BREVIORA 
Museitaim of Connparative Zoology 
Cambridge, ]Mass. December 10, 1965 Number 236 
VARIATION IN THE NUMBER OF 
MARGINAL TOOTH POSITIONS 
IN THREE SPECIES OF IGUANID LIZARDS 
By 
Clayton E. Ray 
U. S. National Museum 
Although tooth counts (more accurately, the number of tooth 
positions) are customarily given in the description of fossil lizards, 
their possible taxonomic value is generally unassessable for want 
of quantitative data on adequate samples of modern lizards. 
Further, the possible developmental and adaptive implications of 
tooth number have been inadequately explored. Edmund (1960, 
p. 66) notes an apparent lack of correlation between wave length 
in tooth replacement and number of tooth positions in iguanids, but 
the condition in juveniles of species with long wave lengths, such as 
Ctenosaura pectinata with 11 teeth per wave, would be especialh' 
interesting. Hotton (1955, p. 97) in a study of adaptive relations 
of dentition to diet in iguanids, including C. similis, states regard- 
ing tooth number only that there are "slight tendencies toward 
fewer teeth in the maxillary row ... in smaller individuals." 
The availabihty of adequate, though not impeccable, samples 
of cranial material (assembled for other purposes) representing 
Ctenosaura similis, C. pedinata, and Anolis carolinensis suggested 
the desirability of characterizing these samples statisticalh^, in 
order to compare ontogenetic and taxonomic variability in number 
of tooth positions, and strength of correlation between number of 
tooth positions and length of tooth row among two closely related 
taxa and one distantly related taxon. 
2 BREVIOKA No. 236 
MATERIAL 
Ctenosaura similis: Well-preserved, dissociated elements repre- 
senting no less than 223 individuals, collected from deposits in the 
ruins of Mayapan (20°37'48"N, 89°27'42"E), Yucatan, during 
expeditions (1950-1956) of the Department of Archaeology, 
Carnegie Institution of Washington. Among the more than 6000 
bones identified from these deposits, over a third pertain to 
C. similis, the only lizard represented (Pollock and Ray, 1957). 
The material utilized in the present study undoubtedly dates 
almost entirely from the period of major occupation of the city, 
the Maya Resurgence, estimated (and in part confirmed through 
radiocarbon dating) to have extended from about 1200 through 
1450 A.D. Thus the sample was accumulated during an interval 
of perhaps 250 years or more. The specimens probably were 
secured within close proximity of the city. The sample is undoubt- 
edly biased in favor of large individuals. Early juveniles are absent 
altogether. Although there is certainly high representation of left 
and right sides of single individuals among the dissociated maxillae 
and dentaries, the correspondence is far from complete and can be 
demonstrated in no given instance. The material was made avail- 
able for study and was deposited in the Museum of Comparative 
Zoology at Harvard College through the courtesy of Dr. H. E. D. 
Pollock. 
Ctenosaura jpectinata: Cleaned skulls of 17 modern specimens 
collected over many years at widespread localities through the 
range of this nominal species. Specimens are housed in the col- 
lections of the Museum of Comparative Zoology, American Mus- 
eum of Natvu'al History, Chicago Natural History Museum, and 
University of Michigan Museum of Zoology, and were made avail- 
able by the curators of those collections. 
AnoJis caroJinensis: iNIacerated skulls of 58 individuals collected 
during 1955 in the vicinity of Gainesville, Alachua County, Florida. 
This sample, like those of Ctenosaura, does not include very young 
individuals, and thus is not ideal for the study of growth phe- 
nomena. In other respects it is for statistical purposes the most 
desirably constituted sample of the three. The specimens were 
made available by Dr. Walter Auffenberg and are deposited in the 
Museum of Comparative Zoology. 
METHODS 
The following symbols represent the statistics calculated in this 
study : 
N = the number of individuals in the sample. 
1965 TOOTH VARIATION IN IGUANID LIZARDS 3 
X = the mean of x, h(n-e the straight-Hue length of tooth row, 
maxillary or dentary. 
y = the mean of y, here the number of teeth or tooth posi- 
tions per maxilla, dentary, or premaxilla (fused pre- 
maxillae). 
Sx = the standard deviation of x. 
