Mespilus

Phipps, James B., 2016, Studies In Mespilus, Crataegus, And × Crataemespilus (Rosaceae), I. Differentiation Of Mespilus And Crataegus, Expansion Of × Crataemespilus, With Supplementary Observations On Differences Between The Crataegus And, Phytotaxa 257 (3), pp. 201-229 : 202-210

Mespilus distinct from Crataegus ?

Molecular studies

Firstly, I consider evidence from the cladistic literature. Some of this covers Maleae as a whole, while other papers deal only with the Crataegus clade plus necessary outgroups. All molecular papers seen resolve Mespilus as closely related to a monophyletic Crataegus , e.g., Verbylaité et al. (2006: Fig. 1 ), Campbell et al. (2007: Fig. 6 ), Potter et al. (2007), Lo et al. (2009), Lo & Donoghue (2012), and Li et al. (2012). All the papers except Lo et al. (2007) accept Mespilu s as a monotypic genus sister to Crataegus .

Thus it is the paper by Lo et al. (2007) that is of particular interest here. It specifically posed the question ‘ Mespilus and Crataegus , one genus or two?’, and returned the answer ‘one’. Then, to enable fusion of the genera, and rather than make a huge number of new combinations under Mespilus , Crataegus was proposed for conservation ( Talent et al. 2008). Brummitt (2011) recorded the passing of this proposal by the Nomenclature Committee for Vascular Plants (report no. 62), which was approved by the International Botanical Congress in Melbourne ( McNeill et al. 2012). Results similar to Lo et al. (2007) are reported by Lo et al. (2009) but in the latter paper attention is directed to the phylogeography of Crataegus and the hybrid origin of the C. phenopyrum-spathulata-marshallii group of species.

Importantly, the conclusions reached by Lo et al. (2007) were dependent on their including for the first time Crataegus brachyacantha . This taxonomically rather isolated species is a large, black-fruited hawthorn of the southern United States with narrow unlobed leaves and was shown to be on the most basal branch of Crataegus s. str. By contrast, Mespilus germanica is a large, brown-fruited plant also with narrow leaves, native to the Pontic (Black Sea) area. The many other Crataegus species constitute a mix with narrow, unlobed leaves and those with pinnately (usually shallowly) lobed leaves; red is the most common fruit color while a minority are yellowish or blackish. Lo et al. (2007) also included in their analysis the then putative nothospecies M. canescens , a rare Arkansas endemic, in which they had a special interest.

Lo et al. (2007) used both nuclear (ITS and LEAFY) and chloroplast data. These three data sets were analyzed independently and together, using both maximum parsimony and maximum likelihood. My focus here is on the relationship of Crataegus brachyacantha , Mespilus germanica and M canescens in various analyses as set out in their Figs. 1–4 ; Fig. 5 presents the results without M. canescens . In every analysis (i.e., both separate and combined data) that includes M. canescens the critical three taxa C. brachyacantha , M. germanica and M. canescens consistently formed a triad (their Figs. 1–4 ). The order of clustering in this triad varies, the first fusion being always involving M. canescens and either C. brachyacantha or M. germanica , different orders of fusion occurring in the combined nuclear maximum parsimony and the chloroplast maximum parsimony analyses (their Figs. 2 and 3 ). There are also notable discrepancies between analyses in the order of clustering of some other taxon groups. In this respect it is informative to compare their Figs. 4 (containing M. canescens ) and 5 (absent M. canescens ). Figure 4 compares the maximum likelihood tree for the combined nuclear results ( Fig. 4a ) with that for the chloroplast results ( Fig. 4b ). In both these trees the Mespilus-C. brachyacantha triad was sister to C. sect. Crataegus (mainly red-fruited, +/- deeply pinnatelylobed leaves). Also in both, this larger group was sister to the remainder of Crataegus . However, when nuclear and chloroplast data were combined and M. canescens omitted from the analysis (their Fig. 5 ), the sister grouping to the M. germanica and C. brachyacantha pair was the entire remainder of Crataegus , not just C. sect. Crataegus . As well, in Lo et al. ’s Fig. 5a (max. parsimony) M. germanica was shown to be sister to C. brachyacantha , while the same paper’s Fig 5b (max. likelihood) showed M. germanica sister to all Crataegus . The authors concluded that “after removing conflicts due to hybridization or other factors, the analyses of the combined nuclear and chloroplast data (Lo et al., Fig. 5 ) suggest that C. brachyacantha is sister to the remaining Crataegus species, rather than to M. germanica ”. Nevertheless, the authors proceeded with the steps required to combine Mespilus and Crataegus under the latter genus. Why was this?

