No. 7937 7938 8929 8932 8947 TABLE 15 MACROFOSSILS FROM ROTH CAVE Location Burned Area Level Type Wh Fr Surf. Corncob w/stick inserted in end surf. Corn Cob B 1 Corn Kernels B 2 Pinon Nut Hull 1 c 1 Yucca Seed Figure 60. Macrofossils. 126 Unburned Wh Fr 1 1 5 2

PAGE 139

127 its presence. However, Zea has been an important cultivated food for many prehistoric cultures, as well as for many historic and present cultures. The ethnographic literature indicates a number of ways maize was utilized. The kernels were parched, soaked in water with juniper ash and boiled to make hominy; dried kernels were ground into a meal, which was used as a staple. Black corn was used as a dye for basketry and textiles and as a body paint. Whole ears were husked immediately upon harvesting, then boiled and eaten. Clean husks were saved for smoking and other uses, such as wrapping food (Cushing 1920; Robbins et al. 1916; Stevenson 1915; Whiting 1939). There were two complete, unburned specimens of yucca seed present in the collection. Yucca flower stalks and fruits were eaten and were stored for winter use. The yucca roots were used for soap and leaf fiber for weaving cordage, sandals and other items (Woodbury and Zubrow 1979) There. was one charred fragment of a pinyon nut hull present in the collection. Pinus nuts, particularly pinon pine nuts, are noted to have been widely exploited historically. The nuts were harvested in the fall or early winter, and a bumper crop was expected approximately every seven years (Harrington 1967). The nuts were usually collected when the cones opened and the seeds fell to the ground. If the cones were not open, they were roasted to open them. The seeds were eaten raw or

PAGE 140

128 roasted, which preserved them for storage and prepared them for grinding into meal (Colton 1974; Gallagher 1977; Nequatewa 1943; Whiting 1939}. Pine needles have been used to make tea, and the inner bark may have been used as a starvation food. The bark may have also been dried and ground into meal (Harrington 1967}. Pine was also used as a fuel (Robbins et al. 1916). The coprolites were not analyzed because of the fact that there were only two present from the cave, and it would not represent a comparable database. It also did not appear as if the coprolites were human. The coprolites were recovered from Area A, Level 1 and Area C, Level 1. Human Bone There were only three specimens of human bone remaining in the collection at the time of analysis -the child burial was missing, leaving two teeth and a partial maxilla (Figure 61). The two teeth were recovered from Area B, Level 2. One of the teeth was a right 1st or 2nd maxillary molar, worn down to secondary dentine over 3/4 of the tooth, with unequal wear on the lingual side (Figure 60c). There was a very slight formation of calculus near the gumline. The second tooth was an indeterminate maxillary or mandibular premolar, exhibiting wear down to near primary dentine (Figure 60b). There was also a slight formation of calculus near the gumline. It is difficult to assign an age to these two teeth.

PAGE 141

Figure 61. Human bone. Maxilla fragment (a), premolar (b) and maxillary molar (c) The third specimen, a partial.maxilla, was 129 recovered from Area C, Level 1 (Figure 60a). The maxilla was fragmented at the maxillary process with five teeth erupted and evidence of a several more teeth present, but which were not yet erupted. The teeth which were erupted included the upper two left incisors, the canine, and two pre-molars (I 2/x c 1/x M 2/x = 5/x). There was a supernumerary. tooth behind the second incisor, and a molar, both of which had not yet erupted. The first incisor shows signs of wear down to primary dentine, as well as the first premolar. Based on the eruption of the teeth in the maxilla, the maxilla represents a child of approximately five years of age.

PAGE 142

CHAPTER 7 DISCUSSION Therewere several goals important to the research of this project. The first was to review a collection that had been curated for a long period of time to review its condition, as well as to investigate the possibility for tests and procedures that had been developed after the excavation had taken place. Secondly, it was hoped a reanalysis of the artifacts and materials in the collection could add to the present database of the area. The third goal was to determine if the conclusions drawn by Lister and Dick (1952) and Wormington and Lister (1956) remain accurate, given a more extensive collection of typologies for the different artifact types. The collection was in good condition, with less than five percent of the collection missing. The most significant loss of artifacts included the faunal collection, an infant burial from Luster Cave and a child burial from Roth Cave. With the exception of one specimen of basketry none of the artifacts had been treated for long-term preservation. After a review of the collection, it appears that some specimens of cordage and basketry might benefit from treatment for preservation. Unfortunately, there were no field notes available for

PAGE 143

131 perusal, and only two rough sketch maps of Luster cave to look at. All the excavation information was gleaned from Lister's explanations in the published literature (Lister and Dick 1952; Wormington and Lister 1956). The only material that had been curated in the collection that could be utilized for additional testing was charcoal, although location and artifactual associations were sketchy, at best. There were no materials available for flotation, pollen analysis or dendrochronology. There were two small pieces of obsidian present in the collection, although the amount was not enough for sourcing or hydration. Neither Luster or Roth Cave had a large amount of diagnostic material, although there sufficient numbers of projectile points, ceramics and basketry fragments present to provide a rough relative chronological framework. One radiocarbon test was run, with a result of 1850 + 210 years. There was no material available to utilize other absolute dating methods. Table 16 shows the basic areal relationships of the diagnostic material of Luster Cave, although there were no comprehensive field notes with which to compare close associations of artifacts from area to area and level to level. The caves are not large and there do not appear to be any feature differences between areas and levels.

PAGE 144

TABLE 16 DISTRIBUTION OF DIAGNOSTIC ARTIFACTS IN LUSTER CAVE A B C D E F Area Level 1 3 4 1 2 3 5 1 2 3 4 1 1 1 2 3 4 5 Corn X X X X X X X X X X X X' X Unc. Brownware X Fremont Bsktry X X Proj. Points: Elko X Pinto X X Gypsum X X X Desert S-N X X Nawthis S-N X X Uinta S-N X Parowan B-N X X C-14 Sample X TABLE 17 PISTRIBUTION OF DIAGNOSTIC ARTIFACTS IN ROTH CAVE Area A B Level Surface 1 2 3 1 2 3 Corn X X Projectile Points: Elko X X X X Gypsum X Rocker S-N X Basketry X 132

PAGE 145

133 Corn was recovered from all levels and areas of Luster Cave. It was associated with nearly all the diagnostic artifacts recovered from the cave as well. The two corn cobs from Roth Cave were recovered from the surface, while the corn kernels were all recovered from Area B, Level 1 (Table 17). Corn is known to have been introduced into the Southwest by 1000 B.C. (Minnis 1980), and at least as early as 390 B.C. in southeastern Utah (Richens 1988). There is also recent evidence that corn was available to groups in the Great Basin by 175 B.C. (Richens 1988; Wilde and Newman 1989). This suggests that the corn recovered from both Roth Luster Caves may have been traded for as early as 1000 B.C., and may have even been cultivated following that date. Unfortunately, there was no corn associated with the C14 sample taken from the cist in Area B, Level 3 of. Luster Cave, precluding an associational date for the corn. There are no absolute dates for Roth Cave, precluding a further definition of the time span for the use of the corn from Roth Cave. There is a clear division of diagnostic points present in Luster Cave. Both Archaic and Fremont points are represented, although they are associated in only one area of the cave. All the diagnostic points recovered from Roth Cave are associated with the Archaic stage culture. Two Pinto Shoulderless points were recovered from Area C, Levels 1 and 2. There are two Nawthis

PAGE 146

134 Side-Notched points, one Uinta Side-Notched point and a Pesert Side-Notched point associated with the Pinto Shoulderless points by levels. Only a Parowan Basal-Notched point was associated with the Pinto Shoulderlf3SS points in Area c, although the Parowan Basal-Notched point was in Level 4, as opposed to Levels 1 and 2. Because the Pinto points were recovered from the uppermost levels, the possibility exists that these points were curated by later groups. An additional problem which exists is the association of two Parowan Basal-Notched points with an Elko and Gypsum point in Area B, Level 3. There are two solutions to this problem, the first of which is the curation by later groups of Archaic points. The second solution is the possibility that there was a slight change in stratigraphy that went unnoticed during the excavation of the cave. The first solution might be supported by the fact that there was a fragment of Fremont-like basketry and sherds of Uncompahgre Brownware that were also recovered in Area F, Levels 3 and 4. The second solution is supported by the radiocarbon result of 1850 B.C. 210 years, as well as by the fact that two additional Gypsum points were recovered in Level 5 of Areas B and F, which are next to one another. Without more extensive field notes to describe the relationships of the artifacts to one another within the fill of each section, all that can be concluded is the possibility of two distinct occupations of Luster Cave, the first of

PAGE 147

135 which was probably during the Archaic time period. The second occupation was mo.st likely during the early protohistoric, given the association of Ute-like pottery and Formative projectile points. There is greater evidence supporting a single occupation at Roth Cave. All the diagnostic points recovered from Roth Cave are of Archaic origin. The basketry present is similar to that developed during the Archaic period in the Great Basin (Adovasio 1971). There is also evidence to support the presence of corn in the Southwest and surrounding areas by 1000 (Minnis 1980). The presence of a Rocker Side-Notched in association with Elko points and a Gypsum point suggest the possible curation of the older point. The paucity of cultural material recovered from both caves suggests that neither cave was utilized as a residence on a long-term basis. The apparent lack of stratigraphy in Roth Cave suggests that the cave was used over a short period of time, perhaps as a base camp from which the occupants pursued game and gathered local fruits and vegetables. There were only a small number of macrofossils recovered from Roth cave, making any claim as to whether the occupants of the cave were nomadic or sedentary difficult. The presence of corn itself cannot suggest anything further than the fact that corn was in the area. There was only one burned pinon nut hull, the presence of which might suggest that it had been gathered

