The Rhipsalis Riddle - or the day the cacti came down from the trees
Part 1

Copyright, Dr. Phil Maxwell, Bathgate's Road, Waimate, South Canterbury (Copyright November 1998. Reprinted with permission of the author. This article originally appeared in the New Zealand Cactus and Succulent Journal.) email:

"Rhipsalis is a highly specialized genus, unlikely to have developed during the Mesozoic and certainly neither ancestral to the rest of the family nor even related closely to the ancestral stock." Benson 1982: 115

"The Rhipsalidinae certainly yield in antiquity to no other cactus. That they are 'derivative' is plainly impossible." Croizat 1961: 759

Several years ago I was browsing in a bookstore when I came across a book on dinosaurs I hadn't seen before. Most dinosaur books make me yawn, but this one had quite attractive though stylised paintings, and one in particular made me pause as it showed dinosaurs disporting themselves near an Opuntia. This made me sigh as everybody "knows" that cacti weren't around in the Mesozoic. I closed the book hurriedly and left the store, but I couldn't help thinking about that illustration. Could the artist be right after all? Cacti have no fossil record so how would we know if any were contemporaneous with dinosaurs? In fact we can infer a great deal about the phylogeny (ie the genealogy) of organisms without any fossil evidence at all, as I hope to show in this article.

Most books or articles on cacti mention the fact that members of the family are almost entirely restricted to continental North, Central and South America and the islands of the Caribbean. There are only two exceptions - a couple of endemic genera (Brachycereus and Jasminocereus) and several endemic species (or subspecies) of Opuntia on the Galapagos Islands (about 800 km west of Ecuador), and the genus Rhipsalis, which as well as being present in the rain forests of much of South and Central America, occurs throughout a large part of central Africa, Madagascar, the Comores, Seychelles, Mascarenes and Sri Lanka. What is remarkable is that most publications devote little or no space to a discussion of the occurrence of Rhipsalis in the "Old World", or if they do write it off as dispersal by birds. Rowley's discussion (1978) is rather more critical than most, but it is far too brief (and inconclusive). Benson (1982: 114-116) devoted more space to the problem than other authors, and although I don't accept his conclusions he did at least discuss the competing ideas in some detail. However, I think it is time to look afresh at the problem in the light of modern ideas on historical biogeography.

Biogeography is, to put it as simply as possible, the study of the distribution of organisms (including extinct ones). [Some authors use terms such as "phytogeography" and "zoogeography" for the study of the distributions of plants and animals respectively, but what does one do about fungi or bacteria or other groups that are neither plants nor animals?] Biogeography is usually divided into two disciplines that are certainly not completely separate, but tend to be studied by biologists with different agendas. One is "ecological biogeography" which deals with the ecological factors influencing the numbers and types of organisms living in a given area - it has an offshoot devoted to the "evolving relationship between humans and their environment" (Allanby 1985). This has a good deal to do with current concerns about "biodiversity" and deciding how many species can be contained within areas set aside for wildlife and the like.

I certainly don't want to play down the importance of ecological biogeography, but my own interest is in the other branch, "historical biogeography", which seeks to explain how organisms come to have the distributions they do (and by extension, the reasons why extinct groups had particular distributions). It is a study of fundamental importance in evolutionary theory, even if it has been downplayed by some contemporary biologists.

The scientific exploration of the planet by Europeans really got under way during the 18th century, even if the various expeditions were not exactly free from commercial, political and military agendas. Prior to this few biologists ever thought very seriously about biogeography: plants and animals occurred where God created them, or spread after disembarking from Noah's Ark. However, it soon became clear that organisms do not occur randomly across the globe. Why, for instance, are there so many marsupials in Australia and New Guinea, and in South America, but except for the Virginian Opossum in North America, none in the rest of the world? Plants often have strange distributions too, an obvious example being Nothofagus (southern beech), which occurs in New Zealand, Tasmania, New Caledonia and South America. The study of fossils (paleontology) also came to be taken seriously in the late 18th century and only added to the confusion, for it soon became apparent that some groups of plants and animals had distributions in the distant past very different from today.

