On 9/8/23 3:20 PM, Peter Nyikos wrote:
> Family and university needs have really cut into my free time,
> disrupting my posting plans. This is the second and final reply to this post
> of 4 days ago.
>
> On Monday, September 4, 2023 at 9:49:58 PM UTC-4, John Harshman wrote:
>> On 9/4/23 6:25 PM, Peter Nyikos wrote:
>
>
>>> look at what is said in Romer's _Vertebrate Paleontology_
>>> about the phylogeny of the garfish and the bowfin
>>> (in Semionotoidea and Amoidea, respectively).
>
>>> These are "holostean" fish, but the evolutionary tree in fig. 60
>>> shows how Holostei is polyphyletic. In the 1945 edition, it is on p. 76, and
>>> its caption is: "Development of the groups of bony fishes and
>>> amphibians." The polyphyly is [...] explicitly mentioned
>>> [on] p. 87, in the section "Ray-finned fishes," where we read:
>
> <snip of later text, dealt with in my first reply>
>
>>> "the holostean level of organization was surely attained by more
>>> than one group of primitive forms, and the teleosts may be similarly
>>> polyphyletic."
>>
>> Yes, Romer was really into polyphyly.
>
> He explained that in the very next sentence wrt fishes:
>
> "However, this type of classification, even if not entirely natural,
> is one which it is best to preserve until our knowledge of the
> complex evolutionary history of the ray-finned fishes is much more
> adequate than is the case at present."

Of course, our knowledge of all this has increased greatly since then,
partly from improvement of methods, partly from the mass of data.

>> He thought mammals were too.
>
> Why are you so skimpy with the details? Do you believe that
> no one besides the two of us is reading this?
>
> You are making this claim because the tree in the chapter "Primitive Mammals"
> shows monotremes in a lineage disconnected from all the mammals
> known at the time, all subsumed under Theria. These included
> all the Mesozoic mammals known from fossils back then:
> "the aberrant multituberculates," the paraphyletic pantotheres,
> the triconodonts and the symmetrodonts.

In fact I was making the claim purely from memory of Romer's opinion. G.
G. Simpson thought that too, so I presume it wasn't an odd notion back then.

> And guess what: Carroll, in his 1988 _Vertebrate Paleontology and Evolution_,
> has a tree on page 415, which shows monotremes in the same "way out" way
> as Romer did. Even Morganucodon, now believed to be outside the
> mammalian crown group [but WHY?] is closer to those Romer-designated Therians
> in the tree than the monotremes are.

Don't know why. You would have to look at the character analyses. Of
course neither Romer nor Carroll showed anything of the sort. But any
modern analysis would show details of methods, matrix, and results.

>> He was wrong about both.
>
> Then so was Carroll. But what makes you so all-fired certain that these
> experts were wrong? How certain are you, for instance, that
> Theropods are closer to Sauropods than they are to Ornithischians?

I'm certain they were wrong because analyses using rigorous methodology
consistently show it. The base of Dinosauria is much less certain,
though the one that found the sauropod + ornithischian clade has been
criticized on many grounds.

>> If I recall, holosteans are paraphyletic, not
>> polyphyletic. But I don't see the relevance here. Still not getting what
>> you're trying to say.
>
> See my first reply to this same post of yours for both sentences.

See my reply to that reply, which perhaps you didn't notice.

>>>>> [1] is here, courtesy of the Wayback machine:
>>>>>
>> https://web.archive.org/web/20170504142727/https://www.ias.ac.in/article/fulltext/jbsc/034/05/0673-0686
>>>>>
>>>>> Unsurprisingly, Thewissen is one of the co-authors.
>>>>>
>>>>>
>>>>>> And it's mentioned in the article, for example "Within the
>>>>>> Pelagiceti clade, the BTD analysis recovered Basilosauridae as
>>>>>> paraphyletic, with Eocetus as the sister taxon of all other
>>>>>> sampled members of Pelagiceti. The lineage that gave rise to
>>>>>> Eocetus is estimated to have split off from all other members
>>>>>> of Pelagiceti around 45 Ma, or around the start of the middle
>>>>>> Eocene (i.e., during the middle Lutetian substage). All other
>>>>>> basilosaurids (aside from Eocetus) and neocetes are included in
>>>>>> a moderately-supported (PP = 0.67) clade that arises from the
>>>>>> next-most crownward divergence from the cetacean stem lineage.
>>>>>> "
>>>>>
>>>>> The article does not mention *Indohyus*. It roots the tree with
>>>>> the help of a pig and a hippo.
>>>
>>>> True enough.
>>>
>>> By the way, why do you think they used two taxa for the rooting? I
>>> don't recall coming across any trees that used more than one before.
>
>> Don't know why they used those two taxa, but the hippo is obvious, and I
>> supposed they wanted another artiodactyl. Trees with multiple outgroups
>> are not at all rare, though. It's considered good form to break up long
>> branches whenever possible.
>
> Does this have anything to do with long branch attractors?

It can, but I was thinking more of cases in which the outgroup is sister
to the ingroup and many species from that outgroup are sampled. For
example, the best outgroup for an analysis of anseriforms would be a
broad sample of galliforms. But adding taxa also just provides more
information useful in reconstructing ancestral nodes, even when topology
isn't at issue.

> In this case, all it supports is the idea that hippos are the sister group
> of cetaceans, and that entails not just one long branch, but the two we see in the tree.

Not a problem, since we know that on the basis of other data anyway. But
what adding the additional taxon also does is enable the reconstruction
of character states at the hippo/cetacean ancestral node.