Freak Felids - A Discussion of History's Largest Felines

[GrizzlyClaws]

So that should be the rank for the prehistoric felids: Smilodon populator 500kg = Ngandong tiger 500kg > Cromerian lion 430kg > American lion 380kg > Cave lion 350kg Reply

[GrizzlyClaws]

And this spelaea specimen from San River, his skull measures 451mm and his upper canine teeth have exceeded 80mm from the skull, so the modern wild big cat that rivals the size of his canine teeth should be Madla.

*This image is copyright of its original author

Reply

[tigerluver]

The Cromerian/Mosbach Cave lion are the common names for P. fossilis. I guess it would be most accurate to say the 475 mm belong to the evolutionary transition of P. fossilis-spelaea, or a giant P. spelaea specimen. The cats living alone has evolutionary implications I will try to get into later.

The Smilodon puzzle is quite complex. Width estimates are exceeding 600 kg while length estimates sub 300 kg. 450 kg- 500 kg is probably a good range for the highest end. Nevertheless, I am working on a bone density study as we speak. Bone density is the ignored factor which explains the discrepancies between bone length and width and proportional heaviness. Hopefully by summer I'll work with the Smithsonian's fossil collections and get some bone mass data of Smilodon. If we can figure out how dense Smilodon's bones were, we can better assess which bone factor should be given more weight, length or width, and also which species would be proper for derivation of regression estimates. I probably sound a bit convuluted, I apologize. The bone density relationship is something very real but hard for me to articulate, but hopefully in time I'll find the best words. Reply

[GrizzlyClaws]

According to Guate, both 475mm and 451mm skulls belong to the transitional specimens.

While the 475mm one shares more characteristics with fossilis, and the 451mm one shares more characteristics with spelaea.

And the finalized version of spelaea is no larger than Panthera leo with the maximum recorded skull of 417mm. Reply

[WaveRiders]

tigerluver
 
in the last couple of weeks I took some of my spare time to read this thread as well as some parts of old threads in AVA related to body size and mass estimates of Pleistocene Ngandong tiger Panthera tigris soloensis.
 
I would like first to highlight that I appreciated a lot your commitment and I wish you all the best for your studies. However with the hope it could be useful for you I have to say that when scientific analyses are driven not only by state of knowledge, experience and science, but also by some conscious bias and a perhaps unconscious will to arrive at a particular result, these analyses can lead to erroneous considerations and erroneous results even by significant extent. History of science is full of biased analyses, so you should not be too surprised if you fell into this trap as I believe you did.
 
In the analyses I have seen you have been doing in the last couple of years or so and the ones in this thread I noticed a considerable number of conceptual mistakes. I took a list of them and I could go through all of them if you like.
 
You came to the so far conclusion that the 480 mm Ngandong femur tiger individual weighed 500 kg. That individual is estimated by all of us to have been approximately 2300 mm and 2350 mm at most in straight head-and-body length, meaning 100 mm and no more then 150 mm longer then the largest (2210 mm) and biggest from Dunbar Brander. This Bengal tiger was estimated by him to have weighed around 270 kg and I add perhaps up to approximately 290 kg empty for a very bulky tiger of that length.
 
Therefore for an individual of a Pleistocene tiger for which there is no actual definitive evidence it was overall build more massively then modern Indian/Nepal tigers you have added over 200 kg on top of the Dunbar Brander biggest tiger. Or that a tiger approximately 400 mm longer then the Sauraha male would weigh approximately twice as him at empty stomach (likely around 250-260 kg). It is a clear non-sense.
 
You believe such a tiger would be even more massive then a coastal Alaska brown bear as long as 2300-2350 mm in straight line in head-and-body length (for such a length they normally weigh less then half a ton in the Spring unless really huge in body build).
 
Your tiger would likely look as much massive or even more  massively build then him.

*This image is copyright of its original author

An important point is that I am a bit amazed in the last couple of years nobody really contested your results, which is a bit strange, particularly when AVA poster Ursus arctos middendorffi was around as I noticed he has some good knowledge in allometric regressions and general considerations. No contests resulted in the weight estimate of the largest Pleistocene Ngandong tiger fossil individual to inflate from the sensible estimates provided by myself and other posters at around 300-330 kg as it was the case in January 2012 when I stayed around AVA to the 500 kg suggested now, even more then estimates of 478.6 kg from Volmer (2005) and 470 kg from Hertler & Volmer (2007).
 
Furthermore I am very much amazed nobody including you do realize what was the body size of a Pleistocene lion with a 192 mm MT3 as well as the body size and massiveness of very many fossil lions including both Panthera leo fossilis and Panthera leo spelaea. There is no need to study the thousands of fossil lion remains as I did to understand that. I could list and provide you details of very many Panthera leo fossilis, Panthera leo spelaea and Panthera leo atrox bones of very large size. I also state that there is no evidence of higher cursoriality of Panthera leo fossilis and Panthera leo spelaea with respect of modern lions. My studies based on a very large amount of fossil remains suggest that the Eurasian Pleistocene lions were if anything slightly less cursorial then the modern lion. I am also not fully convinced of the significantly higher cursoriality of Panthera leo atrox with respect of modern lion as from conclusion of Wheeler & Jefferson (2009). Although possible I believe further evidence is needed for a firmer conclusion while for Panthera leo spelaea we already know quite a lot.
 
I can see that nearly any comments (not only yours) related to Panthera leo fossilis and Panthera leo spelaea are in the direction to push down their size and massiveness despite the enormous amount of evidence shown in literature suggesting the opposite. I noticed in particular in talks by GuateGojira, GrizzlyClaws and you the continuous dwarfing of Panthera leo spelaea following the most bizarre statements: transitional forms, final forms as large as modern lions or a bit more at most, any large bone or large skull is not spelaea but fossilis, the giant lion genes are lost and are not in the modern lion, etc. etc. I found all these comments very bizarre. The skull of ca. 475 mm and 451 mm lengths, the 470 mm femur are Panthera leo spelaea remains dated not earlier then MIS 6 (Late Saalian). They cannot be Panthera leo fossilis as this form ended with MIS 9 (Holsteinian Interglacial). All “transitional forms” (anything is evolving and is transitional) are Panthera leo spelaea and there are Panthera leo spelaea individuals even larger then the 475 mm skull individual. There are also quite large individuals even till at least the middle of the last Glacial.

I also remind people that there is no general consensus if Pleistocene lions were a different species then modern lions. Sometimes an author comes out with something and you take that for grant if you like it or reject if you do not like it. This is a naïf scientific approach, but I am not surprised. You should have talked or talk with Turner, Kurten, Hemmer and other eminent Palaeontologists before being so sure that the Pleistocene lion was a different species then modern lion.
 
One thing I state for sure is that apart from body size and body build Pleistocene lions show many more close similarities to modern lions then likely any other two felids compared (definitely within pantherinae). Pleistocene lions and modern lions are really remarkable similar in all compartments. Also the concepts of paleontological species and biological species are different and you should not forget that (for instance brown bears and polar bears interbreed and produce fertile offspring despite they split during the Middle Pleistocene and are different species).
 
Finally there is no common name for Pleistocene lions other then cave lion or steppe lion and American lion. How do you want to call them cave or steppe tiger (they would immediately gain 100 or 200 kg in that case)?
 
 
One more note.
 
Somebody stated I mentioned in AVA in January 2012 a ca. 440 mm skull as a Panthera leo spelaea skull. It is clear that I talked it is a Panthera leo fossilis skull and I told the story behind it. Any palaeontologist with carnivore and felid expertise could very easily understand I was talking of the famous Mauer skull, described by Wurm in 1912. The original greatest length of that skull (with an apparent slightly broken sagittal crest) was given by Wurm at 431 mm. Then Kurten came and provided that length at 442 mm (likely estimating the full length of the sagittal crest and not re-measuring the skull). This figure has been universally accepted. My greatest length estimate of that skull based on its consylobasal length or basal length are within less 1 mm from the length provided by Kurten. The Mauer individual was an average size Cromerian lion.

I also estimate the 465 mm ulna Panthera leo fossilis individual from Mosbach as an individual above average size but not particularly large.

The 470 mm femur Panthera leo spelaea individual was a large individual, but not very large. A Panthera leo fossilis individual with such a femur would have been just very slightly above average.

 
                      WaveRiders
 

  Reply

[tigerluver]

General points I took from your posts worth discussing:
You have a habit of taking estimates that fit your definition for the tiger's size as reality, and ignore actual weights at those greatest lengths and you know which ones I'm referring to. The seeds were planted in other threads, and that's for a different topic. Regardless, looks like you're in the trap with me, so let's burn some time while we wait for the crew to get us out. 

