Hitching A Ride: Continental Drift and Aquarium Fish

Ever wonder how the fish in your aquarium got into whatever South American stream, Asian river, or African lake that it came from?

Probably not, but think about it now. For example, why are there lots of tetras from Africa and South America, and a very few from Central America, but none at all from Asia or Australia? Likewise, why are there no barbs native to South America or cichlids native to Florida?

And, on another topic, just why does California keep having all those earthquakes anyway?

Well, believe it or not, we now know that you can’t understand why aquarium fish families come from the places they do and why there are earthquakes in California without discussing the same root cause…continental drift.

Continental drift is the theory that the earth’s crust is made up of many semi-disconnected plates that shift and slide around and over each other over geological time scales. The continents ride on top of these plates, so over millions of years the continents will drift over the surface of the earth. The theory of continental drift was very revolutionary when first proposed by the German astronomer and meteorologist Alfred Wegener in 1912. Now however it is a well-established principle and it is used to explain everything from earthquakes to the distribution of tetras.

But as far-fetched as this theory sounds, we now know it’s basically true because we can actually measure the continents’ movements. Astronomical observations using radio telescopes scattered all around the world can be used to measure the distance between the telescopes to a very high accuracy, just a couple of centimeters. So, we can see that the distance between, say, Antarctica and Algonquin Park, changes by several centimeters every year. And you can also see the results of continental drift at the plate boundaries. For example, there are rows of orange trees that cross the San Andreas fault (the boundary between the North American and Pacific plates) which now have kinks in them because the trees on the west side of the orchard have drifted northward with the Pacific plate.

So how does continental drift explain why there are tetras in both Africa and South America? And how does it explain why Los Angeles periodically gets flattened?

First question: Africa and South America were once a single continent and the tetras hitched a ride on South America as broke away from Africa and drifted west.

Second question: Los Angeles is hitching a ride toward San Francisco as it drifts northward past the North American plate. But the trip is a little bumpy because the two plates’ rocks get caught on each other as they slide past. When the rocks suddenly break free the result is an earthquake.

Simple, isn’t it? Well, maybe not. So let’s back up a bit.

In recent years we have been able to figure out where the continents used to be. This was figured out (after a lot of discussion) from numerous clues in the magnetic orientation, location, and age of rocks; and from the height, location, and age of mountains; and of course from the current location, speed, and direction of the moving continents.

As it turns out, we believe that about 220 million years ago all the continents were gathered together in a single large continent called Pangea (“all earth”). Pangea then split up into two continents by about 150 million years ago. We call the resulting northern continent Laurasia and we call its southern contemporary Gondwanaland. Laurasia is then thought to have broken up into North America and Eurasia, while Gondwana is thought to have broken up into South America, Antarctica, Africa, Saudi Arabia, Madagascar, India, New Zealand, and Australia. Much later, South America is thought to have collided with North America and also Africa, Saudi Arabia, and India are thought to have collided with Eurasia, forming the arrangement of continents we know today.

So very old terrestrial plant and animal families that evolved way back in the Pangea days should be found on all continents because there would have been nothing keeping them from spreading throughout Pangea. This is the case, as the fossils (or in some cases, the still-living representatives) of the most ancient freshwater fish families have been found on all of the continents. The same is also true of course for the dinosaur fossils, which showed up just before Pangea is believed to have broken up in the Triassic period.

Of the later arriving families, only those that evolved in the southern continent of Gondwana should now be found in Australia and Antarctica because these continents have never been in contact with a Laurasian continent. But the flora and fauna of North America, Eurasia, South America, India, and Africa should contain a mixture of Gondwanan and Laurasian families.

And all these predictions are true.

Well, except for the part about Antarctica, which had the misfortune of sitting over the South Pole when the ice ages began and so it now has precious little in the way of either flora or fauna (Gondwanan or otherwise).

But remember that if a creature can either swim or fly between continents, or even catch a ride on a floating log, the gap between disconnected continents can be crossed without having to wait for continental drift to carry you to your new home. And of course, if some helpful human loads you on board an outrigger canoe or a Boeing 747, crossing over to new continents is a snap. So the distribution of plant and animal families are a lot more complicated than the continental drift theory would suggest. There are many distributions that are quite baffling, as animals occasionally show up in places where all logic says they could not. These anomalies must be attributed to incredibly rare events, such as a pregnant lemur floating from Africa to Madagascar forty million years after the island split off from the African mainland (no other explanation for the presence of lemurs on Madagascar has been forth coming).

