3. Biogeography and
phylogeny: The Canary Islands
version (my apologies for some bad grammar)
by E. García-Barros, Dep. Biology (Zool.), UAM
Note for UAM students (Biogeography course): the activities downlines were designed to be developped without direct assistance. In our course, however, the precise sequence, goals and other details may vary between terms. Please pay attention to the guide-lines given by the instructors.
A statistical package which enables you to perform multiple regression analyses (SPSS is currently available in the UAM computer rooms)
A program for phylogenetic inference such as the combination: NONA + WINCLADA (please visit cladistics.com for downloads).
Alternatively, you might want to try PAST
2. Numbers of species on islands
3. Grouping islands
4. What does phylogeny tell about?
5. Reconstructing phylogeny
6. Combining evolutionary histories
7. Determining areas of endemism
8. Towards panbiogeography
9. Bibliographic references
10. Notes (teachers)
0. Preface / (or: better skip)
materials intend to help in developing practical activities related to historical and cladistic biogeography, and island biogeography. It was designed
as a complement to the course on Biogeography (Biogeografia)
held in the Universidad Autonoma de Madrid (UAM), but could be adapted for
other purposes. The main target is to provide a practical (and not too deep)
introduction to biogeographic approaches that make use of phylogenetic
information, and to illustrate connections between these and other parts of the
program in this course. No exhaustive treatment of any of the topics is attempted here,
nor are the underlying theories or philosophy presented.
BASIC REQUIREMENTS (MATERIALS/SKILLS). The exercise is intended for advanced studies on Biology. Some knowledge of frequent biogeographic topics is expected: island biogeography, speciation modes, basic genetics. A degree of skill at using statistical software packages is also required, enough at least to apply bivariate and multivariate regression, and obtaining a dendrogram (UPGMA or similar) based on a matrix of distances or of association coefficients. Finally, a background on cladistics is also needed (just to be able to understand the concepts of character and character state, multistate characters, parsimony, consensus tree, as well as to represent dendrograms by means of parenthetical notation or Venn diagrams). In the UAM, most of this activity is developed in the students computer rooms (students should updated their accounts and access key-word in due time). Some parts may require different efforts depending on software availability and characteristics, which may be very different in other institutions. This version was designed for use under a Windows environment; statistics (and related advise) are mostly done with SPSS for Windows. For cladistics, the combination of Winclada + Nona (or Hennig) is possible, but alternatives exist using the package Phylip (J. Felsenstein) which is free. The format for the files to be analyzed under cladistics is rather primitive (text format) so that they can be edited with a word processor. All this can be done in more sophisticated ways, depending on the user/s resource availability and degree of skill.
ANTECEDENTS. This page was inspired in an exercise proposed by R.P. Filson (Berkeley Univ.), which was drawn to my attention by Dr. Moreno. That was based on the lizard genus Gallotia. I tried to make it a little bit more sophisticated, in agreement with the level of the students to which is addressed. Some parts can easily be omitted, or be treated in a very independent way from others. All the examples are real and verifiable within reasonable limits, so that the students can feel that they are working with data of the same 'quality' than those used by any professional biogeographer. Some data, and some of the phylogenetic hypotheses are soon to be outdated. This should not seriously affect the general sense of the results, and allow for updating by any one interested in these topics. The Canaries have been shown to represent an excellent natural 'laboratory', and hence new data and references on the fauna and flora of this archipelago accumulate at more or less regular rates
I hope that any practitioner of biogeography at a professional level can accept my apologies for any mistakes or oversimplifications. I shall sincerely acknowledge any suggestion or correction from any person in whatever biogeographic region.
The Canaries represent an archipelago of volcanic origin, quite close to the African Continent (110 Km.). Latitudinal differences among the several islands are meaningless in a broad context. It is likely that the origin of these islands is related with a 'hot spot' of magma that shifted from East to West in as a consequence of the activity of the Southern Atlas rift. As a result, the islands have different geological ages (2 to 20 MY), those closer to the continent being the eldest (Fuerteventura, 20 MY), and those farther West being much younger (El Hierro, 1 MY). Some pertinent data are given in Table 1.
Tabla 1. ILargest Islands in the Canarian archipelago (Spain), and some of their features (the exact figures may vary depending on source). HI= El Hierro, PA= La Palma, GO= La Gomera, TE= Tenerife, GC= Gran Canaria, FU= Fuerteventura, LA= Lanzarote. The number of species is a partial estimate based in just a few selected taxa; by no means these figures represent the total known or estimated floral and faunal diversities (see Fernández-Palacios & Whitakker, 2008; Fernández-Palacios & Martín Esquivel, 2001)
|Island area (Km2)||296||708||370||2.035||1560||1655||807|
|Maximum elevation (m.)||1501||2423||1487||3710||1950||807||671|
|Island age (MY.)||1.1||1.7||12||11.5||14.5||21||15.5|
|Distance to Africa (km.)||383||416||333||284||196||94||125|
|"Number of species" (selection)||1439||3858||2042||7789||4823||2509||2023|
Volcanic islands are useful as 'natural laboratories'. The can be assumed to be completely devoid of life in their beginnings. They are then colonized by organisms that are passively carried along by the wind or sea currents (or that can fly actively, like birds or some insects). Speciation events can then occur, giving rise to endemic taxa (species, genera, families), depending on the lapse of time since colonization, and of the features and population dynamics of the invading organisms. The patterns of colonization, the sequential arrival to neighbor islands, and the patterns of isolation may widely differ between taxa, giving rise to complex patterns that might prove difficult to understand. Our guess is that a good knowledge of the present distribution patterns of endemic taxa, combined with some knowledge on their evolutionary history, may help in disentangling the true history of occupation and radiation in the islands.
The actors of this story are a selected set of varied plant and animal species. We cannot deal in detail with all of them here, but all of them are interesting and often look quite different than their nearest continental (African or Palaearctic) relatives. Some are relatively popular, such as the 'giant' lizards of the genus Gallotia, restricted to the Canarias (actually, not all species are so large). Just by the end of the XXth century, a new species of this genus was discovered in the island of La Gomera, with just a few surviving individuals ... (more info?). Other instances include the 'fatty' wingless tenebrionid beetles of the genus Pimelia. These are typical inhabitants of arid environments. Their elytra (forewings) are fused so they cannot fly. Same as the lizards, their ability to disperse between islands is very limited. We shall mention a few other plant and animal taxa. It is frequent that island endemic species are peculiar in some way if compared to their nearest continental relatives: unexpectedly large or small size, vegetative organs highly lignified instead of an herb appearance (e.g. Echium spp), and so on.
2. Numbers of species on islands
The number of species found in any area unit may depend on a large series of environmental and geographic factors. In other parts of this exercise you might find that the number of species recorded from a given island does not fit the amount expected, e.g., in comparison to another island of very similar size. Could we explain these differences by reference to environmental variables, or to any ecological or geographical causal factor? A first hand approach to this problem might consist in determining to what extent one island (or set of islands) actually hosts a number of species that is significantly larger (or smaller) than expected, and what are the variables that might be involved in such effect. This implies, first, determining whether or not there is a general 'between-island' trend, and second, identifying the outliers. The results (whatever they happen to be) may be useful for discussion at a later stage.
