BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 40(3)


Table 6. Details of canonical correlation analysis for three subspecies ofPuma concolor (coryi, stanleyana,
azteca). Coefficients having dominant loadings on a standardized variable are shown in bold type and
discussed in the text


 Females Males

 Canonical variable Canonical variable

 1 2 1 2


Canonical correlation .89 .68 .87 .66
Eigenvalue 3.70 .87 3.07 .76
Proportion of variance explained .81 .19 .80 .20
Cumulative proportion explained .81 1.00 .80 1.00
Standardized canonical coefficients
 Total length -0.84 -0.59 -1.15 -0.07
 Condylobasal length 1.85 0.05
 Zygomatic breadth 0.87 -0.33 1.18 -1.80
 Cranium height -0.67 0.44
 Width of palate 0.50 1.08 0.44 -0.11
 Palatal length -0.22 -0.80
 Maxillary tooth row -0.49 134 0.09 1.14
 Mandibular tooth row 0.62 -0.54
 Mastoid breadth -0.26 0.27 -0.52 -0.04
 Post-orbital process breadth -1.61 -0.01
 Postorbital constriction 0.56 -0.11 0.55 -0.16
 Width at canines -0.30 136
 Pterygoid width 0.72 -0.88 -0.01 -0.15
 Condyle width 0.15 -0.70 -0.11 0.08
 Upper carnassial crown length -0.74 0.00 -
 Upper carnassial crown width -0.50 -0.54




groups so was reclassified as P. c. coryi. It is interesting to note that the historic
Louisiana males remained within the P. c. coryi sample, however the Louisiana
female did not. Both Everglades cats were reclassified.

 DISCUSSION AND CONCLUSIONS

 Small sample sizes and problematic techniques limit the interpretive value of
some of the results. Had it been possible to combine the characters measured,
include a broader representation of populations, and have more robust samples, a
more cogent picture of the relationship between Florida and other populations