Another complication arises because land masses that are now separated by seas may once have been connected simply as a result of lower ocean levels. For example, it is pretty certain that the North America and Eurasia were reconnected when the ocean levels dropped during the last ice age, and a land bridge formed across the Bering Sea. Many land mammals, including humans, seem to have taken advantage of this to cross over into the New Wold. And as many mammals made the reverse trip. This temporary reconnection of the Laurasian continents foreshadowed the predicted collision between North America and Asia. North America continues to drift westward, and so this should eventually close the Bering Sea, thus reuniting the Laurasian continents. At that time, however, the Laurasian continents are going to be joined in reverse order…with North America attached to the east side of Eurasia rather than to the west side as was the case when Laurasia was still whole.

And finally, climate change can also have much more obvious effects on the distribution of families in addition to opening land bridges by changing the sea level. As was already mentioned, Antarctica is a perfect example of climate change affecting distributions, for the Antarctic landmass is now nearly devoid of land species because of the unfortunate presence of a continent-wide glacier. But even a short-term change in climate can eliminate a family from a continent, leaving the paleobiologists of the distant future to wonder why they went extinct.

So let’s take the example of the distribution of tetras and see how we can understand it as a result of continental drift. The tetras are members of the family Characidae, and are believed to have evolved in Gondwana about 100 million years ago. This was after Gondwana is thought to have started to break up, and at that time we believe there was no tropical-climate connection between the still-existing South American/African landmass and the continents of Australia and Antarctica. A connection between these Gondwanan continents is thought to have still existed through the southern tip of South America, but this presented an impenetrable climactic barrier for the tropical tetras as it crossed over the South Pole through Antarctica to Australia. The family Characidae is therefore found only in Africa and South America, with a few species in Central America as well. We would predict that the African and South American species should be only distantly related as they were last in contact in the Cretaceous period. The Central American tetras should however be closely related to South American species, as they could not have crossed into Central America until the Panama land bridge formed, only five million years ago.

This turns out to be true: Central American tetras are closely related to the South Americans, but neither are particularly closely related to the Africans. All the African characins belong to the subfamily Alestinae. The South and Central American tetras are in another subfamily, called Characinae. The long separation between African and South American tetras has therefore resulted in the fish on the two continents developing along demonstrably different lines.

The minnows of the family Cyprinidae are thought to have evolved in Laurasia about the same time as the characins evolved in Gondwana. The cyprinids seem to have then expanded their range out of Laurasia as a result of continental collisions, and so they now are found in the Gondwanan landmasses of Africa, Saudi Arabia, and India. They are however still strangely absent from South America. This may be because North American cyprinids are almost exclusively temperate fishes, perhaps as a result of being unable to compete in tropical environments with the livebearers (family Poecilidae). The presence of poecilids in Central America may have barred the cyprinids from entering South America. The Eurasian cyprinids have no such problem however, and there are many tropical cyprinids in Southeast Asia. Undoubtedly their descendants will invade Australia as soon as that continent collides with Eurasia, which we believe will happen in another fifty million years or so.

The native Australian freshwater fauna that the cyprinids will invade is a relatively primitive group that the cyprinids should have no trouble competing against. The very ancient Pangean-era lungfishes (Ceratodontidae) are present, as well as the bony tongues (Osteoglossidae) and the Gondwanan-era galaxids (Galaxidae). Most of the other Australian fishes, like the barramundi (Centropomidae), rainbow fishes (Melanotaeniidae), and the eel-tail catfishes (Plotosidae), are the descendants of marine fishes that have only relatively recently reentered fresh water. Their ancestors may well have arrived in Australia well after it was isolated from the remainder of Gondwana.

Another great family of aquarium fishes is the cichlids. This family is believed to have evolved in Gondwana just like the tetras. They are now native to South America, Central America, Africa, Madagascar, Saudi Arabia, Sri Lanka, and southern India. They are absent from Australia. So what does this tell us about their evolution?

