14 min read

Selective Behavior as a Component of Speciation

Written by Russell McAndrews

Selective behavior, facultative behavior or behavior of choice are all phrases describing the implementation of memory.  Learning if you will.  Such behavior is present in man and has been exhaustively demonstrated with apes and other phyla.  Among these are a few groups of fishes.  As a rule members of the family Cichlidae, whether new or old world, whether lacustrine or riverine, all exhibit brood care, social structures and a high degrees of territoriality.  In conjunction, extreme plasticity and opportunity have equated to a massive exploitation and subsequent evolutionary radiative explosion of the Haplochromines in the Great Lakes of Africa.

Equatorial E. Africa is a region of rich biodiversity.  The formation of the Great Rift Valley caused upheaval and fragmentation of the land.  This fragmentation and the tropical climate of the region set the stage for a rapid morphological radiation.  The youngest of these radiations is the Lake Victoria Basin which is 150-750 thousand years old.  It is a shallow, saucer-shaped basin comprised of many isolated pockets of fishes.  Within L. Victoria alone, the genus Haplochromis has been estimated to contain 400-1,000 species.  Initial analysis of mtDNA of 14 Haplochromines across 9 genera (Meyer, et. al.) has shown that all were derived from a common ancestor.

Ten years ago, three quarters of a million years would not have seemed to be enough time for the development of an assemblage of this magnitude.  Today, the relationship between catastrophe and evolution is better understood and the concept of variable rates or punctuated evolution is not entirely foreign.  The evolutionary history of these animals may be very difficult to decipher from the end-product.

The newly formed L. Victoria underwent a founder event like few known to date.  While many families of fishes were present, only the cichlids underwent this rapid diversification.

Cichlids are secondary freshwater fishes, having evolved from marine ancestors, and all the Haplochromis (sensu lato) are maternal mouth-brooders.   Inherent physical and behavioral predispositions amounted to unique qualifications for the exploitation of opportunity presented by the creation of this new watershed.  Phenotypic plasticity is exhibited to a high degree by all Haplochromines (indeed all cichlids) and so it follows that the ancestral Haplochromine must have been very plastic itself.  Derived Haplochromine behavioral traits commonly exhibited are; behavioral and physical plasticity, territoriality, tolerance of a high dissolved mineral content trophic plasticity and parental brood care.  All of these would turn out to be pivotal survival traits.

With this initial radiation in place, a fluctuating lake level would have alternately opened new areas for dispersal and condensed existing populations into a diminishing habitat as the shoreline contracted.  Such an event would enhance competition for dwindling resources and result in mass casualties accompanied by some extinctions.

Fluctuating lake levels would impact the fish population, particularly the Haplochromines, in a couple of important ways.  It is probably best to portray each scenario individually to facilitate the description.

A rising level would almost never be catastrophic for a population.  There are some imaginable situations where that could happen, but for the most part, the animals are never even forced to relocate in their lifetime.  In actuality, many new opportunities are made available to any would-be explorers and also to those which may have been displaced.

In the fish’s sense, this is a brave new world.  Experience is missing; the chance is great and the payoff high.  These are especially good times to be an inquisitive fish.

Contrasting horribly is the dropping level scenario.  Just as the rising water expands resources, the shrinking water takes them away.  There is of course a key fundamental difference in that the more unfortunate of the two also, by definition, starts with a significantly higher total population and so carries the potential for more death.  Further exacerbating the problem is the fact that the resources are more heavily impacted often permanently when the level drops.  Aquatic and semi-aquatic plants which are flooded don’t die, during a drought; such species can suffer regional extinctions.  The lowest know water level (corresponding to the minimum surface area) took place 150,000 years ago.

Since then the lake has filled to an area about twice its present size allowing those species that survived the low water to colonize new habitat as it become flooded.  While no referenced authors mention it, I believe that this second dispersal event has profound impact on the phylogeny of the lake’s cichlids.  The decline of this high water to today’s level resulted in the fragmentation of the freshwater ecosystem as ponds and swamps became isolated from the main lake at different points in time while the water level was dropping.  The sequentially created demes provide a series of “snap shots” of evolution for resolving the dynamics in play.

