Regulatory Toxicology and Pharmacology 32, 73–77 (2000)

doi:10.1006/rtph.2000.1403, available online at http://www.idealibrary.com on

Regulatory Action Criteria for Filth and Other Extraneous Materials

IV. Visual Detection of Hair in Food

Patricia Valdes Biles 1 and George C. Ziobro

Microanalytical Branch, U.S. Food and Drug Administration, HFS-315, 200 C Street S.W., Washington, DC 20204

Received February 19, 2000

the product (Eisenberg, 1981). Federal regulations, pol-

Under Section 402(a)(3) of the Food, Drug and Cos-
metic Act the distribution of foods which may contain
repulsive or offensive matter, considered as filth, is
prohibited. Filth includes contaminants such as rat,
mouse, and other animal hairs and excreta; whole in-
sects, insect parts, and excreta; and other extraneous
materials which, because of their repulsiveness,
would not knowingly be eaten or used. The presence of
such filth renders food adulterated. Some foods, even
when produced under good manufacturing practice,
contain low levels of natural or unavoidable defects
such as dirt, insect parts, and hair that are not haz-
ardous to health. Although the health hazard is elim-
inated, most consumers find the presence of “any” vis-
ible filth contaminant such as hair in a food product
objectionable. The objective of this study was to deter-
mine when a defect (hair) in a food (mushrooms) is
visible to a consumer by testing the visual acuity of a
consumer panel under controlled laboratory condi-
tions. Of the panelists participating in the study, when
presented with a single hair on a flat surface with a
solid-colored background, 27% were able to detect a
1-mm hair, 58% were able to detect a 3-mm hair, and
75% were able to detect a 10-mm hair. Twenty-five
percent of the panelists surveyed were able to detect a
5- or 10-mm hair on sliced, canned mushrooms. The
experimental design and results of this study are dis-
cussed.
INTRODUCTION
The Food, Drug, and Cosmetic Act (FD&C Act, 1985)
states, “A food shall be deemed to be adulterated . . . ) (3)
if it consists in whole or in part of any filthy, putrid or
decomposed substance . . . ” FD&C Act as amended Sec
420(a)(3). Filth is defined as any objectionable foreign
matter of animal origin found in a food, drug, or medical
device. This matter may be excreta; maggots, worms,
insects, or insect parts; or any other material that defiles
1
To whom reprint requests should be addressed.
73
icy, and good manufacturing practices (GMP) require the
prevention of gross contamination of food with large,
visibly objectionable adulterants (FDA, 1998).
Hair and hair fragments in a food product are con-
sidered an aesthetic adulterant. The presence of aes-
thetic adulterants may reduce the quality level of a
food product. Aesthetic types of filth and extraneous
materials are categorized as adulterants within the
meaning of section 402(a)(3) of the FD&C Act. Regula-
tory action can involve detention or seizure of the adul-
terated goods (FDA, 1996). Possible sources of hair in
food are from animal grooming, clothing, or domestic
animals (Vasquez, 1981). Many mammals groom their
coats by licking and often swallowing much of the
dislodged hair. These hairs, though modified by the
digestive system and excreted, are often readily recog-
nizable and are sometimes the only evidence that the
food has been contaminated with fecal matter. Cloth-
ing made of wool or rabbit fur sheds hair and hair
fragments. Rabbit fur is used extensively for fur-lined
gloves and hats for workers in the frozen food industry.
In some parts of the world it is still routine to use dogs
and cats in food processing facilities to control rodent
populations. The natural shedding by these animals
can lead to hair contamination in a product.
The presence of any visible filth contaminant such as
hair in a food product is objectionable to most consum-
ers. In this investigation the visual acuity of the con-
sumer to hair on a food product (sliced, canned mush-
rooms) is analyzed. The experimental design and
results of this study are presented. The results of this
investigation provide reference information for the
presence of hair in food. This is the fourth report in a
series in the development of a transparent science base
for a revised FDA regulatory policy in the area of filth
and extraneous materials in food.
METHODOLOGY
A semitrained laboratory panel was used to evaluate
the visual acuity of a consumer to the filth analyte hair.