Lo et al. ’s (2007) primary reason for uniting the two genera is clearly their perception that there are so few morphological characteristics that convincingly differentiate Mespilus from Crataegus , effectively, only the covered pyrenes of Mespilus lacked by Crataegus , which they clarify by reference to Medikus and Lindley. This distinction is undoubtedly correct, as far as it goes. Lo et al. further note that attempts to distinguish the two genera on other characteristics have often failed because “character states supposed to be diagnostic of Mespilus occur in species of Crataegus ”. This is also certainly true, as articulated in my companion paper ( Phipps 2016) and illustrates the confusion of many botanists on the issue. Finally, Lo et al. (2007) state in support of their position: “…because the number of morphological differences between the Mespilus-Crataegus clade and Amelanchie r is considerably greater than those distinguishing Mespilus and Crataegus from each other, it seems more reasonable to sink the smaller genus in the larger...”. Such a statement, besides being not logically compelling, is partly based on incorrect premises. This feeds into the discussion, below, of inflorescence differences among the relevant genera and clades. As a result, Lo et al. have unnecessarily presented new and unfamiliar names to the botanical public.

Of the other cited authors who maintained Mespilus , two also used Crataegus brachyacantha . The first, another paper by the senior author ( Lo et al. 2012), but on the entire tribe Maleae (there subtribe Pyrinae ), also using both chloroplast and nuclear DNA evidence, though excluding the nothospecies Mespilus canescens , made Mespilus sister to Crataegus . Interestingly, in this paper the authors showed no inclination to resuscitate the rarely used combination C. germanica (L.) K. Koch, but rather stuck with the familiar M. germanica .A very similar result was produced by Li et al. (2012), also on the entire tribe Maleae (again, as subtribe Pyrinae ). Here, the authors did use both C. brachyacantha and M. canescens , though the analysis was based on only ITS nuclear ribosomal DNA data. In this paper, Li et al. found Mespilus to be weakly supported, a result which actually strengthens the case for retaining Mespilus , in view of the fact that the presence of the nothospecies M. canescens may be expected to dilute the support for Mespilus . Consequently, had the authors not used M. canescens , M. germanica should have been more strongly supported.

Tree topologies are affected by choice of taxa, and the inclusion for the first time of Crataegus brachyacantha by Lo et al. (2007) was very significant in this respect. Not only was a species from the black-fruited group, in which these authors already had a special interest, included for the first time, but this species was a taxon with a narrow, lobeless leaf type, now established as being plesiomorphic in the Crataegus clade. Thus, one also eagerly awaits investigations which include the taxonomically isolated red/yellow-fruited Chinese species C. scabrifolia (Franchet) Rehder which has a narrow, lobeless leaf type, and which may well alter the basal topology again.

Thus, because Mespilus is always concluded to be sister to Crataegus in the cladistics literature, and because my view, discussed below, is that there is ample difference between the genera, I regard the case for merging Mespilus and Crataegus as at best optional or even unnecessary.

