jm A
The most widely adopted system for classifying
dental fluorosis has been in use since 1942. There
may be a better method.
A new method for assessing the
prevalence of dental fluorosis—the
Tooth Surface Index of Fluorosis
Herschel S. Horowitz, DDS, M PH
W illiam S. Driscoll, D D S, M PH
Rhea J. M eyers, RDH, M PH
F
y -M—l xcessive consum ption of fluoride
while teeth are developing can cause den­
tal flu o ro sis, a form of hyp om in er-
alization of tooth enamel and dentin.
Signs of flu orosis range from barely
noticeable white flecks to confluent pits
in the enamel surface and unsightly dark
brown stains. Although fluoride from any
r source may cause dental fluorosis, high
concentrations of fluoride naturally oc­
curring in water supplies account for
m o st of th e c l in i c a l ly u n d e sira b le
fluorosis seen in this country.
The traditional system for classifying
dental fluorosis, as described by Dean,1
has several shortcomings. The most obvi­
ous is that a single score is given to a tooth
rather than a separate score to each tooth
surface. Thus, any differences in the se­
verity of fluorosis in different tooth sur­
faces cannot be ascertained. This defi­
ciency is especially evident for labial sur­
faces of maxillary anterior teeth because
they have special esthetic importance.
Another troublesome feature of Dean’s
index is its inclusion of a “questionable”
diagnostic category that is difficult to de­
fine or interpret precisely. The distinc­
tions between some of the other diagnos­
Stanley B. Heifetz, DDS, M PH
A lbert Kingm an, PhD
tic categories are also unclear or impre­
cise.
Dean’s index lacks sensitivity in areas
where the concentration of fluoride in the
water is exceptionally high. In these
areas, a large proportion of the scores will
fall into only two categories, “moderate”
and “severe.” Moreover, as a person is
assigned to a fluorosis category based
solely on the two most severely affected
teeth, the score may not reflect the overall
amount or severity of fluorosis in the
mouth.
In an attempt to reduce some of the
shortcomings of Dean’s index, we have
d evel oped a new index, the Tooth Surface
Index of Fluorosis (TSIF). The descriptive
criteria and scoring system for the TSIF
are shown in Table 1, and color photo­
graphs, illustrating exam ples of each
fluorosis category, are presented in the Il­
lustration.
With the TSIF, a separate score is given
to each unrestored tooth surface. Two
scores are assigned to anterior teeth (from
the labial and lingual aspects) and three
to posterior teeth (from the buccal, lin­
gual, and occlusal aspects). The TSIF con-
ta in s no “ q u e s t io n a b le ” ca te g o ry .
Fluorosis diagnosed in categories 1, 2,
and 3 may be confined to a single area of
enamel, or may occur irregularly over an
entire surface. An examiner determines
the extent of affected enamel by estimat­
ing the amount of fluorosis as a fraction of
the total visible enamel surface. The TSIF
permits a distinction between discrete
pitting and more advanced confluent pit­
ting and betw een staining alone and
staining in conjunction with pitting, fea­
tures designed to im prove diagnostic
sensitivity in populations where severe
fluorosis is prevalent. When more than
one category of fluorosis exists in a sur­
face, for example, discrete pitting and
staining (score 6) and confluent pitting
(score 7), the highest numerical score is
assigned to the surface. Because the TSIF
is not an interval scale, the scores are not
averaged. The scores, may, however, be
arrayed in various frequency distribu­
tio n s , w h ic h ca n be co m p a re d by
nonparametric tests.
After pretesting, the TSIF was used in
April 1980, when we conducted a cross-
sectional survey to assess the prevalence
of dental caries and dental fluorosis in
communities having optimal and above-
JADA, Vol. 109, July 1984 ■ 37
A R T IC L E S

