THE APPLICATION OF EQUATION MODELS TO DETERMINE THE AGE OF PINK RIVER DOLPHIN SKULLS - Biology, Evolution and Conservation of River Dolphins Within South America and Asia - page 90

the squamosal bone, with R 2 = 0.4641. The width of premaxillaries at midlength of rostrum
(V6) was the morphometric trait which showed a significant relationship with age, but with
lower correlation.
Table 4. Linear regression equations Y = a + bX . V = Significant variable. P = significant
probability. R 2 = coefficient of determination.

X P R 2 Y = a + b

Age -38.8895 2.6687 V15 0.0000 0.4641

Age -27.1907 5.1523 V21 0.0000 0.4275

Age -30.1253 16.5529 V42 0.0000 0.3854

Age -39.108 13.1322 V30 0.0000 0.3739

Age -36.7558 5.9261 V3 0.0000 0.3637

Age -30.5325 2.7541 V12 0.0001 0.3555

Age -23.2609 2.9087 V11 0.0001 0.3314

Age -29.8114 3.6238 V37 0.0002 0.3160

Age -14.0205 6.5668 V29 0.0003 0.2986

Age -5.9999 5.1261 V26 0.0004 0.2816

Age -12.3174 3.9934 V27 0.0005 0.2730

Age -18.6417 5.1957 V23 0.0005 0.2726

Age -16.961 1.5761 V38 0.0019 0.2498

Age -33.0639 3.9577 V22 0.0014 0.2371

Age -48.0576 14.9523 V41 0.0015 0.2291

Age -21.0493 1.2048 V2 0.0016 0.2233

Age -20.2478 1.0683 V9 0.0039 0.1950

Age -21.7943 0.8342 V1 0.0053 0.1783

Age -17.6538 1.2612 V33 0.0059 0.1745

Age -7.3988 1.942 V13 0.0092 0.1697

Age -19.1031 8.4843 V28 0.0091 0.1657

Age -19.8644 2.665 V19 0.0133 0.1438

Age -3.7023 0.5223 V35 0.0328 0.1307

Age -3.1229 9.093 V6 0.0441 0.0975

Multiple Lineal Regressions

Out of the 24 variables employed (those which were significant in the linear regression
analysis), 8 skull measurements turned out to be highly significant in the multiple regression
(Table 5). That is to say, these were the variables that correlated the best (compared to other
variables) with age. The maximum width between zygomatic processes of squamosal bones
again turned out to be the variable that better correlated with the age, with p = 0.00001 and R 2
= 0.0861. In this case, the smaller diameter of the left temporal fossa proper was the variable
which showed the smallest correlation of the significant variables. R 2 was 0.8171 for the 24
variables employed, which means that these variables explained 81 % of the age variation.
In Table 6, the observed age estimates for the 71 exemplars analyzed were compared
with those estimated throughout the multiple regression equation calculated.