Lifeline book fair

The Lifeline book fair takes place every six months at the exhibition park across the road from us, once in spring and once in autumn. It is a fantastic institution. People donate books that they don't want anymore, and they are sold to help finance the eponymous crisis telephone.

It is an odd mixture of old and new, crappy and good, nonsensical and valuable. I was slightly annoyed to find some creationist literature in the natural sciences section. On the other hand, I was amused that, as seen in the photo below, Left Behind was in the fantasy and science fiction section, something that would probably have annoyed Christians who (like the authors themselves!) actually believe the rapture described in that novel is going to happen and would expect it to be part of the religion section at least.

Ah well. As my wife remarked, there are so many books that the volunteers probably only take one superficial look at most of them to decide where they should go.

The fair starts on Fridays, when masses of people compete to grab the rarest and most interesting books. We came today, on its last day, when they start lowering prices, but that also means that the best may already be gone. Still, we picked up a lot of nice things: An Indian cookbook, the plays of Goethe (in German), Martin Millar's The Good Fairies of New York, three of Jule Verne's novels in one volume, children's books, games and even a nicely illustrated plant anatomy textbook as well as volumes two to four of the Flora of South Australia. Volume one wasn't there, sadly, and obviously being published in the forties and fifties the flora is pretty out of date. But it is nice to have it for my private use, and the most frequent species were usually already described then.

So why is incomplete lineage sorting not an issue at higher taxonomic levels?

When I wrote about incomplete lineage sorting some time back, my main point was that it is an often neglected problem at lower taxonomic levels. One cannot assume that one will infer the true species phylogeny based on only one sample per species and/or only one gene region because many species may have inherited some ancestral polymorphism. Different genes and different individuals may tell conflicting stories. (A good example is the publication from a year ago that found a good portion of the gorilla genome to be more closely related to that of humans or chimpanzees although the overall evidence clearly shows the latter two to be sister species.)

I mentioned also that incomplete lineage sorting is not a problem at higher taxonomic levels, such as when we want to figure out whether a genus or subfamily of plants is monophyletic. The same was recently stated confidently when I was meeting with a few colleagues over lunch.

But why, actually? One might wonder how a problem that makes it harder to infer the true relationships of closely related species A, B, C and D, with some genes saying ((A,B),(C,D)) and others saying ((A,C),(B,D)) would suddenly disappear fifteen millions later. Surely if we want to infer the phylogenetic relationships of four clades A', B', C' and D' that have descended from those four species we will run into precisely the same problem?

Well, I guess so: if these four most closely related species all diversified into clades over those fifteen million years that are happily alive today, then the problem remains because all information that is available to infer the relationships of A'-D' is the information that we could have used to infer the relationships of A-D when they were still only four closely related species. A gene that was fixed in the four lineages so that it tells the story ((A,C),(B,D)) although the real species phylogeny is ((A,B),(C,D)) would mislead us equally in both situations.

No, the real difference between the two situations becomes clear when we think about the likelihood of four species from fifteen million years ago all surviving - it is vanishingly small. Most of everything goes extinct. Just as most seeds do not get to be mature plants and most eggs do not get to be mature animals, most species do not diversify into clades but instead go extinct, and most small clades do not diversify into large clades but instead go extinct.

Because of that, it is quite unlikely that the crown groups of the deeper clades whose relationships to each other we want to infer today are derived from a group of very closely related species (their stem groups necessarily are, by definition, but that is besides the point because we cannot sample anything but the crown group). Instead of clades A'-D' in a relationship of

((A',B'),(C',D'))

what we will mostly find are clades A'-D' in a relationship of

(0,(0,((0,(A',((0,(((0,0),((0,(0,(0,0))),(0,(0,(0,0))))),(0,0))),((0,(0,(0,(0,(0,0))))),((0,(0,(0,(0,(0,(0,(0,(0,(0,(0,(0,(((0,(0,(0,(0,0)))),((0,(0,0)),((B',0),(0,0)))),(0,(0,0)))))))))))))),(0,(0,0))))))),(((0,(0,(0,0))),(0,0)),(0,((0,0),(((0,((0,0),((((0,0),(0,0)),(0,(0,(0,0)))),(0,(0,(0,(0,((C',(0,0)),(0,(0,0)))))))))),((0,(0,(0,((0,(0,0)),(0,0))))),((0,0),(0,(0,0))))),(((D',0),(0,(0,0))),(0,0)))))))))

with "0" representing all the related species that have gone extinct over those fifteen million years and thus will never turn up in a molecular analysis.

Clades (sections, genera, tribes, etc.) that are alive today are in most cases derived from ancestral species that were far enough apart on the phylogeny to have accumulated additional synapomorphies along the branches studded only with extinct side lineages. Ultimately, it is not the completion of lineage sorting over time, i.e. the extinction of gene families within species, but instead the extinction of entire species that ensures that incomplete lineage sorting is not a problem for inferring higher level relationships.

