United States
              Environmental Protection
              Agency
                 Office of
                 Toxic Substances
                 Washington, D.C. 20460
EPA 560/5-90-009
April 1990
&EPA
             Toxic Substances
TEXTILE DYE WEIGHING
MONITORING STUDY

-------
                                   EPA  560/5-90-009
                                   April  1990
TEXTILE DYE WEIGHING MONITORING STUDY
    Exposure Evaluation Division
  Economics and Technology Division
     Office of Toxic Substances
U.S. Environmental Protection Agency
         401 M Street, S.W.
       Washington,  D.C.   20460

-------
     This document has been reviewed and approved for publication by
the Office of Toxic Substances, Office of Pesticides and Toxic
Substances,  U.S. Environmental Protection Agency.   The use of trade
names or commercial products does not constitute Agency endorsement
or recommendation for use.
ii

-------
                              CONTENTS

LIST OF ACRONYMS	    viii

AUTHORS AND CONTRIBUTORS	      ix

ACKNOWLEDGMENT  	    xiii

EXECUTIVE SUMMARY
   I.    INTRODUCTION	      xv
   II.   OBJECTIVES  OF THE STUDY   	     xvi
   III.  APPROACH	     xvi
   IV.   RESULTS   	   xviii
   V.    CONCLUSIONS  	     xix

Chapter 1
INTRODUCTION
   I.    BACKGROUND	     1-1
   II.   STUDY OBJECTIVES  	     1-3
   III.  OVERVIEW OF THE REPORT	     1-3

Chapter 2
RESULTS AND CONCLUSIONS
   I.    PRIMARY FINDINGS  	     2-1
   II.   ADDITIONAL  FINDINGS   	     2-5

Chapter 3
QUALITY ASSURANCE PROGRAM
   I.    INTRODUCTION	     3-1
   II.   STATISTICAL/QUALITY CONTROL PLANNING  	     3-1
         A.  Frame Construction  	     3-2
         B.  Pilot Study of the Industry—Mailed-Out  Survey   .     3-3
   III.  QA  PLANNING FOR THE SITE VISITS   	     3-4
         A.  Optimizing the Monitoring  Participation  Rate  .   .     3-4
         B.  Quality  Assurance  for On-Site  Visits   	     3-5
   IV.   QUALITY ASSURANCE FOR THE ANALYTICAL METHODOLOGY  .   .     3-7
   V.    CONSTRUCTION OF THE COMPUTER  DATA BASE	     3-7

Chapter 4
SURVEY DESIGN AND SAMPLE SELECTION
   I.    DEFINITION  OF TARGET  POPULATIONS  	     4-1
   II.   SAMPLE DESIGN   	     4-1
         A.  First-Phase Sample  	     4-1
         B.  Second-Phase Sample  	     4-3
   III.  DYES STUDY  WEIGHTING  SCHEME   	     4-5
         A.  Calculation of Plant-Level Weights   	     4-6
         B.  Calculation of Worker-Level  Weights  	     4-8
                                                                  iii

-------
                              CONTENTS
                             (Continued)

Chapter 5
FIELD SAMPLING PROCEDURES
   I.    IN-PLANT OBSERVATION PROCEDURES   	     5-1
   II.   DYE  COLLECTION PROCEDURES   	     5-2
         A.  Industrial Hygiene  Air  Monitoring 	     5-2
         B.  Dye Bulk Sampling	     5-3

Chapter 6
ANALYTICAL METHODOLOGY
   I.    INTRODUCTION	     6-1
   II.   SCOPE  OF THE METHOD   	     6-1
   III.  DYE  PURITY -ADJUSTMENT    	     6-3
   IV.   DETAILED DISCUSSION OF THE ANALYTICAL METHODOLOGY   .     6-4

Chapter 7
DATA ANALYSIS AND RESULTS
   I.    INTRODUCTION	     7-1
   II.   ANALYSIS OF CONCENTRATION  DATA	     7-1
         A.  Statistical  Methodology 	     7-1
         B.  Results	     7-6
   III.  SUMMARY OF  CROSS-TABULATIONS   	    7-10
   IV.   CORRELATION OF  CONCENTRATION  WITH VARIOUS FACTORS   .    7-16
   V.    RELATION BETWEEN AIRBORNE  DYE CONCENTRATION AND
         OTHER VARIABLES    	    7-17

Chapter  8
INDIVIDUAL SITE CHARACTERIZATIONS
   i.    INTRODUCTION'	     8-1
   II.   FACILITY CHARACTERISTICS  	     8-1
         A.  General  Characteristics 	     8-1
         B.  Process  Characteristics 	     8-2
         C.  Dye Characteristics	     8-6
         D.  Controls and Safety	     8-6

Appendixes

A. SUMMARY  OF  MEASUREMENTS FOR THE 24 SITES  MONITORED    .  .  .     A-l
B. TEXTILE  DYEING PLANTS:   POPULATION AND
   SUBPOPULATION ESTIMATES 	     B-l
C. ANALYTICAL  METHODS    	     C-l
D. STATISTICAL METHODOLOGY 	     D-l
 iv

-------
                                TABLES

2-1.   STATISTICAL SUMMARY OF ANALYTICAL RESULTS   	     2-2
3-1.   PHASE  II  PARTICIPATION RATES  	     3-5
4-1.   BREAKDOWN OF SAMPLING FOR SITE VISITS:   RESPONSE TO
       SURVEY	     4-4
4-2.   SAMPLE WEIGHTS  FOR CALCULATING POPULATION
       ESTIMATES   	     4-7
7-1.   GRAVIMETRIC WEIGHT OF TOTAL DUST PER VOLUME OF AIR
       SAMPLED   	     7-2
7-2.   SPECTROPHOTOMETRIC ESTIMATES OF THE AVERAGE AIRBORNE
       CONCENTRATION OF  COMMERCIAL DYE   	     7-3
7-3.   SPECTROPHOTOMETRIC ESTIMATES OF THE AVERAGE AIRBORNE
       CONCENTRATION OF  ACTIVE COLORANT  	     7-4
7-4.   AIRBORNE  CONCENTRATION OF TOTAL COMMERCIAL DYE
       BY PLANT	     7-7
7-5.   AIRBORNE  CONCENTRATION OF TOTAL COMMERCIAL DYE
       BY WEIGHER	     7-7
7-6.   AIRBORNE  CONCENTRATION OF TOTAL ACTIVE  COLORANT
       BY PLANT	     7-8
7-7.   AIRBORNE  CONCENTRATION OF TOTAL ACTIVE  COLORANT
       BY WEIGHER	     7-8
7-8.   AIRBORNE  CONCENTRATION OF TOTAL COMMERCIAL DYE
       (UNWEIGHTED ESTIMATES)   	     7-9
7-9.   AIRBORNE  CONCENTRATION OF TOTAL ACTIVE  COLORANT
       (UNWEIGHTED ESTIMATES)   	     7-9
7-10.  NUMBER OF WEIGHERS WHO EXPERIENCE  VARIOUS WORKPLACE
       CONCENTRATIONS—ACTIVE COLORANT BASIS    	    7-13
7-11.  NUMBER OF WEIGHERS WHO EXPERIENCE  VARIOUS WORKPLACE
       CONCENTRATIONS—COMMERCIAL DYE BASIS 	    7-13
7-12.  COMMERCIAL DYE  CONCENTRATION ESTIMATES  BROKEN  DOWN
       BY OTHER  VARIABLES MEASURED DURING MONITORING    ....    7-15
7-13.  SPEARMAN  CORRELATIONS BETWEEN AIRBORNE  DYE
       CONCENTRATION AND SELECTED EXPLANATORY  FACTORS  ....    7-17
7-14.  RESULTS OF STEPWISE REGRESSION OF
       log(Airborne Dye  Concentration)  AGAINST EXPLANATORY
       FACTORS   	    7-19
7-15.  RESULTS OF TWO-VARIABLE REGRESSION ANALYSIS
       FOR COMMERCIAL  DYE CONCENTRATIONS    	    7-20
A-l.   SUMMARY DATA FOR  THE 24  SITES   	     A-l
A-2.   INDIVIDUAL SITE CHARACTERISTICS RECORDED DURING
       EACH MONITORING PERIOD  	     A-2
A-3.   SUMMARY OF SITE CHARACTERISTICS RECORDED DURING
       EACH MONITORING PERIOD  	     A-3
A-4.   SUMMARY OF FIBERS PROCESSED OR DYE CLASSES USED
       PER SITE	     A-4
A-5.   INDIVIDUAL SHIFT  CHARACTERISTICS MONITORED DURING
       EACH MONITORING PERIOD  	     A-5

-------
                               TABLES
                             (Continued)

A-6.  SUMMARY OF SHIFT  CHARACTERISTICS  RECORDED DURING
      EACH MONITORING PERIOD   	     A-6
A-7-  WORKER ACTIVITY    	     A-7
A-8.  SUMMARY OF WORKER ACTIVITY	     A-8
A-9.  DYE FREQUENCY  DURING MONITORING PERIOD,  SITE BASIS  .  .     A-9
A-10. DYE WEIGHING ACTIVITY DURING MONITORING  PERIOD:
      NUMBER OF WEIGHINGS OF EACH DYE CLASS   	    A-10
A-ll. DYE WEIGHING ACTIVITY DURING MONITORING  PERIOD:
      WEIGHT OF WEIGHINGS OF EACH DYE CLASS   	    A-ll
A-12. TEXTILE DYE WEIGHING ROOM  MONITORING STUDY:
      COMMERCIAL POWDER DYES WEIGHED, CLASS BASIS    	    A-12
A-13. TEXTILE DYE "WEIGHING ROOM  MONITORING STUDY:
      COMMERCIAL POWDER DYES WEIGHED, COLOR BASIS    	    A-13
A-14. TEXTILE DYE WEIGHING ROOM  MONITORING STUDY:
      INDIVIDUAL DYES ENCOUNTERED   	    A-16
A-15. CONTROLS AND PERSONAL PROTECTIVE  EQUIPMENT AND
      ENGINEERING CONTROLS  	    A-30
A-16. USE OF PROTECTIVE EQUIPMENT AND PERSONAL HYGIENE
      PRACTICES AT EACH SITE	    A-32
B-l.  COMMERCIAL DYE CONCENTRATION ESTIMATES BROKEN DOWN
      BY OTHER VARIABLES MEASURED DURING MONITORING   ....     B-2
C-l.  RESULTS OF TOTAL  DYE ESTIMATIONS  FOR 6-DYE MIXTURES   .     C-5
C-2.  RESULTS OF TOTAL  DYE ESTIMATIONS  FOR 10-DYE  MIXTURES  .     C-6
C-3.  RESULTS OF TOTAL  DYE ESTIMATIONS  FOR 20-DYE  MIXTURES  .     C-7
C-4.  TOTAL DYE ESTIMATES BASED  ON WEIGHTED AVERAGE
      SPECTRAL ABSORPTIVITY  CONSTANTS    	     C-8
C-5.  DYE PURITY AND ABSORPTIVITIES  	    C-18
D-l(a).  UNWEIGHTED  LOGARITHMIC ANOVA RESULTS:
         ACTIVE  DYE  BASIS	    D-13
D-l(b).  UNWEIGHTED  LOGARITHMIC ANOVA RESULTS:
         COMMERCIAL  DYE  BASIS	    D-13
D-2(a).  ESTABLISHMENT-WEIGHTED LOGARITHMIC ANOVA RESULTS:
         ACTIVE  DYE  BASIS	    D-16
D-2(b).  ESTABLISHMENT-WEIGHTED LOGARITHMIC ANOVA RESULTS:
         COMMERCIAL DYE  BASIS	    D-16
D-3(a).  WORKER-WEIGHTED LOGARITHMIC ANOVA RESULTS:
         ACTIVE  DYE  BASIS	    D-17
D-3(b).  WORKER-WEIGHTED LOGARITHMIC ANOVA RESULTS:
         COMMERCIAL DYE  BASIS	    D-17
D-4.     ESTIMATION OF ACROSS-PLANT MEASUREMENT VARIANCE   .  .    D-18
D-5(a).  CHARACTERISTICS OF THE DISTRIBUTION OF
         ACTIVE INGREDIENTS  	    D-21
D-5(b).  CHARACTERISTICS OF THE DISTRIBUTION OF
         COMMERCIAL DYE	    D-22
 Vl

-------
                               FIGURES

1-1.  TEXTILE DYES  SURVEY:  PROJECT  ORGANIZATION  	     1-2
4-1.  TEXTILE DYEING ROOM MONITORING STUDY SAMPLING
      EXPERIENCE	     4-2
6-1.  FLOW CHART FOR THE ESTIMATION  OF  TOTAL DYES  ON AIR
      FILTER	     6-2
7-1.  AIRBORNE COMMERCIAL DYE  CONCENTRATION   	    7-11
7-2.  AIRBORNE ACTIVE DYE CONCENTRATION   	    7-12
C-l.  SIMULATED COMMERCIAL DYE-BASED AVERAGE
      ABSORPTIVITIES  (WEIGHTED) FOR  FIVE  DYES   	    C-12
D-l.  AIRBORNE COMMERCIAL DYE  CONCENTRATION:
      NORMAL PROBABILITY PLOT   	     D-2
D-2 .  LOG(AIRBORNE  COMMERCIAL  DYE CONCENTRATION) :
      NORMAL PROBABILITY PLOT   	     D-3
D-3(a).  ACTIVE DYE  CONCENTRATIONS MEASURED ON LEFT AND
        RIGHT FILTERS:  FREQUENCY DISTRIBUTION  	     D-6
D-3(b).  Log(ACTIVE  DYE  CONCENTRATION) MEASURED ON
        LEFT AND RIGHT  FILTERS:  FREQUENCY DISTRIBUTION   .  .     D-7
D-4(a).  COMMERCIAL  DYE  CONCENTRATIONS MEASURED ON
        LEFT AND RIGHT  FILTERS:  FREQUENCY DISTRIBUTION   .  .     D-8
D-4(b).  Log(COMMERCIAL  DYE CONCENTRATION)  MEASURED ON
        LEFT AND RIGHT  FILTERS:  FREQUENCY DISTRIBUTION   .  .     D-9
                                                                  vil

-------
                          LIST OF ACRONYMS
ATMI      American Textile  Manufacturers Institute,  Inc.
CEB       Chemical Engineering Branch (within ETD)
CIH       Certified Industrial Hygienist
DDE       Design and Development Branch (within EED)
EED       Exposure Evaluation Division (within OTS)
EPA       U.S.  Environmental Protection Agency
ETD       Economics and Technology Division (within OTS)
ETAD      Ecological and Toxicological Association of the
          Dyestuffs Manufacturing Industry
FSB       Field Studies Branch (within EED)
HHI       Health and Hygiene, Inc.
ICB       Industrial Chemistry Branch (within ETD)
MRI       Midwest Research  Institute
OTS       Office of Toxic Substances (within EPA)
PEI       PEI Associates (formerly PEDCO Environmental,  Inc.)
QAM       Quality Assurance Manager
QAPP      Quality Assurance Project Plan
SAS       Statistical Analysis System
WCG       The Washington Consulting Group
 viii

-------
                      AUTHORS AND CONTRIBUTORS

     This survey of particulate dye levels in air of dye weighing
rooms (drug rooms)  of textile wet processing plants is distinctive
in that it represents the voluntary joint cooperative efforts of
industry and EPA.  Industry was represented by the American Textile
Manufacturers Institute, Inc. (ATMI) and the Ecological and
Toxicological Association of the Dyestuffs Manufacturing Industry
(ETAD).   These industry trade associations were supported by their
contractor,  Health and Hygiene,  Inc. of Greensboro,  N.C.

     EPA participation was from two divisions of the Office of Toxic
Substances (OTS).  This included the Economics and Technology
Division (ETD), with support from the Chemical Engineering Branch
(CEB) and the Industrial Chemistry Branch (ICB)/ and the Exposure
Evaluation Division (EED) with support from the Field Studies Branch
(FSB) and the Design and Development Branch (DDE).   Contract support
to OTS included PEI Associates for ETD and Midwest Research
Institute,  the Washington Consulting Group,  Inc. and Westat, Inc.
for EED.

American Textile Manufacturers Institute, Inc.  (ATMI)

     ATMI, in conjunction with ETAD, proposed the initial study plan
to OTS;  cooperated with OTS in development of the final study plan;
contacted industry sites that were selected for each phase of the
study,  explained the objectives and mechanism and encouraged
participation; provided information on the composition of the
industry; provided technical assistance from design to analysis;
served as a clearinghouse for all industry contacts; edited the
final reports; forwarded reports to each participant.
Key personnel included:

     Maggie Dean, ATMI
     Carlos Moore, ATMI
     O'Jay Niles, ATMI
Eugene Roberts, West Point Pepperell
John Tritsch, ATMI
Nancy Weinberg, ATMI
Ecological and Toxicological Association of the Dyestuffs
Manufacturing Industry (ETAD)

     ETAD, in conjunction with ATMI, proposed the initial study plan
to OTS; cooperated with OTS in development of the final study plan;
supplemented the OTS site identification list;  provided technical
assistance from design to analysis; assisted in the development of
an analytical methodology for the measurement of dye dust
concentrations; contacted industry sites that were selected for dye
                                                                   IX

-------
dust monitoring and encouraged participation; edited the final
reports.  Key personnel included:

           Barry  Bochner,  Atlantic Industries,  Inc.
           Edward Boland,  CIBA-GEIGY Corporation
           Eric Clarke,  ETAD
           Jay Dayan,  BASF Corporation
           David  Hackathorn,  Mobay Corporation
           Tucker Helmes,  ETAD
           Joseph LoMenzo,  CIBA-GEIGY Corporation
           Per Stensby,  CIBA-GEIGY Corporation
           Heinz  Trebeitz,  Hoechst Celanese
           Harshad Vyas, Mobay Corporation
           Wolfe  Wagner, ICI Americas, Inc.

Health & Hygiene, Inc.  (HHI)

      HHI,  in support  of industry groups,  scheduled  dates  for
monitoring with  sites that had agreed to participate; with PEI
representatives,  interviewed site  executives and walked through
sites prior to monitoring;  conducted dye dust air monitoring with
personal and area sampling  pumps;  collected  samples of all powder
dyes encountered and  of appropriate  chemicals;  sketched a floor plan
and dye flow sheet for  each site;  assisted PEI  in on-site
activities; conducted a gravimetric  analysis of dye dust on monitor
filters; assisted in  preparation and review  of  24 individual site
reports.  Key staff included:

           William D,yson              Melvin Witcher, Jr.
           Ronald Hill

Economics  and Technology Division (ETD)

      Based on  its responsibilities in the assessment  of worker
exposure,  ETD introduced the need for a  study to ETAD and ATMI;
jointly with EED provided overall management of planning, design and
implementation of the project.   ETD coordinated work within OTS with
the industry groups;  identified  industry  dye user sites;  supervised
field data collection efforts, which included identifying and
cataloguing all  dyes  encountered at each  site;  identified chemical
structures  of dyes; developed a  data base  characterizing  sites,
workers, workplace activities and industrial hygiene; managed  the
preparation of the 24 individual site reports;  and  participated in
the development  and review of the final  report. Key  staff included:

           William Burch              George  Heath
           Russell Farris

-------
Exposure Evaluation Division  (EED)

     EED participated in development of a final study plan;  jointly
with ETD managed overall project, supervised all aspects of this
study that were related to statistical design, site selection,
questionnaire development,  data collection,  quality assurance  (QA),
chemical analysis of dye dust samples, and statistical analysis of
results; performed field QA audits; supervised preparation of the
final report; edited and finalized the overall report.   Key staff
included:

           Joseph Breen               Thomas  Murray
           Margaret Conomos           Eileen  Reilly-Wiedow
           Mary  Frankenberry          Bradley Schultz
           Joseph Glatz               Sarah Shapley
           Martin Halper              Cindy Stroup
           Richard Kent

PEI Associates  (PEI)

     In support of ETD, PEI with HHI interviewed the site executives
and walked through sites prior to monitoring; gathered on-site data
characterizing dyes encountered, sites,  workers,  workplace
activities and industrial hygiene; assisted HHI in on-site
activities; prepared 24 individual site reports;  prepared drafts of
field sampling and results sections of the final report; prepared
characterization tables for the final report.  Key staff included:

           Tom Corwin                 Leslie  Ungers
           Paula Morelli-Schroth     Donald  Unruh
           Kenneth Troutman           Robert  Willson

The Washington Consulting Group, Inc. (WCG)

     In support of EED, WCG assisted in development of data quality
objectives; designed the survey; prepared the quality assurance
project plans; selected sites for the study; conducted the data
analysis and interpreted the results; prepared drafts of the final
report.  Key staff included:

           Harry Chmelynski           Annie Lo
           David Cox                  Bryan Porter
           Arnold Greenland           Charles Smith
           Ayah  Johnson               Vicki Stoltz
                                                                   XI

-------
Midwest Research Institute (MRI)

     In support of BED, MRI developed an innovative
spectrophotometric method of measuring total levels of several
widely used classes of textile dyes on air monitoring filters;
assisted in the development of the chemical analytical quality
assurance project plan; performed laboratory analyses of plant
samples; prepared individual reports of the analytical results for
each monitored plant; prepared the draft of the chemical analysis
methodology.  Key staff included:

           Jack Balsinger             Cynthia Palmer
           Paul Constant             Robin Paris
           Jairus 'Flora,  Jr.          Roger Rembecki
           John Going                Julie Ryan
           Don Harbin

Westat, Inc.

     In support of EED, Westat,  Inc. conducted the mailing of
questionnaires to plants in the  first phase of the survey.  Key
staff included:

           Stephen Dietz             Sharon Gregory
           Susan Engelhardt          Diane Ward
 XII

-------
                           ACKNOWLEDGMENT

     The authors and contributors of this report express their
sincere appreciation to the management and staff at each of the
textile plants who participated in this survey.   Without their
cooperation this study would not have been possible.   The authors
and contributors convey their gratitude to reviewers  from the
Universities of Georgia and Alabama,  the Georgia Institute of
Technology, the National Institute for Occupational Safety and
Health, and the Amalgamated Clothing and Textile Workers Union for
their helpful comments and suggestions.
                                                                xiii

-------
                          EXECUTIVE SUMMARY

I.  INTRODUCTION

     This report presents the results of a survey conducted  jointly
by the U.S.  Environmental Protection Agency (EPA),  the American
Textile Manufacturers Institute, Inc. (ATMI),  and the Ecological and
Toxicological Association of the Dyestuffs Manufacturing Industry
(ETAD)  to estimate airborne concentrations of dye dust in the dye
weighing rooms (drug rooms)  of plants that use powder dyes in the
dyeing and printing of textiles.  The purpose of the project was to
conduct a well-designed study of representative textile dye weighing
rooms,  in order to improve the assessment of workplace exposure
associated with the use of powder dyes in the American textile
industry.

     More than 1,000 domestic textile processing sites have been
identified where dyeing or printing operations may occur.  However,
the available data on potential exposure levels of workers
associated with the weighing or mixing of powder dyes are limited,
and they are not always representative of textile dyeing operations.
Textile workers may be exposed to powder dyes  via inhalation during
dye weighing or mixing operations,  and the EPA is concerned about a
number of potential health hazards from exposure to dye dust.  For
example, some dyes or some of their metabolites are thought to be
carcinogens or mutagens.

     The distribution of dye dust concentrations obtained in this
survey provides the EPA with improved estimates of occupational
exposures for use in developing risk assessments for powder dyes.
Information about the mass,  frequency,  and number of powder dye
compounds weighed during a typical shift was collected,  along with a
physical characterization of the drug room,  to assist in
understanding the factors related to airborne  dye dust
concentrations.  The survey was based on a probability sample of 24
sites chosen at random from textile plants where powder dyes are
weighed.  Estimates produced from probability  samples in carefully
executed studies, such as this one,  are strongly preferred to case
study evidence of a limited nature,  which has  been the only
information previously available to the EPA on textile dye exposure.
The estimates obtained in this study substantially improve the
credibility of estimates of exposure of textile weighers to dye dust
over those based on previously available data.
                                                                   xv

-------
II.  OBJECTIVES OF THE STUDY

     To address the issue of improving the assessment of human
exposure via inhalation from the use of powder dyes, industry and
EPA agreed that the study should have the following objectives:

     1.   Estimate the distribution of dye concentrations  in dye
          weigher breathing zones  (8-hr  time-weighted average (TWA)
          concentration).

     2.   Determine  factors upon which dye concentrations  in the
          breathing  zone are dependent,  including the amount of dye
          weighed and number of weighings per shift.  Determine
          whether'there  is  a functional  relationship between these
          factors and airborne  dye concentrations.

     3.   Estimate the distribution of dye classes and  individual
          dye compounds  weighed during a shift.

     4.   Summarize  selected drug  room observations and general
          plant information.

     5.   Obtain an  extensive  first-hand qualitative view  of drug
          room operations and  characterize industrial hygiene
          practices  at each site.

III. APPROACH

     The  survey was  conducted in two phases.   The first phase
consisted of  a  survey of 240 plants  selected at random from a list
of 1,390 textile facilities thought  to use powder dyes.   The plants
selected were screened by telephone  to determine  eligibility, and
171 eligible plants received questionnaires in the mail to gather
information on  plant characteristics and determine which plants were
qualified as  candidates  for on-site  monitoring.   The response rate
to the questionnaire mailed in  the  first phase was  47%.

     In the  second phase, both  respondents and nonrespondents to the
questionnaire were selected  for monitoring.  Of 52 plants selected,
24 were monitored, resulting in a  response rate of  46% for phase
two.  While the group of establishments not monitored in phase two
represents an appreciable portion  of the total selected, examination
of the information gathered  on  the  plants in the  two groups did not
reveal any concern for bias.

     Monitoring of airborne dye levels in the plants and observation
of drug room  activities  were included in phase two  of the survey.  A
two-member team of certified industrial hygienists  recorded
 xvi

-------
measurements and observations in each plant to satisfy the study
objectives.  In addition, a more extensive examination of practices
and potential exposure was conducted for one randomly selected dye
weigher at each plant during one randomly selected shift.

     Monitoring of airborne dye levels took place over the course of
an 8-hr work shift.  Personal monitors were used to collect solids
from the air in the breathing zone of the workers wearing the
monitors.  Area sampling was also conducted.   Area samples provide
data on ambient dye levels, while samples from personal monitors
measure potential exposure levels for individual workers, without
consideration of personal protective equipment.

     All the samples collected were analyzed in a laboratory to
determine total dust mass and total dye mass.   The total dust mass
is a simple gravimetric measurement, but it was necessary to develop
a new analytical method to estimate dye levels.  This innovative
method involved a complex process,  since the samples contained
mixtures of several to many dyes in unknown proportions.   The
methodology was developed by testing a set of 23 "typical" textile
dyes selected by ETAD and ATMI.   The relative percent error of the
analytical method was estimated to range from -8 to +32 percent for
mixtures of 10 of these dyes,  and from -41 to +16 percent for
mixtures of 20 dyes.

     For each weigher selected at random for more extensive
monitoring, the team of industrial hygienists recorded the amount of
each powder dye and chemical weighed,  the number of powder dyes
weighed, the total number of weighings,  the amount of time the
weigher spent in the weighing area,  and other qualitative and
quantitative information.  Information was also recorded on the size
of the dyeing operation,  cleanliness,  ventilation,  possible routes
of exposure, and other qualitative characteristics.

     An individual site report was prepared for each plant, and
copies were forwarded to the plant.   In order to provide a context
for the participants  to interpret  their results,  a summary of
unweighted average and range values  for all the sites was also
provided.  A copy of  the summation can be found in Table A-l  of
Appendix A, and a summary of the individual site characterizations
is provided in Table  A-2.
                                                                xvii

-------
IV.  RESULTS

     The mean airborne  concentration of commercial dye dust1 for the
target population of plants monitored was estimated to be  0.18
milligrams per cubic meter  (mg/m3) .   This estimate falls within a
95% confidence interval that ranges  from 0.11 to  0.31  mg/m3.  The
geometric mean of the distribution of commercial  dye dust  was 0.11
mg/m3,  with a geometric standard deviation of 2.80.  The 95th
percentile of this distribution  (representing an  estimate  of the
concentration level that would be exceeded by only 5%  of all textile
dyeing plants) is 0.57 mg/m3.   The mean airborne concentration of
active colorant for the population was  estimated  to be 0.085 mg/m3,
with a 95% confidence interval from  0.049 to 0.15 mg/m3 and a 95th
percentile value of 0.27 mg/m3.   The geometric mean of the
distribution of active  colorant  was  0.049 mg/m3, with a geometric
standard deviation of 2.85.  The estimated values for  concentrations
of both commercial dye  and  active colorant closely followed a
lognormal distribution, as  is typical of occupational  exposure  data.

     Significant  correlations were  observed between dye
concentration and 5 of  21 variables  sampled  during the study, for
both commercial dye and active colorant concentrations.  The five
variables with significant  correlation  coefficients were number of
dyes weighed, mass of dye weighed, number of weighings of  dyes,
number of  suppliers, and number  of dye  classes.   Although  the first
three of these variables  (number of  dyes weighed,  mass of  dye
weighed, and number of  weighings) were  expected to be  influential,
the significance  of the other two  (number of suppliers and number  of
dye classes) was  surprising.
                  •t
     On  the basis of  these findings, several statistical models were
examined for their ability  to predict dye dust  concentrations from
the estimated values of other variables. The best of  these models
explained  up to half of the variability in dye  dust concentrations.
The remaining variability may be due to many factors,  such as
variables  not measured in the survey and random characteristics of
the samples,  including uncertainty  in the estimates of dye
concentrations for each plant.
     Because the response of textile dyes to the analytical method
 used was proportional to the purity  of the dyes measured,  correction
 for  dye purity, or active colorant content, was required.   Therefore,
 results are presented for both total commercial dye dust and active
 colorant.  The results presented in  this  report are based on data for
 samples from personal exposure monitors,  unless otherwise noted.
 xvi 11

-------
V.   CONCLUSIONS

     The major accomplishment of the study is the acquisition of
representative data on concentrations of commercial dye dust and
active colorant in the air of textile wet processing plants.  In
addition, the new analytical method developed to estimate dye levels
from air samples will be useful for future studies of exposure to
textile dyes.  The observations of industry weighing activities and
industrial hygiene practices recorded in this study provide insight
into the factors that contribute to and control exposure of dye
weighers.  The data gathered here will be useful both in EPA's
existing chemicals program and in its premanufacture notification
(PMN) process for new chemicals.  The data also provide an
information base for future studies and for the development of
explanatory and predictive models of airborne dye concentrations and
other exposure-related factors in the textile dyeing process.
                                                                 xix

-------
                              Chapter 1

                            INTRODUCTION
I.   BACKGROUND

     Data which document potential exposure levels of workers
associated with the weighing or mixing of powder dyes are limited,
and not always representative of textile dyeing operations.  The
purpose of this project was to conduct a well-designed study of
representative textile dye weighing rooms (drug rooms)  in order to
improve the assessment of workplace exposure associated with the use
of powder dyes in the American textile industry-  Prior assessments
were based on industrial hygiene survey reports of benzidine-azo
dyes prepared by the National Institute for Occupational Safety and
Health, 1977-1978.  More than 1,000 domestic textile processing
sites where dyeing or printing operations may occur have been
identified.  Textile workers may be exposed to powder dyes via
inhalation during dye weighing or mixing operations,  and the EPA is
concerned about a number of potential health hazards from exposure
to dye dust.  For example,  some dyes or some of their metabolites
are thought to be carcinogens or mutagens.

     In the current study, dyes are divided into two categories, as
follows:

      (1)   Active  Colorant—Undiluted chemical substance (s) that  can
          be  affixed  to a  substrate in order to provide coloring
          effects.

      (2)   Commercial  Dye—Formulated mixture of active colorant(s)
          and one or  more other substances offered to the trade,
          usually under a name specific to the  supplier and often
          identified  by a Color Index Name.  Other components may
          include diluent, dispersing agent, solubility promoter,
          dedusting oil or other chemicals to enhance usage on a
          commercial  scale.  A given supplier may formulate a
          commercial  dye at  several concentration levels of active
          colorant, usually  reflected in  its price.

     The study was initially proposed to the Environmental
Protection Agency (EPA)  by representatives from the American Textile
Manufacturers Institute,  Inc. (ATMI)  and the Ecological and
Toxicological Association of the Dyestuffs Manufacturing Industry
(ETAD) .  The implementation of the study plan was from the beginning
a collaborative effort between the industry representatives and EPA;
                                                                  1-1

-------
representatives of the three organizations met  regularly throughout
the study to review the progress and provide  direction.

     The organization of  the  study team is  illustrated in
Figure 1-1.  As shown  in  the figure, within EPA the study was
managed by the Office  of  Toxic  Substances  (OTS).   Responsibilities
were shared by the Economics and Technology Division (ETD)  and the
Exposure Evaluation Division  (EED).  Four branches in those
divisions were involved in  the  study.   They were the Chemical
Engineering Branch  (CEB)  and Industrial Chemistry Branch (ICB)  from
ETD and the Design and Development Branch  (DDE)  and the  Field
Studies Branch  (FSB)  from EED.  Figure 1-1 also shows the
contractors who provided  support to ATMI, ETAD,  and EPA.  The
                              Figure 1-1

              TEXTILE DYES SURVEY:  PROJECT ORGANIZATION
       ATMI
     ETAD
        U.S. EPA
         Office of
      Toxic Substances
                        ETD
                               EED
                  CEB
     HHI
           ICB
       DDB
PEI
             FSB
WCG
Westat
MRI
 1-2

-------
specific contractors involved as well as all of the contributions of
organizational participants are given in the Authors and
Contributors section of this report.

     The major tasks in the study were survey design, quality
assurance,  drawing a national sample of textile dye plants,
developing data collection procedures and materials, site visits
(including recording of worker activities and industrial hygiene
practices and monitoring and collecting bulk dye samples),  chemical
analysis of collection media (filters), data preparation,
statistical analysis, interpretation of results, and report writing.
Quality assurance plans were written for each aspect of the project.

II.  STUDY OBJECTIVES

     To address the  issue of how to improve the assessment of human
exposure via inhalation from the use of powder dyes, industry and
EPA agreed that the study should have the following objectives:

     1.   Estimate  the distribution of dye weigher  breathing zone
          dye  concentration  (8-hr TWA  concentration).

     2.   Determine factors upon which dye concentrations  in the
          breathing zone are dependent.  Factors to be explored
          will include at least the amount of dye weighed  and the
          number  of weighings per shift.  Determine whether a
          functional relationship exists between concentration and
          these factors.