PAGE 148

136 by the occupants of the cave. The presence of the hull might suggest a fall occupation, but without additional evidence it is difficult to be sure. The presence of a child burial -although it is missing from the collection at the present time -suggests that it may have been a family group occupying the cave rather than a hunting party, although there was not an extensive number of artifacts recovered that could be associated with a female's toolkit. Luster Cave may have alsobeen utilized as a base for pursuing game and gathering the local vegetation, although the presence of disparate projectile points types may suggest it had been utilized at separate times. There were several charred macrofossils recovered, including a wild onion bulb, acorn hulls, a juniper seed and a corn cob, suggesting a fall occupation. The presence of an infant burial suggests a family grouping as opposed to an all-male hunting party. The addition of ceramics and fragments of wooden hunting artifacts also suggest that there were both females and males occupying the cave. The presence.of the infant and child burials from Luster and Roth caves suggest a family grouping occupying the caves as opposed to a hunting party. The paucity of cultural material recovered would suggest brief occupations of the caves. There was only one specimen from Roth Cave that might be construed as having come from out of the area -this was a projectile point constructed

PAGE 149

137 of material similar to that from the northwestern section of Colorado (Piontkowski, personal communication 1988). The presence of projectile points associated with the Fremont suggest the group occupying Luster Cave may have had ties to that area. There are no specimens from either cave that might be considered exotic. Given the presence of the burials, the paucity of cultural material and a lack of exotics, this suggests that the occupants of both caves were small family groups of local origin who used the caves as base camps. It might even be suggested that the occupations of Luster cave, while disparate in time, were similar in activity.

PAGE 150

CHAPTER 8 SUMMARY AND CONCLUSIONS The research in this paper was undertaken in order to answer the question of whether a reanalysis of the artifacts from an archaeological site which had been excavated in the past might add more information to its database. While many of the scientific techniques which have been developed over the past 35 years were beyond the scope of the materials curated in the collection, the projectile point, ceramic and basketry typologies which have been developed in that same time span aided in a refinement of culture affiliation for the two caves. For example, Lister (Wormington and Lister 1956:125) felt that Roth cave antedated Luster cave, and was most likely occupied during the Basketmaker III time period. However, the projectile points from Roth Cave compared more favorably with projectile point typologies for the Archaic time period (Holmer 1978). It may also be possible that a small change in stratigraphy in Luster Cave was overlooked during the first excavation, as the projectile points from Luster Cave represent two distinct occupations. The occupations suggested are from the Archaic time period and early proto-historic. A second occupation is also

PAGE 151

139 supported py the presence of ceramics that are associated with Ute occupation of the Uncompahgre Plateau (Buckles 1971}. Additional excavation in front of the caves may produce evidence as to the cultural affiliation of the occupants of the cave, as well as provide soil samples from which more precise palynological and macrofloral information may be gleaned. Additional excavation might also provide vital artifactual associations with materials that may be absolutely-dated. Reanalysis of archaeological materials that have been recovered from excavations undertaken in the past is an important aspect of archaeological studies. While specimens may be missing, or the fieldnotes are incomplete, the additional information gleaned from the artifacts themselves is an important addition to the database of an area. This project, while it was not able to fully utilize new scientific techniques such as pollen analysis, obsidian hydration and sourcing, or dendrochronology, was able to refine the time periods during which Roth and Luster Caves were occupied utilizing projectile point, ceramic and basketry typologies.

PAGE 152

REFERENCES CITED Adams, Karen R., and Robert E. Gasser 1980 Plant Microfossils from Archaeological Sites: Research Considerations, and Sampling Techniques and Approaches. The Kiva 45(4): 293-300. Adovasio, J.M. 1988 Coiled Basketry and Cordage from Lakeside Cave (42B0385), Utah. Paper presented at the 21st Great Basin Conference. Park City. 1980 Fremont: An Artifactual Perspective. Antiquities Section Selected 7(16):35-40. 1977 Basketry Technology: A Guide to Identification and Analysis. Aldine Publishing Company: Chicago. 1975 Fremont Basketry. Tebiwa 17:67-76. 1971 Some Comments on the Relationship of Great Basin Textiles to Textiles from the Southwest. University of oregon Anthropological Papers 1:103-108. Eugene. 1970 Chipped Stone Artifacts. IN Median Village and Fremont Culture Regional Variation by John P. Marwitt. University of Utah Anthropological Papers 95. Salt Lake City. Aikens, Melvin c. 1966 Fremont-Promontory-Plains Relationships. University of Utah Anthropological Papers 82. Salt Lake City. Angell, Madeline 1981 A Field Guide to Berries and Berrylike Fruits. The Bobbs-Merrill Company, Inc.: New York. Baugh, Timothy G., and Frank W. Eddy 1987 Rethinking Apachean Ceramics: The 1985 Southern Athapaskans Ceramics Conference. American Antiquity 52:793-799. Berry, Michaels., and Claudia H. Berry 1976 Ari Archaeological Survey of the White River Area, Northeastern Utah. Antiquities Section Selected Papers Vol. 2, No. 4. Salt Lake City.

PAGE 153

Breternitz, David A., Arthur H. Rohn, and Elizabeth A. Morris 1974 Prehistoric Ceramics of the Museum of Northern Arizona, Northern Arizona Society of Flagstaff. Brothwell, D.R. Mesa Verde Region. Ceramic Series 5. Science and Art: 141 1981 Digging Up Bones: The Excavation, Treatment and Study of Human Skeletal Remains. Cornell University Press: Ithaca. Buckles, William G. 1988 Discussion. In Archaeology of the Eastern Ute: A Symposium, edited by Paul R. Nickens. CCPA Occasional Papers 1:218-232. 1971 The Uncompahgre Complex: Historic Ute Archaeology and Prehistoric Archaeology on the Uncompahgre Plateau in West-Central Colorado. Ph.D. dissertation, University of Colorado. 1968 The Archaeology of Colorado: Part III, Archaeology in Colorado: Historic Tribes. southwestern Lore 34(3):53-67. Cassells, E. Steve 1983 The Archaeology of Colorado. Johnson Books. : Boulder. Chamberlin, Ralph V. 1964 The Ethnobotany of the Gosiute Indians of Utah. American Anthropological Association Memoirs 2:329-405. Clewlow, C.W., Jr. 1967 Time and Space Relationships of Some Great Basin Projectile Point Types. University of California Archaeological Survey Reports 70:141-150. Berkeley. Colton, Harold S. 1974 Hopi History and Ethnobotany. In Hopi Indians. New York: Garland Publishing Inc. Cosgrove, c. Burton 1947 caves of the Upper Gila and Hueco Areas in New Mexico and Texas. Papers of the Peabody Museum of American Archaeology and Ethnology Crane, 1977 24(2) :1-181. Cathy J. A Comparison of Archaeological Sites on the Uncompahgre Plateau and Adjacent Area. M.A. Thesis, Eastern New Mexico University, Portales.

PAGE 154

cushing, Frank Hamilton 1'920 Zuni Breadstuff. Indian Notes and Monographs. Vol. 8. Museum of the American Indian, Heye Foundation, New York. Dalley, Gardiner F. 1976 Swallow Shelter and Associated Sites. University of Utah Anthropological Papers 96. Salt Lake City. 142 1970a Worked Bone and Antler. In Median Village and Fremont Culture Regional Variation by John P. Marwitt. University of Utah Anthropological Papers 95. Salt Lake City. 1970b Artifacts of Wood. In Hogup Cave, edited c. Melvin Aikens. University of Utah Anthropological Papers 93. Salt Lake City. Densmore, Frances 1974 How Indians Use Wild Plants for Food. Medicine and Crafts. Dover Publications, Inc.: New York. Emery, Irene 1966 The Primary Structures of Fabrics. The Textile Museum: Washington, D.C. Fowler, Don D., David B. Madsen and Eugene M. Hattori 1973 Prehistory of Southeastern Nevada. Desert Research Institute Publications in the Social Sciences 6. Reno. French, David H. 1971 The Ethnobotany of the Umbelliferae. In The Biology and Chemistry of the Urnbe1liferae, edited by V.H. Heywood. Academic Press: New York. Frison, George c. 1971 ShoshoniAntelope Procurement in the Upper Green River Basin, Wyoming. Plains Anthropologist 16:54 pt. I. Lincoln. Gallagher, Marsha V. 1977 Contemporary Ethnobotany Among the Apache of the Clarkdale, Arizona Area Coconino and Prescott National Forests. USDA Forest Service Archaeological Report 14. Gilmore, Melvin R. 1977 Uses of Plants by the Indians of the Missouri __ River Region. Reprint, University of Nebraska Press: Lincoln. Originally published 1919, Bureau of American Ethnology: Washington, D.C.

