Paul Ericson
Dec., 1979*
Boulder, Colorado, USA
Email: barkingpo @

*This was a term paper originally written for an undergraduate/graduate biology class, EPOB 407- “Tropical and Insular Biology”, taught at the University of Colorado by Prof. Alexander Cruz, and continued and finished the next term (Dec., 1979) as an independent study project for Prof. Cruz. It has been reorganized and hopefully made more readable but no additional research has been done since then. It is placed here in case it might be of interest to those working in related fields, or to students looking for a project. Please excuse the lack of proper accenting for varzea, igapo, etc.


Seed dispersal by fish, or ichthyochory, is an overlooked seed dispersal mechanism that may have played a significant part in the evolution of seed plants. In an important paper on ichthyochory Gottsberger (1978) correlates this form of seed dispersal with relatively primitive flowering plant groups and concludes that “Migrating fish might be responsible in a decisive way for dispersal of many diaspores throughout the whole of the Amazon and its tributaries.” In certain environments ichthyochory has been a very efficient and reliable seed dispersal mechanism that may have played a significant role in the development and maintenance of dispersal strategies of the early angiosperms in the Cretaceous and possibly even the early seed plants as far back as the Carboniferous and Devonian.

In this paper I will describe a seasonally flooded environment in which ichthyochory still exists. Also, I will try to isolate some selective variables and develop a model for efficient seed dispersal that will help evaluate the importance of fish in primitive seed dispersal, the development of angiosperms, and competition between angiosperms and gymnosperms.


Varzea is is a term used generally for the Amazonian floodplains and covers approximately 64,000 km or 1 to 2 percent of Amazonia (Sternberg, 1975). This includes forests and savannas that are seasonally flooded, forming lakes and ponds that may dry up in the dry season. The number of months of flooding varies with the height above river level.

Gottsberger (1978), describes a backwater igarape (stream or creek) that drains local savannas in the varzea system of the Madeira River, a major tributary of the Amazon, most of which lies in tropical lowland areas of Brazil and Bolivia. He describes these brown waters as unproductive due to a lack of minerals and other nutrients but nevertheless supporting a large diversity of fish due to the abundance of allochthonous material, especially during flooding. The water level varies by 10 to 12 meters and most tree crowns emerged 5 to 20 meters above high water level while the width of the igarape varied between just wide enough for one canoe to 100 meters.

Lowe-McConnell (1975) describes varzea waters as being generally deoxygenated and very acidic due to decomposing debris and a lack of wind mixing. In open flooded savannas and varzea lakes, sunlight and wind mixing can increase productivity considerably (Lowe-McConnell, 1975). Varzea forests generally support important fisheries in many parts of the tropics (Lowe-McConnell, 1975). She writes that the often violent increase in water velocity may be accompanied by a drop in water temperature as well as changes in water chemistry and water turbidity.

Varzea forest shows less species diversity than terra firme forest and there is considerable difference in species composition between the two (Goodland 1975). Richards (1964) doesn’t know of any tendency toward single species dominance. It differs much in character in different places with an average height from 15 to 30 meters. Conspicuous trees include Ceiba pentandra, Carapa giuanensis, Hevea brasiliensis and Hevea guyanensis, Symphonia globulifera, Virola surinamensis, and many Leguminosae. Palms are frequent. Richards (1964) writes that the varzea forest may be regarded as an edaphic climax.

Two varzea forests in the white water areas of the Amazon valley were described by Irmler (1979) as nutrient sinks. Nearly 14% of the litter fall is presumed to be exported to rivers and lakes during the inundation phase but it is replaced through input of nutrients from inflowing white waters. Irmler (1979) observed that a 1 cm layer of sediments, mainly clay minerals, had been left after inundation of one varzea forest.


The majority of South American and 85% of the Amazonian fishes are of the order Ostariophysi (Lowe-McConnell, 1969; Roberts, 1973). Almost all South American groups are represented in the Amazon basin. The Charachoidei and the Siluroidei (catfishes) suborders of the Cypriniformes dominate the South American ostariophysins.

