This essay is a brief description of a piece of research published just prior to Christmas, showing how butterfly leaf mimicry may have evolved. The relevant research is the following article:-


Gradual and contingent evolutionary emergence of leaf mimicry in butterfly wing patterns


The research relied a lot on Bayesian inference and statistics which is a branch of statistics that has, over the last few decades become increasingly used by mathematicians and scientists. This is because the statistics that is normally taught in schools relies on current data only. And so, the likelihood of one event occurring out of several is determined by merely counting the occurrences of all events, and forming a ratio. To form this ratio, the frequency of occurrence of one event is divided by the frequency of all events, and thus the likelihood is obtained.


With Bayesian statistics however, a bit of smart is incorporated into this procedure. The likelihood of an event may well depend on what happened previously and so with this particular kind of statistic, historical data is taken into account in order to asses the likelihood of a current event happening. In principle, its a lot more realistic way of doing things.


Given its reliance on prior information its a good tool to be used in evolutionary studies when assessing likelihoods of various outcomes.
A good description of the process can be found at the following link:-
Bayesian statistics for dummies
So what was this research all about?


Leaf mimicry is one of those hard evolutionary problems. In large part its hard to understand how, in a classical Darwinian scenario, it could have gradually evolved given that the intermediates would have nowhere near looked like the ultimate target. This has led to a lot of debate, in which, while most opted for gradual evolution, others argued that mimicry evolution was sudden. For example, Richard Goldschmidt offered his Hopeful Monster hypothesis as a potential solution.
In the research reported at the link above, scientists were able to demonstrate that mimicry evolved gradually although they were not able to explain how the intermediates remained viable to the point that their genes remained in the population for additional evolution to occur. (They do offer a solution to this problem, but this will require additional testing.)
For a long time, it has been considered that moth and butterfly wing patterns are simply modifications of a ground plan. In the case of the butterflies studied here, the ground plan is called the Nymphalid ground plan (NGP), and it was recognised in the early part of the last century as underpinning butterfly wing patterning. The diagram in figure 1 at the above link shows it, and essentially it divides the wing into discrete regions which, at its most basic level, gives the fore and hind wings a fairly symmetrical and relatively straightforward appearance. Developmentally, the ground plan keeps the regions between the veins independent of each other and so patterns between different veins can move with respect to each other bringing about wing patterns that are relatively straightforward to patterns which are a kaleidoscope of pattern and colour.



How these kaleidoscopes form can be imagined from the diagram in figure 1, by thinking in terms of the lines or dots between veins moving around with respect to each other. Thus curved lines of pigment can be straightened, or lines of pigment can become confused as between vein pigments move forward and backward with respect to each other.


The veins in leaf mimic butterflies are simply between vein pigments of the NGP that have moved to form straight lines giving the appearance of leaf veins. They are not the butterfly wing veins. So essentially then, the problem of leaf mimicry looks to be a problem of modifications to the NGP.


The research basically looked at a species of leaf mimic and a host of non leaf mimics that are phylogenetically closely related. See figure 3 at the above link for the species examined as well as their phylogeny. It then determined that the NGP underlay these species and selected specific aspects of the NGP which gave rise to the mimic leaf veins. Again, see figure 3a.

Then using Bayesian statistics and inference, they were able to begin with the basic ancestral state and determine the accumulating set of character changes to bring about the final leaf mimic pattern. Figure 3b shows this, with node A being the basic ancestral state and node D being the last common ancestor to the leaf mimic states.
What they were able to show was:-


1) Gradual change as opposed to sudden change.
2) Changes that were dependent on each other as well as the kind of dependency (mutual or temporal). Changes that were independent of each other.


3) The temporal ordering of changes.
What they were not able to show was the selective advantage, if any, for each step along the path from non-leaf to leaf mimicry. That awaits future experimentation.
They conclude:-


Other related papers worth looking at are:-
Takao K Suzuki, Modularity of a leaf moth-wing pattern and a versatile characteristic of the wing-pattern ground plan, BMC Evol Biol. 2013; 13: 158.
Martin A, Reed RD, Wingless and aristaless2 define a developmental ground plan for moth and butterfly wing pattern evolution., Mol Biol Evol. 2010 Dec;27(12):2864-78.


