Chromosome counts and classification

[LJA]

Can someone explain how chromosome counts are used to classify orchids? I'm a little confused here. I thought they were used only to identify ploidy.

[Sue]

Chromosome counts help in establishing relationships between species and between genera. And larger groupings too. Similarity in chromosome number may be evidence in favor of nearness in phylogeny, whereas difference in chromosome numbers may contradict presumed similarities.

For example, genus Oncidium. When we started doing chromosome counts, we discovered that there's more variation than there ought to be, giving evidence for the judgement that genus Oncidium is artificial, and may be formed by converging evolutionary adaptations within two or more disparate lineages.

Like this:


Selective pressure may lead a variety of plants to undergo similar evolutionary adaptations, leading to the development of similarities among otherwise distantly related groups. This evolution of similar traits may obscure an underlying difference, leading to the conclusion that similar offshoots of disparate groups (A and B) are in fact a single group (group A-B), when they are in fact still quite different in underlying structure, even as they have developed apparent similarities. Chromosome count is a pretty good cue that something like this has happened, although it isn't decisive.

Does that explain it?

[LJA]

So then, all plants within a particular genus should have the same number of chromosomes, and when some don't, that points to the possibility that the weirdos may be part of a different genus altogether. Is that right?

[Sue]

Originally posted by lja
So then, all plants within a particular genus should have the same number of chromosomes, and when some don't, that points to the possibility that the weirdos may be part of a different genus altogether. Is that right?

Not quite . . .

If they don't have the same number of chromosomes then this may be evidence that they're not part of a true genus. But probably not. Many orchid genera tend to have at least a couple of different common chromosome numbers.

For example, the preponderance of Bulbophyllum are either 2n=38 or 2n=40. If a specimen measures at 60 or 80 or 76, then we might assume that the plant was polyploidal, or that the species had stabilized in what is tetraploidy or triploidy relative to the norm for the genus. Now, if the species measures consistently at, say, 2n=26, then we've got a problem. n=13 can't easily result from a n=19,20 gene pool, either through polyploidism or haploidism, even when combined with picking up or dropping a chromosome pair or two along the way.

So, variation of chromosome counts within a genus is normal. But some variation is abnormal variation. And abnormal variation can be a sign of an artificial genus.

Sorry I didn't make that more clear in the first place. For that matter, sorry I didn't make it more clear this time either.

[LJA]

I see. So the trick then is to figure out what constitutes normal variation, and what constitutes abnormal variation. And when you and Alex said that Oncidium chromosome counts were all over the place, that meant that the variation was abnormal, and that, in all likelihood, those plants with abnormal variation would not readily interbreed with plants whose variation fell within expected bounds. Is that more correct?

(I don't mean to hound you on this, I'm just really trying to understand.)

[Sue]

Yes, that's exactly right.

Oncidium includes 2n=28, 30, 34, 36, 40, 42, 44, 56, 84.

Now, 28 might be closely related to 56 and 30
28 - 30
56

and 84 is likely an established tetraploidal variation of a normal 2n=42, which may be related, in turn, to 40 and 44.

40 - 42 - 44
.......84

And then we leave 34 and 36 in their own little group.
34 - 36

So, if you split it up like that, you get three genera with more normal types of chromosome count variation.

Now, what I've just done is just playing with numbers, and shouldn't be taken seriously. But that's the reason why Oncidium looks like an artificial genus. I don't know enough about it to know how much stake to put in this.

[LJA]

I think I understand now. Has experiment borne this out? Do crosses between plants of 2n=36 and plants of 2n=56 really fail? If, given the expected variation within a genus, the chromosome numbers don't need to match up exactly anyway for a cross to take, is there some "failsafe point" where, if they differ too greatly, failure is a "sure" bet? Or is that line also fuzzy?

[Sue]

I've had difficulty with getting information about which crosses have been tried several times and failed. I pretty much just have information about what has worked. And to the best of my quite limited knowledge here, I'd say that cross almost certainly wouldn't work.

[LJA]

This helped me a lot. Thanks for explaining all of it.

[amaximia]

I would venture that generally speaking, two orchids for which 2n "consistently measured in nature" equals different numbers then they are not part of of the same genus. Doritis pulcherrima and Doritis buyssioniana belong to the same genus because though the first consistently measures 38 chromosomes and the second consistently measures 72, the first occurs in nature at 2n, while the second is almost always 4n.

The problem is, I believe, different specimens of the same species are measured many times at different chromosome counts, independent of poliploidy. This makes it harder to classify.

In any case, this is not a problem in the Cattleya grouping(2n=40) or the Vanda-Phal grouping (2n=38). There you have tremendous consistency. Taxonomists keep creating new genera in Cattleya (Guarianthe, Cattleyella) and Vanda (Amesiella) but based on things other than chromosome count. Within these groupings orchids may not breed together (i.e., phalaenopsis with paraphalaenopsis) perhaps because of the different "sizes" or "make up" of each chromosome, rather than the number of them. W. W. Moir observed difficulty in crossing Cattleya types of different altitudes and speculated the sizes of the temperature-endurance chromosome might vary, making a match difficult.

So in other words, for breeding purposes Cattleyas and Vandas are great as long as there is no poliploidy (which is frequent in awarded hybrids). Most of the others groupings are "weirder," if I may say so.

Lastly, it seems species with different chromosome counts can cross very rarely. This is bizarre. In 1976 Rumrill, who is known for exotic crosses, registered Notylidium Rumrill, a cross between Onc. sphacelatum pod parent [2N=56] and Notylia barkeri [2N=42]. The resulting hybrid appeared to have characteristics of either parent. Go figure.

I hope I didn't confure the issue more. :-)