|
|
Post by Grey on Oct 31, 2013 3:19:02 GMT 5
Theropod, you have to understand this is not a method given by Kent. A correlation of 0.65 is not really better than what is seen in Gottfried. With another sample of width, we could get again another different correlation than what is proposed by Jeremiah. Use it personnally if you want but don't use it as if it represents a tested method. For the same reason, I don't use Jeremiah's anymore.
For the remaining part, we need datas of modern lamniforms set of teeth to be adequately used. I'd quote Kent :
"I'm most comfortable with size estimates based on tooth width, although I'd be even happier with estimates based on jaw perimeter. Tooth height is pretty variable, depending on life style, but jaw perimeter is more tightly related to body size. In the absence of jaw perimeter tooth width is the next best alternative."
I suggest to wait to see this new stuff using Shimada's method.
|
|
|
|
Post by theropod on Oct 31, 2013 3:50:49 GMT 5
You don't know the correlation in Gottfried et al., 1996, the paper doesn't seem to give R² anywhere.
I didn't know you didn't use Jeremiah anymore, not long ago you and most others argued tooth width was the best method there was. The formula from Kent's plot is as good as or better than any we have to date. Its as tested as it can be, which also is not great but, its the best we can do.
And its the only formula other than Gottfried's that we even got to see a function plot of, and have a decent documentation of how it was done...
What Kent writes in that quote of yours is exactly my point: We have no data on jaw perimeter, since we have no jaw or assembled dentition in 99.99% of megalodon specimens, and also: tooth width is the best we have ("the next best alternative"), better than tooth height. And that despite the relatively low correlation coefficient from his plot.
This obviously makes a lot of sense. A jaw can never be smaller than the sum of all tooth widths, while the same can not be said about tooth height, ie. the correlation of jaw size with tooth size is closer, and obviously a jaw also puts certain constraints on the size of the body it belongs to, while that's not the case for the lenght of a single tooth (Thylacosmilus!).
If you don't use tooth width anymore, what do you use at all? It has been made clear none of the methods that are known are very good, but this one is still the best there is. So if you reject it, you reject all the methods there are (I wouldn't have a problem with that, but still it seems strange).
Of course one could argue that all the methods are likely inaccurate, because all base on animals of questionable relation and suitability as an analogy, but then this whole thread is pointless.
|
|
|
|
Post by Grey on Oct 31, 2013 4:05:27 GMT 5
What I say is that Gottfried's is isometry, as well as tooth width, isometry over a large scaling inducing a huge potential error. Since tooth width has less variability, I can understand one favors it over Gottfried though. Indeed I don't refer anymore to these unpublished methods : theworldofanimals.proboards.com/post/7893/threadWhatever you use the formula of that scatter does not make it a proper formula as it justly shows the weakness of that method, and that anyone with a sample of great white teeth could get differents results with a vast range of probability. That's why I don't use that tooth width range of Jeremiah and Kent as you did as the range could be even wider than this, since this does not represent any reliable data. I don't use Jeremiah's anymore for the same reason and also because we don't know properly the basis of his work (samples, datas), no more than what is written in Renz book. Use it if you want, but such a weak correlation does not interest me to use, so don't make Kent's scatterplot a proper method at all. Kent's goal was to show the weakness of this, not its utility, which is weak. I don't spit on it, but I don't use it as a solid tool. But overall any method is not that bad or that good. We do not have data on jaw perimeter but we do have complete set of teeth from megalodon, we know the dentary formula of the species and the number of teeth and we have modern lamniforms datas sets of teeth, not only the white shark. Siversson explained he used not only white sharks datas, but also makos and porbeagle. This makes this more tenable. The only problem is that this is not published, about meg, at least yet. But there is stuff to use. I'm more interested in that method and in Shimada's method as of now. No wonder, both these methods are just used in the most recent papers relative to fossils sharks. BTW, Brett Kent confirmed me the tooth we've observed previously is most likely a UL5. |
|
|
|
Post by theropod on Nov 1, 2013 4:07:11 GMT 5
I can understand the reasoning behind this, its not surprising considering all those estimates base just on isolated teeth; teeth are a small and variable body part. With a bit of luck well have better material soon anyway.
I'm not making this any more of a proper method than Gottfried or Jeremiah, be assured. I'm equally sceptical regarding shimada tough, it also bases on tooth height and great white sharks alone. Does anyone here have access to the paper describing that method? I know the formulas from the supplement to the nursery study, but not the details of how the method was derived.
Jaw perimeter would be by far the best, if we had it. But extrapolating the relationship of tooth size?jaw size and then jaw size?total body size is no better than directly extrapolating tooth size?body size, it's merely one more step in between, but with the data to do that one can also skip it.
Its basically like extrapolating the lenght of Spinosaurus femur based on Suchomimus and then use the femur lenght to extrapolate its total lenght.
This perhaps gets more relevant in showing the accuracy of portrayals and envisioning the size of body parts to avoid mistakes.
What someone should do is compare the relative tooth widths in a quadrant of a C. megalodon dentition and a C. carcharias or other lamniform, see if there are notable differences in the distribution along the toothrows, but without high-resolution pictures it probably doesn't make much sense.
