How fashions change Solved

[Keith H]

...if you can grow one of these (M. guelzowiana) for a few years you are doing well. Have not seen one for years.LoL. Cheers

This is my M. guelzowiana in flower last August. It's still alive, as I type (or was this morning) I bought it from Rene, whose thread this is.

Mammillaria guelzowiana flowers 16 August 2013 (1).JPG

Superb plant Rod, I have just looked on a couple of lists, M.Guelzowiana would appear to still be offered for sale, another for my list of plants that I am not allowed to buy until I understand what I have left of the old collection.

Talking of fashions changing, my plants are a snap shot of an 80's collection with additions up to around the turn of the millennium when my Father stopped acquiring plants from specialist nurseries. Many of them had almost illegible name tags so I have been comparing the log of plants which he maintained to the plants in my greenhouse, to pictures on the net. Many of my plants now have names on both sides of the label, an attempt to keep the old names that he remembers and the names by which the plants are now known. (Notes on label do say 'now' and 'was' on the reverse)
I will admit to enjoying the detective work though I had not realised how variable some of the plants are, especially as many of mine look more like gnarled habitat plants, very different to the pristine specimens offered on sales benches or seen in shows.

[DaveW]

It is even less helpful to present any classification as being the definitive one Phil. It is simply the classifiers interpretation of the information before them, otherwise classifications would never change with time. I am sure in the years before DNA Sequencing you would have been supporting morphological classifications as obviously correct (as most people did), but now many of their conclusions have been shown to be doubtful. DNA Sequencing is still in it's infancy and possibly in the future it may be shown it made as many mistakes as morphological classifications did.

Even the botanical community is now recognising that changing names for new classifications is not necessarily a good thing:-

"In 1981, the International Botanical Congress voted to allow conservation of names in use for species of major economic importance and in certain other circumstances. One outcome of this was that the names of two commonly cultivated heathers, Erica carnea and E. vagans, were conserved.

At the 1993 International Botanical Congress in Tokyo, the power of such committees was extended further so that:
it is now no longer necessary to prove major economic importance to have the name of a species conserved, and any name (at any rank) that might cause nomenclatural instability can now be proposed for rejection."

"One cannot prevent the reclassification of plants into different genera, e.g. the common chrysanthemum has been put into Dendranthema, the genus Sedum is being split up, Zauschneria has been put into Epilobium, Pratia into Lobelia and Pernettya into Gaultheria. There are even whisperings of threats to lump Mahonia into Berberis! However, there are no botanical rules dictating which classification one should adopt and the above examples are still disputed by some botanists. Not long ago, the garden petunia was allowed to retain the Latin name Petunia, despite a proposal to place it in another genus.

Times are changing, and today the majority of botanists and taxonomists are working towards an objectively argued classification system. Many existing systems are somewhat artificial because they reflect the viewpoints of individual taxonomists, rather than attempting to show the pattern of the way plants actually evolved and how they are related to one another."

http://www.kew.org/science/names.html

Most now agree that ideally a classification should group plants together that share a common ancestor. However unless we consider cacti evolved from more than one source they all already share a common ancestor, so it then becomes simply a matter of opinion as to how many useful groups you split them up into, which are hopefully monophyletic.

John Donald once said to me plant classification is like a tree submerged in a lake so only the tips of its topmost branches are showing and these represent our current species. We can only therefore guess as to how the branches join the other unseen branches and the trunk below the water since we can really never know for certain. Plus also there are no known cactus fossils since the supposed Eopuntia douglasii was later proved not to be a cactus fossil at all.

Classifications go in fads and I doubt there has ever been a time there was only a single classification in use around the world. At present, as you will know Phil from an American site we are both on, some of the American botanists have reservations about the NCL classification of their plants, maintaining European botanists have not spent the time in the field they have to be able to properly classify them.

[CactusFanDan]

All lifeforms on earth share a common ancestor. That doesn't mean it's wise to group everything into one gigantic genus. That wouldn't be useful. The way a genus is constructed nowadays (should ideally, in my opinion) present a pay-off between monophyly (representing evolutionary relationships), and the usefulness of grouping plants in such a way. I've seen this done in quite a few papers. This is where not only are the phylogenetic trees taken into account, but also morphological similarities.

For example, you could propose a monophyletic genus Rapicactus including the tradtional Rapicactus (R. mandragora, R. subterranea etc.), Acharagma, Lophophora and Obregonia. That would be a monophyletic genus, but of course it wouldn't be very useful, since they have obviously different characteristics. Splitting them up in this instance would be more useful. And then there would be instances in which lumping several genera would be more useful, based on morphology and phylogenetics. But as Dave says, classifications are plastic and liable to change.

[DaveW]

The trouble Dan is in recent decades we have moved from ultra splitting to ultra lumping and the latter has produced some super genera like Echinopsis sensu lato that are really no longer useful for quickly identifying plants, one of the main requirements of a workable classification. Is there really any logic in dismissing the remark that you could lump all the Cactaceae into a single genus if you were an ultra lumper and then proposing these super genera like Echinopsis sensu lato other than degree, since neither aids identification or advances classification?