All long bones mentioned are nothing but above average. Untrue, take the 465 mm ulna for example, a 203 kg lion's ulna is 79mm shorter than the P. fossilis specimen, that is a huge difference in terms of mass. The same goes for the other large long bones of Pleistocene cats. On body length, my values are close to yours. You disregard isometric and allometric relations, that probably explains why none of the long dimensions produce a very large specimen in your mind. The MT3 was addressed as a risky size estimate due to the inherent nature of bone in relation to the body. But, considering regular Panthera proportions, my estimate was 390 kg, how does that sound? Considering the increasing robusticity, I'd say another 420 kg P. fossilis specimen joins the ulna. 

Next, I've already went over the evidence for great robusticity across Pleistocene species. Their bones are significantly more robust, does that mean nothing to you? You seem to have more data than the entire professional community, please bring some numbers on your bones in that contradicts databases like that of Christiansen. For references, refer to sources such as Brongersma, Dawkins et al., Christiansen, and your local museum. The bears of the time blew the massive fellow you posted out of the water in terms of robusticity as well.

I just pulled up an old file containing a plethora of bone dimensions. Here's a 436 kg bear of FL = 445.5 mm and least circumference = 126.25 mm, an index of 0.283. Modern mainland tigers have an index of 0.24-0.25. Considering the 8% more robusticity of prehistoric cats, that would put them on the bone level of at least this modern Grizzly (in no way am I saying that cats are equal to bears in robusticity at the highest level, as the highest level of bears are the short-faced species.)

Cursoriality is a difficult one. Our conclusions, or at least mine, have been explained on what basis they are throughout the thread. Looks like a published work agrees. What's your basis? Provide the list of the characteristics you found on "thousands" of fossils, not only your final conclusion. Maybe asking for the details may come off as a bit rude, but when the allotment of entirety of published data stemming from the 19th century comes nowhere close to that number, including Dr. Marciszak's full, thorough review of P. fossilis records, alongside the fact that fossils are scattered all around the world's museum, the claim is suspect with all due respect, even if that is/was your sole profession. We need data to progress as we emphasize over and over, so I ask. Regardless, the bones are robust enough compared to the modern species that cursoriality's negative allometry in terms of mass is offset. 

On P. spelaea/fossilis speciation. Why do you disregard more modern studies on their phylogeny and stick with older studies? Next, we all acknowledge that these two giants are of the lion clad, only that being a subspecies may not be the true relationship as new evidence indicates. If the Ngandong tiger grouping does not fit you, I agree a bit. It might have been its own species in the tiger clad, but the Sumatran tiger indicates there was interbreeding between the Sonda and mainland forms, indicating something more toward a subspecies/ecotype.

Where are we downsizing anything? We discussed transitional forms in the manner Sotnikova, Sabol, and Marciszak did. Your take of the evolutionary situation does not have to be agreed upon by the rest. 

When referecing AVA, please just restate the data, I wasn't around and in depth to the extent I am today back then. 

On models and bias. I am not sure where I can insert personal biases to a straight regression line produced by measured datapoints. Most data comes Christiansen, and a few my own, unless you think I'd go measure the museum records in a way that suits your view of my "biased" personality? I know I have not done that, and my supervisors would agree. I am not sure of what mathematical methods you use, as you have levels for your estimates. Mind shedding some light on that? Is it a regression with the levels indicating the slope's confidence intervals? No, something else? Where do you get your databases from. Releasing this info won't hurt you, as these sources are public anyhow. As I've stated before, most personal ones are from Smithsonian's NMNH.

On bias that you feel we are plagued with and you are much above. I'm sure both you and I would have underlying biases, it's human nature. Though, like we have, don't point fingers and ignore one's owns biases. Up until now, nowhere have I attacked your personality when we disagree, as we do most of the time. You on the other hand like to slide accusations of bias and a lack of brain to those who disagree with you, a bit eloquently to fit the nice writing style you have of course. If you'd like to continue, let's debate and discuss the ideas we present, not who we are. Also, let's not present our take of the situation as the final word and the most "sensible." I recognize your right to have your view and respect it, hopefully you can find it in your heart to respect our take. In other words, write a like a legitimate scientific paper , no adjective to describes the the quality of the idea proposed, refuted, or discussed (at least how they have us do it now). If this style does not suit you, I'll let you be you and move on, no hard feelings.



  Reply

[WaveRiders]

 
tigerluver
 
 
You wrote
 
 
“You on the other hand like to slide accusations of bias and a lack of brain to those who disagree with you”
 
 
I am sorry if this was really meant that way towards you. So far you did nothing wrong against me, you behaved correctly and I can say I definitely appreciate a lot what you are doing, your great interest in these topics and your commitments. You are still a student and it would be totally unfair to criticize a student badly because students are still in a fast learning process and I have no titles to do that in a contest of a forum. I have also an advantage because you disclose nearly every detail of what you do while I do not for a number of reasons.
 
Although there could have likely been an initial impulse coming from a slightly different motivation, my major aim in raising possibly serious doubts about your weight estimate of the 480 mm femur tiger individual from the Pleistocene of Ngandong was and is in hopefully trying to be helpful to your studies following my knowledge and understanding. If this is retained arrogance by somebody nicknamed WaveRiders, and it could be in principle, please consider that Per Christiansen wrote in an e-mail to GuateGojira (or somebody else) dated a few years ago “no tiger was ever 470 kg that is absolutely certain”.



You wrote

“please bring some numbers on your bones in that contradicts databases like that of Christiansen”
 
 
Apart form a few data inconsistencies I do not definitely contest or have contested Christiansen database like it seems to me you you did stating that some tigers are underweight and a lion is overweight and I do not really see where I contradicted Christiansen. At most in some considerations which I have not touched here so far and in the methodologies I developed I could say that I believe to have gone farer then him. If this is actually the case this would not be unusual in research. If I could do it 10 years ago it has also been thanks to the work Christiansen performed between 10 and over 15 years ago. Christiansen did not develop further his methodologies after that as he concentrated on other topics, but he could have easily done if he wanted.
 
You are actually right as I have a really voluminous database in many areas, but not in all of them to the extent I would like of course. There are still margins of uncertainties, sometimes small, sometimes that could be substantial and need experience, careful and expertise to manage them properly hopefully. After all in most sciences nothing is really certain and nothing can be really proved as certain. Making research is a process that needs to be learnt with time. There is the need to open the mind, focus on details that can reveal a lot and bring illuminations and be as less biased as possible. Sometimes one can have things under his nose and does not realize he has.
 
If I raise doubts and force you to check your procedure and find in case some inaccuracies I got my target. The estimate of the 480 mm femur individual at 500 kg is really my major issue and unfortunately it casts possibilities we are both biased in opposite directions. I can understand that, but from my side I can say that I retain myself rather neutral or should I be retained biased because I firmly believe that no tiger was ever close to that weight as also Per Christiansen firmly believes?
 
 
Now let’s move on hopefully constructively for our mutual benefit
 
 
1)   post # 27 and post #108
 
You wrote
 
“You have a habit of taking estimates that fit your definition for the tiger's size as reality, and ignore actual weights at those greatest lengths and you know which ones I'm referring to.”
 
 
I definitely reject the weight of the Hasinger tiger even accounting for a full belly and would never include it in any database to build allometric regressions like you did. I also doubt the length measurements of that tiger as the skull is ridiculously small (14 x 9 inches) for its alleged total length (11 ft 1 in along curves and 10 ft 7 in between pegs).
 
The fact that I refuse to trust all 11 footer tigers along the curves but a few exceptions and all 12 footers along curves do not make me feel at all uncomfortable.
 
Overall I do not feel trapped in underestimating tiger size and weights. Furthermore including extreme unreliable data can screw up a database and any statistical analyses based on that to an extent depending how much this database is small and spread.
 
A database should be randomly taken and made by as many as possible accurate and reliable data hopefully with no outliers. If the datapoints for any reason are undervalued in the low range and overvalued in the high range or viceversa the slope of the regression, most affected from datapoints at the its extremes to an variable extent depending from sample size and data spread, may result to be significant overestimated or underestimated.
 
I am not convinced the tiger database you used in post # 27 is the best possible as there are many ore weights vs total lengths available from literature that would allow you to obtain a more realistic regression slope. Just to mention one among the very many more data available for instance you did not use the tiger shot by Gordon (3226 mm and 222.7 kg) and you instead used the tiger shot by Hasinger (3226 mm and 388.7 kg) which is clearly a not realistic weight (a bias?).
 
In my opinion a somewhat biased database is the reason why you obtained a significantly higher then expectable allometric coefficient of Body Mass vs Total Length of 3,9878. You estimated the 480 mm femur individual to be 347 cm in total length and the body mass based on total length therefore to be 446 kg.
 
 
2)   You wrote
 
“All long bones mentioned are nothing but above average. Untrue, take the 465 mm ulna for example, a 203 kg lion's ulna is 79mm shorter than the P. fossilis specimen, that is a huge difference in terms of mass. The same goes for the other large long bones of Pleistocene cats. On body length, my values are close to yours. You disregard isometric and allometric relations, that probably explains why none of the long dimensions produce a very large specimen in your mind.”
 
