I've left in what text I figured was neccessary for folks to understand
the arguments going on. This'll kill most a ya. :-) Read on only if
you dare. Warning: this message could be dangerous to your health. Go
find someone else to complain to when you go cross eyed trying to read
it! Much of the arguments contained within are all subject to question
and no one has the right answer. We can only sit around and bullsh!t
about it and make ourselves feel important and like we have conquered
something bigger than us all. :-) Mind you, the human imagination is
probably bigger, especially in some folks. Enough dribble, onto
rainbowfish evolution!!
> >> Well, there are several arguments why rainbowfish are a rather young
> >> family.
> >> 1. There relatives are also rather modern fish.
>
> >What do you consider their closest relatives to be? I know there has
> >been some recent work done on the relatedness of the various families we
> >are interested in. However, I have not seen any of it. If anyone has
> >any references to it I would be greatful.
>
> I'm not quite sure what their real closest relatives are but we have
> to look for them in the Telmatherinidae and their relatives.
More for info than anything else. This paper reverses some of the
classification thinking of the past. Past work has tended to look from
the bottom up, ie from species to genera to subfamily to family etc.
This study more looked from the top down, ie series, order, suborder,
infraorder, superfamily, family, subfamily and tribe.
Dyer & Chernoff. 1996. Phylogenetic relationships among atheriniform
fishes. Zoological Journal of the Linnean Society. 117:1-69.
This is in no particular order. This will probably get all screwed up on
some folks email programs, sorry.
Melanotaeniidae|-Melanotaeniinae--------------------------|-Melanotaenia
(Family) | (subfamily) |-Glossolepis
| |-etc etc
|
|-Pseudomugilinae-|-Pseudomugilini (Tribe)-|-Pseudomugil
| | |-Kiunga
| | |-Scaturiginichthys
| |
| |-Telmatherinini (Tribe)-|-Kalyptatherina
| |-Paratherina
| |-Telmatherina
| |-Tominanga
|
|-Bedotiinae-------------------------------|-Bedotia
|-Rheocles
No doubt this will change later as these things always seem to. I know
Dr Ivantsoff's lab will be in disagreeance with it. :-) A
couple of quotes from the text. "Relationships among the melanotaeniids
are only tentative, but they do indicate that pseudomugilins and
telmatherins are sister groups. The relevance of this phylogenetic
hypothesis to the biogeographic history of the Indo-West Pacific region
is significant because Melanotaeniidae is a monophyletic group of
primarily freshwater fishes."
Also from the paper, Melanotaeniinae share a common ancestor with
Pseudomugilinae which branched off after the Bedotiinae did. Thus, the
closest relatives to Melanotaeniinae are the Pseudomugilinae according to
this paper. One thing Dr Ivantsoff mentioned to me is that his PhD
student, Aarn Whitehouse feels that Bedotiadae and Melanotaenidae (using
older classification) split around the Eocene (~50 mya) when Oz, India,
and Africa split off and I guess hadn't drifted too far away.
> >> 2. These fishes are still closely related to other fish which have the
> >> same history, coming from seawater relatives and invading fresh water,
> >> like Telmatherina and Bedotia.
>
> >Do we really know that fish such as Bedotia and Melanotaenids originated
> >from salt water? Based upon what evidence? Allen suggests melanotaenids
> >are freshwater evolved since they have not crossed Webers Line. If they
> >were marine they could have swum it as it is not that far apparently.
>
> Yes, we know. Australia - together with New Guinea - has left the ancient
> continent (forgot the name, the stage before Gondwana) very early. Thus
> if it were primary freshwater fish they must be very very old - like lungfishes
> and species like Scleropages. Sign of old species is, that they don't evolve
> strongly (as the two Australian species didn't do).
Australia and Antartica were last to break away, everyone else had split
before that. Guess they just didn't like their neighbour much? :-)
Actually, up until fairly recently, the last million or less years there
were quite a number of lungfish species which are now extinct (around 9).
You cannot say that _only_ old species don't evolve strongly, sure, that
may be the case in some but not in all.
> Rainbowfish still seem
> to be in the course of evolution. According to all what we know in the
> evolutionary biology that's a sign of a rather young (that's relative, also
> a few million years is young) family.