What proved to be the most important of all the intellectual voyages of discovery didn't even have the pretence of being for scientific purposes. This was of course the voyage of HMS Beagle from 1831-6, which set out to map the coast of South America as accurately as possible for the benefit of the Royal Navy. Charles Darwin was invited to take part as a gentleman companion for Capt. Robert Fitzroy (later to be Governor of New Zealand from 1843-45). He set off with two major pieces of intellectual baggage - the fixity of species, and a stabilist view of geology. One of the books he took with him was the first volume of Charles Lyell's "Principles of Geology", one of the most influential scientific works of the 19th century.

In Patagonia he unearthed the skeleton of a giant sloth and wondered how it was that such a large animal had become extinct so recently. He also collected cacti, including one later described as Opuntia darwini, but in general he was less interested in plants than in animals and geology. His experiences in the Galapagos - where the governor of the islands told him that he could tell just by looking at a giant tortoise which island it came from - eventually led him to abandon the widely held idea that species were immutable and to formulate his theory of evolution by natural selection.

Darwin, however, never seriously questioned geological stabilism. Geologists in the 19th century certainly didn't deny the evidence for vertical movements of the Earth's crust (ie in earthquakes and by extension, mountain building), but maintained that the continents had remained in the same relative position throughout time. This was to pose all sorts of problems for biogeographers throughout the 19th century and beyond. There were only three options open to those with a stabilist view.

First, there was the idea that the present ocean basins were formerly occupied by land-masses which for reasons unknown had later foundered. Oceanographers have searched in vain for evidence for sunken continents, although this has never daunted the more extreme proponents of lost continents such as Atlantis and Mu.

Second, the continents were connected by land-bridges which allowed plants and animals to disperse (or "migrate") before becoming submerged. It wasn't a preposterous idea of course - there is a perfectly good land bridge between North and South America, and the remnants of one between north-west North America and north-east Asia. However, there simply is no evidence for most of the other land-bridges proposed by biologists.

The third approach ignores geological explanations entirely and explains what are known as disjunct distributions by passive dispersal of organisms by such means as wind or ocean currents, or by other organisms such as birds. This is usually known as "jump dispersal" or "long-range dispersal". Not an unreasonable approach of course - there are many seeds with adaptations that allow them to be blown considerable distances, and there are small freshwater bivalves that attach themselves to the legs of waterfowl. Large land-dwelling vertebrates present more of a challenge, but this has not stopped the more extreme dispersalists from coming up with a solution - the favourite is to envisage them being transported on rafts of vegetation carried by convenient ocean currents. This according to some is how giant tortoises and the ancestors of the land and marine iguanas got to the Galapagos. A few years ago a locally produced TV documentary used a similar explanation to account for the presence of an iguana in Fiji - its ancestor was alleged to have drifted on such a raft for several thousand kilometres across the Pacific from South America! Darwin himself spent much time immersing seeds in salt water to see how long they could remain in the oceans until they found suitable landfall. Some seeds have an impervious coating but others soon became water-logged. [Nothofagus seeds fall into the latter group.] Dispersalists admit quite freely that some of the processes they envisage are highly improbable but argue that given enough time almost anything is possible - not a very satisfactory explanation. After all the molecules making up my computer keyboard could in theory all move upwards at the same time and take it through the ceiling, but I have no intention of attaching it to the bench with superglue! The dispersalist theory is often called "centre of origin" biogeography: a species evolves in some restricted area, then disperses far and wide.

Of course, there is a very viable alternative to stabilist geology. The story should be familiar by now, how the German meteorologist Alfred Wegener proposed the idea of Continental Drift in 1912 partly to explain the similarity in shape and geology between the Atlantic coasts of South America and Africa; how it was vilified by most earth scientists for about five decades until the mid-60s when some inspired geophysicists came up with the idea of sea-floor spreading and later, plate tectonics, all to the immense discomfort of traditional geologists. Of course things were never quite that simple - in fact it was a group of geologists in the southern hemisphere who kept Wegener's theory alive; Alexander du Toit in South Africa, Warren Carey and the New Zealand-born Lester King in Australia, and John Bradley in New Zealand. In Britain it was Arthur Holmes, who even came up with a mechanism for continental drift. Geophysicists in the meantime were more noted for "proving" continental drift was impossible. The theory held on in the southern hemisphere because the evidence has always been strongest here - the concept of Gondwana and its break-up have remained the cornerstone of the theory. It is also worth noting that some biologists were aware of the implications of continental drift long before plate tectonics became accepted. One who is relevant to our story is the American botanist W.H. Camp who published a paper in the Journal of the New York Botanical Gardens in 1948 titled "Rhipsalis - and plant distributions in the Southern Hemisphere", in which he explicitly attributed the distribution of this genus to continental drift. Another early proponent of the idea that Rhipsalis is a "Gondwanic" genus was Croizat (1952:362, cited by Hunt 1967:433).