0273-2300/00
74 BILES AND ZIOBRO
FIG. 1. Mushroom panel survey questionnaire.
Laboratory panels are used to study the human per-
ception of food attributes such as visual appearance.
The visual appearance of an object is related to the
illumination by which it is viewed, conditions under
which it is viewed, and the visual response character-
istics of the observer (Hoffman, 1958).
The test design was similar to a dilution or a thresh-
old test used in sensory evaluation. The dilution tech-
nique establishes the smallest amount of unknown
that can be detected when mixed with a standard ma-
terial (Amerine et al., 1965). In this study the analyte
was hair. The standard materials were test panel cards
divided into nine colored squares used in training and
Phase I and sliced, canned mushrooms used in Phase
II.
Panelists
The volunteer panelists were recruited from scien-
tific, clerical, and support staff at the Food and Drug
Administration’s Center for Food Safety and Applied
Nutrition (CFSAN). Each panelist filled out a confiden-
tial demographic survey on known visual problems and
frequency of eating mushrooms (Fig. 1). A special test-
ing area with a separate waiting room was used to
minimize distractions and to control test conditions. No
telephones were present in the waiting room or test
area. The panelists were allowed to choose an appoint-
ment for the tests that would best fit their schedules.
Panelists reported to a separate waiting room where
they were given white laboratory coats to wear and
were escorted to the viewing room. (Canada Depart-
ment of Agriculture, 1977; Jellinek, 1985).
Viewing Room
The test area consisted of a 33 ⫻ 70-in. stainless-
steel examining table covered with white 70-lb, 42-in.-
wide 100% cotton rag paper onto which was mounted
two 31-in.-wide ⫻ 36-in.-deep ⫻ 24-in.-high white
 VISUAL DETECTION OF HAIR IN FOOD
TABLE 1
Panel Square Color Data
Mushroom
Button mushroom, Agaricus bisporis (Lange) Imbach
Wrinkled thimble-cap or early morel, Verpa bohemica Krombh.
Meadow mushroom, Agaricus campestris L. ex Fr.
Cinnabar-red chanterelle, Cantharellus cinnabarinus Schw.
Chanterelle, Cantharellus cibarius Fr.
Rare red reishi, Ganoderma lucidum (Leysser) Karsten
a
Values range is 0 –255 with 0 being the darkest shade of the color and 255 the lightest shade of the color. The values for pure black are
0, 0, 0, and pure white 255,255,255.
foamboard observation booths. A door cutout in the
back of each booth was used to present the samples.
Lighting consisted of an Examoloite Model TC-440
overhead luminaire equipped with Deluxe Examolite
4-ft two-prong F40T12/EXX 40 W fluorescent tubes
and Examolite 33WA19/EX 4.33 W long-life bulbs. The
lighting was measured with a Sper Scientific Ltd. light
meter Model FC 840021 and determined to be 150-ft
candles. The viewing distance ranged from 524 to 554
mm.
Test Panel Card Preparation
A series of cards with nine 1.5 ⫻ 1.5-in. colored
squares arranged in a 3 ⫻ 3 grid were prepared by
using Corel WordPerfect 7 Presentations and printed
on a HP Color LaserJet 5/5M printer onto HP Color
LaserJet 5/5M premium glossy paper and mounted
onto gray poster board. Nine colors were used in the