PAGE 155

143 John D. Gooding, 19. 80 The Durango South Project: An Archaeological Salvage of TwO Late Basketmaker III Sites in the Durango District. Anthropological Papers of the University of Arizona 34. Goodwin, Greenville 1935 The Social Divisions and Economic Life of the Western Apache. Affierican Anthropologist 36:55-64. Harrell, Lynn L. 1983 A Study of the Bone and Antler Tools from Hog Creek Reservoir, Bosque and Coryell Counties, Texas. M.A. thesis on file, University of Texas, Austin. Harrington, H.D. 1972 Seed Storage and Longevity. III, edited by T. Kozlowski, Academic Press: New York. Seed Biology Volume pp. 145240. 1967 Edible Native Plants of the Rocky Mountains. Albuquerque: University of New Mexico Press. Heizer, Robert F., and M.A. Baumhoff 1961 The Archaeology of Two Sites at Eastgate, Churchhill County; Nevada: I. Wagon Jack Shelter. University of California Anthropological Records 20(4). Berkeley. Heizer, Robert F. _and R. Berger 1970 Radiocarbon Age of the Gypsum Culture. University of California Archaeological Research Facility Contributions 7:13-18. Berkeley. Heizer, Robert F., and C.W. Clewlow, Jr. 1968 Projectile Points from Site NV-CH-15, Churchill County, Nevada . University of California Archaeological Survey Reports 71:59-88. Berkeley. Hellson, John c., and Morgan Gadd 1974 Ethnobotany of the Blackfoot Indians. National Museums of canada: Ottawa. Hester, Thomas R., and Robert F. Heizer 1973 Review and Discussion of Great Basin Projectile Points: Form and Chronology. University of California Archaeological Research Facility: Berkeley. Holmer, Richard N. 1986 Common Projectile Points of the Intermountain West. In Anthropology of the Desert West

PAGE 156

144 Essays in Honor of Jesse D. Jennings, edited by C.J. Condie and D.D. Fowler. pp. 89-115. University of Utah Anthropological Papers 110:89-115. Salt Lake City. 1978 A Mathematical Typology for Archaic Projectile Points of the Eastern Great Basin. Ph.D. dissertation, University of Utah. Salt City. Holmer, Richard N., and Dennis G. Weder 1980 Common Post-Archaic Projectile Points of the Fremont Area. In Fremont Perspectives, edited by David B. Madsen. Antiquities Section Selected Papers 16. Salt Lake City. Hurley, W.M. 1979 Prehistoric Cordage: Identification of Impressions of Pottery. Aldine Publishing Company: Chicago. Hurst, C.T. 1940 Preliminary Work in Tabeguache Cave -1939. Southwestern Lore 6(2):4-18. 1943 Preliminary Work in Tabeguache cave II. Southwestern Lore 9(1) :10-16. 1944 1943 in cave II, Tabeguache Canyon, Montrose County, Colorado. Southwestern Lore 10(1):2-14. 1945 Completion of Excavation of Tabeguache Cave II. Southwestern Lore 11(1) :7-12. 1946 The 1945 Tabeguache Expedition. Southwestern 12(1) :7-16. 1947 Excavation of Dolores Cave -1946. Southwestern 12(1) :8-17. 1948 The Cottonwood Expedition, 1947 -A Cave and a Pueblo Site. Southwestern Lore 14(1) :4-19. Huscher, Harold 1939 Influence of the Drainage Pattern of the Uncompahgre Plateau on the Movements of Primitive Peoples. Southwestern Lore 5(2}. Huscher, Betty, and Harold Huscher 1943 The Hogan Builders of Colorado. Southwestern 9(2):1-92. Janetski, Joel c. 1980 Wood and Reed Artifacts. IN cowboy cave, edited

PAGE 157

by J.D. Jennings. University of Utah Anthropological Papers 104:75-95. Jennings, Jesse D. 145 1957 Danger Cave. American Antiquity 23 (2, pt. 2), Memoir 14. 1980 Cowboy Cave. Papers 104. University of Utah Anthropological Salt Lake City. Keepax, Carole 1977 Contamination of Archaeological Deposits by Seeds of Modern Origin with Particular Reference to the Use of Flotation. Journal of Archaeological Sciences 4:221-229. Kehoe, Thomas F. 1966 The Small Side-notched Point System of the Northern Plains. American Antiquity 31 (6): 827-841. Lambert, Marjorie F., and Richard Ambler 1961 A Survey and Excavation of Caves in Hidalgo County, New Mexico. The School of American Research Monograph 25. Santa Fe. Lindsay, A.J., Jr., J.R. Ambler, M.A. Stein and P.M. Hobler 1968 Survey and Excavations North and East of Navajo Mountain 19 59-19 6 2 ...,B""'u""l..,l..,..e,_,t ... i=n........,o..,f"--'t"""h=e...__.M=u=s"""e=-um=-o=f Northern Arizona 45, Glen Canyon Series 8. Flagstaff. Lister, Robert H., and Herbert W. Dick 1952 Archaeology of the Glade Park Area: A Progress Report. Southwestern Lore 17(4): 69-92. Madsen, Rex E. 1977 Prehistoric Ceramics of the Fremont. Museum of Northern Arizona Ceramic Series 6. Flagstaff. Minnis, 1981 Paul E. Seeds in Archaeological Interpretive Problems. 46(1) :143-152. Sites: Sources and Some Affierican Antiquity 1980 Domesticating Plants and People in the Greater Southwest. Paper presented at the Advanced Seminar on the Origins of Plant Husbandry in North America, School of American Research, Santa Fe. Morris, Earl H. and Robert F. .Burgh 1954 Basketmaker II Sites near Durango,Colorado. Carnegie Institution of Washington 601.

PAGE 158

Nequatewa, Edmund 1943 Some Hopi Recipes for the Preparation of Wild Plant Foods. Plateau 16(1) :18-20. O'Connell, J.P. 146 1967 Elko Eared/Elko Corner Notched Projectile Points -Points as Time Markers in the Great Basin. In University of California Archaeological Survey Reports 70:129-140. Pierson, Lloyd M. 1980 Cultural Resource Summary of the East Central Portion of Moab District. Cultural Resource Series 10. Bureau of Land Management, Salt Lake City. Reed, Alan D. 1984 West Central ColOrado Prehistoric Context: Regional Research Design. State Historical Society of Colorado, Denver. Reed, Alan D., and Ronald E. Kainer 1978 The Tarnaron Site, 5LP 326. Southwestern Lore 44:1-47. Richens, Lane D. 1988 Late Archaic/Early Formative Adaptations in the Eastern Great Basin. Paper given at the 21st Great Basin Conference. Park City. Robbins, W.W., J.P. Harrington, and Barbara Freire-Marreco 1916 Ethnobotany of the Tewa Indians. Bureau of Affierican Ethnology Bulletin 55. Rohn, Arthur H. 1971 Wetherill Mesa Excavations, Mug House. Mesa Verde National Park, Colorado. Schroeder, Albert H. 1964 The cultural Position of caves and Pueblo Sites. 29 (4): 77-79. Linda J. Hurst's Tabeguache Southwestern Lore Scott, 1983 Manual for Sampling. Denver. Pollen, Phytolith and Macrofloral Ms on file, PaleoResearch Laboratory, Shutler, Richard, Jr. 1967 Cultural Chronology in Southern Nevada. In Pleistocene Studies in Southern Nevada, Part 6, edited by H. Marie Wormington and D. Ellis. Nevada State Museum Anthropolgical Papers 13:305-308. Reno.

PAGE 159

Smith, 1974 Anne M Ethnography of the Northern Utes. Papers in Anthropology 1.7. Albuquerque: Museum of New Mexico Press. Stevenson, Matilda Coxe 1915 Ethnobotany of the Zuni Indians. Thirtieth Annual Report of the Bureau of American Ethnology. Government Printing Office: Washington, D.C. Underhill, Ruth 147 1944 Pueblo Crafts. u.s. Department of the Interior, Bureau of Indian Affairs: Washington, D.C. Whiting, Alfred F. 1939 Ethnobotany of the Hopi. Museum of Northern Arizina Bulletin 15. Wilde, James D. 1986 Lithic Analysis. IN Clear Creek Canyon Archaeological Project: Proposals for Data Analysis, edited by Joel c. Janetski and James D. Wilde. Brigham Young University Museum of Peoples and Cultures Technical Series 85-98:48-72. Wilde, James D. and Deborah E. Newman 1989 Early Corn and the Late Archaic in the Eastern Great Basin. Affierican Anthropologist in press. Woodbury, 1932 George and Edna Woodbury The Archaeological Survey of Paradox Valley and Adjacent Country in Western Montrose county, Colorado, 1931. Colorado Magazine 9:2-21. Woodbury, R. and E. Zubrow 1979 Agricultural Beginnings, 2000 BC -AD 500. IN Handbook of North American Indians, Vol. 9: Southwest. W.C. Sturtevant, general editor and A. Ortiz, volume editor, pp. 43 60. Smithsonian Institution, Washington, D.C. Wormington, H. Marie 1955 A Reappraisal of the Fremont culture with a Summary of the Archaeology of the Northern Periphery. Proceedings of the Denver Museum of Natural History 1. Denver. Wormington, H. Marie and Robert H. Lister 1956 Archaeological Investigations on the Uncompahgre Plateau in West Central Colorado. Proceeding of the Denver Museum of Natural History 2. Denver.

PAGE 160

148 Yanovsky, E. 1936 Food Plants of the North American Indians. Department of Agriculture Miscellaneous Publication 237:1-83.