I'm now going to digress ever so slightly and say something about one of the most controversial biologists of the 20th century, the aforementioned Leon Croizat. I'm doing this because he has been ignored by the biological establishment for far too long and is overdue for reassessment. The following information on his life is taken from Hull (1988).

Croizat was born in 1894 in Turin, Italy of French parentage; his parents separated when he was 6, and he spent the first half of his life in poverty. He emigrated to the United States in 1923, made a living selling his watercolours until the Crash of 1929, moved to Paris where he found the life as a penniless artist less than satisfying, and then moved back to New York. He was eventually employed as technical assistant to the Director of the Arnold Arboretum at Harvard. He is of particular interest to cactophiles because many of his early publications dealt with succulents and were published in the American Cactus and Succulent Journal. One publication, however, is a 141 page booklet "De Euphorbio antiquorum atque officinarum" (A study of succulent Euphorbiae long known in cultivation), dated 1934, which seems to have been privately published, a harbinger of what was to come. [For the record he proposed the cactus genus Navajoa in 1943.] He had the temerity to publish a paper critical of a leading Kew botanist and was in time sacked, which I suspect left him with an outsized chip on his shoulder. He then emigrated to Venezuela, had several jobs in botany in academia between 1947 and 1952, divorced his first wife and married a Hungarian refugee who ended up owning the most successful landscaping firm in Caracas. This was important because his major works, including "Panbiogeography" (3 large volumes), "Principia botanica" (2 volumes) and "Space, time form: the biological synthesis" were privately published. These books, totalling several thousand pages, are not for the faint-hearted; in fact, it is fair to say that Croizat's vitriolic attacks on his opponents have put many readers off. Attitudes to Croizat have not been helped by some of his disciples - many of whom happen to be New Zealanders - who have adopted his tactics. Croizat, who died in 1982, coined the slogan "earth and life evolve together".

Croizat's great insight came when he plotted distributions of closely related organisms (eg members of a genus) on a map of the world. He soon found that similar patterns ("tracks") emerged for quite disparate groups, whether they were trees, lizards or insects. The conclusion he came to, and the one which is the least controversial of his claims, is that these patterns are not the result of chance dispersal but reflect a far more basic underlying cause. Widely distributed organisms were likely to be of ancient origin. He didn't rule out chance dispersal, but relegated it to a minor role to account for the rare exceptions. Most plants and animals in fact have limited powers of dispersal. In 1974 Croizat published a joint paper with two biologists with the American Museum of Natural History, Gareth Nelson and Don Rosen, called "Centers of origin and related concepts". It was a critique of dispersalist scenarios and introduced the world to what is usually called "vicariance biogeography." This claims that after a species arises it spreads quite rapidly into the available space, bounded only by ecological requirements. Any disjunct distribution arises from "vicariance", splitting the original range by climatic, geographical or geological processes. At the smallest scale this may be the result of a river changing its course, by mountain building, or by drowning of coastal hills to form isolated islands (eg in the Marlborough Sounds). On the largest scale rifting of the range is caused by continental drift. The paper appeared at a good time - plate tectonics was accepted by nearly all earth scientists and biologists by the early 70s, but nonetheless it proved very controversial. One implication of the theory is that - other things being equal - the most widely distributed member of a group of organisms with limited powers of dispersal will be the oldest member of that group.