In Brazil there are over 1380 species of about 46 families of fish, with the largest groups present in approximately these proportions (from Lowe-McConnell, 1969):

Species % of total fish fauna

Siluroidei: 494 -43%
Charachoidei: 553 -40%
Cichlidae: 81 -06% (mostly lakes)


According to Lowe-McConnell (1969) there are two main types of breeding habits in the riverine fishes of South America:

  1. The fecund ‘piracema’ fishes.
  2. The small brood fishes showing parental care and few young at a time.

There is a close relationship between the hatch size, amount of movement, and food resources in seasonally fluctuating environments in these fish (Lowe-McConnell, 1969). Strong selection pressures favor rapid growth, early maturity, and short life cycles where seasonal fluctuations may drastically effect spawning success (Lowe-McConnell, 1969).

At the start of, or early in the rainy season, piracema fishes migrate up rivers and streams, either spawning in the rivers or running out in the savanna or forest to spawn. These fish usually produce large numbers of eggs and lay most or all of them at one time. ‘Piracema’ fishes include most of the South American characoids and many large siluroids (Lowe-McConnell). The eggs usually hatch quickly, within two or three days in the warm shallows and the young are soon free swimming. This strategy decreases the chance of being stranded by fluctuating water levels.

‘Small brood’ fishes exhibit parental care and produce few young at a time. They often start to spawn in the dry season, before the rains, and usually remain in one area (Lowe-McConnell, 1969).

The young fishes grow quickly in the streams and pools during the rainy season and the community is in a continual state of flux, some species moving upstream, or downstream, and others may be territorial, although all are influenced by the water level fluctuations (Lowe-McConnell 1969). Larger young leave the inundated regions at the same time as adults, when the waters begin to fall. Many juveniles are left behind and either live in ponds or creeks until the next floods or die when the waters dry up (Lowe-McConnell, 1975).

Morphology/ Family Characteristics

Characoids are mostly laterally compressed open water fishes, active during the daytime. They usually have well developed jaw teeth, often of a highly complex nature which seem to have undergone evolutionary changes in relation to foods consumed (Roberts, 1967; Lowe-McConnell, 1969). The larger characoids are basically predators, fruigivores, insectivores, or mud suckers and often generalize (Lowe McConnell, 1969). There is much overlap, especially in the dry season when few food resources are available except allochthonous forest products such as fruits, flowers, and flying insects (Lowe-McConnell, 1969). Lowe-McConnell (1975) also writes that this apparent lack of specialization in foods eaten is surprising in view of the extensive adaptive radiations of the characoids.

Siluroids, or catfishes, typically are bottom-dwelling fishes with nocturnal habits and have numerous small, simple, conical teeth (Roberts, 1973).

Knoppel (1970- not available) examined the stomach contents of 11,000 of 32,00 fishes collected from a clear water rivulet-varzea lake (Lago Calado) during the rainy season. The total of 53 species included 22 characoids, 6 gymnotoids, 7 siluroids, 17 percoids, and 1 cypronodont. The results showed little temporal or spatial separation within the species. Most species ingested many different items and the lists of stomach contents for the main families were relatively uniform. Despite the specialized teeth of many characoids, there was very little specialization for food items. The most widely consumed food items were immature stages of aquatic insects, found in 27 species, and plant material (including fruit), found in 26 species (taken from Lowe-McConnell, 1975).

Marlier (1967, 1968) studied the fishes of the varzea Lake Redondo. He found that littoral floating meadows provide the main source of food at the base of the food chain and commented that before the banks were cleared of forest, floating meadows were probably less extensive, and the allochthonous fruits and leaves from trees would have been a main food.


Ridley (1930) and Van der Pijl (1969) consider ichthyochory in several fish/plant associations but Gottsberger (1978) gives the subject a more serious examination. Fish frugivory is well known to the fishermen in the environments in question.

Marlier (1967) credits the shore forest with supporting the rich Amazonian fish faunas that live in those relatively unproductive waters. “Many species feed directly on leaves, seeds, fruits, or on terrestrial insects or other invertebrates which take their subsistence in the riparian vegetation. It is thus the forest which maintains the fish fauna at its present high level.” And Gottsberger (1978) states “The allochthonous productivity of the varzea and igapo vegetation determines the life of the fish and other organisms in the waters. There must therefore be a very strong interdependence of the vegetation and the fish, especially at the moment of seed dispersal.” In these flooded forests he says, “ichthyochory is a common and obligatory phenomenon.”