Evidence from intergeneric hybrids in Maleae , especially × Crataemespilus

I feel that too much can be made of the frequency of intergeneric hybrids in the Maleae as an argument for uniting genera. It is worth noting that although hybrids between many pairs of Maloidean genera are known ( Robertson et al. 1991), except for those involving Sorbus most are very rare in nature else only known in botanic gardens as human creations. Also, except for those alloploid species derived from crosses of Sorbus with Aria (Persoon) Host , Chamaemespilus Medicus , and Torminaria M. Roemer, nearly all are highly sterile. Notably, in the case of Mespilus-Crataegus, western Eurasian wild examples of the nothogenus × Crataemespilus are nearly unknown, even where both genera are common elements of the native flora. The naturally occurring Arkansas endemic Mespilus canescens is also nearly sterile (see discussion under the nothospecies) and only known from one population. One may additionally note the complete absence of natural intergeneric crosses in Maleae between certain pairs of common, sympatric, speciose genera, especially Malus with Pyrus L., or either with Crataegus , as well as between such sister-genera as Cormus Spach and Sorbus , which underscores the contention made at the beginning of the paragraph.

Mespilus and Crataegus distinguished morphologically

Mespilus is distinguished from all species of Crataegus by a number of characters. The presence of several consistent differences is the more notable in view of the considerable variation existing within the large genus Crataegus . Nevertheless, the basic morphology of the two genera is broadly similar, the fruit of both, for instance, being a kind of polypyrenous pome (erroneously, and impossibly, called a polypyrenous drupe by some).

Mespilus s. str., i.e., monotypic, is a large, sometimes thorny, shrub with simple, finely serrate to entire leaves, usually uniflorous inflorescences, large hairy, eglandular bracteoles rather like those of certain other Maloideae outside the Mespilus-Crataegus clade (e.g., Amelanchier ), large, perigynous, hawthorn-like flowers, and large brown fruit in which tissues of the disc completely cover the pyrenes, and through which style remnants penetrate ( Figs. 1 , 2 , 3 ). The flesh of the fruit of M. germanica is also particularly rich in stone cells ( Rohrer et al. 1991).

Crataegus is a large genus of shrubby or subarboreal habit, usually thorny, with entire to deeply pinnatifid leaves, distinctly marginally toothed in most, usually multiflorous inflorescences that are both monopodial and domed, with usually much smaller flowers and fewer stamens than Mespilus , smallish fruit (5–15 mm, except cvs.), most often reddish in color, though many are blackish and some yellow to orange; hypanthium in fruit nearly closed; pyrene tips exposed.

Formal descriptions of Mespilus and Crataegus and key to genera are given under ‘Taxonomy’, below.

Fruit differences

It is in the fruit of Mespilus that the most distinctive characters are found. In fact, the brownish fruit of M. germanica has been shown to be specifically adapted to mammalian carnivore dispersal ( Herrera 1989). This fits well the notion of genus as an adaptive suite and has parallels in other brown-fruited Maloidean genera such as Pyrus , Torminalis , and Cormus , as Herrera (op. cit.) points out. This relationship is also evident in an UPGMA phenogram of an 18 fruitcharacter Manhattan distance of 173 species of Maloideae (Rohrer at al. 1991: Fig. 35). The distinctive features of Mespilus fruit are mirrored in the characteristic flavor of medlar fruit when made into a delicious conserve by specialist firms ( Phipps et al. 2003), or as eaten fresh but sightly decayed (‘bletted’), which is quite different from similar mayhaw ( Crataegus ser. Aestivales ) or tejocote ( C. mexicana ) products. Note also the very large, foliaceous, unlobed, +/- entire fruiting sepals that are erect to connivent and distant in Mespilus ( Fig. 8 ).

Flowering differences

In flower, Mespilus differs from Crataegus both in petal form and bracteole type. The petals of M. germanica are usually notched ( Fig. 2 ) as they are also in its hybrid M. canescens ( Fig. 4 ). With regard to the bracteole distinction from Crataegus , Mespilus has large, densely abaxially hairy and eglandular bracteoles. No Crataegus shares this ensemble, one which can, however, be found in Amelanchier . Most Crataegus have smaller, abaxially glabrous and gland-margined bracteoles while those with larger bracteoles are almost always conspicuously gland-margined, an exception being C. viridis L. (C. sect. Coccineae , C. ser. Virides), which is also glabrous. Those that are significantly abaxially hairy are restricted to C. sect. Crataegus (many), C. sect. Sanguineae ser. Nigrae (most), C. sect. Coccineae ser. Molles s.l. (general), and the apparently closely related series C. ser. Triflorae, Bracteatae , and Parvifoliae , but all the above have bracteoles that are strongly gland-bordered. There remains only C. brachyacantha , which has small, eglandular bracteoles in which the pubescence is very sparse. These differences were not considered by Lo et al. (2007).