Seasonal and historic environmental fluctuations in both lake level, and the existence and migration of a hypolimnion, have acted “naturally” to mold the community.  Anthropogenic perturbations, while not covered in detail here, have been extreme and include the introduction of several foreign species, over fishing, soil erosion/siltation and pollution (primarily fertilizer runoff).

The behavior of choice, or facultative behavior, while extremely difficult to quantify, nonetheless plays important roles in both adaptability and survivability.  Conjuring up an idealistic definition of facultative behavior is actually quite easy - that which is not instinctive.  Unfortunately, the problem with this definition is that it both implies and denies delineation.  By that I mean that such a clear-cut definition implies correctly that instincts and faculties are diametrically opposed to each other.  At what point does the animals behavior become facultative?  What specifically did the subject do to demonstrate the application of learning?  Behavioral assessment must be performed from the context in which it is offered, that is the perspective of the subject, in this case the fish.


I propose that selective behavior is an underestimated and overlooked component of speciation and evolution.



Attempts at resolving the phylogeny of the Haplochromines have highlighted what one might consider a flaw in our definition of a biological species.  The component in conflict with the currently accepted biological definition of a biological species is the supposed sterility of inter-specific hybrids.

The Haplochromine flocks of Victoria are capable of viable, even inter-generic reproduction.  While these flocks may be the exception rather than the rule, an understanding of the earliest evolutionary processes may reveal previously unknown phenomena which may be visible nowhere else.  As science explores new questions and uncovers new data, it must occasionally rework theories and definitions to fit.

P.H. Greenwood, the scientist most familiar with the L. Victoria assemblage, eventually recognized that the extreme plasticity of these fish would cloud and phylogenetic relations because of intra-species, morphological and habitat range overlaps, and that more definitive methods would have to be found.

More recently, allozyme analysis was generally looked upon as the answer.  This did produce bountiful eco-relevant data but was not systematically significant.  Failing to resolve any enzymatic differences, scientists turned to differences in DNA.  The results were the same.

Analyses of base pair substitutions and ergo genetic diversity proved little or no genetic variation within the entire assemblage.  At this point, Haplochromine relations are not yet confidently resolvable.

Behavioral observations of captive populations provide a unique perspective not available to past systematist.  Ranges of behavior and coloration are not typically available from a preserved specimens.

Environmental conditions do not permit field observations.  A good analogy is that of a fish in a bottle.  Taken out of contest it is difficult to discern all that the animal was capable of doing.  Life history, ontogeny, trophic and reproductive behaviors, and agonistics play a crucial role in channeling and organism’s evolution.  Systematists need to examine the physical and behavioral data of the living animal to get a clear and accurate picture.


Behavioral variance is noted to be especially plastic (Goldschmidt, et. al. 1990, West-Eberhard, 1989) as well as instantaneously rewarding (West-Eberharad, 1989).  Taking these traits one at a time I shall attempt to detail a perspective based upon extensive behavioral observations.  Differentiable phenotypes as well as infinitely variable characteristic gradients amount to measurable variance in traits.  Behavior and behavioral plasticity, while not as easily measurable, exhibit a similar pattern of variation within a population.  At any given point in time behavioral variance and mean within a population resembles the distribution of any other trait.  That is to say there may be little of no variation or there may be a great deal.  Behavioral deviation can behave sequentially or divergently and can steer variation or hold it to a stable course.

Variance within a population can exhibit differing dynamics for individual traits or traits may be directly or indirectly linked to one another.  Several phenomena are mentioned by West-Eberhard and Goldschmidt and one, the “allelic switch” theory has been generally accepted.  The allelic switch is hypothesized to be responsible for dimorphisms and dichromatism present in today’s animal populations.  Sexual dimorphism itself represents an early ontogenetic commitment to one set of traits or another.  Obligate behavior also varies with the commitment.  While obligate behavior is by definition fixed, dimorphisms are not and adult Haplochromines have been observed to change sex as well as trophic specialization.

Recent extensive gene sequencing (Sturmbauer, Meyer, 90) of these cichlids fishes have raised questions about the conventional notion of speciation as a process of accruing genetic variation.  Sturmbauer and Meyer demonstrate a critical separation of the processes of genetic and morphological variation.  Relative directions of these two processes need not be related in any way and the rates obviously not the same.  Examples of “living fossil” species which have remained physically unchanged can be said to be in morphological stasis and provide a subject in contrast to the well documented trend for genetic variation to accrue over time.