Gottsberger (1978) cites numerous observations of fish eating fruits and seeds and suggests in some cases a coevolution between fruiting time and fish frugivory on those fruits. He writes that fruiting time seems to be correlated with fish activity. Gottsberger notes that some of the seeds he considered are known to be dispersed by other animal vectors as well as passively by freshwater, and a few have been known to be dispersed by sea. Two species observed by Gottsberger, Virola surinamensis and Havea brasiliensis, are also named by Richards (1950) as conspicuous trees of the varzea forest.

Of the twelve seed-eating fish groups discussed by Gottsberger, 8 are characoids, 3 are siluroids, and 1 is a clupeid. The most notorious seed eaters belong to the genera Brycon and Colossoma.

Gottsberger found that of 33 plant species studied, 16 had seeds that were not regularly broken by fish and 17 had seeds that were. The former seeds belong mainly to very primitive groups whereas the latter belonged mainly to more recent groups.

From Gottsberger (1978): Names of plants whose diaspores seem to be
dispersed by fish (seeds intact) or which are triturated when swallowed
(seeds broken)

Annona Hypoglauca
Duguetia marcgraviana
Unonopsis aff. matthewsii
cf. Unonopsis

Virola cf.surinamensis
Ficus sp.
Cecropia membranacea

cf. Elaeocarpaceae
Neolabatia cuprea
Lucuma cf. dissepala

Licania sp. apetala
Licania sp.

Protium sp.
Simaba cf. guianensis
Astrocaryum jauary
Joarizeiro (not ident.)
Nectandra amazonum
Sorocea duckei

Anacardium cf. microsepalum
Carapa cf. guianensis
Byrsonima cf. amazonica
Byrsonima sp.

Mabea nitida
Hevea brasiliensis
H. cf. spruceana
Piranhea trifoliolata

Symeria paniculata
Genipa americana
cf. Randia armata

Tabebuia barbata
Sipia (not ident.)

He writes that this “confirms that ichthyochory is an archaic form of dispersal within primitive angiosperms that has been maintained in inundated tropical regions. The high frequency of ichthyochory is associated with nutrient-poor Amazonian waters, where fish depend substantially on allochthonous materials such as seeds and fruits which fall into the water. The homogeneity of the inundated vegetation in Amazonia may be the result of migratory fish dispersing diaspores.”

He concludes that ichthyochory may be partly responsible for the high percentage of plant species that are common to all varzeas of the Amazon, especially in light of the refugia theory concerning Pleistocene and post Pleistocene retreats of lowland rainforest during periods of dryness. Gottsberger quotes Prance (1973), “Today, many of the species of the varzea forests are the most widely distributed in Amazonia, partially because of the persistence of gallery forests in dry times and partly because of the ease of diaspore dispersal by water.”


Speculation as to the evolutionary importance of ichthyocory requires 1) an evaluation of its efficiency relative to other seed dispersal mechanisms as well as 2) a search of the fossil record with ichthyochory in mind.

1) Selection for Ichthyochory

Using the previously described ecological setting I will try to categorize some of the selective forces acting between fish and plant and limited by the environment- correlating possible biotic and abiotic selective forces with relevant characteristics of the biotic units involved (ie. seeds, fruit, fish, community structure). Important questions must be answered as to the fate of the seeds that are swallowed whole by the fish.

It should be noted that seeds may be a regular incidental intake in the varzea environment. Many herbivorous and omnivorous fish are known to eat seeds with other vegetable matter but ichthyochory has not usually been considered.

Seed Treatment and Specificity

Although various enzymes that digest specific carbohydrates have been found in the intestines in the pancreatic juice of fish, there is no evidence that fish have endocommensal bacterial fauna to break down cellulose plant materials (Lager et al, 1962). Herbivorous fish apparently have to rely on mechanical break down of plant cell walls. Therefore seeds swallowed whole are probably not damaged.

Some fish in the habitats in question are known to be attracted to certain fruits that may be used by fishermen as bait to catch particular fish (Ridley, 1930; Van der Pijl, 1969; Gottsberger, 1978; Sternberg, 1975). Sternberg (1975) writes that indigenous peoples of the Amazon may name a fruit by their associated fishes’ name, eg. “tambaqui fruits”, in reference to those fruits eaten by Colossoma bidens.