Botany picture #48: Loiseleuria procumbens

Loiseleuria procumbens (Ericaceae) from the European Alps, 2004. A cushion-forming dwarf shrub with a wide northern hemsipheric distribution in alpine and arctic areas. The flowers are mostly 5merous but there is the odd 4merous one.

Botany picture #47: Catalpa

Inflorescence of a Catalpa (Bignoniaceae) in the Botanic Garden of Göttingen, Germany, 2004. Bignoniaceae are a family of mostly tropical to subtropical woody plants - trees, lianas or shrubs. They are often quite stunning, with some trees in dry tropical areas covered in massive displays of flowers even while leafless in the dry season. Obviously, many species are ornamentals such as this one.

I still fondly remember how exasperated our Swiss plant anatomy lecturer was when she was asked by some tourists whether the tree depicted in this photograph was an orchid. I mean, it stands to reason, doesn't it? Pretty flower = orchid, that's all the botanical knowledge one needs, durr...

Unfortunately, another equally pretty Catalpa tree in the same botanic garden was knocked over by a storm around the same time (2005?), right towards the botanical institute, and the branches smashed several windows.

Plagiarism, humanities and sciences

My home country has recently seen a distressing profusion of cases of plagiarism among politicians with academic degrees. It all started, as far as I know, with then defense minister Guttenberg who from what has been reported appears to have plagiarized most of his dissertation. What was new about the case was that the dissertation was, after the first few discoveries by academics, thoroughly and publicly examined by numerous contributors to a wiki ("GuttenPlag") in a crowdsourcing approach. From there on, activists started to examine the dissertations of other high-ranking politicians with doctorates, and soon several more cases of plagiarism were found (and one person was charged but then exonerated).

There are two interesting patterns here. First, those politicians are generally conservatives or liberals* as opposed to social democrats or greens. I can only speculate why that is so. An obvious suspicion that has of course been advanced in discussion threads on the internet is that the anti-plagiarism activists selectively target right wing politicians. Said activists seem keenly aware of that charge and take pains to stress that their work is not party-political.

At least part of the answer might be that conservative and liberal politicians are simply more likely to have doctorate degrees. Here is a statistic for the German federal parliament listing the number of doctorates the members from each party have, and the greens have the lowest proportion, followed by the social democrats. This is probably not because academics are more likely to be conservatives - it is actually the German green party that has the most educated supporters. But there might be a cultural issue here about conservatives placing greater value on official titles. Surely a career politician does not actually need a doctorate to be successful in their chosen career, so the only motivation to obtain one that makes sense to me would be the prestige associated with it. But I will be the first to admit that this is speculation.

The second interesting pattern is that the dissertations in question are generally from the humanities as opposed to the natural sciences. Karl-Theodor zu Guttenberg obtained a doctorate in jurisprudence, Annette Schavan in theology and philosophy, Silvana Koch-Mehrin in macroeconomics and history, Margarita Mathiopoulos on what sounds like a historical topic (?), Jorgo Chaitzimarkakis in political science...

It appears logical to assume that politicians would on average have a preference for the humanities, especially sociology, law, politics and history, but not in all cases. The current German chancellor, Angela Merkel, is famously a physicist. So there are natural scientists in politics, but still so far it appears that plagiarism may not be such an issue among them.

And if you follow the news, and look beyond German politicians and at actual career scientists worldwide, you will notice that plagiarism is simply not a relevant issue in the sciences. Of course there is some fraud, but it isn't plagiarism; the frauds committed by scientists are usually (1) inventing data and (2) manipulating data to produce convenient or spectacular "results". The reason is surely down to a fundamental difference between what scholars in the humanities and natural scientists do.

More from the ANBG

We had another nice walk through the Australian National Botanic Gardens today, together with friends.

You may remember the recent botany picture, Azolla pinnata. This is the second native species of this interesting genus, Azolla filiculoides.

Azolla filiculoides differs from A. pinnata in the less regular branching pattern and in its unbranched roots. Personally, I find A. pinnata more attractive.

Many Banksias are just starting to flower. This is Banksia aemula (Proteaceae).

Lambertia formosa, also of the Proteaceae, a shrub found in coastal heath of New South Wales.

I just realized that this is my hundredth blog post. Hooray!

Botany picture #46: Salix x rubens

Male inflorescence of the hybrid willow Salix x rubens (Salicaceae), if I identified it correctly. Willows are one of the very few genera that are really diverse and hard to identify in botanically impoverished central Europe. You can actually find creeks that you can walk along and find four or five species of the genus in a few minutes. The Rothmaler flora of Germany contains three different identification keys: one for female plants, one for male plants, and one for sterile material. Still, identifying these plants is not easy, not least because they hybridize like crazy. Here in Australia a few species of willows are aggressively invasive weeds, and you don't really want to see them, but they are one of the few genera that I miss a bit (others are Papaver and Allium).