1.1 - Find in Table 1 some factors that could have to do with species richness, as well as an estimate of the number of species per island (based on selected taxa). Let's select am a priori relevant feature, such as the island surface (area). Seek for the relation between island area and number of species:
2.1.1- Build a file (e.g. an Excel
spreadsheet, o directly on the statistical package available such as SPSS,
STATISTICA, or similar). Fill one column with the number of species, other with
the areas, and a third one with the island names (or labels such as: HI, PA, GO,
TE...). The islands will be the cases (rows).
2.1.2- Transform both types of numerical data, e.g. into decimal logarithms (or to square roots). (You or your instructor might eventually wish to further discuss about the convenience of doing this).
2.1.3- Calculate the correlation between area and number of species.
2.1.4- Make a dispersion plot using the transformed values (X= areas, Y= nr species). Check for the significance of the correlation (r). If r is statistically significant, that means that there is an overall relationship linking species richness to the area of the islands. Is this the case? And, if so: Should we predict that the smallest islands in this archipelago should host a lower number of species than the largest ones? Try to discuss this in the light of what you know on island biogeography.
2. 2 -
In a further step, we might wish to describe the relationship we have just
discovered. This would enable us to predict the diversity of an hypothetical
island of any size. To do this, and using the same values as above, let's fit a
regression line to the data (again, X= area, Y= nr species).
2.2.1- With the regression equation (of the kind y= a + bx), predict: How many species should we expect to occur in El Hierro in the long term if, due to global warming, one half of the isle happened to disappear under the sea? How many species could potentially host the island of Tenerife if a sudden elevation of the island due to volcanism increased its surface by a 10%?
2.2.2- In the same ways, we can detect what islands (if any) host a species diversity LARGER or SMALLER than that expected after the general trend. This can be done visually, checking for large residual values (data points that lay away from the regression line). A simple way to do it is to plot in the same graph, e.g., the 95% confidence limits of the regression line. Having done this: Can you detect any such island (with comparatively 'too many' or 'too few' species)? If this was the case: What might be the reason for an island to host more or less species than expected after the island area?
2. 3 - A single
variable shall seldom explain all details concerning the species richness in the
elements of an archipelago. In the present example, other variables may allow
for further insight... let's try!
2.3.1- Predict the effect of each of the variables from Table 1 separately (one at a time by bivariate regression or correlation) on the number of species. What is the direction of each relation (correlation)? Any significant r? Discuss briefly, trying to give arguments related to what you would expect from each one of the variables:
2.3.1a- Maximum altitude?
2.3.1b- Distance to the African Continent? Geologic age? Is there any relationship between the last two?
2.3.1c- Density of human population?
2.3.2- This can be repeated for any number of variables. However, most statistical packages allow for more sophisticated approaches. So, try a multiple regression: the number of species is still the dependent variable, but test the simultaneous effect of two or more of the predictor variables (elevation, area, and so on). This should enable you to speculate about the relative contribution (or explanatory power) of each variable. Is the coefficient of determination (R2) higher than in our trial with just island area? Does any of the variables seem to have a specially remarkable weight in relation to species richness? This process can be complicated when two or more of the environmental/geographic variables are highly correlated to each other, which might be a good basis for thorough discussion... we do not attempt to dive too deep into regression techniques at this point, however.
2.3.3- The equation looks a little bit more complex now. But, if the proportion of variation explained by the model is high, we can attempt more precise (and reliable) predictions. On this light, reconsider the problem of Lanzarote (as formerly proposed: area reduced by 50% due to natural -though human-induced- disaster). If we had to manage Lanzarote's biodiversity in the very long term: Can you propose any reasonable action (that is possible for we, Humans) to compensate the predicted loss in biodiversity, creating conditions that could ensure the persistence of the present number of species? (Highly speculative, but worth trying).
Depending on your supervisor's instructions, use your results to:
3. Grouping islands
One frequent approach on the first steps of to the comparative study of diversity is classification of area units. This helps to identify patterns of similarity between the units classified (islands, quadrates, or other), based in the similarity of their flora and fauna, irrespective of the reasons of that similarity (this is to be evaluated by other means). One most frequently used set of techniques is based on hierarchic procedures. A hierarchical classification can easily be expressed by means of one dendrogram (but also by other means). Here I propose two of the many methods, quite diverse from all respects: one dendrogram based derived from a matrix of association values (or similarities), and another one constructed using the principle of parsimony. We are using just presence/absence data (1/0) in both instances. Even when 'parsimony' may sound 'very phylogenetic', we are not making use of phylogenetic information at this point (however, this way of using parsimony programs may serve as an introduction to some of the current usages and procedures related to PAE, as detailed later on).
First, we have to write a file which is the presence/absence data matrix. We shall the use one statistical package (SPSS or other) to build one dendrogram, and a maximum parsimony program to make a parsimony-based dendrogram. Depending on chances and availability, a purely manual approach is possible, and might prove interesting to learn the basics of agglomerative techniques.
Use the data on endemic plant and animal species (reptiles, insects, plants). The taxa dealt with are generally featured by a poor dispersal ability: reptiles of the genera Gallotia, Chalcides (in the Canaries: lisas) and Tarentola (periquenes), the beetles (O. Coleoptera) of the family Tenebrionidae Pimelia and Hegeter, and Calathus (family Carabidae), and plants in the genera Crambe (Family Brassicaceae) and Echium (Boraginaceae). For data sources: see references. The presence/absence data (species by Island) are given in Table 2, below.