We can date the evolution of cichlids fairly well because of the fact that there are primitive (and closely related) cichlids in Madagascar, Southern India, and Sri Lanka: specifically the Indian genus Etroplus and the Malagasy genus Paretroplus. The family must therefore have evolved just before we think India split from Gondwana, about 100 million years ago. The relatively primitive Malagasy and Indian cichlids then survived in their isolated ranges, while more advanced cichlids took over in Africa and South America.

We would expect the Central American cichlids to be closely related to South American cichlids and South American cichlids to be only distantly related to African cichlids, and this is the case. The Central American cichlids are all members of the heroine lineage, a group related to the South American genus Heros. In contrast, the many African lineages are much more closely related to each other than to the Neotropical heroines. The existing cichlid lineages would therefore have evolved from an Etroplus-like ancestor after South America and Africa split.

But a very puzzling question remains about the distribution of cichlids. How did cichlids arrive on the islands of Cuba and Hispaniola? These islands are the native home of four cichlids of the heroine genus Nandopsis, but the islands are thought to have never been connected to the Central or South American mainland.

A possible explanation for the cichlids’ arrival in the Greater Antilles comes from their ancestry. Cichlids are “secondary-division freshwater fishes”, meaning that they evolved from marine fish that reinvaded fresh water. Cichlids are therefore generally much more salt-tolerant than the characins and cyprinids, neither of which have any marine ancestors. For example, the Etroplus species of India and Sri Lanka are commonly found in brackish estuaries. It is therefore conceivable that a cichlid could simply swim between islands, even though no cichlid has ever been seen in the open ocean.

There are however reasons to think that cichlids don’t cross oceans. Cichlids abhor open water, with the exception of few cichlid species living in the Great Lakes of Africa. But this relatively small group of open-water cichlids only exists because cichlids were the first predatory fishes to arrive in those lakes. They therefore found the open water habitant safe and were thus able to evolve adaptations for an open water lifestyle. But the open ocean is far too heavily populated with sleek, fast predators to be crossed by a coastal cichlid that is not adapted to open water life. Furthermore, the cichlids of the Greater Antilles are ambush predators, closely related to the jaguar cichlid (Nandopsis managuense) of Central America. All of these fish are unaccustomed to prolonged swimming and they have an especially strong attraction to cover. They would therefore seem to be quite unsuited for an open ocean crossing. They also somehow managed to arrive in Cuba without ever subsequently colonizing Florida, a relatively short 150 km away. This further casts doubt on their ability to swim from Central America to the Greater Antilles.

The cichlids’ marine relatives, the damsel fishes, also share their aversion to crossing open water, but in the case of damsel fishes they can colonize new habitats as planktonic larvae. Although damsel fathers will care for their eggs, the fry are abandoned to the ocean currents as soon as they hatch. In contrast, cichlid parents herd and care for their young until they become sub-adults and share their parents’ agoraphobia. We would therefore not expect misplaced fry to make a Caribbean crossing either.

Perhaps the islands’ pre-Columbian Arawak peoples learned of the cichlids from their Central American trading partners and imported the fish as a food source. This seems unlikely however because the Greater Antilles have only been inhabited for some 8000 years and a settled agrarian/trading economy has only existed there for some 2000 years. But the fish have already diverged from their Central American ancestor into four separate species, a feat that would require an unprecedented rate of divergence.

Or perhaps some cichlid eggs stuck to the feet of a migrating shorebird, and then somehow survived desiccation while tucked into the bird’s waterproof feathers for the flight north. The eggs would then have been left behind in an island pool during a rest break.

Or a hurricane may have somehow driven some coastal fish out of their lagoons and across the sea. One idea is that the fish would have stayed near the surface and under the heaviest rain, and thus in the least saline water to avoid the uncomfortably high salinity of the surrounding ocean. If the storm drifted eastward, they could have eventually found themselves in the Greater Antilles. But given the wind and waves that accompanies such storms, it seems unlikely that the cichlids would have found the ocean surface to be a hospitable place, even if it were demonstrated that the salinity under a hurricane is significantly lower than the general ocean. And anyway, almost all Caribbean hurricanes drift north-westward, not eastward, and so would be going in the wrong direction to guide a fish to the Greater Antilles.

Since none of these explanations seem particularly likely, even given the five-million year time span it had to occur in, the presence of the cichlids in the Greater Antilles remains a mystery.?