Sy = the standard de\-iation of y. 
Vx = the coefficient of variation of x. 
Vy = the coefficient of variation of y. 
ORx = the observed range of x. 
ORy = the observed range of }'. 
r = the coefficient of correlation, here between x and y. 
a = the growth ratio. 
(Ta = the standard error of a. 
b = the initial growth index. 
Sd = the standard deviation of the diagonal distances of points 
from the reduced major axis; a measure of absolute dis- 
persion of points about that axis. 
Dd = the coefficient of relative dispersion about the reduced 
major axis, 
z = statistic computed in testing for significant difference in 
slope of growth lines. 
]Most of these are the familiar statistics of univariate and bivari- 
ate analysis dealt with in standard statistical handbooks. However, 
Imbrie's paper (1956) is especially valuable regarding the basis for 
utilization of the reduced major axis rather than regression lines, 
as well as for a lucid interpretation of all of the statistics used 
herein. 
Premaxillae: The fused premaxillae were treated as a single 
median element. Only the number of tooth positions was deter- 
mined for this element, as adeciuate means of measuring the (short) 
length of the tooth row, especially in Anolis, were not at hand when 
the samples were assembled. Thus only the standard univariate 
statistics are presented (Table 1, Figure 1). 
Maxillae and dentaries: All of the statistics listed abov^e were 
computed for Ijoth left and right maxillae and dentaries of the two 
samples of Ctcnosaura, and for the right side only of Anolis (Tables 
2, 3, and 4). As a check on symmetiy, the frequency distribution 
for length of tooth row and number of tooth positions was in 
every case represented by histograms, one example of which is 
presented here (Fig. 2). Scatter diagrams of the two \'ariates were 
prepared in each case, from which it was determined that the dis- 
tribution of points approximates a straight line closel}^ enough to 
4 BREVIORA No. 236 
permit representation of the growth hne by the e(iuation: 
y = ax + b. The scatter diagrams for all elements of the right 
side are presented in Figures 3 and 4. 
RESULTS 
Premaxillae: As might be anticipated, the two species of 
Ctcnosaura are very similar, in fact essentially identical, in statis- 
tical characteristics of tooth number, and differ considerably from 
Anolis carolinensis (Table 1). Actually, these representatives of 
the two genera differ much more strikingly than the statistics alone 
reveal. For practical purposes it can be said that the variate Avill 
never assume an odd value in Anolis^, whereas it does so in 70 per 
cent or more of the Ctenosaura sampled (Fig. 1). The tooth 
count in an Anolis premaxilla more likely than not will be 10, and 
in Ctenosaura 7. The fact that premaxillary tooth counts in Anolis 
can assume only even values might suggest that the coefficient of 
variation (Vy) for Anolis is artificially high. However, if only one 
side of the fused element is considered, with tooth counts of 3-6, 
Vy is not altered. The V for tooth counts here and in the maxillae 
and dentaries is of little interest except that it reflects rather low 
variation in tooth number over the broad size range represented 
in the samples. 
]\Iaxillae and dentaries: The frequency distribution for number 
of tooth positions is in general approximately symmetrical (Fig. 
2B), although it is skewed considerably to the right in the right 
dentary of Anolis. The length of tooth row in no case approaches 
a normal curve, and a polymodal frequency distribution is sug- 
gested for Ctenosaura similis (Fig. 2A) and Anolis, particularly the 
former; N is not adequate to suggest a pattern in C. pectinata. It 
is tempting to suppose that broadly overlapping frequency dis- 
tributions of different sexual or age classes or both are responsible 
for the polymodality, analogous to that demonstrated by Klauber 
(1937, fig. 4) for Crotalus v. viridis, but this could not be tested on 
the basis of samples at hand. 
The number of tooth positions per se is not very useful in dis- 
tinguishing the three taxa tested. It is of no value in separating 
the two species of Ctenosaura, and its power of resolution is very 
low even between taxa as remote as Ctenosaura and Anolis. Tooth 
' Reference to "Anolis" and to "Ctenosaura" is to be understood as a shcrt- 
hand device applying only to the samples tested here, and in no case to either 
genus as a whole. 