Table 2. Distribution of selected endemic plant and animal species across the main Canary Islands. Most of these species (and some whole genera like Gallotia) are restricted to this archipelago. HI= El Hierro, PA= La Palma, GO= La Gomera, TE= Tenerife, GC= Gran Canaria, FU= Fuerteventura, LA= Lanzarote. The numbers in the left column are species labels with no special meaning. Crosses (X) indicate the presence of one species (row) in one island (column). A degree of simplification has been adopted in a few instances, in relation to sub-specific taxa and minor islands. [Get this table in Excel format]
species / ISLAND
|1||Gallotia galloti||X||X||F. Lacertidae|
|2||G. caesaris||X||X||F. Lacertidae|
|3||G. intermedia||X||F. Lacertidae|
|4||G. simonyi||X||F. Lacertidae|
|5||G. bravoana||X||F. Lacertidae|
|6||G. atlantica||X||X||X||F. Lacertidae|
|7||G. stehlini||X||X||F. Lacertidae|
|8||Chalcides viridianus||X||X||X||X||F. Scincidae|
|9||C. sexlineatus||X||F. Scincidae|
|10||C. simonyi||X||X||F. Scincidae|
|11||Tarentola boettgeri||X||X||F. Geckonidae|
|12||T. delalandi||X||X||F. Geckonidae|
|13||T. gomerensis||X||F. Geckonidae|
|14||T. angustimentalis||X||X||F. Geckonidae|
species / ISLAND
|15||Pimelia laevigata||X||X||X||F. Tenebrionidae|
|16||P. estevezi||X||F. Tenebrionidae|
|17||P. granulicollis||X||F. Tenebrionidae|
|18||P. fernandezlopezi||X||F. Tenebrionidae|
|19||P. sparsa||X||F. Tenebrionidae|
|20||P. canariensis||X||F. Tenebrionidae|
|21||P. ascendens||X||F. Tenebrionidae|
|22||P. radula||X||F. Tenebrionidae|
|23||P. lutaria||X||X||F. Tenebrionidae|
species / ISLAND
|24||Hegeter amaroides||X||X||X||F. Tenebrionidae|
|25||H. lateralis||X||F. Tenebrionidae|
|26||H. tenuipunctatus||X||F. Tenebrionidae|
|27||H. brevicollis||X||F. Tenebrionidae|
|28||H. transversus||X||F. Tenebrionidae|
|29||H. intercedens||X||F. Tenebrionidae|
|30||H. gomerensis||X||F. Tenebrionidae|
|31||H. glaber||X||F. Tenebrionidae|
|32||H. webbianus||X||F. Tenebrionidae|
|33||H. costipennis||X||F. Tenebrionidae|
|34||H. impressus||X||F. Tenebrionidae|
|35||H. grancanariensis||X||F. Tenebrionidae|
|36||H. subrotundatus||X||F. Tenebrionidae|
|37||H. politus||X||X||F. Tenebrionidae|
|38||H. plicifrons||X||F. Tenebrionidae|
|39||H. fernandezi||X||F. Tenebrionidae|
species / ISLAND
|40||Calathus angularis||X||F. Carabidae|
|41||C. appendiculatus||X||F. Carabidae|
|42||C. canariensis||X||F. Carabidae|
|43||C. rectus||X||F. Carabidae|
|44||C. abaxoides||X||F. Carabidae|
|45||C. gomerensis||X||F. Carabidae|
|46||C. spretus||X||F. Carabidae|
|47||C. marcellae||X||F. Carabidae|
|48||C. obliteratus||X||F. Carabidae|
|49||C. refleximargo||X||F. Carabidae|
|50||C. pilosipennis||X||F. Carabidae|
|51||C. cognatus||X||F. Carabidae|
|52||C. laureticola||X||F. Carabidae|
species / ISLAND
|53||Crambe arborea||X||F. Brassicaceae|
|54||C. pritzelii||X||F. Brassicaceae|
|55||C. tamadabensis||X||F. Brassicaceae|
|56||C. santosii||X||F. Brassicaceae|
|57||C. strigosa||X||X||X||F. Brassicaceae|
|58||C. wildpretii||X||F. Brassicaceae|
|59||C. scoparia||X||F. Brassicaceae|
|60||C. gomerae||X||F. Brassicaceae|
|61||C. feuillei||X||F. Brassicaceae|
|62||C. laevigata||X||F. Brassicaceae|
|63||C. scaberrima||X||F. Brassicaceae|
|64||C. microcarpa||X||F. Brassicaceae|
|65||C. sventenii||X||F. Brassicaceae|
species / ISLAND
|66||Echium auberianum||X||F. Boraginaceae|
|67||E. wildpretii||X||X||F. Boraginaceae|
|68||E. pininana||X||F. Boraginaceae|
|69||E. descaisnai||X||X||X||F. Boraginaceae|
|70||E. handiense||X||F. Boraginaceae|
|71||E. onosmifolium||X||F. Boraginaceae|
|72||E. callithyrsum||X||F. Boraginaceae|
|73||E. strictum||X||X||X||X||X||F. Boraginaceae|
|74||E. bonnetii||X||X||X||F. Boraginaceae|
|75||E. brevirame||X||F. Boraginaceae|
|76||E. hierrense||X||F. Boraginaceae|
|77||E. giganteum||X||F. Boraginaceae|
|78||E. simplex||X||F. Boraginaceae|
|79||E. virescens||X||F. Boraginaceae|
|80||E. leucophaeum||X||F. Boraginaceae|
|81||E. aculeatum||X||X||X||X||F. Boraginaceae|
|82||E. webbii||X||F. Boraginaceae|
3.1. Elaborate a presence/absence matrix. Species on rows (cases), islands in columns (variables). The presence of one species in one islands is to be denoted by a "1", the absence by "0". With a spreadsheet and a hand computer: open the file (Table 2) and save it in a disk or folder for edition and further use (open: here). Next, replace any "X" by a "1" (menu bar: Edition); then, fill the empty spaces (blanks) with "0". The data matrix obtained in this way should look like that of the example attached here. Depending on the software to be used in later steps, this matrix will require some manipulation (keep one copy of your original data matrix).
hand-made dendrogram. Calculate a matrix
of similarities between the variables. The necessity for this may
depend on your previous experience: you might already know what an association measure
(or index) is, which one you want to use, and how it is calculated, and
perhaps some knowledge of how agglomerative clustering methods work (based on
median, nearest neighbor, or other). If so, you ay wish to skip this part (in
this case: go to 3.3.). Otherwise proceed, to see how to do
this in a fully manual way. This may be informative anyway, for you will seldom
do this by hand whenever your data matrix is of a moderate size.
3.2.1. Build a similarity matrix. A very simple association measure is the so-called "Distance of Czekonowski", which is Manhattan distance averaged for the number of cases. Manhattan distance is the sum of the absolute values of the difference between all cases (species) for each pair of variables (islands); Czekonowski's distance results from dividing this figure by the number of specie.
To avoid formulas here: In the example illustrated few lines above, the distance between El Hierro (HI) and La Palma (PA) should be estimated as |1-0| + |1-0|+|0-0| + |1-0|+|0-0| divided by 5, that is Dc (HI-PA)= 3/5, which is=0.60. The distance between La Palma and La Gomera would be estimated as (1+1+0+0+0)/5= 0.40, and so on. The distance matrix that would result from the above example is shown in the figure attached. Using a measure of distance, large values indicate high dissimilarity. In the example, El Hierro and La Gomera are the less dissimilar pair. This kind of approach is often interesting as a first step.
3.2.1.a. Have a glance at the figure on the right side. We shall now follow an agglomerative method. In successive steps, we shall link the pairs of islands linked by the shortest distances. Following the example, let's take the pair HI+GO (where the distance, D=0.20). We join these two to form the first cluster; the linkage distance (to fit on a horizontal scale) shall be the distance between the two species, which is 0.20.
3.2.1.b. Recalculate the distance between the cluster we have just constructed (HI+GO) and all the other islands (not yet linked). We shall use the median as a linkage criterion, which means the new distances between (HI+GO) and TE should be the median (which in this case is equal to the mean) of the distances HI-TE (=1.00) and GO-TE (=0.80), that is (1.00+0.80)/2= 0.90. Write a new matrix (islands that have been clustered replace the old columns of the respective components).