Some tropical fishes are known to have sensitive well-developed olfactory senses and imprinting has been shown to be possible at an early age in cichlids. In well mixed waters rich in secondary compounds one might expect a certain degree of general desensitization but in slow moving waters scent trails left by ripe fruit might allow some specificity. The acidity of the medium may aid in locating ripe fruit that was leaving a scent trail and perhaps accentuate its attractiveness. It may also be that fruit eaten by fish that live in acidic waters don’t have to be as sweet as fruit growing in less acidic environments in order to be attractive. Sternberg (1975) writes that Ducke (1949, not available) has suggested the reason that flood plain fruit are frequently tart as compared to the more sweet upland fruit is that they may have been selected for by the riverine fauna that eats them and passes the seeds unharmed.

Seed size would not seem to be much of a criteria for such a wide range of sizes of fishes, especially since the elastic intestine intestine and mouth parts of most fishes allow even small fishes to swallow foods of considerable size whole (Lagler et al, 1962). However, Gottsberger (1978) writes “Most of the fish that eat fruits and seeds are among the larger species.”

It has generally been accepted that having fewer larger seeds is an advantageous strategy in unproductive, slow growing, tropical forests (Janzen,1974; Harper et al, (1970); Stebbins, 1971; Smythe, 1970; and Snow, 1970). Smaller seed size and hence probably greater numbers of dispersal units would not be advantageous where there is a possibility of good selection of favorable sites for seedling establishment, and where water relations during initial germination in a damp environment (after waters subside) may select against it (see below).

Shape and color may have little value as criteria for specificity as competition for these resources (the fruit) would be very high here and sight may be difficult anyway in these often muddy waters.

Reliability of Visitation

Reliability of visitation to habitats suitable for seedling establishment is increased by the reliability of changes caused by seasonal flooding and the corresponding seasonal fish behavior.

The important factors here are the timing of release of fruit, the distance and direction traveled while the seed is being carried by the fish, the location of release of the seed, and the fate of the seed after release as compared to the fate of seeds that are dropped in the water and are not eaten by fish.

Regarding seeds dispersed by fresh water Stebbins (1971) says “since fresh water does not damage living tissues, mechanisms for aquatic dispersal of fresh water species can be relatively simple. In many instances, seeds or small fruits fall onto the surface of the water, are transported some distances and finally sink.” And, “A number of freshwater species have bladdery-inflated seeds or fruits that can float on the water for considerable periods of time.”

Timing of Release of Fruit: Timing of release of fruit can be very important for optimum dispersal and in seasonally flooded forests, as mentioned before, often corresponds with either the rising or lowering of the flood waters. Such phenological behavior is not uncommon and flooding or rising water and its attendant physical effects would satisfy the requirements to be a “suitable external environmental cue that is not dramatically damaging but is drastic enough to be sensed equally by all individuals of variable age, health, competitive status, past history and genetic programming.” (Janzen, 1974).

In this case the timing according to physical factors has accompanying biological implications, most importantly being an invasion of possible aquatic dispersal agents. Fruiting while the water is high generally limits the variety a of non-fish dispersal agents. Once the fruit drops into the water fish would have a decided advantage over other animals as effective vectors.

Predator satiation would not appear to be important here except in cases meant for passive water dispersal.

Distance and Direction Traveled: The distance and direction traveled by the dispersal agent is dependent on foraging and reproductive movements and water currents.

Fish migrations upstream may allow dispersal of considerable distances depending on the length of time of retention of seed and the speed of travel. For upstream colonization ichthyochory offers a considerable advantage over passive water dispersal. Non migrating fish may be restricted to relatively short lateral movements and others may remain very local, especially in territorial behavior. Specificity may be selected for in short distance dispersal for greater chance of meeting similar growing conditions.

Location and Release of Seed: The location and release of seed would have an influence on species composition and succession in the vegetation. Ichthyochory in this sense could probably not be distinguished from other dispersal mechanisms, except in lateral range, which would be limited by the extent of flooding. However, seedling establishment, via any dispersal vector, may be highly edaphic.