     3.   Estimate  the distribution of dye classes  and individual
          dye  compounds weighed during a shift.

     4.   Summarize selected drug room observations and  general
          plant information.

     5.   Obtain  an extensive first-hand qualitative view  of drug
          room operations and characterize industrial hygiene
          practices at each site.

III. OVERVIEW OF  THE REPORT

     This report  describes how the study was conducted and presents
the study results.  Chapter 2 describes the study's conclusions.
Chapter 3 provides an overview of the quality assurance program.
Chapter 4 describes the survey design and how the textile plants
were selected.  Chapter 5 describes field sampling procedures which
were followed for collecting in-plant data.  Chapter 6 explains the
                                                                  1-3

-------
methodology used for the chemical analysis of the filters.  Chapter
7 presents the analytical results and statistical analysis of the
survey data.  Chapter 8 discusses the facility operations, worker
activities, powder dyes encountered, and control characteristics of
the textile plants visited.  Detailed tables presenting site-
specific data and summary tables of survey-based estimates for the
national population of textile dyeing plants are included in the
Appendixes to this report.  Copies of study materials such as the
quality control project plans, questionnaires, and other information
can be found in a separate supplemental volume accompanying this
report.
 1-4

-------
                              Chapter 2

                       RESULTS AND CONCLUSIONS
     The overall accomplishment of this study has been the
acquisition of data on airborne concentrations of dye dust in the
dye-weighing areas of textile processing plants which are more
representative of the industry than data previously available to
EPA.  The random site selection process coupled with records of dyes
encountered, masses weighed, weighing frequency and other quantities
collected at each site have allowed a study of factors which affect
dye concentrations in air.  Recorded observations of industry
weighing activities and industrial hygiene practices provided
insight that was previously unavailable into factors contributing to
and controlling exposure of dye weighers.   These values will be
useful in both EPA's existing chemicals program and its
premanufacture notification process for new chemical substances.
The study" has also provided an information base for future studies
and development of explanatory and predictive models of airborne dye
concentration and other exposure-related characteristics of textile
dyeing processes.

     While this survey at both stages had less than optimum response
rate, no evidence of response bias was found.  This is mentioned
here to provide a context within which to interpret the results
presented below.  When there is a small response rate, it is
important to consider the possibility of bias in the estimates.  In
phase one,  of the 171 plants to which questionnaires were mailed,  81
plants, or 47%, responded with completed questionnaires.   Due to
this large nonresponse rate in phase one,  the sites for monitoring
in phase two were selected at random from both respondents and
nonrespondents to the phase one questionnaire.   Phase two had a
response rate of 46% with 24 plants monitored.   An examination of
the respondents and nonrespondents did not reveal any reason for
concern.  No specific evidence,  other than the existence of the
nonresponse, was uncovered to indicate that there was a problem.
There was no perceptible difference between airborne dye
concentration estimates for the plants which responded to the
mailed-out questionnaire and those that did not.   This was taken to
mean that the incidence of plant response may not be related to the
level of airborne dye concentration encountered.

I.   PRIMARY FINDINGS

     Specific  results and conclusions are presented below for each
of the survey's-objectives.
                                                                  2-1

-------
1.   Estimate the distribution of dye weigher breathing zone dye
     concentrations  (8-hour time-weighted average concentration).

     Concentrations  of total dust, commercial dyes, and active
colorants in the breathing zone of 24 textile dye weighers over a
period of one shift have been measured analytically.  The summary
statistics for the distributions of concentration for dye dust,
commercial dye and active colorant measured at monitored sites are
given in Table 2-1, and the analytical measurements are provided in
Tables 7-1, 7-2, and 7-3.  These results demonstrate that average
airborne dye dust exposures are substantially lower than prior
assessments.
                              Table 2-1

              STATISTICAL SUMMARY OF ANALYTICAL RESULTS

                        Concentration in mg/m3
                         (Weighted by Plant)
Parameter
>
,>
Lowest Value
Median
Geometric Standard
Deviation
Mean
85th Percentile
90th Percentile
95th Percentile
Highest Value
Total
Dust
0.023
0.39

2.09
0.51
0.84
1.0
1.3
1.37
Total
Commercial Dyes
0.013
0.11

2.80
0.18
0.31
0.39
0.57-
1.20
Total
Active Colorant
0.007
0.049

2.849
0.085
0.15
0.19
0.27
0.56
   Note:  The  "Lowest"  and "Highest" values are simply the minimum
 and maximum  sample measurements.   These values are not appropriate
 for describing the target  population of plants.  All other
 parameters are population  estimates.  All three of these
 distributions  were found to follow the lognormal distribution, a
 probability  distribution which is  often associated with occupational
 exposure measurements.   The percentile estimates were computed using
 this  assumption.  For this distribution, the median is also an
 estimate of  the geometric  mean.
 2-2

-------
2.   Determine factors upon which dye concentrations in the
     breathing zone are dependent.  Factors to be explored will
     include at least the amount of dye weighed and the number of
     weighings per shift.  Determine whether a functional
     relationship exists between concentration and these factors.

     A statistical analysis was done to identify relationships
between the approximately 21 variables on which data were collected
in the' study and the measured airborne dye concentrations.   The
following five variables were found to have a statistically
significant (at the 5% level)  correlation:

     •  Number of Dye Suppliers;
     •  Number of Dyes Weighed;
     •  Mass of Dye Weighed;
     •  Number of Weighings; and
     •  Number of Dye Classes.

Two of the variables,  Number of Suppliers and Number of Dye Classes,
displayed significant correlations to airborne concentrations of
commercial dye.   These results are counter to intuition,  but the
variables may be surrogates for other variables not  measured in this
study.

     Further analysis was carried out using correlation and stepwise
regression techniques to investigate interrelationships among
various study variables and airborne commercial dye  concentrations.
The "best" regression equation selected by a stepwise regression
procedure was that which included only Number of Suppliers,  with an
R2 value  of  0.56.   (This value of R2, the coefficient of
determination,  implies that Number of Suppliers explains 56% of the
observed variation across plants in airborne dye concentration.)
This finding does not have a plausible explanation.   The next "best"
regression equation included only Mass Weighed,  with an R2 value of
0.39.

     While the stepwise regression procedure selected only
1-variable models,  2-variable regression models were also
considered.   When Number of Dye Weighings was included in addition
to Mass Weighed,  this second variable was found to have marginal
significance,  with an increase in the R2 for the regression to 0.47.

3.   Estimate the distribution of dye classes and individual dye
     compounds weighed during a shift.

     Besides active colorant and commercial dye airborne
concentration,  rrtany other variables toere collected from the study.
These ranged from the total number, mass and frequency of dye
                                                                  2-3

-------
compounds weighed to the availability and use of personal protective
equipment by workers in the drug room.  Results for variables of
particular interest are provided below.  Descriptive statistics for
each dye encountered and all the other variables are presented in
Appendix A.
                       POWDER COMMERCIAL DYES

VARIABLE                WEIGHTED MEAN VALUE          RANGE OF VALUES

                                                      Low      High

Mass Weighed            58.2 kg/shift/site            2.1      283.9
Number of Dyes Used     17.1 dyes/shift/site          2         46
Number of Weighings     60.3 weighings/shift/site     7        259
4.    Summarize  selected drug room observations  and general plant
      information.

      At  each  site,  drug room and general  facility characteristics
were  recorded for use in the  evaluation of drug room operational
factors which may affect the  potential for worker exposure.

      Although some  sites were quite  similar  to  others, textile wet
processing operations in general were found to vary considerably in
size, scope,  and processing equipment, and some were atypical among
those monitored. * Examples of similarities among 24 sites included
vertical management  (20), location in EPA Region 4 (19), 24-hr
operation per day  (18),  and batch processing equipment  (18).
Examples of variables exhibiting wide variability included
production volumes  (0.3  to 25 million pounds per year), number of
dyeing machines in operation  (1 to 75), number of dye weighings made
per shift  (7  to 259), mass of powder commercial dye weighed per
shift (2 to 284 kg), number of dyes weighed per shift  (2 to 46), and
a variety of  end products.  Each  site is  described fully in
individual site reports  and  summations, which are contained in
Appendix A.

5.    Obtain an  extensive first-hand  qualitative view of drug  room
      operations and characterize industrial  hygiene practices at
      each  site.

      Dye house  operations  were found to vary considerably.  However,
among the many  operational procedures described by the industrial
hygienists, the mechanism  for weighing  and mixing powder dyes was
notably  consistent  at the  24  sites.  Other than a universal absence
 2-4

-------
of engineering controls to control dye dust exposure, industrial
hygiene practices also displayed wide variation.

II.  ADDITIONAL FINDINGS

     Finally, although not directly related to the survey's
objectives, there were important results related to the chemical
analytical methodology that was used.   Many different analytical
approaches were evaluated to find one that could reliably measure
dye mixture concentrations in drug room samples.  Air sample filters
collected during the study could contain as many as 46 different
textile dyes from up to 5 dye classes.   Conventional methods of dye
quantification were found to be limited in this regard,
necessitating the development of a novel analytical technique.

     An innovative method developed for the study was a
spectrophotometric procedure by which the weighted average of the
individual dye's "spectral" absorptivity constants were used to
derive a constant of the mixture of dyes trapped on a filter.  This
average absorptivity was then used to estimate the amount of dye
material on the filter.   Overall,  this  analytical procedure
successfully met the study objectives,  although analytical
difficulties were encountered with a small group of dyes that are
infrequently used.  Prior to beginning  the full-scale study,  the
procedure was evaluated using known quantities of 20  commercial dyes
representing the five most frequently used dye classes.   The
relative error for total dye measured was found to be within +_ 40%;
this level of accuracy was determined to be acceptable for the
purpose of this study.  It should be emphasized,  however,  that this
figure is not necessarily representative for all 24 plant sites in
the study.  The levels of accuracy for  the analytical method can be
so variable that a general value for the accuracy cannot be stated
prior to obtaining information about the absorption characteristics
of the specific dyes being analyzed.  In addition,  a  few
infrequently used classes of dyes  cannot be measured  by this
procedure.
                                                                  2-5

-------
                              Chapter 3

                      QUALITY ASSURANCE PROGRAM
I.    INTRODUCTION
     Every aspect of the project was subject to quality assurance
(QA) considerations.  Such measures as thoroughly reviewing all work
and assuring that all project personnel have adequate experience and
training for their responsibilities were followed throughout.  Some
aspects of the project work, however, had specific QA procedures,
which are summarized in this chapter.

     The approach to QA planning defined for the study by the study
work group was the preparation of Quality Assurance Project Plans
(QAPPs) setting forth strategies for producing error-free and highly
reliable data.  A QAPP was prepared for each of the three major
components of the survey:

     •  Statistical Design, Data Objectives, and Data Analysis;

     •  Field Sample and Data Collection at the Plants; and

     •  Analytical Methodology Development and Analysis of Field
        Samples.

     These plans  (Section A of Supplement) complement each other and
together constitute a complete quality assurance plan for this
study.   Methods for in-plant monitoring,  recording-observations,  and
making the chemical analyses were pretested at a pilot textile
dyeing site.

     The following sections discuss the QA procedures associated
with the survey design,  data collection,  and creation of the study
data base,  and the quality control activities used before and during
site visits,  laboratory analysis,  and data analysis.
                                                              i
II.  STATISTICAL/QUALITY CONTROL PLANNING

     The data presented in this report are based on a sample survey.
In common with all survey data,  they are subject to sampling and
nonsampling errors.  These two types of error are discussed below.
Before presenting that discussion, the advantages of data collected
in a survey over anecdotal evidence, expert opinion,  and attempted
complete enumerations which include a very low percent of the
universe are reviewed.
                                                                  3-1

-------
     The advantages of survey-based data are that it is known what
the final estimates represent.  Specifically, the universe of plants
to which the estimates apply can be stated; the type of monitoring
conducted is specified and uniform; the same data items are
collected for all selected plants using common definitions; the
laboratory analyses are conducted in a known fashion with specified
quality control procedures; the tabulations are made using
established definitions and the level of precision of the estimates
can be measured by estimating the sampling error from the data.

     Errors encountered in sample  surveys  are commonly classified
into two groups:  sampling and nonsampling errors.  Sampling errors
are discrepancies between the sample estimates and the actual
population values being estimated.  If the statistical procedures
for selecting the elements of the survey are carefully controlled,
the sampling errors for a probability sample can be estimated.

     In contrast, nonsampling errors are those which result from
sources other than those attributable to sampling.  There were
various potential sources of nonsampling error in this survey.
Although the impact of such errors on the estimates is not generally
quantifiable, it is important to acknowledge these sources so that
users of the survey data may be aware of their possible effects.

     Potential  sources of  nonsampling errors include:  nonresponse
bias (discussed below in III.A); failing to sample a representative
group of textile dyeing plants; errors in laboratory analysis of
bulk samples; and errors in data collection, transcription,
keypunching, or computer manipulations.  The QAPP given in Section A
of the Supplement addresses each of these potential sources of
error.  Although such errors may still have occurred, there is no
evidence to suggest that they introduced bias into the survey
results.

     A. Frame Construction

     Frame  errors are those  caused by incorrectly including or
excluding units on the list of plants from which the sample was
selected.  To minimize such errors, a list of textile plants, both
known and suspected users of powder dyes, was constructed using
Davison's Textile Blue Book of Manufacturers.1   The  list  underwent
an extensive verification process which included:
     xNealy BN.   1983.  Davison Textile  Blue Book, 117th Edition.
 Ridgewood, NJ:  Davison Publishing Company.
 3-2

-------
      •  Cross-checking the list with the Standard Industrial
        Classification listing;

        Cross-checking the list with a list of plants known to
        discharge dye-containing effluent;

      •  Additions to the list by EPA personnel familiar with the
        textile dyeing industry;

      •  Revising the list using intimate knowledge of the industry
         (this was done by ETAD and ATMI); and

      •  Conducting a pilot study (Phase I of the survey) to ensure
        that there were no systematic omissions.

      These activities resulted in a frame with 1,390 members.  A
telephone screening survey was then conducted on a random sample of
240 plants selected from the list of 1,390.   Of these plants,  a
total of 171 (71% of 240)  was found to be eligible for the study.
Eligibility was determined primarily on the  basis of sites being
still in business and by their use  of manual weighing of powder dyes
for textile dyeing or printing operations.

      B. Pilot Study of the Industry—Mailed-Out Survey

      A survey of the 171 eligible plants was then conducted by mail
to accomplish a number of quality control tasks.   In particular,  the
survey planning team wished to collect and verify information so
that frame and measurement biases would be reduced.   In addition,
data were requested which would contribute to the design of on-site
monitoring protocols,  thus reducing measurement errors.   The mail
survey had the following results:

      •  The construction of the sample list was verified.  The 171
        randomly selected plants did not deviate in any systematic
        manner from the known make-up of the textile wet processing
        industry.  The random sample of 171 plants provided a
        manageable list for careful examination.

      •  Knowledge of the textile dyeing/printing industry for
        preparation of in-plant monitoring was enhanced.

      •  Stratification of the sample for in-plant monitoring was
        shown to be possible  (the entire list of 1,390 was too large
        to be accurately classified with reasonable cost) .
                                                                  3-3

-------
III.  QA PLANNING FOR THE SITE VISITS

     Prior to the in-plant personal and area monitoring, various
activities were conducted to ensure the quality of the data.  The
questionnaire for the in-plant monitoring study was designed and
tested in a pilot site.  The various quality control activities for
the site visits and the chemical analysis were completed.

     A. Optimizing the Monitoring Participation Rate

     One  of the most critical components of a  sound study is a high
participation rate, since each refusal can add an uncontrolled bias
to the results.  In conformance with government standards and good
scientific practice, a target of 75-80% participation was set.
However, due to the inherent complexity of this particular study,
characterized in large part by the fact that it was a voluntary
program, obtaining participation rates lower than the target was
considered quite likely.

     A total of  62 plants was selected at  random to be  contacted for
in-plant monitoring.  To encourage a high participation rate, all
selected companies were contacted prior to initiation of the site
visits.  A carefully crafted letter was sent by ATMI describing the
study,  guaranteeing confidentiality for the participants, asking for
permission to monitor, and describing the potential benefits of
their  participation  (in an industry-wide sense, and with respect to
free plant monitoring and written reports on the site visit).
                  >
                  ,-»
     Where  initial  resistance to participation was encountered,  the
following actions were options to improve the  response  rate:

      •  Other  appropriate  trade  groups,  e.g.,  the  Carpet and Rug
         Institute (CRI), the National Association  of  Hosiery
         Manufacturers  (NAHM),  or the  Carpet Manufacturers
         Association of the West  (CMAW)  were asked  to  contact plants
         and encourage  them to participate.

         Personal contact was made  with  the Chief Executive  Officer
         (CEO)  or other high official  of the selected  company,  by a
         CEO or other high  official of an ETAD firm participating in
         the study.

      These letters are included in this report in  Section C of the
Supplement.

      The actual response rates achieved for this  stage of the study
are  shown in Table  3-1.
 3-4

-------
                              Table 3-1

                    PHASE II PARTICIPATION RATES
                           Number        Number
Stratum                   Eligible*     Monitored      Percent
A.  Respondents to
    Mailed-Out Survey         23            15           65

B.  Nonrespondents to
    Mailed-Out Survey         29             9           31

    Total                     52            24           46
   *This number excludes two sites contacted which agreed to be
monitored but then were not (see discussion in Chapter 4)  and eight
sites which were found to be ineligible because they had gone out of
business or did not weigh powder dyes manually.


     Table 3-1 reveals that the actual participation rate was 46%,
lower than the target of 75-80%.  A participation rate this low
leaves open the possibility of bias in estimates produced by the
survey, although precision of estimates is preserved.   There was no
appreciable difference in estimates of airborne dye concentration
for the respondents and nonrespondents to the first phase
questionnaire (discussed in Chapter 7),  which was taken as an
indication that nonparticipation in phase two of the study might not
introduce a bias.  Unfortunately,  the nature of nonparticipation is
such that conclusions as to the impact on estimates produced cannot
be drawn.

     B. Quality Assurance for On-Site Visits

     The major steps for assuring quality for on-site visits were:

        Detailed planning of the site visit protocol; and
     •  QA visits by EPA team members during on-site visits.

           1.  Protocol

     The details of the protocol fo± site visits are described  in
Chapter 5 of this report and will not be presented here.  The reader
                                                                  3-5

-------
is referred to Section E of the Supplement for the form used by
industrial hygienists during the on-site visits.

           2.   Quality Assurance Visits

     To assure that all procedures defined in the QAPPs were
followed,  three QA visits were made by the EPA team.  This allowed a
further characterization of the data collected, especially in terms
of site variability.  One site visit each was conducted by FSB, DDE,
and CEB.

     The results of the QA  site visits were summarized as trip
reports with the following observations:

     •  The selection of the  shift and/or the weighers was done
        according to protocol.

     •  All air sampling instruments were calibrated prior to  field
        use.

     •  The personal samplers operated at a flow rate between  2.0
        and 2.5 L/min.

     •  The area sampling pumps operated at a  flow  rate between 10.1
        and 12.6 L/min.

     •  The area samplers were located according to the study
        protocol. ?

        Correct labeling of each  container of  bulk  dye used during
        the sampling period was verified and identified on the field
        sampling form.

     •  Bulk  dyes were  collected  in an unobtrusive  way, with care
        taken to limit  dust generation.

     •  All entries into and  exits out  of the  drug  room by weighers
        were  recorded.

     •  The name of the dye was recorded for all weighings and
        cross-checked with  the bulk dye  sampled.

      •  Management  representatives and  the weighers were  informed of
        the objectives  of  the study prior to  sampling.

      •  To the maximum  extent practicable, the work activities of
        the weighers were  not altered or interrupted by the visiting
        industrial  hygienists.
 3-6

-------
IV.  QUALITY ASSURANCE FOR THE ANALYTICAL METHODOLOGY

     The quality control program relating to the analytical
methodology used in measuring airborne dye concentrations included
the following four main components:

     1. Preparation of a QAPP;

     2. System audits conducted by the quality control coordinator
        at MRI;

     3. Periodic analysis of performance audit samples (PAS); and

     4. Audit by the QA Coordinator.

     For a detailed description of these components, please refer to
Appendix C.

V.   CONSTRUCTION OF THE COMPUTER DATA BASE

     The data collected in the field and in the laboratory were
forwarded to WCG and to CEB.   WCG entered some of the information
provided by MRI on the dye concentrations observed in the laboratory
for each of the participants.   CEB entered the data from  the on-site
monitoring field visits on a PC-based spreadsheet and forwarded this
file to WCG.   The data were verified again by WCG personnel to
ensure that no errors in data entry were made.  A SAS data base was
created by uploading the data from the PC-based spreadsheet to the
EPA mainframe.
                                                                  3-7

-------
                              Chapter 4

                 SURVEY DESIGN AND SAMPLE SELECTION
I.   DEFINITION OF TARGET POPULATIONS

     Two target populations were defined for the study.  The first
consisted of textile dyeing plants where powder dyes are weighed,
and the second consisted of workers who weigh powder dyes.   These
workers have the potential for inhalation and dermal exposure to
powder dyes; however, this study concentrated on inhalation exposure
only.  Population estimates, required for both target populations,
were derived from the sample data by the use of appropriate sample
weights, derived from the properties of the sample design described
in Section II, following.  A discussion of sample weights for the
plant and worker populations is contained in Section III.

II.  SAMPLE DESIGN

     A two-phase survey design was used for the study.  For the
first phase the goals were to (a)  ensure the accuracy of the frame;
(b) provide valuable general information on textile wet processing
operations including data useful to industrial hygienists in
preparing for in-plant monitoring;  and (c)  obtain information for
use during the second phase of the study.   The primary goals of the
second phase included breathing zone monitoring of the worker
conducted in-plant,  recording of typical worker activities,  and
characterization of industrial hygiene practices.   Figure 4-1
displays the two-phase plant selection process; the following
subsections describe the design in detail.

     A. First-Phase  Sample

     Of the 1,390 plants in the United States with the potential for
weighing textile dyes, 240 were chosen by simple random sampling.
Telephone calls were made to all 240 units to determine eligibility
for inclusion in the survey.  Eligibility depended on the plant
still being in the business of dyeing or printing textiles using
manually weighed powder dyes.   The current address was confirmed.
Of the 240 contacted, 171 sites were found eligible.  Questionnaires
were mailed to each; a copy can be found in Section D of the
Supplement.  The questionnaire requested information on the location
of the facility; its ownership (public or private); type of
operation  (vertical or commission); product lines;  the fibers
processed on site; classes of powder (solid) dyes weighed;  number
and type of dyeing or printing equipment used; frequency and volume
                                                                  4-1

-------
                                     Figure  4-1


       TEXTILE DYEING ROOM MONITORING STUDY  SAMPLING EXPERIENCE
 First-Phase Sample
  Second-Phase Sample
              Sampling Frame
              of Textile Plants:
                                      First Sample Drawn:

                                         240 Plants
                         Respondents to
                     Mailed-Out Questionnaire
                             on-respondents to
                                                  Mailed-Out Questionnaire
                                                     Plants Sampled:

                                                          35*
Plants Sampled:

     25*
                                   Out of Scope:
          Monitored:

             15
 Not Monitored

      8
       Not Monitored

            20
                                                            ot Monitore
                                                           by Cutoff Date:
                                                               10
        Dye Analysis Valid:

               15
            Dye Analysis valid:
ye Analysis Invalid
        *Does not include one plant in each category which had agreed to be monitored
    but was not (see Table 4-1).
4-2

-------
of powder dye handled; and presence or absence of drug room exposure
controls when powder dyes are weighed.

     Of 171 plants queried, 81 responded, and 90 failed to respond
by the predetermined cutoff date, producing a response rate of 47%.

     B. Second-Phase  Sample

     The sample  for the second phase of this study was drawn as a
subsample of those included in phase one.   Of several options
considered for the design of the subsample,  response to the mailed
questionnaire or lack thereof were selected as the only strata.
This stratification plan was selected because it:

     •  Allowed  measurement of differences in the nature of exposure
        levels between first-phase respondents and nonrespondents;

     •  Ensured  that the final representation from the two groups
        reflected the actual population of dyeing plants; and

     •  Reduced  the impact of refusal for in-plant monitoring bias
         (the act of responding to the questionnaire is seen to
        separate the two groups based on the likelihood of in-plant
        monitoring being allowed and thus reducing the refusal bias
        in the second phase).

     The stratification plan resulted in demonstration of the two
strata shown in Figure 4-1.   Stratum A contained the 81 plants  that
responded and Stratum B contained the 90 nonrespondents to the
mailed-out survey.

     The sample  size for in-plant monitoring was initially set at 30
but later reduced to 24,  based on a combination  of concern for
statistical accuracy and the limits of the EPA budget for in-plant
monitoring.  Fourteen plants out  of a total  of 81  were to be  drawn
from Stratum A,  and 16 plants  out of a total of  90 were to be
selected from Stratum B.

     While 30 plants were targeted for monitoring, the problem of
scheduling on-site visits and the likelihood that  some plants would
refuse to participate, dictated a strategy of over-sampling,  so that
62 plants were contacted.   The results of sampling at this stage are
summarized in Table 4-1.   Within Stratum A,  26 plants were
contacted.   Of these,  two were judged ineligible based on new
information that was not obtained through the initial telephone
contact,  and one which agreed to participate was not visited.
Within that group of remaining plants which were contacted, 15 sites
were monitored.
                                                                  4-3

-------
                              Table 4-1

     BREAKDOWN OF SAMPLING FOR SITE VISITS:  RESPONSE TO SURVEY



                          Respondent    Nonrespondent    Total


A.  Monitored sites           15              9*           24

B.  Refused to allow
    monitoring                 3             10            13

C.  Willing to be
    monitored, not
    called upon**              112

D.  No response—
    early selection but
    not pinned down            224

E.  Not resolved—late
    selection—little
    or no communication        3              8            11

F.  Ineligible                 2              6             8
                  >

         Total                26             36            62
    *Exposure level data for two sites were unusable because of
analytical  complications.  See Appendix C.

   **These two plants were not visited because of time and monetary
constraints.
      In Stratum B,  36 plants were contacted.   Six were  judged to be
 ineligible,  and one which had agreed to participate was not visited.
 Within the  group judged to be eligible in Stratum B, 9 plants were
 monitored.

      Response rates can easily  be derived from Table  4-1.   Response
 rate is defined as  the number of plants monitored divided by the
 total of number of  plants which are in scope.   "In scope" indicates
 4-4

-------
that the plants were engaged in textile dyeing operations which
required manual weighing of powder dyes.  In each stratum, there was
one plant that agreed to be monitored,  but which was not monitored
because of time and cost constraints.   These two plants were also
excluded from the "in scope" category because their data were not
included in the numerator of the response rate fraction.  For
stratum A, 15 plants were monitored from a total of 23 eligible
plants resulting in a response rate in that stratum of 65%.   In
stratum B, 9 were monitored out of 29 eligible resulting in a
stratum response rate of 31%.  The overall response rate was 46%.

           1.   In-Plant Monitoring

     A two-member team conducted an on-site survey of dye weighing
operations at each plant and recorded measurements and observations
on a questionnaire (see Section E of the Supplement).   One worker
from each plant was monitored.  The shift and the particular worker
within the shift to be monitored were selected at random following
the procedure specified in the QAPPs.   Field sampling procedures are
described in Chapter 5,  and the QAPP is described in Section A of
the Supplement.

III. DYES  STUDY WEIGHTING SCHEME

     Estimates of the parameters of the target populations of plants
and workers were developed using sample weights appropriate for each
population.  Because there are two separate target populations,
textile plants and dye workers, estimates for total numbers of
plants and estimates for total numbers  of workers in a particular
category must be developed separately.   Consider first estimates for
numbers of plants.   The original intention was to have 14 plants
monitored in Stratum A and 16 in Stratum B.  Since these numbers are
roughly in the proportion of the populations of interest (81 in A
and 90 in B), the resulting sample would have been a self-weighting
sample.  As discussed earlier in this chapter and illustrated in
Table 4-1, the actual numbers in the sample turned out to be
different from the targeted numbers due to differential levels of
nonparticipation in the two strata.  In such situations it is
required to develop a "nonresponse adjustment" factor to account for
the fact that the numbers of sample units in Stratum A and Stratum B
are not in proportion to the numbers for those two groups in the
population as a whole.  If estimates were produced without using a
weighting adjustment, the resulting estimators could be biased.
Bias is introduced as follows.  Stratum A represents 81 out of 171
or 47.4% of the population as a whole.   However, 15 out of 24 (or
62.5%) in the sample are in Stratum A.   When estimates within these
two strata differ in a substantial way, a biased estimate for the
overall population could be produced unless a weighting adjustment
                                                                  4-5

-------
is used.  Therefore, all estimates included in this report of
quantities for the universe of textile dyeing plants are calculated
using the nonresponse adjustment to produce a weighting scheme.

     For estimates produced for the population of dye weighers it is
even more crucial to apply a carefully designed weighting scheme,
because only one dye weigher was monitored at each site.  In such a
situation, the single worker represented all weighers at a site.
Total numbers of weighers per plant varied in this data base from
one to eight.  Without accounting for the differences between plants
due to the number of weighers, a biased estimate could be produced.

     The last point  can be illustrated by an example.  The
assumptions used are that there were only two plants in the entire
population—one large and one small; that the large plant had four
weighers and the small plant only one weigher; and that the exposure
of the worker sampled at the large plant was small, say 0.01 mg/m3,
while the exposure at the small plant for the worker monitored was
large, 0.31 mg/m3.   Using the  arithmetic  average  of the  two  values
 (0.16 mg/m3)  as  the estimator  for the  mean  exposure of weighers is
obviously biased.  Too much "weight" is given in this case to the
worker at the small plant.  If, however, the measurements in each
plant are weighted by the number of workers as follows:

     [4  x  (0.01  mg/m3) + 1 x (0.31 mg/m3)] /  (4  +  1)  =0.07 mg/m3,

the weighted mean v,alue produced is more representative of the
average of the population of five weighers, which is the target
population.  The above calculation is the essence of the weight
adjustment for  "workers" which we define below.  Consequently, all
estimates for the universe of weighers will be weighted using the
"worker" adjusted weights.

      Each  set of weights  was  obtained using a  careful accounting of
the number of textile plants in the original sampling frame and the
number  sampled.  In  addition, the set of weights derived for the dye
workers was based on the total number of weighers working normally
at each plant on any given day, a figure provided by management
during  the site visit.

      A.  Calculation of Plant-Level  Weights

      Plant-level  and worker-level weights  are  listed in Table  4-2.
The plant-level weights were computed as the reciprocal of the
probability of  being selected for inclusion in the  study.  The
probability calculation can be easily tracked by reference to
 4-6

-------
                              Table 4-2
         SAMPLE HEIGHTS FOR CALCULATING POPULATION ESTIMATES
           Weights  for  Estimates
            Based on 22 Plants*
Weights for Estimates
 Based on  24  Plants**
Site
10
16
21
24
27
30
33
38
41
43
46
49
52
54
59
62
65
66
77
79
80
86
88
91
Plant
Weights
28.77
28.77
61.70
61.70
61.70
28.77
28.77
28.77
61.70
28.77
28.77
28.77
28.77
28.77
28.77
61.70
61.70


28.77
28.77
61.70
28.77
28.77
Worker
Weights
172.64
86.32
185.10
61.70
246.80
172.64
172.64
86.32
123.40
172.64
57.55
86.32
57.55
115.09
86.32
185.10
123.40


28.77
230.18
123.40
86.32
28.77
Plant
Weights
28.77
28.77
47.99
47.99
47.99
28.77
28.77
28.77
47.99
28.77
28.77
28.77
28.77
28.77
28.77
47.99
47.99
47.99
47.99
28.77
28.77
47.99
28.77
28.77
Worker
Weights
172.64
86.32
143.96
47.99
191.95
172.64
172.64
86.32
95.98
172.64
57.55
86.32
57.55
115.09
86.32
143.96
95.98
287.93
143.96
28.77
230.18
95.98
86.32
28.77
   *Commercial and active dye concentrations were obtained from 22
plants.  Population estimates for plants and workers are based on
these weights for 22 plants.

  **Total dust concentrations were measured in 24 plants.
Population estimates for plants and workers are based on these
weights for 24 plants.
                                                                 4-7

-------
Figure 4-1.  Consider first the  probability (P)  of selection of the
final 15 from Stratum A:

          P(selection)  =   (240/1390) (25/81) (15/23)
                        =  0.0348.

Therefore, the weight for  these  cases  is 1/0.0348 = 28.77.

     For  Stratum B,  the calculation is:

          P(selection)  =   (240/1390) (35/90) (9/29) (7/9)
                        =  0.0162.

The resulting weight is 1/0.0162 =  61.70.   These weights are
constant  for each element  in  the stratum,  as always for stratified
samples.

     B.  Calculation of Worker-Level Weights

     The  worker-level  weights (WGTHorker)  are defined  as the product
of the weight for the plant  (WGTplant) times the total number of
weighers  at the  plant during  a typical day (NW).  Thus,  the general
formula for worker weights is:

               WGTHorker  =  (WGTplant) (NW) .

Unlike plant-level ^weights, it is possible for worker-level weights
to differ for each element in a stratum of the sample,  if the number
of workers  differs across  plants in each stratum.
 4-8

-------
                              Chapter 5

                      FIELD SAMPLING PROCEDURES
I.   IN-PLANT OBSERVATION PROCEDURES

     The purpose of each site survey was threefold:   (1) to conduct
personal and area industrial hygiene air monitoring to determine the
potential for textile worker exposure to airborne dye particulates
associated with the weighing and mixing of powder dyes; (2) to
determine the dye weighers'  assigned duties and to observe their
activities in the performance of those duties; and (3) to record
drug room and general facility characteristics for use in the
evaluation of drug room operational factors which may affect the
potential for worker exposure.

     The on-site activities were performed by a field  survey team
consisting of two board-certified1  industrial  hygienists, one  from
Health and Hygiene, Inc. (HHI),  representing ATMI and ETAD, and one
from PEI Associates,  Inc.,  representing EPA.  Prior to each
monitoring survey,  the survey team conducted a presurvey meeting
with company representatives at each site,  during which they
described the objectives of the study and the procedures that would
be followed during the site survey.  A synopsis of the plant
operations was provided by the plant supervisor at that time.

