Freak Felids - A Discussion of History's Largest Felines

[tigerluver]

Upon Sanjay's request, I'll start this thread. I've some spare time these days, should be fun!

I think it would be interesting to discuss the evolution and morphology of prehistoric cats.

I'll start by transferring my P. s. fossilis post here, and be back later.

(06-22-2014, 10:27 AM)'tigerluver' Wrote: Dug up an old post I wrote from yuku, I haven't worked with cave lion data in a while.

"tigerluver wrote:

The Skull
This post will shortly discuss the theoretical mass tied with the large skull (GSL = 484.7 mm, theoretical CBL = 433 mm). 


The Mazak et al. (2011) gave an estimate of 445 kg, and also seems to overestimate mass. The reasoning behind this is explained in Christiansen and Harris (2005), as follows:
    "A data sample with many small species would introduce a size-related bias, producing unreliably high body mass estimates for large species."

Mazak et al. used the average body mass and condylobasal lengths of each specie as the database to derive the equation. Thus, from the sample size of 6 data points (n=6), 4 were representative of relatively smaller species (P. pardus, N. nebulosa, P. onca, and P. uncia) while 2 were representative of the large species (P. leo and P. tigris). Graphically, there was an uneven distribution of data points, with the smaller species being represent on one extreme and the large on another. Therefore, the data sample had too many small species relative to the amount of large species represented, and thus there was, "a size-related bias, producing unreliably high body mass estimates for large species" (Christiansen and Harris, 2005). 

Mazak et al. (2011) used a species averaged database to prevent confusion between intra- and inter-specific allometry. Though, in reducing the sample size, the distribution of data became uneven, causing the size-related bias mention above. 

I constructed a logarithmically scaled graph using the same database of specimens from Mazak et al. (2011), but had each individual specimen to represent a data point rather than a specie average representing a data point. This produced a plot with an even distribution of data points. The resulting equation:
log(body mass in kg) = 2.6725*log(condylobasal length in mm) - 4.4587

An implication of this equation is that skull size grows more rapidly than body mass. Furthermore, the data sample used can be more safely applied to P. spelaea as P. spelaea is a distinct species, rather than a subspecie of anomalous species in terms of relative proportions and body mass (e.g. P. t. soloensis to P. tigris), and thus one can assume P. spelaea follows the growth trend of Panthera in general. I realize the wording in this paragraph may be a bit confusing, so just ask if any further clarification is needed on the point I am making.

Finally, the equation discussed yields a theoretical body mass for the 484.7 mm skull of approximately 387 kg.

The Femur

The femur estimate you got is similar to the one I have found with regression. I assumed that P. spelaea had a build midway between tigers and lions and thus based the regression off a database of only tigers and lions. The database for the formula is based off of 6 specimens, the equation:
log(mass) = 3.6775*log(femur length) - 7.2568
The 470 mm femur would have a mass of 371 kg accordingly.

The Ulna
Finally, I will go over the ulna in this short post. 


As I stated before, an ulna of 465 mm is certainly from a record breaking specimen. To predict the body mass without encounter false negative allometry, I again used a database of tigers and lions, with six specimens in total. The equation:
log(mass) = 2.8965*log(ulna length) - 5.1318

The R-squared value was .9, weaker than my other equations. This is because the tiger and lions are significantly different in ulna to body mass proportions, with the former being relatively heavier. Again, I assumed P. spelaea fossilis had a built between the tiger and the lion. The resulting estimate, 393 kg. Putting the ulna into perspective with the Ngandong tiger femur, this specimen probably had a femur of 480 mm as well, give or take. Its mass would be slightly less than the Ngandong specimen (as this specimen is classed as a member of the tiger species, c. 409 kg) again assuming it was not built like a tiger, rather midway between lions and tigers. 

I am looking into evidence to help figure the built of P. spelaea. Two things support it being very lion-like in built, if not synonymous, genetic data and robusticity of the bones, which fall into the range of modern lions. Furthermore, it is likely P. spelaea was morphologically lion-like as both species lived in similar, open landscapes, calling for greater cursoriality, explaining the relatively great width of the long bones.

end post"

 


 

Another old post of mine which also discusses size estimation methods and the derivation of my Ngandong tiger estimate. 

Exceptions to Isometry

I'd like to point out why the isometric approach is often times invalid. Data shows that not everything is isometric. Here is an example, based on tiger femur measurements and respective body masses:

We will estimate the mass of CN5698 (P. tigris, FL = 411 mm, actual mass = 230 kg) with the isometric method using 3 other specimens.
 
Estimations using isometry:
1. Specimen used for comparison measurements are FL = 408.5 mm and M = 225 kg. Applying isometry, the estimated mass is 229 kg.
2. Specimen used for comparison measurements are FL = 360.5 mm and M = 145 kg. Applying isometry, the estimated mass is 215 kg.
3. Specimen used for comparison measurements are FL = 341.5 mm and M = 115 kg. Applying isometry, the estimated mass is 200 kg.

I won't bother giving the total average, as that is not helpful to the point being explained. Note, the smaller the individual used for comparison, the smaller the estimated mass. This means that body mass grows at a greater rate than cubically in respect to femur length.

Now let's estimate the mass of small specimen with isometry, the specimen is labeled as CN5669 (P. tigris, FL = 341.5 mm, M = 115 kg).
1. Specimen used for comparison measurements are FL = 408.5 mm and M = 225 kg. Applying isometry, the estimated mass is 131 kg.
2. Specimen used for comparison measurements are FL = 360.5 mm and M = 145 kg. Applying isometry, the estimated mass is 123 kg.
3. Specimen used for comparison measurements are FL = 411 mm and M = 230 kg. Applying isometry, the estimated mass is 132 kg.

Note that now the larger specimens are giving an overestimation of mass when compared with the small specimen. Had body mass scaled to femur length cubically, this would not have been the case. Though, again, as body mass grows a greater rate than cubically in respect to femur length, larger specimens give overestimations when compared with smaller specimens. 

This is where regression comes into play. The scale factor (which is the slope of line of best fit if scaled logistically) represents whether or not mass grows isometrically (cubically). Logistically scaling and graphing the data produces the scale factor of 3.6865. Using the cube law, substitute this scale factor for 3, and note how the estimates become much more accurate. What I have just explained is part of the basis for the use of linear regression rather than isometry in recent documents. It is interesting to note that within species(not genus or family, thus why Christiansen and Harris (2005) formula had scale factors of less than 3) from bears to tigers, mass to femur length seems to grow at a greater rate than cubically. 

Finally, the full equation derived from the 4 specimens cited in this posted is as follows: log(body mass) = 3.6865*log(femur length) - 7.273.

Pertaining to our topic, this produces a mass of 409 kg for our largest specimen of P. t. soloensis. 

Also, a little fun fact, Panthera tigris soloensis roughly translates into "solo tiger," similar to how Homo erectus soloensis mean solo man. Reply