Floods often overturn trees, leaving depressions which may then be expanded or deepened by erosion. Such clearings are essential to the establishment of seedlings in mature rain forest. Air breathers and surface swimmers, common to Amazonian waters (Roberts, 1973) are likely to travel further laterally and stay longer after the waters begin to drop but most of the spawning fish are likely to spawn in depressions, holes, and ponds to ensure longer lasting water sources after they start to recede. Those fish that provide parental care and/or are territorial may possibly carry several or many seeds to the same nesting site, either by eating single seeds or fruit with multiple seeds.

After Release of Seed by Fish: After release, a seed would preferably sink either by absorbing fluid or being denser than the medium, perhaps so as not to void selection pressures that may have brought it to a suitable habitat. The seed could then ‘wait’ for waters to recede and ‘hide’ it from predators with a layer of mud and organic material, or it could float, like water dispersed seeds and immediately get caught on something as one extreme or eventually end up at sea as the other extreme. Receding waters depositing organic matter and sediments on top of the seed would provide valuable protection from predators, as well as providing a suitable moist, soft medium in which to germinate when conditions are favorable. An abundance of stored food in large seeds enables rapid root growth so it can be well established before producing large leaf surfaces (an obvious advantage in a location that is likely to have considerable runoff at any time). There would seem to be selection for large ichthyochorous seeds in this kind of situation. According to Harper et al (1970) “Large seeds may have difficulty in obtaining sufficient water for germination from temporarily available water supplies because of their low ratio of surface to volume” and “even when fully buried, large seeds may not extract sufficient water for germination because the volume of soil from which the water must be extracted becomes appreciable, hence the common garden practice of soaking large seeds.”

2) Evolution and Ichthyochory

Water is well recognized as a dispersal agent (Ridley, 1930; Van der Pijl, 1969) in both the present and in the early evolution of seed plants but its significance may have been overestimated. Stebbins (1971) states, presumably without consideration for ichthyochory, that adaptation for transportation by water cannot be regarded as a major factor for controlling the distribution or evolutionary trends in the great majority of flowering plants since water bodies of fresh water are usually limited in extent and need additional mechanisms for wide dispersal. Perhaps cases of passive water dispersal should be reevaluated with past ichthyochory in mind.

With the radiation of birds and mammals it is likely that ichthyochory has been unable to compete except in seasonally flooded habitats such as varzea. It is possible that ichthyochory may only be an efficient mechanism for certain plants whose adaptive capability has not allowed them to compete with other plants for limited numbers of more efficient vectors (such as birds or mammals) that only recently radiated to an extent that allows them to out-compete fish in those and other riparian habitats. However, I think it would be advantageous for highly edaphic varzea plants to use vectors that were restricted to the flood plain and its related soils.

It would be difficult to speculate on the degree of fish/plant coevolutionary specificity in a situation where seeds may be a regular incidental intake but long term plant strategies and evolutionary trends may have been strongly affected by a stable water-dependent relationship lasting many millions of years.

Transitions to other vectors may have been easy under suitable conditions but large river basin environments could have provided long term stability for the development and maintenance of ichthyochory regardless of global environmental variations that effected the evolution of other flora and fauna more dramatically.

According to Valentine (1978) “The details of the diversity and abundance of plant species through the Paleozoic and Mesozoic eras are largely unknown. The major transitions in the dominant floral elements resemble what was happening to land animals, but as far as one can tell they do not correspond to the events that were affecting the animals.”

Stebbins (1971) writes that changes in character syndromes of plants are more likely to occur than in individual characters and that interrelationships between characters are so complex that successful adaptive shifts usually demand simultaneous and harmonious changes in many parts of the plant. This suggests that a highly successful adaptation is likely to be retained in large part even after the selection pressure is removed. Traces of ichthyochorous adaptations might become obvious in other (dryer) habitats if plants and fishes were examined in that light. Previous adaptations for ichthyochory may show compromise adaptations to other forms of dispersal or may be more obvious than one might expect.

For example, Kramer (1978) proposes, as a possible explanation for unexplainable reproductive seasonality in fishes studied in Panama, that it may be a reflection of earlier evolution for spawning under particular conditions. Three of the six species studied ate fruits and seeds. One genus was Brycon. One of the other three was mentioned as being a fruit and seed eater by Zaret and Rand (1971).

Fossil evidence suggests periods of close correlation between the evolution of seed plants and fish.

