|
|
Post by creature386 on Feb 15, 2014 17:30:16 GMT 5
Yeah, but the Therrien and Henderson method was quickly noted and criticized.
|
|
|
|
Post by theropod on Feb 15, 2014 17:43:00 GMT 5
Yes, by some. Some also criticise tooth-lenght based size-estimate methods for megatoothed sharks.
|
|
|
|
Post by Grey on Feb 15, 2014 23:39:01 GMT 5
Well, They must have included something bigger than that specimen, based on Gottfried et al.’s formula a lenght of 17.9m would imply a weight of 69t. Based on my measurement from the graph (74.131t), the individual with that body mass would have been 18.27m long. That's what I think, and I think that Balk graph is anterior to Pimiento's additionnal sample. That those data are preliminary affects upper and lower bounds even more. If the sample gets enhanced, specimens are going to be added on either side of the median, and will probably expand the size range in the sample. The median will remain RELATIVELY stable, unless there is a massive bias in the data collected so far. The point is at this time, whatever the final results, we already know there is an individual at around 70 tonnes and one below 100 kg. Either way, Balk said me the final results will be differents, some time range seems to have a short sample. This is pointless to predict anything at this stage. But at this time we already know the highest results. And as Shimada's method is somewhat conservative, I doubt there will be Megalodons above 70 tonnes. Of course I can be wrong. We'll see. Fair enough on the first part, as long as you keep it in mind. As soon as we find some additional white-shark dentitions with measurements we can verify it more extensively. Don’t forget you were the one to first note this discrepancy. And it’s not as if scientists always complained as soon as there is a problem with a method, many flaws have simply been overlooked over the years (European Torvosaurus), undoubtedly by numerous researchers, and it’s not as if a method that was not rebutted was automatically considered flawless. I remember I was the first to note this, but I've not stated something as definitive as I couldn't find Kent's corroboration to this when asking him. What I say is that it is possible we ignore others factors which justify the assumption. The results by Shimada seem to be tenable as they are not heavily criticized and even favored to Gottfried's method at now. This method was probably used simply because it provides the most convenient way of estimating the shark’s size based on every position in the toothrow. But at the same time, the resulting figures are also highly uncertain (even ignoring the difference between GWS and meg dentition). For example, the size figures in Pimiento et al. 2010 (and thus 2013 too) are the mean figures of tooth-position-ranges that frequently suggest a 2- to 3-fold difference in terms of TL (that is, literally, even worse than Amphicoelias fragillimus). For example, the specimen estimated at 16.8m, based on a crown height of 28.1mm, could have been a lower lateral 4-6, i.e., it could ahve come from an individual of 9.4m, 14.8m, or 26.0m. Yes, this method is used because it can be used in any position. And the results seem to be acceptable, otherwise the method would not be used repeateadly on Megalodon. The fact they use the middle estimate between a too low and too high is justifiably assumed. I have not seen big criticism about it by fellow sharks dentition experts. For the same tooth it is quite more conservative than Gottfried's, which is considered as conservative in its results itself, and considered conservative by Ehret. Jaws perimeter seems to give sizes up to above 19 m. Jeremiah's exceed 20 m while using Hubbell's arge tooth. Direct sizing gives easily sizes above 18-19 m. With Shimada to date I did not see anything yet above 18 m, whatever the tooth position (let alone the potential 18.27 which is still closer to 18 m). Compared to others methods, it is more conservative. Still I don't reject it. Of course if we see in the final study Megs above 19 m, I'll revise that statement, but that'd be still in the acceptable modern range. They favor tooth width or jaw perimeter in particular cases. There's overall a better correlation we're agreed on that. But that does not mean that any results from Shimada's are necessarily far off. In short, I'm fairly aware of this, but coudn't find strong support to this in litterature and discussions, that's why I consider that we perhaps ignore others factors in these studies. And by sheer modesty I don't consider myself more clever than the people working extensively on this all along the year. Shimada or not Shimada, the main lesson to remember is that all modern rigorous methods must be used with a degree of caution. But I'd ask simply to Kent : "Are the posterior teeth using Shimada's are likely to prone overestimates and underestimates using anterior teeth for Megalodon because of the apparent teeth differences with the white shark ?". My guess is that perhaps we are too much based on specific white sharks teeth set, ignoring all the variations which are included in Shimada's method. Edit : I had contacted Shimada years ago and found a response from him about this : Although I have never published papers that specifically discuss the
megalodon dental pattern, we do have a reasonable idea about the
variation range in a single megalodon individual based on:
Uyeno, T., O. Sakamoto and H. Sekine: Description of an almost complete
tooth set of Carcharodon megalodon from a Middle Miocene bed in Saitama
Prefecture, Japan. Bulletin of Saitama Museum of Natural History
7:73-85, 1989.