optimal concentrations of natural fluo­
ride in their drinking waters. Findings of
this study on dental caries prevalence and
fluorosis derived with Dean’s index have
already been reported.2 This paper pre­
sents findings for the same study popula­
tion using the TSIF.
Methodology
T he survey took place in Illinois in seven
c o m m u n itie s w ith s im ila r d e m o g ra p h ic
c h a r a c te r is tic s . T h e c o m m u n itie s w e re
grouped into four categories according to the
relation of their w ater-fluoride concentration
to the recom m ended optim al fluoride concen­
tration for the area (1-, 2-, 3-, or 4-times opti­
mal). The study population consisted of 807
children in grades 3 to 10 (aged 8 to 16) who,
according to questionnaires com pleted by
th eir parents, had always lived in their respec­
tive com m unities and h ad always used the
com m unity w ater su p p ly as th eir prim ary
source of drinking water.
The TSIF exam inations w ere made at the
participating schools by tw o National Caries
Program dentists, w ho used portable dental
chairs, artificial lights, and plane dental m ir­
rors. Each exam iner applied the index to ap­
proxim ately half of the study participants and
u sed criteria described by Russell3 to distin ­
g u ish b e tw e e n flu o ro sis an d n o n flu o rid e
opacities. To be scored, at least one surface of a
tooth had to be com pletely erupted.U nerupted
tooth surfaces and surfaces th a t contained res­
torations w ere excluded.
Findings
Prim ary teeth were not exam ined. The findings
th u s pertain only to fluorosis that occurred in
perm anent teeth. F indings for both examiners w hich are esthetically conspicous. Of these
have been pooled. (The percentages show n in surfaces, 79% received scores of 0 in the opti­
Tables 2 through 5 have been rounded to the mal fluoride area, whereas, at fluoride levels of
nearest tenth of a percent.) 2- and 3-times optimal, 44% and 34.1% of the
The num bers of study participants, along surfaces, respectively, show ed no fluorosis. At
w ith distributions of TSIF scores by percentage 2-times optimal, there was no pitting of the
for all tooth surfaces in each water-fluoride enamel (scores 5 to 7), although 1.3% of the
category, are show n in Table 2. In the optimal surfaces did show evidence of staining. At
fluoride area, fluorosis w as absent in 84.5% of 3-times optimal, 5.4% of the surfaces showed
the tooth surfaces. Of th e surfaces affected, staining, pitting, or both. At the 4-times opti­
m ost show ed only the m ildest form of the con­ mal fluoride level, only 15.8% of the tooth sur­
dition (score 1). In contrast, at the 4-times op­ faces were free of fluorosis, w hitish discolora­
tim al level, only 31.9% of the tooth surfaces tion alone appeared in 73.5% of the surfaces,
w ere free of fluorosis, and 3.4% received scores and 10.9% of the surfaces showed staining,
of 4 through 7. Prevalences and degrees of se- pitting, or both.
The percentages of all tooth surfaces affected by
fluorosis characterized by staining, pitting, or both
w ere 0%, 0.2%, 1.4%, and 3.4%, respectively, in
communities with 1-, 2-, 3-, and 4-times optimal
fluoride.
verity of fluorosis at the 2- and 3-times optimal Table 4 shows distributions of TSIF scores
levels were betw een those extremes. The dif­ by percentages based on the maximum score
ferences in fluorosis scores betw een the 2- and assigned to an individual child for all tooth
3-times optim al levels were slight, although surfaces and for the labial surfaces of the maxil­
the scores consistently show ed more fluorosis lary anterior teeth. TSIF scores of 1, 2, and 3
at the higher fluoride level. have been com bined and expressed as a single
Table 3 shows percentage distributions of percentage figure, as have the scores 4, 5, 6,
TSIF scores for labial surfaces of the six m axil­ and 7. These com binations are logical because
lary an terio r te e th (incisors an d canines), the former scores represent various degrees of

