host-associated differentiation

Now that's a mouthful!

I have moved away for looking at local adaptation and my new focus is host-associated differentiation or HAD. What does HAD mean? Although the focus of my reseach has previously been examing at adaptation from a tri-trophic perspective (the plant, the herbivore, the parasitoid), two recent papers have taken a tri-trophic perspective with a famous (well, at least in my world) system (Stireman et al. 2005, 2006). The famous system of which I speak is the golden rod system, Solidaginis altissima and S. gigantean.
Golden rod pictured here:

(from Thomas Schöpke - Institut für Pharmazie)

The stem galling tephritid, Eurosta solidaginis,

(from Tim Craig's Lab)

feeds on golden rod and produce
galls that look like this:

(from Vassar Biology Department)

This is a classic and well-studied system because it presents a case of both phenotypic and genetic host-associated differentiation. There are host-races of tephritids, Eurosta solidaginis, associated with each species of goldenrod (Solidaginis altissima, S. gigantean). So flies on S. altissima have a different genetic profile than flies on S. gigantean. In this study phenotypic differences in adult preference (female fly laying eggs on the plant) and larval performance (maggot eating the plant) were reinforced by genetic differences using both mitochondrial and allozymes (Waring et al. 1990, Brown et al. 1996).

So who cares?

Let's start with what a host race is...
Diehl and Bush provide the most succinct definition of host races;
a population of a species that is partially reproductively isolated from other conspecific populations as a direct consequence of adaptation to a specific host (1984).
If we see differences in different populations on the same species, this may mean that the plant host is driving diversification of this species, i.e. if we give these tephritid flies enough time (in evolutionary terms) then they may become 2 different species. This is a good indicator that plant hosts may be the force driving new species and it may explain why insects that feed on plants are the most diverse organism (Strong et al. 1994), Mitter et al. 1988, Jaenike 1990). This concept is the foundation of my research.

Are plant hosts driving herbivores to diversify and as the herbivores diversify, do the parasitoids also diversify?

I am hoping that if I find host-associated differentiation in a caterpillar, i will also find host-associated differentiation in the parasioid. Host-associated differentiation can be phenotypic or genotypic. This means I could find differences in performance of the caterpillar (i.e. growth rate, pupal mass, # of eggs laid by female) and/or I could find differences in the genetic sequence. Local adaptation examines variation (or differentiation) only at the phenotypic level (more precisely, differentiation at a few loci and not genetically apparent). Looking at both phenotypic and genotypic differentiation will allow me to get a better idea of the extent to which my caterpillars (and then my parasitoids) have specialized on their hosts.

Later this week... my caterpillars and what I am doing during the summer of 2006. Yikes! - which is already 2/3 over.

to be mobile or not to be mobile

First and foremost, a big thanks to you all for reading this blog. I know that it is probably not the most riveting blog to read but I think it is a good experiment in how we go about shaping our ideas and hypotheses. I believe in a community of science. That means all of you. I feel that it is invaluable to reach across labs, across research topics and across universities to overcome a lot of boundaries that previous academic generations have built. I am not sure this will make me a good scientist but this community interaction will make the process more bearable for me. If I get to the end that great PhD tunnel, I hope this leads to future collaborations.

Next I would like to answer some of the questions raised about my last two entries. They are all really good questions because either I was unclear or I just have not been able to get to the next nuance of local adaptation. This is where Julie is particularly helpful because she works on host-race formation, which on my concept of an adaptation continuum, would be the next ‘big step’ (depending on where biotypes fit in).

A population or deme can be adapted but not locally adapted. When I mean adapted to extant conditions, I mean that they have a range of fitness in their present environment that is the same mean fitness as other population in the surrounding environment. For example, in the following schematic of circles, imagine the white circles are adapted populations, the red circle is a maladapted population and the blue circle is a locally adapted population (I know that you are thinking about gene flow, but hold off on that for a moment).