The first seed plants developed in the Devonian and underwent considerable diversification in the Carboniferous (Valentine, 1978). These included the Pteridosperms, the first plants to have developed true ovules (Janzen, 1966) with seed coats and the ability to become detached (Delevoryas, 1966). These megaspores developed floating forms according to Potonie (1955). Van der Pijl (1969) mentions secondary characteristics in Pteridosperms which may have been attractive. This diversification mostly died out in the beginning of the Permian, and was followed by the expansion of conifers, associated with the appearance of dryer climates (Valentine, 1978). Pteridosperms were major contributors to the coal deposits of the Carboniferous.

During the Devonian there was a divergence in one evolutionary line radiating from the early lungfishes, according to Romer (1959), which may be of particular importance with respect to ichthyochory. Both adapted to shallow margins of lakes and streams (supposedly) in different ways with respect to their diet. One line, the lobefinned fishes, retained predatious habits and dentition, gave rise to amphibians, and became extinct. The other line evolved crushing dentition for feeding on shellfish and vegetation. This second line of evolution supposedly developed rapidly at first and then stabilized and remained relatively constant, giving rise to the modern lungfishes. One group can be found in South America, one in central Africa, and one in South Australia.

Radiation of the early seed plants corresponds temporally to the development of jaws (in fish), which greatly expanded the choice of food items, causing a “spectacular diversification” (Valentine, 1978). These fishes diversified into ray-fined fishes which are the ancestors of most modern fish, and the lobe-finned fishes which took several forms, most of which declined at the end of the Carboniferous.

Van der Pijl (1969) gives major credit to reptiles for dispersal of Pteridosperms in the Carboniferous but ichthyochory may have developed long before then. Regal (1978) suggests that the reptiles of the Mesozoic were not reliable visitors and sufficiently diverse to provide “quality dispersal”. Harsh chemical digestion suggests many reptiles could not have been good dispersal agents. Long distance dispersal by reptiles and early mammals (later) in wet, lowland environments might also be questioned.

Carter (1967) writes that vertebrates may have lived in freshwater environments from a very early stage in their evolution. The Rhipidistia (which were ancestral to the Amphibians), and other early osteichthyans lived in the shallow waters of fresh water swamps and pools rather than in large and deep lakes, under the warm and even tropical climates of Devonian times. Also, he writes that all the living osteichthyans (with the exception of Latimeria), and many amphibians, live in shallow tropical freshwaters.

As mentioned earlier, the characoids, with their extensive radiation and dental specificity that can’t be adequately explained, seem most closely associated with seed dispersal. The characoids are believed to be the most primitive (Greenwood et al, 1966) of the ostariophysins, with the siluroid offshoot appearing very soon after the characoids, in Gondwanaland and fresh waters probably in the Jurassic and the Triassic and prior to the Cretaceous when the continents split. The characoids provide strong evidence for the former connection of Africa to South America; the genus Brycon, the notorious Amazonian seed eater mentioned by Gottsberger (1978), and the generalized African genus Alestes appear to be the most closely related (Eigenmann, 1917; Myers, 1957; Weitzman, 1962).

The close parallel in behavior of Brycon and Alestes may be significant. According to Daget (1952), Alestes change their food habits seasonally, feeding heavily on west African flood plains on seeds, other plant matter, and insects, but subsisting on zooplankton in dry season pools. Alestes is known, at least in some species, to be potamodromous, as is Brycon, relying heavily on flood plains as feeding areas during the rainy season. They may have adapted similarly to similar conditions or perhaps they maintained previous behavior and adaptations since the Triassic or Jurassic with similar plant/fish coevolution. More information on Alestes is needed though, in relation to ichthyochory. Study and comparison of the two geographically isolated systems might be revealing.

In the Cretaceous the major speciation of the modern angiosperms occurred to spread throughout terrestrial environments, with gymnosperms declining. This period also corresponds closely with the major speciation of modern fishes and insects. By the end of the Cretaceous the reptiles and many of the large land mammals (largely undiversified) were suffering major waves of extinctions (Valentine, 1978) and the first modern birds begin to show. Most modern mammals begin to appear later, in the Paleocene and Eocene of the Tertiary period.