     During each monitoring survey, the survey team followed a
prescribed procedure for taking measurements and making related
observations of the dye weighing activities and the general facility
operation.   Personal and area air monitoring was conducted over
approximately an 8-hr period during a single work shift at each
facility.  All facility information and observations were recorded
on standard site survey forms developed for this study (see Section
E of the Supplement).   The facility characteristics,  sampling area
characteristics (including a sketch of the drug room area and
materials flow patterns),  the number of dyeing/printing units
(available and in operation),  the overall appearance of the
monitored area,  and the engineering controls in place were recorded.
Also recorded were the number of dye weighers employed at the
facility and the survey team's observations of employee work
practices and use of personal protective equipment.  A record was
made of the number of weighings and mass weighed of each powder dye
and chemical that was weighed by the person monitored.  A work
history was taken of each monitored dye weigher.
     1American Board of Industrial Hygiene.


                                                                  5-1

-------
II.  DYE COLLECTION PROCEDURES

     A. Industrial Hygiene Air  Monitoring

     The air monitoring  was  conducted using standard industrial
hygiene sampling equipment.  The personal  sampling  train used in the
surveys consisted of an  open-faced  filter2 connected to a Gilian
personal sampling pump with  a length  of Tygon tubing.   The  pumps
were calibrated to a prescribed flow  rate  of approximately  2  L/min.
The flow rates and start and stop times of the  sampling pumps were
recorded on air sampling data sheets.  The individual  (dye  weigher)
to be monitored was selected in accordance with the method  described
in the QAPPs  (see Section A  of  the  Supplement).   One dye weigher was
monitored for one work shift per site, except at Site  5/4.3  The
monitored dye weigher at each site  wore two sampling trains,  with
inlets located in the worker's  breathing  zone on each  lapel.   Two
area samples were collected  at  each site  on a similar  sampling train
with stationary high-volume  pumps calibrated to a flow rate of 5 to
12 L/min, which exceeded slightly the recommended flow rate in the
QA report of 5 to 8 L/min.   One area  sampling apparatus was located
near the drug room weighing  station and the other in the dye  drum
storage area within the  drug room,  in an  area remote from where the
monitored dye weigher was most  active.

     All  samples  (personal  and area)  were collected for
approximately an 8-hr period.   Sampling time intervals were recorded
on the appropriate air sampling data  sheets. Temperature,  relative
humidity, and barometric pressure measurements  of the  drug  room were
monitored and recorded hourly during  most  of the monitoring surveys.
When barometric pressure was not recorded during the monitoring
period, it was later obtained by contacting the weather station at
the local airport for that  facility.4  Field blanks, which were
     2The filter  used was a Metricil VML polyvinyl  chloride
 37-mm-diameter filter with  a  5-^.m pore  size, manufactured by Gelman
 Filtration Products.

     JAt Site 5/4,  two dye weighers were monitored  consecutively over
 the latter part of the first  work shift  (3 p.m. to 7 p.m.)  and the
 start of the second work shift (7 p.m. to 11 p.m.)  of two 12-hr
 shifts.  Both monitored workers wore the same sampling apparatus  so
 that the samples would be obtained over a consecutive 8-hr time
 period.

     4The failure to record data  was usually the result  of
 communication breakdown or equipment malfunction.


 5-2

-------
handled in the same manner as the sampling filters except that  no
air was drawn through them, were also submitted for analysis.

     Although the  sampling protocol described above was  followed in
general at each of the monitored sites, slight deviations did occur,
most notably at Sites 5/4 and 7/7.  At Site 5/4, previously
mentioned practical problems associated with scheduling required
sampling over two shifts.  At Site 7/7, the personal samples were
obtained with closed-faced filter cassettes.  Work practices at this
site resulted in excessive spraying of water.  The survey team  felt
that this precaution was necessary in order to avoid aspiration of.
water and complete invalidation (or destruction) of the sample.  In
addition, one of the filter cassettes used at Site 7/7 fell off the
sampling train and was inadvertently inserted in a backward position
by the dye weigher where it remained for 30 minutes of the
monitoring period before the survey team discovered the error and
returned the cassette to its correct position.

     B. Dye Bulk Sampling

     Bulk samples of the dyes weighed during each monitoring period
were obtained during each of the site surveys.   Samples of each dye
were placed in special containers5 and  labeled with the following
information:  bulk sample identification number, dye name, dye
manufacturer,  name of the drug room operations  and operator, name of
the person responsible for obtaining the sample, and the date the
sample was collected.  The survey team recorded the bulk sample
identification number, batch ticket name,  full  trade name, batch or
lot number,  and supplier name for each bulk sample on the
appropriate site survey form.  After the survey, bulk samples were
forwarded to the laboratory for analysis.
     5The bulk samples were collected in 2- to 4-ounce glass or
nonreactive plastic containers, amber  in  color  (to screen degradative
light),  which were purchased by Health and Hygiene, Inc.
                                                                  5-3

-------
                              Chapter 6

                       ANALYTICAL METHODOLOGY
I.   INTRODUCTION

     An analytical method was needed for very complex samples
containing trace levels of textile dyes.  Conventional methods of
dye quantitation would probably have required much longer analysis
times and might not have permitted a complete analysis of the very
complex samples.  The requirements of the analyses encountered in
the drug room survey necessitated the development of a novel
analytical technique to ensure that the requirements of the study
could be achieved.   This new technique arose from the conclusion
that sample complexity would preclude the determination of
individual dye concentrations.  The method therefore focused on
making an estimate of the total amount of dye present on air
filters.

II.  SCOPE OF THE METHOD

     The method which was developed is basically a
spectrophotometric procedure.  The spectral absorptivity (as)
constant of each individual dye handled by the monitored worker is
determined.  The analytical method uses a weighted average of the
individual dye as constants  to represent the  as constant of the
mixture of dyes trapped on the air filter.   This average
absorptivity is divided into the total absorbance measured for the
filter extract to estimate the amount of dye material on the filter.
The average extraction efficiency (adjustment for recovery)  of the
dyes from the air filters is determined by spiking a known dye
mixture onto a blank air filter and exposing the filter to constant
temperature, humidity, and fluorescent lighting for 8 hr.  After
exposure,  the spiked air filter is extracted and the total
absorbance is compared to that of a reference standard solution
which contains an identical quantity of the dye mixture and has
undergone the same exposure conditions.  This dye recovery
determination is then employed in the dye estimate calculation to
correct for dye losses in the filter extraction process.   The steps
involved in the analytical procedure for air filters, bulk dyes, and
spiked air filters are outlined in Figure 6-1.

     A  solution of dimethyl  sulfoxide  (DMSO) and pH  7.0 buffer  (9:1,
v/v) was used to extract all air filters and dissolve all bulk dye
samples.  With the exception of the trial plant and the first three
plant sites in the study, all absorbance measurements were taken
from 45 to 60 minutes after initial contact with the dye solvent.
                                                                  6-1

-------
                                                    Figure  6-1

          FLOW  CHART  FOR  THE  ESTIMATION  OF  TOTAL  DYES  ON AIR  FILTER
                                                  FMd Air Filter Samples
                                                and Air Sampling Cassette*
            Individual Dye
          a$ Determination
           Prepare solutions of each
         individual commercial dye at
           known concentrations in
                dye solvent.

             Calculate the spectral
          absorptivity constant (as)
                for each dye.
                   I
         Calculate the weighted average
         spectral absorptivity constant
               (weighted "a"5)
           for the dye analysis set
                                                           I
Extract dyes from each filter
 or cassette using a known
  volume of dye solvent
                                                Scan solution over 800-330 nm
                                                 range on spectrophotometer
                                                     and store spectrum.
                                                   Integrate area beneath the
                                                    spectrum to obtain total
                                                     absorbance (A TOT)
   Total Dye Estimate
     on Air Filter or
    Cassette Extract
                 asvalue
Using the weighted
calculate the total commercial
   dye estimate on the air
filters and sampling cassettes.
                                                                                                Dye Recovery
                                                                                                 Calculation
                                         Spike blank air filters with a
                                        known commercial dye mixture
                                        (e.g., spiking standard solution)
                                        at levels bracketing the amount
                                          estimated on the personal
                                         air filters. Pull air through the
                                            filters for 6 to 8 hours.
                                        Prepare reference standards
                                           using known amounts
                                        of the same spiking standard
                                           solution used to spike
                                            the blank air filters.
                                          Compare total absorbance
                                          of each spiked filter extract
                                          to that of the corresponding
                                         reference standard. Calculate
                                             the average percent
                                                recovery. R.
                                                  Calculate the level of active
                                                     colorant by adjusting
                                                   air filter dye estimates by
                                                  using dye recovery and dye
                                                      purity information.
                                                     Adjust cassette dye
                                                    estimates by using only
                                                    dye purity information.
                                                            I
                                                  Obtain estimates of mg total
                                                   commercial dye and active
                                                    colorant dye/m3of air.
6-2

-------
The instrumentation which was employed in the analyses consisted  of
a dual-beam spectrophotometer and a microcomputer-based
chromatography data system.

     Confidence  intervals  for the total dye estimates and estimated
average airborne total dye concentrations were obtained through a
microcomputer-based simulation program.  This program simulated
probable dye mixtures which could occur at a particular plant site.
By performing a statistical analysis on the simulated probable dye
mixtures which were generated, the uncertainty of the total dye
estimate could be evaluated for each plant site, and confidence
intervals could be assigned around the point estimate.

     The procedure was evaluated prior to commencing the full-scale
study using known quantities of 20 commercial dyes representing the
five most frequently used classes.  The relative error for total dye
measured was found to be within ± 40%.  This level of accuracy was
determined to be acceptable for the purpose of this study,  although
it should be emphasized that this figure is not necessarily
representative for all 24 plant sites in the study.  The levels of
accuracy for the analytical method can be so variable that  a general
value for the accuracy cannot be stated prior to obtaining
information about the absorption characteristics of the specific
dyes being analyzed.

     Analytical difficulties were encountered for samples from three
of the plant sites,  but most of those problems occurred in  the case
of dye classes that had never been analyzed using the established
procedure.   When it was observed that triphenyl methane (TPM)  basic
dyes reacted with a component of the solvent,  the procedure was
modified.   In addition, a few infrequently used classes of  dyes
cannot be measured by this procedure.   The insolubility of  vat and
sulfur dyes and the reactivity of naphthol dye components in the
solvent mixture precluded analytical detection.   However, the
analytical method was successful overall in meeting the study
objectives.

III. DYE PURITY ADJUSTMENT

     In addition to adjustments made based on extraction efficiency
discussed earlier (Section II),  adjustments were also required to
account for dye purities.  These are discussed briefly below.
Table C-5 in Appendix C contains estimated dye purities for all dyes
weighed at the 24 sites included in the survey.

     The response of textile dyes to the analytical method is
directly proportional to the purity of the measured dyes.
Correcting for the dye purity,  or active colorant content of each
                                                                  6-3

-------
textile dye was therefore required to compensate for the significant
variability in purity values which occurs from class to class,  as
well as within any particular dye class.  The active colorant
content of the textile dyes handled by the monitored workers was
obtained from dye manufacturers by ETAD.  In circumstances where dye
purity information was not readily available, ETAD provided an
estimate of the active colorant content.

IV.  DETAILED DISCUSSION  OF THE ANALYTICAL METHODOLOGY

     Appendix C of this report contains  a detailed discussion of the
development of the analytical methodology including the theoretical
basis, the description of the statistical methods used to produce
confidence bounds about dye concentration values for individual
sites, the quality control procedures in place for the analytical
work, and further discussion of the methodology used.
 6-4

-------
                              Chapter 7

                      DATA ANALYSIS AND RESULTS
I.    INTRODUCTION
     Tables 7-1, 7-2, and  7-3  contain the air monitoring  results  for
the 24 sites included in the survey.  The tables list gravimetric
mass of total dust, airborne concentrations of commercial dyes, and
airborne concentrations of active colorants.1   Total  dust
measurements were available at all 24 sites, while commercial dye
and active colorant concentrations were available at 22 sites.  In
the three tables, results are shown for the four separate monitors
used at each site and described in Section 5.I.A.  Two personal
monitors (denoted A and B)  were worn by the selected dye weigher.
The other two monitors were located in stationary positions:  one at
the weigh station and one at a remote storage area.  The statistical
analysis in this chapter is confined to the personal monitoring
data, specifically, the average of monitors A and B.

     Section II  of this chapter presents a statistical analysis of
the concentration data in Tables 7-1, 7-2,  and 7-3.  Survey-based
estimates are provided for the population mean, median, standard
deviation and selected percentiles of concentration,  for both
commercial dye and active colorant.  Selected confidence intervals
are also presented.  Section III concerns tabulation of dye
concentrations broken down by various categorized variables
collected during the on-site monitoring.  Section IV contains an
analysis of the correlation of concentration data with factors (such
as Mass of Dye Weighed and Number of Weighings) suspected a priori
to influence concentration.  Finally, Section V presents a
discussion of the regression procedures.

II.  ANALYSIS OF CONCENTRATION DATA

     A. Statistical Methodology

     Since only  22 values  for airborne dye concentration were
available,  and it was desired to estimate upper percentiles of
concentration,  a model-based analysis of the data was considered the
most appropriate approach.   In a model-based approach, the lognormal
     *Dye concentrations represent the total amount found on the  air
filter and its associated plastic cassette.  Dust  concentrations
represent the amount  found on the air filter alone.
                                                                  7-1

-------
                             Table 7-1

    GRAVIMETRIC WEIGHT OF TOTAL DUST PER VOLUME OF AIR SAMPLED

                                (mg/m3)
                 Personal Filters*
Area Filters
Site
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6**

7/7
7/9
8/0
8/6
8/8
9/1
(A)
0.75
0.27
1.07
0.30
0.48
0.47
0.18
0.26
0.74
0.62
0.37
1.71
0.33
0.02
0.44
0.30
0.52
1.08

0.29
0.20
0.22
0.49
0.52
0.45
(B)
0.70
0.23
0.82
0.33
0.44
0.35
0.09
0.33
0.73
0.24
0.57
1.04
0.40

0.22
0.25
0.53
0.49

0.26
0.18
0.20
0.62
0.55
0.45
Average
0.73
0.25
0.94
0.31
0.46
0.41
0.13
0.29
0.73
0.43
0.48
1.37
0.36
0.02
0.33
0.27
0.53
0.77

0.27
0.19
0.21
0.55
0.53
0.45
Weigh
Station
0.31
0.15
0.17
0.15
0.23
0.30
0.05
0.16
0.59
0.09
0.47
0.47
0.30
0.04
0.12
0.15
0.14
0.07
0.36
0.08
0.11
0.13
0.11
0.16
0.28
Remote
Area
0.19
0.04
0.14
0.10
0.25
0.09
0.05
0.21
0.19
0.05
0.11
0.12
0.15
0.04
0.05
0.04
0.14


0.06
0.10
0.16
0.23
0.16
0.18
 * (A),  Personal Canister A; (B), Personal Canister B.
**The area monitors were placed at two separate weigh stations in
  the two adjacent drug rooms.   The dye weigher weighed dyes at
  both weigh stations during the shift.
7-2

-------
                              Table 7-2

 SPECTROPHOTOMETRIC ESTIMATES OF THE AVERAGE AIRBORNE CONCENTRATION
                          OF COMMERCIAL DYE

                                (mg/m3)
Personal Filters*
Site
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6
7/7
7/9
8/0
8/6
8/8
9/1
(A)
0.46
0.16
0.02
0.08
0.08
0.27
0.02
0.18
0.12
0.08
0.09
1.43
0.07
0.03
0.08
0.22
0.09


0.01
0.05
0.47
0.10
0.20
(B)
0.46
0.16
0.02
0.15
0.09
0.35
0.02
0.14
0.14
0.04
0.08
0.97
0.07
0.03
0.05
0.14
0.11


0.01
0.03
0.56
0.10
0.24
Average
0.46
0.16
0.02
0.12
0.08
0.31
0.02
0.16
0.13
0.06
0.09
1.20
0.07
0.03
0.07
0.18
0.10


0.01
0.04
0.51
0.10
0.22
Area Filters
Weigh
Station
0.20
0.12
0.11
0.01
0.07
0.30
0.01
0.07
0.37
0.04
0.15
0.30
0.25
0.02
0.07
0.08
0.04


0.01
0.03
0.09
0.11
0.16
Remote
Area
0.10
0.00
0.00
0.01
0.06
0.05
0.00
0.07
0.09
0.01
0.02
0.05
0.09
0.01
0.01
0.01
0.01


0.00
0.04
0.16
0.03
0.01
*(A),  Personal Canister A; (B),  Personal Canister B.
                                                                  7-3

-------
                             Table 7-3

 SPECTROPHOTOMETRIC ESTIMATES OF THE AVERAGE AIRBORNE CONCENTRATION
                         OF ACTIVE COLORANT

                                (mg/m3)
                 Personal Filters*
Area Filters
Site
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6
7/7
7/9
8/0
8/6
8/8
9/1
(A) '
0.22
0.09
0.01
0.05
0.04
0.09
0.01
0.09
0.06
0.04
0.06
0.66
o.oe
0.01
0.02
0.07
0.07


0.01
0.01
0.17
0.02
0.09
(B)
0.22
0.09
0.01
0.10
0.05
0.11
0.01
0.07
0.07
0.02
0.06
0.45
0.05
0.01
0.02
0.04
0.09


0.01
0.01
0.21
0.02
0.10
Average
0.22
0.09
0.01
0.08
0.04
0.10
0.01
0.08
0.06
0.03
0.06
0.56
0.06
0.01
0.02
0.06
0.08


0.01
0.01
0.19
0.02
0.09
Weigh
Station
0.09
0.07
0.06
0.01
0.03
0.09
0.00
0.03
0.18
0.02
0.10
0.14
0.19
0.01
0.02
0.03
0.03


0.00
0.01
0.03
0.03
0.07
Remote
Area
0.05
0.00
0.00
0.01
0.03
0.01
0.00
0.03
0.04
0.01
0.01
0.03
0.07
0.00
0.01
0.00
0.01


0.00
0.01
0.06
0.01
0.01
Note:  Spectrophotometric analysis calculated on the basis of the
       best estimate of dye purities.
 *(A), Personal Canister A;  (B), Personal Canister B.
7-4

-------
distribution is considered a good approximation to the data, as  in
many environmental studies.  Using the mean and standard deviation
of the data collected at the 22 plants, the upper percentiles were
calculated on the basis of the fitted distribution.  For simplicity,
model determination was done using the raw, unweighted data, as
discussed in Section V of this chapter.  This approach greatly
simplifies the statistical calculations and is considered to be  an
excellent approximation (since, as discussed in Chapter 4, the
maximum disparity between plant-level sampling weights was at most
roughly a factor of 2:  28.8 for the "respondent" stratum and 61.7
for the "nonrespondent" stratum; worker-level weights were only
slightly more variable).

     Figure D-l in Appendix D  shows a normal probability plot of the
airborne commercial dye concentration data from Table 7-2.  A
Kruskal-Wallis test2  rejected the  hypothesis  that  the  data  came  from
a normal distribution, with a p-value less than 0.01.   Indeed, the
plot shows the characteristic concave shape that one would expect if
the underlying distribution were lognormal rather than normal.  This
is confirmed by Figure D-2 in Appendix D, which shows a normal
probability plot of the (natural)  logarithm of the data.   The plot
closely approximates a straight line,  indicating that the logarithm
of the data is normally distributed, i.e., the data itself is
lognormally distributed.  This is confirmed by a Kruskal-Wallis test
of the log data, which does not indicate any significant departure
from normality, with p = 0.98.   Similar results were obtained for
the active colorant data.

     It was concluded, based on the above considerations, that the
data for both active colorant and commercial dye follow a lognormal
distribution.   This distribution is characterized by the two
parameters m and s.  If Y follows a lognormal distribution, then m
is the mean value of log(Y) and s is the standard deviation of
log(Y).  The mean value of Y itself is:

                      v = exp  (m + s2/2)    ,

while its standard deviation is:

                      u = v(exp(s2)  -  I)0'5

The median of Y is exp(m), while the gth percentile of Y is
exp(m + zs), where z is the gth percentile of the standard normal
distribution.   These formulas were applied as follows.  First, a
     2Daniel WW.  1978.  Applied Nonparametric Statistics.  Boston:
Houghton Mifflin, p.  200.
                                                                  7-5

-------
standard statistical software package, the Statistical Analysis
System (SAS),3 was used to compute weighted means and standard
deviations on the log scale for active colorant and commercial dye
concentration at both the plant and worker level.  The weights used
were those described in Chapter 4.  In other words, the assumed
lognormal distribution was fitted using the appropriate weights for
the data.  The above formulas were then applied to estimate mean,
median, standard deviation, and 85th, 90th, and 95th percentiles on
the scale of the original measurements.

     Next, it  was  of interest to  compute  approximate  confidence
intervals for the mean concentration and  for the upper percentiles
of concentration.  The details of the computation of these
confidence intervals are provided in Appendix D.

     B.  Results

     Tables  7-4,  7-5,  7-6, and  7-7  summarize the  results of applying
statistical analysis to the airborne dye  concentration data.
Recalling that airborne dye concentration estimates represent a
composite of all dyes used at plants monitored and not any
individual dye, the following observations can be made.  First,
there is little difference between estimated concentration at the
plant and worker level for either active  colorant or commercial dye.
Estimates weighted by plant are very slightly higher, perhaps
indicating that plants with more weighers tend to exhibit slightly
lower concentrations; however, this effect could also easily be
explained by  sampling variability alone.  Secondly, airborne dye
concentrations for c6mmercial dye are consistently roughly twice the
corresponding values of active colorant.  The specific factor
difference between concentration  for active colorant and commercial
dye can be interpreted as saying that the average strength of
commercial dye preparations is approximately 50%.  Thirdly, the
estimated confidence intervals for the high percentiles do not
appear to be  so wide as to preclude the use of these percentiles for
regulatory purposes.

     The estimates contained  in Tables  7-4  through 7-7 can also be
calculated using  simple unweighted estimates, as shown in Tables 7-8
and 7-9.  A comparison was made between the weighted and unweighted
estimates to provide a check of the possible nonresponse bias that
was noted as  a cause of concern in Chapter  4.  Comparison
demonstrated  that  the unweighted  values were close to the weighted
estimates for  all  variables in the tables; they were generally
     3SAS Institute, Inc.  1985.  SAS User's Guide, Version 5.   Gary,
 NC:  SAS Institute,  Inc.
 7-6

-------
                              Table 7-4

       AIRBORNE CONCENTRATION OP TOTAL COMMERCIAL DYE BY PLANT
Concentration
(mg/m3)
Median
Geometric Standard Deviation
Mean
Standard Deviation
Percentiles
85th
90th
95th
0.11
2.80
0.18
0.26
0.31
0.39
0.57
95% Confidence
Interval

(0.11, 0.31)
(0.18, 0.53)
(0.22, 0.70)
(0.29, 1.11)
                              Table 7-5

     AIRBORNE CONCENTRATION OF  TOTAL COMMMERCIAL DYE BY WEIGHER
                                Concentration    95% Confidence
                                   (mg/m3)           Interval
Median                              0.10
Geometric Standard Deviation        2.90

Mean                                0.17          (0.10,  0.30)
Standard Deviation                  0.25

Percentiles
     85th                           0.30          (0.17,  0.52)
     90th                           0.38          (0.21,  0.70)
     95th                           0.57          (0.29,  1.14)
                                                                  7-7

-------
                              Table 7-6

      AIRBORNE CONCENTRATION OF  TOTAL ACTIVE COLORANT BY PLANT
                                Concentration    95% Confidence
                                   (mg/m3)           Interval
Median                              0.049
Geometric Standard Deviation        2.849

Mean                                0.085         (0.049, 0.147)
Standard Deviation                  0 .12

Percentiles
     85th                           0.15          (0.086, 0.26)
     90th                           0.19          (0.10,  0.34)
     95th                           0.27          (0.14,  0.53)
                              Table 7-7

     AIRBORNE CONCENTRATION OF TOTAL ACTIVE COLORANT BY WEIGHER
                                Concentration    95% Confidence
                                    (mg/m3)           Interval
Median                              0.042
Geometric  Standard Deviation       3.075

Mean                                0.079          (0.043, 0.145)
Standard Deviation                 0.12

Percentiles
      85th                           0.13           (0.072, 0.23)
      90th                           0.18           (0.095, 0.34)
      95th                           0.27           (0.13,  0.56)
 7-8

-------
                    Table  7-8

 AIRBORNE CONCENTRATION OP TOTAL COMMERCIAL DYE
             (UNWEIGHTED ESTIMATES)
Concentration
(mg/m3)
Median
Geometric Standard Deviation
Mean
Standard Deviation
Percentiles
85th
90th
95th
0.10
3.03
0.19
0.29
0.32
0.42
0.63
95% Confidence
Interval

(0.11, 0.34)
(0.18, 0.57)
(0.22, 0.79)
(0.31, 1.29)
                   Table 7-9

AIRBORNE CONCENTRATION OF TOTAL ACTIVE COLORANT
             (UNWEIGHTED ESTIMATES)
                     Concentration    95% Confidence
                         (mg/m3)           Interval
Median
Geometric Standard Deviation
Mean
Standard Deviation
Percentiles
85th
90th
95th
0.046
3.167
0.089
0.15

0.15
0.20
0.31


(0.048,


(0.082,
(0.10,
(0.15,


0.17)


0.27)
0.27)
0.65)
                                                        7-9

-------
higher than the weighted estimates, indicating that the
nonrespondent stratum exhibited slightly lower average concentration
levels than the respondent stratum.  This mitigates a potential
source of nonresponse bias, i.e., the concern that plants with
higher concentration levels might have been less likely to cooperate
with the study.

     The fitted lognormal  distributions of airborne commercial  and
active colorant concentrations are given in Figures 7-1 and 7-2, and
parameters of the distributions such as the median, mean, and
selected percentiles are also indicated.  The same information  is
displayed in tabular form  in Tables 7-10 and 7-11, which present
estimates of the total number of weighers exposed to various levels
of active colorant and commercial dye concentrations, respectively.
From Table 7-10 there are  134 weighers  (5% of the estimated weigher
population) exposed to active colorant concentrations which exceed
0.27 mg/m3;  and the same number of weighers are exposed to active
colorant concentrations between 0.18 and 0.27 mg/m3.   Half of the
weighers (an estimated 1,345 persons) are exposed to active colorant
concentrations which exceed 0.042 mg/m3.   Corresponding estimates on
the commercial dye basis in Table 7-11 are:  134 weighers  (5%)  are
exposed to commercial dye  concentrations greater than 0.57 mg/m3;
another 134 weighers are exposed to concentrations between 0.38  and
0.57 mg/m3;  half of the population (estimated to be 1,345 persons)
are exposed to concentrations greater than 0.10 mg/m3.

III.  SUMMARY OF  CROSS-TABULATIONS
                      >
      In addition  to  tabulating the airborne  dye concentration  for
the plants as a whole, the population was broken down into groups
defined by variables collected  during monitoring.  These variables
are among those collected  on the in-plant questionnaire form  (see
sample in Section E  of the Supplement).  Data on variables such as
the type of management of  the plant, number of  dyeing machines,  type
of shifts used, etc., were collected; it is interesting to see  the
airborne dye concentrations tabulated according to the groups so
defined.  A complete listing of these tabulations is shown in
Appendix B.

      It  is  important to note,  at the outset  of this  discussion, that
the sample size  (though  quite  adequate  for total industry estimates)
will  substantially limit the precision  of estimates of subgroups of
the population.   For that  reason,  no attempt will be made to perform
statistical tests of significance  for the differences between the
values found.  Rather,  they will be  considered for the general
information which can be obtained.
 7-10

-------
       8



I
o>     6
i
I     5
0)
                               Figure 7-1

               AIRBORNE COMMERCIAL DYE CONCENTRATION



                    Fined Lognormal Probability Density
=    2
A
S
I    1
   MEDIAN=0.10
       7
  1345
Weighers
 below
  0.10
         MEAN-0.17
             V
                             85%S0.30
                                V
                           90%S0.38
                              V
            1076 Weighers
                                                   95%S0.57

                                         ,  134 Weighers'
                                                                  134 Weighers over 0.57
               0.2
                                        0.4             0.6

                                    CONCENTRATION (mg/m3)
                                                                        0.8
1.0
The relative frequencies of total commercial dyestuff concentrations are shown. Note that of an
estimated 2688 dye weighers, 134 weighers experience average concentrations above 0.57 mg/m3,
268 weighers experience concentrations above 0.38 mg/m3, and 1345 weighers experience
concentrations above 0.10 mg/m3. There may be more weighers than shown here if the sampling
list missed some plants.
                                                                           7-11

-------
o>
I
o
o>
I
0)
0)
o
I
20
19
18
17
16
15
14
13
12
11
10
 9
 8
 7
 6
 5
 4
 3
 2
 1
 0
                                       Figure 7-2

                         AIRBORNE ACTIVE DYE CONCENTRATION


                            Fitted Lognormal Probability Density
IMEDIAN-0.042
     \y
           1345
         Weighers
          below
          0.042
       MEAN-0.079
                     1076 Weighers
                         90%£0.18
                                       95%S0.27
                              134 Weighers T-
                        0.1
                                  0.2
134 Weighers over 0.27
                                                               0.4
                       0.5
                                    CONCENTRATION (mg/m3)
         The relative frequencies of total active colorant concentrations are shown. Note that of an estimated
         2688 dye weighers, 134 weighers experience average concentrations above 0.27 mg/m3,268
         weighers experience concentrations above 0.18 mg/m3, and 1345 weighers experience
         concentrations above 0.042 mg/m3. There may be more weighers than shown here if the sampling
         list missed some plants.
    7-12

-------
                             Table 7-10

                  NUMBER OF WEIGHERS WHO EXPERIENCE
       VARIOUS WORKPLACE CONCENTRATIONS—ACTIVE COLORANT BASIS
Concentration Range                              Number  of Weighers
      (mg/m3)              Percentile Range         in  Given Range
Less than 0.042                Oth -  50th               1,345

0.042 - 0.087                 50th -  75th                 672

0.087 - 0.13                  75th -  85th                 270

0.13  - 0.18                  85th -  90th                 134

0.18  - 0.27                  90th -  95th                 134

Greater than 0.27             95th and  above               134



                             Table 7-11

                  NUMBER OF WEIGHERS  WHO EXPERIENCE
       VARIOUS WORKPLACE CONCENTRATIONS—COMMERCIAL DYE BASIS
Concentration Range                             Number of Weighers
      (mg/m3)              Percentile Range        in Given Range
Less than 0.10                 Oth  -  50th                1,345

0.10  - 0.20                  50th  -  75th                  672

0.20  - 0.30                  75th  -  85th                  270

0.30  - 0.38                  85th  -  90th                  134

0.38  - 0.57                  90th  -  95th                  134

Greater than 0.57             95th  and above               134
                                                                 7-13

-------
     The tabulations are shown in Table B-l of Appendix B.  Several
important types of estimates can be found in the table.  First there
are estimates of the numbers of plants and weighers in the
population of interest.  From Table B-l and based upon the survey,
it is estimated that there are a total of 863 textile dyeing plants
regularly weighing powder dyes.  In addition, the results of the
survey indicate that an estimated total of 2,688 weighers are
engaged in weighing powder dyes at those establishments.  These two
totals are shown broken down by a number of auxiliary variables in
Table B-l.

     In addition to estimates  for numbers of weighers  and plants,
Table B-l also contains, estimates for the commercial dye
concentrations  (weighted by both plant and worker weights) and the
associated standard errors of the estimate.  The standard error of
the estimate is a measure of the precision of the estimate and is a
function of both the variability found in the concentrations and the
size of the subgroups.  Since the number of plants represented in
each group is small, relatively large standard errors are to be
expected.

     To illustrate the use of  the tables, two variables from the
tables are reproduced  in this section as Table 7-12.  Included in
that table are results for the variables Type of Ownership  (private
versus public) and Number of Dyes Weighed  (in four categories).
Consider first the case of the type of ownership.  Note that
estimates for the airborne dye concentrations are almost identical
in the two subgroups of vthe population.  For the plant weighted
data, the group of privately owned establishments had an airborne
dye concentration of 0.19 mg/m3 compared to a value of 0.17  mg/m3 for
the publicly held establishments, indicating that the variable Type
of Ownership is not a  factor that explains differences in airborne
dye concentration.

     In contrast,  consideration  of the variable  Number of Dyes
Weighed gives a different picture.  As can be seen from Table 7-12,
for a group of plants  weighing larger numbers of dyes there is a
steady increase in that group's mean airborne dye concentration  (it
increases from  0.046 mg/m3 for the plants weighing 0 to 10 dyes to
0.52 mg/m3 in the group weighing 31 to 46 dyes) .   The small  sample
sizes in subgroups suggests that making statistically based
comparisons is  unwise; however,  it is still interesting that the dye
concentration increases as the number of dyes weighed  increases.

      There  are no  major  findings of  a statistical  nature  in these
tables, and therefore, they are  not  included in  total  in the body of
the  report.  However,  Appendix B contains tabulations  similar to
those shown in  Table  7-12 for  11 cross-tabulation variables.
 7-14

-------
                             Table 7-12

         COMMERCIAL DYE CONCENTRATION ESTIMATES BROKEN DOWN
            BY OTHER VARIABLES  MEASURED DURING MONITORING
                      	Variable	

                         Ownership         Number of Dyes Weighed

                      Private  Public   0-10   11-20   21-30   31-46
Number of
Sampling Cases*
   12
10
Estimated Universe

  No. of Plants         444       419

  No. of Workers      1,529     1,159
                    144     391    243      86

                    547   1,225    715     201
Plant-Weighted

  Total Commercial
   Dye Concentration
   in Dye Weighing
   Room Air (mg/m3)

  Standard Error
 0.19     0.17    0.046   0.083   0.30    0.52

0.053    0.096    0.013   0.014  0.077    0.28
Worker-Weighted

  Total Commercial
   Dye Concentration
   in Dye Weighing
   Room Air (mg/m3)

  Standard Error
 0.18     0.17    0.040   0.077   0.33    0.61

0.056     0.10    0.012   0.014  0.066    0.32
   *Includes only the 22 plants with a valid concentration
measurement.
   Note:  This table'is an abridged version of Table B-l in
Appendix B.
                                                                7-15

-------
IV.  CORRELATION OF CONCENTRATION WITH VARIOUS FACTORS

     As a first step in analyzing the data,  correlation  coefficients
between airborne dye concentrations and various factors potentially
influencing concentration were calculated.  To mitigate the
potential effect of outliers on the analysis and to detect nonlinear
associations, the Spearman rank correlation coefficient4  was  used as
a convenient quantitative measure of the relationship between
airborne dye concentration and other factors.  Table 7-13 shows
Spearman correlations between active and commercial dye basis and 14
factors of interest.  An asterisk (*) indicates that an estimated
correlation was statistically significantly different from zero
under a two-sided test.of significance at the 5% level.  Under the
assumption of no correlation, the magnitude of the correlation for
the survey data would represent an event which would happen only 5%
of the time or less.  This is a commonly used level at which to
accept the proposition that the correlation is not likely to have
occurred by chance alone.  Note that, although a number of factors
other than those listed were measured in the study, none of the
Spearman correlations with airborne dye concentration were close to
being statistically significant.

