This article introduces the diversity of Abies in Mexico and Mesoamerica, one of two major centres of diversity in the genus. An introduction to the diversity that exists in the second major centre, the Sino-Himalaya, may be found under ‘Abies in the Sino-Himalaya’.
Abies diversity in Mexico and Mesoamerica
The firs of Mexico and Mesoamerica include several apparent relicts with narrow distributions, and other more widespread species that may still be segregating. Besides their apparent diversity, the confusion surrounding them is also attributable to the fact they are not nearly so well-known, nor as widely cultivated, as their northern and Old World cousins. Only one species, A. vejarii, may reasonably be said to be widespread in our area; several others are occasionally encountered in specialist collections but otherwise they are virtually unseen. Whilst they cannot yet rival their Sino-Himalayan cousins for the number of names published, approaches to their circumscription and classification vary considerably; infra- and interspecific diversity, both morphological and molecular, continues to fuel taxonomic debate (Grimshaw & Bayton 2009).
Many conifers have relatively high levels of genetic diversity across their range, but individual populations are often only weakly differentiated at the molecular level – for example Knees (2011) was not able to demonstrate significant genetic differentiation between the various Abies species of the Mediterranean basin, despite their discrete distributions and morphological distinctions. In several Abies species in Mexico and Mesoamerica the reverse is true: despite low levels of overall genetic diversity, population differentiation is stronger than would be expected in species from more northern regions (e.g. Heredia-Bobadilla et al. 2012). This is partly explained by their long and recurring periods in isolation – a result of the cycle of glacial and interglacial periods that has driven latitudinal (and more locally altitudinal) migrations of greater complexity than those typical of more northern regions – but also by the extraordinarily rapid fragmentation of some populations due to human activity (Strandby et al. 2006; Aguirre-Planter, Furnier & Eguiarte 2000). This last point is particularly important to keep in mind, for extensive deforestation in the region has robbed us of context that would otherwise have been valuable in our attempts to better understand this diversity.
Another obstacle, often overlooked, is the importance of observing material at key stages of the reproductive cycle. Phenology has been demonstrated to be an important trait of several species in this region, despite the fact it is seldom referenced elsewhere in the genus: ‘The maintenance of very different morphologies and phenologies in trees of A. flinckii and A. religiosa intermixed at the same site in the Sierra de Manantlán strongly suggests that these are reproductively isolated separate species’ (Aguirre-Planter, Furnier & Eguiarte 2000).
Until 1932 only A. religiosa (Kunth) Schltdl. & Cham. was known from the region, having been described in 1830. Over 100 years would pass before A. hickelii Flous & Gaussen was described, in 1932, closely followed by A. guatemalensis Rehder in 1939, A. durangensis Martínez and A. vejarii Martínez in 1942, and A. coahuilensis I.M. Johnst. in 1943. Martínez made an important revision of the Mexican Firs in 1948 and at this time described A. oaxacana Martínez and A. guatemalensis var. jaliscana Martínez. A. religiosa var. emarginata Loock was also published at about this time. In 1963 Martínez reduced A. coahuilensis to a variety of A. durangensis and described A. mexicana Martínez (Farjon 2001).
In his 1971 monograph Liu interpreted the Mexican-Mesoamerican species with characteristic economy, treating only 5 of the 6 species that had been described up until that point (reducing Martínez’s A. mexicana to a variety of A. vejarii) and recognising only one other infraspecific taxon, A. durangensis var. coahuilensis (Liu 1971). Following fieldwork during the 1980s Rushforth described two new species in 1989: A. colimensis Rushforth & Narave and A. flinckii Rushforth – in the latter name he re-circumscribed the entities previously represented by the names A. guatemalensis var. jaliscana and A. religiosa var. emarginata (Rushforth 1989).
It is only in the relatively recent past that all these firs have been subject to increasing scrutiny. Whilst fieldwork to study them with sufficient rigour remains beset by a unique set of challenges, the publication of new names over the last c. 30 years generally reflects differing taxonomic concepts rather than truly new discoveries (Farjon 2017). In 1990, Farjon & Silba reduced A. oaxacana to a variety of A. hickelii: A. hickelii var. oaxacana (Martínez) Farjon & Silba (Farjon 2001). Debreczy & Rácz described A. guatemalensis var. longibracteata Debreczy & Rácz in 1995, and in the same year they co-authored three new species names: A. hidalgensis Debreczy, Rácz & Guízar, A. neodurangensis Debreczy, Rácz & Salazar, and A. zapotekensis Debreczy, Rácz & Ramírez (Debreczy & Rácz 2011). In 1997 Silba described A. guatemalensis var. tamaulipasensis, and three years later made several other additions to the A. guatemalensis complex including the publication of three new varietal names – see A. guatemalensis for further discussion.