TOOTH SURFACE INDEX OF FLUOROSIS AND ASSIGNED SCORES

0 1

38 ■ JADA, Vol. 109, July 1984

 A R T IC L E S

w hitish discoloration only, whereas the latter
scores generally represent the more u ndesir­
able sequelae of staining, pitting, or both. The
percentages of subjects assigned a score of 0
dropped progressively and dram atically for
both categories of tooth surfaces as the fluoride
levels increased above optim al. The percent­
ages of subjects w ho had m axim um scores of 4
to 7 increased for both tooth surface categories
as the w ater-fluoride concentrations increased
above the optim al level.
D istributions of m axim um scores for each
subject are more amenable to statistical testing
th an are data based on total tooth surface
scores, w hich lack independence. The tw o sets
of data on m axim um scores show n in Table 4
w ere analyzed by an extended x 2 statistic4
after a set of ordinal w eights based on rid its5 of
the m arginal distribution for each examiner
w as assigned to each of the three categories of
fluorosis. The analysis show ed that the scores
across fluoride levels w ere statistically dif­
ferent (P < .001), both for all surfaces and for
the labial surfaces of maxillary anterior teeth.
Also, scores betw een all pairs of fluoride levels
were statistically different (P < .05), except for
the difference betw een 3- and 4-times optimal
for facial surfaces of maxillary anterior teeth (P
= .23).
Table 5 presents a percentage distribution of
TSIF scores for perm anent first molars of chil­
dren aged 8 to 10 and 13 to 16 years for each of
the w ater-fluoride levels. First molars had been
in the m outh for different periods in the two
age groups. In general, the data show a greater
prevalence and intensity of fluorosis among
first m olars of the younger children than in the
sam e teeth of the older children. The same
relation by age group w as apparent for the la­
bial surfaces of the eight perm anent central
an d lateral incisors, w h ich also h ad been
erupted for different periods in the two groups
(data n o t show n but available from the au­
thors).
Because the TSIF is a new index, a special
effort w as m ade to determ in e the level of
agreem ent achieved by the examiners in diag­
nosing fluorosis. They conducted repeat and
d u plicate exam inations to assess intra- and in­
terexam iner agreement, respectively, on a total
of 111 subjects selected random ly from the
study populations of the participating com-
The p reva len ce and severity o f fluorosis at the
optimal water fluoride level w ere typically low, with
almost 85% of tooth surfaces showing no visible signs
o f the condition.
m unities, w ith o u t know ing w hich children
w ere being reexam ined. A greement was as­
sessed for total and m axim um scores for all
to o th surfaces com bined, and for total and
m axim um scores for the labial surfaces of the
six m axillary anterior teeth.
Levels of agreem ent on m axim um scores
w ere tested using kappa statistics.6 Kappa val­
ues for intraexam iner agreem ent ranged from
0.66 to 0.83 w hich, according to Landis and
K och,6 represent strengths of agreem ent rang­
ing from “substantial” to “alm ost perfect.”
K appa v alues for interexam iner agreem ent
w ere predictably lower at 0.35 and 0.54 for all
tooth surfaces and for maxillary anterior sur­
faces, respectively. The respective strengths of
ag reem ent for th ese v alues are “ fair” an d
“ m oderate.” (More detailed inform ation on
exam iner agreem ent may be obtained from the
authors.)
D iscussion
F in d in gs previously reported for this
population,2 as well as those from several
other studies in which Dean’s index was
used,1,7*11 have shown that the prevalence
and severity of dental fluorosis increases
as the fluoride concentration of the drink­
ing water increases. Our findings using
the TSIF clearly demonstrate the same re­
lation.
T h e p r e v a le n c e an d s e v e r ity of
fluorosis at the optimal water fluoride
level were typically low, with almost 85%
of tooth surfaces showing no visible signs
of the condition. Nearly all the surfaces
that did show fluorosis had only its m il­
dest form (score 1). In contrast, the preva-

T a b le l * Descriptive criteria and scoring system fo r the Tooth Surface Index

of Fluorosis (TSIF).____________________________________________________
N um erical D escriptive
score criteria

Enam el shows no evidence of fluorosis.