     The  same  five  factors exhibit  statistically  significant
correlations with dye concentration for both commercial dye and
active colorant.  All other factors have much smaller estimated
correlations, none of which approach statistical significance.  The
factor Number of Suppliers, with the highest correlation with dye
concentration, measures, the number of suppliers the monitored plant
has for those dyes weighed during the monitoring period.  The high
correlation exhibited was rather unexpected  and has no apparent
plausible basis.

     The  other four factors  with significant correlations with  the
airborne  dye concentration were  expected.  In particular, Mass of
Dye Weighed and Number of Weighings had been expected
a  priori  to be the most highly correlated with airborne  dye
concentrations.  Two other factors—Number of Dye Classes and Number
of Dyes Weighed—exhibit high correlations in Table 7-13.  Their
correlation with airborne dye concentration  might be a surrogate  for
some other variables,  such as Number of Weighings.
      4Johnston J.   1972.  Econometric Methods,  2nd Edition.  New York:
 McGraw-Hill, p. 219.
 7-16

-------
                              Table  7-13

      SPEARMAN CORRELATIONS BETWEEN AIRBORNE DYE  CONCENTRATION
                  AND  SELECTED EXPLANATORY FACTORS
                                                Correlation with:
Factor
Number of Suppliers
Number of Dyes Weighed
Mass of Dye Weighed
Number of Weighings of Dyes
Number of Dye Classes
Number of Fibers in Final Product
Number of Minutes Weigher Was in Drug Room
Production Volume of Product
Age of Person Monitored
Average Humidity in Drug Room
Average Temperature in Drug Room
Number of Shifts Site Operates
Number of Entries into Drug Room by Weigher
Number of Minutes Weigher Was Monitored
Commercial
Dye
0.76*
0.71*
0.64*
0.62*
0.51*
0.30
0.29
0.29
0.26
0.22
0.17
0.16
-0.11
0.05
Active
Colorant
0.68*
0.64*
0.60*
0.58*
0.51*
0.30
0.08
0.19
0.24
0.28
0.24
0.10
0.05
-0.02
   *Statistically significant at 0.05 level
V.   RELATION BETWEEN AIRBORNE DYE CONCENTRATION AND OTHER VARIABLES

     The first step  in this  analysis was the development of  "best"
regression equations5 of  airborne dye  concentration based on  the
study data.  A forward stepwise regression procedure was used, under
which variables are introduced to the model based on their
incremental contribution to its explanatory power.   The final
equation arrived at by this procedure is "best" in the sense of
maximum explanatory power with a minimum number of variables.
Explanatory power per se can be misleading if too many independent
variables are used.  With 22 data points a perfect fit but useless
     5See Draper N, Smith H.   1981.  Applied Regression Analysis,
Second Edition,  Section  6.  New York:   Wiley.
                                                                 7-17

-------
equation could be obtained by using any 21 independent variables
plus a constant term.  The natural logarithm of airborne dye
concentration was examined closely for use as the scale of the
dependent variable because it provides a better model fit and
reduces the potential over-influence of outliers.  For the sake of
simplicity, the data were unweighted.  This was considered
reasonable because the results of weighted and unweighted analyses
reported in Section B are so close.  Finally, all equations were
developed using commercial dye basis, because this was the measure
of greater potential usefulness for the complex dye formulations
assessed, and was what was directly measured by the analytical
methodology used in the study.

     The stepwise  regression procedure was  first applied with the
set of explanatory variables in Table 7-13, using both measurement
and log scales for these variables.  Then the analysis was repeated,
omitting Number of Suppliers, because this was not a logical
parameter in an exposure assessment.  Finally, both analyses were
repeated, omitting data from the plant corresponding to the
observation from Plant 4/9 in Tables 7-1, 7-2, and 7-3.  The
airborne dye concentration for this plant was by far the highest
observed and there was concern it might be an unusually influential
observation in the regressions.  The results indicated no
substantial differences whether plant 4/9 was included or not.  The
results with site  4/9 included are shown in Table 7-14.

     When  Number of  Suppliers  is omitted from the explanatory
variable set, the  stepwise regression always picks a single-variable
equation, but the variable chosen depends on the scale of the
explanatory variables.  The equation with the highest R2 value is:

            log (estimated airborne dye concentration)   =
               a + b (log of dye mass weighed out)

That is, the airborne dye concentration  is best predicted by a
linear relationship using the log of the dye mass weighed out as the
sole predictor.  The estimates for the coefficients are given below,
along with their standard errors in parentheses:

                           a  = -4.13  (0.55)
                           b  =  0.54  (0.15).

This means that if 10 kg of dye were weighed out during a shift, the
average  airborne dye concentration during the 8-hr shift was found
to be 0.056 mg/m3.   The R2 value is 0.39, indicating that 39% of the
original variation in the data was explained by this equation.
 (Note that some of the variation,  such as that  due to  sampling and
laboratory variability, cannot be  explained by  the explanatory
 7-18

-------
                             Table 7-14

                  RESULTS OF STEPWISE REGRESSION OF
                   log(Airborne Dye Concentration)
               AGAINST EXPLANATORY FACTORS  (ALL DATA)
Include
Number of
Suppliers
Yes
Yes
No
No
Scale of the
Explanatory
Variables
Natural Log
Measurement
Natural Log
Measurement
Selected
Equation
log (Number of Suppliers)
Number of Suppliers
log (Mass Weighed)
Number of Dyes Weighed
R2
0.52
0.56
0.39
0.38
variables.)  Plots of residuals from this regression versus
predicted values and other possible variables revealed no unusual
features.  This can be taken as an indication that the one-variable
equation is adequate.

     When Number of  Suppliers is included in the stepwise
regression, the procedure selects a single-variable equation with
that variable as the best (or its natural log when looking at
logarithms of the predictors as inputs).   The equation with the
highest R2  is  that  with Number of Suppliers  (untransformed)  as  a
predictor:

            log (estimated airborne dye concentration)   =
                   a + b (Number of Suppliers)

The estimates for the coefficients are given below, along with their
standard errors in parentheses:

                     a = -4.03  (0.38)
                     b =  0.33  (0.065)

     For this  case the value of R2 is 0.56,  meaning that the
explanatory variables explain 56% of the variability of airborne dye
concentration.  Although inclusion of this variable in the stepwise
                                                                 7-19

-------
equation improves the R2,  this  variable's  relationship to  airborne
dye concentration is difficult to explain.

     In addition to the single-variable models selected by the
stepwise regression procedure,  regression models with two
explanatory variables were also examined.   Of the five variables
with significant correlations in Table 7-13, Mass of Dye Weighed and
Number of Weighings of Dyes are two variables which have a direct
causal relationship with airborne concentrations.  Hence,  2-variable
regression models were estimated for these variables, first using
the original data and again using logarithms of the data.   The
results of these regressions are shown in Table 7-15.

     The linear model was  specified as follows:

        Commercial Dye Concentration =  a + b  (mass weighed)
                                         + c (number of weighings)  .

Estimates of the coefficients a, b, and c are shown in Table 7-15,
with the standard errors of the estimates included in parentheses.
The specification of the log-log regression model in Table 7-15 was
similar to that for the linear model, except that logarithms of all
variables were used.
                             Table 7-15

             RESULTS OF,'TWO-VARIABLE REGRESSION ANALYSIS
                  FOR COMMERCIAL DYE CONCENTRATIONS
                    (DATA FROM PERSONAL MONITORS)
                                 Coefficient Estimates
                                  (and Standard Errors)
Type of Model
Linear
Log-Log
Intercept
R2 (a)
0.39 -0.0010
0.47 -5.361
Mass Weighed
(b)
0.0020
(0.0007)
0.400
(0.165)
Number of
Weighings
(c)
0.0011
(0.0008)
0.447
(0.256)
 7-20

-------
      In both the  linear and log-log regressions, the estimated
coefficients for the Mass Weighed variable are highly significant,
since these estimates are larger than twice the standard error.   The
estimated coefficient for the Number of Weighings variable is not
significant in the linear model.  With a t-statistic of 1.74, this
coefficient is not significantly different from zero at the 0.05
probability level in the log-log model.  However, at the 0.10
probability level, there is evidence that the coefficient is
significantly different from zero.  Hence, there is only marginal
evidence for including Number of Weighings in the model.

      The R2 of the linear regression model with only Mass Weighed as
a single variable is 0.34, and inclusion of the Number of Weighings
variable increased the R2  to 0.39.   In  the log-log  regression model,
the R2 increases  from 0.39 for  the single-variable  model  with Mass
Weighed (discussed previously)  to 0.47 when Number of Weighings  is
included.   The increase of R2 due  to the Number  of  Weighings
variable is insignificant at the 0.05 level in the linear model.
Based on the regression F test, the increase in R2  has only marginal
significance when the second variable is added in the log-log form
of the model.

      As previously noted, residual analysis of the log-log model
with Mass of Dye Weighed as the only explanatory variable indicated
that the model is "adequate";  that is,  no other variable provides
significant additional information for predicting airborne dye
concentration.   Furthermore,  using a jackknife procedure which
examined the ability of each model to predict individual site values
when these sites are omitted from the data set,  the 2-variable model
with Mass of Dyes Weighed and Number of Weighings provided only
slightly higher predictive power versus the Mass of Dyes Weighed
model alone.  (The predictive performances were compared using the
mean absolute predictive error and the mean square predictive error
as criteria.)

      To summarize, the set of explanatory factors for concentration
on which data was collected in the study appears to have moderate
explanatory power.  A (forward) stepwise regression procedure always
picks single-variable equations under a variety of scenarios, with
the best of these simple equations explaining 56% of the variability
in the airborne dye concentration data.  However, the equation with
the highest R2 is  for Number of Suppliers,  a variable which cannot
logically be expected to have a causal relationship with airborne
dye concentration.  When Number of Suppliers is omitted from the
stepwise regression procedure,  a log-log equation with Mass Weighed
as the explanatory variable is selected by the stepwise procedure.
                                                                 7-21

-------
For this regression the functional relationship between the
variables is

                 log (airborne dye concentration)   =
           -4.13 + (0.54)(log of dye mass weighed, in kg),

or, equivalently,

                    airborne dye concentration  =
         exp[-4.13  +  (0.54)(log of  dye  mass  weighed,  in  kg)].

Residual analysis  indicated that this  1-variable  explanatory was
adequate.  This equation has an R2 of 0.39,  and the chosen variable
has a clear causal relationship with airborne dye concentration.
 7-22

-------
                              Chapter 8

                  INDIVIDUAL SITE CHARACTERIZATIONS
I.   INTRODUCTION

     Information was collected during the 24 site visits to obtain  a
better understanding of the operations associated with the use of
powder dyes in the textile industry.  At each site,  the information
collected included the facility s physical and operational
characteristics, the dye weighers' activities,  and the engineering
controls and personal protective equipment used to limit worker
exposure to chemicals and dyestuffs.

II.  FACILITY CHARACTERISTICS

     A. General Characteristics

     Dyehouses may be owned by the textile manufacturer  (vertical
basis), or operate independently  (commission basis).   Vertical
dyehouses may also operate on a commission basis.  Most of the
facilities surveyed (16 of 24) operated solely on a vertical basis.
Fourteen of the facilities were privately-owned and 10 were publicly
owned.   The annual production volume of the facilities ranged from
300,000 pounds to 25 million pounds of various textile products.
While processes included batch dyeing, continuous dyeing, and
printing operations, most of the facilities monitored (18 of 24)
were engaged in batch dyeing.  Dyeing equipment included machines in
which the dye liquor was transported through the stationary textile
substrate (package, beam, skein,  and stock machines); machines in
which the textile material was transported through essentially
dormant dye liquor  (jig, beck, and pad machines); and machines in
which both the dye liquor and the material were in motion (jet,
paddle, rotary, and some skein machines).   The number of dyeing
machines available for operation at each site varied from 1 to 75,
and all but one of the sites were operating at 50% or greater
capacity during the monitoring period.

     Fibers dyed during  the monitoring period were not recorded.
Site end products were composed of acrylic/modacrylic, acetate,
rayon,  wool, nylon, polyester, cotton, silk, and flax fibers,  either
neat or as blends.  Variations in fiber suppliers and types were not
noted.   Almost all of the sites dyed more than one fiber; one site
reported up to nine fibers in the various end products.  Product
lines of the facilities  surveyed included raw stock  (staple);
apparel, space dyed'and  carpet yarns; hosiery and intimate wear;
woven and knitted apparel fabrics for indoor and/or outerwear uses;
                                                                  8-1

-------
automotive and upholstery fabrics; garments, woven industrial
fabrics; and carpets.  Table A-2 in Appendix A presents a summary of
the specific facility characteristics recorded during the monitoring
period, and Table A-3 summarizes these characteristics.  Table A-4
summarizes the number of fibers processed annually at a facility and
the number of dye classes encountered during the monitoring period.

     B. Process  Characteristics

           1.   Drug Room Operations

     For  purposes  of this  study,  the area in  which powder dyes  were
weighed was considered the  "drug  room" of concern  at each facility.
The drug  room was usually well-defined within a walled area, but at
some sites the lack  of'physical barriers  necessitated a definition
of the drug room boundaries by the survey team.  At most  facilities,
dyes were stored in  the  area where they were weighed.  Generally,
the drug  rooms where dyestuffs and chemicals were  stored  and weighed
were rooms or areas  separated from areas  where mixing equipment and
dye machines were  located.  Mixing facilities were located within
the drug  room at six sites.  The  physical characteristics of two
sites  were significantly different than those of the other 22
sites.1  Dyes were stored in closed drums varying in size from 50-
to 100-pound containers, to 200-  to 275-pound barrels.  At some
sites,  liquid dyes  and dry  chemicals were also stored and weighed in
the drug  room.   At  three sites, drum storage  areas were separate
from the  drug room.   Open dye drums were  usually located  on the
floor  and along  the walls of the  room with  the lids loosely in
place;  however,  at  many  sites, frequently used drums of dyes were
placed in an area  close  to  the weighing station.   The weighing
station included at least one and often several scales  for weighing
the dyes  or  chemicals.

      Some of the drug rooms were temperature- and humidity-
controlled by ventilation  systems independent of the general
facility  ventilation system.  Most of those observed during the
surveys,  however,  had either no  controlled ventilation  or only  the
passive ventilation provided by  the general building or area
ventilation  system.   At  one of the sites  visited,  the drug room was
      *At Site 8/0,  the  dye weighing station was contained within the
 production area where mixing and dyeing  occurred.  The entire
 production area was considered the drug  room at that facility for
 purposes  of  recording the dye weigher's  time in the drug room.  At
 Site 6/6, two drug rooms were in operation  and both were used by the
 dye weigher  monitored at that site.
 8-2

-------
previously used as a vault; thus the only ventilation available in
that drug room was through the door, which was kept open.

     About half of the drug rooms were relatively clean of visible
accumulated dye materials; at 11 of the sites visited, a significant
accumulation of dye was observed on the drug room floor, dye drums,
and equipment, indicative of dye spillage.  The amount of
accumulated dye particulates within the drug room did not appear to
be related to the number of dye weighings performed or the amount of
dye weighed on the day of the survey.  The accumulation of dye in
these drug rooms may contribute to the amount of dyestuff collected
on the filter.

           2.   Dye Weigher Activities

     Most of the facilities visited  (13) operated under a three 8-hr
work shift schedule.   Five facilities operated on two 12-hr shifts,
three operated two 8-hr shifts, and only three operated a single
work shift of 8-10 hr.  Monitoring was conducted on the first work
shift (day) at 12 facilities, on the second work shift (evening)  at
8 facilities, and on the third work shift (night)  at 3 facilities.
Parts of the first and second shifts were monitored at one of the
sites (see Chapter 1, Section 8).   At facilities where the dye
weighers' responsibilities were limited to weighing dyes (and
sometimes chemicals), one or two dye weighers were typically
employed per shift.   At other facilities, dye weighers also mixed
the weighed dyes.  At some sites,  dye machine operators also had the
responsibility for weighing and mixing dyes.  Table A-5 in Appendix
A presents individual shift characteristics, and Table A-6
summarizes these characteristics.

     Duties of the dye weigher typically included transporting full
and empty drums both into and out of the storage area; relocating
drums within the storage area and drug room and recording the amount
weighed on batch tickets; transferring the weighed dye from the
scale pan to a transfer container, usually a metal bucket,  and then
transferring the container of weighed material to the mixing area;
and cleaning the dye storage and weighing areas.

     The amount of time that the dye weigher spent in the drug room
varied according to his assigned duties.  The time the dye weigher
spent in the drug room was reduced if some of the worker's assigned
job responsibilities took place in other areas of the plant.  The
monitored dye weighers spent less than 25% of the monitoring period
in the drug room at six sites, and more than 25% but less than 50%
of the monitoring period at seven sites.  The dye weighers at the
remaining sites spent more than 50% of the monitoring period in the
drug room.  At one site the dye weigher spent 99% of the time in the
                                                                  8-3

-------
drug room.  At the four sites where the personal air monitoring
results indicated the highest concentrations of both formulated dye
and active colorant in worker's breathing zones, the dye weighers
spent more than 60% of the respective monitoring periods inside the
drug room (84%, 95%, 84%, and 62%, respectively).   However, no
correlation between exposure and time in the drug room was found.
Table A-7 presents individual worker activities during the
monitoring period and Table A-8 summarizes these activities.

     Regardless of  assigned  duties and  frequency of weighing, the
actual dye weighing operation was similar at all monitored sites.
The typical dye weigher's activities in filling each batch order
were as follows:

      (1)   The weigher obtained the dyes by walking to a drum or
           container of the appropriate  dye and, with  a hand scoop,
           removing an approximate quantity of dye.

      (2)   The weigher then transferred  the scoop  of dye to the
           weighing station,  where the required amount of dye was
           poured into the scale pan.

      (3)   Any unused portion of the dye left in the hand scoop  was
           poured back into the dye container from which it came.

      (4)   The weighed portion of the dye was poured into a transfer
           container,  most often a metal bucket.   (For purposes  of
           this study, each dye transfer as described  in steps 1
           through 4 was^ recorded by the survey team as a separate
           weighing.)    ;

      (5)   Steps 1 through 4 were repeated until all the weighings
           specified on the batch ticket had been  completed.   At
           times,  the dye weigher accumulated several  dyes on the
           scale pan prior to transferring them into the transfer
           bucket.

     At  most  facilities,  the dye weigher then transported the bucket
of weighed materials  directly to the mixing  area, where  he either
left it  near  the  tank and stirring equipment  (usually a  rotary blade
mixer  or  homogenizer) for dissolution in hot water by another
operator, or  added  it to  the tank and dissolved/ dispersed the
weighed  dyes  himself.   The dye  solution was  then transferred
manually or by pipeline from the mixing tank to the dyeing machine.
At nine  sites, the  dye  weigher  actually added liquid  (hot water  in
most cases) directly  to the  transfer bucket  containing  the weighed
dye.   At one  site (Site 6/5),  in a unique  departure from the
behavior observed at  other sites, the dye  weigher manually added the
 8-4

-------
dry dye to the addition spout of the "washing machine-like" rotary
dyeing machines.  This method of operation—the direct addition of
dry dye into dyeing machines—differed significantly from operations
at other monitored sites where the dry dyes were first dissolved or
dispersed in  water (often with the aid of a hand mixer or
homogenizer) before addition to the dyeing machines.  Five of the
monitored weighers were actually dyeing machine operators who
operated dyeing equipment, and in two of those cases also loaded and
unloaded fabrics.   Dye weighing was but one of several diverse
responsibilities for these operators.

     At some facilities, when a barrel or drum of dye was almost
empty,  the dye weighers inverted the spent barrel over a new open
barrel of dye to transfer the dregs from the old container into the
new one.  Each such transfer created an additional potential for
exposure to the dyestuffs through possible airborne contamination
and dermal contact.  This activity was observed at five sites.

     At several sites, variations in work practices and operating
procedures were observed which could affect the dye weigher's level
of exposure.  Specific observations were noted as follows:

     •  Site 6/6—Ninety-five percent of all dyes stored and handled
        at this site were liquid dyes.  Most of the dyes used on the
        day of monitoring were liquid Vat dyes.  Eight weighings of
        only two powder substances  (a Naphthol and a. Naphthol Salt)
        were performed on the day of monitoring.  In addition, a
        process malfunction during the monitoring period at this
        facility resulted in a steam loss and forced postponement of
        several scheduled dyelots.

     •  Site 1/0—The monitored dye weigher performed several
        activities that could increase the amount of airborne dye
        particulates, such as banging the scale pan on the table to
        loosen the dye material.

     •  Site 3/3—At this, the only monitored screen printing site,
        the weighing activities were different from those at
        comparable size dyehouses in three ways.  First, all
        weighings were at the beginning of the shift, and the dye
        weigher was in the drug room for only 125 minutes of the 447
        minutes that he was monitored.  Second, the dye weigher
        performed  42 dye weighings and only 3 chemical weighings,
        yet the mass of chemicals weighed amounted to more than 75%
        of the total mass weighed.  Third, following the weighing
        activity,  th.e dye weigher was stall being monitored but he
        was engaged in other pursuits.
                                                                  8-5

-------
     •  Site 7/9—After the appropriate dyes for a particular batch
        were weighed, the container into which the weighed material
        was transferred was sealed with a plastic lid, reducing the
        potential for the dyestuffs to become airborne during
        transfer to the mixing tanks or dye machines.

     •  Site 6/2—Some dyes remaining in almost depleted supply
        barrels were transferred to unused replacement barrels of
        the same dye by inverting or dumping.  The dregs in other
        depleted barrels were permitted to remain in place and the
        containers were used as trash receptacles in the drug room.

     C. Dye Characteristics

     There was a wide'variety at the monitored sites in the fiber
composition of textiles that were processed, production volume,  and
product lines.   These parameters evidently governed the classes of
dyes and individual dyes within classes that were used, quantities
weighed, number of weighings, and shades produced.  Table A-9 in
Appendix A catalogues the number of powder dyes from each dye class
that were used.  Tables A-10 and A-ll provide the number of
weighings and the mass weighed, respectively, from each dye class,
and Table A-12 summarizes these statistics by dye class.  Table A-13
is a summary of commercial dye weighed by dye class, broken down
into the three categories:  All Dyes, All But Black Dyes, and Black
Dyes Only.

     Table A-14 lists by Color Index Name2 all individual dyes that
were encountered and the number of weighings and quantities that
were weighed.

     D. Controls and Safety

     The  presence of engineering controls and the use  of personal
protective equipment were recorded during each monitoring survey.
One of the drug rooms at Site 6/6 was equipped with an overhead
exhaust hood for the removal of airborne dye particles.  At Site
5/2, a small 10-inch exhaust fan was located in the wall of the drug
room near the weighing station to exhaust airborne particles.  None
of the other drug rooms was equipped with local exhaust ventilation
for control of worker exposure to the dye material.
     2Dye identification postscripts commencing with U or M refer to
 unidentified or mixture dyes, respectively, for which  no  color index
 number had been assigned.
 8-6

-------
     As previously described, more than 50% of the drug rooms were
observed to have a relatively low amount of visible accumulated dye
on the drug room walls, floors,  equipment, and storage containers.
Almost all of the drug rooms reported using wet mopping techniques
to clean up spilled dye material.  At several locations,  mopping was
replaced or augmented with direct water washing using hoses and
spray nozzles.  Direct water washing was conducted at one site
during the monitoring period (necessitating the previously described
alteration in sampling procedure at that site); other facilities
reportedly use direct water washing procedures either as part of the
drug room cleaning between shifts or during weekly cleaning
operations.   Wet mopping on a regular basis controls the
accumulation of dye material in the drug room,  reduces the potential
for the dye material to become airborne and inhaled by workers,  and
the possible damage to the textile materials through contamination.

     The monitored dye weighers' work practices use of personal
protective equipment and general safety and personal hygiene
practices were recorded at each site.  Table A-15(a)  and (b) in
Appendix A presents individual facility characterization,  and Table
A-16 summarizes them.  (At one site,  during an approximately 8-hr
apparatus monitoring period, the pump was worn by two individuals).
At 12 of the 24 monitored sites, dye weighers wore disposable dust-
mask respirators while performing dye weighing operations,  at 6 they
wore negative pressure air-purifying cartridge respirators,  and at 1
the weigher wore a powered, positive pressure air-purifying
respirator.   Several of the workers that were using air-purifying
respiratory protective equipment during the surveys were observed as
not using the respirators properly to attain optimum protection.
For example, the monitored dye weigher at one site who wore a dust-
mask respirator had a full beard; facial hair greatly increases the
air leakage around the respirator and renders it less effective.  At
two sites, the monitored dye weigher wore his dust-mask respirator
in such a way that, at times, both elastic straps were not
positioned for proper fit and mask-to-face seal.   Dye weighers at
the site which had the greatest concentration of dye dust in the air
used powered, positive pressure air-purifying respirators.   Only
four of the monitored dye weighers who were using respirators
indicated they had received training in the proper use and
maintenance of respiratory protective equipment (Sites 1/0,  1/6,
2/4, and 7/7).

     The monitored dye weighers used various forms of protective
equipment to control dermal exposure to the powder dyes.   They wore
gloves while weighing dye batches at 15 of the 24 facilities
monitored.  Leather work gloves were worn at one of the facilities;
eight facilities provided the dye weighers with latex or butyl
rubber gloves.  The nature of the glove material was not recorded
                                                                  8-7

-------
for six of the facilities where glove use was noted.  Because the
dye weighers must often reach deep into a dye barrel to retrieve the
dyestuffs, the use of protective gloves did not always prevent
dermal contact with the dye materials, especially on the lower and
upper arms.   Significant dermal contact with the dye material was
observed on the hands and arms of the dye weighers who were not
using protective gloves.  Eight of the monitored dye weighers wore
protective clothing  (in the form of a protective apron or smock), 9
wore protective boots or steel-toed safety shoes, and 6 of the 24
monitored dye weighers wore safety glasses or goggles to prevent eye
exposure to the dyestuffs.

      The  dye weighers'  eating, drinking, and  smoking habits while in
the drug room were noted during the monitoring surveys to assess the
additional potential for exposure to the dye material through
ingestion.  Monitored dye weighers were observed eating and/or
drinking in five of the facilities monitored, although almost all of
the facilities stated that company policy prohibited such activities
in the drug room.  Eight of the 24 monitored dye weighers smoked
while in the drug room.  This practice can result in the ingestion
of dye material through transfer from hands that have not been
thoroughly washed prior to cigarette handling or exposure through
inhalation as a combustion product in the air moving through the
cigarette.  Potential for ingestion of the dye material was also
noted at one site, where the monitored dye weigher used chewing
tobacco while in the drug room.
 8-8

-------
                    Appendix A
SUMMARY OF MEASUREMENTS FOR THE 24 SITES MONITORED

-------
                              Table A-l

                    SUMMARY DATA FOR THE 24 SITES
                         Average
     Range
Units
Dye Usage

Number'of Individual
Powder Dyes Handled         17.4

Number of
Powder Dye Weighings        59.5

Total Weight of
Commercial Powder Dyes
Weighed                     56.7

Air Monitoring Results

Concentration of Total
Particulates in Drug Room
Air, Gravimetric Analysis

  Average of 2 Personal
  Monitors                  0.46

  Weighing Station Area     0.21

  Drum Storage Area         0.12

Concentration of Total
Commercial Powder Dyes
in Drug Room Air,
Spectrophotometric Analysis

  Average of 2 Personal
  Monitors                  0.19

  Weighing Station Area     0.12

  Drum Storage Area         0.04
      2-46


     7-259



 2.1-283.9
 0.02-1.37

 0.04-0.59

 0.04-0.25
 0.01-1.20

 0.01-0.37

<0.01-0.16
No. of Dyes


No. of Weighings



kg per shift
mg/m3

mg/m3

mg/m3
mg/m3

mg/m3

mg/m3
   Note:  This table summarizes site data sent with individual site
reports to each study site to provide a context for the
interpretation of individual results.  These are not survey
estimates and do not represent the population of dye houses as a
whole.
                                                                  A-l

-------
                                      Table A-2

                  INDIVIDUAL SITE  CHARACTERISTICS  RECORDED
                         DURING EACH  MONITORING PERIOD
Site Product
ID Line
1/0 piece goods/outerwear
1/6 piece goods/outerwear
2/1 apparel/knit goods
2/4 stock
2/7 apparel/knit goods
3/0 upholstery yarn
3/3 upholstery
3/8 yarn
4/1 stock
4/3 yarn
4/6 carpet yarn
4/9 yarn
5/2 carpet yarn/roll goods
5/4 sheets-woven piece goods
5/9 woven piece goods
6/2 knit goods-tricot/auto
6/5 garments-socks
6/6 apparel/piece goods
7/7 yarn/apparel-sweaters
7/9 garments
8/0 woven seat belts
8/6 novelty yarn
8/8 warp yarns
9/1 garments-socks
Plant Weighted Mean
Unweighted mean
Minimum Value
Maximum Value
Produc-
tion
volume,
mm Ibs
18.0
5.0
3.5
2.5
6.0
25.0
2.2
24.0
6.0
1410
2.0
12.0
4.8
13.0
5.5
23.0
0.4
1.2
1.5
0.9
6.0
3.0
2.2
0.3
7.1
7.6
0.3 ,
25.0 '
Type of
Busi-
ness*
V
C
V
V
C
V
V
V/C
V
V/C
V
V
V
V
V/C
V
V
C
V
C
V
V/C
V
V




Type of
Manage-
ment**
priv.
publ.
publ.
publ.
priv.
priv.
priv.
priv.
publ.
publ.
priv.
publ.
publ.
publ.
publ.
publ.
priv.
priv.
priv.
priv.
priv.
priv.
priv.
priv.




Dyeing
Type of Pro-
Equipment cess***
jet /beck
jig/beam/pad
beck
stock
jet
package
rotary screen
package/skein
stock/atmosph
package
KDK space dye
package
skein/beck
beck
jet/beck/pad
jet/beam
rotary
pad
package/paddle
paddle
pad
space dye
beam
rotary




B
B/C
B
B
B
B
P
B
B
B
C
B
B
B
B/C
B
B
C
B
B
C
B
B
B




Units
Avail-
able
75
17
11
33
11
12
2
35
7
15
2
15
8
11
18
33
5
3
10
7
4
9
4
11
14.6
14.9
2.0
75.0
Units Percent
in of Capa—
Opera- city in
tion Operation
75
14
11
20
11
12
1
31
7
15
1
15
7
11
9
33
5
2
10
7
1
7
4
8
12.9
13.2
1.0
75.0
100.0
82.4
100.0
60.6
100.0
100.0
50.0
88.6
100.0
100.0
50.0
100.0
87.5
100.0
50.0
100.0
100.0
66.7
100.0
100.0
25.0
77.8
100.0
72.7
84.9
83.8
25.0
100.0
   *Type of Business:  V,  vertical; C, commission; V/C, both vertical and commission.
  **Type of Management:  priv., private; publ.,  public.
 ***Dyeing Process: B, batch; C, continuous; P, print; B/C, both batch and continuous.
A-2

-------
                     Table A-3

     SUMMARY OF SITE CHARACTERISTICS RECORDED
           DURING EACH MONITORING PERIOD
      Variable
                               Number
Production Volume (million pounds)
         0.0 to 5.0
         5.1 to 10.0
         10.1 to 20.0
         20.1 to 25.0
                              13  sites
                               4  sites
                               4  sites
                               3  sites
Management Type
         Vertical
         Commision
         Both

Ownership
         Private
         Public

Dyeing Processes Performed
         Batching Dyeing
         Continuous Dyeing
         Both
         Printing

Number of Dyeing Machines
         1 to 5
         6 to 10
         11 to 15
         16 to 35
         75
Location  (EPA Region)
   Region 1 (New England)
   Region 3
   Region 4
   Region 5
   Region 7
(Middle-Atlantic)
(Southeast)
(Great  Lakes)
(Central)
                              16  sites
                               4  sites
                               4  sites
                              14  sites
                              10  sites
                              18  sites
                               3  sites
                               2  sites
                               1  site
                               6  sites
                               5  sites
                               7  sites
                               5  sites
                               1  site
 1 site
 2 sites
19 sites
 1 site
 1 site
                                                         A-3

-------
                             Table A-4

     SUMMARY OF FIBERS PROCESSED  OR  DYE CLASSES USED PER SITE
      Variable
Number
      Number of Fibers Processed
           1                                   2 sites
           2                                   3 sites
           3                                   6 sites
           4                                   3 sites
           5                                   4 sites
           6                                   4 sites
           7-9                                 2 sites

      Fibers Processed
           Acetate                             3 sites
           Acrylic/modacrylic                 13 sites
           Cotton                             17 sites
           Nylon                              17 sites
           Polyester                          18 sites
           Rayon                              11 sites
           Wool                               12 sites
           Other                               5 sites

      Number of Dye Classes  (powder  dyes  only)
           1                                   9 sites
           2                                   6 sites
           3                                   4 sites
           4                                   4 sites
           5           .>                        1 site
                       •»

      Dye  Classes Encountered  (powder dyes  only)
           Acid                               14 sites
           Chrome                              2 sites
           Disperse                           15 sites
           Basic                               9 sites
            Reactive
            Direct
            Vat
            Naphthol
6 sites
6 sites
1 site
1 site
A-4

-------
                Table A-5

INDIVIDUAL SHIFT CHARACTERISTICS MONITORED
      DURING EACH MONITORING PERIOD
Typical Daily
Dye Weighing
Activities

Site No.