[as a side note, I interacted with all of these authors quite
extensively when I was a high school student in Japan]
Although the megalodon dentition was obviously not identical to that of
C. carcharias (otherwise they must be considered the same species, which
is clearly not the case), Uyeno et al.'s paper, combined with the
general pattern of dentitions among all the modern macrophagous
lamniforms (see attachment), does strongly suggest that it is reasonable
to assert that we can use C. carcharias as a model to talk about the
tooth pattern at least in a general way (e.g., how tooth sizes decrease
from the front to the back of the jaws).
I should note that when I was a doctoral student, I spent some time at
Gordon Hubbell's house and have actually examined his megalodon tooth
set and attempted to reconstruct the dentition with him. My take from
that experience was that the variation range seen in his tooth set was
similar to the range shown by Uyeno et al. Note that in his paper, Shimada also considers his estimate for Megalodon as "conservative". |
|
|
|
Post by creature386 on Feb 16, 2014 0:01:27 GMT 5
Yes, this method is used because of it can be used in any position. Remember that these "applicable to everything" methods everyone wants (including me) often simply aren't good for everything because there are always variations and different irregularities. |
|
|
|
Post by Grey on Feb 16, 2014 1:11:02 GMT 5
They are not good for everything but very simply they are the best available in any case. Quoting Kent : "[When we use any of the published methods of estimating meg body size based on extant lamnids, we are using the best available evidence...}". And I simply remark that Shimada's, despite the potential problem we are justly discussing with Theropod, still yeilds into acceptable sizes ranges for megalodon by modern view, and is even slightly more conservative than others available methods.
The best is again to wait, because discussing of this at lenght will be pointless. We potentially ignore a number of factors and there are reviewers and colleagues comments to follow, hoping it to be published on Plosone like the nursery related paper.
|
|
|
|
Post by theropod on Feb 16, 2014 2:35:47 GMT 5
|
|
|
|
Post by Grey on Feb 16, 2014 2:38:00 GMT 5
Google translate is our friend.
|
|
|
|
Post by creature386 on Feb 16, 2014 2:41:09 GMT 5
If you can understand it.
Remember that Japanese has a different sentence structure than English (verbs in the end, from what I remember). I doubt I would be able to understand it, especially because the text is likely very complex.
EDIT: I can't even copy the text to paste it in google translator.