PAGE 161

APPENDIX A ARTIFACT CATALOG

PAGE 162

150 FS # FIELD # PRESENT AREA LEVEL MUSEUM. CATALOG DESC LUSTER 8572.01-.05 LBO A 3 Yucca Quids 8577.01-.03 L86 A 4 Yucca Quid 8578.01-.02 L122 B 4 Yucca Quid 8579.01-.02 L191 c 3 Yucca Quid 8580.01-.07 L229,L248 E 1 Yucca Quid 8581.01-.02 L259 F 1 Yucca QLiid 8582.01-.02 L268 F 2 Yucca Quid 8605.01-.02 L70 A 3 Reed Sect. 8608.01-.03 L60, L58 A 2 Arw Sft Frag 8617.01-.02 L174 c 2 Worked Sticks 8619.01-.02 L203 c 3 Stick Frag. 8620.01-.02 L226 E 1 Bark-wrap Stick 8625.01-.05 L285 F 4 Potsherds 8626.01-.19 L286 F 4 Fur Cord 8628.01-.05 L289 F 4 Yucca Cord 8629.01-.03 L290 F 5 Sinew Cord 8631.01-.02 L294 F 5 Leather 8632.01-.16 L295 F 4 Uti 1 Flake 8634.01-.03 L298 F 5 Uti 1 Flake 8635.01-.03 L299 F C' -' Bifacial if e 8637.01-.02 L301 F 5 Rubbing Stones 8638.01-.04 L302 NO F 5 Animal Teeth 8641.01-.02 L305 F 5 Acorn Shell 8642.01-.03 L309 F 5 Yucca Cord 8644.01-.03 L2 SUF:F Fur g( Yucca Cordage 8648.01-.12 L10 A 1 Utilized Flakes 8653.01-.07 L17 NO A 1 Animal Bone Frag 8655.01-.07 L20 A 1 Cordage 8657.01-.02 L9 A 1 Con ve::Base Knife 8658.01-.02 L26 A 2 Flake 8660.01-.09 L29 A 2 Ycca Cord 8663.01-.03 L35 B 1 Ut. Flake 8673.01-.03 L49 B 2 Flake Knife 8674.01-.02 L47 B 2 Bifacial Knife 8679.01-.03 L54 A 2 Flake Scraper 8693.01-.02 L77 A 3 Ut Flake 8695.01-.02 L79 A 3 Yucca Knot L85 A 3 Ut Flake 8700.01-.03 L92 A 4 Ut Flake 8707.01-.13 L98 B 3 Ut Flake 8712.01-.05 L105 B 3 Ceramic Sherd 8716.01-.02 L108 B 4 Ut Flake 8724.01-.08 L123 1 NO B 5 Ceramic Sherd 8725.01-.02 L127 B 5 Proj Point 8729.01-.02 L133 NO c 1 Faunal Teeth 8734.01-.04 L148 NO c 2 Faunal Teeth 8737.01-.02 L152 c 2 Ut Flake 8742.01-.02 L162 c 2 Chopping Tool

PAGE 163

151 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8743.01-.03 L163 c 2 Ut Flake 8744.02-.03 L165 NO c 2 Animal Teeth 8746.01-.03 L167 c 2 YLlCCa Cord 8752.01-.02 L124 B 5 YLlcca Knot 8755.01-.02 L178 c 3 Bone Bead 8758.01-.05 L182 c 3 Ut Flake 8770.01-.06 L201 c 3 YLlCCa Cord 8772.01-.04 L206 c 4 Ut Flake 8773.01-.03 L207 NO c 4 FaLtna 1 Teeth 8779.01-.04 L216 NO [I 1 FaLmal Teeth 8784.01-.02 L225 E 1 Cord 8785.01-.02 L228 E 1 FL!r Cord 8790.01-.02 L237 E 1 Pl no-cnv: : Scraper 8793.01-.03 L240 E 1 Bifacial Knife 8795.01-.02 L242 NO E 1 Deer Dew Claws 8796.01-.03 L243 NO E 1 FaLtnal Teeth 8801.01-.02 L250 E 1 FLtr Cord 8803.01-.02 L253 NO F 1 FaLmal Teeth 8804.01-.02 L254 F 1 Bifac ial Knife Frag 8805.01-.05 L255 F 1 Ut Flake 8807.01-.05 L260 F 1 YLlcca Cord 8810.01-.04 L263 F 2 YLlcCa Fiber 8818.01-.17 L273 F 3 Ut Flake 8822.01-.04 L277 F 3 FLlr Cord 8824.01-.02 L281 F 3 YLlCCa Cord 8828.01-.03 L312 SURF Ut Flake 8831.01-.02 L13? A 1 Charcoal 8832.01-.06 L114 B 3 Charcoal 8573.01 L215 [I 1 YLlcca OLlid 8574.61 L157 c 1 YLICCa QLlid 8575.01 L291 F 4 YLlCCa QLiid 8576.01 L125 B 5 YLICC. a QLiid 8583.01 L44 B 2 3 Corn Kernels 8584 L69,L82 A 3 6 Corn Kernels 8585 L88 A 4 2 Corn Kernels 8586 L126 B 5 3 Corn Kernels 8587 L146 c 2 6 Corn Kernels 8588 L156 c 1 1 Corn Kernel 8589 L175 c 2 2 Corn Kernels 8590 L224 E 1 9 Corn Kernels 8591 L257 F 1 3 Corn Kernels 8592 L265 F 2 5 Corn Kernels 8593 L283 F 3 7 Corn s 8594 L292 F 4 3 Corn Kernels 8595 L306 F 5 2 Corn Kernels 8596.01 L18 A 1 Corn Cob Frag 8597.01 L44 B 2 JLiniper Seed 8598.01 L18 A 1 JLtniper Seed

PAGE 164

152 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8598.02 ?? Seed 8599.01 L113 B 4 Juniper Seed 8600.01 L280 F 3 Seed Pod 8601.01 L307 F C" -' Yucca Seed 8602.02 L175 c 2 Reed 8603.01 L23 A 1 Reed Sect. 8604.01 L59 A 2 Reed Frag. 8606.01 L198 c 3 Reed Sect. 8607.01 L21 A 1 Arw Sft. Frag. 8609.01 L139 c 2 Arw Sft Frag 8609.02 L144 c 2 Arw Sft Frag 8610.01 L22 A 1 Cut Wood 8611.01 L57 A 2 Cut Wood 8612.01 L81 A 3 CLlt Yucca 8613.01 L83 A 3 CLlt Wood 8614.01 L138 c 2 Twig 8615.01 L140 c 2 Shaped Stick 8616.01 L141 c 2 Cut Stick 8616.02 L145 c 2 Cut Stick 8616.03 L145 c 2 Cut Stick 8618.01 L202 c 3 Crushed Stick 8621.01 L227 E 1 Cut Stick 8622.01 L258 F 1 Worked Wood 8622.02 L258 NO F 1 Worked Wood 8623 .o1 L287 F 4 Cut Wood 8624.01 L308 F 5 CLlt Wood 8627.01 L288 F 4 Limonite 8630.01 L293 F 5 Obs. Flake 8633.01 L296 F 4 Bifac ial Knife 8636.01 L300 F 5 Proj. Point 8639.01 L303 F 5 Antler Tine 8640.01 L304 F C" ._I Bone Awl 8643.01 L1 SURF Proj Point Fragment 8645.01 L3 A 1 Oval Scraper 8646.01 L4 A 1 Conve:: l
PAGE 165

153 FS # FIE:LD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8661.01 L30 B 1 Proj Point 8661 .02 L31 B 1 Proj Point 8661 .03 L32 B 1 Proj Point 8661.04 L33 B 1 Proj Point 8662.01 L34 B 1 Stone l
PAGE 166

154 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8705.01 L96 B 3 Yucca Cord 8706.01 L97 B 3 Leather Frag 8708.01 L99 B 3 Dr i 11 Point 8709.01 L100 B 3 End Scraper 8710.01 Ll:01 B 3 Proj Point 8710.02 L102 B 3 Proj Point 8710.03 L103 B 3 Proj Point 8710.04 L103 NO B 3 F'roj Point 8710.05 L103 B 3 Proj Point 8711.01 L104 B 3 Bi facial Knife 8713.01 L106 B 4 Proj Point 8713.02 L109 B 4 F'roj Point 8714.01 L107 B 4 Bifac ial t
PAGE 167

155 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8754.01 L177 c 3 Gaming Piece 8756.01 L179 c 3 Mano 8757.01 L181 c 3 Hammers tone 8759.01 L183 c 3 Dr i 11 8760.01 L184 c 3 Knife Blade Base 8761.01 L185 c 3 Proj Point 8762.01 L186 c 3 Cedar Bark Mat 8763.01 L188 c 3 Basket Frag 8764.01 L189 c 3 Weaving Frag 8765.01 L193 c 3 Bndle Lthr Fbr 8766.01 L194 c 3 Bndle Twstd Cdr Brk 8767.01 L196 c 3 Bundle Fiber 8768.01 L197 c 3 Human Hair Cord 8769.01 L200 c 3 Bndl snw 8( ycca cord 8771.01 L204 c 4 Proj Point 8771.02 L205 c 4 Proj Point 8774.01 L208 c 4 Gaming Piece .8775.01 L209 c 4 Yucca Cord Frag 8776.01 L210 c 4 Acorn 8777.01 L212 [I 1 Proj Point 8777.02 L213 [I 1 F'roj Point 8778.01 L214 [I 1 Stone Flake 8780.01 L218 E 1 Ut Flake 8781.01 L219 E 1 Stone Knife 8781 .02 L220 E 1 Stone if e 8782.01 L221 E 1 Proj Point 8782.02 L222. E 1 Proj Point 8783.01 L223 NO E 1 An Rib w/cut marks 8786.01 L230 E 1 Fur Frag 8787.01 L231 [I 1 Bone Bead Frag 8788.01 L233 E 2 Proj Point 8788.02 L234 E 2 Proj Point 8789.01 L236 E 1 Bifacial Knife 8791.01 L232 E 2 Hammers tone 8792.01 L239 E 1 Hammers tone Frag 8794.01 L241 E 1 Proj Point 8797.01 L244 E 1 Bone Awl 8798.01 L? E 1 Gaming Piece 8799.01 L247 E 1 Yucca Quid 8800.01 L249 E 1 Yucca Cord 8802.01 L252 E 1 Leather Frag 8806.01 L256 F 1 Proj Point 8808.01 L261 F 1 Limonite 8809.01 L262 F 1 Sinew Cord 8811.01 L264 F 2 Basketry Frag 8812.01 L266 F 2 Resin 8813.01 L267 F 2 Yucca Cord 8814.01 L269 F 2 Corn Cob Frag