 
The fundamental aspects to consider for a correct application of allometric scaling in zoology are theoretical, mathematical, biomechanical and database related.
 
                                                                                           
Below I will discuss a few among the considerations you did that in my opinion need to be addressed differently.
 
 
2A)
 
Humerus, femur and tibia lengths in Christiansen & Harris (2005) are articular (functional) lengths and not greatest lengths. Greatest length is only the one used for ulna. In the case of tiger femur the articular length in a few cases may be about the greatest length, but normally is marginally smaller. Humerus and ulna articular lengths are generally smaller then greatest lengths.
 
All weight estimates using equations and database from Christiansen & Harris (2005) that I have seen done by you and GuateGojira so far are therefore conceptually not correct as you always used greatest length (the error is generally however marginal). Christiansen (1999) regressions instead used greatest lengths of limb bones.
 
 
2B)
 
Humerus and femur distal widths in Christiansen & Harris (2005) are articular widths (mediolateral width across condyles) and not epicondylar (greatest mediolateral) widths.
 
All weight estimates using equations and database from Christiansen & Harris (2005) that you have done so far using epicondylar (greatest mediolateral) of fossil specimens are therefore conceptually not correct.
 
 
2C)   Your post #1
 
“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”


 
It seems that you are convinced that building an allometric relationship mixing individuals of different species can prevent a size-related bias when a species average allometry create it. Well, as Christiansen and Mazak suggest, I also advice not to mix intraspecific and interspecific allometry using single individuals even if taking care to the same number of individuals for each species like you did for the femur and the ulna. The fundamental reason is that it is conceptually wrong, can lead to substantial errors and may not be necessary for what you want to achieve.
 
 
Condylobasal Length of Skull
 
Using Panthera leo, tigris, onca, pardus, uncia species average (neglecting Neofelis nebulosa) you got a body mass vs condylobasal regression nearly identical to the one you used. The two equations becomes basically the same enlarging the Christiansen database. For the 484.7 mm skull from Chateau you calculated 387 kg (myself with your equation 372 kg likely due to a smaller estimated CbL). while the species average gives (using CN5698 BM = 220 kg) 365 kg (allometric scaling exponent 2.63792) to 385 kg (allometric scaling exponent 2.75170) depending from the database.
 
The Mauer 442 mm skull individual was pretty much an average size Cromerian male lion I estimate at around 310 kg. There is no much question on its about average size considering the size margin given by the 465 mm and the 484.7 mm skulls from Chateau. That individual has generally been retained an Upper Middle Pleistocene lion of large size even because since 1912 and for over 60 years it has been the only Cromerian lion skull relatively complete unearthed (to date I am aware there are only 5 of them of which 3 actually substantially intact and complete).
 
My body mass estimate for the 484.7 mm skull from Chateau is around 355 kg.
 
 
Femur
 
For the femur you obtained a very positive allometric coefficient of body mass vs length of 3.6775 using 3 P. leo individual and 3 P. tigris altaica individuals. Inevitably you made a choice and you introduced a bias related to the individuals you chose. Assuming P. (leo) spelaea spelaea had a built between the tiger and the lion you therefore obtained a body mass of 371 kg assuming 470 mm is the articular length (which conceptually is not although very close) or nearly a couple of % less estimating the articular length.
 
Considering instead species average allometry using P. leo, P. tigris tigris and P. tigris altaica the allometric coefficient results to be 3.01601 for CN5698 body mass = 230 kg or 3.08298 for CN5698 BM = 220 kg. The body mass estimate for the P. (leo) spelaea spelaea 470 mm femur results to be 318 kg or 311 kg respectively based on femur length only and considering the articular length estimated from the greatest length.
 
The Panthera (leo) spelaea spelaea 470 mm femur individual from Mladec Cave was a male individual that within the spelaea form as a whole spanning over 300 KY is rated by me to be no more then of large size. Following biostratigraphy analysis and looking at him under the perspective to actually likely be a Late Pleistocene Eurasian Panthera (leo) spelaea spelaea it could then be classified as a very large male. With my methodology my body mass estimate for the 470 mm femur individual is 295+ kg.
 
Within the spelaea form there are quite a number of individuals larger and even significantly larger then the 470 mm femur individual. I estimate its greatest skull length around 445 mm, therefore likely nearly the size of the San River Skull (451 mm), but significantly smaller then the massive ca. 475 mm Ural skull (CbL = 422.0 mm, BM = ca 320 kg, HBLstraight approaching 2350 mm). The largest spelaea individuals I am aware would have had a greatest skull length likely in the range 490-500 mm with a couple of them seemingly even up to around 510 mm and head-and-body length straight around 2450 mm (it must be considered that the number of specimens of Panthera (leo) spelaea spelaea unearthed is much higher then any other fossil lion form, particularly of fossilis and vereshchagini).
 
No Panthera (leo) atrox fossil individual unearthed so far that I am aware reaches this size, being the largest and heaviest the 467.5 mm skull 14001 (CbL = 424.3 mm), approaching 2350 mm in straight head-and-body length and ca. 320 kg in body mass.
 
 
Ulna
 
For the ulna you obtained a nearly isometric coefficient of body mass vs length of 2.8965 using again a bias choice of 3 P. leo individual and 3 P. tigris altaica individuals and you obtained a body mass of 393 kg assuming again that P. (leo) spelaea fossilis had a built between the tiger and the lion. In your last computation (post #94) and assuming isometry with modern P. leo from database of Christiansen & Harris (2005) and applying the robustness index you calculated comparing the fossil ulna with a single modern ulna you obtained 347 * 1.0837^2.5 = 423 kg.
 
Considering instead P. leo, P. tigris tigris and P. tigris altaica species average the allometric coefficient of ulna length results to be 2.50902 for CN5698 body mass = 220 kg and a body mass estimate of 340 kg.
 
My methodology estimates the 465 mm ulna P. (leo) spelaea fossilis individual body mass to be around 325 kg.
 
You retain the 465 mm ulna individual a huge one, while among Cromerian male lions I retain it an individual above average size but not properly considerable of large size, definitely not very large or huge. I estimate the greatest skull length of this individual to have likely been 455+ mm and possibly approached 460 mm, but unlikely more then that. Therefore in my opinion it was likely slightly smaller in size then the 465 mm skull from Chateau, significantly smaller then the other skull from Chateau measuring 484.7 mm and much smaller then the 192 mm MT3 individual. However without a doubt the bone is quite more robust then an ulna of a modern lion.

 
2D)   192 mm Metatarsal 3
 
In post #89 you wrote
 
“On the MT3 reported by WaveRiders of 192 mm of P. fossilis, in accord with the ratios published by Day and Jane (2006), that specimen would be of 115 cm at the hip and 97 cm at the shoulder, assuming lion built, and a bit shorter assuming the other species' builts. Day and Jane (2006) use a pictoral method for their data, so there might be room for error as those numbers look too small to me. Furthermore, higher MT/Femur ratios is directly correlated to increased cursoriality, another piece suggesting how cursorial P. fossilis was.”
 

It seems you forgot to include the scapula contribution for the shoulder height and the distance from the acetabulum to the top of the ilium in its anatomical position for the hip height. Considering the anatomical position of the forelimb and the scapula, the skin and pad thicknesses and the proper contribution of the forefoot an Eurasian Pleistocene lion individual with a 192 mm MT3 should have had a standing shoulder height between 1350+ and 1400 mm, a head-and-body length in straight line most likely pretty much around 2500 mm and a greatest skull length most likely around 525 mm -10 mm / +5 mm. In spite of his huge size I estimate such a monstrous Cromerian male lion to have weighed in the region of 400 kg (at empty stomach of course).
 
 
2E)   Assumed Outlier
 
To build the BM vs Femur Length intraspecific regression for tigers you took out the presumed oddity tiger datapoint CN5697 (femur length 429.5 mm, body mass 230 or 220 kg). You assumed it is an outlier following the fact that the remaining 4 datapoints are well correlated while CN5697 substantially deviates. In my opinion in this case removing this presumed outlier datapoint contributed to lead you to unrealistic results as the regression departs from slight positive body mass allometry to femur length (exponent 3.14751 using 230 kg or  3.08233 using 220 kg) to an hardly justifiable significant positive allometry of 3.68654. It is a quite hazardous to remove a data in a sample of only 5 datapoints with the result to get farer and quite far from isometry and assume that having a sample of well fit 4 individuals the intraspecific allometry is more realistic. If you would have got closer to isometry instead of quite farer, removing an outlier could have cautiously been acceptable. The point is that CN5697 femur is not retainable a genuine outlier.
 
Furthermore that body mass of the other individual CN5698 is given by Christiansen in two papers at 220 kg and in other 2 papers at 230 kg (evisceration does account for nothing in this case as the difference would be much higher and the year of issue of the papers seems not correlated to one figure or the other). Therefore there is no evidence that 230 kg must be the correct figure and not 220 kg. Although the difference is marginal using 220 kg for CN5698 would make the suspected oddity CN5697 less odd.
 