3.2.1.c. Repeat the procedure: Select the smallest distance value in the new matrix, join the variables (islands) at that distance... In the end of this process, we should have obtained a dendrogram based on Czekonowski distances, an agglomerative procedure, and the median as the linkage criterion (this is what UPGMA means). The bifurcations correspond to the distances we calculated.
a dendrogram based on an index of association - machine-made.
Several statistical software packages can do this. It can be done in several
ways, either starting from the raw data, or from a matrix of distances (or
similarities) that was previously calculated by other means (this thing here is:
does the program you have calculate the index you need, or do you have to
feed the program with the distances). First, determine what measure of
association is required (there are a lot, and there might be reasons to prefer
one over another). Examples of these indexes or coefficients are: Simpson's,
Jaccard's, Baroni-Urbani-Busher's, Sorensen's). Second: what basic procedure is
to be applied, agglomerative or divisive? And third: What kind of linkage method
is going to be used (average, median, nearest neighbor, centroid)? Most
statistical programs will prompt you for this in one or another way, hence it is
important to know that one always has to take these three decisions (explicitly,
The package SPSS is available in the computer rooms of UAM, so this is what I recommend to our students. Using this you can build a dendrogram based on Jaccard's distances following an agglomerative procedure and based on median-linkage criterion (this is what technical texts refer to by UPGMA).
3.3.1. Import the data matrix Excel into SPSS (first, please activate the program)
3.3.2. Alternatively, hand-key the file in the program spreadsheet (this is easy for a short file).
3.3.3. Save the SPSS (or whatever) format file in a disk or temporary folder.
3.3.4. If everything is OK: click the menu bar (above): ANALYZE, and there: CLASSIFY, and then HIERARCHIC AGGLOMERATES (or similar terms in other languages).
3.3.5. A dialog appears; select "Graphs", then "dendrogram". In "Method", select Measure= binary, Method= Jaccard, and = between-groups linkage procedure. A dendrogram is obtained (either in vertical or horizontal layout). You can export this in a convenient image format, so it could be used to insert in a report, and for further discussion.
3.4. A phenetic dendrogram based in the principle of maximum parsimony. Your 0/1 matrix is in all respects similar to the kind of data frequently used by cladists for phylogenetic inference. If we add an outgroup (consisting of one hypothetical continent where all species are missing=0), we can process the data by assuming that '0' is the plesiomorphic state. Even if making use of parsimony, what we do has little to do with phylogenetic methods: note that there is no information on the species phylogeny, and the single assumption is that all species are restricted to the islands and never found elsewhere. The results may be broadly comparable to those based in association measures; here, the measure of association is maximum parsimony.
NOTE: A detailed explanation of the topics related to parsimony and phylogenetic reconstruction is out of place here. The method will work whenever the data matrix is in adequate format. In general use is required of any of the programs that are specifically designed for parsimony analysis (Mix from the Phylip package, Hennig, Paup, or any other. The following instructions refer to Nona + Winclada).
3.4.1. Prepare/edit a data matrix
for use with Nona-Winclada:
One simple way is to edit a pre-existing matrix, adding a few specifications required by the program. Note that now the units to be classified (islands) are in rows, and the species (these are going to be used as characters) in columns. You can use the 'transpose' facility of Excel (paste-special). You can edit the file with Word or another word-processor. Whatever the case, it is very important that you save it as MS DOS TEXT (never as a *.doc file). Do Never use tabulations. The file must look exactly like this (yellow block in the left, the line to the right side are comments):
...this command must go in the first line
'a:canarias' ...Optional, in hyphens?, name of the file
10 5 ...Important: number of characters / blank space / nr of islands (in your data)
Gext 0000000000 ...Island name or code / blank / states of the characters 1 to n
Hierro 0000000000 (better use short island names, of up to 10 characters)
Palma 1 0 0 1 0 0 1 0 0 0 ...Character states could go in a block, or separated by blanks (space bars)
Gomera ...(Is not mandatory that the island name is in the same row as its data,
1001001010 you could use the format applied here to Gomera)
Tenerife 1001011001 ...Avoid long names, and do not insert blanks withim them. For instance,
; write 'La_Gomera' or 'LaGomera' instead of 'La Gomera'
proc/; ...Mandatory: file MUST end like this: "; proc/;"
(You can get this example as a text file by: clicking here)
Practical hints: This file con be processed with Winclada, or with the program Hennig. Besides the data themselves, the specifications in the third row (10 5 in the example) are very important for the programs to work. Do not forget to fit these figures to your data. "Gext" (or any other name) is an hypothetic outgroup, where all characters are 0. One way to make the file is copying from the original spreadsheet to Word, then saving as 'MS DOS TEXT' . ¡Do never use tabs, just the space bar (a long key in the lower part of the keyboard). If using Windows: Do never edit the file with the Windows Notebook, use either Wordpad o Word!
Using phylogenetic inference programs, you might happen to get multiple solutions for a single data set. If so, simply select the first tree. This will suffice at this stage of the present exercise.
Reaching this point means
that you have obtained at least one dendrogram. This dendrogram is an hypothesis
on the relations between the islands (based on their overall similarity from a
faunal and floral point of view). Keep the graph, it can be compared with other
results. Discussing it -and its possible implications- might prove interesting.
Following this idea:
Write a brief report on the results of section 3, including your dendrogram/s, and at least a comment on:
4. What does phylogeny tell about? (On the way to phylogeography)
We can find relations between islands, but we cannot but speculate on the nature of that relationships... What are the causes of the similarities we detected? Do they depend just on the geographic distances, elevations, and so on? Do the endemic species from these islands share a common evolutionary history which determines island relatedness, or is this relatedness just a question of chance or of colonization sequence? Can we test the hypothesis that a majority of the endemic species originated from continental species that reached the Canaries in an East-West direction? This is not so straightforward from our former approaches, due to the fact that we lack: 1) evidence for the sense (direction) of the relations, and 2) any evidence that the same pattern of diversification (e.g., our former dendrogram) has been followed not just by the islands, but by the endemic species that inhabit them.
Phylogeny might give some information on both 1) and 2) (above). In theory, if we knew the pattern of descent of an endemic taxon (at a species level at least), we would be able to induce 'who was first there', 'who may have derived from whom', and hence 'what was occupied by which species, and in what sequence'. That is: the history of occupation and diversification in the archipelago. In very short, the idea is that schematized in the next figure. Given a known distribution of the species A-E (an endemic monophyletic taxon), and knowing their pattern of descent, we can hypothesize on the history of the geographic distribution of the species A-E.
To do this, we need phylogenetic reconstructions, and hence some elements of cladistics would not be so bad... We have a chance to reconstruct the phylogeny of one taxon whenever we identify one homologous structure in a group of species, and we can say what has been the direction of evolutionary change in that structure: which is its ancestral state (primitive, plesiomorphic), and which the derived (apomorphic) state. If we can discover this for many features, better the best. If all evidences point in the same direction, then perfect! It is also useful that we understand the equivalences between several forms of representing nested hierarchies: dendrograms, parentheses (or brackets), Venn diagrams. This facilitates some of later work. [ EXAMPLE].