If you Google both titles, then you will be able to find a link that takes you to the full article, online.




Gradual and contingent evolutionary emergence of leaf mimicry in butterfly wing patterns


The research relied a lot on Bayesian inference and statistics which is a branch of statistics that has, over the last few decades become increasingly used by mathematicians and scientists. This is because the statistics that is normally taught in schools relies on current data only. And so, the likelihood of one event occurring out of several is determined by merely counting the occurrences of all events, and forming a ratio. To form this ratio, the frequency of occurrence of one event is divided by the frequency of all events, and thus the likelihood is obtained.


With Bayesian statistics however, a bit of smart is incorporated into this procedure. The likelihood of an event may well depend on what happened previously and so with this particular kind of statistic, historical data is taken into account in order to asses the likelihood of a current event happening. In principle, its a lot more realistic way of doing things.


Given its reliance on prior information its a good tool to be used in evolutionary studies when assessing likelihoods of various outcomes.
A good description of the process can be found at the following link:-
Bayesian statistics for dummies
So what was this research all about?


Leaf mimicry is one of those hard evolutionary problems. In large part its hard to understand how, in a classical Darwinian scenario, it could have gradually evolved given that the intermediates would have nowhere near looked like the ultimate target. This has led to a lot of debate, in which, while most opted for gradual evolution, others argued that mimicry evolution was sudden. For example, Richard Goldschmidt offered his Hopeful Monster hypothesis as a potential solution.
In the research reported at the link above, scientists were able to demonstrate that mimicry evolved gradually although they were not able to explain how the intermediates remained viable to the point that their genes remained in the population for additional evolution to occur. (They do offer a solution to this problem, but this will require additional testing.)
For a long time, it has been considered that moth and butterfly wing patterns are simply modifications of a ground plan. In the case of the butterflies studied here, the ground plan is called the Nymphalid ground plan (NGP), and it was recognised in the early part of the last century as underpinning butterfly wing patterning. The diagram in figure 1 at the above link shows it, and essentially it divides the wing into discrete regions which, at its most basic level, gives the fore and hind wings a fairly symmetrical and relatively straightforward appearance. Developmentally, the ground plan keeps the regions between the veins independent of each other and so patterns between different veins can move with respect to each other bringing about wing patterns that are relatively straightforward to patterns which are a kaleidoscope of pattern and colour.



How these kaleidoscopes form can be imagined from the diagram in figure 1, by thinking in terms of the lines or dots between veins moving around with respect to each other. Thus curved lines of pigment can be straightened, or lines of pigment can become confused as between vein pigments move forward and backward with respect to each other.


The veins in leaf mimic butterflies are simply between vein pigments of the NGP that have moved to form straight lines giving the appearance of leaf veins. They are not the butterfly wing veins. So essentially then, the problem of leaf mimicry looks to be a problem of modifications to the NGP.


The research basically looked at a species of leaf mimic and a host of non leaf mimics that are phylogenetically closely related. See figure 3 at the above link for the species examined as well as their phylogeny. It then determined that the NGP underlay these species and selected specific aspects of the NGP which gave rise to the mimic leaf veins. Again, see figure 3a.

Then using Bayesian statistics and inference, they were able to begin with the basic ancestral state and determine the accumulating set of character changes to bring about the final leaf mimic pattern. Figure 3b shows this, with node A being the basic ancestral state and node D being the last common ancestor to the leaf mimic states.
What they were able to show was:-


1) Gradual change as opposed to sudden change.
2) Changes that were dependent on each other as well as the kind of dependency (mutual or temporal). Changes that were independent of each other.


3) The temporal ordering of changes.
What they were not able to show was the selective advantage, if any, for each step along the path from non-leaf to leaf mimicry. That awaits future experimentation.
They conclude:-


link above said:
Other related papers worth looking at are:-
Takao K Suzuki, Modularity of a leaf moth-wing pattern and a versatile characteristic of the wing-pattern ground plan, BMC Evol Biol. 2013; 13: 158.
Martin A, Reed RD, Wingless and aristaless2 define a developmental ground plan for moth and butterfly wing pattern evolution., Mol Biol Evol. 2010 Dec;27(12):2864-78.


If you Google both titles, then you will be able to find a link that takes you to the full article, online.