Anyway, why you convinced yourself that the scaling exponent 3.68654 should be more correct then 3.14751? I found out and below is I reconstruct the procedure you have gone through.
 
I have seen that in AVA thread “Body size of the Ngandong tiger (Panthera tigris soloensis“ you were using power law instead of the logarithmic transformations that you are using recently.
 
In your post #346 sent on Fri, Jul 19, 2013 1:28 AM of thread “Body size of the Ngandong tiger (Panthera tigris soloensis)” in http://animalsversesanimals.yuku.com/top...PNmFS5mr5w pag 18
 
You wrote the comment
 
 
“Significant error, significant underestimation for both lion and tiger data”
 
 
after applying the formula you got from body mass vs femur length regression for tigers on the tiger dataset used to get the regression as well as to the lion individuals. The formula is shown in the graph you posted and it is Y= 1E-06 x^3,1475. I checked your tiger individual BM estimates and it resulted you actually used exactly that formula.
 
Being unsatisfied of the results you then removed tiger datapoint CN5697 and you found that with the new formula Y= 5E-08 x^3,6865 you got better results when applied on the database used to get the regression (which is a procedure to be done to get statistical parameters of accuracy of the prediction such as PE% and SEE%).
 
For this latest regression, you wrote
 
 
“Overall, not too bad for lions and tigers. The equation supports well the little data we have when applied only to lions and tigers. Now, is it a safe bet to extend it over to the larger prehistoric felids? This is open to discussion.
Interestingly, neither equation applied properly to jaguars. The tiger equations gave severe underestimates.
So, questions, comments, corrections?”
 
 
Now it comes the correction from me with the aim to be useful for you in the future as nobody told you that (apparently even your professor).
 
 
A much more correct formula using 5 datapoints (and 230 kg) then the one you wrote Y= 1E-06 x^3,1475
 
is
 
Y= 0.0000012738 x^3.1475073474
 
 
How I got it? Simply using 10 decimals in the Excel graph instead of the 6 only that you used.
 
It is immediate to understand that depending from chosen units and figures involved it is sometimes necessary to use a very high number of decimals above all when non log-transforming. Using only 6 instead of the 10 I showed you (much more correct) in this case you underestimated body masses by 27.38% (coefficient of 0.000001 instead of 0.0000012738) !!!
 
 
Using BM = 220 kg the formula for tiger femur length would be BM = 0.0000018622 FL^3.0823275490
 
 
With 6 decimals you would have had an overestimation by about 13.78% as your constant would have appeared in the Excel graph as 0.000002.
 
In recent posts I noticed you started to use logarithmic transformations. Data transformation are performed for inference statistics. Logarithmic transformation “linearizes” datapoints and improve homoscedasticity to apply statistics linear analysis methods. You also got rid from those small numbers and avoided the aforementioned kind of problem, but you did not check your previous results and stick to not consider the presumed oddity tiger datapoint CN5697 keeping the scaling exponent of 3.68654 which is far too high (either 3.08233 or 3.14751 appear to me a more accurate solution within the tiger database from Christansen & Harris, 2005).
 
 
2F)   480 mm Pleistocene Tiger Femur from Ngandong

Your posts # 72 and #96
 

“I've more indicators on the sheer robustness of P.t. soloensis after some more research. For one, a value I did not discuss that was published by vK himself was the AP diameter of the head. It was 59 mm. That is extremely massive. Let's put it into perspective. P. leo and its clad (P. atrox, P. speleaea, P. fossilis) is at least 10%-20% relatively more robust than P. tigris in the proximal end of the femur (Dawkins et al. 1866). An exceptionally robust P. atrox specimen had an AP diameter of head of 54.3 mm and length of 455 mm, an index of 0.119. The Ngandong tiger specimen had an index of 0.123. Considering how much more slender tiger bones are, the Ngandong tiger proximal AP diameter is absurdly thick. This published AP diameter of the head also supports my personal extropolated DAW and greatst proximal breadth, as a 59 mm diameter can only be of a bone with such dimensions (beyond the fact that vK actually differentaited between DAW and distal diameter in his book). But in terms of mass estimation, I don't have enough specimens to produce a good trendline, so no number will be produced by the AP diameter of the femoral head. Another measurement published by vK was the diameter of the distal notch of 23 mm. The DAW scales about 4.5-4.8 times the notch diameter, within the range of my DAW measurement of 107 mm.
 
On a good note, I was able to find a third comparitive diameter of the Ngandong tiger femur, the lateromedial diameter of the midshaft. Christiansen and Harris (2005) list this diameter at the point of least circumference. The word "least" is a bit misleading. Egi (2001) clarifies the point of measurement, which is the 1/2 point of the length of the bone (AKA the midshaft) in femur and tibia and the 55% length point in the humerus as follows:
 
Thus, I extropolated the third Ngandong femur measurement compatible with published databases, LM diameter of the shaft. The measure was 42 mm, much more robust than modern tigers and a bit shy of the most robust P. atrox femur.
 
Thus, from the measurements of the femur length (480 mm, est. 409 kg), DAW (107 mm, est. 531 kg), and LM diameter of shaft (42 mm, 559 kg), the calculated mass of this specimen is 500 kg.”
 
 
The picture of the 480 mm femur is much likely distorted because the minimum diameter of the femur results to be close to the proximal epiphysis while it should be around mid-shaft and the distal epiphysis appears much larger then the proximal epiphysis. It seems to me that the plane of the bone is not perpendicular to the sight direction. There might also be a minor rotation along its major axis.
 
The supposed proximal and distal widths based on reconstruction from the picture assuming the 480 mm length is correct should not be considered as in my opinion is a significant speculation. The distal mediolateral width you estimated at 107 mm is far above any regression of modern tiger and Pleistocene tiger femora giving a robustness with no corroboration from any other bone either femur or anything else.
 
I therefore retain your estimate of distal width 107 mm and minimum mediolateral diaphisis width of 42 mm the result of a speculation on a picture assuming that the length 480 mm is correct. As I pointed out above in my opinion the proximal epiphysis is farer from the observer then the distal epiphysis. While for sure there is something wrong in either the angles of the pictures or in the length or widths, first of all it could be the length. Considering either modern tigers femora and other Pleistocene tiger femora solidity regressions, the distal width provided by Koenigsvald (1933) at 88 mm matches pretty much a femur length of ca. 430-440 mm and with reasonable probability up to 460 mm. Similar trend is considering the proximal epiphysis mediolateral diameter given at 94 mm (estimated femur length 415-450 mm and with reasonable probability up to 460 mm). Considering the other dimensions (proximal sagittal diameter of 59 mm and distal sagittal diameter of 82 mm) that femur appears of build lighter then average, particularly as a male, although within the typical solidity range of modern tigers (therefore not an outlier).
 
My conclusion is that while I definitely do not rule out that the length figure of 480 mm or one or more diameters might be wrong, till updated info for the large 480 mm femur from Ngandong will emerge I would stick with all the measurements provided by Koenigsvald (1933) meaning femur length of 480 mm, proximal sagittal diameter of 59 mm, distal sagittal diameter of 82 mm and distal mediolateral width of 88 mm. The tiger individual owner of that 480 mm femur should therefore be retained as a very large but not heavily build individual. For sure other individuals as large and larger and bulkier then him existed, but that particular individual was not and it must not be forgotten that the weight estimate must refer to that particular fossil individual.
 
I still retain that there is no clear evidence the Ngandong Pleistocene tiger form was any more robust then modern Indian/Nepal tigers if possibly marginally in the forelimbs (see my comments in AVA in January 2012). In the Pleistocene lions compared to modern lions this feature is very much evident.
 
I currently estimate the 480 mm Ngandong tiger femur individual to have weighed around 315 kg (most likely range 300-335 kg) with a head-and-body length of 2300+ mm (up to 2350 mm) and a standing shoulder height of around 1200 mm. There is no question that the occurrence of such a fossil remain much larger then the average within a very small sample of various kind of cranial and postcranial remains is a very high unusual circumstance that would require much deeper investigation.
 
 
 
I do realize my body mass estimates could be retained underestimates. Perhaps some or all of them are like that by 10% or something and even 20% in a few circumstances and they are therefore pretty much stable for quite a number of years. I do feel comfortable with my methodologies and I am quite confident in what I have done despite the inevitable benefit of doubt and the only thing I really know for sure being that they can be improved.
 
 
I hope my discussion and observations have been useful to you at least to force you to reflect. This is what I really would like you to do it. I normally do it when something does not match my thoughts. I have no ambition and/or expectation to illuminate you and this is also the reason why I refrain to open my big book however worth it can be. It is going to be up to you to build yours, but I can guarantee you that once you will do it by your own you will have greater satisfactions.
 
 
I may come back to cursoriality and speciation at a later stage.


