[Summary and Contents]

[CHROMOSOMAL SPECIATION]


 

PHYLOGENETIC PATTERNS PREDICTED BY THE CASCADING SPECIATION MODEL

The patterns of chromosomal phylogeny provided by the 5 sequences of chromosomal derivation seen in Sceloporus show some common features which may result from basic properties of the cascading speciation process: 

  1. All sequences of derivation except the 2n=40 clarki-melanorhinus complex end with populations which have either (to a first approximation) completely fused or completely fissioned (at least as polymorphisms) karyotypes. 
  2. In all sequences except for the very recent grammicus case, there are few or no chromosomally intermediate survivors, and even in grammicus at least 2 chromosomal intermediates towards the end of the sequence of derivation seem to be missing (or have such small ranges that they were missed by my extensive geographic sampling [Hall and Selander, 1973; Hall, 1973]). 
  3. 3. In the two sequences where there definitely are intermediate survivors (qrammicus and the sequence to the 2n=22 horridus), the karyotypes of these survivors suggest that long sequences of derivation have almost or strictly linear phylogenies. The only apparent exceptions to this linearity might be: 
    1. a) The P1 grammicus population (Hall and Selander, 1973), which may be derived from F6 by a re-fusion of the fissioned pair 6, rather than from the karyotypically more similar S populations; or 
    2. b) the 2n=26 mainland magister, which is derived away from its presumed place in the linear sequence by a pericentric inversion (Cole, 1970; Hall, 1973).
  4. The only known cases of Robertsonian polymorphism in Sceloporus involve the definitely terminal and near terminal FM populations of grammicus (Hall, 1973), or the P1 grammicus (Hall and Selander, 1973), which may also possibly terminate a short sequence of derivation. Significantly, all polymorphisms in grammicus involve only the largest and most exactly metacentric chromosomes. These chromosomes would be expected to show the most regular chiasma formation and orientation on the meiotic spindle. 
  5. Lastly, but most interestingly, in all sequences of derivation except for the very recently terminated grammicus (which may have reached its culmination only with the development of native agriculture in central Mexico--Hall, 1973), the chromosomally terminal forms are widespread and ecologically important radiations: 
    1. The 2n=46 merriami has 4 subspecies and is common in areas of major rock outcropping from the Big Bend region of the Rio Grande Valley south to the barrier of the sandy deserts of south-central Coahuila (Olson, 1973). 
    2. The 2n=40 clarki-melanorhinus complex occupies tree and rock habitats from Guatemala north into the SW US along the Mexican Pacific and Gulf of California coasts and slopes (Smith, 1939; Smith and Taylor, 1950). 
    3. c) The 2n=24 scalaris-aeneus complex occupies grassy habitats over most of the Mexican Plateau from Oaxaca north to southeastern Arizona and adjacent New Mexico, 
    4. Finally, the 2n=22 complex of mostly parapatric large-scaled species is the dominant or near dominant trunk-ground-rock lizard in all habitats of almost all North America from western Panama to the northern limit of lizard distribution. These lizards are missing only from the majority of Baja California, the most extreme deserts, the southern part of the lower Mississippi Valley, and the lowlands of Veracruz.

All of the above features common to the 5 sequences of chromosomal derivation are predicted and can be explained by logical extensions of the cascading speciation model. 

Specific predictions of the model

The cascading speciation model discussed above generates a series of verifiable predictions about patterns of evolutionary relationships and the final states of the genetic systems of the chromosomally derived species. These are outlined below. 

  1. Sequences of chromosomal derivation should be mainly linear from origin to termination

In lineages where chromosomal derivation is frequent, i.e. where one would expect to see effects of the cascading amplification process, sequences of chromosomal and phylogenetic derivation should be mainly linear from origin to termination, rather than highly branched: Until stopped by a termination situation, because of the amplification process each most recently derived species (at derivation level L1) should be more likely to give rise to still further derivatives (derivation level L2) than would be the direct ancestor of this recent derivation (level L0). Since each new species at level L1 will most likely occupy a niche which is parapatrically or sympatrically adjacent to its direct L0 ancestor, it will then be more difficult for the L0 species to form a second L1 species in the adjacent environmental space occupied by the first L1 species. On the other hand, at least until an L2 species is actually formed, the L1 should have one side of its environmental space free of ancestral competitors. In other words, a chromosomally derived population will have a greater probability of surviving in an area of the range of the ancestral species where the newly isolated population can readily develop an optimal adaptation to a niche or environmental conditions which are not so effectively being used by either the ancestral population or by other close relatives. 