|
|
|
|
Post by theropod on Feb 16, 2014 3:09:01 GMT 5
Grey: Nah, I don’t like google translator anyway, even though it works reasonably well with some languages. and with a scanned paper from 1987 in a writing system I don’t even know how to copy it won’t help much. We have measurements from one white shark tooth set, and you initially noted the differences in another one, which shows the same features. Carcharodon’s teeth appear to be more variable in terms of size, the anteriors being taller and the posteriors lower. Unless these two somehow represent extreme outliers, it would be ridiculous to not at least take that into account as long as there is nothing that contradicts the implications this has. Note that Pimiento et al. (2010, 2013) were not all about producing exact size estimates, merely differentiating between probable adult and probable subadult and juvenile specimens. The method is certainly good enough for that, but I’m not very confident in it for size estimates that are supposed to be as good as possible (or, the least bad possible). Here we should eliminate any source of error that we can possibly eliminate, the uncertainties remain big enough without leaving a likely bias that we know unaccounted for. Shimada’s method is thus likely (=for all we know) overconservative for anteriors, and liberal for posterior teeth. What you may have noted too is that the largest size estimates derived from it are typically from relatively small teeth (i.e. posterior ones). In a normal sample, one would expect a bigger percentage of anterior teeth, which are those most prone to falling out, since they are the tallest-crowned, most transversely oriented and those that contact the prey first. And of course every point of criticism related to the dental proportions applies to Gottfried et al.’s method just as much. Shimada’s is nowadays more commonly used, since it provides an easy way to get a figure for any tooth based on a (guesstimated) assignment to a position. Being favoured to another (equally problematic) method because of being more convenient is not really a strong support imo. Of course we should, perhaps most importantly, continue to test this, but as it is now it is an observation that is yet to be falsified. On the other point; the estimate used in Pimiento et al. was not the intermediate estimate. It was the average of the three. e.g. estimates for UF 237956 (28.1mm CH) are as follows: l4 position ? 937.64mm l5 position ? 1479.74mm l6 position ? 2607.43mm Where the figure used was the mean, 1674.94mm TL, while the middle estimate (i.e. the median) would have been 1479.74mm. The former corresponds to none of the possible tooth positions, which, in a sample of three and with such great disparity, is quite a significant difference, that between ~16.8 and ~14.8m. On the date: The upcoming research by Pimiento and Balk was presented at the SVP meeting in late October (which is probably where that poster is from), it is in the accompanying abstract book. The newer paper about elasmobranchs from the Gatun Formation was published in September, so in all likelyhood it incorporates data from a recent paper by one of the authors. We only know two random specimen sizes, and that only roughly. But the median size is not likely to be subject to considerable changes with the addition of further specimens on either side of it.
|
|
|
|
Post by Grey on Feb 16, 2014 11:27:55 GMT 5
Grey: We have measurements from one white shark tooth set, and you initially noted the differences in another one, which shows the same features. Carcharodon’s teeth appear to be more variable in terms of size, the anteriors being taller and the posteriors lower. Unless these two somehow represent extreme outliers, it would be ridiculous to not at least take that into account as long as there is nothing that contradicts the implications this has. It seems that these variations have been taken into account by Shimada in his methodology. I take it into account but instead a make any conclusion, I keep in mind this method is complex and has been used by several sharks experts since a decade. So I very much doubt this problem has been ignored by rigorous guys such as Kent or Hubbell. Still I'm going to investigate this. But Shimada's results appear to be all justifiable and reasonnable. The problem of Shimada remains the same than in any method of estimate in not complete taxa, the ignorance of allometry. Yes but to determine the different life stage they had to make precise sizes estimates. I won't eliminate anything in these papers until I get more clues about the variations used in the original method and more solid oppostion. Otherwise would be premature. Kent did not say anything about this, neither Ehret, Hubbell, Pimiento, MacFadden when they did the study. No reason I do it too because I've remarked differences with the white shark dentition, differences kept in mind by Shimada when he used the method on megalodon like explained in his mail. You have missed something : you focus on the posterior and lateral yeilding big sizes but you forget that most of the posterior and lateral in the study actually result in small sizes under 10 m, including teeth with several possible positions even with a greater range of position than the 16.8 m one. And there's no reason to consider that we should have more numerous anteriors for big specimens. It is true that upper anteriors are most likely to be preserved because of greater calcification but there's no reason to consider that adults in the areas, which were there occasionnally, in an area not frequented by large potential preys and for a short time period, necessarily should have lost their upper there. Also, the total sample is relatively small. Well