Croizat later repudiated his paper with Nelson and Rosen, claiming that his particular kind of biogeography, which he called "panbiogeography", was not the same as vicariance biogeography. In fact, vicariancism tends to focus on the continental masses, whereas panbiogeography regards the ocean basins as all-important. Croizat had pointed out that the margins of continents were at least in some cases composite, that they seemed to have been added to over long periods and retained their distinctive biotas. One example was the Pacific coast of South America. During the 1980s geologists came up with the concept of "terranes". Work in Alaska revealed a large chunk of the country (called Wrangelia) that seemed to have a very different geological history from that of adjacent rocks. It was suggested that Wrangelia had formed a long way (thousands of kilometres) from its present position and had been "accreted" much later. It was soon found that similar "exotic" or "suspect" terranes are widespread, particularly around the Pacific. The terrane concept has since become part of geological orthodoxy, and the adjectives have long since been dropped. New Zealand itself is now thought to be made up of several terranes added onto a small continental core.

Now that I have given the background, let's take a look at Rhipsalis. The Rhipsalis occurring outside the Americas are closely related to the very widely distributed R. baccifera (J.S. Mueller) Stearn [long known by its synonym R. cassutha (or cassytha) Gaertner]. In fact, Barthlott & Taylor (1995: 63-65) regard R. baccifera as a polytypic species which they subdivide into 6 subspecies, including:

  • R. baccifera subsp. baccifera (Caribbean, eastern Mexico, Florida, Central America and northern South America southwards to north-east Brazil)
  • R. baccifera subsp. mauritiana (De Candolle) Barthlott (tropical Africa, Madagascar, Mascarenes, Comores, Seychelles and Sri Lanka)
  • R. baccifera subsp. erythrocarpa (K. Schumann) Barthlott (East Africa), and
  • R. baccifera subsp. horrida (Baker) Barthlott which is known only from Madagascar.

[The other subspecies are restricted to South America.] They do not give reasons for this classification, and I prefer to regard these "subspecies" as distinct species if only to avoid clumsy trinomials. In fact, R. mauritiana is tetraploid (ie has twice the normal number of chromosomes, which is 22 in this group) whereas R. baccifera includes both diploid and tetraploid forms; according to Barthlott & Taylor (1995:64) the former also differs in "micromorphological epidermal characters and in having generally larger fruits". R. horrida may be tetraploid or octoploid (with four times the complement of chromosomes) and differs in having ribbed adult stems with bristle-like spines.

There are three suggests that have been made to account for the distribution of Rhipsalis. These are (a) introduction by humans, (b) long-range dispersal by natural processes, and (c) vicariance.

(a) Introduction by Humans
This is assumed to have taken place in post-Columbian times (ie within the last 500 years), as there is no evidence that Amazonian Indians ever crossed the Atlantic (or that Africans made return trips to Brazil). Benson (1982: 115) took this possibility seriously enough to discuss it in some detail, and commented "Rhipsalis is both a beautiful plant and a curiosity, as a succulent, leafless epiphyte living on and dangling gracefully from tree branches, logs, or cliffs. It was among the first plants to capture the attention of explorers of the tropics." [Was it? - Benson does not back up this claim.] He went on to suggest that R. baccifera is: "the most widespread and abundant species in Latin America and the one universally cultivated, its escape from cultivation in two widely separated parts of the Eastern Hemisphere tropics, and its rapid dispersal there seems likely. R. baccifera could have invaded disturbed areas or even native forests with ease. The small, fleshy mucilaginous fruits of Rhipsalis are eaten or carried on the bills, shanks, feet, and feathers of nonmigratory as well as migratory birds, and the escape of this genus from cultivation to the surrounding forests is likely. There have been nearly five centuries during which this could have occurred."

Of course there are plenty of examples of introduced plants and animals invading native habitats, and New Zealanders certainly don't have to look far for examples. However, Benson's argument makes a lot of assumptions that are not documented. One is that R. baccifera is "universally cultivated". Is there any evidence that Rhipsalis of American origin was ever cultivated in west Africa prior to the 20th century? What about Madagascar? And what about the Comores, Mascarenes, Seychelles and Sri Lanka, most of which are a long way from standard shipping routes? R. baccifera is in fact a rather inconspicuous plant with tiny flowers and white fruits, and surely would not have been the species of choice for cultivators. [It certainly doesn't seem to be at present either - most popular books on cacti barely mention Rhipsalis - if at all - and if they do usually illustrate other species.] More likely candidates for cultivation would be species with more striking flowers such as R. grandiflora or R. megalantha.