The situation then remained relatively static up until about the time New Trees was being prepared for publication. Various studies were undertaken in the late 1990s and 2000s, but the first that might have been expected to have significant implications for this group came too late to be considered in New Trees. In 2008, Jaramillo-Corea et al. (2008) demonstrated that in southern Mexico and Guatemala populations of A. guatemalensis, A. hickelii and A. religiosa form ‘a relatively homogenous group’. The following year Strandby, Christensen & Sørensen (2009) tackled this complex and concluded that although differences exist, these taxa do not merit continued recognition as distinct species. They propose that they are merged, the oldest name (A. religiosa) taking priority. A. hickelii is treated as A. religiosa subsp. hickelii (Flous & Gaussen) U. Strandby, K.I. Chr. & M. Sørensen, A. guatemalensis as A. religiosa subsp. mexicana (Martínez) U. Strandby, K.I. Chr. & M. Sørensen, while A. religiosa s.s. is the nominate subsp. religiosa. Whilst noting insufficient data to determine the status of A. flinckii, they further conclude that A. vejarii ‘cannot retain its status as a separate taxon as the material studied is nested within [their concep of] A. religiosa subsp. mexicana’ (Strandby, Christensen & Sørensen 2009).
A later study (Vázquez-García et al. 2014) noted that the approach proposed by Strandby et al. (2009) would involve combining into a single variable species taxa that have traditionally been treated in different Sections of the genus, based on the widely accepted generic classification of Farjon & Rushforth. According to that classification, two sections are present in Mexico and Mesoamerica: Section Grandis and Section Oiamel. The publication of a new infrageneric classification (Xiang et al. 2018) removes this technical impediment, but several problems remain with Strandby et al.’s findings, and thanks are due to Keith Rushforth for sharing his insights as this account was being prepared. Perhaps the most significant concern is the statement ‘we were able to demonstrate that 15 of the 30 measured characters were correlated with altitude (9), latitude (8) or longitude (11)’ (Strandby, Christensen & Sørensen 2009). The altitudinal correlations are particularly troublesome because any student of Abies – and indeed of so many other genera – should be aware that altitude and exposure have significant impact on morphological expression: the results attached to these characters were a foregone conclusion, an issue generally compounded by an apparently narrow sampling method and failure to take the cone rachis into consideration. Another problem concerns the validity of their combination A. religiosa subsp. mexicana. The epithet mexicana was presumably used because, following the lumping of A. vejarii, it was assumed to be oldest name validly published at subspecies rank for their new, expanded taxon. However, for any given species to have an infraspecific taxon, in this example A. vejarii subsp. mexicana, it is at the point of publication universally assumed that the type infraspecific already exists, and this is de facto the older name. Because Farjon published A. vejarii subsp. mexicana in 1990 there exists, technically, subsp. vejarii (Farjon 2001) and as the older valid name for the taxon in question it is this that should have been used in a new combination as A. religiosa subsp. vejarii (K. Rushforth pers. comm. 2020).
For these and other reasons – not least the risk of adding further confusion to a poorly understood group of limited representation in cultivation – Strandby et al.’s conclusions will not be adopted here. Instead, we will maintain the taxonomy applied to this suite of species in Farjon (2017) with amendments as outlined below, at least until such a time as their underlying tenets have been more convincingly proven.
Vázquez-García et al. (2014) concluded that A. flinckii Rushforth merits recognition as a distinct species, but that the taxon previously treated as A. guatemalensis var. jaliscana, which Rushforth circumscribed within A. flinckii, is distinct but not attributable to either A. guatemalensis or A. flinckii; they therefore elevate it to species rank as A. jaliscana (Martínez) Mantilla, Shalisko & A. Vázquez. We will afford A. flinckii full treatment, but A. jaliscana is maintained at infraspecific rank within A. guatemalensis.
Debreczy and Rácz discuss A. colimensis Rushforth in appendices where they caution that it ‘appears to be just a form of A. religiosa’ (Debreczy & Rácz 2011), a view shared by Farjon (2017) who synonymises it with that species, but Rushforth suggests that if anything its cone suggests an affinity with Section Nobilis (A. procera and A. magnifica) (K. Rushforth pers. comm. 2020) which distances it from A. guatemalensis. Certainly, Debreczy and Rácz’s image of the cone supports this inferred relationship and distinctiveness (Debreczy & Rácz 2011, p. 1006). It is in cultivation, albeit in small numbers, and is treated here in full.
Farjon affords A. hidalgensis ‘the benefit of the doubt’ (Farjon 2017) but it is probably not in cultivation and thus will not be treated here. Also absent, so far as can be ascertained, are A. neodurangensis (see A. durangensis) and A. zapotekensis (see A. guatemalensis). The status of these last two species remains unresolved; some consider them good species, others regard them as synonyms of the names under which we discuss them.
Within Mexico and Mesoamerica, then, the following taxa are treated here: A. durangensis; A. durangensis var. coahuilensis; A. colimensis; A. flinckii; A. guatemalensis subsp. guatemalensis; A. guatemalensis subsp. jaliscana; A. guatemalensis subsp. tamaulipasensis; A. hickelii; A. religiosa; A vejarii; A. vejarii subsp. mexicana.