The physical, and therefore visible, aspects of morphology would seem to have little directly to do with evolution and yet it does a great deal.  Evolution is a process and so perspective is crucial.  The apparent pace of evolution varies directly with the alignment of these two, directional processes.  The concurrent alignment, via behavior of both phenotypic potential and genetic variation, exhibit’s a synergistic-like response.  In a very short period of time species seem to diverge a great deal.  To lend a real advantage in survivorship, behavior needs to be ultraconservative for obligate (instinctive) patterns and flexible as well as curious for facultative ones.  The instantaneous nature of behavior and learning implies operation on a different time scale than genetic variation.  Behavioral flexibility enables these animals to gain maximum use of their extraordinary plasticity to minimized competition by fine tuning or shifting trophic preferences many times within the normal life expectancy of the organism.  This is in contrast to the Darwinian notion requiring a multitude of generations to affect change.



Proportionality Effect

Small populations respond to chance events which cause fluctuations in allelic frequencies.  Population gene pool size and therefore diversity plays a critical role in the overall rate of speciation.  For a given species, the smaller the population, the greater its potential rate of evolution.  However, this is not an individually variable trait and is only discussed as a population characteristic for distribution to all genetic rate terms.

Population Structure

A variety of micro-demes exist for the Haplochromines.  Many are highly localized, often monotypic genera with territorial behavior and are therefore effectively quite isolated (Goldschmidt, et. al. 1990).  Genetic loads of Haplochromine populations, while considerably variable, have been shown to vary inversely with the size of the population.  Under the pressures of extreme territoriality exhibited by most of the Haplochromini deleterious alleles seem to have purged themselves.  Most Haplochromines exhibit no inbreeding depression through five or more generations (pers. obs.).  Opposingly, those species with pelagic natures seem to suffer from a high genetic load which exhibits itself as inbreeding depression within two generations.

A population structure intermediate to the extremes mentioned above are the numerous species complexes which respond like a meta-population.  Additionally, some complexes are localized while others are lake-wide.  It is likely that these complexes arose and are maintained from ancestral, inter-specific variation in predisposition for dispersal.  Alternately, these complexes may represent the second, more recent radiation which occurred as the lake refilled.


It is important to note that genetic drift is independent of morphological variation.  Genetic drift is another basal mechanism which is active only at the population level.  While its importance is key and well studied, its obligate nature excludes its detailed coverage here.


Instinct, although extremely important, from an evolutionary and behavioral standpoint is by definition fixed.  In this sense obligate behavior will be viewed as an across the board modifier which applies evenly to all con-specifics.  This is not necessarily true but represent an assumption of simplification.  For example, an individual may have slightly different instinctive behavioral patterns than its fellows.  Also, individuals and populations may exhibit differences in plasticity of the “fixed” traits.

The flexibility of these fixed traits represents opportunity for the individual.  It follows that within population variation, some individuals would have more or less rigid behavior and that this rigidity/flexibility is itself inheritable.

Ritualized behavior such as sexual selection and hierarchical structure is also of particular interest as a point of diversification.  All Haplochromine males exhibit egg-spots.  These small yellow to orange ovoid spots are specific enough in size, color and arrangement to allow identification of most species outright.  The role of these egg-mimics as an isolation mechanism (Goldschmidt, 1990, 1991) is not new.  Investigation, observation and correlation of egg-spots of L. Victorian cichlids for over a decade by the Haplochromine Ecological Survey Team of Leiden Univ. provided some interest and unforeseen interactions while supporting the reliability of egg-spots as tools of identification.  It seems that egg-spots do not always mimic the egg of the animal.  Variation of color and size of individual occeli occur on two related clines.  Generally, species from dark water (deep or turbid) display egg-spots which are larger and more brightly colored than the actual egg.  Those species known to spawn in shallow, brightly lit waters have egg-spots which are smaller and more pale than real thing, and those species inhabiting and spawning in the intermediate zone exhibit spots which most closely resemble the egg of the species.

Variation or variability of instinct may manifest itself and would be indistinguishable from truly facultative behavior.