Enam el shows defin ite evidence of fluorosis, nam ely areas
with parchm ent-w hite colo r th a t total less than one-
thircLof the visible enam el surface. This category includes
fluorosis confined only to incisal edges of ante rio r teeth
and cusp tips of posterior teeth f's n o w c a p p in g ” ).
P archm ent-w hite fluorosis totals at least one-third of the
visible surface, but less than tw o-thirds.
Parchm ent-w hite fluorosis totals at least tw o -thirds of the
visible surface.
Enam el shows staining in conjunction with any of the pre­
ceding levels of fluorosis. Staining is defined as an area
of defin ite discoloration that m ay ran ge from light to
very da rk brown.
Discrete pitting of the enam el exists, unaccom panied by
evidence of staining of in tact enam el. A pit is defined as a
defin ite physical defect in th e enam el surface w ith a rough
flo or th a t is surrounded by a wall of intact enam el. The
pitted area is usually stained o r differs in colo r from the
surrounding enamel.
Both discrete pitting and staining of the in tact enam el exist.
C onfluent pitting of the enam el surface exists. Large areas of
enam el may be m issing and the anatom y of the tooth m ay be
altered. Dark-brown stain is usually present.__________________

Horowitz-Others : METHOD FOR A SSESSIN G DENTAL FLUOROSIS ■ 39

A R T IC L E S

clusive; the latter are merely a component

of the former.
Table 2 ■ Percentage distribution of TSIF scores for all perm anent tooth surfaces accord- Use of a single tooth surface (maximum
ing to w ater-fluoride level, Illinois com m unities. score) to categorize the fluorosis status of <
Percentage distribution of T SIF scores a subject and, subsequently, the status of a
Water-fluoride No. of
level children 0 1 2 3 4 5 6 7 community showed a remarkably clear-
cut, dose-response pattern among the Il­
Optimal 336 84.5 12.4 2.0 1.1 0.0 0.0 * 0.0 0.0
2 x optimal 143 58.1 28.4 7.6 5.6 0.1 0.1 0.0 0.0 linois communities, as did the findings
3 x optimal 192 50.4 25.7 13.2 9.3 0.4 0.8 0.0 0.2 derived from the more comprehensive
4 x optimal 136 31.9 27.0 17.1 20.5 0.4 2.1 0.1 0.8 distribution of scores for all tooth sur­
*Fou r affected surfaces. faces. The definitive dose-response pat- .»
tern is readily apparent in Table 4; each
su ccessiv e increase in w ater-fluoride
Table 3 ■ Percentage distribution of TSIF scores for labial surfaces of perm anent level is accompanied by both a steady de­
m axillary anterior teeth according to water-fluoride level, Illinois com m unities. cline in the percentage of subjects having
a maximum score of 0 and an equally
Percentage distribution of T SIF scores
W ater-fluoride steady increase in the percentage of sub­
level 0 1 2 3 4 5 6 7 jects with scores of 4 to 7. The sensitivity
Optimal 79.0 18.8 1.7 0.6 0.0 0.0* 0.0 0.0 of the TSIF is clearly demonstrated by the
2 x optimal 44.0 39.1 9.6 6.0 1.3 0.0 0.0 0.0 ability to differentiate statistically be­
3 x optimal 34.1 37.0 13.9 9.9 3.6 1.1 0.5 0.2
4 x optimal 15.8 26.2 20.6 26.7 4.1 2.9 0.7 3.2
tween each pair of communities with dif­
feren t w ater-flu o rid e co n cen tration s
*O ne affected surface.
based on all tooth surfaces, and to nearly
do so (except for 3- and 4-times optimal)
based on facial surfaces of maxillary an-
Table 4 ■ Percentage distribution of subjects’ m axim um TSIF terior teeth.
scores for all perm anent tooth surfaces and for labial surfaces of The distributions of fluorosis scores for
perm anent m axillary anterior teeth according to w ater-fluoride first molars and incisors in the two age
level, Illinois comm unities. groups are particularly interesting be­