And is a period of five centuries really long enough for an epiphyte to spread across tropical Africa (a distance of about 3500 km)? In fact, the time available is a lot less as the occurrence of Rhipsalis in the eastern hemisphere has been known for more than 170 years. The real problem, however, is that the Old World Rhipsalis are not identical with typical R. baccifera, and that there are in fact good grounds for regarding them as distinct species (see above).

There is a variant of the human-vector scenario that I will dispose of quickly and mention only for its risibility. I don't know who first suggested it but it gets mentioned every now and then if only for laughs (eg Rowley 1978: 5). It is suggested that homesick sailors used Rhipsalis as a substitute for mistletoe on long sea voyages; after all as Rowley points out, there is a superficial similarity between R. baccifera and common mistletoe. [Rowley does not specify the nationality of these mythical seamen, but Cullmann, Götz & Gröner (1986: 49) claim they were English!]. The mind boggles at the thought of marines attaching Rhipsalis in cabin doorways, but you know what sailors are! The main problems are the wide distribution of Rhipsalis in places like central Africa, far from any port, and on tiny islands such as the Comores, and of course the fact that several different taxa (probably distinct species) are involved.

(b) Dispersal by Birds or by Rafting
The most popular explanation, if the books available to me are any guide, is that birds dispersed the seeds of Rhipsalis across the Atlantic then to Madagascar, Comores, Seychelles, Mascarenes and Sri Lanka. This certainly has a superficial plausibility - Australian birds often get blown across the Tasman Sea to New Zealand (a distance of about 1700 km) but I don't know of any evidence that they bring seeds with them. One who explicitly attributed the disjunct distribution of Rhipsalis to long-range transport by birds was R. Roland-Gosselin whose original paper appeared in 1912 and was reprinted in 1947. Anthony (1948, cited by Benson 1982: 115) supported this theory in a paper published in the same issue of the Journal of the New York Botanical Gardens as Camp's. Similar ideas have been espoused - invariably with little or no critical discussion - by many subsequent authors, although Benson (1982: 115) was commendably cautious when he commented "transportation by birds is a somewhat plausible explanation, though there is no evidence that it is the correct one". The leading worker on the genus, Wilhelm Barthlott (1979: 23) stated "migrating birds must presumably have brought seeds to West Africa, some thousands (or millions) or years ago, and these plants spread from there". In their generally excellent book "The Cactus Primer" Gibson & Noble (1986: 250) mentioned the fact that some biogeographers had pointed to the distribution of Rhipsalis as evidence for continental drift, then stated:

"This example was thoroughly discredited by biologists, who noted that the species involved, R. baccifera, is widespread in the New World and the Old and because birds feed on the small, sticky, whitish fruits and then deposit the seeds with their feces a long distance from where the fruits were eaten."

There are two points to make about this claim - the first of course is that the New World Rhipsalis differ from the typical R. baccifera (see above). The other point is that Gibson and Nobel do not provide any documentation for the efficacy of bird dispersal of Rhipsalis seeds. Exactly how far is "a long distance" - 10 km, 50 km or 1000 km? I for one would like to know if any experiments have in fact been carried out to answer this question. Rowley (1978: 4-5) was one of the very few botanists to have devoted more than a few words to the Rhipsalis problem - he mentioned the three possibilities and commented that "Roland-Gosselin rather too blandly accepted birds as the agency responsible" for its presence in the Old World, but he did not come down firmly in favour of any theory.