PAGE 168

156 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC LUSTER 8815.01 L270 F 2 Charcoal Frag 8816.01 L? F 2 BLmdl e Human Hair 8817.01 L272 F 2 Limonite 8819.01 L274 F 3 Bifacial Knife 8820.01 L275 F 3 Proj Point 8821.01 L276 F 3 .Bund 1 e Human Hair 8823.01 L278 F 3 Skin Frag 8825.01 L282 F 3 Leather Thong 8826.01 L297 F 4 Proj Point 8827.01 L311 SURF Mano 8829.01 L313 SURF Hammers tone 8830.01 L315 SURF Ceramic Sh.erd 8833.01 L? F 3 Coprolites ROTH 7930.01-.10 LUSTER 1 NO Al SURF Ceramic Sherds 7932.01-.05 LUSTER SURF Ut. Flakes 7934.01-.04 ROTH SURF Ut. Flakes 7935.01-.05 ABOVE ROTH SPUR SITE Ceramic Sherds 7936.01-.02 ROTH SURF Coiled Basket Frag 7938.01-.09 ROTH SURF Corn Cobs Kernels 7939.01-.03 ROTH 1 NO SURF Fiber Cord 7940.01-.04 ROTH SURF 1-Hand Manos 8897.01-.08 R3 R4 1 NO SURF Yucca Cord 8899.01-.(17 R6 A 1 Ut. Flake 8900.01-.03 R7 A 1 Proj. Point 8901.01-.02 R8 A 1 Blade Fragment 8903.01-.03 RiO A 1 Yucca Cord 8908.01-.06 R16 A 2 Ut. Flake 8909.01-.04 R17 A 2 Bifac ial Knife 8915.01-.02 R28 NO A 2 Bone Fragments 8918.01-.02 R31 SURF Hammers tone 8919.01-.08 R32 SURF Ut. Flake 8922.01-.02 R35 A 3 Ut. Flake 8928.01-.02 R42 B 1 Cord 8933.(11-.02 R48 B 2 Human Teeth 8937.01-.02 R52 NO B 3 Animal Bone Frags 8940.01-.04 R56 c 1 YLICCa Cord 8941.01-.02 R57 c 1 YLtcca Quid 8942.01-.02 R58 c 1 Leather Fragment 7931.01 LUSTER SURF Bone Gaming Piece 7933.01 ROTH SURF F'roj. Point 7937.1 ROTH Corn Cob 7941.1 ROTH NO SURF Metate, Basin-type 8896.01 R2 SURF Ut. Flake 8898.01 R5 SURF Wooden Plaque 8902.01 R9 A 1 Dr i 11 Tip

PAGE 169

157 FS # FIELD # PRESENT AREA LEVEL MUSEUM CATALOG DESC ROTH 8904.01 R11 A 1 Yucca Leaf Bundle 8905.01 R12 A 1 13 Corn Kernels 8906.01 R14 A 1 Bone Awl 8907.01 R15 NO A 1 Faunal Tooth 8910.01 R18 A 2 Proj. Point 8910.02 R19 A 2 Proj. Point 8910.03 R20 A 2 F'roj. Point 8910.04 R21 A 2 Proj. Point 8910.05 R21 A 2 F'roj Point 8911.01 R23 A 2 Feces 8912.01 R24 A 2 Resin 8913.01 R25 A 2 Yucca QLiid 8914.01 R26 A 2 Yucca Leaf Bundle 8916.01 R29 A 2 Hair 8917.01 R30 NO c FLOOR Child Burial 8920.01 R33 A 3 Proj. Point 8921.01 R34 A 3 Bifacial Knife 8923.01 R37 A 3 Bone Awl Fragment 8924.01 R38 B 1 Hammerstone 8925.01 R39 B 1 Mano Fragment 8926.01 R40 B 1 Bifac ial Knife 8927.01 R41 B 1 End Scraper 8929.01 R43 B 1 5 Corn Kernels 8930.01 R44 B 1 Charcoal 8931.01 R45 B 2 Charcoal 8932.01 R47 B 2 Pinyon NLit 8934.01 R49 B 2 Mano Fragment 8935.01 R50 B 3 Ut. Flake 8936.01 R51 B 3 Proj. Point 8938.01 R53 NO B 3 Tooth 8939.01 R54 B 3 Charcoal 8943.01 R59 c 1 Yucca Cord 8944.01 R60 c 1 Feces 8945.01 R61 c 1 HL1man Mandible Frag 8946.01 R62 c 1 Charcoal 8947.01 R63 c 1 Yucca Seeds 8948.01 R64 NO c 1 Reed Fragment 8949.01 R65 c 16" Charcoal 8950.01 R66 NO c FLOOR Horn Core



PAGE 1

DYNAMIC TEST SIMULATIONS FOR DETERMINING GROUND QUADRILATERAL FAULT IMPEDANCE MEASUREMENTS by Karl M. Smith BSc., Colorado School ofMines, 1986 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Science Electrical Engineering Spring 2000

PAGE 2

This thesis for the Master of Science degree by Karl M. Smith has been approved 4-(-00 Date

PAGE 3

Smith, Karl M. (M.S., Electrical Engineering) Dynamic Test Simulations for Determining Ground Quadrilateral Fault Impedance Measurements Thesis directed by Professor Pankaj K. Sen ABSTRACT Complex algorithms in todays microprocessor based relays are capable of determining ground quadrilateral fault impedance, fault location and direction by unique measurement calculations from settings and inputs. This complexity, however, increases the challenge for developing test procedures for these relays. This is in part due to the limitations of the test equipment and the high volume of complex calculations required for dynamic testing. Therefore it is necessary to provide advanced mathematical models and simulations of the power system to demonstrate how the fundamentals are applied when implementing test procedures that conform to more rigorous standards than outlined in the test manuals To automate the calculation process, a spreadsheet ha been developed to determine ground quadrilateral measurements by conveniently entering in relay settings and inputs. This abstract accurately represents the content of the candidate's thesis. I recommend its iii

PAGE 4

ACKNOWLEDGEMENTS I would like to thank New Century Energies for the vast amount of financial and technical resources made available to me for the completion of this thesis and my education. In particular, I would like express my appreciation for the following individuals; George Laughlin, a system protection engineer at New Century Energies, for his guidance and providing engineering support. Chris Gallegos, coworker, for providing the necessary test data for the extensive numerical examples throughout this thesis. Ken Behren, SEL developer, for discussion on reactance element error. Karl Zimmerman, Brad Heilman and Bill Flemming, SEL application engineers, for discussion on directional elements. Edi VonEngeln, classmate, for help in the development of Excel spreadsheets. Carlene Stroh, coworker, for help in using Microsoft Word commands. Mickey Pitt, a friend and fellow student who has taken many classes with me, for his time and commitment in helping me to complete the graduate program. Thanks for your encouragement. Dean Nester, my supervisor, for his support and encouragement. Finally, I would like to thank Dr. P.K. Sen, who has taught the majority of my courses, for helping me realize my potential as an electrical power systems engineer.

PAGE 5

CONTENTS F .. 1gures ................... .... ................................................................................ vn Tables ....................................................................................................... viii Abbreviations and tenns ............................................................................ ix Chapter 1. Introduction ...................................................................................... 1 2. Theory and fundaillentals ............................ .......................... .......... 4 2.1 Reactance measurement .............. ..................................................... 4 2.2 Resistance measurement ................................................................. 11 2.3 Negative sequence impedance measurement. ....... ............. ........... 14 3. Modeling a nonhomogeneous system ....... ..................................... 19 3.1 Reactance measurement calculation for the .......................... ......... 21 zone 1 fault location with 5 n fault resistance (T=O) 3.2 Reactance measurement calculation for the .................................... 22 zone 1 fault location with 5 n fault resistance (T=5) 3.3 Resistance measurement calculation for the ................................... 22 zone 1 fault location (true resistance from the fault study= 50) 3.4 Negative sequence impedance measurement threshold .................. 24 calculations for a bolted fault at the zone 1 fault location 3.5 Negative sequence impedance measurement threshold .............. ... 25 calculations for 5 n fault resistance at the zone 1 fault location v

PAGE 6

4. Dynamic test simulations ................................................................ 27 5. Results and Conclusion ................................................................... 31 Appendices A. Dynamic test simulation spreadsheet (T=O) ................................... 33 B. Dynamic test simulation spreadsheet (T=5) .................................... 34 C. Residual current compensation factor derivation ........................... 35 D. Zero sequence current compensation factor derivation .................. 38 E. Effect of transmission line conductor changes on KN and Ko . ........... 39 References ................ ................................................................................ 40 vi

PAGE 7

FIGURES Figure 1.1 Ground Quadrilateral characteristics ................................................ 1 2.1 Reactance reach dimension .............................................................. 4 2.2 Sequence network for SLG fault ...................................................... 5 2.3 Measurement error for a two source system .................................... 7 2.4 Voltage phasors for a resistive fault in a nonhomogeneous system. 8 2.5 Zero sequence network for a single line to ground fault.. ................ 9 2.6 Resistance reach dimensions ................... ...................................... 11 2. 7 Negative sequence network for a SLG fault .................................. 16 2.8 Z2 characteristics, MTA = 90 ........................................................ 17 2.9 Z2 charcteristics, MTA <90 ............................................................ 18 3.1 Nonhomogeneous system-345 KV line from Rifle to Craig ....... 19 4.1 PROTEST 3 Z plot ......................................................................... 27 4.2 Reactance measurement error (T=O) .............................................. 29 4.3 Reactance measurement error (T=5) .............................................. 30 C.l Sequence network for SLG fault ................................................... 35 vii