Resting bud differences

The resting buds of medlar and hawthorns are basically similar though conical, dull and gray-brown in Mespilus as opposed to +/- globular, shiny and reddish in Crataegus . Resting bud differences were not considered by Lo et al. (2007).

Other evidence

Folk-taxonomy always distinguishes Mespilus from Crataegus , as reviewed in detail in a companion paper ( Phipps 2016). Thus, the distinctive folk names of Mespilus , as well as its cultural significance as reviewed by Baird & Thieret (1989), represent additional reasons for retaining Mespilus . Further, the academic taxonomy traced since Linnaeus ( Phipps 2016) demonstrates a consistent preference for two genera, though often with an inadequate understanding of their differences.

Molecular analyses routinely indicate a Mespilus-Crataegus clade, and all researchers but Lo et al. (2007, 2009), who favor its union with Crataegus , maintain Mespilus . Lo et al. (2007) was significant for including C. brachyacantha (which comes from a basal branch) and M. canescens for the first time. Natural hybrids between Crataegus and Mespilus are very rare and more or less sterile—in the Maloidean context not a strong argument for generic fusion. The fundamental differences between M. germanica and Crataegus are largely in the fruit but there are also lesser differences in the flowers, bracteoles and dormant buds, though not in inflorescence position. I regard the suite of fruit characters as central, especially in view of Herrera’s demonstration of adaptation to carnivore frugivory. In my view the differences found between Mespilus and Crataegus are more than ample for the continued recognition of Mespilus as a distinct genus.

However, as will be demonstrated in a companion paper ( Phipps 2016) it is the inadequate or even quite wrong understandings of these differences that has sometimes lead to the idea that the differences are slight, and as such, might justify combining these genera. Thus, Lo et al. ’s (2007) arguments that significant character overlap can be found between these two genera carry little weight; indeed character overlap is to be expected between related genera. Similarly, their view that there is little actual difference between Mespilus and Crataegus must be rebuffed as the distinctions between Mespilus and Crataegus are evident and sufficient. Finally, the close contact of rural people with plants that impact their lives, and that has lead them always to distinguish medlar from hawthorn, makes an interesting validation of the thesis presented here, and is discussed in the companion paper ( Phipps 2016).

Figures

FIGURE 1. Mespilus germanica. Fruiting branches of wild type. Note fruit about 15 mm diam. with wide hypanthial openings. Cult., Kew. (photo K.R. Robertson).

FIGURE 2. Mespilus germanica. Wild type in flower, diam. ca. 30 mm. Note some petals notched, long sepals, 28–30 stamens, red anthers. Cult., Kew.

FIGURE 3. Mespilus germanica. Fruit of cv. Nottingham. Note proportionately broad leaves, very large fruit (ca. 35 mm diam.), widespaced, very long sepals. Arrowed are: (a) style projecting through disc tissue; (b) undulating surface of disc tissue covering pyrenes; (c) split in side of fruit caused by inadequate allometric growth. Cult., Ailsa Craig, Ontario.

FIGURE 4. Crataemespilus ×canescens flowers. Note large size (ca. 25 mm diam.), ca. 20 stamens, yellow anthers, notched petals. Arkansas.

FIGURE 5. Crataegus brachyacantha infructescence showing blackish fruit with small hypanthial opening. Texas. (photo R.W. Lance).

FIGURE 6. Crataegus brachyacantha inflorescences showing numerous small flowers on type A inflorescence. Texas.

FIGURE 8. Amelanchier arborea. Herbarium specimen in flower. Note sylleptic growth (at a), and types A and B inflorescence origin, arrowed (respectively at c2 and b). Ontario.