In the example of Sceloporus grammicus, areas of excessively high or low humidity, although occupied by the ancestral species in areas where derived populations are absent, are probably suboptimal for it. Once chromosomal differentiation blocks gene introgression from the widespread ancestral form, the derived population is able to optimize its adaptation to the originally suboptimal habitat, and can then spread through it to displace or exclude the ancestral form from it. Once this habitat is efficiently filled by the derived form, further speciation into it from the ancestral species will be much more difficult. It is also quite likely that the derived form can spread in this new optimal habitat beyond the original range of the ancestral species. In the case of the grammicus radiation, sympatric habitats are throughly saturated by other Sceloporus and sceloporines, so the sequence of chromosomal speciation has been limited to geographic exclusion along climatic axes of the environment.

 

  1. The sequence of ecological or geographic derivation will closely parallel the sequence of karyotypic derivation

    Directly following from the same arguments used above to make prediction #1, above, sequences of ecological derivation (evolution along environmental axes due to character displacement in sympatry) or geographic relationships (geographic exclusion and parapatry along geographic and climatic axes) should closely parallel sequences of chromosomal derivation. Geographic relationships in the grammicus sequence to the FM populations fit this, but other sequences in Sceloporus are too old and too many intermediates are missing to test the predictions. 

 

  1. Terminal species in a sequence will either be a) ecologically very specialized or b) ecologically dominant over near relatives. 

In cases involving early or even eventual sympatry, terminal species will be either: 

  1. very specialized ecologically, as a result of being excluded into extreme niches through competition with the series of ancestral species; or 
  2. especially good competitors, as a result of having retained many ancestral adaptations and of having ecologically dominated these ancestors. 

Which alternative is realized will depend considerably on the initial ecological specializations of the ancestral species and on how saturated the environment is with competitors when the sequence of derivation begins. As I show below, given appropriate conditions (e.g. initiation of chromosomal speciation well within the periphery of the ancestral species), contact hybridization in the chromosomal speciation process will provide a group selection process which very strongly favors chromosomally derived populations with higher general fitnesses than their direct ancestors. The 2n=22 Sceloporus have completely dominated the North American continent, while grammicus are most likely constrained to remain wood-crevice users. 

  1. Side branches in a sequence will usually either be proximal or terminal. 

If a lineage does have side branches, these will most likely be either early derivatives or else derivatives of the chromosomally terminal populations: Early in a radiation the environmental hyperspace should be comparatively far from saturation, and an early species in a sequence should find it easier to form a second derivative at any given level than would a species formed later in a sequence of derivation, when the environmental space would be more nearly saturated. 

In the large-scaled Sceloporus there were two early derivations from the 2n=32 ancestor: the fissioning sequence to clarkii and the fusion sequence to the 2n=22 horridus. In grammicus, P1 and F5 are early branches into spaces not occupied by more derived forms. As noted in 3b above, at least some terminal derivatives are expected to be superior competitors. Given that cascades are likely to take place and to be terminated in geologically short periods of time, and that usually we will be looking at the results of such a cascade some time after its completion; the terminal, highly competitive species may have had opportunity to overlap its ancestors and to spread and speciate geographically. The Pleistocene geographic speciation of the 2n=22 horridus very clearly fits this pattern. In grammicus, it appears that chromosomally derived forms F5+6 through FM2 exclude S from habitats that appear to be optimal for S where the derived forms are absent. The 2n=26 magister, which differs from the strictly linear sequence by a pericentric inversion is the only mid-sequence branch known in Sceloporus, and here we have no evidence that fixation of a pericentric inversion would facilitate speciation .

  1. Missing species in a sequence will usually be chromosomally intermediate

If species are missing (i.e. extinct) from an obvious sequence of derivation, the missing species will most likely be intermediate in the sequence: The ancestral species, which probably is relatively old and evolutionarily conservative, will most likely be well enough adapted and widespread enough geographically to survive competition with its derivatives. As noted above in 3b, terminal derivatives may be exceptionally good competitors, and by being terminal (assuming termination for intrinsic reasons) they should also have no closely related competitors on one side of their ecological space. However, two factors work against the prolonged survival of intermediate species: 

  1. they have closely related competitors on two (or more) sides of their ecological space, and 
  2. they may give rise to derivatives that are better competitors than they are before having much time to spread geographically. The intermediate with a limited geographic range could then be easily exterminated by its competetively superior derivative. 