More notes and questions:

How close is the correlation between seed size and fish size in modern cases of ichthyochory? How important was passive water dispersal and the more efficient ichthyochory in selection for larger seed size and its implications for angiosperm evolution?

Why don’t most herbivorous fish have the endocommensal organisms for digesting cellulose? Perhaps there is a correlation between the chemistry of the water and the number of fish with or without endocommensal organisms. In an environment with many secondary compounds in the water and in the diet, especially with the periodic changes in concentrations during the year, the change in chemistry of the water or diet of the fish may be too harsh for such bacteria to adapt to.

It is also interesting that most fishes, if not all, apparently lack the ability to synthesize vitamin C (Chatterjee, 1973). Although it is not clear what requirements the fish have for the vitamin, salmon and trout are dependent on dietary vitamin C.

Can fossil seeds be found in physical proximity to fossil fish in any frequency?

What kind of seed bank exists in a varzea forest after a seasonal inundation?


Ichthyochory may have been a dominant seed dispersal mechanism effecting the evolution of seed plants for many millions of years. In light of the importance of water relations to both fish and plant and the importance of plants as a food source for fish, fish living in shallow water environments could have provided a very reliable and stable long-term influence in the evolution of seed plants. In tropical lowland forests during the Cretaceous gradually changing water courses and flood plains and occasional extensive flooding that would extend over savannas to connect river tributaries and spill into lakes could provide extensive lateral colonization and isolation of populations of plants from a genetic base being efficiently maintained and distributed by ichthyochory. This model would allow efficient competition of angiosperms with gymnosperms, especially in an environment where seasonal flooding would exclude many forms of low, dense gymnosperms and provide a means of genetic exchange between widely spaced individuals or populations along water courses. Further and more detailed investigation of the history and development of early seed plants and angiosperms in relation to corresponding changes in faunal and floral compositions, and floral vectors, and in correlation to geographical and climatic changes in the tropics where early seed plants are considered to have first appeared, may provide a range of correlations far back into the Mesozoic and Paleozoic.


Thanks to Libby Ericson (Mom) for last minute all-hours typing and emergency editing help on the original paper. Thanks to Eric Ericson (Dad) for the story about catching fish with pieces of banana while he was supposed to be doing geology field work in the Bolivian Amazon in the 1950s. Thanks to Prof. Alex Cruz for his encouragement and suggestions and allowing me to continue working on this very interesting topic for two semesters. Thanks also to G.T. Prance and W.M. Lewis Jr. for managing to read this in its original form, and for their encouragement.