I can't say I've heard of this before. Do you have a reference for that
statement? I also can't say I agree with it either. This is also fairly
subjective and thus not going to be resolved by us either. :-)
Based upon what evidence are they thought to be young? They show
tremendous variation in colouration. Other than that they are relatively
similar morphologically. They all share the same basic body plan.
Craterocephalus, Ambassis are all just as hard to distinuish
morphologically. If you look at the key to rainbowfishes it's amazing
how important colour is in distinuishing them. Now, I'm not doubting
present rainbowfish taxonomy, but if they were a plain grey fish (largely
like Craterocephalus and Ambassis are) would we have the same taxonomy
today? Would we still think they are young species? Take a look at a
key to Ambassis or Craterocephalus, no colour, and it's damn hard to tell
many of them apart. All I am trying to point out here is that we have a
highly variable trait (a single characteristic--colour) that we seem to
regularly give considerable weight to while forgetting the rest of the
fishes characteristics.
As far as using evolutionary theory goes, I think this is a little
dangerous or potentially misleading. On it's own it is pretty weak.
Combine that with geological evidence and I think you have a very strong
case. To me, the geological evidence suggests a very old lineage of
fish. There are still many questions left unanswered regarding the
geological record too, but I think it has more credence than evolutionary
theory.
> >> 4. Strong speciation in M. trifasciata and M. splendida and others,
> >> which is a sign for modern fishes.
>
> >Why? Pupfish (Cyprinodon) as a genus have been around well in excess of
> >10 million years, probably more like 30 and they are very speciose.
>
> Are they? Or have they been? Most pupfish are at the end of their way. If you
> have such an extreme isolation under extreme conditions like in pupfish
> you will find speciation of extinction. I don't think they are much older than
> 10 million years - and thus a rather young family.
Pupfish are presently more widely spread than rainbowfishes are. I'd
hardly call them on the way out. They will be around for a few years to
come since they tend to specialise in habitats that other fish cannot
survive in. As long as these habitats remain so will pupfish. Not all
pupfish have severe isolation either. C. variegatus occurs from the
north eastern US all the way through the Gulf of Mexico.
> >> Rainbowfish are thaught to have an age of a few million years,
> >> maybe two or three.
>
> >Define rainbowfish. Do you mean Melanotaenidae, Melanotaenia, or
> >something else? According to paleontologists the average age of a
> >species in the fossil record is 6 million years. Who thinks they are 2-3
> >million years old? What is your source of this information?
>
> E.g. look into Jerry's new book.
You still didn't define what level of taxonomy you consider to be 2-3
million years old. :-) This is an important distinction to make.
I took a look at Jerry's books. A few quotes from them;
"There are no known fossils of rainbowfishes, although insufficient work
has been carried out on more recent fossil strata of the Aust-NG region."
"Eventually it (the ancestor) evolved a purely freshwater life cycle, thus
setting the stage for the explosive speciation of rainbowfishes in
relatively recent times, perhaps within the past 200,000 years."
"There appears to be little doubt that the group originated in fresh
waters of the Australian continent as there are no closely related forms
occurring outside of this region. Although some species of Pseudomugil
are found in salt water or brackish estuaries, as previously mentioned,
the family as a whole has had little success in crossing marine barriers,
even if very narrow."
Note that there is very little geology mentioned.
As an alternative;
Crowley. 1990. Biogeography of the endemic freshwater fish
Craterocephalus. Memoirs of the Qld Museum. 28(1): 89-98.
She proposes that there were two invasions into Australian freshwaters,
the first was the "eyresii" group around 110-95 million years ago, the
second was the "stercusmuscarum" group around 35-25 mya.
"Although Whitley (1959) and later authors (e.g. Allen & Cross 1982;
Merrick & Schmida 1984) considered the freshwater fishes of Australia to
be of recent origin, I suggest that despite the present lack of atherinid
fossils in Australia ... hardyheads at least (and most probably
rainbowfishes with similar species distributions) are not the result of
recent speciation." She bases this statement on the geological history
of Australia and the potential routes for dispersal across the
continent. It is very difficult to get a species such as splendida from
on end of Oz to the other in 3 million years.