PAGE 8

TABLES Table 3.1 Relay inputs from fault data (Rr= 0 Q) ........................................ 20 3.2 Relay inputs from fault data (Rr= 5 Q) ........................................ 20 Vlll

PAGE 9

ABBREVIATIONS AND TERMS ZuPositive sequence line impedance at the zone 1 fault location. ZuNegative sequence line impedance at the zone 1 fault location. Z1..0Zero sequence line impedance at the zone 1 fault location. ZtANGPositive sequence line impedance angle MTA-Maximum torque angle RF Fault resistance IF-Total fault current IR Residual relay current ( 3Io) KN-Residual current compensation factor Zero sequence current compensation factor T-Nonhomogeneous compensation correction factor angle ALT-Reactance element error term Z2M-Negative sequence impedance measurement ( a) z2F, z2R-Negative sequence impedance for forward, reverse faults Z 2 rr, Z 2RT Negative sequence impedance forward and reverse thresholds SEL-Schweitzer Engineering Laboratories IX

PAGE 10

1. Introduction Ground quadrilateral distance elements provide more fault resistance coverage than ground mho units when protecting transmission lines from ground faults. Fault resistance coverage in a ground quadrilateral is controlled by the resistive reach setting which defines the side boundaries of the quadrilateral in the RX plane. The upper boundary of the quadrilateral is controlled by the reactive reach setting. In recent versions of microprocessor based relays, the reactive reach setting is defined in terms of the positive sequence line impedance in the R-X plane instead of the X-axis. The lower boundary defining direction, resides in the negative sequence impedance plane and is controlled by negative sequence impedance threshold settings [ 1]. jX R FORWARD DIRECTION -----....... ... __ REVERSE DIRECTION Figure 1.1: Ground quadrilateral dimensions. 1

PAGE 11

When power systems have unequal source and line impedance angles, the fault resistance measurement will have a reactive component that is introduced into the reactance measurement that will cause the relay to overreach or underreach. This type of system is defined as nonhomogeneous. The difference between the source and line impedance angles determines the value of the nonhomogeneous compensation setting required for correcting the reactance measurement error To determine the reactance measurement error, a fault study needs to be conducted to model a non-homogeneous system for increasing increments of fault resistance at the reactive reach point. The apparent impedances for each value of fault resistance then need to be converted to current compensated reactance measurements that would occur in the relay. The difference between the current compensated reactance measurement and the reactive reach boundary from relay test results defines the reactance measurement error. Test results for the reactive reach boundary can be obtained by running a computer generated impedance search plot. Apparent impedances from the search plot are then converted to current compensated reactance measurements to confirm that these measurements are within close tolerances to the reactive reach setting boundary. To simulate the relays conversion of reactance and resistance measurements, a spread sheet will be created that expedites the calculation for these measurements from relay settings, voltages and currents. This method of dynamic testing is necessary since 2

PAGE 12

there is no way to determine the reactance measurement error from a computer generated search plot since the relay is underreaching. 3

PAGE 13

2. Theory and Fundamentals 2.1 Reactance Measurement The ground reactance measurement calculates the positive sequence impedance of the line at the fault location [ 1]. The reactive reach dimension of the ground quadrilateral can be represented by lines sloped at the positive sequence line angle that originate from the R-axis (figure 2.1 ). The distance of these lines depends on the zone of protection which designates the percentage of the line length expressed in terms of positive sequence impedance. jX R Figure 2.1: Reactance reach dimension 4

PAGE 14

If the reactance reach measurement is within the reactive reach defined by the top portion of the the reactive reach element will assert. For a trip output to occur, however, all ground quadrilateral elements must assert. To derive the expression for the reactance measurement, a voltage drop expression needs to be written from a single line to ground sequence network (figure 2.2), that includes a fault resistance term. RELAY LOCATION FAULT LOCATION Figure 2.2: Sequence network for a SLG fault (Phase A) 5

PAGE 15

From the A phase relay voltage equation written in terms of the symmetrical impedances, and, 31o = IA+ Is+ Ic The voltage drop equation can be expressed in terms of the residual current compensation factor KN. VA= Zu(IA + KN) (2.1) (2.2) Derivations of residual and zero sequence current compensation factors including effects of conductor line changes are shown in appendices C-E. Adding a fault resistance term to the voltage drop equation, the expression now yields, VA= Zu(IA + KNIR) + IFRF (2.3) Multiplying both sides of the equation by the complex conjugate ofiRLT to make the voltage across the fault resistance appear real yields, V A(IRLT )* = Zu (lA + KNIR)(IRLT )* + IFRF (IRLT )* where T is the nonhomogeneuos compensation factor correction angle. Taking the imaginary component of both sides of the equation, the expression now yields, Im[V A(IRLT )*] = Im[Zu(IA + KNIR)(IRLT )*] + Im[IFRF (IRLT )*] where, 6

PAGE 16

Multiply both sides of the equation by IZul, IZul = Im[V A(IRLT )*] Im[ Zu/IZul (IA + KNIR)(IRLT )*] where IZul is the reactive reach referenced to the positive sequence line angle. Therefore Zu/!Zul = lLZtANG Substituting, (2.4) If the T setting is not used for nonhomogeneous systems, an error term will be introduced into the reactance measurement since the voltage across the fault resistance now appears reactive (figure 2.3) [3]. Figure 2.3: Measurement error for a two source system 7

PAGE 17

The voltage drop from equation 2.3, now has an error term in the relay voltage (figure 2.4). jiX underreach ALT-REACTANCE ERROR ALT overreach IR Figure 2.4: Voltage phasors for a resistive fault in a nonhomogeneous system 8

PAGE 18

The reactance element error term, ALT, can be determined from zero sequence currents and impedances shown in the zero sequence network (figure 2.5). I Zso ZRO I so IRO I ... - I ZLO (\..IF=Iso+IRo Figure 2.5: Zero sequence network for a SLG fault There are two methods of calculating ALT. One method requires a theoretical approach while the other takes a more practical approach. ALT = ALT = IFo Iso (2.5) (2.6) The reactance measurement error, 11IZLII, can now be expressed in terms of ALT from the following equation; 9

PAGE 19

AIZul = (2.7) The reactance measurement error AIZul can easily be obtained from fault studies and relay settings to determine the compensation required to prevent the relay from overreaching or underreaching. 10

PAGE 20

2.2 Resistance Measurement The resistance measurement calculates fault resistance for line to ground faults from relay inputs and settings. The resistance reach dimension is defined by any point on the reactive reach line that intersects the origin in the R-X plane to the side boundaries of the quadrilateral. [ 1] jX Figure 2.6: Resistance reach dimensions 11 R

PAGE 21

If the resistance measurement is within the resistive reach boundary of the quadrilateral, the resistive reach element will assert. Like the reactance measurement derivation, deriving the resistance measurement can also be accomplished by writing a voltage drop expression (equation 2.3) from the sequence network found in figure 2.2. [2] Multiplying both sides of the equation by the complex conjugate of ZLI (IA + KNIR) to make the positive sequence line impedance appear real, Taking the imaginary component of both sides of the equation Im[V A (ZLICIA + KNIR))*] = Im[ZLICIA + KNIR)(ZLICIA + KNIR))*] + .......... where, Solving for Rp, R Im[V A (ILZIANGCIA + KNIR))*] F Im [lp(lLZIANG(IA + KNIR))*] .. ........ Im [lpRp(ZLICIA + KNIR))*] Since IF includes current contributions from both ends of the line, fault current seen from the relay, 3Io, is approximated in terms of zero and negative sequence currents to eliminate the effects ofload current, i.e., 12

PAGE 22

for a SLG fault, Therefore, 3Io = 3/2( 12 + Io ) Substituting, RF = Im[VA (1LZtANG0A + KNIR))*] (2.8) Im [3/2(I2+Io){lLZtANG(IA + KNIR))*] From the derivation it is evident that Rr will only be half the actual fault resistance if source impedances are equal for both fault contributions. To calculate the true fault resistance, the ratio of total zero sequence fault current to relay fault current, 3Io, must be calculated [3,4] The measured fault resistance then needs to be multiplied by this ratio to determine the true value of fault resistance. (2.9) 13

PAGE 23

2.3 Negative Sequence Impedance Measurement The negative sequence impedance measurement was designed to overcome several limitations imposed by the negative sequence directional relay [5]. These limitations occur whenever the negative sequence voltage is minimized by strong negative sequence sources behind the relay or when the negative sequence current is minimized due to fault resistance. The negative sequence impedance measurement is derived by compensating the negative sequence voltage in the traditional negative sequence torque equation shown below ([3,4]); where MTA is typically set to the positive sequence impedance line angle, LZtANG Expressing equation 2.10 in complex conjugate form, V2 is compensated by rewriting the expression as V2-a.LZ1ANGI2 where a = the negative sequence impedance measurement. Substituting, The term will now increase V2 for forward faults and decreaseh for reverse faults. Setting the torque equal to zero to define the balance point, 14 (2.10) (2.11) (2.12)

PAGE 24

Rearranging, 0 = Re[ Vz(hLZtANG)*] -(ZzM)Re[(lzLZtANG) (hLZtANG)*] Solving for ZzM, From the identity, the expression now yields, Re[ V,(l,LZJ ANal*] (2.13) IIi For the relay to sense fault direction, ZzM must be compared to negative sequence impedance threshold quantities. These threshold quantities are a function of ZzM and the negative sequence voltage and currents for faults at the relay point. To determine threshold quantities, a fault study must be conducted to arrive at Zz in the forward (ZzF) and reverse (Z2 R) directions at the relay point. The negative sequence network for ground faults shown in figure 2. 7 indicates Is2 for forward faults and IR2 for reverse faults. 15