No.
ID Hrs. Shifts
1/0 24
1/6 24
2/1 24
2/4 8
2/7 24
3/0 24
3/3 24
3/8 24
4/1 16
4/3 24
4/6 24
4/9 24
5/2 24
5/4 24
5/9 24
6/2 24
6/5 16
6/6 24
7/7 24
7/9 8
8/0 24
8/6 16
8/8 24
9/1 8
Plant Weighted
Unweighted Mean
Minimum Value
Maximum Value
3
3
3
1
2
3
3
3
2
3
2
3
2
2
3
3
2
3
3
1
2
2
3
1
Mean




Shift
Monitored
3
2
2
1
2
1
3
1
1
2
1
2
.. 1
-•-' 1-2
2
1
1
2
3
1
1
1
2
1




No. of Dye
Weighers
Working
During
Monitoring
Period
2
1
1
1
1
2
1
1
1
1
1
1
1
4
1
1
1
2
1
1
3
1
1
1
1.3
1.3
1.0
4.0
No. of
No. of
Fibers Powder Dye
Processed
(per year)
5
4
3
2
3
2
5
6
3
6
1
6
1
4
7
4
5
6
3
2
3
9
3
5
4.1
4.1
1.0
9.0
Classes
Used
3
2
3
1
3
2
1
4
1
2
1
4
1
2
4
2
3
1
1
1
1
4
2
5
2.2
2.3
1.0
5.0

No. of
Powder Dyes
Weighed
30
11
15
21
16
21
8
46
17
18
6
31
3
16
18
30
12
2
7
9
3
25
11
41
17.1
17.4
2.0
46.0
No. of
Solid State
Chemicals
Weighed
1
1
0
0
0
0
1
1
0
4
1
2
0
0
1
0
8
2
0
0
0
0
3
1
1.1
1.1
0.0
8.0
                                                    A-5

-------
                             Table A-6

             SUMMARY OF SHIFT CHARACTERISTICS RECORDED
                   DURING EACH MONITORING PERIOD
       Variable
  Number
       Shifts Regularly  in Operation
                 3  shifts, 8  hours  each
                 2  shifts, 12 hours each
                 2  shifts, 8  hours  each
                 1  shift,  8-10  hours each

       Shift Monitored
                 First  Shift  (7am - 3 pm)
                 Second Shift (3  pm - 11 pm)
                 Third  Shift  (11  pm - 7 am)

       Average  Number  of Dye Weighers
       Working  on  a Shift
  13 sites
   5 sites
   3 sites
   3 sites
12.5 sites
 8.5 sites
   3 sites

 1.3 dye weighers
     per shift
     per site
A-6

-------
   Table A-7
WORKER ACTIVITY
Time
Site Monitored
ID (minutes)
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6
7/7
7/9
8/0
8/6
8/8
9/1
Plant-Weighted Mean
Unweighted Mean
Minimum Value
Maximum Value
449
464
367
440
475
428
447
469
445
457
447
426
454
400>
452
459
463
446
369
435
474
434
458
465
440.7
442.6
367.0
475.0
Minutes
in
Weighing
Room
376
407
84
213
124
263
125
342
94
96
46
359
137
287
403
456
36
217
16
320
437
413
146
127
220.9
230.2
16.0
456.0
Number of
Entries
into
Weighing
Room
11
10
12
18
16
18
18
12
14
9
16
4
19
7
7
4
9
18
3
6
8
4
8
12
10.9
11.0
3.0
19.0
Number of
Weighings
of
Powder
Dyes
97
54
20
108
27
72
42
149
38
29
15
62
84
46
44
88
15
8
7
33
11
259
17
103
60.3
59.5
7.0
259.0
Number of
Weighings
of All
Solid
Substances
98
56
20
108
27
72
45
150
38
51
23
65
84
46
54
88
35
14
7
33
11
259
26
104
63.7
63.1
7.0
259.0
Weight of
All
Powder Dyes
Weighed
(kg)
121.8
56.4
64.0
44.7
25.6
97.9
4.5
74.6
30.9
60.0
54.2
197.8
5.2
10.7
51.1
283.9
6.4
39.6
6.0
15.5
2.1
73.5
10.1
24.1
58.1
56.7
2.1
283.9
Weight of
All Solid
Substances
Weighed
(Xg)
123.0
56.7
64.0
44.7
25.6
97.9
22.9
74.6
30.9
556.8
83.9
206.7
5.2
10.7
143.6
283.9
107.1
223.3
6.0
15.5
2.1
73.5
58.8
24.1
97.1
97.6
2.1
556.8
                                       A-7

-------
                             Table A-8
                    SUMMARY OF WORKER ACTIVITY
   Variable
                              Number of Sites
   Number of Total Weighers
   Normally Working on Any Day
              1
              2
              3
              4
              5
              6
              8

   Kilograms of Dye Weighed
             0 to 10
         10.01 to 30
         30.01 to 60
         60.01 to 80
         80.01 to 284.3

   Number of Dyes Weighed
            0-10
           11 - 20
           21 - 30
           31 - 46

   Number of Dye Weighings
            0
           21
           41
           61
           81
          100
20
40
60
80
100
259
    Hours in the Drug Room
         0.00-2   ( < 25% of monitoring period)
         2.01 - 4   (25-50% of monitoring period)
         4.01 - 6   (50-75% of monitoring period)
         6.01-8   ( > 75% of monitoring period)
                                     3
                                     5
                                     7
                                     2
                                     1
                                     5
                                     1
                                     5
                                     5
                                     6
                                     4
                                     4
                                     7
                                     9
                                     5
                                     3
7
4
4
2
3
4
                                     6
                                     7
                                     5
                                     6
A-8

-------
                               Table A-9



         DYE FREQUENCY DURING MONITORING PERIOD, SITE BASIS




        (Number of Powder Dyes  Encountered by Class  of Dyes)


Site
ID
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6
7/7
7/9
8/0
8/6
8/8
9/1
Total
Number of
Unique Dyes
Total Number
of Dye
Encounters
Number of
Sites Where
Dye Class
Was Weighed
Average
Number
of Dyes
in Class
Per Site*



Acid
0
10
0
21
0
7
8
16
17
0
6
13
3
0
3
8
8
0
0
0
0
4
0
17


101


141



14




10



Chrome
0
0
0
0
0
0
0
1
0
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0


3


4



2




2



Disperse
8
1
5
0
4
14
0
11
0
6
0
8
0
6
7
22
0
0
0
0
3
5
5
5


62


110



15




7
Dye


Basic
11
0
2
0
6
0
0
18
0
0
0
0
0
0
4
0
1
0
7
0
0
4
0
4


34


57



9




6
Class


Reactive
0
0
8
0
0
0
0
0
0
12
0
7
0
0
0
0
3
0
0
0
0
12
0
6


42


48



6




8



Direct
11
0
0
0
6
0
0
0
0
0
0
0
0
10
4
0
0
0
0
9
0
0
0
9


38


49



6




8



Vat
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6
0 ,


6


6



1




6



Naphthol
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0


2


2



1




2

Total No.
of Dyes at
Each Site
30
11
15
21
16
21
8
46
17
18
6
31
3
16
18
30
12
2
7
9
3
25
11
41


288


417



24




17
*Includes only sites where the class of dyes was weighed during the monitoring period.
                                                                      A-9

-------
                                Table A-10

           DYE NEIGHING  ACTIVITY  DURING MONITORING PERIOD:
                 NUMBER OF WEIGHINGS OF EACH DYE CLASS
Site
ID
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6*
7/7
7/9
8/0
8/6
8/8
9/1
Total
Average

Acid Chrome

53

108

13
42
24 2
38

15
26 5
84

6
8
11




18

61
507 7
36 4

Disperse
20
1
5

4
59

11

7

22

23
24
80




11
33
9
5
314
21
Dye Class
Basic Reactive Direct Vat Naphthol
21 56

2 13

17 6


112

22

9

23
5 9

1 3
8
7
33

58 150
8
4 7 26
227 204 153 8 8
25 34 26 8 8
Total Number
of Weighings
at Each Site
97
54
20
108
27
72
42
149
38
29
15
62
84
46
44
88
15
8
7
33
11
259
17
103
1,428
60
*Site was operating under atypical conditions during the monitoring period.
A-10

-------
                                   Table A-ll

             DYE NEIGHING ACTIVITY DURING MONITORING PERIOD:
                   WEIGHT OF WEIGHINGS OF EACH DYE CLASS
Site
ID
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2
6/5
6/6**
7/7
7/9
8/0
8/6
8/8
9/1
Total
Number
of sites
Average

Acid Chrome

56.30

44.70

10.80
4.50
13.80 19.00
30.80

54.20
60.60 51.30
5.20

8.30
0.50
3.30




2.40

7.40
303.0 70.2

14 2
21.6 35.1

Disperse
38.10
0.10
22.30

1.90
87.10

18.70

8.90

85.00

5.80
40.90
283.40




2.10
4.20
1.20
8.10
608.5

15
40.6
Dye Class
Basic Reactive Direct Vat Naphthol
11.40 72.30

5.10 36.60

2.10 21.60


23.20

51.20

0.80

4.90
1.60 0.30

0.30 2.80
39.6
6.00
15.50

15.80 51.00
8.80
0.70 2.90 5.00
66.2 145.3 119.6 8.8 39.6

9 6611
7.4 24.2 9.5 8.8 39.6
Total Weight
of Weighings
at Each Site*
121.768
56.410
63.961
44.674
25.592
97.926
4.489
74.577
30.846
60.029
54.245
197.837
5.239
10.708
51.082
283.912
6.404
39.585
6.034
15.510
2.114
73.496
10.077
24.082
1,360.597*


56.691*
 *Totals  for each site may not be consistent with the sum of each dye class for a particular
  site due to rounding errors.
**Site was operating under atypical conditions during the monitoring period.
                                                                             A-ll

-------
                            Table A-12

            TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
            COMMERCIAL POWDER DYES WEIGHED,  CLASS BASIS
Powder
Number
of
Dyes
101
3
62
34
38
42
6
1
1
288
Class Number
of of
Dye Sites
Acid
Chrome
Disperse
Basic
Direct
Reactive
Vat
Naphthol Salt
Naphthol
TOTAL
14
2
15
9
6
6
1
1
1
24
Number
of
Weighings
507
7
314
227
153
204
8
2
6
1428
Total
Weighed
(kg)
303
70
608
66
119
145
8
24
14
1361
.0
.2
.5
.2
.6
.3
.8
.9
.7
.3
Dyes Average Weighed
(kcr)
Per
Site Per Per Dye
Used Weighing Encountered
21
35
40
7
19
24
8
24
14
56
.6
.1
.6
.4
.9
.2
.8
.9
.7
.7
0
10
1
0
0
0
1
12
2
0
.60
.03
.94
.29
.78
.71
.11
.47
.45
.95
3
23
4
1
3
3
1
24
14
4
.0
.4
.5
.9
.1
.5
.5
.9
.7
.7
A-12

-------
                Table A-13

TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
COMMERCIAL POWDER DYES WEIGHED,  COLOR BASIS






Dye Category


Class
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Chrome
Chrome
Chrome
Chrome
Chrome
Chrome
Chrome
Chrome
Chrome
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse


Color
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total


Total
Mass of
Category
Weighed
(kg)
63.995
29.177
40.152
13.617
40.127
20.354
29.047
66.575
303.044
0.000
1.770
0.000
0.000
0.000
0.000
0.000
68.441
70.211
35.013
136.067
131.817
13.636
181.632
0.272
0.000
110.027
608.464


Number
of Dyes
Encountered
in
Category
21
10
24
4
22
4
7
9
101
0
1
0
0
0
0
0
2
3
13
4
23
3
14
1
0
4
62
Average
Mass
Weighed
of Each
Dye in
Category
(kg/dye)
3.05
2.92
1.67
3.40
1.82
5.09
4.15
7.40
3.00
0.00
1.77
0.00
0.00
0.00
0.00
0.00
34.22
23.40
2.69
34.02
5.73
4.55
12.97
0.27
0.00
27.51
9.81


Number
of Sites
Where
Dye Class
Used
14
14
14
14
14
14
14
14
14
2
2
2
2
2
2
2
2
2
15
15
15
15
15
15
15
15
15
Average
Mass of
Category
Weighed per
Site Where
Class Used
(kg/site)
4.57
2.08
2.87
0.97
2.87
1.45
2.07
4.76
21.65
0.00
0.89
0.00
0.00
0.00
0.00
0.00
34.22
35.11
2.33
9.07
8.79
0.91
12.11
0.02
0.00
7.34
40.56
                                                    A-13

-------
                             Table A-13

            TEXTILE DYE NEIGHING ROOM MONITORING STUDY:
            COMMERCIAL PONDER DYES NEIGHED, COLOR BASIS
                             (Continued)






Dye Category


Class
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Reactive
Reactive
Reactive
Reactive
Reactive
Reactive
Reactive
Reactive
Reactive


Color
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total


Total
Mass of
Category
Weighed
(kg)
23.441
1.767
24.292
1.302
10.217
2.843
0.000
2.329
66.191
6.418
8.778
22.318
0.002
32.471
0.000
8.103
41.554
119.644
17.413
5.012
56.865
0.897
55.978
0.000
0.000
9.113
145.278


Number
of Dyes
Encountered
in
Category
11
2
7
3
9
1
0
1
34
4
5
10
1
11
0
3
4
38
10
3
11
2
14
0
0
2
42
Average
Mass
Weighed
of Each
Dye in
Category
(kg/dye)
2.13
0.88
3.47
0.43
1.14
2.84
0.00
2.33
1.95
1.60
1.76
2.23
0.00
2.95
0.00
2.70
10.39
3.15
1.74
1.67
5.17
0.45
4.00
0.00
0.00
4.56
3.46


Number
of Sites
Where
Dye Class
Used
9
9
9
9
9
9
9
9
9
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
Average
Mass of
Category
Weighed per
Site Where
Class Used
(kg/site)
2.60
0.20
2.70
0.14
1.14
0.32
0.00
0.26
7.35
1.07
1.46
3.72
0.00
5.41
0.00
1.35
6.93
19.94
2.90
0.84
9.48
0.15
9.33
0.00
0.00
1.52
24.21
A-14

-------
                              Table A-13

             TEXTILE DYE NEIGHING ROOM MONITORING STUDY:
             COMMERCIAL PONDER DYES WEIGHED,  COLOR BASIS
                              (Continued)






Dye Category


Class
Vat
Vat
Vat
Vat
Vat
Vat
Vat
Vat
Vat
Naphthol
Naphthol
Naphthol


Color
Yellow
Orange
Red
Violet
Blue
Green
Brown
Black
Total
Salt
Base
Dye


Total
Mass of
Category
Weighed
(kg)
0.785
2.352
0.000
2.887
0.861
0.000
1.962
0.000
8.847
24.932
0.000
14.700


Number
of Dyes
Encountered
in
Category
1
1
0
2
1
0
1
0
6
1
0
1
Average
Mass
Weighed
of Each
Dye in
Category
(kg/dye)
0.79
2.35
0.00
1.44
0.86
0.00
1.96
0.00
1.47
24.93
0.00
14.70


Number
of Sites
Where
Dye Class
Used
1
1
1
1
1
1
1
1
1
1
1
1
Average
Mass of
Category
Weighed per
Site Where
Class Used
(kg/site)
0.79
2.35
0.00
2.89
0.86
0.00
1.96
0.00
8.85
24.93
0.00
14.70
Naphthol   Total
  39.632
           19.82
                    39.63
  GRAND TOTAL
1361.31
288
4.73
24
56.72
                                                                   A-15

-------
                            Table A-14

            TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
                    INDIVIDUAL DYES ENCOUNTERED
Average Weighed
(kg)
No.
Name of of
Commercial Dye Sites
ACID

Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid

Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid

DYES (used at 14

Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow

Orange
Orange
Orange
Orange
Orange
Orange
Orange
Orange
Orange
Orange
Orange


17
19
40
49
65
79
99
116
121
127
129
135
151
159
198
216
218
219
235
241
U-l
Subtotal

3
10
51
60
74
116
142
149
156
U-l
Subtotal

of 24

1
1
1
5
1
2
2
1
2
1
2
1
2
3
1
1
1
2
1 >
1
1
33

1
1
1
1
1
2
1
1
2
1
12

No. of
Weigh-
ings
sites)

3
12
2
14
1
3
9
1
3
1
3
10
2
15
4
3
1
22
5
2
2
118

7
1
1
1
1
5
1
1
20
1
39

Total
Weighed
(kg)


0.
3.
0.
7.
0.
2.
16.
0.
0.
0.
5.
2.
0.
1.
0.
0.
0.
1.
18.
1.
0.
63.

1.
0.
0.
0.
0.
0.
0.
4.
21.
0.
29.



437
774
001
814
175
746
352
002
765
002
840
258
048
176
145
868
025
663
500
334
070
995

911
067
320
003
432
086
247
192
850
069
177

Per
Site


0.
3.
0.
1.
0.
1.
8.
0.
0.
0.
2.
2.
0.
0.
0.
0.
0.
0.
18.
1.
0.
1.

1.
0.
0.
0.
0.
0.
0.
4.
10.
0.


437
774
001
563
175
373
176
002
383
002
920
258
024
392
145
868
025
832
500
334
070
939

911
067
320
003
432
043
247
192
925
069
2.431


Per
Weigh-
ing


0.
0.
0.
0.
0.
0.
1.
0.
0.
0.
1.
0.
0.
0.
0.
0.
0.
0.
3.
0.
0.
0.

0.
0.
0.
0.
0.
0.
0.
4.
1.
0.


146
315
001
558
175
915
817
002
255
002
947
226
024
078
036
289
025
076
700
667
035
542

273
067
320
003
432
017
247
192
093
069
0.748


Number
Encountered
Sites: 4/9,3/8,
5/2,8/6,4/1,
3/0,1/6,2/4,
5/9,3/3,6/2,
4/6,6/5,9/1.


















21 Yellow
Acid Dyes










10 Orange
Acid Dyes
A-16

-------
                Table A-14

TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
        INDIVIDUAL DYES ENCOUNTERED
                (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
Acid Red 1
Acid Red 52
Acid Red 57
Acid Red 111
Acid Red 143
Acid Red 158
Acid Red 182
Acid Red 186
Acid Red 194
Acid Red 259
Acid Red 260
Acid Red 266
Acid Red 299
Acid Red 337
Acid Red 357
Acid Red 359
Acid Red 360
Acid Red 361
Acid Red 396
Acid Red 399
Acid Red M-2
Acid Red U-3
Acid Red U-5
Acid Red U-6
Acid Red Subtotal

Acid Violet 7
Acid Violet 48
Acid Violet 90
Acid Violet 121
Acid Violet Subtotal

No.
of
Sites
1
1
1
1
1
1
2
1
1
1
2
3
3
1
1
1
1
2
1
1
1
1
1
1
31

1
2
1
1
5

No. of
Weigh-
ings
2
1
6
1
6
1
3
1
1
2
5
16
7
2
6
1
1
8
1
6
35
2
2
12
128

2
6
1
1
10

Total
Weighed
(kg)
0.081
0.100
9.051
0.368
1.075
1.840
1.221
0.481
2.086
0.046
0.382
7.928
3.099
0.173
0.702
3.457
0.364
3.134
0.001
0.201
2.476
0.014
1.614
0.258
40.152

0.008
0.715
0.760
12.134
13.617


Per
Site
0.081
0.100
9.051
0.368
1.075
1.840
0.611
0.481
2.086
0.046
0.191
2.643
1.033
0.173
0.702
3.457
0.364
1.567
0.001
0.201
2.476
0.014
1.614
0.258
1.295

0.008
0.358
0.760
12.134
2.723

Per
Weigh-
ing
0.041
0.100
1.509
0.368
0.179
1.840
0.407
0.481
2.086
0.023
0.076
0.496
0.443
0.087
0.117
3.457
0.364
0.392
0.001
0.034
0.071
0.007
0.807
0.022
0.314

0.004
0.119
0.760
12.134
1.362


Number
Encountered
























24 Red
Acid Dyes




4 Violet
Acid Dyes
                                                    A-17

-------
                             Table A-14

            TEXTILE DYE WEIGHING ROOM MONITORING  STUDY:
                    INDIVIDUAL DYES ENCOUNTERED
                             (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
Acid Blue 7
Acid Blue 25
Acid Blue 40
Acid Blue 45
Acid Blue 62
Acid Blue 80
Acid Blue 90
Acid Blue 102
Acid Blue 113
Acid Blue 158
Acid Blue 177
Acid Blue 205
Acid Blue 239
Acid Blue 258
Acid Blue 264
Acid Blue 277
Acid Blue 281
Acid Blue 284
Acid Blue 290
Acid Blue 324
Acid Blue 335
Acid Blue 345
Acid Blue Subtotal

Acid Green 25
Acid Green 28
Acid Green 104
Acid Green 108
Acid Green Subtotal

No.
of
Sites
1
'4
2
1
2
3
1
1
4
2
1
1
1
1
1
1
1
1
1,
2
1
1
34

3
1
1
1
6

No. Of
Weigh-
ings
1
31
9
4
3
11
1
1
9
6
1
1
3
6
1
1
6
1
4
34
1
1
136

7
3
1
1
12

Total
Weighed
(kg)
0.001
15.050
0.501
0.270
0.004
5.747
0.007
0.368
9.972
3.519
0.015
0.454
0.092
0.607
0.182
0.110
0.174
0.995
0.129
1.477
0.034
0.419
40.127

8.161
0.466
3.718
8.009
20.354


Per
Site
0.001
3.763
0.251
0.270
0.002
1.916
0.007
0.368
2.493
1.760
0.015
0.454
0.092
0.607
0.182
0.110
0.174
0.995
0.129
0.739
0.034
0.419
1.180

2.720
0.466
3.718
8.009
3.392

Per
Weigh-
ing
0.001
0.485
0.056
0.068
0.001
0.522
0.007
0.368
1.108
0.587
0.015
0.454
0.031
0.101
0.182
0.110
0.029
0.995
0.032
0.043
0.034
0.419
0.295

1.166
0.155
3.718
8.009
1.696


Number
Encountered






















22 Blue
Acid Dyes




4 Green
Acid Dyes
A-18

-------
                                Table A-14

              TEXTILE DYE WEIGHING  ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                                (Continued)
Average
(kg)
Name of
Commercial
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid

Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Acid
Brown
Brown
Brown
Brown
Brown
Brown
Brown
Brown

Black
Black
Black
Black
Black
Black
Black
Black
Black
No.
Of
Dye Sites
45
227
298
330
384
M-l
U-2
Subtotal

52
58
60
107
172
187
M-l
M-2
M-3
1
2
2
1
1
1
1
9

1
2
1
2
1
1
1
1
1
No. of Total
Weigh- Weighed
ings
2
2
10
3
6
3
4
30

3
2
3
13
4
1
2
2
4
(kg)
0.
3.
21.
2.
1.
0.
0.
29.

6.
7.
4.
12.
21.
2.
7.
0.
4.
035
496
937
183
059
045
292
047

499
933
020
112
429
121
520
481
460
Per
Weighed
Per
Weigh- Number
Site
0
1
10
2
1
0
0
3

6
3
4
6
21
2
7
0
4
.035
.748
.969
.183
.059
.045
.292
.227

.499
.967
.020
.056
.429
.121
.520
.481
.460
0
1
2
0
0
0
0
0

2
3
1
0
5
2
3
0
1
ing Encountered
.018
.748
.194
.728
.177
.015
.073
.968 7 Brown
Acid Dyes
.166
.967
.340
.932
.357
.121
.760
.241
.115
Acid Black Subtotal    11     34      66.575
TOTAL—ACID DYES      141    507     303.044
         6.052  1.958   9 Black
                         Acid Dyes

         2.149  0.598 101 Acid Dyes,
                         all colors
CHROME DYES  (used at 2 of 24  sites)

Mordant Orange 3        11
1.770
1.770   1.770
Sites:  4/9,3/8.

 1 Orange
   Chrome  Dye
Mordant Black 9 1
Mordant Black 11 2
Mordant Black Subtotal 3

TOTAL — CHROME DYES '4

1 15.400 15.400 15.400
5 53.041 26.521 10.608
6 68.441 22.814 11.407 2 Black
Chrome Dyes
7 70.211 17.553 10.030 3 Chrome Dyes,
all colors
                                                                      A-19

-------
                             Table A-14

            TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
                    INDIVIDUAL DYES ENCOUNTERED
                             (Continued)
Average Weighed
(kg)
Name of
Commercial Dye
DISPERSE

Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
No.
of
Sites
DYES (used at

Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow

3
23
42
54
64
67
86
93
108
114
184:1
198
218

Subtotal

Disperse
Disperse
Disperse
Disperse
Disperse

Orange
Orange
Orange
Orange
Orange

29
30
37
41

Subtotal

Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse

Red 4
Red 43
Red 55
Red 60
Red 65
Red 72
Red 73
Red 82
Red 86
Red 88
Red 91












No. of
Weigh-
ings
Total
Weighed
(kg)
Per
Site
Per
Weigh- Number
ing Encountered
15 of 24 sites)

2
1
3
2
1
1
1
1
1
1
1
1
2

18
-»
1 *
4
1
2

8

1
1
1
7
1
1
3
2
1
1
2

2
1
14
3
1
1
1
3
3
1
1
3
8

42

3
35
4
5

47

1
5
4
17
1
6
7
3
3
2
5

0.
13.
12.
0.
0.
0.
4.
0.
0.
0.
0.
1.
1.

35.

0.
90.
4.
41.

136.

0.
1.
1.
5.
1.
0.
4.
0.
0.
2.
7.

080
276
194
761
054
109
646
620
171
114
100
360
528

013

081
587
126
273

067

021
370
198
717
552
808
185
268
440
734
,451

0.
13.
4.
0.
0.
0.
4.
0.
0.
0.
0.
1.
0.

1.

0.
22.
4.
20.

17.

0.
1.
1.
0.
1.
0.
1.
0.
0.
2.
3.

040
276
065
381
054
109
646
620
171
114
100
360
764

945

081
647
126
637

008

021
370
198
817
552
808
395
134
440
734
,726

0.
13.
0.
0.
0.
0.
4.
0.
0.
0.
0.
0.
0.

0.

0.
2.
1.
8.

2.

0.
0.
0.
0.
1.
0.
0.
0.
0.
1.
Sites: 4/9,3/8,
2/1,8/6,2/7,
040 3/0,5/4,1/0,
276 1/6,5/9,8/0,
871 8/8,4/3,
254 6/2,9/1.
054
109
646
207
057
114
100
453
191

834 13 Yellow
Disperse Dyes
027
588
032
255

895 4 Orange
Disperse Dyes
021
274
300
336
552
135
598
089
147
367
1.490
A-20

-------
                 Table A-14

TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
        INDIVIDUAL DYES ENCOUNTERED
                 (Continued)
Average
(kg)
No.
Name of of
Commercial Dye Sites
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse

Disperse
Disperse
Disperse
Disperse
Red 135
Red 151
Red 159
Red 167
Red 167:1
R^d 177
Red 211
Red 263
Red 305
Red 333
Red 338
Red U-2
Red Subtotal

Violet 26
Violet 48
Violet 57
Violet
Subtotal

Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse
Disperse


Blue 3
Blue 27
Blue 56
Blue 60
Blue 73
Blue 77
Blue 79
Blue 87
Blue 109
Blue 139
Blue 281
Blue 337
Blue M-3
Blue U-4
Blue Subtotal

2
1
1
1
1
1
1
2
1
1
1
1
35

1
1
1

3

2
3
8
7
5
2
3
1
2
1
2
1
1
2
40

No. of Total
Weigh- Weighed
ings
2
1
6
5
4
4
9
4
2
2
1
1
95

1
2
5

8

2
12
20
19
13
5
13
3
2
1
6
6
1
5
108


11
5
15
32
5
5
0
12
2
13
0
1
131

0
0
12

13

0
19
35
18
13
26
10
8
4
0
16
7
13
5
181

(kg)
.604
.045
.039
.350
.657
.126
.914
.224
.036
.937
.793
.348
.817

.159
.680
.797

.636

.017
.893
.494
.818
.849
.009
.870
.374
.770
.078
.991
.479
.349
.641
.632

Per
Weighed
Per
Weigh-
Site
5
5
15
32
5
5
0
6
2
13
0
1
3

0
0
12

4

0
6
4
2
2
13
3
8
2
0
8
7
13
2
4

.802
.045
.039
.350
.657
.126
.914
.112
.036
.937
.793
.348
.766

.159
.680
.797

.545

.009
.631
.437
.688
.770
.005
.623
.374
.385
.078
.496
.479
.349
.821
.541

5
5
2
6
1
1
0
3
1
6
0
1
1

0
0
2

1

0
1
1
0
1
5
0
2
2
0
2
1
13
1
1

ing
.802
.045
.507
.470
.414
.282
.102
.056
.018
.969
.793
.348
.388

.159
.340
.559

.705

.009
.658
.775
.990
.065
.202
.836
.791
.385
.078
.832
.247
.349
.128
.682

Number
Encountered












23 Red
Disperse Dyes




3 Violet
Disperse Dyes














14 Blue
Disperse Dyes
                                                    A-21

-------
                                Table A-14

              TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                                (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
Disperse Green 9

Disperse Black M-l
Disperse Black M-2
Disperse Black M-3
Disperse Black M-4
No.
of
Sites
1

1
1
1
1
No. Of
Weigh-
ings
1

4
2
6
1
Total
Weighed
(kg)
0.272

41.645
11.900
53.693
2.789

Per
Site
0.272

41.645
11.900
53.693
2.789
Per
Weigh-
ing
0.272

10.411
5.950
8.949
2.789

Number
Encountered
1 Green
Disperse Dye




Disperse Black Subtotal 4     13
TOTAL—DISPERSE DYES  109    314
110.027   27.507  8.464
4 Black
  Disperse Dyes
608.464    5.582  1.938 62 Disperse Dyes,
                           all colors
BASIC DYES (used at 9 of 24 sites)
                        Sites: 3/8,7/7,
                        2/1,8/6,2/7,
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic

Basic
Basic
Basic

Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow
Yellow

Orange
Orange
Orange

11
13
21
24
25
28
29
40
51
87
91
Subtotal

21
30
Subtotal

2
1 >
1
1
1
4
1
1
1
1
1
15

2
1
3

29
1
1
2
3
16
2
1
1
8
1
65

2
1
3

8
0
0
1
2
5
1
3
0
0
0
23

1
0
1

.045
.018
.068
.072
.662
.009
.635
.747
.420
.432
.333
.441

.707
.060
.767

4
0
0
1
2
1
1
3
0
0
0
1

0
0
0

.023
.018
.068
.072
.662
.252
.635
.747
.420
.432
.333
.563

.854
.060
.589

0.277
0.018
0.068
0.536
0.887
0.313
0.818
3.747
0.420
0.054
0.333
0.361

0.854
0.060
0.589

1/0,5/9,
6/5,9/1.









11 Yellow
Basic














Dyes


2 Orange
Basic
Dyes
A-22

-------
                Table A-14

TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
        INDIVIDUAL DYES ENCOUNTERED
                (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
Basic Red 14
Basic Red 15
Basic Red 29
Basic Red 46
Basic Red 49
Basic Red 51
Basic Red U-2
Basic Red Subtotal

Basic Violet 14
Basic Violet 16
Basic Violet 37
Basic Violet Subtotal

Basic Blue 3
Basic Blue 21
Basic Blue 41
Basic Blue 45
Basic Blue 54
Basic Blue 69
Basic Blue 124
Basic Blue 141
Basic Blue U-l
Basic Blue Subtotal

Basic Green 4

Basic Black M-l

TOTAL — BASIC DYES

No.
Of
Sites
3
2
2
4
1
1
1
14

1
2
1
4

4
1
3
1
3
1
1
1
1
16

4

1

•57

NO. Of
Weigh-
ings
5
8
25
16
1
16
5
76

1
6
1
8

9
1
19
2
6
16
13
1
2
69

5

1

227

Total
Weighed
(kg)
1.543
6.050
11.821
1.081
0.866
2.835
0.096
24.292

0.074
1.183
0.045
1.302

0.975
0.015
4.390
0.030
2.312
2.071
0.286
0.068
0.070
10.217

2.843

2.329

66.191


Per
Site
0.514
3.025
5.911
0.270
0.866
2.835
0.096
1.735

0.074
0.592
0.045
0.326

0.244
0.015
1.463
0.030
0.771
2.071
0.286
0.068
0.070
0.639

0.711

2.329

1.161

Per
Weigh-
ing
0.309
0.756
0.473
0.068
0.866
0.177
0.019
0.320

0.074
0.197
0.045
0.163

0.108
0.015
0.231
0.015
0.385
0.129
0.022
0.068
0.035
0.148

0.569

2.329

0.292


Number
Encountered







7 Red
Basic Dyes



3 Violet
Basic Dyes









9 Blue
Basic Dyes
1 Green
Basic Dye
1 Black
Basic Dye
34 Basic Dyes,
all colors
                                                    A-23

-------
                               Table A-14

             TEXTILE DYE NEIGHING ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                               (Continued)
Average Weighed
(kg)
No.
Name of of
Commercial Dye Sites
DIRECT DYES (used at 6

Direct Yellow 44
Direct Yellow 58
Direct Yellow 106
Direct Yellow 142
Direct Yellow Subtotal

Direct Orange 34
Direct Orange 72
Direct Orange 80
Direct Orange M-2
Direct Orange M-3
Direct Orange Subtotal

Direct Red 9
Direct Red 72
Direct Red 75
Direct Red 80
Direct Red 89
Direct Red 149
Direct Red 224
Direct Red 227
Direct Red 243
Direct Red U-l
of 24

1
1
5
1
8

1
1
1
1
1
5

2 *
1
1
2
1
1
1
1
2
1
No. Of
Weigh-
ings
sites)

1
1
23
6
31

1
3
1
5
1
11

2
6
9
5
7
1
1
8
6
1
Total
Weighed
(kg)


1.362
0.078
4.318
0.660
6.418

0.147
0.013
1.345
7.243
0.030
8.778

2.274
8.151
2.278
0.150
3.023
0.117
2.316
2.565
0.371
1.073

Per
Site


1.362
0.078
0.864
0.660
0.802

0.147
0.013
1.345
7.243
0.030
1.756

1.137
8.151
2.278
0.075
3.023
0.117
2.316
2.565
0.186
1.073
Per
Weigh- Number
ing Encountered
Sites: 2/7,5/4,
1/0,5/9,
1.362 7/9,9/1.
0.078
0.188
0.110
0.207 4 Yellow
Direct Dyes
0.147
0.004
1.345
1.449
0.030
0.798 5 Orange
Direct Dyes
1.137
1.359
0.253
0.030
0.432
0.117
2.316
0.321
0.062
1.073
Direct Red Subtotal
Direct Violet 9
13     46      22.318    1.717  0.485  10 Red
                                       Direct Dyes

 1      1       0.002    0.002  0.002    1 Violet
                                       Direct Dye
A-24

-------
                               Table A-14

              TEXTILE DYE NEIGHING ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                               (Continued)
Average
Weighed
(kg)



Name of
Commercial Dye
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct
Direct

Direct
Direct
Direct
Direct

Direct
Direct
Direct
Direct
Blue
Blue
Blue
Blue
Blue
Blue
Blue
Blue
Blue
Blue
Blue
Blue

Brown
Brown
Brown
Brown

Black
Black
Black
Black
25
78
80
106
160
189
191
218
251
U-l
M-2
Subtotal

113
115
116
Subtotal

2
62
80
M-l
No.
of
Sites
1
1
2
1
1
1
2
1
1
1
1
13

1
2
1
4

1
1
2
1
No. of
Weigh-
ings
1
4
7
1
2
1
5
4
2
1
8
36

3
12
1
16

1
1
9
1
Total
Weighed

0
0
1
0
3
0
9
0
5
7
3
32

0
8
0
8

0
0
32
8
(kg)
.019
.051
.085
.381
.895
.026
.744
.463
.926
.045
.836
.471

.049
.042
.012
.103

.644
.038
.196
.676
Per
Per
Weigh-
Site
0.
0.
0.
0.
3.
0.
4.
0.
5.
7.
3.
2.