  1. Carter, G.S. 1967. Structure and Habit and Vertebrate Evolution. Wash. U. Press.
  2. Chatterjee, I.B. 1973. Evolution and the biosynthesis of ascorbic acid. Science. Vol. 182: 1271-1272.
  3. Daget, J. 1952. ‘Memoire sur la biologie des poissons du Niger Moyen. I. Biologie et croissance des especes du genre Alestes.’, Bull. Inst. fr. Afric. noire, 14: 192-225. (From Lowe-McConnell, 1975. Not available.)
  4. Ducke, A. 1949. Arvores Amazonicas e sua Propagacas. Bol. Mus. para. E. Goeldi. Hist.nat. Ethnogr. 10: 81-92. (From Lowe-McConnell, 1975. Not available).
  5. Eigenmann, C.H. 1917. The American Characidae, Mem. Mus. comp. Zool. Harvard, 43: 1-558.
  6. Goodland, R.J.A. and Irwin, H.S. 1975. Amazon Jungle: Green Hell to Red Desert. An ecological discussion of the environmental impact of the highway construction program in the Amazon Basin. Developments in landscape management and urban planning. 1. Elsevier Scientific Publishing Company. Oxford, N.Y.
  7. Gottsberger, G. 1978. Seed dispersal by fish in the inundated regions of Humaita, Amazonia. Biotropica 10(3): 170-183.
  8. Greenwood, et al. 1966. Phyletic studies of Teleostean fishes with provisional classification of living forms. Bull. Am. Nat. Hist., 131: 339-455.
  9. Harper, J.L., et al. 1970. The shapes and sizes of seeds.
  10. Irmler, U. 1979. Considerations on structure and function of the Central Amazonian inundation forest ecosystem with particular emphasis on selected soil animals. Oikos.
  11. Janzen, D. H. 1966. Coevolution of mutualism between ants and acacias in Central America. Evolution 20:249-75.
  12. Janzen, D.H., 1974 Tropical backwater rivers, animals, and mast fruiting by the Dipterocarpacea. Biotropica 6(2): 69-103.
  13. Knoppel, H-A. 1970. Food of the Central Amazonian fishes. Amazoniana,2: 257-352. (From Lowe-McConnell. Not available).
  14. Kramer, D.L., 1978. Reproductive seasonality in the fishes of a tropical stream. Ecology, 59(5): 976-985.
  15. Lagler, et al. 1962. “Ichthyology.” John Wiley and Sons, Inc., New York and London.
  16. Lowe-McConnell, R.H. 1975. “Fish Communities in Tropical Freshwaters.” Longman, London and New York.
  17. Marlier, G. 1967. Ecological studies of some lakes of the Amazon valley. Amazoniana, I: 91-115.
  18. Marlier, G. 1968. Les poissonsdu lac Redondo et leur regine alimentaire; les chaines trophiques du lac Redondo; les poissons du Rio Preto da Amazonia, Manaus, Brazil. (From Lowe-McConnell. 1975. Not available).
  19. Myers, G.S. 1958. Trends in the evolution of Teleostean fishes. Stanford Ichthyol. Bull. 7: 27-30.
  20. Nikolsky, G. 1963. “The Ecology of Fishes.” London, New York, Academic Press.
  21. Pijl, L. van der. 1969. “Principles of Dispersal in Higher Plants.”
  22. Potonie, R. 1955. Zur biologie der karbonischen sporen. Palaont. Zs. Bd. 29, no. 1-2, p.27-32, illus. (ABSTRACT). Discusses the morphologic specializations developed in the Carboniferous spores as adaptations to wet and flooded terrain. Megaspores developed floating forms, while microspores specialized in equipment for anchoring to the ground or coupling with megaspores. (From: “Bibliography and Index of Geology Exclusive of North America.” Vol. 20. 1955. The Geological Society of America).
  23. Prance, G.T. 1973. Phytogeographic support for the theory of Pleistocene forest refuges in the Amazon basin, based on evidence from distribution patterns Caryocaracae, Chrysobalanacae, Dichapetalacea, and Lecythidacae. Acta Amazonica 3: 5-28. (Not available).
  24. Regal, P.J. 197x. Ecology and evolution of flowering plant dominance.
  25. Richards, P.W. 1964. The Tropical Rain Forest. Cambridge U. Press, Cambridge.
  26. Ridley, H.N. 1930. “The Dispersal of Plants Throughout the World.” Reeve, Ashford.
  27. Roberts, T.R. 1973. Ecology of fishes in the Amazon and Congo basin. IN, B.J. Meggers et al. (Eds). Tropical Forest Ecosystems in Africa and South America: A Comparative Review. Smthsonian Inst. Press, Wash: 239-254.
  28. Romer, A.S. 1959. The Vertebrate Story. Chicago U. Press.
  29. Simpson, B.B. and Haffer, J. (1978). Speciation patterns in the Amazonian forest biota. Ann. Rev. Ecol. Syst. 9: 497-518.
  30. Smythe, N. 1970. Relationships between fruiting seasons and seed dispersal methods in a neotropical forest. The Amer. Nat. Vol. 104: 25-35.
  31. Snow, D.W. 1970. Evolutionary aspects of fruit eating by birds.
  32. Stebbins, G.L. 1971. Adaptive radiation of reproductive characteristics in angiosperms, II: seeds and seedlings.
  33. Sternberg, H.O. 1975. The Amazon River of Brazil. Franz Steiner Verlag G.M.B.H. Weisbaden.
  34. Valentine, J.W. 1978. The evolution of multicellular plants and animals. Scientific American 239, no. 3.
  35. Weitzman, S.H. 1962. The osteology of Brycon meeki, a generalized characoid fish, with an osteological definition of the family. Stanford Ichthyol. Bull. 8: 1-77.
  36. Zaret, T.M. and Rand, A.S. 1971. Competition in tropical stream fishes; support for the competitive exclusion principle. Ecology 52: 336-342.

Copyright. All rights reserved.