In this schematic, the white populations all have around the same mean fitness. Let us say for these populations we are using the fitness measure of # eggs laid by a female. The white populations all lay an average of 60 eggs. The maladapted red population is not doing very well and only lays an average of 25 eggs. The blue population has ‘reached’ a higher level of fitness due to some phenotypic adaptation that may only be genetic changes at a single locus and difficult to find with genome-wide genetic analysis. This adaptation means that this locally adapted population of females lays 100 eggs on average. This is what I meant by the graph below:



Is the only difference between locally adaptation and host-race formation gene flow. First let me quickly explain host-race formation in brief so we are all on the same page. Julie, feel free to jump in on this because you are much more an expert than I am.
Julie and I both agree that the best paper to explain host-race formation is one by Dres & Mallet 2002. If you think that local adaptation is controversal than host-race formation is an Africa-hot topic. This is a simplified list of the criteria for host races according to (Dres & Mallet 2002):

1. The species uses different hosts in the wild and some individuals seem to “prefer” one host.
ex. the moth caterpillar feeds on two trees, box elder and willow. Some caterpillars prefer to feed on willow, some prefer to feed on box elder OR some female adult moths prefer to lay their eggs on box elder, while others choose willow.

2. The populations are found where the 2 host trees exist together.
ex. the moth caterpillar is found in ranges where both box elder and willow are found side-by-side or near to each other.

3. The individuals on one host have significant genetic differentiation from the individuals on the other host.
ex. the caterpillars on box elder are significantly (defined as more than one locus) genetically different from the caterpillars on willow.

4A. There is a connection between the preference for the host and the preference for a mate who likes the same host.
ex. moths that like to lay their eggs on box elder also like to mate with males who grew up eating box elder.

4B. There is some gene flow between the individuals groups on the different hosts. Because there are still the same species, there needs to be some mating between moths on box elder with moths on willow but not enough so that the genetic differentiation is lost.
ex. box elder moths occasionally mate with willow moths.

5. This is the same criteria as local adaptation. Individuals of the different host races have higher fitness on their own (natal) plant. ex. box elder caterpillars have higher fitness on box elder relative to their fitness on willow.

Okay, so now that I have gotten host-races out of the way, let’s get back to whether the only difference between host-races and local adaptation is the amount of gene flow.

Certainly in the Dres & Mallet paper they say the only difference between host-race and host-associated groups is gene flow (are host-associated and locally adapted the same thing? Initially I would say yes…). I would argue that it is not just gene flow. It really depends on what and how strong the selection pressure is. If there is a very strong parasitism pressure on the caterpillars on box elder, e.g. if you are willow caterpillar and you end up on box elder you are going to get parasitized, than the system can tolerate relatively high gene flow. So, yes, gene flow is important. In order for a group or population on a host to become a host-race, they must be genetically differentiated and so gene flow intuitively would be low. But then this connection between gene flow and host race formation (or local adaptation) would beg the question, are insects with low mobility more likely to be locally adapted (or host races)?

Well, Astrid was right on top of that question! It is true that the classic case of local adaptation was with scales which are pretty non-mobile (Edmunds & Alstad 1978). In my last post I referred to a meta-analysis examining 17 independent studies on local adaptation (Van Zandt and Mopper 1998). In this study they were interested in several trends in who was and who wasn’t locally adapted. One question they did look into was whether insects that were more sedentary were more likely to be locally adapted (Van Zandt and Mopper 1998). Well, the results were surprising… of the 12 insects used in the meta-analysis (4 mobile, 8 sessile), they did not find that to be the case. An insect was equally likely to be locally adapted if it was sessile, as if it was mobile.
Now I did a little eyebrow raising here because I have read enough papers on gene flow (a favorite topic of mine) to know it is very difficult to measure gene flow. In addition, quantifying gene flow can get particularly murky when you are dealing with an insect with a sessile caterpillar and a mobile adult or a sessile female and a mobile male, e.g your favorite and mine – Orgyia spp.(see poor wingless female below).


photo by Jeremy B. Tatum

Van Zandt and Mopper’s study was done over 8 years ago now. It would be interesting to look at where we are now with evidence of locally adapted species. Maybe we would get a different picture with respect to mobility. But with the information up to this point, I cannot say that there is evidence that sessile species are more likely to be locally adapted, although intuitively we would think this would be the case.

On a side tangent – If local adaptation is a necessary step before host race formation, then maybe it would be worthwhile to see if the same life history characteristics that are found to be correlated with local adaptation would also be correlated to host-race formation. It would be interesting to discover that only a subset of characteristics in locally adapted species are found to be present in host races. Then maybe one could predict what subset of locally adapted species would have a higher probability of becoming host-races. This is just an idea that I have been toying around with and will discuss with my host-race pal, Julie.