The first problem with the dispersalist scenario is to identify a suitable vector. There are of course plenty of migratory birds, but most of these fly more or less along the meridian (ie north-south and back). After all they migrate to seek food and a suitable breeding site, leaving the northern hemisphere in the winter to take advantage of the southern summer, and vice versa. This is at right angles to the kind of migration we are seeking. Some oceanic birds such as albatrosses of course fly in a more or less latitudinal direction, but they are not noted for eating berries of any description, let alone those on epiphytes growing in rain forests. For the sake of argument, however, let us imagine a small fructivirous bird eating Rhipsalis berries somewhere in Brazil, conveniently near the coast, and then being caught up in a violent westerly storm. This bird is flown across the Atlantic (a minimum distance of nearly 3000 km) without voiding the ingested seeds, alights in a tree in a rain forest in west Africa and promptly regurgitates or defecates them into a convenient crevice. After a time they germinate, other birds eat the fruits and disperse the seeds so Rhipsalis eventually becomes established in the west African rain forests after which it spreads across the continent to the east coast. The whole process is then repeated, and Rhipsalis colonizes Madagascar, the Comores, Mascarenes, Seychelles and Sri Lanka. Even if we overlook the fact that the Old World Rhipsalis differ from those in the Americas, it doesn't require much thought to realise just how unlikely this scenario becomes; if the only occurrence of Rhipsalis outside of the Americas was in west Africa then it might be plausible, but to account for its presence elsewhere requires repetition of an already highly improbable event! [As an alternative, it could be argued that the seeds were stuck to the birds' feet or feathers, even though birds take great pains to preen themselves to maximise their aerodynamic efficiency. The problem is otherwise exactly the same - the process has to be repeated several times.] Of course the Atlantic was much narrower in the past - Africa and South America started to part about 130 Ma (million years ago), with the equatorial region parting company a few tens of millions of years later, but if bird-assisted dispersal took place when they were much closer, exactly when was this - when they were 1000 km apart, 500 km, or 50 km? Why not while they were still in contact?

There is another aspect to this scenario that hasn't been addressed to my knowledge. A bird flying or blown across the Atlantic would have the advantage of a large "target", ie it couldn't avoid making landfall somewhere on the west coast of Africa (assuming it could last the distance). Madagascar is of course a lot smaller than Africa but is as little as 425 km from the African coast, and bird-assisted dispersal across the intervening Mozambique Channel is not too implausible. [It is worth noting that until humans arrived there (probably less than 2000 yeas ago) Madagascar was home to a pigmy hippo. Its presence there suggests either that its ancestor swam from Africa when Madagascar was a lot closer or that there was some sort of short-lived land connection to Africa.] It is the other occurrences of Rhipsalis that pose a real problem for the bird dispersal theory. Consider first of all the Comores, tiny islands about 300 km east of Mozambique. They certainly don't present much of a target for any bird heading eastwards across the Indian Ocean. The same can be said for the Mascarenes (Mauritius and Reunion) which are 850 km east of Madagascar, and the Seychelles which are 1800 km east of Kenya and 1200 km north-east of the northern tip of Madagascar. Sri Lanka is a lot larger than these tiny islands, but it is a staggering 3600 km north-east of Madagascar.

Yet another question must be asked: Why, of all the cacti that have juicy fruits (and are therefore potentially attractive to birds) should it be Rhipsalis that has this wide distribution? Gibson and Nobel, quoted above about the fondness some birds have for Rhipsalis fruits, go on to say a mere two pages lager (1986: 252): "even though cacti generally have juicy fruits, long-distance dispersal of most tribes of Cactoideae has apparently not been effective in causing these cacti to range widely between the continental areas or even within a continent."

Quite! Croizat's comment is particularly relevant to this discussion: "it remains to be seen why 'casual dispersal' could not fill with Rhipsalidinae the Andes throughout, where ecology to suit is certainly not wanting" (Croizat 1961: 759).

If there is one cactus genus that might plausibly be subject to trans-oceanic dispersal by birds it is surely Melocactus, some species of which conveniently live in coastal areas, and of course have fruits which are not only juicy but brightly coloured as well and are therefore presumably even more likely to be eaten.

The alternative method of dispersing Rhipsalis across the Atlantic ocean is by rafting on vegetation washed into the sea by storms. This has had much less attention that bird dispersal, but Benson (1982: 115) thought that "Rhipsalis would be more likely than most plants to cling to a tree trunk or to its branches, above water". Well, maybe, but I think it highly improbable.

In any case, as Benson admits, there are not suitable ocean currents that would take Rhipsalis into the Indian Ocean to Madagascar, let alone Sri Lanka and assorted islands.

It should be clear from the preceding discussion that I consider the bird-dispersal scenario to be as dead in the water as any rain forest bird that tries to fly the Atlantic. This really leaves only the vicariance explanation as a viable option, and I will discuss this scenario and its implications for cactus evolution in the second part of this article.

Part 2

Previous Page