training panel of which six were used in Phase I. The
background colors chosen for the training and Phase I
portion of the survey were picked by comparing color
charts to commercially prepared, canned mushrooms.
The names of the commercial mushrooms; the red,
green, and blue color values; and common name of the
colors used in Phase I are listed in Table 1. Onto each
square was mounted a single hair of a known length (0,
1, 3, 5, 7, 10 mm) and type [human black, human
brown, human white, cat/dog (cat or dog), rat/mouse
(rat or mouse)] using clear fingernail polish. Hair
placement upper, lower, right and left, and center was
randomly chosen by rolling a 20-sided percentile die.
Multiples of numbers 1–5 were used for a defined po-
sition on the square. The test panel cards were labeled
with a unique, random three-digit number. The cat/dog
hair was obtained from personal pets; human hair
white/brown/black, from staff; and rat/mouse hair,
from authentic animal pelts.
Training
Visual perception and recognition may be considered
as behavioral processes (Ryback et al., 1998). Ten test
75
Values a
Red Green Blue Color
255 247 239 Seashell
81 43 20 Dark chocolate brown
224 176 144 Beige
236 117 85 Red-orange
255 255 132 Light yellow
255 145 35 Pumpkin
panel cards were used as a training exercise prior to
Phase I of the survey to help reinforce this visual
process. Test subjects were told a brief review of the
study and given up to 60 s to observe a panel card for
the presence of hair and record their results on a stan-
dardized form. Participants were then asked for any
comments or suggestions that might help with the
testing. This training helped the panel leader in direct-
ing the participants and helped the participants to
understand the evaluation process. Based on training
observations, the study parameters of hair length, hair
colors, and test panel card square colors were refined
and the observation time was reduced to 30 s per card
in Phase I of the study.
Mushroom Test
A single layer of sliced commercially canned Agari-
cus spp. mushrooms was arranged in a 100 ⫻ 15-mm
plastic petri dish which had been divided into four
quadrants. A total of 16 different petri dishes were
prepared. In each quadrant, a single hair was ran-
domly placed on a mushroom slice. Unbeknownst to
the panelists, the same hair type and length were
placed in all four quadrants on the same petri dish.
Each petri dish was given a random three-digit num-
ber. Four petri dishes were arranged on a white Sty-
rofoam tray. The trays were placed carefully on the
table prior to examination (two trays per booth). Pan-
elists were not allowed to handle the petri dishes as
this could compromise the position of the mushrooms
and/or hair. Care was taken to prevent the desiccation
of the mushrooms during the test by misting the mush-
rooms with distilled water as necessary and keeping
the petri dishes covered with lids. The petri dish lids
were removed just prior to testing. Panelists were
given 1 min to observe each tray and record their
results on a standardized form.
Statistical Design
Experiments were designed to generate a balanced
data set where there were an equal number of obser-
76 BILES AND ZIOBRO