                                 WaveRiders
 

  Reply

[tigerluver]

On tiger lengths, the 388 kg specimen was not used in the length estimate in the final table. By greatest length, I was referring to the 305-315 cm tigers that were around 300-320 kg. Regardless, I concluded a while back that long bones aren't that well correlated with total length, and one can't be certain of the bone robusticities of those specimens, so reconstruction estimates were discarded. Pleistocene cats differ a bit too much.

I see you have looked through the old discussions on yuku. The old equations are outdated at this point, including the early ones here. For lions, at the time, I only had the three poorly correlated datapoints of Christiansen and Harris. Same for tigers. Using tigers with lions does put some false allometry (overestimate lions, underestimate tigers) if you are strictly applying the equation to only tigers or only lions. The same false negative allometry occurs when you use jaguars with lions when applying it to a certained species (overestimate jagaurs, underestimate lions). Cave lions were given the benefit of the doubt in their proportion, and a midway was assumed. All the equations I use now are slightly different and also log scaled (power approach causing asymptotic issues), thanks to the newer database, so smaller species don't have to be relied on (as they can cause strong issues with allometry, as Anyonge and Christiansen showed). The outlier was removed after email contact with a staff at the museum regarding the specimen's health at death, when it was weighed. The 220 kg tiger is actually 230 kg, confirmed by future works of Christiansen and another author (Campione, 2012) who posted a huge database on animal femurs and humerus. I can attach that file here if people would be interested. Other than that, not much to be discussed, no point in going over what's outdated and weaker in power. Points I did not address about the equation were already addressed, I had the same doubts you did and fixed them. Not every post's equations and estimates have been updated here. The equations at the start of the thread were copied from yuku and not altered. Once things get locked into publishing, I can't update equations anyhow for obvious copyright reasons. This thread and old threads are a timeline of the evolution of our understanding of paleobiology, with like science, recent posts applying changes to previous status quo.

Day and Jane's equation was based on the animal in the flesh, so all that you have brought up had been accounted for. What basis did you use your body length estimation, as based on published data, 2500 mm is a bit of an overestimate. Here I could slide in a "bias" accusation as you did again a few times, but I'd rather leave space for explanation. Lastly, the point goes back to the high variability of MT proportions. If long bones are not a great predictor of body length, MTs and MCs are even worse, being one possible explanation to why published data does not match your estimate.

What build would you say cave lions were? Tigers are proportionately heavier than lions. I used long bone ratios (they reflect proportional weight) and late cave lions match lions exactly, i's here on this thread, but I haven't given the new regression estimates, which result in a bit lower masses, with lions only due to the doubts. In the latest posts, I kept it to isometry off lions for simplicity.

Another note, bones closer to the body (femur, humerus) scale more positively allometrically in terms of length than bones further away (ulna, tibia, etc.). Closer bones are more incorporated in the core body, and maybe that explains correlations. Lower bones could simply elongate for cursoriality purposes and not robusicity purposes in the taller species, explaining the negative allometry. Turner's table might support this. Epiphysis diameters also scaled quite negatively while shaft diameters positively.

On articular and greatest length. Christiansen's humerus length actually included the entirety of the femur, under the assumption that the protruding spur of the distal humerus is articular as most definitions says, test from his images yourself. Femur, tibia, etc. articular and greatest lengths are also the same, the articular surfaces are at the official ends of such bones. Ulna length had to be differentiated due to the nature of the bone itself, as you know, and is why Christiansen took the time to go over the ulna. All in all, his lengths are greatest lengths, refer to Christiansen 1999b for the assertion from the man himself.

On the tiger's femur bone itself. I also recognize the distortion possibility. The bone is straightened in the manner vK measured it, the condyles are even. All the doubts you have stated are reasonable, I've had the same ones before I posted anything. So here's how I tried to address them. I tested distortion out by measuring myself other pictures in vK's book with measurements he published, and everything matched very well with ratios being equal to the hundredths place , thus distortion did not look to be there. The femoral thinning above midshaft from this view is actually seen in some tigers, such as the Trinil femur. Albeit, this is not considered the point of least circumference as the posterior side is actually more robust there. A greatest LM width of 94 mm for a femur of 480 mm would require the lengthening the femur vK's photo shows to an irrational amount. Plus, as long as the item is in the center of the frame the distortion is left out to the far sides of the frame. Christiansen used photos to measure as well with likely the same methods, so that would balance that out as well. Albeit, my measurements are measured by hand, but I don't think this would be a problem as I've just described. Next, vK actually differentiated between diameter and greatest width in his book from proximal/distal. Furthermore, I went into Dawkins to extrapolate vK's measurements. They were at odd places, and not the greatest width whatsoever. Actually Dawkin's himself measured cave lion bones awkwardly too, where proximal and distal diameters were smaller than they would have been if it was DAW, making it look like they were thinner than modern species. For example, a 422 mm femur had LM proximal width of 76.2 mm. Such a value cannot be of greatest proximal width on a lion bone of that length, something around the 100 mm is more realistic. vK seems to have followed suit. Next, Christiansen's DAW scales to equally the distal frontal breadth as long as there's no spurs and frontal plane widening as some femurs have, especially in cats due to the nature of the curvature. You can test it on pictoral femurs such as in Brongersma or yourself or on any femur of any species without a spur or frontal widening (occurs in some species like humans).

Supplemental picture straight from Brogersma, as the wording may not have explained everything clearly:

*This image is copyright of its original author

On the proximal widths of the tiger femur. A LM proximal width of 94 mm cannot have a AP diameter of 59 mm in any cat or bear, it's impossible. Proximal AP/LM is in the low .40s for bears and high .40s for all cats, including P. atrox. The ratio itself is wrong and one must remember tigers have thin bones compared to the rest of Panthera, in both the LM and AP angles. 59 mm to the 110 mm shown in the picture is at the high end of cat range, makes more sense. The distal diameter vK gave of 88 mm cannot be DAW either, as explained by the notch width he also wrote on. Lastly, the proximal, distal, and midshaft widths are very similar to that of the modern Javan tiger. 

On that note, if I did not address something, please let me know. I spend a good hour plus writing this and your posts are mighty long (not a bad thing by any means), and it makes it easy to miss words. Also, we have continuously answered your questions, if you could please respond to our's as well. Especially, what is your mathematical model, as based on your Crater example, it seems more prone to interpretation than regression. Again, there is no way giving this general information can hurt, especially, as it looks to be, there is no expectations of getting anything published. You must admit, a large portion of the data you have found, as in the case of all of our members here, has come from this community. Collectively, posters have put the entirety of sources together and have all corners covered.

  Reply

[tigerluver]

Two very interesting papers:

On sexual selection in sabertooths: http://journals.plos.org/plosone/article...ne.0072868

On extinction of the great North American felids: http://journals.plos.org/plosone/article...ne.0052453


 

  Reply

[WaveRiders]

 
1)         Christiansen database
 
“The outlier was removed after email contact with a staff at the museum regarding the specimen's health at death, when it was weighed. The 220 kg tiger is actually 230 kg, confirmed by future works of Christiansen and another author (Campione, 2012) who posted a huge database on animal femurs and humerus.”
 
 
You have to ask yourself why Christiansen used what you believe is an outlier. It seems that you have a animal stereotype, tigers, that all must be bulky otherwise there is something wrong with them. It also seems that you assume Christiansen has no way to judge the range of individual variability in tigers. Statistics cannot introduce bias to aim generating realistic and sensible results. It also seems sometimes you used that individual in your database (body mass vs condylobasal length of skull) other times you did not (body mass vs femur length) seemingly by judging how well the data fits the regression.
 
I stick with Christiansen consideration to consistently include tiger CN5697 in my database.
 
Concerning the tiger CN5698 220 kg / 230 kg issue
 
Christiansen & Harris (2005) uses 230 kg
Christiansen & Adolffsen (2007) uses 220 kg
Christiansen & Harris (2007) uses 220 kg
Mazak, Christiansen, & Kitchener (2011) uses 230 kg
 
Campione (2012) felid data have been taken by Christiansen & Harris (2005) database. I bet he did not bother to check any inconsistencies as he should have done for several more among the 255 animals and 1020 data he included as he took part of his data from existing literature.
 
It therefore seems to me that 220 kg or 230 kg is given on a random basis. There are a few other inconsistencies of Christiansen database

 
 
2)  Shoulder height
 
“Day and Jane's equation was based on the animal in the flesh, so all that you have brought up had been accounted for.”
 
It seems you have not followed me. The definition of the forelimb in Day & Janes (2007) does not include the scapula. Let’s do the math and the add the scapula you forgot to include..
 
From Day & Janes (2007)
 
Lion Forelimb %
 
Humerus -  37.4%
Radius 34.9%
Metacarpal   15.0%
Phalanges  12.8%
 
Lion Forelimb %
 
Femur   36.4%
tibia   37.3%
Metatarsal   16.7%
Phalanges   9.7%
 
Lion Hindlimb / Forelimb %   =  1.19% (1.1884%)
 
 
For a 192 mm MT3 fossil lion and assuming fossil lion had the same limb proportions of modern lions you calculated in post # 89 a hip height of 115 cm (192/0.167) and a shoulder height of 97 cm (1150/1.1884=968 mm) with a comment that those numbers looked too small for you.
 