4.1. Design a phylogenetic hypothesis. Below these lines are represented (very inaccurately, I am afraid) the several species of Gallotia lizards, each in its island (labeling coincides with that in Table 2). The size of the drawings is roughly proportional to the species actual mean size (males, tail excluded). Inter-specific size variation is remarkable. Let's presume that the extant species nearest to Gallotia (and hence, likely to be related to the antecessor of this genus) is a member of the genus Psammodromus, with a size comparable to that of Gallotia atlantica from Fuerteventura. With this information, construct a phylogenetic hypothesis on Gallotia using an intuitive graphic approach, based on the size of the species as the sole character.
There are several possibilities:
4.1.1) Order the species into a lineal sequence according to the character you know (size). The line should start in the more primitive point (that is, a Psammodromus size). Afterwards, you could translate this line into one cladogram [See how?: Example] (Note: Technically, we should not assume that a lineal ordination of species represents the phylogeny, for we do not know whether one extant species really IS ancestral to other, or if instead both share a common ancestor).
4.1.2) Use Venn diagrams. Starting from the species displaying the more primitive character state (that is, 'Psammodromus size') identify the more evident shifts in that character (size). Include the species sharing the same character state into one set (this is a graphic way of identifying autoapomorphies). Proceed gradually like that, just avoid any intersection of the subsets (that is not allowed). Once this is done, translate the hierarchies into a dendrogram: Example.
4.1.3) Whatever the result was, formalize it as a cladogram (if you did not before). Remember, there is a close relation between the several modes of expressing nested hierarchies: Example. Write the species names (Table 2) in the branch tips. The result is your hypothesis on the phyllogeny of Gallotia (and this is a cladogram, just because it was built following the basic principles of cladistic analysis: homology, sinapomorphies, outgroup...).
4.2. The taxonomic area cladogram and the resolved area cladogram. Before a little more of digging into phylogenetics, let's appreciate the way in which phylogenetic information can be used in combination with distribution data from endemic species. The Taxonomic Area Cladogram is the more immediate way in which both things can be related. The result is an hypothesis on the history of distributions. What is a T.A.C.? In a cladogram of the taxon we are interested in, replace the name of each species by the name of the area unit (in this example, one island) where that species occurs. If the species inhabits two or more islands, write those names in the same tip. We are going to need several of these TACs later on. Let's warm up with the next exercise:
the cladogram obtained in section 4.1 (Gallotia). Let us assume that we
rely on this phylogenetic hypothesis, so let's make use of it.
4.2.2. Substitute each species name by the name of the island where that species occur. This is a taxonomic area cladogram, and represents the history of the distribution of Gallotia across the Canary Islands. If this dendrogram is developed completely (by dividing the branch tips until there is just one island name in each tip) we would obtain what is called a Resolved Area Cladogram (RAC). However, there are some complications (two species in an island, a species in more than one island, island with no species). In these instances, apply the following recipes:
4.2.3. Problem? / Example - recipe:
x What to do when there are two species in the same island?: redundant distributions
x ...one species in two (or more) islands? widely distributed taxa
x .. and island with no species?: missing areas
4.2.4. Now, represent the resolved area cladogram for Gallotia.
5. Reconstructing phylogeny
Our former attempt of phylogenetic reconstruction might
not be very sound... There is room for discussion: Is body size a reliable index
of phylogenetic relatedness? Arguments for, or against?
Whatever the answer, a phylogenetic hypothesis is more reliable the more it is based on a wider and varied evidence. There is some amount on available data (morphology, ecology, genetics) on some of the groups of organisms we are dealing with here; this could be used to construct a -probably more solid- phylogenetic hypothesis.
A potentially relevant source of characters for phylogenetic reconstruction are the nucleotide sequences from DNA and RNA. When a researcher sequences a fragment of a gene and publishes the results, the basic data (nucleotide sequences) are submitted to a database which is accessible on the Internet (Genbank) so that other scientists can make use of the information.
In this section we shall produce a new phylogenetic hypothesis for Gallotia, this time based on molecular data: 12S RNA and mtDNA (cit. B), using one species of the genus Psammodromus as the outgroup.
A nucleotide sequences is, for practical purposes, a code of four letters (A, C, G, T). The position of one given nucleotide in the chain can be regarded as homologous between species, and the nucleotide is the character state. For related species, most nucleotides in homologous positions will be the same, but hopefully some will vary -thus providing potentially phylogenetically sound information. Oversimplifying some details, we can regard nucleotide sequences as series of multistate (in fact, four-state) unordered characters. This can be processed with parsimony-based programs. We shall avoid several potential complications in this approach. Even so, several levels of difficulty are possible, so I propose two possibilities (basic and advanced: it is up to you, and the time you can or want to invest here).
For a quick trial, you can use an ad-hoc file I prepared. Open it, and save a copy in your folder or disk: GALLOTIA
This permits going directly to data analysis, skipping the process of previous preparation of the data. However, the results are not 100% guaranteed. Go to: 5.2.7.
SLOW (but more informative): Alternatively, if you can spend 30 minutes to see how GenBank looks like, go to the next section (5.1).
5.1. How do we get molecular data published by other biologists? Try Internet: GenBank. Write a name, e.g. Gallotia, in the search path and click on "GO". This may probably take you to several species, choose one. You'll arrive to a sheet with varied data, with a path named "Entrez Records" which includes, among other things, the number of direct links with nucleotide sequences. By clicking that number you should get a list of available sequences, each headed by a code: the accession number (looks like, e.g.: Z48035). These codes are worth noting, for these are the codes quoted in the original publications to label the sequences stored. Double-click one, just to see how the data are presented. The nucleotide sequence is at the bottom of the screen. You can copy and paste it in a word processor. You need to locate comparable sequences of all the species you want to analyze (that is, same gene, same part, same length of the sequence). This is not always feasible. There are tricks you can use, such as first getting a published reference on this kind of work, where you can know what species were sequenced, and what are the precise codes you need. Also, the process of alignment can be done directly in the net. However, all this takes a time, and we shall better simplify a bit (try whenever you have got some more time to spend with this). Now you know more about GenBank, go to section 5.2.
Get a file with nucleotide sequences from Gallotia,
and edit it for a cladistic approach. The sample data are slightly pre-edited,
which means some time-saving in comparison with the whole process. We shall use
a word processor, paste there your sequences. Step by step:
5.2.0. Download the file: [ here ] (save it in your folder or disk).
5.2.1. Delete the arbitrary blanks (their single function is facilitating nucleotide counts).
5.2.2. Try to align the sequences (when two sequences of related species are aligned, most positions show the same nucleotide).
