Far from dismantled: Reply to “dispersal syndromes of Vachellia caven” by Velasco et al., 2023

Introduction

Sir,

Velasco et al. (2023) conclude that the only reasonable explanation for the presence of Vachellia [Acacia] caven in Chile (west of the Andes) is human transport during the Holocene, making it an invasive archeophyte. They present a poorly argued review which would have benefited from a better empirical grounding. A systematic review, rather than a traditional review, might have helped to get a broader overview of the literature. At the same time, the way in which the evidence was evaluated is also problematic. We believe that the question of Vachellia [Acacia] caven's origins remains entirely unresolved.

The authors point out that there is no fossil evidence that V. caven existed west of the Andes (in what is today Chile) before the Holocene. However, as they review, there is also no fossil evidence for V. caven east of the Andes during that period. The authors note correctly that Acacia/Vachellia pollen is rarely fossilized in lake beds. Furthermore there is only one paleoecological ancient lake site within the range of V. caven in Chile (Lago Tagua Tagua) that goes back to the late Pleistocene. The lack of fossil pollen evidence for V. caven on both sides of its range is meaningless and cannot be interpreted as evidence for its initial distribution. Lack of evidence for something is not evidence for anything.

The authors cite two different estimates for the speciation of V. caven, one situating it around 10 Mya, the other around 6 Mya. According to them, the second estimate is too late in the orogeny of the Andes to allow the possibility of natural dispersal of the species across the mountains. We find a series of problems with this interpretation.

First, the speciation estimates cited are less precise than indicated. The 6 Mya estimate [1] is actually reported as 6.3±2.5, min 2, max 13 Mya. This range does not a priori rule out a speciation event before the Andes achieved its maximum height [2].

Secondly, the Andes is not a barrier to seed dispersal except when mediated by modern humans, as the authors seem to imply. There are two issues here. One is about the nature of long-distance plant dispersal. The other is about the plausibility of a non-human dispersal mechanism across the Andes. We address each in order.

The authors assume that it is necessary to establish that long-distance dispersal mechanisms are likely, “probable”, or “parsimonious”. To the contrary, long-distance dispersal is inherently improbable. Seeds are known to make extremely long journeys across inhospitable deserts and oceans, sometimes without any obvious dispersal adaptations [3,4]. Over very long time periods and long distances, dispersal is essentially trait-neutral [5]. It is thus irrelevant to make arguments about the probability of dispersal events when discussing long-distance dispersal across the Andes. At the same time, we know that numerous trees and other plants have dispersed crossed the Andes [[5], [6], [7]].

We also believe that there are plausible, if inherently low-probability or low-frequency non-human dispersal mechanisms by which V. caven could have crossed the Andes. Assessing plausibility requires adequately assessing minimum requirements for possibility. The author's implication that an animal disperser would have to carry a seed continuously for 165 km before defecating, corresponding to the distance between the current distributions of V. caven, seems to be an irrelevant and inaccurate criterion. The use of the 165 km criterion assumes that the current distribution of V. caven has been stable for at least several thousand years and has not been affected by land-use change or loss of its main disperser species; it also assumes that the current distribution is identical to the species' tolerances. These are all unreasonable assumptions. The authors state that if the seed did not travel the total distance between the current distribution in a single digestive event (eating to defecation), we would rather see a patchy distribution of V. caven within the Andes, which we do not, and that consequently neither shorter nor longer dispersal events ever happened.