For instance, consider fluviatilis and duboulayi. The former is inland,
the latter is coastal. There are many species pairs which occur on both
sides of the dividing range. Southern and Oxylean pygmy perch, mountain
galaxias, two species of Murray cod, Murrray and Marjorie's hardyheads,
two carp gudgeons and others. It is also worth noting that several
species come down the east coast as far as the Clarence River (or around
about) and then jump into the Murray system. There are known river
captures in the headwaters of the Clarence when the area was last
volcanic 20-23mya. No doubt river captures continued to occur as
rivers cut through various lava flows. Irrespective, there don't appear
to have been any recent captures indicating that the opportunity for
dispersal have been limited for a long time, seemingly more than a couple
of million years. That area appears to have been an important dispersal
route based upon todays distribution patterns. Unfortunately, we don't
know how well this reflects past distributions and we will never know for
sure. That's were genetic data can really assist in biogeographic
interpretations. Now, even I have trouble accepting these dates as it
just seems too old. However, you have to go with the best supporting
evidence of the day and try to further refine and test that hypothesis
and improve on it/change it etc.
Further to this, the apparent closer relationship between Bedotiinae and
other Melanotaeniidae to me suggests old age. Madagaskar has been
isolated for a long time from everything (I don't know how long, anyone
know?). If there was a recent ancestor it would have to have been very
widespread across the Indian Ocean. Why would it's distrition be so
disjunct today? Obviously there are many possibilities here, none of
which are testable. It's just that the two places are a long way apart.
Also, if it were a recent marine ancestor why does it only occur (or at
least primarily) on the Australian side of Weber's Line? Especially
since that line has existed for several millions years or more.
Also, if it were a recent marine ancestor why can't all but a couple of
species (blue eyes) of the Melanotaenidae (as defined by Dyer and Chernoff)
tolerate more than 50% salinity levels? Now, many folks have suggested
that the higher than normal salinity tolerance of Australian fish in
general is indicative of a recent marine origin, however, this overlooks
that fact that over Australia's history salinity has been somewhat of an
evolutionary force that fish have had to adapt to, if they couldn't they
went extinct. Few continents have "fresh" water that gets as saline as
regularly as parts of Australia.
> >Only if the genus Melanotaenia is monophyletic. This just means they all
> >came from one ancestor, vs polyphyletic, meaning that there is more than
> >one ancestor to the genus. There is some evidence that the genus is not
> >monophyletic as we presently recognise it. The work of folks at Qld Uni
> >found that M. affinis was as distinctive as G. incisus from the
> >Melanotaenia spp they examined. This suggests there is something wierd
> >going on (and also that M. affinis aren't very closely related to
> >goldies, herbies, or tris as Allen first proposed).
>
> I still believe Melanotaeniidae are monophyletic. What work?
> And most of all, what method? Do you remember my critics
> on some of the "modern" methods which create results they want to have?
Zhu, Jamieson, Hugall, & Moritz. 1994. Sequence evolution and
phylogenetic signal in control-region and cytochrome b sequences of
rainbow fishes. Molecular Biology and Evolution. 11(4): 672-683.
You can critise genetics all you want, the fact remains that the genetic
code doesn't lie, only the interpretation can. While their results may
be considered somewhat tentatively (since it is a first up effort and
not every species was examined) that was one of their most conclusive
results.
> >> The history tells us that M. trifasciata and M. herbertaxelrodi have been
> >> isolated for a very long time. While M. trifasciata developed a great lot
> >> of varieties M. herbertaxelrodi did not - because there may have been no
> >> isolation in the Lake Tebera area..
>
> >That was certainly not one of my arguements. You wouldn't expect to see
> >differentiation when you only have one population, ie Lake Tebra.
>
> Wrong, see Lake Malawi and Lake Tanganjica and, as far as we know today,
> M. boesemani and M. ajamaruensis as well as C. fasciata and C. sentaniensis.
