Pitfalls of the long branch

Long branches in phylogenetic trees represent lineages of  organisms that have been around for a long time, but exist today as only one or a few species.  A few years ago, I discussed two examples in detail: the monocot genus, Acorus, and the archaic angiosperm species Amborella trichopoda.  I feel that the topic is worthy of a review, especially for newer readers who may not have gone back to the older posts. 

In both cases, these lineages branched off very early, over 100 million years ago, but have left no fossils, and have no close living relatives.  The Amborella branch is the earliest surviving lineage of angiosperms in general, while the Acorus branch is the earliest surviving lineage of monocots.  Expressed in a different way, Amborella is the sister group to all other angiosperms, and Acorus is the sister group to all other monocots.

At the level of phylogenetic analysis, such long branches have often been problematical, with "long branch attraction" leading occasionally to errors in the resulting phylogenetic tree.  This has been much discussed, and there are ways to correct for it, but this is a very technical issue. If you want to learn more, you might begin with .  Begin with this Wikipedia article, and go from there.

 In both cases, however, many phylogenetic analyses have confirmed the ancient position, and length of these two branches, so that is not a question  here..

Such long branches can lead to errors of interpretation at another level, however. A common misconception is that what we see in the current species, which occupy the very tips of these ancient lineages, will be similar to the  ancestors from which the lineage began, i.e. that these are archaic or primitive species.

But think about it.  These lineages have been around for more than 100 million years  (140 million for Amborella, 120 million  for Acorus).  Isn't it likely that the occupants of these lineages have changed somewhat over all those years?

Amborella fruits are single-seeded drupes, adapted for
dispersal by fruit-eating birds.  This is a specialization
that has evolved many times among angiosperms, including
most famously, cherries. Early angiosperms most
likely had fruits that split open to release several to many
seeds (see Were the first carpels plicate or ascidiate? . 
Small, unisexual flowers in dense clusters
 is also a specialization. Photo courtesy Joel McNeal.

Modern phylogenetic analyses are based primarily on molecular (DNA) comparisons, so in-and-of themselves tell us nothing about changes in the characteristics of the plants occupying the lineages.  So there is no direct basis for inferring what the first species in a lineage looked like or in  what ways their modern descendants may have changed.

As I argued in the previous posts, both Acorus and Amborella, as they exist today, exhibit a mix of ancient and specialized characteristics. They are both well-adapted to their environments, and have some distinctive specialized characteristics, particularly in their adaptations for pollination and seed dispersal. The Acorus and Amborella lineages have been around for such a long time, that it is rather absurd to think that they have not changed at all during that time. For groups that have good fossil records, we can trace such changes.  Fossils, for example, tell us that we modern humans have changed a great deal from the first members of our genus, even more from the ancestral genus Australopithecus!

The spadix-like inflorescence of Acorus led early
taxonomists to classify this genus with the Aroids.
Since the two families are not closely related, it is likely 
that the similarity is due to convergent evolution, driven by
adaptations for pollination. A spadix is a highly specialized
 way to arrange flowers and has evolved independently in a
number of families, including the Aroid, Palm, and 
Cyclanthus families. It is likely that the early monocots had
looser arrangements of flowers, more like those in most
Alismatales, and that dense flower spikes were not
characteristic of the first members of the lineage.
The folded and fused (equitant) leaves of Acorus, are
also a specialized adaptation that has occurred in many
unrelated families, most famously in several members
of the Iris Family.

How do we know, or at least develop hypotheses, as to what changes have taken place in a lineage in the absence of any fossils?  We can look at the characteristics of other early branches to see what they have in common, and hypothesize that the shared characteristics were present in their common ancestor.

 We can also analyze how particular characteristics might have arisen as adaptations to natural selective pressures, and determine which are most likely ancestral, and which are more specialized. Adaptations arise in logical sequences and often become canalized in non-reversible directions (see What is an adaptation? and G. L. Stebbins and the process of adaptive modification)

Both in comparison with other related groups, and in considering likely sequences of adaptations, Amborella and Acorus are specialized in some ways. For Amborella, small, numerous, unisexual flowers in clusters, and red, single-seeded fruits are both features that are more specialized than in other archaic angiosperms. For Acorus, the dense spikes of flowers with fused carpels (see also Were the first moncots syncarpous?) and the the leaves with the two sides fused together (equitqnt are specialized features, that have evolved independently in a number of families from more generalized types.