0.
4.
0.
2.

0.
0.
16.
8.
019
051
543
381
895
026
872
463
926
045
836
498

049
021
012
026

644
038
098
676
0
0
0
0
1
0
1
0
2
7
0
0

0
0
0
0

0
0
3
8
ing
.019
.013
.155
.381
.948
.026
.949
.116
.963
.045
.480
.902

.016
.670
.012
.506

.644
.038
.577
.676

Number
Encountered











11 Blue
Direct Dyes



3 Brown
Direct Dyes




Direct Black Subtotal
TOTAL—DIRECT DYES
49
      12
153
        41.554
119.644
          8.311  3.463
2.442  0.782
 4 Black
   Direct Dyes

38 Direct Dyes,
   all colors
                                                                     A-25

-------
                             Table A-14

            TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
                    INDIVIDUAL DYES ENCOUNTERED
                             (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
REACTIVE DYES (used

Reactive Yellow 3
Reactive Yellow 15
Reactive Yellow 25
Reactive Yellow 27
Reactive Yellow 37:1
Reactive Yellow 58
Reactive Yellow 64
Reactive Yellow 125
Reactive Yellow 160
Reactive Yellow U-2
Reactive Yellow
Subtotal

Reactive Orange 16
Reactive Orange 70
Reactive Orange 82
Reactive Orange
Subtotal

Reactive Red 40
Reactive Red 43
Reactive Red 94
Reactive Red 120
Reactive Red 152
Reactive Red 168
Reactive Red 180
Reactive Red 198
Reactive Red U-2
Reactive Red U-3
Reactive Red U-4
No.
of ,r
Sites
at 6 of

1
1
1
2
1
1
1
1
1
1

11

1
i;
i

3

2
2
1
2
1
1
1
1
1
1
1
Reactive Red Subtotal 14


No. of
Weigh-
ings
24 sites)

1
3
2
8
1
2
1
39
1
1

59

3
1
1

5

11
2
2
2
1
2
3
2
1
1
41
68

Total
Weighed
(kg)


0.692
0.870
0.230
1.447
0.043
0.642
0.117
13.267
0.101
0.004

17.413

3.832
1.112
0.068

5.012

1.196
2.949
16.991
3.270
1.602
0.247
5.820
9.060
0.030
0.036
15.664
56.865


Per
Site


0.692
0.870
0.230
0.724
0.043
0.642
0.117
13.267
0.101
0.004

1.583

3.832
1.112
0.068

1.671

0.598
1.475
16.991
1.635
1.602
0.247
5.820
9.060
0.030
0.036
15.664
4.062

Per
Weigh-
ing


0.692
0.290
0.115
0.181
0.043
0.321
0.117
0.340
0.101
0.004

0.295

1.277
1.112
0.068

1.002

0.109
1.475
8.496
1.635
1.602
0.124
1.940
4.530
0.030
0.036
0.382
0.836


Number
Encountered
Sites: 4/9,2/1,
8/6,4/3,
6/5,9/1.










10 Yellow
Reactive Dyes




3 Orange
Reactive Dyes











11 Red
Reactive Dyes
A-26

-------
                               Table A-14

              TEXTILE DYE WEIGHING ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                               (Continued)
Average Weighed
(kg)

Name of
Commercial Dye
Reactive Violet 5
Reactive Violet 33
Reactive Violet
Subtotal

Reactive Blue 7
Reactive Blue 10
Reactive Blue 18
Reactive Blue 21
Reactive Blue 27
Reactive Blue 29
Reactive Blue 52
Reactive Blue 114
Reactive Blue 116
Reactive Blue 137
Reactive Blue U-l
Reactive Blue U-3
Reactive Blue U-4
Reactive Blue U-5
Reactive Blue
Subtotal

Reactive Black 5
Reactive Black U-l
No.
of
Sites
1
1

2

1
1
1
2
2
1
1
1
1
1
1
1
1
1

16

1
1
No. of
Weigh-
ings
1
2

3

2
5
1
4
2
5
1
1
2
1
2
32
3
1

62

1
6
Total
Weighed
(kg)
0.045
0.852

0.897

10.352
5.990
1.007
5.930
0.394
0.810
0.166
0.960
13.161
11.570
0.202
4.130
0.034
1.272

55.978

8.787
0.326

Per
Site
0.045
0.852

0.449

10.352
5.990
1.007
2.965
0.197
0.810
0.166
0.960
13.161
11.570
0.202
4.130
0.034
1.272

3.499

8.787
0.326
Per
Weigh-
ing
0.045
0.426

0.299

5.176
1.198
1.007
1.483
0.197
0.162
0.166
0.960
6.581
11.570
0.101
0.129
0.011
1.272

0.903

8.787
0.054

Number
Encountered



2 Violet
Reactive Dyes















14 Blue
Reactive Dyes


Reactive Black
  Subtotal
TOTAL—REACTIVE DYES   48
204
  9.113    4.557  1.302    2 Black
                          Reactive Dyes

145.278    3.027  0.712  42 Reactive Dyes,
                         all colors
                                                                     A-27

-------
                                Table A-14

              TEXTILE DYE WEIGHING ROOM MONITORING  STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                                (Continued)
Average Weighed
(kg)
Name of
Commercial
VAT
Vat

Vat

Vat
Vat
Vat

Vat

Vat

No.
of
Dye Sites
No. of
Weigh-
ings
Total
Weighed
(kg)
Per
Site
Per
Weigh-
ing
DYES (used at 1 of 24 sites)
Yellow

Orange

Violet
Violet
Violet

Blue 6

Brown

2

2

1
13
Subtotal



M-l

1

1

1
1
2

1

1

1

1

1
2
3

2

1

0

2

1
1
2

0

1

.785

.352

.562
.325
.887

.861

.962

0

2

1
1
1

0

1

.785

.352

.562
.325
.444

.861

.962

0

2

1
0
0

0

1

.785

.352

.562
.663
.962

.431

.962

Number
Encountered
Site: 8/8.
1 Yellow
Vat Dye
1 Orange
Vat Dye


2 Violet
Vat Dyes
1 Blue
Vat Dye
1 Brown
Vat Dye
TOTAL—VAT DYES
 8.847
1.475  1.106
6 Vat Dyes,
  all colors
NAPHTHOL DYES (used at 1 of 24 sites)

Azoic Diazo
   Compound 13         1      2

Azoic Coupling
   Compound 17         1      6
                       Site:  6/6.


24.932   24.932 12.466   1 Naphthol  Salt


14.700   14.700  2.450   1 Naphthol  Dye
FLUORESCENT WHITENING AGENT  (used at 1 of 24 sites)

FWA  (non-dye)  61        11       0.045
                       Site: 3/8.
          0.045   0.045  1 Fluorescent
                         Whitening
                         Agent
A-28

-------
                                Table A-14

              TEXTILE DYE NEIGHING  ROOM MONITORING STUDY:
                      INDIVIDUAL DYES ENCOUNTERED
                                (Continued)
                                              Average Weighed
                                                  (kg)
Name of
Commercial Dye
All Black Dyes
No.
Of
Sites
26
No. Of
Weigh-
ings
73
Total
Weighed
(kg)
298.04
Per
Site
11.46
Per
Weigh-
ing
4.08
Number
Encountered
22 Black
Dyes
                                              13.55 kg Weighed
                                                   per Average
                                                   Black Dye
All Non-black Dyes    390    1355     1063.27
 2.73   0.78    266 Non-black
                   Dyes

 4.00 kg Weighed
      per Average
      Non-black Dye
TOTAL—ALL DYES
416
1428
1361.31
3.27
0.95
288 Total
Dyes
                      24 sites
        4.73 kg Weighed,
            Average per Dye

56.72 kg Weighed,
      Average per Site
                                                                      A-29

-------
                                       Table A-15

                 CONTROLS AND  PERSONAL PROTECTIVE EQUIPMENT
                             AND  ENGINEERING CONTROLS
      Respiratory Protective Equipment
      Used During Dye Handling*
                                                  Dermal Protection Used
                                                  During Dye Handling
Site  Disposable
ID    Dust Mask
                  Neg.  Press.  Pos. Press  Gloves
                  AP-Resp.     AP-Resp.
                                                  Apron/  Boots/        Safety Glasses/
                                                  Smock  Safety Shoes  Goggles
1/0   yes                                  yes     no      no
1/6               yes                     no      no      no
2/1               yes                     yes     yes     no
2/4   no respiratory protection used        no      no      no
2/7               yes                     no      no      yes
3/0   yes                                  yes     no      yes
3/3   yes                                  no      no      yes
3/8   yes                                  yes     yes     yes
4/1   yes                                  yes     no      no
4/3   no respiratory protection used        **      no      yes
4/6   no respiratory protection used        no      no      no
4/9                           yes         yes     yes     yes
5/2   yes                                  yes     no      no
5/4   yes                                  yes     no      no
5/9   yes                                  no      no      yes
6/2   yes                                  yes     yes     yes
6/5               yes                     yes     yes     no
6/6   yes                                  yes     no      no
7/7               yes                     yes     yes     no
7/9   no respiratory protection used        yes     yes     no
8/0               yes                     no      no      yes
8/6   yes                                  yes     yes     no
8/8   yes                                  no      no      no
9/1   no respiratory protection used        yes     no      no
                                                                       no
                                                                       no
                                                                       yes
                                                                       no
                                                                       no
                                                                       yes
                                                                       no
                                                                       no
                                                                       no
                                                                       **
                                                                       no
                                                                       yes
                                                                       no
                                                                       no
                                                                       no
                                                                       yes
                                                                       yes
                                                                       no
                                                                       no
                                                                       no
                                                                       no
                                                                       yes
                                                                       no
                                                                       no
 *Neg.  Press. AP-Resp. — negative pressure air-purifying respirator; Pos. Press.  AP-Resp. —
  positive pressure air-purifying respirator.
**These protective items were worn only when weighing caustics,  not when handling dyestuffs.
A-30

-------
                                   Table A-15

                CONTROLS  AND  PERSONAL PROTECTIVE  EQUIPMENT
                           AND  ENGINEERING CONTROLS
                                   (Continued)
      Dye Weigher Activities
Site   Eat/Drink in   Smoke in
ID    Weighing Room  Weighing Room
Engineering Controls
Local exhaust  Wet mop floors
ventilation
1/0
1/6
2/1
2/4
2/7
3/0
3/3
3/8
4/1
4/3
4/6
4/9
5/2
5/4
5/9
6/2

6/5
6/6
7/7
7/9
8/0
8/6
8/8
9/1
no
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
yes

no
yes
no
no
yes
yes
no
no
yes
***
no
no
yes
no
yes
yes
no
yes
yes
no
no
no
no
no

no
yes
no
no
yes
no
no
no
no
no
no
no
no
no
no
no
no
no
no
no
yes
no
no
no

no
yes
no
no
no
no
no
no
yes, reportedly between shifts
information not recorded
information not recorded
information not recorded
information not recorded
yes, reportedly between shifts
yes, reportedly weekly
information not recorded
information not recorded
yes, reportedly between shifts
yes, frequency not reported
yes, reportedly between shifts
information not recorded
yes, observed during monitoring period
yes, reportedly between shifts
yes, reportedly weekly.
dry swept between shifts
information not recorded
information not recorded
information not recorded
information not recorded
yes, reportedly weekly
information not recorded
no, reportedly dry swept only
no, reportedly dry swept weekly
***Employee used chewing tobacco while in the drug room.
                                                                              A-31

-------
                             Table A-16

              USE OF PROTECTIVE EQUIPMENT AND PERSONAL
                   HYGIENE PRACTICES AT EACH SITE
                                              Number of
Variable                                        Sites
Use of Respirators
     Disposable Dust-Mask                        11
     Negative Pressure Air, Purifying Type        6
     Positive Pressure Air, Purifying Type        1
     None                                         6

Dermal Protection
     No  Protection                                4
     Gloves  Only                                  6
     Shoes Only                                   4
     Gloves  and Apron                             2
     Gloves, Apron, Eyewear                       2
     Gloves, Apron, Shoes                         1
     Gloves, Shoes, Eyewear                       3
     All Four Types                               2

Weigher  Activities in Drug Room
     Eat/Drink Only                               2
     Smoke Only                                   6
     Both                                         3
     None                                        13
                     >
                     >
Engineering  Controls in Place
     Local Exhaust                                2
     Wet Mop Observed                            13
     Wet Mop Indicated/Not Observed              5
     None                                         4
 A-32

-------
              Appendix B
        TEXTILE DYEING PLANTS:
POPULATION AND SUBPOPULATION ESTIMATES

-------
     This appendix contains two types of estimates.  First there  are
estimates for the numbers of plants and weighers included in certain
categories.   These estimates are produced by summing the  plant and
weigher level weights (described in Chapter 7)  to produce estimates
for the categories.   The categories break the population into groups
based upon information obtained from the in-plant monitoring.  For
example, in Table B-l, the first variable in the table is Management
Type (Vertical, Commission or Both).   The table shows the number of
sample cases represented (for example 15 for the Vertical only
Management Type), the estimated number of plants in the population
having vertical management — 596,  and an estimate of the total
number of weighers at such plants — 1,858.  The remainder of the
line for each of the entries contains estimates of commercial dye
concentration for the specific subgroup of the population defined in
the far left column.

     It is important to remember that the survey was not designed to
produce accurate estimates in subgroups of the population so finely
configured.   In many of the subgroups shown there are only 2 or 3
cases;  in some there are only 1 case.  Having such small numbers in
the individual cells makes the prospect of making inferences from
the data very remote.  However, the data are useful to present from
a general informational point of view.

     The second type of estimate included in Table B-l is for
concentration of commercial dye.   These estimates are broken down by
subgroups of the population.  Continuing the example started above,
in the first row of  the table (vertically managed plants),  the
commercial dye concentration for the  population of plants is 0.18
mg/m3.   This means that  for  the universe of plants, the  survey
produced an estimate for the typical  vertically managed plant as
having an airborne dye concentration  for commercial dye of 0.18
mg/m3.   The estimate  for the population of weighers in vertically
managed plants is 0.19 mg/m3.  Again, this can be interpreted as
saying that the typical weigher in one of these plants would be
exposed to that much commercial dye per air volume breathed.

     The final comment to make relates to the inclusion of standard
errors of estimates.   These are included for estimates of commercial
dye concentration for both the plant  and weigher populations.  They
are used as follows.   Construct an interval about the estimates
which is equal to approximately twice the standard error.   This
interval is a 95% confidence interval.   Roughly speaking,  if two
intervals for two different classes of the same variable do not
overlap, the estimates for these classes are significantly
different.  Again,  the idea is illustrated with an example.   For the
variable, Management Type,  and for the groups vertical versus
commission,  we have  the following situation:
                                                                 B-l

-------
                  Estimate
                     Standard  Error
                       Interval
Vertical

Commission
0.18

0.084
  0.070

  0.031
 (0.04  - 0.32)

 (0.022 - 0.15)
It  is  clear  from this example  that  the intervals overlap.
Therefore, there is  not a  significant difference uncovered.

       As mentioned above,  the  sample sizes per cell are so  small  that
it  is  not appropriate to highlight  such  analyses;  and  this analysis
was, therefore,  not  included in Chapter  7.   It is  provided,  here,
for the readers7  use to facilitate  their understanding of the table.
It  can be considered as producing qualitative information about  the
data which can help  the reader better understand the population  of
interest.
                                    Table B-l

                   COMMERCIAL DYE CONCENTRATION ESTIMATES
       BROKEN DOWN BY  OTHER VARIABLES  MEASURED  DURING MONITORING
Variable
                           Estimated Universe
                                              Plant-Weighted
                                               Worker-Weighted
 No. of
Sampling  No. of   No. of
 Cases    Plants   Workers
Commercial Dye
Concentration
   (mg/m')
                                                       Standard
                                                        Error
Commercial Dye
Concentration
  (mg/m')
Standard
Error
Management Type
    Vertical            15       596    1,858       0.18
    Commission           3       119      362       0.084
    Both                4       148      469       0.27

Ownership
    Private             12       444     1,529      0.19
    Public             10       419     1,159      0.17

Number of
Dyeing Machines
    1 to 5              5       177      670      0.074
    6 to 10              4       181      333      0.23
    11 to 15             7       267     1,007      0.22
    16 to 75             6       238      678      0.18

Production Volume
 (million pounds)
    0.0 to 5.0          11       448     1,011      0.15
    5.1 to 10.0           4       181      687      0.089
    10.1 to 20.0          4       115      547      0.44
    20.1 to 25.0          3       119      444      0.21

Dyeing Processes
Performed
    Batch Dyeing         17       720     2,056      0.20
    Continuous Dyeing      2       58      288      0.063
    Both                2       58      173      0.11
    Printing             1       29      173      0.019
                                     0.07
                                     0.031
                                     0.11
                                     0.053
                                     0.096
                                     0.015
                                     0.11
                                     0.14
                                     0.048
                                     0.050
                                     0.017
                                     0.24
                                     0.037
                                     0.065
                                     0.017
                                     0.032
                                      NA
                        0.19
                        0.095
                        0.2
                        0.18
                        0.17
                        0.056
                        0.25
                        0.2
                        0.23
                        0.13
                        0.074
                        0.36
                        0.23
                         0.21
                         0.048
                         0.11
                         0.019
              0.074
              0.029
              0.1
              0.056
              0.10
              0.016
              0.11
              0.13
              0.063
              0.052
              0.017
              0.21
              0.041
              0.069
              0.016
              0.032
                NA
 B-2

-------
                                Table B-l

                COMMERCIAL DYE CONCENTRATION ESTIMATES
     BROKEN DOWN BY OTHER VARIABLES  MEASURED  DURING MONITORING
                               (Continued)
No. of
Sampling
Variable Cases
Shifts
3-8 hour shifts
2-12 hour shifts
2 8 hour shifts
1-10 hour shifts
1-8 hour shifts
Time Weigher in Drug Room
Less than 25%
Between 25% and 49.99%
Between 50% and 74.99%
At least 75%
Number of Weighers
1
2
3
4
6
8
Mass weighed (kg)
0 to 10 kg
10.01 to 30 kg
30.01 to 60 kg
60.01 to 80 kg
80.01 to 284.3 kg
Number of dyes weighed
0 to 10
11 to 20
21 to 30
31 to 46
Number of dye weighings
0 to 20
21 to 40
41 to 60
61 to 80
81 to 100
101 to 249

11
4
4

2

5
6
4
7

3
5
7
2
4
1

4
5
5
4
4

5
9
5
3

5
4
4
2
3
4
Estimated
No. of
Plants

382
148
214
29
90

243
238
115
267

119
243
267
90
115
29

148
177
210
181
148

144
391
243
86

210
181
115
58
119
181
Universe
Plant -Weighted
Commercial Dye
No. of Concentration
Workers (mg/m3)

1,492
650
428
29
90

662
654
403
970

119
485
802
362
691
230

584
506
415
567
617

547
1,225
715
201

683
572
460
259
415
300

0.22
0.065
0.22
0.013
0.15

0.08
0.10
0.13
0.37

0.12
0.21
0.23
0.07
0.21
0.039

0.066
0.088
0.11
0.22
0.46

0.046
0.083
0.30
0.52

0.066
0.084
0.069
0.76
0.22
0.27
Worker-Weighted
Commercial Dye
Standard Concentration
Error (mg/mj)

0.097
0.013
0.096
NA
0.037

0.019
0.023
0.060
0.14

0.044
0.085
0.14
0.018
0.091
NA

0.017
0.031
0.013
0.12
0.20

0.013
0.014
0.077
0.280

0.016
0.021
0.027
0.31
0.088
0.093

0.22
0.059
0.22
0.013
0.15

0.071
0.076
0.18
0.32

0.12
0.21
0.23
0.07
0.21
0.039

0.049
0.078
0.11
0.16
0.44

0.040
0.077
0.33
0.61

0.057
0.081
0.057
0.61
0.28
0.31
Standard
Error

0.089
0.013
0.096
NA
0.037

0.018
0.022
0.071
0.13

0.044
0.085
0.14
0.018
0.091
NA

0.016
0.025
0.014
0.099
0.17

0.012
0.014
0.066
0.32

0.016
0.017
0.027
0.31
0.097
0.098
*Includes only the 22 plants with a valid concentration measurement.
                                                                        B-3

-------
    Appendix C
ANALYTICAL METHODS

-------
     This appendix contains the details of the development  of  the
analytical methodology.  This includes the evolution of the method,
statistical techniques developed to evaluate the accuracy of
estimates, quality control procedures followed and some analysis of
the methodology.  For additional information, the interested reader
may refer to the article "A Spectral Photometric Method of Total
Levels of Textile Dyes in Air Monitoring Filters" in the Journal of
the American Industrial Hygiene Association.1

I.   INITIAL EVALUATION OF ANALYTICAL METHODS

     Many different analytical approaches were evaluated with  regard
to their applicability to the anticipated complexity and low levels
of dye mixtures in the drug room samples.   The air filters could
contain as many as 50 different textile dyes from up to six
different dye classes.  There were several general requirements of
the analytical method to be used in the study.  One of the most
important requirements was the ability to detect total dye amounts
of less than 20 micrograms (p.g) , since low levels were anticipated
on many of the air filters.  Because of these possible low levels,
another crucial requirement of the method was to have dye recoveries
of at least 60% from the air filters (but no greater than 140% of
the theoretical values).   An additional prerequisite of the
analytical method was to minimize the amount of interference caused
by the presence of any non-dye compounds on the air filters.
Finally, the size of the study made it essential that the analysis
time and cost of the procedure not be too great for each plant site.

     The most conventional, straightforward analytical approach
would be to utilize high performance liquid chromatography (HPLC) to
separate the component dyes in the sample and employ a programmable
variable wavelength ultraviolet/visible absorbance detector to
quantitate each dye at its most sensitive wavelength.  Using HPLC,
however, was very problematic for the drug room study.  The foremost
problem was the anticipated need to develop multiple chromatography
systems to analyze dyes from different dye classes.  Coupled with
that requirement was the probable need to "fine tune" these
chromatography systems to optimize the recovery and separations
required for each unique plant site.  The total number of plant
sites to be sampled and the sheer number of dyes which could be
encountered led to the conclusion that HPLC,  and chromatography in
general, would not be a viable analytical technique for the drug
room study-

     Due to the sample complexity problems, a nonspecific, general
dye screening procedure appeared to present the best solution to the
     1Harbin DN, Going JE,  Breen JJ.  1990.  A spectrophotometric
method of estimating'total levels  of textile  dyes on air monitoring
filters.   J Am Ind Hyg Assoc 51(4):185-193.
                                                                  C-l

-------
analytical challenge.  This approach satisfied the needs of the
study, since individual dye quantitations were not essential in the
overall results.  A method based on integrated absorbance
measurements over the visible wavelength range was judged to be
subject to few interferences and sufficiently sensitive to detect
levels of dyes in the range of 10 to 20 jig.

II.   DEVELOPING  THE  ANALYTICAL METHOD

      A.  Theoretical  Concepts

      Conventional  dye quantitation methods which  employ
spectrophotometry are based on the Beer-Lambert law.  This law is
often expressed  by the following formula:

                            A = a x b x c

where A  is the absorbance of the absorbing species at a specified
wavelength, a is the absorptivity constant of the absorbing species
at the specified wavelength, b is the absorption pathlength, and c
is the concentration of the absorbing species in the sample.  For
the  case of spectrophotometric measurements on solutions of the same
absorbing species, the a and b terms will be constant and the
Beer-Lambert law can be simplified to the statement that absorbance
at the specified wavelength is directly proportional to the
concentration of the absorbing species.  Since dye mixtures having
absorbance across  the entire visible region could be encountered in
the  drug room study, it is necessary to measure absorbances
accordingly to  ensure that all dyes present are detected.
Therefore, a total integrated absorbance from 380 to 750 nm was
used.

      The approach used in this  study basically  extended the
Beer-Lambert law over a designated wavelength interval:

                               =  a x b x c
 where A^,. is the total measured absorbance of the sample solution
 over a specified wavelength interval and as is the "spectral"
 absorptivity constant for the absorbing species  over the  specified
 wavelength  interval.   The total  absorbance of  a  sample solution  can
 be obtained by integrating the area beneath the  absorbance  spectrum
 over the specified wavelength interval and converting the area to
 absorbance  units.   In a mixture  of n absorbing components,  the
 following relationship will exist if Beer's law  is  obeyed:
                                   n
                              = b    asi x
 C-2

-------
where ATO,, is the total measured absorbance of the mixture, b is the
absorption pathlength, asl is the spectral absorptivity  constant  for
the ith component, and c± is the concentration of the ith component.

     Taken  as a whole, the  absorbance  characteristics of a dye
mixture can be  considered to be a composite of the individual  dye
.absorptivity constants.  As a result, the average value  of  the
individual dye  absorptivity constants  (as)  is a reasonable
approximation of the overall absorptivity of the dye mixture,
provided that no single component dye has a grossly disproportionate
presence in the mixture.   Similarly, the individual dye
concentrations  in the mixture can be summed together to  result in a
total dye concentration  (CTOT) .   In  applying this quantitation
method, the individual as constant  for  each of the possible
component dyes  is experimentally determined.  An average as value
(as)  for the entire  group of possible component dyes  is  then
calculated.  By measuring the observed A^T for the sample solution
and knowing the as value  for the dye group and the pathlength,  the
concentration of the total amount of dyes present in solution  (CTOT)
can be calculated using the equation:

                                  •"TOT

                                 b x as

The total weight of the dyes in solution therefore can be determined
from the concentration and the volume of the sample solution.

     Because the method  is based on  an average spectral  absorptivity
constant from many different dyes,  the result of the calculations
will be considered to be a total dye estimate rather than a
conventional quantitation.  The advantages that this method has over
more conventional analytical approaches are substantial in terms of
analysis time and cost.

     B. Method  Development

     Method development work was initiated to determine  the
magnitude of the uncertainties of the total dye estimate method.  In
addition, a great many practical aspects to the development of a
general estimation method for textile dyes were investigated in the
initial phase of the developmental work.  ETAD provided samples of
23 "typical" textile dyes from the following dye classes:  acid,
basic,  direct,  disperse,  and reactive.   Initial work was spent
devising a solvent mixture which could dissolve all 23 dyes.  A
mixture of dimethylsulfoxide (DMSO) and water, 9:1 (v/v), was
discovered to be an excellent solvent for all of the dyes.
Subsequent laboratory work was directed towards interfacing a
spectrophotometer and computer-based integration hardware.
                                                                  C-3

-------
     Laboratory work then focused on the spectra from individual
dyes.  Analysis of individual dyes confirmed that A^o? values are
linear over a wide range of dye concentrations.  The observed range
for the experimentally determined as constants  in the visible
wavelength region  (380-750 nm) for each of the 23 test dyes was
approximately an order of magnitude.  There was no apparent
relationship between the as  value and either  dye color or class.   If
these dyes are truly typical, then use of an average as  value to
estimate a group of dyes is judged to be sufficiently accurate in
most circumstances.

     Validation of the analytical method consisted of three phases.
The goal of the first phase was to determine the magnitude  of the
relative errors when various dye mixtures were prepared and
analyzed.  The criteria for acceptable performance of the method was
that the relative  errors not exceed 50% of the true value.  The
second phase of the validation process consisted of developing a
satisfactory filter extraction procedure which would be used to
analyze the plant  site air filters.  The validation criterion was
that dye recoveries be in the range of 60 to 140% of theoretical
values.  The final phase of method validation was to successfully
analyze spiked filters and performance audit samples associated with
a pilot study site.

     One of the characteristics  of  using an  average  as constant (as)
derived from all possible component dyes, is the fact that  errors
can arise whenever predominant components on the air filter consist
of "outlier" dyes  (i.e., dyes with as constants that  are
significantly different from the overall as value).   In  these
instances the errors incurred from basing the total dyes' estimation
on the as constant can be substantial.   Many times  such  a "worst
case" scenario can betidentified by comparing the absorbance
spectrum of the sample with that of the individual component dyes.
If the predominance of outlier as dyes on the air  filters can be
confirmed by such  a comparison, then the numerical value of as could
be adjusted appropriately.  The  "worst case" scenario is more
problematic if it  cannot readily be identified from the  sample
absorbance spectrum.  For this reason a significant proportion of
the experimental work in developing the method was performed on
known dye mixtures prepared to simulate some "worst case" scenarios.
The resulting data were used to gauge the magnitude of the  errors
arising when the as constant was employed to make  the total dye
calculation.

     Experimental work on known dye mixtures began with relatively
simple solutions  and eventually  grew  in complexity to better
simulate the expected composition of  actual  samples.  Work  with
mixtures containing six  dyes  is  summarized in  Table  C-l.  Six-dye
mixtures exhibited relative errors  ranging from -11  to  +17% in  the
total dye estimate based on mean as value when the total dye amount
was  approximately 160 to 200  ^.g.  A 10-fold dilution of one of  these
 C-4

-------
                              Table C-l

        RESULTS OF  TOTAL DYE  ESTIMATIONS FOR 6-DYE MIXTURES"

Dye
Mixture
No.
1
2
3
4
5
6
7
8
9
Estimated
Total Dye
Concentration
(jig/ml)b'c
53.2
46.3
36.7
39.5
38.7
49.2
43.1
4.69
2.44
Actual
Total Dye
Concentration
(ng/ml)c
50.0
49.6
41.1
41.1
41.4
41.9
41.2
4.10
1.63

Percent
Relative
Error
+ 6
- 7
- 11
- 4
- 6
+ 17
+ 5
+ 14
+ 50
"Dye mixture  composition:  Acid Blue  40,. Basic Blue  3,  Direct
 Red 80, Disperse Orange 29, Disperse Yellow 23, Reactive
 Violet 5.
bEstimate  based on mean  as constant.
"Solution  volume is  approximately  4 ml.
mixtures produced a 3-fold increase in the relative error of the
quantitation, while a 25-fold dilution resulted in an increase in
the relative error by an order of magnitude.  These increases in the
relative error of the total dye estimate with decreasing
concentration are ascribed to the very low absorbances being
measured and the difficulty in integration of the resulting
absorbance spectrum.

     Work with 10-dye mixtures, as shown in Table C-2, concentrated
on the simulation of "worst case" scenarios.  Total dye amounts
varied from 88 to 164 jig.  The observed relative errors for the dye
estimates based on mean
-39 to +40%.
as  constants  for these mixtures  ranged from
     Twenty-dye mixtures were prepared to simulate the complex
samples which could be encountered in actual plant analyses.  Within
these mixtures the concentration ratios of the component dyes varied
significantly, ranging from a factor of 6 in some mixtures to as
much as a factor of 25 in others.  The total dye amounts were
similar to those employed for previous experiments (i.e., 144 to
220 jig).  The results of these experiments are shown in Table C-3.
                                                                  C-5

-------
                              Table  C-2

        RESULTS OF TOTAL DYE ESTIMATIONS FOR 10-DYE MIXTURES"

Dye
Mixture
No.
1
2
3
4
5
6
Estimated
Total Dye
Concentration
(ng/ml)b'°
43.0
25.2
38.3
55.2
31.8
14.5
Actual
Total Dye
Concentration
(^g/ml)c
40.6
41.1
40.0
40.5
22.6
22.5

Percent
Relative
Error
+ 6
- 39
- 4
+ 36
+ 40
- 36
"Dye mixture  composition:   Acid Blue 40,  Basic Yellow 11,  Direct
 Blue 15, Direct Yellow 4, Disperse Brown 1, Disperse Orange 29,
 Disperse Red 17, Disperse Red 177, Reactive Black 5, Reactive
 Violet 5.
"Estimate based on mean as constant.
°Solution volume is approximately 4 ml.


The observed relative errors of the total dye estimates based on
mean as constants varied from -20 to +26%.