Percentage distribution of subjects’ maximum T S IF scores cause they provide an in d ica tio n of
changes that may have taken place in the
Labial surface of m axillary
Water- All tooth surfaces anterior teeth prevalence of fluorosis over time. In this
study, the approximate difference be- >
level 0 1-3 4-7 0 1-3 4-7 tween the mean ages of the two groups of
Optimal 40.2 58.9 0.9 67.3 32.4 0.3 children in each of the four water-fluoride
2 x optimal 11.9 80.4 7.7 28.0 68.5 3.5 areas was about 5 years. The data showed
3 x optimal 9.4 71.4 19.3 19.8 68.2 12.0
4 x optimal 2.9 58.8 38.3 7.4 70.6 22.1 that the younger children had distinctly
more fluorosis than the older children in
both first molars and incisors at all four
water-fluoride levels. At least two factors
Table 5 ■ Percentage distribution of TSIF scores for p er­ may help to explain this difference. One,
m anent first molars according to age group and water- suggested by Aasenden and Peebles12 ’
fluoride level, Illinois comm unities. after a long-term evaluation of dental
Percentage distribution of TSIF scores fluorosis associated with dietary fluoride
supplementation, is that abrasion or con­
Water- Aged 8 to 10 years Aged 13 to 16 years
fluoride ------------------------------------ ---------------------------------- tinued mineralization may diminish the
level 0 1-3 4-7 0 1-3 4-7 milder forms of fluorosis with time. That
Optimal 77.6 22.1 0 .0 ’ 92.0 7.8 0.0 abrasion might lower fluorosis scores has
2 X optimal 43.2 56.4 0.0 ’ 68.8 31.1 0.0 also been suggested by Thylstrup and
3 X optimal 41.6 56.4 2.0 66.4 31.9 1.7
4 X optimal 24.8 68.4 6.8 58.9 38.8 2.4
Fejerskov,13 who found that occlusal sur­
faces of posterior teeth generally showed
* Three affected surfaces.
less fluorosis than the buccal and lingual
surfaces of the same teeth.
T he oth er ex p la n a tio n is th at the
younger group of children may have con­
lence of fluorosis increased decidedly higher than the corresponding percent­ sumed greater amounts of fluoride during
and successively at each higher water- ages for all tooth surfaces (Table 2). The tooth development as a result of fluoride
fluoride level. The severity of the condi­ differences were more pronounced at the from multiple sources being more widely
tion at all three higher than optimal levels higher than optimal fluoride levels than available to them than they were to the
was limited to whitish discoloration in at the optimal level. Although a different older children when their corresponding
the majority of tooth surfaces. Even at the r e la t io n e x is te d b e tw e e n th e tw o teeth developed.14 A few anecdotal re­
4-times optimal level, less than 4% of all categories of tooth surfaces when the data ports have also suggested that fluorosis
tooth surfaces showed staining or pitting. were analyzed according to the maximum may be more prevalent in recent years.
The labial surfaces of the maxillary in­ TSIF score assigned to individuals (Table Because these observations have been re-
cisors and canines are usually the most 4), this may be an aberration because a p o rte d fro m b o th f lu o r id a te d and
esthetically important tooth surfaces in subject’s maximum score for all tooth sur­ nonfluoridated areas, excessive fluoride
the mouth. At all water-fluoride levels, faces must be as high or higher than that may be coming from sources other than
the percentages of these surfaces with for facial surfaces of maxillary anterior drinking water. These sources may in­
definite signs of fluorosis (Table 3) were teeth because they are not mutually ex­ c lu d e in f a n t f o r m u la s an d o th e r