Whilst I am not a splitter, there are advantages to smaller genera for classifiers since they can be moved around like pieces on a chess board and put into different relationships with each other in view of new knowledge without producing a host of unnecessary new combinations that in the future will often be relegated to synonymy. We are already starting to see some of the NCL over lumping being dismantled in the light of new knowledge, so those combinations will no doubt in future simply clutter up the synonymies of our plants. I think what the recent exercise has proved is you can't produce satisfactory classifications by committee's.

Regarding DNA Sequencing and genetics for classification, how sure can we be that present methods are fool proof, or we yet understand them?

http://www.nature.com/news/2005/050321/ ... 321-8.html

The following statement in the above link could have relevance for our plants:-

"Indeed, the process could exist because it helps plants to survive whenever they find themselves in difficult condition, such as when water or nutrients become scarce. Such stress could trigger plants to revert to the genetic code of their ancestors, which is perhaps more hardy than that of their parents. To test this, Pruitt is examining whether stressful situations do indeed prompt the same phenomenon."

http://www.purdue.edu/uns/html4ever/200 ... tance.html

[Wizard]

I have some interesting picture of Krainzia guelzowiana:

[CactusFanDan]

Regarding DNA Sequencing and genetics for classification, how sure can we be that present methods are fool proof, or we yet understand them?

http://www.nature.com/news/2005/050321/ ... 321-8.html

The following statement in the above link could have relevance for our plants:-

"Indeed, the process could exist because it helps plants to survive whenever they find themselves in difficult condition, such as when water or nutrients become scarce. Such stress could trigger plants to revert to the genetic code of their ancestors, which is perhaps more hardy than that of their parents. To test this, Pruitt is examining whether stressful situations do indeed prompt the same phenomenon."

http://www.purdue.edu/uns/html4ever/200 ... tance.html

Turns out those findings were due to contamination of the samples.

http://www.nature.com/news/2006/060925/ ... 925-7.html

We have a good understanding of genetics and underlying genetic principles. Our knowledge is nowhere near complete, but it's good enough to get quite reliable results in terms of classification. Currently, the number of genes that are sequenced in angiosperms for classification is being increased. The more genes you use, the more you can be sure of the relationships between species. And new genes that are ideal for DNA barcoding are being searched for. It's a very powerful tool for the taxonomist, but not necessarily always the most appropriate on it's own.

[DaveW]

Yes one should never trust scientific results from one team that are not corroborated by other non connected workers. Interesting though is the other example quoted. Has this been corroborated by others?

http://www.nature.com/news/2006/060522/ ... 22-13.html

I suppose therefore that proves we should never trust classifications based on DNA Sequencing by one team unless they are backed up by similar results from other non connected workers?

[CactusFanDan]

Yes one should never trust scientific results from one team that are not corroborated by other non connected workers. Interesting though is the other example quoted. Has this been corroborated by others?

http://www.nature.com/news/2006/060522/ ... 22-13.html

I suppose therefore that proves we should never trust classifications based on DNA Sequencing by one team unless they are backed up by similar results from other non connected workers?

It's okay to trust results from such a study group to some extent, but you should always view their findings with scepticism, especially if their paper hasn't been through the rigorous peer-review process. With papers like the one above, where the results are totally unexpected, but the method used and the conclusions drawn from the data are sound, then other research groups will try and repeat the experiment. If results are unrepeatable, then there's obviously a problem. If a paper has been peer-reviewed and has sound results, then there's no reason why you shouldn't trust the findings of such a paper. However, looking upon any scientific paper with scepticism can be useful.

Mouse genetics is interesting, but some basic genetic principles don't cross over between mammals and plants. Genome structure is different in flowering plants, along with the frequency of genetic recombination and the ways by which gene products are organised post-translation.

With scientific studies, repeats and statistical tests are useful to drive down error. Now classification experiments tend to be a whole different kettle of fish and it's to do with how a phylogenetic tree is constructed. In order to sequence a gene you need millions or billions of copies of the DNA comprising that gene and in doing so you essentially find the average sequence out of millions of copy for that gene from the individual being examined. Within a single organism there might be occasional inconsistencies in a gene due to error-prone DNA replication and stuff, but they will be present in tiny numbers (maybe 100 or less out of 1 million copies). That helps to reduce the error in your findings. Once you've got your sequences you're ready to construct a tree and one way to do so is by what's called 'bootstrapping'. Each 'bootstrap' uses a very complex mathematical model which tests how likely it is that certain relationships occur. You often see the numbers on phylogenetic tree diagrams (eg 98/100 = 98 bootstrap repeats out of 100 agree with this relationship) representing bootstrap repeats and the number of these repeats which agree that that relationship should be there. You could do 1000's of bootstrap repeats if you wanted, but they take a heck of a lot of computing power and time. Basically, the more bootstraps, the more you can be sure of a relationship. And then going back to what happens if you get an unusual or unexpected result: you can repeat your experiment.