Javascript Disabled Detected

You currently have javascript disabled. Several functions may not work. Please re-enable javascript to access full functionality.

Snubnose Darter (Etheostoma simoterum) species diversity

Started by Guest_TomNear_* , Oct 13 2011 10:58 PM

Reply to this topic

16 replies to this topic

#1 Guest_TomNear_*

Guests

Posted 13 October 2011 - 10:58 PM

Attached is a paper in press that uses genetic data and information on male nuptial coloration to argue that there are three, and not six, species in the Etheostoma simoterum complex. Photos of nuptial males are attached at the end of the PDF as an embedded jpg (hence the large size of the file).

Attached Files

Harrington&NearInPress.pdf 3.02MB 102 downloads

#2 Guest_blakemarkwell_*

Guests

Posted 14 October 2011 - 12:01 AM

Thanks for the paper -- I definitely saw a prelude to this in your most recent nuclear gene phylogenies. Out with E. occidentale, E. orientale, and E. tennesseense and E. atripinne, E. planasaxatile, and E. simoterum stay. Three species of the E. simoterum complex in the Cumberland always did cause me to scratch my head, but haven't you found a similar level of diversity with N. sanguifluus in the Cumberland? It appears you found molecular support for all the recent E. spectabile splits, which were also largely delimited on the basis of nuptial male pigmentation patterns. The E. spectabile complex is definitely a bit more widespread than the E. simoterum complex! Man, those nuclear gene trees were worthless for supporting the monophyly of any of the E. simoterum species.

Edited by blakemarkwell, 14 October 2011 - 12:33 AM.

#3 Guest_TomNear_*

Guests

Posted 14 October 2011 - 09:22 AM

Blake-You bring up a good point that allows me to offer some clarity. I do not think that species have to exhibit substantial genetic divergence, just like some species can exhibit genetic divergence but little morphological differences. However, when species exhibit neither morphological or genetic divergence, species delimitation for those groups is less likely. If you look at the meristic data presented in Powers and Mayden (2007) there is substantial overlap among the three species delimited in the Cumberland (E. atripinne) and the there is more variance in the counts for E. tennessenese than any other species. The reasons for the lack of variation in the Cumberland species is because they are not different and the wide variance in E. tennessense is driven by the realization that populations of E. tennessense in the Upper Tennessee River are more similar (lower counts) than populations in the lower part of the system. This is not clear in the tables in Powers and Mayden (2007), because they did not break up the counts by regions of the Tennessee River System.

I think that male nuptial color patterns are valid and fine characters to delimit darter species. However, Rich's analysis of 104 photographs of male specimens show that the diagnostic traits are not quite diagnostic in this clade.

-Tom

#4 Guest_bpkeck_*

Guests

Posted 17 October 2011 - 09:50 AM

Three species of the E. simoterum complex in the Cumberland always did cause me to scratch my head, but haven't you found a similar level of diversity with N. sanguifluus in the Cumberland?

There is a lot of diversity in the Cumberland River and many groups have diversified within the drainage, just look at the Oopareia, Barcheek Darters! Way more diversity in that group than most other groups within the Cumberland.

The N. sanguifluus and N. microlepidus relationships are a good illustration of what Tom is saying. These two Nothonotus have very distinct and different male nuptial coloration, with N. microlepidus having bright green median fins and N. sanguifluus having reddish median fins; however, N. microlepidus and N. sanguifluus from the Rockcastle River and Buck Creek share mitochondrial haplotypes (not many) suggesting a very close relationship. In this case there are two species that are genetically pretty similar relative to genetic similarity observed between other species of darters, but have very distinct coloration. Interestingly, the N. sp. cf. sanguifluus from the Caney Fork looks very similar in color to the N. sanguifluus from the Rockcastle, Big South Fork, etc., but shares less genetic similarity to the other N. sanguifluus and N. microlepidus than the latter two share with each other. In fact, the estimated phylogeny for these species and populations has the N. sp. cf. sanguifluus from the Caney Fork as sister to a clade comprised of all the other N. sanguifluus and N. microlepidus. So, in this instance we have a species, N. sp. cf. sanguifluus (description complete and waiting to be submitted), that is genetically identifiable, but the coloration differences are not obvious.