that's your opinion, but apparently not the opinion of Pimiento et al. I advise to investigate this further before make any conclusion. And I don't believe we have the talent yet to call for a revision of the paper. It seems that this method and its results are reasonnable. An observation based on one particular white shark set. I think Shimada has seen more white sharks and megalodon sets than us (just reading the mail above..). This observation is interesting (after all it was mine  ) but you give it probably too much importance for now. For sure it deserves to be investigated. I know, error of tipping. That's a reasonnable assumption. Also, you can try it with others posterior or lateral teeth yeilding smaller estimates, despite the possible range of position is even wider in some (l1-l5 for UF 237953, resulting in 7.2 m). I say you, there's something we don't know yet in these calculations. That's a good thing to have verified the dates but this does not mean Balk kept it in mind in her graph which was perhaps made weeks or months before her talk. And it does not seem there is the 17.9 m TL individual at 10 MY in the graph, the largest at this time would be more around...16.8 m ? Any change even small is important. Balk repeated me to keep in mind the study is not finished. So, again, no conclusion. |
|
|
|
Post by theropod on Feb 16, 2014 20:28:14 GMT 5
I’ve measured the whole thing as good as possible, here’s the results: Untitled 1.ods (14.57 KB) It’s so small it’s difficult to read, but doesn’t the poster say "Gottfried et al. 19 98"? If so, is that just another date of publishing for the same paper (1996) or is it yet another one? It seems unlikely the same author would have published two different lenght-weight regressions for the same taxon in rapid sucession. That is because the area was a nursery, with the vast majority of specimens being relatively small. But our subject here are the large ones. ![]() At least it is odd that none of the really large specimens from Pimiento et al 2010 and 2013 seem to be known from anterior teeth, they are all at least middle laterals or even further posterior. On the other hand, numerous specimens are known from anterior teeth, the largest of those is 8.6m long, most are below 6m. Curious, isn’t it? We don’t know Pimiento et al.’s opinion on that matter, we cannot look into their heads. Their papers are not predominantly about megalodon’s size, but about palaeoecology and population structures. I will test this further, but what is apparent so far should not be neclected. On two sets. we have no reason as of now to suspect those are exceptional specimens, but we will search for more specimens. How? It’s a documented method using linear equations based on great white sharks. You can do it too, it’s not complicated at all. Enter the formulas from the supplementary material of the nursery study in a spreadsheet, estimate lenght for every tooth position based on the given crown height, and then compare the mean and the median of the estimates. |
|
|
|
Post by Grey on Feb 16, 2014 21:25:44 GMT 5
I cannot open ods file.
That's a typo.
I'm aware it's a nursery that's not the point. You claim that posterior teeth are likely to prone overestimates especially when several positions are suspected. That would mean that the young individuas at less than 10 m (UF 237951 L1-L5 is only 3.1 m), which are numerous in that case, are likely overestimated in size because Shimada produces overestimates with posterior and lateral ? Not logical. We cannot decide arbitrary that this posterior tooth gives an overestimate when this one gives an accurate size.
It is not odds because adults were rare and made short occurences in these waters, not hunting whereas neonates and juveniles lived exclusively there during years.
And frankly you really think this kind of reflection would not come to mind of the sharks experts having collected material there during years ? It of course came to my mind since a long : that's a megalodon nursery.
I have discussed with Catalina Pimiento and Dana Ehret. And they certainly not have written the paper saying "don't believe our estimates for the large adults megalodons, that's hogwash, but the datas about paleoecology and populations structure are good stuff !" And I won't suggest authors make works they don't consider.
I don't neclect it but as I'm certain there are some values in that methodology we might ignore I won't test it further. I'm going to simply discuss it with the specialists.
That's not a matter of excpetionnal specimens but of individuals variations in white sharks that have been initially taken into account by Shimada in his formula.
How? It’s a documented method using linear equations based on great white sharks.
It's been years I use Shimada's formula. But in some cases I got strange results, including in the white shark, because I ignored how to finish my calculations. We don't know all the data variations used in the initial formula.
|
|
|
|
Post by theropod on Feb 16, 2014 21:42:39 GMT 5
Can you open an excel-file?
two times?
No, not especially. The problem to make an exact assignment is another one.
Why not? How much is a different story, I have not measured mid-laterals, just posterior laterals. Don’t forget the size estimate is the mean of all those positions, not the estimate for the more posterior positions.
17.6mm CH suggests the following (L3 being the median in this case):
L1?255.4072
L2?241.3318
L3?256.544
L4?316.3498
L5?458.5276
EDIT:
The first lateral will likely give an underestimate, the fifth lateral an overestimate. L5 is only 70% the lenght in the greath white, while L1 is 114% the lenght at toothrow-lenght parity. L3 will also give an underestimate, it’s even ~117% the lenght.