Two sequences from two species may not start nor end at exactly the same points. One trick: if there is not very evident equivalence, copy a sequence of six or seven bases (that look unusual, say: tacataca) in the 'search' option of the 'edition' menu of the word processor, then seek for the same sequences in the other species. You could mark in a different colour when located (e.g., with the 'replace' option). After a few trials, you may have discovered fragments that are homologous in all the species you have. Taking these as a reference, and with some patience, you can then know what are exactly the parts that are really comparable in all the species. It would be perfect that the several homologous sequences start and end in the same point, same line... fragments at the beginning or the end which are not found in some species represent material that cannot be compared across species. The easiest option is to delete such non-comparable parts of the sequences...
Here there is and illustrated example. You have to practice your Spanish, but the translation is straightforward (think about, especially if you were considering a visit to the Canaries) (e.g., eliminar= eliminate; caracter= character; comparable=comparable, and so on... and Paso= Step).
Ejemplo, ver: Paso I Paso II Paso III Paso IV Paso V Paso VI
NOTE: A word processor and a manual method are not a very professional approach. This can be done much faster. However, it is a good thing to try with your own hands, once in your life. For more details, have a look at Baldauf (2003) [see reference].
If several nucleotide sequences are to be combined, repeat the process. Place
all the sequences of each species one after another, taking care to keep the same
order for each species. Remember, the outgroup (Psammodromus algirus) has
to be in the first place (this is for practical reasons).
5.2.4. Check that the number of nucleotides (bases) is the same for all species.
5.2.5. Now the file needs some edition. Prepare a file exactly in the same format as we used in a former step for a parsimony program (follow the same instructions as then: example). Do not forget writing the correct number of species (including the outgroup) and of characters.
5.2.6. Normal: ¿Normal? In these circumstances, I think that trying to process the data would be 'normal' (don't you?). The details may depend on the precise method to follow, ask your instructor. Yes?
Processing the data matrix with Winclada (+ Hennig/Nona).
Brief instructions. It is assumed that the data file fills all specifications, is in MS-DOS TEXT FORMAT (or ONLY-TEXT FORMAT), and that you installed the programs somewhere in your computer (preferably, all related files in the same folder). Right? Then:
Open the program:
Double click (i.e., click-click) on the Winclada icon.
Upper Menu Bar: Push on "FILE", "open file", specify the location of your data file (in "file types-formats", specify: "all files", then "Open").
Menu bar: "CHARS", then "Select all chars" (you select all the characters).
Menu bar: "CHARS" again, then: "Make sel chars NONADDITIVE", and "ACCEPT" (treats all characters as disordered).
Menu bar: "Analyze": "Heuristics" (let's ask for a cladogram). A dialog opens:
Write "100" in the three upper left paths:
Maximum.... = 100
Number of... = 100
Starting... = 100
Click on "Search"
The result should look (more or less) like this:
The number of trees found, ci (consistency
index) and ri (retention index) are given in the lower part of the screen.
Check that you obtained one single cladogram (left and down, it says "Tree 1 of
1"; there might be more than one, then it should read "Tree
1 of 7". "Tree 1 of 22", and so on).
If there is more than one tree, ask for a consensus tree: "trees" in the menu bar, "Consensus-compromise", and then "Consensus (strict)".
Once this has been done, check for the number of synapomorphies is supporting each branch (one of the icons in the upper left side of the screen, looking like a small red square). The tree becomes of this style:
...where the small black (filled) dots are synapomorphies (free of homoplasy), the empty circles are homoplastic characters (denoting convergence or reversal).
If everything was right, this part of the activity is over now. We have got an original hypothesis of Gallotia phylogeny, based on molecular characters.,..
(!) If you enjoyed it, and want a little bit more, click: here (you'll have to read some Spanish)
Otherwise, with your results: Prepare a taxonomic area cladogram (sect. 4.2), using this new phylogenetic inference. Does any new inference on the history of Gallotia en these islands arises?
Perhaps other organisms evolved (speciated) in parallel with Gallotia. This is likely, to the extent that they arrived to the islands in similar ways and times... If so, how could we treat information on the phylogeny of varied taxa, to get a single historical 'summary'? Perhaps combining... What?
6. Combining evolutionary histories
6.1. Brief reasoning
We seek for information on possible historic relations between the biotas of the Canary Islands. Taking in mind the results from sections 3 and 4, try to find a short answer to the questions:
x In Table 2: What species are most informative for any relatedness between islands? Why?
x Imagine that every endemic species inhabits just one island, no more. What kind of information could let us calculate distances (faunistic, floristic) between islands? How could we obtain such informations?
x Finally: If we can gather such information in any way, how could we process it?
6.2. Preparatory activity
We shall try to combine information from the cladograms calculated for several taxa. We assume that, in each case, the cladogram is a reliable hypothesis. Cladograms for these diverse taxa are presented in Table 3 (these are the same organisms quoted in Table 2, with the exception of Gallotia for which we already have a cladogram).
Now we need the following materials:
Drawings or printed versions of all the cladograms.
Using the cladograms and Table 2: make resolved area cladograms for each genus (combine data from Tables 2 and 3) (see section 4.2).
Table 3. Other taxa, other phylogenies... The next data are the topologies of maximum parsimony cladograms for the genera of animals and plants of Table 2 (see the references section for any further detail). Species names have been replaced by the number associated to that species in Table 2. The best way is to draw manually all the cladograms (Chalcides and Pimelia may serve as examples)
Some white (even re-cycled) paper and a pencil are of
much use now:
6.2.1. Draw the cladograms on a paper
6.2.2. Modify the cladograms to design the taxonomic area cladograms (substitute the species by the islands where they occur)
6.2.3. Now prepare the resolved area cladograms.
The whole process can be summarized as follows (again, try the Spanish - there is a slight Canarian accent, not easy to detect in the written version)
6.3. The general area
This is, essentially, one dendrogram that shows the information shared by (or dominating in) a series of resolved area cladograms. There is some diversity of methods and ways of thinking about this, which results in an amount of methodological complexity and varied software applications. We shall restrict ourselves to an intuitive search for shared patterns, and Brooks parsimony method.
6.4. Intersection of two conjuncts:
in the search for shared elements.
Take your set of resolved area cladograms as a working material. Search parts of these trees that are shared by all of them (if any; this is the best possibility). Otherwise, seek for parts that are repeated in a majority off the trees. There are several graphic approaches that are of use. In general, Venn's diagrams are the most useful in an intuitive approach. The figure below gives the basic idea:
Note that the clade (GC+TE), in blue, is present in the three cases. In the same way, all islands belong to the same set (black) in all cladograms. The relationship between the 'blue' clade and "LA" or "FU" varies depending on the tree, as does the link between the last two islands. Thus, the shared elements in these three trees are the clades (LA+FU+GC+TE) (black) and (GC+TE) (blue). Both subsets are compatible (there is no intersection of their border lines) and can be combined into the following General Area Cladogram: (LA, FU (GC, TE)).