     Total dye  estimates were  also calculated on the basis of a
weighted average as constant for the group  of dyes  rather  than a
mean value.  The results are shown in Table C-4.  This approach
takes the relative proportions of the component dyes in the mixture
into account and therefore should be more accurate.  The weighted
average-based calculations produced mixed results for the 10- and
20-dye mixtures.  Not surprisingly, little improvement in the
relative errors of the total dye estimates was observed when the
numerical value of the weighted average as  constant (weighted as)  was
similar to that for the mean as constant.   Substantial improvements
were generally  seen when the value of the weighted as was
significantly different from the mean as.   In a few cases  the
relative errors were larger when the weighted as was used  to  make
the total dye estimation compared  to the result obtained  from using
the mean as.   In these latter cases there was some  circumstantial
evidence that the measured absorbances were not obeying Beer's law.
The most likely explanation for this behavior is dye reaction or
interaction in  solution.  It should be emphasized that the magnitude
of the relative errors arising from the apparent deviations  from
Beer's law was  no  larger than  that observed when the mean as
constant was used  to make all  of the dye estimate calculations.
 C-6

-------
                              Table C-3

        RESULTS OP TOTAL DYE ESTIMATIONS FOR 20-DYE MIXTURES'
Estimated
Dye Total Dye
Mixture Concentration
Actual
Total Dye
Concentration
No. (p.g/ml)b'c (p,g/ml)°
1 44
2 43
3 49
4 45
5 45
6 45
7d 21
8" 12
.12
.77
.89
.36
.14
.99
.88
.18
52
54
53
36
43
43
21
10
.03
.95
.91
.1
.1
.1
.5
.76
Percent
Relative
Error
_
-
-
+
+
+
+
+
15
20
7
26
5
7
2
13
"Dye mixture  composition:  Acid Blue  40, Acid Red  337, Acid
 Yellow 151,  Basic Blue 3, Basic Red 15, Basic Yellow 11, Direct
 Blue 15, Direct Red 80, Direct Yellow 4,  Disperse Blue 56,
 Disperse Blue 79, Disperse Brown 1,  Disperse Orange 29, Disperse
 Orange 30, Disperse Red 60, Disperse Red 177,  Disperse Yellow
 23,  Reactive Black 5, Reactive Blue 4,  Reactive Violet 5.
bEstimate based on mean as constant.
°Solution volume  is  approximately  4 ml.
dl:l dilution of  Mixture 6.
•1:3 dilution of  Mixture 6.
     On the basis of the results obtained from analyzing the known
dye mixtures,  it was concluded that using as  constants  to estimate
total dye amounts was a viable method.  The relative errors of the
results were in the 10 to 40% range when the mean as value  was  used.
If additional information about the relative amounts of the
component dyes can be obtained, then the accuracy of the method can
be improved in many instances.  In the case of the air filters,
knowledge about the amounts of each dye which were handled during
the air monitoring period allows a weighted as constant to  be
calculated and used to make the total dye estimation.   The
underlying assumption for this approach is that all dyes are equally
dusty and that dyes will appear on the air filters in proportion to
their use at the drug room site.  Although this assumption is, in
reality, not always true, its basic premise is nevertheless the most
logical one given the lack of any other data.  Therefore, weighted
as constants were used for  all dye  estimate  calculations in the
study.
                                                                  C-7

-------
                              Table C-4

            TOTAL DYE ESTIMATES BASED ON WEIGHTED AVERAGE
                   SPECTRAL ABSORPTIVITY CONSTANTS
No. of Dyes                                Percent Relative Error
in Mixtures           Mixture No."         of Total Dye Estimate
10 1
2
3
4
5
6
20 ' 1
2
3
4
5
6
7
8
+ 5
+20
+32
+17
- 8
- 5
-15
+16
- 1
-41
-19
-12
-16
- 7
"For concentration and composition of the mixtures,  see Tables C-2
and C-3.
      Extraction efficiencies of dyes from PVC air filters  (personal
 sampling pump  size, 37 mm) were determined by spiking blank filters
 with  known dye mixtures  and  extracting them with  dye solvent.  The
 recovery of the dyes was measured in terms of the ratio of the A^
 values obtained from the reference standard and the spiked filter
 extract.  Recoveries ranged  from 63 to 108% when  the spike level was
 40 to 160 p.g.   The average relative difference in recovery values
 between duplicate  spiked filters was 7%  at the 40- to 160-jig  spike
 levels.  Both  the  precision  and the magnitude of  the dye recoveries
 were  observed  to decrease  with  decreasing dye concentrations.  No
 dye class appeared to have a characteristic recovery from PVC
 filters.

      A final validation of the proposed analytical method took place
 when  samples  from  a pilot  study (Plant  0/0) were  analyzed using the
 previously developed procedures.  A total of  33 bulk dye samples
 from  four dye  classes were analyzed along with six PVC air filters
 and their associated  cassettes.  No dye  solubility problems were
 encountered  and the air  filter  extracts  had more  than adequate
 absorbance to  calculate  total dye estimates.   The total dye
 C-8

-------
estimates obtained from the Plant 0/0 air filters were compared to
the gravimetric weights of the collected participates and were found
to be reasonable and internally consistent.

     C. Modifications of the Analytical Method

     It was discovered early in the study that many dyes from  the
basic class were subject to significant fading (i.e.,
decomposition)  over time when dissolved in the standard dye  solvent
prior to analysis.   Although a change in the pH of the dye solvent
is sufficient to stabilize these basic dyes, the stability of the
remaining dye classes at this different pH is very uncertain.  As an
alternative it was decided to maintain the standard dye solvent
composition and instead initiate a narrow analysis time window into
the analytical procedure.   As a result, any dye decomposition
occurring in the air filter extracts would presumably occur at the
same rate as that observed for the corresponding bulk dye samples.

     The basic dye reactivity also required a modification of the
filter spiking procedure used to determine the overall dye recovery
of the group of dyes being analyzed for each plant  site.   The
modification ensured that the reference standard used to calculate
dye recoveries was exposed to the same potential dye fading
conditions as was the spiked filters.   This eliminated a major
source of potential bias in the dye recovery experiments for each
plant site.

III. OBTAINING UNCERTAINTY RANGES FOR THE DYE ESTIMATES

     Because of a variety of factors, the actual number and
proportions of dyes found on an air filter will vary.   As the
number and proportions of the dyes on the filter vary, the overall
as  constant of the mixture will vary.  Differences of the actual as
constant from the value represented by the weighted average will
lead to errors in the determination of the total concentration of
the dyes.   In order to estimate the magnitude of this  source of
error,  a method to account for this variability was sought.

     Each combination, C, of a number of dyes, r  (less than or equal
to the total number handled by the worker,  n), that could  be
trapped on the filter could give a different overall as constant.
In  addition,  even for a specified number of dyes, r,  on the filter,
there could be different overall as  constants  depending on which  set
of the dyes was found on the filter.  Some idea of the possible
variability of the method can be found by considering the different
subsets of dyes that could be present and the resulting average as
constant of each set of dyes.

     For a moderate to large number of. dyes,  the number of
combinations of subsets that could be on the filter becomes quite
large.  If a worker handled a total of n dyes during the sampling
                                                                  C-9

-------
period, there are nCr different subsets of r dyes.  The sum of all
possible subsets of n dyes is  (2n - 1),  which rapidly becomes too
large to work with.  Because of the large number of possible
combinations of dyes that could be found on the filter, it is not
practical to calculate all the possible combinations of dyes and the
resulting overall as constants that could be found on the filter.

      In order to address the  problem, a  simulation routine was
programmed.  This simulation randomly selects a sample of size r of
the n dyes handled by the worker and calculates the mean as
constant for that mixture.  This is repeated 250 times to give a
distribution of values for the average as constant for a mixture of
r dyes.  The simulation is run for each possible value of r from 1
to n for each plant.  The result is a set of numbers that  represent
possible mean as constants  for dye mixtures,  including  different
numbers of the dyes handled by the worker.

      This  set of numbers is used to estimate the  variability  that
one can expect to see in the average value of the overall as
constant for the dye mixture.  Upper and lower limits for the as
value can be estimated from the distribution.  These are then used
to calculate corresponding limits on the concentration of dye
material in the air estimated by the method.

      The  simulation assumes that  the probability  that  a  dye  will be
in the air and will be collected on the filter is proportional to
the amount of dye used.  Thus, in selecting a sample of dyes, the
dyes that were used in larger amounts have a higher probability of
being included in the sample.  The simulation uses probability
sampling with replacement and so generates samples that have a
composition of the dyes proportional to the amount of each dye used.
The distribution of the mean  as constants that results from the
simulation is used to' obtain confidence intervals for the mean as
constant for samples of each  size.  Typically, these intervals are
not symmetric about either the mean as constant or the resulting
total dye estimate.

      The  results of the  simulation are  empirical  error bounds on the
average as constant for each possible number of dyes  on the  filter.
In order to summarize the error estimates, a convention  was
established to use the results for the sample of  dyes that had the
smallest number of dyes in it that accounted for  at least 80% of the
total quantity of dyes used during the sampling period.  This
provides an error estimate based on the major use dyes.  Thus, the
summary error estimate is based on the simulation for the dyes that
account for 80% of  the dye material weighed out during the sampling
period.

      The  simulation is  run separately for the as constants of the
commercial dyes and for the as constants of the active colorant.
The a., constants of the active colorant of the dye are based on the
 C-10

-------
reported purity of the dye.  Any errors in the reported purity will
add to the uncertainty in the active colorant concentration
estimates.

     The  simulation routine is written in Basic and implemented on
an IBM-PC computer.  The program writes a file of the average as
constants to a floppy disk.  These numbers (250 for each number of
dyes) are then sorted, and summary statistics are prepared using a
commercial program to give the distribution of possible values used
to estimate the errors.  Use of the simulation is recommended with
the method to give an approximate range for the dye estimate, since
the uncertainty has been found to differ substantially depending on
the individual dye as  constants and the amounts of  each dye used.
An example of a statistically derived distribution of probable
commercial dye-based weighted as values for Plant 4/1  is shown in
Figure C-l.

IV.  APPLYING TOTAL DYE ESTIMATION METHOD TO THE ANALYSIS OF ACTUAL
     AIR  FILTERS

     The  absorbance spectrum of a known dye mixture can be used to
evaluate the general accuracy of a total dye estimate made on actual
air filters which contain those same dyes.   Since the total dye
estimate is calculated from the weighted average  as constant of all
the possible component dyes in the mixture,  the validity of using a
weighted average value can be assessed by a comparison of the
spectra of the sample filter and that obtained from a weighted
average dye mixture.  Such a dye mixture is composed of all the
possible component dyes in proportion to their use at the plant
site.  If the resulting spectrum of this mixture  is similar to that
of the actual air filter,  then the basic assumption that dyes are
present on the air filters in proportion to their usage at the plant
site will be confirmed.

     Obvious dissimilarities between the spectra indicate that this
basic assumption may not apply,  in which case the use of a weighted
average as constant may lead to greater than  expected  errors in the
total dye estimate.

     In the case of the 25 plant sites analyzed in the drug room
study,  the spectrum of the dye mixture used for dye recovery
experiments was employed to appraise the use of a weighted average
as constant.  The dye  recovery mixture contained the dyes
comprising 80% of the total quantity handled by the worker.   The
dyes were present at levels proportionate to their usage,  and
therefore the mixture was a weighted average solution.  The lower
use dyes were not included in the dye mixture for logistical
reasons, as well as from the judgment that these  components would
not significantly influence the overall absorbance spectrum of the
solution.
                                                                 C-ll

-------
                                     Figure C-l
             SIMULATED COMMERCIAL DYE-BASED AVERAGE  ABSORPTIVITIES
                              (WEIGHTED)  FOR  FIVE DYES
3
V
150
140 -
130 -
120 -
110 -
100 -
 90 -
 BO -
 70 -
 60 -
 50 -J
 40 -
 30 -
 20 -
 10 -
  0
            Plant 4—1  Commercial  Dye Absorptivities
                             Weighted Sampling with Replacement
           Statistical
           Commercial
           Dye as (5%)
                        Weighted Average
                        Commercial Dye 8,
         0.25
                            Statistical
                            Commercial
                            Dye 5, (95%)
                                                Weighted Average Commercial Dye 5j =  1.99
                                                Statistical Commercial Dye 5s (5%) = 1.58
                                                Statistical Commercial Dye 3S (95%) = 2.46
                                                            -e-
                                               -B-
1.75
                         3.25        4.75        6.25
                          Mean Absorptivities for Five Dyes
7.75
9.25
    C-12

-------
     Several plant sites in the study had dissimilarities between
the dye recovery and the air filter spectra.  These dissimilarities
varied from moderate to substantial.   In at least two instances
(Plants 5/4 and 3/3)  low sample absorbances contributed heavily to a
significant lack of agreement between the air filter and the dye
recovery spectra.  In addition to low absorbances, the most common
cause of spectral dissimilarities is  disproportionate dye amounts on
the air filters compared to dye handling information.  Disparities
between filter amounts and the amounts expected based on dye
handling information are usually the  result of differences in dye
dustiness or the number of dye weighing operations for a few dyes in
the group.  In most cases it was possible to identify which dyes
were the most likely cause of the dissimilarity and what the general
consequences were for the accuracy of the total dye estimate.  The
more systematic approach of experimentally determining a logical dye
mixture which closely matches the spectrum of the air filter was not
performed due to time and cost constraints.   It would be a feasible
approach for a smaller study.

V.   DYE PURITY  INFORMATION TO CORRECT FOR ACTIVE COLORANT CONTENT
     IN THE DYESTUFF

     The use of  dye purity information for each of the possible dye
mixture components is an effective means of addressing some of the
uncertainties of the total dye estimate.  The active colorant
content of the dyestuffs can vary substantially for textile dyes,
ranging from very low to very high.  The commercial dye as constants
are experimentally determined with the inert constituents present,
so calculations made using a commercial dye-based weighted average
as constant  will  have  the inert  components incorporated  into  the
total dye estimate.  The use of dye purity information is essential
to compensate for this source of error in the total dye estimate.
Dye purity values are used to calculate an active colorant-based
weighted average as value for the  particular  group of dyes being
analyzed.  In this manner the total dye estimate is made in terms of
the active colorant content alone.

     There were  several disadvantages in using dye purities,
however.  One of the primary disadvantages is the problem of
obtaining this information from the dye manufacturers.   Dye purity
is often considered by the manufacturers to be proprietary
information, and consequently was often difficult and time-consuming
to obtain.  As a result, considerable time delays occurred before
enough information became available to allow active colorant-based
calculations.  Another disadvantage in using dye purity information
was the fact that the methods used to determine active colorant
content of a dye are not standardized throughout the dye industry.
Since the value  for active colorant content is dependent on the
particular assay technique which is employed, the dye purity values
are subject to varying degrees of uncertainty.  The additive effect
of these uncertainties can therefore produce a significant potential
                                                                 C-13

-------
error for active colorant-based total dye estimates when large
groups of dyes are involved.

     As discussed in Chapter  6, the  spectrophotometric analytic
method used to calculate active colorant required estimates of dye
purity values and the percent of active colorant in the commercial
dyes.  Table C-5 (at the end of this appendix) shows estimated dye
purities for all dyes weighed at the 24 sites included in the
survey.

VI.  MONITORING  LABORATORY  PERFORMANCE BY QUALITY ASSURANCE/ QUALITY
     CONTROL MEASURES

     There were  four primary  components of the quality
assurance/quality control (QA/QC) program which was used  throughout
this study.  The first component consisted of the  preparation of a
Quality Assurance Project Plans  (QAPPs) before the study began.  The
QAPPs contained a detailed  summary of the data quality objectives,
the analytical procedures which were to be followed,
responsibilities of key project personnel, and appropriate actions
to be taken when any QA/QC  requirements were not met.  The second
component of the QA/QC program involved a series of comprehensive
system audits scheduled at  the beginning, middle, and end of the
analytical work  for the study.  The  system audits were conducted by
the quality control coordinator  (QCC) and confirmed that all
experimental procedures and records were  consistent with the
requirements outlined in the QAPPs.

     The most extensive part  of  the  QA/QC program was the  periodic
analysis of performance audit samples  (PAS) during the  course of
all plant sample analyses.  Three PAS samples were to be analyzed
for each plant site. , The QCC prepared the samples and submitted
them to the laboratory technician for total absorbance  (ATOT)
determinations.  The PAS  samples were composed of individual dyes
from the plant site being analyzed and were unknown to the
technician performing the analyses.  Based on the  measured A^,.
value and the previously  established as constant  for the dye used,
the quality assurance coordinator  (QAC) calculated the found
concentration and compared  the result to the actual value.  The
results for a PAS sample were judged to be acceptable if the found
value was ±30% of the actual value.  Of the 82 PAS results obtained
during the study, only two  samples failed to meet these data quality
objectives.  This success rate was fully acceptable for the study-

     The  final  component  of the  QA/QC  program was an  audit  by  the
QCC of the raw data generated for the analyses of samples from each
plant site.  The QCC documented  that all of the  reported data met
the project's data quality  objectives.  In addition,  each plant
report was reviewed by the  QCC to assure consistency  with the
associated data.
 C-14

-------
     Due to rigorous analysis time windows specified in the QAPPs
for all sample analyses, extensive documentation was kept showing
that this time requirement was met for all of the samples.  In
addition, documentation of sample traceability was maintained.  A
bar code sample identification system was employed to facilitate
this process.

VII. ADVANTAGES OF THE TOTAL DYE ESTIMATE METHOD

     The total dye estimation method has several advantages
compared to conventional quantitation techniques.   Probably the
foremost advantage is the capability of analyzing highly complex
mixtures at very low levels.  The accuracy of any quantitation
method which is based on an average value for an intrinsic  property
of a class of compounds will tend to increase as the  number of
compounds being averaged increases.   Conversely, any  quantitation
method which is based on the determination of the  levels  for each
of the components in the mixture will tend to  decrease in accuracy
as the number of components in the mixture  increases.   This
decrease in accuracy, which is due to the  additivity of the
uncertainties for each component being  quantitated, will  be
magnified at trace levels, where uncertainties inevitably  increase.
For this reason the accuracy of the total dye estimation method can
conceivably be better than  that obtainable for conventional methods
when analyzing complex mixtures at trace levels.

     Another important advantage of the estimation method is the
relative simplicity of the technique compared to conventional
quantitation methods.  Because all components are analyzed
simultaneously, the total analysis time is considerably less than
that for other methods, which would probably require multiple
analytical systems to handle certain dye classes.   A direct benefit
of simultaneous analysis is the fact that the amount of associated
data will be at least an order of magnitude less than that derived
from individual component quantitation.

     Finally, another asset of the total dye estimate method is the
lack of sophisticated equipment or personnel training required.
Practically any spectrophotometer or chromatography data system
would be sufficient to perform the analysis,  and a laboratory
technician can easily be trained to do all of the work.   The
individual component quantitation methods would require
sophisticated detection equipment and undoubtedly a higher level of
trained personnel.  Such requirements could severely limit the
number of laboratories which could perform the analyses.

VIII. LIMITATIONS OF THE TOTAL DYE ESTIMATE METHOD

     The total dye estimate method is  affected by several factors
which cannot always be controlled, and consequently the method is
subject to limitations.  Some of these limitations are unique to the
                                                                 C-15

-------
dye estimate method, while others would affect almost any analytical
method employed.

     One of the most  important  factors, which uniquely affects the
accuracy of the total dye estimate method, is the dustiness of the
individual dyes.  Because this  information is not generally
available  (or easily measurable), the relative dustiness of
different powder dyes cannot be taken into account in the
calculations.   The assumption that the composition of the unknown
dye mixture on  the air filter will always be proportional to the
quantity of dyes handled  (or the number of dye weighings) is
therefoxe made  by default.  As  long as the inherent dustiness of
powder dyes does not differ substantially between many components of
the group of dyes being analyzed, the assumption of equal dustiness
will not lead to significant errors in the total dye estimate.
Observations by the field sampling teams do not suggest that there
were extremely  large differences in dustiness between the powder
dyes sampled in the plant survey.

     One of the factors which would affect virtually any  analytical
method is. the behavior of non-dye compounds which may be present on
the air filters.  The impact of these compounds is  difficult to
assess.  While  it is easy to measure the absorbance of a non-dye
compound alone, it is not at all easy to determine the exact effect
which that compound can have on the absorbance of a dye mixture.
This is primarily because the magnitude of the influence, if any,
will usually be proportional to the amount of the non-dye compound
which is present.  Since this is unknown in the case of the air
filters, as are the combined effects of other non-dye compounds
which may be present, it is practically impossible to determine the
extent of potential interference which the non-dye compounds
represent.  Not many non-dye compounds encountered in the plant
survey have a powerful capability to influence dye absorbance in the
dye solvent.  In addition, the  low relative dustiness of a great
many non-dye compounds encountered in the survey makes it unlikely
that many of these compounds would predominate on the air filters.
For these reasons, non-dye compounds have been judged to represent
no substantial  interference to  the analytical method for the textile
plant survey.

     Another  factor affecting most analytical methods  is the
behavior of dyes from untested  dye classes.  The total dye
estimation method was validated for dyes from the acid, basic,
direct, disperse, and reactive  classes.  The applicability of the
method to analyzing dyes  from other classes is unknown and would
require additional method development work in many cases.   Untested
dye classes constitute potential interferences due to the
possibility of  dye  interaction  or instability.  It is often
difficult to identify dye interaction occurring in solution since
these processes can be very  subtle at times.  Interaction is most
likely to  occur between dyes from different classes and  can result
 C-16

-------
in total absorbances that deviate from Beer's law.  Evidence of dye
interaction was observed at only three plant sites (Plants 2/1, 2/7,
and 6/6).   The spectrophotometric data from one of these sites
(Plant 6/6) was dropped from the study for other reasons, and the
results from the other two sites were not seriously compromised by
the degree of dye reactivity which was observed.  The effects from
dye interaction are minimized in cases where a large number of dyes
are being analyzed and the concentrations are low.

     The final factor which uniquely affects the total dye
estimation method is the number of dyes in the particular group
being analyzed.  In general, the accuracy of the dye estimation
method will improve as the number of dyes increases.   This is due
to the fact that the adverse effects from dyes with "outlier" as
values will tend to be moderated by a much larger group of dyes.
Conversely, the potential effects of outlier dyes will be increased
as the number of dyes being analyzed decreases.   In such cases, the
use of a mean or weighted average as value to make a dye  estimate
calculation will be subject to greater amounts of uncertainty.   This
trend is exactly opposite of that for conventional methods of
analysis (e.g., HPLC), where the overall accuracy tends to decrease
as the number of compounds in the mixture increases.
                                                                 C-17

-------
                             Table C-5
                   DYE PURITY AND ABSORPTIVITIES
ABSORPTIVITY (as)
ID
4/9
3/3
2/4
4/9
4/6
3/8
4/1
4/9
3/0
9/1
9/1
6/5
9/1
3/0
1/6
6/5
8/6
3/8
6/2
6/2
9/1
2/4
1/6
5/9
3/3
4/1
6/5
4/1
9/1
4/1
2/4
5/2
2/4
6/5
9/1
2/4
9/1
3/8
3/8
4/1
FINAL
PURITY
COLOR INDEX NAME (PERCENT)
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
BLACK
BLACK
BLACK
BLACK
BLACK
BLACK
BLACK'
BLACK
BLACK
BLACK
BLACK
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
107
107
172
187
52
58
58
60
M-l
M-2
M-3
102
113
113
113
113
158
158
111
205
239
>
^5
25
25
25
258
264
277
281
284
290
324
324S
335
345
40
40
45
7
80
44
40
85
68
85
55
33
70
40
47
50
50
70
66
66
81
50
76
24
47
41
33
65
33
33
50
50
60
47
43
26
80
77
58
31
54
30
60
70
40
ACTIVE
COLORANT
5.76
3.94
4.12
5.02
3.60
5.00
4.24
2.28
7.42
6.88
6.33
7.04
18.18
7.80
7.34
8.78
7.56
4.93
15.31
2.46
2.67
3.76
5.72
5.68
3.94
4.48
2.96
4.92
2.72
8.05
3.94
3.69
3.79
4.81
8.63
3.22
5.07
3.23
8.54
2.44
COMMERCIAL
DYE
2.53
2.28
3.50
3.41
3.06
2.75
1.40
1.60
2.97
3.23
3.16
3.52
5.45
5.15
4.84
7.11
3.78
3.74
3.67
1.16
1.10
1.99
3.72
1.87
1.97
2.24
1.48
2.95
1.28
3.46
1.02
2.95
2.92
2.79
2.67
2.57
1.52
1.94
5.98
0.98
C-18

-------
          Table C-5

DYE PURITY AND ABSORPTIVITIES
         (Continued)
ABSORPTIVITY (as)
ID
2/4
4/9
3/8
3/0
6/2
4/1
3/8
3/3
4/6
4/9
8/6
4/1
9/1
9/1
4/9
4/9
1/6
2/4
9/1
2/4
3/8
9/1
6/2
4/9
1/6
4/6
2/4
3/3
9/1
3/3
3/8
1/6
3/8
6/5
2/4
6/5
2/4
3/8
3/8
COLOR INDEX NAME
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
BLUE
BLUE
BLUE
BLUE
BLUE
BROWN
BROWN
BROWN
BROWN
BROWN
BROWN
BROWN
BROWN
BROWN
GREEN
GREEN
GREEN
GREEN
GREEN
GREEN
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
RED
RED
RED
RED
RED
RED '
RED
80
80
90
U-l
U-l
227
227
298
298
330
384
45
M-l
U-2
104
108
25
25
25
28
10
116
116
142
149
156
156
3
51
60
74
U-l
1
111
143
158
182
186
194
FINAL
PURITY
(PERCENT)
60
32
67
47
47
30
61
52
50
54
60
29
46
40
80
62
83
67
60
41
88
46
46
75
85
67
67
50
64
35
76
58
56
72
65
82
70
100
38
ACTIVE
COLORANT
2.
0.
1.
3.
3.
4.
3.
6.
7.
4.
3.
4.
6.
6.
4.
4.
3.
3.
3.
4.
3.
5.
4.
3.
2.
5.
7.
3.
4.
4.
3.
15.
4.
3.
1.
2.
4.
2.
3.
44
98
98
10
22
03
67
94
10
55
22
01
06
57
03
17
16
61
05
16
62
51
70
60
74
47
84
95
36
32
46
46
63
46
17
61
74
36
24
COMMERCIAL
DYE
1
0
1
1
1
1
2
3
3
2
1
1
2
2
3
2
2
2
1
1
3
2
2
2
2
5
5
1
2
1
2
8
2
2
0
2
3
2
1
.47
.79
.32
.46
.51
.21
.24
.61
.55
.46
.93
.16
.79
.63
.22
.59
.62
.42
.83
.71
.18
.20
.16
.70
.33
.25
.26
.97
.79
.51
.63
.97
.59
.49
.76
.14
.32
.36
.23
                                             C-19

-------
                             Table C-5

                   DYE PURITY AND ABSORPTIVITIES
                             (Continued)
ABSORPTIVITY
ID
4/1
6/5
4/1
2/4
1/6
4/6
2/4
3/0
4/6
2/4
8/6
4/1
3/0
4/1
4/6
6/2
3/3
1/6
2/4
5/2
4/9
3/0
5/9
9/1
4/1
2/4
4/1
3/8
4/9
6/2
4/1
8/6
3/8
3/8
4/9
1/6
3/3
2/4
6/2
COLOR INDEX NAME
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
RED
RED
RED
RED
RED
RED '
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
VIOLET
VIOLET
VIOLET
VIOLET
VIOLET
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
259
260
260
266
266
266
299
299
299
337
357
359
360
361
361
396
399
52
57
M-2
JJ-1
U-3
U-5
U-6
121
48
48
7
90
116
121
121
127
129
129
135
151
151
159
FINAL
PURITY
(PERCENT)
30
70
70
73
36
32
46
91
65
81
91
60
47
50
43
75
62
45
40
85
83
50
40
50
55
65
33
75
86
56
25
25
83
35
35
95
63
86
35
ACTIVE
COLORANT
2.
1.
2.
4.
5.
4.
7.
3.
4.
4.
3.
4.
7.
5.
3.
2.
4.
6.
7.
3.
3.
6.
4.
8.
3.
2.
1.
6.
3.
3.
5.
6.
1.
2.
2.
1.
3.
3.
4.
97
71
09
29
47
79
93
20
46
01
31
91
39
89
33
27
96
70
21
20
81
61
14
88
87
39
92
40
47
32
19
01
61
92
67
06
68
12
25
(as)
COMMERCIAL
DYE
0.
1.
1.
3.
1.
1.
3.
2.
2.
3.
3.
2.
3.
2.
1.
1.
3.
3.
2.
2.
3.
3.
1.
4.
2.
1.
0.
4.
2.
1.
1.
1.
1.
1.
0.
1.
2.
2.
1.
89
19
46
13
97
53
65
91
90
25
01
95
47
95
43
70
08
01
88
72
16
31
66
44
13
55
77
80
99
86
30
50
33
02
93
00
32
68
49
C-20

-------
          Table C-5

DYE PURITY AMD ABSORPTIVITIES
         (Continued)
ABSORPTIVITY (aj
ID
9/1
3/0
3/8
2/4
2/4
4/1
2/4
5/9
5/2
4/9
4/9
9/1
1/6
6/2
3/3
4/1
2/4
9/1
4/1
6/5
1/6
3/8
9/1
7/7
3/8
1/0
5/9
2/7
7/7
3/8
1/0
6/5
5/9
3/8
8/6
1/0
3/8
9/1
8/6
COLOR INDEX NAME
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
ACID
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
BLACK
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE'
BLUE
159
159
17
19
198
216
218
219
219
235
241
40
49
49
49
49
49
65
79
79
99
99
U-l
M-l
124
141
21
3
3
3
3
41
41
41
45
54
54
54
69
FINAL
PURITY
(PERCENT)
35
66
78
60
56
45
58
70
71
50
87
60
62
62
78
50
78
70
75
75
50
92
78
75
52
39
30
55
71
60
27
19
35
47
51
8
40
20
58
ACTIVE
COLORANT
4
5
3
2
4
2
1
5
5
2
3
2
3
4
3
4
2
5
2
2
7
2
1
3
9
5
4
1
14
17
15
35
21
15
2
19
4
12
0
.58
.53
.08
.36
.62
.45
.60
.50
.50
.43
.16
.88
.91
.05
.18
.72
.98
.95
.11
.09
.72
.70
.60
.19
.22
.06
.29
.71
.10
.24
.53
.60
.50
.31
.73
.77
.89
.70
.36
COMMERCIAL
DYE
1
3
2
1
2
1
0
3
3
1
2
1
2
2
2
2
2
4
1
1
3
2
1
2
4
1
1
0
9
10
4
6
7
7
1
1
1
2
0
.60
.65
.40
.41
.56
.10
.93
.85
.91
.21
.75
.73
.42
.51
.48
.36
.33
.17
.58
.57
.86
.48
.25
.39
.79
.97
.29
.94
.99
.34
.19
.77
.52
.20
.39
.58
.96
.54
.21
                                             C-21

-------
                             Table C-5

                   DYE PURITY AND ABSORPTIVITIES
                             (Continued)
ABSORPTIVITY
ID
3/8
9/1
1/0
7/7
3/8
1/0
9/1
5/9
3/8
2/7
2/1
1/0
3/8
3/8
8/6
1/0
2/7
3/8
7/7
7/7
3/8
1/0
9/1
2/7
7/7
2/7
3/8
8/6
1/0
5/9
3/8
3/8
7/7
3/8
1/0
2/7
3/8
2/1
1/0
COLOR INDEX NAME
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BASIC
BLUE
GREEN
GREEN
GREEN
GREEN
ORANGE
ORANGE
ORANGE
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
VIOLET
VIOLET
VIOLET
VIOLET
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
U-l
4
4
4
4
21
21
30
14
14
14
15
15
29
29
46
46
46
46
49
>*51
U-2
14
16
16
37
11
11
13
21
24
25
28
28
28
28
29
40
51
FINAL
PURITY
(PERCENT)
5
93
90
99
100
28
62
30
64
23
46
21
44
24
24
15
75
90
72
35
66
50
93
77
77
7
48
50
23
48
11
50
42
46
48
42
34
24
25
ACTIVE
COLORANT
7.
0.
0.
0.
0.
7.
8.
13.
3.
3.
3.
1.
3.
2.
4.
12.
9.
7.
8.
2.
8.
7.
13.
19.
8.
60.
5.
1.
3.
4.
4.
1.
5.
4.
4.
1.
88
01
01
00
00
39
45
78
08
37
08
29
28
83
22
84
93
24
69
78
62
09
11
57
10
48
36
35
14
33
23
81
11
60
36
68
3.96
3.70
6.07
(a,)
COMMERCIAL
DYE
0.
0.
0.
0.
0.
2.
5.
4.
1.
0.
1.
0.
1.
0.
1.
1.
7.
6.
6.
0.
5.
3.
12.
15.
6.
4.
2.
0.
0.
2.
0.
0.
2.
2.
2.
39
01
01
00
00
07
24
14
97
78
42
27
44
68
01
93
45
52
25
97
69
54
19
07
24
23
57
68
72
08
47
91
15
12
09
0.70
1.35
0.89
1.52
C-22

-------
          Table C-5

DYE PURITY AND ABSORPTIVITIES
         (Continued)
ABSORPTIVITY (as)
ID
3/8
3/8
9/1
5/4
9/1
1/0
2/7
2/7
7/9
7/9
1/0
5/4
7/9
7/9
1/0
9/1
5/9
5/4
1/0
2/7
9/1
5/4
1/0
5/4
5/4
9/1
9/1
1/0
5/9
9/1
2/7
1/0
5/9
5/4
7/9
7/9
9/1
1/0
1/0
COLOR INDEX NAME
BASIC
BASIC
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
YELLOW
YELLOW
BLACK
BLACK
BLACK
BLACK
BLACK
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BROWN
BROWN
BROWN
BROWN
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
RED
RED
RED
RED
RED
RED
RED
RED.
RED'
RED
87
91
2
62
80
80
M-l
106
160
189
191
191
218
25
251
78
80
80
M-2
U-l
113
115
115
116
34
72
80
M-2
M-3
149
224
227
243
243
72
75
80
80
89
FINAL
PURITY ACTIVE
(PERCENT ) COLORANT
63
62
50
29
36
76
58
34
45
46
45
45
52
35-40
50
40
28
28
28
38
33
17
17
93
34
59
35
40
40
60
S9
58
67
67
57
42
41
42
34
4
6
11
5
10
5
9
9
8
6
4
4
4
16
8
4
1
1
6
1
8
17
19
4
6
3
2
3
5
5
5
2
3
3
5
5
8
6
4
.64
.36
.27
.16
.50
.69
.00
.08
.96
.97
.62
.47
.44
.44
.60
.19
.16
.14
.09
.47
.83
.71
.25
.54
.12
.88
.64
.01
.98
.16
.17
.80
.25
.60
.79
.42
.90
.09
.46
COMMERCIAL
DYE
2
3
5
1
3
4
5
3
4
3
2
2
2
2
4
1
0
0
1
0
2
3
3
4
2
2
2
1
2
3
2
1
2
2
3
2
3
2
1
.93
.94
.64
.50
.78
.32
.22
.09
.03
.21
.08
.01
.31
.30
.30
.68
.32
.32
.70
.56
.91
.01
.27
.22
.08
.29
.11
.21
.39
.10
.02
.62
.18
.41
.30
.28
.65
.56
.52
                                             C-23