We don’t do that.
That does not explain why in the several big specimens found there with that method none are known from anterior teeth
Well?
As I wrote, their goal definitely was not that of establishing exact size estimates for their specimens, at least not so far. But I don’t really think I have to justify why I favour other methods any further.
But what you are talking about would require these two to both be on one extreme of the scope of variation. It seems much more reasonable that great white sharks simply have somewhat different dental proportions.
I’m not sure I get what you mean, can you elaborate please?
|
|
|
|
Post by Grey on Feb 16, 2014 22:35:41 GMT 5
Yes.
Yes, the litterature reference is at the bottom of the page, 1996. Gottfried wrote nothing about Megalodon in 1998.
Hence averaging the TL yeilded from the various position is a fair assumption.
You have not measured laterals, but the 17.9 m individual justly comes from a lateral.
And the mean of all these estimates is indeed the most reasonnable assumption ! When the size estimate with a small posterior tooth results in a large TL means that even though there are several position, that tooth most likely was far at the back of the dentition. It is not logical to arbitrary decide that the average yeilded from several positions here IS an overestimate. That simply means it was most likely a very small tooth in a large shark.
Well you're doing a trial to the 16.8 individual. Just kidding. But the 17.9 m individual is not from a posterior but a lateral tooth. UF 245996 is also yeilded from several tooth positions and result in 13.1 m, not really an overestimate.
No adult in the area had the time to lost an upper because they were few, not for a long time and not hunting big game in the area. And even if it lost it, nothing says that the upper was necessarily preserved and if so not collected by fossils hunters long ago who tend to ignore small, less valuable and less striking teeth. That explains this.
Well nuff said in my last responses, in a nursery area, in a sample of 40 teeth, you're not necessarily going to find a large upper from a massive adult, even less than in other areas. Upper from massive adults are rare and almost always in the hands of privates.
They needed to make precise estimates here in the goal to establish life stage of the individuals. I favor other methods in others instances, not there.
We cannot make any conclusion based on two sets whereas Shimada's formula is based on numerous various individuals.
Some results were not logical (far too small or far too high) for individuals teeth, despite I did the calculation several times. That's why I think in some instances others factors that we ignore are taken into account.
The main problem I try to explain is that you give far too much importance to the data we know about Megalodon teeth compared to the white shark applied here whereas we ignore how much we are right compared to the individuals variations used in Shimada's formula.
I was the first to found these differences and already sent questions to Kent, without satisfying response. I'm going to try it again. The main thing is that whatever the results by Shimada would be differents compared to others methods, there's no reason to think they would be far off as they still fall in the acceptable modern size range for Megalodon and are close to those of Gottfried et al.
There's a better correlation in jaws perimeter than in crown height, this does not mean the crown height correlation is utter hogwash and deserves to spit on the datas by Pimiento et al. No professionnal did it so far, there's no reason to do it because we have measured two white sharks sets in px on Paint. We need more clues in both directions.
For short, whatever the accuracy of Shimada regression when applied to Megalodon, it always fall in an acceptable size range for the species and it is not that important to always keep in mind others methods while Shimada is the only one to be useful while using teeth that are not UA. Otherwise, all the published sizes of Megalodon using Shimada should be rejected, which I don't agree with. This is not because upper jaws perimeter is overall better than Shimada's regression must be ignored. On a rigorous line, all the published methods are subject to potential criticism. But all the published methods are the best available. Unless studies by Shimada yeild in unreasonnable Megalodons under 1 m TL or over 25 m TL, this method is good enough.
|
|
|
|
Post by creature386 on Feb 16, 2014 23:01:23 GMT 5
It’s so small it’s difficult to read, but doesn’t the poster say "Gottfried et al. 19 98"? If so, is that just another date of publishing for the same paper (1996) or is it yet another one? It seems unlikely the same author would have published two different lenght-weight regressions for the same taxon in rapid sucession. This has to be a typo. There is a function in google scholar where you can determine the year of the publication and there were no Gottfried publications in 1998. |
|
) but you give it probably too much importance for now. For sure it deserves to be investigated.