As further exercises:
a) use this procedure with the resolved cladograms of Tarentola and Pimelia, and
b) devote five minutes to search for elements shared by all the resolved area cladograms you have. This last could result remarkably complex, depending on the patterns of distribution and the cladogram topology, so I think that it should be enough if you get the basic concept here. Is there any way to do this in a scientific way?
This method was designed in order to summarize the relations among area units, using the phylogenetic relations of endemic taxa as characters. More specifically, the clades of each taxon cladogram as characters (sounds terrible, doesn't it?). If a majority of the extant species did follow a more or less parallel speciation pattern, this method should permit to discover such 'general pattern'. This is more or less what we were looking for at the end of section 6.4. The method is as follows:
In the resolved area cladogram of each taxon, one has to 'label' the nodes (that is, the clades) with one letter or numeral. For practical reasons, in this exercise we shall use numbers that are consecutive to those found in Table 2, that is, equal or higher than '83'.
We shall then build one table of the kind of those in section 3 (that is, islands in rows and characters in columns). However, the clades or nodes (not the species) will be the characters here. In each case we shall code with a '1' the presence of one clade in one island, and the absence with an '0'. One clade is 'present' in one of the islands when one of the terminal taxa (species) belonging to that clade does occur in that island. Or, in other words, when the basal node of that clade drives (up the tree) to a tip or terminal branch where the island name is written. The next figure gives an idea of the process (however, read the 'Notes' down lines before starting to fill the matrix.
The species are themselves informative (we might think of each of them as one clade). I would not be so bad to combine these 'clades' with the 'supra-specific' ones, as in fact we already have this information coded (=presence/absence of species on islands). It is very convenient that, when doing this, we use the first (left) side of the matrix for the species (this has no effect in the analyses, but will enable us to recognize species by their numbers in the analyses).
One must insert and hypothetic outgroup with a '0' state in all characters.
One problem: "Missing areas". When none of the species of one endemic taxon occurs in a given area (missing areas in the area cladogram) we cannot really deny that part of the evolution actually happened in that area (followed by extinction). There are no species at present but they could have been there in the past... In such instances, we shall not assign a '0' value, but a question mark ('?'). See, for instance, the case of Crambe: following Table 2 there are no species in Lanzarote. It would be correct to write '0' in the column corresponding to the species, but we should assign a '?' to the clades of level higher to species in the other columns relating Crambe to Lanzarote.
This work is easily done as team work; whenever the number of students in your group is n>1, you have got at least two working teams. Each team may take charge of coding one area cladogram. Nearly 60 clades are regularly obtained. These, when added to the formerly available information on species, throws a total of ca. 140 characters.
Edit a file similar to that of section 3.4. Remember all details of the file format! (Exact number of characters, areas including the outgroup...).
Process the file with a program for phylogenetic inference based on parsimony (as before). If many dendrograms are obtained, perhaps a consensus tree is the best option. If only two or three cladograms arise, then it could be a good idea to study all of them.
With the results:
Onto a map of the Canary Islands, plot a phylogram representing the dominating relations we have just inferred. Note that this is based on both, the distribution of extant species, and on their phylogeny. Can you detect any directionality in the relations between islands? Do these results support the idea of a colonization from the African Continent? Same, an increasing diversification in westwards direction?
Can you propose any hypothesis on the past presence of any taxon in areas where it is absent at present? (Perhaps, any unknown ancestor of the species we know? For doing this it is interesting 'to map' some of the clades (characters, in our former analysis) on the general area cladogram (this can be done with Winclada, or just by hand following the instructor's instructions). More specifically: Could an ancestor of the extant Crambe have once occurred in Lanzarote? An ancestor of Calathus in La Palma?
7. DETERMINING AREAS OF ENDEMISM
We formerly applied the principles of cladistics (3.4) to data of presence/absence of species on islands. One extension of this doing may be applied to identify potential areas of endemism, as far as we can accept that such areas are those with at least two species not found anywhere else. This is one of the ideas underlying the method called PAE (P.A.E.= Parsimony Analysis of Endemicity, o "analysis of simplicity of endemisms"): One builds a cladogram of areas based on the presence/absence data; on the consensus tree obtained (or the single cladogram, if that is the case), those clusters of areas (clades) features by two exclusive species are designated as potential endemism areas.
The data matrix is just that used in section 3.4, and has to be processed in the same manner (the program Winclada is used for reference on procedures, as before). Whenever more than two trees are obtained, get the strict consensus of them (Trees, select all, consensus, strict).
Observe the distribution of synapomorphies (to see them, click on the red square icon, upper menu bar, the fifth one starting from your left) exactly as we did when working with Gallotia. Black (filled) circles are homoplasy-free synapomorphies. These are species that are present in all the islands included in that clade, and absent elsewhere. Only these are of interest now: every subset of islands featured by two or more of these species, are the areas of endemism we are looking for. We can know what are the species that identify each of those areas: click on the small red square in the menu (sixth one, starting from the left). Beware: depending on options, many times the program starts counting the characters from '0' instead as from '1' (this can be amended).
Given the high concentration of endemisms in the Canaries, it is easy to demonstrate that not just every island, but also some subsets of them, represent potential areas of endemism. Consider, for instance, the case of Lanzarote: Any comment? Is this result similar to that obtained using Brooks parsimony?
8. TOWARDS PANBIOGEOGRAPHY
I have not yet tried to develop this section.... One group of islands could certainly be used as an example. However, to the extent that this was originated by volcanic phenomena, the example would not be very realistic... As an emergency resource, the consensus obtained in the PAE analysis could be used as an analogy for drawing "generalised tracks". Then, choose the areas of endemism, and plot them on a map of the archipelago (The results of a compatibility analysis should generally be preferred to those of a regular parsimoiny approach). Feel free to try, depending on time- and software availability.
9. BIBLIOGRAPHIC REFERENCES
Espinosa Organista, D.; Morrone, J.J.; Llorente Bousquets, J. & Flores Villela, O., 2002. Introducción al análisis de patrones en biogeografía histórica. Las Prensas de Ciencias, UNAM, México. 133 pp. ISBN 968-36-9912-X.
C.J. & Parenti, L.R., 2001. Cladistic
Biogeography. Second edition. Oxford University Press, 187 pp.
B.C., 2002. Evolution on oceanic islands: moleculart phylogenetic approaches to
understanding pattern and process. Molecular
Ecology, 11: 951-966.
J.M. & Martín Esquivel, J.L. (Eds.), 2001. Naturaleza
de las Islas Canarias. Ecología y Conservación. Publicaciones Turquesa,
S.L., Santa Cruz de Tenerife. ISBN 84-95412-18-7.
Fernandez-Palacios, J.M. & Whittaker, R.J., 2008. The Canaries: an important biogeographical meeting place. Journal of Biogeography, 35: 379–387. (download PDF)
J.M.; Márquez, R. & Lizana, M. (Eds.), 2002. Atlas y Libro Rojo de los
anfibios y reptiles de España. Ministerio de Medio Ambiente. Madrid, 585
Pleguezuelos, J.M.; Márquez, R. & Lizana, M. (Eds.), 2002. Atlas y Libro Rojo de los anfibios y reptiles de España. Ministerio de Medio Ambiente. Madrid, 585 pp.