PAGE 25

+ R ---1 FORWARD FAULT Figure 2.7: Negative sequence network for a SLG fault From V 2 and I2 at the relay point, ZzF = -V2 -Zs2 ZzR = -Vz = Zu+ ZR2 -IR2 To obtain Z2F and Z2R from the fault study, several contingencies are required (2.14) (2.15) to obtain the strongest negative sequence source behind the relay. If this method is not convenient, a simplified approach is to set Z2 F equal to half of the positive sequence line impedance and add 0.1 to arrive at ZzR From z2F and z2R threshold settings, the negative sequence impedance threshold quantities can now be calculated [ 1]. 16

PAGE 26

THRESHOLD CONDTIONS THRESHOLD QUANTITIES ZzF < 0; ZzFr = 0.75( ZzF)0.25 I ZzM I (2.16) ZzF > 0; ZzFr = 1.25(ZzF)0.25 I ZzM I (2.17) ZzR> 0; ZzRT = 0.75(ZzR) + 0.251 ZzM I (2.18) ZzR ZzRT a reverse fault is declared. The characteristics for the threshold quantities for a 90 degree MTA are shown in the Z2 plane in figure 2.8. I Z2PLANE I REVERSE FAULT ll!///1/1 FORWARD FAULT Figure 2.8: Z2 characteristics, MTA = 90 17 Z2RT Rz Z2FT

PAGE 27

For MTA's other than 90 degrees, the negative sequence impedance threshold characteristics are perpendicular to the positive sequence line impedance which is typically set at the MTA. This is required to maintain the same threshold characteristics since positive sequence line and fault impedances remain unchanged when mapped in the Zz plane. I Z2PLANE I Figure 2.9: Z2 characteristics, MTA <90 18 REVERSE FAULT FORWARD FAULT Z2RT R2 Z2FT

PAGE 28

3. Modeling a Nonhomogeneous System To verify the accuracy of ground quadrilateral measurements, a fault study was conducted using "ASPEN" software on the 345 kV 81.7 mile PSCOTri State transmission line from Rifle to Craig substations to model a nonhomogeneous system. MODELING A NONHOMOGENEOUS SYSTEM RIFLE SUBSTATION T = L3Io-LIF=S 345 KV LS 3Io RELAY PT 3000: CT; 300:1 ....... f-:=...-.... 1 I SEL 321 I :,. ZLI = 4.1 L86 n secondary IF l CRAIG SUBSTATION 345KVLO ... RF=50 primary --Figure 3 .I: Nonhomogeneous system-345 KV line from Rifle to Craig 19

PAGE 29

Fault data for zone 1 faults ( 85 % of the line ) were obtained for increasing 1.0 ohm increments of fault resistance to determine how fault resistance effects the reactance measurement (appendix A). Relay inputs for 0 ohm and 5 ohm fault resistances are shown in tables 3.1 and 3.2. (S d ) econ ary ( Secondary ) VOLTAGE(V) CURRENT(A) v A= 37.52L-3.2 5.53L-81.6 Vs = 63.87L-117.1 .027L96.1 Vc = 65.0Lll6.6 .027L96.1 Table 3.1 :Relay inputs from fault data (Rr= 0 0) ( Secondary ) ( Secondary ) VOLTAGE(V) CURRENT(A) VA= 40.8L-ll.l 5 24L-69.2 Vs = 63.7L-117.7 .026L108.5 Vc = 65.9Lll6.6 .026L108.5 Table 3.2 : Relay inputs from fault data (RF =50) For bolted faults in a nonhomogeneous system the reactance measurement is unaffected. However if a fault resistance is introduced, the reactance measurement will have an error term. From fault data and relay settings the reactance measurement without compensation can be calculated and compared to the reactive reach of the relay. The difference from this comparison is the reactance measurement error. When the compensation factor T, is used in the calculation, the reactance measurement should be within close tolerances to the reactive reach setting for a zone 1 fault. 20

PAGE 30

3.1 Reactance Measurement Calculation for the Zone 1 Fault Location with 5 n Fault Resistance (T = 0) IZul = Im[V A(IRLT )*] lm[lLZrANG (IA+KNIR)(IRLT )*) IZul = Im[V A0AL-9(1LT ))*] Im[ 1 LZrANG(IA(l + KN)(IRLT )*) From the following identity, the equation can be simplified to IZul = Im[V A(IIAIL-9(1LT ))*] Im[lLZtANGIIAI2(1 + KN)(lLT )*] = Im[V A(1L-9(1LT ))*] IIAI Im[lLZtANGCl+ KN)(lLT )*] = lm[40 8L-11.1C1L-69.2 ClLO ))*] 5.24 Im[1L86(1+ 75L-19)(1LO )*] = 3.92 n 21

PAGE 31

3.2 Reactance Measurement Calculation for the Zone 1 Fault Location with 5 n Fault Resistance ( T=5 ) IZul = Im[40.8L-11.1(1L-69.2 C1L5 ))*] 5.24 Im[1L86(1+ .75L-19)(1L5 )* = 3.77 n It is also important to model the system since the true fault resistance from the fault study does not equal the measured fault resistance. The measured fault resistance depends only on the relays contribution of zero sequence fault current. The true fault resistance measurement will depend on the ratio of total zero sequence current to the relay zero sequence fault current. From this ratio, the true fault resistance can be calculated and checked with the resistance from the fault study. 3.3 Resistance Measurement Calculation For the Zone 1 Fault Location (True Resistance From Fault Study = 5 Q) R Im[V A (1 LZIANG(IA + KNIR))*] FJm [3/2(IA2 + IAo)(lLZIANG(IA + KNIR))*] Since, IR, I IA I= IAL-8 IA2 = lAo = IA/3 then, 22

PAGE 32

Rp = Im[V A (IIAI L-8(1 + KN)ILZIANo)*] Im [IA(IA(I+KN) ILZIANG)*] From the following identity, The equation can be simplified, Rp = Im[V A (I L-8(1 + KN)ILZIANG)*J IIAI Im [((I+KN) ILZIANG)*] RF = Im[40.8L-Il.I (1 L-69.2 (1 + .75L-19)IL86)*] 5.24 Im [((I+ .75L-I9) 1L86)*] = 2.6940 To calculate the true fault resistance the following equation is used Rp(true) = 2.694 x 2885.4 I4I8.3 = 5.48 n The fault resistance measurement is now consistent with the resistance from the .. fault study when this approximation method is used. To verify fault direction for the system model, negative sequence impedance measurements and threshold quantities should be calculated from the fault study at the zone I fault location. Directional quantities should be calculated for both bolted faults and faults with resistance at the resistive reach setting to understand how the threshold characteristics are affected by fault resistance in the plane. 23

PAGE 33

3.4 Negative Sequence Impedance Measurement Threshold Calculations for a Bolted Fault at the Zone 1 Location From table 3.1 relay inputs, = 113(37.52L-3.2 + 1L240(63.87 L-117.1) + 1L120(65.0L116.6)) = 11.04L-175 V = 1/3(5.53L-81.6 + 1L240(0.27L96.1) + 1L120(0.27L96.0)) = 1.93L-81.7 A 312 = IA = 5.79 A (exceeds typical3I2 supervision setting of0.5A) = Re[ 11.04L-175(1.93(L-81.7(1L86))*] 11.9312 =5.72 n = 1.25(Z2F)0.251Z2ml = 1.25(3.6)0.25(5.72) = 3.07 n Z2m < Z2Fr ; forward fault declared 24

PAGE 34

3.5 Negative Sequence Impedance Measurement Threshold Calculations for 5 n Fault Resistance at the zone 1 Location From table 3.2 relay inputs, = l/3(40.8L-11.1 + 1L240(63.7L-117.7) + 1L120(65;9Ll16.6)) = 10.53L-163.4 V = 1/3(5 24L-69.2 + 1L240(0.026L108.5) + 1L120(0.026L108.5)) = .1.76L-69.2 A 312 = IA = 5.28 A (exceeds typical3I 2 supervision setting of0.5A) Ref 10.53L-163.4(1.76L-69.2ClL86))*] 11.7112 =-5.98 n Z 2Fr = 1.25(Z2F)0.25IZ2ml = 1.25(3.6)0.25(5.98) =3.0.0. Z2M < z2Ff ; forward fault declared 25

PAGE 35

The margin between Z2M and Z2Fr is slightly greater for the 5 ohm fault resistance, therefore, the fault will appear more in the forward direction when fault resistance is introduced. Creating a region between the threshold settings will prevent high fault resistances from causing a fault behind the relay to appear in the forward direction. For this example, a 5 ohm fault resistance does not significantly affect the directional capabilities of this model. 26

PAGE 36

4. Dynamic Test Simulations To establish the reactance measurement base line for a dynamic test simulation, an impedance search plot test was conducted for a zone 1 ground quadrilateral in a SEL 321 (Schweitzer Engineering Laboratories ) relay at Rifle substation using "PROTEST 3" software. The search plot calculates values of apparent ground fault impedance by setting the equation in the "PROTEST 3" Z PLOT binary search macro to Z=VII. Apparent impedances are plotted at specified increments of test current angles (figure 4.1 ). -6 -2 2 6 H Figure 4.1: PROTEST 3 Z PLOT [ 6] The voltage and currents from relay test results were then entered into an "Excel" spreadsheet to convert the apparent ground fault impedances to current compensated resistance and reactance measurements that correspond to reach settings of the 27