40 ■ JADA, Vol. 109, July 1984


foodstuffs processed with fluoridated wa­
ter, fluoride den tifrices in g ested by
young children, dietary fluoride supple­
ments given according to the dosage
schedule formerly recommended by the
American Academy of Pediatrics, 15 and
dietary fluoride supplements prescribed
without knowledge of or without regard
for the amount of fluoride already present
in drinking water. Although the manufac­
turers of infant formulas, baby cereals,
and juices took steps in the late 1970s to
limit the fluoride content of those prod­
ucts to negligible levels16 (Foman, S., per­
sonal com m unication, 1982), and, in
1979, the American Academy of Pediat­
rics17 adopted a new schedule specifying
lower fluoride supplement dosages for
children younger than 2 years of age, the
teeth examined in this survey would have
been at risk to fluorosis before these
changes.
Because definitive information is lack­
ing on whether, or to what extent, these
factors contributed to the differences in
fluorosis scores betw een the two age
groups, their impact remains uncertain.
The findings, however, emphasize a need
for further research, including longitudi­
nal studies in which fluorosed incisors
and first molars are periodically reevalu­
ated to determine the effects of abrasion
and mineralization on fluorosis. It should
be stressed, however, that the higher
levels of fluorosis in the younger children
o ccu rred a lm o s t e n tir e ly in TSIF
categories 1 ,2 , and 3, which reflect only
varying degrees of white blemishes. Al­
though these degrees of fluorosis may be
esthetically objectionable in some instan­
ces, they are frequently detectable only on
close visual inspection. Thus, the higher
p r ev a len ce of flu o r o s is am ong the
younger children should not cause undue
concern for the appearance of their teeth.
The TSIF helps to overcome many of
the shortcomings associated with Dean’s
index. Because the data can be analyzed
on a surface-specific basis, the TSIF is
especially useful for determining the
public health effect of fluorosis in a popu­
lation. Clearly, the TSIF was sufficiently
sensitive in this study to distinguish the
communities with four different fluoride
levels with regard to both prevalence and
severity of fluorosis. In fact, with the use
of Dean’s index, there was no obvious
dose-response relation between the water
fluoride levels of 2- and 3-tim es op­
timum2 whereas, with the TSIF, there was
a distinct relation. Use of the TSIF in the
field did not disclose any particular need
to modify or refine it, although categories
5 and 6 might be combined into a single
category because relatively small per­
centages of tooth surfaces received scores
of 6.
The TSIF is not difficult to learn, nor is
it time-consuming or difficult to apply.
The examiners, after a period of training,
were able to achieve acceptable levels of
intra- and inter-exam iner diagnostic
agreement, although they did have some
difficulty agreeing on whether fluorosis
was present, that is, distinguishing be­
tween scores of 0 and 1. Therefore, initial
standardization should stress the point at
w hich fluorosis is considered present.
The examiners showed excellent agree­
ment on the more advanced forms of
fluorosis. We have found the new index to
be sensitive and useful for assessing den­
tal fluorosis. Before the TSIF can be
unequivocally recommended, however,
it should be used and evaluated by other
investigators under varied conditions.
Summary
A new index for measuring the preva­
lence of dental fluorosis, the Tooth Sur­
face Index of Fluorosis (TSIF), was used
to assess the condition in the permanent
teeth of 807 children, aged 8 to 16, who
had resided all their lives in one of seven
Illinois com m unities w ith an optimal
concentration of fluoride in its water or
with 2, 3, or 4 times the optimal concen­
tration. Fluorosis was absent in 84.5% of
all tooth surfaces examined in the com­
munity with optimal fluoride. In contract,
only 31.9% of tooth surfaces had no
fluorosis in the community with 4-times
the optimal fluoride level. In the optimal
fluoride area, 79% of facial surfaces of
maxillary anterior teeth, which are esthet­
ica lly co n sp icou s, had no fluorosis,
whereas only 15.8% of these surfaces in
the 4-times optimal area were unaffected.
The percentages of all tooth surfaces af­
fected by fluorosis characterized by stain­
ing, pitting, or both were 1%, 8%, 19%,
and 38%, respectively, in communities
with 1-, 2-, 3-, and 4- times optimal fluo­
ride. Differences in fluorosis based on
maximum score for all tooth surfaces in a