EDIT: Indeed it seems what we have been ignoring up to this point is that the relative sizes (especially the height, but also the width to some degree) of teeth in parts of the dentition are quite different between C. megalodon and C. carcharias. If we use an associated megalodon tooth set as a template, thatÂ’s a good reason to favour this method over others which do not manage to account for these differences.
Didn't you talk about some giant-predator-comparison you received from Siversson?
|
|
|
|
Post by Grey on Nov 1, 2013 4:31:12 GMT 5
I have the Shimada's paper. I'll post it.
Jaws perimeter is the best method, but not much better than the others. And I have a problem in tha jaws perimeter gives potentially too gigantic sizes (if I've made the calculation right using Hubbell's tooth).
Siversson's comparison was not a scale picture. I can show it privately, I have to ask him before post it on a forum.
|
|
|
|
Post by theropod on Nov 1, 2013 16:31:14 GMT 5
Really? The figures Siversson stated in his talk didn't seem too gigantic, albeit not very conservative either.
How did you do the calculation?
|
|
|
|
Post by Grey on Nov 1, 2013 18:11:12 GMT 5
I'm not sure if Mike Siversson used Hubbell's tooth (which is in fact 15 cm wide).
If you use it with Jeremiah's (yeah I had written I wouldn't use it anymore...), you get 20.55 m TL.
I've simply scaled this figure based on the difference seen between Jeremiah's and jaws perimeter seen in Kent's scale and calculated a jaws width perimeter yielding 22 m TL for the owner of that tooth (which is actually 15.24 cm wide).
But I guess my calculation is wrong as there are parameters specific in Parotodus dentition which might explain that difference between jaw perimeter and Jeremiah's method...
|
|
|
|
Post by theropod on Nov 1, 2013 19:01:14 GMT 5
That in one case jaw perimeter yields a result higher for a certain percentage does not mean it always does, there are variables to consider in the tooth width method, the used tooth was an UA1 and apparently it is a good deal smaller by width than L1-L3 in P. benedeni.
Also, are you sure the formula was the same as published in Renz (2001) and not the one based on the regression Kent performed himself? We can check if someone knows the exact width of the upper anterior in question, but strangely no first UA seems to be figured in the description.
What seems to be clear is that tooth height scales well below isometry, while width is a more stable indicator.
|
|
|
|
Post by Grey on Nov 1, 2013 19:37:36 GMT 5
Kent specified he used Jeremiah's method, I recall he did not propose "his" own regression and this scale comparison was a picture from his talks about Parotodus, which date back years ago.
Fair enough, like I said, particularities in Parotodus dentition may explain this vast difference with Jeremiah's method.
However, I'm not sure that Mike Siversson when he estimates meg size, was aware or used the 6 inches wide Hubbell's tooth.
|
|
|
|
Post by theropod on Nov 1, 2013 22:30:26 GMT 5
I know, "Jeremiah's method" may simply refer to the root width method in general, that's why I wrote "formula".
Siversson talks about the refers to the largest known teeth in his talk, its unlikely he'd ignore or not be aware of Hubbel's tooth.
Isn't there any high-quality shot of Hubbel's meg dentition (something fit for measuring)? For perimeter, we need, the combined tooth widths (we can measure those from a picture of the assembled dentition), the spacing (based on great whites, as in Kent's works) and the lenght of tooth-free space in the back of the mouth (~17%). How much space is typical in lamniform toothrows?
|
|
|
|
Post by Grey on Nov 1, 2013 22:42:17 GMT 5
I'm not sure he's really certain that Hubbell's tooth is 15 cm in width as I've only realized this through my communication with Renz.
You have Hubbell's set measurements here
CW (mm) CH (mm)
A1 82.7 82.3
A2 78.8 75.4
A3 76.7 80.2
L1 77.6 79.2
L2 77.5 87.8
L3 73.7 81.6
L4 68.2 71.7
L5 53.7 62.2
L6 36.3 50.6
L7 32.2 48.2
L8 19.9 34.1
L9 14.2 21.0
a1 68.7 59.8
a2 72.0 67.9
a3 74.0 64.7
l1 67.2 63.0
l2 67.7 66.8
l3 63.1 65.8
l4 55.8 63.9
l5 46.7 58.8
l6 34.4 48.4
l7 21.3 32.8
l8 10.8 22.2
I've calculated a total upper jaw perimeter for the young Hubbell's specimen to be about 187 cm, including the 17 percent in more but I don't know if it includes the spacing.
Needs Mollet's data.
|
|
|
|
Post by Grey on Nov 1, 2013 22:48:03 GMT 5
|
|
|
|
Post by theropod on Nov 1, 2013 22:57:10 GMT 5
|
|
|
|
Post by creature386 on Nov 1, 2013 22:59:34 GMT 5
Theropod, for some reason, your file doesn't work.
|
|
|
|
Post by theropod on Nov 1, 2013 23:04:11 GMT 5
Did you download it? It appears to be protected from edits if you just click "open" in the browser popup, you first have to save the file on your computer.
|
|