We argue that a seed could potentially travel half-way over the Andes by endozoochory and then be washed the rest of the way in a flood event or along a river. Many rivers and quebradas connect the Andes to the central Chilean plains where V. caven is now found. Rivers are in fact closely associated with known precolumbian Andean passes, and these areas are also currently used and inhabited by camelids [8]. The following of large herbivore migration patterns is thought to be a determinant of early human migration routes, suggesting that herbivores were likely using the same Andean passes long before humans arrived (e.g. Ref. [9]). Historical records indicate that Andean passes leading to central Chile can be crossed by humans on foot with pack animals in 7 days or less [8]. Although we have not found data on the time it takes for guanaco to cross the Andes during migration, we assume it is comparable. Compared to a gut passage time of 1–3 days for an herbivore the size of a guanaco (following [10]), we suggest that it is within the realm of the possible for a guanaco or larger herbivore to eat a V. caven seed and defecate it near a high-Andean river from which it would be washed into central Chile. Thus we suggest that a combination of large herbivores using Andean passes during seasonal migrations and hydrochory provides a plausible explanation for long-distance dispersal across the Andes (if indeed we are looking at a dispersal and not a vicariance scenario).

Furthermore, the authors’ incredulity regarding the various non-human dispersal mechanisms indicates a lack of knowledge about V. caven and the animal species they consider. For example, the discussion of the role of water in bringing V. caven into disturbed areas cites a personal observation stating that V. caven seems to occur where disturbance “reduces competition with other forest species”. This is a misleading statement and there is no reason to cite a personal observation when this issue has been studied by numerous authors (e.g. Ref. [[11], [12], [13], [14]]). V. caven is a pioneer species establishing in open patches post disturbance. It dominates an early-succession formation and facilitates forest trees which later overshade it. The primary mechanism by which V. caven colonises disturbed areas is currently through livestock [15,16]. It is not difficult to find V. caven seedlings growing literally out of cattle dung. If, as the authors state, cattle disperse but do not “enhance” the germination of seeds, this is entirely beside the point (but see the discussion of scarification below). The cattle endochory mechanism has been adequate to lead to post-disturbance woodland succession across central Chile (e.g. Ref. [13,14]).

Cattle may of course act as proxies for extinct megafauna, as the authors acknowledge, but discarding the possibility of megafauna as dispersers because megafaunal dispersal events could only have occurred in the late Pleistocene or early Holocene—when the megafauna were already in decline and would have thus supposedly resulted in a patchy V. caven distribution—is premature, for all the same reasons stated above. Furthermore, it is important to consider that the Pleistocene and late Pleistocene assemblage of large mammals was quite diverse for the area that is the current distribution of V. caven. This large vertebrate assemblage did not only include proboscideans of the genus Notiomastodon (not Stegomastodon as it is erroneusly metioned by the authors) and several species of giant ground sloths, but also species of extinct horses (genera Hippidion and Equus), extinct camelids (i.e. Palaeolama), the litoptern Macrauchenia and extinct cervids such as Antifer sp. While we can speculate a bit about the feeding behavior for some of these taxa by observing closely related species still living today, this is not the case, for example, for Machrauchenia and giant ground sloths, therefore it is hard to assess and dismiss their possible role as seed dispersers. Giant sloths, horses and proboscideans, among other extinct megafauna had high-Andean distributions during the Pleistocene [17,18] late Pleistocene [19,20] and further back in the Pliocene (e.g. Ref. [21]) and could thus be imagined as potential trans-Andean dispersers.

Finally, the understanding of guanaco ecology is shallow. Guanacos are flexible generalists and their foraging habits in one habitat do not predict their foraging habits in another habitat [22,23]. The fact that they mainly eat herbs (in some habitat) has no bearing on whether they ever eat tree seeds, especially at times of the year when there are few herbs. We agree that further research on guanaco- V. caven interactions is needed, given that their habitat overlap has been lost across most of its range for several hundred years. The authors suggest that our report (Guerrero-Gatica & Root-Bernstein 2018) of a V. caven seedling growing out of a guanaco dung midden is unreliable as the seedling may have been growing from under the dung midden, and should be discounted as evidence of possible guanaco dispersal of V. caven. We think this reinterpretation of our observation is implausible, and we include photographic evidence here (Fig. 1). We do not have direct proof that the seed was ingested by the guanacos and arrived at their dung midden by defecation, although this is the simplest explanation. However, we can at least conclude that the dung midden provided the necessary temperature and water requirements for germination [24,25]. Germination also requires some mechanism of scarification. Simulated digestion in the form of an acid treatment is insufficient to stimulate V. caven seeds to germinate, while manual scarification, that is, physical abrasion, significantly increases germination rates [24,25]. This implies that whether potential dispersers chew the seeds or swallow them whole is more important than digestion. Potentially, the seed could have been abraded by invertebrates inhabiting the guanaco dung midden, but we do not believe that the possibility of guanaco ingestion can be ruled out without experiments.