1965 TOOTH VARIATION IN IGUANID LIZARDS 5 
count alone in an isolated fossil specimen would be of taxonomic 
interest only if it lay within the portion of the Ctenosaura dis- 
tribution well above the OR of Anolis. The sample of Anolis is 
separable from those of Ctenosaura on the basis of absolute length 
of toothrow, which reflects the valid biological distinction that 
Ctenosaura is a much larger lizard, although the true population 
parameters unquestionably overlap. The two samples of Ctenosaura 
are not clearly distinguishable on the basis of any statistic pre- 
sented in Tables 3 and 4. The relatively high values of Vx, Vy, sj, 
and Dd in Ctenosaura pectinata reflect principally its greater ORx 
but may result in part also from geographic variation, not a factor 
in the other two samples. 
The most interesting features of the relation between length of 
tooth row and number of tooth positions in the samples studied 
are the relative strength of correlation between the two variates 
and the slope of their growth line-. The two samples of Ctenosaura 
are essentially identical in these features. Both samples exhibit 
a strong correlation between the variates, both visually (Figs. 3A, 
B, and 4A, B) and statistically (minimum value of r = .83). The 
probability is greater than .05 that the minor deviations in slope 
of the growth lines are attributable to chance, for the values of z 
are in every case less than 1.96 (Table 4). 
Anolis, on the other hand, differs radically from Ctenosaura 
both in strength of correlation and in slope of growth line. Cor- 
relation is not visually detectable (except in a mildly suggestive 
linear clustering of points within the wide scatter) in Figures 3C 
and 4C, and is only weakly indicated by the correlation coefficient. 
However, the probability of obtaining a calculated r as great as 
.24 if the sample was drawn from a population in which p = is 
less than .10, or, if, as is reasonable on developmental grounds, it 
is assumed that r is necessarily positive, less than .05 (Simpson, 
Roe, and Lewontin, 1960, appendix table V, p. 426). The slope of 
the growth line is very steep in Anolis, very gentle in Ctenosaura. 
The difference in slope is statistically highly significant (Table 4), 
but biologically means only that the teeth of Atiolis are con- 
siderably smaller than those of Ctenosaura. Thus, more teeth per 
unit increase in length of tooth row can be accommodated in 
Anolis than in Ctenosaura. 
- Strictly speaking, tlie line is not purely a "growth" line as the scatter of 
points on which it is based undoubtedly represents in part non-ontogenetic 
variation. Designation of the line as a growth line rests on the assumption 
that regression in individuals of different sizes is the same as in individuals of 
different ages. 
6 BItEVIORA No. 236 
Although so few samples provide the basis for no more than 
suggested interpretations, it may be supposed that a difference such 
as that observed in correlation between number of tooth positions 
and length of tooth row has adaptive significance. The most 
fundamental adaptive distinction relating to dentition between 
Anolis and Ctenosaura is in diet; Anolis is a predatory insect 
eater, and Ctenosaura an extreme vegetarian. In Ctenosaura the 
maxillary and dentary tooth rows constitute continuous dental 
palisades in which the broadly expanded, strongly cusped crowns 
(more pronounced posteriorly) succeed one another in close 
order, and tend to overlap, an arrangement apparently advan- 
tageous in leaf-chopping, whereas in Anolis the tooth rows con- 
stitute open series in which the broadly based, apically narrow 
teeth with weak cusps are separated by variable gaps, possibly 
adaptive in the apprehension and puncturing of insect prey (Fig. 5). 
If in a strict vegetarian there is selective advantage in maintaining 
a continuous tooth row, it would be reasonable to expect a con- 
sistent addition of tooth positions with increasing size, i.e., a 
strong positive correlation between the two phenomena. The 
slope of the growth line would depend in part on the size of addi- 
tional teeth and on the degree of increase in size of successive 
replacement teeth. In a form such as Anolis carolinensis with no 
apparent selection pressure toward a closed tooth row, and perhaps 
with positive selection for an open tooth row, one might expect 
only a loose correlation between number of tooth positions and 
length of tooth row. Obviously these suggestions require testing 
on broader ontogenetic series of the taxa studied, and on similar 
series of other taxa. 
I wish to thank Drs. Thomas Frazzetta, Nicholas Hotton, 
George Simpson, and Ernest Williams for reading the manuscript. 