I raised the point that you forgot to include the scapula as a contribution to estimate the shoulder height and the vertical distance between the acetabulum and the top foremost point of ilium to estimate the hip height.
 
I can tell you that a scapula of a such a huge fossil lion was likely to measure between ca. 390 mm and ca. 420 mm with 407 mm a likely best estimate.
 
Therefore 968+407 = 1375 mm
 
In reality it should be considered the anatomical position of the forelimb and scapula, the fact that longest phalanges should not be considered in length but in thickness, the pad thickness and the skin thickness. At the end a reasonable calculation gives a range 1350-1400 mm making therefore the 1377 mm figure pretty much a good estimate as a standing shoulder height.
 
I have to say that the above calculation following Day & Janes (2007) producing what in my opinion is a sensible shoulder height estimate accounting for the scapula is more the fruit of circumstance then anything else because the limb ratios % given by Day & Janes (2007) are approximated as based on stickers. For instance if you have to calculate the crural index in each felid species the results would be completely wrong as the tibias % appear consistently much longer then the respective femur % but in the ocelot suggesting the position of the knee sticker (or perhaps of the hip sticker) does not reflect the actual osteological ratio between femur and tibia. Should this have happened for the metatarsal sticker the above calculation could have produce a hindlimb, a forelimb length and a standing shoulder height rather different then 1150 mm, 968 mm and 1350-1400 mm respectively.
 
My model does not use Day & Janes (2007) data. I believe it is as much rigorous and accurate as reasonably possible. Yet I obtain a result nearly identical of 1375 mm, but honestly I did not even care to compare them.
 
 
3)   Metatarsal
 
“What basis did you use your body length estimation, as based on published data, 2500 mm is a bit of an overestimate. Here I could slide in a "bias" accusation as you did again a few times, but I'd rather leave space for explanation. Lastly, the point goes back to the high variability of MT proportions. If long bones are not a great predictor of body length, MTs and MCs are even worse, being one possible explanation to why published data does not match your estimate.
 
……
 
Another note, bones closer to the body (femur, humerus) scale more positively allometrically in terms of length than bones further away (ulna, tibia, etc.). Closer bones are more incorporated in the core body, and maybe that explains correlations. Lower bones could simply elongate for cursoriality purposes and not robusicity purposes in the taller species, explaining the negative allometry. Turner's table might support this. Epiphysis diameters also scaled quite negatively while shaft diameters positively.”


 
I know very well the individual variability of metapodials and the fact that they cannot be as accurate predictors of body size and mass epipodials and propodials. The foot anatomy and the load sharing among several small bones makes managing metacarpals and metatarsals quite tricky. Humerus and femur are generally the best limb long bone body mass predictors as they are single weight bearing bones. Radius and ulna for instance share loads and are generally not as good as humerus. Tibia is generally a predictor comparable or nearly comparable to the femur as the fibula takes a minor part of the total dynamic and static load in the distal hindlimb. No doubt that there is a significantly greater lever of uncertainty managing metapodials. Also a matter is estimating body size, a quite different and more complex matter is estimating body mass.
 
As far as I know there has never been an attempt to estimate body size and mass of the 192 mm MT3. Therefore I cannot see how published results contradict mine on this individual. The fist time the 192 mm appeared in published literature after the original report has been in Marciszak et al. (2014). I mentioned about this huge fossil in AVA on January 2012 without providing any detail on the original source and the kind of bone it was. Body size estimates of fossil lions found in literature are based on other bones and not on the 192 mm MT3. You assume any scientist knows any bone at any time of publication which is far from being true. Fossil felid body size estimates are traditionally generally simplified or relatively simplified and I found flaws in some of them. Ten years I went significantly more in depth and complexity in body size and mass estimates and although there is always room for inaccuracies and mistakes I trust my models more then anything else.
 
Make sure I have a voluminous database suited for the purposes of my studies, I accounted for any allometric phenomena and anything is done at professional level within the theory I am aware and the database I have. I have studied allometric scaling and allometric patterns very much in depth and since long time.
 
I estimate the straight head-and-body length for the 192 mm MT3 Cromerian lion to be most likely around 2500 mm. Obviously the level of uncertainty for a metapodial makes the possible range quite wide and depending by the model and confidence level on a statistical basis it could roughly be as low as 2300 mm or lower and as high as 2700 mm, whose upper side I obviously retain it totally unrealistic in practise. Even accounting for a diminutive scenario the size margin of a Pleistocene lion 192 mm MT3 is very high, and there are several other lion fossils confirming the size of that individual was not a genuine outlier. Even considering 2400-2450 mm it would still be a very large individual.
 
In my opinion HBL_straight around 2500 mm is not an overestimate due to bias and is in line with my estimate of  2300-2350 mm for the 480 mm Ngandong tiger femur individual. It is not too difficult to closely assess it even without the approach as rigorous as possible that I use.
 
 

4)   Articular and greatest length
 
 
“Christiansen's humerus length actually included the entirety of the femur, under the assumption that the protruding spur of the distal humerus is articular as most definitions says, test from his images yourself. Femur, tibia, etc. articular and greatest lengths are also the same, the articular surfaces are at the official ends of such bones."
 
 
I insist that the definition and his definition of articular and greatest lengths are different and to be rigorous there is the need to apply articular length of humerus, femur and tibia and greatest length of ulna when using Christiansen & Harris (2005) database. By their definition the difference in the case of femur and tibia is small, not so small in the case of humerus.
 
The proximal point of the articular surface is in proximal most point of the caput humeri while that for the greatest length is the most proximal one between the greater tuberosity and the caput humeri. The distal humeral articular surface include the trochlea and the capitulum. The humeral articular length is the distance from the caput humeri to the trochlear notch. Because of the trochlea feature the resulting lengths are normally markedly different even if the caput humeri is the top most point of the proximal epiphysis.
 
The articular length of radius is the distance from the medial point of the head to the middle of the lunate surface or the shelf of the carpal articular surface while the greatest length include the whole styloid process and the most proximal point of the head.. Because of the styloid process being a notch it implicates as for the humeral trochlea a marked difference between the two lengths.
 
In the case of the femur the reference point in the proximal epiphysis is the most proximal point of the caput femuri while in the distal articular surface can be either the mid-point of the condyles or the mid-point of the trochlear surface. Even when considering the mid-point of the condyles for the definition of the distal articular point (as Christiansen & Harris, 2005, do) if either the distal condyles are not both levelled with respect of the axis transversal to the axial principal bone direction and/or the greater trochanter is higher then the caput femuri the articular length is smaller then the greatest length.
 
The articular length of tibia is the distance between the most proximal point of the medial talar trochlear articulation and the medial condyle point.
 
I have both measurements of many different bones of many different carnivores species and they are slightly different. If Christiansen (1999) switched from greatest length of humerus and femur to articular lengths in Christiansen & Harris (2005) it means that he clearly meant the two lengths are conceptually different and serve different purpose in principle although generally they are close in the case of femur and tibia. I can guarantee you that for Christiansen the two lengths are not the same although very close.
 
The greatest lengths of bones include processes that are not closely related to weight bearing surfaces but are related to out-in lever features serving the purpose to move limb extremities with either greater strength or greater speed. These processes are taxa related and get rid of their size improve correlation and prediction abilities of bone length to body mass.
 
In the specific case of the 480 mm femur it is difficult to judge the ratio of the articular length to the greatest length particularly in the the proximal extremity due to the likely distorted pictures and the inconsistencies of the two views. However from the distal extremity it appears that the mid-point of the condyles is consistently closer in the axial projection to the caput femori then the lower most point of the condyles. The top part of the proximal epiphysis can be either the caput femuri or the greater trochanter depending by the view of the two pictures. You can try to do it by yourself
 
 
The total articular width of the distal humerus and femur exclude the epicondyles. It is clearly stated in Christiansen & Harris (2005). The two widths have not exactly the same measurements even in felids.
                 
 
Conclusively we can argue that the systematic errors accounting for greatest length and epicondylar mediolateral diameters instead of articular lengths (except ulna) and articular width may produce small body mass overestimates in some cases when using database of Christiansen & Harris (2005) and not always significant overestimates, but whenever possible it would be better to use the correct measurements to obtain more realistic results.
 
 
 
5)   Pictures of the 480 mm tiger femur from Ngandong Pleistocene


I have never seen such an amount of thinning of a tiger femur diaphysis towards the proximal epiphysis as in the 480 mm pictured from Von Koenigsvald (1933). Both pictures appear to be likely distorted.
 
I agree with the 59 mm proximal sagittal diameter is unrealistic and has no consistency with any other measurement. However reconstruct the likely correct width measurements from those pictures assuming the length is the correct one and generating as a result such a high increase in tiger limb long bone robustness that has NO corroboration from any other tiger limb bone being a femur or any other major limb bone either from Pleistocene and Holocene is a speculative exercise that in my opinion goes beyond a realistic and sensible analysis.
 