Fig. 1.

Left, the Vachellia [Acacia] caven seedling observed growing out of the top of a guanaco dung midden, observed October 16, 2017 (photo (c) MR-B). The “soil” seen under the whole pellets is a layer of decomposed pellets. This was the only V. caven seedling that we observed on the site, in any of the dung middens (which we monitored regularly) or in the randomly placed herbaceous samples which we collected regularly. We also found no larger seedlings (<40 cm) on the site. There were no other tree species with this leaf shape in the enclosure or the vicinity. The guanacos were naive to this habitat when they entered the enclosure in 2014, having previously only ever eaten alfalfa, and we did not observe them to immediately eat V. caven seeds. Our interpretation is that they started to eat them only in 2017, when the experiment ended. The dung middens were subsequently collected for use as garden fertiliser and we were not able to continue monitoring them. Right, for reference, the size of dung middens on the site in 2016.

In addition, the authors claim that the role of guanacos can be discounted because of “the low abundance of these mammals in major parts of their historical distribution where they were extirpated almost entirely from central Chile by indigenous man” for which they cite González et al. (2006). González et al. (2006) makes no such claim, because it is false. Discussing the entire range of guanacos in South America, González et al. instead state “During the rest of the Holocene, small climatic changes and human pressure probably affected dispersion of the guanaco, with alternating periods of expansion and colonization events with contraction and extinction of local populations” (González et al., 2006, p. 158). Recently, the Metro of Santiago (in central Chile) announced the discovery of human remains, hearths, and guanaco remains with a probable age around 11,000 ybp at a site along the future line 7 of the metro [26]. There were still wild guanaco populations seasonally migrating from the Andes into central Chilean lowlands in the early 1700s [27]. It is well-established that there were populations of wild guanacos and herded varieties (chilihueques) in central Chile throughout the precolumbian period [28].

Finally, it is confusing that the authors appear to support the idea that other tree species, e.g. Prosopsis spp., may be have been dispersed across the Andes, but state that V. caven could not have been because it is less appetising. It is well-established that herbivores eat both palatable and unpalatable plants because mixing them improves their digestion [29]. Surely if Prosopsis spp. could cross the Andes by endozoochory, so could V. caven.

Fundamentally, the authors misunderstand that long-distance dispersal is not a highly likely event but rather a trait-neutral fundamentally improbable event. All of their arguments about low probabilities are thus irrelevant. Moreover, the suggestion that it is implausible to imagine any mechanism of nonhuman dispersal simply displays a lack of knowledge of the species and mechanisms involved, along with a lack of imagination. It remains possible that A. caven experienced vicariance or spread across the Andes to the west during or after its orogeny, and could have been dispersed through a combination of several mechanisms including human dispersal: there is no reason to believe that only one dispersal event ever took place or that dispersal only ever occurred in one direction.

All scenarios for the origin of Vachellia [Acacia] caven remain plausible and there is no evidence justifying recategorising the species as an invasive archeophyte. Such a recategorization risks exposing the species to lowered conservation protections. Only recently has V. caven been recognized as covered by the Native Forest law in Chile. Arguments allowing land conversion where V. caven occurs on the grounds that it is technically invasive would be welcomed by land developers, but would put in danger the natural successional process to native forest that it facilitates.

CRediT authorship contribution statement

Meredith Root-Bernstein: Writing – review & editing, Writing – original draft, Conceptualization. Matías Guerrero-Gatica: Writing – review & editing, Writing – original draft. Natalia A. Villavicencio: Writing – review & editing, Writing – original draft.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.