J.C.; Valido, A.; Nogales, M. & Martín, A., 2000. Lagarto
gigante de La
Gomera. Un fósil que vuelve a la vida. Quercus,
C.; Emerson, B.M.; Oromí, P. & Hewitt, G.M., 2000. Colonization and
diversification: towards a synthesis for the Canary Islands. TREE,
San Martín, I.; van der Mark, P. & Ronquist, F., 2008. Inferring dispersal: a Bayesian approach to phylogeny-based island biogeography, with special reference to the Canary Islands. Journal of Biogeography, 35: 428-499. (download PDF)
the organisms, phylogenetic hypotheses... (2003, not
Baldauf, S.L., 2003. Phylogeny for the faint of the heart: a tutorial. Trends in Genetics, 19(6): 345-351.
U.R.; Hilger, H.H. & Martin, W.F., 1996. Island colonization and evolution
of the insular woody habit in Echium
L. (Boraginaceae). Proc. Natl. Acad. Sci.
USA, 93: 11740-11745.
R.P. & Pestano, J., 1998. Phylogeography of skinks (Chalcides)
in the Canary Islands inferred from mitochondrial DNA sequences. Molecular
Ecology, 7: 1183-1191.
C.F.A. & Hurst, G.D.D., 1998. Mitochondrial DNA phylogeny of Brimstone
butterflies (genus Gonepteryx) from
the Canary Islands and Madeira. Biol. J.
Linn. Soc., 63: 69-79.
S.; Arnold, E.N.; Mateo, J.A. & López-Jurado, L.F., 2000. Long-distance
colonization and radiation in gekkonid lizards, Tarentola
(Reptilia: Gekkonidae), revealed by mitochondrial DNA sequences. Proc.
R. Soc. lon. B, 267: 637-649.
B.C.; Oromí, P. & Hewitt, G.M., 1999. MtDNA phylogeography and recent
intra-island diversification among Canary Island Calathus
beetles. Molecular Phylogenetics and
Evolution, 13(1): 149-158.
Fuertes-Aguilar, J.; Kim, S.C.; Santos-Guerra, A.; Crawford, D.J. & Jansen, R.K.,
2002. Phylogeny of the Macaronesian endemic
Crambe section Dendrocrambe
(Brassicaceae) based on based on internal transcribed spacer sequences of nuclear
ribosomal RNA. American Journal of Botany,
C.; Oromí, P. & Hewitt, G.M., 1995. Mitochondrial DNA phylogeny and
sequential colonization of Canary Islands by darkling beetles of the genus Pimelia.
Proc. R. Soc. Lond. B, 261:
C.; Oromí, P. & Hewitt, G.M., 1997. Molecular phylogeny of darkling beetles
from the Canary Islands: Comparison of Inter Island colonization patterns in two
genera. Biochemical Systematics and
Ecology, 25(2): 121-130.
J.; Brown, R.P.; Suárez, N.M. & Fajardo, S., 2003. Phylogeography of
pipistrelle-like bats within the Canary Islands, based on mtDNA sequences. Molecular Phylogenetics and Evolution, 26: 56-63.
D.J.; Emerson, B.C.; Oromí, P. & Hewitt, G.M., 2001. The diversification of
the genus Nesotes (Coleoptera: Tenebrionidae) in the Canary Islands: evidence
from mtDNA. Molecular Phylogenetics and
Evolution, 21: 321-326.
10. Notes (instructors)
(Student: If you were able to reach this point, you can also read this... there are no secrets here).
This materials were originally intended for students in the last course (5th of 5, first semester) of the carrier of Biology, in the UAM. The students of the course on Biogeografía are hence expected to have a relatively important background in general biology topic, and some technical skills related to data processing and statistical packages.
Text/net. A simplified version of the text to be printed is 'expected', but not yet ready. In the meanwhile, and provided that the software is available, this 'online-available' version can be used.
The selection of activities is by no means complete, but allow for a selection of them, or for variations on some of them.
It was designed in three basic block:
a) Correlation between species richness and biotic/abiotic factors
b) Classification of area units based on absence/presence data (species)
c) phylogenetic inference
d) phylogeography and basic cladistics
All are very simplified, theoretical explanations should be provided in advance.
Some parts can be easily organized for working in teams of 2-3 persons. For the more simplified approach, a minimum of 3 sessions of 2 hour is required; a more thorough work may well fill twice that time.
There is an ample selection of software that can be applied in this activity. I
suggested Winclada (+ Nona or Hennig) for a Windows environment, for it is more
'user-friendly' on the screen. There might be reasons to select packages that
are accessible for free, like PHYLIP). This should induce many variations
related to the edition of the files, and in the file formats, as well as in the
details of the analyses themselves.
Phylogenetic reconstruction: In the text, I have suggested to try a cladogram for the butterflies in the genus Gonepteryx. The reasons are that there are very few species, and few nucleotide sequences should be found (and highly compatible, at least by the date when these lines were written). Alternatively, a more complex approach to Gallotia is feasible, but requires much more time and detail. This could include something on the 'almost-fossil' G. gomerana/bravoana... from which there is genetic information in GenBank (a.n. AF306569). Related references are González et al., 1996, Mol. Phyl. Evol., 6: 63-71; Carranza et al., 1999, Herpetol. J., 9: 83-86.
The data: I have simplified them to some extent, avoiding some information on subspecies. The 'pre-edited' sequences have a couple of 'gaps' I added so that they are more easily aligned. The method proposed, as explicitly recognized, is rather primitive. You might wish to paste and align the sequences directly in GenBank (with the software associated), this may require a working e-mail account. The original Gallotia data from GenBank have codes: AJ238178, Z48040, AF019647, Z79499, AF306568, Z49751, AJ272395, AF080311, Z48036, Z48035, AF206534, Z79498, AF306659, ZZ49752, AJ272396, AF080312).
The body sizes of Gallotia in section 4.1 may not be completely correct, especially those of the Eastern species group. Alternative -more precise- approaches could be easily implemented, based in e.g. data from available field guide (the results shall prove somewhat more complex!)
Suggestions...: Do not hesitate to send whatever you thing, including suggestions, corrections, variations, ammendments, and the like. And: Make a trip to the Canaries! ¡Gracias!
11. Acknowledgements, for varied reasons to: Dr. J.C. Moreno, Dr. J. Martín Martín, Dr. C. Juan, Dr. A. Valido...
2. Numbers of species in islands
3. Grouping islands
4. What does the phylogeny tell about?
5. Reconstructing phylogeny
6. Combining evolutionary histories
7. Determining areas of endemism
8. Towards panbiogeography
9. Bibliographic references
10. Notes (teachers/supervisors)
go to: Main page (E. Garcia-Barros) Zoology/UAM (2nd year) Taxonomy Lab. (UAM) (4th year)