On January 2012 in AVA I wrote I had analyzed tiger fossil remains from Ngandong and came to the conclusion of how many individuals were likely included among the remains. I considered 7 fossil remains and sexed 5 males and 2 females with no doubt as it is very much clear in my opinion. I have estimated body size and mass of all of them for a long time. Either I consider the male sample or the pooled sample of both sexes, when considering the average size and mass and the standard deviation removing the very large individual, it results that an individual of the size of the 480 mm femur (using VK measurements) has a probability to occur not higher then around 2% and as small as much less then 1%.
 
Another suggestion why the 500 kg estimate is totally unrealistic in my opinion. You know that a realistic ultimate body mass of the largest panthera felid and many carnivores can be retained as around 150 % of the average body mass of the same sex at the very best and in my opinion is actually less then that in most cases. Ultimate means exceptional among many hundreds as a minimum and possibly thousands of adult individuals of that sex. If you therefore obtained 500 kg, it means that you should obtain an average body mass of at least around 340 kg among hundreds/thousands of Ngandong adult tiger individuals  You do not have hundreds as you have only 5 of them. Therefore the average body mass on a statistical basis should be much closer to the highest body mass. Perhaps 400 kg? I suggest you to never loose the sense of reality and the big picture.
 
In my case I have best estimates of average body mass of those 5 males between 245 kg and 250 kg and somewhat over 135 kg for those 2 females with the heaviest male (the 480 mm femur of course) around 315 kg. I definitely stick with my figures although I do not rule out the 480 mm femur individual could have been somewhat heavier (but as you can see my body mass estimates are always conservative even for the fossil lions, Smilodon and other extinct carnivores).
 
I would never dare to write a scientific paper with all the assumptions you made based on the length being the correct one, a quite unusual proximal AP diameter and generating a bone robustness on the basis of a distorted picture you cannot have a precise way to fix. This presumed robustness has no corroboration with any other Pleistocene and Holocene propodials and epipodials (major limb long bones) and come out with a 500 kg Pleistocene tiger particularly among 7 individuals only. When I posted the picture of that very massive brown bear from Alaska Peninsula taken in the Spring it was not only to joke a bit. It was to call to the reality because a brown bear that massive at a straight HBL length around 2300-2350 mm is not going to scale unrealistic weights of 600-700 kg in the Spring, but much likely no more then 500 kg. I am aware of only very few reliable and accurate Spring weights of brown bears in excess of 500 kg and these kind of weights are rather exceptional even in Kodiak Archipelago, Alaska Peninsula and Kamchatka. How you can conceive a panthera felid like that I have no idea.
 
Given the exceptional size of the 480 mm femur among the very few unearthed remains compared to the others the only way to get rid of any doubt is to find that femur and re-measure it. Whatever the measurement would be I personally can very hardly justify weights of 450 kg and more among panthera felids unless a bone would prove to be absolutely huge in size as a panthera felid cannot have the build of that massive brown bear I posted.

 
 
6)   My Database

 
“if you could please respond to our's as well. Especially, what is your mathematical model, as based on your Crater example, it seems more prone to interpretation than regression. Again, there is no way giving this general information can hurt, especially, as it looks to be, there is no expectations of getting anything published. You must admit, a large portion of the data you have found, as in the case of all of our members here, has come from this community. Collectively, posters have put the entirety of sources together and have all corners covered.”
 
 
Feel free to believe I just interpret things and do not rely on complex mathematical models and extensive database. On my side I do instead believe that matters in a web community are nearly always treated quite superficially, lacking a lot of theroretical background in many areas and with much insufficient data apart from hunting records of tiger weights and lengths in AVA and this website and to a lesser extent of those of a few other felids. I did however read some posters showing good knowledge in several areas in the months I spent in the community over the last 11 years.
 
The voluminous amount of material and the database I have collected have been very nearly entirely done with my own effort and research only over the last 20 years (and I mean more then 95%) and I can guarantee you that it has been a very big effort and very much time consuming. In the last 3 years I have to notice that I lost some pace in keeping it updated in the topics we are discussing in this community due to a number of reasons. I am however try now to go back to normal.

The fact that you can assume a large portion of the data I have come from this or any other web community make me think you have not yet realized how big can become a very large personal professional library and database. I have both and good links in the zoology and palaeontology professional community for reasons I am not interested to disclose. I hate to highlight things about myself and get personal, but who is telling you this is the one who posted in AVA many scientific and hunting records of lions and tigers back in early 2004 that nobody in the community knew about, data that are now very popular but that at that time were not. Who is telling you this is who privately received in my email box the Kerley et al. (2005) and Slaght et al. (2005) papers not later then early 2007 while no web community knew anything about those papers and data for likely more then another couple of years. Who is telling toy this was who first talked in January 2012 about an absolutely huge Pleistocene lion fossil remain never mentioned in scientific literature after the first original report of 23 years ago and that after 2 years has been finally mentioned in a peer-reviewed scientific paper. May be I am aware now of things the community will not know about before a few years.
 
At your age and still being a university student I also had probably not a fully clear idea what is a professional library and database, but with time I realized it and build one. Unfortunately there are always areas that need improvements in a process that never ends as I always would like to have more and more data to rely with even more confidence in my results. You are on the way to make a career in biology so I am totally confidence you will get to a professional level in all areas you need to. At that point may be we will meet, take a beer together and discuss privately of many topics I do not discuss in a web under the name of WaveRiders unless they have been covered by published research and these topics are being discussed.
 
I have been a quite irregular poster and reader since I joined the community in early 2004, earlier then any of you who are around here by this time but perhaps a very few exceptions. This is the third time interval spanning in total no more then 6 or 7 months so far in which I have been involved. I definitely want to say that for the benefits I received, and I definitely did benefit for a few things no doubt about it, I am grateful to those guys provide those info. Very much unfortunately the very most of those guys do not appear to be around here and around in general any more. At present time I am grateful to @peter as without a doubt he showed some old historical books on tigers I was not aware. Being present in the community and contribute for the months I did over the last 11 years has been for me a way to repay the community and I believe I have done it. Even this third period is a fruit of unusual circumstances of my life. I have no idea if my involvement can last much beyond this time when things will go back to normal at some point. I believe it will also depend from the level of discussions, the topics and how they are afforded.
 
The points I raised were not to compare my models with yours. Forget my estimates if you want as I did not say they must for sure the correct ones. I stated they are the ones I personally trust more, even more then those from Christiansen, but this it is up to me.
 
My critical analyses has been based to what you showed you have done without comparing it to what I did many years ago with little refinements since then and only minor updates as I have basically frozen my studies in these topics for a few years. I am now focusing my attention mostly in quite different areas. If my words have a weight they have for how I discuss your models, not how I have done mine or what I took into account.
 
In our discussion there is one major point and a minor one.
 
The major point is that in my opinion there is absolutely no way a 480 mm Pleistocene tiger femur can realistically generate a body mass of 450-500 kg and I share the same opinion of Christiansen while you seem to follow those of Hertler and Volmer, who are for sure nice and kind ladies and good researchers, but for their Pleistocene tiger weight estimates that I firmly reject.
 
The minor point is that in my opinion the Cromerian 192 mm MT3 is the largest felid individual ever existed and possibly even the heaviest one (as an individual) as the heaviest Smilodon populator in my very large database is likely to weigh around 370-380 kg according to my estimates. I first stated this in January 2012 in AVA. Several years before I had finally let Argant notice my feeling with that bone in a discussion we had and Jacqueline Argant confirmed me in writing she had no issue about that measurement avoiding me to ask them I would like to re-measure it. I also asked about another MT3 appearing to be as a different individual missing the distal epiphyses and I was confirmed of a length estimate at ca. 190 mm. Finally and with my great relief in late 2013/2014 Marciszak et al. (2014) stated the 192 mm MT3 had an incredible length. Have you ever heard in a peer-reviewed palaeontology publications of recent years to define a felid bone of  i n c r e d i b l e size? Incredible may mean hardly credible but it does not seem the case for that small piece of bone apparently actually measuring 192 mm.
 
 
                               WaveRiders
 

  Reply

[tigerluver]

1) 220 kg or 230 kg, it does not really matter, as there are some individuals of greater weight (250+ kg) at the same limb bone lengths in collections. Christiansen 1999a, the original source of the figure, has it down as 230 kg also.

2) My mistake here, I agree with you on this point assuming equal scaling proportions with modern lions, maybe not to the height you gave, but that's of no significance here.

3) Never made any assumption like that. I just stated the issues with bone measurements. Most don't know much about the Ngandong specimens either. Even the North American short faced bear seems to have gotten the short end of the stick in recent years after the hype of the South American form. In other words, most just know about the North and South American species. That's why we're here spreading the information. The weight of P. fossilis no one has a problem with, and actually my estimate a bit greater than yours, which may due to the inclusion of tigers in the genus database.