PAGE 37

ground quadrilateral. These results were then checked for accuracy and plotted. Relay voltage and currents from the fault study were also entered in the spreadsheet. Fault reactance and true resistance measurements were then plotted against the results from the search plot to graphically show reactance measurement error when the T setting was set to zero (figure 4.2 ). When the T setting was changed to 5, the reactance measurement was reduced significantly (figure 4.3) 28

PAGE 38

,..,.------------------------ZL1 (OHMS) FIGURE 4.2: REACTANCE MEASUREMENT ERROR (T=O) SEL 321 ZONE 1 GROUND QUADRILATERAL FOR 345 KV LINE TO CRAIG 5.0 3.93 3.93 ..... 3 93 3.92 3.92 4Jr. . . . 3.0 2.0 1.0 .-------'-----.-------.--------10:9 -----r .. _ .. ___ 1 ----...... 1 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 FAULT RESISTANCE (OHMS) 0\ N

PAGE 39

ZL1 (OHMS) ---------------.......... --.. FIGURE 4.3: REACTANCE MEASUREMENT ERROR (T=S) SEL 321 ZONE 1 GROUND QUADRILATERAL FOR 345 KV LINE TO CRAIG 5 0 3 0 2.0 1 0 . ------0;0--t --,------1 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0 FAULT RESISTANCE (OHMS) L___________________ -------------0 ("f)

PAGE 40

5. Results and Conclusion Reactance measurements for ground faults are set in tenns of the positive sequence line impedance to detennine the fault location. This is convenient since the positive sequence line impedance is directly proportional to the length of the line. However, it is evident from the fault study that the reactance measurement will not be exactly proportional to the length of the line if the system is nonhomogeneous and includes fault resistance. A system is nonhomogeneous when the total zero sequence fault current angle is not equal to relay zero sequence fault current angle This angular difference is defined as the nonhomogeneous compensation correction factor angle setting T. The reactance measurement line from figure 4.2 tilts at this angle when referenced to the reactive reach line generated from the impedance search plot. If the correction factor T is applied, the reactance measurement error is now proportional to the length of the line for ground faults. This method of testing requires a high volume of complex calculations. By creating a spreadsheet, protection engineers and technicians can conveniently enter relay settings, impedance search plot points and fault data to augment testing and verify relay accuracy. 31

PAGE 41

Appendices 32

PAGE 42

Appendix A Dynamic Test SimulationReactance Measurement Error (T = 0) SEL 321 Zone 1 Ground Quadrilateral345 KV line to Craig IMPEDANCE SEARCH PLOT TEST PROTEST RESULTS Ia angle Ia magnitude 0 20 40 48.2 60 80 90 105 120 135 165 SElTINGS kOM kOA Z1ANG T RG XG FAULT STUDY 5.98 5.16 3.74 3 .04 3.64 4 3 4.44 4.38 4 12 5 .16 6 .12 0.75 -19 86 5 5 4.1 ASPEN RESULTS Va@ o deg 30 T 0 Ia angie I a magnitude Va angle Va magni!llde Fault Z .S1.6 5.53 .J.2 37.52 5.49 5.1 37. 94 1 76 .5 5.44 .0 9 38.5 2 -74 5.38 .a.5 39.18 3 -71.6 5 .31 .9 39 95 4 -
PAGE 43

Appendix B Dynamic Test Simulation-Reactance Measurement Error (T = 5) SEL 321 Zone 1 Ground Quadrilateral345 KV line to Craig IMPEDANCE SEARCH PLOT TEST PROTEST RESULTS Ia (calc) l'ro ERROR R (calc) X (calc) Region Ia angie Ia magnitude Va@O deg R canst Xconst 0 5 98 30 6 000 0 33% 5 017 -0.265 20 5.16 5.197 0.71% 5.036 0.912 40 3.74 3 767 0 72% 5 037 2.789 48.2 3 .04 3.038 0 07% 4.996 4 094 60 3.64 3.472 4 84% 2.587 4 092 80 4.3 4.094 5.03% -0.265 4.085 90 4.44 4.222 5 15% .452 4 080 10. 5 4.38 4.174 4 93% -3.195 4.088 120 4.12 3.841 7.25'ro -4.996 4 000 135 5.16 -5.155 0.11'ro -4.995 2.699 165 6.12 .129 0 15% 5.008 1 016 SETTINGS kOM 0.75 1+KO 1+k01 kOA Magnitude Angle Z1ANG 86 1.726493 .13054 T 5 RG 5 XG 4 1 FAULT STUDY ASPEN RESULTS T Ia (calc) l'ro ERROR R (calc) X (calc) Region I a angle Ia magnitude Vaangle Va magnitude Fautt Z 5.53 ..:3.2 37.52 0 5 080 8.86% -0. 064 3 941 -79 5.49 1 37.94 I 5.001 9.78% 0.489 3 908 5 5.44 -6. 9 38.5 2 4.914 10. 71'r. 1.041 3 875 -74 5.38 -8.5 39.18 3 4.818 11.67% 1.596 3 842 6 5.31 -9. 9 39.95 4 4.718 12.56% 2.143 3 .811 -69. 2 5.24 .1 40.8 5 4 610 13. 67'l'. 2.694 J .n4 34

PAGE 44

Appendix C C.l Derivation of Residual Current Compensation Factor KN [7 ,8] RELAY LOCATION FAULT LOCATION Figure C.l: sequence network for a SLG fault (Z1=Z2, phase A) 35

PAGE 45

V1F =CV2F+VoF) V1F = V A-I1 (Zu) V 2F = l2(Zu) VoF = -Io(ZLO) Substituting equations C.2, C.3 and C.4 into C. I yields, VAI1(Zu) = h(Zu)+Io(ZLO) For an "a" phase to ground fault, Is= Ic = 0 Therefore, II = I2 = Io = I,J3 Then, VA= II(Zu+Zu+ZLO) Substituting Zu = Zu, the expression now yields, VA= II(Zu)+I2(Zu)+Io(ZLO) Rearranging this expression by substitution from the equation, 3Io= IA+Is+Ic, yields, VA= I1Zu+I2Zu+(IA +Is +Ic)/3ZLO = Zu[ II+I2+(IA+Is+lc)Zw/3Zu] since, IA = I1+I2+Io, or I1 = IA-Irlo 36 (C. I) (C.2) (C.3) (C.4)

PAGE 46

Then, = Zu[ IA-lo+3Io(ZL0/3Zu)] = Zu[ IA+3Io(-1/3+Zw/3Zu)] where ZLO-Zu is the residual curr3Zu ent compensation factor KN 3Zu Z1.0-Zu KN = 3Zu and IA+Ia+Ic is the residual current, Substituting, Dividing by the sum of A phase current, IA and compensated residual current, Ioc, the expression now yields = Zu(IA +KNIR) = IA+KNIR Zu KN can also be written, KN = Ko -1 3 where Ko is the zero sequence compensation factor Z1.0/Zu 37

PAGE 47

AppendixD Derivation of Zero Sequence Current Compensation Factor, Ko [7 ,8] Ko is derived in a similar manner to KN. From the SLG fault sequence network in Appendix C, Rearranging, where Z10/ZLI is the zero sequence current compensation factor Ko Ko=Zw/ZLI and Io is the zero sequence current substituting and dividing by the compensated A phase current, lAc, in terms of symmetrical quantities, the expression now yields, v A = ZLI(Il+l2+Kolo) = ZLI lAc l1+I2+Kolo 38

PAGE 48

AppendixE Effect of Transmission Line Conductor Changes on KN and Ko [9] From the following equations, Zu = jroJ.to[l/4+1n(D/R)] 21t Zw = jroJ.to[l ] 21t J.lO = 41t xl0-7 ro = 21tf where, D = Geometric mean distance of phase conductors DN = Average distance of neutral conductor to phase conductors R = Geometric mean radius of phase conductors RN = Geometric mean radius neutral conductor And, K Zw-Zu N3Zu Ko=Zw/Zu It is evident that changes in conductor spacing, radius and bundling result in impedance changes that are not proportional to the length of the line and that vary from positive sequence to negative sequence. Therefore KN and Ko do not remain constant over the length of the transmission line. Settings on the SEL 321 address this problem by assigning a dedicated residual compensation factor for zone 1. 39

PAGE 49

References [ 1 ] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8 ] [ 9 ] Schweitzer Engineering Laboratories Inc. SEL 321 Instruction Manual Pullman, Washington, March 20, 1998 edition S.E. Zocholl. Three Phase Circuit Analysis and the Mysterious KO factor Presented before the 22nd annual Western Protective Relay Conference Spokane, Washington, October 24-26, 1995 J. Mooney, P.E., J. Peer. Application Guidelines for Ground Fault Protection Presented before the 241h annual Western Protective Relay Conference Spokane, Washington, Oct 21-23, 1997 E.O. Schweitzer, III, J. Roberts. Distance relay element design Presented before the 46th annual conference for protective relay engineers Texas A&M University, April12-14, 1993 B. Flemming. Negative Sequence hnpedance Directional Element Presented before the 101h Annual Protest User Group Meeting Pasadena, California, February 24-26, 1998 C. Gallegos. Provided 321 test results using the "Protest 3" relay testing software for the SEL 321 installation at Rifle substation, 1999 GE Multilin-Distance Relays Fundamentals GER-3966 GEC Measurements. Protective Relays Application Guide Published GEC Measurements, Stafford England, 1987 0. Elgerd. Electric Energy Systems Theory-an Introduction, 2nd edition McGraw-Hill, Inc. 1982 40