LITERATURE CITED 
Edmund, A. G. 
19()0. Tootli replacement phenomena in the lower vertel)rates. Roy. 
Ontario Mus., Life Sri. Div., Contrib. No. 52: 1-190, 5S fig-s. 
Hotton, Nicholas, III 
1955. A survey of adaptive relationships of dentition to diet in the North 
American Iguanidae. Am. Midland Nat., 53: 88-114. 
Imbrie, John 
1956. Biometrical methods in the study of invertebrate fossils. Bull. Am. 
.Mus. Nat. Hist., 108: 211-252, lU figs. 
Klauber, L. AI. 
1937. A statistical study of the rattlesnakes. IV. The growth of the 
rattlesnake. Occ. Pap. San Diego Soc. Nat. Hist., 3: 1-56, 6 figs. 
1965 
TOOTH VARIATION IN IGUANID LIZARDS 
Pollock, H. E. D., and C. E. Ray 
1957. Notes on vertebrate animal remains from Mayapan. Carnegie 
Inst. Washington, Dept. Archaeol., Current Repts., 41: 633-656. 
Simpson, G. G., Anne Roe, and R. C. Lewontin 
1960. Quantitative zoology. Harcourt, Brace and Co., New York, 
vii + 440 pp., rev. ed. 
(Received 24 August 1965) 
Table 1. Univariate statistical characterization of tooth coinit (y) 
in the fused premaxillae of three iguanid lizards. 
8 
BREVIORA 
No. 236 
03 
> 
(M 
03 
H 
w 
S 
h4 
O 
o 
C-4 
GO 
t^ r^ 
o 
o 
^^ -H -H -H 
-rt< CO (N C-. 
CO CM C^ -* C5 
i-~; od T- 1 T-i 
t-^ lO rt< CO CO CM 
rt CO lO '^^ GO 05 lO 
CM ' ^^ ■^' i-I lO 
4] ^^ ^1 ^^ ^i -H i 
^ CM CO CO »0 t^ 
-O:' QO 05 CO fC T-i -* 
^' CO GO CM 00 .-i O 
rt CO CM 
C2 
CM i-H .-1 
r-H 05 CM CO GO 
CO CO CM 
-H -H -H -H -ti -H 
c^i 
,3 rv) lO t^ 
CO CO CM 
CO c^ 
-^ 66 
c^i oi 
CM 
c^i 
CO 
CO CM 
^ ^ 
CO 
1^ 
o; o CO 
'^ X i^ 
^ 41 ^ 
CO CO CO 
CO CO lO 
I 
CO (M 
t^ '>\ 
CM Oi 
CM 
C5 
CM 
I 
03 
X x_ 
X CM —I 
^ O 
X CO 
o 
-Tfl -^ 03 CO 
O CO CM CO 
" T^jH lO CO 
.^ CO CO 
t^ lo X CO CO 
<Y QC CM_ O ^_ 
O CM 
O X 
lo cm' 
X CM CM 
X CO o 
o; 
o 
CO 
I 
C3 O 
LO 0^1 O^J 
X CO o 
CO 
CO 
rt_ CM_ 
^ d 
GO 
> > Ph tf 
o o 
C3 ^ 
■73 -a 
1965 
TOOTH VARIATION IN IGUANID LIZARDS 
9 
o 
o 
o 
o 
C3 
O 
O o3 
-1^ c 1 
O .^ 
C5 O 
LO lO 
— _-, I >-'• ^fc^ l^ ^ti "■•' 
"-^ 'M C^ ^ O !M CO 
o 
•M (M CI 
-= bC 
s. 
3 
CO 
-— t- 
C^ LO IM 
10 BREVIORA No. 236 
Table 4. Values of the statistic z in tests for significant difference 
in slope of pairs of growth lines. 
Maxilla Dentary 
C. si mil is 
left -right .22 - .84 
C. pectinata 
left -right - .:?17 - .08 
('. similis - C. pedinata 
left -left 1.12 - 1.28 
right - right ..52 - 1.02 
('. sitnilis - Anolis 
right - right 5.83 5.28 
( '. pectinata - Anolis 
right - right 5.82 4.70 
1965 
TOOTH VARIATION IN IGUANID LIZARDS 
11 
5 6 7 8 9 10 
TOOTH COUNT 
5 6 7 8 9 10 
TOOTH COUNT 
5 6 7 8 9 10 
TOOTH COUNT 
II 12 
Fig. 1. Histograms illustrating the frequency distribution of tooth counts 
in the fused ])renia\illae of (A) Ctcnos^aura sinnlif^, (B) ('. pcrtiiiata, and 
(C) AnoUs carolineiisis. 