TABLE 2
Percentage of Panelist Observations of Hair on Colored Squares

Human
All hairs black/brown b Human white Cat/dog Rat/mouse
a
Sig % Sig % Sig % Sig % Sig %

10 mm A 75.2 A 38.9 A 90.1 A 90.1 A 81.5

5 mm B 56.6 B 21.0 B 75.9 B 81.5 B 48.4
3 mm B 58.5 B 20.4 AB 83.3 AB 83.3 B 46.9
1 mm C 27.0 C 7.4 C 53.7 C 30.9 C 16.0
None D 9.0 B 17.9 D 5.6 D 6.2 D 6.2
a
% observed with the same letter are not significantly different at P ⬎ 0.05.
b
(/) ⫽ OR.

vations for every combination of hair lengths, types,
and background colors. ANOVA T test was calculated
by using PC SAS 6.12 (SAS Institute, Cary, NC) with
significance set at P ⬎ 0.05.
RESULTS AND DISCUSSION
The goal of this study was to test the ability of an
average consumer to detect the presence of hair in food
using sliced, canned mushrooms as the test media. The
test design was similar to a dilution test used in sen-
sory evaluation. On the basis of the recommendations
of the Committee on Sensory Evaluation of the Insti-
tute of Food Technologists (Amerine et al., 1965), 48
people were recruited for a semitrained sensory panel
of which 27 people completed both phases of the study.
The surveyed population was drawn from the scien-
tific, clerical, and support staff at the Food and Drug
Administration’s CFSAN. Of the participants that
completed all phases of the study, the population was
74.1% male and 25.9% female, 81.5% of whom wore
corrective lenses and 29.6% of the population had
astigmatism. No one was color blind. The majority
(70.4%) of people surveyed were between the ages of 40
and 59 years old followed by 22.2% between the ages of
25 and 39 years old and 7.4% were more than 60 years
old. There were no participants younger than 25 years
old. This closely resembles the age distribution of the
CFSAN. A minority of the participants (14.8%) did not
eat mushrooms. Of those eating mushrooms 25.9% ate
them one to five times a week; 55.6%, one to five times
a month; and 3.7%, one to five times a year. Of those
that ate mushrooms, the most commonly consumed
type was fresh (85.2%), canned (59.3%), and dried
(37.0%). The percentage total for the types of mush-
rooms consumed exceeds 100% because the partici-
pants eat more than one type of mushroom (Fig. 1).
The training panel showed that there was no statis-
tical difference in response to one pair and one triplet
of the nine background colors. The Phase I survey
showed that there was no statistical difference in re-
sponse to human black and human brown hair and
that there was no difference in the ability to observe
the hair lengths 7 and 10 mm.
Tables 2 and 3 summarize the data from the survey
listing the percentage of panelists that observed the
presence of hair by length, type, and overall appear-
ance. Those percentages with the same letter in the
same column are not significantly different from each
other at P ⬎ 0.05. There was no significant difference
between the different types of hair. As expected it was
easier to see a longer hair on a solid-colored back-
ground than a shorter hair, with no statistical differ-