Historic introductions to cultivation
As noted above, A. religiosa became the first Mexican-Mesoamerican fir to be described to science when it was published in 1830, and a gap of over 100 years would follow before the next species was described. In the intervening century, all the various introductions from the region were made under this name, despite numerous expressions of doubt in early literature. This raises several questions about the early introductions of Mexican-Mesoamerican firs, most particularly: is A. religiosa the true identity of the early introductions to Europe, as widely recorded? By scrutinising historic literature and overlaying a modern understanding of fir diversity in the region, it now seems certain that several species were involved.
The first introduction is unanimously said to be Hartweg’s, gathered in 1838 when he was collecting for the Horticultural Society of London. Hartweg’s record of his 1836–1843 expedition mentions several encounters with ‘A. religiosa’, but only one can unequivocally be said to be this species, those trees Hartweg saw ‘at the highest point of the mountains of Angangueo’ (Hartweg 1848). The forests here are nowadays much studied for they are one of the hibernation areas for the Monarch Butterfly (Danaus plexippus) whose annual migration is one of the great spectacles of the natural world, and the identity of the firs here is beyond doubt. Most modern accounts agree, however, that A. religiosa does not occur any further south than northernmost Oaxaca (e.g. Farjon 2017; Debreczy & Rácz 2011), and the firs Hartweg later saw there most likely belong to A. hickelii; the fir he saw near Totonicapán in Guatemala – well outside the range of A. religiosa – would have been A. guatemalensis. Although Hartweg says that none of his collections from Angangueo ever reached England, there must have been two dispatches from here, for in early 1839 he ‘proceeded again to Angangueo, where by that time I found the pine cones, which I left ripening on my first visit, in a fit state for transmission’ (Hartweg 1848). It cannot be proven absolutely, but this is the most likely source of the original introduction of genuine A. religiosa, for all those encounters with firs he reports on elsewhere took place in 1839 or later, yet A. religiosa is included in his list of those 1838 collections that were successfully raised in England (Hartweg 1848, p. 125).
Elwes & Henry mention several specimens of A. religiosa, both living and dead, although none are specifically linked to Hartweg’s introduction, nor to two others implicit in their description of its range: ‘throughout the mountains of Mexico, from near Durango in the Sierra Madre range (lat. 24°), where it was collected by Seeman, to the mountains of northern Guatemala (lat. 15°), where it was observed by Hartweg and collected by Skinner’ (Elwes & Henry 1906–1913). As has been said the Guatemalan populations belong to A. guatemalensis. Similarly, most modern works agree that A. religiosa does not occur as far north as Durango state; Seeman’s collections most likely represented A. durangensis. Clearly there was a greater diversity of Mexican-Mesoamerican firs in cultivation, and from a far earlier date, than is generally considered the case.
Unfortunately, this diminishes the value of most early accounts from cultivation, for in most instances individual trees are not linked to collectors or provenances, hence it is impossible to determine their true identities. Exceptions exist, of course, such as the beautiful illustration of a cone and foliage in the Gardeners’ Chronicle (1876, p. 561) which clearly depicts A. religiosa. Another is confirmed by extension, for one of two trees at Fota was later reported to be identical with this illustration (Gardeners’ Chronicle 1885, p. 56). The source of the illustrated material was probably the first example to cone in cultivation; the specimen was cut from a tree growing in a garden in Cornwall, about 30 years old at the time, but as usual no provenance is given in Murray’s accompanying discussion. It is telling that even early accounts such as this hypothesise the existence of multiple species hiding among the cultivated population of A. religiosa.
We must accept, in the absence of compelling evidence, that records even of ‘famous’ trees cited in early literature must now be reappraised. It is not in doubt that Mexican firs grew in the mid to late 19th century in gardens such as those listed by Elwes & Henry, including Trevince, Lamorran, Tregothnan, Fota, Woodstock, and Castle Kennedy, as well as in milder districts of France, such as at Cherbourg, and around the Italian Lakes, but their true identities may never be known. The two that grew at Fota seem very likely to have represented two different species, for they differed markedly in several important characters, as well as hardiness (Gardeners’ Chronicle 1885, p. 56).
Being species of relatively southern distribution, often over large altitudinal ranges, provenance would seem to be key to success with these firs. None of the large trees reported in the early 20th century seemed to have survived much beyond the 1930s, when they would all have been less than 100 years old. Early literature is peppered with notes suggesting that besides being prone to frost, many were also inclined to develop multiple, competing leaders, rendering trees susceptible to damage and even complete failure in high winds (Bean 1976). Writing in 1972, Mitchell noted that ‘A. religiosa’ was ‘almost extinct’ in UK and Irish gardens before fresh introductions were made in the early 1960s (Mitchell 1972). Even now, in the third decade of the 21st century, all these firs remain far rarer in collections than they deserve to be, on account of the twin-hindrances of lack of supply and their short-lived nature. Those that have found their way into gardens show promise, especially in a warming climate, and were they more readily available they would no doubt be widely planted. The taxonomic structure applied to them remains far from perfect and will doubtless be subject to ongoing debate and change; perhaps when they have been grown as widely and studied as long as their Sino-Himalayan counterparts, we will be approaching a more useful and lasting consensus.