4) In simple words, the articular surface is where bones meet. Midpoint of the condyles, etc. are not the articular end. The bottom of the distal end and top of proximal caput/femoral head. The ulna had to be differentiated as total length as a sector of the proximal side does not meet another bone, and by definition is not part of the articular region. Look at any cat skeleton:

*This image is copyright of its original author

\

There are logical connections one can make to also prove the equality between full length of proximal long bones to Christiansen and Harris' (2005) measurements. For example, if Christiansen's length were not total, the Smilodon in his work are oddly much less robust than Merriam and Stock's specimens.

Campione agrees. Finally, Christiansen uses a few specimens in his 2005 from his earlier 1999a/b works. The lengths of the specimen used are the same. That ends that.

5) You've ignored the entirety of my explanation on the bone and vK's methods and the like and just repeated yourself. I'd expect one like you to understand the examples and relationships of the bone diameters as you've measured many. You can't just pass off inconsistencies if they don't fit your viewpoint. The specimen number is very low, and even then the averages do hover around 300 kg. The bones are also more robust, and interestingly enough, the Ngandong locality bones are more robust than the Wateolang locality bones. However, I agree that the femur specimen is likely at the very top of the species range, and coincidentally it was unearthed. 

I sense an implication that vK's femur measurements were wrong. I've seen this before and this has no basis either. There are a few instances throughout the book where he emphasizes how much larger the 480 mm femur is than the rest of the cat fossils. 

I've only posted one picture here of the femur from my copy. The two pictures on yuku that you are referring to are extremely distorted, but that is not what i used. The posterior view is still sitting snug in the book, whereby the condyles are seen. I'm not shredding up the entirety of a $100 book. Diameters came out equal. Easy explanation to that is that the tiger's epicondyles are not pronounced. Even in the yuku distorted images one can trace the sides of the femur from front to back. I am certain the images are not distorted in vK's publishing, as I've already explained. I'll say it again. There are quite a few fossils of where his measurements were of obvious areas, and the photographic extrapolation was on point to the hundredths place. To me at least, it makes no sense that one select bone he'll have photographed that poorly. I'll repeat another point, the extrapolated measurements fit the measurements that are comparable to modern methods.

We can end it with this. The rejection or acceptance of the femur, like in every aspect of the human scope, comes down to preference or disgust. There was once a group that photoshopped the tables for it to read 408 mm, at least you have doubts but are not stooping to that point.  

6)
Unfortunately, you get a bit aggressive when you are questioned. My questioning is not for derogatory purposes. Rather, it's for learning. I was not even looking for numbers, just the general model. There's a limit of what models are acceptable, and if someone has derived a new one we'd like to assess it, and maybe even apply it to a broader range. Instead, we get the tribute to the mighty WaveRiders and then some more accusations. I'm sure you are knowledgeable. It'd be improper to debate the tribute points anywhere on this forum, so moving on. Only remark on this will be, knowing this is the internet, you should expect a strong amount of distrust, especially considering the unlikelihood of the situation. Write with that in mind and don't be offended if someone questions the anonymity. For the sake of the forum, you'll likely respond with another tribute, but I will not respond back to keep it to the content of interest. Also, please don't put words in another's mouth as they say. I see that can occur when posts must be that word heavy, but debating words that are not mine is futile. Finally, as the debate has run its course, ideas are being repeated. So to finish, I apologize if anything I've said to you was taken as rude or negatively, especially with the lack of tone on the internet. I respect you, disagree with you, but negativity is not the intention. Good luck in your studies.

Next post will be on Homotherium, stay tuned. Reply

[tigerluver]

I've been to quiet lately, my apologies. In my free time, I've been digging through literature to uncover the Wahnsien tiger better. Remember those extremely robust canines, well at least one source has recorded something similar:

*This image is copyright of its original author

That's the only mention of tigers in the document by C.C. Young (1936?), the rest are other species. If one would like to see it, I can PM it or post it here (I'm opting not to as it isn't really about prehistoric felids). Wahnsien was teeming with species.

Anyhow, this may be of interest to some, this is a comparison of the skull of the Ngandong tiger (left) with that of a Sze-Chuan tiger (right) (presumably P.t. acutidens).

*This image is copyright of its original author

The Ngandong skull is 380 mm x 250 mm (vK has it down as 240 mm, proven to be an error by his own math and personal measurements). The Ngandong tiger skull seems very different with regard to the dentition proportions from the rest of tigers. Dentitions seem rather small, completely opposite of the modern Javan tiger.

I was left to find the length of the China skull myself. This is likely a female's, of greatest length 269 mm. The dentitions are proportionately larger than the other tiger subspecies. 

Based on these holotypes, the Wahnsien tiger is actually not very large and the seemingly small Ngandong tiger specimens are of respectable size when accounting for gender and the species exception robusticity not seen in the Wahnsien bones (bar metapodials). Of course, one specimen isn't necessarily representative of an entire species, but its the best we have. I will say this. As I continue examining fossil "tigers", the morphological differences may be strong enough to speciate, similar to the P. spelaea system. Of course, a lot of math is going into it to prove the speciation concept (e.g. ANOVA tests), and whenever all the code is written into R (somehow, the program isn't too user friendly), hopefully the manuscript will go through without being shredded too much. Reply

[tigerluver]

Unfortunately, the Homotherium review has yet to come. But until then, attached readings we can discuss... To start, the 400 kg estimate floating around is likely much too high based on these records.
 
I'm linking this paper instead of as an attachment: http://www.academia.edu/1194408/2008_-_D...e_Noordzee

 

 

  Reply

[tigerluver]

Homotherium crenatidens measurements from provided scale images:

*This image is copyright of its original author

*This image is copyright of its original author

Length: 347 mm
Distal AW: 67.7 mm 
Midshaft DLM: 29.8 mm

Conclusion:
It's difficult to prove that H. crenatidens was more robust than modern Pantherines. The Ballesio illustration shows a humerus on the less robust side of modern cats. The North Sea fossils show a 248 kg and 242 kg based on the Christiansen and Harris (2005) equations. The Ballesio specimen, with regard to all three measures, estimates to ((197 + 195 + 159) / 3) 184 kg. I did not use the weighting method as there's a major flaw to it if percent errors are below a certain value compared to others. As I've stated before, these equations seem to underestimate often. Homotherium latidens seems to be a ways smaller based on the mandible data, its total skull size would not be much more than the smaller tiger subspecies. Reply

[GuateGojira]

(05-26-2015, 10:40 AM)'tigerluver' Wrote: Anyhow, this may be of interest to some, this is a comparison of the skull of the Ngandong tiger (left) with that of a Sze-Chuan tiger (right) (presumably P.t. acutidens).

*This image is copyright of its original author

The Ngandong skull is 380 mm x 250 mm (vK has it down as 240 mm, proven to be an error by his own math and personal measurements). The Ngandong tiger skull seems very different with regard to the dentition proportions from the rest of tigers. Dentitions seem rather small, completely opposite of the modern Javan tiger.

I was left to find the length of the China skull myself. This is likely a female's, of greatest length 269 mm. The dentitions are proportionately larger than the other tiger subspecies. 

Based on these holotypes, the Wahnsien tiger is actually not very large and the seemingly small Ngandong tiger specimens are of respectable size when accounting for gender and the species exception robusticity not seen in the Wahnsien bones (bar metapodials). Of course, one specimen isn't necessarily representative of an entire species, but its the best we have. I will say this. As I continue examining fossil "tigers", the morphological differences may be strong enough to speciate, similar to the P. spelaea system. Of course, a lot of math is going into it to prove the speciation concept (e.g. ANOVA tests), and whenever all the code is written into R (somehow, the program isn't too user friendly), hopefully the manuscript will go through without being shredded too much.

 

Some time ago, I made my own estimation of the size of the Wanhsien tiger skull. I get to the conclusion that it was probably a female, no larger than modern South China tigers.

My estimation was of 326 cm in greatest length and 289 cm in condylobasal length. This was based in the fact that the mandible of this specimen was of 215 kg (Hooijer, 1947). Latter, based on this and using the equations of Christiansen & Harris (2009), Sorkin (2008) and Van Valkenburgh (1990), I get to a body mass of 157.9 kg for this particular specimen.

I most confess that I have not compared the pictures of the skull of the Ngandong tiger, in order to see if the zygomatic wide of 250 mm was correct or not. The first time that I made the original review of the Ngandong tiger in AVA, I believed that it was 250 mm, but latter Waveriders corrected me that it was 240 cm. Now that you say that it is 250 mm, I will like to see how do you get to that conclusion. I am going to make a comparison to see this too.

Other thing, the dentition of this particular Ngandong tiger skull is similar to that of modern Bengal tigers of similar size, in fact, using the equations of Christiansen & Harris (2009), Sorkin (2008) and Van Valkenburgh (1990), I get to a body mass of 255.8 kg, which seems more or less accurate for a specimen of that skulls size.
  Reply