Now I have only begun to scratch the surface of some of your questions. I still need to address questions about spatial scale. But I imagine that you are as tired as I am. So I will leave that for another day.

the controversy

Remember how I told you that local adaptation is a slippery slope?
There are several reasons why local adaptation can be a difficult concept with which to work. Here are at least 3 that I have been grappling with:

1. Local adaptation is still a very controversal subject.
Remember that Edmunds & Alstad paper on on the black pine leaf scale that I mentioned in my last post (1978)? I said that it is a classic example of local adaptation. Well, that is not entirely correct. I should have said it was considered a classic example of local adaptation. Alstad has since recanted the results and conclusions of the study due some problems with the experimental methods (Alstad 1998). Other sceintist criticized the experimental methods, in particular the locations of the natal and novel plots relative to each other (Unruh & Luck 1987). Although this particular study has been strongly criticized, other scientists have gone on to design experiments that took these errors into consideration. Some of them were successful in finding evidence for local adaptation, while others were not.

2. The empirical evidence for local adaptation is not overwhelming.
By 1998 there were at least 17 studies on local adaptation in natural populations (Van Zandt & Mopper 1998). Even with more empirical evidence, the status on local adaptation remains controversal because the evidence is split in support and refuting the existence of local adaptation. This low number of empirical evidence in support of local adaptation may be due to the issues that confront those of us interesting in finding it. It is difficult to do reciprocal transplants and other experiments testing for local adaptation on natural populations in the field. Many problems are due to underreplication (Boecklen & Mopper 1998). The statistical power to detect a difference between the fitness on a natal or novel host is extremely difficult with low replication. Considering the difficulties in finding statistical evidence for local adaptation, I find it encouraging that it has been found. I also encouraged by recent evidence for genetic differentiation in a suite of insect species on 2 different species of goldenrod (Stireman et al. 2005). I will talk about this study at great length in future posts (I can tell you are at the edge of your seat).


Goldenrod photo from usn. This one is for you scott.

3. Local adaptation and adaptive deme formation - Is it merely a question of spatial scale?
Local adaptation is also termed adaptive deme formation. The Adaptive Deme Formation Hypothesis (aka ADF) was coined by the aforementioned Edmunds and Alstad in their 1978 paper. A deme is usually defined as an interbreeding population of the same species. ADF and local adaptation are used interchangabley by some. It is confusing whether they really mean the same thing. Afterall, when Edmunds and Alstad used it, they were referring to localized demes adapting to an individual tree. Local adaptation, on the other hand, can refer to any spatial scale. An organism can be locally adapted to a tree, to a host plant species or to a site. Are these terms interchangable? Or can you use ADF as long as you define a deme in your particular case?

That is all for today. Please leave comments and questions.

local adaptation

Note: I am inexperienced at blogging my research concepts and ideas, so please bear with me. I really appreciate any comments or questions. Please don't be shy.

The main concept behind my proposal is local adaptation.
Local adaptation is when a population evolves and adaptation to the ecological conditions in its local environment. The adaptation to these ecological conditions is driven by natural selection. The classic example of local adaptation comes from work done on black pine leaf scale insects, Nuculaspis californica, on pine trees. In the pine stands where these scales occur some trees have high infestations of scales while others are left unscathed. This lead to the hypothesis that these scales (and their descendants) were adapting to individual trees. To test this hypothesis scales were transferred from their natal tree to a novel tree. The scales experienced high mortality on the novel host. After these reciprocal transplant experiments, it was concluded that N. californica scale populations had adapted to survive on individual trees within a pine stand. These immobile scales had adapted to to the defensive properties of their natal tree (Edmunds & Alstad 1978).

One of the reason there is a great deal of interest in local adaptation is that it has been considered a stepping stone to speciation (Dres & Mallet 2002). I have made sense of the slippery slope of local adaptation by visualizing an adaptation continuum:




It is important to point out is that not all organisms will follow this continuum. Some species may never become locally adapted or speciate. This leads to another tangent that I will discuss in future posts – why are some groups of organisms so specious while others (say, for example, insects), while others have only one of a few species (sponges, for example). In this continuum an population can be adapted to extant conditions but not locally adapted. The population is not locally adapted because the mean fitness of the population is the same as any other population in that environment. The population is locally adapted when its mean fitness is higher than the mean fitness of other populations:



Here you can see that the fitness of populations fluctuate over time but a locally adapted population experiences higher mean fitness relative to the other populations.
How's that for a start?

the beginning...

this is the beginning of a beautiful relationship.