Selective behavior represents the most probable chance of an individual getting ahead.  The application of learning has been demonstrated on various levels and by various forms of life.  Without the application of learning warning coloration would impart no benefit.  How an animal responds to its niche particularly at critical times that really matters.  Choices represent opportunity and threat.  When memory provides input, a facultative animal can make a beneficial decision which an inexperienced sibling would not necessarily have made.  The greater the surrounding level of environmental and biological heterogeneity, the more stimulating the habitat.  The greater the number and diversity of stimuli, the lower the probability that a fixed, instinctive reaction will yield maximum benefit to the individual.  In some such cases the potential resource can only be obtained by the atypical response.  Behaviorally-limited resources provide their own incentive and reward.  Behavioral mimicry does play a role in the learning of the masses but observations in captivity have shown that it is not a prerequisite for shifting trophic behavioral patterns.

Having established the need for flexibility of choice and the relatively instantaneous outcome (rewarding, threatening or neither) of that choice, the important role of memory cannot be over stated.  Without memory there can be only random choices and therefore no net benefit.  With memory, initial experiences with strong reinforcement (negative or positive) can be recalled improving perception and quickening that selection.  Only extremes seemed to be learned in such an acute way, that is, the most telling issue, reward or threat, is fixed for future responses while options of less obvious issues remain flexible.

Seasonal fluctuations in food resources produce alternate facultative phenotypes.  Some individuals choose to shift to another trophic resource which is presumably more abundant thereby lessening pressure on the initial food source.  Ergo, behavioral divergence leads and controls physical divergence within the conscripts of the physio-chemical environment.  A divergent trend in morphology is predisposed to continue diverging as this shift in niche ultimately reduces competition for both morphs.  Applied evolutionarily, this means there is no turning back for a species once it has begun divergence.  Obviously, this assumes that the instigating event is not of a cyclic nature and that no other events occur to redirect divergence into convergence.

Inter-specific competition when not mutually exclusive implies a slight variation in niches (Ricklefs, 1993).  Indeed, the concepts of the niche and competition are inseparable.  Goldshmidt and others have gone one step further to tie the ecological segregation of a niche to reproductive isolation.  Behavioral plasticity facilitates the breakdown of this and other isolations mechanisms in captivity.    In situ behavioral observations are inadequate at this point to shed definite light on the situation but it is believed that competition as well as a relatively rigid social structure prevent these isolation breakdowns in nature.

Facultative survivorship can easily be broken into two contributing areas.  Physical adaptability of the individual which can perhaps best be perceived as morphological plasticity, i.e., the sum of the potentials of all traits.  The potential itself is the derivative of curve plotted by the population distribution of a character variation and is represented by the area under the curve.  In the case of physical traits, total morphological diversity is proportional to the summation of these areas of potential.

Behavioral adaptability is believed to exhibit similar dynamics but on a much accelerated time scale.  Total behavioral diversity is, likewise, the sum of all behavioral potentials.  That is, the areas under all behavioral characteristic curves.  Additionally, each behavioral trait is capable of exerting a sphere influence over each physical trait (West-Eberhard, 1989).

The step-wise sequence as it is now understood assigns a preliminary importance to behavioral adaptability.


While many of the earlier works supported my hypothesis, the bulk of, and the newest of the referenced works exhibit an increased tendency toward behavioral factors.  I reject my initial hypothesis.  Science is addressing issues of behavior and evolution.  I feel that my hypothesis was flawed from the point of perception.  My common perspective at the time lagged several years behind current scientific effort.  While the ignorance of that common perspective is real, it does not follow that the ignorance is inclusive.

Behavioral plasticity controls physical plasticity in response to the environment.  Evolution, like life itself, is interactive.  Individuality and those mechanisms outlined above combine to yield an astronomical number of interdependent possibilities.

A significant rise in lake level affords a slight to modest survival advantage.  Inquisitive aggression, learning and memory, phenotypic plasticity and behavioral plasticity are all evolutionarily significant malleable traits.

A drastic drop in lake level exerts enormous stress on a population.  Those animals that best handle the stress will be highly selected for.  This, in turn, imparts an evolutionary advantage to those disproportionately few survivors.