#5 Guest_blakemarkwell_*

Guests

Posted 17 October 2011 - 02:38 PM

Thanks for the information! However, it leaves me with quite a few questions off the top of my head:

How much older/younger of a clade is Oopareia in relation to Nothonotus and Adonia/Ulocentra? What's responsible for the disparity in species diversity within the same drainage if they've had similar life histories (selection pressures, dispersal capabilities, etc)?

As for N. microlepidus and N. sanguifluus: is the mitochondrial introgression unidirectional, and if so, what's the direction? Have you guys tried microsatellite data for some of these species groups? I know as the genetic distance increases, so does the amount of noise to signal with microsatellites, but I'd think it would work fine for species this closely related. It seems most nuclear genes don't provide the desired resolution with species this closely related, and mtDNA seem to fail due to introgression. It'd be interesting to see how E. orientale, E. occidentale, and E. atripinne line up with microsatellites. It seems y'all have had good node-support using nDNA for less specific groupings (subgenera, etc) and similar support using mtDNA for species-level phylogenies -- at least that's intuitive.

#6 Guest_farmertodd_*

Guests

Posted 17 October 2011 - 03:34 PM

Some evidence from ecology.... The mid-domain of clades Oopareia and Nothonotus correspond to the same underlying geology. There's a fold of calcareous siltstone that runs from French Creek in PA across OH, KY and TN to Shoal and Bear Creeks in AL, but that also branches northward into IL on the Vermillion from KY around the Green. Both species require large, plucked clasts for nest habitat - which is provided by this rock type. The rock type would have presumably been available along each of these rivers due to first uplift and then erosion since the time of the Appalachian orogen. Who knows what localized Oopareia were lost during glaciation from Illinois, Indiana and Ohio, it makes for fascinating speculation!

The Duck and Cumberland would be excellent places to start quantifying processes at local scales. Hypotheses related to habitat partitioning in feeding morphology, time of breeding, and longitudinal zonation should be entertained.

I don't have my head around snubs

Todd

#7 Guest_blakemarkwell_*

Guests

Posted 17 October 2011 - 04:11 PM

Thanks for the input, Todd. A few questions from this perspective as well....

Why isn't Ooparia found in Indiana, the Middle-Fork Vermillion of Illinois, or anywhere in the Interior Highlands? N. camurus and N. maculatus sure made it into Indiana (and N. camurus into IL via the Wabash). N. juliae and N. moorei also made it into the Interior Highlands, although N. juliae is the basal taxon of Nothonotus, so saying 'made it into' isn't quite appropriate in this context. I realize these questions don't have an answer, but it begs the question when looking at ecological similarities and distribution.

Yeah, the snubs drive me crazy as well. Once again, I can't figure out why they aren't in the Interior Highlands either.

Edited by blakemarkwell, 17 October 2011 - 04:50 PM.

#8 Guest_fundulus_*

Guests

Posted 17 October 2011 - 05:58 PM

My experience with local snubs in north 'bama is that they're apparently the least picky, most broadly tolerant of the local darters. Our (just finished) habitat partitioning data from the last year in Estill Fork of the Paint Rock, and the main stem of the Flint River east of Huntsville, show large niche overlaps of black snubs with banded darters in the Flint but not with redlines. In Estill Fork the tennessee snubs show moderate to large niche overlaps with rainbows, stripetails and fantails, but again very low niche overlap with redlines. And in both systems the snubs are about the more abundant of the darter species, although in the Flint the bandeds are almost as common.