child were statistically significant (an ex­
tend ed x 2 sta tistic) am ong all com ­
munities.
First molars and incisors in children 8
to 10 years old were affected by more
fluorosis than were the same teeth in
children 13 to 16 years old. These teeth
had been erupted for about 5 years longer
in the older age group. The difference
might result from abrasion or reminerali-
zation of these teeth in the older children
or from greater consumption of fluoride
by the younger children during tooth de­
velopment. Plans have been made to as­
certain the reason for the difference.
Because the TSIF permits a surface-
specific assessment of fluorosis, it is es­
pecially useful for determining the public
health effect of fluorosis in a population.
It proved to be sufficiently sensitive in
this study to distinguish the communities
with four different fluoride levels with
H orow itz-O thers: METHOD FOR ASSESSING DENTAL FLUOROSIS ■ 41
A RTICLES
regard to both prevalence and severity of
dental fluorosis.
___________________________ J&DA
T he informed consent of all hum an subjects who
participated in the investigation described in this
paper was obtained after the procedures and possible
discom forts and risks had been fu lly explained.
T he authors thank M s. Sh aron P ierce and Dr.
Raymond Kuthy, division of dental health, Illinois
Department of Public Health, for coordinating the
study locally. T he authors also thank the students,
faculty, and adm inistrative personnel of the par­
ticipating schools for their cooperation and assis­
tance.
Drs. Horowitz and Driscoll, Ms. Meyers, and Dr.
Heifetz are w ith the clin ical trials section and Dr.
Kingman is w ith the biometry section, Epidemiology
and Oral Disease Prevention Program, National Insti­
tute of Dental Research, National Institutes o f Health,
W estw ood B ld g , Rm 5 3 8 , 5 3 3 3 W estbard A ve,
Bethesda, MD 20205. Address requests for reprints to
Dr. Horowitz.
1. Dean, H.T. T he investigation of physiological
effects by th e epidem iological method. In Moulton,
F.R., ed. Fluorine and dental health. Am erican Asso­
ciation for th e Advancement o f Science, pub no. 19,
W ashington, DC, 1942, pp 23-31.
2. Driscoll, W .S., and others. Prevalence of dental
caries and dental fluorosis in areas w ith optim al and
above-optimal water fluoride concentrations. JADA
107(l):42-47, 1983.
3. Russell, A.L. T he differential diagnosis o f fluo­
ride and nonfluoride enam el opacities. Public Health
Dent 21:143-146, 1961-62.
4. Landis, J.R.; Heyman, E.R.; and Koch, G.G. Av­
erage partial association in the three-way contin­
gency tables: a review and discussion of alternative
tests. Int Stat Rev 237-254, 1978.
5. Fleiss, J.F.; Chilton, N.W.; and W allenstein, S.
Ridit analysis in dental clin ical studies. ) Dent Res
58:2080-2084, 1979.
6. Landis, J.R., and Koch, G.G. T he measurement of
observer agreement for categorical data. Biom etrics
33:159-174, 1977.
7. Dean, H .T. E p id em io lo g ical studies in the
United States. In Moulton, F.R., ed. Dental caries and
fluorine. Am erican Association for the Advancement
of Science, Lancaster, PA, Scien ce Press, 1946, pp
5-31.
8. Galagan, D.J., and Lamson, G.G. Climate and
endemic dental fluorosis. Public Health Rep 68:497-
508, 1953.
9. R ichards, L.F., and others. D eterm ining op­
timum fluoride levels for com m unity water supplies
inrelation to temperature. JADA 74(2):389-397,1967.
10. Kunzel, W „ and Soto Padron, F. Caries and
d ental flu o ro sis in Cuban ch ild ren . C aries R es
10(2):104-112, 1976.
11. W enzel, A., and Thylstrup, A. Dental fluorosis
and lo calized enam el o p acities in flu orid e and
n o n flu o rid e D a n ish c o m m u n itie s . C a rie s R es
16:340-348, 1982.
12. Aasenden, R ., and Peebles, T.C. Effects of fluo­
ride supplem entation b om birth on dental caries and
flu oro sis in teen aged c h ild re n . A rch O ral B io l
23:111-115, 1978.
13. Thylstrup, A., and Fejerskov, O. Clinical ap­
pearance of dental fluorosis in permanent teeth in
relation to histologic changes. Community Dent Oral
Epidemiol 6 :3 1 5 -3 2 8 ,1 9 7 8 .
14. Leverett, D.H. Flu orides and the changing
prevalence o f dental caries. Scien ce 2 1 7 :26-30,1982.
15. Committee on Nutrition, Am erican Academy
o f P ed iatrics. Flu orid e as a n utrien t. P ed iatrics
49:456-459, 1972.
16. Tinanoff, N.; Pinkerton, R.; and Ramanan, C.
Connecticut physician’s role in preventing dental
caries. Conn Med 45:141-143, 1981.
17. Committee on Nutrition, Am erican Academy
of Pediatrics. Fluoride supplementation: revised dos­
age schedule. Pediatrics 6 3 :1 5 0 -1 5 2 ,1 9 7 9 .