§6 
m 
I 
20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 
LENGTH OF TOOTH ROW (MM) 
20 
18 
16- 
14- 
12 
10 
8 
6 
4 
2 
I 
I- 
f 
n': 
22 24 26 28 30 32 34 36 38 
TOOTH COUNT 
Fig. 2. Histogianis illu.strating the frequency distribution of (A) length 
of tooth row and (li) tooth counts in the right dentary of Ciencsaiira sitnilis. 
12 
BREVIORA 
No. 236 
16 
28 
h- 
226 
O 
X 
fe2 
"20 
18 
18 20 22 24 26 28 30 32 34 36 38 
LENGTH OF TOOTH ROW (MM) 
40 42 44 46 
40 42 44 46 
16 18 20 22 24 26 28 30 32 34 36 38 
LENGTH OF TOOTH ROW (MM ) 
22r 
21 
20 
19 
0,8 
17 
16 
/ 
./ 
/ 
/ 
6 7 8 9 
LENGTH OF TOOTH ROW (MM ) 
10 
Fig. 3. Scatter diagrams showing tooth counts (y) and length of tooth 
row (x) in the right maxillae of (A) Ctenosaura similis, (B) C. pcciinata, and 
(C) Anolis carolinensis. The broken line represents the reduced major axis; 
the open circle, the joint mean. 
1965 
TOOTH VARIATION IN IGUANID LIZARDS 
13 
38- 
36 
1-34 
Z 
32 
O 
o 
028 
26 
24 
2 -•'••• 
. 3 •• -• 
\ 
_J i \ [ L_ 
I. I A- t I I 1 1 1 1 L- 
36 
34 
32 
■30 
28 
026 
I- 
24 
22 
20- 
18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 
LENGTH OF TOOTH ROW (MM) 
,o^ 
-1 — I — J — I — \ — I — I — I — I — 1 — i — i — I — I — \ — I — \ — I — I I 1 1 I I I I 1 I 1 j_ I '_ 
18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 
LENGTH OF TOOTH ROW (MM ) 
25r 
24 
23 
22 
O 
O 
21 
20 
19 
/ 
/ 
/ 
/ 
/ 
; / 
/ 
/ 
•/ 
A/. 
/ 
6 7 8 9 10 
LENGTH OF TOOTH ROW (MM) 
Fig. 4. Scatter diagrams showing tooth counts (y) and length of tooth 
row (x) in the right dentaries of (A) Ctenosaura similis, (B) C. pectinata, and 
(C) Anolis carolinensis. The broken line represents the reduced major axis; 
the open circle, the joint mean. 
14 
BREVIORA 
No. 236 
1965 TOOTH VARIATION IN IGUANID LIZARDS 15 
o3 -tj' '-5 •— ij 
fS S c3 is c3 
£ =^ >^ ^ . 
— ■M -I-' -li 
- c >^ = J: 
.s •- ^ r 2 
~ =- .2 -►^ u= 
;o -"-^ ■■^ ^ o 
"^ >-• ^ s ^ 
fc«^ ^ C O >j 
S '5 
c3 
— >j 
^ "S ^ "C - 
•2 "m "£ is ^ 
05 "i" — '^ ^ 
^ .3 5 "I "5b 
/-^ -*^ — '^X 'i^ 
■2 -c :r ^^ o 
-; ->! ~ ?^ -^ 
2 =^ .s ^ '-z 
-r f-^ is <^ a 
— « CO a 
5 S3 _- ;c: -g 
c3 O -w '~ z. 
^ *5 ~ —^ -^ 
§ § t= a S:' 
«i 
S c 
CJ ~ T3 -kJ u^. 
■^ £ "S I I 
P*' -^'" ~ "^ -2 
5? ^ .5 ~ 
.M -^ N ^ 'TJ 
ii, •:= U -g ^