#9 Guest_farmertodd_*

Guests

Posted 17 October 2011 - 07:11 PM

Juliae and moorei weren't in glaciated regions. Camurus, maculatus and tippecanoe cover some of the largest geographic area among the clade, presumably because they had refugia in the Kentucky, Licking and Kanawah, which allowed them to recolonize. However, you don't find any Ooparia in those river systems (unless I'm forgetting something in the Kentucky).

Ben, what were the genetic distances of these things? Like it was one big population, wasn't it? I'm sure there was some distance for stuff in the Green, and maybe up the Elk, but I'd guess that'd be about it.

This reminds me, I need to use GIS to figure out where Glacial Lake Tight would have been. There needs to be maps, dangit.

I look forward to seeing your results Bruce!

Todd

#10 Guest_blakemarkwell_*

Guests

Posted 17 October 2011 - 07:42 PM

Juliae and moorei weren't in glaciated regions. Camurus, maculatus and tippecanoe cover some of the largest geographic area among the clade, presumably because they had refugia in the Kentucky, Licking and Kanawah, which allowed them to recolonize. However, you don't find any Ooparia in those river systems (unless I'm forgetting something in the Kentucky).
Todd

Yeah, I guess I shouldn't have added species examples in my first response. What I was getting at is when looking at the most speciose clades in Etheostoma, nearly all of them (Nothonotus, Catonotus, Doration, Poecilichthys, Neoetheostoma, Oligocephalus, Litocara, etc) have species in both the Interior and Eastern Highlands. Yet, Adonia/Ulocentra (snubs) and Oopareia (although these guys, like you said are dependent on specific nesting habitat) are restricted east of the Mississippi. It just begs the question why, but I realize that it's hard to do a vicariance study when they simply aren't present west of the Mississippi. Like Bruce has stated, they aren't picky (from E. pyrrhogaster in the Hatchie to E. simoterum in the mountains), and they hold their own where they're present. I guess what they lack in the Interior Highlands, they make up for in Alabama.

Edited by blakemarkwell, 17 October 2011 - 08:31 PM.

#11 Guest_TomNear_*

Guests

Posted 18 October 2011 - 08:17 AM

My thoughts is that the geographic patterns of clades are quite idiosyncratic. We spent a lot of time looking for common patterns and shared vicariant mechanisms, but I think in the end the shared history of isolation will be minimal. One idea that I had not explored very thoroughly is treating the disjunct highland areas as "islands" and thinking about patterns of biodiversity in the context of island theory, where island size and distance from the mainland are predictive of patterns of immigration, extinction, and speciation.

Blake, you should come join our group here at Yale and help us figure out these long-standing problems!

#12 Guest_farmertodd_*

Guests

Posted 19 October 2011 - 08:30 AM

Blake,

Here's a hypothesis: Ulocentra evolved in the Tennessee/Alabama River system before it jumped Walden's Gorge, and the species outside the mid-domain that you're focused on are the radiations from different capture events. The Duck/Collins/Elk, apparently, are the only other HUC 8's with 2 species, and they share a fair amount of connectivity on a geological scale (and if I'm correctly HUC categorizing the Barren and Green).

Doration follows a similar pattern, although stigmaeum is more tolerant of low productivity, and found a nice home in the southern Ozarks. And pyrrhogaster, for example, just never found a nice, big, productive alluvial system on the west side of the river with all that cold muddy water coming down the pipe for 2.5 mya.

There really isn't good siltstone in the surficial bedrock immediately west of the Mississippi. A lot of it is dolostone that fractures on a very short A axis and the Ozarks are volcanic, old and very very weathered.

Run with it!

Todd

Edited by farmertodd, 19 October 2011 - 08:31 AM.

#13 Guest_TomNear_*

Guests

Posted 19 October 2011 - 12:14 PM

A man much wiser than me once said "darters know rocks."