  • Booton, G.C. & P. Guerst, 1991. Molecular Biology and the Conservation of Endangered  Species of  Lake Victoria Cichlids.  The Ohio State University.
  • Foose, T.J. 1986. Raiders of the Last Ark: The Role of Captive Breeding in Conservation  Strategies,  The Last Extinction.  The MIT Press.
  • Fryer, G. & T.D. Iles, 1972. The Cichlid Fishes of the Great Lakes of Africa. TFH Pub.
  • Goldschmidt, T. 1991. Egg Mimics in Haplochromine Cichlids (Pisces, Perciformes) from  Lake  Victoria. Ethology 88: 177-190
  • Goldschmidt, T. & J. de Visser, 1990. On the Possible Role of Egg Mimics in Speciations,  Acta  Biotheretica. Kluwer Academic Publishers.
  • Goldschmidt, T., F. Witte, & J. de Visser, 1990. Ecological Segregation of  Zooplanktivorous Haplochromine Species (Pisces: Cichlidae) from Lake Victoria  Oikos 58: 343-355.
  • Goldschmidt, T. & F. Witte, 1990. Reproductive Strategies of Zooplanktivorous  Haplochromine Cichlids (Pisces) from Lake Victoria Before the Nile Perch Boom,  Oikos 58: 356-368.
  • Graham, M.A. 1929. The Victoria Nyanza and its Fisheries, Crown Agent for the Colonies.  London
  • Greenwood, P.H. 1981. The Haplochromine Fishes of the East African Lakes. Cornell Univ.  Press.
  • Hert, E. 1991. Female Choice Based on Egg-spots in Pseudotropheus aurora. Burgess  1976, a Rock-dwelling Cichlid of Lake Malawi, Africa. Journal of Fish Biology  38: 951-953.
  • Hert, E. 1992. Homing and Home-site Fidelity in Rock-dwelling Cichlids (Pisces: Teleostei)  of Lake Malawi, Africa. Environmental Biology of Fishes 33: 229-237
  • Kaufman, L.S. 1986. Why the Ark is Sinking, The Last Extinction. The MIT Press.
  • Kaufman, L.S. 1991. A Fish Faunal Conservation Program: The Lake Victoria Cichlids,  Endangered Species UPDATE Vol. 8, 1: 72-75.
  • Kaufman, L.S. 1992. Catastrophic Change in Species-Rich Freshwater Ecosystems: The  Lessons of  Lake Victoria, Bioscience Vol. 42, 11: 846-858.
  • Kosswig, C. 1963. Ways of Speciation in Fishes, Copeia 2: 238-243.
  • Meyer, A., T. Kocker, P. Basasibwaki & A. Wilson. 1990. Monophyletic Origin of Lake
  • Victoria Cichlid Fishes Suggested by Mitochondrial DNA Sequences, Nature
  • 347: 550-553.
  • Meyer, A., T.D. Kocher, A.C. Wilson. 1190. Mitochondrial DNA Phylogeny of East African  Cichlid Fishes: Single Ancestor for the Species Flocks in Lake Victoria, Nature.
  • Ricklefs, R.E. 1993. The Economy of Nature, W.H. Freeman. New York.
  • Sturmbauer, C. 1992. The Evolution of Mouth-brooding in African Cichlid Fishes Inferred
  • from Mitochondrial DNA Sequences, Research proposal submitted to the American
  • Cichlid Assoc.
  • Sturmbauer, C. & A. Meyer. 1992. Genetic Divergence, Speciation and Morphological  Stasis in a Lineage of African Cichlid Fishes, Nature 358: 578-581.
  • West-Eberhard, M.J. 1989. Phenotypic Plasticity and the Origins of Diversity, Annu. Rev.  Ecol.  Syst. 20: 249-278.
  • Witte, F., T. Goldschmidt, J. Wanink, M. van Oijen, K. Goudswaard, E. Witte-Maas & N.  Bouton. 1992. The Destruction of the Endemic Species Flock: Quantitative Data on  the Decline of the Haplochromine Cichlids of Lake Victoria, Environmental Biology of  Fishes 34: 1-28.
  • Witte, F. & M.J.P. van Oijen. 1990. Taxonomy Ecology and Fishery of Lake Victoria  Haplochromine Tropic Groups, Zoologische Verhandelingen. National Natuurhistorisch  Museum, Leinden.