TABLE 3
Percentage of Panelist Observations of Hair on Mushroom Slices

Human
All hairs black/brown b Human white Cat/dog Rat/mouse

Sig a % Sig % Sig % Sig % Sig %

10 mm A 24.3 A 28.7 B 3.7 A 27.8 A 37.0

5 mm A 25.5 A 35.2 A 20.4 A 38.7 B 17.6
1 mm B 15.7 A 35.2 B 9.3 B 14.8 C 3.7
None C 8.8 B 10.2 B 7.4 B 13.0 C 4.6
a
As in Table 2.
b
As in Table 2.
 VISUAL DETECTION OF HAIR IN FOOD
ence in observation between a 3- and 5-mm hair. To
simplify further testing the 3-mm hair was eliminated
from the sliced, canned mushroom panel. The percent-
age of hair observations dropped when sliced mush-
rooms were used instead of flat, solid-colored squares.
For example, 75.2% of the people observed a 10-mm
hair on a flat, solid colored background. This dropped
to an observation rate of 24.3% when sliced, canned
mushrooms were used. At first glance a mushroom
slice appears to be uniform in color, but color and
specular reflectance differences exist between the
stem, cap, and gills of the mushroom. Furthermore,
within a container of mushrooms, there is a range of
colors depending on the age of the mushroom, when it
was picked, and how long it was held before processing.
The panel had a more difficult time observing human
white hair and cat/dog hair than human black/brown
or rat/mouse hair. The cat/dog hairs were light in color
comparable to human white hair while; the rat/mouse
hair had similar coloration to human black/brown hair.
For observation of hair on sliced mushrooms, the ori-
entation of the hair was a factor in the visibility to the
panelist. For instance, a hair placed perpendicular to a
gill or striation is more easily seen than a hair placed
parallel to the same. This factor, as well as differences
in specular reflectance, luminance, chromaticity, and
transparency may help to explain why the 5-mm hair
was observed more than the 10-mm hair on the mush-
rooms (Table 3). There was a readily apparent differ-
ence in the specular reflectance and transparency be-
tween human white hairs and the darker hairs used in
this study. On a sliced mushroom the human white
hair appeared to reflect the color and luster of the
mushroom. This characteristic made them much more
difficult to observe. This was not so when human white
hair was viewed on a flat, solid-colored square. The
panelists were able to detect the dark hair on a mush-
room slice more often than on the colored squares. The
color contrast between the mushroom slices and dark
hair was greater than the color contrast between the
dark hair and the colored squares. Observation of hairs
either on flat colored squares or on mushroom slices
gave similar rates of false positives (9.0 vs 8.8%).
CONCLUSIONS
In this laboratory panel survey study more than 50%
of the participants were able to detect as small as a
single 3-mm hair placed on a flat, solid-colored surface.
The positive observation rate of approximately 20%
(e.g., the false-positive rate multiplied by the normal
distribution constant, 1.96) is near the visual detection
77
limit found in this study. Observation of hair on a food
matrix is more difficult than observation on a flat col-
ored surface as was demonstrated by the panelists who
detected only 25% of the 5- and 10-mm hairs on the
canned, sliced mushrooms. We conclude 5 and 10 mm
to be a visible length of hair in food. The results of this
study are consistent with existing United States De-
partment of Agriculture guidelines that define a 7-mm
strand of hair as readily discernible and/or highly ob-
jectionable (USDA, 1996).
ACKNOWLEDGMENTS
The authors thank the employees of the U.S. Food and Drug
Administration Center for Food Safety and Applied Nutrition who
participated in this study.
REFERENCES
Amerine, M., Pangborn, R. M., and Roessler, E. B. (1965). Principles
of Sensory Evaluation of Food, pp. 220 –225, 275–313, 417– 423,
and 428. London, England.
Canada Department of Agriculture (1977). Laboratory Methods for
Sensory Evaluation, Publication 1637, pp. 6 –7. Canada Dept. of
Agriculture, Ottawa, Canada.
Eisenberg, W. V. (1981). Sources of Food Contaminants. In FDA
Technical Bulletin No. 1: Principles of Food Analysis for Filth,
Decomposition, and Foreign Matter (J.R. Gorham, Ed.), pp. 11–25.
U.S. Gov. Printing Office, Washington, DC.
Food, Drug, and Cosmetic Act, as Amended, and Related Laws.
(1985). U.S. Gov. Printing Office, Washington, DC.
Food and Drug Administration (FDA) (1980). Laboratory Procedures
Manual, Chapter 18, part 1. U.S. Department of Health and Hu-
man Services, Washington, DC.
Food and Drug Administration (FDA) (1996). FDA Compliance Pol-
icy Guides, Chapter 5. U.S. Department of Health and Human
Services, Rockville, MD.
Food and Drug Administration (FDA) (1998). Natural or Unavoid-
able Defects in Food for Human Use That Present no Health Haz-
ard, Title 21 Code of Federal Regulations, Section 110.110. U.S.
Department of Health and Human Services, Washington, DC.
Hoffman, J. A. (1958). Principles of Psychological Test Methods
(Judgemental Methods—Appearance), Basic Principles of Sensory
Analysis, ASTM STP 433, pp. 98 –105. Am. Soc. Testing Materials,
Philadelphia, PA.
Jellinek, G. (1985). Sensory Evaluation of Food, p. 322. Chichester,
England.
Ryback, I. A., et al. (1998). BMV Behavioral Model of Visual Percep-
tion and Recognition, scientific pages of Ilya Rybak http://
www.voicenet.com⬃rybak/vnc.html. p. 1.
United States Department of Agriculture (1996) Inspection Proce-
dures for Foreign Material, File Code 172-A-1. pp. 1–18. U.S. Gov.
Printing Office, Washington, DC.
Vazquez, A. W. (1981). Hairs. In FDA Technical Bulletin No. 1:
Principles of Food Analysis for Filth, Decomposition, and Foreign
Matter (J.R. Gorham, Ed.), pp. 125–170. U.S. Gov. Printing Office,
Washington, DC.