#14 Guest_bpkeck_*

Guests

Posted 19 October 2011 - 01:07 PM

I'll try to get at a few of these:

As for N. microlepidus and N. sanguifluus: is the mitochondrial introgression unidirectional, and if so, what's the direction? Have you guys tried microsatellite data for some of these species groups? I know as the genetic distance increases, so does the amount of noise to signal with microsatellites, but I'd think it would work fine for species this closely related. It seems most nuclear genes don't provide the desired resolution with species this closely related, and mtDNA seem to fail due to introgression. It'd be interesting to see how E. orientale, E. occidentale, and E. atripinne line up with microsatellites. It seems y'all have had good node-support using nDNA for less specific groupings (subgenera, etc) and similar support using mtDNA for species-level phylogenies -- at least that's intuitive.

This is the Nothonotus paper discussing N. microlepidus and N. sanguifluus:

Keck&Near2010b.pdf 401.55KB 309 downloads

The mitochondrial haplotype sharing is probably due to unsorted ancestral polymorphism and not due to introgression, because these two species have never been collected in the same river systems. Introgression can't be completely ruled out, but patterns in the nuclear genes support a very recent divergence as well supporting ancestral polymorphism. The nuclear genes we have sampled do not resolve these, or any relationship among the species in the N. maculatus species group, if analyzed individually, but in a species tree method they do pretty well. Microsats, AFLPs, and SNPs would probably identify lots of structure, but inference is up to the author.

The attached paper points out the disparity in ages between the Oopareia and Nothonotus, but we basically say that the Cumberland is just damn old and fish have a tendency to not move around much once they establish. The Cumberland has the highest number of endemic darters per area than any other river system, it'll be interesting to see what we find in the other major fish groups.

Juliae and moorei weren't in glaciated regions. Camurus, maculatus and tippecanoe cover some of the largest geographic area among the clade, presumably because they had refugia in the Kentucky, Licking and Kanawah, which allowed them to recolonize. However, you don't find any Ooparia in those river systems (unless I'm forgetting something in the Kentucky).

Ben, what were the genetic distances of these things? Like it was one big population, wasn't it? I'm sure there was some distance for stuff in the Green, and maybe up the Elk, but I'd guess that'd be about it.

Todd

I've got N. camurus and N. maculatus sampled throughout their ranges and they are pretty much genetic duplicates throughout the previously glaciated region. The highest genetic diversity (granted just for cytb) for N. maculatus is in the Green River, KY, and I forget the actual distances, but they were really small, less than 1% for sure. The Tennessee R. is the source population for all the northern N. camurus. There's only one haplotype (by my sampling) found in populations from IL, IN, OH, PA, NY, WV, and KY excluding the Cumberland and Tennessee rivers. This haplotype is also found throughout the Tennessee R, even all the way up the Clinch.

#15 Guest_blakemarkwell_*

Guests

Posted 20 October 2011 - 10:25 PM

Thanks, y'all! Yeah, sorry about the first question. I realized after posting that they were allopatric, so by sharing you meant recent ancestry and not introgression (I guess I had the other Nothonotus infamous for introgression on my mind!). Seeing the redundant haplotypes among the northern Nothonotus makes sense, given that it would be a rapid recolonization (presumably). They've found similar results with Ambystoma laterale (Blue-spotted Salamander) -- I guess specimens collected from Maine and Illinois differed by a single nucleotide! Although a bit more recent, as they occur primarily in areas covered by the LIS.

Thanks Todd -- that's definitely a start! :biggrin:

Edited by blakemarkwell, 20 October 2011 - 10:27 PM.

#16 Guest_farmertodd_*

Guests

Posted 21 October 2011 - 01:08 PM

Thanks for the response Ben.

That's ridiculous! My buddy did work with 4-toed salamanders and only found 2 clades across the northern part of the range. But a nucleotide? Wow.

Todd

#17 Guest_TomNear_*

Guests

Posted 21 October 2011 - 01:27 PM

Percina macrocephala from the Barren River in Kentucky and French Creek in PA share mtDNA haplotypes, meaning not a single difference. Recent and sudden expansion!