You can't compare cichlid evolution to rainbow evolution. Cichlids are
very outstanding and unusual when compared to most other fishes. Sure,
you give two examples where I am incorrect, but how many examples are
there where this is not the case? I would suggest that your two examples
are out of the ordinary. Mind you, this is kinda off topic and a mute
point anyway (as is most of this since no one can actually PROVE
anything!). :-)
> >> In a family were species tend do develop varieties very quickly (and the
> >> tri's varieties had only about 10.000 years)
>
> >Sure, rainbows tend to develop _colour_ varieties "relatively" quickly.
> >So what? Why did the tris have only about 10 000 years to do this? This
> >may be true to a degree _if_ they occupied the lake that once existed in
> >the Gulf of Carpentaria. However, I don't think that is the reason you
> >are suggesting. You are presumably saying they only invaded Australian
> >waters 10 000 years ago? Why couldn't they have done it earlier though?
>
> I'm quite sure that a common ancestor of all tri's we know lived in the waters
> which are now partially filled with the Gulf of Carpentaria. I don't know if
> there already were varieties (I haven't been there, see above) but there was
> even theoretically a gene flow between all of the varieties living today in
> rivers ending in the Gulf. The M. trifasciata ancestor has surely been older,
> but I think that the varieties are really young, about after last ice age.
> One sign (no proof) is that you can't find two rivers which aren't connected
> having the same variety.
The varieties that live on the eastern side of Cape York and probably
some around Darwin didn't have recent contact with the lake in the Gulf
of Carpentaria. There aren't any known recent connections between these
areas, especially eastern Cape York. Bruce, wanna comment on Harro's
last comment re two rivers which aren't connected having the same (or at
least very similar) colour varieties? Aren't some of the fish around
Nhullumby similar to one another?
> >> Tri's are rather deepbodied at least when they become old. Herb's are
> >> never as deepbodied. And this is (!) astonishing as a lake dwelling species
> >> which would just be a variety of a stream dwelling should be much more
> >> deepbodied!
>
> >Perhaps tris were once a lake adapted fish, thus the reason for their
> >deep, rather than slender bodies. It's interesting to note that those on
> >the eastern side of Cape York tend to be more slender. It gets more
> >interesting when you consider that around 10000 years ago there was a
> >massive freshwater lake in the Gulf of Carpentaria (three times bigger
> >than Lake Bonneville which once covered ~1/5 to 1/4 of Utah!) to which
> >many of the rivers presently containing tris would have drained into.
>
> 10.000 years are long enough to loose a deep body. However, it's still
> a hypothesis.
How many tris actually live in an environment with strong enough flow to
select for a less deep body? Many of them tend to be pretty slow in the
dry season aren't they? Several bows, such as M. oktediensis and M.
monticola can both be relatively deep bodied yet live in quite high
gradient streams. Again, kind of a moot point that drifts off the track.
> >Further food for thought on rainbowfish ages. M. exquisita primarily
> >(only??) exist above waterfalls in the Katherine and Alligator Rivers
> >(and maybe elsewhere in that vicinity). How old are these waterfalls? I
> >would argue that exquisita were present before the waterfalls. At a
> >later point, a splendida invaded the area and exquisita disappeared either
> >through competition or hybridization below these waterfalls. It is
> >notable that the Qld Uni folks found that M.s. australis from the NT had
> >mtDNA's very similar to exquisita which they felt was was likely to
> >introgression (hybridisation) between them. And some folks firmly
> >believe rainbows don't hybridise in the wild. :-) Although, you should
> >note, this probably happened a long time ago.
>
> You're surely right with your idea on the distribution of the species.
> However, we also find this in other fish and I'm quite sure the waterfalls
> aren't very old. The are in the tropics, were landscape changes very quickly.
> However, we have to know more about the geology to be sure.
Hey Bruce (or anyone else) if ya happen to talk to Dave Wilson see if he
knows anyone that may know some age of those waterfalls above which
exquisita occur. I might try and blip Helen Larson and see if she knows.
> I think that the RML should not only consist of SPAM ;-) but also
> some more sophisticated items :->> Let's work for a good mix.
Life on the RML without spam is just not immaginable. :-) Spam is
good! spam spam spam spam spam spam spam spam spam spam.......
You get the picture..... :-)
Tootles
Peter Unmack