This is especially well demonstrated by the sequence of derivation in Sceloporus grammicus, where the maximum range that could have been occupied by intermediates between the F5+5 and FM populations is very small. 

  1. Many terminal species will have "used up" their chromosomal substrate for speciation

Many sequences may terminate with species which have no more possibilities for chromosomal mutations of the kind used in their sequence of derivation. This is justified under the discussion of the first kind of cascade termination above. In Sceloporus, this is true for all sequences except that leading to clarki

  1. Species polymorphic for the kind of chromosomal mutation involved in a sequence of derivation will usually be terminal. 

Chromosomal polymorphism for mutations normally differentiating species in a sequence will most frequently be found in species which terminate sequences of derivation: Cascade amplification will frequently boost chromosome mutation rates to sufficiently high levels that species at the end of a sequence of derivation will be forced to evolve mechanisms to obviate the fitness reducing effects of these mutations. Terminal species would therefore be left with high mutation rates and meiotic adaptations to eliminate fitness reducing effects in heterozygotes, hence there would no longer be selection to prevent polymorphism for the chromosomal rearrangements being introduced by the high mutation rates. FMl and FM2 grammicus, which are at the end of a long sequence of derivation are both polymorphic for fissions. P1 may be an exception to this rule in grammicus, if it is assumed to derive directly from S; on the other hand, if it derives from F6 it does terminate a short sequence of derivation. 

 

Testing the predictions of the cascading chromosomal speciation model

As shown, all of the above predictions are demonstrated by one or more lineages within Sceloporus. However, since the Sceloporus data were used to suggest many of the components of the basic chromosomal and cascading speciation models, true verification should be based on the examination of other radiations to avoid circularity. Paull et al. (1976) suggest that patterns similar to those documented in Sceloporus are also found in other iguanid lineages, but most of the available detail is either insufficient, as with the hints of chromosomal variability found in interestingly large proliferations of South American branches of the family, or the radiations are too old and/or confused by other factors, as with the beta Anolis and the alpha Anolis radiation in Puerto Rico and its derivatives. 

On the other hand, although it too is much less known than might be wished, the seemingly recent sequence of chromosomal derivation noted in the Anolis monticola complex of Haiti (Webster, et al., 1972; Williams and Webster, 1974) provides one iguanid example that seems to fit several of the predictions and may be used to demonstrate how the cascading speciation model can be tested. This complex is restricted to high elevations of the Tiburon Penninsu and includes three species which are clearly derived from a conservative 2n=36 ancestry. Quite unusually for such closely related species of Anolis, the species in the complex are either definitely syntopic or are geographically very close to one another (collection records are sparse). A. monticola, the most highly derived species, shows 2n's of 46 to 48 due to fixation for fissions of 5 of the 6 ancestral metacentric macrochromosomes, and polymorphism for the 6th fission. This fits predictions 6 and 7. It is also geographically the most widespread and ecologically diverse species in the radiation (prediction 3b). Two other derived species have been described: koopmani, with a 2n=40 (2 fissions fixed), which is known from only 2 closely adjacent localities; and rupinae, with 2n's of 39 and 41, which is known from only one locality. A. rupinae may be polymorphic for 2 different fissions (observed macro-chromosome numbers are 7 and 9 based only on meiotic preparations from only 2 individuals--a B-chromosome or extra microchromosomal bivalent is also seen in these two individuals). Three other unkaryotyped specimens from this radiation may belong to a closely related but still undescribed species or may represent geographically separated populations of rupinae. The chromosomally intermediate koopmani and rupinae both appear to have limited ranges and to be early derivatives in the sequence to monticola (prediction 4). Also many chromosomal intermediates in the radiation appear to be missing (prediction 5). Only the polymorphism in rupinae, assuming that the limited karyological data are valid, does not fit the predictions of the cascading speciation model. 

However, as interesting as the Anolis monticola radiation is, it should be reemphasized that the area in which it occurs is still poorly sampled, that the species collected are still karyologically poorly known, and that their habits have undoubtedly suffered extreme modifications (degredation) in recent history due to Haitian land use practices. Obviously the predictions need to be tested in many other radiations of chromosomally variable organisms.


[THE ROLE OF CONTACT HYBRIDIZATION AS A BARRIER TO GENE FLOW BETWEEN PARAPATRIC POPULATIONS]

[Summary and Contents]