-------
                             Table C-5

                   DYE PURITY AND ABSORPTIVITIES
                             (Continued)
ABSORPTIVITY
ID
7/9
5/4
1/0
7/9
1/0
9/1
2/7
5/9
7/9
5/4
2/7
5/4
4/9
1/0
6/2
9/1
8/8
9/1
3/8
5/9
8/0
6/2
3/0
1/0
3/8
9/1
3/0
9/1
6/2
2/1
4/9
5/9
5/4
3/8
2/7
8/0
5/9
1/0
8/6
COLOR INDEX NAME
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DIRECT
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
RED
RED
RED
VIOLET
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
BLACK
BLACK
BLACK
BLACK
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
9
9
u-i
9
106
106
106
106
106
142
44
58
M-l
M-2
M-3
M-4
109
109
139
27
27
27
281
281
3
3
337
56
56
56
56
56
56
56
56
60
60
60
60
FINAL
PURITY
(PERCENT)
55
80
45
56
41
52
41
41
41
60
44
36
25-30
35
30
69
32
25
28
18
18
18
37
35
36
44
22
28
28
29
25
28
27
27
34
25
23
44
25
ACTIVE
COLORANT
2.
1.
4.
7.
3.
3.
4.
3.
3.
3.
4.
3.
25.
9.
6.
4.
3.
3.
6.
5.
7.
7.
10.
7.
4.
6.
13.
7.
7.
6.
7.
7.
7.
5.
9.
28
46
35
81
57
31
65
78
54
76
85
70
81
05
44
43
72
72
50
49
15
52
98
34
70
51
24
48
46
43
03
46
93
93
25
4.27
4.35
2.
,02
4.03
(as)
COMMERCIAL
DYE
1.
1.
1.
4.
1.
2.
1.
1.
1.
2.
2.
1.
1.
3.
1.
3.
1.
0.
1.
1.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
1.
2.
1.
2.
2.
26
17
96
37
46
12
91
55
45
26
13
33
81
17
93
06
19
93
82
37
29
35
74
57
35
87
91
09
09
19
90
09
98
,02
,59
1.07
1.
,00
2.02
1.01
C-24

-------
          Table C-5

DYE PURITY AND ABSORPTIVITIES
         (Continued)
ABSORPTIVITY (as)
ID
3/8
3/0
6/2
4/3
8/8
3/0
2/1
1/0
4/9
6/2
6/2
8/6
4/9
6/2
2/1
1/0
6/2
2/7
5/4
6/2
8/6
3/0
5/9
4/9
6/2
1/0
6/2
4/3
4/3
3/0
6/2
3/0
5/9
8/6
3/0
6/2
3/0
6/2
5/4
FINAL
PURITY ACTIVE
COLOR INDEX NAME (PERCENT) COLORANT
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
GREEN
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
ORANGE
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED'
RED
60
60
60
73
73
73
73
73
77
77
79
79
79
87
M-3
U-4
U-4
9
29
30
30
30
30
37
41
41
135
135
151
159
167
167:1
177
211
263
263
305
333
338
24
23
44
28
28
54
26
25
40
40
60
25
50
23
36
13
28
15
44
30
29
30
30
35
31
25
41
40
34
24
45
30
25
27
25
25
38
28
25
3
4
4
6
7
7
5
6
4
4
7
7
7
4
6
11
5
12
11
5
5
5
5
7
4
5
6
6
8
3
6
9
15
9
2
3
10
10
12
.99
.09
.59
.41
.53
.25
.50
.45
.13
.04
.54
.70
.63
.02
.94
.13
.02
.56
.98
.83
.06
.59
.77
.11
.62
.47
.14
.18
.13
.35
.71
.57
.40
.30
.74
.43
.87
.60
.90
COMMERCIAL
DYE
0
0
2
2
2
3
1
1
1
1
4
1
3
0
2
1
1
1
5
1
1
1
1
2
1
1
2
2
2
0
3
2
3
2
0
0
4
2
3
.96
.94
.02
.12
.11
.91
.93
.81
.65
.62
.53
.92
.81
.93
.50
.45
.41
.88
.27
.75
.47
.68
.73
.49
.43
.37
.52
.47
.76
.80
.02
.87
.85
.51
.68
.86
.13
.97
.23
                                             C-25

-------
                             Table C-5

                   DYE PURITY AND ABSORPTIVITIES
                             (Continued)
ID
3/8
3/0
6/2
4/9
9/1
5/9
4/9
5/4
2/1
2/7
3/8
3/8
8/0
8/0
1/0
6/2
4/9
8/6
3/8
8/8
4/3
3/0
6/2
2/1
5/4
3/0
6/2
8/8
4/3
1/6
3/0
3/8
5/4
3/8
6/2
3/8
4/3
5/9
2/7
•'k
FINAL
PURITY
COLOR INDEX NAME (PERCENT)
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
RED
RED
RED
RED
RED
RED '
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
RED
VIOLET
VIOLET
VIOLET
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
4
43
55
60
60
60
60
60
60
60
60
65
72
72
73
73
73
82
82
86
.88
91
91
U-2
26
48
57
108
114
184:1
198
218
218
23
3
3
42
42
54
17
25
23
27
25
23
23
28
20
27
28
59
20
22
22
22
44
30
30
41
24
25
25
41
34
25
33
30
45
10
21
25
25
38
55
49
29
59
53
ABSORPTIVITY (as)
ACTIVE
COLORANT
4.
17.
4.
3.
3.
3.
3.
2.
2.
3.
2.
10.
9.
11.
6.
6.
12.
10.
12.
3.
14.
3.
3.
9.
3.
11.
3.
5.
5.
18.
6.
7.
7.
13.
4.
47
83
22
91
99
42
74
05
98
97
35
58
61
86
51
70
86
83
31
03
03
62
50
47
06
81
93
23
13
96
59
17
13
81
17
4.32
2.
99
1.50
5.29
COMMERCIAL
DYE
0
4
0
1
1
0
0
1
0
1
1
6
2
2
2
2
5
3
3
1
3
0
0
3
1
2
1
1
2
1
1
1
1
5
2
2
0
0
.76
.46
.97
.05
.00
.92
.86
.11
.80
.07
.13
.24
.40
.61
.87
.95
.66
.25
.69
.24
.37
.90
.88
.88
.04
.95
.30
.57
.31
.90
.38
.79
.78
.25
.29
.11
.87
.88
2.81
C-26

-------
          Table C-5

DYE PURITY AMD ABSORPTIVITIES
         (Continued)
ABSORPTIVITY (as)
ID
6/2
1/0
6/2
8/8
6/2
3/0
4/9
3/8
4/9
4/9
4/3
8/6
2/1
8/6
8/6
2/1
9/1
4/3
8/6
4/9
4/3
8/6
9/1
2/1
4/9
8/6
4/3
6/5
4/3
2/1
4/3
2/1
9/1
2/1
4/3
4/3
6/5
8/6
2/1
COLOR INDEX NAME
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
DISPERSE
MORDANT
MORDANT
MORDANT
MORDANT
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
BLACK
BLACK
BLACK
ORANGE
BLACK
BLACK
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
BLUE
ORANGE
ORANGE
ORANGE
RED
RED
RED
RED
RED
RED •
RED
RED
54
54
64
67
86
93
11
11 -
9
3
5
U-l
10
114
116
137
18
21
21
27
27
29
52
7
U-l
U-3
U-4
U-5
16
70
82
120
120
152
180
198
40
40
43
FINAL
PURITY
(PERCENT)
29
40
19
24
46
45
58
58
30
75
60
35
91
30
30
91
50
60
25
50
50
53
60
99
45
60
50
84
62
71
40
96
96
94
50
40
70
70
72
ACTIVE
COLORANT
2
2
2
3
1
6
12
12
14
4
7
8
3
6
11
1
15
4
10
2
2
2
4
2
2
6
3
2
3
2
5
2
2
3
4
4
3
2
4
.99
.22
.84
.53
.72
.46
.74
.82
.18
.31
.06
.35
.06
.11
.64
.38
.45
.46
.67
.81
.86
.20
.69
.85
.77
.90
.79
.60
.55
.46
.72
.72
.36
.10
.18
.87
.23
.66
.04
COMMERCIAL
DYE
0
0
0
0
0
2
7
7
4
3
4
2
2
1
3
1
3
2
2
1
1
1
2
2
1
4
1
2
2
1
2
2
2
2
2
1
2
1
2
.87
.89
.54
.85
.79
.91
.39
.43
.25
.24
.24
.92
.78
.83
.49
.26
.86
.68
.67
.40
.43
.16
.81
.83
.25
.14
.90
.18
.20
.75
.29
.61
.27
.91
.09
.95
.26
.86
.91
                                             C-27

-------
                              Table C-5

                    DYE  PURITY AND ABSORPTIVITIES
                             (Continued)
K ABSORPTIVITY (as)
ID
9/1
4/3
4/9
4/9
4/9
8/6
8/6
4/3
8/6
4/3
4/3
8/6
6/5
8/6
2/1
9/1
9/1
4/9
4/9
8/8
8/8
8/8
8/8
8/8
8/8
COLOR INDEX NAME
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
REACTIVE
VAT
VAT
VAT
VAT
VAT
VAT
RED
RED
RED
RED
RED
RED
VIOLET
VIOLET
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
YELLOW
BLUE
BROWN
ORANGE
VIOLET
VIOLET
YELLOW
43
94
U-l
U-2
U-3
U-4
33
5
125
15
160
25
27
27
3
58
64
U-l
U-2
,6
M-l
2
1
13
2
FINAL
PURITY
(PERCENT)
72
75
35
35
45
35
35
55
50
50
65
70
58
58
80
55
60
45
50
14
25
14
16
31
13
ACTIVE
COLORANT
3
3
3
5
4
4
3
2
5
3
1
2
2
2
2
2
2
2
5
0
0
1
3
1
0
.90
.96
.59
.11
.39
.45
.50
.51
.15
.96
.70
.74
.93
.88
.13
.24
.41
.59
.15
.39
.42
.76
.10
.07
.38
COMMERCIAL
DYE
2
2
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
0
0
0
0
0
0
.81
.97
.26
.79
.98
.56
.22
.38
.58
.98
.10
.92
.70
.67
.71
.23
.45
.16
.58
.05
.11
.25
.50
.33
.05
Note:  10 grams =  0.3527 ounce; 1,000 grams =  2.2046 pounds.
C-28

-------
      Appendix D
STATISTICAL METHODOLOGY

-------
I.   NORMAL PROBABILITY PLOTS

     A normal probability plot of  the  commercial  dye concentrations
observed in the 22 plants is shown in Figure D-l.  As discussed in
Section VII, the plot indicates significant departures from
normality.  A normal probability plot of the  (natural) logarithms  of
the commercial dye concentrations is shown in Figure D-2.  The
approximately straight-line shape of this plot indicates that a
lognormal distribution will provide a good fit to the observed
airborne concentrations.

II.  COMPUTATION OF PERCENTILE ESTIMATES FOR  THE  STATISTICAL
     ANALYSIS OF DATA

     Procedures for estimating the mean, variance, and percentiles
of the lognormal distributions were presented in Chapter 7.  In this
appendix, we discuss derivation of confidence intervals for the
estimated percentiles of the lognormal distribution.

     First consider the percentiles exp(m + zs).  Our approach  was
to compute a 95% confidence interval for m + zs on the log scale and
then exponentiate back to the measurement scale.  In estimating the
sampling variance of the estimate m + zs we treated m and s as  if
they were computed from a simple random sample from a normal
distribution.   This leads to the approximation

                   Var(m + zs)  = Var(m) + z2Var(s)

            = s2(n-1 + z2(l - 2G(n/2)2/(n - l)G((n -l)/2)2)

where G is the gamma function1  and  n (=  22 here)  is the sample size.

     Now, an approximate 95% confidence interval  for m + zs  can be
computed as
                    m + zs + 1.96  (Var(m + zs))
                                               0.5
This interval can be exponentiated to give a confidence interval for
the percentile exp(m + zs) of the measurement data.  In a similar
fashion, we find the formula

                 Var(m = s2/2) = s2/n +  s'/2 (n - 1),

treating s2 as having a Chi-squared distribution with (n - 1)
degrees of freedom.2  A confidence interval for m + s2/2 and hence
     xSee Theorems 1.3.1 and 1.3.3 in Bickel PJ, Doksum, KA.   1977
Mathematical Statistics.  San Francisco.

     2Ibid.,  Theorem 1.3.3.
                                                                  D-l

-------
                                          Figure  D-l

                         AIRBORNE COMMERCIAL DYE CONCENTRATION:
                                  NORMAL PROBABILITY PLOT
   1.5
N
V
E
P
S
E

N
0
P
v
*
L
   1.0
0.5
0.0
  -0.5
  -1.0
                                        T«mt of Normality:
                                        A«*umDt1on o* Normality R»J«ct«cf
                                        («1th o l«55 then .01)
  -1.5
  -2.0
      0.0     0.1     0.2     0.3     0.4     0.5     0.6     0.7     0.8

                                    AIRBORNE COMMERCIAL DVE CONCENTRATION
                                                                     0.9
                                                                            1.0
                                                                                          1.2
       D-2

-------
                                       Figure D-2

                    LOG(AIRBORNE COMMERCIAL  DYE CONCENTRATION):
                               NORMAL PROBABILITY PLOT
  1.5
  t.O
  0.5 *
»  0.0
I

D
I
S
1 -0.5
p
I

u
  1.5
  J.O
T«st O* Normal < ty :
Not s
-------
for v = exp(m + s2/2),  the  mean of the data on the measurement
scale, can then be constructed as before.

III.  SENSITIVITY ANALYSIS  OF DYE  CONCENTRATION ESTIMATES DUE TO
      MEASUREMENT ERROR

      A.  Introduction

      This section is a sensitivity analysis of the effect of
measurement error for  airborne dye concentration.  In this analysis
ANOVA techniques were  applied to  estimate the  component of total
sample variance due to measurement errors and  to the underlying
population variance.   The  estimated population variance is then
applied to generate percentiles of the population distribution which
are not biased upward  by the additional variance due to measurement
errors.

      While many ANOVA  techniques  have been proposed, the standard
one-way, random effects model will be  considered here.3  The one-way
model is applied to separate the  total sample  variance into two
components:  one for the variation of  exposure levels within-plants;
and another for variations across (or  between) plants.  In the
former case, it is appropriate to consider the deviations of each
observation from the mean  of all  observations  for that plant.  In
the latter case, we consider the  squared deviation of the plant
means from the overall mean for all plants.

      Measurement errors are  shown to  affect both  the within-plant
and across-plant components of variance in the survey-  Within-plant
variance is affected by the traditional field/laboratory airborne
chemical measurement sources of variation  such as variations in flow
rates of collection devices, filter efficiency, efficiency of sample
recovery and extraction, and errors inherent  in laboratory
measurements using spectrometric  devices.  These errors add to the
spatial variation of dye concentration within  the weighing area and
the variation due to left- and right-handed workers.  The total of
such  within-plant variations is collectively  estimated in the
within-plant component of  variance.

      In  the  survey,  there  is also an  effect of measurement  error in
the estimated across-plant variance.   This second type of
measurement error relates  to the  problem of measuring the total
concentration of multiple  dyes with a single  spectrophotometric
measurement.  To accomplish this, it  is necessary to know the
relative amounts of each dye in the collected sample.  In the
survey,  these relative amounts are based on physical measurement of
the amount of each dye weighed.   Due  to variations in  "dustiness" of
     3See,  for example, Snedecor and Cochran,  Statistical Methods,
 6th Ed., Chapter 10.
 D-4

-------
each dye weighed, the true airborne relative concentrations may
differ from the weighed concentrations.

     This type of measurement error affects all measurements  within
a plant in a similar fashion, and is not a component of the within-
plant variation.   This type of measurement error is contained in the
estimated across-plant component of variance.   Hence, the total
variance across plants is larger than the true population variance
across plants.

     The purpose of ANOVA is to identify the within- and across-
plants components of variance.   Additional simulations4  were
necessary to estimate the amount of measurement error affecting the
across-plants component of variance.  A final estimate of the
underlying population variance is obtained by subtracting the
variance due to the across-plant type of measurement error from the
total across-plant variance.

     In the following section, the ANOVA procedures for estimating
total,  within, and across-plant variances are summarized.   In the
final section of this report, the across-plant component of variance
is decomposed into one component which represents across-plant
measurement error, and the remaining component is identified as the
underlying population variance.   Population statistics are then
presented based on the estimated population mean and variance.

     B. ANOVA Results

     The final survey data  set contains measurements of airborne
concentrations tabulated both on an active dye basis and on a
commercial dye basis.  The analysis of variance was conducted
separately for each set of measurements.  While the general
discussion is phrased in terms of active dye concentrations,
analogous results hold for the commercial dye concentration
measurements.   The active dye measurements contain one set of two
observations (left and right) in each of 22 plants, yielding a total
of 44 data points.  Sampling weights are available for estimating
population parameters for all plants and separately for all workers.

     Frequency plots of the unweighted observations on an active dye
basis are shown in Figures D-3(a)  and D-3(b).   Figure D-3(a)  depicts
the original data, while the frequency plot of the natural logarithm
of the observations is shown in Figure D-3(b).  In both figures,
left and right measurements are shown separately.  Figures D-4(a)
and D-4(b)  show equivalent plots on a commercial dye basis.
Examination of the figures demonstrates that the distribution of
the logarithm of the observations is approximately a normal
     4These simulations,  performed by Midwest Research Institute,  are
discussed in  Appendix C of this report.
                                                                  D-5

-------
                            Figure D-3(a)
    ACTIVE DYE CONCENTRATIONS MEASURED ON LEFT AND RIGHT FILTERS;
                        FREQUENCY DISTRIBUTION
                            ACTIVE  DYE
  12
  11 -

  10-

   t-

   • -

   7-
   4-

   3-

   t-

   1 -

   0
          <.03
     IITIIIIIIIi
<.23 ,<.3 <.33 <.4 <.45  <.3 <.33 <.f <.€5 <.7 <.79 >.7S
                            LEFT
                  WCHT
D-6

-------
                          Figure D-3(b)
Log (ACTIVE DYE CONCENTRATION)  MEASURED ON LEFT AND RIGHT FILTERS:
                     FREQUENCY DISTRIBUTION
                     LOG (ACTIVE  DYE)





-






4.
™





~

2m
m



f_
•

o J


























1





















^^
/
/
A
/
/
/
/

























_
I
p

























/
/
^




















H



















^
k


>

^
'
r
/
/
A
/
/
kl
,/





















*•
HPI
H [<£
.^IK!
r ;! /
/ ' /
« r/ra r/l
i ^ ^^
r ^
y >
i ^ , ^









































/

f
/

/
\/b
/
/

/
/
/
/
/
/
/
n


















i
1

1

1
i
;





^
*







',

r
/

\\
S
Y
/

/
/
/
/
/
/
/













\



M

1
1 ^^^^»
l ^^ 1 1
' / ''
! /iji
9 KKI
/ ,',
. / i|i r i
• / i'i /
' ^ ' ' ' A ' '
! / !•! / •!
-9 <-4.S
  <-3,5 <-3 <-2.3 <-2 <-
                                             -1  <-O.3 0.3
.___  LDCtCOMCOfflWnON)
P7I UEFT       pffW
                                       MGHT
                                                               D-7

-------
                          Figure D-4(a)
 COMMERCIAL DYE CONCENTRATIONS MEASURED ON LEFT AND RIGHT FILTERS:
                      FREQUENCY DISTRIBUTION
  10
                      COMMERCIAL DYE
   7-
   3

   4

   3

   1

   1

   0
<05
<.35
<.45
                   <.33
                                                          <.7S >.73
LEFT
                            CONCENTRM
                                        MCHT
D-8

-------
                Figure D-4(b)
 Log(COMMERCIAL DYE CONCENTRATION) MEASURED ON
LEFT AND RIGHT FILTERS:  FREQUENCY DISTRIBUTION
         LOG(COMMERCIAL DYE)
9-






4-








3 -






f.




1 -






































1 f i
<-5 <-4.9





























y
x
/
s
<—




























B


'

» 4
















r^n

'y
x
~

/
y
X
/
r
J
"

/
/
*
/
s
/
C-3
















••
i


..
!'

_-. r-
X /
/ /
y y
i X X
! '81 /
i > i1 y
/ i'i X
;' ' ' i r .
i y i , y
r ^ \ v
f 1,1 r
' X '' /
r I r
,3 <-3 <-2







•







1
(

,
I


1


1






1

.9
M



1


T- i
/ '






X i
y |
X i
y ,
x
r '
/!

^^ i

/:
y i
X i
/'
y ,
X i
/'
" ]
^r
<_ '
^*^
n
•





i













,

!
1
'


1




I 4




V


/
/






/

^r
>


f

/
/
/
A
X


>
x









•
>






}
<
<
1


«,
1

I X X !'
•' X X* '*'
", y y iS
i X X i'i
•' X n X < in BO
', > ! y « • y ',
i / i'i / I'M X i

i dL i « £. 1 1 1 ^ '
V
.9 <-1 <-O.9 <0.0 <0.9 >O.9
                tDiKCOMCtMIHATIOM)
                LIFT       ESS!
                                                   D-9

-------
distribution.  The figures also demonstrate that  there  is  no
apparent difference between the left  and  right measurements in terms
of mean or variance.

     Based on  the  above observations, the following random effects
log-linear model is adopted:

     1) Natural  numbers - multiplicative  model:

            Yitl = M A! Elfj (1  = 1,...,  NI;  j =  left,  right)

where the Y1(j denote the original airborne concentration
measurements in plant i,  on  side  j, using the active dye basis.   The
fixed parameter M  will  be discussed below.  The terms AA denote
random  (lognormal) multiplicative effects for variation across
plants, and the Elrj denote random (lognormal)  variations within
plants.  Taking the logarithms of equation (1) yields:

     2) Logarithms -  additive model:

                        YI,J = m + ai  +  elfj

where                   yltj = log  (Y1>:)) ,
                         m = log  (M),
                         aL = log  (AJ, and
                        ei.3 = log  (Eifj) .

     In Equation (2),  the ai are  normally distributed random  effects
across plants  with mean zero  and  variance sa2. The elrj  are normally
distributed error  terms within plants with mean zero and variance
s2.   The fixed parameter m represents the overall mean  of the
logarithms of  the  data.   By  the theory of the lognormal
distribution,  the  fi^ed parameter M  (=em)  in  Equation (1)  above
represents the overall  median of  the  observations expressed in
original numerical form.   The parameter M is  also an estimate of the
median of the  underlying population distribution,  because  the a±  and
elfj have mean  zero.  The  symbol sa2 denotes the across-plant variance
component, while the  symbol  s2 denotes the within-plant component of
variance.  As  noted in  the introduction above, sa2 includes both the
variance due to the population distribution and due to  the across-
plant component of measurement error.  The purpose of the  ANOVA
presented in this  section is  to estimate  s2 and sa2.   In the
following section, our  simulation-based estimate  of the across-
plant measurement  error is removed  from sa2 to yield an  unbiased
estimate of the population variance.

     Note  that all estimates produced by the ANOVA procedure are
statistics calculated from the  logarithms of  the  observations.  Use
of  logarithms  implies that sa2 estimates the across-plant  variance in
the logarithms and s2 estimates the within-plant  variance  in  the
logarithms.  Use  of  logarithms  in ANOVA is equivalent to  analyzing
 D-10

-------
the percent variations in the original observations.  Thus,  the
variance of the logarithms translates to the mean square percent
variation in the original numbers, and the mean of the  logarithm
translates to the median of the original numbers.  The  mean  of the
original numbers is a function of both the median and the variance
estimates, due to the skewness of the lognormal distribution.  All
results presented in this section are in terms of the logarithms.
Similarly, the across-plant component of measurement error estimated
in the following section is presented in terms of logarithms.  The
final results for population characteristics will be presented in
original numerical form, however.

     The  log-linear random effects ANOVA model in equation  (2) above
was estimated using both weighted and unweighted logarithms.  The
use of sampling weights detracts from the simplicity of the  required
calculations for unweighted ANOVA.  First, we discuss the unweighted
analysis procedure, and then present these results.   This section
concludes with an analysis of the weighted results.

Calculation Procedures for Unweighted ANOVA

     A.   Within plants:

        1.  Plant  i Mean

                       Yif. = (Yi. right + Yl,left)/2

        2.  Squared Deviations

                    di.j2 =  (yi(j - yir.)2

        3.  Mean Squared Error Estimate

                    MSEH =  SSW  d.   ,

              where SSW is the simple  average  of the  squared
              deviations within-plants

                           1
                    SSW	£ £ di>:12    ,
                          2X1,   1  '

              and  d,, is the correction factor for the available
              within-plant degrees  of  freedom,

                      d. =  n^/n^n^ - 1)  = 2

        4.  Within-plant Component  of Variance

                      s2 =  MSEH
                                                                 D-ll

-------
     B.  Across plants:

        1.   Overall Mean
         \

                            1
                     y... = —
                            Hi

        2.   Squared Deviations

                    di..2 =  (ylf. - y...)2    .
                           j

        3.   Mean Squared  Error Across  Plants

                    MSEa = SSA da

               where  SSA is the  simple average of the squared
               deviations across plants

                           1
                    SSA =	  £ dlf.   ,
                           n,   i

               and da is the correction factor for the available
               across-plants degrees of freedom

                   I   da = niiij/dii  - 1)  = 44/21 = 2.095

        4.  Across-plant  Component of Variance
                     _>
                     ' sa2 =  (MSEa - s2) /2

     Inspection of the  unweighted ANOVA results in Tables D-l(a) and
D-l(b)  confirms the following conclusions:

     a. The within-plant  variation is small compared to across-
        plant  variation (3.4  percent on an active dye basis and 3.7
        percent on a commercial dye basis).

     b. The variance within plants is approximately equal, in both a
        commercial or active dye basis.

     c. The variance component  across plants is 1.306 on an active
        dye basis  and 1.202 on  a commercial dye basis.  This
        component  includes both the population variance and the
        across-plant measurement error.
D-12

-------
                            Table D-l(a)

       UNWEIGHTED LOGARITHMIC ANOVA RESULTS:
             ACTIVE DYE BASIS

Source of Variation
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component
Percent of Total Variance
Across
Plants
21
55.81
2.658
1.306
96.6
Within
Plants
22
1.019
0.0463
0.0463
3.4

Total
43
56.83
—
1.352
100.0
Overall Mean = -3.0894
                            Table D-l(b)

     UNWEIGHTED LOGARITHMIC ANOVA RESULTS:
           COMMERCIAL DYE BASIS
Source of Variation
Across
Plants
Within
Plants
Total
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component
Percent of Total Variance
21
51.47
2.451
1.202
96.3
22
1.020
0.0464
0.0464
3.7
43
52.49

1.249
100.0
Overall Mean = -2.2871
                                                                 D-13

-------
Calculation Procedures for Weighted ANOVA

     The  four  steps  described above  for obtaining the across-plant
and within-plant variance components proceed  in an analogous  fashion
for weighted ANOVA, with the  simple averages  replaced by weighted
averages, and  squared deviations  from the mean replaced with
weighted  squared deviations from  the weighted mean.   Because  the
weights w± on both observations within a plant are identical,  the
weighted  mean  within plants equals the  unweighted mean within
plants .

     The  following steps  summarize the  calculations performed for
the weighted analysis of variance:

     A.   Within plants:

         1.   Plant i Weighted Mean

                     y*i.. = YI..

         2.   Squared Deviations

                  (d*lfj)2 = 
-------
             Squared  Deviations

                  (d*t,.)2 = (y*if. - y*.,.)2    .

             Weighted Mean Squared Error,  Across  Plants

                    MSEa = SSA*   da    ,

               where  SSA* is the weighted average of the squared
               deviations across plants
                    SSA* =   wt(d*lf.)V £ Wl   ,
                           i            i

               and da is the same correction factor for the available
               across-plants degrees of freedom as in the unweighted
               analysis of variance.

         4 .  Across-plant  Component  of Variance

                     sa2 = (MSE*a -  s2)/2

     Results for  the weighted ANOVA based on  establishment weights
on an active and  commercial dye basis  are presented in Tables D-2 (a)
and Table D-2 (b) .  Similar results  using weigher level weights are
presented in Tables D-3 (a) and D-3 (b) .

III. PRESENTATION OF FINAL  RESULTS

     The across-plant  variance estimated in Section  II. B above
contains both the population and the  across-plant measurement
variances .

     Simulations  by MRI  of  the errors induced by  possible variations
of the airborne dye mixture from the  weighed  dye proportions were
used to provide an estimate of the  variance due to measurement
error.   These calculations are shown  in Table D-4 .

     In  this table, an estimate of  the across-plant  measurement
variance is computed from the .05 and .95 percentiles of the
simulation results .  An estimate of the  across-plant measurement
variance was obtained  based on the  lognormal  model using the formula

                     sn,i = log [Pi (.95) /Pi (.05)] /2(1.64)

for each simulation in plant i.  The  resulting estimates of the
standard deviations were then averaged to produce an average
standard deviation
                                                                 D-15

-------
         ,                   Table D-2(a)

         ESTABLISHMENT-WEIGHTED LOGARITHMIC ANOVA RESULTS:
                          ACTIVE DYE BASIS

Source of Variation
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component .
Percent of Total Variance
Across
Plants
21
48.111
2.291
1.121
96.8
Within
Plants
22
1.071
0.0487
0.0478
3.2

Total
43
49.18
—
1.169
100.0
                            Table D-2(b)

          ESTABLISHMENT-WEIGHTED LOGARITHMIC ANOVA RESULTS:
                        COMMERCIAL DYE BASIS
Source of Variation
Across
Plants
Within
Plants
Total
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component
Percent of Total Variance
21
46.45
2.212
1.082
95.8
22
1.054
0.0479
0.0479
4.2
43
47.50

1.130
100.0
D-16

-------
                            Table D-3(a)

             WORKER-WEIGHTED LOGARITHMIC ANOVA RESULTS:
                          ACTIVE DYE BASIS

Source of Variation
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component
Percent of Total Variance
Across
Plants
21
55.67
2.651
1.300
96.1
Within
Plants
22
1.148
0.0522
0.0522
3.9

Total
43
56.82
—
1.352
100.0
                            Table D-3(b)

             WORKER-WEIGHTED LOGARITHMIC ANOVA RESULTS:
                        COMMERCIAL DYE BASIS
Source of Variation
Across
Plants
Within
Plants
Total
Degrees of Freedom
Sums of Squares
Mean Square Error
Variance Component
Percent of Total Variance
21
50.02
2.382
1.165
95.7
22
1.157
0.0526
0.0526
4.3
43
51.18

1.217
100.0
                                                                 D-17

-------
                             Table D-4

          ESTIMATION OF ACROSS-PLANT MEASUREMENT VARIANCE
Plant
A. Commercial
10
16
21
24
27
30
33
38
41
43
46
49
52
54
59
62
65
79
80
86
88
91
P(.05)
Dye Basis
2.13
2.51
1.67
2.16
1.39
•1.55
1.98
3.25
1.58
2.34
2.66
2.11
2.72
1.22
1.71
1.45
,»
2.6
1.18
1.42
1.99
P(.95)

3.36
4
2.63
3.69
5.31
2.86
3.43
5.81
2.46
3.29
4.68
4.59
3.59
2.26
3.71
2.2
3.79
2.61
2.76
3.52
Sm,i

0.139
0.142
0.138
0.163
0.409
0.187
0.168
0.177
0.135
0.104
0.172
0.237
0.085
0.188
0.236
0.127
0.115
0.242
0.203
0.174
                                                   3.540 =  Sum
                                                      20 =  n
                                                   0.177 =  sm
D-18

-------
                              Table D-4

           ESTIMATION OF ACROSS-PLANT MEASUREMENT VARIANCE
                             (Continued)
Plant              P(.05)          P(.95)
B.  Active Dye Basis

10                 6.51            9.32            0.241
16                 5.63            7.02            0.159
21                 3.98            5.41            0.219

24                 4.54            5.82            0.179
27                 5.95            9.42            0.292
30                 6.9             8.94            0.182

33                 4.85            5.94            0.125
38                 8.88            11.2            0.151
41                 4.11            5.12            0.128

43                 5.43            7.33            0.171
46                 5.7             6.56            0.114
49                 6.79            8.93            0.191

52                 3.93            4.9             0.130
54                 4.93            7.53            0.313
59                 7.34           11.01            0.228

62                 5.73            6.54            0.092
65
79                 6.41            7.91            0.137

80                 8.47           11.86            0.223
86                 5.75            8.06            0.250
88

91                 6.45            8.27            0.170
                                                   3.694 = Sum
                                                      20 = n
                                                   0.185 = sn
                                                                 D-19

-------
The measurement error variance was then obtained as sm2.  As shown in
Table D-4, these variances are quite small, roughly the same size as
the within-plant variance component obtained in Section II.B above.

     To generate the  final population  statistics,  the  quantity  sra2
was subtracted from the across-plant variance component from the
ANOVA:

                            s  2  =  s2 -s2
                            °pop    °a   °ra    r

where all terms are variances of logarithms.

     The  theory of the  lognormal distribution was  then applied  to
estimate population parameters from the overall mean of the
logarithms and the population variance spop2.  Results of these
calculations are shown  in Tables D-5(a) and D-5 (b), for active  and
commercial dye, respectively.  The difference between  adjustment for
across-plant measurement error and no adjustment is high-lighted by
including the estimates obtained directly  from the across-plant
variance, sa2.  The effects of adjustment for the across-plant
measurement variance  is quite small.

     After  reviewing  the  compact of measurement  error  estimates it
was decided not to adjust estimates of airborne dye concentration
because of measurement  error.  This was done for two reasons.

     First, the impact  of measurement  error was  quite  small  in  this
      case;

      In addition, the adjustment technique described above is very
      complicated.   It was decided  that the extra complexity
      introduced by  the  method was  not  worth the  increased difficulty
      in explaining  the  results  considering that  the estimates in
      general were changed to  two significant digits.
 D-20

-------
                             Table D-5(a)

      CHARACTERISTICS OF THE DISTRIBUTION OF ACTIVE  INGREDIENTS
Exposure  (mg/m3)
  Corrected for
Measurement Error
      (spop2)
  Uncorrected
for Measurement
  Error (sa2)
A. Plant Level
- Median
- Average
- Standard Deviation
- Percentiles
.85th
.90th
,95th
B. Weigher Level
- Median
- Average
- Standard Deviation
- Percentiles
.85th
.90th
.95th

.0484
.0833
.1168

.1425
.1840
.2688

.0413
.0777
.1239

.1324
.1744
.2624

.0484
.0847
.1218

.1449
.1879
.2760

.0413
.0790
.1290

.1345
.1778
.2690
                                                                  D-21

-------
                             Table  D-5(b)

        CHARACTERISTICS OF THE DISTRIBUTION OF COMMERCIAL DYE
Exposure  (mg/m3)
  Corrected for
Measurement Error
      (spop2)
  Uncorrected
for Measurement
  Error (sa2)
A.  Plant Level

      - Median              '   .1045
      - Average                .1767
      - Standard Deviation    .2410
      - Percentiles
           .85th    •          .3024
           .90th              .3888
           .95th              .5642

B.  Weigher Level

      - Median                 .0970
      - Average                .1710
      - Standard Deviation    .2481
      - Percentiles
           .85th              .2924
           ,90th              .3796
           .95th              .5589
                              ,1045
                               1795
                              ,2508

                              ,3072
                              ,3964
                              ,5784
                               0970
                              ,1737
                              ,2579

                              ,2969
                              ,3868
                              ,5725
 D-22

-------